1. health and fitness
2. disease

SPECIAL COMMUNICATION
ONLINE FIRST
Acute Respiratory Distress Syndrome
The Berlin Definition
The ARDS Definition Task Force*
V
ALID AND RELIABLE DEFINI -
tions are essential to conduct epidemiological studies succ e s s f u lly a n d t o
facilitate enrollment of a consistent patient phenotype into clinical trials.1 Clinicians also need such definitions to
implement the results of clinical trials,
discuss prognosis with families, and
plan resource allocation.
Following the initial description of
acute respiratory distress syndrome
(ARDS) by Ashbaugh et al2 in 1967,
multiple definitions were proposed and
used until the 1994 publication of the
American-European Consensus Conference (AECC) definition.3 The AECC
defined ARDS as the acute onset of hypoxemia (arterial partial pressure of
oxygen to fraction of inspired oxygen
[PaO2/FIO2] ⱕ200 mm Hg) with bilateral infiltrates on frontal chest radiograph, with no evidence of left atrial hypertension. A new overarching entity—
acute lung injury (ALI)—was also
described, using similar criteria but with
less severe hypoxemia (PaO2/FIO2 ⱕ300
mm Hg).3
The AECC definition was widely
adopted by clinical researchers and
clinicians and has advanced the
knowledge of ARDS by allowing the
acquisition of clinical and epidemiological data, which in turn have led to
improvements in the ability to care
for patients with ARDS. However,
after 18 years of applied research, a
number of issues regarding various
criteria of the AECC definition have
emerged, including a lack of explicit
For editorial comment see p 2542.
2526
JAMA, June 20, 2012—Vol 307, No. 23
The acute respiratory distress syndrome (ARDS) was defined in 1994 by the
American-European Consensus Conference (AECC); since then, issues regarding the reliability and validity of this definition have emerged. Using a consensus process, a panel of experts convened in 2011 (an initiative of the European Society of Intensive Care Medicine endorsed by the American Thoracic
Society and the Society of Critical Care Medicine) developed the Berlin Definition, focusing on feasibility, reliability, validity, and objective evaluation of
its performance. A draft definition proposed 3 mutually exclusive categories
of ARDS based on degree of hypoxemia: mild (200 mm Hg⬍PaO2/FIO2 ⱕ300
mm Hg), moderate (100 mm Hg⬍PaO2/FIO2 ⱕ200 mm Hg), and severe (PaO2/
FIO2 ⱕ100 mm Hg) and 4 ancillary variables for severe ARDS: radiographic severity, respiratory system compliance (ⱕ40 mL/cm H2O), positive endexpiratory pressure (ⱖ10 cm H2O), and corrected expired volume per minute
(ⱖ10 L/min). The draft Berlin Definition was empirically evaluated using patientlevel meta-analysis of 4188 patients with ARDS from 4 multicenter clinical data
sets and 269 patients with ARDS from 3 single-center data sets containing physiologic information. The 4 ancillary variables did not contribute to the predictive validity of severe ARDS for mortality and were removed from the definition. Using the Berlin Definition, stages of mild, moderate, and severe ARDS
were associated with increased mortality (27%; 95% CI, 24%-30%; 32%; 95%
CI, 29%-34%; and 45%; 95% CI, 42%-48%, respectively; P⬍.001) and increased median duration of mechanical ventilation in survivors (5 days; interquartile [IQR], 2-11; 7 days; IQR, 4-14; and 9 days; IQR, 5-17, respectively;
P⬍.001). Compared with the AECC definition, the final Berlin Definition had
better predictive validity for mortality, with an area under the receiver operating curve of 0.577 (95% CI, 0.561-0.593) vs 0.536 (95% CI, 0.520-0.553;
P⬍.001). This updated and revised Berlin Definition for ARDS addresses a number of the limitations of the AECC definition. The approach of combining consensus discussions with empirical evaluation may serve as a model to create
more accurate, evidence-based, critical illness syndrome definitions and to better inform clinical care, research, and health services planning.
JAMA. 2012;307(23):2526-2533
Published online May 21, 2012. doi:10.1001/jama.2012.5669
criteria for defining acute, sensitivity
of PaO2/FIO2 to different ventilator settings, poor reliability of the chest
radiograph criterion, and difficulties
distinguishing hydrostatic edema
(TABLE 1).4
www.jama.com
*Authors/Writing Committee and the Members of the
ARDS Definition Task Force are listed at the end of
this article.
Corresponding Author: Gordon D. Rubenfeld, MD,
MSc, Program in Trauma, Emergency, and Critical Care,
Sunnybrook Health Sciences Center, 2075 Bayview
Ave, Toronto, ON M4N 3M5, Canada (gordon
[email protected]).
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME
For these reasons, and because all
disease definitions should be reviewed periodically, the European Society of Intensive Care Medicine convened an international expert panel to
revise the ARDS definition, with endorsement from the American Thoracic Society and the Society of Critical Care Medicine. The objectives were
to update the definition using new data
(epidemiological, physiological, and
clinical trials) to address the current
limitations of the AECC definition and
explore other defining variables.
Methods
Consensus Process. Three co-chairs
were appointed by the European Society of Intensive Care Medicine, who in
turn selected panelists based on their
work in the area of ARDS and to ensure
geographic representation from both Europe and North America. An overview
of the consensus process used by the
panel is outlined in the FIGURE. In revising the definition of ARDS, the panel
emphasized feasibility, reliability, face
validity (ie, how clinicians recognize
ARDS), and predictive validity (ie, ability to predict response to therapy, outcomes, or both). In addition, the panel
determined that any revision of the definition should be compatible with the
AECC definition to facilitate interpretation of previous studies. After initial
preparations and an in-person consensus discussion, a draft definition was
proposed,13 which underwent empirical evaluation. The definition was further refined through consensus discussion informed by these empirical data.
Empirical Evaluation of Draft
Definition.
Cohort Assembly. Through the review
of the literature presented at the consensus meeting, discussions with other
experts, and review of personal files, the
panel identified studies that met the following eligibility criteria: (1) large, multicenter prospective cohorts, including consecutive patients or randomized
trials, or smaller, single-center prospective studies with unique radiological or
physiological data that enrolled adult
patients with ALI as defined by AECC;
Table 1. The AECC Definition3—Limitations and Methods to Address These in the Berlin Definition
AECC Definition
Acute onset
AECC Limitations
No definition of acute4
ALI category
All patients with PaO2/
FIO2 ⬍300 mm Hg
Oxygenation
PaO2/FIO2 ⱕ300
mm Hg (regardless of PEEP)
Misinterpreted as
PaO2/FIO2 = 201-300,
leading to confusing
ALI/ARDS term
Inconsistency of PaO2/
FIO2 ratio due to the
effect of PEEP and/or
FIO25-7
Chest radiograph
Bilateral infiltrates observed on frontal
chest radiograph
PAWP
PAWP ⱕ18 mm Hg
when measured or
no clinical evidence of left atrial
hypertension
Risk factor
None
Timing
Poor interobserver
reliability of chest
radiograph
interpretation8,9
High PAWP and ARDS
may coexist10,11
Poor interobserver
reliability of PAWP and
clinical assesments of
left atrial
hypertension12
Not formally included in
definition4
Addressed in
Berlin Definition
Acute time frame
specified
3 Mutually exclusive
subgroups of
ARDS by severity
ALI term removed
Minimal PEEP level
added across
subgroups
FIO2 effect less
relevant in severe
ARDS group
Chest radiograph
criteria clarified
Example radiographs
created a
PAWP requirement
removed
Hydrostatic edema
not the primary
cause of
respiratory failure
Clinical vignettes
created a to help
exclude
hydrostatic edema
Included
When none
identified, need to
objectively rule out
hydrostatic edema
Abbreviations: AECC, American-European Consensus Conference; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, fraction of inspired oxygen; PaO2, arterial partial pressure of oxygen; PAWP, pulmonary artery
wedge pressure; PEEP, positive end-expiratory pressure.
aAvailable on request.
(2) studies collected granular data necessary to apply the individual criteria
of both the draft Berlin Definition and
the AECC definition; and (3) authors
of these original studies were willing to
share data and collaborate. The panel
identified 7 distinct data sets (4 multicenter clinical studies for the clinical
database14-17 and 3 single-center physiological studies for the physiological database18-20) that met these criteria. Further details of these studies are included
in the eMethods (http://www.jama
.com).
Variables. Studies provided data on
hospital or 90-day mortality. Ventilatorfree days at 28 days after the diagnosis
of ALI were calculated as a composite
measure of mortality and duration
of mechanical ventilation. Duration of
mechanical ventilation in survivors was
selected as an indirect marker of severity of lung injury because this outcome
is not biased by mortality or decisions
©2012 American Medical Association. All rights reserved.
related to the withdrawal of lifesustaining treatments.21 Progression of severity of ARDS within 7 days was assessed using the longitudinal data
collected within each cohort. We distinguished patients with more extensive involvement on the frontal chest radiograph (3 or 4 quadrants) from those with
the minimal criterion of “bilateral opacities” (2 quadrants).
Static compliance of the respiratory
system (CRS) was calculated as tidal volume (mL) divided by plateau pressure
(cm H2O) minus positive endexpiratory pressure (PEEP) (cm H2O).
The corrected expired volume per minute (V̇ECORR) was calculated as the measured minute ventilation multiplied by
the arterial partial pressure of carbon
dioxide (PaCO2) divided by 40 mm Hg.22
Total lung weight was estimated from
quantitative computed tomography
(CT) images.23 Shunt was calculated at
one site as previously reported.24
JAMA, June 20, 2012—Vol 307, No. 23 2527
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME
Figure. Outline of Consensus Process
Premeeting preparations
(May to September 2011)
Selection of panelists by chairs
Precirculation of key topics for discussion
Preparation of background material by
panelists
In-person discussions
(September 30 to October 2, 2011, Berlin,
Germany)
Presentations of key background material
Development of the conceptual model of
ARDS
Draft of Berlin Definition based on informal
consensus discussions
Empirical evaluation of draft definition
(October 2011 to January 2012)
Assembling clinical and physiologic cohorts
Demonstration of patient characteristics
and distribution according to definition
categories
Evaluation of impact of ancillary variables
for severe ARDS subgroup
Follow-up of consensus discussions and
analysis
(February 2012 by multiple teleconferences)
Presentation of empirical evaluation
Final definition created based on further
informal consensus discussions
Decision to present the results of a
post hoc higher-risk subset
Testing of predictive validity
ARDS indicates acute respiratory distress syndrome.
Analytic Framework and Statistical
Methods. The analytic framework for
evaluating the draft Berlin ARDS Definition was to (1) determine the distribution of patient characteristics across
the defined severity categories; (2)
evaluate the value of proposed ancillary variables (more severe radiographic criterion, higher PEEP levels,
static respiratory compliance, and
V̇ECORR) in defining the severe ARDS
subgroup in the draft definition; (3) determine the predictive validity for mortality of the final Berlin Definition; and
(4) compare the final Berlin Defini2528
JAMA, June 20, 2012—Vol 307, No. 23
tion to the AECC definition. In addition, in a post hoc analysis, we sought
thresholds for C RS and V̇ E CORR that
would identify a severe group of patients with ARDS who had more than
50% mortality and include more than
10% of the study population.
We did not evaluate other PaO2/FIO2
cutoffs or the requirement of a minimum PEEP level (5 cm H2O) as they
were selected by the panel using face
validity criteria and to ensure compatibility with prior definitions. Similarly, we did not explore other variables that might improve predictive
validity, such as age and severity of nonpulmonary organ failure, because they
were not specific to the definition of
ARDS.25
To compare the predictive validity of
the AECC definition and the Berlin
Definition, we used the area under the
receiver operating curve (AUROC or C
statistic) in logistic regression models
of mortality with a dummy variable for
the ARDS definition categories.26 Because this technique requires independent categories to create the dummy
variable and the AECC definition for
ARDS is a subset of ALI, we could not
compare the AECC definition as specified. Therefore, we modified the AECC
definition and divided ALI into the independent categories of ALI nonARDS (200 mm Hg⬍PaO2/FIO2 ⱕ 300
mm Hg) and ARDS alone (Pa O 2 /
FIO2 ⱕ200 mm Hg). Although the category of ALI non-ARDS is not explicitly described by the AECC, it has been
used by many investigators.27,28
P values for categorical variables were
calculated with the ␹2 test; P values for
continuous variables were estimated
with the t test, Mann-Whitney, analysis of variance, or Kruskal-Wallis, depending on the distribution and number of variables. The receiver operating
curve statistical analyses were performed by using MedCalc for Windows version 12.1.4.0 (MedCalc Software) and other statistical tests were
performed with SAS/STAT for Windows version 9.2 (SAS Institute Inc).
Statistical significance was assessed at
the 2-sided P ⬍.05 level.
Results
Draft Consensus Definition.
The ARDS Conceptual Model. The panel
agreed that ARDS is a type of acute diffuse, inflammatory lung injury, leading to increased pulmonary vascular
permeability, increased lung weight, and
loss of aerated lung tissue. The clinical hallmarks are hypoxemia and bilateral radiographic opacities, associated
with increased venous admixture, increased physiological dead space, and
decreased lung compliance. The morphological hallmark of the acute phase
is diffuse alveolar damage (ie, edema,
inflammation, hyaline membrane, or
hemorrhage).29
Draft Definition Criteria. Following 2
days of consensus discussions, the panel
proposed a draft definition with 3 mutually exclusive severity categories (mild,
moderate, and severe) of ARDS. A set of
ancillary variables was proposed to further characterize severe ARDS and these
were explicitly specified for further empirical evaluation.13
Timing. Most patients with ARDS are
identified within 72 hours of recognition of the underlying risk factor, with
nearly all patients with ARDS identified within 7 days.30 Accordingly, for a
patient to be defined as having ARDS,
the onset must be within 1 week of a
known clinical insult or new or worsening respiratory symptoms.
Chest Imaging. The panel retained bilateral opacities consistent with pulmonary edema on the chest radiograph as defining criteria for ARDS, but
also explicitly recognized that these
findings could be demonstrated on CT
scan instead of chest radiograph. More
extensive opacities (ie, 3 or 4 quadrants on chest radiograph) were proposed as part of the severe ARDS
category and identified for further
evaluation.
Origin of Edema. Given the declining use of pulmonary artery catheters
and because hydrostatic edema in the
form of cardiac failure or fluid overload may coexist with ARDS,10,11 the
pulmonary artery wedge pressure criterion was removed from the defini-
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME
Table 2. Exploration of Proposed Variables to Define Severe ARDS a
Mild
Severe ARDS Definition
Consensus panel draft
PaO2/FIO2 ⱕ100 mm Hg ⫹ chest
radiograph of 3 or 4 quadrants ⫹
PEEP ⱖ10 cm H2O ⫹ (CRS ⱕ40 mL/cm
H2O or V̇ECORR ⱖ10 L/min)
Consensus panel final
PaO2/FIO2 ⱕ100 mm Hg
Moderate
Severe
No. (%) of
Patients
% Mortality
(95% CI)
No. (%) of
Patients
% Mortality
(95% CI)
No. (%) of
Patients
220 (22)
27 (24-30)
2344 (64)
35 (33-36)
507 (14)
220 (22)
27 (24-30)
1820 (50)
32 (29-34)
1031 (28)
% Mortality
(95% CI)
45 (40-49) b
45 (42-48) b,c
Abbreviations: ARDS, acute respiratory distress syndrome; CRS, compliance of the respiratory system; FIO2, fraction of inspired oxygen; PaO2, arterial partial pressure of oxygen;
PEEP, positive end-expiratory pressure; V̇ECORR, corrected expired volume per minute.
a The moderate group includes patients with PaO /FIO ⱕ200 mm Hg and patients with PaO /FIO ⱕ100 mm Hg who do not meet the additional criteria for severe ARDS in the draft
2
2
2
2
definition. All patients are receiving at least 5 cm H2O PEEP and have bilateral infiltrates on chest radiograph.
b P⬍.001 comparing mortality across stages of ARDS (mild, moderate, severe) for draft and final definitions.
c P=.97 comparing mortality in consensus draft severe ARDS to consensus final severe ARDS definitions.
tion. Patients may qualify as having
ARDS as long as they have respiratory
failure not fully explained by cardiac
failure or fluid overload as judged by
the treating physician using all available data. If no ARDS risk factor (eTable
1) is apparent, some objective evaluation (eg, with echocardiography) is required to help eliminate the possibility of hydrostatic edema.
Oxygenation. The term acute lung injury as defined by the AECC was removed, due to the perception that clinicians were misusing this term to refer to
a subset of patients with less severe hypoxemia rather than its intended use as
an inclusive term for all patients with the
syndrome. Positive end-expiratory pressure can markedly affect PaO2/FIO25,6;
therefore, a minimum level of PEEP (5
cm H2O), which can be delivered noninvasively in mild ARDS, was included
in the draft definition of ARDS. A minimum PEEP level of 10 cm H2O was proposed and empirically evaluated for the
severe ARDS category.
Additional Physiologic Measurements.
Compliance of the respiratory system
largely reflects the degree of lung volume loss.2 Increased dead space is common in patients with ARDS and is associated with increased mortality. 2 4
However, because the measurement of
dead space is challenging, the panel chose
minute ventilation standardized at a
PaCO2 of 40 mm Hg (V̇ECORR = minute
ventilation ⫻ Pa CO 2 /40) as a surrogate.22 The draft definition of severe
ARDS included the requirement of either
a low respiratory system compliance
(⬍40 mL/cm H2O), a high V̇ECORR (⬎10
L/min), or both. These variables were
identified for further study during the
evaluation phase.
The panel considered a number of
additional measures to improve specificity and face validity for the increased pulmonary vascular permeability and loss of aerated lung tissue that
are the hallmarks of ARDS, including
CT scanning, and inflammatory or genetic markers (eTable 2). The most
common reasons for exclusion of these
measures were lack of routine availability, lack of safety of the measure in
critically ill patients, or a lack of demonstrated sensitivity, specificity, or both
for use as a defining characteristic for
ARDS.
Empirical Evaluation of the Draft
Definition.
Patients. A total of 4188 patients in the
clinical database had sufficient data to
classify as having ARDS by the AECC
definition. Of these patients, 518 (12%)
could not be classified by the draft Berlin Definition because PEEP was missing or was less than 5 cm H2O. Patients who could not be classified by the
draft Berlin Definition had a mortality
rate of 35% (95% CI, 31%-39%), a median (interquartile range [IQR]) of 19
(1-25) ventilator-free days, and a median (IQR) duration of mechanical ventilation in survivors of 4 (2-8) days.
These patients were excluded from
analyses of the draft Berlin Definition
and comparisons between the AECC
©2012 American Medical Association. All rights reserved.
definition and the draft Berlin
Definition.
Compared with patients from the
population-based cohorts, patients from
clinical trials and the academic centers cohorts were younger, had more severe hypoxemia, and had more opacities on chest radiographs. The cohort
of patients from the clinical trials had
the lowest mortality, likely reflecting the
inclusion and exclusion criteria of the
trials.31 The cohort of patients from academic centers had the highest mortality and the lowest percentage of trauma
patients, reflecting the referral population (eTable 3).
There were 269 patients in the physiological database with sufficient data to
classify ARDS by the AECC definition, although the numbers of patients in each cohort were small. Patients in the Turin cohort had worse
PaO2/FIO2 ratios and had higher mortality than the other studies (eTable 4).
Evaluation of Ancillary Variables. The
draft Berlin Definition for severe ARDS
that included a PaO2/FIO2 of 100 mm Hg
or less, chest radiograph with 3 or 4
quadrants with opacities, PEEP of at least
10 cm H2O, and either a CRS of 40 mL/cm
H2O or less or a V̇ECORR of at least 10
L/min identified a smaller set of patients with identical mortality to the simpler severe ARDS category of PaO2/FIO2
of 100 mm Hg or less (TABLE 2). To address the possibility that the CRS and
V̇ECORR thresholds might be different in
patients with higher body weight, we
evaluated weight-adjusted cutoffs for
JAMA, June 20, 2012—Vol 307, No. 23 2529
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME
these variables in one of the cohorts.
There was no significant difference in the
predictive validity of the weightadjusted criteria. The consensus panel reviewed these results and considered the
lack of evidence for predictive validity of
these ancillary variables and their potential contribution to face validity and
construct validity and decided to use the
simpler definition for severe ARDS that
relied on oxygenation alone.
The Berlin Definition. The final Berlin Definition of ARDS is shown in
TABLE 3. Twenty-two percent (95% CI,
21%-24%) of patients met criteria for
mild ARDS (which is comparable with
the ALI non-ARDS category of the
AECC definition; TABLE 4), 50% (95%
CI, 48%-51%) of patients met criteria
for moderate ARDS, and 28% (95% CI,
27%-30%) of patients met criteria for
severe ARDS. Mortality increased with
stages of ARDS from mild (27%; 95%
CI, 24%-30%) to moderate (32%; 95%
CI, 29%-34%) to severe (45%; 95% CI,
42%-48%). Median (IQR) ventilatorfree days declined with stages of ARDS
from mild (20 [1-25] days) to moderate (16 [0-23] days) to severe (1 [020] day). Median (IQR) duration of mechanical ventilation in survivors
increased with stages of ARDS from
mild (5 [2-11] days) to moderate (7 [414] days) to severe (9 [5-17] days).
Using the Berlin Definition, 29% (95%
CI, 26%-32%) of patients with mild
ARDS at baseline progressed to moderate ARDS and 4% (95% CI, 3%-6%) progressed to severe ARDS within 7 days;
and 13% (95% CI, 11%-14%) of pa-
Table 3. The Berlin Definition of Acute Respiratory Distress Syndrome
Timing
Chest imaging a
Origin of edema
Oxygenation b
Mild
Moderate
Severe
Acute Respiratory Distress Syndrome
Within 1 week of a known clinical insult or new or worsening respiratory
symptoms
Bilateral opacities—not fully explained by effusions, lobar/lung collapse, or
nodules
Respiratory failure not fully explained by cardiac failure or fluid overload
Need objective assessment (eg, echocardiography) to exclude hydrostatic
edema if no risk factor present
200 mm Hg ⬍ PaO2/FIO2 ⱕ 300 mm Hg with PEEP or CPAP ⱖ5 cm H2O c
100 mm Hg ⬍ PaO2/FIO2 ⱕ 200 mm Hg with PEEP ⱖ5 cm H2O
PaO2/FIO2 ⱕ 100 mm Hg with PEEP ⱖ5 cm H2O
Abbreviations: CPAP, continuous positive airway pressure; FIO2, fraction of inspired oxygen; PaO2, partial pressure of
arterial oxygen; PEEP, positive end-expiratory pressure.
a Chest radiograph or computed tomography scan.
b If altitude is higher than 1000 m, the correction factor should be calculated as follows: [PaO /FIO ⫻(barometric pressure/
2
2
760)].
c This may be delivered noninvasively in the mild acute respiratory distress syndrome group.
tients with moderate ARDS at baseline
progressed to severe ARDS within 7 days.
All differences between outcome variables across categories of modified AECC
(ALI non-ARDS and ARDS alone) and
across categories of Berlin Definition
(mild, moderate, and severe) were statistically significant (P⬍.001).
Compared with the AECC definition, the final Berlin Definition had better predictive validity for mortality with
an AUROC of 0.577 (95% CI, 0.5610.593) vs 0.536 (95% CI, 0.5200.553; P⬍.001), with the difference in
AUROC of 0.041 (95% CI, 0.0300.050). To ensure that missing PEEP
data in one of the cohorts did not bias
the results, the regression analysis was
repeated without this cohort and
yielded similar results.
The Berlin Definition performed similarly in the physiological database as in
the clinical database (TABLE 5, eFigure
1, and eFigure 2). Twenty-five percent
(95% CI, 20%-30%) of patients met
criteria for mild ARDS, 59% (95% CI,
54%-66%) of patients met criteria for
moderate ARDS, and 16% (95% CI, 11%21%) of patients met criteria for severe
ARDS. Mortality increased with stages of
ARDS from mild (20%; 95% CI, 11%31%) to moderate (41%; 95% CI, 33%49%) to severe (52%; 95% CI, 36%68%), with P = .001 for differences in
mortality across stages of ARDS. Median (IQR) ventilator-free days declined with stages of ARDS from mild
Table 4. Predictive Validity of ARDS Definitions in the Clinical Database
Modified AECC Definition a
No. (%) [95% CI] of patients
Progression in 7 d from mild,
No. (%) [95% CI]
Progression in 7 d from moderate,
No. (%) [95% CI]
Mortality, No. (%) [95% CI] b
Ventilator-free days, median (IQR) b
Duration of mechanical ventilation in
survivors, median (IQR), d b
ALI Non-ARDS
1001 (24) [23-25]
ARDS
3187 (76) [75-77]
336 (34) [31-37]
Berlin Definition ARDS a
Mild
819 (22) [21-24]
Moderate
1820 (50) [48-51]
234 (29) [26-32]
Severe
1031 (28) [27-30]
33 (4) [3-6]
230 (13) [11-14]
263 (26) [23-29]
20 (2-25)
5 (2-10)
1173 (37) [35-38]
12 (0-22)
7 (4-14)
220 (27) [24-30]
20 (1-25)
5 (2-11)
575 (32) [29-34]
16 (0-23)
7 (4-14)
461 (45) [42-48]
1 (0-20)
9 (5-17)
Abbreviations: AECC, American-European Consensus Conference; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, fraction of inspired oxygen; IQR, interquartile range; PaO2, arterial partial pressure of oxygen; PEEP, positive end-expiratory pressure.
a The definitions are the following for ALI non-ARDS (200 mm Hg⬍PaO /FIO ⱕ300 mm Hg, regardless of PEEP), ARDS (PaO /FIO ⱕ200 mm Hg, regardless of PEEP), mild Ber2
2
2
2
lin Definition (200 mm Hg⬍PaO2 /FIO2 ⱕ300 mm Hg with PEEP ⱖ5 cm H2O), moderate Berlin Definition (100 mm Hg⬍PaO2 /FIO2 ⱕ200 mm Hg with PEEP ⱖ5 cm H2O), and
severe Berlin Definition (PaO2 /FIO2 ⱕ100 mm Hg with PEEP ⱖ5 cm H2O).
b Comparisons of mortality, ventilator-free days, and duration of mechanical ventilation in survivors across categories of modified AECC (ALI non-ARDS and ARDS) and across
categories of Berlin Definition (mild, moderate, and severe) are all statistically significant (P⬍.001).
2530
JAMA, June 20, 2012—Vol 307, No. 23
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME
Table 5. Predictive Validity of ARDS Definitions in the Physiologic Database
Modified AECC Definition a
No. (%) [95% CI] of patients
Mortality, No. (%) [95% CI] b
Ventilator-free days
Median (IQR)
Missing, No.
Duration of mechanical ventilation in
survivors, median (IQR), d
Lung weight, mg c
Mean (SD)
Missing, No.
Shunt, mean (SD), % c,d
ALI Non-ARDS
66 (25) [19-30]
13 (20) [11-31]
8.5 (0-23.5)
10
6.0 (3.3-20.8)
1371 (360.4)
16
21 (21)
Berlin Definition ARDS a
ARDS
203 (75) [70-80]
84 (43) [36-50]
Mild
66 (25) [20-30]
13 (20) [11-31]
Moderate
161 (59) [54-66]
62 (41) [33-49]
0 (0-16.0)
26
13.0 (5.0-25.5)
8.5 (0-23.5)
10
6.0 (3.3-20.8)
0 (0-16.5)
25
12.0 (5.0-19.3)
0 (0-6.5)
1
19.0 (9.0-48.0)
1556 (469.7)
32
29 (11)
1828 (630.2)
16
40 (16)
1602 (508.1)
48
32 (13)
1371 (360.4)
16
21 (12)
Severe
42 (16) [11-21]
22 (52) [36-68]
Abbreviations: AECC, American-European Consensus Conference; ALI, acute lung injury; ARDS, acute respiratory distress syndrome; FIO2, fraction of inspired oxygen; IQR, interquartile range; PaO2, arterial partial pressure of oxygen; PEEP, positive end-expiratory pressure.
a The definitions are the following for ALI non-ARDS (200 mm Hg⬍PaO /FIO ⱕ300 mm Hg, regardless of PEEP), ARDS (PaO /FIO ⱕ200 mm Hg, regardless of PEEP), mild Ber2
2
2
2
lin Definition (200 mm Hg⬍PaO2 /FIO2 ⱕ300 mm Hg with PEEP ⱖ5 cm H2O), moderate Berlin Definition (100 mm Hg⬍PaO2 /FIO2 ⱕ200 mm Hg with PEEP ⱖ5 cm H2O), and
severe Berlin Definition (PaO2 /FIO2 ⱕ100 mm Hg with PEEP ⱖ5 cm H2O).
b Eight patients are missing in the moderate Berlin Definition ARDS group. P=.001 for difference in mortality across Berlin stages of ARDS.
c Comparisons of lung weight and shunt across categories of modified AECC (ALI non-ARDS and ARDS) and across categories of Berlin Definition (mild, moderate, and severe)
are statistically significant (P⬍.001).
d Only available at 1 site.
(8.5 [0-23.5] days) to moderate (0 [016.5] days) to severe (0 [0-6.5] days),
with P=.003 for differences in ventilatorfree days across stages of ARDS. Median (IQR) duration of mechanical ventilation in survivors increased with stages
of ARDS from mild (6.0 [3.3-20.8] days)
to moderate (12.0 [5.0-19.3] days) to severe (19.0 [9.0-48.0] days), with P=.045
for differences in duration of mechanical ventilation in survivors across stages
of ARDS.
Using the Berlin Definition, stages of
mild, moderate, and severe ARDS had increased mean lung weight by CT scan
(1371 mg; 95% CI, 1268-1473; 1556 mg;
95% CI, 1474-1638; and 1828 mg; 95%
CI, 1573-2082; respectively) and increased mean shunt (21%; 95% CI, 16%26%; 29%; 95% CI, 26%-32%; and 40%;
95% CI, 31%-48%; respectively). Comparisons of lung weight and shunt (from
the single site providing these data)
across categories of modified AECC (ALI
non-ARDS and ARDS alone) and across
categories of Berlin Definition (mild,
moderate, and severe) were statistically
significant (P⬍.001) (Table 5, eFigure
3, and eFigure 4).
In a post hoc analysis, combining a
PaO2/FIO2 of 100 mm Hg or less with
either a Crs of 20 mL/cm H2O or less or
a V̇ECORR of at least 13 L/min identified
a higher-risk subgroup among pa-
tients with severe ARDS that included
15% of the entire ARDS population and
had a mortality of 52% (95% CI, 48%56%). Patients with severe ARDS who
did not meet the higher-risk subset criteria included 13% of the entire ARDS
population and had a mortality rate of
37% (95% CI, 33%-41%). The difference between the mortality of patients
with higher-risk severe ARDS and patients with severe ARDS who did not
meet these criteria was statistically significant (P ⬍ .001).
Comment
Developing and disseminating formal
definitions for clinical syndromes in
critically ill patients are essential for research and clinical practice. Although
previous proposals have relied solely on
the consensus process, this is to our
knowledge the first attempt in critical
care to link an international consensus panel endorsed by professional societies with an empirical evaluation.
The draft Berlin Definition classified patients with ARDS into 3 independent categories but relied on ancillary variables (severity of chest
radiograph, PEEP ⱖ10 cm H2O, CRS
ⱕ40 mL/cm H 2 O, and V̇ E CORR ⱖ10
L/min) in addition to oxygenation to define the severe ARDS group. When the
ancillary variables selected by the panel
©2012 American Medical Association. All rights reserved.
were subjected to evaluation, these
parameters did not identify a group of
patients with higher mortality and were
excluded from the final Berlin Definition after further consensus discussion. Without this evaluation, a needlessly complex ARDS definition would
have been proposed. However, static respiratory system compliance and an understanding of minute ventilation are
important variables for clinicians to
consider in managing patients with
ARDS, even though those variables were
not included as part of the definition.32
The Berlin Definition addresses some
of the limitations of the AECC definition, including clarification of the exclusion of hydrostatic edema and adding minimum ventilator settings, and
provides slight improvement in predictive validity. Our study presents data
on the outcomes of patients with ARDS
defined according to the Berlin Definition in a large heterogeneous cohort of
patients including patients managed
with modern approaches to lung protective ventilation. Estimates of the
prevalence and clinical outcomes of
mild, moderate, and severe ARDS can
be assessed from this database for research and health services planning.
Acute respiratory distress syndrome is
a heterogeneous syndrome with comJAMA, June 20, 2012—Vol 307, No. 23 2531
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME
plex pathology and mechanisms. The
proposed definition does not resolve this
problem. Investigators may choose to design future trials using 1 or more of the
ARDS subgroups as a base study population, which may be further refined
using criteria specific to the putative
mechanism of action of the intervention (eg, IL-6 levels for an anti–IL-6 trial
or more stringent hypoxemia criteria for
a study on extracorporeal membrane
oxygenation). Furthermore, some variables that were excluded from the Berlin Definition because of current feasibility and lack of data on operational
characteristics may become more useful in the future. We anticipate that clinical research using our model of definition development will be used to revise
the definition in the future.
There are limitations to our approach. First, although the Berlin Definition had statistically significantly superior predictive validity for mortality
compared with the modified AECC definition, the magnitude of this difference
and the absolute values of the AUROC
are small and would be clinically unimportant if the Berlin Definition was designed as a clinical prediction tool. However, predictive validity for outcome is
only one criterion for evaluating a syndrome definition and the purpose of the
Berlin Definition is not a prognostication tool.33 Although the Berlin Definition was developed with a framework
including these criteria, we did not empirically evaluate face validity, content
validity, reliability, feasibility, or success at identifying patients for clinical trial
enrollment.
Second, it is possible that our results
are not generalizable because of the data
sets we studied. This seems unlikely because patients from a broad range of
populations, including clinical trials, academic centers, and community patients, were included in the analyses.
Third, some variables (eg, CRS and
PEEP) were missing in some patients in
the data sets we used, either due to the
mode of mechanical ventilation that precluded their measurement or the practicalities of population-based research.
However, bias due to cohort selection or
2532
JAMA, June 20, 2012—Vol 307, No. 23
missing data seem unlikely because our
results were robust to sensitivity analyses that excluded individual cohorts.
Fourth, it is possible that the ancillary
variables did not identify a higher-risk
subset because the number of quadrants
on the chest radiograph cannot be assessed reliably, PEEP was not used in a
predictable fashion, or CRS and V̇ECORR
were not accurately measured. However,
if this is true, it is likely also to be true in
future studies and in clinical practice because the study database was constructed
from clinical trial, academic, and community sites reflecting practice in the real
world of clinical research. In addition, we
evaluated PEEP and CRS as used by clinicians in practice and not as a test of prespecified ventilator settings that may be
betterthanthevariablesevaluatedherein,
but may not be practical, particularly in
observational cohort studies.5,6
Fifth, because our study was not an
exercise in developing a prognostic
model for ARDS, we only considered
the variables and cutoffs proposed by
the consensus panel. We could not
compare this definition directly to the
AECC definition because the categories of that definition overlap. It is possible that the outcomes as well as the
relative proportion of patients within
each category of ARDS will change if
the underlying epidemiology of the syndrome evolves due to changes in clinical practice or risk factors.34 This is
particularly true for the post hoc higherrisk subset reported, for which the cut
points were derived from the data sets.
Conclusion
In conclusion, we developed a consensus draft definition for ARDS with an international panel using a framework that
focused on feasibility, reliability, and validity. We tested that definition using empirical data on clinical outcome, radiographic findings, and physiological
measures from 2 large databases constructed from 7 contributing sources to
assess the predictive value of ancillary
variables, refine the draft definition, and
compare the predictive validity of the
definition to the existing AECC definition. This approach for developing the
Berlin Definition for ARDS may serve as
an example for linking consensus definition activities with empirical research
to better inform clinical care, research,
and health services planning.
Published Online: May 21, 2012. doi:10.1001
/jama.2012.5669
Authors/Writing Committee: V. Marco Ranieri, MD
(Department of Anesthesia and Intensive Care Medicine, University of Turin, Turin, Italy); Gordon D.
Rubenfeld, MD, MSc (Program in Trauma, Emergency, and Critical Care, Sunnybrook Health Sciences Center, and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto,
Ontario, Canada); B. Taylor Thompson, MD (Department of Medicine, Massachusetts General Hospital and
Harvard Medical School, Boston); Niall D. Ferguson,
MD, MSc (Department of Medicine, University Health
Network and Mount Sinai Hospital, and Interdepartmental Division of Critical Care Medicine, University
of Toronto, Toronto, Ontario, Canada); Ellen Caldwell,
MS (Division of Pulmonary and Critical Care Medicine, University of Washington, Seattle); Eddy Fan, MD
(Department of Medicine, University Health Network and Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada); Luigi Camporota,
MD (Department of Critical Care, Guy’s and St.
Thomas’ NHS Foundation Trust, King’s Health Partners, London, England); and Arthur S. Slutsky, MD
(Keenan Research Center of the Li Ka Shing Knowledge Institute of St. Michael’s Hospital; Interdepartmental Division of Critical Care Medicine, University
of Toronto, Toronto, Ontario, Canada).
Author Contributions: Dr Rubenfeld and Ms Caldwell
had full access to all of the data in the study and take
responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Ranieri, Rubenfeld,
Thompson, Ferguson, Caldwell, Camporota.
Acquisition of data: Ranieri, Rubenfeld, Thompson.
Analysis and interpretation of data: Rubenfeld,
Thompson, Ferguson, Caldwell, Fan, Slutsky.
Drafting of the manuscript: Rubenfeld, Ferguson,
Caldwell, Slutsky.
Critical revision of the manuscript for important intellectual content: Ranieri, Rubenfeld, Thompson,
Ferguson, Caldwell, Fan, Camporota, Slutsky.
Statistical analysis: Rubenfeld, Caldwell, Slutsky.
Obtained funding: Ranieri.
Administrative, technical, or material support:
Rubenfeld, Thompson, Fan, Camporota.
Study supervision: Ranieri, Rubenfeld, Thompson,
Slutsky.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure
of Potential Conflicts of Interest. Dr Ranieri reported
receiving consulting fees or honoraria from Maquet
and Hemodec and board membership from Faron. Dr
Rubenfeld reported receiving consulting fees or honoraria from Ikaria, Faron, and Cerus. Dr Thompson reported receiving support for travel from European Society of Intensive Care Medicine; being an advisory
board member of Hemodec and AstraZeneca; receiving consultancy fees from US Biotest, Sirius Genetics,
sanofi-aventis, Immunetrics, Abbott, and Eli Lilly; and
receiving grants from the National Heart, Lung, and
Blood Institute. Dr Slutsky reported receiving support for travel expenses from European Society of Intensive Care Medicine; board membership from Ikaria;
receiving consultancy fees from GlaxoSmithKline and
Tarix; having stock/stock options with Apeiron and
Tarix; and sitting on advisory boards for Maquet Medical and NovaLung and steering committees for
HemoDec and Eli Lilly. No other authors reported any
financial disclosures.
Members of the ARDS Definition Task Force: V. Marco
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
THE BERLIN DEFINITION OF ACUTE RESPIRATORY DISTRESS SYNDROME
Ranieri, MD (Department of Anesthesia and Intensive Care Medicine, University of Turin, Turin, Italy);
Gordon D. Rubenfeld, MD, MSc (Program in Trauma,
Emergency, and Critical Care, Sunnybrook Health Sciences Center and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada); B. Taylor Thompson, MD (Department
of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston); Niall D. Ferguson, MD,
MSc (Department of Medicine, University Health Network and Mount Sinai Hospital, and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada); Ellen Caldwell, MS
(Division of Pulmonary and Critical Care Medicine, University of Washington, Seattle); Eddy Fan, MD (Department of Medicine, University Health Network and
Mount Sinai Hospital, University of Toronto, Toronto,
Ontario, Canada); Luigi Camporota, MD (Department of Critical Care, Guy’s and St. Thomas’ NHS
Foundation Trust, King’s Health Partners, London, England); and Arthur S. Slutsky, MD (Keenan Research
Center of the Li Ka Shing Knowledge Institute of St.
Michael’s Hospital; Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto,
Ontario, Canada); Massimo Antonelli, MD (Dipartimento di Anestesia e Rianimazione, Universita Cattolica
del Sacro Cuore, Rome, Italy); Antonio Anzueto, MD
(Department of Pulmonary/Critical Care, University
of Texas Health Sciences Center, San Antonio); Richard
Beale, MBBS (Department of Critical Care, Guy’s and
St. Thomas’ NHS Foundation Trust, King’s Health Partners, London, England); Laurent Brochard, MD (Medical-Surgical Intensive Care Unit, Hopitaux Universitaires de Geneve, Geneva, Switzerland); Roy Brower,
MD (Division of Pulmonary and Critical Care Medicine, Johns Hopkins University, Baltimore, Maryland); Andrés Esteban, MD, PhD (Servicio de Cuidados Intensivos, Hospital Universitario de Getafe,
CIBERES, Madrid, Spain); Luciano Gattinoni, MD (Istituto di Anestesiologia e Rianimazione, Universita degli Studi di Milano, Milan, Italy); Andrew Rhodes, MD
(Department of Intensive Care Medicine, St. George’s
Healthcare NHS Trust, London, England); Jean-Louis
Vincent, MD (Department of Intensive Care, Erasme
University, Brussels, Belgium); Provided data for the
empiric evaluation of the definition but were not part
of the consensus development: Andrew Bersten, MD
(Department of Critical Care Medicine, Flinders University, Adelaide, South Australia); Dale Needham, MD,
PhD (Outcomes After Critical Illness and Surgery Group
[OACIS], Division of Pulmonary and Critical Care Medicine and Department of Physical Medicine and Rehabilitation, Johns Hopkins University, Baltimore, Maryland); and Antonio Pesenti, MD (Department of
Anesthesia and Critical Care, Ospedale San Gerardo,
Monza, Italy; and Department of Experimental Medicine, University of Milano Bicocca, Milan, Italy).
Funding/Support: This work was supported by the Eu-
ropean Society of Intensive Care Medicine and grant
R01HL067939 from the National Institutes of Health
(Dr Rubenfeld). Dr Ferguson is supported by a Canadian Institutes of Health Research New Investigator
Award (Ottawa, Canada).
Role of the Sponsors: The European Society of Intensive Care Medicine, the National Institutes of Health,
the Canadian Institutes of Health Research, and the endorsing professional societies had no role in the design
and conduct of the study, in the collection, management, analysis, and interpretation of the data, or in the
preparation, review, or approval of the manuscript.
Online-Only Material: The eMethods, eReferences,
eTables 1 through 4, and eFigures 1 through 4 are available at http://www.jama.com.
Additional Contributions: Salvatore Maggiore, MD,
PhD (Department of Anesthesiology and Intensive Care,
Agostino Gemelli University Hospital, Università Cattolica
del Sacro Cuore, Rome, Italy), and Anders Larsson, MD,
PhD (Department of Surgical Sciences, Anesthesiology and Critical Care Medicine, Uppsala University,
Uppsala, Sweden), attended the roundtable as representatives of the European Society of Intensive Care
Medicine. Drs Maggiore and Larsson received no
compensation for their roles. Karen Pickett, MB BCh
(Department of Intensive Care, Erasme Hospital, Université Libre de Bruxelles, Brussels, Belgium), provided
technical assistance. Dr Pickett received compensation for her role in the conference.
terpretation of pulmonary artery catheter pressure
tracings. Chest. 1991;100(6):1647-1654.
13. ESICM Congress Highlights. http://www.esicm
.org/07-congresses/0A-annual-congress/webTv
.asp. Accessed April 23, 2011.
14. Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J
Med. 2005;353(16):1685-1693.
15. Bersten AD, Edibam C, Hunt T, Moran J; Australian and New Zealand Intensive Care Society Clinical
Trials Group. Incidence and mortality of acute lung injury and the acute respiratory distress syndrome in three
Australian States. Am J Respir Crit Care Med. 2002;
165(4):443-448.
16. Needham DM, Dennison CR, Dowdy DW, et al.
Study protocol: The Improving Care of Acute Lung Injury
Patients (ICAP) study. Crit Care. 2006;10(1):R9.
17. Britos M, Smoot E, Liu KD, Thompson BT, Checkley
W, Brower RG; National Institutes of Health Acute Respiratory Distress Syndrome Network Investigators. The
value of positive end-expiratory pressure and Fio criteria in the definition of the acute respiratory distress
syndrome. Crit Care Med. 2011;39(9):2025-2030.
18. Bellani G, Guerra L, Musch G, et al. Lung regional metabolic activity and gas volume changes induced by tidal ventilation in patients with acute lung
injury. Am J Respir Crit Care Med. 2011;183(9):
1193-1199.
19. Terragni PP, Rosboch G, Tealdi A, et al. Tidal hyperinflation during low tidal volume ventilation in acute
respiratory distress syndrome. Am J Respir Crit Care
Med. 2007;175(2):160-166.
20. Terragni PP, Del Sorbo L, Mascia L, et al. Tidal
volume lower than 6 mL/kg enhances lung protection: role of extracorporeal carbon dioxide removal.
Anesthesiology. 2009;111(4):826-835.
21. Spragg RG, Bernard GR, Checkley W, et al. Beyond mortality: future clinical research in acute lung
injury. Am J Respir Crit Care Med. 2010;181(10):
1121-1127.
22. Wexler HR, Lok P. A simple formula for adjusting arterial carbon dioxide tension. Can Anaesth Soc J.
1981;28(4):370-372.
23. Gattinoni L, Caironi P, Pelosi P, Goodman LR. What
has computed tomography taught us about the acute
respiratory distress syndrome? Am J Respir Crit Care
Med. 2001;164(9):1701-1711.
24. Nuckton TJ, Alonso JA, Kallet RH, et al. Pulmonary dead-space fraction as a risk factor for death in
the acute respiratory distress syndrome. N Engl J Med.
2002;346(17):1281-1286.
25. Cooke CR, Kahn JM, Caldwell E, et al. Predictors
of hospital mortality in a population-based cohort of
patients with acute lung injury. Crit Care Med. 2008;
36(5):1412-1420.
26. DeLong ER, DeLong DM, Clarke-Pearson DL.
Comparing the areas under two or more correlated
receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837845.
27. Luhr OR, Antonsen K, Karlsson M, et al; The ARF
Study Group. Incidence and mortality after acute respiratory failure and acute respiratory distress syndrome in Sweden, Denmark, and Iceland. Am J Respir
Crit Care Med. 1999;159(6):1849-1861.
28. Brun-Buisson C, Minelli C, Bertolini G, et al; ALIVE
Study Group. Epidemiology and outcome of acute lung
injury in European intensive care units: results from
the ALIVE study. Intensive Care Med. 2004;30
(1):51-61.
29. Katzenstein AL, Bloor CM, Leibow AA. Diffuse alveolar damage—the role of oxygen, shock, and related factors: a review. Am J Pathol. 1976;85(1):
209-228.
30. Hudson LD, Milberg JA, Anardi D, Maunder RJ.
Clinical risks for development of the acute respiratory distress syndrome. Am J Respir Crit Care Med.
1995;151(2 pt 1):293-301.
31. Phua J, Badia JR, Adhikari NKJ, et al. Has mortality from acute respiratory distress syndrome decreased over time? a systematic review. Am J Respir
Crit Care Med. 2009;179(3):220-227.
32. Hager DN, Krishnan JA, Hayden DL, Brower RG;
ARDS Clinical Trials Network. Tidal volume reduction in patients with acute lung injury when plateau
pressures are not high. Am J Respir Crit Care Med.
2005;172(10):1241-1245.
33. Rubenfeld GD. Epidemiology of acute lung injury.
Crit Care Med. 2003;31(4)(suppl):S276-S284.
34. Li G, Malinchoc M, Cartin-Ceba R, et al. Eightyear trend of acute respiratory distress syndrome: a
population-based study in Olmsted County,
Minnesota. Am J Respir Crit Care Med. 2011;
183(1):59-66.
REFERENCES
1. Streiner D, Norman G. Health Measurement Scales.
4th ed. New York, NY: Oxford University Press; 2008.
2. Ashbaugh DG, Bigelow DB, Petty TL, Levine BE.
Acute respiratory distress in adults. Lancet. 1967;
2(7511):319-323.
3. Bernard GR, Artigas A, Brigham KL, et al. The
American-European Consensus Conference on ARDS:
definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;
149(3 pt 1):818-824.
4. Phua J, Stewart TE, Ferguson ND. Acute respiratory distress syndrome 40 years later: time to revisit
its definition. Crit Care Med. 2008;36(10):29122921.
5. Villar J, Pérez-Méndez L, López J, et al; HELP
Network. An early PEEP/FIO2 trial identifies different degrees of lung injury in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med.
2007;176(8):795-804.
6. Ferguson ND, Kacmarek RM, Chiche J-D, et al.
Screening of ARDS patients using standardized ventilator settings: influence on enrollment in a clinical trial.
Intensive Care Med. 2004;30(6):1111-1116.
7. Gowda MS, Klocke RA. Variability of indices of hypoxemia in adult respiratory distress syndrome. Crit
Care Med. 1997;25(1):41-45.
8. Rubenfeld GD, Caldwell E, Granton JT, Hudson
LD, Matthay MA. Interobserver variability in applying a radiographic definition for ARDS. Chest. 1999;
116(5):1347-1353.
9. Meade MO, Cook RJ, Guyatt GH, et al. Interobserver variation in interpreting chest radiographs for
the diagnosis of acute respiratory distress syndrome.
Am J Respir Crit Care Med. 2000;161(1):85-90.
10. Ferguson ND, Meade MO, Hallett DC, Stewart
TE. High values of the pulmonary artery wedge pressure in patients with acute lung injury and acute respiratory distress syndrome. Intensive Care Med. 2002;
28(8):1073-1077.
11. Wheeler AP, Bernard GR, Thompson BT, et al; National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials
Network. Pulmonary-artery versus central venous catheter to guide treatment of acute lung injury. N Engl J
Med. 2006;354(21):2213-2224.
12. Komadina KH, Schenk DA, LaVeau P, Duncan
CA, Chambers SL. Interobserver variability in the in-
©2012 American Medical Association. All rights reserved.
JAMA, June 20, 2012—Vol 307, No. 23 2533
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 07/02/2013
ORIGINAL ARTICLE
Surgeon-Reported Conflict With Intensivists
About Postoperative Goals of Care
Terrah J. Paul Olson, MD; Karen J. Brasel, MD, MPH; Andrew J. Redmann, BA, BS;
G. Caleb Alexander, MD, MS; Margaret L. Schwarze, MD, MPP
Objective: To examine surgeons’ experiences of conflict with intensivists and nurses about goals of care for
their postoperative patients.
Design: Cross-sectional incentivized US mail-based
survey.
Setting: Private and academic surgical practices.
Participants: A total of 2100 vascular, neurologic, and
cardiothoracic surgeons.
Main Outcome Measures: Surgeon-reported rates of
conflict with intensivists and nurses about goals of care
for patients with poor postsurgical outcomes.
Results: The adjusted response rate was 55.6%. Forty-
three percent of surgeons reported sometimes or always
experiencing conflict about postoperative goals of care
with intensivists, and 43% reported conflict with nurses.
Younger surgeons reported higher rates of conflict than
older surgeons with both intensivists (57% vs 32%;
P = .001) and nurses (48% vs 33%; P = .001). Surgeons
practicing in closed intensive care units reported more
frequent conflict than those practicing in open intensive care units (60% vs 41%; P=.005). On multivariate
analysis, the odds of reporting conflict with intensivists
were 2.5 times higher for surgeons with fewer years of
experience compared with their older colleagues (odds
ratio, 2.5; 95% CI, 1.6-3.8) and 70% higher for reporting conflict with nurses (odds ratio, 1.7; 95% CI, 1.12.6). The odds of reporting conflict with intensivists about
goals of postoperative care were 40% lower for surgeons
who primarily managed their intensive care unit patients than for those who worked in a closed unit (odds
ratio, 0.60; 95% CI, 0.40-0.96).
Conclusions: Surgeons regularly experience conflict with
critical care clinicians about goals of care for patients with
poor postoperative outcomes. Higher rates of conflict are
associated with less experience and working in a closed
intensive care unit.
JAMA Surg. 2013;148(1):29-35
C
Author Affiliations are listed at
the end of this article.
ONFLICT IN THE INTEN sive care unit (ICU) is a
significant public health
problem, as more than
70% of ICU clinicians report experiencing conflict weekly.1,2 The
combination of caring for acutely ill patients, end-of-life decision making, and coordination of large multidisciplinary teams
can lead to frustration, communication
breakdown, and discord among members of a health care team. The epidemiology of ICU conflict is well described.1
This conflict has been associated with
lower-quality patient care,3,4 higher rates
of medical error,5 higher levels of staff
burnout,6,7 and greater direct and indirect costs of care.2,8 Intensive care unit conflict can occur between the health care
team and patients’ families; among members of the intensive care team (intrateam conflict); and between different
JAMA SURG/ VOL 148 (NO. 1), JAN 2013
29
groups of clinicians caring for the same patient (interteam conflict), most notably between surgeons and intensivists.1-4,9
See Invited Critique
at end of article
Two primary contributors to ICU conflict are particularly relevant to surgeons:
patient-doctor relationships formed prior
to the ICU admission and discussions of
end-of-life care.3,4,10 Others have shown
that surgeons have a strong sense of personal responsibility for patient outcomes
that may influence surgeons’ interactions
with critical care clinicians as well as discussions about end-of-life care.3,10-15 Surgeons are often reluctant to switch goals
of care from cure to comfort, particularly
in the postoperative period. 3,10,14 Although these sources of conflict have been
WWW.JAMASURG.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
Author Aff
Departmen
University
Hospital an
Olson); Dep
Division of
(Dr Schwar
Wisconsin
and Public
Redmann),
Departmen
College of W
Milwaukee
Wisconsin;
Medicine an
for Clinical
University
Departmen
Practice, Un
Chicago Sc
(Dr Alexan
Illinois; and
Epidemiolo
Health and
Medicine, J
University,
(Dr Alexan
well described by intensivists, it is unknown whether surgeons appreciate these conflicts.
We examined whether surgeons recognized and reported conflict with intensivists and nurses about goals
of care for their patients, specifically in the setting of a
poor postoperative outcome. In addition, we explored how
often surgeons reported conflict with ICU clinicians as
well as surgeon factors associated with such conflict.
METHODS
PARTICIPANTS
We selected a random sample of neurologic, vascular, and cardiothoracic surgeons. We chose these specialties because they
were likely to perform high-risk operations and have patients
who frequently require intensive care postoperatively. We excluded other surgeons who routinely care for patients in the
ICU to avoid specific confounding issues. Trauma surgeons were
excluded owing to the routine performance of emergency surgery, transplant surgeons were excluded owing to their concern for resource allocation, and surgical oncologists were excluded owing to the heterogeneous nature of surgical oncology
whereby surgical oncologists who specialize in breast or endocrine surgery would be unlikely to care for patients in the
ICU. We randomly selected participants from the membership lists of the American Association for Neurological Surgery Cerebrovascular Division, regional vascular surgery societies (Midwestern, New England, Eastern, and Western
societies), and the Society for Thoracic Surgery.
We mailed a total of 2100 surveys, 700 to each specialty,
via the US Postal Service. The survey packet also contained a
stamped addressed return envelope and a laser-pointer pen valued at $2.85 as an incentive to complete the survey. In March
2010, we mailed the first round of surveys. We sent a second
mailing (including a return envelope but no pen incentive) to
nonrespondents. Finally, because of a high proportion of nonrespondents from the neurosurgical group owing to incorrect
addresses, we mailed a third survey with an additional laserpointer pen as well as a letter of support from a key neurosurgical opinion leader. Prior to this third mailing, we validated
these addresses through an Internet search. We could not verify
the addresses of 180 members of the original neurosurgical cohort; therefore, we replaced these members with 180 randomly selected new participants. The survey was completed in
August 2010.
This study was approved as exempt, including a waiver of
written consent, by the institutional review boards of the University of Wisconsin and University of Chicago.
SURVEY
We designed a survey to assess surgeon attitudes and practices regarding advanced directives, withdrawal of lifesupporting therapies, and decision making surrounding highrisk operations. First, we performed a qualitative study using
semistructured interviews of surgeons and other physicians involved in perioperative care to examine surgeons’ attitudes and
routines surrounding the use of advanced directives and withdrawal of life-supporting therapies postoperatively.15,16 In addition to work by others, this study identified conflict between surgeons and ICU clinicians that stems from decisions
about the use or withdrawal of life-supporting therapies in postoperative patients.3,4,10,14,15 Using these results, we designed survey questions to explore the validity and generalizability of our
qualitative findings to a larger group of surgeons.
To assess the face validity of the survey questions, we performed cognitive interviews with 10 surgeons. To avoid interviewing actual survey recipients, the surgeons who were interviewed performed high-risk operations routinely (transplant,
surgical oncology, and trauma) but were not members of the
subspecialties included in the study target population. To inform questionnaire design, we asked respondents to think aloud
as they read survey items and explain their interpretation of
each item to ensure that the intended focus of the question was
clearly understood.17,18 We incorporated each respondent’s input in a stepwise fashion.
Our final survey included items about the surgeons’ experiences with conflict during the care of patients who had poor postoperative outcomes. Using a 4-point Likert scale of never, rarely,
sometimes, or always, respondents were asked to rate how often
they experienced conflict about the goals of care for their postoperative patients with poor outcomes with critical care physicians, nursing staff, and other groups. In addition, we gathered
data on surgeon sex, specialty, number of years in practice, practice setting (academic vs private practice), number of high-risk
procedures performed per month (defined as procedures having
⬎1% operative mortality or significant morbidity), percentage of
patients routinely requiring ICU care postoperatively, and the administrative model for the main ICU where the surgeon practices (open, closed, mixed, or other). We defined a closed ICU
as a unit where an intensive care physician was primarily responsible for all patients and an open ICU as a unit where the operative surgeon was primarily responsible for his or her patients. A
mixed ICU combines elements of both open and closed units.
STATISTICAL ANALYSIS
After removing all surveys that were returned to sender and
surveys completed by ineligible respondents (junior residents
and nonphysicians), an Internet search was used to estimate
the percentage of nonrespondents owing to inaccurate contact information. We used a sample of 60 respondents (20 from
each specialty) and 60 nonrespondents to estimate the percentage of nonrespondents owing to faulty contact information. Using the American Association for Public Opinion Research guidelines, we calculated the adjusted response rate with
the following formula: response rate=R/([R]⫹e[T⫺R⫺NE]),
where R = eligible respondents, e = proportion of nonrespondents estimated to be ineligible, T=total number of surveys,
and NE=ineligible respondents (including return to sender).19
As a surrogate marker for nonresponse bias, we looked for evidence of forward response wave bias. To do this, we calculated response time for each survey and identified clusters of
early and late respondents. Rates of surgeon-reported conflict
were compared between the early and late respondents.
We defined our primary outcome as surgeon-reported conflict with critical care physicians and nursing staff. We collapsed the response frame for conflict into dichotomous variables (never/rarely vs sometimes/always). On sensitivity analysis,
we found no difference in our outcomes between the collapsed responses and the 4-point response frame. We performed bivariate analyses using the ␹2 test. We next developed a multivariate logistic regression model including basic
demographic characteristics, such as sex and surgical subspecialty (vascular, cardiothoracic, or neurologic surgery); variables that were significant on bivariate analysis at P = .10 included years in clinical practice, practice type (private, academic,
or private with academic affiliation), and ICU administrative
model (closed, open, or mixed). Of note, we did not find that
the ICU administrative model was a statistically significant variable for conflict with nursing staff on bivariate analysis (P=.85),
but we included it in our multivariate model of conflict with
JAMA SURG/ VOL 148 (NO. 1), JAN 2013
30
WWW.JAMASURG.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
Table 1. Characteristics of 912 Survey Respondents a
Characteristic
Table 2. Rates of Surgeon-Reported Conflict and Challenges
No. (%)
Specialty
Vascular
Neurologic
Cardiothoracic
Time in practice, y
0-10
11-20
21-30
ⱖ31
Sex
Male
Female
Practice type
Private
Academic
Private with academic affiliation
Other
ICU administrative model
Closed
Open
Mixed
Other
No. of high-risk procedures per mo
0
1-5
6-10
ⱖ11
Question
327 (36)
273 (30)
312 (34)
196 (23)
215 (25)
233 (27)
223 (26)
850 (94)
50 (6)
343 (38)
323 (36)
167 (19)
61 (7)
96 (11)
277 (32)
500 (57)
4 (0.5)
34 (4)
311 (37)
256 (30)
239 (28)
Abbreviation: ICU, intensive care unit.
a Numbers do not sum to 912 because not all respondents answered all
questions.
nursing staff because this was a variable of considerable interest for our explanatory model of ICU conflict. All statistical analyses were performed using SAS version 9.1 (SAS Institute Inc).
RESULTS
Answer, %
At times, conflicts may arise among different
parties involved in the care of a patient who has
a poor postsurgical outcome. How frequently,
if ever, do you experience conflict with each of
these groups about the goals of care for your
postoperative patients?
Critical care physicians
Nursing staff
Primary care physicians
Ethics consultants
Family members of the patient
Surgical colleagues
How challenging, if at all, do you find each of the
following aspects of surgical practice?
Communicating with family and/or patient
about poor outcomes
Addressing patient fears regarding morbidity
and mortality
Managing personal discomfort about poor
outcomes
Managing clinical aspects of poor outcomes
Emphasis on outcome measures such as
physician profiling
Sometimes/
always
43
43
23
16
60
18
Somewhat/very
challenging
64
42
73
62
50
cians, and a similar percentage (43%) reported sometimes or always experiencing conflict with nursing staff
about the goals of postoperative care (Table 2).
Although 50% of our respondents felt that the emphasis on outcome measures such as physician profiling was a challenging aspect of surgical practice, a greater
number of surgeons identified the difficulty of managing clinical aspects related to poor outcomes or communicating with patients and their families as a somewhat
or very challenging aspect of surgical practice. Moreover, 73% of respondents noted that managing their own
discomfort about poor outcomes was a significant challenge (Table 2).
PARTICIPANTS
We received completed surveys from 912 respondents,
and 203 were returned to sender. The adjusted response rate as calculated by the American Association for
Public Opinion Research guidelines was 55.6%. There was
no significant difference in the rates of reported conflict
between early and late respondents, suggesting no evidence for forward response wave bias, our surrogate measure for nonresponse bias.
Survey respondents were evenly distributed with respect to years in practice (Table 1). There were approximately equal proportions of respondents who practiced in
an academic vs a private setting. Most respondents (57%)
described the administrative model for their main ICU as
mixed. About one-third of respondents practiced primarily with an open ICU model (32%), whereas a smaller proportion worked in hospitals with a closed ICU model (11%).
SURGEON-REPORTED CONFLICT
Forty-three percent of surgeons reported sometimes or
always experiencing conflict with critical care physi-
CONFLICT WITH
CRITICAL CARE PHYSICIANS
On bivariate analysis (Table 3), surgeons with less than
10 years of experience were significantly more likely than
surgeons with more than 30 years of experience to report conflict with critical care physicians (57% vs 32%;
P ⬍ .001). In addition, surgeons in academic practices
reported significantly more conflict compared with those
in private practice (52% vs 36%; P ⬍ .001). Forty-one
percent of surgeons practicing in an open ICU and 42%
practicing in a mixed ICU reported conflict with critical
care doctors, whereas 60% of surgeons practicing within
a closed ICU model reported conflict with critical care
doctors (P = .005).
On multivariate logistic regression, a strong and statistically significant association persisted between surgeon experience and surgeon-reported conflict with critical care physicians. The odds of surgeons reporting conflict
with critical care physicians were 2.5 times higher for surgeons with less than 10 years of experience than for those
with more than 30 years of experience (odds ratio, 2.5;
JAMA SURG/ VOL 148 (NO. 1), JAN 2013
31
WWW.JAMASURG.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
Table 3. Surgeon Characteristics Associated With Conflict With Intensivists a
Characteristic
Sex
Female
Male
Specialty
Neurologic
Vascular
Cardiothoracic
Time in practice, y
ⱖ31
21-30
11-20
0-10
Practice type
Private
Academic
Private with academic affiliation
Other
ICU administrative model
Closed
Open
Mixed
Other
Total No.
Reported Conflict
With Intensivists, %
Bivariate
P Value
Multivariate Odds Ratio
(95% CI)
50
838
54
43
.12
1.0 [Reference]
0.8 (0.4-1.6)
269
323
308
38
47
45
.08
1.0 [Reference]
1.6 (1.1-2.3)
1.5 (1.0-2.1)
217
231
213
196
32
40
45
57
⬍.001
1.0 [Reference]
1.5 (1.0-2.2)
1.8 (1.2-2.8)
2.5 (1.6-3.8)
340
318
163
61
36
52
44
41
⬍.001
1.0 [Reference]
1.4 (1.0-2.0)
1.3 (0.8-1.9)
1.1 (0.6-2.0)
96
272
496
4
60
41
42
25
.005
1.0 [Reference]
0.6 (0.4-1.0)
0.60 (0.40-0.96)
0.20 (0.02-2.30)
Abbreviation: ICU, intensive care unit.
a Number does not sum to 912 because not all respondents answered all questions.
Table 4. Surgeon Characteristics Associated With Conflict With Nursing Staff a
Characteristic
Sex
Female
Male
Specialty
Neurologic
Vascular
Cardiothoracic
Time in practice, y
ⱖ31
21-30
11-20
0-10
Practice type
Private
Academic
Private with academic affiliation
Other
ICU administrative model
Closed
Open
Mixed
Other
Total No.
Reported Conflict
With Nurses, %
Bivariate
P Value
Multivariate Odds Ratio
(95% CI)
50
838
54
42
.11
1.0 [Reference]
0.7 (0.4-1.3)
269
323
308
39
47
43
.17
1.0 [Reference]
1.6 (1.1-2.2)
1.3 (0.9-1.9)
217
232
213
195
33
42
50
48
.001
1.0 [Reference]
1.4 (0.9-2.1)
1.8 (1.2-2.8)
1.7 (1.1-2.6)
341
317
163
61
39
50
43
34
.02
1.0 [Reference]
1.4 (1.0-2.0)
1.1 (0.7-1.6)
0.8 (0.5-1.5)
95
273
496
4
45
44
43
25
.85
1.0 [Reference]
1.3 (0.8-2.2)
1.1 (0.7-1.7)
0.40 (0.03-3.90)
Abbreviation: ICU, intensive care unit.
a Numbers do not sum to 912 because not all respondents answered all questions.
95% CI, 1.6-3.8). In addition, the odds of surgeons reporting conflict with critical care physicians was 40%
lower for surgeons practicing within a mixed or open
model of ICU administration compared with those practicing in a closed ICU (mixed: odds ratio, 0.60; 95% CI,
0.40-0.96; open: odds ratio, 0.6; 95% CI, 0.4-1.0).
CONFLICT WITH NURSING STAFF
On bivariate analysis (Table 4), an association between surgeon experience and reported conflict with
nurses was demonstrated. Surgeons with less than 10
years of experience were significantly more likely than
JAMA SURG/ VOL 148 (NO. 1), JAN 2013
32
WWW.JAMASURG.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
surgeons who had been in practice for more than 30
years to report conflict with nursing staff regarding
goals of care for postoperative patients who had poor
postsurgical outcomes (48% vs 33%; P = .001). Surgeons in academic settings reported more conflict with
ICU nurses than their counterparts in private practice
(50% vs 39%; P = .02). In contrast to the experience of
conflict with critical care physicians, the model of ICU
administration was not associated with significantly
higher rates of conflict with nurses (open: 44%, mixed:
43%, closed: 45%; P = .85).
On multivariate analysis, a statistically significant association was found between surgeons’ experiences and
reported rates of conflict with nursing staff. The odds
of surgeons reporting conflict with nursing staff was
70% greater in surgeons with less than 10 years of experience compared with surgeons with more than 30
years in practice (odds ratio, 1.7; 95% CI, 1.1-2.6). The
administrative model of ICU care was not associated
with increased rates of conflict reported by surgeons
with nurses.
COMMENT
More than 40% of surgeons who routinely perform highrisk operations reported conflict with critical care physicians and nurses regarding the goals of care for their
patients with poor postoperative outcomes. Surgeons who
reported higher rates of conflict had fewer years in practice and worked in an academic setting. Additionally, surgeons who practiced in a closed ICU reported higher rates
of conflict about the goals of care with critical care physicians, but not with critical care nurses. These findings
have important implications for surgeons, critical care
clinicians, and their patients.
For surgeons, these findings suggest that learning how
to manage personal discomfort in the setting of a poor
postoperative outcome may take time and experience.
Prior work by our group and others has shown that surgeons have a particularly difficult time switching goals
of care from cure to comfort postoperatively, especially
given the perception of a direct link between the surgeon’s performance and the patient’s outcome.11,13,14,20,21
When critical care clinicians who have had no direct role
in the patient’s operation suggest a change in the goals
of care from survival to comfort, conflict may ensue. Surgeons with more experience may be more accepting of
the limits of surgical therapy and may have developed
more robust coping strategies for the inevitable unwanted outcome.
For critical care clinicians, it is important to note
that the ICU administrative model can contribute to
conflict with surgeons. Surgeons typically have a relationship with their patient that begins preoperatively
and extends throughout the patient’s recovery. Because
of this longitudinal relationship, surgeons see themselves as the primary decision makers for their postoperative patients. The structure of a closed ICU likely
poses a barrier to the continuity of the surgeon-patient
relationship and contributes to conflict when the surgeon is replaced as the primary decision maker for his
or her patient by an intensivist.3,13-15,21 As Cassell and
colleagues14 noted, intensivists working in a closed ICU
are charged with allocation of scarce resources for the
good of the entire unit, while surgeons focus primarily
on their covenantal responsibilities to individual patients. The administrative model of a closed ICU thus
promotes conflict with its juxtaposition of clinicians
with competing viewpoints.
Patients and their families are likely affected by the
conflict that surgeons report with critical care clinicians. This can impact the overall experience of patients
and their families in the ICU, with perceived conflict leading to decreased satisfaction with care and increased stress
for families.1-4,9,22,23 Changing the goals of care postoperatively may trigger conflict when those caring for the
patient try to determine what is most in line with the patient’s preferences. Our findings about surgeonreported conflict with intensivists underscore how important it is for patients to discuss their goals and values
with their surgeon and surrogate decision makers preoperatively to inform postoperative decisions in the event
of an undesired outcome.15,16,24,25 Patients and their surrogates can be referred to other available resources for
conflict adjudication to help guide treatment decisions
such as a second opinion from a panel of senior clinicians or the hospital ethics committee.3,9,13,21,23,26,27
Our study has several limitations. Survey-based research can be subject to nonresponse bias, where nonrespondents differ systematically from respondents. However, our response rate was robust, and there was no
evidence of forward response wave bias, making the possibility of nonresponse bias low. Because conflict is not
considered acceptable behavior, questions about conflict are subject to social desirability bias. As such, the
true rate of conflict might be higher than our respondents reported. This survey also targeted 3 surgical subspecialties (vascular, cardiothoracic, and neurologic surgery) to ensure the respondents routinely care for critically
ill patients. It is unclear whether our results are generalizable to other surgical subspecialties.
Our survey did not specifically define the parameters
of a poor postoperative outcome, thus we cannot specify
a clinical threshold that may prompt such conflicts. Also,
we did not distinguish between closed ICUs run by surgeons vs closed ICUs managed by nonsurgeons, although this may prove to be an important determinant
in rates of surgeon-reported conflict.21,26,28 Finally, although we noted a high rate of surgeon-reported conflict between surgeons and patients’ family members, we
chose not to focus on this finding here. The primary focus of our survey was physician practices and, as such,
we did not generate enough information for a complete
discussion of this important topic.
Surgeons frequently experience conflict with critical
care physicians and nurses about the goals of care for their
postoperative patients with poor outcomes. Higher rates
of conflict are reported by surgeons with fewer years of
experience and those working in institutions with a closed
model of ICU administration. This conflict is a significant public health problem that diminishes quality of care
for critically ill patients and their families. Given the
myriad challenges inherent in delivering the highest qual-
JAMA SURG/ VOL 148 (NO. 1), JAN 2013
33
WWW.JAMASURG.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
ity of care in these settings, clinicians from all backgrounds should focus on eliminating these interteam conflicts to allow energies to be spent more productively on
other clinical issues affecting safety and quality. Interventions directed at the individual level as well as the system level will be important to mitigate conflict to provide better care for our critically ill postoperative patients.
tistical analyses as well as Ann Chodara at the University of Wisconsin School of Medicine and Public
Health for her assistance with data collection and
management.
Accepted for Publication: April 30, 2012.
Author Affiliations: Department of Surgery, University
of Wisconsin Hospital and Clinics (Dr Paul Olson); Department of Surgery, Division of Vascular Surgery (Dr
Schwarze), University of Wisconsin School of Medicine
and Public Health (Mr Redmann), Madison; Department of Surgery, Medical College of Wisconsin, Milwaukee (Dr Brasel), Wisconsin; Department of Medicine and
MacLean Center for Clinical Medical Ethics, University
of Chicago; Department of Pharmacy Practice, University of Illinois at Chicago School of Pharmacy (Dr Alexander), Chicago, Illinois; and Departments of Epidemiology, School of Public Health and Medicine, School of
Medicine, Johns Hopkins University, Baltimore, Maryland (Dr Alexander).
Correspondence: Margaret L. Schwarze, MD, MPP, Department of Surgery, Division of Vascular Surgery, University of Wisconsin, G5/315 CSC, 600 Highland Ave,
Madison, WI 53792 ([email protected]).
Author Contributions: Dr Schwarze had full access to
all data in this study and takes responsibility for the integrity of the data and accuracy of the data analysis. Study
concept and design: Brasel, Alexander, and Schwarze. Acquisition of data: Redmann and Schwarze. Analysis and
interpretation of data: Paul Olson, Brasel, Redmann, Alexander, and Schwarze. Drafting of the manuscript: Paul
Olson and Schwarze. Critical revision of the manuscript
for important intellectual content: Brasel, Redmann, Alexander, and Schwarze. Statistical analysis: Paul Olson,
Brasel, Redmann. Obtained funding: Schwarze. Administrative, technical, and material support: Schwarze. Study
supervision: Alexander and Schwarze.
Conflict of Interest Disclosures: None reported.
Funding/Support: Dr Paul Olson’s work is supported by
grant T32 CA090217 from the National Institutes of
Health. Mr Redmann’s work was supported by a Shapiro Summer Research Award from the University of Wisconsin School of Medicine and Public Health. Dr Alexander’s work is supported by grants K08 HS15699 and
R01 HS0189960 from the Agency for Healthcare Research and Quality. Dr Schwarze’s work is supported by
a Greenwall Faculty Scholars Award and the Department of Surgery at the University of Wisconsin.
Role of the Sponsors: The funding sources had no role
in the design and conduct of the study; collection, management, analysis, or interpretation of the data; and
preparation, review, or approval of the manuscript for
publication.
Previous Presentation: This work was presented in part
at the 7th Annual Academic Surgical Congress; February 16, 2012; Las Vegas, Nevada.
Additional Contributions: We thank Glen Leverson,
PhD, in the Department of Surgery at the University of
Wisconsin, Madison, for his invaluable help with sta-
1. Azoulay E, Timsit JF, Sprung CL, et al; Conflicus Study Investigators and for the
Ethics Section of the European Society of Intensive Care Medicine. Prevalence
and factors of intensive care unit conflicts: the Conflicus Study. Am J Respir Crit
Care Med. 2009;180(9):853-860.
2. Fassier T, Azoulay E. Conflicts and communication gaps in the intensive care unit
[published online October 7, 2012]. Curr Opin Crit Care. doi:10.1097/MCC
.0b013e32834044f0.
3. Danjoux Meth N, Lawless B, Hawryluck L. Conflicts in the ICU: perspectives of
administrators and clinicians. Intensive Care Med. 2009;35(12):2068-2077.
4. Studdert DM, Mello MM, Burns JP, et al. Conflict in the care of patients with prolonged stay in the ICU: types, sources, and predictors. Intensive Care Med. 2003;
29(9):1489-1497.
5. Baldwin DC Jr, Daugherty SR. Interprofessional conflict and medical errors: results of a national multi-specialty survey of hospital residents in the US. J Interprof Care. 2008;22(6):573-586.
6. Embriaco N, Azoulay E, Barrau K, et al. High level of burnout in intensivists: prevalence and associated factors. Am J Respir Crit Care Med. 2007;175(7):686692.
7. Poncet MC, Toullic P, Papazian L, et al. Burnout syndrome in critical care nursing staff. Am J Respir Crit Care Med. 2007;175(7):698-704.
8. Brinkert R. A literature review of conflict communication causes, costs, benefits
and interventions in nursing. J Nurs Manag. 2010;18(2):145-156.
9. Breen CM, Abernethy AP, Abbott KH, Tulsky JA. Conflict associated with decisions to limit life-sustaining treatment in intensive care units. J Gen Intern Med.
2001;16(5):283-289.
10. Larochelle MR, Rodriguez KL, Arnold RM, Barnato AE. Hospital staff attributions of the causes of physician variation in end-of-life treatment intensity. Palliat Med. 2009;23(5):460-470.
11. Bosk CL. Forgive and Remember: Managing Medical Failure. 2nd ed. Chicago,
IL: The University of Chicago Press; 1979.
12. Katz P. The Scalpel’s Edge: The Culture of Surgeons. Needham Heights, MA: Allyn and Bacon; 1999.
13. Buchman TG, Cassell J, Ray SE, Wax ML. Who should manage the dying patient? rescue, shame, and the surgical ICU dilemma. J Am Coll Surg. 2002;
194(5):665-673.
14. Cassell J, Buchman TG, Streat S, Stewart RM. Surgeons, intensivists, and the
covenant of care: administrative models and values affecting care at the end of
life: updated. Crit Care Med. 2003;31(5):1551-1557, discussion 1557-1559.
15. Schwarze ML, Bradley CT, Brasel KJ. Surgical “buy-in”: the contractual relationship between surgeons and patients that influences decisions regarding lifesupporting therapy. Crit Care Med. 2010;38(3):843-848.
16. Bradley CT, Brasel KJ, Schwarze ML. Physician attitudes regarding advance directives for high-risk surgical patients: a qualitative analysis. Surgery. 2010;
148(2):209-216.
17. Conrad F, Blair J. From impression to data: increasing the objectivity of cognitive interviews. In: Proceedings of the Survey Research Methods Section of the
American Statistical Association. Alexandria, VA: American Statistical Association; 1996.
18. Drennan J. Cognitive interviewing: verbal data in the design and pretesting of
questionnaires. J Adv Nurs. 2003;42(1):57-63.
19. American Association for Public Opinion Research. Standard Definitions: Final
Dispositions of Case Codes and Outcome Rates for Surveys. 7th ed. Deerfield,
IL: American Association for Public Opinion; 2011.
20. Schwarze ML, Redmann AJ, Brasel KJ, Alexander GC. The role of surgeon error
in withdrawal of postoperative life support. Ann Surg. 2012;256(1):10-15.
21. Penkoske PA, Buchman TG. The relationship between the surgeon and the intensivist in the surgical intensive care unit. Surg Clin North Am. 2006;86(6):
1351-1357.
22. Abbott KH, Sago JG, Breen CM, Abernethy AP, Tulsky JA. Families looking back:
one year after discussion of withdrawal or withholding of life-sustaining support.
Crit Care Med. 2001;29(1):197-201.
23. Buchman TG, Ray SE, Wax ML, Cassell J, Rich D, Niemczycki MA. Families’ perceptions of surgical intensive care. J Am Coll Surg. 2003;196(6):977-983.
24. Redmann AJ, Brasel KJ, Alexander CG, Schwarze ML. Use of advance directives
for high-risk operations: a national survey of surgeons. Ann Surg. 2012;255
(3):418-423.
REFERENCES
JAMA SURG/ VOL 148 (NO. 1), JAN 2013
34
WWW.JAMASURG.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
25. Sudore RL, Fried TR. Redefining the “planning” in advance care planning: preparing for end-of-life decision making. Ann Intern Med. 2010;153(4):256-261.
26. Eachempati SR, Miller FG, Fins JJ. The surgical intensivist as mediator of endof-life issues in the care of critically ill patients. J Am Coll Surg. 2003;197(5):
847-853, discussion 853-854.
27. Schaefer KG, Block SD. Physician communication with families in the ICU: evidencebased strategies for improvement. Curr Opin Crit Care. 2009;15(6):569-577.
28. Whitman GJ, Haddad M, Hirose H, Allen JG, Lusardi M, Murphy MA. Cardiothoracic surgeon management of postoperative cardiac critical care. Arch Surg. 2011;
146(11):1253-1260.
INVITED CRITIQUE
Patient Ownership and Conflict
in the Intensive Care Unit
Mine, Yours, or Ours?
P
aul Olson and colleagues1 attempt to answer the
question of patient ownership with a survey of
surgeons who frequently have patients in the intensive care unit. In view of the increased prevalence of
hospitalists—whether they be intensivists, trauma surgeons, anesthesiologists, or internists—and the current
resident work-hour restrictions, the issue of continuity
of care is an increasingly important one. In the first few
sentences of their article, the authors set the stage, citing recent critical care publications in which 70% of intensive care unit clinicians reported experiencing conflict on a weekly basis!
This subject is an important one, particularly as it pertains to surgeons’ difficulty accepting bad outcomes in
their patients and de-escalating care. No event in the life
of a surgeon surpasses the physical and emotional impact of a bad surgical outcome. We are, by training, taught
to help others. When as a consequence of our direct action (an operation or treatment) an adverse outcome ensues, there is often difficulty in giving up—a difficulty
in admitting that this situation cannot be salvaged. The
article by Paul Olson et al1 provides valuable insight. One
of the better messages in this article is the no doubt guilt-
accentuated concerns about the patient with a less-thandesired outcome; predictably, the operating surgeon is
less prompt to move toward palliative care.
Just who is the more consistent and ethical conservator of resources varies widely in the hospitals in our
university medical center. I do not believe intensivists
are more protective of those resources; more likely, the
opposite is true. The reader must also be wondering how
intensive care unit nurses view these complex and intense interactions. These kinds of questions are hard to
pose meaningfully and even more difficult to answer.
Susan Galandiuk, MD
Author Affiliation: Department of Surgery, University of
Louisville, Kentucky.
Correspondence: Dr Galandiuk, Department of Surgery, University of Louisville, Louisville, KY 40292
([email protected]).
Conflict of Interest Disclosures: None reported.
1. Paul Olson TJ, Brasel KJ, Redmann AJ, Alexander GC, Schwarze ML. Surgeonreported conflict with intensivists about postoperative goals of care. JAMA
Surg. 2013;148(1):29-35.
JAMA SURG/ VOL 148 (NO. 1), JAN 2013
35
WWW.JAMASURG.COM
©2013 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
Factors Associated With Elevated Plateau Pressure in
Patients With Acute Lung Injury Receiving Lower Tidal
Volume Ventilation
Hallie C. Prescott, MD1,2; Roy G. Brower, MD3; Colin R. Cooke, MD, MSc4; Gary Phillips, MAS5;
James M. O’Brien, MD, MSc6; for the National Institutes of Health Acute Respiratory Distress Syndrome
Investigators
Objectives: Lung-protective ventilation with lower tidal volume and
lower plateau pressure improves mortality in patients with acute lung
injury and acute respiratory distress syndrome. We sought to determine the incidence of elevated plateau pressure in acute lung injury
/acute respiratory distress syndrome patients receiving lower tidal
volume ventilation and to determine the factors that predict elevated
plateau pressure in these patients.
Patients: We used data from 1398 participants in Acute Respiratory
Distress Syn­drome Network trials, who received lower tidal volume
ventilation (≤ 6.5 mL/kg predicted body weight).
Design: We considered patients with a plateau pressure greater than
30 cm H2O and/or a tidal volume less than 5.5 mL/kg predicted body
weight on study day 1 to have “elevated plateau pressure.” We used
logistic regression to identify baseline clinical variables a­ ssociated
with elevated plateau pressure and to develop a model to predict
elevated plateau pressure using a subset of 1,188 patients. We
validated the model in the 210 patients not used for model development.
Setting: Medical centers participating in Acute Respiratory Distress
Syn­drome Network clinical trials.
1Department of Internal Medicine, The Ohio State University Medical Center,
Columbus, OH.
2Division of Pulmonary and Critical Care Medicine, University of Michigan
Health System, Ann Arbor, MI.
3Pulmonary and Critical Care Medicine, John Hopkins University School of
Medicine, Baltimore, MD.
4University of Michigan, Robert Wood Johnson Clinical Scholars Program
and Center for Healthcare Outcomes & Policy, Ann Arbor, MI.
5The Ohio State University Center for Biostatistics, Columbus, OH.
6Division of Pulmonary, Allergy, Critical Care & Sleep Medicine, Department
of Internal Medicine, The Ohio State University Medical Center, Columbus,
OH.
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions
of this article on the journal’s Web site (http://journals.lww.com/ccmjournal).
Supported, in part, by National Heart, Lung, and Blood Institute (NHLBI) contract NO1-HR-46054–46064 and NO1-HR 56165–56179.s
The authors have not disclosed any potential conflicts of interest.
For information regarding this article, E-mail: [email protected]
Critical Care Medicine
Interventions: None.
Measurements and Main Results: Of the 1,398 patients in our study,
288 (20.6%) had elevated plateau pressure on day 1. Severity of
illness indices and demographic factors (younger age, greater body
mass index, and non-white race) were independently associated
with elevated plateau pressure. The multivariable logistic regression
model for predicting elevated plateau pressure had an area under
the receiving operator characteristic curve of 0.71 for both the
developmental and the validation subsets.
Conclusions: acute lung injury patients receiving lower tidal volume
ventilation often have a plateau pressure that exceeds Acute Respiratory Distress Syn­drome Network goals. Race, body mass index,
and severity of lung injury are each independently associated with
elevated plateau pressure. Selecting a smaller initial tidal volume for
non-white patients and patients with higher severity of illness may
decrease the incidence of elevated plateau pressure. Prospective studies are needed to evaluate this approach. (Crit Care Med
2013; 41:0–0)
Key Words: acute lung injury; acute respiratory distress syndrome;
mechanical ventilation; plateau pressure
A
cute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are common and deadly diagnoses that
affect 190,000 patients (1) each year in the United States
and are associated with a 36–44% mortality rate (1–3). Mechanical ventilation strategies in ALI patients aim to minimize
ventilator-induced lung injury while providing adequate oxygenation and ventilation. Accurate bedside measures of lung stress
and lung strain are not readily available, so clinicians rely on surrogate measures of transpulmonary pressure to guide ventilator
management. Despite its limitations (4), plateau pressure is used
commonly as a surrogate measure of transpulmonary pressure.
The association between plateau pressure (Pplat) and mortality, as
Copyright © 2013 by the Society of Critical Care Medicine and Lippincott
Williams & Wilkins
DOI:10.1097/CCM.0b013e3182741790
www.ccmjournal.org
1
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Prescott et al
has been demonstrated in retrospective analyses using ARDS Network trial participants, supports this practice (5, 6).
In 2000, the ARDS Network published a multicenter trial comparing lower tidal volume (VT) ventilation (6 mL/kg predicted
body weight [PBW]) to traditional ventilation (12 mL/kg PBW) in
ALI patients. The lower VT arm had an 8.8% lower mortality rate
compared with the control arm (7). As a means of scaling VT to
lung size, the ARDS Network protocol calculated a PBW for each
patient based on sex and height, and then used this PBW to determine a VT (7). VT was then adjusted based on estimates of lung
stress (second order scaling). Plateau pressure was monitored every 4 hrs, and VT reduced if Pplat exceeded 30 cm H2O in the lower
VT arm or 50 cm H2O in the traditional ventilation arm (7).
Since publication of the ARDS Network lower VT trial, physicians have prescribed smaller VTs (5, 8). However, Pplat monitoring
is also part of the ARDS Network ventilation protocol, there is
less evidence to suggest that Pplat monitoring or reduction of VT
in response to elevated Pplat has been adopted on a widespread
level. Several studies of physician practice patterns for treating
ALI do not comment on Pplat (9, 10) because this variable is frequently unavailable or underreported in datasets (8, 10–13), suggesting that Pplat monitoring does not occur regularly in clinical
practice. Moreover, the initial VT often remains in place for over
24 hrs (9, 14) even when the VT exceeds ARDS Network goals,
suggesting that physicians do not routinely reassess ventilator
settings.
Given the association between higher Pplat and mortality, and
the concern that Pplat monitoring has not been widely adopted, we
sought to determine the incidence of elevated Pplat in ALI patients
receiving a VT of 6 mL/kg PBW. Our second objective was to identify factors that predict elevated Pplat in patients receiving lower VT
ventilation.
MATERIALS AND METHODS
Study Subjects
We performed a secondary analysis of data collected from ALI/
ARDS patients enrolled in three ARDS Network clinical trials between 1996 and 2005, including a trial of different VTs (7), a trial
of different positive end-expiratory pressures (15), and a trial of
different hemodynamic management strategies (16, 17). Patients
with VT > 6.5 mL/kg PBW and those missing VT or Pplat data for
study day 1 were excluded from our study.
Definitions
The outcome of interest for this analysis was an “elevated Pplat”
on study day 1 (postenrollment). We defined “elevated Pplat”
as a measured Pplat > 30 cm H2O or a VT < 5.5 mL/kg PBW.
The reduction of VT to less than 6.0 mL/kg PBW was assumed to
be the result of a prior Pplat measurement greater than 30 cm H2O.
Statistical Analyses
Descriptive statistics used counts and frequencies for categorical
data and means and standard deviations for continuous variables.
To determine whether differences existed between the study population and those excluded for missing data, we used Pearson’s
2
www.ccmjournal.org
chi-square for categorical variables and Wilcoxon rank-sum test
for continuous variables. We used a two-sample t test to evaluate whether differences in Pplat existed across VT and Pearson’s chisquare to test if Pplat > 30 cm H2O was different across VT.
Modeling Approach
To determine the factors independently associated with elevated
Pplat, we used logistic regression to model the effect of baseline (preenrollment) variables on elevated Pplat. We performed two distinct
approaches to model development and validation: cross-validation
and bootstrap resampling (Online supplement, Supplemental Digital
Content 1, http://links.lww.com/CCM/A561). In the cross-validation
approach, the study population was randomly divided into development (85%) and validation (15%) groups. The regression model was
derived in the development sample, and model performance was assessed in the validation sample. The size of the validation group (210
subjects) was selected in order to provide enough data to perform a
goodness-of-fit test while still allowing the majority (1188 subjects)
to be used in estimating the model coefficients (18).
Model Selection
We first conducted a preliminary univariable analysis to determine the associations between baseline variables and the odds of
elevated Pplat on study day 1. Candidate variables for the multivariable logistic regression included: age, sex, race, body mass index
(BMI; calculated from enrollment height and weight), PaO2/Fio2
ratio, cirrhosis, chronic dialysis, serum albumin, net fluid balance over the 24 hrs preceding study enrollment, use of sedation
or neuromuscular blockade (NMB), number of quadrants of infiltrate on enrollment chest radiograph, APACHE III score, and
diabetes. We evaluated BMI as both a continuous variable and a
categorical variable using the previously defined classifications:
underweight < 18.5 kg/m2, normal 18.5–24.9 kg/m2, overweight
25–29.9 kg/m2, obese 30–34.9 kg/m2, and severely obese ≥ 35 kg/
m2. Race was recorded in the ARDS Network trials as white nonHispanic, black non-Hispanic, and other; we reclassified race as
white or non-white for our analysis.
We included all variables in the multivariable logistic regression
if they achieved a p value of less than or equal to 0.2 in a univariable
logistic regression with elevated Pplat as the outcome. The form of
the continuous variables was assessed using fractional polynomials
to ensure that the logit was linear (19). All variables considered for
model inclusion were tested for all possible two-way interactions.
All of the candidate variables and their interactions were included
in the multivariable model. Interactions were dropped from the
model one at a time if they were no longer statistically significant
(p > 0.050). This step was followed by dropping main effect
variables if they were no longer significant. The predictor variables
were also assessed for collinearity using variance inflation factor.
Model Performance
We evaluated the discrimination of the cross-validation model
using AUC in both the development and the validation cohorts.
We assessed calibration using the Hosmer-Lemeshow goodnessof-fit test (20, 21). Given the low power of the Hosmer-Lemeshow
goodness-of-fit test to detect poor performance of models with
March 2013 • Volume 41 • Number 3
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Clinical Investigation
RESULTS
Study Sample
The three ARDS Network studies
enrolled a total of 2,452 patients. Of
these, 785 (32.0%) were excluded
from our analysis for VT > 6.5 mL/kg
PBW, and 269 (11.0%) were excluded for missing day 1 Pplat and/or day 1
VT data (Fig. 1). Of those with missing data, 185 (68.8%) were missing
both Pplat and VT, 60 (22.3%) were
missing just Pplat, and 24 (8.9%) were
missing just VT. In the VT trial, 429
patients were randomized to a VT of
12 mL/kg (7); these patients account
for 54.6% of the patients excluded
from this study for VT > 6.5 mL/kg.
The 269 patients excluded for
missing data had similar baseline
characteristics to the 1,398 patients
(57.0%) included in the analysis
(data not shown), with the exception of a greater mean PaO2/Fio2
ratio (150.4 mm Hg vs. 126.9 mm
Hg, p < 0.001). Also, the patients excluded for missing data had a higher
proportion enrolled in the VT trial
(28.6% vs. 24.5%) and the hemodynamic management trial (49.1%
vs. 45.6%) and a smaller proportion
enrolled in the PEEP trial (22.3% vs.
30.0%) than those included in the
Figure 1. Flow chart of study. All patients enrolled in three Acute Respiratory Distress Syndrome (ARDS)
analysis (p = 0.031).
Network trials were screened for inclusion in our study. Patients receiving tidal volume (VT) > 6.5 mL/kg
Summary statistics for entire
predicted body weight (PBW) and patients missing plateau pressure (Pplat) and/or VT data for study day 1 were
excluded for our analysis. The remaining 1,398 subjects were categorized as having target vs. elevated Pplat on
study population, the developmenstudy day 1.
tal cohort, and the validation cohort
are provided in Table 1. The develsmall sample sizes (22), we also compared actual and predicted
opment and validation cohorts had
frequencies of elevated Pplat by decile of predicted probability.
similar baseline characteristics.
The overall study population of 1,398 patients had a mean Pplat
Sensitivity Analyses
of 24.3 ± 6.4 cm H2O on day 1.
In order to test the assumptions made in building the regression
The 183 patients with VT < 5.5 mL/kg PBW had a greater mean
Pplat and a higher frequency of Pplat > 30 cm H2O than patients with
model, we performed two sensitivity analyses. In the first, we removed patients with VT < 5.5 mL/kg PBW to evaluate whether VT 5.5–6.5 mL/kg PBW (Table 2). Histograms of Pplat by VT demour classification of these subjects as having elevated Pplat was ap- onstrate that the frequency distribution for Pplat is shifted upward
propriate. In the second, we removed patients with barotrauma about 5 cm H2O for patients receiving a VT < 5.5 mL/kg PBW in
comparison with patients receiving a VT of 5.5–6.5 mL/kg PBW
because we were unsure if or how barotrauma may have affected
(Fig. 2). The frequency of elevated Pplat did not differ by study
Pplat. We refit these revised datasets to the model from the develenrollment: (63/342) 18.4% of patients in the VT trial, (87/419)
opmental dataset.
20.8% of patients in the PEEP trial, and (138/637) 21.7% of paThe Institutional Review Board for The Ohio State University
tients in the hemodynamic management trial had elevated Pplat
approved the study. All statistical analyses were conducted using
(p = 0.487). On average, patients gained a net 2240 mL of fluid in
Stata 10.1, Stata Corporation, College Station, Texas.
the day preceding enrollment.
Critical Care Medicine
www.ccmjournal.org
3
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Prescott et al
Table 1.
Baseline Characteristics of Patients by Study Population
Develop­mental
Cohort
Validation
Cohort
Characteristic
Entire Study
Subjects, n (%)
1398 (100)
1188 (85)
210 (15)
Pplat (cm H2O), mean ± sd
24.3 ± 6.4
24.3 ± 6.4
24.6 ± 6.4
Elevated Pplat, %
288 (20.6)
240 (20.2)
48 (22.9)
Pplat > 30 cm H2O, n (%)
179 (12.8)
147 (12.4)
32 (15.2)
VT < 5.5 mL/kg PBW, n (%)
183 (13.1)
153 (13.1)
27 (12.9)
74 (5.3)
63 (5.3)
11 (5.2)
49.9 ± 16.7
49.9 ± 16.6
50.3 ± 17.3
Male, n (%)
810 (57.9)
685 (57.7)
125 (59.5)
Non-white, n (%)
420 (30.0)
357 (30.1)
63 (30.0)
White, non-Hispanic
978 (70.0)
831 (70.0)
147 (70.0)
Black, non-Hispanic
253 (18.1)
219 (18.4)
34 (16.2)
Other
167 (11.9)
138 (11.6)
29 (13.8)
Pplat > 30 cm H2O and VT < 5.5 mL/kg PBW, n (%)
Age (years), mean ± sd
Race, n (%)
BMI (kg/m ), mean ± sd
27.8 ± 7.1
2
27.8 ± 7.0
27.4 ± 7.3
Underweight BMI (<18.5 kg/m ), n (%)
60 (4.4)
50 (4.3)
Normal BMI (18.5–24.9 kg/m ), n (%)
482 (35.7)
403 (35.0)
79 (39.3)
Overweight BMI (25.0–29.9 kg/m2), n (%)
394 (29.1)
336 (29.1)
58 (28.9)
Obese BMI (30.0–34.9 kg/m2), n (%)
228 (16.8)
197 (17.1)
31 (15.4)
Severely obese BMI (≥ 35.0 kg/m2), N (%)
190 (14.0)
167 (14.5)
23 (11.4)
126.9 ± 59.1
127 ± 59.1
126.0 ± 59.2
91.5 ± 30.5
91.3 ± 30.6
92.8 ± 30.3
Four-quadrant infiltrate on chest radiograph, n (%)
1,061 (75.9)
902 (75.9)
159 (75.7)
Fluid balance during preceding 24 hrs (mL), mean ± sd
2,240 ± 3,727
2,231 ± 3,750
2,286 ± 3,608
Sedation or NMB, n (%)
1,215 (86.9)
1,031 (86.8)
184 (87.6)
2
2
PaO2/Fio2 ratio, mean ± sd
Acute Physiology and Chronic Health Evaluation III, mean ± sd
10 (5.0)
Cirrhosis, n (%)
42 (3.0)
34 (2.9)
7 (3.3)
Diabetes, n (%)
219 (15.7)
186 (15.7)
33 (15.7)
Chronic dialysis, n (%)
20 (1.4)
17 (1.4)
3 (1.4)
Serum albumin (g/dL), mean ± sd
2.2 ± 0.7
2.3 ± 0.7
2.1 ± 0.6
ARMA, n (%)
342 (24.5)
289 (24.3)
53 (25.2)
ALVEOLI, n (%)
419 (30.0)
252 (21.2)
67 (31.9)
FACTT, n (%)
637 (45.6)
547 (46.0)
90 (42.9)
BMI = body mass index; NMB = neuromuscular blockade; Pplat = plateau pressure; sd = standard deviation; VT = tidal volume.
Plateau Pressure Outcome Groups
According to our study definitions, 1110 patients (79.4%) had ontarget Pplat and 288 patients (20.6%) had elevated Pplat on study day
1 (Fig. 1). Of the 288 patients with elevated day 1 Pplat, 105 (36.5%)
did not have VT reduced below 5.5 mL/kg in response to their Pplat
recording. Three of these 105 patients had an arterial pH < 7.15,
which was a protocol exception allowing higher Pplat; for the other
4
www.ccmjournal.org
102 patients, there appears to have been nonadherence with the intended study protocol. Following these patients to study day 3, 86 of
the 102 patients were still alive and had Pplat and VT data recorded: 33
of 86 (38.4%) still had Pplat > 30 cm H2O, and 25 (75.8%) of these 33
patients with Pplat > 30 cm H2O still had VT ≥ 5.5 mL/kg PBW. None
of these 25 patients had protocol-approved exceptions allowing Pplat
> 30 cm H2O, suggesting nonadherence to the study protocol.
March 2013 • Volume 41 • Number 3
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Clinical Investigation
Table 2.
Plateau Pressure Statistics by Tidal Volume Cohort
Entire Study
VT < 5.5
VT 5.5–6.5
pa
1398 (100)
183 (13.1)
1215 (86.9)
—
Pplat (cm H2O), mean ± sd
24.3 ± 6.4
30.2 ± 6.5
23.5 ± 5.9
< 0.001
Pplat > 30 cm H2O, n (%)
288 (20.6)
74 (40.4)
105 (8.6)
< 0.001
Subjects, n (%)
Pplat = plateau pressure; VT = tidal volume.
a
p values based on comparison of VT < 5.5 vs. VT 5.5–6.5 based on either the two-sample t test or the Pearson’s chi-square test.
Univariable Analysis
Summary statistics by outcome group are presented in Table 3.
Compared with patients with on-target Pplat, those with elevated
Pplat were more likely to be non-white, severely obese (BMI ≥ 35 kg/
m2), have four quadrants of infiltrate on chest radiograph, to have
received sedation or NMB, and to be receiving chronic dialysis.
Patients with elevated Pplat were also younger, had a lower PaO2/
Fio2 ratio, a higher APACHE III score, a higher serum albumin,
and received more fluid in the 24 hrs preceding study enrollment
(+2093 mL for Pplat < 30 cm H2O vs. +2803 mL for elevated Pplat, p
< 0.001).
Multivariable Analysis
Age, BMI, race, PaO2/Fio2, APACHE III, and four-quadrant
chest radiograph infiltrates remained significant in multivariable
analysis (Table 4). Each of these variables had a directly or inversely
linear relationship with elevated Pplat, with the exception of BMI,
which had a J-shaped relationship. Severely obese BMI was associated with increased odds of elevated Pplat (OR = 1.85, p < 0.001),
whereas underweight BMI trended towards greater odds of elevated
plateau pressures (OR = 1.87, p = 0.080).
indicating good discrimination of the model. The HosmerLemeshow goodness-of-fit test showed good calibration with no
evidence of inadequate fit for the development or validation models (χ2df=8 = 4.57, p = 0.803; χ2df=10 = 12.97, p = 0.227). Comparing
observed vs. predicted rates of elevated Pplat by decile of predicted probability also demonstrated good calibration of the model
(Fig. 3). Using conservative estimates, there was no evidence of
collinearity in our model.
The bootstrap model had the same final variables as the crossvalidation model and similar variable coefficients, discrimination,
and calibration (Supplemental Table 1, Supplemental Digital
Content 1, http://links.lww.com/CCM/A561).
Removing subjects with barotrauma or VT < 5.5 mL/kg
PBW and refitting the remaining subjects to the logistic regression model did not significantly alter the model or the variable
coefficients (Supplemental Tables 2 and 3, Supplemental Digital
Content 1, http://links.lww.com/CCM/A561).
DISCUSSION
Using data collected from a large sample of participants in ARDS
Network trials conducted across multiple centers, we determined
that approximately one in five patients with ALI has elevated day 1
Model Performance
Pplat despite receiving lower VT ventilation. The rate of elevated Pplat
The AUC was 0.71 (95% CI: 0.67–0.74) for the developmental
likely varies by patient population, and the incidence observed in
dataset and 0.71 (95% CI: 0.62–0.79) for the validation dataset,
our study may have been inflated by
excluding patients with missing day 1
VT and/or Pplat data who had a greater
mean PaO2/Fio2 than the patients included in the study. Still, the 20% incidence seen in our study indicates that a
sizeable minority of patients may have
elevated Pplat despite receiving lower
VT ventilation. This suggests that setting a VT of 6 mL/kg PBW alone does
not ensure ALI patients receive the full
benefit of lung-protective ventilation.
An interesting finding in our
study was the incidence of nonadherence to the ventilation protocol.
A third of the patients with elevated
Pplat did not have VT reduced according to study design. In routine clinical practice, noncompliance with the
Figure 2. Histogram of plateau pressure by tidal volume (VT). This histogram shows the distribution of maxiARDS Network ventilation protocol
mum plateau pressures recorded on study day 1 for the 1215 patients with VT 5.5–6.5 mL/kg predicted body
weight (PBW) and the 183 patients with VT < 5.5 mL/kg PBW.
is likely higher than the rate observed
Critical Care Medicine
www.ccmjournal.org
5
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Prescott et al
Table 3.
Baseline Characteristics of the Patients by Outcome Group
On-target Pplat
Characteristic
Elevated Pplat
pa
1110 (79.4%)
288 (20.6%)
Pplat (cm H2O), mean ± sd
22.4 ± 4.7
32.0 ± 6.2
Unacceptable Pplat, n (%)
0 (0.0)
1398 (100)
N/Ab
Pplat > 30 cm H2O, N (%)
0 (0.0)
179 (62.2)
N/Ab
VT < 5.5 mL/kg PBW, n (%)
0 (0.0)
183 (63.5)
N/Ab
Pplat > 30 cm H2O and VT < 5.5 mL/kg PBW, n (%)
0 (0.0)
74 (25.7)
N/Ab
51.2 ± 16.7
45.1 ± 15.8
<0.001
Subjects, n (%)
Age (years), mean ± sd
<0.001
Male, n (%)
654 (58.9)
156 (54.2)
0.145
Non-white, n (%)
297 (26.8)
123 (42.7)
<0.001
27.6 ± 6.9
28.4 ± 7.8
Underweight BMI (<18.5 kg/m ), n (%)
43 (3.9)
17 (6.0)
Normal BMI (18.5–24.9 kg/m ), n (%)
389 (36.4)
93 (33.0)
Overweight BMI (25.0–29.9 kg/m2), n (%)
321 (29.9)
73 (25.9)
Obese BMI (30.0–34.9 kg/m2), n (%)
178 (16.6)
50 (17.7)
Severely obese BMI (≥ 35.0 kg/m2), n (%)
141 (13.2)
49 (17.4)
PaO2/Fio2 ratio, mean ± sd
131.1 ± 60.0
110.6 ± 52.8
<0.001
89.1 ± 29.8
100.8 ± 31.4
<0.001
246 (85.4)
<0.001
2802 ± 3398
<0.001
BMI (kg/m ), mean ± sd
2
2
2
Acute Physiology and Chronic Health Evaluation III, mean ± sd
0.178
0.122
Four-quadrant infiltrate on chest radiograph, n (%)
815 (73.4)
Fluid balance (mL) during preceding 24 hrs, mean ± sd
2093 ± 3796
Sedation or NMB, n (%)
950 (85.6)
265 (92.0)
0.004
Cirrhosis, n (%)
33 (3.0)
9 (3.1)
0.893
Diabetes, n (%)
166 (15.0)
54 (18.8)
0.115
19 (1.7)
0 (0.0)
0.025
2.2 ± 0.6
0.039
Chronic dialysis, n (%)
Albumin (g/dL), mean ± sd
2.3 ± 0.7
ARMA, n (%)
279 (25.1)
63 (21.9)
ALVEOLI, n (%)
332 (29.9)
87 (30.2)
FACTT, n (%)
499 (45.0)
138 (47.9)
0.208
BMI = body mass index; NMB = neuromuscular blockade; Pplat = plateau pressure; sd = standard deviation; VT = tidal volume.
a
p values are based on Pearson’s chi-square for categorical variables or Wilcoxon rank-sum test for continuous variables
b
Different by definition
in clinical trial participants, as supported by recent observational
study where only 41% of ventilator settings in ALI patients were
adherent to ARDS Network goals (13). In this study, each additional ventilator setting adherent to ARDS Network goals was associated with a measureable decrease in 2-year mortality (13). Thus, it
is important to identify patients at higher risk of elevated Pplat and
to consider additional methods to limit exposure to elevated Pplat.
Our regression model identified several factors that predict elevated Pplat. As would be expected, measures of the severity of lung
injury and overall illness were associated with elevated Pplat. We
also identified that race and BMI are both independently associated with plateau pressure.
6
www.ccmjournal.org
In our model, non-white race was associated with an almost
two-fold increase in the risk of elevated Pplat. Since the ARDS
Network lower VT ventilation protocol was developed, reference
equations for normal lung function have been revised to include
race (23, 24). For a given height and gender, African Americans
have 12% smaller total lung capacity than people of white
race (23, 25), and limited available data suggests that other
non-white racial groups also have smaller lung volumes (23,
26). Accounting for differences in lung volume by race when
selecting VT may limit exposure to elevated Pplat in non-white
patients. For example, VT could be set at 5 mL/kg PBW in African
Americans to account for the expected 12% reduction in TLC.
March 2013 • Volume 41 • Number 3
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Clinical Investigation
Logistic Regression to Predict
Elevated Plateau Pressure Using the
Developmental Dataset
lung capacity and functional residual capacity (27–29), likely
due to abdominal and intrathoracic fat limiting full diaphragm
and lung expansion (30). Thus, calculating VT by PBW results in
relatively larger VT (expressed as a proportion of lung capacity) as
BMI increases. However, because obesity is also associated with
Odds Ratio
(95% Confidence
decreased respiratory system compliance and elevated pleural
Variable
Interval)
p
pressure (31), elevated Pplat may not signify higher transpulmonary
pressure. Knowing that pleural pressure is elevated in obese patients
Age, 10 yr increase
0.79 (0.72–0.87)
<0.001
and patients with elevated intra-abdominal pressure for other
Body mass index
reasons, physicians may choose to accept a higher Pplat threshold
in these patients, use esophageal pressure monitoring for a better
1.87 (0.93–3.87)
0.08
Underweight
(<18.5 kg/m2)
estimation of transpulmonary pressure, or simply target a VT of
6 mL/kg until further prospective studies provide guidance for
1
—
Normal, referent
optimizing lung-protective ventilation in this patient population.
(18.5–24.9 kg/m2)
There are several important limitations to this study. As a retro1.23 (0.83–1.83)
0.31
Overweight
spective
analysis, this study was limited by available data. Race was
2
(25.0–29.9 kg/m )
reported only as white non-Hispanic, black non-Hispanic, and
1.31 (0.83–2.06)
0.242
Obese (30.0–34.9 kg/m2)
other, so we could not evaluate associations between specific races
and elevated Pplat. Furthermore, we could not assess whether the
2.11 (1.34–3.33)
0.001
Severely obese
(≥35.0 kg/m2)
differences in Pplat by race are due to differences in genetic ancestry
vs. confounding socioeconomic factors. Similar limitations have
Non-white race
1.85 (1.35–2.53)
<0.001
been encountered when evaluating the association between race
0.95 (0.92–0.97)
<0.001
PaO2/Fio2, 10 unit increase
and normal lung function (32, 33). Yet, reference equations for
Acute Physiology and
1.12 (1.06–1.18)
<0.001
normal lung function are based on self-reported race because this
Chronic Health Evaluation
information does improve prediction of lung function (32, 33).
III, 10 unit increase
BMI was calculated from weight measured after the onset of
Four-quadrant infiltrate on
1.87 (1.24–2.82)
0.003
critical illness, so fluid balance may have affected these calculations.
chest radiograph
A previous analysis of subjects from the ARDS Network lower VT
trial demonstrated that adjusting for fluid balance changed the BMI
The benefit of this approach would need to be confirmed in a
category in 12.9% of subjects although the correlation between
prospective study.
fluid-adjusted and unadjusted BMI was strong (r = 0.978 (34)).
Patients with a severely obese BMI (≥35.0 kg/m2) had a more
We chose a threshold of 30 cm H2O to define elevated Pplat bethan two-fold increase in risk of elevated Pplat. Obesity, and in
cause it prompted VT reduction according to the ARDS Network
particular severe obesity, has been associated with decreased total protocol (7) and still remains the point at which VT reduction is
recommended (7, 35). However, this
is merely a surrogate marker for lung
strain. Regional hyperinflation and
lung injury are still possible with Pplat
< 30 cm H2O (36, 37). Other measurements of lung injury, such as biochemical markers, histopathologic findings, and transpulmonary pressure,
determined with esophageal pressure
monitoring may be more accurate but
are less widely used in clinical practice
and were not available in our datasets.
Finally, we developed a predictive
rather than an explanatory model of
factors associated with elevated Pplat.
Because of this approach, there may
be confounding factors associated
with Pplat that are not included in the
model. The beneficial or harmful efFigure 3. Predicted vs. observed rates of elevated plateau pressure (Pplat) for the cross-validation logistic
fects of hypercarbia as a result of
regression model. Patients were stratified according to their probability of elevated Pplat predicted by the
decreased VT, which could be more
regression model and then divided into 10 groups. The predicted rate of elevated Pplat was compared with the
common in certain patient groups
observed rate of elevated Pplat for each group and the overall population.
Table 4.
Critical Care Medicine
www.ccmjournal.org
7
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Prescott et al
based on physiologic parameters, was beyond the scope of our
analysis.
CONCLUSION
A ventilation strategy incorporating lower VT with further VT
reductions based on Pplat measurements decreases mortality in
patients with ALI, yet one in five ALI patients has elevated day
1 Pplat while receiving a VT of 6 mL/kg PBW. Race, BMI, and
severity of lung injury are independently associated with Pplat
in patients receiving lower VT ventilation. Selecting a smaller
initial VT for non-white patients and patients with higher
severity of lung injury may decrease the incidence of elevated
Pplat. However, we do not suggest this approach in severely obese
patients because higher P plats may be necessary to maintain
appropriate transpulmonary pressure in this population.
Prospective studies are needed to confirm the benefit of
lowering initial VTs in non-white patients and patients with
higher severity of lung injury.
REFERENCES
1.Rubenfeld GD, Caldwell E, Peabody E, et al: Incidence and outcomes of
acute lung injury. N Engl J Med 2005; 353:1685–1693
2. Zambon M, Vincent JL: Mortality rates for patients with acute lung injury/
ARDS have decreased over time. Chest 2008; 133:1120–1127
3.Phua J, Badia JR, Adhikari NK, et al: Has mortality from acute respiratory
distress syndrome decreased over time?: A systematic review. Am J Respir
Crit Care Med 2009; 179:220–227
4.Chiumello D, Carlesso E, Cadringher P, et al: Lung stress and strain during
mechanical ventilation for acute respiratory distress syndrome. Am J Respir
Crit Care Med 2008; 178:346–355
5.Checkley W, Brower R, Korpak A, et al; Acute Respiratory Distress Syndrome Network Investigators: Effects of a clinical trial on mechanical ventilation practices in patients with acute lung injury. Am J Respir Crit Care Med
2008; 177:1215–1222
6.Brower RG, Matthay M, Schoenfeld D: Meta-analysis of acute lung injury
and acute respiratory distress syndrome trials. Am J Respir Crit Care Med
2002; 166:1515–1517
7. The Acute Respiratory Distress Syndrome Network: Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung
injury and the acute respiratory distress syndrome. N Engl J Med 2000;
342:1301–1308
8. Esteban A, Ferguson ND, Meade MO, et al; VENTILA Group: Evolution of
mechanical ventilation in response to clinical research. Am J Respir Crit
Care Med 2008; 177:170–177
9.Umoh NJ, Fan E, Mendez-Tellez PA, et al: Patient and intensive care unit
organizational factors associated with low tidal volume ventilation in acute
lung injury. Crit Care Med 2008; 36:1463–1468
10.Brun-Buisson C, Doyon F, Sollet JP, et al: Prevention of intravascular
catheter-related infection with newer chlorhexidine-silver sulfadiazinecoated catheters: a randomized controlled trial. Intensive Care Med
2004; 30:837–843
11.Akhtar SR, Weaver J, Pierson DJ, et al: Practice variation in respiratory therapy documentation during mechanical ventilation. Chest 2003;
124:2275–2282
12.Cooke CR, Watkins TR, Kahn JM, et al: The effect of an intensive care unit
staffing model on tidal volume in patients with acute lung injury. Crit Care
2008; 12:R134
13.Needham DM, Colantuoni E, Mendez-Tellez PA, et al: Lung protective mechanical ventilation and two year survival in patients with acute lung injury:
prospective cohort study. BMJ 2012; 344:e2124
14. Gajic O, Dara SI, Mendez JL, et al: Ventilator-associated lung injury in
patients without acute lung injury at the onset of mechanical ventilation. Crit
Care Med 2004; 32:1817–1824
8
www.ccmjournal.org
15. Brower RG, Lanken PN, MacIntyre N, et al; National Heart, Lung, and Blood
Institute ARDS Clinical Trials Network: Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome.
N Engl J Med 2004; 351:327–336
16.Wheeler AP, Bernard GR, Thompson BT, et al; National Heart, Lung, and
Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials
Network: Pulmonary-artery versus central venous catheter to guide treatment of acute lung injury. N Engl J Med 2006; 354:2213–2224
17.Wiedemann HP, Wheeler AP, Bernard GR, et al; National Heart, Lung, and
Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network: Comparison of two fluid-management strategies in acute lung
injury. N Engl J Med 2006; 354:2564–2575
18.Hosmer DW, Lemeshow S, Wiley online library: Applied Logistic Regression. Second Edition. New York, Wiley, 2000, pp 373
19.Royston P, Altman DG: Regression using fractional polynomials of continuous covariates: parsimonious parametric modelling. J Roy Stat Soc Ser C
(Applied Statistics) 1994; 43:429–467
20.Hanley JA, McNeil BJ: The meaning and use of the area under a receiver
operating characteristic (ROC) curve. Radiology 1982; 143:29–36
21. Lemeshow S, Hosmer DW Jr: A review of goodness of fit statistics for use in the
development of logistic regression models. Am J Epidemiol 1982; 115:92–106
22. Zhu BP, Lemeshow S, Hosmer DW, et al: Factors affecting the performance
of the models in the Mortality Probability Model II system and strategies of
customization: a simulation study. Crit Care Med 1996; 24:57–63
23.Stocks J, Quanjer PH: Reference values for residual volume, functional
residual capacity and total lung capacity. ATS Workshop on Lung Volume
Measurements. Official Statement of The European Respiratory Society. Eur
Respir J 1995; 8:492–506
24. Hankinson JL, Odencrantz JR, Fedan KB: Spirometric reference values from
a sample of the general U.S. population. Am J Respir Crit Care Med 1999;
159:179–187
25.Lapp NL, Amandus HE, Hall R, et al: Lung volumes and flow rates in black
and white subjects. Thorax 1974; 29:185–188
26.Quanjer PH, Tammeling GJ, Cotes JE, et al: Lung volumes and forced ventilatory flows. Report Working Party Standardization of Lung Function Tests,
European Community for Steel and Coal. Official Statement of the European Respiratory Society. Eur Respir J Suppl 1993; 16:5–40
27.Jones RL, Nzekwu MM: The effects of body mass index on lung volumes.
Chest 2006; 130:827–833
28.Ray CS, Sue DY, Bray G, et al: Effects of obesity on respiratory function.
Am Rev Respir Dis 1983; 128:501–506
29. Porhomayon J, Papadakos P, Singh A, et al: Alteration in respiratory physiology in obesity for anesthesia-critical care physician. HSR Proceedings in
Intenstive Care and Cardiovascular Anesteria 2011; 3:109–118
30.Watson RA, Pride NB, Thomas EL, et al: Reduction of total lung capacity
in obese men: Comparison of total intrathoracic and gas volumes. J Appl
Physiol 2010; 108:1605–1612
31.Behazin N, Jones SB, Cohen RI, et al: Respiratory restriction and elevated
pleural and esophageal pressures in morbid obesity. J Appl Physiol 2010;
108:212–218
32.Kumar R, Seibold MA, Aldrich MC, et al: Genetic ancestry in lung-function
predictions. N Engl J Med 2010; 363:321–330
33.Scanlon PD, Shriver MD: “Race correction” in pulmonary-function testing.
N Engl J Med 2010; 363:385–386
34.O’Brien JM Jr, Welsh CH, Fish RH, et al; National Heart, Lung, and
Blood Institute Acute Respiratory Distress Syndrome Network: Excess
body weight is not independently associated with outcome in mechanically ventilated patients with acute lung injury. Ann Intern Med 2004;
140:338–345
35. Hager DN, Krishnan JA, Hayden DL, et al: Tidal volume reduction in patients
with acute lung injury when plateau pressures are not high. Am J Respir Crit
Care Med 2005; 172:1241–1245
36. Mattingley JS, Holets SR, Oeckler RA, et al: Sizing the lung of mechanically
ventilated patients. Crit Care 2011; 15:R60
37.Terragni PP, Rosboch G, Tealdi A, et al: Tidal hyperinflation during low tidal
volume ventilation in acute respiratory distress syndrome. Am J Respir Crit
Care Med 2007;175:160–166
March 2013 • Volume 41 • Number 3
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
original article
High-Frequency Oscillation for Acute
Respiratory Distress Syndrome
Duncan Young, D.M., Sarah E. Lamb, D.Phil., Sanjoy Shah, M.D.,
Iain MacKenzie, M.D., William Tunnicliffe, M.Sc., Ranjit Lall, Ph.D.,
Kathy Rowan, D.Phil., and Brian H. Cuthbertson, M.D.,
for the OSCAR Study Group*
A bs t r ac t
Background
From John Radcliffe Hospital (D.Y.) and
the University of Oxford (D.Y., S.E.L.), Oxford; the Bristol Royal Infirmary, Bristol
(S.S.); Queen Elizabeth Hospital, Birmingham (I.M., W.T.); Warwick Clinical Trials
Unit, University of Warwick, Warwick
(R.L.); and Intensive Care National Audit
and Research Centre, London (K.R.) — all
in the United Kingdom; and the Department of Critical Care Medicine, Sunnybrook Health Sciences Centre and the
Department of Anesthesia, University of
Toronto, Toronto (B.H.C.). Address reprint requests to Dr. Young at the Adult
Intensive Care Unit, John Radcliffe Hospital, Headley Way, Oxford OX3 9DU,
United Kingdom, or at duncan.young@
nda.ox.ac.uk.
*Investigators in the Oscillation in ARDS
(OSCAR) study group are listed in the
Supplementary Appendix, available at
NEJM.org.
This article was published on January 22,
2013, and updated on January 31, 2013, at
NEJM.org.
N Engl J Med 2013;368:806-13.
DOI: 10.1056/NEJMoa1215716
Copyright © 2013 Massachusetts Medical Society.
806
Patients with the acute respiratory distress syndrome (ARDS) require mechanical
ventilation to maintain arterial oxygenation, but this treatment may produce secondary lung injury. High-frequency oscillatory ventilation (HFOV) may reduce this
secondary damage.
Methods
In a multicenter study, we randomly assigned adults requiring mechanical ventilation for ARDS to undergo either HFOV with a Novalung R100 ventilator (Metran) or
usual ventilatory care. All the patients had a ratio of the partial pressure of arterial
oxygen (Pao2) to the fraction of inspired oxygen (Fio2) of 200 mm Hg (26.7 kPa) or
less and an expected duration of ventilation of at least 2 days. The primary outcome
was all-cause mortality 30 days after randomization.
Results
There was no significant between-group difference in the primary outcome, which
occurred in 166 of 398 patients (41.7%) in the HFOV group and 163 of 397 patients
(41.1%) in the conventional-ventilation group (P = 0.85 by the chi-square test). After
adjustment for study center, sex, score on the Acute Physiology and Chronic Health
Evaluation (APACHE) II, and the initial Pao2:Fio2 ratio, the odds ratio for survival
in the conventional-ventilation group was 1.03 (95% confidence interval, 0.75 to
1.40; P = 0.87 by logistic regression).
Conclusions
The use of HFOV had no significant effect on 30-day mortality in patients undergoing mechanical ventilation for ARDS. (Funded by the National Institute for Health
Research Health Technology Assessment Programme; OSCAR Current Controlled
Trials number, ISRCTN10416500.)
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation for Respir atory Distress
T
he acute respiratory distress syndrome (ARDS) is a severe, diffuse inflammatory lung condition caused by a range
of acute illnesses. Mortality in affected patients
is high,1 and survivors may have functional limitations for years.2,3 Although mechanical ventilation can initially be lifesaving in patients with
ARDS, it can also further injure the patients’
lungs and contribute to death.4
High-frequency oscillatory ventilation (HFOV)
was first used experimentally in the 1970s to
minimize the hemodynamic effects of mechanical
ventilation.5 Patients’ lungs are held inflated to
maintain oxygenation, and carbon dioxide is
cleared by small volumes of gas moved in and
out of the respiratory system at 3 to 15 Hz. This
action is thought to minimize the repeated process of opening and collapsing of lung units that
causes the secondary lung damage during mechanical ventilation. On the basis of small trials
with outdated controls6 and the commercial availability of HFOV equipment, many clinicians use
HFOV for patients who have hypoxemia despite
the use of standard approaches for improving
arterial oxygenation. The increasing use of HFOV
in the absence of good evidence of effectiveness
led the National Institute for Health Research in
the United Kingdom to commission a study to
determine the effectiveness of HFOV as a treatment for ARDS.
manuscript preparation. The study was approved
by national ethics review committees and research governance departments at each center.
Patients or their representatives provided written
informed consent.
Patients
Patients who were undergoing mechanical ventilation were eligible for the study if they had a
ratio of the partial pressure of arterial oxygen
(Pao2) to the fraction of inspired oxygen (Fio2) of
200 mm Hg (26.7 kPa) or less while receiving a
positive end-expiratory pressure (PEEP) of 5 cm
of water or greater, if bilateral pulmonary infiltrates were visible on chest radiography without
evidence of left atrial hypertension, and if they
were expected to require at least 2 more days of
mechanical ventilation.
Patients were excluded if they had undergone
mechanical ventilation for 7 or more days, if
they were under the age of 16 years, if they
weighed less than 35 kg, if they were participating in other interventional studies, if they had
lung disease characterized by airway narrowing
or air trapping, or if they had undergone recent
lung surgery.
An independent telephone randomization system assigned patients to either HFOV or conventional mechanical ventilation in a 1:1 ratio. Randomization was by permuted block stratified
according to study center, Pao2:Fio2 ratio
(≤113 mm Hg [15 kPa] or >113 mm Hg), age
Me thods
(≤55 years or >55 years), and sex. Each center
Study Design
had one HFOV ventilator, so recruitment could
We conducted a randomized, controlled trial of not take place if the device was in use for anHFOV, as compared with conventional mechani- other study patient.
cal ventilation. Patients were recruited from adult
general intensive care units (ICUs) in 12 university Study Treatments
hospitals, 4 university-affiliated hospitals, and Patients in the HFOV group were treated with the
13 district general hospitals in England, Wales, use of a Novalung R100 ventilator (Metran)7 until
and Scotland. Three hospitals had previous expe- the start of weaning. The initial settings were a
rience with HFOV with the use of SensorMedics ventilation frequency of 10 Hz, a mean airway
3100B ventilators (CareFusion), and the remain- pressure of 5 cm of water above the plateau airder had limited experience (in 6 hospitals) or no way pressure at enrollment, bias flow rate of
experience (in 20 hospitals) with HFOV. Details 20 liters per minute, a cycle volume of 100 ml
regarding HFOV training are provided in the Sup- (the volume of gas used to move the oscillating
plementary Appendix, available with the full text diaphragm; the tidal volume delivered to the alof this article at NEJM.org. The full protocol is veoli is a fraction of this volume), and an inspired
also available at NEJM.org.
oxygen fraction of 1. This ventilator has a fixed
The ventilators were purchased from Inspira- 1:1 inspiratory:expiratory time ratio.
tion Healthcare. The company had no role in the
Two algorithms were used to determine changstudy design, data acquisition, data analysis, or es in HFOV settings (for details, see the Supplen engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
807
The
n e w e ng l a n d j o u r na l
mentary Appendix). The partial pressure of arterial carbon dioxide (Paco2) was controlled to
maintain an arterial pH above 7.25 by increasing
the cycle volume to the maximum at each frequency. If this was insufficient, the frequency
was reduced by 1 Hz. If the minimum frequency
(5 Hz) was reached, the on-call study clinician
would suggest other measures to control the
Paco2 level (see the Supplementary Appendix).
The Pao2 level was maintained between
60 mm Hg and 75 mm Hg (8 kPa to 10 kPa).
Hypoxemia was treated by increasing the mean
airway pressure and then by increasing the Fio2
level. If a patient reached a mean airway pressure
of 24 cm of water, at an Fio2 level of 0.4 or less,
with a Pao2 level of 60 mm Hg or greater, for
12 hours or more, he or she was switched to
pressure-controlled ventilation for weaning from
mechanical ventilation, since there was no facility
to accommodate patients’ spontaneous respiratory
efforts during HFOV. Patients could be restarted
on HFOV up to 2 days after the start of weaning.
Patients in the conventional-ventilation group
were treated according to local practice in the
participating ICUs. The participating units were
encouraged to use pressure-controlled ventilation
at 6 to 8 ml per kilogram of ideal body weight
and to use the combinations of PEEP and Fio2
values that were used in the Acute Respiratory
Distress Syndrome Network study.4 All other
treatment was determined by the patients’ physicians on the basis of assessment of clinical need.
Data Collection
At the time of enrollment, we recorded data with
respect to the patients’ demographic characteristics, ventilation before enrollment, physiology and
other data required to calculate the score on the
Acute Physiology and Chronic Health Evaluation
(APACHE) II, coexisting medical conditions, the
use of sedatives and muscle relaxants, and ventilator settings. For each day that a patient was
treated in the ICU, we recorded data with respect
to the use of antibiotics, sedatives, and muscle
relaxants during the previous day or since enrollment on the first day. Data regarding support for
respiratory and cardiovascular organ systems
were recorded daily during treatment in the ICU
with the use of the United Kingdom’s criticalcare minimum data set.8 Vital status at 30 days
was known for all patients, but causes of death
were not recorded.
808
of
m e dic i n e
Outcomes
The primary outcome, vital status at 30 days,
was obtained from hospital records and verified
with the use of a national database. Secondary
outcomes were all-cause mortality at the time of
discharge from the ICU and the hospital, the
duration of mechanical ventilation, and the use
of antimicrobial, sedative, vasoactive, and neuromuscular-blocking drugs. We recorded the
duration of treatment in both the ICU and the
hospital.
Statistical Analysis
Recruitment-rate estimates and sample-size calculations were performed after a systematic review of the incidence and outcome of ARDS, national audits in the United Kingdom, and two
randomized, controlled trials of HFOV.9,10 We
determined that the enrollment of 503 patients
per study group would provide a power of 80% to
identify a change of 9 percentage points in an
estimated rate of death of 45% in the control
group at a P value of 0.05. At a planned interim
review, the sample size was revised to 401 patients per group on the basis of accumulated
mortality data in the control group and an effect
size of 10 percentage points (80% power at
P = 0.05).
All analyses were conducted on an intentionto-treat basis. Three planned interim analyses
were conducted by an independent data and safety
monitoring committee after the recruitment of
100, 340, and 640 patients. Formal stopping
rules were not specified. Instead, the committee
assessed whether the randomized comparisons
provided “proof beyond reasonable doubt” that
for all or some patients the treatment was
clearly indicated or clearly contraindicated and
provided evidence that might reasonably be expected to influence future patient treatment.
We used chi-square tests to compare betweengroup rates of death at 30 days and among patients in ICU and hospital settings. We performed an analysis of mortality after adjustment
for study center, sex, Pao2:Fio2 ratio, and
APACHE II score using logistic regression. Continuous variables were compared with the use of
Student’s t-tests. Since both the rate and timing
of death were similar in the two study groups,
data for survivors and those for nonsurvivors
were not analyzed separately. All P values are
two-sided.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation for Respir atory Distress
R e sult s
Trial Progression and Recruitment
We trained 2306 intensive care nurses, medical
staff, physiotherapists, and technicians in 198
face-to-face training sessions. Patients were recruited from December 7, 2007, until the end of
July 2012. Of the 2769 patients who were
screened, 795 (28.7%) underwent randomization
(Fig. 1). The study had 968 ICU-months of recruitment averaging 0.82 patients per ICUmonth. (A graphical summary of recruitment is
provided in Figure S3 in the Supplementary Appendix.) The baseline characteristics of the patients at randomization were similar in the two
study groups (Table 1).
Ventilation
HFOV was used for a median of 3 days (interquartile range, 2 to 5) in 388 patients. The longest
initial period of receipt of HFOV was 24 days.
Figure 2 shows the use of HFOV in the two study
groups. Ten patients in the conventional-ventilation group underwent HFOV at some point during the study period, and 10 patients who were
assigned to the HFOV group never received this
treatment. Table 2 shows ventilatory and other
variables for the first 3 days of the study period.
Neuromuscular-blocking drugs were used for
a mean (±SD) of 2.0±3.4 days in the conventionalventilation group and for 2.5±3.5 days in the
HFOV group (P = 0.02). Sedative drugs were used
for 8.5±6.9 days in the conventional-ventilation
group and for 9.4±7.2 days in the HFOV group
(P = 0.07).
The patients had 17.6±8.8 ventilator-free days in
the conventional-ventilation group and 17.1±8.6
ventilator-free days in the HFOV group (P = 0.42).
Mechanical ventilation (including HFOV but excluding noninvasive ventilation) was used for
14.1±13.4 days in the conventional-ventilation
group and 14.9±13.3 days in the HFOV group
(P = 0.41).
the odds ratio for survival in the conventionalventilation group, as compared with the HFOV
group, was 1.03 (95% CI, 0.75 to 1.40; P = 0.87 by
logistic regression) (Fig. 3). Similar proportions
of patients died at each time point in each group.
The rates of death at first discharge from the
ICU were 42.1% in the conventional-ventilation
group and 44.1% in the HFOV group, for an absolute difference of 2.0 percentage points (P = 0.57).
At first hospital discharge, 48.4% of patients in
the conventional-ventilation group and 50.1% of
those in the HFOV group had died, for an absolute difference of 1.7 percentage points (P = 0.62).
Data are not provided with respect to the duration of care for survivors and nonsurvivors, since
the proportions of patients who died in each
study group over time were nearly identical. The
total duration of ICU stay was 16.1±15.2 days in
the conventional-ventilation group and 17.6±16.6
2769 Patients were screened
1974 Were excluded
913 (33.0%) Did not meet
inclusion criteria or met
exclusion criteria
134 (4.8%) Did not have
access to ventilator
282 (10.2%) Did not provide
consent
645 (23.3%) Had other
reasons
795 (28.7%) Underwent randomization
397 Were assigned to receive
conventional intervention
397 (100%) Received assigned
intervention
10 (2.5%) Received HFOV
at some point after
randomization
398 Were assigned to receive HFOV
388 (97.5%) Received assigned
intervention
10 Did not receive assigned
intervention
3 Died
4 Had ventilator malfunction
1 Recovered
2 Were treated by clinician
who did not comply with
assigned therapy
397 (100%) Were included in
primary analysis
398 (100%) Were included in
primary analysis
Outcomes
The primary outcome occurred in 166 of 398 patients (41.7%) in the HFOV group and in 163 of
397 patients (41.1%) in the conventional-ventilation group (P = 0.85), for an absolute difference
of 0.6 percentage points (95% confidence interval [CI], −6.1 to 7.5). After adjustment for study
center, sex, APACHE II score, and Pao2:Fio2 ratio,
Figure 1. Enrollment and Outcomes.
HFOV denotes high-frequency oscillatory ventilation.
n engl j med 368;9 nejm.org february 28, 2013
4 col
22p3
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
809
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
Table 1. Baseline Characteristics of the Patients.*
Conventional
Ventilation
(N = 397)
HFOV
(N = 398)
All Patients
(N = 795)
55.9±16.2
54.9±18.8
55.4±16.2
Male sex — no. (%)
239 (60.2)
256 (64.3)
495 (62.3)
APACHE II score†
21.7±6.1
21.8±6.0
21.8±6.1
Probability of in-hospital death (as calculated from APACHE II score)
0.43±0.19
0.44±0.19
0.43±0.19
PaO2:FiO2 ratio — mm Hg
113±38
113±37
113±38
Exhaled tidal volume — ml
505±173
541±271
523±228
Exhaled tidal volume — ml/kg of ideal body weight‡
8.3±3.5
8.7±3.5
8.5±3.9
10.17±3.46
10.41±3.25
10.29±3.35
11.3±3.3
11.4±3.5
11.4±3.4
2.1±2.1
2.2±2.3
2.2±2.2
304 (76.6)
302 (75.9)
606 (76.2)
Characteristic
Age — yr
Exhaled minute ventilation — liters/min
Positive end-expiratory pressure — cm of water
Duration of mechanical ventilation before randomization — days
Pulmonary cause of ARDS — no. (%)
*Plus–minus values are means ±SD. There was no significant difference between groups except for exhaled tidal volume
(P = 0.04). ARDS denotes acute respiratory distress syndrome, FiO2 fraction of inspired oxygen, and PaO2 partial pressure
of arterial oxygen.
†Scores on the Acute Physiology and Chronic Health Evaluation (APACHE) II scale range from 0 to 71, with higher scores
indicating more severe disease.11
‡Ideal body weight was calculated as 2.3 kg for each inch of height above 60 in. added to 50 kg for men or 45.5 kg for women.
days in the HFOV group (P = 0.18); the total durations of hospital stay were 33.1±44.3 days and
33.9±41.6 days, respectively (P = 0.79). As of October 1, 2012, the date that the database was
closed, 7 patients remained in acute hospital care.
Patients received antimicrobial drugs for
12.4±10.3 days in the conventional-ventilation
group and for 12.8±12.0 days in the HFOV group
(P = 0.56); 67.5% and 64.4% of these drugs, respectively, were administered to treat pulmonary
infections.
There was no significant difference in the
number of days on which patients received inotropic agents or pressor infusions, with 2.8±5.6
days in the conventional-ventilation group and
2.9±4.5 days in the HFOV group (P = 0.74).
Discussion
This study, which was designed to help practitioners choose between options for care, met 7 of
the 10 criteria of the Pragmatic–Explanatory
Continuum Indicator Summary (PRECIS).12 The
results were not totally pragmatic because of the
tight protocol-specified restrictions on the use of
HFOV, protocol-compliance monitoring, and additional follow-up. We found no significant be810
tween-group difference in the primary outcome
of mortality up to 30 days after randomization.
Our estimate of the 95% confidence interval for
the treatment excludes the treatment effect we
specified in both the initial and revised samplesize estimates. Since data collection is ongoing, we
cannot yet report the longer-term outcomes (including survival and health-related quality of life).
We recruited patients with moderate-tosevere ARDS, with an average Pao2:Fio2 ratio of
113 mm Hg (15.1 kPa). The study-entry criterion
was a Pao2:Fio2 ratio of less than 200 mm Hg
(26.7 kPa), which was in line with the agreed
definition of ARDS,13 but the additional requirement of a further 48 hours or more of mechanical ventilation may have excluded milder cases
of ARDS. The average Pao2:Fio2 ratio is nearly
identical to the mean of 112 mm Hg reported in
the recent systematic review of HFOV6 and is
similar to the mean values reported in studies of
other treatments for ARDS.14-16 The patients had
a high severity of illness, as evidenced by the
APACHE II scores, which also were nearly identical to those reported in the two other multicenter
studies of HFOV in adults.9,10 Thus, we appear to
have recruited patients who were similar to those
in previous randomized, controlled trials of HFOV.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation for Respir atory Distress
100
Patients Undergoing HFOV (%)
HFOV improved oxygenation as expected. The
Paco2 value increased as a predicted result of the
HFOV treatment algorithms, resulting in a modest
respiratory acidosis. A similar effect was seen in
the larger of the two reported studies of HFOV
in adults10 but not in the smaller study9 or the
meta-analysis.6 The conventional-ventilation group
was treated with tidal volumes at the upper end
of the accepted range of 6 to 8 ml per kilogram
of ideal body weight.
The use of HFOV was initially associated with
an increased use of neuromuscular-blocking
drugs, probably because the R100 ventilator has
no facility to allow the patient to breathe spontaneously. HFOV has been reported to cause a
reduction in cardiac output,17 but as indicated by
the use of vasoactive and inotropic drugs, that
did not occur in this study.
Our results are at variance with the latest metaanalysis of HFOV,6 which showed a reduced risk
of death (risk ratio, 0.77; 95% CI, 0.61 to 0.98),
as compared with conventional ventilation. This
may be simply that our study recruited more than
90
80
70
HFOV
60
50
40
30
20
Conventional
ventilation
10
0
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
Day
Figure 2. Proportions of Patients Undergoing High-Frequency Oscillatory
Ventilation (HFOV) during the First 30 Days, According to Study Group.
Shown are the percentages of patients in each study group who underwent
HFOV. Ten patients in the conventional-ventilation group underwent HFOV
at some point during their treatment, and 10 patients who were assigned to
the HFOV group never received this treatment. Day 0 is the day of randomization, and subsequent determinations were made at 8 a.m. each day.
Table 2. Ventilatory Variables during the First 3 Study Days.*
Variable
Day 1
Day 2
Day 3
HFOV
Conventional
Ventilation
HFOV
Conventional
Ventilation
HFOV
Conventional
Ventilation
370
392
326
374
240
348
Mean airway pressure (HFOV) or plateau pressure
(conventional ventilation) — cm of water
26.9±6.2
30.9±11.0
25.3±5.5
29.5±10.7
25.1±5.4
28.5±11.2
Total respiratory frequency — Hz (HFOV) or
breaths/min (conventional ventilation)
7.8±1.8
21.7±8.4
7.5±1.8
22.7±9.0
7.2±1.8
23.3±8.2
Cycle volume (HFOV) or tidal volume (conventional
ventilation) — ml (HFOV) or ml/kg of ideal
body weight (conventional ventilation)
213±72
8.3±2.9
228±75
8.2±2.5
240±75
8.3±3.0
NA
11.4±3.6
NA
11.0±3.6
NA
10.5±3.7
PaO2:FiO2 ratio — mm Hg
192±77
154±61
212±69
163±66
217±69
166±63
PaCO2 — mm Hg
55±17
50±19
56±16
49±13
56±17
48±13
7.30±0.10
7.35±0.10
7.32±0.09
7.37±0.10
7.34±0.10
7.39±0.09
Neuromuscular-blocking agent
209 (52.5)
165 (41.6)
147 (36.9)
115 (29.0)
110 (27.6)
77 (19.4)
Vasoactive or inotropic agent
173 (43.5)
177 (44.6)
158 (40.0)
146 (36.8)
126 (31.7)
112 (28.2)
Sedative agent
390 (98.0)
388 (97.7)
371 (93.2)
363 (91.4)
341 (85.7)
335 (84.4)
No. of patients
Positive end-expiratory pressure — cm of water
(conventional ventilation only)
Arterial pH
Medication use — no. (%)†
*Measurements were taken at 8 a.m. Day 1 values were recorded the morning after recruitment. The values for high-frequency oscillatory
ventilation (HFOV) are only for patients who actually underwent the treatment. The values for conventional ventilation are for all patients
assigned to receive conventional ventilation who were receiving any mechanical ventilation. NA denotes not applicable, and PaCO2 partial
pressure of arterial carbon dioxide.
†Percentages were calculated on the basis of the 398 patients in the HFOV group and the 397 patients in the conventional-ventilation group
who underwent randomization.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
811
The
n e w e ng l a n d j o u r na l
Probability of Survival
1.00
Conventional
ventilation
0.75
HFOV
0.50
0.25
0.00
0
5
10
15
20
25
30
Days
No. at Risk
Conventional
ventilation
HFOV
397
351
312
281
259
243
236
398
349
311
280
253
241
233
Figure 3. Kaplan–Meier Survival Estimates during the First 30 Study Days.
twice the number of patients who were included
in the meta-analysis. Adding our results to the
meta-analysis changes the estimated risk ratio
from the pooled studies to 0.90 (95% CI, 0.76 to
1.07), indicating no significant benefit for HFOV.
The use of HFOV is a lung-protection strategy,
which may be ineffective if it is used for too brief
a period. We used it up to the point at which the
HFOV design hindered weaning. In the two other
multicenter studies of HFOV in adults,9,10 the
duration of ventilation was not reported.
In the HFOV group in our study, we used the
Novalung R100 ventilator, a device that had not
been used before in clinical trials. To date, all
studies have used the SensorMedics 3100B ventilator, a device that has an electromechanically
driven diaphragm, which normally oscillates with
an inspiratory:expiratory time ratio of 1:2. The
R100 ventilator uses a pneumatically driven diaphragm with a fixed 1:1 ratio. It seems unlikely
that these differences would explain the difference in mortality between our study and the
pooled results of studies to date, but it remains
a possibility.
We recruited patients who met the definition
of ARDS13 that was in place at the time the study
was planned, and the entry criteria match the
812
of
m e dic i n e
“moderate” and “severe” categories in the recently revised definition.18 The study has good
internal and external validity. Bias was minimized by using centers with equipoise, by concealing treatment assignments before randomization, by concealing interim analyses from all
study investigators except for the data and safety
monitoring committee, and by using an analysis
plan that was agreed on before study closure.
There was no loss to follow-up, crossovers were
minimal, and the study recruited 99.1% of the
planned sample size. External validity was maintained by using a large number of different-sized
ICUs spread across the United Kingdom. Most of
the centers in this trial were inexperienced with
the intervention at the start, but this was unavoidable, since few centers in the United Kingdom
have experience with the use of HFOV. We invested heavily in training at each study center.
The consent-refusal rate was low.
Our report coincides with the publication in
the Journal of the results of a large multicenter
efficacy study of HFOV, the Oscillation for Acute
Respiratory Distress Syndrome Treated Early
(OSCILLATE) trial.19 This study showed 47% mortality in the HFOV group and 35% in the control
group. The patients who were recruited in both
studies were broadly similar. The OSCILLATE
trial used the 3100B ventilator, maneuvers to reexpand collapsed areas of lung before HFOV, and
a protocol-specified high-PEEP strategy for conventional ventilation. In that study, the patients
undergoing HFOV required more inotropic and
pressor support than did those in the control
group. It is possible that the HFOV strategy used
in the OSCILLATE trial was injurious, but the low
mortality in their control group also raises the possibility that the control treatment was a very effective ventilation strategy in patients with ARDS.
In conclusion, in a large effectiveness study, we
were unable to find any benefit or harm from the
use of HFOV in adult patients with ARDS. We
recommend that this mode of ventilation not be
used for routine care.
Supported by the National Institute for Health Research
Health Technology Assessment Programme (project number
06/04/01).
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation for Respir atory Distress
References
1. Erickson SE, Martin GS, Davis JL,
Matthay MA, Eisner MD. Recent trends in
acute lung injury mortality: 1996-2005.
Crit Care Med 2009;37:1574-9.
2. Herridge MS, Tansey CM, Matté A, et
al. Functional disability 5 years after acute
respiratory distress syndrome. N Engl J
Med 2011;364:1293-304.
3. Herridge MS, Cheung AM, Tansey CM,
et al. One-year outcomes in survivors of
the acute respiratory distress syndrome.
N Engl J Med 2003;348:683-93.
4. The Acute Respiratory Distress Syndrome Network. Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute lung injury and
the acute respiratory distress syndrome.
N Engl J Med 2000;342:1301-8.
5. Lunkenheimer PP, Rafflenbeul W,
Keller H, Frank I, Dickhut HH, Fuhrmann
C. Application of transtracheal pressure
oscillations as a modification of “diffusing respiration.” Br J Anaesth 1972;44:
627.
6. Sud S, Sud M, Friedrich JO, et al. High
frequency oscillation in patients with
acute lung injury and acute respiratory
distress syndrome (ARDS): systematic review and meta-analysis. BMJ 2010;340:
c2327.
7. Metran. Product information: R100.
2012 (http://www.metran.co.jp/en/product/
r100/index.html).
8. Critical care minimum dataset, version
8.0, revised March 2010 (http://www.isb
.nhs.uk/documents/isb-0153/amd-812010/0153812010spec.pdf).
9. Bollen CW, van Well GT, Sherry T, et
al. High frequency oscillation ventilation
compared with conventional mechanical
ventilation in adult respiratory distress
syndrome: a randomized controlled trial
[ISRCTN24242669]. Crit Care 2005;9:
R430-R439.
10. Derdak S, Mehta S, Stewart TE, et al.
High-frequency oscillatory ventilation for
acute respiratory distress syndrome in
adults: a randomized, controlled trial.
Am J Respir Crit Care Med 2002;166:
801-8.
11. Knaus WA, Draper EA, Wagner DP,
Zimmerman JE. APACHE II: a severity of
disease classification system. Crit Care
Med 1985;13:818-29.
12. Thorpe KE, Zwarenstein M, Oxman
AD, et al. A Pragmatic-Explanatory Continuum Indicator Summary (PRECIS): a
tool to help trial designers. CMAJ 2009;
180:E47-E57.
13. Bernard GR, Artigas A, Brigham KL,
et al. The American-European Consensus
Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical
trial coordination. Am J Respir Crit Care
Med 1994;149:818-24.
14. Villar J, Pérez-Méndez L, Basaldúa S,
et al. A risk tertiles model for predicting
mortality in patients with acute respiratory distress syndrome: age, plateau pressure, and PaO2/FIO2 at ARDS onset can
predict mortality. Respir Care 2011;56:
420-8.
15. Villar J, Blanco J, Añón JM, et al. The
ALIEN study: incidence and outcome of
acute respiratory distress syndrome in the
era of lung protective ventilation. Intensive Care Med 2011;37:1932-41. [Erratum,
Intensive Care Med 2011;37:1942.]
16. Perkins GD, Gates S, Lamb SE, McCabe C, Young D, Gao F. Beta Agonist
Lung Injury TrIal-2 (BALTI-2) trial protocol: a randomised, double-blind, placebocontrolled of intravenous infusion of salbutamol in the acute respiratory distress
syndrome. Trials 2011;12:113.
17. Mehta S, Lapinsky SE, Hallett DC, et
al. Prospective trial of high-frequency oscillation in adults with acute respiratory
distress syndrome. Crit Care Med 2001;
29:1360-9.
18. Ranieri VM, Rubenfeld GD, Thompson
BT, et al. Acute respiratory distress syndrome: the Berlin definition. JAMA 2012;
307:2526-33.
19. Ferguson ND, Cook DJ, Guyatt GH, et
al. High-frequency oscillation in early acute
respiratory distress syndrome. N Engl J
Med 2013;368:795-805.
Copyright © 2013 Massachusetts Medical Society.
an nejm app for iphone
The NEJM Image Challenge app brings a popular online feature to the smartphone.
Optimized for viewing on the iPhone and iPod Touch, the Image Challenge app lets
you test your diagnostic skills anytime, anywhere. The Image Challenge app
randomly selects from 300 challenging clinical photos published in NEJM,
with a new image added each week. View an image, choose your answer,
get immediate feedback, and see how others answered.
The Image Challenge app is available at the iTunes App Store.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
813
new england
journal of medicine
The
established in 1812
february 28, 2013
vol. 368 no. 9
High-Frequency Oscillation in Early Acute Respiratory
Distress Syndrome
Niall D. Ferguson, M.D., Deborah J. Cook, M.D., Gordon H. Guyatt, M.D., Sangeeta Mehta, M.D., Lori Hand, R.R.T.,
Peggy Austin, C.C.R.A., Qi Zhou, Ph.D., Andrea Matte, R.R.T., Stephen D. Walter, Ph.D., Francois Lamontagne, M.D.,
John T. Granton, M.D., Yaseen M. Arabi, M.D., Alejandro C. Arroliga, M.D., Thomas E. Stewart, M.D.,
Arthur S. Slutsky, M.D., and Maureen O. Meade, M.D., for the OSCILLATE Trial Investigators
and the Canadian Critical Care Trials Group*
A bs t r ac t
Background
Previous trials suggesting that high-frequency oscillatory ventilation (HFOV) reduced
mortality among adults with the acute respiratory distress syndrome (ARDS) were limited by the use of outdated comparator ventilation strategies and small sample sizes.
Methods
In a multicenter, randomized, controlled trial conducted at 39 intensive care units
in five countries, we randomly assigned adults with new-onset, moderate-to-severe
ARDS to HFOV targeting lung recruitment or to a control ventilation strategy targeting
lung recruitment with the use of low tidal volumes and high positive end-expiratory
pressure. The primary outcome was the rate of in-hospital death from any cause.
Results
On the recommendation of the data monitoring committee, we stopped the trial after
548 of a planned 1200 patients had undergone randomization. The two study groups
were well matched at baseline. The HFOV group underwent HFOV for a median of
3 days (interquartile range, 2 to 8); in addition, 34 of 273 patients (12%) in the
control group received HFOV for refractory hypoxemia. In-hospital mortality was
47% in the HFOV group, as compared with 35% in the control group (relative risk
of death with HFOV, 1.33; 95% confidence interval, 1.09 to 1.64; P = 0.005). This
finding was independent of baseline abnormalities in oxygenation or respiratory
compliance. Patients in the HFOV group received higher doses of midazolam than
did patients in the control group (199 mg per day [interquartile range, 100 to 382]
vs. 141 mg per day [interquartile range, 68 to 240], P<0.001), and more patients in the
HFOV group than in the control group received neuromuscular blockers (83% vs.
68%, P<0.001). In addition, more patients in the HFOV group received vasoactive
drugs (91% vs. 84%, P = 0.01) and received them for a longer period than did patients in the control group (5 days vs. 3 days, P = 0.01).
Conclusions
In adults with moderate-to-severe ARDS, early application of HFOV, as compared with
a ventilation strategy of low tidal volume and high positive end-expiratory pressure, does
not reduce, and may increase, in-hospital mortality. (Funded by the Canadian Institutes of Health Research; Current Controlled Trials numbers, ISRCTN42992782 and
ISRCTN87124254, and ClinicalTrials.gov numbers, NCT00474656 and NCT01506401.)
From the Interdepartmental Division of
Critical Care Medicine (N.D.F., S.M., J.T.G.,
T.E.S., A.S.S.), Departments of Medicine
and Physiology (N.D.F.), the Department
of Medicine, Division of Respirology (S.M.,
J.T.G., T.E.S.), and the Departments of
Medicine, Surgery, and Biomedical Engineering (A.S.S.), University of Toronto, University Health Network and Mount Sinai
Hospital (N.D.F., J.T.G.), Mount Sinai Hospital (S.M., T.E.S.), St. Michael’s Hospital
(A.S.S.), and the Critical Care Program, University Health Network (A.M.), Toronto; the
Interdepartmental Division of Critical Care,
Hamilton Health Sciences (D.J.C., M.O.M.),
the Departments of Medicine and Clinical
Epidemiology and Biostatistics (D.J.C.,
G.H.G., S.D.W., M.O.M.), and the CLARITY
Research Centre (D.J.C., G.H.G., S.D.W.,
M.O.M., L.H., P.A., Q.Z.), McMaster University, Hamilton, ON; and the Centre de
Recherche Clinique Étienne-Le Bel, Université de Sherbrooke, Sherbrooke, QC
(F.L.) — all in Canada; the Intensive Care
Department, King Saud bin Abdulaziz
University for Health Sciences, Riyadh,
Saudi Arabia (Y.M.A.); and the Department of Medicine, Scott and White Healthcare and Texas A&M Health Science Center College of Medicine, Temple (A.C.A.).
Address reprint requests to Dr. Meade at
1280 Main St. W., Hamilton, ON L8N 3Z5,
Canada, or at [email protected].
*A complete list of the investigators in the
Oscillation for Acute Respiratory Distress
Syndrome Treated Early (OSCILLATE)
trial is provided in the Supplementary
Appendix, available at NEJM.org.
This article was published on January 22,
2013, at NEJM.org.
N Engl J Med 2013;368:795-805.
DOI: 10.1056/NEJMoa1215554
Copyright © 2013 Massachusetts Medical Society.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
795
The
n e w e ng l a n d j o u r na l
T
he acute respiratory distress syndrome (ARDS) is a common complication
of critical illness.1,2 Mortality is high, and
survivors often have long-term complications.3,4
Although mechanical ventilation is life-sustaining
for patients with ARDS, it can perpetuate lung
injury. Basic research suggests that repetitive
overstretching or collapse of lung units with each
respiratory cycle can generate local and systemic
inflammation, contributing to multiorgan failure and death.5 Consistent with these findings
are data from clinical trials that support the use
of smaller tidal volumes (6 vs. 12 ml per kilogram of predicted body weight)6 and higher levels of positive end-expiratory pressure (PEEP).7-10
Mortality remains high, however, and additional
therapies are needed to protect the lung in cases
of severe ARDS.11,12
One such approach is high-frequency oscillatory ventilation (HFOV), which delivers very small
tidal volumes (approximately 1 to 2 ml per kilogram13) at very high rates (3 to 15 breaths per
second).14-19 Previous randomized trials of the use
of HFOV in adults with ARDS have suggested that
this strategy results in improvements in oxygenation and survival, but the trials were limited by
small sample sizes and outdated ventilation strategies for the control group.20-22 Consequently, despite the frequent use of HFOV in patients who
do not have an adequate response to conventional
mechanical ventilation and the increased use of
HFOV earlier in the course of the disease, this
approach remains an unproven therapy for adults
with ARDS.23-26 We therefore compared HFOV
with a conventional ventilation strategy that used
low tidal volumes and high levels of PEEP in patients with new-onset, moderate-to-severe ARDS.
Me thods
Study Oversight
For the pilot phase of the study, we enrolled patients at 11 centers in Canada and 1 in Saudi
Arabia from July 2007 through June 2008; for the
main trial, we enrolled patients at the same centers and at an additional 27 centers in Canada, the
United States, Saudi Arabia, Chile, and India from
July 2009 through August 2012 (see the Supplementary Appendix, available with the full text of
this article at NEJM.org). The trial protocol, which
is available at NEJM.org, was approved by the research ethics board at each participating site.
796
of
m e dic i n e
The first and last author vouch for the accuracy
and completeness of the reported data and for
the fidelity of this report to the study protocol.
For HFOV, we used the SensorMedics 3100B
High-Frequency Oscillatory Ventilator (CareFusion);
the manufacturer loaned nine ventilators and
provided technical support but had no role in the
design of the study, the collection or analysis of
the data, or the preparation of the manuscript.
Patients
Patients were eligible for inclusion if they had
had an onset of pulmonary symptoms within the
previous 2 weeks, had undergone tracheal intubation, had hypoxemia (defined as a ratio of the
partial pressure of arterial oxygen [Pao2] to the
fraction of inspired oxygen [Fio2] of ≤200, with
an Fio2 of ≥0.5), and had bilateral air-space opacities on chest radiography. Patients were excluded
if they had hypoxemia primarily related to left
atrial hypertension, suspected vasculitic pulmonary hemorrhage, neuromuscular disorders that
are known to prolong the need for mechanical
ventilation, severe chronic respiratory disease, or
preexisting conditions with an expected 6-month
mortality exceeding 50%; if they were at risk for
intracranial hypertension; if there was a lack of
commitment to life support; if the expected duration of mechanical ventilation was less than 48
hours; if they were younger than 16 years of age
or older than 85 years of age; or if their weight
was less than 35 kg or more than 1 kg per centimeter of height. We did not enroll patients who
had already met the eligibility criteria for more
than 72 hours, those who were already receiving
HFOV, or those whose physicians declined to enroll them.
After enrollment, standardized ventilator settings were used for all the patients: pressurecontrol mode, a tidal volume of 6 ml per kilogram, and an Fio2 of 0.60 with a PEEP level of
10 cm of water or higher if needed for oxygenation. After 30 minutes, if the Pao2:Fio2 ratio
remained at 200 or lower, patients underwent
randomization; otherwise the standardized ventilator settings were maintained, and the patients
were reassessed at least once daily for up to
72 hours. Eligible patients were randomly assigned in a 1:1 ratio to the HFOV group or to
the conventional-ventilation group. Randomization was performed in undisclosed block sizes of
2 and 4 with the use of a central Web-based
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation in Early ARDS
Table 1. Ventilator Protocols.*
Component Variable
Ventilator mode
HFOV
Control Ventilation
High-frequency oscillatory
ventilation
Pressure control
Tidal volume target (ml/kg of predicted body weight)
NA
6
Tidal volume range (ml/kg of predicted body weight)
NA
4–8
Plateau airway pressure (cm of water)
NA
≤35
Positive end-expiratory pressure (cm of water)
NA
Adjusted according to
oxygenation†
Adjusted according to
oxygenation†
Measured but not adjusted
3–12 Hz
≤35 breaths/min
Mean airway pressure (cm of water)
Respiratory frequency
Pressure amplitude target (cm of water)
90
NA
Partial pressure of arterial oxygen (mm Hg)
55–80
55–80
Oxygen saturation by pulse oximetry (%)
88–93
88–93
7.25–7.35
7.30–7.45
Ratio of inspiratory-to-expiratory time
Arterial blood pH
1:2
1:1–1:3
Recruitment maneuvers
Yes
Yes
*The full version of the study protocol is available at NEJM.org. HFOV denotes high-frequency oscillatory ventilation, and NA
not applicable.
†For more information on the protocol for adjustment, see Table 2.
randomization system, stratified according to
center. All patients or their legal surrogates provided written informed consent for participation
in the study.
Table 2. Usual Combinations of the Fraction of Inspired
Oxygen (Fio2) and Positive End-Expiratory Pressure (PEEP)
or Mean Airway Pressure Used to Adjust Ventilators.
HFOV
HFOV Protocol
The HFOV protocol was designed on the basis of
the results of pilot testing and consensus guidelines.24,27 We first conducted a recruitment maneuver, by applying 40 cm of water pressure for
40 seconds to the airway opening in an effort to
reopen closed lung units. We then initiated HFOV
with a mean airway pressure of 30 cm of water,
adjusting the pressure thereafter according to the
protocol, targeting a Pao2 of 55 to 80 mm Hg (Tables 1 and 2). We minimized HFOV tidal volumes
by using the highest possible frequency that
would maintain arterial blood pH above 7.25.13,28
After 24 hours of HFOV, conventional ventilation could be resumed if the mean airway pressure was 24 cm of water or less for 12 hours.
This transition was mandatory when airway
pressures reached 20 cm of water. Thereafter,
mechanical ventilation followed the control protocol. Over the next 48 hours, if an Fio2 of more
than 0.4 or a PEEP level of more than 14 cm of
water was required for more than 1 hour to
achieve oxygenation targets, HFOV was resumed.
Fio2
Control Ventilation
Mean Airway
Pressure
Fio2
cm of water
PEEP
cm of water
0.4
20
0.3
0.4
22
0.3
5
8
0.4
24
0.3
10
0.4
26
0.4
10
0.4
28
0.4
12
0.4
30
0.4
14
0.5
30
0.4
16
0.6
30
0.4
18
0.6
32
0.5
18
0.6
34
0.5
20
0.7
34
0.6
20
0.8
34
0.7
20
0.9
34
0.8
20
1.0
34
0.8
22
1.0
36
0.9
22
1.0
38
1.0
22
1.0
24
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
797
The
n e w e ng l a n d j o u r na l
m e dic i n e
Control Ventilation Protocol
Statistical Analysis
The control ventilation protocol, which was adapted from an earlier trial,9 called for a target tidal
volume of 6 ml per kilogram, with plateau airway
pressure of 35 cm of water or less and high levels
of PEEP. After an initial recruitment maneuver
(the same as that used for the HFOV group), clinicians applied ventilation using pressure-control
mode with a PEEP level of 20 cm of water and
then adjusted the PEEP level and the Fio2 according to the protocol (Tables 1 and 2). The protocol
permitted the use of volume-assist control mode
or pressure-support mode with the same limits
for tidal volumes and airway pressures. For patients receiving pressure support with PEEP levels of 10 cm of water or less and an Fio2 of 0.4 or
less, there were no limits on tidal volume or airway pressures. The weaning protocol, which has
been published previously, included daily trials of
spontaneous breathing.9,29
We anticipated that mortality in the control group
would be 45%. Assuming a two-sided alpha level
of 0.05, we calculated that enrollment of 1200
patients would provide at least 80% power to detect a relative-risk reduction with HFOV of 20%,
even if mortality in the control group was as low
as 37%.
Investigators reviewed feasibility data from the
pilot phase, which involved 94 patients, but remained unaware of the clinical outcomes. The
independent data monitoring committee reviewed
the clinical outcomes from the pilot phase and
recommended that the trial continue to the next
phase. As originally planned, data from the patients involved in the pilot phase were included
in the current analyses. In addition to an interim
analysis after 800 patients had undergone randomization, safety analyses of physiological data
at the initiation of the study were planned after
300, 500, and 700 patients had undergone randomization. After reviewing these safety data, the
data monitoring committee could request analyses of in-hospital mortality, which they did after
both the 300-patient and 500-patient safety
analyses. With plans to stop the study early only
in response to a strong signal of harm in association with the use of HFOV, we used the
O’Brien–Fleming method to calculate alpha
spending and generated one-sided P values for
considering early stopping after random assignment of 300 patients (P≤0.00001), 500 patients
(P≤0.0001), and 700 patients (P≤0.0064).
We used SAS software, version 9.2, for the
statistical analyses. We summarized data using
means with standard deviations, medians and
interquartile ranges, or proportions. Normally
distributed data were compared with the use of
Student’s t-test, nonnormally distributed data
with the use of the Wilcoxon rank-sum test, and
proportions with the use of the Mantel–Haenszel
chi-square test, with stratification according to
center. We analyzed data from all patients according to their assigned group.
The primary outcome was in-hospital mortality, with the outcome compared between the two
groups stratified according to center. Other than
recording whether death occurred as a result of
withdrawal of life support, we did not record
specific causes of death. As a sensitivity analysis, we used logistic regression to adjust the
treatment effect for prespecified baseline vari-
Procedures in Both Groups
When hypoxemia persisted despite increases in
PEEP or mean airway pressure, or when, on the
basis of radiographic or clinical evidence, physicians judged that the lungs were over-distended,
they could reduce PEEP or mean airway pressure
to a level below that indicated in the assigned
protocol (Table 2).
For patients with hypoxemia who required an
Fio2 of 0.9 or greater, clinicians could institute
therapies for hypoxemia (e.g., prone positioning
or inhaled nitric oxide) that did not interfere
with the assigned ventilator protocols. Physicians
could institute any alternative therapy (including
HFOV in the control group) for patients who met
any one of the following criteria: refractory hypoxemia (Pao2 <60 mm Hg for 1 hour with an
Fio2 of 1.0 and neuromuscular blockade), refractory barotrauma (persistent pneumothorax or increasing subcutaneous emphysema despite two
thoracostomy tubes on the involved side), or refractory acidosis (pH of ≤7.05 despite neuromuscular blockade).
Physicians prescribed fluids, sedatives, and
neuromuscular blockers at their discretion. We
recorded cardiorespiratory variables daily as well
as data on cointerventions applied while patients
were undergoing mechanical ventilation for up to
60 days. Intensivists reviewed chest radiographs
for evidence of new barotrauma. Patients were
followed until their discharge from the hospital.
798
of
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation in Early ARDS
ables: age, the Acute Physiology Score component of the Acute Physiology and Chronic Health
Evaluation (APACHE) II score,30 the presence or
absence of sepsis, and the duration of hospitalization before randomization.9 To compare the
two groups with respect to the time to death, we
used a survival analysis, in which patients who
were discharged alive from the hospital were assumed to be alive at day 60.
We conducted prespecified subgroup analyses
to determine whether there were interactions of
the treatment effect with baseline severity of lung
injury (in quartiles of the Pao2:Fio2 ratio) or with
center experience with HFOV and study protocols (in thirds of number of patients recruited).
In addition, we studied interactions of the treatment effect with baseline dynamic compliance
measured from tidal breaths during conventional
ventilation (in quartiles), baseline body-mass index (in quartiles), and receipt or no receipt of
vasopressors at baseline — all post hoc analyses.
R e sult s
Early Termination of the Trial
After the 500-patient analysis, the steering committee terminated the trial, acting on a unanimous recommendation from the data monitoring committee, although the threshold P value
for stopping had not been reached. At the time of
termination, 571 patients had been enrolled, of
whom 548 had undergone randomization: 275 to
the HFOV group and 273 to the control-ventilation group (Fig. 1). Important prognostic factors
were similar in the two groups at baseline (Table 3,
and Table S1 in the Supplementary Appendix).
Mortality
A total of 129 patients (47%) in the HFOV group,
as compared with 96 patients (35%) in the control
group, died in the hospital (relative risk of death
with HFOV, 1.33; 95% confidence interval, 1.09 to
1.64; P = 0.005) (Table 4 and Fig. 2). The results
were consistent in a multivariable analysis (Table
S2 in the Supplementary Appendix), in an analysis of mortality in the intensive care unit (ICU),
and in an analysis of 28-day mortality. Subgroup
analyses showed no interaction of mortality with
baseline severity of hypoxemia, respiratory compliance, body-mass index, or use or nonuse of
vasopressors or with center experience in the
trial (Fig. S1 in the Supplementary Appendix).
Early Physiological Responses to Ventilation
Table S3 in the Supplementary Appendix shows
early physiological responses to HFOV and to
control ventilation. The use of vasopressors was
similar in the HFOV and control groups before
the initiation of ventilation (66% and 61%, respectively; P = 0.24) but increased in the HFOV
group as compared with the control group within 4 hours after initiation (73% vs. 62%, P = 0.01)
and increased even more in the HFOV group by
the following day (78% vs. 58%, P<0.001). The
use of neuromuscular blockers followed a similar
pattern: 27% of patients in the HFOV group and
29% of those in the control group received neuromuscular blockers before the initiation of ventilation (P = 0.66), 46% as compared with 31%
received them within 4 hours after initiation
(P<0.001), and 46% as compared with 26% received them the next day (P<0.001). The mean Fio2
at these time points decreased to a similar extent
in both groups: the Fio2 was 0.75 in the HFOV
group and 0.73 in the control group before initiation (P = 0.93); 0.62 and 0.64 in the two groups, respectively, 4 hours after initiation (P = 0.94); and
0.51 and 0.50, respectively, the next day (P = 0.97).
Cardiorespiratory Results
Table S4 in the Supplementary Appendix shows
cardiorespiratory data from the first week of the
study. On day 1, the mean (±SD) of the mean airway pressure in the HFOV group was 31±2.6 cm
of water, with a frequency of 5.5±1.0 Hz; patients
in the control group underwent ventilation with
a tidal volume of 6.1±1.3 ml per kilogram, PEEP
of 18±3.2, and plateau pressure of 32±5.7 cm
of water. The mean Fio2 in the control group
was similar to or lower than that in the HFOV
group, despite lower mean airway pressures. The
net fluid balance was higher in the HFOV group
than in the control group, but the difference was
not significant. In the HFOV group, 270 of the
275 patients (98%) underwent HFOV for a median
of 3 days (interquartile range, 2 to 8); a total of
222 patients (81%) survived and were transitioned
to conventional ventilation for a further 5 days
(interquartile range, 2 to 7). In the control group,
34 patients (12%) crossed over to HFOV (31 according to protocol and 3 in violation of protocol) for
7 days (interquartile range, 5 to 15), beginning
2 days (interquartile range, 1 to 4) after randomization; 24 of those 34 patients (71%) died in the
hospital.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
799
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
2720 Patients met inclusion criteria
1571 Were excluded (some patients may
have more than 1 exclusion criterion)
411 Had primary cardiac failure
316 Had chronic respiratory disease
259 Had condition with expected >50%
6-mo mortality
193 Were considering palliation rather than
aggressive care
118 Were at risk for intracranial hypertension
103 Had vasculitic pulmonary hemorrhage
373 Had other reasons
1149 Were eligible
574 Were not enrolled
254 Did not provide consent
130 Were withdrawn by physician
82 Were in ICU >72 hr
75 Were already undergoing HFOV
24 Were enrolled in related trial
9 Had other reasons
571 Were enrolled
23 Did not undergo randomization
19 Had Pao2:FIo2 ratio >200 mm Hg
on standard settings
4 Had other reasons
548 Underwent randomization
273 Were assigned to receive control
ventilation
273 Received assigned intervention
275 Were assigned to receive HFOV
270 Received assigned intervention
2 Died before HFOV could be started
2 Withdrew consent after randomization
1 Had approval withdrawn by physician
1 Had premature termination of assigned
strategy after withdrawal of consent
3 Had premature termination of assigned
strategy after withdrawal of consent
273 Were included in primary analysis
275 Were included in primary analysis
Figure 1. Screening, Randomization, and Follow-up.
HFOV denotes high-frequency oscillatory ventilation, ICU intensive care unit, and Pao2:Fio2 the ratio of the partial
pressure of arterial oxygen to the fraction of inspired oxygen.
800
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation in Early ARDS
Table 3. Baseline Characteristics of the Patients.*
HFOV Group
(N = 275)
Characteristic
Age — yr
Female sex — no. (%)
Control Group
(N = 273)
Patients Eligible but
Not Enrolled
(N = 472)†
P Value‡
55±16
54±16
53±16
0.18
108 (39)
120 (44)
198 (42)
0.42
26±8
<0.001
APACHE II score§
29±8
29±7
Duration of hospital stay — days
5.6±8.0
4.9±8.0
Duration of mechanical ventilation — days
2.5±3.3
1.9±2.3
Risk factors for ARDS — no. of patients (%)
Sepsis
128 (47)
130 (48)
193 (41)
0.01
Pneumonia
155 (56)
164 (60)
289 (61)
0.37
49 (18)
44 (16)
51 (11)
0.02
Gastric aspiration
Trauma
10 (4)
5 (2)
24 (5)
0.07
Other
71 (26)
67 (25)
137 (29)
0.34
Tidal volume — ml/kg of predicted body weight
7.2±1.9
7.1±1.8
Plateau pressure — cm of water
29±6
29±7
27±7
<0.001
11±4
<0.001
13±3
13±4
Minute ventilation — liters/min
Set PEEP — cm of water
11.3±3.1
11.2±3.3
Oxygenation index
19.6±11.2
19.9±9.3
17.8±10.2
0.002
Pao2:Fio2 ratio — mm Hg
121±46
114±38
118±47
0.17
PaCo2 — mm Hg
Arterial pH
Barotrauma — no. of patients (%)
46±13
47±14
45±14
0.01
7.32±0.10
7.31±0.10
7.32±0.12
0.06
19 (7)
14 (5)
184 (67)
171 (63)
Cointerventions — no. of patients (%)
Inotropes or vasopressors
Renal-replacement therapy
29 (11)
28 (10)
Glucocorticoids
93 (34)
96 (35)
Neuromuscular blockers
84 (31)
94 (34)
*Plus–minus values are means ±SD. There were no significant differences between the two study groups in any of the baseline characteristics listed here, with the exception of duration of mechanical ventilation, for which P = 0.003. ARDS denotes
acute respiratory distress syndrome, and Pao2 partial pressure of arterial oxygen.
†Not all centers had approval from an ethics committee to collect data on patients who were eligible but not enrolled in
the study.
‡The P values are for the comparison of patients who were eligible but not enrolled with all patients who underwent randomization, with adjustment for stratification according to center.
§ Scores on the Acute Physiology and Chronic Health Evaluation II (APACHE II) range from 0 to 71, with higher scores
indicating greater severity of illness.
Cointerventions
During the course of the study, larger proportions
of patients in the HFOV group than in the control
group received vasoactive drugs (91% vs. 84%,
P = 0.01) and neuromuscular blockers (83% vs.
68%, P<0.001); vasoactive drugs were administered for an average of 2 days longer in the HFOV
group than in the control group, and neuromus-
cular blockers were administered for an average
of 1 day longer in the HFOV group (Table S5 in the
Supplementary Appendix). Sedatives and opioids
(most commonly midazolam and fentanyl) were
administered for the same duration in the two
groups (median, 10 days [interquartile range, 6 to
18] and 10 days [interquartile range, 6 to 17], respectively; P = 0.99), but during the first week the
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
801
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
Table 4. Outcomes.
HFOV Group
(N = 275)
Control Group
(N = 273)
Relative Risk
(95% CI)
P Value
Death in hospital — no. (%)
129 (47)
96 (35)
1.33 (1.09–1.64)
0.005
Death in intensive care unit — no. (%)
123 (45)
84 (31)
1.45 (1.17–1.81)
0.001
Death before day 28 — no. (%)
111 (40)
78 (29)
1.41 (1.12–1.79)
0.004
New barotrauma — no./total no. (%)*
46/256 (18)
34/259 (13)
1.37 (0.91–2.06)
0.13
New tracheostomy — no./total no. (%)†
59/273 (22)
66/267 (25)
0.87 (0.64–1.19)
0.39
19 (7)
38 (14)
0.50 (0.29–0.84)
0.007
Outcome
Refractory hypoxemia — no. (%)
Death after refractory hypoxemia — no./total no. (%)
15/19 (79)
25/38 (66)
1.20 (0.87–1.66)
0.31
Refractory acidosis — no. (%)
9 (3)
8 (3)
1.12 (0.44–2.85)
0.82
Refractory barotrauma — no. (%)
2 (<1)
2 (<1)
0.99 (0.14–7.00)
0.99
Use of mechanical ventilation, among survivors
— days
Median
Interquartile range
0.59
11
10
7–19
6–18
15
14
9–25
9–26
Stay in intensive care, among survivors — days
Median
Interquartile range
0.93
Length of hospitalization, among survivors — days
Median
Interquartile range
0.74
30
25
16–45
15–41
*Barotrauma was defined as pneumothorax, pneumomediastinum, pneumopericardium, or subcutaneous emphysema
occurring spontaneously or after a recruitment maneuver. Excluded from this category were patients who had barotrauma
at baseline.
†Excluded from this category were patients who had a tracheostomy at baseline.
median doses of midazolam were significantly
higher in the HFOV group than in the control
group (199 mg per day [interquartile range, 100
to 382] vs. 141 mg per day [interquartile range,
68 to 240], P<0.001), and there was a trend toward higher doses of fentanyl equivalents in the
HFOV group (2980 μg per day [interquartile range,
1258 to 4800] vs. 2400 μg per day [interquartile
range, 1140 to 4430], P = 0.06) (for daily doses of
selected sedative and analgesic drugs, see Fig. S2
in the Supplementary Appendix). The rates of use
of other cointerventions, including glucocorticoids, renal-replacement therapy, and prone positioning, were similar in the two groups (Table S5
in the Supplementary Appendix).
Other Outcomes
Refractory hypoxemia developed in significantly
more patients in the control group than in the
HFOV group; however, the total number of deaths
after refractory hypoxemia was similar in the
802
two groups (Table 4). The proportion of deaths
after withdrawal of life support was similar in
the two groups (55% [71 of 129 patients] in the
HFOV group and 49% [47 of 96 patients] in the
control group, P = 0.12). The rate of new-onset
barotrauma was higher in the HFOV group than
in the control group, but the difference was not
significant (18% and 13%, respectively; P = 0.13).
Among survivors, the duration of ventilation and
the length of stay in the ICU were similar in the
two groups (Table 4).
Discussion
The main finding of this multicenter, randomized trial is that among patients with moderateto-severe ARDS, early application of HFOV was
associated with higher mortality than was a ventilation strategy that used small tidal volumes
and high PEEP levels, with HFOV used only in
patients with severe refractory hypoxemia. HFOV
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation in Early ARDS
1.0
0.9
Probability of Survival
was associated with higher mean airway pressures
and with greater use of sedatives, neuromuscular
blockers, and vasoactive drugs.
We stopped the trial early on the basis of a
strong signal for increased mortality with HFOV,
even though the prespecified stopping thresholds
had not been reached. Studies that are stopped
early on the basis of harm (or benefit) typically
overestimate the magnitude of effect.31 We chose
to terminate the study for three reasons: there
was a consistent finding of increased mortality
with HFOV in three consecutive analyses that were
conducted after enrollment of 94, 300, and 500
patients; the increased need for vasoactive drugs
in the HFOV group suggested a mechanism of
harm that was not offset by better oxygenation
and lung recruitment; and the effect size was
sufficiently large that we concluded that even if
early HFOV did not increase mortality, it would
be very unlikely to decrease mortality. We believe that continued enrollment would have put
patients at risk with little likelihood of benefit.
Our results are inconsistent with the physiological rationale for HFOV and with the results
of studies in animals. In studies in animals in
which benefits of HFOV were observed, lung
injury was induced with the use of saline lavage
— a highly recruitable model of surfactant deficiency — which our results suggest does not
translate directly to human adults with ARDS, in
whom recruitability can be heterogeneous.32 Our
results also contrast with those of prior randomized trials involving adults.22 A possible explanation, which provided motivation for our trial, is
that prior studies used control ventilation strategies that are now known to be potentially harmful.20,21 We found no benefit with HFOV when a
current ventilation strategy was used as a control. This finding of no benefit with respect to
mortality is consistent with the results of another trial now reported in the Journal; in that
trial, conducted in the United Kingdom, current
standards for lung protection were suggested
but not mandated.33 More surprising was our
finding of harm. Several plausible mechanisms
may contribute to increased mortality with
HFOV. Higher mean airway pressures may result
in hemodynamic compromise by decreasing venous return or directly affecting right ventricular
function.34 Increased use of vasodilating sedative agents may also contribute to hemodynamic
compromise. Moreover, we cannot exclude the
0.8
0.7
Control
0.6
0.5
0.4
HFOV
0.3
0.2
P=0.004 by log-rank test
0.1
0.0
0
15
30
45
60
54
54
26
39
Days since Randomization
No. at Risk
HFOV
Control
275
273
169
181
98
92
Figure 2. Probability of Survival from the Day of Randomization to Day 60
in the HFOV and Control Groups.
possibility of increased barotrauma in association with HFOV.
The HFOV strategy that we chose, which was
supported by preclinical data15,16 and a prospective physiological study,24 aimed to adjust mean
airway pressure on the deflation limb of the
volume-pressure curve and use the highest frequency possible to limit oscillatory volumes.
This approach led to relatively high mean airway
pressures, even considering that when mean airway pressures are delivered with a ratio of inspiratory-to-expiratory time of 1:2, as in our study,
the pressures measured at the airway opening
during HFOV are somewhat higher than those
measured in the trachea.35-37 It is possible that an
HFOV protocol that uses lower mean airway pressures, a different ratio of inspiratory-to-expiratory
time, or a lower oscillatory frequency might have
led to different results.
The strengths of this trial include its methodologic rigor, the application of protocols designed
to open lung units in patients in both groups on
the basis of the best available evidence, and enrollment at centers in several countries, which
enhances the generalizability of our findings.
Because we were cognizant that there is a learning curve associated with the use of HFOV,38,39
we enrolled most patients at centers that were
experienced with HFOV, and we did not detect
an interaction between treatment effect and the
number of enrolled patients per site.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
803
The
n e w e ng l a n d j o u r na l
Our results raise serious concerns about the
early use of HFOV for the management of ARDS
in adults. The results of this study increase the
uncertainty about possible benefits of HFOV even
when applied in patients with life-threatening
refractory hypoxemia.
In conclusion, in adults with moderate-tosevere ARDS, the early application of HFOV
targeting lung recruitment — as compared with
a ventilation strategy that uses low tidal volume
and high PEEP and that permits HFOV only in
cases of refractory hypoxemia — does not reduce mortality and may be harmful.
Supported by the Canadian Institutes of Health Research,
Randomized Controlled Trial (RCT) Program (Ottawa) and the
King Abdullah International Medical Research Center (Riyadh,
of
m e dic i n e
Saudi Arabia). Dr. Ferguson is supported by a Canadian Institutes of Health Research New Investigator Award; Dr. Cook
holds a Canada Research Chair; Dr. Lamontagne is supported by
a Fonds de Recherche de Québec–Santé Research Career Award;
and Drs. Ferguson and Meade and Drs. Lamontagne and Meade
were supported by Canadian Institutes of Health Research RCT
Mentorship awards.
Dr. Ferguson reports receiving grant support through his institution from the Physicians Services Incorporated Foundation;
Dr. Granton, receiving consulting fees from Ikaria, lecture fees
from Actelion and Eli Lilly, grant support through his institution from Pfizer, Actelion, Eli Lilly, GlaxoSmithKline, and Bayer, payment through his institution from Telus for the sale of a
software site license, and support from Actelion for his hospital
foundation for pulmonary hypertension research and providing
expert testimony for Pfizer regarding patent legal action; and
Dr. Slutsky, receiving consulting fees from Maquet Medical,
Novalung, Gambro, Ikaria, and Hemodec. No other potential
conflict of interest relevant to this article was reported.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
References
1. ARDS Definition Task Force, Ranieri
VM, Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin Definition. JAMA 2012;307:2526-33.
2. Rubenfeld GD, Caldwell E, Peabody E,
et al. Incidence and outcomes of acute
lung injury. N Engl J Med 2005;353:
1685-93.
3. Phua J, Badia JR, Adhikari NKJ, et al.
Has mortality from acute respiratory distress syndrome decreased over time? Am J
Respir Crit Care Med 2009;179:220-7.
4. Herridge MS, Tansey CM, Matté A, et
al. Functional disability 5 years after acute
respiratory distress syndrome. N Engl J
Med 2011;364:1293-304.
5. Tremblay LN, Slutsky AS. Ventilatorinduced lung injury: from the bench to
the bedside. Intensive Care Med 2006;32:
24-33.
6. The Acute Respiratory Distress Syndrome Network. Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute lung injury and
the acute respiratory distress syndrome.
N Engl J Med 2000;342:1301-8.
7. Brower RG, Lanken PN, MacIntyre N,
et al. Higher versus lower positive endexpiratory pressures in patients with the
acute respiratory distress syndrome.
N Engl J Med 2004;351:327-36.
8. Mercat A, Richard J-CM, Vielle B, et
al. Positive end-expiratory pressure setting in adults with acute lung injury and
acute respiratory distress syndrome: a
randomized controlled trial. JAMA 2008;
299:646-55.
9. Meade MO, Cook DJ, Guyatt GH, et al.
Ventilation strategy using low tidal volumes, recruitment maneuvers, and high
positive end-expiratory pressure for acute
lung injury and acute respiratory distress
syndrome: a randomized controlled trial.
JAMA 2008;299:637-45.
804
10. Briel M, Meade M, Mercat A, et al.
Higher vs lower positive end-expiratory
pressure in patients with acute lung injury
and acute respiratory distress syndrome:
systematic review and meta-analysis.
JAMA 2010;303:865-73.
11. Terragni PP, Rosboch G, Tealdi A, et
al. Tidal hyperinflation during low tidal
volume ventilation in acute respiratory
distress syndrome. Am J Respir Crit Care
Med 2007;175:160-6.
12. Terragni PP, Del Sorbo L, Mascia L, et
al. Tidal volume lower than 6 ml/kg enhances lung protection: role of extracorporeal carbon dioxide removal. Anesthesiology 2009;111:826-35.
13. Hager DN, Fessler HE, Kaczka DW, et
al. Tidal volume delivery during high-frequency oscillatory ventilation in adults
with acute respiratory distress syndrome.
Crit Care Med 2007;35:1522-9.
14. Ferguson ND, Slutsky AS, Kacmarek
R. High-frequency ventilation is/is not the
optimal physiological approach to ventilate ARDS patients. J Appl Physiol 2008;
104:1230-1.
15. Froese AB. High-frequency oscillatory
ventilation for adult respiratory distress
syndrome: let’s get it right this time. Crit
Care Med 1997;25:906-8.
16. Rimensberger PC, Pache JC, McKerlie
C, Frndova H, Cox PN. Lung recruitment
and lung volume maintenance: a strategy
for improving oxygenation and preventing lung injury during both conventional
mechanical ventilation and high-frequency
oscillation. Intensive Care Med 2000;26:
745-55.
17. Imai Y, Nakagawa S, Ito Y, Kawano T,
Slutsky AS, Miyasaka K. Comparison of
lung protection strategies using conventional and high-frequency oscillatory
ventilation. J Appl Physiol 2001;91:
1836-44.
18. Muellenbach RM, Kredel M, Said HM,
et al. High-frequency oscillatory ventilation reduces lung inflammation: a largeanimal 24-h model of respiratory distress.
Intensive Care Med 2007;33:1423-33.
19. Sedeek KA, Takeuchi M, Suchodolski
K, et al. Open-lung protective ventilation
with pressure control ventilation, highfrequency oscillation, and intratracheal
pulmonary ventilation results in similar
gas exchange, hemodynamics, and lung
mechanics. Anesthesiology 2003;99:
1102-11.
20. Derdak S, Mehta S, Stewart TE, et al.
High frequency oscillatory ventilation for
acute respiratory distress syndrome: a
randomized controlled trial. Am J Respir
Crit Care Med 2002;166:801-8.
21. Bollen CW, van Well GT, Sherry T, et
al. High frequency oscillatory ventilation
compared with conventional mechanical
ventilation in adult respiratory distress
syndrome: a randomized controlled trial.
Crit Care 2005;9:R430-R439.
22. Sud S, Sud M, Friedrich JO, et al. High
frequency oscillation in patients with
acute lung injury and acute respiratory
distress syndrome (ARDS): systematic review and meta-analysis. BMJ 2010;340:
c2327.
23. Fort P, Farmer C, Westerman J, et al.
High-frequency oscillatory ventilation for
adult respiratory distress syndrome — a
pilot study. Crit Care Med 1997;25:937-47.
24. Ferguson ND, Chiche J-D, Kacmarek
RM, et al. Combining high-frequency oscillatory ventilation and recruitment maneuvers in adults with early acute respiratory distress syndrome: the Treatment
with Oscillation and an Open Lung Strategy (TOOLS) Trial pilot study. Crit Care
Med 2005;33:479-86.
25. Mehta S, Granton J, MacDonald RJ, et
al. High-frequency oscillatory ventilation
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
High-Frequency Oscillation in Early ARDS
in adults: the Toronto experience. Chest
2004;126:518-27.
26. Adhikari NKJ, Bashir A, Lamontagne
F, et al. High-frequency oscillation in
adults: a utilization review. Crit Care Med
2011;39:2631-44.
27. Fessler HE, Derdak S, Ferguson ND,
et al. A protocol for high-frequency oscillatory ventilation in adults: results from a
roundtable discussion. Crit Care Med
2007;35:1649-54.
28. Fessler HE, Hager DN, Brower RG.
Feasibility of very high-frequency ventilation in adults with acute respiratory distress syndrome. Crit Care Med 2008;36:
1043-8.
29. Esteban A, Frutos F, Tobin MJ, et al. A
comparison of four methods of weaning
patients from mechanical ventilation.
N Engl J Med 1995;332:345-50.
30. Knaus WA, Draper EA, Wagner DP,
Zimmerman JE. APACHE II: a severity of
disease classification system. Crit Care
Med 1985;13:818-29.
31. Montori VM, Devereaux PJ, Adhikari
NK, et al. Randomized trials stopped
early for benefit: a systematic review.
JAMA 2005;294:2203-9.
32. Gattinoni L, Caironi P, Cressoni M, et
al. Lung recruitment in patients with the
acute respiratory distress syndrome.
N Engl J Med 2006;354:1775-86.
33. Young D, Lamb SE, Shah S, et al. Highfrequency oscillation for acute respiratory
distress syndrome. N Engl J Med 2013;
368:806-13.
34. Guervilly C, Forel J-M, Hraiech S, et
al. Right ventricular function during
high-frequency oscillatory ventilation in
adults with acute respiratory distress syndrome. Crit Care Med 2012;40:1539-45.
35. Hatcher D, Watanabe H, Ashbury T,
Vincent S, Fisher J, Froese A. Mechanical
performance of clinically available, neo-
natal, high-frequency, oscillatory-type
ventilators. Crit Care Med 1998;26:1081-8.
36. Pillow JJ, Neil H, Wilkinson MH,
Ramsden CA. Effect of I/E ratio on mean
alveolar pressure during high-frequency
oscillatory ventilation. J Appl Physiol
1999;87:407-14.
37. Mentzelopoulos SD, Malachias S,
Kokkoris S, Roussos C, Zakynthinos SG.
Comparison of high-frequency oscillation
and tracheal gas insufflation versus standard high-frequency oscillation at two
levels of tracheal pressure. Intensive Care
Med 2010;36:810-6.
38. The HiFi Study Group. High-frequency oscillatory ventilation compared with
conventional mechanical ventilation in the
treatment of respiratory failure in preterm
infants. N Engl J Med 1989;320:88-93.
39. Bryan AC, Froese AB. Reflections on
the HIFI trial. Pediatrics 1991;87:565-7.
Copyright © 2013 Massachusetts Medical Society.
n engl j med 368;9 nejm.org february 28, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
805
Clinical Review & Education
Special Communication | LESS IS MORE
“Less Is More” in Critically Ill Patients
Not Too Intensive
Matthijs Kox, PhD; Peter Pickkers, MD, PhD
The current view in intensive care medicine is that very sick patients need very intensive
treatment. However, in this group of highly vulnerable patients, more intensive treatment
may promote the chances of unwanted adverse effects and hence, iatrogenic damage.
Therefore, we state that critically ill patients probably benefit from a more cautious approach.
Using data from large clinical trials of previous years, we exemplify that less intensive
treatment is associated with a better outcome in intensive care patients and suggest that we
reappraise patient management as well as trial design in intensive care medicine while
bearing in mind the “less is more” paradigm. We illustrate our case by describing the intensity
of the most relevant treatment options for patients with septic shock, including mechanical
ventilation, fluid management, blood pressure–targeted therapy, corticosteroids, patient
monitoring, sedation, and nutrition. We conclude that treatment of critically ill patients while
keeping in mind the “less is more” paradigm might not only benefit the patient but could also
have a notable impact on the ever-increasing intensive care–related health care costs.
JAMA Intern Med. 2013;173(14):1369-1372. doi:10.1001/jamainternmed.2013.6702
Published online June 10, 2013.
T
he phrase “less is more” was first popularized by the German architect Ludwig Mies van der Rohe, one of the pioneers of modern architecture. Because of financial shortages after World War II, the modernist style originally conceived for
aesthetic reasons was quickly adopted as a practical program of inexpensive construction. Although not directly related to the current financial crisis, we propose that “less is more” also applies to
the treatment of critically ill patients. The current view is that these
very sick patients need very intensive treatment. Indeed, the addition of several treatments such as mechanical ventilation and improved hemodynamic monitoring has revolutionized the care of
medical and surgical patients who are critically ill. However, because more intensive treatment may promote the chances of unwanted adverse effects, iatrogenic damage is more likely to occur,
resulting in unfavorable sequelae and worse outcome in this highly
vulnerable group of patients. Therefore, in keeping with the Hippocratic oath (primum non nocere), we state that critically ill patients
probably benefit from a more “cautious” approach. In this Special
Article, based on data from large clinical trials conducted in the previous years, we show that less intensive treatment is indeed associated with a better outcome in critically ill patients (Table). To improve outcome, we should therefore reappraise patient management
as well as trial design in intensive care medicine, bearing in mind with
the principles of “less is more.”
Respiratory Therapy, Fluids, and Transfusion
To illustrate our case, we describe the intensity of the most relevant treatment options for septic shock, the number 1 cause of
jamainternalmedicine.com
Author Affiliations: Department of
Intensive Care Medicine, Radboud
University Nijmegen Medical Center,
Nijmegen, the Netherlands (Kox,
Pickkers); Department of
Anesthesiology, Radboud University
Nijmegen Medical Center, Nijmegen,
the Netherlands (Kox); Nijmegen
Institute for Infection, Inflammation,
and Immunity (N4i), Nijmegen, the
Netherlands (Kox, Pickkers).
Corresponding Author: Matthijs
Kox, PhD, Department of Intensive
Care Medicine, Internal Mail 710,
Radboud University Nijmegen
Medical Center, Geert Grooteplein 10,
6500 HB Nijmegen, the Netherlands
([email protected]).
death in noncardiac intensive care units (ICUs). 1,24 In most
patients with septic shock, mechanical ventilation is required. The
first Acute Respiratory Distress Syndrome (ARDS) Network trial
has clearly shown that ventilation with lower tidal volumes (6 vs
12 mL/kg) reduces lung injury and improves outcome.1 Furthermore, higher levels of positive end-expiratory pressure (13 cm of
water) did not improve clinical outcome compared with lower levels (8 cm of water),2 indicating that “more” is not “better.”2 Of
interest, recently high-frequency ventilation was found to
increase mortality in ARDS patients, illustrating that complicated
tertiary approaches to ventilation do not seem to help.3 Related
to respiratory therapy, too-liberal supplementation of oxygen
resulting in high P O 2 values is associated with, albeit slightly,
increased mortality in observational cohorts.4 Randomized trials
to investigate the effects of higher and lower oxygen levels are
currently not available. In addition, following the initial fluid
resuscitation phase, it has been demonstrated that a more conservative fluid management results in improved lung function and
shorter stay in the ICU compared with a more liberal regime.5
Recently, an extreme example of the benefits of less fluid administration emerged in a study performed in African children presenting with febrile illness and impaired perfusion.6 Strikingly, a
survival benefit was found in children who were withheld from
fluid administration compared with those receiving, by US
and European standards, conservative fluid boluses. 6 Also, a
restrictive strategy of red blood cell transfusion is superior to
a liberal one.7 It is clear that for respiratory and hemodynamic
management, less is more in the critically ill patient. Remarkably,
the same holds true for many adjunctive treatments applied in
the ICU.
JAMA Internal Medicine July 22, 2013 Volume 173, Number 14
Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013
1369
Clinical Review & Education Special Communication
“Less Is More” in Critically Ill Patients
Table. Interventions for Which “More” Has Been Shown to Be Associated
With Worse Outcome
Intervention
Respiratory therapy
Higher tidal volume
(12 vs 6 mL/kg)1;
higher PEEP (13 vs 8
cm of water)2;
high-frequency
ventilation3
High PO24
Effect
Type of Study
1,3
1-3
Increased mortality
and days on ventilator1;
no benefit2
RCT
Increased mortality4
Observational study4
Worsened lung
function,5 increased
days on ventilator, and
ICU stay5; increased
mortality6
Increased mortality7
RCT5,6
No benefit8; increased
mortality9,11; increased
mortality (only in less
severe septic shock)10
RCT8,10; observational
study9; post hoc analysis of RCT11
High-dose
corticosteroids12
Cardiac therapy
Increased mortality12
RCT12
Prophylactic lidocaine treatment to
reduce VF13
Monitoring
Increased mortality13
RCT13
Swan-Ganz
cathether14
Daily routine x-ray
examinations15
Antibiotics
No benefit14
RCT14
No benefit15
Observational study15
Longer duration of
antibiotic
treatment16
Renal therapy
No benefit16
RCT16
High vs low volume
renal replacement
therapy17
Sedation
No benefit17
RCT17
Continuous sedation
vs daily interruption
of sedatives (lower
sedative use)18; daily
interruption of sedatives vs no sedation19
Nutrition and glucose
control
High vs moderate
caloric intake20; target feeding vs permissive
underfeeding21; full
enteral nutrition vs
trophic nutrition22
Strict vs less strict
glucose control in
patients that are predominantly fed
enterally23
Increased days on
ventilator18,19 and ICU
stay18
RCT18,19
Increased
mortality20,21;
no benefit22
Observational study20;
RCT21,22
Fluids and transfusion
Liberal fluid management vs strict
regime5; conservative
vs no fluid
administration6
Liberal vs restrictive
red blood cell
transfusion7
Blood pressure–targeted
therapy
Increasing MAP >65
mm Hg8; higher vasopressor use9; higher
vasopressor infusion
rates vs lower
rates + vasopressin10;
higher vs lower dosage of L-NMMA11
Corticosteroids
RCT7
Patients with septic shock require vasopressor therapy by definition. Although the use of different vasopressor agents appears to
result in similar outcome, increasing mean arterial pressure above
65 mm Hg showed no benefit.8 Furthermore, vasopressor load was
shown to be associated with mortality in an observational study.9
While there is no direct evidence for adverse effects of high vasopressor use (there are no randomized clinical trials comparing different dosages), most pharmacological adverse effects are dose dependent. Hence, combinations of (lower dosages) of vasopressors
might be beneficial compared with the use of higher dosages of a
single vasopressor, and indeed there is some evidence for this. For
instance, vasopressin therapy in combination with norepinephrine
resulted in a significant survival benefit in patients with mild septic
shock compared with patients treated with norepinephrine alone,
although this was not the case in patients with severe shock.10 Also,
while the nitric oxide synthase inhibitor N G -monomethyl-Larginine (L-NMMA) was associated with worse outcome in the complete group of patients with septic shock, those who received a lower
infusion rate (<5 mg/kg/h) had an improved survival rate relative to
placebo-treated patients, suggesting that lower dosages are
beneficial.11
Corticosteroids, Monitoring, and Antibiotics
Patients with refractory shock who require high dosages of vasopressor therapy might benefit from corticosteroids. However, in the
early trials, high dosages of corticosteroids resulted in adverse
outcome.12 In later trials it was demonstrated that lower corticosteroid dosages may improve outcome, especially when administrated early in those with refractory shock.25,26 In case of impending myocardial infarction, prophylactic treatment with lidocaine,
while reducing the incidence of ventricular fibrillation, actually increases mortality.13 Another argument in support of “less is more”
in critically ill patients is the fact that studies have failed to demonstrate that more invasive hemodynamic monitoring using SwanGanz catheters is beneficial for the ICU patient.14 Furthermore, daily
routine x-ray examinations, another form of monitoring, were not
shown to be useful,15 and even a shorter duration of antibiotic treatment (guided by procalcitonin levels) was not associated with worse
outcome,16 while liberal use of antibiotics has many disadvantages, among which increased risk of the development of resistance is the most important one.
Renal Therapy, Sedation, Nutrition,
and Glucose Control
Increased mortality23
RCT23
Abbreviations: ICU, intensive care unit; L-NMMA, NG-monomethyl-L-arginine;
MAP, mean arterial pressure; PEEP, positive end-expiratory pressure;
RCT, randomized clinical trial; VF, ventricular fibrillation.
1370
Blood Pressure–Targeted and Cardiac Therapy
Patients with septic shock often develop acute kidney injury, for
which renal replacement therapy can be initiated, but increasing ultrafiltrate volume has not showed any benefit.17 Most patients with
septic shock will be sedated. The old paradigm was to deeply sedate to facilitate adequate patient management. It is now becoming increasingly clear that less18 or even no19 sedation is superior. Finally, patients need nutrition, and it has emerged that high caloric
JAMA Internal Medicine July 22, 2013 Volume 173, Number 14
Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013
jamainternalmedicine.com
“Less Is More” in Critically Ill Patients
Special Communication Clinical Review & Education
intake is associated with increased mortality,20 whereas low caloric intake via the enteral route (permissive underfeeding) reduces hospital mortality in critically ill patients.21 In line, it was recently demonstrated in ARDS patients that full enteral feeding is not
superior compared with trophic feeding in terms of time on ventilator or mortality.22 Furthermore, early initiation of parenteral nutrition worsens outcome,27 suggesting it is better to limit caloric intake via the parenteral route until enteral feeding is possible. Related
to this issue, beneficial effects of intensive glucose control therapy
were found in ICU patients who receive larger amounts of parenteral calories,28 while it was found to increase mortality in patients
who are predominantly fed via the enteral route.23 From these studies, it might be concluded that very strict glucose regulation is not
beneficial by itself and only nullifies the deleterious effects of early
parenteral administration of calories.
Exceptions to the Rule
Naturally, there are some notable exceptions to the rule; sometimes “more is more.” For instance, early mobilization of critically ill
patients,29 infection prevention by means of decontamination of the
digestive tract and oropharynx,30 and high-intensity ICU physician31
and nurse32 staffing is associated with better outcome. Furthermore, hemodynamic optimization and/or early goal-directed therapy
is associated with better outcome following major surgery33 and in
the early phase of sepsis.34 Nevertheless, equipoise still remains, and
3 trials (United States, United Kingdom, and Australia) are under way
to evaluate the true effectiveness of early goal-directed therapy. FurARTICLE INFORMATION
Accepted for Publication: March 16, 2013.
Published Online: June 10, 2013.
doi:10.1001/jamainternmed.2013.6702.
Author Contributions: Study concept and design:
Both authors.
Drafting of the manuscript: Kox.
Critical revision of the manuscript for important
intellectual content: Pickkers.
Study supervision: Pickkers.
Conflict of Interest Disclosures: None reported.
REFERENCES
1. The Acute Respiratory Distress Syndrome
Network. Ventilation with lower tidal volumes as
compared with traditional tidal volumes for acute
lung injury and the acute respiratory distress
syndrome. N Engl J Med. 2000;342(18):1301-1308.
2. Brower RG, Lanken PN, MacIntyre N, et al;
National Heart, Lung, and Blood Institute ARDS
Clinical Trials Network. Higher versus lower positive
end-expiratory pressures in patients with the acute
respiratory distress syndrome. N Engl J Med.
2004;351(4):327-336.
3. Ferguson ND, Cook DJ, Guyatt GH, et al;
OSCILLATE Trial Investigators; Canadian Critical
Care Trials Group. High-frequency oscillation in
early acute respiratory distress syndrome. N Engl J
Med. 2013;368(9):795-805.
4. de Jonge E, Peelen L, Keijzers PJ, et al.
Association between administered oxygen, arterial
partial oxygen pressure and mortality in
jamainternalmedicine.com
thermore, a highly intensive and invasive treatment such as extracorporeal circulation membrane oxygenation has proved its value
during the 2009 H1N1 influenza pandemic.35 However, a recent observational study indicates that the effectiveness of extracorporeal circulation membrane oxygenation in ARDS patients is highly
dependent on chest wall mechanics and that inappropriate use of
extracorporeal circulation membrane oxygenation can be avoided
in many patients by partitioning respiratory system mechanics between lung and chest wall.36
Conclusions
It is becoming clear that following decades of increasingly more
intensive treatment of critically ill patients, a more cautious
approach for the vulnerable ICU patient is in many cases associated with improved outcome. While it may prove to be a difficult
task to convince more and more ICU physicians to do less, implementation of evidence-based reduction of therapy intensities is
of vital importance. With regard to ICU trial design, instead of
investigating the effects of altering a single therapy, it would be of
interest to evaluate the efficacy of combinations of lower dosages
or intensities of certain therapies to reduce the chances of
adverse effects to occur. Last but not least, keeping the “less is
more” paradigm in mind in the treatment of critically ill patients
might not only benefit the patient but could also have a notable
impact on the ever-increasing ICU-related health care costs and
thereby benefit the society as a whole, especially in these times
of financial hardship.
mechanically ventilated intensive care unit patients.
Crit Care. 2008;12(6):R156.
5. Wiedemann HP, Wheeler AP, Bernard GR, et al;
National Heart, Lung, and Blood Institute Acute
Respiratory Distress Syndrome (ARDS) Clinical
Trials Network. Comparison of two
fluid-management strategies in acute lung injury.
N Engl J Med. 2006;354(24):2564-2575.
6. Maitland K, Kiguli S, Opoka RO, et al; FEAST Trial
Group. Mortality after fluid bolus in African children
with severe infection. N Engl J Med. 2011;364(26):
2483-2495.
7. Hébert PC, Wells G, Blajchman MA, et al;
Transfusion Requirements in Critical Care
Investigators for the Canadian Critical Care Trials
Group. A multicenter, randomized, controlled
clinical trial of transfusion requirements in critical
care. N Engl J Med. 1999;340(6):409-417.
8. Bourgoin A, Leone M, Delmas A, Garnier F,
Albanèse J, Martin C. Increasing mean arterial
pressure in patients with septic shock: effects on
oxygen variables and renal function. Crit Care Med.
2005;33(4):780-786.
9. Dünser MW, Ruokonen E, Pettilä V, et al.
Association of arterial blood pressure and
vasopressor load with septic shock mortality: a post
hoc analysis of a multicenter trial. Crit Care.
2009;13(6):R181.
10. Russell JA, Walley KR, Singer J, et al; VASST
Investigators. Vasopressin versus norepinephrine
infusion in patients with septic shock. N Engl J Med.
2008;358(9):877-887.
11. López A, Lorente JA, Steingrub J, et al.
Multiple-center, randomized, placebo-controlled,
double-blind study of the nitric oxide synthase
inhibitor 546C88: effect on survival in patients with
septic shock. Crit Care Med. 2004;32(1):21-30.
12. Bone RC, Fisher CJ Jr, Clemmer TP, Slotman GJ,
Metz CA, Balk RA. A controlled clinical trial of
high-dose methylprednisolone in the treatment of
severe sepsis and septic shock. N Engl J Med.
1987;317(11):653-658.
13. Sadowski ZP, Alexander JH, Skrabucha B, et al.
Multicenter randomized trial and a systematic
overview of lidocaine in acute myocardial
infarction. Am Heart J. 1999;137(5):792-798.
14. Harvey S, Harrison DA, Singer M, et al; PAC-Man
study collaboration. Assessment of the clinical
effectiveness of pulmonary artery catheters in
management of patients in intensive care
(PAC-Man): a randomised controlled trial. Lancet.
2005;366(9484):472-477.
15. Graat ME, Choi G, Wolthuis EK, et al. The clinical
value of daily routine chest radiographs in a mixed
medical-surgical intensive care unit is low. Crit Care.
2006;10(1):R11.
16. Bouadma L, Luyt CE, Tubach F, et al; PRORATA
trial group. Use of procalcitonin to reduce patients’
exposure to antibiotics in intensive care units
(PRORATA trial): a multicentre randomised
controlled trial. Lancet. 2010;375(9713):463-474.
17. Bouman CS, Oudemans-Van Straaten HM,
Tijssen JG, Zandstra DF, Kesecioglu J. Effects of
early high-volume continuous venovenous
hemofiltration on survival and recovery of renal
JAMA Internal Medicine July 22, 2013 Volume 173, Number 14
Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013
1371
Clinical Review & Education Special Communication
31. Pronovost PJ, Angus DC, Dorman T, Robinson
KA, Dremsizov TT, Young TL. Physician staffing
patterns and clinical outcomes in critically ill
patients: a systematic review. JAMA. 2002;288(17):
2151-2162.
function in intensive care patients with acute renal
failure: a prospective, randomized trial. Crit Care
Med. 2002;30(10):2205-2211.
severe sepsis in the United States: analysis of
incidence, outcome, and associated costs of care.
Crit Care Med. 2001;29(7):1303-1310.
18. Kress JP, Pohlman AS, O’Connor MF, Hall JB.
Daily interruption of sedative infusions in critically ill
patients undergoing mechanical ventilation. N Engl
J Med. 2000;342(20):1471-1477.
25. Annane D, Sébille V, Charpentier C, et al. Effect
of treatment with low doses of hydrocortisone and
fludrocortisone on mortality in patients with septic
shock. JAMA. 2002;288(7):862-871.
19. Strøm T, Martinussen T, Toft P. A protocol of no
sedation for critically ill patients receiving
mechanical ventilation: a randomised trial. Lancet.
2010;375(9713):475-480.
26. Sprung CL, Annane D, Keh D, et al; CORTICUS
Study Group. Hydrocortisone therapy for patients
with septic shock. N Engl J Med. 2008;358(2):
111-124.
20. Krishnan JA, Parce PB, Martinez A, Diette GB,
Brower RG. Caloric intake in medical ICU patients:
consistency of care with guidelines and relationship
to clinical outcomes. Chest. 2003;124(1):297-305.
27. Casaer MP, Mesotten D, Hermans G, et al. Early
versus late parenteral nutrition in critically ill adults.
N Engl J Med. 2011;365(6):506-517.
33. Pearse R, Dawson D, Fawcett J, Rhodes A,
Grounds RM, Bennett ED. Early goal-directed
therapy after major surgery reduces complications
and duration of hospital stay: a randomised,
controlled trial [ISRCTN38797445]. Crit Care.
2005;9(6):R687-R693.
21. Arabi YM, Tamim HM, Dhar GS, et al. Permissive
underfeeding and intensive insulin therapy in
critically ill patients: a randomized controlled trial.
Am J Clin Nutr. 2011;93(3):569-577.
28. Vlasselaers D, Milants I, Desmet L, et al.
Intensive insulin therapy for patients in paediatric
intensive care: a prospective, randomised
controlled study. Lancet. 2009;373(9663):
547-556.
34. Rivers E, Nguyen B, Havstad S, et al; Early
Goal-Directed Therapy Collaborative Group. Early
goal-directed therapy in the treatment of severe
sepsis and septic shock. N Engl J Med.
2001;345(19):1368-1377.
22. National Heart, Lung, and Blood Institute Acute
Respiratory Distress Syndrome (ARDS) Clinical
Trials Network. Initial trophic vs full enteral feeding
in patients with acute lung injury: the EDEN
randomized trial. JAMA. 2012;307(8):795-803.
29. Schweickert WD, Pohlman MC, Pohlman AS,
et al. Early physical and occupational therapy in
mechanically ventilated, critically ill patients: a
randomised controlled trial. Lancet.
2009;373(9678):1874-1882.
35. Noah MA, Peek GJ, Finney SJ, et al. Referral to
an extracorporeal membrane oxygenation center
and mortality among patients with severe 2009
influenza A(H1N1). JAMA. 2011;306(15):1659-1668.
23. NICE-SUGAR Study Investigators. Intensive
versus conventional glucose control in critically ill
patients. N Engl J Med. 2009;360(13):1283-1297.
30. de Smet AM, Kluytmans JA, Cooper BS, et al.
Decontamination of the digestive tract and
oropharynx in ICU patients. N Engl J Med.
2009;360(1):20-31.
24. Angus DC, Linde-Zwirble WT, Lidicker J,
Clermont G, Carcillo J, Pinsky MR. Epidemiology of
1372
“Less Is More” in Critically Ill Patients
32. Cho SH, Hwang JH, Kim J. Nurse staffing and
patient mortality in intensive care units. Nurs Res.
2008;57(5):322-330.
36. Grasso S, Terragni P, Birocco A, et al. ECMO
criteria for influenza A (H1N1)-associated ARDS: role
of transpulmonary pressure. Intensive Care Med.
2012;38(3):395-403.
JAMA Internal Medicine July 22, 2013 Volume 173, Number 14
Downloaded From: http://archinte.jamanetwork.com/ by a University of California - Los Angeles User on 10/01/2013
jamainternalmedicine.com
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
original article
A Trial of Intraoperative Low-Tidal-Volume
Ventilation in Abdominal Surgery
Emmanuel Futier, M.D., Jean-Michel Constantin, M.D., Ph.D.,
Catherine Paugam-Burtz, M.D., Ph.D., Julien Pascal, M.D.,
Mathilde Eurin, M.D., Arthur Neuschwander, M.D., Emmanuel Marret, M.D.,
Marc Beaussier, M.D., Ph.D., Christophe Gutton, M.D., Jean-Yves Lefrant, M.D., Ph.D.,
Bernard Allaouchiche, M.D., Ph.D., Daniel Verzilli, M.D., Marc Leone, M.D., Ph.D.,
Audrey De Jong, M.D., Jean-Etienne Bazin, M.D., Ph.D., Bruno Pereira, Ph.D.,
and Samir Jaber, M.D., Ph.D., for the IMPROVE Study Group*
A BS T R AC T
From the Département d’Anesthésie et Ré­
animation, Hôpital Estaing (E.F., J.-M.C.,
J.P., J.-E.B.), Université de Clermont-Ferrand, Retinoids, Reproduction, and Developmental Diseases Unit, Équipe Accueil
7281 (E.F., J.-M.C.), and the Biostatistics
Unit, Direction de la Recherche Clinique
(B.P.), Centre Hospitalier Universitaire
(CHU) de Clermont-Ferrand, ClermontFerrand; Assistance Publique–Hôpitaux de
Paris (AP-HP), Département d’Anesthé­
sie et Réanimation, Hôpital Beaujon, Hôpitaux Universitaires Paris Nord Val de Seine
and Université Paris Diderot, Sorbonne
Paris Cité (C.P.-B., M.E., A.N.), Départe­
ment d’Anes­thé­sie et Réanimation, Hôpital Tenon (E.M.), and AP-HP, Départe­
ment d’Anes­thé­sie et Réanimation,
Hô­pi­tal Saint-Antoine (M.B., C.G.), Paris;
CHU de Nîmes, Section d’Anesthésie
and Dé­partement d’Anesthésie et Réanimation, Nîmes (J.-Y.L.); CHU de Lyon,
Département d’Anesthésie et Réanimation, Hôpi­tal Edouard Herriot, Lyon (B.A.);
CHU de Montpellier, Département d’Anes­
thésie et Réanimation B, Hôpital SaintEloi, and INSERM Unité 1046 and Université Montpellier 1, Montpellier (D.V., A.D.J.,
S.J.); and Assistance Publique–Hôpital
de Marseille, Département d’Anesthésie
et Réanimation, Hôpital Nord, Marseille
(M.L.) — all in France. Address reprint requests to Dr. Jaber at the Département
d’Anesthésie et Réanimation B (DAR B),
80 Ave. Augustin Fliche, 34295 Montpellier,
France, or at [email protected].
*Additional investigators in the Intraoperative Protective Ventilation (IMPROVE)
Study Group are listed in the Supplementary Appendix, available at NEJM.org.
N Engl J Med 2013;369:428-37.
DOI: 10.1056/NEJMoa1301082
Copyright © 2013 Massachusetts Medical Society.
428
BACKGROUND
Lung-protective ventilation with the use of low tidal volumes and positive endexpiratory pressure is considered best practice in the care of many critically ill
patients. However, its role in anesthetized patients undergoing major surgery is
not known.
METHODS
In this multicenter, double-blind, parallel-group trial, we randomly assigned 400
adults at intermediate to high risk of pulmonary complications after major abdominal surgery to either nonprotective mechanical ventilation or a strategy of
lung-protective ventilation. The primary outcome was a composite of major pulmonary and extrapulmonary complications occurring within the first 7 days after
surgery.
RESULTS
The two intervention groups had similar characteristics at baseline. In the intention-to-treat analysis, the primary outcome occurred in 21 of 200 patients (10.5%)
assigned to lung-protective ventilation, as compared with 55 of 200 (27.5%) assigned
to nonprotective ventilation (relative risk, 0.40; 95% confidence interval [CI], 0.24
to 0.68; P = 0.001). Over the 7-day postoperative period, 10 patients (5.0%) assigned
to lung-protective ventilation required noninvasive ventilation or intubation for
acute respiratory failure, as compared with 34 (17.0%) assigned to nonprotective
ventilation (relative risk, 0.29; 95% CI, 0.14 to 0.61; P = 0.001). The length of the hospital stay was shorter among patients receiving lung-protective ventilation than among
those receiving nonprotective ventilation (mean difference, −2.45 days; 95% CI,
−4.17 to −0.72; P = 0.006).
CONCLUSIONS
As compared with a practice of nonprotective mechanical ventilation, the use of
a lung-protective ventilation strategy in intermediate-risk and high-risk patients
undergoing major abdominal surgery was associated with improved clinical outcomes and reduced health care utilization. (IMPROVE ClinicalTrials.gov number,
NCT01282996.)
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Intr aoper ative Low-Tidal-Volume Ventilation
W
orldwide, more than 230 million
patients undergoing major surgery each
year require general anesthesia and
mechanical ventilation.1 Postoperative pulmonary complications adversely affect clinical outcomes and health care utilization,2 so prevention of these complications has become a
measure of the quality of hospital care.3 Previous, large cohort studies have shown that 20 to
30% of patients undergoing surgery with general anesthesia are at intermediate to high risk
for postoperative pulmonary complications.4,5
Mechanical ventilation with the use of high
tidal volumes (10 to 15 ml per kilogram of predicted body weight) has traditionally been recommended to prevent hypoxemia and atelectasis in anesthetized patients.6 There is, however,
considerable evidence from experimental and
observational studies that mechanical ventilation — in particular, high tidal volumes that
cause alveolar overstretching — can initiate
ventilator-associated lung injury7 and contribute to extrapulmonary organ dysfunction
through systemic release of inflammatory mediators.8,9
Lung-protective ventilation, which refers to
the use of low tidal volumes and positive endexpiratory pressure (PEEP), and which may also
include the use of recruitment maneuvers (periodic hyperinflation of the lungs),10 has been
shown to reduce mortality among patients with
the acute respiratory distress syndrome11 and is
now considered best practice in the care of
many critically ill patients.12 Although this approach may be beneficial in a broader population,13,14 some physicians have questioned the
benefits of using lung-protective ventilation in
the surgical setting,15-18 especially since the use
of high tidal volumes and no PEEP is still commonplace and less than 20% of patients receive
protective ventilation in routine anesthetic
practice.19,20
We conducted the Intraoperative Protective
Ventilation (IMPROVE) trial to determine whether a multifaceted strategy of prophylactic lungprotective ventilation that combined low tidal
volumes, PEEP, and recruitment maneuvers
could improve outcomes after abdominal surgery, as compared with the standard practice of
nonprotective mechanical ventilation.
ME THODS
TRIAL DESIGN AND OVERSIGHT
The IMPROVE trial was an investigator-initiated,
multicenter, double-blind, stratified, parallel-group,
clinical trial. Randomization was performed with
the use of a computer-generated assignment sequence and a centralized telephone system. The
study protocol and statistical analysis plan were
approved for all centers by a central ethics committee (Comité de Protection des Personnes Sud-Est I,
Saint-Etienne, France) according to French law.
The protocol, including the statistical analysis
plan, is available with the full text of this article at
NEJM.org. An independent data and safety monitoring committee oversaw the study conduct and
reviewed blinded safety data. The members of
the steering committee (see the Supplementary
Appendix, available at NEJM.org) vouch for the
accuracy and completeness of the data and analyses and the fidelity of the study to the protocol.
There was no industry support or involvement in
the trial.
Patients were screened and underwent randomization between January 31, 2011, and August 10,
2012, at seven French university teaching hospitals.
Written informed consent was obtained before
randomization from each patient, on the day before surgery. Randomization was stratified according to study site and the planned use or nonuse of
postoperative epidural analgesia, which is a factor
that may influence outcomes.21 Treatment assignments were concealed from patients, research staff,
the statistician, and the data and safety monitoring committee. Although the staff members who
collected data during surgery were aware of the
group assignments, outcome assessors were unaware of these assignments throughout the study.
PATIENTS
Patients were eligible for participation in the
study if they were older than 40 years of age,
were scheduled to undergo laparoscopic or nonlaparoscopic elective major abdominal surgery1
with an expected duration of at least 2 hours, and
had a preoperative risk index for pulmonary
complications5 of more than 2. The risk index
uses risk classes that range from 1 to 5, with
higher risk classes indicating a higher risk of
postoperative pulmonary complications (see the
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
429
The
n e w e ng l a n d j o u r na l
Supplementary Appendix). Patients were ineligible
if they had received mechanical ventilation within
the 2 weeks preceding surgery, had a body-mass
index (the weight in kilograms divided by the
square of the height in meters) of 35 or higher,
had a history of respiratory failure or sepsis within the 2 weeks preceding surgery, had a requirement for intrathoracic or emergency surgery, or
had a progressive neuromuscular illness.
INTERVENTIONS
Patients were assigned to receive volume-controlled mechanical ventilation according to one
of two strategies: nonprotective ventilation with
a tidal volume of 10 to 12 ml per kilogram of
predicted body weight, with no PEEP and no recruitment maneuvers, as previously described20
(the nonprotective-ventilation group), or lungprotective ventilation with a tidal volume of 6 to
8 ml per kilogram of predicted body weight, a
PEEP of 6 to 8 cm of water, and recruitment maneuvers repeated every 30 minutes after tracheal
intubation (the protective-ventilation group). Each
recruitment maneuver consisted of applying a
continuous positive airway pressure of 30 cm of
water for 30 seconds. During anesthesia, a plateau
pressure of no more than 30 cm of water was
targeted in each group. All other ventilation procedures were identical in the two study groups
(see the Supplementary Appendix).
The predicted body weight was calculated for
each patient with the use of previously defined
formulas.11 For episodes of arterial desaturation
(defined as a peripheral oxygen saturation of
≤92%), a transient increase in the fraction of inspired oxygen (Fio2) to 100% was permitted, and
in patients assigned to nonprotective ventilation,
the use of PEEP, recruitment maneuvers, or both
was allowed, if required. Decisions about all
other aspects of patient care during the intraoperative and postoperative periods, including general anesthesia, administration of fluids, use of
prophylactic antibiotic agents, and postoperative
pain management, were made by the attending
physician according to the expertise of the staff
at each center and routine clinical practice.
OUTCOMES
The primary outcome was a composite of major
pulmonary and extrapulmonary complications
occurring by day 7 after surgery. Major pulmonary complications were defined as pneumonia
430
of
m e dic i n e
(defined according to standard criteria; see the
Supplementary Appendix) or the need for invasive
or noninvasive ventilation for acute respiratory failure. Major extrapulmonary complications were defined as sepsis, severe sepsis and septic shock (defined according to consensus criteria),22 or death.
Secondary outcomes within the 30-day follow-up period were the incidence of pulmonary
complications due to any cause, graded on a scale
from 0 (no pulmonary complications) to 4 (the
most severe complications)23 (see the Supplementary Appendix); ventilation-related adverse events
during surgery; postoperative gas exchange; unexpected need for admission to the intensive
care unit (ICU); extrapulmonary complications;
durations of ICU and hospital stays; and the rate
of death from any cause 30 days after surgery.
Pulmonary complications were analyzed separately; in particular, the need for invasive or
noninvasive ventilation because of acute respiratory failure, the development of postoperative
atelectasis, pneumonia, acute lung injury, and the
acute respiratory distress syndrome, defined according to standard criteria (see the Supplementary
Appendix). Extrapulmonary complications in­cluded
the systemic inflammatory response syndrome
(SIRS); sepsis; severe sepsis and septic shock; and
surgical complications, including intraabdominal abscess, anastomotic leakage, and unplanned
reoperation (all defined according to consensus
criteria22,24).
STATISTICAL ANALYSIS
We calculated that a sample of 400 patients
would provide 80% power to detect a relative difference of 50% in the primary outcome, at a twosided alpha level of 0.05, assuming a 20% rate of
postoperative complications in the nonprotectiveventilation group.25 For safety reasons, an interim
analysis was conducted after the enrollment of
the first 200 patients, according to the a priori
statistical analysis plan. The data and safety monitoring committee did not recommend discontinuation of the trial on the basis of that analysis,
and 400 patients were therefore included. A total
of 3 patients were excluded after randomization;
surgery was stopped prematurely in 2 of the 3 patients because of extensive illness (duration of surgery, <2 hours), and 1 had undergone randomization in error (violation of exclusion criteria).
An additional 3 patients were thus randomly assigned to a study group to obtain the full sample.
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Intr aoper ative Low-Tidal-Volume Ventilation
All analyses were conducted on data from the
modified intention-to-treat population, which included all patients who underwent randomization
except the three who were excluded (Fig. 1). An un­
adjusted chi-square test was used for the primary
outcome analysis. Multiple logistic-regression analysis was used to identify relevant baseline covariates
associated with the primary outcome, in addition
to the stratification variables (use or nonuse of
epidural analgesia and study center). Variables
tested in the model were selected if the P value
was less than 0.10 and if they were clinically relevant. Adjusted analyses were performed with the
use of robust Poisson generalized-linear-model re­
gression26 and are presented as relative risks with
95% confidence intervals. A chi-square test (or
Fisher’s exact test, as appropriate) was used for
secondary binary outcomes. The Hochberg procedure was used to adjust for multiple testing of
components of the composite primary outcome.27
Adjusted analyses were performed with the use
of the same adjustment variables that were used
in the robust Poisson regression analysis. Continuous variables were compared with the use of
an unpaired t-test or the Mann–Whitney U test.
Adjusted analyses were performed with the use
1803 Patients were scheduled to undergo
elective abdominal surgery
1202 Were ineligible
47 Were ≤40 yr of age
102 Had expected duration of surgery of <2 hr
1053 Had preoperative risk index for pulmonary
complications of <2
601 Underwent screening
198 Were excluded
49 Declined to participate
149 Met exclusion criteria
400 Underwent randomization
3 Were excluded after randomization
2 Had extensive illness (surgery <2 hr)
1 Had violation of exclusion criteria (age ≤40 yr)
An additional 3 underwent randomization
200 Were assigned to receive nonprotective
mechanical ventilation
200 Were assigned to receive lung-protective
mechanical ventilation
200 (100%) Were included in 30-day analysis
200 (100%) Were included in 30-day analysis
Figure 1. Assessment, Randomization, and Follow-up of Patients.
A total of 1803 patients awaiting abdominal surgery were assessed preoperatively for trial eligibility by the research
staff. A total of 400 patients were included in the modified intention-to-treat analysis and were followed for 30 days
after surgery. After randomization, 3 patients were excluded; 2 patients were excluded because surgery was stopped
prematurely (duration, <2 hours), owing to extensive illness, and 1 had undergone randomization in error. An additional 3 patients were then enrolled in the study.
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
431
The
n e w e ng l a n d j o u r na l
Table 1. Baseline Characteristics of the Patients.*
Nonprotective
Ventilation
(N = 200)
Characteristic
Lung-Protective
Ventilation
(N = 200)
Age — yr
63.4±10.0
61.6±11.0
Male sex — no. (%)
121 (60.5)
116 (58.0)
Height — cm
169.5±9.0
169.1±8.8
Actual
71.3±13.9
71.4±14.2
Predicted†
63.8±9.9
63.3±9.7
24.7±3.8
24.8±3.8
Body weight — kg
Body-mass index‡
Mean
25–35 — no. (%)
88 (44.0)
99 (49.5)
Risk class 2
100 (50.0)
101 (50.5)
Risk class 3
94 (47.0)
93 (46.5)
6 (3.0)
6 (3.0)
Current smoking
50 (25.0)
51 (25.5)
Any alcohol intake
10 (5.0)
21 (10.5)
8 (4.0)
8 (4.0)
Chronic obstructive pulmonary
­disease
20 (10.0)
20 (10.0)
Loss of >10% of body weight in
previous 6 mo
44 (22.0)
40 (20.0)
4 (2.0)
7 (3.5)
44 (22.0)
41 (20.5)
Pancreaticoduodenectomy
80 (40.0)
84 (42.0)
Liver resection
52 (26.0)
44 (22.0)
Gastrectomy
17 (8.5)
15 (7.5)
Colorectal resection
40 (20.0)
47 (23.5)
Other procedure
11 (5.5)
10 (5.0)
164 (82.0)
155 (77.5)
36 (18.0)
45 (22.5)
Preoperative risk index — no. (%)§
Risk class 4 or 5
Coexisting condition — no. (%)¶
Not fully independent in activities
of daily living
Long-term glucocorticoid use
Laparoscopic surgery — no. (%)
Type of surgery — no. (%)
Diagnosis — no. (%)
Cancer
Diagnosis other than cancer
*Plus–minus values are means ±SD. There were no significant differences between
the two groups (P>0.05).
†The predicted body weight was calculated as follows: for men, 50+0.91(height in
centimeters − 152.4); and for women, 45.5 + 0.91(height in centimeters − 152.4).11
‡The body-mass index is the weight in kilograms divided by the square of the
height in meters.
§ The preoperative risk index for pulmonary complications5 uses risk classes
that range from 1 to 5, with higher risk classes indicating a higher risk of
postoperative complications. Patients with a risk class of 2 or more were eligible for participation in the study.
¶All factors listed as coexisting conditions were included in the preoperative
risk index as predictors of postoperative pulmonary complications.
432
of
m e dic i n e
of the same adjustment variables that were used
in the linear-regression model. The time-to-event
curves were calculated with the Kaplan–Meier
method. Details regarding the handling of missing
data are provided in the Supplementary Appendix.
All analyses were conducted with the use of
Stata software, version 12 (StataCorp). A two-sided
P value of less than 0.05 was considered to indicate
statistical significance.
R E SULT S
STUDY POPULATION
From January 2011 through August 2012, a total
of 1803 patients awaiting abdominal surgery were
assessed for trial eligibility. A total of 400 patients were included in the modified intentionto-treat analysis and were followed for 30 days
after surgery (Fig. 1). One patient in the nonprotective-ventilation group received lung-protective
ventilation but was included in the analysis for
the group to which he was assigned. Data on the
primary outcome were available for all patients.
Baseline characteristics were similar between the
two groups (Table 1). Open laparotomy, mainly
for cancer resection, was performed in 156 patients
(78.0%) in the nonprotective-ventilation group and
in 159 (79.5%) in the protective-ventilation group
(P = 0.80).
INTRAOPERATIVE PROCEDURES
Table 2 shows the distribution of the main intraoperative procedures. Mean (±SD) tidal volumes
were 11.1±1.1 ml per kilogram in the nonprotective-ventilation group, as compared with 6.4±0.8
ml per kilogram in the protective-ventilation
group (P<0.001), and values remained within target ranges throughout the intraoperative period.
In the protective-ventilation group, the median
PEEP was 6 cm of water (interquartile range, 6 to 8),
and the median number of recruitment maneuvers was 9 (interquartile range, 6 to 12); in the
nonprotective-ventilation group, the value for each
of these measures was 0 (interquartile range, 0 to
0) (Table 2). There were no significant betweengroup differences in type and duration of surgery, use or nonuse of epidural analgesia, blood
loss, volume of fluids administered, and need for
vasopressor administration. Five patients in the
nonprotective-ventilation group required at least
one intraoperative rescue therapy for arterial de-
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Intr aoper ative Low-Tidal-Volume Ventilation
Table 2. Intraoperative Procedures.*
Variable
Tidal volume — ml
Tidal volume — ml/kg of predicted body weight
Nonprotective
Ventilation
(N = 200)
Lung-Protective
Ventilation
(N = 200)
P Value
719.0±127.8
406.7±75.6
<0.001
11.1±1.1
6.4±0.8
<0.001
PEEP — cm of water
Baseline
<0.001
Median
Interquartile range
0
6
0–0
6–8
End of surgery
<0.001
Median
Interquartile range
0
6
0–0
6–8
No. of recruitment maneuvers
<0.001
Median
0
9
0–0
6–12
Baseline
20.1±4.9
18.9±3.6
0.04
End of surgery
20.6±4.4
20.0±4.0
0.15
Baseline
16.1±4.3
15.2±3.0
0.02
End of surgery
16.6±3.5
15.2±2.6
<0.001
Baseline
48.4±17.8
55.2±26.6
0.06
End of surgery
45.1±12.9
55.2±26.7
<0.001
47.2±7.6
46.4±7.3
0.27
Interquartile range
Peak pressure — cm of water
Plateau pressure — cm of water
Respiratory system compliance — ml/cm of water
Fio2 — %
Volume of fluids administered — liters
Crystalloid
0.47
Median
Interquartile range
2.0
1.5
1.5–3.5
2.0–3.0
Colloid
0.97
Median
Interquartile range
0.5
0.5
0.25–1.0
0.50–1.0
Duration of surgery — no./total no. (%)†
0.95
2–4 hr
76/192 (39.6)
75/195 (38.5)
>4–6 hr
75/192 (39.1)
76/195 (39.0)
>6 hr
41/192 (21.4)
44/195 (22.6)
Duration of mechanical ventilation — min
344±127.9
319±139.4
0.84
Epidural analgesia — no. (%)
77 (38.5)
83 (41.5)
0.61
*Plus–minus values are means ±SD. Detailed data on intraoperative procedures are given in Table S2 in the Supplementary
Appendix. Fio2 denotes inspired oxygen fraction, and PEEP positive end-expiratory pressure.
†The duration of surgery was calculated as the time between skin incision and closure of the incision.
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
433
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
saturation (PEEP in one patient, recruitment ma- ventilation (P<0.001 by the log-rank test) (Fig. S1
neuvers in two, and both in two), as compared in the Supplementary Appendix).
with no patients in the protective-ventilation
There was no significant difference between
group (P = 0.06).
the protective-ventilation group and the nonprotective-ventilation group with respect to the proOUTCOMES
portion of patients who were unexpectedly adPrimary Outcome
mitted to the ICU during the 30-day period after
Major pulmonary and extrapulmonary complica- surgery (11.0 and 12.5%, respectively; adjusted relations occurred within the first 7 days after sur- tive risk with protective ventilation, 0.88; 95% CI,
gery in 21 patients (10.5%) in the protective-ven- 0.49 to 1.59; P = 0.67), nor was there a significant
tilation group, as compared with 55 (27.5%) in the between-group difference in the rate of adverse
nonprotective-ventilation group (adjusted relative events (Table S3 in the Supplementary Appendix).
risk, 0.40; 95% confidence interval [CI], 0.24 to Mortality at 30 days in the protective-ventilation
0.68; P = 0.001) (Table 3). The results of associated group was similar to that in the nonprotectiveunivariate and multivariate analyses are provided ventilation group (3.0% and 3.5%, respectively;
in Table S1 in the Supplementary Appendix.
adjusted relative risk with protective ventilation,
1.13; 95% CI, 0.36 to 3.61; P = 0.83). However, the
median hospital stay was shorter in the protecSecondary Outcomes
One or more pulmonary complications developed tive-ventilation group than in the nonprotectivewithin the first 7 days after surgery in 35 patients ventilation group (Table 3).
(17.5%) in the protective-ventilation group, as compared with 72 (36.0%) in the nonprotective-ventilaDISCUSSION
tion group (adjusted relative risk, 0.49; 95% CI,
0.32 to 0.74; P<0.001). More patients in the non- In this trial, intraoperative lung-protective meprotective-ventilation group than in the protec- chanical ventilation, as compared with non­
tive-ventilation group had major (grade ≥3) pul- protective ventilation, led to improved clinical
monary complications (Table 3, and Tables S3 and outcomes and reduced health care utilization
S4 in the Supplementary Appendix) and major after abdominal surgery. The observed rate of
pulmonary and extrapulmonary complications dur- postoperative complications in our study was
ing the 30 days after surgery (P<0.001 by the log- slightly higher than predicted.25 This was due,
rank test) (Fig. 2). There were no relevant between- in part, to the exclusion of patients with a low
group differences in gas exchange after extubation risk of complications, as well as the large proand on day 1 after surgery (Table S5 in the Supple- portion of patients who underwent major abmentary Appendix).
dominal procedures, which are associated with
The proportion of patients who required post- increased morbidity rates. Of the 400 patients
operative ventilatory assistance (noninvasive ven- en­rolled, 19 had postoperative pneumonia and
tilation or intubation) for acute respiratory failure 47 had respiratory failure requiring intubation
was lower in the protective-ventilation group than or noninvasive ventilation. These rates are conin the nonprotective-ventilation group during the sistent with previously reported rates of pulmofirst 7 days after surgery (10 of 200 patients [5.0%], nary complications25,28 and mortality.29 Our lungvs. 34 of 200 [17.0%]; adjusted relative risk, 0.29; protective ventilation strategy resulted in a 69%
95% CI, 0.14 to 0.61; P = 0.001), and the proportion reduction in the number of patients requiring
was also lower with protective ventilation during ventilatory support within the first 7 days after
the first 13 days after surgery (6.5% vs. 18.5%; surgery.
adjusted relative risk, 0.36; 95% CI, 0.19 to 0.70;
Several hypotheses could explain some of
P = 0.003) (Table 3). In addition, the cumulative the differences between the results of the pres30-day probability of an event requiring intuba- ent study and findings in other trials of lungtion or noninvasive ventilation for postoperative protective ventilation during high-risk surgery.
acute respiratory failure was lower among pa- Previous trials have included small numbers of
tients who received lung-protective ventilation patients, have focused on different (and not
than among those who received nonprotective necessarily clinically relevant) outcomes,17 and
434
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Intr aoper ative Low-Tidal-Volume Ventilation
Table 3. Results of Unadjusted and Adjusted Outcome Analyses.*
Nonprotective Lung-Protective
Ventilation
Ventilation
(N = 200)
(N = 200)
Variable
Unadjusted
Relative Risk or
­Between-Group
­Difference
(95% CI)
P Value†
Adjusted
Relative Risk or
­Between-Group
­Difference
(95% CI)‡
P Value
Primary composite outcome — no. (%)
Within 7 days§
55 (27.5)
21 (10.5)
0.38 (0.24–0.61)
<0.001
0.40 (0.24–0.68)
0.001
Within 30 days
58 (29.0)
25 (12.5)
0.43 (0.28–0.66)
<0.001
0.45 (0.28–0.73)
<0.001
Grade 1 or 2
30 (15.0)
25 (12.5)
0.69 (0.42–1.13)
0.14
0.67 (0.39–1.16)
0.16
Grade ≥3
42 (21.0)
10 (5.0)
0.24 (0.12–0.46)
<0.001
0.23 (0.11–0.49)
<0.001
Atelectasis within 7 days‖
34 (17.0)
13 (6.5)
0.38 (0.21–0.70)
0.001
0.37 (0.19–0.73)
0.004
Pneumonia within 7 days
16 (8.0)
3 (1.5)
0.19 (0.05–0.63)
0.01
0.19 (0.05–0.66)
0.009
6 (3.0)
1 (0.5)
0.17 (0.02–1.37)
0.12
0.21 (0.02–1.71)
0.14
7 (3.5)
2 (1.0)
0.29 (0.06–1.36)
0.51
0.40 (0.08–1.97)
0.26
29 (14.5)
9 (4.5)
0.31 (0.15–0.64)
0.006
0.29 (0.13–0.65)
0.002
100 (50.0)
86 (43.0)
0.86 (0.70–1.06)
0.16
0.87 (0.65–1.17)
0.37
29 (14.5)
13 (6.5)
0.45 (0.24–0.84)
0.04
0.48 (0.25–0.93)
0.03
9 (4.5)
8 (4.0)
0.89 (0.35–2.26)
0.80
1.48 (0.51–4.32)
0.47
7 (3.5)
6 (3.0)
0.86 (0.29–2.51)
0.80
1.13 (0.36–3.61)
0.83
13
11
8–20
8–15
Secondary outcomes — no. (%)
Pulmonary complication within 7 days¶
Acute lung injury or ARDS within 7 days
Need for ventilation within 7 days
Invasive
Noninvasive
Extrapulmonary complication within 7 days
SIRS
Sepsis
Severe sepsis or septic shock
Death within 30 days
Duration of stay in hospital and ICU — days
Hospital
Median
Interquartile range
0.02
−2.25 (−4.04 to −0.47)
ICU
Median
Interquartile range
0.006
−2.45 (−4.17 to −0.72)
0.58
7
6
4–9
4–8
−1.48 (−6.87 to 3.91)
0.69
−1.21 (−4.98 to 7.40)
*All postoperative complications were defined according to consensus criteria (see the Supplementary Appendix). For additional data on
postoperative outcomes, see Tables S3 and S4 in the Supplementary Appendix. ARDS denotes acute respiratory distress syndrome, CI confidence interval, ICU intensive care unit, and SIRS systemic inflammatory response syndrome. Relative risks are shown for outcome variables,
and differences between groups are shown for the duration of stays in the hospital and ICU.
†Adjustment was performed for stratification variables (use or nonuse of epidural analgesia and study center), preoperative risk index for
postoperative pulmonary complications, sex, duration of surgery, and need for blood transfusion (yes or no).
‡The Hochberg procedure was used to adjust for multiple testing of components of the composite primary outcome.27
§ The primary outcome was a composite of major pulmonary complications (defined as pneumonia or need for invasive or noninvasive ventilation
for acute respiratory failure) and extrapulmonary complications (defined as sepsis, septic shock, or death) within the first 7 days after surgery.
¶Postoperative pulmonary complications were scored with the use of a graded scale23 from 0 (no pulmonary complications) to 4 (the most
severe complications) (see the Supplementary Appendix).
‖Atelectasis was defined as opacification of the lung with shift of the mediastinum, hilum, or hemidiaphragm toward the affected area and
compensatory overinflation in the adjacent, nonatelectatic lung.
have used either very low levels of PEEP or no
PEEP.15,16,18 One strength of the present trial is our
use of a robust composite outcome that is highly
pertinent to this high-risk surgical population.5
Mechanical ventilation itself can induce an inflammatory response30 and can synergize with
the response induced by major surgery at both
local and systemic levels. This amplification of
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
435
The
n e w e ng l a n d j o u r na l
0.50
Probability of Event
0.40
Nonprotective ventilation
0.30
0.20
Lung-protective ventilation
0.10
0.00
1
3
7
15
30
Days since Randomization
No. at Risk
Nonprotective
ventilation
Lung-protective
ventilation
200 182 163
145
142
142
200 192 184
179
176
175
Figure 2. Kaplan–Meier Estimates of the Probability of the Composite
­Primary Outcome.
Data for the Kaplan–Meier estimates of the probability of the composite
primary outcome of major pulmonary or extrapulmonary complications
were censored at 30 days after surgery. Major pulmonary complications
­included pneumonia or the need for invasive or noninvasive ventilation for
acute respiratory failure. Major extrapulmonary complications were ­sepsis,
severe sepsis, septic shock, and death. P<0.001 by the log-rank test for the
between-group difference in the probability of the primary outcome.
the inflammation cascade contributes to the
subsequent development of lung injury31 and sys­
temic organ failure.8,32
The use of very low levels of PEEP in previous
trials may have promoted the repeated opening
and closing of small airways, leading to atelectasis, which can precipitate the development of
pulmonary complications.6,33 We used a multifaceted strategy of lung-protective ventilation
that combined low tidal volumes, recruitment
maneuvers to open collapsed alveoli, and moderate levels of PEEP to prevent further collapse.34
Other strengths of the present trial include the
methods used to minimize bias (blinded and
centralized randomization, complete follow-up,
and intention-to-treat analyses); the pragmatic
nature of the trial protocol, with routine practice
being maintained; and the enrollment of patients with characteristics similar to those of
patients enrolled in other studies analyzing outcomes after major surgery.29
Our findings are consistent with the observation of transient arterial hypotension during recruitment maneuvers.35 Consequently, recruit436
of
m e dic i n e
ment maneuvers, in which hemodynamic effects
are potentially influenced by the applied level of
alveolar pressure,36 should be used with caution
in patients with hemodynamic instability.
There are several limitations to our study. The
trial design did not include standardization of
the administration of fluids. However, this limitation is unlikely to have affected our results,
since the volume of fluids administered was
similar in the two groups. The definition of
nonprotective ventilation was arbitrary but is
supported in the literature.19,20 The trial protocol
did not include standardization of requirements
for noninvasive ventilation; however, it was recommended that the study centers follow clinicalpractice guidelines,37,38 and postoperative care
was conducted by health care workers who were
unaware of the study assignments. The utilization of noninvasive ventilation in our trial is
close to that reported in earlier studies.37 We
therefore consider it unlikely that any imbalance
in interventions affected our results.
In conclusion, our study provides evidence that
a multifaceted strategy of prophylactic lung-protective ventilation during surgery, as compared
with a practice of nonprotective mechanical ventilation, results in fewer postoperative complications and reduced health care utilization.
Dr. Futier reports receiving consulting fees from General
Electric Medical Systems, lecture fees from Fresenius Kabi, and
reimbursement of travel expenses from Fisher and Paykel
Healthcare. Dr Constantin reports receiving consulting fees
from Baxter, Fresenius Kabi, Dräger, and General Electric Medical Systems, payment for expert testimony from Baxter, Dräger,
and Fresenius Kabi, lecture fees from General Electric Medical
Systems, Dräger, Fresenius Kabi, Baxter, Hospal, Merck Sharp &
Dohme, and LFB Biomedicaments, payment for the development
of educational presentations from Dräger, General Electric
Medical Systems, Baxter, and Fresenius Kabi, and reimbursement of travel expenses from Bird, Astute Medical, Astellas,
Fresenius Kabi, Baxter, and Hospal. Dr. Paugam-Burtz reports
receiving consulting fees from Fresenius Kabi, lecture fees and
reimbursement of travel expenses from Astellas, and payment
for the development of educational presentations from LFB Biomedicaments and Merck Sharp & Dohme. Dr. Allaouchiche reports receiving consulting fees from Fresenius Kabi and lecture
fees from Novartis and Astellas. Dr. Leone reports receiving
consulting fees from LFB Biomedicaments and lecture fees from
Fresenius Kabi and Novartis. Dr. Jaber reports receiving consulting fees from Dräger France and Maquet France, lecture fees
from Fisher and Paykel Healthcare, Abbott, and Philips, and reimbursement of travel expenses from Pfizer. No other potential
conflict of interest relevant to this article was reported.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
We thank all the patients who participated in the study; the
clinical and research staff at all the trial sites, without whose assistance the study would never have been completed; and Mervyn
Singer for valuable advice during the preparation of the manuscript.
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Intr aoper ative Low-Tidal-Volume Ventilation
References
1. Weiser TG, Regenbogen SE, Thomp-
son KD, et al. An estimation of the global
volume of surgery: a modelling strategy
based on available data. Lancet 2008;372:
139-44.
2. Khuri SF, Henderson WG, DePalma
RG, et al. Determinants of long-term survival after major surgery and the adverse
effect of postoperative complications. Ann
Surg 2005;242:326-41.
3. Shander A, Fleisher LA, Barie PS, Bigtello LM, Sladen RN, Watson CB. Clinical
and economic burden of postoperative
pulmonary complications: patient safety
summit on definition, risk-reducing interventions, and preventive strategies. Crit
Care Med 2011;39:2163-72.
4. Arozullah AM, Daley J, Henderson
WG, Khuri SF. Multifactorial risk index
for predicting postoperative respiratory
failure in men after major noncardiac surgery: the National Veterans Administration Surgical Quality Improvement Program. Ann Surg 2000;232:242-53.
5. Arozullah AM, Khuri SF, Henderson
WG, Daley J. Development and validation
of a multifactorial risk index for predicting postoperative pneumonia after major
noncardiac surgery. Ann Intern Med
2001;135:847-57.
6. Bendixen HH, Hedley-Whyte J, Laver
MB. Impaired oxygenation in surgical patients during general anesthesia with con­
trolled ventilation: a concept of atelectasis. N Engl J Med 1963;269:991-6.
7. Serpa Neto A, Cardoso SO, Manetta
JA, et al. Association between use of lungprotective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress
syndrome: a meta-analysis. JAMA 2012;
308:1651-9.
8. Imai Y, Parodo J, Kajikawa O, et al.
Injurious mechanical ventilation and endorgan epithelial cell apoptosis and organ
dysfunction in an experimental model of
acute respiratory distress syndrome.
JAMA 2003;289:2104-12.
9. Lellouche F, Dionne S, Simard S, Bussieres J, Dagenais F. High tidal volumes in
mechanically ventilated patients increase
organ dysfunction after cardiac surgery.
Anesthesiology 2012;116:1072-82.
10. Amato MB, Barbas CS, Medeiros DM,
et al. Effect of a protective-ventilation
strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;
338:347-54.
11. The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional
tidal volumes for acute lung injury and the
acute respiratory distress syndrome. N Engl
J Med 2000;342:1301-8.
12. Putensen C, Theuerkauf N, Zinserling
J, Wrigge H, Pelosi P. Meta-analysis: ven-
tilation strategies and outcomes of the
acute respiratory distress syndrome and
acute lung injury. Ann Intern Med 2009;
151:566-76. [Erratum, Ann Intern Med
2009;151:897.]
13. Schultz MJ, Haitsma JJ, Slutsky AS,
Gajic O. What tidal volumes should be
used in patients without acute lung injury?
Anesthesiology 2007;106:1226-31.
14. Severgnini P, Selmo G, Lanza C, et al.
Protective mechanical ventilation during
general anesthesia for open abdominal
surgery improves postoperative pulmonary
function. Anesthesiology 2013 March 29
(Epub ahead of print).
15. Wrigge H, Uhlig U, Zinserling J, et al.
The effects of different ventilatory settings on pulmonary and systemic inflammatory responses during major surgery.
Anesth Analg 2004;98:775-81.
16. Wrigge H, Uhlig U, Baumgarten G,
et al. Mechanical ventilation strategies and
inflammatory responses to cardiac surgery: a prospective randomized clinical
trial. Intensive Care Med 2005;31:1379-87.
17. Hong CM, Xu DZ, Lu Q, et al. Low
tidal volume and high positive end-expiratory pressure mechanical ventilation results in increased inflammation and ventilator-associated lung injury in normal
lungs. Anesth Analg 2010;110:1652-60.
18. Treschan TA, Kaisers W, Schaefer MS,
et al. Ventilation with low tidal volumes
during upper abdominal surgery does not
improve postoperative lung function. Br J
Anaesth 2012;109:263-71.
19. Jaber S, Coisel Y, Chanques G, et al. A
multicentre observational study of intraoperative ventilatory management during
general anaesthesia: tidal volumes and relation to body weight. Anaesthesia 2012;
67:999-1008.
20. Hess DR, Kondili D, Burns E, Bittner
EA, Schmidt UH. A 5-year observational
study of lung-protective ventilation in the
operating room: a single-center experience. J Crit Care 2013 January 29 (Epub
ahead of print).
21. Pöpping DM, Elia N, Marret E, Remy
C, Tramèr MR. Protective effects of epidural analgesia on pulmonary complications
after abdominal and thoracic surgery: a
meta-analysis. Arch Surg 2008;143:990-9.
22. Bone RC, Sibbald WJ, Sprung CL. The
ACCP-SCCM consensus conference on
sepsis and organ failure. Chest 1992;101:
1481-3.
23. Hulzebos EH, Helders PJ, Favié NJ,
et al. Preoperative intensive inspiratory muscle training to prevent postoperative pulmonary complications in high-risk patients
undergoing CABG surgery: a randomized
clinical trial. JAMA 2006;296:1851-7.
24. Dindo D, Demartines N, Clavien PA.
Classification of surgical complications: a
new proposal with evaluation in a cohort
of 6336 patients and results of a survey.
Ann Surg 2004;240:205-13.
25. Thompson JS, Baxter BT, Allison JG,
Johnson FE, Lee KK, Park WY. Temporal
patterns of postoperative complications.
Arch Surg 2003;138:596-602.
26. Zou G. A modified Poisson regression
approach to prospective studies with binary data. Am J Epidemiol 2004;159:702-6.
27. Hochberg Y. A sharper Bonferroni
procedure for multiple tests of significance. Biometrika 1988;75:800-2.
28. Lawrence VA, Dhanda R, Hilsenbeck
SG, Page CP. Risk of pulmonary complications after elective abdominal surgery.
Chest 1996;110:744-50.
29. Pearse RM, Moreno RP, Bauer P, et al.
Mortality after surgery in Europe: a 7 day
cohort study. Lancet 2012;380:1059-65.
30. Ranieri VM, Suter PM, Tortorella C,
et al. Effect of mechanical ventilation on
inflammatory mediators in patients with
acute respiratory distress syndrome: a
randomized controlled trial. JAMA 1999;
282:54-61.
31. Mascia L, Pasero D, Slutsky AS, et al.
Effect of a lung protective strategy for
­organ donors on eligibility and availability of lungs for transplantation: a randomized controlled trial. JAMA 2010;304:
2620-7.
32. Bouadma L, Dreyfuss D, Ricard JD,
Martet G, Saumon G. Mechanical ventilation and hemorrhagic shock-resuscitation
interact to increase inflammatory cytokine release in rats. Crit Care Med 2007;
35:2601-6.
33. Duggan M, Kavanagh BP. Pulmonary
atelectasis: a pathogenic perioperative entity. Anesthesiology 2005;102:838-54.
34. Futier E, Constantin JM, Pelosi P, et al.
Noninvasive ventilation and alveolar recruitment maneuver improve respiratory
function during and after intubation of
morbidly obese patients: a randomized
controlled study. Anesthesiology 2011;114:
1354-63.
35. Fan E, Wilcox ME, Brower RG, et al.
Recruitment maneuvers for acute lung injury: a systematic review. Am J Respir Crit
Care Med 2008;178:1156-63.
36. Lim SC, Adams AB, Simonson DA,
et al. Transient hemodynamic effects of
recruitment maneuvers in three experimental models of acute lung injury. Crit
Care Med 2004;32:2378-84.
37. Jaber S, Chanques G, Jung B. Post­
operative noninvasive ventilation. Anesthesiology 2010;112:453-61.
38. Keenan SP, Sinuff T, Burns KE, et al.
Clinical practice guidelines for the use of
noninvasive positive-pressure ventilation
and noninvasive continuous positive airway pressure in the acute care setting.
CMAJ 2011;183(3):E195-E214.
Copyright © 2013 Massachusetts Medical Society.
n engl j med 369;5 nejm.org august 1, 2013
The New England Journal of Medicine
Downloaded from nejm.org by DARRYL SUE on July 31, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
437
CARING FOR THE
CRITICALLY ILL PATIENT
Association Between Use of Lung-Protective
Ventilation With Lower Tidal Volumes
and Clinical Outcomes Among Patients
Without Acute Respiratory Distress Syndrome
A Meta-analysis
Ary Serpa Neto, MD, MSc
Sérgio Oliveira Cardoso, MD
José Antônio Manetta, MD
Victor Galvão Moura Pereira, MD
Daniel Crepaldi Espósito, MD
Manoela de Oliveira Prado
Pasqualucci, MD
Maria Cecı́lia Toledo
Damasceno, MD, PhD
Marcus J. Schultz, MD, PhD
M
ECHANICAL VENTILATION
is a life-saving strategy
in patients with acute
respiratory failure. However, unequivocal evidence suggests
that mechanical ventilation has the
potential to aggravate and precipitate
lung injury.1 In acute respiratory distress syndrome (ARDS), and in a
milder form of ARDS formerly known
as acute lung injury (ALI),2 mechanical ventilation can cause ventilatorassociated lung injury. Ventilatorassociated lung injury is a frequent
complication in critically ill patients
receiving mechanical ventilation, and
its development increases morbidity
and mortality.1
Higher tidal volume (VT) ventilation causes the alveoli to overstretch
in a process called volutrauma, and
this overstretching is the main cause
of ventilator-associated lung injury.3
The use of a lower VT was shown to
reduce morbidity and mortality in
For editorial comment see p 1689.
Context Lung-protective mechanical ventilation with the use of lower tidal volumes
has been found to improve outcomes of patients with acute respiratory distress syndrome (ARDS). It has been suggested that use of lower tidal volumes also benefits
patients who do not have ARDS.
Objective To determine whether use of lower tidal volumes is associated with improved outcomes of patients receiving ventilation who do not have ARDS.
Data Sources MEDLINE, CINAHL, Web of Science, and Cochrane Central Register
of Controlled Trials up to August 2012.
Study Selection Eligible studies evaluated use of lower vs higher tidal volumes in patients without ARDS at onset of mechanical ventilation and reported lung injury development, overall mortality, pulmonary infection, atelectasis, and biochemical alterations.
Data Extraction Three reviewers extracted data on study characteristics, methods,
and outcomes. Disagreement was resolved by consensus.
Data Synthesis Twenty articles (2822 participants) were included. Meta-analysis using
a fixed-effects model showed a decrease in lung injury development (risk ratio [RR], 0.33;
95% CI, 0.23 to 0.47; I2, 0%; number needed to treat [NNT], 11), and mortality (RR,
0.64; 95% CI, 0.46 to 0.89; I2, 0%; NNT, 23) in patients receiving ventilation with lower
tidal volumes. The results of lung injury development were similar when stratified by
the type of study (randomized vs nonrandomized) and were significant only in randomized trials for pulmonary infection and only in nonrandomized trials for mortality. Metaanalysis using a random-effects model showed, in protective ventilation groups, a lower
incidence of pulmonary infection (RR, 0.45; 95% CI, 0.22 to 0.92; I2, 32%; NNT, 26),
lower mean (SD) hospital length of stay (6.91 [2.36] vs 8.87 [2.93] days, respectively;
standardized mean difference [SMD], 0.51; 95% CI, 0.20 to 0.82; I2, 75%), higher mean
(SD) PaCO2 levels (41.05 [3.79] vs 37.90 [4.19] mm Hg, respectively; SMD, −0.51; 95%
CI, −0.70 to −0.32; I2, 54%), and lower mean (SD) pH values (7.37 [0.03] vs 7.40 [0.04],
respectively; SMD, 1.16; 95% CI, 0.31 to 2.02; I2, 96%) but similar mean (SD) ratios of
PaO2 to fraction of inspired oxygen (304.40 [65.7] vs 312.97 [68.13], respectively; SMD,
0.11; 95% CI, −0.06 to 0.27; I2, 60%). Tidal volume gradients between the 2 groups
did not influence significantly the final results.
Conclusions Among patients without ARDS, protective ventilation with lower tidal
volumes was associated with better clinical outcomes. Some of the limitations of the
meta-analysis were the mixed setting of mechanical ventilation (intensive care unit or
operating room) and the duration of mechanical ventilation.
www.jama.com
JAMA. 2012;308(16):1651-1659
Author Affiliations: Department of Critical Care Medicine, ABC Medical School, Santo André, São Paulo,
Brazil (Drs Serpa Neto, Cardoso, Manetta, Pereira, Espósito, and Damasceno); Department of Internal Medicine, Hospital das Clı́nicas, University of São Paulo, São
Paulo (Dr Damasceno); and Department of Intensive
Care Medicine and Laboratory of Experimental Intensive Care and Anesthesiology, Academic Medical
©2012 American Medical Association. All rights reserved.
Center, University of Amsterdam, Amsterdam, the
Netherlands (Dr Schultz).
Corresponding Author: Ary Serpa Neto, MD, MSc, Avenue Lauro Gomes, 2000 São Paulo, Brazil (aryserpa
@terra.com.br).
Caring for the Critically Ill Patient Section Editor: Derek
C. Angus, MD, MPH, Contributing Editor, JAMA
([email protected]).
JAMA, October 24/31, 2012—Vol 308, No. 16
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
1651
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
patients with ARDS or ALI, thus justifying the progressive decrease in VT
used by clinicians over the past
decades.4-6 However, in critically ill
patients without ALI, there is little
evidence regarding the benefits of
ventilation with lower V T , partly
because of a lack of randomized controlled trials evaluating the best ventilator strategies in these patients.7
Some observational studies have
suggested that use of higher V T in
patients without ARDS or ALI, at the
initiation of mechanical ventilation,
increases morbidity and mortality.8-10
As suggested by the “biotrauma
hypothesis,” ventilation with higher
V T and peak pressures may lead to
recruitment of neutrophils and local
production and release of inflammatory mediators. 11 We conducted a
meta-analysis to determine whether
conventional (higher) or protective
(lower) tidal volumes would be associated with lung injury, mortality,
pulmonary infection, and atelectasis
in patients without lung injury
at the onset of mechanical ventilation.
METHODS
Studies were identified by 2 authors
through a computerized blinded search
of MEDLINE (1966-2012), Cumulative Index to Nursing and Allied Health
Literature (CINAHL), Web of Science, and Cochrane Central Register of
Controlled Trials (CENTRAL) using a
sensitive search strategy combining the
following Medical Subject Headings and
keywords (protective ventilation [text
word] OR lower tidal volumes [text
word]). All reviewed articles and crossreferenced studies from retrieved articles were screened for pertinent information.
Selection of Studies
Articles were selected for inclusion in
the systematic review if they evaluated 2 types of ventilation in patients
without ARDS or ALI at the onset of
mechanical ventilation. In 1 group of
the study, ventilation was protective
(lower VT). Then, this protective ven1652
JAMA, October 24/31, 2012—Vol 308, No. 16
tilation group was compared with
another group using conventional
methods (higher V T ). A study was
deemed eligible if it evaluated patients
who did not meet the consensus criteria for ARDS or ALI at baseline.12 We
included randomized trials as well as
observational studies (cohort, before/
after, and cross-sectional), with no
restrictions on language or scenario
(intensive care unit or operating
room). We excluded revisions and
studies that did not report the outcomes of interest. When we found
duplicate reports of the same study in
preliminary abstracts and articles, we
analyzed data from the most complete
data set. When necessary, we contacted the authors for additional
unpublished data.
Data Extraction
Data were independently extracted from
each report by 3 authors using a data
recording form developed for this purpose. After extraction, data were reviewed and compared by the first author. Instances of disagreement between
the 2 other extractors were solved by a
consensus among the investigators.
Whenever needed, we obtained additional information about a specific study
by directly questioning the principal investigator.
Validity Assessment
In randomized trials, we assessed allocation concealment, the baseline similarity of groups (with regard to age, severity of illness, and severity of lung
injury), and the early stopping of treatment. We used the GRADE approach
to summarize the quality of evidence
for each outcome.13 In this approach,
randomized trials begin as highquality evidence but can be rated down
for apparent risk of bias, imprecision,
inconsistency, indirectness, or suspicion of a publication bias.
Definition of End Points
The primary end point was the development of lung injury in each group of
the study. Secondary end points included overall survival, incidence of
pulmonary infection and atelectasis, intensive care unit (ICU) and hospital
length of stay, time to extubation,
change in PaCO2, arterial pH values, and
change in the ratio of PaO2 to fraction
of inspired oxygen (FIO2).
Statistical Analysis
We extracted data regarding the
study design, patient characteristics,
type of ventilation, mean change in
arterial blood gases, lung injury
development, ICU and hospital
length of stay, time to extubation,
overall survival, and incidence of
atelectasis. For the analysis of lung
injury development, mortality, pulmonary infection, and atelectasis, we
used the most protracted follow-up
in each trial up to hospital discharge.
We calculated a pooled estimate of
risk ratio (RR) in the individual studies using a fixed-effects model
according to Mantel and Haenszel
and graphically represented these
results using forest plot graphs.
We explored the following variables as potential modifiers: incorporation of “open lung” techniques (using
the authors’ definitions) into experimental strategies, between-group gradients in tidal volumes and plateau pressures, and case mix effects. We reasoned
that each of these might influence the
effect of protective ventilation on outcome. To explore whether these variables modified the outcome, we compared pooled effects among studies with
and without them. For continuous variables, we used the standardized mean
difference (SMD), which is the difference in means divided by a standard
deviation.
The homogeneity assumption was
measured by the I2, which describes the
percentage of total variation across studies that is due to heterogeneity rather
than chance. I2 was calculated from basic results obtained from a typical metaanalysis as I2 =100%⫻(Q−df)/Q, where
Q is the Cochran heterogeneity statistic. A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity.
When heterogeneity was found
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
(I2 ⬎ 25%) we presented the randomeffects model results as primary
analysis.
A sensitivity analysis was carried
out by recalculating pooled RR estimates for different subgroups of studies based on relevant clinical features.
This analysis demonstrates whether
the overall results have been affected
by a change in the meta-analysis
selection criteria. Also, a sensitivity
analysis about the treatment effect
according to quality components of
the studies (concealed treatment allocation, blinding of patients and caregivers, blinded outcome assessment)
was conducted. A potential publication bias was assessed graphically
with funnel plots, as well as by a Begg
and Mazumdar rank correlation and
an Egger regression. Interrater reliability was determined by comparing
the number of studies included by
one author with those of another
author in each stage of the search
using ␬ coefficients.
Parametric variables were presented as the mean and standard deviation, and nonparametric variables
were presented as the median and interquartile range (IQR). All analyses
were conducted with Review Manager
version 5.1.1 (The Cochrane Collaboration) and SPSS version 16.0.1 (IBM
SPSS). For all analyses, 2-sided P values less than .05 were considered significant.
RESULTS
Our initial search yielded 2122 studies (458 from MEDLINE, 141 from
CENTRAL, 885 from CINAHL, and 638
from Web of Science). After removing
711 duplicate studies, we evaluated the
abstracts of 1411 studies. After evaluating the abstract of each study, we excluded 1364 studies because they did
not meet inclusion criteria. Subsequently, we carefully read the full text
of each of the remaining 47 studies and
excluded 27 for the following reasons:
no data on outcome of interest in 20
studies and same cohort previously analyzed in 7. Twenty references (2822 participants) were included in the final
analysis (FIGURE 1 and TABLE 1). For
the comparisons of interrater reliability in each stage of the search, the ␬ coefficient was 0.91 in the citation stage
(P = .004), 0.86 during the abstract review (P =.03), and 0.90 in the full-text
stage (P=.006).
Figure 1. Literature Search Strategy
2122 Articles identified
458 From MEDLINE
141 From CENTRAL
885 From CINAHL
638 From Web of Science
711 Excluded (duplicate studies)
Study Characteristics
Table 1 summarizes the studies’ characteristics. All but 5 studies16,22,23,26,29
were randomized controlled trials, and
median follow-up time was 21.0 hours
(IQR, 6.28-54.60 hours). The median
time of per-protocol mechanical ventilation was 6.90 hours for protective
and 6.56 hours for conservative strategy. The development of lung injury
was the primary outcome in 4 studies.
Eight studies evaluated the levels of
inflammatory mediators in bronchoalveolar lavage or blood. Tidal volume
was set to 6 mL/kg of ideal body
weight (IBW) in the protective group
of 13 studies; only in 1 study was the
tidal volume in the protective ventilation group above 8 mL/kg IBW. Four
studies did not report what weight
was used to calculate the tidal volume,14,15,21,25 1 study used the measured weight,19 and 15 studies used
the predicted weight.9,16-18,20,22-24,26-32
Of these, 7 used the ARDSnet formula
to calculated the predicted body
weight.16,18,20,24,28-30
The tidal volume gradient between
protective and conventional ventilation ranged from 2 to 6 mL/kg IBW,
with a mean (SD) of 4.15 (1.42)
mL/kg IBW. The tidal volume gradient
was less than 4 mL/kg IBW in 30.0%
of the studies, between 4 and 5 mL/kg
IBW in 40% of the studies, and above
5 mL/kg IBW in 30% of the studies. In
15 studies, the reason for intubation
was scheduled surgery,9,15,17-22,24,25,29-32
and in 5, the reason was mixed (medical or surgery).14,16,23,24,28 Lung injury
was diagnosed according to the
American-European Consensus Conference definition in 6 of the 8 trials
that assessed this outcome.16,23,26,27,31,32
The diagnosis of infection was
made by clinical assessment plus laboratory, radiological, and microbiologi-
©2012 American Medical Association. All rights reserved.
1411 Potentially relevant articles
screened based on abstracts
1364 Excluded
576 ARDS/ALI at onset of
mechanical ventilation
487 Reviews
227 Experimental studies
33 Secondary analysis
21 Older version of a study
20 Other
47 Full-text articles assessed
for eligibility
27 Excluded
20 No data on outcome
of interest
7 Same cohort previously
analyzed
20 Articles included in meta-analysis
(2822 study participants)
ALI indicates acute lung injury; ARDS, acute respiratory distress syndrome; CENTRAL, Cochrane Central
Register of Controlled Trials; CINAHL, Cumulative Index to Nursing and Allied Health Literature.
cal evaluation in 2 studies14,26 ; was
made by decrease in PaO2/FIO2 plus
radiological assessment in 1 study31;
and was not specified in the last
study.20
eTable 1 (available at http://www.jama
.com) summarizes study methods, highlighting features related to the risk of
bias. Randomization was concealed in
11 of 15 randomized controlled trials included, and follow-up was excellent with
minimal loss. Limitations included a lack
of blinding (all trials), a lack of intention-to treat analysis (12 trials), and early
stopping for benefit (1 trial). Age, weight,
minute-volume (product of respiratory rate and tidal volume), and PaO2/
FIO2 were all similar between the 2
groups analyzed (TABLE 2 and eTable 2).
As expected, VT and plateau pressure
were lower and positive end-expiratory pressure (PEEP) and respiratory rate
were higher in the protective group.
PaCO2 was higher in the protective group
JAMA, October 24/31, 2012—Vol 308, No. 16
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
1653
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
but remained within normal limits
(35-45 mm Hg). Acidosis (pH ⬍7.35)
was found in the protective group in 3
studies, and the pH level in the protective group was similar to that of the conventional group. The mechanical ventilation settings for each study are
provided in eTable 3.
sult of the overall test for heterogeneity was not statistically significant,
and the I2 was 0% (no sign of heterogeneity) (FIGURE 2). When stratified by
the tidal volume gradient between the
2 groups, the RR for lung injury decreased from 0.35 (95% CI, 0.230.51) in the group with less than 4
Primary Outcome
Forty-seven of 1113 patients (4.22%)
assigned to protective ventilation and
138 of 1090 patients (12.66%) assigned to conventional ventilation developed lung injury during follow-up
(RR, 0.33; 95% CI, 0.23-0.47; number
needed to treat [NNT], 11). The re-
Table 1. Characteristics of the Included Studies and Summary of Continuous Variables
Protective
Source a
Lee et al,14
1999
Chaney et al,15
2000
Gajic et al,16
2004
Koner et al,17
2004
Wrigge et al,9
2004
Wrigge et al,18
2005
Zupancich et
al,19 2005
Michelet et al,20
2006
Cai et al,21 2007
Wolthuis et al,22
2007
Yilmaz et al,23
2007
Determann et
al,24 2008
Lin et al,25 2008
No. of
Patients
103
VT,
mL/kg
6
No.
47
Duration of MV,
Mean (SD), h
Conservative
VT,
mL/kg
12
No.
56
Setting
SICU
Protective
2.30 (0.5)
Conservative
3.90 (0.8)
Dis
ST ⫹ 1
ST ⫹ 1
NS
NS
12
9.90 (1.0)
10.0 (1.4)
3
NS
NS
16.1 (10.2)
12.9 (4.4)
Primary
Outcome
Duration of MV
Jadad
Score
3
Pulmonary
mechanics
LI
2
Cytokines in
blood
Cytokines in
BAL
Cytokines in
BAL
Cytokines in
BAL
Cytokines in
blood
CT atelectasis
Sedative use
1
25
6
12
12
13
166
9
66
12
100
44
6
15
10
29
CABG
62
6
30
12
32
Surgical
44
6
22
12
22
CABG
40
8
20
10
20
CS
6
NS
NS
52
5
26
9
26
OS
18
7.06 (1.81)
7.76 (1.85)
16
36
6
8
8
23
10
10
8
13
Neurosurgery
ICU
7.15
6.90 (2.2)
NS
7.4 (3.1)
NS
375
8
163
11
212
NS
NS
LI
40
6
21
12
19
Surgical
5
ST
ST
40
5
20
9
20
OS
24
4.33 (0.9)
4.23 (0.71)
1091
6
558
9
533
OS
2.93 (1.2)
2.76 (1.0)
Cytokines in
BAL
Cytokines in
blood
LI
150
6
76
10
74
ICU
672
NS
NS
20
6
10
12
10
SICU
672
168.0
72.0
229
8
154
10
75
Surgical
NS
NS
Sundar et al,30
2011
Yang et al,31
2011
149
6
75
10
74
CS
672
7.50
10.71
100
6
50
10
50
OS
168
2.00 (0.68)
Weingarten et
al,32 2012
Total, Mean
(SD)
40
6
20
10
20
Surgical
Dis
21.0
(6.28-54.6) b
Licker et al,26
2009
Determann et
al,27 2010
de Oliveira et
al,28 2010
FernandezBustamante
et al,29 2011
2822
6.45
(1.09)
1416
10.60
(1.14)
1406
CABG
Follow-up,
h
168
ICU
Dis
ICU
Cytokines in
BAL
Cytokines in
BAL
3
1
1
3
2
3
1
3
3
Duration of MV;
ICULS;
mortality
Duration of MV
3
2.11 (0.8)
LI
3
5.13 (1.86)
5.73 (1.71)
Oxygenation
3
6.90
(2.93-9.90) b
6.56
(3.61-10.17) b
2.33
(0.89)
Abbreviations: BAL, bronchoalveolar lavage; CABG, coronary artery bypass graft surgery; CS, cardiac surgery; CT, computed tomography; Dis, until patient’s discharge; ICU, intensive
care unit; ICULS, ICU length of stay; LI, lung injury; MV, mechanical ventilation; NS, not specified; OS, oncology surgery; SICU, surgical intensive care unit; ST, surgery time; VT, tidal
volume.
a Most of the studies were randomized controlled trials. The exceptions are as follows: Gajic et al,16 Yilmaz et al,23 and Licker et al,26 were cohort studies; Wolthuis et al22 had a beforeand-after design; and Bustamante et al29 had a cross-sectional design.
b Median (interquartile range).
1654
JAMA, October 24/31, 2012—Vol 308, No. 16
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
mL/kg IBW to 0.26 (95% CI, 0.100.66) in the group with 4 to 5 mL/kg
IBW (eFigure 1). The RR for the development of lung injury with conventional ventilation, analyzing only randomized controlled trials, was 0.26
(95% CI, 0.10-0.66; NNT, 10).
Secondary Outcomes
Overall mortality was lower in patients receiving protective ventilation
(RR, 0.64; 95% CI, 0.46 to 0.89; NNT,
23). The incidence of pulmonary infection (using the authors’ definition)
and atelectasis were lower in the group
receiving ventilation with a lower VT
(RR [random-effect], 0.45; 95% CI, 0.22
to 0.92; NNT, 26; and RR, 0.62; 95%
CI, 0.41 to 0.95, respectively)
(Figure 2). The I2 test indicated moderate heterogeneity only in the analysis of pulmonary infection (32%). Protective ventilation was associated with
a shorter mean (SD) hospital stay (6.91
[2.36] vs 8.87 [2.93] days, respectively; SMD, 0.51; 95% CI, 0.20 to 0.82),
and showed no difference in ICU stay
(3.63 [2.43] vs 4.64 [3.29] days, respectively; SMD, 0.37; 95% CI, −0.53
to 1.27) and time of mechanical ventilation (51.07 [58.08] vs 47.12 [45.00]
hours, respectively; SMD, 0.48; 95% CI,
−0.27 to 1.23).
Mean (SD) levels of Pa CO 2 were
higher in the protective ventilation
group (41.05 [3.79] vs 37.90 [4.19]
mm Hg, respectively; SMD, −0.51; 95%
CI, −0.70 to −0.32), and mean (SD) pH
levels were lower (7.37 [0.03] vs 7.40
[0.03], respectively; SMD, 1.16; 95% CI,
0.31 to 2.02). The mean (SD) PaO2/
F IO 2 ratio was similar between the
groups (304.40 [65.70] vs 312.97
[68.13], respectively; SMD, 0.11; 95%
CI, −0.06 to 0.27). All these analyses
yield significant heterogeneity and were
analyzed by random-effects model (I2
for hospital stay, ICU stay, time of mechanical ventilation, PaCO2, pH, and
PaO2/FIO2 of 75%, 95%, 92%, 54%, 96%,
and 60%, respectively) (eFigures 2, 3,
4, 5, 6, and 7 and eTable 4).
In eTable 5, the GRADE evidence
profile is provided. This profile evaluates the effect of protective ventilation
Table 2. Demographic, Ventilation, and Laboratory Characteristics of the Patients at the Final
Follow-up Visit
Mean (SD)
Protective
Ventilation
(n = 1416)
Age, y
59.97 (7.92)
Conventional
Ventilation
(n = 1406)
60.22 (7.36)
Weight, kg
Tidal volume, mL/kg IBW a
PEEP, cm H2O a
Plateau pressure, cm H2O a
72.71 (12.34)
6.45 (1.09)
6.40 (2.39)
72.13 (12.16)
10.60 (1.14)
3.41 (2.79)
16.63 (2.58)
18.02 (4.14)
21.35 (3.61)
13.20 (4.43)
.006
.01
8.46 (2.90)
304.41 (65.74)
41.05 (3.79)
7.37 (0.03)
9.13 (2.70)
312.97 (68.13)
37.90 (4.19)
7.40 (0.03)
.72
.51
.003
.11
Respiratory rate,
breaths/min a
Minute-volume, L/min a,b
PaO2/FIO2 a
PaCO2, mm Hg a
pH a
P
Value
.93
⬍.001
.01
.93
Abbreviations: FIO2, fraction of inspired oxygen; IBW, ideal body weight; PEEP, positive end-expiratory pressure.
a At the final follow-up visit.
b Minute-volume is the product of respiratory rate and tidal volume.
in patients without ARDS or ALI, only
from a systematic review and a metaanalysis of randomized controlled trials.
The findings for lung injury, mortality, and pulmonary infection were considered moderate, high, and low quality, respectively, by the GRADE profile.
Sensitivity analyses according to quality components of each study are shown
in eTable 6.
In addition, we excluded each trial
one at a time and assessed the results.
In lung injury and pulmonary infection analyses, the results were always
significant despite the exclusion of any
trial. After we excluded the trial by
Yilmaz et al,23 the analysis of mortality
was no longer significant.
Sensitivity Analysis
To explore these results, we performed a stratified analysis across a
number of key study characteristics and
clinical factors, and this analysis is
shown in TABLE 3. Protection from lung
injury, in the protective group, was
more pronounced in studies that were
not randomized controlled trials performed in the ICU. These trials did not
incorporate recruitment maneuvers,
had a higher plateau pressure gradient, and a smaller tidal volume gradient. In the survival analysis, we found
significant changes in studies without
recruitment maneuvers, in studies that
©2012 American Medical Association. All rights reserved.
were not randomized trials, and in studies performed in the ICU with a lower
tidal volume gradient.
For pulmonary infections, we
found no statistically significant association in studies that were not randomized trials, a tidal volume gradient less than 4 mL/kg IBW, and the
use of recruitment maneuvers. A tidal
volume gradient from 4 to 5 mL/kg
IBW and a randomized controlled
trial performed in surgical patients
were each associated with a significant reduction in pulmonary infections in the protective group.
Publication Bias
Funnel-plot graphical analysis (eFigure 8), Begg and Mazumdar rank correlation, and Egger regression did not
suggest a significant publication bias for
the analyses conducted in Figure 2
(Kendall ␶=0.17, P=.63; Egger regression intercept=0.24, P =.68).
COMMENT
We found evidence that a ventilation
strategy using lower tidal volumes is associated with a lower risk for developing ARDS. Furthermore, the strategy
was associated with lower mortality,
fewer pulmonary infections, and less atelectasis when compared with higher
tidal volume ventilation in patients
without lung injury at the onset of meJAMA, October 24/31, 2012—Vol 308, No. 16
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
1655
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
ciated with the use of lower tidal volumes. Notably, differences in the levels of PEEP and plateau pressure did not
influence the final results of the metaanalysis.
Previously, Esteban et al33 showed
plateau pressures above 35 cm H2O to
be associated with an increased risk of
death in ICU patients. Although not definitive, this study at least suggested that
chanical ventilation. These benefits
were associated with a shorter hospital length of stay. Protective ventilation was associated with higher PaCO2
levels and lower pH values, but no difference in the incidence of acidosis was
found. In all studies, although the primary goal of the investigators was to
compare 2 different tidal volumes, other
ventilator strategy elements were asso-
higher VT has the ability to exaggerate
lung injury and maybe even cause death
in patients who require mechanical ventilation for days. Fernández-Pérez et al34
showed higher VT to be associated with
postoperative respiratory failure in patients receiving ventilation for only a
few hours in the operating room. In
light of this information, over the past
decade, V T has progressively de-
Figure 2. Effect of Ventilation With Smaller Tidal Volume in Patients With Healthy Lungs at the End of the Follow-up Period for Each Study
High VT, No.
Events
Lung injury
Gajic et al,16 2004
Michelet et al,20 2006
Yilmaz et al,23 2007
Licker et al,26 2009
Determann et al,27 2010
Yang et al,31 2011
Fernandez-Bustamante et al,29 2011
Weingarten et al,32 2012
Subtotal (95% CI)
Total events
32
6
60
20
10
4
5
1
Total
100
26
212
533
74
50
75
20
Low VT, No.
Events
12
3
17
5
2
1
7
0
1090
138
Total
Weight, %
RR (95% CI)
66
26
163
558
76
50
154
20
18.1
4.6
40.7
17.7
8.6
3.4
5.6
1.3
0.47 (0.22-1.00)
0.43 (0.10-1.97)
0.29 (0.16-0.53)
0.23 (0.09-0.62)
0.17 (0.04-0.82)
0.23 (0.03-2.18)
0.67 (0.20-2.17)
0.32 (0.01-8.26)
1113
100.0
0.33 (0.23-0.47)
Favors Low VT
Favors High VT
47
Heterogeneity: χ27 = 3.74; P = .81, I 2 = 0%
Test for overall effect: z = 6.06; P<.001
0.01
0.1
1.0
10
100
10
100
10
100
10
100
RR (95% CI)
Mortality
Michelet et al,20 2006
Wolthuis et al,22 2007
Yilmaz et al,23 2007
Licker et al,26 2009
Determann et al,27 2010
Fernandez-Bustamante et al,29 2011
Sundar et al,30 2011
Yang et al,31 2011
Weingarten et al,32 2012
Subtotal (95% CI)
Total events
1
2
69
15
23
1
2
1
1
26
13
212
533
74
75
74
50
20
2
3
27
13
24
3
1
0
1
1077
115
26
23
163
558
76
154
75
50
20
1.0
2.5
55.7
16.7
17.7
1.5
2.2
1.7
1.1
2.08 (0.18-24.51)
0.82 (0.12-5.71)
0.41 (0.25-0.68)
0.82 (0.39-1.75)
1.02 (0.51-2.04)
1.47 (0.15-14.38)
0.49 (0.04-5.48)
0.33 (0.01-8.21)
1.00 (0.06-17.18)
1145
100.0
0.64 (0.46-0.86)
74
Heterogeneity: χ28 = 6.94; P = .54, I 2 = 0%
Test for overall effect: z = 2.68; P = .007
Pulmonary infection
Lee et al,14 1999
Michelet et al,20 2006
Licker et al,26 2009
Yang et al,31 2011
Subtotal (95% CI)
Total events
0.01
Subtotal (95% CI)
Total events
1.0
RR (95% CI)
10
10
30
7
56
26
533
50
2
6
23
1
665
57
47
26
558
50
16.6
14.6
55.8
13.0
0.20 (0.04-0.99)
0.48 (0.14-1.60)
0.72 (0.41-1.26)
0.13 (0.01-1.06)
681
100.0
0.52 (0.33-0.82)
32
Heterogeneity: χ23 = 4.39; P = .22, I 2 = 32%
Test for overall effect: z = 2.79; P = .005
Atelectasis
Lin et al,25 2008
Cai et al,21 2007
Licker et al,26 2009
Yang et al,31 2011
Weingarten et al,32 2012
0.1
0.01
0.1
1.0
RR (95% CI)
2
5
47
3
5
20
8
533
50
20
3
7
28
1
4
631
62
Heterogeneity: χ24 = 3.76; P = .44, I 2 = 0%
Test for overall effect: z = 2.18; P = .03
20
8
558
50
20
3.1
1.1
83.1
5.4
7.3
1.59 (0.24-10.70)
4.20 (0.33-53.12)
0.55 (0.34-0.89)
0.32 (0.03-3.18)
0.75 (0.17-3.33)
656
100.0
0.62 (0.41-0.95)
43
0.01
0.1
1.0
RR (95% CI)
A pooled estimate of risk ratio (RR) was calculated in the individual studies using a fixed-effects model according to Mantel and Haenszel. The size of the data markers
indicates the weight of the study in the final analyses. VT indicates tidal volume.
1656
JAMA, October 24/31, 2012—Vol 308, No. 16
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
creased from greater than 12 to 15
mL/kg IBW to less than 9 mL/kg
IBW.6,35 The results of the present metaanalysis support this change in ventilation practice. Our results may even
suggest that VT should be further reduced.
Protective ventilation in patients
with ALI or ARDS is already well
established; however, physicians do
not always adhere to such guidelines.
Mikkelsen et al 36 reported that approximately one-third of the patients were receiving protective ventilation at 48 hours, and the main
reason for poor adherence was the
uncertainty about the diagnosis of
ARDS. Another possible reason is
that 82% of the patients who never
received protective ventilation had a
plateau pressure below 30 cm H2O.
However, it is well established that
reducing the VT in patients with plateau pressures below 30 cm H2O is
associated with a survival benefit.10
In this context, the adoption of protective ventilation in patients without
lung injury may be even more difficult.
It is possible that the beneficial effects of protective ventilation, regarding the development of lung injury, are
even greater than what is suggested by
the current analysis. Mechanical ventilation can damage the lung, cause inflammation, and release cytokines into
the systemic circulation.20,25 This process may cause fever, leukocytosis, and
new pulmonary infiltrates, which could
be interpreted as ventilator-associated
pneumonia instead of ventilatorassociated lung injury. The absence of
strict criteria for the diagnosis of pneumonia, such as microbiological identification in blood and bronchoalveolar
lavage, in the studies evaluated may lead
to an incorrect diagnosis. Ventilatorassociated lung injury may be incorrectly diagnosed as pneumonia in many
cases, underestimating the true incidence of lung injury. It is difficult to diagnose pneumonia in the presence of
ARDS or ALI, with a quoted sensitivity using conventional clinical criteria
of less than 50%.37
Table 3. Summary of Stratified Analyses of Pooled Relative Risks
Stratified Analysis
Acute Lung Injury
Recruitment maneuvers
Yes
No
Tidal volume gradient, mL/kg IBW
⬍4
4-5
Randomized
Yes
No
Setting
Operation room
ICU
Plateau pressure gradient, cm H2O
⬍4
4-8
Diagnosis
AECCD
Other
Mortality
Recruitment maneuvers
Yes
No
Tidal volume gradient, mL/kg IBW
⬍4
4-5
Randomized
Yes
No
Setting
Operation room
ICU
Plateau pressure gradient, cm H2O
⬍4
4-8
Pulmonary Infection
Recruitment maneuvers
Yes
No
Tidal volume gradient, mL/kg IBW
⬍4
4-5
⬎5
Randomized
Yes
No
Setting
Operation room
ICU
Plateau pressure gradient, cm H2O
⬍4
4-8
Infection diagnosis
Not specified
Specified
CLCXR
PaO2/FIO2 ⫹ x-ray
No. of
Trials
No. of
Patients
Risk Ratio
(95% CI)
P
Value
Heterogeneity,
Q
1
7
1091
1112
0.23 (0.09-0.62)
0.35 (0.24-0.52)
.004
⬍.001
0.80
4
4
1861
342
0.35 (0.23-0.51)
0.26 (0.10-0.66)
⬍.001
.004
0.43
.87
4
4
342
1861
0.26 (0.10-0.66)
0.35 (0.23-0.51)
.004
⬍.001
0.87
0.43
5
3
1512
691
0.34 (0.18-0.63)
0.33 (0.21-0.51)
⬍.001
⬍.001
0.73
0.43
3
1
368
1091
0.38 (0.21-0.71)
0.23 (0.09-0.62)
.002
.004
0.52
6
2
1922
281
0.30 (0.21-0.45)
0.56 (0.22-1.41)
⬍.001
.22
0.83
0.66
1
8
1091
1131
0.82 (0.39-1.75)
0.60 (0.42-0.87)
.61
.006
0.49
4
5
1731
491
0.54 (0.36-0.79)
0.97 (0.53-1.78)
.002
.92
0.35
0.89
5
4
491
1731
0.97 (0.53-1.78)
0.54 (0.36-0.79)
.92
.002
0.89
0.35
6
3
1661
561
0.86 (0.46-1.60)
0.57 (0.38-0.84)
.63
.005
0.94
0.10
3
1
351
1091
1.02 (0.54-1.92)
0.82 (0.39-1.75)
.95
.61
0.71
1
3
1091
255
0.72 (0.41-1.26)
0.27 (0.12-0.64)
.25
.003
1
2
1
1091
152
103
0.72 (0.41-1.26)
0.31 (0.11-0.86)
0.20 (0.04-0.99)
.25
.02
.05
3
1
255
1091
0.27 (0.12-0.64)
0.72 (0.41-1.26)
.003
.25
0.48
3
1
1243
103
0.59 (0.36-0.95)
0.20 (0.04-0.99)
.03
.05
0.27
2
1
155
1091
0.33 (0.13-0.85)
0.72 (0.41-1.26)
.02
.25
0.40
1
3
2
1
52
1294
1194
100
0.48 (0.14-1.60)
0.53 (0.32-0.87)
0.60 (0.36-1.01)
0.13 (0.01-1.06)
.23
.01
.05
.06
.48
0.28
0.11
0.14
Abbreviations: AECCD, American-European Consensus Conference definition; CLCXR, clinical⫹laboratory⫹culture⫹xray; FIO2, fraction of inspired oxygen; IBW, ideal body weight; ICU, intensive care unit.
©2012 American Medical Association. All rights reserved.
JAMA, October 24/31, 2012—Vol 308, No. 16
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
1657
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
Our findings are in line with a recently published retrospective study of
cardiac surgery patients.38 Although it
should be noted that the lower tidal volumes in that study were much higher
than those used in the protective groups
of the studies analyzed in this metaanalysis, a tidal volume of more than
10 mL/kg was found as a risk factor for
organ failure and prolonged ICU stay
after cardiac surgery.
The results of this meta-analysis
should be interpreted within the context of the included studies. Systematic reviews are subject to publication
bias, which may exaggerate the study’s
conclusion if publication is related to
the strength of the results. Additionally, it may be important to distinguish between mechanical ventilation
performed in the operating room and
that performed in the ICU. Patients in
the operating room receive mechanical
ventilation for a much shorter time
than those in the ICU. Both surgical
patients and critically ill patients are at
risk for several causes of lung injury.
However, these may not be the same
for both patient groups, and mechanical ventilation may have different
effects on both groups. In addition,
although our meta-analysis found
decreased mortality rate with protective ventilation, the interpretation of
this finding should be considered cautiously because it was discovered only
after the addition of the study by
Yilmaz et al. 23 Also, one important
limitation is that the patients received
ventilation for a relatively short time
in most studies, which complicates the
extrapolation of the results for patients
receiving ventilation for long periods
in the ICU. For the lung injury analysis, 4 of 8 studies (accounting for
85.4% and 87.2% of the events in the
conservative and protective groups,
respectively) were not randomized
controlled trials, and the randomized
controlled trials were of moderate
quality. Furthermore, funnel plots are
limited as a test for publication bias
for a small number of studies.
All the dichotomous analyses yielded
significant results, and with the excep1658
JAMA, October 24/31, 2012—Vol 308, No. 16
tion of pulmonary infection, all the results showed no heterogeneity (I2 =0%).
Pulmonary infection yielded moderate heterogeneity (I2 = 32%), but the
analysis with a random-effects model
showed similar results. However, all the
continuous analyses showed significant heterogeneity (all I2 ⬎60%) and
with the use of a random-effects model
only differences in pH level, PaCO2 level,
and hospital length of stay showed significant results. Therefore, continuous analyses need to be interpreted with
caution because of the heterogeneity.
In conclusion, our meta-analysis suggests that among patients without lung
injury, protective ventilation with use
of lower tidal volumes at onset of mechanical ventilation may be associated
with better clinical outcomes. We believe that clinical trials are needed to
compare higher vs lower tidal volumes in a heterogeneous group of patients receiving mechanical ventilation for longer periods.
Author Contributions: Dr Serpa Neto had full access
to all of the data in the study and takes responsibility
for the integrity of the data and the accuracy of the
data analysis.
Study concept and design: Serpa Neto, Cardoso,
Manetta, Pereira, Espósito, Schultz.
Acquisition of data: Serpa Neto, Pereira, Espósito,
Pasqualucci.
Analysis and interpretation of data: Serpa Neto,
Cardoso, Manetta, Pereira, Espósito, Damasceno,
Schultz.
Drafting of the manuscript: Serpa Neto, Pereira,
Damasceno, Schultz.
Critical revision of the manuscript for important intellectual content: Serpa Neto, Cardoso, Manetta,
Pereira, Espósito, Pasqualucci, Damasceno, Schultz.
Statistical analysis: Serpa Neto, Pereira.
Administrative, technical, or material support:
Cardoso, Manetta, Espósito, Pasqualucci, Damasceno,
Schultz.
Study supervision: Damasceno, Schultz.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure
of Potential Conflicts of Interest and none were reported.
Online-Only Material: The eTables and eFigures are
available at http://www.jama.com.
REFERENCES
1. Plötz FB, Slutsky AS, van Vught AJ, Heijnen CJ. Ventilator-induced lung injury and multiple system organ failure: a critical review of facts and hypotheses.
Intensive Care Med. 2004;30(10):1865-1872.
2. The ARDS Definition Task Force; Ranieri VM,
Rubenfeld GD, et al. Acute respiratory distress syndrome: the Berlin Definition. JAMA. 2012;307
(23):2526-2533.
3. Carney D, DiRocco J, Nieman G. Dynamic alveolar mechanics and ventilator-induced lung injury. Crit
Care Med. 2005;33(3)(suppl):S122-S128.
4. Amato MB, Barbas CS, Medeiros DM, et al. Effect
of a protective-ventilation strategy on mortality in the
acute respiratory distress syndrome. N Engl J Med.
1998;338(6):347-354.
5. The Acute Respiratory Distress Syndrome Network.
Ventilation with lower tidal volumes as compared with
traditional tidal volumes for acute lung injury and the
acute respiratory distress syndrome. N Engl J Med.
2000;342(18):1301-1308.
6. Brun-Buisson C, Minelli C, Bertolini G, et al; ALIVE
Study Group. Epidemiology and outcome of acute lung
injury in European intensive care units: results from
the ALIVE study. Intensive Care Med. 2004;30
(1):51-61.
7. Schultz MJ, Haitsma JJ, Slutsky AS, Gajic O. What
tidal volumes should be used in patients without acute
lung injury? Anesthesiology. 2007;106(6):12261231.
8. Gajic O, Frutos-Vivar F, Esteban A, Hubmayr RD,
Anzueto A. Ventilator settings as a risk factor for acute
respiratory distress syndrome in mechanically ventilated patients. Intensive Care Med. 2005;31(7):
922-926.
9. Wrigge H, Uhlig U, Zinserling J, et al. The effects
of different ventilatory settings on pulmonary and systemic inflammatory responses during major surgery.
Anesth Analg. 2004;98(3):775-781.
10. Hager DN, Krishnan JA, Hayden DL, Brower RG;
ARDS Clinical Trials Network. Tidal volume reduction in patients with acute lung injury when plateau
pressures are not high. Am J Respir Crit Care Med.
2005;172(10):1241-1245.
11. de Prost N, Ricard JD, Saumon G, Dreyfuss D. Ventilator-induced lung injury: historical perspectives and
clinical implications. Ann Intensive Care. 2011;
1(1):28-42.
12. Bernard GR, Artigas A, Brigham KL, et al. The
American-European Consensus Conference on ARDS:
definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;
149(3 pt 1):818-824.
13. Guyatt GH, Oxman AD, Kunz R, Vist GE,
Falck-Ytter Y, Schünemann HJ; GRADE Working Group.
What is “quality of evidence” and why is it important to clinicians? BMJ. 2008;336(7651):995-998.
14. Lee PC, Helsmoortle CM, Cohn SM, Fink MP. Are
low tidal volumes safe? Chest. 1990;97(2):430434.
15. Chaney MA, Nikolov MP, Blakeman BP, Bakhos
M. Protective ventilation attenuates postoperative pulmonary dysfunction in patients undergoing cardiopulmonary bypass. J Cardiothorac Vasc Anesth. 2000;
14(5):514-518.
16. Gajic O, Dara SI, Mendez JL, et al. Ventilatorassociated lung injury in patients without acute lung
injury at the onset of mechanical ventilation. Crit Care
Med. 2004;32(9):1817-1824.
17. Koner O, Celebi S, Balci H, Cetin G, Karaoglu K,
Cakar N. Effects of protective and conventional mechanical ventilation on pulmonary function and systemic cytokine release after cardiopulmonary bypass.
Intensive Care Med. 2004;30(4):620-626.
18. Wrigge H, Uhlig U, Baumgarten G, et al. Mechanical ventilation strategies and inflammatory responses to cardiac surgery: a prospective randomized clinical trial. Intensive Care Med. 2005;31
(10):1379-1387.
19. Zupancich E, Paparella D, Turani F, et al. Mechanical ventilation affects inflammatory mediators in
patients undergoing cardiopulmonary bypass for cardiac surgery: a randomized clinical trial. J Thorac Cardiovasc Surg. 2005;130(2):378-383.
20. Michelet P, D’Journo XB, Roch A, et al. Protective ventilation influences systemic inflammation after esophagectomy: a randomized controlled study.
Anesthesiology. 2006;105(5):911-919.
21. Cai H, Gong H, Zhang L, Wang Y, Tian Y.
Effect of low tidal volume ventilation on atelectasis
in patients during general anesthesia: a computed
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
PROTECTIVE VENTILATION AND LOWER TIDAL VOLUMES
tomographic scan. J Clin Anesth. 2007;19(2):125129.
22. Wolthuis EK, Veelo DP, Choi G, et al. Mechanical ventilation with lower tidal volumes does not influence the prescription of opioids or sedatives. Crit
Care. 2007;11(4):R77-R85.
23. Yilmaz M, Keegan MT, Iscimen R, et al. Toward
the prevention of acute lung injury: protocol-guided
limitation of large tidal volume ventilation and inappropriate transfusion. Crit Care Med. 2007;35
(7):1660-1666.
24. Determann RM, Wolthuis EK, Choi G, et al. Lung
epithelial injury markers are not influenced by use of lower
tidal volumes during elective surgery in patients without
preexistinglunginjury.AmJPhysiolLungCellMolPhysiol.
2008;294(2):L344-L350.
25. Lin WQ, Lu XY, Cao LH, Wen LL, Bai XH, Zhong
ZJ. Effects of the lung protective ventilatory strategy
on proinflammatory cytokine release during onelung ventilation [in Chinese]. Ai Zheng. 2008;27
(8):870-873.
26. Licker M, Diaper J, Villiger Y, et al. Impact of intraoperative lung-protective interventions in patients
undergoing lung cancer surgery. Crit Care. 2009;
13(2):R41-R50.
27. Determann RM, Royakkers A, Wolthuis EK, et al.
Ventilation with lower tidal volumes as compared with
conventional tidal volumes for patients without acute
lung injury: a preventive randomized controlled trial.
Crit Care. 2010;14(1):R1-R17.
28. Pinheiro de Oliveira R, Hetzel MP, dos Anjos Silva
M, Dallegrave D, Friedman G. Mechanical ventilation with high tidal volume induces inflammation in
patients without lung disease. Crit Care. 2010;
14(2):R39-R47.
29. Fernandez-Bustamante A, Wood CL, Tran ZV,
Moine P. Intraoperative ventilation: incidence and risk
factors for receiving large tidal volumes during general anesthesia. BMC Anesthesiol. 2011;11(1):
22-29.
30. Sundar S, Novack V, Jervis K, et al. Influence of
low tidal volume ventilation on time to extubation in
cardiac surgical patients. Anesthesiology. 2011;
114(5):1102-1110.
31. Yang M, Ahn HJ, Kim K, et al. Does a protective
ventilation strategy reduce the risk of pulmonary complications after lung cancer surgery? a randomized controlled trial. Chest. 2011;139(3):530-537.
32. Weingarten TN, Whalen FX, Warner DO, et al.
Comparison of two ventilatory strategies in elderly patients undergoing major abdominal surgery. Br J
Anaesth. 2010;104(1):16-22.
33. Esteban A, Anzueto A, Frutos F, et al; Mechanical Ventilation International Study Group. Character-
©2012 American Medical Association. All rights reserved.
istics and outcomes in adult patients receiving mechanical ventilation: a 28-day international study.
JAMA. 2002;287(3):345-355.
34. Fernández-Pérez ER, Keegan MT, Brown DR,
Hubmayr RD, Gajic O. Intraoperative tidal volume as
a risk factor for respiratory failure after
pneumonectomy. Anesthesiology. 2006;105(1):
14-18.
35. Checkley W, Brower R, Korpak A, Thompson BT;
Acute Respiratory Distress Syndrome Network
Investigators. Effects of a clinical trial on mechanical
ventilation practices in patients with acute lung injury.
Am J Respir Crit Care Med. 2008;177(11):12151222.
36. Mikkelsen ME, Dedhiya PM, Kalhan R, Gallop
RJ, Lanken PN, Fuchs BD. Potential reasons why physicians underuse lung-protective ventilation: a retrospective cohort study using physician documentation.
Respir Care. 2008;53(4):455-461.
37. Baudouin SV. Ventilator induced lung injury and
infection in the critically ill. Thorax. 2001;56(suppl 2):
ii50-ii57.
38. Lellouche F, Dionne S, Simard S, Bussières J,
Dagenais F. High tidal volumes in mechanically ventilated patients increase organ dysfunction after cardiac surgery. Anesthesiology. 2012;116(5):10721082.
JAMA, October 24/31, 2012—Vol 308, No. 16
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/18/2013
1659
Evolution of Mortality over Time in Patients
Receiving Mechanical Ventilation
Andrés Esteban1, Fernando Frutos-Vivar1, Alfonso Muriel2, Niall D. Ferguson3, Oscar Peñuelas1, Victor Abraira2,
Konstantinos Raymondos4, Fernando Rios5, Nicolas Nin1, Carlos Apezteguı́a5, Damian A. Violi6, Arnaud W. Thille7,
Laurent Brochard8, Marco González9, Asisclo J. Villagomez10, Javier Hurtado11, Andrew R. Davies12, Bin Du13,
Salvatore M. Maggiore14, Paolo Pelosi15, Luis Soto16, Vinko Tomicic17, Gabriel D’Empaire18, Dimitrios Matamis19,
Fekri Abroug20, Rui P. Moreno21, Marco Antonio Soares22, Yaseen Arabi23, Freddy Sandi24, Manuel Jibaja25, Pravin Amin26,
Younsuck Koh27, Michael A. Kuiper28, Hans-Henrik Bülow29, Amine Ali Zeggwagh30, and Antonio Anzueto31
1
Hospital Universitario de Getafe & Centro de Investigación Biomédica en red de Enfermedades Respiratorias, Spain; 2Department of Clinical
Biostatistics, Hospital Ramón y Cajal, Instituto Ramón y Cajal de Investigación Sanitaria & Centro de Investigación Biomédica en red de
Epidemiologı́a y Salud Pública, Spain; 3Interdepartmental Division of Critical Care Medicine, and Departments of Medicine and Physiology,
University of Toronto; 4Medizinische Hochschule Hannover, Germany; 5Hospital Nacional Alejandro Posadas, Buenos Aires, Argentina; 6Hospital
Interzonal General de Agudos Dr. Luis Güemes, Haedo, Argentina; 7Cenre Hospitalier Universitaire Henri Mondor, Creteil, France; 8Geneva
University Hospital, Geneva, Switzerland; 9Clı́nica Medellı́n & Universidad Pontificia Bolivariana, Medellı́n, Colombia; 10Hospital Regional 1o de
Octubre, Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado, México DF, México; 11Hospital de Clı́nicas de Montevideo,
Uruguay; 12Alfred Hospital & Monash University, Melbourne, Australia; 13Peking Union Medical College Hospital, Beijing, Popular Republic of China;
14
Policlinico “Agostino Gemelli,” Università Cattolica Del Sacro Cuore, Roma, Italy; 15Dipartimento di Scienze Chirurgiche e Diagnostiche Integrate,
Universita’ degli Studi, Genoa, Italy; 16Instituto Nacional del Tórax de Santiago, Chile; 17Clı́nica Las Lilas de Santiago, Chile; 18Hospital de Clı́nicas de
Caracas, Venezuela; 19Papageorgiou Hospital, Thessaloniki, Greece; 20Hospital Fattouma Bourguina, Monastir, Tunisia; 21Unidade de Cuidados
Intensivos Neurocrı́ticos, Hospital de São José, Centro Hospitalar de Lisboa Central, Lisbon, Portugal; 22Hospital Universitário São José, Belo
Horizonte, Brazil; 23King Saud Bin Abdulaziz University for Health Sciences, Riyadh, Saudi Arabia; 24Hospital Obrero No.1, La Paz, Bolivia; 25Hospital
Eugenio Espejo, Quito, Ecuador; 26Bombay Hospital Institute of Medical Sciences, Mumbai, India; 27Asan Medical Center, University of Ulsan, Seoul,
Republic of Korea; 28Medical Center Leeuwarden, Leeuwarden, The Netherlands; 29Holbak Hospital, Region Zealand University of Copenhagen, Denmark;
30
Hôpital Ibn Sina, Rabat, Morocco; and 31South Texas Veterans Health Care System and University of Texas Health Science Center, San Antonio, Texas
Rationale: Baseline characteristics and management have changed
over time in patients requiring mechanical ventilation; however, the
impact of these changes on patient outcomes is unclear.
Objectives: To estimate whether mortality in mechanically ventilated
patients has changed over time.
Methods: Prospective cohort studies conducted in 1998, 2004, and
2010, including patients receiving mechanical ventilation for more
than 12 hours in a 1-month period, from 927 units in 40 countries. To
examine effects over time on mortality in intensive care units, we
performed generalized estimating equation models.
(Received in original form December 5, 2012; accepted in final form May 1, 2013)
Supported by Centro de Investigación Biomédica en red Enfermedades Respiratorias, CIBER en Epidemiologı́a y Salud Pública. Instituto de Salud Carlos III,
Madrid, Spain, Instituto Ramón y Cajal de Investigación Sanitaria, Madrid, Spain,
and by a Canadian Institutes of Health Research (Ottawa, ON, Canada) New
Investigator Award (N.D.F.).
Author Contributions: A.E., as principal investigator, had full access to all the data
in the study and takes responsibility for the integrity of the data and the accuracy
of the data analysis. Study concept and design: A.E., F.F.-V., O.P., N.D.F., A.A.
Coordination of data acquisition: N.D.F., K.R., F.R., N.N., C.A., D.A.V., A.W.T., L.B.,
A.J.V., J.H., M.G., A.R.D., B.D., S.M.M., P.P., L.S., V.T., G.D., D.M., F.A., R.P.M.,
M.A.S., Y.A., F.S., M.J., P.A., Y.K., M.A.K., H.-H.B., A.A.Z., A.A. Analysis and interpretation of data: F.F.-V., O.P. Drafting of the manuscript: F.F.-V., O.P., N.D.F. Critical
revision of manuscript: A.E., K.R., F.R., N.N., C.A., D.A.V., A.W.T., L.B., A.J.V., J.H.,
M.G., A.R.D., B.D., S.M.M., P.P., L.S., V.T., G.D., D.M., F.A., R.P.M., M.A.S., Y.A.,
F.S., M.J., P.A., Y.K., M.A.K., H.-H.B., A.A.Z., A.A. Statistical analysis: A.M., V.A.
The funding organizations had no role in the design or conduct of the study,
collection, management, analysis, or interpretation of the data, or preparation,
review, or approval of the manuscript.
Correspondence and requests for reprints should be addressed to Andrés Esteban,
M.D., Ph.D., Intensive Care Unit, Hospital Universitario de Getafe, Carretera de
Toledo, Km 12,500, 28905 Madrid, Spain. E-mail: [email protected]
This article has an online supplement, which is accessible from this issue’s table of
contents at www.atsjournals.org
Am J Respir Crit Care Med Vol 188, Iss. 2, pp 220–230, Jul 15, 2013
Copyright ª 2013 by the American Thoracic Society
Originally Published in Press as DOI: 10.1164/rccm.201212-2169OC on April 30, 2013
Internet address: www.atsjournals.org
AT A GLANCE COMMENTARY
Scientific Knowledge on the Subject
Baseline characteristics and management have changed over
time in patients requiring mechanical ventilation; however,
the impact of these changes on patient outcomes is unclear.
What This Study Adds to the Field
The implementation of mechanical ventilation has significantly changed in the last decade. These changes have likely
resulted in a significant decrease in short-term mortality.
Measurements and Main Results: We included 18,302 patients. The
reasons for initiating mechanical ventilation varied significantly
among cohorts. Ventilatory management changed over time (P ,
0.001), with increased use of noninvasive positive-pressure ventilation (5% in 1998 to 14% in 2010), a decrease in tidal volume (mean
8.8 ml/kg actual body weight [SD ¼ 2.1] in 1998 to 6.9 ml/kg [SD ¼
1.9] in 2010), and an increase in applied positive end-expiratory
pressure (mean 4.2 cm H2O [SD ¼ 3.8] in 1998 to 7.0 cm of H2O [SD ¼
3.0] in 2010). Crude mortality in the intensive care unit decreased in
2010 compared with 1998 (28 versus 31%; odds ratio, 0.87; 95% confidence interval, 0.80–0.94), despite a similar complication rate. Hospital
mortality decreased similarly. After adjusting for baseline and management variables, this difference remained significant (odds ratio, 0.78;
95% confidence interval, 0.67–0.92).
Conclusions: Patient characteristics and ventilation practices have
changed over time, and outcomes of mechanically ventilated patients
have improved.
Clinical trials registered with www.clinicaltrials.gov (NCT01093482).
Keywords: mechanical ventilation; mortality; epidemiology; cohort study
Mechanical ventilation remains a cornerstone in the supportive
management of patients with acute respiratory failure, and is one
Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation
221
TABLE 1. COMPARISON OF GENERAL CHARACTERISTICS OF PARTICIPATING UNITS AND BASELINE CHARACTERISTICS
OF PATIENTS
Participating units, n
Medical, %
Medical–Surgical, %
Surgical, %
Neurological, %
Other, %
Number of beds, median (IQR)
Proportion of mechanically ventilated patients, %
Patients included, n
Geographical area, n (%)
Europe
USA/Canada
Latin America
Africa
Asia
Oceania
Age, mean (SD), yr
Female sex, n (%)
Weight, mean (SD), kg
Body mass index, mean (SD), kg/m2
Simplified Acute Physiology Score II on admission,
mean (SD), points
Previous tracheotomy, n (%)
Noninvasive ventilation at home, n (%)
Bilevel positive airway pressure, n (%)
Continuous positive airway pressure, n (%)
Main reason for mechanical ventilation, n (%)†
Chronic obstructive pulmonary disease
Asthma
Other chronic pulmonary disease
Neurologic disease
Metabolic, n (% of neurologic)
Overdose/intoxication, n (% of neurologic)
Hemorrhagic stroke, n (% of neurologic)
Ischemic stroke, n (% of neurologic)
Brain trauma, n (% of neurologic)
Other cause, n (% of neurologic)
Neuromuscular disease
Acute respiratory failure
Postoperative
Pneumonia
Community acquired, n (% of pneumonia)
Hospital acquired, n (% of pneumonia)
Sepsis
Acute respiratory distress syndrome
Congestive heart failure
Cardiac arrest
Trauma
Aspiration
Other cause of acute respiratory failure
Cohort 1998
Cohort 2004
Cohort 2010
361
19
77
4
—
—
10 (8, 14)
33%
5,183
349
16
68
13.5
2
0.4
10 (8, 15)
25%
4,968
494
17
73
5
1.5
3
12 (9, 19)
35%
8,151
2,387
1,455
1,222
119
(46)
(28)
(24)
(2)
0
0
59 (17)
1,985 (39)
72 (17)
n.d.
44 (17)
102
73
38
35
(2)
(1)
(52)
(48)
2,133
1,341
1,306
110
78
59
1,967
76
27
42
(43)
(27)
(26)
(2)
(2)
0
(17)
(40)
(20)
(7)
(18)
123 (2)
n.d.
P Value*
,0.001
,0.001
,0.001
(42)
(11)
(21)
(2)
(15)
(9)
(17)
(38)
(20)
(6.5)
(18)
,0.001
180
213
81
132
(2)
(3)
(38)
(62%)
0.22
,0.001
3,407
929
1,692
167
1,242
714
61
3,105
75
27
45
,0.001
0.23
,0.001
0.122
,0.001
522 (10)
79 (2)
60 (1)
864 (17)
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
94 (2)
267
63
85
938
129
148
310
119
228
4
58
(5)
(1)
(2)
(19)
(14)
(16)
(33)
(13)
(24)
(0.4)
(1)
524
98
142
1,574
265
211
470
214
302
112
74
(6)
(1)
(2)
(19)
(17)
(13)
(30)
(13.5)
(19)
(7)
(1)
,0.001
0.26
0.019
,0.001
1,080 (21)
721 (14)
n.d.
n.d.
458 (9)
231 (5)
539 (10)
100 (2)
407 (8)
129 (3)
367 (7)
1,053
528
376
152
449
148
285
239
284
139
432
(21)
(11)
(71)
(29)
(9)
(3)
(6)
(5)
(6)
(3)
(9)
1,750
819
539
280
726
281
617
473
367
200
506
(21)
(10)
(66)
(34)
(9)
(3)
(8)
(6)
(4.5)
(2.5)
(6)
0.68
,0.001
,0.001
0.93
,0.001
,0.001
,0.001
,0.001
0.44
,0.001
Definition of abbreviations: IQR ¼ interquartile range; n.d. ¼ no data available.
* Chi-square tests were used for the comparison of categorical variables between groups; Student’s t test and ANOVA were used for the
comparison of continuous variables.
y
Because of rounding, percentages may not total 100. In 1998, more than one cause of acute respiratory failure per patient was permitted.
of the defining interventions of intensive care medicine as a specialty. Basic science, translational, and physiological studies have
all informed clinical trials, which, in turn, have shown significant
impact in areas such as avoidance of ventilator-induced lung injury (1–5), liberation from mechanical ventilation (6–10), and
improved outcomes using noninvasive positive-pressure ventilation (11, 12). Little is known, however, about the impact of
these interventions on mortality in a real-world setting.
In 1998 (13) and 2004 (14), we performed two observational
studies on the use of mechanical ventilation and its associated
outcomes. In the first, we focused on describing mortality outcomes and identifying factors that impacted survival (13). In the
second, we systematically reviewed clinical trials published in the
prior 6 years and examined the implementation of this published
evidence (14). The main finding of this study was the increasing use
of noninvasive positive-pressure ventilation, a reduction in tidal
volumes for patients with acute respiratory distress syndrome
(ARDS), and more pressure support and less synchronized intermittent mandatory ventilation being used for liberation from mechanical ventilation. Despite these encouraging findings, our data
failed to demonstrate a significant improvement in outcomes. Possible reasons for these results include lack of statistical power, or
a true lack of outcome improvements due to differences in intensive care unit (ICU) admission patterns over time, an insufficient
magnitude of practice change, or a lack of efficacy of these therapies in unselected populations outside the setting of a clinical trial.
222
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
VOL 188
2013
TABLE 2. COMPARISON OF VARIABLES RELATED TO MANAGEMENT
Noninvasive positive-pressure ventilation before admission in
the ICU, n (%)
Noninvasive positive-pressure ventilation at admission in the
ICU, n (%)
Mode of ventilation, days of use per 1,000 d of invasive
mechanical ventilation (excluding days during weaning
from mechanical ventilation process)
Assist-control
SIMV
SIMV-PS
Pressure support
PCV
APRV/BIPAP
PRVC
Other mode
Ventilator settings over the course of invasive ventilation
Tidal volume, mean (SD)
ml/kg actual body weight
ml/kg predicted body weight
Positive end-expiratory pressure, cm H2O, mean (SD)
Total respiratory rate, breaths per minute, mean (SD)
Sedation, n (%)
Analgesia, n (%)
Neuromuscular blocking, n (%)
Liberation from mechanical ventilation
Met criteria, n/eligible patients (%)†
Scheduled extubation, n/patients met criteria (%)
Unplanned extubation, n/patients at risk (%)‡
Noninvasive positive-pressure ventilation after extubation,
n/eligible patients (%)x
Reintubation, n/ patients at risk (%)jj
After scheduled extubation
After unplanned extubation
Tracheotomy, n/patients at risk (%)¶
Percutaneous, n (%)
Cohort 1998 (n ¼ 5,183)
Cohort 2004 (n ¼ 4,968)
n.d.
n.d.
256 (5)
479 (10)
627
66
132
65
75
n.d.
n.d.
35
8.8
n.d.
4.2
18
3,164
2,241
686
412
21
132
125
125
66
98
21
(2.1)
(3.8)
(11)
(61)
(43)
(13)
3,640/5,008 (73)
2,858/3,640 (78.5)
179/4,906 (4)
n.d.
424/3,037
350/2,858
74/179
546/4,906
n.d.
(14)
(12)
(41)
(11)
7.6
9.3
5.4
18
3,486
n.d.
524
(2.1)
(2.3)
(4.6)
(6)
(70)
3,005/4,682
2,714/3,005
135/4,559
308/2,849
(64)
(90)
(3)
(11)
319/2,849
278/2,714
41/135
664/4,559
362
(10.5)
(11)
(10)
(30)
(15)
(54.5)
Cohort 2010 (n ¼ 8,151)
P Value*
429 (5)
—
1,169 (14)
,0.001
330
34
94
237
130
91
61
23
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
6.9
8.2
7.0
19
5,755
5,043
890
(1.9)
(2.0)
(3.0)
(6)
(71)
(62)
(11)
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
5,111/7,323
4,151/5,111
654/7,143
473/4,805
(70)
(81)
(9)
(10)
,0.001
,0.001
,0.001
0.17
643/4,805
511/4,151
132/654
993/7,143
526
(13)
(12)
(20)
(14)
(53)
0.005
0.001
,0.001
,0.001
0.53
Definition of abbreviations: APRV/BIPAP ¼ airway pressure release ventilation/biphasic positive airway pressure; ICU ¼ intensive care unit; PCV ¼ pressure-controlled
ventilation; PRVC ¼ pressure regulated volume control; PS ¼ pressure support; n.d. ¼ no data available; SIMV ¼ synchronized intermittent mandatory ventilation.
* Chi-square tests were used for the comparison of categorical variables between groups; ANOVA was used for the comparison of continuous variables.
y
Eligible patients for criteria for liberation from mechanical ventilation: overall cohort excepting patients with successful noninvasive positive-pressure ventilation.
z
Patients at risk for unplanned extubation: overall cohort excepting patients with a previous tracheotomy and patients with successful noninvasive positive-pressure ventilation.
x
Eligible patients for noninvasive positive-pressure ventilation after extubation: scheduled and unplanned extubated patients.
jj
Patients at risk for reintubation: scheduled and unplanned extubated patients.
¶
Patients at risk for tracheotomy: overall cohort excepting patients with a previous tracheotomy and patients with successful noninvasive positive-pressure ventilation.
At 6 years after the last study, we have now conducted a new
international study with the primary objective of estimating
whether mortality in mechanically ventilated patients has changed
over time. We reasoned that combining this new study with our
two prior databases would give us more statistical power to demonstrate potential improvements in survival. Furthermore, allowing more time for incorporation of previous research findings into
clinical practice, and performing a multivariable analysis that
accounted for potential changes in case mix over time should improve the study importance. Our secondary aims were to describe
the changes in the baseline and management characteristics of
ventilated patients over this 12-year period.
METHODS
Design
As in previous studies (13, 14), we conducted a prospective use review
of patients receiving invasive mechanical ventilation for at least 12 hours
or noninvasive positive-pressure ventilation for at least 1 hour during a
1-month period starting in March 2010. National coordinators recruited
local investigators (see full list of Investigators in the Third International
Study on Mechanical Ventilation (2010) at the end of this article for 2010
and the online supplement for 2004 and 1998) from eligible ICUs (see
Table E1 in the online supplement). Only the investigator and research
coordinators at each site were aware of the exact purpose and timing of
the study to minimize practice changes in response to observation. The
research ethics board of each participating institution approved the protocol and waived the need for informed consent.
Protocol
We collected baseline characteristics, and daily ventilator, gas exchange,
clinical management, and complications data while patients were ventilated or until Day 28. Patients were followed in hospital for mortality
and length-of-stay outcomes.
For this analysis, we combined these 2010 data with those from our
studies in 1998 and 2004 (13, 14). Detailed descriptions of the variables
collected in each study, along with their definitions, are shown in
Tables E2 and E3. One notable variable not collected in 1998 was
height; we therefore express tidal volume as ml/kg actual body weight
for comparisons across all three studies.
Statistical Analysis
Data are expressed as mean (6SD), median (interquartile range), and
absolute and relative frequencies, as appropriate. ANOVA or KruskalWallis tests were used to compare continuous variables, and Chi-square
tests were used for categorical variables. We rejected the null hypothesis of no difference among cohorts at a nominal significance level of
Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation
0.05. These analyses were performed using SPSS 17.0 (SPSS Inc., Chicago, IL).
To estimate the effect of the year of study on ICU mortality, we used
generalized estimating equation models with a binomial distribution and
logit link function to measure the impact of time (by cohort year), accounting for the effects related to the clustering of patients receiving
care from the same ICU (15). For the purpose of the analysis,
the overall population of patients from the three studies (n ¼ 18,302)
was randomly divided into two sets: modeling group (n ¼ 13,644) and
validation group (n ¼ 4,658). A first generalized estimating equation
estimative model was derived entering the following variables: year of
study (reference year, 1998), geographical area (Europe, USA–
Canada, Latin American, Africa, Asia, and Oceania), age, sex, ICU
characteristics (type of unit [medical, surgical, medical–surgical, neurological, other] and size of the unit [fewer than 10 beds, 10–20 beds,
more than 20 beds]), severity at ICU admission (Simplified Acute
Physiology Score II), and reason for mechanical ventilation. We built
on this model by entering daily variables related to management over the
course of ventilation: type of ventilation (invasive or noninvasive), inspired fraction of oxygen, positive end-expiratory pressure (PEEP), tidal
volume (ml/kg actual body weight), prone position (yes/no), minute ventilation, use of continuous intravenous sedatives (yes/no), and use of
neuromuscular blockers (yes/no). These two models were then evaluated
for consistency of effects in the validation set. As a sensitivity analysis,
we ran these models in the cohort of patients included from ICUs that
participated in all three studies. These analyses were performed using
Stata Software 11.0 (StataCorp LP, College Station, Texas).
RESULTS
Characteristics of Participating ICUs and Included Patients
The characteristics of the 927 participating ICUs and 18,302
patients across the studies are shown in Table 1. These include
2,913 patients from 55 ICUs that participated in all three studies. Further details regarding the ICU distribution by country
across studies are shown in Table E4.
Management during Mechanical Ventilation
Variables related to management in the ICU are shown in Table 2.
The use of noninvasive positive-pressure ventilation increased
significantly over the 12-year period: 5% in 1998, 10% in 2004,
and 14% in 2010 (P , 0.001). Meanwhile, the median duration of
noninvasive positive-pressure ventilation decreased over time
(3 d [interquartile range, 2–6] in 1998 versus 2 d [interquartile
range, 2–4] in 2004 and 2 d [interquartile range, 1–3] in 2010;
P , 0.001). The need for intubation among these patients was
not higher in 2010: 32% in 1998, 40% in 2004, and 29% in 2010
(P , 0.001).
During invasive ventilation, volume-cycled assist-control ventilation remained the most common ventilator mode, but its use
223
has decreased over time in favor of other modes, particularly
pressure support (Table 2; see also Figure E1). Among all
patients, we observed a significant decrease in tidal volume
and an increase in PEEP (P , 0.001) (Table 2).
In the subgroup of patients with ARDS, tidal volume
(expressed as ml/kg actual body weight) was also significantly
(P , 0.001) reduced over time (Figure 1). The proportion of
patients receiving a ventilation strategy with pressure/volume
limitation (arbitrarily defined as tidal volume below 6 ml/kg
actual body weight; or tidal volume below 8 ml/kg actual body
weight and plateau or peak inspiratory pressure less than 30 cm
H2O [13]) increased significantly over time: 29% in 1998, 57%
in 2004, and 67.5% in 2010 (P , 0.001). In addition, PEEP
levels increased significantly (P , 0.001; Figure 2). The use of
adjunctive therapies, such as prone positioning, was infrequent,
and remained stable: 9% in 1998, 5% in 2004, and 7% in 2010
(P ¼ 0.07).
A process of weaning or liberation from mechanical ventilation began in 73% of patients in 1998, 64% in 2004, and 70% in
2010 (Table 2). The proportion of patients who successfully
completed their first attempt of liberation from mechanical ventilation increased over time (49% in 1998, 55% in 2004, and
63.5% in 2010; P , 0.001). Among patients who failed their first
attempt at weaning, there was a significant increase in the subsequent use of pressure support as the weaning mode, with
a concomitant significant reduction in the use of synchronized
intermittent mandatory ventilation with or without pressure
support (Table E5). The duration of weaning in these patients,
although statistically different (P , 0.001), was not clinically
different among the three time periods (median, 3 [interquartile
range, 2–5] d in 1998, 3 [interquartile range, 2–4] d in 2004, and
3 [interquartile range, 2–5] d in 2010).
The rate of reintubation after scheduled extubation remained
similar over time, at around 12% (Table 2). This occurred most
commonly within the first 12 hours after extubation (63%
of reintubations in 1998, 57% in 2004, and 52% in 2010;
P ¼ 0.01).
More tracheotomies were performed in 2004 (15%) and
2010 (14%) than in 1998 (11%) (P , 0.001). Median time to
tracheotomy from intubation, however, remained stable (12
[interquartile range, 7–17] d in 1998, 9 [interquartile range,
6–14] d in 2004, and 11 [interquartile range, 6–16] d in 2010
(P ¼ 0.73).
Outcomes
A comparison of patient outcomes across the studies is shown in
Table 3. Half of the patients included had at least one complication over the course of mechanical ventilation (50% in 1998,
Figure 1. Evolution of the set tidal volume (ml/kg actual
body weight) in patients with criteria of acute respiratory
distress syndrome (ARDS). Left: mean (6SD); P , 0.001 for
all comparisons with ANOVA test. Right: proportion of the
time of mechanical ventilation in each category of tidal
volume; P , 0.001 for all comparisons with chi-square test.
224
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
VOL 188
2013
Figure 2. Evolution of the applied positive end-expiratory
pressure (PEEP; cm H2O) in patients with criteria of acute
respiratory distress syndrome (ARDS). Left: mean (6SD);
P , 0.001 for comparisons, with ANOVA, between 1998
versus 2004 and 1998 versus 2010; P ¼ 0.05 for comparison, with ANOVA, between 2004 and 2010. Right: proportion of the time of mechanical ventilation ventilated
in each category of PEEP. P , 0.001 for all comparisons
with chi-square test. ZEEP ¼ zero positive end-expiratory
pressure.
48% in 2004, and 54% in 2010; P , 0.001). We note higher rates
of sepsis and cardiovascular failure in 2010, and a reduction
over time in ventilator-associated pneumonia. We observed significant differences in the duration of ventilatory support, ICU,
and hospital length of stay.
Analysis of Mortality in the ICU
There was a decline in the crude ICU mortality, from 31% in
1998 to 28% in 2010 (odds ratio, 0.87; 95% confidence interval,
0.80–0.94; P , 0.001).
After adjustment for baseline variables, mortality in 2010
remained lower than in 1998 (odds ratio, 0.74; 95% confidence
interval, 0.66–0.84; Table 4). This result was similar after management variables were added to the model (odds ratio, 0.78;
95% confidence interval, 0.67–0.92). These results were quantitatively similar in the validation group (final model odds ratio,
0.86; 95% confidence interval, 0.67–1.09; Table 4).
The observed mortality in the 2,913 patients from ICUs that
participated in all three studies showed a similar trend: 34% in
1998, 34% in 2004, and 27% in 2010 (P ¼ 0.002). Sensitivity
analysis of the final model showed robust results in this subgroup, with a final adjusted odds ratio for 2010 ICU mortality
of 0.58 (95% confidence interval, 0.45–0.75) relative to 1998.
DISCUSSION
Our analysis of these international multicenter studies suggests
that, in mechanically ventilated patients, short-term mortality
has decreased over time. Although many factors may have contributed to these results, several factors related to mechanical
ventilation management may have contributed, including reducing the need for invasive mechanical ventilation, and improving
both mechanical ventilation safety and the process of liberation
from mechanical ventilation. This study confirms our previous
work showing that changes in mechanical ventilation practices
continue to follow findings of positive clinical trials, and shows,
for the first time, a significant association with reduced mortality.
We have observed a significant change of mortality over time
after adjustment for case mix and for variables related to ventilatory settings. Despite a similar proportion of patients with complications related to mechanical ventilation and organ dysfunctions,
the intensive care mortality in 2010 was lower than in previous
studies. The most likely explanation is that the overall management of mechanically ventilated patients has improved over the
last 12 years. The implementation of different integrated management strategies for precipitating conditions, such as sepsis (16),
protocols related to ventilator management, including sedation
(10, 17, 18), and incorporating bundles of care to prevent
TABLE 3. COMPARISON OF EVENTS EMERGING OVER THE COURSE OF MECHANICAL VENTILATION AND IN THE OUTCOMES
Cohort 1998
(n ¼ 5,183)
Events emerging over the course of mechanical ventilation
Acute respiratory distress syndrome, n (%)
Acquired ICU pneumonia, n (%)
Sepsis, n (%)
Barotrauma, n (%)
Cardiovascular failure, n (%)
Renal failure, n (%)
Hepatic failure, n (%)
Hematological failure, n (%)
Outcomes
Duration of ventilatory support†, median (IQR), days
Length of stay in the ICU, median (IQR), days
Length of stay in the hospital‡, median (IQR) days
Mortality, n (% [95% confidence interval])
In the ICU
At Day 28 after admission in the ICU
In the hospital‡
218
438
457
154
1,145
971
326
552
(4)
(8.5)
(9)
(3)
(22)
(19)
(6)
(11)
4 (2, 8)
7 (4, 14)
16 (9,29)
1,590 (31 [29–32])
1,719 (33 [32–34])
1,876 (40 [38–41])
Cohort 2004
(n ¼ 4,968)
279
265
400
157
1,193
948
691
795
(6)
(5)
(8)
(3)
(24)
(19)
(14)
(16)
6 (3, 11)
8 (4, 15)
17 (9,31)
1,533 (31 [29–32])
1,605 (32 [31–34])
1,759 (37 [36–38])
Cohort 2010
(n ¼ 8,151)
P Value*
(6)
(4)
(18)
(2)
(39)
(22)
(7)
(8)
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
,0.001
5 (3, 10)
7 (4, 14)
17 (9,31)
,0.001
,0.001
0.002
495
359
1,473
140
3,145
1,775
555
662
2,269 (28 [27–29])
2,445 (30 [29–31])
2,733 (35 [34–36])
,0.001
,0.001
,0.001
Definition of abbreviations: ICU ¼ intensive care unit; IQR ¼ interquartile range.
* Chi-square tests were used for the comparison of categorical variables between groups; ANOVA or Kruskall-Wallis test was used for the comparison of continuous
variables.
y
Including duration of liberation from mechanical ventilation in the extubated patients.
z
Patients whose date and status at discharge from hospital was unknown were not included in the calculation (465 patients in 1998, 211 patients in 2004, and 377
patients in 2010).
Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation
TABLE 4. ESTIMATION OF RELATION OF YEAR OF STUDY
TO MORTALITY IN THE INTENSIVE CARE UNIT
Study Year
Crude odds ratio
1998
2004
2010
After adjustment for ICU and baseline
variables
1998
2004
2010
After adjustment for ICU and baseline
variables and for daily management
variables
1998
2004
2010
Modeling Group
(n ¼ 13,644)
Validation Group
(n ¼ 4,658)
1
1.01 (0.90–1.13)
0.83 (0.75–0.93)
1
1.00 (0.82–1.20)
0.90 (0.75–1.07)
1
1.01 (0.89–1.15)
0.74 (0.66–0.84)
1
1.05 (0.90–1.24)
0.78 (0.67–0.92)
1
1.05 (0.90–1.24)
0.78 (0.67–0.92)
1
1.06 (0.83–1.34)
0.86 (0.67–1.09)
Definition of abbreviation: ICU ¼ intensive care unit.
Data presented are odds ratios (95% confidence intervals).
nosocomial infections (19, 20) could all possibly contribute to
these findings.
Throughout our studies, we observed a significant decrease in
delivered tidal volume, both in the general mechanically ventilated population and particularly among those with ARDS. This
lung-protective strategy has been evaluated in several randomized clinical trials (1, 2, 21–23), the largest of which showed
reduced mortality when tidal volume and plateau pressure were
controlled in patients with acute lung injury (2). These data
suggest that routine implementation of a pressure–volume limited ventilation strategy in clinical practice could reduce mortality in these patients. In a recent observational ARDS study
(24), adherence to lung-protective ventilation was associated
with a 3% decrease in the mortality risk at 2 years (hazard ratio,
0.97; 95% confidence interval, 0.95–0.99; P ¼ 0.002). These
results, like ours, suggest that these clinical trial results are
robust even applied to a less selective population, as in usual
clinical practice. There are also suggestions that lung-protective
strategies may be important in patients who do not yet have
ARDS (25–28). The fact that the odds ratio for mortality moved
closer to 1 in all datasets after accounting for management variables, including tidal volume, is in keeping with the hypothesis
that lung-protective ventilation is a key intervention across
a broad range of mechanically ventilated patients, not just those
with ARDS.
PEEP is another essential component of the ventilatory strategy in patients with ARDS. From the first report of this syndrome (29), it has been observed that PEEP can improve
oxygenation by keeping recruited alveoli open and decreasing
intrapulmonary shunt (30). In addition, PEEP-induced lung recruitment can decrease alveolar overdistention, because the volume of each subsequent tidal breath is shared by more open
alveoli. PEEP may also decrease repetitive alveolar opening
and closing during the respiratory cycle, thereby preventing
lung injury (31). Observational studies have reported an independent association between zero PEEP and mortality (32)
in a heterogeneous cohort of patients who were mechanically
ventilated and between low values of PEEP and mortality in
patients with ARDS (33). An individual patient data metaanalysis (3, 4) from randomized trials (1, 34–38) comparing
higher versus lower PEEP levels while controlling tidal volume
in patients with acute lung injury overall showed no statistically
significant difference in hospital mortality (random effects
model relative risk, 0.85; 95% confidence interval, 0.71–1.01).
225
However, in patients with moderate to severe ARDS (39),
higher levels of PEEP were associated with a significant 4%
absolute mortality reduction (40). In our analysis, we observed
a marked reduction in the use of zero end-expiratory pressure
and a significant increase in the use of higher PEEP, especially in
patients with ARDS. Nevertheless, the mean PEEP level used in
the first week of ventilatory support in 2010 (9.3 6 3.3 cm H2O)
was still lower than that set in the higher PEEP groups of some
randomized clinical trials.
The use of noninvasive positive-pressure ventilation at ICU
admission has almost doubled every 6 years. This finding is similar to that reported by Chandra and colleagues (41), who observed
a 46% increase in noninvasive positive-pressure ventilation use
and a 42% decline in invasive mechanical ventilation use between
1998 and 2008 in patients with exacerbations of chronic obstructive pulmonary disease. Encouragingly, despite this increased use,
the proportion of patients intubated after failing a trial of noninvasive positive-pressure ventilation has remained stable over time.
An important observation is that patients who required invasive
ventilation after noninvasive positive-pressure ventilation had
higher mortality than patients who were directly intubated. This
finding was described previously by our group (13, 14, 42) and
others (43), and has several possible explanations. First, failure
of noninvasive positive-pressure ventilation could identify
patients who are difficult to ventilate or who have a higher
severity of illness. In our study, patients who failed noninvasive
positive-pressure ventilation had significantly higher Simplified
Acute Physiology Score II scores (mean ¼ 42) than successful
patients (mean ¼ 37), but they had lower scores than patients
invasively ventilated from the beginning (mean ¼ 45). Another
explanation is that failing noninvasive ventilation is, in itself,
harmful, and that delayed recognition of failure may exacerbate
this harm (41).
Our study has several limitations. This is an analysis of prospectively collected clinical data from a wide variety of ICUs,
patient conditions, and clinical practices, some of which are related to outcome (43). We have performed an analysis accounting for clustering by the ICUs, and we have included ICU type
and size in the model to try to minimize the possible bias related
to those variables. Although we used multivariable modeling,
unmeasured confounders may remain, which could have impacted our results. We constructed an estimative model to
evaluate the decreasing ICU mortality in the last 12 years, and
we have outlined plausible explanations for this. This model has
been validated and adjusted, and the results appear consistent.
However, it is not possible to make a causal inference model
from our observational data. In addition, we have no information about withdrawal and/or withholding of life support and its
possible impact on our results. Another limitation is that we did
not have all the same ICUs participating in all three studies. To
some extent, this was unavoidable due to closure of some previously participating hospitals, changes in resources available
for data collection at participating centers, and our goal to expand the generalizability of results by including ICUs from
more countries.
In conclusion, our analysis shows that the implementation of
mechanical ventilation has significantly changed in the last decade; we speculate that these changes have resulted in the significant decrease in short-term mortality.
Author disclosures are available with the text of this article at www.atsjournals.org.
Investigators in the Third International Study on Mechanical Ventilation
(2010):
Argentina: Coordinators: Fernando Rı́os (Hospital Nacional Alejandro Posadas),
Damian Violi (Hospital Hospital Interzonal General de Agudo Profesor Dr .Luis
Guemes, Haedo). Marisol Rodrı́guez-Goñi, Roger Lamoglie and Fernando Villarejo
226
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
(Hospital Nacional Profesor A. Posadas, Buenos Aires); Norberto Tiribelli and
Santiago Ilutovich (Sanatorio de la Trinidad, General Mitre); Matı́as Brizuela
and Mariana Monllau (Hospital Tránsito Cáceres de Allende, Córdoba); Fernando
Saldarini and Silvina Borello (Hospital General de Agudos donación Francisco
Santojanni, Buenos Aires); Alberto Marino and Norberto Tiribelli (Hospital
Churruca-Visca, Buenos Aires); Mauricio Vinzio and Karina Bonasegla (Sanatorio
de la Trinidad, San Isidro); Julián Hernández and Marı́a Belén Yapur (Sanatorio
Nuestra Señora del Rosario, Jujuy); Marı́a Eugenia González (Hospital Privado de
Comunidad, Mar del Plata); Sebastián E. Mare (Sanatorio Dr. Julio Méndez, Buenos
Aires); Judith Sagardı́a and Marco Bezzi (Hospital General de Agudos P. Piñeiro,
Buenos Aires); Cecilia Pereyra and Julian Strati (Hospital Interzonal General de
Agudo Profesor Dr. Luis Guemes, Haedo); Daniel Vargas and Claudia Diaz (Hospital
Pablo Soria, Jujuy); Pablo Gómez and Marcelo Palavecino (Sanatorio Juncal, Temperley);
Graciela Elizabeth and Aguilera Garcı́a, M. Eugenia (Hospital de San Luis); Luis Pablo
Cardonnet and Lisandro Betttini (Hospital Provincial del Centenario, Rosario); Hernán
Nuñez and Lucas Vallejo (Hospital General de Agudos Juan A. Fernández, Buenos Aires);
Fernando Fernández and Jorge Arroyo (Hospital Central, Mendoza); Daniel Duarte and
Gerardo Filippa (Hospital Regional Rı́o Grande, Tierra del Fuego); Cayetano Galetti and
Hernan Nunia (Sanatorio Allende, Córdoba); Fernando Lambert and Elisa Estenssoro
(Hospital Interzonal de Agudos San Martin, La Plata); Marina Busico and Fernando
Villarejo (Clı́nica Olivos, Vicente López); Javier Horacio Álvarez (Hospital Universitario
Austral, Pilar); Alejandro Raimondi and Gustavo Badariotti (Sanatorio Mater Dei, Buenos
Aires); Martı́n Lugaro (Sanatorio Las Lomas, San Isidro); Fernando Lipovestky (Clı́nica
Santa Isabel; Buenos Aires); Alan Javier Zazu and Hugo Capponcelli (Clı́nica Privada de
Especialidades de Villa Marı́a); Patricia Vogl and Cristina Orlandi (Hospital Zonal
Francisco López Lima, General Roca); Alejandro Gómez and Gustavo Jannello
(Sanatorio de los Arcos; Buenos Aires); Alejandro Risso (Sanatorio Otamendi y
Miroli, Buenos Aires); Leticia Rapetti and Guillermo Chiappero (Hospital Universitario,
Universidad Abierta Interamericana, Buenos Aires); Juan Domingo Fernández (Hospital
Regional de Comodoro Rivadavia, Chubut); Rodrigo E. Gómez-Paz (Hospital Español,
Buenos Aires); Marcos Juan Zec and Pascual Valdez (Hospital General de Agudos
Dalmacio Vélez Sársfield, Buenos Aires); Jorgelina Guyon and Ariel Chena (Hospital
Lagomaggiore, Mendoza); Sergio Lasdica (Hospital Municipal Coronel Suárez, Buenos
Aires); Martin Deheza and Schimdt Alejandra (Hospital General de Agudos Bernardino
Rivadavia, Buenos Aires); Francisco Criado (Hospital Naval Puerto Belgrano, Bahı́a
Blanca); Norma Beatriz Márquez (Policlı́nico Atlántico del Sur, Rı́os Gallegos); Pablo
Desmery and José Luis Scapellato (Sanatorio Anchorena, Buenos Aires); Gonzalo Javier
Rı́os and Cristian Casabella (Clı́nica Bazterrica, Buenos Aires)
Australia: Coordinators: Jasmin Board and Andrew R. Davies (Alfred Hospital,
Melbourne). Andrew Bersten, Elisha Matheson, and Amy Waters (Flinders Medical
Center, Adelaide); John Santamaria and Jennifer Holmes (St. Vincent’s Hospital,
Melbourne); Cartan Costello, Manoj K Saxena, and John Myburgh (St. George
Hospital, Sydney); Ellen Kinkel and Forbes McGain (The Western Hospital, Melbourne);
Claire Cattigan and Allison Bone (Barwon Health, Geelong Hospital, Geelong); Ian
Seppelt, Leonie Weisbrodt and Cheryl Cuzner (Nepean Hospital, Sydney); Christopher
MacIsaac, Deborah Barge and Tania Caf (Royal Melbourne Hospital, Melbourne);
Cameron Knott and Graeme Duke (The Northern Hospital, Melbourne); Imogen
Mitchell, Helen Rodgers, Rachel Whyte, and Elisha Fulton (Canberra Hospital,
Canberra); Jasmin Board, Andrew Davies, and Alistair Nichol (Alfred Hospital,
Melbourne); Hergen Buscher, Priya Nair, and Claire Reynolds (St. Vincent’s Hospital, Sydney); Simon J. G. Hockley, Ian Moore, and Katherine Davidson (Calvary
Wakefield Hospital, Adelaide); David Milliss, Raju Pusapati, and Helen Wong
(Concord Hospital, Sydney); Jason Fletcher and Julie Smith (Bendigo Hospital,
Bendigo); Paul Goldrick, Dianne Stephens, and Jane Thomas (Royal Darwin Hospital, Darwin); Anders Aneman, Sutrisno Gunawan, and Tom Cowlam (Liverpool
Hospital, Liverpool); George Lukas and Rick McAllister (Royal Hobart Hospital,
Hobart); Minka Springham, Joanne Sutton, and Jeff Presneill (Mater Health Services,
Brisbane); Tony Sutherland and Dianne Hill (Ballarat Health Services, Ballarat);
Howard Connor, Jenny Dennett, and Tim Coles (Central Gippsland Hospital, Sale).
Bolivia: Coordinator: Freddy Sandy (Hospital Obrero No. 1, La Paz). Sandro Chavarria
and Marcelo Choque (Hospital Obrero No. 1, La Paz); Ronald Pairumani and Juan
Guerra (Instituto Gastroenterológico boliviano japonés, Santa Cruz).
Brazil: Coordinator: Marco Antônio Soares Reis (Hospital Universitário São José,
Belo Horizonte). José Carlos Versiani (Hospital Madre Teresa, Belo Horizonte);
Eduardo Fonseca Sad (Hospital Luxemburgo, Belo Horizonte); Maria Aparecida
Braga (Hospital Dia e Maternidade Unimed-BH, Belo Horizonte); Dinalva Aparecida
Gomes (Hospital Vera Cruz, Belo Horizonte); Fernando Antônio Botoni (Hospital de
Pronto Socorro Risoleta Tolentino Neves, Belo Horizonte); Maurı́cio Meireles Góes
(Hospital da Baleia, Belo Horizonte); Frederico Costa Val Barros (Hospital da Polı́cia
Militar, Belo Horizonte); Rogério de Castro Pereira (Hospital Felı́cio Rocho, Belo
Horizonte); Hugo Urbano (Hospital Vila da Serra, Belo Horizonte); Valéria de
Carvalho Magela (Hospital Santa Rita, Contagem); Aline Camile Yehia (Hospital
Júlia Kubitschek, Belo Horizonte); Bruno Bonaccorsi Fernandino (Hospital São
Francisco-Setimig, Belo Horizonte); Marco Antônio Ribeiro Leão (Hospital São
João de Deus, Divinópolis); Rovı́lson Lara (Hospital Arnaldo Gavazza Filho, Ponte
Nova); Rovı́lson Lara (Hospital São João Batista, Viçosa); Rubens Altair Amaral de
Pádua (Hospital Vaz Monteiro, Lavras); Janine Dias Alves (Santa Casa de Misericórdia
de Ouro Preto, Ouro Preto); Aloı́sio Marques do Nascimento (Hospital Nossa
Senhora das Graças, Sete Lagoas); Bruno do Valle Pinheiro (Hospital Universitário da
Universidade Federal de Juiz de Fora, Juiz de Fora); Carlos Alberto Studart Gomes
(Hospital de Messejana, Fortaleza); Marcelo Alcântara Holanda (Hospital
Universitário Walter Cantı́dio, Fortaleza); Frederico Rodrigues Anselmo
(Hospital Nossa Senhora Aparecida, Belo Horizonte).
VOL 188
2013
Canada: Coordinator: Niall D. Ferguson (Mount Sinai Hospital and University
Health Network, Toronto). Neill Adhikari, Damon Scales, Robert Fowler, Cheromi
Sittambalam, Mehar-Un-Nisa Raja, and Nicole Marinoff (Sunnybrook Health Science
Centre, Toronto); Lauralyn McIntryre, Shawna Reddie, Laura Jones, and Irene
Watpool (Ottawa Hospital, Ottawa); Jeffrey Singh and Madison Dennis (Toronto
Western Hospital, Toronto); Andrew Steel, Andrea Matte, Marc Lipkus, Ryan
Albert, and Emily Stern (Toronto General Hospital, Toronto); Michael Miletin,
Antonio Raso, and Robyn Klages (William Osler Health Science Centre, Brampton);
Jan Friedrich, Orla Smith and Laura Wilson (St. Michael’s Hospital, Toronto);
Deborah Cook and Mark Bailey (St. Joseph’s Hospital, Hamilton); Sangeeta
Mehta, Stephen Lapinsky, Hannah Mathers, Cheryl Ethier, Stephanie Lubchansky,
and Samer Haj-Bakri (Mount Sinai Hospital, Toronto); Dietrich Henzler and Lisa
Julien (Queen Elizabeth II Health Sciences Centre, Halifax).
Chile: Coordinator: Luis Soto-Román (Instituto Nacional del Tórax, Santiago).
Juan Carlos Maurelia (Hospital de Copiapo, Copiapo); César Antonio Maquilon
(Instituto Nacional del Tórax, Santiago); Luis Soto-Germani (Hospital de Coquimbo,
Coquimbo).
China: Coordinator: Bin Du (Peking Union Medical College Hospital, Beijing). Yan
Kang and Bo Wang (West China Hospital, Chengdu); Fachun Zhou and Fang Xu
(Chongqing Medical University 1st Hospital, Chongqing); Haibo Qiu and Yi Yang
(Southeast University Zhongda Hospital, Nanjing); Qingyuan Zhan and Bing Sun
(Beijing Chaoyang Hospital, Beijing); Zhenjie Hu and Bin Yu (Hebei Medical
University 4th Hospital, Shijiazhuang); Xi Zhu and Yu Bai (Peking University 3rd
Hospital, Beijing); Gang Li and Yi Li (Sino-Japanese Friendship Hospital, Beijing);
Geng Zhang and Jianbiao Meng (Zhejiang Tongde Hospital, Hangzhou); Xiaobo
Huang and Hong Pu (Sichuan Provincial Hospital, Chengdu); Bin Du and Daxing
Yu (Peking Union Medical College Hospital, Beijing); Chuanyun Qian and Wei
Zhang (Kunming Medical College 1st Hospital, Kunming); Yongjie Yin and
Debiao Song (Jilin University 2nd Hospital, Changchun); Yunxuan Yue and
Zhengxuan Lv (Kunming City 3rd People’s Hospital, Kunming); Chengmin Yu
and Qunmei Yao (Yunnan Chuxiong People’s Hospital; Chuxiong); Xue Wang
(Xi’an Jiaotong University 1st Hospital, Xi’an); Yuan Xu and Wei He (Beijing
Tongren Hospital, Beijing); Mian Chen and Zhihua Hu (Hainan Medical College
Hospital, Haikou); Dongpo Jiang and Jian Huang (Daping Hospital, Chongqing);
Wei Yu (Yantai Yuhuangding Hospital, Yantai); Juanxian Gu (Zhejiang Haining
People’s Hospital, Naining); Yangong Chao (Beijing Huaxin Hospital, Beijing);
Zhixiang Li (Fengrun District People’s Hospital, Tangshan); Zhicheng Zhang
(PLA Navy General Hospital, Beijing); Wanxia Li (Nanchang University 2nd Hospital, Nanchang); Zhenyang He (Hainan Provincial People’s Hospital, Haikou);
Jianguo Li and Chang Liu (Wuhan University Zhongnan Hospital, Wuhan); Tiehe
Qin and Shouhong Wang (Guangdong General Hospital, Guangzhou); Feng Li
(Nantong 1st People’s Hospital, Nantong); Jun Jin and Jianhong Fu (Suzhou
University 1st Hospital, Suzhou); Hongyang Xu (Wuxi Municipal People’s Hospital, Wuxi); Hongyuan Lin and Jianying Guo (PLA 304 Hospital, Beijing); Yalin Liu
and Jinghua Wang (Beijing Hospital, Beijing); Maoqin Li and Jiaqiong Li (Xuzhou
Central Hospital, Xuzhou); Lei Chen (Sun Yet-Sen University 6th Hospital,
Guangzhou); Qing Song and Liang Pan (PLA General Hospital, Beijing); Xianyao
Wan and Jiuzhi Zhang (Dalian Medical University 1st Hospital, Dalian); Weihai
Yao and Yuhong Guo (Beijing TCM Hospital, Beijing); Pang Wing Yan (Prince
Margaret Hospital, Hong Kong); Kelly Choy (Queen Elizabeth Hospital, Hong
Kong); Kwan Ming Chit (Pamela Youde Nethersole Eastern Hospital, Hong
Kong); Patricia Leung (Prince of Wales Hospital, Hong Kong); Chau Chin Man
(North District Hospital, Hong Kong).
Colombia: Coordinator: Marco González (Clı́nica Medellı́n and Universidad Pontificia Bolivariana, Medellı́n). Ricardo Buitrago (Clı́nica Shaio, Bogotá); Marcela
Granados (Clı́nica Fundación Valle Lili, Cali); Guillermo Ortiz (Hospital Santa
Clara, Bogotá); Cesar Enciso (Grupo Cimca Hospital San José, Bogotá); Mario
Gómez (Grupo Cimca Hospital San José and Clı́nica Fundadores, Bogotá); Bladimir
Alejandro Gil (Clı́nica Las Américas, Medellı́n); Juan Pablo Sedano, Luis Fernando
Castro (Centro Medico Imbanaco, Cali); Carlos Alberto Acosta (Hospital Federico Lleras Acosta, Ibague); Marco Gonzalez A.(Hospital San Rafael, Itagui);
Francisco Molina (Clı́nica Universitaria Bolivariana, Medellı́n); Camilo Pizarro
(Fundación Cardiovascular Colombiana, Bucaramanga); Mario Villabon (Grupo
Cimca Hospital de Suba, Bogota); Carmelo Dueñas (Nuevo Hospital Bocagrande,
Cartagena de Indias); Carlos Andrés Dı́az (Hospital General de Medellı́n, Medellı́n);
Nelson Fonseca (Corbic, Medellı́n); Rubén Camargo (Clı́nica General del Norte,
Barranquilla); Juan David Uribe (Clı́nica Cardiovascular, Medellı́n).
Denmark: Coordinator: Hans-Henrik Bülow (Holbak). Simona Beniczky and Jens
Brushoj (Naestved); Mikkel Præst and Henrik Guldager (Nykobing Falster); Birgitte
Viebaek and Sine Wichman (Roskilde); Anette Mortensen (Holbak); Susanne Andi
Iversen (Slagelse); Bo Broberg and Lone M. Poulsen (Koge).
Dominican Republic: Edgard Luna (Hospital Universitario José Marı́a Cabral y Báez,
Santiago).
Ecuador: Coordinator: Manuel Jibaja (Hospital Eugenio Espejo, Quito). Leonardo
Pazmiño, Katty Trelles, and Fabricio Picoita (Hospital Eugenio Espejo, Quito);
Gustavo Paredes and Vanesa Ramı́rez (Hospital Enrique Garcés, Quito); Guillermo
Falconı́, Cristian Cevallos and Boris Villamagua (Hospital Carlos Andrade Marı́n,
Quito); Marco Escobar and Freddy Sánchez (Hospital de la Policı́a, Quito); Miguel
Llano and Miguel Lazcano (Hospital General de las Fuerzas Armadas, Quito);
Ramiro Puetate and José Miguel Guerrero (Hospital Pablo Arturo Suárez, Quito);
Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation
Mijail Játiva and Myriam Montalvo (Hospital de los Valles, Quito); Franklin Villegas
(Hospital Metropolitano, Quito); Luis González Zambrano, Ronnie Mantilla, Gina
Quinde, Andrea Gimenez, and Luis Gonzalez Mosquera (Hospital Luis Vernaza,
Guayaquil); Henry Caballero and Marı́a Fernanda Garcı́a (Hospital de SOLCA,
Quito); Marcelo Ochoa, Soraya Puertas, and Jackeline Coello (Hospital José Carrasco
Arteaga, Cuenca); Mario Acosta (Hospital San Vicente de Paul, Ibarra).
Egypt: Medhat Soliman (Cairo University Hospitals, Cairo).
France: Coordinator: Arnaud W. Thille (CHU Henri Mondor, Créteil). Achille
Kouatchet and Alain Mercat (CHU d’Angers); Laurent Brochard (CHU Henri Mondor,
Créteil); François Collet (Centre Hospitalier De Saint-Malo); Guillaume Marcotte
(Hôpital Édouard Herriot, Lyon); Pascal Beuret (Centre Hospitalier De Roanne);
Jean-Christophe M. Richard, Gaëtan Bedunaeu, Pierre-Gildas Guitard, Dorothée
Carpentier, Benoit Veber, and Fabien Soulis (CHU Charles Nicolle, Rouen); Frédéric
Bellec (Centre Hospitalier De Montauban); Philippe Berger (Centre Hospitalier de
Châlons en Champagne); Salem Ould Zein, Géraldine Dessertaine (CHU De
Grenoble); C. Canevet (Hôpital d’Armentières); Fabien Grelon (Centre Hospitalier
Du Mans).
Germany: Coordinator: Konstantinos Raymondos (Medizinische Hochschule
Hannover, Hannover). Rolf Dembinski and Rolf Rossaint (Universitaetsklinikum
Aachen, Aachen); Steffen Weber-Carstens (Charité Universitaetsklinikum, Berlin);
Christian Putensen (Universitaetsklinikum Bonn, Bonn); Maximillian Ragaller (Universitaetsklinikum Carl Gustav Carus, Dresden); Michael Quintel (Universitaetsklinikum
der Georg-August-Universitaet, Goettingen); Winfried Schubert (Carl-Thiem-Klinikum
Cottbus gGmbH, Cottbus); Thomas Bein and Heinrich Paulus (Klinikum der
Universitaet Regensburg); Walter Brandt (Universitaetsklinikum, Magdeburg);
Lutz Pfeiffer and Silke Frenzel (Hufeland Klinikum GmbH, Mühlhausen); Thoralf
Kerner and P. Kruska (Allgemeines Krankenhaus Harburg, Hamburg); Leila Eckholt
and Joachim Hartung (Vivantes Krankenhaus am Urban, Berlin-Kreuzberg); Harald
Fritz and Monika Holler (Staedtisches Krankenhaus Martha-Maria Halle-Doelau
gGmbH, Halle-Doelau); Johannes Busch and Andreas Viehöfer (Evangelisches
Waldkrankenhaus Bad Godesberg gGmbH, Bonn); Jens Buettner (Evangelisches
Krankenhaus Elisabethenstift gGmbH, Darmstadt); Jörn Schlechtweg and Achim
Lunkeit and Roland Schneider (Klinikum Bad Salzungen GmbH, Bad Salzungen); Maria
Wussow, Nils Marquardt, and Christian Frenkel (Staedtisches Klinikum Lueneburg,
Lueneburg); Falk Hildebrandt (Dietrich-Bonhoeffer Klinikum Neubrandenburg,
Neubrandenburg); Tumbass Volker and Thomas Lipp (Ermstalklinik Bad Urach,
Bad Urach); Cezar Mihailescu and Thomas Moellhoff (Katholische Stiftung Marienhospital
Aachen, Aachen); Thomas Steinke (Universitaetsklinikum der Martin-Luther-Universitaet
Halle-Wittenberg, Halle-Wittenberg); Oliver Franke and Marcus Ruecker (Lungenklinik
Heckeshorn, Berlin); Markus Schappacher and Steffen Appel (Ermstalklinik
Staedtisches Krankenhaus Sindelfingen, Sindelfingen); Heinz Kerger (Evangelisches
Diakoniekrankenhaus, Freiburg); Andreas Schwartz (Bundeswehrkrankenhaus Ulm,
Ulm); Jan Dittmann and Jörg Haberkorn (Georgius-Agricola-Klinikum Zeitz, Zeitz);
Wolfgang Baier (St. Nikolaus-Stiftshospital GmbH, Andernach); Walter Seyde
(Staedtisches Klinikum Wolfenbuettel, Wolfenbuettel).
Greece: Coordinator: Dimitrios Matamis (Papageorgiou Hospital, Thessaloniki).
Eleni Antoniadou (Gennimata Hospital, Thessaloniki); Pertsas Evangelos (Agios
Pavlos General Hospital, Thessaloniki); Maria Giannakou (Ahepa Hospital,
Thessaloniki).
Hungary: Zoltan Szentkereszty (Kenezy Hospital, Debrecen); Zsolt Molnar (University
of Szeged, Szeged).
India: Coordinator: Pravin Amin (Bombay Hospital Institute of Medical Sciences,
Mumbai). Farhad N. Kapadia (Hinduja Hospital, Mumbai); Nagarajan Ramakrishnan
(Apollo Hospitals, Chennai); Deepak Govil (Artemis Health Institute, Gurgaon
Haryana); Anitha Shenoy and Goneppanavar Umesh (Kasturba Medical College, Manipal); Samir Sahu, (Kalinga Hospital, Odisha); Sheila Nainan Myatra
(Tata Memorial Hospital, Mumbai); Subhash Kumar Todi (AMRI Hospitals Kolkata,
West Bengal); Sanjay Dhanuka (CHL Apollo Hospital, Indore); Mayur Patel (Saifee Hospital, Mumbai); P. Samaddar (Tata Main Hospital, Jamshedpur); Dhruva
Chaudhry (PGIMS, Rohtak); Vivek Joshi and Srinivas Samavedam (CARE Hospitals, Surat); Ankur Devendra Bhavsar (Spandan Multispeciality Hospital, Vadodar); Prachee Sathe (Ruby Hall Clinic, Pune); Sujoy Mukherjee (Calcutta
Medical Research Institute, Kolkat).
Italy: Coordinator: Salvatore Maurizio Maggiore (Policlinico “Agostino Gemelli,”
Università Cattolica Del Sacro Cuore, Rome). Francesco Idone and Federica Antonicelli
(Policlinico “Agostino Gemelli,” Università Cattolica Del Sacro Cuore, Rome); Paolo
Navalesi, Rosanna Vaschetto, and Arianna Boggero (Ospedale Maggiore Della Carità,
Università Del Piemonte Orientale “Amedeo Avogadro,” Novara); Rosalba Tufano,
Michele Iannuzzi, and Edoardo De Robertis (Ospedale Policlinico “Federico II,”
Università Di Napoli, Naples); Romano Tetamo and Andrea Neville Cracchiolo
(Ospedale “Arnas Civico, Di Cristina, Benfratelli,” Palermo), Antonio Braschi,
Francesco Mojoli, and Ilaria Currò (ICU 1, Fondazione IRCCS Policlinico “S. Matteo,” Università Di Pavia, Pavia); Mirko Belliato, Chiara Verga, and Marta Ferrari
(ICU 2, Fondazione IRCCS Policlinico “S. Matteo,” Università Di Pavia, Pavia); Erika
Mannelli, Valerio Mangani, and Giorgio Tulli (Ospedale “San Giovanni Di Dio,”
Florence); Francesca Frigieri and Armando Pedullà (Ospedale “Santa Maria Annunziata,” Florence); Monica Rocco, Giorgia Citterio, and A. Di Russo (Policlinico
“Umberto I,” Università La Sapienza, Rome); Gaetano Perchiazzi and Loredana
Pitagora (Ospedale Policlinico, Università Di Bari, Bari); Antonio Pesenti and Michela
227
Bombino (Ospedale “San Gerardo,” Università Di Milano Bicocca, Monza); Davide
Chiumello, Federica Tallarini, and Serena Azzari (Fondazione IRCCS “Cà Grande”
Ospedale Maggiore Policlinico, Università Di Milano, Milan); Antonina Pigna, Ivano
Aprile, and Marco Adversi (Policlinico Universitario “S. Orsola-Malpighi,” Bologna);
Antonio Corcione, Marianna Esposito, and Annunziata Mattei (Ospedale “V. Monaldi,”
Naples); Vito Marco Ranieri, Rosario Urbino, and Ilaria Maria Mastromauro (Ospedale
“San Giovanni Battista–Molinette,” Università Di Torino, Torino); Antonino Giarratano,
Maurizio Raineri Santi and Ambrogio Sansone (Policlinico “P. Giaccone,” Università Di
Palermo, Palermo).
Japan: Toru Katani (Tokyo Women’s Medical University, Tokyo).
Korea: Coordinator: Younsuck Koh (Asan Medical Center, University of Ulsan,
Seoul). Moo Suk Park (Hospital Severance, Yonsei University Health System,
Seoul); Je Hyeong Kim (Hospital Korea University of Ansan, Ansan); Kyung Chan
Kim (Hospital Catholic University of Daegu, Daegu); Hye Sook Choi (Hospital
Dongguk University of Gyeongju, Gyeongju); Yun Seong Kim (Hospital Pusan
National University of Yangsan, Yangsan); Jin Hwa Lee (Hospital Ewha Womans
University Mokdong, Seoul); Myung-Goo Lee (Hospital Chuncheon Sacred Heart
Hospital, Hallym University Medical Center, Chuncheon); Won-Yeon Lee (Hospital Yonsei University Wonju Christian, Wonju); Jin Young An (Hospital Chungbuk
National University, Chenogiu); Gee Young Suh (Samsung Medical Center,
Sungkyunkwan University, Seoul); Ki-Suck Jung (Hallym University Medical
Center, Anyang).
Mexico: Coordinator: Asisclo J Villagómez Ortiz (Hospital Regional 18 de Octubre,
ISSSTE, México DF). César Cruz Lozano (Hospital Regional de Pemex, Ciudad
Madero); Zalatiel Maycotte Luna (Hospital Ángeles de las Lomas, Mexico DF);
José Francisco López Baca (Hospital Regional de Zona No.1 del IMSS, Mexico
DF); José Elizalde (Instituto Nacional De Ciencias Médicas y Nutrición Salvador
Zubirán, Mexico DF); Guillermo Cueto Robledo (Hospital General de México,
Mexico DF); Mario Alonso Treviño Salinas (Hospital Universitario de Nuevo León
’Dr. Eleuterio González,’ Nuevo León); Ricardo Martinez Zubieta (Hospital Español
de México, Miguel Hidalgo); Claudia Olvera-Guzman and Marco Montes De Oca
(Centro Médico ABC, Mexico DF), Silvio A. Ñamendys-Silva (Instituto Nacional de
Cancerologı́a, Mexico DF); José Salvador Martı́nez Cano (Centenario Hospital
Miguel Hidalgo, Aguascalientes); Jose Angel Baltazar Torres (Umae Hospital De
Especialidades Dr. Antonio Fraga Mouret, Mexico DF); Gustavo Morales Muñoz
(Hospital Regional de Alta Especialidad de la Mujer, Villahermosa); Antonio Villa
Delgado (Hospital Mérida Yucatán; Mérida); Javier Ladape Martinez (Hospital Juárez
de México, Mexico DF).
Morocco: Coordinator: Amine Ali Zeggwagh (Hôpital Ibn Sina, Rabat). Tarek
Dendane (Hôpital Ibn Sina, Rabat); Abderrahim Azzouzi (Hôpital Ibn Sina, Rabat);
Ahmed Sbihi (Hôpital Ibn Sina, Rabat); Wajdi Maazouzi and Mourad Amor (Hôpital
des Specialites, Rabat); Charki Haimeur (Hôpital Militaire D’Instruction Mohamed V,
Rabat).
Netherlands: Coordinator: Michael A. Kuiper (Medical Center Leeuwarden (MCL),
Leeuwarden). Matty Koopmans (MCL, Leeuwarden); Uli Strauch, Dennis Bergmans,
and Serge Heines (Universitair Medisch Centrum Maastricht, Maastricht); Sylvia den
Boer (Spaarneziekenhuis, Hoofddorp); Bas M. Kors (Kennemer Gasthuis, Haarlem);
Peter van der Voort (Onze Lieve Vrouwe Gasthuis, Amsterdam); Paul J. Dennesen
(Medisch Centrum Haaglanden, Den Haag); Bert Beishuizen, Ingrid van den Hul,
Erna Alberts, Harry PPM Gelissen, and Eduard Bootsma (Vrije Universiteit Medisch
Centrum, Amsterdam); Auke Reidinga (Tjongerschans Ziekenhuis, Heerenveen).
New Zealand: Coordinators: Jasmin Board and Andrew R. Davies (Alfred Hospital,
Melbourne). Kim Heus, Diane Mackle, and Paul Young (Wellington Hospital,
Wellington); Rachael Parke, Eileen Gilder, and Jodi Brown (CVICU Auckland City
Hospital, Auckland); Lynette Newby and Catherine Simmonds (DCCM Auckland
City Hospital, Auckland); Jan Mehrtens and Seton Henderson (Christchurch
Hospital, Christchurch); Tony Williams, Judi Tai, and Chantal Hogan (Middlemore, Auckland); Mary La Pine, John Durning, and Sheree Gare (Waikato,
Hamilton); Troy Browne, Shirley Nelson, and Jennifer Goodson (Tauranga
Hospital, Tauranga).
Panama: Julio Osorio (Hospital Rafael Hernández, Chiriquı́).
Peru: Coordinator: Isabel Coronado Campos (Hospital Guillermo Almenara Irigoyen, Lima). Rollin Roldán Mori (Hospital Edgardo Rebagliatti Martins, Lima);
Rosa Luz López Martı́nez (Hospital Guillermo Almenara Irigoyen, Lima).
Poland: Adam Mikstacki and Barbara Tamowicz (Karol Marcinkowski University of
Medical Sciences, Poznan).
Portugal: Coordinator: Rui Moreno (UCINC, Hospital de São José, Centro Hospitalar de Lisboa Central, E.P.E., Lisbon). Eduardo Almeida (Hospital Garcia de Orta,
Almada); Joana Silvestre (Centro Hospitalar de Lisboa Central, Lisbon); Heloisa
Castro, Irene Aragão, and Susana Alves Ferreira (Centro Hospitalar Do Porto–Hospital
Geral De Santo António, Porto); Nelson Barros (Centro Hospitalar Trás-Os-Montes E
Alto Douro, Vila-Real); Filomena Faria (Instituto Portugues de Oncologia, Porto);
Carlos André Correia Casado (Hospital Da Luz, Lisboa); Fausto Fialho Moura (Hospital
de Cascais, Cascais); Paulo Marcal (Hospital De São Sebastião, São Sebastião); Ricardo
Matos (Hospital de Santo Antonio dos Capuchos, Centro Hospitalar de Lisboa Central,
E.P.E., Lisbon); António Alvarez (Centro Hospitalar de Lisboa Norte, Lisbon).
228
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
Russian Federation: Coordinator: Edward Nicolayenko (Hospital No.1, Moscow).
Mikhail Kirov (Hospital No. 1, Arkhangelsk); Andrey Yaroshetskiy (Hospital No.7,
Moscow); Andrei Piontek (Hospital No.14, Ekaterinburg); Valery Subbotin (National
Institute of Surgery Named After A. V. Vishnevskij, Moscow).
Saudi Arabia: Yaseen Arabi, Olivia Dulfo, Charina Marie Olay, Edgardo E. Tabhan
(King Saud Bin Abdulaziz University for Health Sciences, Riyadh).
Spain: Coordinator: Nicolas Nin (Hospital Universitario de Getafe, Madrid).
Alfonso Muñoz and César Aragón (Hospital Carlos Haya, Málaga); Ana Villagrá
(Corporación Sanitaria y Universitaria Parc Taulı́, Sabadell); Ainhoa Rosselló and
Joan Maria Raurich (Hospital Universitario Son Espasses, Mallorca); Marı́a Garitacelaya,
Miguel Ángel González-Gallego, and Francisco Ortuño (Hospital Clı́nico Universitario
San Carlos, Madrid); Miguel Fernández-Vivas (Hospital Virgen de la Arrixaca, Murcia);
David Freire (Hospital Juan Canalejo, A Coruña); Francisco Guerrero and Francisco
Manzano (Hospital Virgen de las Nieves, Granada); Juan Carlos Sotillo (Hospital
Universitario Gregorio Marañón, Madrid); Alejandra Bustos (Hospital de Torrevieja,
Torrevieja); Alfredo Padrón, Pedro Rosas, Rafael Morales, and Liliana Caipe (Hospital
Doctor Negrı́n, Las Palmas de Gran Canaria); Maurizio Bottiroli (Hospital de la Santa
Creu i Sant Pau, Barcelona); José Marı́a Nicolás (Hospital Clinic-IDIBAPS, Barcelona);
Marta Ugalde (Hospital de Cruces, Barakaldo); Javier Ruiz (Hospital Sagrado
Corazón, Barcelona); Lucia Capilla (Hospital Morales Meseguer, Murcia); Guillermo
Muñiz (Hospital Central de Asturias, Oviedo); Jesús Sánchez-Ruiz (Hospital General
de Jerez de la Frontera, Jerez de la Frontera); Javier Cebrián, Begoña Balerdi, Elena
Parreño, and Alvaro van Bommel (Hospital Universitario La Fe, Valencia); César
Pérez-Calvo (Fundación Jiménez Dı́az, Madrid); Irene Dot (Hospital del Mar,
Barcelona); Javier Blanco (Complejo Hospitalario de Ciudad Real, Ciudad Real);
Raquel Manzanedo and José J. Blanco (Hospital Insular de Gran Canaria, Gran
Canaria); Daniel Fontaneda, Raúl González, and Javier Dı́az Domı́nguez (Complejo Hospitalario de León, León); Alfonso Moreno (Hospital San Pedro, Logroño); Antonio Reyes and Ian Carrasco (Hospital de la Princesa, Madrid); Itziar
Mintegui, Rosa Sebastián, and Javier Garcı́a-Alonso, (Complejo Hospitalario de
Donostia, Donostia); Carolina Lorencio and Josep Maria Sirvent (Hospital Universitario Dr.Josep Trueta, Girona); Patricia Jimeno (Hospital General de Segovia, Segovia); Miguel León (Hospital Arnau de Vilanova, Lleida); Pedro Galdos (Hospital
Universitario Puerta de Hierro, Majadahonda); Nuria Alonso (Hospital Universitario
Nuestra Santa Marı́a del Rosell, Cartagena); Julia López-Dı́az (Hospital Universitario
La Paz, Madrid); Marı́a Victoria de la Torre, Jorge Vidal Hernández, and Nicolás
Zamboschi (Hospital Universitario Virgen de la Victoria, Málaga); Francisco Lucena
(Hospital Universitario de Valme, Seville); Gemma Rialp (Hospital Son Llatzer, Mallorca); Raquel Montoiro (Hospital Clı́nico Lozano Blesa, Zaragoza); Victoria Goñi,
Marı́a Ángeles Pena, and Antonio Maestre (Hospital Virgen del Rocı́o, Seville); Marc
Fabra, Jacinto Baena and Eva Benveniste (Hospital Germans Trias I Pujol, Badalona);
Susana Temprano (Hospital 12 de Octubre, Madrid); Jesús Sánchez (Hospital de
Rio Hortega, Valladolid); Carmen Campos (Hospital Universitario Dr.Peset, Valencia); Sara Cabañes (Hospital Santiago Apóstol, Vitoria); Marı́a Elena Pérez-Losada,
Javier González-Robledo, and José Claudio Leo (Complejo Hospitalario de Salamanca, Salamanca); Enrique Piacentini (Hospital Mutua de Terrassa, Terrassa); Marı́a del Carmen de la Torre (Hospital de Mataró, Mataró); Laura Álvarez-Montero
and Fernando Sánchez (Hospital Xeral Calde, Lugo); Antonio Viñuales (Hospital
Lluis Alcanyis, Xàtiva); Bernabé Álvarez (Hospital General de Alicante, Alicante);
Javier Castañeda (Hospital Clı́nico de Valladolid, Valladolid); Ángela Alonso (Hospital de Fuenlabrada, Fuenlabrada); Marı́a Isabel Ruiz (Complejo Hospitalario de Jaén,
Jaén); Pedro Jesús Domı́nguez (Hospital Juan Ramón Jiménez, Huelva); Marcos
Delgado (Complejo Hospitalari de Manresa, Manresa); Eugenio Palazón (Hospital
Universitario Reina Sofı́a, Murcia); Antonio Garcı́a-Jiménez (Hospital Arquitecto
Marcide, Ferrol); Rosa Álvaro (Hospital de La Plana, Castellón); Clara Laplaza, Eva
Regidor and Enrique Maravı́ (Complejo Hospitalario de Navarra, Pamplona); José
Marı́a Quiroga (Hospital de Cabueñes, Gijón); Amalia Martı́nez de la Gandara
(Hospital Infanta Leonor, Madrid); Cecilia Carbayo (Hospital Torrecárdenas, Almerı́a); Marı́a Luisa Navarrete (Hospital San Juan, Alicante); Manuel Valledor and
Raquel Yano (Hospital San Agustı́n, Avilés); José Marı́a Gutiérrez (Hospital General
de Albacete, Albacete); Amparo Ferrandiz, Alberto Belenguer, and Lidón Mateu
(Hospital General de Castellón, Castellón); Laura Sayagues (Complejo Hospitalario
de Santiago de Compostela, Compostela); Marı́a José Tolón (Hospital Royo Vilanova, Zaragoza); Nieves Franco (Hospital de Móstoles, Móstoles); Elena Gallego
(Hospital San Pedro de Alcántara, Cáceres); Félix Lacoma (Hospital Quirón,
Madrid); Patricia Albert (Hospital del Sureste, Arganda); Vicente Arraez (Hospital
Universitario General, Elche); Mar Gobernado (Hospital General de Soria, Soria);
Susana Moradillo (Hospital Rı́o Carrión, Palencia); Carolina Gı́menez-Esparza (Hospital de la Vega Baja, Orihuela); Teresa Sánchez de Dios (Complejo Hospitalario
Montecelo, Pontevedra); Carlos Marian Crespo (Hospital General de Guadalajara, Guadalajara); Cecilia Hermosa and Federico Gordo (Hospital del Henares,
Coslada); Genis Carrasco (Hospital SCIAS, Barcelona); Marı́a Ángeles Alonso
(Trauma ICU, Hospital 12 de Octubre, Madrid); Alejandro Algora (Fundación Hospital Universitario de Alcorcón, Madrid); Raúl de Pablo (Hospital Prı́ncipe de Asturias, Alcalá de Henares); Sofı́a Garcı́a (Hospital del Poniente, El Ejido); Ana Carolina
Caballero (Hospital de Zamora, Zamora); José Marı́a Montón (Hospital Obispo
Polanco, Teruel); Teresa Mut (Hospital Provincial de Castellón, Castellón); Eva Manteiga (Hospital Infanta Cristina, Parla); Alejandro de la Serna (Hospital de Galdakao,
Galdakao); Ana Esther Trujillo (Hospital General de La Palma, La Palma); Rafael
Blancas (Hospital del Tajo, Aranjuez); Inmaculada Vallverdú (Hospital Universitario San Juan, Reus); José Manuel Serrano (Hospital Universitario Reina Sofı́a,
Córdoba); Miquel Ferrer (Hospital Clinic-IDIBAPS, Barcelona); Juan Diego Jiménez (Hospital de Don Benito, Don Benito); Carlos Gallego (Hospital Infanta
Elena, Valdemoro); Luis Marina (Complejo Hospitalario de Toledo, Toledo).
VOL 188
2013
Taiwan: Chen Chin-Ming and Ai-Chin Cheng (Chi Mei Medical Center, Tainan
City).
Tunisia: Coordinator: Fekri Abroug (Hospital Fattouma Bourguina, Monastir).
Besbes Mohamed (Hospital Abderrahmane Mami, Ariana); Imed Chouchene
(Hospital Universitarie Farhat Hached, Sousse); Mounir Bouaziz (CHU Habib
Bourguiba, Sfax); Stambouli Neji and Islem Ouanes (Hospital Fattouma Bourguina,
Monastir); Ayed Samia (Hospital Taher Sfar City, Mahdia).
Turkey: Coordinator: Nahit Cakar (Istanbul Medical Faculty, Istanbul). Ismail Kati
(Medical Faculty of Yuzuncu Yil University, Van) Ali Aydım Altunkan (Faculty of
Medicine, Mersin University, Mersin); Remzi Iscimen (Uludag University Faculty
of Medicine, Bursa); Zafer Dogan (Sutcu İmam University, Kahramanmaras);
Bilge Çetin (Erciye Üniversity, Kayseri); Tayfun Adanir (Atatürk Hospital, Izmir);
Sabriye Guvenc (Anadolu Medical Center Istanbul, Istanbul); Unase Büyükkoçak
(Kirikkale University, Kirikkale).
United States: Coordinator: Antonio Anzueto (University Hospital San Antonio
and South Texas Veterans Health Care System, San Antonio, TX). Ashley Ellis
and Gary Kinasewitz (Oklahoma University Health Science Center and VA medical
Center, Oklahoma City, OK); Allan Walkey and Phil Alkana (Boston Medical Center, Boston, MA); Gregory A. Schmidt, Susan Gillen, Kathleen Lilli, Jennifer
Twombley, Denice Wells, and Larry Welder; (University of Iowa Hospitals and
Clinics, Iowa City, IA); Alejandro Arroliga, Alfredo Vasquez-Sandoval, Vincent
John Scott, Craig Cernosek, and Christopher Spradley (Temple Clinic, Scott
and White Healthcare, Temple, TX); Dimple Tejwani, Sindhaghatta Venkatram,
and Gilda Diaz-Fuentes (Bronx Lebanon Hospital Center, New York, NY); Amber
Monson, Anthony Saleh, Madhav Gudi, and George Liziamma (New York
Methodist Hospital, New York, NY); Mohamed A. Saad, Crissie De Spirito,
Bryan Beatty, Samir Vermani, and Crissie Despirito (University of Louisville
School of Medicine Hospital, Louisville, KY); Zaza Cohen, Amee Patrawalla,
Samir Abdelhadi, Rupesh Vakil, and Steven Y. Chang (UMDNJ–New Jersey
Medical School, Newark, NJ); Brian Sherman, Rosanna Del Giudice, and John
Oropell (Mount Sinai Medical Center, New York, NY); Timothy D Girard, Cayce
Strength, Joyce Okahashi, Leanne Boehm, and Matthew Kirchner (Vanderbilt
University School of Medicine, Nashville, TN); Ashley Ellis and Gary Kinasewitz
(Oklahoma City VA Medical Center, Oklahoma City, OK); Erwin J. Oei, Sebastian
Circo, Nelson Medina, and Mohammed Al-Jagbeer (Morristown Medical Center,
Morristown, NJ); V. J. Cardenas, Jr. and Smyth Smith (University of Texas Medical
Branch at Galveston, Galveston, TX); Shelby Sutton, Marcela Canola-Mazo, Tim
Houlihan, Yogeet Kaur, and Travis Parry (University Hospital San Antonio and South
Texas Veterans Health Care System, San Antonio, TX); Craig A. Piquette and Kerry
Canady (Omaha VA Medical Center, Omaha, NE); Rahul Nanchal and Dana K.
Soetaert (Medical College of Wisconsin, Milwaukee, WI); Maria del Mar TorresPerez, Carlos Robles-Arias, and William Rodriguez-Cintron (VA Caribbean
Health Care System, San Juan, Puerto Rico); Mark Tidswell, Jennifer Germain, LoriAnn Kozikowski, and Erin Braden (Baystate Medical Center, Springfield, MA); Geneva
Tatem (Henry Ford Hospital, Detroit, MI).
Uruguay: Coordinator: Javier Hurtado (Hospital Español and CUDAM, Montevideo).
Alberto Deicas (CASMU No. 2, Montevideo); Daniel Weiss (Hospital Pasteur, Montevideo); Marta Beron (Hospital Maciel, Montevideo); Román Garrido (Hospital
Evangélico, Montevideo); Cristina Santos and Mario Cancela (Hospital de Clı́nicas,
Montevideo); Raúl Lombardi (Impasa, Montevideo); Pedro Alzugaray (CAAMOC,
Carmelo, Sanatorio Americano, Montevideo and Orameco, Colonia); Jorge Gerez
(Hospital Policial, Montevideo); Silvia Mareque (Sanatorio CAMS, Mercedes);
Graciela Franca (Circulo Católico, Montevideo); Oscar Cluzet (Sanatorio Americano,
Montevideo); Edgardo Nuñez (Sanatorio Mautone and Hospital de Maldonado,
Maldonado); Julio Pontet (Hospital de Florida, Florida); Sergio Cáceres (Centro
Cardiológico Sanatorio Americano, Montevideo); Elia Caragna (CRAMI, Las
Piedras); Alberto Soler (COMEPA and Hospital Escuela del Litoral, Paysandú);
Frank Torres (Sanatorio Cantegril, Punta del Este); Gustavo Pittini (CAAMEPA,
Pando).
Venezuela: Coordinator: Gabriel d‘Empaire (Hospital de Clı́nicas de Caracas, Caracas).
Stevens Salva and Fernando Pérez (Hospital de Clı́nicas de Caracas, Caracas); Clara
Pacheco and Zoraida Parra (Hospital Clı́nico Universitario de Caracas, Caracas); Ingrid
Von der Osten (Hospital Miguel Pérez Carreño, Caracas); Luis Williams and José Salinas
(Hospital Centro de Especialidades de Anzoategui, Lecherı́a).
Vietnam: Do Danh Quynh, Pham Thi Van Anh, Nguyen Huu Hoang, and Nghuyen
ba Tuan (Viet Duc Hospital, Hanoi).
References
1. Amato MB, Barbas CS, Medeiros DM, Magaldi RB, Schettino GP,
Lorenzi-Filho G, Kairalla RA, Deheinzelin D, Munoz C, Oliveira R,
et al. Effect of a protective-ventilation strategy on mortality in the acute
respiratory distress syndrome. N Engl J Med 1998;338:347–354.
2. The Acute Respiratory Distress Syndrome Network. Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute
lung injury and the acute respiratory distress syndrome. N Engl J Med
2000;342:1301–1308.
Esteban, Frutos-Vivar, Muriel, et al.: Evolution of Mortality in Mechanical Ventilation
3. Gordo-Vidal F, Gómez-Tello V, Palencia-Herrejón E, Latour-Pérez J,
Sánchez-Artola B, Díaz-Alersi R. PEEP alta frente a PEEP convencional en
el síndrome de distrés respiratorio agudo: revisión sistemática y metaanálisis.
Med Intensiva 2007;31:491–501.
4. Gordo-Vidal F, Gómez-Tello V, Palencia-Herrejón E, Latour-Pérez J.
Impacto de dos nuevos estudios en el resultado de un metaanálisis
sobre la aplicación de presión positiva al final de la espiración alta
a pacientes con síndrome de distrés respiratorio agudo. Med Intensiva
2008;32:316–317.
5. Protti A, Cressoni M, Santini A, Langer T, Mietto C, Febres D, Chierichetti
M, Coppola S, Conte G, Gatti S, et al. Lung stress and strain during
mechanical ventilation: any safe threshold? Am J Respir Crit Care
Med 2011;183:1354–1362.
6. Brochard L, Rauss A, Benito S, Conti G, Mancebo J, Rekik N, Gasparetto
A, Lemaire F. Comparison of three methods of gradual withdrawal
from ventilatory support during weaning from mechanical ventilation.
Am J Respir Crit Care Med 1994;150:896–903.
7. Esteban A, Frutos F, Tobin MJ, Alía I, Solsona JF, Valverdú I, Fernández R,
de la Cal MA, Benito S, Tomás R, et al.; Spanish Lung Failure Collaborative
Group. A comparison of four methods of weaning patients from mechanical
ventilation. N Engl J Med 1995;332:345–350.
8. Esteban A, Alía I, Gordo F, Fernández R, Solsona JF, Vallverdú I,
Macías S, Allegue JM, Blanco J, Carriedo D, et al.; The Spanish
Lung Failure Collaborative Group. Extubation outcome after spontaneous
breathing trials with T-tube or pressure support ventilation. Am J Respir Crit
Care Med 1997;156:459–465.
9. Esteban A, Alía I, Tobin MJ, Gil A, Gordo F, Vallverdú I, Blanch L,
Bonet A, Vázquez A, de Pablo R, et al.; Spanish Lung Failure Collaborative
Group. Effect of spontaneous breathing trial duration on outcome of
attempts to discontinue mechanical ventilation. Am J Respir Crit Care
Med 1999;159:512–518.
10. Girard TD, Kress JP, Fuchs BD, Thomason JW, Schweickert WD, Pun
BT, Taichman DB, Dunn JG, Pohlman AS, Kinniry PA, et al. Efficacy
and safety of a paired sedation and ventilator weaning protocol for
mechanically ventilated patients in intensive care (Awakening and
Breathing Controlled Trial): a randomised controlled trial. Lancet
2008;371:126–134.
11. Ram FS, Picot J, Lightowler J, Wedzicha JA. Non-invasive positive
pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2004;CD004104.
12. Vital FM, Saconato H, Ladeira MT, Sen A, Hawkes CA, Soares B,
Burns KE, Atallah AN. Non-invasive positive pressure ventilation
(CPAP or bilevel NPPV) for cardiogenic pulmonary edema. Cochrane
Database Syst Rev 2008;CD005351.
13. Esteban A, Anzueto A, Frutos F, Alía I, Brochard L, Stewart TE, Benito
S, Epstein SK, Apezteguía C, Nightingale P, et al.; Mechanical
Ventilation International Study Group. Characteristics and outcomes in
adult patients receiving mechanical ventilation: a 28-day international
study. JAMA 2002;287:345–355.
14. Esteban A, Ferguson ND, Meade MO, Frutos-Vivar F, Apezteguia C,
Brochard L, Raymondos K, Nin N, Hurtado J, Tomicic V, et al.;
VENTILA Group. Evolution of mechanical ventilation in response
to clinical research. Am J Respir Crit Care Med 2008;177:170–177.
15. Hanley JA, Negassa A, Edwardes MD, Forrester JE. Statistical analysis
of correlated data using generalized estimating equations: an orientation.
Am J Epidemiol 2003;157:364–375.
16. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J,
Gea-Banacloche J, Keh D, Marshall JC, Parker MM, et al. Surviving
Sepsis Campaign guidelines for management of severe sepsis and
septic shock. Intensive Care Med 2004;30:536–555.
17. Shehabi Y, Bellomo R, Reade MC, Bailey M, Bass F, Howe B, McArthur
C, Seppelt IM, Webb S, Weisbrodt L; Sedation Practice in Intensive
Care Evaluation (SPICE) Study Investigators; ANZICS Clinical Trials
Group. Early intensive care sedation predicts long-term mortality in
ventilated critically ill patients. Am J Respir Crit Care Med 2012;186:
724–731.
18. Mehta S, Burry L, Cook D, Fergusson D, Steinberg M, Granton J, Herridge
M, Ferguson N, Devlin J, Tanios M, et al.; SLEAP Investigators; Canadian
Critical Care Trials Group. Daily sedation interruption in mechanically
ventilated critically ill patients cared for with a sedation protocol:
a randomized controlled trial. JAMA 2012;308:1985–1992.
229
19. Pronovost P, Needham D, Berenholtz S, Sinopoli D, Chu H, Cosgrove S,
Sexton B, Hyzy R, Welsh R, Roth G, et al. An intervention to
decrease catheter-related bloodstream infections in the ICU. N Engl
J Med 2006;355:2725–2732.
20. Resar R, Pronovost P, Haraden C, Simmonds T, Rainey T, Nolan T.
Using a bundle approach to improve ventilator care processes and
reduce ventilator-associated pneumonia. Jt Comm J Qual Patient
Saf 2005;31:243–248.
21. Stewart TE, Meade MO, Cook DJ, Granton JT, Hodder RV, Lapinsky
SE, Mazer CD, McLean RF, Rogovein TS, Schouten BD, et al.;
Pressure- and Volume-Limited Ventilation Strategy Group. Evaluation of a ventilation strategy to prevent barotrauma in patients at high
risk for acute respiratory distress syndrome. N Engl J Med 1998;338:
355–361.
22. Brochard L, Roudot-Thoraval F, Roupie E, Delclaux C, Chastre J,
Fernandez-Mondéjar E, Clémenti E, Mancebo J, Factor P, Matamis
D, et al. Tidal volume reduction for prevention of ventilator-induced
lung injury in acute respiratory distress syndrome. The Multicenter
Trial Group on Tidal Volume Reduction in ARDS. Am J Respir Crit
Care Med 1998;158:1831–1838.
23. Brower RG, Shanholtz CB, Fessler HE, Shade DM, White P Jr, Wiener
CM, Teeter JG, Dodd-o JM, Almog Y, Piantadosi S. Prospective,
randomized, controlled clinical trial comparing traditional versus reduced
tidal volume ventilation in acute respiratory distress syndrome patients.
Crit Care Med 1999;27:1492–1498.
24. Needham DM, Colantuoni E, Mendez-Tellez PA, Dinglas VD, Sevransky
JE, Dennison Himmelfarb CR, Desai SV, Shanholtz C, Brower RG,
Pronovost PJ. Lung protective mechanical ventilation and two year
survival in patients with acute lung injury: prospective cohort study.
BMJ 2012;344:e2124.
25. Gajic O, Frutos-Vivar F, Esteban A, Hubmayr RD, Anzueto A. Ventilator
settings as a risk factor for acute respiratory distress syndrome in
mechanically ventilated patients. Intensive Care Med 2005;31:922–926.
26. Gajic O, Dara SI, Mendez JL, Adesanya AO, Festic E, Caples SM, Rana
R, St Sauver JL, Lymp JF, Afessa B, et al. Ventilator-associated lung
injury in patients without acute lung injury at the onset of mechanical
ventilation. Crit Care Med 2004;32:1817–1824.
27. Determann RM, Royakkers A, Wolthuis EK, Vlaar AP, Choi G, Paulus
F, Hofstra JJ, de Graaff MJ, Korevaar JC, Schultz MJ. Ventilation
with lower tidal volumes as compared with conventional tidal volumes
for patients without acute lung injury: a preventive randomized controlled
trial. Crit Care 2010;14:R1.
28. Serpa Neto A, Cardoso SO, Manetta JA, Pereira VG, Espósito DC,
Pasqualucci MdeO, Damasceno MC, Schultz MJ. Association between
use of lung-protective ventilation with lower tidal volumes and clinical
outcomes among patients without acute respiratory distress syndrome:
a meta-analysis. JAMA 2012;308:1651–1659.
29. Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute respiratory
distress in adults. Lancet 1967;2:319–323.
30. Gattinoni L, Pelosi P, Crotti S, Valenza F. Effects of positive endexpiratory pressure on regional distribution of tidal volume and recruitment in adult respiratory distress syndrome. Am J Respir Crit
Care Med 1995;151:1807–1814.
31. Caironi P, Cressoni M, Chiumello D, Ranieri M, Quintel M, Russo SG,
Cornejo R, Bugedo G, Carlesso E, Russo R, et al. Lung opening and
closing during ventilation of acute respiratory distress syndrome. Am
J Respir Crit Care Med 2010;181:578–586.
32. Metnitz PG, Metnitz B, Moreno RP, Bauer P, Del Sorbo L, Hoermann
C, de Carvalho SA, Ranieri VM; SAPS 3 Investigators. Epidemiology
of mechanical ventilation: analysis of the SAPS 3 database. Intensive
Care Med 2009;35:816–825.
33. Ferguson ND, Frutos-Vivar F, Esteban A, Anzueto A, Alía I, Brower
RG, Stewart TE, Apezteguía C, González M, Soto L, et al.; Mechanical
Ventilation International Study Group. Airway pressures, tidal volumes,
and mortality in patients with acute respiratory distress syndrome. Crit
Care Med 2005;33:21–30.
34. Ranieri VM, Suter PM, Tortorella C, De Tullio R, Dayer JM, Brienza A,
Bruno F, Slutsky AS. Effect of mechanical ventilation on inflammatory
mediators in patients with acute respiratory distress syndrome: a randomized
controlled trial. JAMA 1999;282:54–61.
35. Brower RG, Lanken PN, MacIntyre N, Matthay MA, Morris A, Ancukiewicz
M, Schoenfeld D, Thompson BT; National Heart, Lung, and Blood
230
36.
37.
38.
39.
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
Institute ARDS Clinical Trials Network. Higher versus lower positive
end-expiratory pressures in patients with the acute respiratory distress
syndrome. N Engl J Med 2004;351:327–336.
Villar J, Kacmarek RM, Pérez-Méndez L, Aguirre-Jaime A. A high
positive end-expiratory pressure, low tidal volume ventilatory strategy
improves outcome in persistent acute respiratory distress syndrome:
a randomized, controlled trial. Crit Care Med 2006;34:1311–1318.
Meade MO, Cook DJ, Guyatt GH, Slutsky AS, Arabi YM, Cooper DJ,
Davies AR, Hand LE, Zhou Q, Thabane L, et al.; Lung Open
Ventilation Study Investigators. Ventilation strategy using low tidal
volumes, recruitment maneuvers, and high positive end-expiratory
pressure for acute lung injury and acute respiratory distress syndrome:
a randomized controlled trial. JAMA 2008;299:637–645.
Mercat A, Richard JC, Vielle B, Jaber S, Osman D, Diehl JL, Lefrant
JY, Prat G, Richecoeur J, Nieszkowska A, et al.; Expiratory Pressure
(Express) Study Group. Positive end-expiratory pressure setting in
adults with acute lung injury and acute respiratory distress syndrome:
a randomized controlled trial. JAMA 2008;299:646–655.
Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E,
Fan E, Camporota L, Slutsky AS; ARDS Definition Task Force.
40.
41.
42.
43.
VOL 188
2013
Acute respiratory distress syndrome: the Berlin definition. JAMA
2012;307:2526–2533.
Briel M, Meade M, Mercat A, Brower RG, Talmor D, Walter SD, Slutsky
AS, Pullenayegum E, Zhou Q, Cook D, et al. Higher vs lower positive
end-expiratory pressure in patients with acute lung injury and acute
respiratory distress syndrome: systematic review and meta-analysis.
JAMA 2010;303:865–873.
Chandra D, Stamm JA, Taylor B, Ramos RM, Satterwhite L, Krishnan
JA, Mannino D, Sciurba FC, Holguín F. Outcomes of noninvasive
ventilation for acute exacerbations of chronic obstructive pulmonary
disease in the United States, 1998–2008. Am J Respir Crit Care Med
2012;185:152–159.
Esteban A, Frutos-Vivar F, Ferguson ND, Arabi Y, Apezteguía C, González
M, Epstein SK, Hill NS, Nava S, Soares MA, et al. Noninvasive positivepressure ventilation for respiratory failure after extubation. N Engl
J Med 2004;350:2452–2460.
Wunsch H, Linde-Zwirble WT, Angus DC, Hartman ME, Milbrandt EB,
Kahn JM. The epidemiology of mechanical ventilation use in the
United States. Crit Care Med 2010;38:1947–1953.
new england
journal of medicine
The
established in 1812
september 16, 2010
vol. 363 no. 12
Neuromuscular Blockers in Early Acute Respiratory
Distress Syndrome
Laurent Papazian, M.D., Ph.D., Jean-Marie Forel, M.D., Arnaud Gacouin, M.D., Christine Penot-Ragon, Pharm.D.,
Gilles Perrin, M.D., Anderson Loundou, Ph.D., Samir Jaber, M.D., Ph.D., Jean-Michel Arnal, M.D., Didier Perez, M.D.,
Jean-Marie Seghboyan, M.D., Jean-Michel Constantin, M.D., Ph.D., Pierre Courant, M.D., Jean-Yves Lefrant, M.D., Ph.D.,
Claude Guérin, M.D., Ph.D., Gwenaël Prat, M.D., Sophie Morange, M.D., and Antoine Roch, M.D., Ph.D.,
for the ACURASYS Study Investigators*
A bs t r ac t
Background
In patients undergoing mechanical ventilation for the acute respiratory distress syndrome (ARDS), neuromuscular blocking agents may improve oxygenation and decrease ventilator-induced lung injury but may also cause muscle weakness. We evaluated clinical outcomes after 2 days of therapy with neuromuscular blocking agents in
patients with early, severe ARDS.
Methods
In this multicenter, double-blind trial, 340 patients presenting to the intensive care
unit (ICU) with an onset of severe ARDS within the previous 48 hours were randomly assigned to receive, for 48 hours, either cisatracurium besylate (178 patients)
or placebo (162 patients). Severe ARDS was defined as a ratio of the partial pressure
of arterial oxygen (PaO2) to the fraction of inspired oxygen (FiO2) of less than 150,
with a positive end-expiratory pressure of 5 cm or more of water and a tidal volume of
6 to 8 ml per kilogram of predicted body weight. The primary outcome was the proportion of patients who died either before hospital discharge or within 90 days after
study enrollment (i.e., the 90-day in-hospital mortality rate), adjusted for predefined
covariates and baseline differences between groups with the use of a Cox model.
Results
The hazard ratio for death at 90 days in the cisatracurium group, as compared with
the placebo group, was 0.68 (95% confidence interval [CI], 0.48 to 0.98; P = 0.04), after
adjustment for both the baseline PaO2:FIO2 and plateau pressure and the Simplified
Acute Physiology II score. The crude 90-day mortality was 31.6% (95% CI, 25.2 to 38.8)
in the cisatracurium group and 40.7% (95% CI, 33.5 to 48.4) in the placebo group
(P = 0.08). Mortality at 28 days was 23.7% (95% CI, 18.1 to 30.5) with cisatracurium
and 33.3% (95% CI, 26.5 to 40.9) with placebo (P = 0.05). The rate of ICU-acquired
paresis did not differ significantly between the two groups.
Conclusions
From Assistance Publique–Hôpitaux de
Marseille Unité de Recherche sur les Maladies Infectieuses et Tropicales Émergentes (URMITE), Centre National de la
Recherche Scientifique–Unité Mixte de
Recherche (CNRS-UMR) 6236 (L.P., J.-M.F.,
A.R.) and Faculté de Médecine (A.L.), Université de la Méditerranée Aix–Marseille
II; Hôpital Sainte-Marguerite (C.P.-R.);
Assistance Publique–Hôpitaux de Marseille (G. Perrin); Hôpital Ambroise Paré
(J.-M.S.); and Centre d’Investigations Cliniques, Assistance Publique–Hôpitaux de
Marseille, INSERM 9502 (S.M.) — all in
Marseille; Hôpital Pontchaillou, Rennes
(A.G.); Hôpital Saint Eloi, Montpellier
(S.J.); Hôpital Font-Pré, Toulon (J.-M.A.);
Hôpital Jean Minjoz, Besançon (D.P.);
Hôpital Hôtel-Dieu, Clermont-Ferrand
(J.-M.C.); Centre Hospitalier, Avignon
(P.C.); Hôpital Caremeau, Nîmes (J.-Y.L.);
Hôpital de la Croix-Rousse, Lyon (C.G.);
and Hôpital de Cavale Blanche, Brest (G.
Prat) — all in France. Address reprint requests to Dr. Papazian at Service de Réani­
mation Médicale, Hôpital Nord, Chemin
des Bourrely, 13009 Marseille, France, or
at [email protected].
*The ARDS et Curarisation Systematique
(ACURASYS) study investigators are
listed in the Appendix.
N Engl J Med 2010;363:1107-16.
Copyright © 2010 Massachusetts Medical Society.
In patients with severe ARDS, early administration of a neuromuscular blocking
agent improved the adjusted 90-day survival and increased the time off the ventilator
without increasing muscle weakness. (Funded by Assistance Publique–Hôpitaux de
Marseille and the Programme Hospitalier de Recherche Clinique Régional 2004-26 of
the French Ministry of Health; ClinicalTrials.gov number, NCT00299650.)
n engl j med 363;12 nejm.org september 16, 2010
1107
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
T he acute respiratory distress syndrome (ARDS) is characterized by hypoxemic respiratory failure; it affects both
medical and surgical patients.1 Despite rigorous
physiological management,2 in most studies,
ARDS has been fatal in 40 to 60% of patients.3-7
Neuromuscular blocking agents are used in a
large but highly variable proportion of patients
with ARDS.8-12 Current guidelines indicate that
neuromuscular blocking agents are appropriate for
facilitating mechanical ventilation when sedation
alone is inadequate, most notably in patients with
severe gas-exchange impairments.10 In a fourcenter randomized, controlled trial of gas exchange in 56 patients with ARDS,13 infusion of
a neuromuscular blocking agent for a period of
48 hours was associated with improved oxygenation and a trend toward lower mortality in the
intensive care unit (ICU) (46%, vs. 71% among
patients who did not receive a blocking agent;
P = 0.06). However, this study was not designed or
powered to evaluate mortality. Thus, the benefits
and risks of adjunctive therapy with neuromuscular blocking agents in patients with ARDS who
were receiving lung-protective mechanical ventilation14 require further evaluation.
We conducted a multicenter, randomized, placebo-controlled, double-blind trial to determine
whether a short period of treatment with the neuromuscular blocking agent cisatracurium besylate
early in the course of severe ARDS would improve
clinical outcomes.
Me thods
Patients
Patients were enrolled from March 2006 through
March 2008 at 20 ICUs in France (see the Appendix). Eligibility criteria were the receipt of
endotracheal mechanical ventilation for acute hypoxemic respiratory failure and the presence of
all of the following conditions for a period of no
longer than 48 hours: ratio of the partial pressure of arterial oxygen (PaO2, measured in millimeters of mercury) to the fraction of inspired oxygen (FiO2, which is unitless) of less than 150 with
the ventilator set to deliver a positive end-expiratory pressure of 5 cm of water or higher and a
tidal volume of 6 to 8 ml per kilogram of predicted body weight, and bilateral pulmonary infiltrates that were consistent with edema. An additional eligibility criterion was the absence of
1108
of
m e dic i n e
clinical evidence of left atrial hypertension — that
is, a pulmonary-capillary wedge pressure, if available, of less than 18 mm Hg. If the pulmonarycapillary wedge pressure was not available,
echocardiography was performed if the patient had
a history of, or risk factors for, ischemic heart
disease or had crackles on auscultation. Exclusion
criteria are listed in Figure 1.
The trial was monitored by an independent data
and safety monitoring board. Randomization and
blinding regarding the study-group assignments
were performed according to Consolidated Standards for the Reporting of Trials (CONSORT)
guidelines, as indicated in the Supplementary Appendix (available with the full text of this article
at NEJM.org). The study protocol and statistical
analysis plan (also available at NEJM.org) were
approved for all centers by the ethics committee of
the Marseille University Hospital (Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale), according to French law. The
study was conducted in accordance with the protocol and statistical analysis plan. Written informed consent was obtained from the patients
or their proxies.
Study Treatment
Cisatracurium besylate (150-mg formulation,
GlaxoSmithKline) and placebo were prepared in
identical separate 30-ml vials for intravenous infusion. Peripheral-nerve stimulators were not permitted. The Ramsay sedation scale was used to
adapt sedative requirements. The scale assigns the
conscious state a score of 1 (anxious, agitated, or
restless) to 6 (no response on glabellar tap). Once
the assigned Ramsay sedation score was 6 and
the ventilator settings were adjusted (Table 1), a
3-ml rapid intravenous infusion of 15 mg of cis­
atracurium besylate or placebo was administered,
followed by a continuous infusion of 37.5 mg per
hour for 48 hours. This regimen was based on the
results of two studies of a total of 92 patients monitored for paralysis.13,15
Ventilation and Weaning Protocol
The volume assist–control mode of ventilation was
used, with a tidal volume of 6 to 8 ml per kilogram
of predicted body weight (Table 1). The goal was
a saturation of peripheral blood oxygen (SpO2) as
measured by means of pulse oximetry of 88 to
95% or a PaO2 of 55 to 80 mm Hg. To achieve this
goal, FiO2 and the positive end-expiratory pres-
n engl j med 363;12 nejm.org september 16, 2010
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Neuromuscular Blocking Agents in ARDS
1326 Patients were assessed for eligibility
986 Were excluded
340 Underwent randomization
178 Were assigned to receive
cisatracurium
162 Were assigned to receive
placebo
1 Withdrew consent
177 Received cisatracurium
162 Received placebo
177 Were included in the analysis
162 Were included in the analysis
Figure 1. Randomization and Follow-up
the Patients, According to Study
Group. 1st
RETAKE:
AUTHOR:ofPapazian
For the 986 patients who were assessed for eligibility but excluded, the reasons for 2nd
exclusion were as follows: age
FIGURE:
1 of
2 of consent (185 patients, 18.8%), continuous
3rd
younger than 18 years (19 patients,
1.9%),
lack
infusion of a neuroRevised
muscular blocking agent at enrollment
ARTIST:(42
ts patients, 4.3%), known pregnancy (19 patients, 1.9%), enrollment in
SIZE
­another trial within the previous 30 days, (57 patients, 5.8%), increased intracranial
6 col pressure (18 patients, 1.8%),
Line
Combo
4-C
H/T
TYPE:
33p9
­severe chronic respiratory disease requiring long-term oxygen therapy or mechanical
ventilation at home (95 patients, 9.6%), actual body weight exceeding 1 kgAUTHOR,
per centimeter
of height, (20 patients, 2.0%), severe chronic liver
PLEASE NOTE:
Figure8.3%),
has been
redrawn
and type
has been reset.or chemotherapy-induced neutro­
disease (Child–Pugh class C) (82 patients,
bone
marrow
transplantation
Please check carefully.
penia (97 patients, 9.8%), pneumothorax (18 patients,
1.8%), expected duration of mechanical ventilation of less
than 48 hours (15 patients, 1.5%),
(168 patients, 17.0%), other reason
JOB:decision
36312 to withhold life-sustaining treatment
ISSUE: 09-16-10
(103 patients, 10.4%), and time window missed (48 patients, 4.9%).
sure were adjusted as in the Prospective, Randomized, Multi-Center Trial of 12 ml/kg Tidal Volume
Positive Pressure Ventilation for Treatment of
Acute Lung Injury and Acute Respiratory Distress
Syndrome (ARMA).14
An open-label, rapid, intravenous injection of
20 mg of cisatracurium was allowed in both
groups if the end-inspiratory plateau pressure
remained greater than 32 cm of water for at least
10 minutes despite the administration of increasing doses of sedatives and decreasing tidal volume
and positive end-expiratory pressure (if tolerated).
If this rapid, intravenous injection resulted in a
decrease of the end-inspiratory plateau pressure
by less than 2 cm of water, a second injection of
20 mg of cisatracurium was allowed. If after the
injection, the end-inspiratory plateau pressure did
not decrease or decreased by less than 2 cm of
water, cisatracurium was not administered again
during the following 24-hour period.
Organ or System Failure
Patients were monitored daily for 28 days for signs
of failure of nonpulmonary organs or systems.14
Circulatory failure was defined as systolic blood
pressure of 90 mm Hg or less or the need for vasopressor therapy. Coagulation failure was defined
as a platelet count of 80,000 or less per cubic millimeter. Hepatic failure was defined as a serum
bilirubin level of 2 mg per deciliter (34 μmol per
liter) or higher. Renal failure was defined as a serum creatinine level of 2 mg per deciliter (177 μmol
n engl j med 363;12 nejm.org september 16, 2010
1109
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
n e w e ng l a n d j o u r na l
The
of
m e dic i n e
Weaning attempt: starting on day 3, if FiO2 ≤0.6
findings, and relevant therapeutic interventions
were also recorded just before starting the studydrug infusion and again at 24, 48, 72, and 96 hours.
Physiological variables were also measured daily
between 6 a.m. and 10 a.m. until day 90 or until
hospital discharge of a patient who could breath
spontaneously.
Opioid doses were converted to morphine
equivalents. The equivalencies were as follows:
0.01 mg of sufentanil = 10 mg of morphine = 0.1 mg
of fentanyl = 0.1 mg of remifentanil.16
Barotrauma was defined as newly developed
pneumothorax, pneumomediastinum, subcutaneous emphysema, or pneumatocele larger than 2 cm
in diameter. Muscle strength was evaluated with
the use of the Medical Research Council (MRC)
scale, a previously validated scale that assesses
three muscle groups in each arm and leg. The
score for each muscle group can range from 0
(paralysis) to 5 (normal strength), with the overall
score ranging from 0 to 60.17 The definition of
ICU-acquired paresis was an MRC score of less
than 48.17
Goals during weaning procedure: SpO2 ≥88% and respiratory rate 26–35 cycles
per min
Study Outcomes
Weaning procedure: decrease PEEP over 20–30 min to 5 cm of water
Primary Outcome
Table 1. Summary of the Ventilation Procedure.*
Variable
Ventilator mode: volume assist–control
Initial tidal volume: 6–8 ml/kg of predicted body weight
Plateau pressure: ≤32 cm of water
Oxygenation goal: PaO2 of 55–80 mm Hg or SpO2 of 88–95%
Permitted combinations of FiO2 and PEEP, respectively (cm of water): 0.3
and 5, 0.4 and 5, 0.4 and 8, 0.5 and 8, 0.5 and 10, 0.6 and 10, 0.7 and 10,
0.7 and 12, 0.7 and 14, 0.8 and 14, 0.9 and 14, 0.9 and 16, 0.9 and 18, 1.0
and 18, 1.0 and 20, 1.0 and 22, and 1.0 and 24
pH goal: 7.20–7.45
Procedure when oxygenation goal not achieved despite adjustments to FiO2
and PEEP: use inhaled nitric oxide, almitrine mesylate, prone positioning,
or any combination thereof
Procedure when plateau pressure is >32 cm of water for at least 10 min (in
the following order, as needed): increase sedation, reduce tidal volume to
4 ml/kg, decrease PEEP by decrements of 2 cm of water, and perform injection of cisatracurium in a bolus of 20 mg (not to be given again if plateau pressure decreased by <2 cm of water because further doses would
probably be futile, but permitted if the drug had its intended effect)
Procedure to correct hypercapnia when pH is <7.20 (in the following order,
as needed): connect Y-piece directly to endotracheal tube, increase respiratory rate to a maximum of 35 cycles per min, and increase tidal volume
to a maximum of 8 ml/kg
The primary outcome was the proportion of patients who died before hospital discharge and
If weaning procedure fails at a pressure-support ventilation level of 20 cm
within 90 days after study enrollment (the 90-day
of water, switch to volume assist–control mode of ventilation
mortality). Patients who were outside the hospital
After at least 2 hr of successful pressure-support ventilation at a level of 5 cm
(including those in other types of health care faciliof water, disconnect patient from the ventilator
ties) and who were able to breathe spontaneously
*FiO2 denotes fraction of inspired oxygen, PaO2 partial pressure of arterial oxy- at day 90 were considered to have been discharged
gen, PEEP positive end-expiratory pressure, and SpO2 saturation of peripheral home. Because we anticipated that there would
blood oxygen, as measured by means of pulse oximetry.
be an imbalance in at least one key risk factor at
baseline, the primary outcome was derived from
per liter) or higher. The number of days without a Cox regression model in which we adjusted for
organ or system failure was calculated by subtract- such imbalance. We also report the crude moring the number of days with organ failure from tality at day 90.
28 days or from the number of days until death, if
death occurred before day 28. Organs and systems Secondary Outcomes
were considered to be free of failure after hospital Secondary outcomes were the day-28 mortality, the
discharge. There was no recommendation regard- numbers of days outside the ICU between day 1 and
day 28 and between day 1 and day 90, the number
ing volume-resuscitation goals.
of days without organ or system failure between
Data Collection
day 1 and day 28, the rate of barotrauma, the rate of
During the 24-hour period before randomization, ICU-acquired paresis, the MRC scores on day 28
we recorded data on demographic characteristics, and at the time of ICU discharge, and the numphysiological variables, relevant interventions per- bers of ventilator-free days (days since successful
formed in the ICU, radiographic findings, coexist- weaning from mechanical ventilation) between day
ing conditions, and medications. Data on ventila- 1 and day 28 and between day 1 and day 90. It was
tor settings, physiological variables, radiographic required that the patient breathe spontaneously,
Pressure-support ventilation levels used during weaning procedure: 20, 15,
10, and 5 cm of water
1110
n engl j med 363;12 nejm.org september 16, 2010
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Neuromuscular Blocking Agents in ARDS
without the aid of a ventilator, for a period of at
least 48 hours for weaning from the ventilator to be
considered successful. The number of ventilatorfree days was considered to be zero for patients
who were weaned from mechanical ventilation but
who died before day 28 or day 90.18
Statistical Analysis
Assumptions for the sample-size calculation were
based on our previous studies13,15 that used the
same inclusion criteria and on the European epidemiologic study Acute Lung Injury Verification
(ALIVE).4 Assuming a 50% mortality at 90 days in
the placebo group, we calculated that 340 patients
would need to be enrolled to detect a 15% absolute
reduction in the 90-day mortality in the cisatra­
curium group as compared with the placebo group,
with 80% statistical power and a two-sided alpha
value of 0.05. No interim analysis was performed.
We assessed the differences between the groups
using Student’s t-test, the Wilcoxon test, the chisquare test, or Fisher’s exact test, as appropriate.
All reported P values are two-sided and have not
been adjusted for multiple comparisons. Kaplan–
Meier curves were plotted to assess the time from
enrollment to death and the time to disconnection from the ventilator for a period of at least
48 hours.
The primary analysis consisted of evaluating
the effect of cisatracurium on the primary outcome
(i.e., 90-day mortality), with adjustment by means
of a Cox multivariate proportional-hazards model
that included two predefined covariates: the baseline Simplified Acute Physiology Score (SAPS) II
and the baseline plateau pressure.19 SAPS II is calculated from 12 physiological measurements during a 24-hour period, information about previous
health status, and some information obtained at
admission. This score ranges from 0 to 163, with
higher scores indicating more severe disease. We
planned to include all the variables for which there
was an imbalance between the two groups at baseline, but the only imbalanced variable was the
PaO2:FiO2 ratio. Therefore, we also conducted an
analysis based on the baseline PaO2:FiO2 ratio, in
which the two thirds of patients with a ratio below 120 (indicating hypoxemia) were compared
with the third with a higher ratio. A total of 12
secondary analyses of prespecified outcomes were
performed, and results of 9 of these are reported.
Only one post hoc analysis was conducted; the results are reported.
R e sult s
Baseline Characteristics
We enrolled 340 patients, of whom 178 were randomly assigned to cisatracurium and 162 to placebo. We excluded 986 patients (Fig. 1). One patient
in the cisatracurium group withdrew consent before treatment was started, and data for this patient were therefore not included in the analysis.
The median time from the diagnosis of ARDS to
study inclusion was 16 hours (interquartile range,
6 to 29) in the study population and did not differ significantly between the cisatracurium group
(median, 18 hours; interquartile range, 6 to 31) and
the placebo group (median, 15 hours; interquartile
range, 7 to 27; P = 0.45). The median time from
initiation of mechanical ventilation to study inclusion did not differ significantly between the cis­
atracurium group (22 hours; interquartile range,
9 to 41) and the placebo group (21 hours; interquartile range, 10 to 42; P = 0.91). The only significant difference between the two groups at baseline
was a lower mean PaO2:FiO2 value in the cisatracurium group (P = 0.03) (Table 2, and Table 1 in
the Supplementary Appendix).
Outcomes
Primary Outcome
The Cox regression model yielded a hazard ratio
for death at 90 days in the cisatracurium group, as
compared with the placebo group, of 0.68 (95%
confidence interval [CI], 0.48 to 0.98; P = 0.04), after adjustment for the baseline PaO2:FiO2, SAPS II,
and plateau pressure (Fig. 2). The crude 90-day
mortality was 31.6% (95% CI, 25.2 to 38.8) in the
cisatracurium group and 40.7% (95% CI, 33.5 to
48.4) in the placebo group (P = 0.08).
Secondary Prespecified Outcomes
The beneficial effect of cisatracurium on the 90day survival rate was confined to the two thirds of
patients presenting with a PaO2:FiO2 ratio of less
than 120. Among these patients, the 90-day mortality was 30.8% in the cisatracurium group and
44.6% in the control group (P = 0.04) (Fig. 2 in the
Supplementary Appendix). The absolute difference
in 28-day mortality (mortality in the cisatracurium
group minus mortality in the placebo group) was
−9.6 percentage points (95% CI, −19.2 to −0.2;
P = 0.05) (Table 3).
The cisatracurium group had significantly more
ventilator-free days than the placebo group during
n engl j med 363;12 nejm.org september 16, 2010
1111
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
Table 2. Baseline Characteristics of the Patients, According to Study Group.*
Cisatracurium
(N = 177)
Characteristic†
Age — yr
Placebo
N = 162)
P Value
58±16
58±15
0.70
Tidal volume — ml/kg of predicted body weight
6.55±1.12
6.48±0.92
0.52
Minute ventilation — liters/min
10.0±2.5
10.1±2.2
0.83
PEEP applied — cm of water
9.2±3.2
9.2±3.5
0.87
Plateau pressure — cm of water
25.0±5.1
24.4±4.7
0.32
Respiratory-system compliance — ml/cm of water
31.5±11.6
31.9±10.7
0.71
FiO2
0.79±0.19
0.77±0.20
0.33
PaO2:FiO2‡
106±36
115±41
0.03
pH
7.31±0.10
7.32±0.10
0.11
PaO2 — mm Hg
80±24
85±28
0.09
PaCO2 — mm Hg
47±11
47±11
0.62
33 (18.6)
23 (14.2)
0.31
50±16
47±14
0.15
133 (75.1)
125 (77.2)
0.66
129 (72.9)
113 (69.8)
0.52
Surgical, emergency
27 (15.3)
31 (19.1)
0.34
Surgical, scheduled
21 (11.9)
18 (11.1)
0.83
70 (39.5)
73 (45.1)
0.30
142 (80.2)
123 (75.9)
0.34
Prone position or inhaled nitric oxide or almitrine mesylate — no. (%)
SAPS II§
Nonfatal condition according to McCabe–Jackson score — no. (%)¶
Main reason for ICU admission — no. (%)
Medical
Corticosteroids for septic shock — no. (%)
Direct lung injury — no. (%)
*Plus–minus values are means ±SD. FiO2 denotes fraction of inspired oxygen, ICU intensive care unit, PaCO2 partial
pressure of arterial carbon dioxide, PEEP positive end-expiratory pressure, and SpO2 saturation of peripheral oxygen as
measured by means of pulse oximetry.
†All variables listed except age, nonfatal condition according to McCabe–Jackson score, and main reason for ICU admission were inclusion criteria.
‡Partial pressure of arterial oxygen (PaO2) was measured in millimeters of mercury.
§ The Simplified Acute Physiology Score (SAPS) II is calculated from 12 physiological measurements during a 24-hour
period, information about previous health status, and some information obtained at admission. Scores can range from
0 to 163, with higher scores indicating more severe disease.
¶Possible McCabe–Jackson scores for medical condition are 1 (nonfatal), 2 (ultimately fatal), and 3 (fatal).
the first 28 and 90 days (Table 3, and Fig. 3 in the
Supplementary Appendix). The Cox regression
model yielded an adjusted hazard ratio for weaning from mechanical ventilation by day 90, in the
cisatracurium group as compared with the placebo
group, of 1.41 (95% CI, 1.08 to 1.83; P = 0.01).
The cisatracurium group had more days free of
failure of organs, other than the lungs, during the
first 28 days (15.8±9.9 days, vs. 12.2±11.1 days in
the placebo group; P = 0.01). There were nearly significant between-group differences in the numbers
of days without coagulation abnormalities, hepatic
failure, and renal failure (Table 3). No patient
required dialysis after hospital discharge during
the first 28 days. Significantly more days were
spent outside the ICU between day 1 and day 90 in
the cisatracurium group.
1112
Pneumothorax occurred in a larger proportion of patients in the placebo group (11.7%, vs.
4.0% in the cisatracurium group; P = 0.01) and
tended to develop earlier in the placebo group
(Fig. 4 in the Supplementary Appendix). During
the 48-hour period of study-drug infusion, pneumothorax occurred in one patient (0.6%) in the
cisatracurium group as compared with eight patients (4.9%) in the placebo group (P = 0.03). The
plateau pressures and minute ventilations for the
nine patients are presented in Table 5 in the
Supplementary Appendix. Before the development
of pneumothorax, none of these patients had an
elevated plateau pressure necessitating changes
in the mechanical-ventilation settings, changes in
the sedation regimen, or open-label administration of cisatra­curium.
n engl j med 363;12 nejm.org september 16, 2010
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Neuromuscular Blocking Agents in ARDS
The incidence of ICU-acquired paresis, as
evaluated on the basis of the MRC score on day
28 or at the time of ICU discharge, did not differ
significantly between the two groups (Table 3).
Corticosteroids were used during the ICU stay in
189 patients. There was no significant effect of
cisatracurium use on the 90-day mortality in the
subgroup of patients given corticosteroids (Fig. 6
in the Supplementary Appendix).
Ventilator Settings and Lung Function
Ventilator settings and lung-function variables during the first week are given in Table 7 in the Supplementary Appendix. On day 7, the PaO2:FiO2
ratio was higher, and the PaCO2 value lower, in the
cisatracurium group than in the placebo group.
0.9
0.8
Probability of Survival
Secondary Post Hoc Outcome
1.0
Cisatracurium
0.7
0.6
Placebo
0.5
0.4
0.3
0.2
0.1
0.0
0
10
20
30
40
50
60
70
80
90
Days after Enrollment
Figure 2. Probability of Survival through Day 90, According to Study Group.
Cointerventions
During the ICU stay, there were no significant
between-group differences in the incidence of
cointerventions. A total of 42% of patients in the
cisatracurium group and 48% in the placebo group
were treated with the use of prone positioning,
inhaled nitric oxide, intravenous almitrine mesylate, or a combination of these (Table 8 in the Supplementary Appendix). The criteria for using these
interventions were the same in the two groups.
Open-label cisatracurium was given more frequently in the placebo group than in the cisatra­
cu­rium group during the first 48 hours after
enrollment. However, the two groups did not differ significantly with respect to the number of patients given at least one open-label cisatracuri­um
bolus during the entire ICU stay after enrollment
(Table 8 in the Supplementary Appendix). The required dose of sedatives or analgesics was similar
in the two groups during the first week of the
study (Table 9 in the Supplementary Appendix).
Safety
Bradycardia developed during the cisatracurium
infusion in one patient. No other side effects were
reported.
Discussion
Treatment with the neuromuscular blocking agent
cisatracurium for 48 hours early in the course of
severe ARDS improved the adjusted 90-day survival rate, increased the numbers of ventilatorfree days and days outside the ICU, and decreased
the incidence of barotrauma during the first 90
days. It did not significantly improve the overall
90-day mortality.
Strengths of this trial include the methods used
to minimize bias (blinded randomization assignments, a well-defined study protocol, complete
follow-up, and intention-to-treat analyses). The recruitment of a large number of patients from 20
multidisciplinary ICUs where international standards of care are followed suggests that our data
can be generalized to other ICUs.
Limitations of the trial include the fact that
our results were obtained for cisatracurium bes­
ylate and may not apply to other neuromuscular
blocking agents. Furthermore, we did not assess
the use of a neuromuscular blocking agent late in
the course of ARDS or use on the basis of plateaupressure or transpulmonary-pressure measurements.20 Another limitation is the absence of data
on conditions known to antagonize or potentiate
neuromuscular blockade. However, any condition
that increases the duration of neuromuscular
blockade would have adversely affected the patients
receiving the neuromuscular blocking agent, in
particular by increasing the duration of mechanical ventilation.
The sample-size calculation was based on our
two previous studies performed in four ICUs13,15
that used the same inclusion criteria as were used
in the current trial and on the European epidemiologic study ALIVE.4 However, the mortality in the
placebo group in this study (40.7%) is lower than
n engl j med 363;12 nejm.org september 16, 2010
1113
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
Table 3. Secondary Outcomes, According to Study Group.*
Cisatracurium
(N = 177)
Placebo
(N = 162)
Relative Risk with
Cisatracurium
(95% CI)
P Value
42 (23.7 [18.1–30.5])
54 (33.3 [26.5–40.9])
0.71 (0.51–1.00)
0.05
In the ICU
52 (29.4 [23.2–36.5])
63 (38.9 [31.7–46.6])
0.76 (0.56–1.02)
0.06
In the hospital
57 (32.2 [25.8–39.4])
67 (41.4 [34.1–49.1])
0.78 (0.59–1.03)
0.08
Outcome
Death — no. (% [95% CI])
At 28 days
No. of ventilator-free days†
From day 1 to day 28
10.6±9.7
8.5±9.4
0.04
From day 1 to day 90
53.1±35.8
44.6±37.5
0.03
No cardiovascular failure
18.3±9.4
16.6±10.4
0.12
No. of days without organ failure, from day 1 to day 28
No coagulation abnormalities
22.6±8.9
20.5±9.9
0.05
No hepatic failure
21.3±9.6
19.1±10.6
0.05
No renal failure
20.5±10.1
18.1±11.6
0.05
None of the four
15.8±9.9
12.2±11.1
0.01
6.9±8.2
5.7±7.8
0.16
39.5±35.6
0.03
No. of days outside the ICU
From day 1 to day 28
From day 1 to day 90
Hospital survivors admitted to other health care
facilities from day 1 to day 90 — % (95% CI)
47.7±33.5
22.3 (15.8–30.5)
18.8 (12.2–27.8)
0.52
Barotrauma — no. (% [95% CI])‡
9 (5.1 [2.7–9.4])
19 (11.7 [7.6–17.6])
0.43 (0.20–0.93)
0.03
Pneumothorax — no. (% [95% CI])
7 (4.0 [2.0–8.0])
19 (11.7 [7.6–17.6])
0.34 (0.15–0.78)
0.01
At day 28
55 (46–60)
55 (39–60)
1.07 (0.80–1.45)
0.49
At ICU discharge
55 (43–60)
55 (44–60)
0.92 (0.71–1.19)
0.94
MRC score — median (IQR)§
Patients without ICU-acquired paresis¶
By day 28 — no./total no. (% [95% CI])
By ICU discharge — no./total no. (% [95% CI])
68/96 (70.8 [61.1–79.0]) 52/77 (67.5 [56.5–77.0])
0.64
72/112 (64.3 [55.1–72.6]) 61/89 (68.5 [58.3–77.3])
0.51
*Plus–minus values are means ±SD. ICU denotes intensive care unit, and IQR interquartile range.
†The number of ventilator-free days was defined as the number of days since successful weaning from mechanical ventilation after a period
of spontaneous breathing lasting at least 48 consecutive hours.
‡Barotrauma was defined as any new pneumothorax, pneumomediastinum, subcutaneous emphysema, or pneumatocele larger than 2 cm in
diameter.
§ The Medical Research Council (MRC) scale is a previously validated scale that assesses strength in three muscle groups in each arm and
leg. The score for each muscle group can range from 0 (paralysis) to 5 (normal strength), with the overall score ranging from 0 to 60.17
¶ICU-acquired paresis was defined as an overall MRC score of less than 48.
that in the control groups in the earlier studies.
Given the observed mortality in our placebo group,
the current study was underpowered. Indeed, 885
patients would have been needed to be enrolled to
achieve 80% statistical power with a two-sided
alpha value of 0.05.
Finally, all our patients had severe ARDS. Additional work is needed to determine whether the
1114
use of neuromuscular blocking agents for only
24 hours is beneficial in selected patients. In our
general analysis, which was prespecified but with
post hoc determination of the threshold value for
classifying subgroups, we found that the beneficial
effect of the neuromuscular blocking agent on
survival was confined to the two thirds of patients
with a PaO2:FiO2 ratio below 120.
n engl j med 363;12 nejm.org september 16, 2010
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Neuromuscular Blocking Agents in ARDS
The mechanisms underlying the beneficial
effect of neuromuscular blocking agents remain
speculative. A brief period of paralysis early in
the course of ARDS may facilitate lung-protective
mechanical ventilation by improving patient–
ventilator synchrony and allowing for the accurate
adjustment of tidal volume and pressure levels,
thereby limiting the risk of both asynchronyrelated alveolar collapse and regional alveolarpressure increases with overdistention. Another
possible mechanism of the benefit involves a decrease in lung or systemic inflammation.15
The main safety concern with the use of a
neuromuscular blocking agent is muscle weakness; the risk varies among agents.21,22 Steroidal
compounds (vecuronium, pancuronium, and ro­
cu­ronium) may carry the highest risk of myopathy,23 although myopathy has also been reported
with benzylisoquinolines, including cisatracurium besylate.24,25 Muscle weakness was not in-
creased significantly by the use of the neuromuscular blocking agent in our study. The short
duration of use of the neuromuscular blocking
agent probably explains this result.
In conclusion, this multicenter trial provides
evidence that the administration of a neuromuscular blocking agent early in the course of severe
ARDS managed with low-tidal-volume ventilation
may improve outcomes. Future studies are needed to replicate and expand these findings before
they can be widely adopted in clinical practice.
Supported by the Assistance Publique–Hôpitaux de Marseille
and a grant from the Ministère de la Santé (Programme Hospitalier de Recherche Clinique Régional 2004-26). GlaxoSmithKline
France provided the study drugs.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
We thank Kahena Amichi, Daniel Pulina, François Voillet,
Patrick Sudour, Teodora Iordanova, Jean-Charles Reynier, Mohamed Fathallah, Karine Barrau, Pascal Auquier, Yann Lebras,
Nathalie Boggi, Ammar Zerrar, Cécile Roché-Thiaux, Fabienne
Brégeon, Didier Dreyfuss, Marie-Josée Bonavita, Marie-Dominique Chollet, Didier Sanchez, and Antoinette Wolfe.
Appendix
The ARDS et Curarisation Systematique (ACURASYS) study investigators are as follows (all in France): Marseille: Hôpital Nord — J.
Albanese, V. Blaso, M. Leone, F. Antonini, P. Visintini; Hôpital Sainte-Marguerite, Réanimation des Urgences — G. Perrin, D. Blayac, B.
Eon, P. Michelet, P. Saux, D. Lambert, V. Fulachier; Hôpital Sainte-Marguerite, Réanimation Médicale — J.M. Forel, A. Roch, C. Guervilly, J.
Allardet-Servent, D. Demory, N. Embriaco, M. Gainnier, L. Papazian; Hôpital Ambroise Paré — J.M. Seghboyan, N. Beni Chougrane, R.
Soundaravelou, A. Piera, M. Bonnetty. Paris: Hôpital Saint-Louis — E. Azoulay. Besançon: Hôpital Jean Minjoz — D. Perez, G. Capellier, P.
Midez, C. Patry, E. Laurent, E. Belle, J.C. Navellou. Clermont-Ferrand: Hôpital Hôtel-Dieu — J.M. Constantin, S. Cayot-Constantin, R.
Guerin. Avignon: Centre Hospitalier — P. Courant, T. Signouret, K. Pavaday, P. Garcia, L. Delapierre, K. Debbat. Toulon: Hôpital Font-Pré
— J.M. Arnal, J. Durand-Gasselin, G. Corno, A. Orlando, I. Granier, S.Y. Donati. Rennes: Hôpital Pontchaillou — A. Gacouin, Y. Le
Tulzo, S. Lavoue, C. Camus. Nîmes: Hôpital Caremeau — C. Gervais, C. Bengler, C. Arich, J.Y. Lefrant, P. Poupard, L. Muller, P. Joubert,
R. Cohendy, G. Saissi, P. Barbaste. Lyon: Hôpital de la Croix-Rousse — C. Guérin, F. Bayle, J.C. Richard, J.M. Badet. Bordeaux: Hôpital
Pellegrin — G. Hilbert, H. Bruno, E. Rosier, W. Pujol, H.N. Bui. Montpellier: Hôpital Saint-Eloi — S. Jaber, B. Jung, M. Sebbane, G.
Chanques. Brest: Hôpital La Cavale Blanche — G. Prat, E. L’Her, J.M. Tonnelier. Aix-en-Provence: Centre Hospitalier du Pays d’Aix — O.
Baldesi, C. Leroy, L. Rodriguez, J.L. Le Grand, B. Garrigues. Nice: Hôpital de l’Archet — G. Bernardin, J. Dellamonica. Grenoble: Hôpital
Sud — C. Schwebel, J.F. Timsit, R. Hamidfar, L. Hammer, G. Dessertaine, C. De Couchon, A. Bonadona. Saint-Etienne: Hôpital Bellevue
— F. Zeni, D. Thevenet, C. Venet, S. Guyomarc’h.
References
1. Ware LB, Matthay MA. The acute re-
spiratory distress syndrome. N Engl J Med
2000;342:1334-49.
2. Malhotra A. Low-tidal-volume ventilation in the acute respiratory distress syndrome. N Engl J Med 2007;357:1113-20.
3. Bernard GR. Acute respiratory distress
syndrome: a historical perspective. Am J
Respir Crit Care Med 2005;172:798-806.
4. Brun-Buisson C, Minelli C, Bertolini
G, et al. Epidemiology and outcome of
acute lung injury in European intensive
care units: results from the ALIVE study.
Intensive Care Med 2004;30:51-61.
5. Esteban A, Anzueto A, Frutos F, et al.
Characteristics and outcomes in adult patients receiving mechanical ventilation: a
28-day international study. JAMA 2002;
287:345-55.
6. Esteban A, Ferguson ND, Meade MO,
et al. Evolution of mechanical ventilation
in response to clinical research. Am J Respir Crit Care Med 2008;177:170-7.
7. Rubenfeld GD, Caldwell E, Peabody E,
et al. Incidence and outcomes of acute lung
injury. N Engl J Med 2005;353:1685-93.
8. Hansen-Flaschen JH, Brazinsky S,
Basile C, Lanken PN. Use of sedating
drugs and neuromuscular blocking agents
in patients requiring mechanical ventilation for respiratory failure: a national survey. JAMA 1991;266:2870-5.
9. Mehta S, Burry L, Fischer S, et al. Canadian survey of the use of sedatives, analgesics, and neuromuscular blocking
agents in critically ill patients. Crit Care
Med 2006;34:374-80.
10. Murray MJ, Cowen J, DeBlock H, et al.
Clinical practice guidelines for sustained
neuromuscular blockade in the adult critically ill patient. Crit Care Med 2002;30:
142-56.
11. Samuelson KA, Larsson S, Lundberg
D, Fridlund B. Intensive care sedation of
mechanically ventilated patients: a national Swedish survey. Intensive Crit Care
Nurs 2003;19:350-62.
12. Vender JS, Szokol JW, Murphy GS, Nitsun M. Sedation, analgesia, and neuromuscular blockade in sepsis: an evidencebased review. Crit Care Med 2004;32:
Suppl:S554-S561.
13. Gainnier M, Roch A, Forel JM, et al.
Effect of neuromuscular blocking agents
on gas exchange in patients presenting
with acute respiratory distress syndrome.
Crit Care Med 2004;32:113-9.
n engl j med 363;12 nejm.org september 16, 2010
1115
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Neuromuscular Blocking Agents in ARDS
14. The Acute Respiratory Distress Syn-
drome Network. Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute lung injury and
the acute respiratory distress syndrome.
N Engl J Med 2000;342:1301-8.
15. Forel JM, Roch A, Marin V, et al. Neuro­
muscular blocking agents decrease inflammatory response in patients presenting
with acute respiratory distress syndrome.
Crit Care Med 2006;34:2749-57.
16. Reisine T, Pasternak G. Opioid analgesics and antagonists. In: Hardman J,
Limbird L, eds. The pharmacological basis of therapeutics. 9th ed. New York:
McGraw-Hill, 1996:521-55.
17. De Jonghe B, Sharshar T, Lefaucheur
JP, et al. Paresis acquired in the intensive
care unit: a prospective multicenter study.
JAMA 2002;288:2859-67.
18. Schoenfeld DA, Bernard GR. Statisti-
cal evaluation of ventilator-free days as an
efficacy measure in clinical trials of treatments for acute respiratory distress syndrome. Crit Care Med 2002;30:1772-7.
19. Hager DN, Krishnan JA, Hayden DL,
Brower RG. Tidal volume reduction in patients with acute lung injury when plateau
pressures are not high. Am J Respir Crit
Care Med 2005;172:1241-5.
20. Talmor D, Sarge T, Malhotra A, et al.
Mechanical ventilation guided by esophageal pressure in acute lung injury. N Engl
J Med 2008;359:2095-104.
21. Segredo V, Caldwell JE, Matthay MA,
Sharma ML, Gruenke LD, Miller RD. Persistent paralysis in critically ill patients
after long-term administration of vecuronium. N Engl J Med 1992;327:524-8.
22. Murray MJ, Coursin DB, Scuderi PE, et
al. Double-blind, randomized, multicenter study of doxacurium vs. pancuronium in intensive care unit patients who
require neuromuscular-blocking agents.
Crit Care Med 1995;23:450-8.
23. Testelmans D, Maes K, Wouters P, et
al. Rocuronium exacerbates mechanical
ventilation-induced diaphragm dysfunction in rats. Crit Care Med 2006;34:301823.
24. Davis NA, Rodgers JE, Gonzalez ER,
Fowler AA III. Prolonged weakness after
cisatracurium infusion: a case report. Crit
Care Med 1998;26:1290-2.
25. Leatherman JW, Fluegel WL, David
WS, Davies SF, Iber C. Muscle weakness
in mechanically ventilated patients with
severe asthma. Am J Respir Crit Care Med
1996;153:1686-90.
Copyright © 2010 Massachusetts Medical Society.
posting presentations at medical meetings on the internet
Posting an audio recording of an oral presentation at a medical meeting on the
Internet, with selected slides from the presentation, will not be considered prior
publication. This will allow students and physicians who are unable to attend the
meeting to hear the presentation and view the slides. If there are any questions
about this policy, authors should feel free to call the Journal’s Editorial Offices.
1116
n engl j med 363;12 nejm.org september 16, 2010
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on July 16, 2013. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
new england
journal of medicine
The
established in 1812
november 13, 2008
vol. 359 no. 20
Mechanical Ventilation Guided by Esophageal Pressure
in Acute Lung Injury
Daniel Talmor, M.D., M.P.H., Todd Sarge, M.D., Atul Malhotra, M.D., Carl R. O’Donnell, Sc.D., M.P.H.,
Ray Ritz, R.R.T., Alan Lisbon, M.D., Victor Novack, M.D., Ph.D., and Stephen H. Loring, M.D.
A bs t r ac t
Background
Survival of patients with acute lung injury or the acute respiratory distress syndrome
(ARDS) has been improved by ventilation with small tidal volumes and the use of
positive end-expiratory pressure (PEEP); however, the optimal level of PEEP has been
difficult to determine. In this pilot study, we estimated transpulmonary pressure
with the use of esophageal balloon catheters. We reasoned that the use of pleuralpressure measurements, despite the technical limitations to the accuracy of such
measurements, would enable us to find a PEEP value that could maintain oxygenation
while preventing lung injury due to repeated alveolar collapse or overdistention.
Methods
We randomly assigned patients with acute lung injury or ARDS to undergo mechanical ventilation with PEEP adjusted according to measurements of esophageal pressure
(the esophageal-pressure–guided group) or according to the Acute Respiratory Distress Syndrome Network standard-of-care recommendations (the control group).
The primary end point was improvement in oxygenation. The secondary end points
included respiratory-system compliance and patient outcomes.
Results
The study reached its stopping criterion and was terminated after 61 patients had
been enrolled. The ratio of the partial pressure of arterial oxygen to the fraction of
inspired oxygen at 72 hours was 88 mm Hg higher in the esophageal-pressure–
guided group than in the control group (95% confidence interval, 78.1 to 98.3;
P = 0.002). This effect was persistent over the entire follow-up time (at 24, 48, and 72
hours; P = 0.001 by repeated-measures analysis of variance). Respiratory-system compliance was also significantly better at 24, 48, and 72 hours in the esophagealpressure–guided group (P = 0.01 by repeated-measures analysis of variance).
From the Department of Anesthesia,
Critical Care, and Pain Medicine, Beth Israel Deaconess Medical Center (D.T.,
T.S., R.R., A.L., S.H.L.); the Division of
Pulmonary and Critical Care and the Division of Sleep Medicine, Brigham and
Women’s Hospital (A.M.); the Division of
Pulmonary, Critical Care, and Sleep Medicine, Beth Israel Deaconess Medical
Center (C.R.O.); the Harvard Clinical Research Institute (V.N.); and Harvard
Medical School (D.T., T.S., A.M., C.R.O.,
A.L., S.H.L.) — all in Boston. Address reprint requests to Dr. Talmor at the Department of Anesthesia, Critical Care,
and Pain Medicine, Beth Israel Deaconess Medical Center, 1 Deaconess Rd., CC470, Boston, MA 02215, or at dtalmor@
bidmc.harvard.edu.
This article (10.1056/NEJMoa0708638) was
published at www.nejm.org on November
11, 2008.
N Engl J Med 2008;359:2095-104.
Copyright © 2008 Massachusetts Medical Society.
Conclusions
As compared with the current standard of care, a ventilator strategy using esophageal
pressures to estimate the transpulmonary pressure significantly improves oxygenation and compliance. Multicenter clinical trials are needed to determine whether this
approach should be widely adopted. (ClinicalTrials.gov number, NCT00127491.)
n engl j med 359;20 www.nejm.org november 13, 2008
2095
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
The
R
n e w e ng l a n d j o u r na l
ecent changes in the practice of
mechanical ventilation have improved survival in patients with the acute respiratory
distress syndrome (ARDS), but mortality remains
unacceptably high. Whereas low tidal volumes
are clearly beneficial in patients with ARDS, how
to choose a positive end-expiratory pressure (PEEP)
is uncertain.1-4 Ideally, mechanical ventilation
should provide sufficient transpulmonary pressure (airway pressure minus pleural pressure) to
maintain oxygenation while minimizing repeated alveolar collapse or overdistention leading to
lung injury.5 In critical illness, however, there is
marked variability among patients in abdominal
and pleural pressures6,7; thus, for a given level of
PEEP, transpulmonary pressures may vary unpredictably from patient to patient.7
We estimated pleural pressure with the use of
an esophageal balloon catheter. Although this
technique has been validated in healthy human
subjects and animals, it has not been systematically applied in patients in the intensive care setting. We reasoned that we could adjust PEEP
according to each patient’s lung and chest-wall
mechanics.8-10 We speculated that in patients with
high estimated pleural pressure who are undergoing ventilation with conventional ventilator
settings, underinflation may cause hypoxemia. In
such patients, raising PEEP to maintain a positive
transpulmonary pressure might improve aeration
and oxygenation without causing overdistention.
Conversely, in patients with low pleural pressure,
maintaining low PEEP would keep transpulmonary pressure low, preventing overdistention and
minimizing the adverse hemodynamic effects of
high PEEP.11
We report the results of a randomized, controlled pilot trial involving patients with acute
lung injury or ARDS. The trial compared mechan­
ical ventilation directed by esophageal-pressure
measurements with mechanical ventilation managed according to the Acute Respiratory Distress
Syndrome Network (ARDSNet) recommendations.12 We tested the hypothesis that oxygenation
in patients can be improved by adjusting PEEP to
maintain positive transpulmonary pressures.
of
m e dic i n e
was approved by the institutional review board of
the center, and written informed consent was obtained from the patients or their nearest relatives.
No commercial entities providing equipment or
devices had a role in any aspect of this study.
Patients were included in the study if they
had acute lung injury or ARDS according to the
American–European Consensus Conference definitions.13 The exclusion criteria included recent
injury or other pathologic condition of the esophagus, major bronchopleural fistula, and solidorgan transplantation.
Measurements and Experimental Protocol
While undergoing treatment, the subjects were
supine, with the head of the bed elevated to 30
degrees. Airway pressure, tidal volume, and air
flow were recorded during mechanical ventilation.
An esophageal balloon catheter was passed to a
depth of 60 cm from the incisors for measurement of gastric pressure and then withdrawn to
a depth of 40 cm to record esophageal pressure
during mechanical ventilation. Placement of the
balloon in the stomach was confirmed by a transient increase in pressure during a gentle compression of the abdomen and by a qualitative
change in the pressure tracing (i.e., an increased
cardiac artifact) as the balloon was withdrawn
into the esophagus. In approximately one third
of the patients, the balloon could not be passed
into the stomach, and esophageal placement was
confirmed by the presence of a cardiac artifact
and the changes in transpulmonary pressure during tidal ventilation. The mixed expired partial
pressure of carbon dioxide was measured to allow calculation of physiological dead space. After
these initial measurements, patients were randomly assigned with the use of a block-randomization scheme to the control or esophagealpressure–guided group.
Each patient, while under heavy sedation or
paralysis, underwent a recruitment maneuver to
standardize the history of lung volume,14 in which
airway pressure was increased to 40 cm of water
for 30 seconds. If needed, a lower pressure was
used to keep the transpulmonary pressure (the
difference between the airway pressure and the
esophageal pressure) in the physiologic range
Me thods
(<25 cm of water while the patient is in the supine
Patients
position).15 After the recruitment maneuver, the
We performed the trial in the medical and surgi- patient underwent mechanical ventilation accordcal intensive care units (ICUs) of Beth Israel Dea- ing to the treatment assignment.
coness Medical Center in Boston. The protocol
The patients in the esophageal-pressure–guided
2096
n engl j med 359;20 www.nejm.org november 13, 2008
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Mechanical Ventilation in Acute Lung Injury
group underwent mechanical ventilation with settings determined by the initial esophageal-pressure measurements. Tidal volume was set at 6 ml
per kilogram of predicted body weight. The predicted body weight of male patients was calculated as 50 + 0.91 × (centimeters of height – 152.4)
and that of female patients as 45.5 + 0.91 × (centimeters of height − 152.4). PEEP levels were set to
achieve a transpulmonary pressure of 0 to 10 cm
of water at end expiration, according to a sliding
scale based on the partial pressure of arterial
oxygen (PaO2) and the fraction of inspired oxygen
(FiO2) (Fig. 1). We also limited tidal volume to
keep transpulmonary pressure at less than 25 cm
of water at end inspiration, although this limit
was rarely approached, and tidal volume was
never reduced for this purpose.
Patients in the control group were treated according to the low-tidal-volume strategy reported
by the ARDSNet study of the National Heart,
Lung, and Blood Institute.12 This strategy specifies that the tidal volume is set at 6 ml per kilogram of predicted body weight and PEEP is based
on the patient’s PaO2 and FiO2 (Fig. 1).
In both groups, the goals of mechanical ventilation included a PaO2 of 55 to 120 mm Hg or a
pulse-oximeter reading of 88 to 98%, an arterial
pH of 7.30 to 7.45, and a partial pressure of arterial carbon dioxide (PaCO2) of 40 to 60 mm Hg,
according to the sliding scales in Figure 1. To
reduce the need for frequent manipulation of the
ventilator settings, the goals for oxygenation in
both groups were relaxed from the narrow range
of PaO2 values in the ARDSNet study (55 to 80
mm Hg) to a broader range of 55 to 120 mm Hg.
All measurements were repeated 5 minutes
after the initiation of experimental or control ventilation and again at 24, 48, and 72 hours. Measurements were also performed as needed after
changes were made to ventilator settings because
of any clinically significant change in the patient’s condition.
Therapies other than mechanical ventilation
were administered by members of the primary
ICU team, who were unaware of the results of the
esophageal-pressure measurements. To avert complications, these team members used protocols
to guide hemodynamic resuscitation,16 sedation,
weaning from ventilation, and other standard
interventions related to ventilator care.17 These
care standards were aggressively applied in both
groups. After the measurements at 72 hours, the
results of pressure measurements were made
available to the caregivers, who were free to use
or not use them for decisions concerning treatment and ventilator management.
The primary end point of the study was arterial oxygenation, as measured by the ratio of PaO2
to FiO2 (PaO2:FiO2) 72 hours after randomization.
The secondary end points included indexes of
lung mechanics and gas exchange (respiratorysystem compliance and the ratio of physiological
dead space to tidal volume), as well as outcomes
of the patients (the number of ventilator-free days
at 28 days, length of stay in the ICU, and death
within 28 days and 180 days after treatment).
Esophageal-Pressure–Guided Group
FIO2
PLexp
0.4
0
0.5
0
0.5
2
0.6
2
0.6
4
0.7
4
0.7
6
0.8
6
0.8
8
0.9
8
0.9
10
1.0
10
Control Group
FIO2
PEEP
0.3
5
0.4
5
0.4
8
0.5
8
0.5
10
0.6
10
0.7
10
0.7
12
0.7
14
0.8
14
0.9
14
0.9
16
0.9
18
1.0
20–24
Figure 1. Ventilator Settings According to the Protocol.
For the intervention group, keep the partial pressure of arterial oxygen (PaO
55 and 120 mm Hg or keep
2) between
1st
RETAKE
AUTHOR: Talmor
ICM
the oxygen saturation, as measured
by pulse oximeter, between 88 and 98% by using
2nd the ventilator settings in one
FIGURE:
1
of
2
REG F
column at a time. Set the positive end-expiratory
pressure (PEEP) at such a level that
3rd transpulmonary pressure durRevised
ing end-expiratory occlusion (PL exp)CASE
stays between 0 and 10 cm of water, and
keep transpulmonary pressure during
Line
4-C
EMail25 cm of water. For the control group, keep
SIZE PaO2 between 55 and 120 mm Hg
end-inspiratory occlusion at less than
ARTIST: ts
H/T
H/T
Enonto pulse oximeter between 88 and 98%)33p9
(or keep oxygen saturation according
by using the ventilator settings in one
Combo
column at a time. Set the PEEP and tidal volume at such levels that the airway pressure during end-inspiratory ocAUTHOR, PLEASE NOTE:
clusion stays at less than 30 cm of water.
In has
both
groups,
apply
ventilation
Figure
been
redrawn
and type
has beenwith
reset.either pressure-control ventilation or
Please
check
carefully. time between 1:1 and 1:3 to minimize dysvolume-control ventilation with a ratio of inspiratory
time
to expiratory
synchrony between the patient and the ventilator while achieving a tidal volume of 6±2 ml per kilogram of predicted
JOB:of35915
ISSUE: 10-09-08 maneuvers are permitted to
body weight and a respiratory rate
35 breaths per minute or less. Lung-recruitment
reverse episodic hypoxemia after suctioning or inadvertent airway disconnection, but not on a routine basis.
n engl j med 359;20 www.nejm.org november 13, 2008
2097
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
Statistical Analysis
In evaluating the PaO2:FiO2 at 72 hours, we decided a priori that a clinically important change
in the PaO2:FiO2 would be approximately 20%,
with measurement error taken into account. To
determine sample size, we chose a minimal average between-groups difference of 40 in the
PaO2:FiO2. We conservatively estimated the standard deviation to be 100 (equivalent to a coefficient of variation of 250%); on the basis of this
estimate, a sample of 100 patients per group would
be required to detect a difference of 40 in the
PaO2:FiO2 with 80% power and a two-tailed alpha
value of 0.05. Because of the uncertainty in the
estimate of standard deviation, we designed the
study with the aid of a data safety and monitoring board, whose members were not involved in
patient care or data gathering. The board members were instructed to perform an interim analysis after 60 patients had been enrolled, at which
point they could recommend stopping the trial if
an overwhelming effect was detected on the basis
of the critical significance level (P≤0.02), as adjusted for the Lan–DeMets alpha-spending function with Pocock boundary. The members of this
board also participated in the writing of this
article.
Continuous variables with normal distribution
are presented as means (±SD) and compared with
the use of Student’s t-test. Continuous variables
with non-normal distributions are presented as
medians and interquartile ranges and compared
with the use of the Mann–Whitney test. Dichotomous or nominal categorical variables are compared with the use of the chi-square test with
normal approximation or Fisher’s exact test, as
appropriate. We assessed the trend over time in
respiratory measurements by comparing the control group and the esophageal-pressure–guided
group at 24, 48, and 72 hours with the use of the
F test with one degree of freedom for a general
linear model with repeated measures. We used
sequential hypothesis testing for the assessment
of differences between the groups at 72 hours
and 24 hours. When a statistically significant difference was found at 72 hours, we performed a
repeated-measures analysis and then compared
the values at 24 hours. Kaplan–Meier analysis
with the log-rank test was applied to compare
survival at 180 days between the groups.
In a single prespecified analysis, we adjusted
the relative risk by using the Acute Physiology and
Chronic Health Evaluation (APACHE II) score to
2098
of
m e dic i n e
estimate the effect of study group on the risk of
death within 28 days after treatment. The relative
risk was estimated by Poisson regression with
conservative robust error variance.18,19 For death
within 180 days after treatment, we used a Cox
proportional-regression model to compare the
control and treatment groups, with adjustment
for the APACHE II score at admission. A twotailed P value less than 0.05 was considered to
indicate statistical significance.
R e sult s
The characteristics of the patients in the two
groups were well matched at baseline (Table 1).
Most patients in both groups were severely ill,
with a mean (±SD) APACHE II score of 26.6±6.4
and a median of two failed organs (interquartile
range, one to three). We were unable to sedate
one patient in the esophageal-pressure–guided
group sufficiently to obtain stable esophagealpressure measurements; this patient is included
in the analysis on the basis of the intention-totreat principle. There were no adverse events or
incidents of barotrauma in either group.
We stopped the study after 61 patients had
been enrolled, because the planned interim analy­
sis showed that it had reached the prespecified
stopping criterion. The PaO2:FiO2 at 72 hours was
88 mm Hg higher in patients treated with mechanical ventilation with esophageal balloons
than in control patients (95% confidence interval
[CI], 78.1 to 98.3; P = 0.002) (Table 2).
Physiological Measurements
The ventilator settings and physiological measurements at baseline were similar in the two groups
(Table 2). Forty-nine patients (80%), including
the one patient that we were unable to sedate,
met the criteria for ARDS (PaO2:FiO2 <200 mm Hg)
(see Table 1 in the Supplementary Appendix,
available with the full text of this article at www.
nejm.org), and there was no significant difference in baseline PaO2:FiO2 between the groups.
The average tidal volume was reduced during the
first day of therapy by 67 ml in the control group
(P<0.001 by paired t-test) and by 44 ml in the
esophageal-pressure–guided group (P<0.001 by
paired t-test).
Oxygenation and respiratory-system compliance improved in the esophageal-pressure–guided
group as compared with the control group, whereas the ratio of dead space to tidal volume did not
n engl j med 359;20 www.nejm.org november 13, 2008
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Mechanical Ventilation in Acute Lung Injury
Table 1. Baseline Characteristics of the Patients.*
Esophageal-Pressure–Guided
(N = 30)
Characteristic
Male sex — no. (%)
Conventional Treatment
(N = 31)
P Value
0.44
19 (63)
17 (55)
54.5±16.1
51.2±23.0
0.52
26 (87)
27 (87)
0.96
Predicted body weight — kg
67.1±8.9
63.2±11.1
0.14
APACHE II score at admission
26.3±6.4
26.8±6.5
0.76
Age — yr
White race — no. (%)†
Primary physiological injury — no. (%)‡
0.54
Pulmonary
7 (23)
5 (16)
Abdominal
13 (43)
11 (35)
Trauma
6 (20)
9 (29)
Sepsis
3 (10)
2 (6)
Other
1 (3)
4 (13)
Organ failure at baseline — no. (%)
Cardiac
10 (33)
10 (32)
0.93
Renal
19 (63)
16 (52)
0.36
Neurologic
12 (40)
12 (39)
0.92
Hepatic
11 (37)
10 (32)
0.72
7 (23)
5 (16)
0.48
7.34±0.09
7.32±0.08
0.34
Hematologic
Arterial blood gases at baseline
pH
PaCO2 — mm Hg
42±8
40±8
0.23
PaO2 — mm Hg§
91±25
107±44
0.09
Bicarbonate — mmol/liter
24±5
22±4
0.05
3.1±3.5
3.4±3.3
0.83
Hemodynamic variables at baseline
Lactate — mg/dl
Heart rate — beats/min
98±26
100±19
0.71
Systolic blood pressure — mm Hg
108±18
107±18
0.80
Diastolic blood pressure — mm Hg
58±11
54±11
0.20
Central venous pressure — mm Hg
16±5
16±4
0.96
*Plus–minus values are means ±SD. APACHE denotes Acute Physiology and Chronic Health Evaluation, PaCO2 the partial pressure of arterial carbon dioxide, and PaO2 the partial pressure of arterial oxygen.
†Race was determined by the investigators.
‡Pulmonary injury included pneumonia (nine patients), aspiration pneumonitis (two), and smoke inhalation (one). Abdom­
inal injury included bowel obstruction (four patients), abdominal surgery (four), pancreatitis (four), cholangitis (two),
small-bowel perforation (three), ruptured aortic aneurysm or surgery for aortic aneurysm (two), gastrointestinal bleeding
(one), Crohn’s disease (one), end-stage liver disease (one), ischemic bowel (one), and perforated viscus (one). Trauma
included motor vehicle accident (five patients), multiple trauma (eight), abdominal gunshot wound (one), and traumatic
brain injury (one). Other injuries included drug overdose (three patients), intraventricular hemorrhage (one), and hypoxic
respiratory failure (one).
§ For the PaO2, the values for the fraction of inspired oxygen were as follows: for the esophageal-pressure–guided group,
the median was 0.6 and the interquartile range was 0.5 to 0.8; for the conventional-treatment group, the median was 0.7
and the interquartile range was 0.6 to 1.0.
significantly differ between the groups during
the first 72 hours (Fig. 2C). The PaO2:FiO2 improved during the first 72 hours by 131 mm Hg
(95% CI, 79 to 182) in the esophageal-pressure–
guided group and by 49 mm Hg (95% CI, 12 to
86) in the control group (Table 2). The higher
value of PaO2:FiO2 in the esophageal-pressure–
guided group than in the control group was evident at 24 hours (P = 0.04) (Fig. 2A). Respiratorysystem compliance was significantly improved
and was higher in the esophageal-pressure–
guided group than in the control group (P = 0.01;
n engl j med 359;20 www.nejm.org november 13, 2008
2099
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
Table 2. Measurements of Ventilatory Function at Baseline and 72 Hours.*
Measurement
Baseline
EsophagealConventional
Pressure–Guided
Treatment
(N = 30)
(N = 31)
PaO2:FiO2
Respiratory-system compliance
(ml/cm of water)
Ratio of physiological dead space to tidal
volume
PaO2 (mm Hg)
FiO2
PEEP (cm of water)
Tidal volume (ml)
Tidal volume (ml per kg of predicted body
weight)
Respiratory rate (breaths/min)
Inspiratory time (sec)
PEEPtotal (cm of water)
Peak inspiratory pressure (cm of water)
Mean airway pressure (cm of water)
Plateau pressure (cm of water)
Transpulmonary end-inspiratory pressure
(cm of water)
Transpulmonary end-expiratory pressure
(cm of water)
Esophageal end-inspiratory pressure
(cm of water)
Esophageal end-expiratory pressure
(cm of water)
72 Hr†
P Value
EsophagealPressure–Guided
(N = 29)
Conventional
Treatment
(N = 29)
147±56
36±12
145±57
36±10
0.89
0.94
280±126
45±14
191±71
35±9
P Value
0.002
0.005
0.67±0.11
0.67±0.09
0.95
0.61±0.09
0.64±0.10
0.27
91±25
0.66±0.17
13±5
484±98
7.3±1.3
107±44
0.77±0.18
13±3
491±105
7.9±1.4
0.09
0.02
0.73
0.80
0.12
124±44
0.49±0.17
17±6
472±98
7.1 ±1.3
101±33
0.57±0.18
10±4
418±80
6.8±1
0.03
0.07
<0.001
0.03
0.31
26±6
0.8±0.1
14±5
35±8
20±6
29±7
7.9±6.0
24±6
0.9±0.2
15±4
35±7
20±4
29±5
8.6±5.4
0.32
0.19
0.67
0.85
0.88
0.79
0.61
26±6
0.8±0.1
18±5
32±8
22±6
28±7
7.4±4.4
28±5
0.8±0.1
12±5
28±7
16±5
25±6
6.7±4.9
0.20
0.27
<0.001
0.007
0.001
0.07
0.58
−2.8±5.0
−1.9±4.7
0.49
0.1±2.6
−2.0±4.7
0.06
21.2±4.9
20.7±5.1
0.68
21.7±7.2
17.9±5.2
0.03
17.2±4.4
16.9±5.0
0.79
18.4±5.9
14.3±4.9
0.008
*Plus–minus values are means ±SD. FiO2 denotes the fraction of inspired oxygen, PaO2 the partial pressure of arterial oxygen, PEEP positive
end-expiratory pressure applied by the ventilator, and PEEPtotal airway pressure measured during end-expiratory occlusion.
†The values are given for the 29 surviving patients in each treatment group.
repeated-measures analysis of variance at 24, 48,
and 72 hours) (Fig. 2B).
On the first therapeutic day, PEEP was
changed by less than 5 cm of water in all but one
of the control patients, whereas patients in the
esophageal-pressure–guided group had variable
and often substantial increases in PEEP (Table 3)
and significantly higher PEEP at 24, 48, and 72
hours (Fig. 2D, and Fig. 1 in the Supplementary
Appendix). At 24 hours, the difference in PEEP
between the groups reached 7.7 cm of water
(95% CI, 5.5 to 9.9), with a mean PEEP in the
esophageal-pressure–guided group of 18.7±5.1 cm
of water, although in 3 of the 31 patients in this
group, the initial PEEP level was decreased on the
basis of initial transpulmonary pressure. At 24,
48, and 72 hours, the mean transpulmonary endexpiratory pressure remained above zero in the
esophageal-pressure–guided group, whereas it remained negative in the control group (P<0.001 by
2100
repeated-measures analysis of variance) (Fig. 2E).
The plateau airway pressure during end-inspiratory occlusion was higher in the esophagealpressure–guided group than in the control group
(P = 0.003 by repeated-measures analysis of variance) (Fig. 2F, and Fig. 1 in the Supplementary
Appendix). However, transpulmonary pressures
during end-inspiratory occlusion never exceeded
24 cm of water and did not differ significantly
between the groups (P = 0.13 by repeated-measures
analysis of variance) (Fig. 2G).
Clinical Outcomes
Table 4 presents the clinical outcomes, all of
which were prespecified secondary outcomes.
There was no significant difference between the
groups in ventilator-free days at day 28 or length
of stay in the ICU. The 28-day mortality rate in
the entire study cohort was 17 of 61 patients
(28%). As would be expected, the APACHE II
n engl j med 359;20 www.nejm.org november 13, 2008
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Mechanical Ventilation in Acute Lung Injury
B
Respiratory-System Compliance
(ml/cm of water)
A
350
PaO2:FIO2
300
Esophageal pressure
250
200
150
Conventional treatment
P=0.002
100
50
0
Baseline
24 Hr
48 Hr
72 Hr
0.70
20.0
P=0.01
10.0
0.0
Baseline
24 Hr
48 Hr
72 Hr
PEEP
(cm of water)
0.60
Esophageal pressure
0.55
P=0.29
15.0
10.0
Conventional treatment
5.0
0.50
0
Esophageal pressure
20.0
Conventional treatment
0.65
Baseline
24 Hr
48 Hr
0.0
72 Hr
Baseline
24 Hr
48 Hr
P<0.001
72 Hr
F
3.0
2.0
Esophageal pressure
1.0
0.0
−1.0
−2.0
−3.0
−4.0
−5.0
Conventional treatment
Baseline
24 Hr
P<0.001
48 Hr
Transpulmonary End-Inspiratory
Pressure (cm of water)
Transpulmonary End-Expiratory
Pressure (cm of water)
Conventional treatment
30.0
25.0
E
G
Esophageal pressure
40.0
D
0.75
Plateau Pressure (cm of water)
Ratio of Dead Space to Tidal
Volume Ratio
C
50.0
72 Hr
35.0
Esophageal pressure
30.0
25.0
Conventional treatment
20.0
15.0
P=0.003
10.0
5.0
0.0
Baseline
24 Hr
48 Hr
72 Hr
12.0
Esophageal pressure
10.0
8.0
6.0
4.0
Conventional treatment
P=0.13
2.0
0.0
Baseline
24 Hr
48 Hr
72 Hr
1st Esophageal-Pressure–Guided Groups.
Figure 2. Respiratory Measurements at Baseline
andAUTHOR:
at 24, 48,Talmor
and 72 Hours in theRETAKE
Control and
ICM
2nd
2 of 2by repeated-measures analysis
Means and standard errors are shown. P values
calculated
of variance. Panel A shows the ratio of
REG were
F FIGURE:
3rd
the partial pressure of arterial oxygen to the fraction
of
inspired
oxygen
(PaO
:F
i
O
),
Panel
B
respiratory-system compliance, Panel C the
2
2
CASE
Revised
Line
4-C
ratio of dead space to tidal volume, Panel D positive
end-expiratory pressure
(PEEP),
Panel
EMail
SIZEE transpulmonary end-expiratory pressure,
ARTIST:
ts
H/T
H/T
Panel F plateau pressure, and Panel G transpulmonary
end-inspiratory
pressure.
36p6
Enon
Combo
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
patients survivors
(153.2±53.7 and
Please check carefully.
score at admission was higher among
143.8±58.0 mm Hg,
who died than among those who survived (31.5±4.5 respectively; P = 0.56).
35915
ISSUE:
JOB: baseline
vs. 24.7±6.1, P<0.001). However, the
The mortality rate
at 10-09-08
28 days was lower among
PaO2:FiO2 was similar among survivors and non- patients in the esophageal-pressure–guided group
n engl j med 359;20 www.nejm.org november 13, 2008
2101
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
Table 3. Changes in PEEP at the Initiation of Ventilation According to the Protocol.*
Treatment Group
Change in PEEP
−1 to −6 cm
of Water
Esophageal-pressure–guided group
Control group
0 to 5 cm
of Water
6 to 10 cm
of Water
11 to 15 cm
of Water
16 to 20 cm
of Water
3
9
no. of patients
12
4
2
12
18
1
0
0
*PEEP denotes applied positive end-expiratory pressure.
than among control patients, although the difference was not significant (relative risk, 0.43;
95% CI, 0.17 to 1.07; P = 0.06). Multivariable
analysis showed that after adjustment for baseline APACHE II score (relative risk per point of
score, 1.16; 95% CI, 1.09 to 1.23; P<0.001), the
esophageal-pressure protocol was associated
with a significant reduction in 28-day mortality
as compared with conventional treatment (relative risk, 0.46; 95% CI, 0.19 to 1.0; P = 0.049).
The mortality rate at 180 days did not differ
significantly between the treatment groups; the
point estimate for the relative risk of death in
the esophageal-pressure–guided group was 0.59
(95% CI, 0.29 to 1.20) as compared with the control group. However, a Kaplan–Meier survival plot
(see Fig. 2 in the Supplementary Appendix) shows
separation between the curves that persists at
180 days. Cox regression modeling showed that
after adjustment for baseline APACHE II score
(hazard ratio per point, 1.12; 95% CI, 1.04 to
1.22), the hazard ratio for 180-day mortality was
0.52 in the esophageal-pressure–guided group
(95% CI, 0.22 to 1.25) as compared with the
control group.
Discussion
We found that it is feasible to make repeated
measurements of esophageal pressure that are of
adequate fidelity and quality to be used to manage
the treatment of patients requiring mechanical
ventilation. Patients with acute lung injury or ARDS
treated in this way had significantly improved
oxygenation, as measured by the PaO2:FiO2, and
significantly improved respiratory-system compliance. Moreover, these improvements were achieved
without elevating transpulmonary pressure at end
inspiration above the physiologic range. Finally,
these improvements in lung function were associated with a trend toward improved 28-day survival in this group of very sick patients.
2102
Numerous animal models of acute lung injury
have shown that reducing end-expiratory lung
volume or pressure can be injurious, even when
tidal volume or peak pressure is controlled.20-24
In these models, increasing PEEP can be protec­
tive.25,26 However, in patients with ARDS, effective
adjustment of PEEP to the physiological features
of the individual patient has been difficult to
achieve. For example, in the ARDSNet study of low
tidal volume, PEEP and FiO2 were adjusted according to arterial oxygenation without reference
to chest-wall or lung mechanics.12 The subsequent
Assessment of Low Tidal Volume and Elevated
End-Expiratory Volume to Obviate Lung Injury
(known as the ALVEOLI trial) (ClinicalTrials.gov
number, NCT00000579) compared an increased
level of PEEP with standard PEEP, with both levels
adjusted according to the patient’s oxygenation,
and showed no benefit.1 The recent Lung Open
Ventilation Study (NCT00182195) used a similar
approach to adjustment of PEEP, without benefit.2 The Expiratory Pressure Study Group trial
(NCT00188058) increased PEEP in the intervention
group to reach a plateau pressure of 28 to 30 cm
of water. This study showed improvements in ventilator-free and organ-failure-free days, oxygenation,
and respiratory-system compliance but showed no
significant change in sur­vival.3 Other studies, including those using the lower point of maximum
curvature on the pressure–volume curve or the
stress index, have had mixed results.27-31
The disappointing results of these previous
studies may be due in part to their inclusion of
patients with elevated pleural or intraabdominal
pressure32,33 and elevated esophageal pressure.7
The lungs of such patients may be effectively compressed by high pleural pressures, and their alveoli may collapse at end expiration, despite levels
of PEEP that would be adequate in other patients.
By using esophageal-pressure measurements to
determine PEEP, we may have prevented repeated
alveolar collapse or overdistention.5 In the pres-
n engl j med 359;20 www.nejm.org november 13, 2008
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Mechanical Ventilation in Acute Lung Injury
Table 4. Clinical Outcomes.*
Esophageal-Pressure–Guided
(N = 30)
Conventional Treatment
(N = 31)
P Value
28-Day mortality — no. (%)
5 (17)
12 (39)
0.055
180-Day mortality — no. (%)
8 (27)
14 (45)
0.13
15.5
13.0
10.8–28.5
7.0–22.0
Outcome
Length of ICU stay — days
0.16
Median
Interquartile range
No. of ICU-free days at 28 days
Median
Interquartile range
0.96
5.0
4.0
0.0–14.0
0.0–16.0
11.5
7.0
0.0–20.3
0.0–17.0
12.0
16.0
7.0–27.5
7.0-20.0
No. of ventilator-free days at 28 days
Median
Interquartile range
0.50
No. of days of ventilation among survivors
Median
Interquartile range
0.71
*For patients who were deceased at day 28, a value of 0 days was assigned. ICU denotes intensive care unit.
ent pilot study, PEEP was lowered in 3 of the 30
patients treated with the use of esophageal pressure to determine PEEP and in 12 of the 31 patients treated according to the ARDSNet protocol. More importantly, PEEP was increased by
more than 5 cm of water in 18 patients treated
with the use of esophageal pressure to determine PEEP and in only 1 patient treated according to the ARDSnet protocol (Table 3). Thus, the
key difference between the two approaches appears to be that measurement of esophageal pressure identifies patients who derive benefit from
higher levels of PEEP than would ordinarily be
used. Although no adverse events resulting from
this strategy were observed, we would have been
able to observe only adverse events that occurred
at very high frequency, because of the small size
of the trial.
There is currently mistrust of the use of esophageal-pressure measurements in supine, critically
ill patients, largely because of possible artifacts
associated with body position and lung pathologic conditions.34 Although transpulmonary
pressure–volume curves have been used to characterize lung disease, esophageal-pressure measurements are not usually used to manage mechanical ventilation in patients with acute lung
injury or ARDS.35 However, artifacts in esophageal pressure may not be large enough to obscure
differences in esophageal and pleural pressures
among patients with acute lung injury or ARDS.
For example, the average difference in esophageal
pressure measured in the upright and the supine
position that was imposed by cardiac weight was
2.9±2.1 cm of water,10 and mechanical abnormalities in diseased lungs may reduce tidal excursions in esophageal pressure by a few centimeters of water.34 By contrast, in patients with
acute respiratory failure, end-expiratory esophageal pressures ranged from 4 to 32 cm of water.7
Our study has several limitations. It was a
single-center study with physiologically expert
staff and a small sample, and although we enrolled medical and surgical patients with various
diseases, the findings cannot be generalized until
they are confirmed in a larger trial powered to
detect changes in appropriate clinical end points.
Since our primary end point was oxygenation,
which is known to be improved by applied PEEP,36
and improvements in patient oxygenation have
been associated with unchanged or increased
mortality when these were obtained at the cost
of higher airway pressures, one cannot be sure
of a favorable outcome until a larger trial has
been completed.2,3,12
In conclusion, adjustment of the settings of
mechanical ventilation for patients with acute
lung injury or ARDS on the basis of the patients’
estimated transpulmonary pressure may have
clinical benefit. This approach shows promise for
improvement in lung function and survival that
warrants further investigation.
n engl j med 359;20 www.nejm.org november 13, 2008
2103
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Mechanical Ventilation in Acute Lung Injury
Supported in part by a grant from the National Heart, Lung,
and Blood Institute (HL-52586).
Dr. Malhotra reports receiving consulting fees from Respironics. Mr. Ritz reports receiving consulting fees from INO
Therapeutics and lecture fees from Tri-anim and serving on
the advisory boards of Cardinal Health and Respironics. No
other potential conflict of interest relevant to this article was
reported.
References
1. Brower RG, Lanken PN, MacIntyre N,
et al. Higher versus lower positive endexpiratory pressures in patients with the
acute respiratory distress syndrome.
N Engl J Med 2004;351:327-36.
2. Meade MO, Cook DJ, Guyatt GH, et al.
Ventilation strategy using low tidal volumes, recruitment maneuvers, and high
positive end-expiratory pressure for acute
lung injury and acute respiratory distress
syndrome: a randomized controlled trial.
JAMA 2008;299:637-45.
3. Mercat A, Richard JC, Vielle B, et al.
Positive end-expiratory pressure setting in
adults with acute lung injury and acute respiratory distress syndrome: a randomized
controlled trial. JAMA 2008;299:646-55.
4. Grasso S, Fanelli V, Cafarelli A, et al.
Effects of high versus low positive endexpiratory pressures in acute respiratory
distress syndrome. Am J Respir Crit Care
Med 2005;171:1002-8.
5. Slutsky AS. Lung injury caused by mechanical ventilation. Chest 1999;116:Suppl:
9S-15S.
6. Malbrain ML, Chiumello D, Pelosi P,
et al. Incidence and prognosis of intra­
abdominal hypertension in a mixed population of critically ill patients: a multiplecenter epidemiological study. Crit Care Med
2005;33:315-22.
7. Talmor D, Sarge T, O’Donnell CR, et al.
Esophageal and transpulmonary pressures
in acute respiratory failure. Crit Care Med
2006;34:1389-94.
8. Milic-Emili J, Mead J, Turner JM,
Glauser EM. Improved technique for estimating pleural pressure from esophageal
baloons. J Appl Physiol 1964;19:207-11.
9. Pelosi P, Goldner M, McKibben A, et al.
Recruitment and derecruitment during
acute respiratory failure: an experimental
study. Am J Respir Crit Care Med 2001;
164:122-30.
10. Washko GR, O’Donnell CR, Loring
SH. Volume-related and volume-independent effects of posture on esophageal and
transpulmonary pressures in healthy subjects. J Appl Physiol 2006;100:753-8.
11. Beyer J, Beckenlechner P, Messmer K.
The influence of PEEP ventilation on organ
blood flow and peripheral oxygen delivery.
Intensive Care Med 1982;8:75-80.
12. The Acute Respiratory Distress Syndrome Network. Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute lung injury and
the acute respiratory distress syndrome.
N Engl J Med 2000;342:1301-8.
13. Bernard GR, Artigas A, Brigham KL,
et al. The American-European Consensus
2104
Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical
trial coordination. Am J Respir Crit Care
Med 1994;149:818-24.
14. Nishida T, Suchodolski K, Schettino
GP, Sedeek K, Takeuch M, Kacmarek RM.
Peak volume history and peak pressurevolume curve pressures independently
affect the shape of the pressure-volume
curve of the respiratory system. Crit Care
Med 2004;32:1358-64.
15. Colebatch HJ, Greaves IA, Ng CK. Exponential analysis of elastic recoil and aging in healthy males and females. J Appl
Physiol 1979;47:683-91.
16. Shapiro NI, Howell MD, Talmor D,
et al. Implementation and outcomes of the
Multiple Urgent Sepsis Therapies (MUST)
protocol. Crit Care Med 2006;34:1025-32.
17. Resar R, Pronovost P, Haraden C,
Simmonds T, Rainey T, Nolan T. Using
a bundle approach to improve ventilator
care processes and reduce ventilator-associated pneumonia. Jt Comm J Qual Patient
Saf 2005;31:243-8.
18. McNutt LA, Wu C, Xue X, Hafner JP.
Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003;157:940-3.
19. Zou G. A modified Poisson regression
approach to prospective studies with binary data. Am J Epidemiol 2004;159:702-6.
20. Chiumello D, Pristine G, Slutsky AS.
Mechanical ventilation affects local and
systemic cytokines in an animal model of
acute respiratory distress syndrome. Am J
Respir Crit Care Med 1999;160:109-16.
21. Faridy EE, Permutt S, Riley RL. Effect of
ventilation on surface forces in excised dogs’
lungs. J Appl Physiol 1966;21:1453-62.
22. Muscedere JG, Mullen JB, Gan K,
Slutsky AS. Tidal ventilation at low airway
pressures can augment lung injury. Am J
Respir Crit Care Med 1994;149:1327-34.
23. Webb HH, Tierney DF. Experimental
pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures: protection by positive endexpiratory pressure. Am Rev Respir Dis
1974;110:556-65.
24. Wyszogrodski I, Kyei-Aboagye K,
Taeusch HW Jr, Avery ME. Surfactant inactivation by hyperventilation: conservation by end-expiratory pressure. J Appl
Physiol 1975;38:461-6.
25. Tremblay L, Valenza F, Ribeiro SP, Li J,
Slutsky AS. Injurious ventilatory strategies
increase cytokines and c-fos m-RNA expression in an isolated rat lung model.
J Clin Invest 1997;99:944-52.
26. Imai Y, Parodo J, Kajikawa O, et al.
Injurious mechanical ventilation and endorgan epithelial cell apoptosis and organ
dysfunction in an experimental model of
acute respiratory distress syndrome. JAMA
2003;289:2104-12.
27. Villar J, Kacmarek RM, Pérez-Méndez
L, Aguirre-Jaime A. A high positive endexpiratory pressure, low tidal volume ventilatory strategy improves outcome in persistent acute respiratory distress syndrome:
a randomized, controlled trial. Crit Care
Med 2006;34:1311-8.
28. Maggiore SM, Jonson B, Richard JC,
Jaber S, Lemaire F, Brochard L. Alveolar
derecruitment at decremental positive endexpiratory pressure levels in acute lung injury: comparison with the lower inflection
point, oxygenation, and compliance. Am J
Respir Crit Care Med 2001;164:795-801.
29. Jonson B, Richard JC, Straus C, Mancebo J, Lemaire F, Brochard L. Pressurevolume curves and compliance in acute
lung injury: evidence of recruitment above
the lower inflection point. Am J Respir
Crit Care Med 1999;159:1172-8.
30. Amato MB, Barbas CS, Medeiros DM,
et al. Effect of a protective-ventilation
strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;
338:347-54.
31. Harris RS, Hess DR, Venegas JG. An
objective analysis of the pressure-volume
curve in the acute respiratory distress syndrome. Am J Respir Crit Care Med 2000;
161:432-9.
32. Malbrain ML. Abdominal pressure in
the critically ill: measurement and clinical relevance. Intensive Care Med 1999;25:
1453-8.
33. Chieveley-Williams S, Dinner L, Puddi­
combe A, Field D, Lovell AT, Goldstone JC.
Central venous and bladder pressure reflect
transdiaphragmatic pressure during pressure support ventilation. Chest 2002;121:
533-8.
34. de Chazal I, Hubmayr RD. Novel aspects of pulmonary mechanics in intensive care. Br J Anaesth 2003;91:81-91.
35. Gattinoni L, Carlesso E, Cadringher P,
Valenza F, Vagginelli F, Chiumello D.
Physical and biological triggers of ventilator-induced lung injury and its prevention. Eur Respir J Suppl 2003;47:15s-25s.
36. Villar J, Pérez-Méndez L, López J, et
al. An early PEEP/FIO2 trial identifies different degrees of lung injury in patients
with acute respiratory distress syndrome.
Am J Respir Crit Care Med 2007;176:795804.
Copyright © 2008 Massachusetts Medical Society.
n engl j med 359;20 www.nejm.org november 13, 2008
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on October 1, 2013. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Effects of Prone Positioning on Lung Protection in
Patients with Acute Respiratory Distress Syndrome
Rodrigo A. Cornejo1, Juan C. Dı́az2, Eduardo A. Tobar1, Alejandro R. Bruhn3, Cristobal A. Ramos2,
Roberto A. González1, Claudia A. Repetto1, Carlos M. Romero1, Luis R. Gálvez1, Osvaldo Llanos1,
Daniel H. Arellano1, Wilson R. Neira1, Gonzalo A. Dı́az1, Anı́bal J. Zamorano1, and Gonzalo L. Pereira2
1
Unidad de Pacientes Crı́ticos, Departamento de Medicina, Hospital Clı́nico Universidad de Chile; 2Departamento de Radiologı́a, Hospital Clı́nico
Universidad de Chile, Santiago, Chile; and 3Departamento de Medicina Intensiva, Facultad de Medicina, Pontificia Universidad Católica de Chile,
Santiago, Chile
Rationale: Positive end-expiratory pressure (PEEP) and prone positioning may induce lung recruitment and affect alveolar dynamics in
acute respiratory distress syndrome (ARDS). Whether there is interdependence between the effects of PEEP and prone positioning on
these variables is unknown.
Objectives: To determine the effects of high PEEP and prone positioning on lung recruitment, cyclic recruitment/derecruitment,
and tidal hyperinflation and how these effects are influenced by
lung recruitability.
Methods: Mechanically ventilated patients (VT 6 ml/kg ideal body
weight) underwent whole-lung computed tomography (CT) during
breath-holding sessions at airway pressures of 5, 15, and 45 cm H2O
and Cine-CTs on a fixed thoracic transverse slice at PEEP 5 and 15 cm
H2O. CT images were repeated in supine and prone positioning. A
recruitment maneuver at 45 cm H2O was performed before each
PEEP change. Lung recruitability was defined as the difference in
percentage of nonaerated tissue between 5 and 45 cm H2O. Cyclic
recruitment/derecruitment and tidal hyperinflation were determined as tidal changes in percentage of nonaerated and hyperinflated tissue, respectively
Measurements and Main Results: Twenty-four patients with ARDS
were included. Increasing PEEP from 5 to 15 cm H2O decreased nonaerated tissue (501 6 201 to 322 6 132 grams; P , 0.001) and increased tidal-hyperinflation (0.41 6 0.26 to 0.57 6 0.30%; P ¼ 0.004)
in supine. Prone positioning further decreased nonaerated tissue (322
6 132 to 290 6 141 grams; P ¼ 0.028) and reduced tidal hyperinflation
observed at PEEP 15 in supine patients (0.57 6 0.30 to 0.41 6 0.22%).
Cyclic recruitment/derecruitment only decreased when high PEEP and
prone positioning were applied together (4.1 6 1.9 to 2.9 6 0.9%; P ¼
0.003), particularly in patients with high lung recruitability.
Conclusions: Prone positioning enhances lung recruitment and
decreases alveolar instability and hyperinflation observed at high
PEEP in patients with ARDS.
Ventilator-induced lung injury (VILI) seems to play an important role in patients with acute respiratory distress syndrome
(Received in original form July 23, 2012; accepted in final form January 12, 2013)
This work was supported by FONDECYT grant 11070156, Chile.
Author Contributions: Substantial contributions to conception and design: R.A.C.
Acquisition of data: R.A.C., J.C.D., C.M.R., R.A.G., C. A. Repetto, D.H.A., W.R.N.,
G.A.D., A.J.Z., G.L.P. Analysis and interpretation of data: R.A.C., E.A.T., A.R.B.,
C.M.R., L.R.G., O.L. Drafting the article or revising it critically for important
intellectual content: R.A.C., J.C.D., E.A.T., A.R.B., C.M.R., R.A.G., C. A. Repetto,
C. A. Ramos, L.R.G., O.L., D.H.A., W.R.N., G.A.D., A.J.Z., G.L.P. Final approval of
the version to be published: R.A.C., J.C.D., E.A.T., A.R.B., C.M.R., R.A.G., C. A.
Repetto, C. A. Ramos, L.R.G., O.L., D.H.A., W.R.N., G.A.D., A.J.Z., G.L.P.
Correspondence and requests for reprints should be addressed to Rodrigo
Cornejo, M.D., Universidad de Chile, Santos Dumont 999, Independencia, Santiago, Chile. E-mail: [email protected]
This article has an online supplement, which is accessible from this issue’s table of
contents at www.atsjournals.org
Am J Respir Crit Care Med Vol 188, Iss. 4, pp 440–448, Aug 15, 2013
Copyright ª 2013 by the American Thoracic Society
Originally Published in Press as DOI: 10.1164/rccm.201207-1279OC on January 24, 2013
Internet address: www.atsjournals.org
AT A GLANCE COMMENTARY
Scientific Knowledge on the Subject
Experimental and clinical studies suggest that high levels of
positive end-expiratory pressure (PEEP) and prone positioning may favor protective mechanical ventilation in
patients with acute respiratory distress syndrome. High
PEEP may induce lung recruitment and decrease cyclic
recruitment/derecruitment; however, increasing PEEP may
increase hyperinflation. Prone positioning could have synergistic effects with high PEEP by providing a more uniform
recruitment and better distribution of lung stress.
What This Study Adds to the Field
In ventilated patients with acute respiratory distress syndrome, prone positioning enhances the effects of high PEEP
in terms of lung recruitment and reduction of cyclic
recruitment/derecruitment and prevents the negative impact of PEEP on tidal hyperinflation.
(ARDS) (1). The mechanisms by which mechanical ventilation
exerts its detrimental effect are not completely understood, but
it appears that hyperinflation of lung units and shear forces
generated during cyclic recruitment/derecruitment of unstable
alveoli exacerbate, or even initiate, lung injury (1).
High levels of positive end-expiratory pressure (PEEP) and
prone positioning have been demonstrated to reduce VILI in experimental models of acute lung injury (ALI) (2–5). However,
analysis of several large clinical trials in patients with ALI or
ARDS suggests that these interventions may be effective only
in patients with severe ARDS (6–9). Patients who seem to benefit
from prone positioning are frequently subjected to higher levels
of PEEP. Thus, there may be a potential interaction between the
effects of both interventions on the mechanisms of VILI.
Regarding VILI, PEEP may have a protective effect by favoring lung recruitment and by reducing cyclic recruitment/
derecruitment (1, 6), but other mechanisms (e.g., redistribution
of extravascular lung water, redistribution of pulmonary blood
flow to better aerated units, or preservation of surfactant activity) may be involved. However, PEEP may induce hyperinflation and increase the risk of VILI, especially in patients
with low recruitability or lobar ARDS (10–12). In fact, some
patients exhibit tidal hyperinflation despite using low VT and
moderate PEEP levels according to the ARDS-Net strategy
(13, 14).
Prone positioning may influence mechanisms of VILI. By
recruiting nonaerated tissue and by reducing the vertical pleural
pressure gradient, prone positioning may provide a more uniform
distribution of transpulmonary pressures during mechanical
Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation
ventilation (15–21). Therefore, prone positioning may act synergistically with high PEEP to protect the lungs from VILI by enhancing lung recruitment and decreasing the risk of PEEP-induced
hyperinflation.
The objectives of this study were to determine the combined
effects of high PEEP and prone positioning on lung recruitment,
cyclic recruitment/derecruitment, and tidal hyperinflation, as
assessed by static and dynamic computed tomography (CT) in
patients with ARDS. In addition, we sought to determine whether
these responses are influenced by lung recruitability (22). Some of
the results of this study have been previously reported in the form
of abstracts (23, 24).
METHODS
Study Population
Twenty-four patients were studied in a university hospital. Institutional and governmental ethical committees granted their approval.
Informed consent was obtained from the patients’ next of kin. Adult
patients fulfilling ARDS criteria (25) and on mechanical ventilation
for 24 to 72 hours who required lung CT scan for clinical purposes
were enrolled. Patients younger than 18 years of age, who were pregnant, or for whom prone positioning was contraindicated (26) were
excluded.
Study Protocol
Patients were evaluated in the ICU and CT room (Figure 1). During the
protocol they were kept under deep sedation and neuromuscular paralysis and ventilated in volume-controlled mode with VT of 6 ml/kg
ideal body weight.
ICU Assessments
Patients were ventilated with PEEP 5 and 15 cm H2O, for 20 minutes
each, in supine and prone positions (PEEP changes and positioning
were applied in random order). Respiratory mechanics, oxygenation,
and hemodynamic parameters were assessed at the end of each setting.
441
Before PEEP changes, a recruitment maneuver at 45 cm H2O airway
pressure was performed to standardize volume history. Quasistatic
compliance of the respiratory system (“compliance”) was calculated
by dividing VT by the difference between plateau pressure and total
PEEP.
CT Assessments
Patients underwent whole-lung CT during breath-holding sessions at
three different airway pressures: 45 cm H2O end-inspiratory airway
pressure and at 5 and 15 cm H2O PEEP. CT scanning (Somaton Sensation, Siemens, Germany) was performed under the following protocol: voltage 120 kVp, current 200 mA, mAs 100, rotation time
0.5 seconds, matrix 512 3 512. Lung compartments were defined according
to their CT density in hyperinflated (2901 to 21,000 Hounsfield units
[HU]), well aerated (2501 to 2900 HU), poorly aerated (2101 to –500
HU), and nonaerated tissue (2100 to 1100 HU) (22). Lung weight for
each compartment was calculated as 1 2 (mean CT number/21,000) 3
volume, where CT number represents lung density (HU). Percentage
of potentially recruitable lung was defined as (nonaerated tissue at
5 cm H2O 2 nonaerated tissue at 45 cm H2O)/total weight and high lung
recruitability as a percentage of potentially recruitable lung greater
than 13.9% in supine, which corresponds to the median value observed
in 49 patients with ARDS in a previous study of lung recruitability (22).
Contiguous axial sections (5 mm thick) were reconstructed from the
volumetric data using a high-definition filter.
A 2.4-mm-thick supradiaphragmatic CT slice was selected for dynamic CT. Respiratory rate was transiently decreased to 10 breaths
per minutes during dynamic CT capture. Cine-CTs of 12 seconds were
performed under the following protocol: voltage 100 kVp, current
80 mA, mAs 40; rotation time 0.5 seconds, 24 images; matrix 512 3
512. Lung compartments were expressed as percentage of tissue weight
in the transverse slice. Cyclic recruitment/derecruitment and tidal hyperinflation were determined as tidal changes in percentage of nonaerated and hyperinflated tissue, respectively.
CT images were repeated in supine and prone positioning, and the
sequence of positions and PEEP levels was applied in random order.
Images were analyzed manually by radiologists using Pulmo (Siemens,
Germany) and MALUNA (University of Gottingen, Germany) software. Further details are provided in the online supplement.
Figure 1. Study protocol. The protocol was performed in the ICU and CT rooms. In the ICU, patients were ventilated with positive end-expiratory
pressure (PEEP) 5 and then 15 cm H2O for 20 minutes each, starting with a recruitment maneuver at 45 cm H2O before each period. Respiratory
mechanics, oxygenation, and hemodynamic parameters were assessed at the end of each setting. Thereafter, in the CT room, patients underwent
whole-lung CT during breath-holding sessions at consecutive airway pressures of 5, 45, and 15 cm H2O; afterward, Cine-CTs were performed on
a fixed thoracic transverse slice at PEEP 5 and 15 cm H2O. CT images were repeated in supine and prone positioning, and the sequence of positions
and PEEP levels was applied in random order.
442
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
VOL 188
2013
TABLE 1. BASELINE CHARACTERISTICS OF THE STUDY POPULATION
Patient no.
Age (yr)
Gender
APACHE
SOFA
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Mean 6 SD
25
67
47
78
66
80
64
83
30
66
41
63
39
36
52
45
53
74
17
47
65
64
51
37
53 6 18
F
F
F
F
M
M
F
M
M
M
F
M
M
M
F
M
M
M
F
M
M
M
F
M
18
14
22
26
29
27
12
20
31
23
18
21
16
6
18
10
17
22
22
19
19
24
19
15
20 6 6
5
7
10
12
17
13
9
8
15
8
8
8
8
5
9
9
12
11
12
10
10
11
8
8
10 6 3
Lung
Infiltrates*
Diffuse
Patchy
Diffuse
Diffuse
Patchy
Lobar
Lobar
Patchy
Diffuse
Diffuse
Patchy
Patchy
Patchy
Patchy
Diffuse
Patchy
Patchy
Patchy
Patchy
Lobar
Lobar
Patchy
Patchy
Patchy
MV (h)
62
28
32
64
68
66
67
26
38
40
69
26
33
29
52
36
34
31
24
38
44
67
69
33
45 6 17
Prestudy
position
Etiology
ARDS
PaO2/FIO2
(mm Hg)
PEEP
(cm H2O)
Compliance†
(ml/cm H2O)
SP
SP
SP
SP
SP
SP
SP
SP
PP
PP
SP
SP
PP
PP
PP
PP
PP
SP
SP
SP
SP
PP
SP
SP
SS1Asp. P
SS
SS
HS
SS
BP
SS
BP
Asp. P
SS
SS
BP
H1N1
H1N1
H1N1
BP
H1N11BP
BP
SS
AP
AP
H1N1
SS
H1N1
195
178
154
88
190
170
186
175
101
155
177
183
100
97
78
198
91
185
195
193
105
112
183
86
149 6 44
12
8
10
12
8
8
6
12
14
12
8
8
12
14
13
6
12
7
8
10
13
12
8
11
10 6 3
30
37
39
44
40
43
28
28
32
32
32
35
43
43
29
30
45
38
35
35
29
25
32
43
35 6 6
Definition of abbreviations: AP ¼ acute pancreatitis; APACHE ¼ Acute Physiology and Chronic Health Evaluation II score; ARDS ¼ acute respiratory distress syndrome;
Asp. P ¼ aspirative pneumonia; BP ¼ bacterial pneumonia; HS ¼ hemorrhagic shock; MV ¼ mechanical ventilation; PEEP ¼ end-expiratory pressure; PP ¼ prone position;
SOFA ¼ Sequential Organ Failure Assessment score; SP ¼ supine position; SS ¼ severe sepsis.
* Type of lung infiltrates according to computed tomography.
y
Quasistatic compliance of respiratory system.
Statistical Analysis
Results are expressed as mean (6 SD) or median (interquartile range).
The effect of PEEP level and position was assessed by applying a twoway ANOVA for repeated measurements with Sidak’s post hoc analysis.
Comparisons between patients with low and high lung recruitability were
performed with Student’s t test or Wilcoxon Mann-Whitney test. SPSS
20.0 software (Chicago, IL) was used for statistical calculations. Twosided P , 0.05 was considered statistically significant.
RESULTS
We included 24 patients (15 male; 54 6 18 yr of age) with
ARDS. Fourteen patients presented with patchy, six with diffuse, and four with lobar CT attenuations. Baseline characteristics are presented in Table 1.
Prone positioning had no significant effect on gas exchange or
compliance when compared with supine position at PEEP 5 (Table 2). However, it induced a decrease in nonaerated tissue and
an increase in well aerated tissue (Table 3).
Increasing PEEP from 5 to 15 cm H2O increased oxygenation
and compliance (Table 2), decreased nonaerated tissue, and
increased well aerated tissue (Table 3). However, it markedly
increased hyperinflation. These effects were observed at supine
and prone positioning. Nevertheless, compared with supine, at
prone positioning oxygenation and compliance tended to be
higher at PEEP 15, whereas nonaerated tissue and hyperinflation were lower (Table 3; Figure 2).
The effects of prone positioning and PEEP level on determinants of VILI, namely cyclic recruitment and derecruitment
and tidal hyperinflation, are shown in Figure 3. Compared with
supine position at PEEP 5, neither prone positioning nor increasing PEEP had a significant effect on cyclic recruitment
and derecruitment. However, both strategies applied together
(prone positioning at PEEP 15) significantly decreased cyclic
recruitment and derecruitment (4.1 6 1.9 to 2.9 6 0.9%;
P ¼ 0.003). Tidal hyperinflation was not affected by prone positioning at PEEP 5, but it increased in response to PEEP 15 in
supine positioning (0.41 6 0.26 to 0.57 6 0.30%; P ¼ 0.004).
The PEEP-induced tidal hyperinflation observed in supine positioning was markedly decreased by prone positioning (0.57 6
0.30 to 0.41 6 0.22%; P ¼ 0.01).
Potentially recruitable lung was 18.3 6 11% in supine positioning and 14.9 6 8% in prone positioning (P ¼ 0.036). Figure
4 shows CT images from representative patients obtained at
airway pressures of 5 and 45 cm H2O in supine and prone
positioning. According to our predefined threshold, 14 patients
had high lung recruitability.
Increasing PEEP to 15 cm H2O improved oxygenation and
compliance only in patients with high lung recruitability. This
subgroup exhibited a 48% relative decrease in nonaerated
TABLE 2. EFFECTS OF POSITIVE END-EXPIRATORY PRESSURE
LEVELS AND PRONE POSITION ON RESPIRATORY VARIABLES
Supine
PEEP 5
PaO2:FIO2, mm Hg
PaO2, mm Hg
PaCO2, mm Hg
Compliance,‡ ml/cm H2O
Plateau pressure, cm H2O
Driving pressure,jj cm H2O
143
90
47
34
18
13
6
6
6
6
6
6
58
34
7
6
3
3
Prone
PEEP 15
235
155
46
40
27
12
6
6
6
6
6
6
68*
73*
8
9*
3*
3
PEEP 5
170
114
46
35
18
13
6
6
6
6
6
6
71
68
6
8
2
2
PEEP 15
259
175
46
44
26
11
6
6
6
6
6
6
63†
79†
7
10†x
2†x
2†x
Definition of abbreviations: PEEP 5 ¼ positive end-expiratory pressure at 5 cm
H2O; PEEP 15 ¼ positive end-expiratory pressure at 15 cm H2O.
* P , 0.05 between parameters at supine PEEP 5 and 15 cm H2O.
y
P , 0.05 between parameters at prone PEEP 5 and 15 cm H2O.
z
Compliance of the respiratory system.
x
P , 0.05, between parameters at supine PEEP 15 and prone PEEP 15 cm H2O.
jj
Difference between plateau pressure and positive end-expiratory pressure.
Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation
443
TABLE 3. EFFECTS OF POSITIVE END-EXPIRATORY PRESSURE LEVELS AND PRONE POSITION ON LUNG
COMPARTMENT DISTRIBUTION
Supine
PEEP 5
Total lung volume, ml
Total lung weight, g
Volume of nonaerated compartment, ml
Weight of nonaerated tissue, g
Volume of poorly aerated compartment, ml
Weight of poorly aerated tissue, g
Volume of well aerated compartment, ml
Weight of well aerated tissue, g
Volume of hyperinflated compartment, ml
Weight of hyperinflated tissue, g
2,140
1,201
503
501
623
446
872
254
144
5.6
6
6
6
6
6
6
6
6
6
6
643
307
201
201
210
149
406
103
160
6
Prone
PEEP 15
3,271
1,282
324
322
753
517
1,767
485
434
19
6
6
6
6
6
6
6
6
6
6
900
316
133*
132*
314
216
661*
148*
318*
15*
PEEP 5
2,240
1,216
431
431
657
467
1,027
324
125
4.4
6
6
6
6
6
6
6
6
6
6
614
341
193†
192†
207
157
456†
148†
105
4
PEEP 15
3,268
1,261
291
290
724
497
1,893
535
362
14
6
6
6
6
6
6
6
6
6
6
821
321
143‡x
141‡x
289
203
617‡
152‡
220‡x
8‡
Definition of abbreviations: PEEP 5 ¼ positive end-expiratory pressure at 5 cm H2O; PEEP 15 ¼ positive end-expiratory pressure
at 15 cm H2O.
* P , 0.05 between parameters at supine PEEP 5 and 15 cm H2O.
y
P , 0.05, between parameters at supine PEEP 5 and prone PEEP 5 cm H2O.
z
P , 0.05 between parameters at prone PEEP 5 and 15 cm H2O.
x
P , 0.05 between parameters at supine PEEP 15 and prone PEEP 15 cm H2O.
tissue when increasing PEEP; this effect was seen in only 22%
of patients with low lung recruitability (Table 4). In contrast to
patients with high lung recruitability who showed no additional
effect of prone positioning on nonaerated tissue at PEEP 15,
patients with low lung recruitability exhibited a significant decrease in nonaerated tissue when subjected to prone positioning
(20% additional relative decrease).
Cyclic recruitment/derecruitment was significantly lower at
baseline in patients with low lung recruitability (2.8 6 1.3%
vs. 5.1 6 1.8% in patients with high lung recruitability; P ¼
0.002), and neither increasing PEEP nor prone positioning
had a significant effect on this variable in the former subgroup
(Figure 3). In contrast, cyclic recruitment and derecruitment
decreased in patients with high lung recruitability when increasing PEEP to 15 cm H2O, and a significant additional effect was
obtained after prone positioning at the same PEEP level. Regarding tidal hyperinflation, patients with low and high lung
recruitability had similar values at baseline (0.39 6 0.25% vs.
0.45 6 0.24; P ¼ 0.568). The effect of high PEEP level on increasing tidal hyperinflation was prevented by prone positioning
in both subgroups of lung recruitability.
hyperinflated tissue or plateau pressures. This effect may be
explained in part by the suppression of the compressive force
of the heart on dorsal lung regions caused by prone positioning
(31, 32). Patients with predominantly basal consolidations, such
as patients A and B from Figure 4, were those who experienced
higher recruitment induced by prone positioning (data not
shown), which is in line with the findings of Galiatsu and colleagues (15). These observations may complement the concept
by Rouby and colleagues about lung morphology as a predictor
of the response to increasing airway pressures (33), although
this requires confirmation by further studies.
In patients with low potentially recruitable lung, prone positioning was able to recruit a significant amount of nonaerated
tissue in addition to that already recruited by high PEEP (Table
4). Therefore, the assessment of potentially recruitable lung
between 5 to 45 cm H2O may correctly predict lung recruitability to increasing airway pressures but not the response to prone.
Thus, prone positioning may be considered in patients with
severe ARDS even if they have poor response to recruitment
maneuvers in supine position as assessed by CT.
Effects of PEEP
DISCUSSION
The main finding of the present study is that prone positioning
enhances the effects of high PEEP on lung recruitment and cyclic
recruitment/derecruitment, whereas it prevents the effects of high
PEEP on tidal hyperinflation. In addition, we found that lung
recruitability, assessed by CT at high airway pressures (22), does
not predict lung recruitment induced by prone positioning.
Effects of Prone Positioning
Although it has been reported that oxygenation improves after
prone positioning (7–9), we only found a trend that did not
reach statistical significance. This may be explained by the short
sampling period. Data obtained from different series of prolonged prone positioning in patients with ARDS found that
oxygenation improves several hours after prone positioning
begins (26–30). Another possible explanation is that oxygenation variation and lung recruitment may be dissociated because
“anatomical” lung recruitment may differ from “functional”
recruitment (22).
We observed that prone positioning was an effective recruitment strategy that, in contrast to high PEEP, did not increase
Although increasing PEEP from 5 to 15 cm H2O decreased
nonaerated tissue and increased oxygenation, it had no consistent effect on cyclic recruitment/derecruitment in the overall
population. However, in the subgroup of patients with higher
lung recruitability, cyclic recruitment/derecruitment significantly decreased when increasing PEEP to 15 cm H2O (Figure
3). Similar findings were reported in a recent study in which
cyclic recruitment/derecruitment was assessed indirectly with
static CT images (34).
As expected, in supine positioning high PEEP resulted in
higher plateau pressures, hyperinflated tissue, and tidal hyperinflation. Hyperinflation is a morphologic description of the lung
areas that appear overfilled with gas in CT images, and it has
been associated with VILI (14). The low values of tidal hyperinflation may seem irrelevant (,1%). However, fractional analysis of CT data was based on lung weight instead of volume, so
the magnitude of real hyperinflated tissue may have been systematically underestimated (35, 36). We used weight for fractional analysis in dynamic CT because total volume is changing
throughout the respiratory cycle. If the results were expressed
as percentage of volume, tidal hyperinflation would have been
10 times greater (4–6%). PEEP-induced tidal hyperinflation
444
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
VOL 188
2013
Figure 2. Distribution of lung compartments expressed in weight and volume.
Proportion of total lung weight (A) and
volume (B) of each compartment classified as nonaerated tissue (NAT), poorly
aerated tissue (PAT), well aerated tissue
(WAT), and hyperinflated tissue (HIT),
at end expiration with positive endexpiratory pressure (PEEP) 5 and 15 cm
H2O, assessed in supine (supine 5 and supine 15) and prone positioning (prone 5
and prone 15). Data are presented for the
overall population (left; n ¼ 24), for the
subgroup of patients with high lung
recruitability (center; n ¼ 14), and for
the subgroup of patients with low lung
recruitability (right; n ¼ 10). *P , 0.05
between parameters at supine 5 and supine 15 cm H2O. yP , 0.05 between
parameters at prone 5 and prone 15 cm
H2O. zP , 0.05 between parameters at
supine 5 and prone 5 cm H2O. xP ,
0.05 between parameters at supine 15
and prone 15 cm H2O.
could be an important drawback of high PEEP strategies (14),
which may explain the lack of consistency in the results of recent clinical trials (6).
The fact that high PEEP levels could favor hyperinflation
without a consistent decrease in cyclic recruitment/derecruitment
supports the notion that PEEP should be titrated individually
based not only on oxygen exchange but also on other variables,
such as lung recruitability.
Interaction of Prone Positioning with PEEP
Few studies have addressed the issue of potential interactions
between prone positioning and PEEP (37, 38). In the present
study, we found that prone positioning has a synergistic effect
with high PEEP in increasing respiratory system compliance,
likely because of an increase in well aerated lung tissue. Previous
studies have shown contradictory data about the effects of prone
positioning on respiratory system compliance (39). These discrepancies may be due to differences in chest wall compliance and
lung recruitability of the study groups.
Prone positioning and high PEEP induced lung recruitment,
resulting in the least amount of nonaerated tissue. Prone positioning also reinforced the effect of high PEEP on cyclic recruitment/
derecruitment, especially in patients with higher lung recruitability.
This interaction did not seem to be a simple additive effect because in the whole population neither prone at low PEEP nor
high PEEP on supine decreased cyclic recruitment/derecruitment.
Thus, prone positioning and high PEEP together showed a synergistic effect on cyclic recruitment/derecruitment.
Prone positioning prevented the increase in hyperinflated tissue and in tidal hyperinflation induced by high PEEP levels in
supine positioning. Likewise, the increase in plateau pressures
Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation
445
Figure 3. Effects of positive end-expiratory pressure (PEEP)
and prone positioning on cyclic recruitment/derecruitment (R/D) and tidal hyperinflation (TH). Cyclic R/D and
TH with PEEP 5 and 15 cm H2O, assessed in supine (supine
5 and supine 15) and prone positioning (prone 5 and
prone 15). Data are presented for the overall population
(n ¼ 24) (A), for the subgroup of patients with high lung
recruitability (n ¼ 14) (B), and for the subgroup of patients
with low lung recruitability (n ¼ 10) (C). *P , 0.05 between parameters at supine 5 and prone 15 cm H2O. yP ,
0.05 between parameters at supine 5 and supine 15 cm
H2O. zP , 0.05 between parameters at supine 15 and
prone 15 cm H2O.
induced by high PEEP was lower in prone positioning. These
findings may be related to the effects of prone positioning in decreasing pleural pressure gradients and homogenizing transpulmonary pressures in the lungs of patients with ARDS (15, 16, 18,
20). A regional analysis of CT images, as performed by Grasso
and colleagues (40), would be a valuable complement to the
present study to assess the effects of prone positioning on inhomogeneity and recruitment of individual lung regions.
The results of the present study suggest that a high PEEP
strategy applied in prone positioning, instead of supine positioning, could have more beneficial and less adverse effects in terms
of respiratory mechanics and determinants of VILI. These findings are consistent with the observations of a metaanalysis, which
indicate that patients with severe forms of ARDS, who are usually ventilated with high levels of PEEP, may have a survival benefit when treated in prone positioning (7–9).
Recruitment and Cyclic Recruitment/Derecruitment:
Methodological Issues
CT has been the gold standard to assess lung recruitment, although different definitions have been applied (10, 41). We chose
the original definition of recruitment based on the decrease of
nonaerated tissue expressed in lung weight as we have used in
the past (22, 42) because it may be applied to analyze cyclic
recruitment/derecruitment in dynamic CT. Other authors have
defined recruitment as the reaeration of the nonaerated and
poorly aerated compartment (10). Because we applied a definition
limited to the nonaerated compartment, we acknowledge that
our results for PEEP-induced recruitment may appear as subestimated compared with studies that include the poorly aerated
compartment, as shown in a recent study using transthoracic ultrasound (43).
The threshold of 13.9% used in the present study to classify
patients as having high or low lung recruitability in supine position was predefined arbitrarily based on the median value of the
subgroup of 49 patients with ARDS from Gattinoni’s study. The
original lung recruitability threshold of Gattinoni’s study was
9%, which corresponded to the median value of the whole
ALI/ARDS population (68 patients). By applying a different
threshold, our subgroups of higher and lower lung recruitability
are not comparable to the subgroups defined in the original
study by Gattinoni (22).
There is controversy regarding whether cyclic recruitment/
derecruitment, as assessed by CT, corresponds to intratidal opening and closing of lung units or to flooded alveoli that become
partially inflated during inspiration (44, 45). Whatever the underlying phenomenon (cyclic mechanical deformation or cyclic
recruitment/derecruitment of lung units), a reduction of instability produced by prone positioning at high PEEP, as supported by our study, appears as theoretically positive in terms
of lung protection.
Several approaches to assess cyclic phenomena have been
used (46). A cine-CT analysis of a fixed transverse slice allows
dynamic imaging without mechanical ventilation interruption.
This method has been recently validated by experimental and
clinical studies to determine cyclic recruitment/derecruitment,
tidal hyperinflation, and dynamic lung strain (18, 35, 36, 47, 48).
446
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
VOL 188
2013
Figure 4. Representative chest CT images obtained during
breath-holding sessions. Representative CT slices of the
lungs obtained 2 cm above the diaphragm dome at airway
pressures of 5 cm H2O (left) and 45 cm H2O (right) from
four patients in supine (upper) and prone positioning
(lower). The percentage of potentially recruitable lung
was defined as the proportion of nonaerated tissue in
which aeration is restored when increasing airway pressures from 5 to 45 cm H2O. In patient A, the percentage
of potentially recruitable lung was 24% in supine and 18%
in prone positioning; in patient B, the percentage of potentially recruitable lung was 24% in supine and 15% in
prone; in patient C, the percentage of potentially recruitable lung was 14% in supine and 21% in prone; and in
patient D, the percentage of potentially recruitable lung
was 27% in supine and 24% in prone.
The main limitations of this method are: 1) The more inhomogeneous the lung impairment, the less representative the slice
may be (this handicap is particularly true in patients with lobar
pattern, but only four of our patients had such pattern); 2) It is
impossible for dynamic CT to scan exactly the same anatomical
structure in different settings, although the careful definition of
anatomical landmarks and fractional analysis used for dynamic
CT may avoid artifacts created by the cranio-caudal motion;
and 3) the absolute amount of grams of lung tissue undergoing
recruitment/derecruitment or tidal hyperinflation cannot be
determined. The alternative approach of using static CT images
of the whole lung after end-expiratory and end-inspiratory
breath holds (15, 16, 18–20, 22, 34) also has several limitations,
the most important being the time dependency of the recruitment and derecruitment phenomena (49, 50). Despite the limitations of both methods, a previous study comparing them
showed no major differences (48).
In conclusion, prone positioning induces lung recruitment
even in patients classified as having low potential for lung recruitment. In addition, prone positioning applied together with high
TABLE 4. COMPARISONS BETWEEN PATIENTS WITH HIGH AND LOW LUNG RECRUITABILITY
High Lung Recruitability
PaO2/FIO2 at PEEP 5, mm Hg
PaO2/FIO2 at PEEP 15, mm Hg
PaCO2 at PEEP 5, mm Hg
PaCO2 at PEEP 15, mm Hg
Compliance at 5 cmH2O, ml/cm H2O
Compliance at 15 cmH2O, ml/cm H2O
Total lung weight at 5 cm H2O, ml
NAT at 5 cm H2O, g
NAT at 15 cm H2O, g
PAT at 5 cm H2O, g
PAT at 15 cm H2O, g
WAT at 5 cm H2O, g
WAT at 15 cm H2O, g
HIT at 5 cm H2O, g
HIT at 15 cm H2O, g
Low Lung Recruitability
Supine
Prone
Supine
Prone
128 6 61
230 6 69*
49 6 7
47 6 8
34 6 6
40 6 7*
1,307 6 240
576 (410–797)
301 (230–400)*
454 (360–546)
486 (422–786)
208 (163–347)
468 (388–593)*
3.6 (2.1–5.4)
15 (7.3–18)*
156 6 62
272 6 70*†
47 6 7
46 6 8
33 6 8
45 6 11*†
1,312 6 292
448 (354–646)x
296 (231–329)*†
443 (411–561)
459 (410–681)
255 (191–434)x
534 (458–630)*†jj
2.7 (2.0–4.9)
12 (9.0–18)*
164 6 49
244 6 70
45 6 6
46 6 6
35 6 6
40 6 12
1,052 6 334‡
372 (305–472)‡
290 (250–334)*
387 (281–431)
378 (241–572)
240 (210–351)
461 (377–581)*
5.8 (4.7–6.9)
17 (14–26)*
191 6 80
241 6 49†
45 6 6
46 6 5
37 6 9
43 6 10†
1,083 6 376
325 (181–390)‡
215 (163–330)*†jj
457 (275–561)
407 (263–577)
324 (239–343)x
483 (336–593)*
6.0 (4.8–7.9)‡
15 (10–24)*
Definition of abbreviations: HIT ¼ hyperinflated tissue; NAT ¼ nonaerated tissue; PAT ¼ poorly aerated tissue; PEEP 5 ¼ positive
end-expiratory pressure 5 cm H2O; PEEP 15 ¼ positive end-expiratory pressure 15 cm H2O; WAT ¼ well aerated tissue.
* P , 0.05 between parameters at 5 cm H2O (or PEEP 5) and 15 cm H2O (or PEEP 15).
y
P , 0.05 between parameters at 5 cm H2O (or PEEP 5) in supine positioning and 15 cm H2O (or PEEP 15) in prone
positioning.
z
P , 0.05 comparing patients with high lung recruitability versus low lung recruitability at the same level of PEEP and position.
x
P , 0.05 comparing parameters at 5 cm H2O (or PEEP 5) between supine and prone positioning.
jj
P , 0.05 comparing parameters at 15 cm H2O (or PEEP 15) between supine and prone positioning.
Cornejo, Dı́az, Tobar, et al.: Prone Decreases Instability and Hyperinflation
PEEP levels in patients with ARDS act synergistically to
decrease mechanical determinants of VILI such as cyclic
recruitment/derecruitment and tidal hyperinflation.
Author disclosures are available with the text of this article at www.atsjournals.org.
Acknowledgment: The authors thank Dr. Jerónimo Graf, Dr. Gastón Murias, and
Dr. Guillermo Bugedo for critical comments and suggestions for the manuscript;
the nurses, respiratory therapists, medical staff, and medical technologists from
Hospital Clı́nico Universidad de Chile for support during the execution of the
studies; Dr. Hector Gatica for the statistical advice in the present article; Daniel
Castro, medical biophysics and radiation protection officer, who assessed and
adjusted the level of radiation per CT, allowing that the total dose per patient did
not exceed the equivalent of one coronary CT angiography study.
References
1. Dos Santos CC, Slutsky AS. Invited review: mechanisms of ventilatorinduced lung injury. A perspective. J Appl Physiol 2000;89:1645–1655.
2. Argiras EP, Blakeley CR, Dunnill MS, Otremski S, Sykes MK. High
peep decreases hyaline membrane formation in surfactant deficient
lungs. Br J Anaesth 1987;59:1278–1285.
3. Broccard A, Shapiro RS, Schmitz LL, Adams AB, Nahum A, Marini JJ.
Prone positioning attenuates and redistributes ventilator-induced lung
injury in dogs. Crit Care Med 2000;28:295–303.
4. Dreyfuss D, Soler P, Basset G, Saumon G. High inflation pressure
pulmonary edema: respective effects of high airway pressure, high
tidal volume, and positive end-expiratory pressure. Am Rev Respir
Dis 1988;137:1159–1164.
5. Webb HH, Tierney DF. Experimental pulmonary edema due to
intermittent positive pressure ventilation with high inflation pressures:
protection by positive end-expiratory pressure. Am Rev Respir Dis
1974;110:556–565.
6. Briel M, Meade M, Mercat A, Brower RG, Talmor D, Walter SD,
Slutsky AS, Pullenayegum E, Zhou Q, Cook D, et al. Higher vs
lower positive end-expiratory pressure in patients with acute lung
injury and acute respiratory distress syndrome: systematic review and
meta-analysis. JAMA 2010;303:865–873.
7. Sud S, Friedrich JO, Taccone P, Polli F, Adhikari NK, Latini R, Pesenti
A, Guerin C, Mancebo J, Curley MA, et al. Prone ventilation reduces
mortality in patients with acute respiratory failure and severe
hypoxemia: systematic review and meta-analysis. Intensive Care
Med 2010;36:585–599.
8. Gattinoni L, Carlesso E, Taccone P, Polli F, Guerin C, Mancebo J. Prone
positioning improves survival in severe ARDS: a pathophysiologic
review and individual patient meta-analysis. Minerva Anestesiol
2010;76:448–454.
9. Abroug F, Ouanes-Besbes L, Dachraoui F, Ouanes I, Brochard L. An
updated study-level meta-analysis of randomised controlled trials on
proning in ARDS and acute lung injury. Crit Care 2011;15:R6.
10. Malbouisson LM, Muller JC, Constantin JM, Lu Q, Puybasset L, Rouby
JJ. Computed tomography assessment of positive end-expiratory
pressure-induced alveolar recruitment in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 2001;163:1444–
1450.
11. Nieszkowska A, Lu Q, Vieira S, Elman M, Fetita C, Rouby JJ. Incidence
and regional distribution of lung overinflation during mechanical
ventilation with positive end-expiratory pressure. Crit Care Med 2004;
32:1496–1503.
12. Rouby JJ. Lung overinflation: the hidden face of alveolar recruitment.
Anesthesiology 2003;99:2–4.
13. Grasso S, Terragni P, Mascia L, Fanelli V, Quintel M, Herrmann P,
Hedenstierna G, Slutsky AS, Ranieri VM. Airway pressure-time
curve profile (stress index) detects tidal recruitment/hyperinflation
in experimental acute lung injury. Crit Care Med 2004;32:1018–1027.
14. Terragni PP, Rosboch G, Tealdi A, Corno E, Menaldo E, Davini O,
Gandini G, Herrmann P, Mascia L, Quintel M, et al. Tidal hyperinflation
during low tidal volume ventilation in acute respiratory distress syndrome.
Am J Respir Crit Care Med 2007;175:160–166.
15. Galiatsou E, Kostanti E, Svarna E, Kitsakos A, Koulouras V, Efremidis
SC, Nakos G. Prone position augments recruitment and prevents
alveolar overinflation in acute lung injury. Am J Respir Crit Care
Med 2006;174:187–197.
447
16. Mentzelopoulos SD, Roussos C, Zakynthinos SG. Prone position reduces
lung stress and strain in severe acute respiratory distress syndrome. Eur
Respir J 2005;25:534–544.
17. Mure M, Domino KB, Lindahl SG, Hlastala MP, Altemeier WA, Glenny
RW. Regional ventilation-perfusion distribution is more uniform in
the prone position. J Appl Physiol 2000;88:1076–1083.
18. Perchiazzi G, Rylander C, Vena A, Derosa S, Polieri D, Fiore T, Giuliani
R, Hedenstierna G. Lung regional stress and strain as a function of
posture and ventilatory mode. J Appl Physiol 2011;110:1374–1383.
19. Richter T, Bellani G, Scott Harris R, Vidal Melo MF, Winkler T,
Venegas JG, Musch G. Effect of prone position on regional shunt,
aeration, and perfusion in experimental acute lung injury. Am J
Respir Crit Care Med 2005;172:480–487.
20. Valenza F, Guglielmi M, Maffioletti M, Tedesco C, Maccagni P, Fossali
T, Aletti G, Porro GA, Irace M, Carlesso E, et al. Prone position
delays the progression of ventilator-induced lung injury in rats: does
lung strain distribution play a role? Crit Care Med 2005;33:361–367.
21. Wiener CM, Kirk W, Albert RK. Prone position reverses gravitational
distribution of perfusion in dog lungs with oleic acid-induced injury.
J Appl Physiol 1990;68:1386–1392.
22. Gattinoni L, Caironi P, Cressoni M, Chiumello D, Ranieri VM, Quintel
M, Russo S, Patroniti N, Cornejo R, Bugedo G. Lung recruitment in
patients with the acute respiratory distress syndrome. N Engl J Med
2006;354:1775–1786.
23. Cornejo R, Diaz J, Repetto C, Suarez P, Bozan F, Arellano D, Rouliez
K, Diaz G, Romero C, Mezzano E, et al. Comparison of potentially
recruitable lung (PRL) in supine versus prone position: preliminary
data [abstract]. Intensive Care Med 2009;35:S126.
24. Cornejo R, Diaz J, Repetto C, Suarez P, Bozan F, Neira W, Zamorano
A, Tobar E, Castro D, Monsalve J, et al. Effect of prone position and
peep on cyclic recruitment-derecruitment (r/d), and tidal hyperinflation (th) using cine-ct: preliminary data [abstract]. Intensive Care Med
2009;35:S103.
25. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L,
Lamy M, Legall JR, Morris A, Spragg R. The American-European
consensus conference on ARDS: definitions, mechanisms, relevant
outcomes, and clinical trial coordination. Am J Respir Crit Care Med
1994;149:818–824.
26. Romero CM, Cornejo RA, Galvez LR, Llanos OP, Tobar EA, Berasain
MA, Arellano DH, Larrondo JF, Castro JS. Extended prone position
ventilation in severe acute respiratory distress syndrome: a pilot
feasibility study. J Crit Care 2009;24:81–88.
27. Chan MC, Hsu JY, Liu HH, Lee YL, Pong SC, Chang LY, Kuo BI,
Wu CL. Effects of prone position on inflammatory markers in
patients with ARDS due to community-acquired pneumonia. J Formos Med Assoc 2007;106:708–716.
28. Charron C, Bouferrache K, Caille V, Castro S, Aegerter P, Page B,
Jardin F, Vieillard-Baron A. Routine prone positioning in patients
with severe ARDS: feasibility and impact on prognosis. Intensive
Care Med 2011;37:785–790.
29. Cornejo R, Tobar E, Diaz G, Romero C, Llanos O, Galvez LR,
Zamorano A, Fabrega L, Neira W, Arellano D, et al. Systematic
approach for severe respiratory failure due to novel a (H1N1)
influenza. Minerva Anestesiol 2011;77:510–521.
30. Mancebo J, Fernandez R, Blanch L, Rialp G, Gordo F, Ferrer M,
Rodriguez F, Garro P, Ricart P, Vallverdu I, et al. A multicenter
trial of prolonged prone ventilation in severe acute respiratory
distress syndrome. Am J Respir Crit Care Med 2006;173:1233–1239.
31. Malbouisson LM, Busch CJ, Puybasset L, Lu Q, Cluzel P, Rouby JJ; CT
scan ARDS study group. Role of the heart in the loss of aeration
characterizing lower lobes in acute respiratory distress syndrome.
Am J Respir Crit Care Med 2000;161:2005–2012.
32. Albert RK, Hubmayr RD. The prone position eliminates compression of
the lungs by the heart. Am J Respir Crit Care Med 2000;161:1660–1665.
33. Puybasset L, Cluzel P, Gusman P, Grenier P, Preteux F, Rouby JJ; CT
scan ARDS study group. Regional distribution of gas and tissue in
acute respiratory distress syndrome: I. Consequences for lung morphology.
Intensive Care Med 2000;26:857–869.
34. Caironi P, Cressoni M, Chiumello D, Ranieri M, Quintel M, Russo SG,
Cornejo R, Bugedo G, Carlesso E, Russo R, et al. Lung opening and
closing during ventilation of acute respiratory distress syndrome. Am
J Respir Crit Care Med 2010;181:578–586.
448
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE
35. Bruhn A, Bugedo D, Riquelme F, Varas J, Retamal J, Besa C, Cabrera
C, Bugedo G. Tidal volume is a major determinant of cyclic recruitmentderecruitment in acute respiratory distress syndrome. Minerva Anestesiol
2011;77:418–426.
36. David M, Karmrodt J, Bletz C, David S, Herweling A, Kauczor HU,
Markstaller K. Analysis of atelectasis, ventilated, and hyperinflated
lung during mechanical ventilation by dynamic CT. Chest 2005;128:
3757–3770.
37. Gainnier M, Michelet P, Thirion X, Arnal JM, Sainty JM, Papazian L.
Prone position and positive end-expiratory pressure in acute respiratory distress syndrome. Crit Care Med 2003;31:2719–2726.
38. Walther SM, Johansson MJ, Flatebo T, Nicolaysen A, Nicolaysen G.
Marked differences between prone and supine sheep in effect of
peep on perfusion distribution in zone II lung. J Appl Physiol 2005;
99:909–914.
39. Pelosi P, Brazzi L, Gattinoni L. Prone position in acute respiratory
distress syndrome. Eur Respir J 2002;20:1017–1028.
40. Grasso S, Stripoli T, Sacchi M, Trerotoli P, Staffieri F, Franchini D, De
Monte V, Valentini V, Pugliese P, Crovace A, et al. Inhomogeneity
of lung parenchyma during the open lung strategy: a computed
tomography scan study. Am J Respir Crit Care Med 2009;180:415–
423.
41. Gattinoni L, Pelosi P, Crotti S, Valenza F. Effects of positive endexpiratory pressure on regional distribution of tidal volume and recruitment in adult respiratory distress syndrome. Am J Respir Crit
Care Med 1995;151:1807–1814.
42. Bugedo G, Bruhn A, Hernandez G, Rojas G, Varela C, Tapia JC,
Castillo L. Lung computed tomography during a lung recruitment
43.
44.
45.
46.
47.
48.
49.
50.
VOL 188
2013
maneuver in patients with acute lung injury. Intensive Care Med
2003;29:218–225.
Bouhemad B, Brisson H, Le-Guen M, Arbelot C, Lu Q, Rouby JJ.
Bedside ultrasound assessment of positive end-expiratory pressureinduced lung recruitment. Am J Respir Crit Care Med 2011;183:341–
347.
Hubmayr RD. Perspective on lung injury and recruitment: a skeptical
look at the opening and collapse story. Am J Respir Crit Care Med
2002;165:1647–1653.
Tschumperlin DJ, Oswari J, Margulies AS. Deformation-induced injury
of alveolar epithelial cells: effect of frequency, duration, and amplitude. Am J Respir Crit Care Med 2000;162:357–362.
Ghadiali S, Huang Y. Role of airway recruitment and derecruitment in
lung injury. Crit Rev Biomed Eng 2011;39:297–317.
Doebrich M, Markstaller K, Karmrodt J, Kauczor HU, Eberle B, Weiler
N, Thelen M, Schreiber WG. Analysis of discrete and continuous
distributions of ventilatory time constants from dynamic computed
tomography. Phys Med Biol 2005;50:1659–1673.
Zinserling J, Wrigge H, Neumann P, Muders T, Magnusson A,
Hedenstierna G, Putensen C. Methodologic aspects of attenuation
distributions from static and dynamic thoracic ct techniques in
experimental acute lung injury. Chest 2005;128:2963–2970.
Albert SP, DiRocco J, Allen GB, Bates JH, Lafollette R, Kubiak BD,
Fischer J, Maroney S, Nieman GF. The role of time and pressure on
alveolar recruitment. J Appl Physiol 2009;106:757–765.
Massa CB, Allen GB, Bates JH. Modeling the dynamics of recruitment
and derecruitment in mice with acute lung injury. J Appl Physiol 2008;
105:1813–1821.
new england
journal of medicine
The
established in 1812
june 6, 2013
vol. 368 no. 23
Prone Positioning in Severe Acute Respiratory Distress
Syndrome
Claude Guérin, M.D., Ph.D., Jean Reignier, M.D., Ph.D., Jean-Christophe Richard, M.D., Ph.D., Pascal Beuret, M.D.,
Arnaud Gacouin, M.D., Thierry Boulain, M.D., Emmanuelle Mercier, M.D., Michel Badet, M.D.,
Alain Mercat, M.D., Ph.D., Olivier Baudin, M.D., Marc Clavel, M.D., Delphine Chatellier, M.D., Samir Jaber, M.D., Ph.D.,
Sylvène Rosselli, M.D., Jordi Mancebo, M.D., Ph.D., Michel Sirodot, M.D., Gilles Hilbert, M.D., Ph.D.,
Christian Bengler, M.D., Jack Richecoeur, M.D., Marc Gainnier, M.D., Ph.D., Frédérique Bayle, M.D.,
Gael Bourdin, M.D., Véronique Leray, M.D., Raphaele Girard, M.D., Loredana Baboi, Ph.D., and Louis Ayzac, M.D.,
for the PROSEVA Study Group*
A bs t r ac t
Background
Previous trials involving patients with the acute respiratory distress syndrome (ARDS)
have failed to show a beneficial effect of prone positioning during mechanical ventilatory support on outcomes. We evaluated the effect of early application of prone
positioning on outcomes in patients with severe ARDS.
Methods
In this multicenter, prospective, randomized, controlled trial, we randomly assigned 466 patients with severe ARDS to undergo prone-positioning sessions of at
least 16 hours or to be left in the supine position. Severe ARDS was defined as a
ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen
(Fio2) of less than 150 mm Hg, with an Fio2 of at least 0.6, a positive end-expiratory pressure of at least 5 cm of water, and a tidal volume close to 6 ml per kilogram
of predicted body weight. The primary outcome was the proportion of patients who
died from any cause within 28 days after inclusion.
The authors’ affiliations are listed in the
Appendix. Address reprint requests to Dr.
Guérin at Service de Réanimation Médicale, Hôpital de la Croix-Rousse, 103 Grande
Rue de la Croix-Rousse, 69004 Lyon,
France, or at [email protected].
*The Proning Severe ARDS Patients
(PROSEVA) study investigators are listed
in the Supplementary Appendix, available
at NEJM.org.
This article was published on May 20, 2013,
at NEJM.org.
N Engl J Med 2013;368:2159-68.
DOI: 10.1056/NEJMoa1214103
Copyright © 2013 Massachusetts Medical Society.
Results
A total of 237 patients were assigned to the prone group, and 229 patients were assigned to the supine group. The 28-day mortality was 16.0% in the prone group and
32.8% in the supine group (P<0.001). The hazard ratio for death with prone positioning was 0.39 (95% confidence interval [CI], 0.25 to 0.63). Unadjusted 90-day mortality was 23.6% in the prone group versus 41.0% in the supine group (P<0.001), with a
hazard ratio of 0.44 (95% CI, 0.29 to 0.67). The incidence of complications did not
differ significantly between the groups, except for the incidence of cardiac arrests,
which was higher in the supine group.
Conclusions
In patients with severe ARDS, early application of prolonged prone-positioning sessions significantly decreased 28-day and 90-day mortality. (Funded by the Programme
Hospitalier de Recherche Clinique National 2006 and 2010 of the French Ministry
of Health; PROSEVA ClinicalTrials.gov number, NCT00527813.)
n engl j med 368;23 nejm.org june 6, 2013
2159
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
P
rone positioning has been used for
many years to improve oxygenation in patients who require mechanical ventilatory
support for management of the acute respiratory
distress syndrome (ARDS). Randomized, controlled trials have confirmed that oxygenation is
significantly better when patients are in the prone
position than when they are in the supine position.1,2 Furthermore, several lines of evidence have
shown that prone positioning could prevent ventilator-induced lung injury.3-6 In several previous
trials, these physiological benefits did not translate into better patient outcomes, since no significant improvement was observed in patient survival
with prone positioning.7-10 However, meta-analyses2,11 have suggested that survival is significantly
improved with prone positioning as compared
with supine positioning among patients with severely hypoxemic ARDS at the time of randomization. We conducted a prospective, multicenter,
randomized, controlled trial to explore whether
early application of prone positioning would improve survival among patients with ARDS who,
at the time of enrollment, were receiving mechanical ventilation with a positive end-expiratory pressure (PEEP) of at least 5 cm of water and in whom
the ratio of the partial pressure of arterial oxygen
(Pao2) to the fraction of inspired oxygen (Fio2) was
less than 150 mm Hg.
Me thods
Patients
of
m e dic i n e
Randomization was computer-generated and stratified according to ICU. Patients were randomly
assigned to the prone group or supine group
with the use of a centralized Web-based management system (Clininfo). The protocol, available at
NEJM.org, was approved by the ethics committee
Comité Consultatif de Protection des Personnes
dans la Recherche Biomedicale Sud-Est IV in Lyon,
France, and by the Clinical Investigation Ethics
Committee at Hospital de Sant Pau in Barcelona.
Written informed consent was obtained after the
patients’ next of kin read the informational leaflet. If patients were able to read the leaflet at
some point after inclusion in the study, they were
approached to confirm participation in the trial.
An investigator at each center was responsible for
enrolling patients in the study, following the protocol, and completing the case-report form. Centers were regularly monitored by research fellows.
Data collectors were aware of the study-group assignments, but outcomes assessors were not.
The trial was overseen by a steering committee that met monthly. An independent data and
safety monitoring board, comprising three experts
in the field, was also set up (a list of board members is provided in the Supplementary Appendix).
There was no commercial support. No one who is
not listed as an author contributed to the writing
of this manuscript. All authors vouch for the accuracy of the data and analysis and the fidelity of
the study to the protocol.
Protocol
We included in the study adults who met the following criteria: ARDS, as defined according to the
American–European Consensus Conference criteria12; endotracheal intubation and mechanical
ventilation for ARDS for less than 36 hours; and
severe ARDS (defined as a Pao2:Fio2 ratio of
<150 mm Hg, with an Fio2 of ≥0.6, a PEEP of
≥5 cm of water, and a tidal volume of about 6 ml
per kilogram of predicted body weight; the criteria were confirmed after 12 to 24 hours of mechanical ventilation in the participating intensive
care unit [ICU]). Exclusion criteria are listed in
the Supplementary Appendix, available with the
full text of this article at NEJM.org.
After a patient was determined to be eligible, a
stabilization period of 12 to 24 hours was mandated. Inclusion in the study was confirmed only
at the end of this period (Fig. S1 in the Supplementary Appendix).
Patients assigned to the prone group had to be
turned to the prone position within the first hour
after randomization. They were placed in a completely prone position for at least 16 consecutive
hours. Participating centers were given guidelines
(see the Supplementary Appendix) to ensure standardization of prone placement. Standard ICU
beds were used for all patients. Patients assigned
to the supine group remained in a semirecumbent
position.
Trial Design
Mechanical ventilation13 was delivered in a
Patients were recruited from 26 ICUs in France volume-controlled mode with constant inspiraand 1 in Spain, all of which have used prone po- tory flow, with tidal volume targeted at 6 ml per
sitioning in daily practice for more than 5 years. kilogram of predicted body weight13 and the PEEP
2160
n engl j med 368;23 nejm.org june 6, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Prone Positioning in Severe ARDS
level selected from a PEEP–Fio2 table14 (Table S1
in the Supplementary Appendix). The goal was to
maintain an end-inspiratory plateau pressure of
the respiratory system (PplatRS), measured after
a 1-second period of no air flow, of no more than
30 cm of water and an arterial plasma pH of 7.20
to 7.45. Physiological variables were measured at
predetermined times in both groups. In the supine group, measurements were performed every
6 hours; in the prone group, measurements were
performed just before the patient was turned to
the prone position, after 1 hour of prone positioning, just before the patient was turned back to
the supine position, and 4 hours after the patient
was returned to the supine position. Adjustments
of ventilator settings in specific situations are detailed in the Supplementary Appendix.
The criteria for stopping prone treatment were
any of the following: improvement in oxygenation
(defined as a Pao2:Fio2 ratio of ≥150 mm Hg, with
a PEEP of ≤10 cm of water and an Fio2 of ≤0.6; in
the prone group, these criteria had to be met in the
supine position at least 4 hours after the end of
the last prone session); a decrease in the Pao2:Fio2
ratio of more than 20%, relative to the ratio in the
supine position, before two consecutive prone sessions; or complications occurring during a prone
session and leading to its immediate interruption.
Complications leading to the immediate interruption of prone treatment included nonscheduled
extubation, main-stem bronchus intubation, endotracheal-tube obstruction, hemoptysis, oxygen
saturation of less than 85% on pulse oximetry or a
Pao2 of less than 55 mm Hg for more than 5 minutes when the Fio2 was 1.0, cardiac arrest, a heart
rate of less than 30 beats per minute for more
than 1 minute, a systolic blood pressure of less
than 60 mm Hg for more than 5 minutes, and
any other life-threatening reason for which the
clinician decided to stop the treatment.
After patients in the prone group were turned
to the supine position, the prone session could
be resumed at any time before the planned assessment at 4 hours in the supine position if the
criteria for oxygen saturation level, Pao2, or both
were met. The prone-positioning strategy was applied every day up to day 28, after which it was
used at the clinician’s discretion. Patients in the
supine group could not be crossed over to the
prone group except as a rescue measure in case of
life-threatening hypoxemia when all the following
criteria were met simultaneously: a Pao2:Fio2 ratio
of less than 55 mm Hg, with an Fio2 of 1.0;
maximal PEEP according to the PEEP–Fio2 table;
administration of inhaled nitric oxide at a concentration of 10 ppm; infusion of intravenous
almitrine bismesylate at a dose of 4 μg per kilogram per minute; and performance of respiratory recruitment maneuvers to increase the amount
of aerated lung.
Weaning from mechanical ventilation was conducted in the same way for both groups (see the
Supplementary Appendix). Details regarding the
management of sedation and the use of neuromuscular blocking agents are also provided in
the Supplementary Appendix. The investigators
assessed patients at least every morning until day
28 or discharge from the ICU.
Data Collection
At the time of admission, we recorded data on age,
sex, the setting from which the patient was admitted to the ICU, the context for admission to the
ICU, McCabe score14 (which ranges from A to C,
with A indicating no underlying disease that compromises life expectancy, B an estimated life expectancy with the chronic disease of <5 years, and C an
estimated life expectancy with the chronic disease
of <1 year), ventilator settings, time from intubation to randomization, height, predicted body
weight, and the Simplified Acute Physiology Score
(SAPS) II15 (which ranges from 0 to 164, with higher scores indicating greater severity of symptoms).
We also recorded the number of lung quadrants
involved on chest radiography, results of measurements of arterial blood gases, PplatRS, arterial
blood lactate levels, the cause of ARDS, the Sepsisrelated Organ Failure Assessment (SOFA) score16
(which ranges from 0 to 24, with higher scores indicating more severe organ failure), the lung injury
score (which ranges from 0 to 4, with higher scores
indicating more severe lung injury),17 and the time
at which the first prone session was started.
The following events were recorded daily until day 28: attempts at extubation, administration
of inhaled nitric oxide, infusion of almitrine bismesylate, use of extracorporeal membrane oxygenation (ECMO), infusion of sedatives and neuromuscular blockers, complications, and the SOFA
score. Ventilator settings, PplatRS, static compliance of the respiratory system, and the results of
measurements of arterial blood gases were recorded daily during the first week as indicated
above. Data quality was verified by the research
n engl j med 368;23 nejm.org june 6, 2013
2161
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
51,189 Patients were admitted to 27 ICUs in the
study period, Jan. 1, 2008–July 25, 2011
47,740 Did not have ARDS
3449 Had ARDS
2015 Were not screened
1434 Were screened
858 Were not eligible
576 Were eligible
102 Were excluded
37 Had improved symptoms after 12–24 hr
55 Had organizational problems in study center
10 Were withdrawn by physician
474 Underwent randomization
234 Were assigned to supine group
240 Were assigned to prone group
5 Were excluded
3 Had PaO2:FIO2 >150 mm Hg
1 Was enrolled before 12-hr
stabilization period was over
1 Had guardianship issues
3 Were excluded
2 Were enrolled before 12-hr
stabilization period was over
1 Received NIV >24 hr
466 Were included in the intention-to-treat analysis
229 Were in supine group
237 Were in prone group
466 Were included in the 90-day follow-up
229 Were in supine group
237 Were in prone group
Figure 1. Enrollment, Randomization, and Follow-up of the Study Participants.
ARDS denotes the acute respiratory distress syndrome, ICU intensive care unit, NIV noninvasive ventilation, and
Pao2:Fio2 the ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen.
2162
n engl j med 368;23 nejm.org june 6, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Prone Positioning in Severe ARDS
fellows, and data were stored in a database (Clin­
info) that was specifically developed for the study
with the use of Epi Info software, version 3.4.3.
Table 1. Characteristics of the Participants at Inclusion in the Study.*
Supine Group
(N = 229)
Prone Group
(N = 237)
60±16
58±16
152 (66.4)
166 (70.0)
Emergency room
98 (42.8)
101 (42.6)
Acute care facility
87 (38.0)
86 (36.3)
Home
26 (11.4)
31 (13.1)
ICU
9 (3.9)
11 (4.6)
Other
9 (3.9)
8 (3.4)
A
183 (79.9)
197 (83.1)
B
45 (19.7)
39 (16.5)
C
1 (0.4)
1 (0.4)
Diabetes
39 (17.0)
50 (21.1)
Renal failure
12 (5.2)
10 (4.2)
Hepatic disease
16 (7.0)
15 (6.3)
Coronary artery disease
24 (10.5)
24 (10.1)
Cancer
30 (13.1)
24 (10.1)
COPD
29 (12.7)
23 (9.7)
38 (16.6)
32 (13.5)
Characteristic
Outcome Measures
Age — yr
The primary end point was mortality at day 28.
Secondary end points were mortality at day 90, the
rate of successful extubation, the time to successful
extubation, the length of stay in the ICU, complications, the use of noninvasive ventilation, the tracheotomy rate, the number of days free from organ
dysfunction, and ventilator settings, measurements
of arterial blood gases, and respiratory-system mechanics during the first week after randomization.
Successful extubation was defined as no reintubation or use of noninvasive ventilation in the
48 hours after extubation. In patients who had
undergone a tracheotomy, successful weaning
from the ventilator was defined as the ability to
breathe unassisted through the tracheostomy
cannula for at least 24 hours.
Male sex — no. (%)
Statistical Analysis
The expected 28-day mortality in the supine
group was 60%. We estimated that with a sample
of 456 patients, the study would have 90% power
to detect an absolute reduction of 15 percentage
points (to 45%) with prone positioning, at a onesided type I error rate of 5%.
An interim analysis was planned 28 days after half the patients had been enrolled, and two
analyses were scheduled, each with a type I error
rate set to 2.5% to maintain an overall type I error
rate of 5%. The statistician sent the data from
the interim analysis to the data and safety monitoring board, which had to decide whether to
continue or discontinue the trial. An absolute difference in mortality of 25 percentage points or
more between groups at the time of the interim
analysis was the only criterion for early trial termination. There was no stopping rule for futility.
The analysis was performed on an intentionto-treat basis. Continuous variables were expressed
as means with standard deviations. Data were
compared between groups with the use of the
chi-square test or Fisher’s exact test and analysis
of variance as indicated. Patient survival was
analyzed with the use of the Kaplan–Meier method and compared between groups with the use of
the log-rank test. Cox proportional-hazards regression, with stratification according to center,
Setting from which patient was admitted
to ICU — no. (%)
McCabe score — no. (%)†
Coexisting conditions — no. (%)
Immunodeficiency — no. (%)
SAPS II‡
Sepsis — no./total no. (%)§
SOFA score¶
47±17
45±15
195/229 (85.2)
194/236 (82.2)
10.4±3.4
9.6±3.2
133 (58.1)
148 (62.4)
29±7
28±6
Vasopressors
190/229 (83.0)
172/237 (72.6)
Neuromuscular blockers
186/226 (82.3)
212/233 (91.0)
39/228 (17.1)
27/237 (11.4)
101/225 (44.9)
91/230 (39.6)
ARDS due to pneumonia
Body-mass index‖
Other interventions — no./total no. (%)
Renal-replacement therapy
Glucocorticoids
*Plus–minus values are means ±SD. There were no significant differences between the groups in any of the characteristics listed, with the exception of the
Sepsis-related Organ Failure Assessment (SOFA) score, the use of vasopressors, and the use of neuromuscular blockers. ARDS denotes the acute respiratory distress syndrome, COPD chronic obstructive pulmonary disease, and
ICU intensive care unit. A version of this table with additional information is
available as Table S2 in the Supplementary Appendix.
†A McCabe score of A indicates no underlying disease that compromises life expectancy, B an estimated life expectancy with the chronic disease of less than 5 years,
and C an estimated life expectancy with the chronic disease of less than 1 year.
‡The Simplified Acute Physiology Score (SAPS) II ranges from 0 to 164, with
higher scores indicating greater severity of symptoms.
§ Sepsis was defined according to the American–European Consensus
Conference criteria.
¶SOFA scores range from 0 to 24, with higher scores indicating more severe
organ failure.
‖The body-mass index is the weight in kilograms divided by the square of the
height in meters.
n engl j med 368;23 nejm.org june 6, 2013
2163
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
Table 2. Ventilator Settings, Respiratory-System Mechanics, and Results
of Arterial Blood Gas Measurements at the Time of Inclusion in the Study.*
Supine Group
(N = 229)
Prone Group
(N = 237)
Tidal volume (ml)
381±66
384±63
Tidal volume (ml per kg of PBW)
6.1±0.6
6.1±0.6
Respiratory frequency (breaths per min)
27±5
27±5
PEEP (cm of water)
10±4
10±3
0.79±0.16
0.79±0.16
23±5
24±5
Variable
Fio2
PplatRS (cm of water)
CstRS (ml per cm of water)
35±15
36±23
Pao2 (mm Hg)
80±18
80±19
100±20
100±30
52±32
50±14
7.30±0.10
7.30±0.10
25±5
25±5
Pao2:Fio2 (mm Hg)
Paco2 (mm Hg)
Arterial pH
Plasma bicarbonate (mmol per liter)†
*Plus–minus values are means ±SD. CstRS denotes static compliance of the respiratory system, Fio2 the fraction of inspired oxygen, Paco2 partial pressure
of arterial carbon dioxide, Pao2 partial pressure of arterial oxygen, PBW predicted body weight, PEEP positive end-expiratory pressure, and PplatRS endinspiratory plateau pressure of the respiratory system.
†Data are for 227 participants in the supine group and 236 participants in the
prone group.
was planned to adjust the between-group differences in mortality at day 28 and day 90 for significant baseline covariates. The statistical analysis was performed with the use of SPSS software
(SPSS for Windows, version 17.0). The investigators
had no access to the database until the study was
completed. All reported P values are two-sided,
and have not been adjusted for multiple comparisons. A P value of less than 0.05 was considered
to indicate statistical significance.
R e sult s
Participants
From January 1, 2008, through July 25, 2011, a
total of 3449 patients with ARDS were admitted
to the participating ICUs, and 474 underwent
randomization (Fig. 1). Eight patients were subsequently excluded (Fig. 1), and 466 patients were
included in the analysis: 229 in the supine group
and 237 in the prone group. After the interim
analysis, the data and safety monitoring board
recommended that the trial be continued.
of
m e dic i n e
for the SOFA score and the use of neuromuscular
blockers and vasopressors (Table 1). In more
than half the cases, the main cause of ARDS was
pneumonia (Table 1). Influenza A (H1N1) virus
infection was the main cause of ARDS in 28 patients, with no significant difference between the
groups in the rate (5.7% in the supine group and
6.3% in the prone group, P = 0.85). The mean (±SD)
time from intubation to randomization was 31±26
hours in the supine group and 33±24 hours in
the prone group (P = 0.66). The lung injury score
was 3.3±0.4 in both groups, and the rate of use
of noninvasive ventilation in the 24 hours before
inclusion was similar in the two groups (29.3%
and 30.8% in the supine and prone groups, respectively). Ventilator settings, respiratory-system mechanics, and results of arterial blood-gas
measurements were also similar in the two
groups (Table 2).
Prone Positioning
Patients in the prone group underwent their first
prone-positioning session within 55±55 minutes
after randomization. The average number of sessions was 4±4 per patient, and the mean duration
per session was 17±3 hours. All the patients in
this group underwent at least one prone-positioning session. In the prone group, patients were ventilated in the prone position for 73% of the 22,334
patient-hours spent in the ICU from the start of
the first session to the end of the last session.
Adjunctive Therapies
The rates of the use of rescue therapies in the supine and prone groups were 2.6% versus 0.8% for
ECMO (P = 0.14), 15.7% versus 9.7% for inhaled
nitric oxide (P = 0.05), and 6.6% versus 2.5% for
almitrine bismesylate (P = 0.04). Neuromuscular
blockers were used for 5.6±5.0 days in the supine
group and 5.7±4.7 days in the prone group (P = 0.74),
and intravenous sedation was given for 9.5±6.8
and 10.1±7.2 days in the two groups, respectively
(P = 0.35). The use of antiviral therapy for H1N1
virus infection was similar in the two groups.
Ventilator Settings and Lung Function
during the First Week
The Pao2:Fio2 ratio recorded in the supine position was significantly higher in the prone group
than in the supine group at days 3 and 5, whereCharacteristics at Inclusion
as the PEEP and Fio2 were significantly lower
The characteristics of the patients at inclusion in (Table S3 in the Supplementary Appendix). The
the study were similar in the two groups except PplatRS was 2 cm of water lower by day 3 in the
2164
n engl j med 368;23 nejm.org june 6, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Prone Positioning in Severe ARDS
Table 3. Primary and Secondary Outcomes According to Study Group.*
Supine Group
(N = 229)
Prone Group
(N = 237)
75 (32.8 [26.4–38.6])
38 (16.0 [11.3–20.7])
Outcome
Hazard Ratio
or Odds Ratio
with the Prone
Position (95% CI)
P Value
Mortality — no. (% [95% CI])
At day 28
Not adjusted
Adjusted for SOFA score†
0.39 (0.25–0.63)
<0.001
0.42 (0.26–0.66)
<0.001
0.44 (0.29–0.67)
<0.001
0.48 (0.32–0.72)
<0.001
0.45 (0.29–0.70)
<0.001
At day 90
Not adjusted
94 (41.0 [34.6–47.4])
56 (23.6 [18.2–29.0])
145/223
(65.0 [58.7–71.3])
186/231
(80.5 [75.4–85.6])
Survivors
19±21
17±16
Nonsurvivors
16±11
18±14
Survivors
26±27
24±22
Nonsurvivors
18±15
21±20
Adjusted for SOFA score†
Successful extubation at day 90 —
no./total no. (% [95% CI])
Time to successful extubation,
­assessed at day 90 —
days
0.87
Length of ICU stay, assessed at
day 90 — days
0.05
Ventilation-free days
At day 28
10±10
14±9
<0.001
At day 90
43±38
57±34
<0.001
13 (5.7 [3.9–7.5])
15 (6.3 [4.9–7.7])
0.89 (0.39–2.02)
Pneumothorax — no. (% [95% CI])
0.85
Noninvasive ventilation — no./
total no. (% [95% CI])
At day 28
10/212 (4.7 [1.9–7.5])
4/228 (1.8 [0.1–3.5])
0.36 (0.07–3.50)
0.11
At day 90
3/206 (1.5 [0.2–3.2])
4/225 (1.8 [0.1–3.5])
1.22 (0.23–6.97)
1.00
Tracheotomy — no./total no.
(% [95% CI])
At day 28
12/229 (5.2 [2.3–8.1])
9/237 (3.8 [1.4–6.0])
0.71 (0.27–1.86)
0.37
At day 90
18/223 (8.1 [4.5–11.7])
15/235 (6.4 [3.3–9.5])
0.78 (0.36–1.67)
0.59
*Plus–minus values are means ±SD. Hazard ratios are shown for mortality and successful extubation; odds ratios are
shown for other outcomes. CI denotes confidence interval.
†There were no significant differences between the groups in organ dysfunction as assessed from the SOFA score (Table S4
in the Supplementary Appendix).
prone group than in the supine group. The partial pressure of arterial carbon dioxide and static
compliance of the respiratory system were similar in the two groups.
Primary and Secondary Outcomes
Mortality at day 28 was significantly lower in the
prone group than in the supine group: 16.0% (38
of 237 participants) versus 32.8% (75 of 229)
(P<0.001) (Table 3). The significant difference in
mortality persisted at day 90 (Table 3). A comparison of the two survival curves showed the
same significant difference (Fig. 2). After adjustment for the SOFA score and the use of neuromuscular blockers and vasopressors at the time
of inclusion, mortality remained significantly lower in the prone group than in the supine group
(Table S5 in the Supplementary Appendix). The
rate of successful extubation was significantly
higher in the prone group (Table 3). The duration
n engl j med 368;23 nejm.org june 6, 2013
2165
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
The
n e w e ng l a n d j o u r na l
Cumulative Probability of Survival
1.0
Prone group
0.8
0.6
Supine group
0.4
0.2
0.0
P<0.001
0
10
20
30
40
50
60
70
80
90
Days
No. at Risk
Prone group
Supine group
237
229
202
163
191
150
186
139
182
136
Figure 2. Kaplan–Meier Plot of the Probability of Survival from RandomizaRevised
AUTHOR: Guérin
tion to Day 90.
FIGURE: 2 of 2
ARTIST: ts
SIZE
4 col
TYPE:of invasive
Line
Combo
4-C
H/T
22p3 length
mechanical
ventilation,
JOB:
of stay
in the ICU,
incidence
of pneumothorax, rate of use
AUTHOR,
PLEASE NOTE:
Figure
has been redrawnventilation
and type has been
reset.
of noninvasive
after
extubation, and
Please check carefully.
tracheotomy rate did not differ significantly be36823
tween the two groups (TableISSUE:
3). 06-06-13
Complications
A total of 31 cardiac arrests occurred in the supine group, and 16 in the prone group (P = 0.02).
There were no significant differences between
the groups with respect to other adverse effects
(Table S6 in the Supplementary Appendix).
Discussion
Videos showing
prone positioning
of patients with
ARDS are
available at
NEJM.org
2166
Survival after severe ARDS was significantly higher in the prone group than in the supine group.
Furthermore, the effect size was large despite the
fact that mortality in the supine group was lower
than anticipated.
Our results are consistent with findings from
previous meta-analyses2,11 and an observational
study,18 even though prior randomized trials have
failed to show a survival benefit with prone positioning. Meta-analyses of ARDS studies have
suggested that the outcomes with prone positioning are better in the subgroup of patients with
severe hypoxemia.2,11 However, when we stratified
our analysis according to quartile of Pao2:Fio2
of
m e dic i n e
ratio at enrollment, we found no significant differences in outcomes (Table S8 in the Supplementary Appendix).
Several factors may explain our results. First,
patients with severe ARDS were selected on the
basis of oxygenation together with PEEP and Fio2
levels. Second, patients were included after a
12-to-24-hour period during which the ARDS
criteria were confirmed. This period may have
contributed to the selection of patients with more
severe ARDS19 who could benefit from the advantages of the prone positioning, such as relief
of severe hypoxemia and prevention of ventilatorinduced lung injury. A previous study has shown
that prone positioning, as compared with supine
positioning, markedly reduces the overinflated
lung areas while promoting alveolar recruitment.5
These effects (reduction in overdistention and
recruitment enhancement) may help prevent ventilator-induced lung injury by homogenizing the
distribution of stress and strain within the
lungs. In our trial, alveolar recruitment was not
directly assessed. However, studies have shown
that lung recruitability correlates with the extent
of hypoxemia20,21 and that the transpulmonary
pressure along the ventral-to-dorsal axis is more
homogeneously distributed in the prone position
than in the supine position.22 We therefore suggest that prone positioning in our patients induced
a decrease in lung stress and strain.
Third, as in previous investigations,9,10 we
used long prone-positioning sessions. Fourth, the
prone position was applied for 73% of the time
ascribed to the intervention and was concentrated
over a period of a few days. Fifth, in our trial,
the tidal volume was lower than in previous trials,9,10 and the PplatRS was kept below 30 cm of
water. However, because all patients were returned
to the supine position at least once a day, the
effect of the prone position itself cannot be distinguished from the effects of being moved from
the supine to the prone position over the course
of a day.
We should acknowledge that the technical
aspects of prone positioning are not simple and
that a coordinated team effort is required (see
Videos 1 and 2, available at NEJM.org). All centers participating in this study were skilled in
the process of turning patients from the supine
to the prone position, as shown by the absence
of adverse events directly related to repositioning.
Because the experience of the units may explain
n engl j med 368;23 nejm.org june 6, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Prone Positioning in Severe ARDS
the low rate of complications, our results cannot
necessarily be generalized to centers without
such experience. We should also emphasize that
our results were obtained in the subgroup of severely ill patients with ARDS.
It could be argued that our results can be explained by higher mortality in the control group.
However, mortality at day 28 in the supine group
was similar to that among controls in recent trials.23,24 Furthermore, although the mortality in
the control group was lower than that used to
compute the power of this study, we calculated
that the power of our study was 99%.
The study has several limitations. Although
we planned to record the data of patients who
were eligible but not included, only a few ICUs
complied with this request, making it impossible to fully appreciate the physiological condition
of the excluded patients. In addition, fluid balance and the cumulative dose of catecholamines
were not assessed. The imbalance between the
groups in baseline SOFA score, vasopressor use,
and the use of neuromuscular blockers could also
have influenced the results. However, even after
adjustment for these covariates, mortality was
significantly lower in the prone group.
In conclusion, this trial showed that patients
with ARDS and severe hypoxemia (as confirmed
by a Pao2:Fio2 ratio of <150 mm Hg, with an
Fio2 of ≥0.6 and a PEEP of ≥5 cm of water) can
benefit from prone treatment when it is used
early and in relatively long sessions.
Dr. Guérin reports receiving grant support from Air Liquide;
Dr. Mercat, receiving consulting fees from Faron Pharmaceuticals, grant support from Covidien and General Electric, patent
royalties on a method for evaluating positive end-expiratory pressure that is licensed to General Electric, and reimbursement for
travel expenses from Covidien and Maquet; Dr. Jaber, receiving
consulting fees from Maquet and Dräger, lecture fees from Fisher
and Paykel, Abbott Laboratories, and Philips Respironics, and reimbursement for travel expenses from Pfizer; and Dr. Mancebo,
receiving fees for serving on the data and safety monitoring board
of Air Liquide, consulting fees from Faron Pharmaceuticals,
ALung, and Philips Respironics, and grant support to his institution from Covidien and General Electric. No other potential conflict of interest relevant to this article was reported.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
We thank all the physicians, including those on night duty,
and nurses in the participating centers for the care provided to
patients during the study; the Réseau Européen en Ventilation
Artificielle network; and Carolyn Newey for help in editing an
earlier version of the manuscript.
Appendix
The authors’ affiliations are as follows: Réanimation Médicale, Hôpital de la Croix-Rousse, Hospices Civils de Lyon; Université de Lyon;
and Creatis INSERM 1044, Lyon (C.G., J.-C.R., F.B., G.B., V.L., L.B.); Réanimation Polyvalente, and Clinical Research in Intensive Care
and Sepsis (CRICS) Group, La Roche-Sur-Yon (J. Reignier); Réanimation Polyvalente, Roanne (P.B.); Réanimation Médicale, Hôpital
Pontchaillou, Rennes (A.G.); Réanimation Polyvalente, and CRICS Group, Hôpital d’Orléans, Orleans (T.B.); Réanimation Médicale,
Hôpital Bretonneau, CRICS Group, and Université de Tours, Tours (E.M.); Réanimation Polyvalente, Hôpital de Chambéry, Chambery
(M.B.); L’Université Nantes Angers Le Mans, Université d’Angers, Centre Hospitalier Universitaire Angers, Réanimation Médicale, Angers (A.M.); Réanimation Polyvalente, and CRICS Group, Hôpital d’Angoulême, Angouleme (O.B.); Réanimation Polyvalente Centre
d’Investigation Clinique 0801 and CRICS Group, Hôpital de Limoges, Limoges (M.C.); Réanimation Médicale, Hôpital de Poitiers, and
CRICS Group, and University of Poitiers, Poitiers (D.C.); Réanimation Chirurgicale, Hôpital Saint Eloi, INSERM Unité 1046, and Université de Montpellier, Montpellier (S.J.); Réanimation Polyvalente, Hôpital Saint Joseph et Saint Luc, Lyon (S.R.); Réanimation Polyvalente,
Hôpital d’Annecy, Annecy (M.S.); Réanimation Médicale, Hôpital Pellegrin, and Université de Bordeaux, Bordeaux (G.H.); Réanimation
Polyvalente, Hôpital de Nîmes, and Université de Nîmes-Montpellier, Nimes (C.B.); Réanimation Polyvalente, Hôpital de Cergy-Pontoise, Cergy-Pontoise (J. Richecoeur); Réanimation des Urgences, Hôpital de la Timone, and Université de la Méditerranée, Marseille
(M.G.); Service d’Hygiène Hospitalière, Groupement Hospitalier Lyon Sud, Hospices Civils de Lyon, Pierre Bénite (R.G.); and Centre de
Coordination et de Lutte contre les Infections Nosocomiales Sud-Est, Hôpital Henri Gabrielle, Saint Genis-Laval (L.A.) — all in France;
and Servei de Medicina Intensiva, Hospital de Sant Pau, Barcelona (J.M.).
References
1. Abroug F, Ouanes-Besbes L, Elatrous
S, Brochard L. The effect of prone positioning in acute respiratory distress syndrome or acute lung injury: a meta-analysis:
areas of uncertainty and recommendations for research. Intensive Care Med
2008;34:1002-11.
2. Sud S, Friedrich JO, Taccone P, et al.
Prone ventilation reduces mortality in patients with acute respiratory failure and
severe hypoxemia: systematic review and
meta-analysis. Intensive Care Med 2010;
36:585-99.
3. Broccard A, Shapiro RS, Schmitz LL,
Adams AB, Nahum A, Marini JJ. Prone
positioning attenuates and redistributes
ventilator-induced lung injury in dogs.
Crit Care Med 2000;28:295-303.
4. Mentzelopoulos SD, Roussos C, Zakynthinos SG. Prone position reduces lung
stress and strain in severe acute respiratory distress syndrome. Eur Respir J 2005;
25:534-44.
5. Galiatsou E, Kostanti E, Svarna E, et
al. Prone position augments recruitment
and prevents alveolar overinflation in
acute lung injury. Am J Respir Crit Care
Med 2006;174:187-97.
6. Papazian L, Gainnier M, Marin V, et
al. Comparison of prone positioning and
high-frequency oscillatory ventilation in
patients with acute respiratory distress
syndrome. Crit Care Med 2005;33:2162-71.
7. Gattinoni L, Tognoni G, Pesenti A, et
al. Effect of prone positioning on the survival of patients with acute respiratory
failure. N Engl J Med 2001;345:568-73.
8. Guerin C, Gaillard S, Lemasson S, et
al. Effects of systematic prone positioning in hypoxemic acute respiratory failure: a randomized controlled trial. JAMA
2004;292:2379-87.
n engl j med 368;23 nejm.org june 6, 2013
2167
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Prone Positioning in Severe ARDS
9. Mancebo J, Fernández R, Blanch L, et
al. A multicenter trial of prolonged prone
ventilation in severe acute respiratory distress syndrome. Am J Respir Crit Care
Med 2006;173:1233-9.
10. Taccone P, Pesenti A, Latini R, et al.
Prone positioning in patients with moderate and severe acute respiratory distress
syndrome: a randomized controlled trial.
JAMA 2009;302:1977-84.
11. Gattinoni L, Carlesso E, Taccone P,
Polli F, Guérin C, Mancebo J. Prone positioning improves survival in severe ARDS:
a pathophysiologic review and individual
patient meta-analysis. Minerva Anestesiol
2010;76:448-54.
12. Bernard GR, Artigas A, Brigham KL,
et al. The American-European Consensus
Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical
trial coordination. Am J Respir Crit Care
Med 1994;149:818-24.
13. The Acute Respiratory Distress Syndrome Network. Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute lung injury and
the acute respiratory distress syndrome.
N Engl J Med 2000;342:1301-8.
14. McCabe WR, Treadwell TL, De Maria
A Jr. Pathophysiology of bacteremia. Am J
Med 1983;75:7-18.
15. Le Gall JR, Lemeshow S, Saulnier F.
A new Simplified Acute Physiology Score
(SAPS II) based on a European/North American multicenter study. JAMA 1993;270:
2957-63. [Erratum, JAMA 1994;271:1321.]
16. Vincent JL, Moreno R, Takala J, et al.
The SOFA (Sepsis-related Organ Failure
Assessment) score to describe organ dysfunction/failure. Intensive Care Med 1996;
22:707-10.
17. Murray JF, Matthay MA, Luce JM, Flick
MR. An expanded definition of the adult
respiratory distress syndrome. Am Rev
Respir Dis 1988;138:720-3. [Erratum, Am
Rev Respir Dis 1989:139:1065.]
18. Charron C, Bouferrache K, Caille V, et
al. Routine prone positioning in patients
with severe ARDS: feasibility and impact
on prognosis. Intensive Care Med 2011;
37:785-90.
19. Villar J, Pérez-Méndez L, López J, et
al. An early PEEP/FiO2 trial identifies different degrees of lung injury in ARDS patients. Am J Respir Crit Care Med 2007;
176:795-804.
20. Gattinoni L, Caironi P, Cressoni M, et
al. Lung recruitment in patients with the
acute respiratory distress syndrome. N Engl
J Med 2006;354:1775-86.
21. Ranieri VM, Rubenfeld GD, Thompson BT, et al. Acute respiratory distress
syndrome: the Berlin Definition. JAMA
2012;307:2526-33.
22. Mutoh T, Guest RJ, Lamm WJ, Albert
RK. Prone position alters the effect of volume overload on regional pleural pressures and improves hypoxemia in pigs in
vivo. Am Rev Respir Dis 1992;146:300-6.
23. Papazian L, Forel J-M, Gacouin A, et
al. Neuromuscular blockers in early acute
respiratory distress syndrome. N Engl J
Med 2010;363:1107-16.
24. Ferguson ND, Cook DJ, Guyatt GH, et
al. High-frequency oscillation in early acute
respiratory distress syndrome. N Engl J
Med 2013;368:795-805.
Copyright © 2013 Massachusetts Medical Society.
an nejm app for iphone
The NEJM Image Challenge app brings a popular online feature to the smartphone.
Optimized for viewing on the iPhone and iPod Touch, the Image Challenge app lets
you test your diagnostic skills anytime, anywhere. The Image Challenge app
randomly selects from 300 challenging clinical photos published in NEJM,
with a new image added each week. View an image, choose your answer,
get immediate feedback, and see how others answered.
The Image Challenge app is available at the iTunes App Store.
2168
n engl j med 368;23 nejm.org june 6, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 17, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Risk factors for the development of acute lung injury in patients
with septic shock: An observational cohort study*
Remzi Iscimen, MD; Rodrigo Cartin-Ceba, MD; Murat Yilmaz, MD; Hasrat Khan, MD; Rolf D. Hubmayr, MD;
Bekele Afessa, MD; Ognjen Gajic, MD, MSc
Objective: Almost half of the patients with septic shock develop acute lung injury (ALI). The understanding why some patients do and others do not develop ALI is limited. The objective of
this study was to test the hypothesis that delayed treatment of
septic shock is associated with the development of ALI.
Design: Observational cohort study.
Setting: Medical intensive care unit in a tertiary medical center.
Patients: Prospectively identified patients with septic shock
who did not have ALI at the outset, excluding those who denied
research authorization.
Measurements and Main Results: High frequency cardio-respiratory monitoring, arterial gas analysis, and portable chest radiographs were reviewed to identify the timing of ALI development.
Risk factors present before ALI development were identified by
review of electronic medical records and analyzed in univariate
and multivariate analyses. Seventy-one of 160 patients (44%)
S
eptic shock is one of the most
common, life-threatening medical conditions and is frequently
complicated by organ failures,
especially acute lung injury (ALI). Development of ALI is associated with short
and long term morbidity, mortality, prolonged hospitalization, and high costs
(1). While septic shock has long been
recognized as a trigger of ALI and of
acute respiratory distress syndrome
(ARDS) (2, 3), our understanding of why
*See also p. 1666.
From the Department of Internal Medicine, Division
of Pulmonary and Critical Care Medicine, Mayo Clinic
(RI, RC-C, MY, HK, RDH, BA, OG), Rochester, Minnesota;
Department of Anesthesiology and Intensive Care, Uludag University (RI), Medical Faculty, Bursa, Turkey;
Department of Anesthesiology and Intensive Care, Akdeniz University (MY), Medical Faculty, Antalya,
Turkey.
Dr. Hubmayr has consulted for Novartis and the
DSMB. The remaining authors have not disclosed any
potential conflicts of interest.
Supported by NIH grant: NHLBI 1 K23 HL087843
For information regarding this article, E-mail:
[email protected]
Copyright © 2008 by the Society of Critical Care
Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/CCM.0b013e31816fc2c0
1518
developed ALI at a median of 5 (range 2–94) hours after the onset
of septic shock. Multivariate logistic regression analysis identified
the following predictors of ALI development: delayed goal-directed resuscitation (odds ratio [OR] 3.55, 95% confidence interval [CI] 1.52– 8.63, p ⴝ .004), delayed antibiotics (OR 2.39, 95% CI
1.06 ⴚ5.59, p ⴝ .039), transfusion (OR 2.75, 95% CI 1.22– 6.37,
p ⴝ .016), alcohol abuse (OR 2.09, 95% CI .88ⴚ5.10, p ⴝ 0.098),
recent chemotherapy (OR 6.47, 95% CI 1.99ⴚ24.9, p ⴝ 0.003),
diabetes mellitus (OR .44, 95% CI .17ⴚ1.07, p ⴝ .076), and
baseline respiratory rate (OR 2.03 per SD, 95% CI 1.38ⴚ3.08,
p < .001).
Conclusion: When adjusted for known modifiers of ALI expression, delayed treatment of shock and infection were associated
with development of ALI. (Crit Care Med 2008; 36:1518–1522)
KEY WORDS: shock; metabolic acidosis; hyperventilation; epidemiology; antibiotic; resuscitation
some patients with septic shock do and
others do not develop ALI has been limited. Previous studies have identified a
history of chronic alcohol use (4), hypoalbuminemia (5), transfusions (2, 6),
pulmonary source of infection (2), and
the absence of diabetes mellitus (7) as
significant modifiers of ALI expression.
However, the timing of ALI development
has generally been poorly defined, making it difficult to separate cause from
effect of specific risk factors, particularly
those related to treatment.
This study aims to assess the effect of
delayed treatment of shock and infection
on the development of ALI using data
from prospectively identified consecutive
patients with septic shock. We hypothesized that delayed goal-directed resuscitation and delayed antibiotic administration are risk factors for the development
of ALI in patients with septic shock.
MATERIALS AND METHODS
The Mayo Clinic Institutional Review
Board approved the study protocol and waived
the need for informed consent in this observational study. Daily screening identified con-
secutive adults (⬎18 yrs of age) with septic
shock who were admitted to the medical intensive care unit (ICU) of a tertiary care center
between March 2004 and April 2007. We excluded patients with preexisting ALI or cardiogenic pulmonary edema at the onset of septic
shock, those who denied research authorization, patients in whom care was withdrawn
within 6 hrs of onset of septic shock, and those
in whom the time of onset of septic shock
started before hospital admission or could not
be accurately determined (i.e., patients transferred from another facility). The characteristics
of the ICU have been previously described (8).
The onset of septic shock was determined
according to the standards of the American
College of Chest Physicians/Society of Critical
Care Medicine consensus conference criteria
(9) when, in a patient with suspected infection, two consecutive measurements revealed
the following:
1) Two systemic inflammatory response
syndrome (SIRS) criteria (temperature
⬎38.3°C or ⬍35.6°C, heart rate ⬎90 beats/
min, respiratory rate ⬎20/min or white blood
cell [WBC] count ⬎12.0 ⫻ 103 or ⬍4.0 ⫻
103); and 2) Hypoperfusion as evidenced by a
systolic blood pressure ⬍90 mmHg or mean
arterial pressure (MAP) ⬍60 mmHg or a fall of
⬎40 mmHg from baseline despite 20 mL/kg
Crit Care Med 2008 Vol. 36, No. 5
fluid bolus, and/or serum lactate ⱖ4 mmol/L
regardless of the blood pressure (10).
Primary outcome was the development of
ALI, defined according to the standard American-European Consensus Conference (AECC)
criteria (11) as acute onset of: a) New or worsening hypoxemia with partial pressure of oxygen in arterial blood divided by inspired oxygen concentration (PaO2/FIO2) ⬍300 mmHg;
b) New bilateral pulmonary infiltrates on chest
radiography consistent with edema; and c)
The absence of clinical signs of left atrial hypertension as the principal explanation for
pulmonary edema. Secondary outcome measures included hospital mortality and length
of stay. To increase interobserver reliability,
outcome assessors reviewed a structured ALI
tutorial (12) before study onset.
Outcome assessors determined the onset of
septic shock and ALI using high frequency
cardio-respiratory monitoring as previously
described (13). Study investigators identified
newly admitted patients with septic shock at 8
a.m. every morning including weekends. The
investigators reviewed monitoring logs capturing respiratory rate, heart rate, arterial and
central venous pressures, oxygen saturation,
FIO2, laboratory findings, ventilator settings,
infusions, and other treatments to determine
the time when the criteria for septic shock and
for ALI (if present) were met. Bedside nurses
validated all of the measurements. The diagnosis of ALI was confirmed at the time of the
qualifying portable chest radiograph. The absence of ALI at baseline was assured by the
absence of hypoxemia (PaO2/FIO2 ⬎300 mm
Hg) and/or the presence of chest radiograph
without bilateral infiltrates (at least one lung
had to be clear).
Risk factors for development of ALI were
grouped as follows: 1) Underlying condition
before development of septic shock: demographics, source of infection, aspiration, history of chronic alcohol abuse, diabetes mellitus,
tobacco use, preexisting chronic obstructive
lung disease, and recent (⬍6 months) exposure
to chemotherapy. Chronic alcohol abuse was
defined as a previously established diagnosis of
chronic alcoholism, a prior admission for alcohol detoxification, or alcohol withdrawal (14).
Tobacco use was defined as history of cigarette
smoking based on the information provided in
the standardized form on hospital admission.
Aspiration was defined as aspiration of gastric
contents into the airway witnessed or strongly
suspected by the bedside provider.
2) Severity of illness, inflammation, and
hypoperfusion at the onset of septic shock
including white cell count, platelet count,
temperature, respiratory rate, lactate, and base
excess. The degree of baseline hypoxemia was
estimated from the PaO2/FIO2 (if arterial
blood gas analysis was obtained at the time of
septic shock onset) or O2sat/FIO2 ratio (15).
Crit Care Med 2008 Vol. 36, No. 5
The baseline Acute Physiologic and Chronic
Health Evaluation (APACHE III) scores were
calculated from the data available at the onset
of septic shock.
3) Treatment during the first six hours of
septic shock: time to the administration of
antibiotic therapy (16), source control (17),
goal-directed fluid resuscitation (10), transfusion of blood products, use of vasopressors,
steroid administration, and use of activated
protein C. Adequate antibiotic therapy was defined as empirical use of broad spectrum antibiotics that would cover gram positives,
gram negatives, and anaerobes according to
the suspected site of infection (16). Blood
transfusion was defined as infusion of any
plasma blood product including red blood cell
(RBC), platelets, fresh frozen plasma (FFP), or
cryoprecipitate, during the 48 hrs before the
development of ALI within 48 hrs of septic
shock in patients who did not develop ALI.
Adequate early goal-directed resuscitation was
defined as central venous oxygen saturation
ⱖ70 percent (10) and/or a combination of
clinical factors (18): central venous pressure
⬎8 mmHg, MAP ⱖ65 mmHg, urine output
ⱖ.5 mL/kg/hour, and/or improvement in
mental state (Glasgow Coma Scale-GCS), base
excess, or lactate. The achievement of resuscitation goals at any point within the first six
hours was considered as adequate goaldirected resuscitation; achievement of resuscitation goals after the first 6 hrs was considered delayed goal-directed resuscitation.
Adequate antibiotic administration ⬎3 hrs after the onset of septic shock was considered as
delayed antibiotic therapy. Monitoring logs
were reviewed and only the predictors present
at baseline (before any worsening in oxygenation— decrease in oxygen saturation to
⬍90% requiring an increase in FIO2 by the
bedside providers) were considered for the
analysis. In patients in whom FIO2 was not
directly measured, O2 supplementation was
determined according to following approximations: room air .21, 1L/min .24, 2L/min .28,
3L/min .32, 4L/min .35, ⬎4L/min .4.
Statistical Analysis. The risk factors were
compared between patients who did and did
not develop ALI in a univariate followed by
multivariate analysis. Continuous and categorical variables were compared using a Wilcoxon rank sum, Fisher’s exact test, or chisquare test, as appropriate. The variables were
considered for multivariable logistic regression models if they occurred before or during
the first six hours of septic shock resuscitation
and before the development of ALI, had less
than 10% missing data and the following: a)
had p values ⬍.1; b) had high odds ratios (OR
ⱖ2 for categorical variables or per SD of a
continuous variable); and c) were biologically
plausible. In addition to our main predictors
of interest, delayed goal-directed resuscitation
and delayed antibiotic therapy that were al-
ways in the model, previously identified modifiers of ALI (references 2– 6, see the Results
section) were included in the multivariate
analysis as covariates. The final model was
determined based on a combination of forward
selection and backward elimination taking
into consideration colinearity, interaction,
and the number of patients who experienced
outcome of interest. In a case of colinearity
between some of the variables (aspiration and
pulmonary source) the variable with “stronger” association (based on forward selection
process) was used in multivariate analysis.
JMP statistical software (version 6.0, SAS institute, Cary, NC) was used for all analyses.
RESULTS
Seventy-one of the 160 patients (44%)
developed ALI at a median of 5 (range
2–94) hrs after the onset of septic shock.
Ninety percent of patients developed ALI
during the first 12 hrs of septic shock, 64
of whom (90%) met the criteria for
ARDS. Table 1 presents univariate comparisons of chronic conditions, baseline
characteristics, resuscitation parameters,
co-exposures, and co-interventions between patients with septic shock who did
and did not develop ALI.
When adjusted for known modifiers of
ALI expression (transfusion, aspiration,
alcohol abuse, chemotherapy, and diabetes mellitus) in a multivariate logistic regression analysis (Table 2), delayed goaldirected resuscitation, delayed antibiotics, and high baseline respiratory rate
were significantly associated with development of ALI in patients with septic
shock.
When the analysis was restricted to
112 patients who had PaO2/FIO2 measurements at baseline, delayed resuscitation (odds ratio [OR] 5.0, 95% confidence
interval [CI] 1.8 –15) and delayed antibiotics (OR 2.6, 95% CI 1.0 –7.8) were associated with development of ALI.
In a subgroup analysis of 97 patients
with nonpulmonary source of sepsis, delayed resuscitation (OR 4.8, 95% CI 1.5–
16.8) but not delayed antibiotics (OR 1.9,
95% CI .62–5.74) were significantly associated with development of ALI.
To determine whether the high respiratory rate was more likely to represent
hyperventilation from shock and metabolic acidosis versus early pulmonary
dysfunction in our study, we posthoc correlated baseline respiratory rate with the
degree of baseline hypoxemia and pulmonary impairment. Baseline respiratory
rate was significantly associated with the
presence of metabolic acidosis, base def1519
Table 1. Risk factors for development of ALI in patients with septic shock
ALI (n ⫽ 71)
Underlying condition prior to development of septic shock
Age, years (IQR)
Female gender, n (%)
Diabetes mellitus, n (%)
Chronic alcohol use, n (%)
History of tobacco smoking, n (%)
COPD, n (%)
Recent chemotherapy, n (%)
ACE inhibitors (%)
Amiodarone (%)
Source of sepsis, n (%)
Pulmonary
Abdomen
Urine
Skin and soft tissue
Other
Positive blood cultures (%)
Aspiration (%)
Baseline severity of illness, inflammation, and hypoperfusion
APACHE III (IQR) at the onset of septic shock
Temperature, C (IQR)
GCS (IQR)
WBC count, per mm3 (IQR)
Platelet count, per mm3 (IQR)
Respiratory rate (IQR)
O2 sat/FIO2 (IQR)
PaO2/FIO2 (IQR) n ⫽ 112
pH (IQR) n ⫽ 153
Base excess (IQR), n ⫽ 150
Lactate (IQR), n ⫽ 148
Serum albumin, mg/dL (IQR), n ⫽ 79
Serum creatinine, mg/dL (IQR)
Treatment during the first 6 hours of septic shock
Time to antibiotic administration, minutes (IQR)
Delayed antibiotic administration
(⬎3 hours), n (%)
Source control, n (%)
Time to goal-directed resuscitation endpoints,
minutes (IQR)
Delayed goal-directed resuscitation
(⬎6 hours), n (%)
Mean MAP mm Hg (IQR)
Mean CVP (IQR), n ⫽ 121
Mean SvcO2 (IQR), n ⫽ 83
Fluid intake, mL first 6 hrs (IQR)
Fluid intake, mL 6–12 hrs (IQR)
Colloid (5% albumin) use, n (%)
Any transfusion, n (%)
RBC only, n (%)
FFP or platelets, n (%)
Dobutamine, n (%)
Vasopressors, n (%)
Corticosteroids, n (%)
Mechanical ventilationa, n (%)
Activated protein C, n (%)b
Vt, mL/kg PBW (IQR), n ⫽ 44
Ppk, cm H2O (IQR), n ⫽ 40
Outcome
ICU mortality, n (%)
Hospital mortality, n (%)
No ALI (n ⫽ 89)
68 (56 to 80)
33 (46)
17 (24)
27 (38)
38 (54)
24 (34)
17 (24)
17 (24)
7 (10)
73 (61 to 83)
39 (44)
38 (43)
15 (17)
33 (37)
23 (26)
5 (6)
22 (25)
5 (6)
35 (49)
21 (30)
8 (11)
5 (7)
2 (3)
18 (25)
18 (25)
28 (32)
24 (27)
20 (22)
10 (11)
7 (8)
18 (20)
8 (9)
p Value
.056
.737
.012
.003
.037
.273
⬍.001
.910
.313
.064
.441
.005
61 (49 to 72)
36.8 (36.1 to 38.2)
13 (9 to 15)
16 (9 to 24)
192 (113 to 293)
27 (23 to 29)
405 (263 to 457)
200 (100 to 260)
7.27 (7.19 to 7.34)
⫺8 (⫺12 to ⫺5)
3.2 (1.7 to 5.5)
2.4 (1.9 to 3.2)
2.1 (1.4 to 3.1)
55 (44 to 64)
36.9 (36.4 to 38.2)
14 (13 to 15)
12 (8 to 18)
181 (107 to 283)
20 (18 to 25)
431 (331 to 467)
300 (222 to 400)
7.37 (7.31 to 7.42)
⫺6 (⫺8 to ⫺2)
2.4 (1.4 to 4.1)
2.8 (2.4 to 3.3)
1.8 (1.3 to 3.4)
223 (82.5 to 360)
44 (62)
113 (0 to 235)
36 (40)
.001
.007
12 (17)
505 (329 to 653)
12 (14)
320 (202 to 480)
.550
.001
36 (51)
58 (53 to 62)
6 (4 to 8)
64 (55 to 75)
4518 (2982 to 6723)
2320 (1450 to 3992)
14 (20)
40 (56)
10 (14)
16 (23)
8 (11)
68 (96)
37 (52)
19 (27)
9 (13)
6.9 (6.1 to 7.8)
30 (27 to 35)
27 (38)
36 (51)
20 (22)
58 (55 to 62)
5 (3 to 8)
67 (61 to 72)
5394 (3476 to 6766)
1851 (1072 to 2893)
17 (19)
27 (30)
15 (17)
12 (14)
7 (8)
67 (75)
40 (45)
35 (39)
2 (2)
7.0 (6.5 to 7.9)
26 (23 to 31)
10 (11)
16 (18)
.028
.499
.039
.032
.548
⬍.001
.033
⬍.001
⬍.001
⬍.001
.038
.050
.566
⬍.001
.703
.200
.580
.527
.020
.922
⬍.001
.631
.136
.465
⬍.001
.367
.093
.008
.362
.097
⬍.001
⬍.001
IQR, interquartile range; WBC, white blood cell; ALI, acute lung injury; MAP, mean arterial blood pressure; ACE, angiotensin-converting enzyme; GCS,
Glasgow Coma Scale; RBC, red blood cells; FFP, fresh frozen plasma; COPD, chronic obstructive pulmonary disease; Ppk, peak airway pressure; Vt, tidal
volume; PBW, predicted body weight; CVP, central venous pressure; APACHE, Acute Physiology and Chronic Health Evaluation; ICU, intensive care unit.
a
During the first 6 hours of resuscitation. In patients who developed ALI, only mechanical ventilation started before the development of ALI was
included. bActivated protein C was used after the first 6 hours in all patients and after the development of ALI in patients with ALI.
1520
Crit Care Med 2008 Vol. 36, No. 5
Table 2. Risk factors for development of ALI in patients with septic shock: multiple logistic
regression analysis
Delayed goal-directed resuscitation
Delayed antibiotics
Respiratory rate (per SD)
Chemotherapy
Chronic alcohol use
Transfusion
Aspiration
Diabetes mellitus
Odds Ratio
95% CI
p Value
3.55
2.39
2.03
6.47
2.09
2.75
3.48
.44
1.52–8.63
1.06–5.59
1.38–3.08
1.99–24.9
.88–5.10
1.22–6.37
1.22–10.78
.17–1.07
.004
.039
⬍.001
.003
.098
.016
.024
.076
SD, standard deviation; CI, confidence interval; ALI, acute lung injury.
In addition to delayed goal-directed resuscitation and delayed antibiotic therapy, the following
variables were entered into the stepwise logistic regression procedure: O2 sat /FIO2, diabetes, alcohol,
aspiration, age, pulmonary source, smoking, transfusion, vasopressors, chemotherapy, baseline respiratory rate, baseline Acute Physiology and Chronic Health Evaluation III score, white cell count,
Glasgow Coma Scale, pH, base deficit, lactate.
icit (p ⫽ .004) but not with hypoxemia
(p ⫽ .209), or a pulmonary source of
sepsis (p ⫽ .790).
When adjusted for baseline APACHE
III scores, age, metastatic cancer, and delayed resuscitation, development of ALI
was associated with increased hospital
mortality (OR 4.4, 95% CI 2.0 to 11).
DISCUSSION
In this cohort of patients with septic
shock who did not have ALI at the outset,
nearly half of the patients developed ALI.
As determined from high frequency cardio-respiratory monitoring, the majority
of patients developed ALI within the first
12 hrs after the septic shock onset. When
adjusted for known modifiers of ALI
expression (history of recent chemotherapy, transfusion, aspiration, diabetes mellitus, and alcohol abuse), the development of ALI was associated with delayed
treatment of shock (goal-directed resuscitation), inadequate control of infection,
and high baseline respiratory rate. Patients who developed ALI had higher hospital mortality.
Early antibiotic administration and
early goal-directed therapy have been
identified as the most important treatment options in septic shock and have
been strictly recommended by the Surviving Sepsis Campaign and the Institute
for Healthcare Improvement (17). Our
study adds to the existing knowledge
demonstrating that the development of
one of the most important complications
of septic shock, ALI, is associated with
delayed treatment. In the landmark study
by Rivers et al., delayed goal-directed resuscitation was associated with sudden
cardiovascular collapse and the trend toCrit Care Med 2008 Vol. 36, No. 5
ward development of multiorgan failure
(10). However, the authors did not assess
the development of ALI in this study.
Although short of statistical significance, our study confirmed known associations among the chronic alcohol abuse
and the absence of diabetes mellitus and
the development of ALI (4, 7). Previous
studies have shown that even submassive
transfusions increase the probability of
ALI development acting as one of the
multiple hits occurring during resuscitation of critically ill patients (2, 6). As
previously reported for a general ICU
population, transfusion of plasma containing blood products was associated
with a higher ALI risk than transfusion of
RBCs (6). Multiple studies identified pulmonary aspiration as a significant risk
factor for ALI development (19). In our
study, aspiration appeared to be a more
important risk factor than having identified the lung as a likely source of sepsis.
The strong association between recent
cancer chemotherapy and development of
ALI comes as no surprise since ALI is a
known complication of many cancer chemotherapeutic agents. Indeed, chemotherapy had the highest odds ratio of all
risk factors examined in our study. The
mechanisms and their possible interactions with septic shock and its management are not known. However, they likely
involve the generation of reactive oxygen
species and their effects on alveolar barrier properties.
The association between high baseline
respiratory rate and the development of
ALI confirmed the findings of a recent
large prospective cohort of patients at
risk for ALI (2). Baseline respiratory rate
correlated with metabolic acidosis (but
not with hypoxemia) supporting an interesting experimental observation that
(even spontaneous) hyperventilation may
be a contributing factor in the development and expression of ALI (20). Since
the spontaneous tidal volumes were not
measured, we could not exclude the alternative explanation, i.e., patients who
had a higher respiratory rate at baseline
may have already had a subclinical form
of ALI and were going to develop the full
expression of the syndrome some hours
later. However, the fact that more patients in the control group underwent
mechanical ventilation at baseline 39%
vs. 27%, p ⫽ .09) is not consistent with
this alternative explanation. In a subgroup of patients who were mechanically
ventilated during the first six hours of
resuscitation, tidal volumes were small
and similar between the patients who did
and did not develop ALI. Peak airway pressures tended to be higher in patients who
were to develop ALI, suggesting an already
smaller “baby lung” compartment.
Significant limitations of our study include the inability to assess causality as
opposed to association. For example, the
inability to achieve resuscitation thresholds within a certain time period may
have been due to inadequate and delayed
therapy or to more severe original insult.
Also, while the patients with septic shock
and ALI were prospectively identified, all
of the predictor variables were collected
by the passive review of cardio-respiratory monitoring logs, nurses’ and physicians’ observations, and patient/family
provided information in the electronic
medical records leading to potential measurement error. For example, rather than
on the basis of standard validated questionnaires, the definition of alcohol abuse
was based on “a known diagnosis of
chronic alcoholism or a previous admission for alcohol detoxification or alcohol
withdrawal” (14). In addition, because a
significant number of patients did not
have arterial blood gas performed at the
time of septic shock onset, baseline PaO2/
FIO2 was substituted by O2sat/FIO2 (15),
which was available in all patients. Nevertheless, our study is one of the first
attempts to assess the exact timing of
exposures of interest in relationship to
the development of critical care syndromes, such as ALI, using data from
continuous monitoring available through
high resolution clinical information systems. An additional limitation of our
study stems from the complex and some1521
what imprecise definitions of both critical
care syndromes (septic shock and ALI)
and their supportive interventions (timing of adequate antibiotics and resuscitation). While we paid attention to use precise definitions, formally trained the
observers (12), and independently assessed predictors and the outcome, full
blinding of the observers was not possible
within the integrated electronic information system. Finally, the generalizability
of our results is limited as it was done in
the medical ICU of a single institution.
The fact that patients who developed
ALI in our study had higher adjusted hospital mortality is similar to some but not
all previous reports. Eggimann et al. recently found that ALI per se was not associated with an increased mortality
when adjusted to baseline characteristics
and evolving organ failures (21). Since we
limited our predictors to only those
present before ALI development, we did
not adjust for subsequent development of
nonpulmonary organ failures. It is not
known to what extent nonpulmonary organ failures are caused or contributed by
underlying septic shock and its management as opposed to biotrauma related to
ALI development and its treatment with
mechanical ventilation (22).
In conclusion, ALI is a common complication and occurs early in the course of
septic shock. In addition to the known
modifiers of ALI expression such as transfusion, aspiration, chemotherapy, alcohol
use, and diabetes, our study identified
delayed goal-directed resuscitation, delayed antibiotic administration, and baseline high respiratory rate as important determinants of ALI development. Targeting
the modifiers of ALI expression may be an
attractive strategy for prevention of this
common and devastating ICU complication.
1522
REFERENCES
1. Rubenfeld GD, Herridge MS: Epidemiology
and outcomes of acute lung injury. Chest
2007; 131:554 –562
2. Gong MN, Thompson BT, Williams P, et al:
Clinical predictors of and mortality in acute
respiratory distress syndrome: potential role
of red cell transfusion. Crit Care Med 2005;
33:1191–1198
3. Hudson LD, Milberg JA, Anardi D, et al: Clinical risks for development of the acute respiratory distress syndrome. Am J Respir Crit
Care Med 1995; 151:293–301
4. Moss M, Burnham EL: Chronic alcohol
abuse, acute respiratory distress syndrome,
and multiple organ disfunction. Crit Care
Med 2003; 31(4 Suppl):S207–S212
5. Mangialardi RJ, Martin GS, Bernard GR, et
al: Hypoproteinemia predicts acute respiratory distress syndrome development, weight
gain, and death in patients with sepsis. Ibuprofen in Sepsis Study Group. Crit Care Med
2000; 28:3137–3145
6. Khan H, Belsher J, Yilmaz M, et al: Freshfrozen plasma and platelet transfusions are
associated with development of acute lung
injury in critically ill medical patients. Chest
2007; 131:1308 –1314
7. Moss M, Guidot DM, Steinberg KP, et al:
Diabetic patients have a decreased incidence
of acute respiratory distress syndrome. Crit
Care Med 2000; 28:2187–2192
8. Afessa B, Keegan MT, Hubmayr RD, et al:
Evaluating the performance of an institution
using an intensive care unit benchmark.
Mayo Clin Proc 2005; 80:174 –180
9. Bone RC, Sibbald WJ, Sprung CL: The ACCPSCCM consensus conference on sepsis and
organ failure. Chest 1992; 101:1481–1483
10. Rivers E, Nguyen B, Havstad S, et al: Early
goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med
2001; 345:1368 –1377
11. Bernard GR, Artigas A, Brigham KL, et al: The
American-European Consensus Conference on
ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J
Respir Crit Care Med 1994; 149:818 – 824
12. Tutorial: Improving the Radiographic Diagnosis of Acute Lung Injury. Tutorial. June
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
20th 2007; http://depts.washington.edu/
kclip/about.shtml
Gajic O, Rana R, Winters JL, et al: Transfusion related acute lung injury in the critically
Ill: prospective nested case-control study.
Am J Respir Crit Care Med 2007; 176:
839 – 840
Moss M, Bucher B, Moore FA, et al: The role
of chronic alcohol abuse in the development
of acute respiratory distress syndrome in
adults. JAMA 1996; 275:50 –54
Rice TW, Wheeler AP, Bernard GR, et al:
Comparison of the SpO2/FiO2 Ratio and the
PaO2/FiO2 Ratio in Patients with Acute Lung
Injury or Acute Respiratory Distress Syndrome. Chest 2007; 132:410 – 417
Kumar A, Roberts D, Wood KE, et al: Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock.
Crit Care Med 2006; 34:1589 –1596
Dellinger RP, Carlet JM, Masur H, et al: Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock.
Crit Care Med 2004; 32:858 – 873
Antonelli M, Levy M, Andrews PJ, et al: Hemodynamic monitoring in shock and implications for management. International Consensus Conference, Paris, France, 27–28
April 2006. Intensive Care Med 2007; 33:
575–590
Garber BG, Hebert PC, Yelle JD, et al: Adult
respiratory distress syndrome: a systemic
overview of incidence and risk factors. Crit
Care Med 1996; 24:687– 695
Mascheroni D, Kolobow T, Fumagalli R,
et al: Acute respiratory failure following
pharmacologically induced hyperventilation:
an experimental animal study. Intensive
Care Med 1988; 15:8 –14
Eggimann P, Harbarth S, Ricou B, et al:
Acute respiratory distress syndrome after
bacteremic sepsis does not increase mortality. Am J Respir Crit Care Med 2003; 167:
1210 –1214
Tremblay L, Valenza F, Ribeiro SP, et al:
Injurious ventilatory strategies increase
cytokines and c-fos m-RNA expression in
an isolated rat lung model. J Clin Invest
1997; 99:944 –952
Crit Care Med 2008 Vol. 36, No. 5
CARING FOR THE
CRITICALLY ILL PATIENT
Scan for Author
Audio Interview
ONLINE FIRST
Daily Sedation Interruption in Mechanically
Ventilated Critically Ill Patients
Cared for With a Sedation Protocol
A Randomized Controlled Trial
Sangeeta Mehta, MD
Lisa Burry, PharmD
Deborah Cook, MD
Dean Fergusson, PhD
Marilyn Steinberg, RN
John Granton, MD
Margaret Herridge, MD
Niall Ferguson, MD
John Devlin, PharmD
Maged Tanios, MD
Peter Dodek, MD
Robert Fowler, MD
Karen Burns, MD
Michael Jacka, MD
Kendiss Olafson, MD
Yoanna Skrobik, MD
Paul Hébert, MD
Elham Sabri, MSc
Maureen Meade, MD
for the SLEAP Investigators and the
Canadian Critical Care Trials Group
C
RITICALLY ILL PATIENTS WEAN
more quickly from mechanical ventilation, with lower
risk of delirium, when clinicians use specific strategies to reduce
excessive sedation. 1 - 3 A nursingimplemented sedation titration protocol that specifies clear targets for level
of awareness is one approach to miniFor editorial comment see p 2030.
Author Audio Interview available at
www.jama.com.
Context Protocolized sedation and daily sedation interruption are 2 strategies to minimize sedation and reduce the duration of mechanical ventilation and intensive care unit
(ICU) stay. We hypothesized that combining these strategies would augment the benefits.
Objective To compare protocolized sedation with protocolized sedation plus daily
sedation interruption in critically ill patients.
Design, Setting, and Patients Randomized controlled trial of 430 critically ill, mechanically ventilated adults conducted in 16 tertiary care medical and surgical ICUs in
Canada and the United States between January 2008 and July 2011.
Intervention Continuous opioid and/or benzodiazepine infusions and random allocation to protocolized sedation (n=209) (control) or to protocolized sedation plus daily
sedation interruption (n=214). Using validated scales, nurses titrated infusions to achieve
light sedation. For patients receiving daily interruption, nurses resumed infusions, if indicated, at half of previous doses. Patients were assessed for delirium and for readiness for
unassisted breathing.
Main Outcome Measure Time to successful extubation. Secondary outcomes included duration of stay, doses of sedatives and opioids, unintentional device removal,
delirium, and nurse and respiratory therapist clinical workload (on a 10-point visual
analog scale [VAS]).
Results Median time to successful extubation was 7 days in both the interruption and
control groups (median [IQR], 7 [4-13] vs 7 [3-12]; interruption group hazard ratio, 1.08;
95% CI, 0.86-1.35; P=.52). Duration of ICU stay (median [IQR], 10 [5-17] days vs 10
[6-20] days; P=.36) and hospital stay (median [IQR], 20 [10-36] days vs 20 [10-48] days;
P=.42) did not differ between the daily interruption and control groups, respectively. Daily
interruption was associated with higher mean daily doses of midazolam (102 mg/d vs 82
mg/d; P=.04) and fentanyl (median [IQR], 550 [50-1850] vs 260 [0-1400]; P⬍.001)
and more daily boluses of benzodiazepines (mean, 0.253 vs 0.177; P=.007) and opiates
(mean, 2.18 vs 1.79; P⬍.001). Unintentional endotracheal tube removal occurred in 10
of 214 (4.7%) vs 12 of 207 patients (5.8%) in the interruption and control groups, respectively (relative risk, 0.82; 95% CI, 0.36-1.84; P=.64). Rates of delirium were not significantly different between groups (53.3% vs 54.1%; relative risk, 0.98; 95% CI, 0.821.17; P=.83). Nurse workload was greater in the interruption group (VAS score, 4.22 vs
3.80; mean difference, 0.41; 95% CI, 0.17-0.66; P=.001).
Conclusion For mechanically ventilated adults managed with protocolized sedation, the addition of daily sedation interruption did not reduce the duration of mechanical ventilation or ICU stay.
Trial Registration clinicaltrials.gov Identifier: NCT00675363
JAMA. 2012;308(19):1985-1992
Published online October 17, 2012. doi:10.1001/jama.2012.13872
Author Affiliations are listed at the end of this
article.
Corresponding Author: Sangeeta Mehta, MD, Mount
Sinai Hospital, 600 University Ave, Ste 18-216, Toronto,
©2012 American Medical Association. All rights reserved.
www.jama.com
ON, M5G 1X5 Canada ([email protected]).
Caring for the Critically Ill Patient Section Editor: Derek
C. Angus, MD, MPH, Contributing Editor, JAMA
([email protected]).
JAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
1985
DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS
mize sedation.4 Daily interruption of
sedative infusions may achieve the
same goal if infusions are resumed
only when necessary and at half the
previous dose.5 Early clinical trials
evaluating each strategy led to strong
recommendations for their use in
practice.6 However, results of subsequent clinical trials varied,7-10 and use
of these strategies in clinical practice
has been inconsistent.11,12 Concerns
about daily interruption of sedation
include patient discomfort, unintentional device removal, and increased
clinician workload.13,14 A systematic
review of 5 trials that evaluated daily
interruption highlighted the need for
further research.15
Avoiding excessive sedation is intuitively appealing. In light of the observed and potential benefits of both
protocolized sedation and daily interruption in some settings, we hypothesized that mechanically ventilated
adults managed with both strategies
would receive less sedation and have a
shorter duration of mechanical ventilation than patients managed with protocolized sedation alone.
METHODS
We conducted this multicenter randomized controlled trial in 16 centers from
January 2008 to July 2011, after approval
from local institutional review boards. In
preparation, we completed a 65-patient,
3-center pilot randomized trial.16
Participants
Eligible critically ill adults were those expected by the intensive care unit (ICU)
team to require mechanical ventilation
for at least 48 hours after enrollment and
for whom the ICU team had decided to
initiate continuous sedative and/or opioid infusion(s). Patients admitted to the
ICU after cardiac arrest or traumatic brain
injury were excluded, as were patients
receiving neuromuscular blocking
agents, those enrolled in another trial,
those previously enrolled in the current
trial, or those for whom there was a lack
of commitment to maximal treatment.
Legally authorized surrogates provided
written informed consent.
1986
JAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
Randomization and Masking
Research staff randomized patients to
protocolized sedation plus daily interruption (interruption group) or protocolized sedation alone (control group),
using an automated telephone system
that stratified by center with undisclosed variable block sizes. None of the
participants, study personnel, clinicians, or investigators analyzing data
was masked to group assignment.
Procedures
Bedside nurses titrated analgesic and
sedative infusions according to a protocol that prioritized pain assessment
(eFigures 1 and 2, available at http:
//www.jama.com). Morphine, fentanyl, or hydromorphone was administered for analgesia; midazolam or
lorazepam, for sedation. Nurses used
the Sedation-Agitation Scale17 (8 sites)
(eTable 1) or the Richmond Agitation
Sedation Scale18 (8 sites) (eTable 2) to
assess sedation needs hourly and titrated infusions to maintain, ideally, a
comfortable yet rousable state equivalent to a Sedation-Agitation Scale score
of 3 or 4 or Richmond Agitation Sedation Scale score of −3 to 0. When the
sedation score directed an increase in
medication, the bedside nurse judged
whether to increase the opioid and/or
benzodiazepine infusions. When patients were oversedated, nurses alternately reduced opioid and benzodiazepine infusions. Midazolam and
morphine were reduced in 1- to 2-mg
decrements, fentanyl in 12.5- to 25-␮g
decrements, and hydromorphone in
0.1- to 0.5-mg decrements at 15- to 30minute intervals. If doses of midazolam, lorazepam, or morphine were
less than 3 mg/h, 0.5-mg decrements
could be used. If Sedation-Agitation
Scale score was 1 to 2 (Richmond Agitation Sedation Scale score −4 or −5),
yet the patient showed signs of agitation or distress, bolus doses were administered as needed. When patients
were extremely agitated (SedationAgitation Scale score 7; Richmond Agitation Sedation Scale score 3 or 4),
nurses could deviate from this protocol. For both groups, infusions were
discontinued when a patient was
oversedated (Sedation-Agitation Scale
score 1 or 2; Richmond Agitation
Sedation Scale score −4 or −5) while
receiving 0.5 to 1 mg/h of midazolam
or morphine (or fentanyl, 12.5-25 ␮g/
h). Intermittent dosing was permitted
for procedures. Propofol, ketamine, and
dexmedetomidine infusions were not
permitted.
In the interruption group, bedside
nurses interrupted benzodiazepine and
opioid infusions daily and assessed
hourly for wakefulness, defined as Sedation-Agitation Scale score 4 to 7
(Richmond Agitation Sedation Scale
score −1 to 4) and ability to perform at
least 3 of the following on request: eye
opening, tracking, hand squeezing, and
toe moving. If the bedside nurse and a
physician agreed that infusions were no
longer required (the patient was free of
discomfort and agitation and the Sedation-Agitation Scale score was between 2 and 5 or the Richmond Agitation Sedation Scale score was between
−4 and 1), oral or bolus intravenous
therapy was used at their discretion. Alternatively, if they judged that ongoing benzodiazepine or opioid infusions were required, nurses resumed
infusions at half of the previous dose
and titrated to achieve the target level
of light sedation. If a patient became agitated (Sedation-Agitation Scale score 6
or 7 or Richmond Agitation Sedation
Scale score 2 to 4) or exhibited signs
of discomfort (respiratory rate ⬎35/
min, oxygen saturation as measured by
pulse oximetry ⬍90%, heart rate ⬎140/
min or a change in heart rate of 20%
in either direction, systolic blood pressure ⬎180 mm Hg, or increased anxiety and diaphoresis) before the physician’s assessment, nurses promptly
resumed infusions at half the previous
rate. Daily interruption could be delayed for procedures. When an interruption was not performed or infusions were not restarted at 50% of the
previous dosage, the primary reason
was documented. We also recorded
any interruption of benzodiazepine
and opioid infusions among control
patients.
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS
Patients were weaned from mechanical ventilation at the discretion of the
ICU team. To standardize the assessment of a patient’s extubation readiness, respiratory therapists evaluated
patients daily at their current ventilator settings for the following criteria:
awake, adequate cough with suctioning, PaO2 greater than 60 mm Hg, oxygen saturation greater than or equal to
90%, fraction of inspired oxygen less
than or equal to 0.4, positive end expiratory pressure less than or equal to
10 cm H2O, respiratory rate less than
or equal to 35/min, ventilation less than
or equal to 15 L/minute, no inotrope
or vasopressor infusions, mean arterial pressure greater than 60 mm Hg,
and no evidence of acute myocardial ischemia (ie, chest pain, consistent electrocardiogram findings, elevated biomarker levels, or new arrhythmia). If
all criteria were met, a 1-hour trial of
unassisted breathing was initiated, during which ventilatory support was withdrawn and the patient breathed spontaneously at the previous fraction of
inspired oxygen through a t-tube circuit, a tracheostomy mask, or the ventilator circuit with continuous positive airway pressure of 5 cm H2O. The
breathing trial could be terminated if
any of the following signs of failure persisted for more than 5 minutes: respiratory rate greater than 35/min, oxygen saturation less than 90%, heart rate
greater than 140/min or a change in
heart rate of 20% in either direction, systolic blood pressure less than 90 or
greater than 180 mm Hg, or increased
anxiety and diaphoresis. A breathing
trial was successful if the patient could
breathe without mechanical assistance for 1 hour. When patients passed
a trial of unassisted breathing, respiratory therapists notified a physician with
a view to extubation. Research staff recorded reasons for delayed extubation, and daily screening continued until extubation. If the patient did not pass
the unassisted breathing trial, the previous ventilator settings were resumed and the screening and breathing trials were repeated daily until
extubation. If reintubation occurred
Figure 1. Flow of Patients in the Trial
2091 Patients screened
1661 Excluded
403 No surrogate decision maker
371 Refused consent
179 Physician refused participation
104 Missed/no research staff
86 Enrolled in another trial
13 Not specified
10 Previously in SLEAP
495 Other a
430 Patients randomized
218 Randomized to receive
protocolized sedation
and daily interruption
212 Randomized to receive
only protocolized sedation
4 Consent withdrawn
3 Consent withdrawn
214 Included in analysis
209 Included in analysis
a Other includes 362 patients receiving propofol, 39 with open abdomen or chest, 33 needing ongoing deep
sedation (because of a plan to return to the operating room, severe agitation, chronic pain, precarious airway,
or hemodynamic instability), and 23 receiving high-frequency ventilation. For the remainder, please see the
supplemental eAppendix.
within 48 hours, study sedation procedures resumed.
Bedside nurses also assessed daily for
delirium with the Intensive Care Delirium Screening Checklist.19 Patients
in both groups were managed “off protocol” during periods of neuromuscular blockade, high-frequency oscillation, or palliative care.
Outcomes
The primary study outcome was time
to successful extubation, defined as time
from randomization to extubation (or
tracheostomy mask) for 48 hours. Secondary outcomes included unintentional device removal (eg, endotracheal tubes), physical restraint use,
delirium, neuroimaging in the ICU, tracheostomy, barotrauma, total doses of
sedatives and analgesics during mechanical ventilation, organ dysfunction, ICU and hospital lengths of stay,
and death. Twice daily, nurses and respiratory therapists recorded their additional clinical workload attributed to
study procedures, using a 10-point visual analog scale (VAS), with 1 corresponding to “very easy” and 10 to “difficult.” For patients assigned to daily
interruption, we measured the propor-
©2012 American Medical Association. All rights reserved.
tion of days during which sedation was
interrupted.
Statistical Analysis
The sample size estimate assumed a median time to successful extubation of 7
days among controls and a 2-day reduction with the addition of daily interruption (hazard ratio 1.4). We determined that 205 patients per group
would provide a power of 90%, with an
␣ level of 5%.
Our primary analysis was based on an
intention-to-treat principle whereby all
patients were analyzed according to their
original group allocation, regardless of
whether they received the allocated treatment. We used the Kaplan-Meier
method to estimate and plot the distributions of time to successful extubation and an unadjusted Cox proportional hazards model to estimate a
hazard ratio. For the time-to-extubation analysis, the event occurred when
a patient was extubated within 28 days
from randomization and remained extubated for more than 48 hours. Patients who died before extubation or who
were transferred to another institution
before 28 days were censored at death
or transfer. Patients undergoing withJAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
1987
DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS
drawal of life support were censored
when that decision was made.
We also conducted a per-protocol
analysis of patients who had interrup-
tions on more than 75% of eligible study
days and 1 prespecified subgroup analysis, according to classification of a patient’s ICU admission as medical vs sur-
Table 1. Baseline Characteristics
No. (%)
Characteristics
Age, median (IQR), y
Women
Type of admission a
Medical
Surgical
Trauma
BMI, median (IQR)
APACHE II score, median (IQR) b
SOFA at day 1, median (IQR) c
Mechanical ventilation, median (IQR), d
Opioid infusions
No. (%)
Days of infusion, median (IQR)
Benzodiazepine infusions
No. (%)
Days of infusion, median (IQR)
ICU admission diagnosis d
Bacterial/viral pneumonia
Nonurinary sepsis
Other respiratory disease
Aspiration pneumonia
COPD
Postoperative respiratory disease
Urinary sepsis
Gastrointestinal perforation/rupture
Hepatic failure
Noncardiogenic pulmonary edema
Other
Pre-ICU conditions e
Alcohol use
Tobacco use
Any psychiatric condition
Any neurologic condition
Respiratory disease
Renal dysfunction
Habitual drug use
Liver disease
Protocolized Sedation
and Daily Interruption
(n = 214)
57 (46-70)
93 (43.5)
175 (81.8)
30 (14.5)
8 (3.7)
28.2 (23.8-34.2)
24.0 (18-28)
7 (5-10)
2 (1-4)
Protocolized Sedation
(n = 209)
60 (49-70)
92 (44.0)
179 (86.1)
22 (11.0)
6 (2.9)
28.6 (25.0-33.2)
23.0 (19-29)
6 (4-9)
2 (1-4)
184 (87)
1 (1-3)
186 (89)
1 (1-3)
169 (81)
1 (1-3)
163 (80)
1 (1-3)
39 (18.2)
40 (18.7)
22 (10.3)
11 (5.1)
4 (1.9)
7 (3.3)
3 (1.4)
6 (2.8)
6 (2.8)
5 (2.3)
71 (33.2)
47 (22.5)
36 (17.2)
21 (10.0)
4 (1.9)
10 (4.8)
7 (3.3)
9 (4.3)
5 (2.4)
4 (1.9)
4 (1.9)
62 (29.7)
49 (23.0)
48 (22.5)
42 (19.6)
33 (15.4)
17 (8.0)
20 (9.4)
14 (6.6)
12 (5.6)
44 (21.2)
40 (19.3)
29 (14.4)
36 (17.2)
26 (12.4)
16 (7.7)
10 (4.8)
11 (5.3)
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation; BMI, body mass index (calculated as weight
in kilograms divided by height in meters squared); COPD, chronic obstructive pulmonary disease; ICU, intensive care
unit; IQR, interquartile range; SOFA, Sequential Organ Failure Assessment.
a Surgical refers to admission from an operating room or postoperative recovery area.
b APACHE II score may range from 0 to 71, with higher scores indicating more severe disease.
c SOFA score may range from 0 to 24 points, with higher scores indicating more severe disease.
d Diagnoses in this category are mutually exclusive. The 10 most frequent diagnoses are listed, and the remainder are
categorized as “other.”
e Pre-ICU conditions are listed in descending frequency: neurologic condition defined as stroke, seizure disorder, dementia, neuromuscular disease, Parkinson disease, or other neurologic condition; psychiatric condition includes depression, bipolar disorder, schizophrenia, anxiety disorder, or other psychiatric condition; respiratory disease defined as home oxygen, carbon dioxide retention at baseline, or home ventilation; renal dysfunction defined as chronic
renal failure with creatinine level greater than 180 ␮mol/L or chronic dialysis; liver disease defined as Child Pugh
Grade C or known esophageal varices; and habitual drug use other than tobacco or alcohol.
1988
JAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
gical/trauma. We hypothesized that
medical patients would benefit more
than surgical patients from daily interruption, given their anticipated longer durations of mechanical ventilation and sedative infusions.
Descriptive data are presented as
percentages, means with standard
deviations for normally distributed
variables, and medians with interquartile ranges for nonnormally distributed variables. Sedative and opioid
doses are presented as midazolam and
fentanyl equivalents, respectively.20
We converted Richmond Agitation
Sedation Scale values to SedationAgitation Scale scores for analyses
(eTable 3).
To examine between-group differences in categorical variables, we used
␹2 or Fisher exact tests, as appropriate.
For dichotomous outcomes, we present relative risks or hazard ratios and
their 95% CIs. If all assumptions were
met for parametric analyses of the continuous variables, we used a 2-sample t
test; otherwise, we used a 2-sample Wilcoxon rank sum test. Mean SedationAgitation Scale and VAS scores per patient and mean differences with 95% CIs
were calculated. All statistical tests were
2-sided and considered statistically significant at ␣⬍.05. SAS version 9.2 and
S-Plus version 7.0 were used for statistical analysis.
An independent data and safety
monitoring committee reviewed trial
progress and adverse events after randomization of 67, 117, and 292 patients. They also reviewed blinded data
for 1 planned interim analysis after enrollment of 211 patients.
RESULTS
Participants
Patients were enrolled in 14 Canadian
and 2 US centers. Of 2091 eligible patients, 1661 were not enrolled, primarily because of lack of an authorized decision maker (24.3%), consent refusal
(22.3%), or physician refusal (10.8%)
(FIGURE 1). Among 430 randomized
patients, 7 withdrew consent in the first
3 days of the study and were excluded
from the analysis.
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS
Patient characteristics were similar
in the 2 groups (TABLE 1). Eightyfour percent received medical diagnoses. At enrollment, 359 (84.9%) patients were receiving midazolam
infusions; 334 (79.0%), fentanyl; 71
(16.5%), morphine; and 41 (9.5%),
propofol. Propofol infusions were discontinued at enrollment according to
the study protocol.
ence in time to successful extubation
between groups. There were no between-group differences in ICU or hospital lengths of stay, hospital mortality, rates of unintentional device
removal, delirium, ICU neuroimaging, barotrauma, tracheostomy, or organ dysfunction (TABLE 2).
TABLE 3 summarizes data related to
sedative and opioid administration. Pa-
Figure 2. Kaplan-Meier Curves for Time to Successful Extubation
1.0
Proportion Intubated
0.8
Outcomes
The median time to successful extubation was 7 days in both groups (hazard
ratio, 1.08; 95% CI, 0.86-1.35; P=.52)
(FIGURE 2). Adjustment for age, body
mass index, Acute Physiology and
Chronic Health Evaluation II score, and
admission type gave consistent results
(adjusted hazard ratio, 1.04; 95% CI,
0.83-1.31). Similarly, in a perprotocol analysis of patients who had
interruptions on more than 75% of eligible study days, there was no differ-
0.6
0.4
P = .52
Protocolized sedation only
0.2
0
Protocolized sedation and
daily interruption
0
5
10
15
20
25
34
28
23
16
30
Time, d
No. at risk
Protocolized sedation only
Protocolized sedation and
daily interruption
209
214
146
140
72
81
49
42
P value calculated from log-rank statistic.
Table 2. Patient Outcomes
Days to successful extubation, median (IQR) a
Days in ICU, median (IQR) a
Days in hospital, median (IQR) a
ICU mortality, No. (%)
Hospital mortality, No. (%)
ICU-acquired organ failure and supportive
therapies, No. (%)
ARDS
Vasopressors/inotropes
Renal replacement
Neuromuscular blockade
Unintentional device removal, No. (%)
Gastric tube
Endotracheal tube
Urinary catheter
Central venous or arterial catheter
Neuroimaging in ICU, No. (%)
Computed tomography
Magnetic resonance imaging
Physical restraint
Patients, No. (%)
Study days, mean (SD)
Delirium, No (%) b
Reintubation within 48 h, No. (%)
Tracheostomy, No (%)
Protocolized Sedation
and Interruption
(n = 214)
7 (4 to 13)
10 (5 to 17)
20 (10 to 36)
50 (23.4)
63 (29.6)
Protocolized Sedation
(n = 209)
7 (3 to 12)
10 (6 to 20)
20 (10 to 48)
52 (24.9)
63 (30.1)
Measure of Effect (95% CI)
HR, 1.08 (0.86 to 1.35)
Mean difference, −3.17 (−6.89 to 0.55)
Mean difference, −8.2 (−17.64 to 1.19)
RR, 0.94 (0.67 to 1.32)
RR, 0.98 (0.73 to 1.31)
P
Value
.52
.36
.42
.72
.89
89 (41.8)
121 (56.8)
50 (23.5)
20 (9.7)
78 (37.3)
130 (62.2)
37 (17.7)
21 (10.2)
RR, 1.12 (0.88 to 1.42)
RR, 0.91 (0.78 to 1.07)
RR, 1.33 (0.91 to 1.94)
RR, 0.94 (0.53 to 1.69)
.35
.26
.14
.84
18 (8.5)
10 (4.7)
29 (13.9)
12 (5.8)
RR, 0.61 (0.35 to 1.07)
RR, 0.82 (0.36 to 1.84)
.08
.64
6 (2.8)
17 (8.0)
13 (6.2)
10 (4.8)
RR, 0.45 (0.17 to 1.17)
RR, 1.68 (0.79 to 3.57)
.09
.18
29 (13.6)
9 (4.2)
33 (15.9)
7 (3.4)
RR, 0.85 (0.54 to 1.35)
RR, 1.25 (0.47 to 3.29)
.53
.64
166 (76.4)
4.71 (5.67)
113 (53.3)
12 (5.6)
166 (79.4)
5.36 (6.14)
113 (54.1)
16 (7.7)
RR, 0.96 (0.87 to 1.07)
Mean difference, −0.70 (−1.84 to 0.43)
RR, 0.98 (0.82 to 1.17)
RR, 0.73 (0.35 to 1.50)
.46
.83
.39
49 (23.2)
54 (26.3)
RR, 0.88 (0.63 to 1.23)
.46
Abbreviations: ARDS, acute respiratory distress syndrome; HR, hazard ratio; ICU, intensive care unit; IQR, interquartile range; RR, relative risk.
a Analyses are measured from enrollment.
b Patients who ever had a score of 4 or more on the Intensive Care Screening Delirium Checklist.19
©2012 American Medical Association. All rights reserved.
JAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
1989
DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS
tients in the interruption group received higher mean daily benzodiazepine doses (102 vs 82 mg/d midazolam
equivalents; median, 8 [IQR, 0-86 vs
median, 0 [IQR, 0-50]; P = .04) and a
greater number of boluses per day
(mean, 0.253 vs 0.177; P = .007). They
also received higher daily opioid doses
(1780 vs 1070 ␮g/d fentanyl equivalents; P⬍.001), both as infusion and
boluses, and a greater number of opioid boluses per day (mean, 2.18 vs 1.79;
P⬍ .001).
Protocol Adherence
and Clinician Workload
Adherence with daily interruption was
72.2% of all eligible study days for an
average patient and 85.6% for all eligible patient-days. Fifty-three percent
of patients missed at least 1 daily interruption, and 6 patients missed every scheduled interruption. The most
common reasons for noninterruption
were related to mechanical ventilation
(38.5%), agitation or pain (16.3%), and
first day of study (14.6%) (eTable 4).
Infusions were reinitiated at a dose exceeding 50% of the previous dose for
30 patients (14.1%) on a total of 47
days. Propofol infusions were administered to 28 patients, accounting for
3.0% of study days.
In the control group, 34 patients
(16.4%) had infusions interrupted on
54 occasions, accounting for 2.3% of
study days. Forty patients receiving
propofol infusions accounted for 2.1%
of study days.
Overall, mean Sedation-Agitation
Scale scores per patient were similar in
the 2 groups (3.28 [95% CI, 2.92 to 3.85]
in the interruption group vs 3.23 in controls; 95% CI, 3.0 to 3.71, respectively;
mean difference, 0.05; 95% CI, −0.10 to
0.19; P = .52). However, nurse workload was significantly higher in the interruption group (mean VAS score, 4.22
vs 3.80; 95% CI, 3.30 to 5.0 vs 2.98 to
4.40; mean difference, 0.41; 95% CI, 0.17
to 0.66; P=.001). Respiratory therapist
workload was similar in the 2 groups
(mean VAS score, 3.69 in the interruption group vs 3.61 in controls; 95% CI,
2.62 to 4.67 vs 2.70 to 4.33, respectively; mean difference, 0.08; 95% CI,
−0.20 to 0.36; P=.57). Adherence with
the performance of spontaneous breathing trials and with extubation after a successful spontaneous breathing trial was
similar in the 2 groups (eTable 5).
Subgroup Analysis
Contrary to our hypothesis, surgical and
trauma patients randomized to daily interruption had significantly shorter time
to successful extubation than those randomized to protocolized sedation alone
(6 vs 13 days; hazard ratio 2.55; 95%
CI, 1.40 to 4.55), whereas there was no
difference among medical patients (9
vs 8 days; hazard ratio, 0.92; 95% CI,
0.72 to 1.18; P value for the interaction=.004). Baseline characteristics and
outcomes of the surgical/trauma patients by randomization group are presented in the supplementary appendix (eTables 6 to 8).
Table 3. Benzodiazepine and Opioid Administration a
Protocolized Sedation
and Interruption
(n = 214)
Midazolam equivalents
Total dose/patient, mg
Dose/patient/d, mg
Dose/patient/d, infusion, mg
Dose/patient/d, bolus, mg
Infusion, d
Boluses/d, No.
Fentanyl equivalents
Total dose/patient, ␮g
Dose/patient/d, ␮g
Dose/patient/d, infusion, ␮g
Dose/patient/d bolus, ␮g
Infusion, d
Boluses/d, No.
1087 (4297)
222 (50 to 734)
102 (326)
8 (0 to 86)
101 (325)
6 (0 to 86)
0.99 (5.9)
0 (0 to 0)
5.73 (6.42)
4 (2 to 7)
0.253 (1.145)
0 (0 to 0)
18 997 (59 928)
5286 (1512 to 16 437)
1780 (4135)
550 (50 to 1850)
1664 (4070)
420 (0 to 1725)
116 (215)
0 (0 to 100)
6.44 (6.86)
5 (2 to 9)
2.18 (2.87)
1 (0 to 4)
Protocolized Sedation
(n = 209)
1038 (4592)
237 (57 to 599)
82 (287)
0 (0 to 50)
82 (287)
0 (0 to 50)
0.49 (2.65)
0 (0 to 0)
5.58 (5.91)
4 (2 to 7)
0.177 (0.808)
0 (0 to 0)
13 532 (23 219)
5936 (2056 to 15 236)
1070 (2066)
260 (0 to 1400)
984 (2002)
80 (0 to 1260)
86 (169)
40 (0 to 150)
6.61 (6.20)
5 (3 to 9)
1.79 (2.67)
0 (0 to 3)
Measure of Effect,
Mean Difference (95% CI)
48.4 (−804.4 to 901.2)
P
Value
.91
19.23 (2.37 to 37.07)
.04
19.22 (1.92 to 36.53)
.03
0.50 (0.23 to 0.76)
⬍.001
0.15 (−1.04 to 1.33)
.81
0.077 (0.020 to 0.134)
.007
5464.6 (−3236.0 to 14 165.2)
.22
709.3 (522.0 to 897.7)
⬍.001
679.7 (495.3 to 864.1)
⬍.001
30.13 (19.15 to 41.11)
⬍.001
−0.17 (−1.42 to 1.09)
.79
0.395 (0.239 to 0.551)
⬍.001
Conversion factors: For conversion of lorazepam to midazolam, 1 mg midazolam=0.5 mg lorazepam. For conversion of opioids to fentanyl equivalents, 10 mg morphine=2 mg hydromorphone=0.1 mg fentanyl.
a Doses are presented as mean (SD) in the first row and median (interquartile range) in the second row.
1990
JAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS
COMMENT
In this multicenter randomized trial, we
found that among mechanically ventilated patients receiving continuous
sedation, the combined use of protocolized sedation and daily sedative
interruption did not improve on the
clinical outcomes observed with protocolized sedation alone. Patients in the
daily interruption group received more
opioids and benzodiazepines, and selfassessed nursing workload was higher
for patients in the daily interruption
group than the control group; however, these findings are of uncertain
clinical importance.
Our results contrast with those of 2
earlier trials supporting daily interruption of sedative infusions in mechanically ventilated adults.5,21 In the original single-center trial comparing daily
interruption with usual care in 128 mechanically ventilated patients receiving sedative and opioid infusions, daily
interruption was associated with shorter
durations of mechanical ventilation and
ICU stay and less neuroimaging.5 In that
trial, research personnel were always
present for sedation interruption and
had decisional authority regarding resumption of infusions. In a 4-center
study, investigators randomized 336
ICU patients receiving ventilation to
daily interruption (with up to 4 hours
of monitoring by research personnel)
or to usual care without a sedation protocol or additional monitoring.21 Patients assigned to daily interruption had
shorter durations of mechanical ventilation and ICU and hospital stay; however, unintentional extubation occurred more frequently.
Our study is distinct from these earlier trials. First, we compared a sedation strategy adding daily interruption to a control group strategy of
protocolized sedation that targeted light
sedation, which is likely superior to
“usual care” of an earlier era. Second,
in this pragmatic trial, sedation was not
directed by research staff but was managed by bedside ICU staff with their
usual patient assignments, according to
well-tested study protocols. Third, the
multicenter design reflects actual prac-
tice in ICUs with variable workloads
and ICU staffing models. Finally, we enrolled surgical patients in addition to
medical patients; in this small prespecified subgroup, daily interruption was
unexpectedly associated with shorter
time to extubation.
The potential benefit of nursedirected sedation titration protocols
to minimize sedation is recognized,6
although early trials testing this strategy have been conflicting. A nursedirected sedation protocol compared
with usual care among 322 medical patients resulted in shorter durations of
mechanical ventilation and ICU and
hospital stay.4 In contrast, no clinical
benefits were found with a different
nurse-directed protocol in another center, potentially related to advancedpractice nurses managing ventilators
who were already routinely using sedation-minimization strategies.7 The effectiveness of any new intervention to
minimize sedation likely depends on
the local usual care.
In this trial, adherence with sedation interruption of 72% compares favorably with that achieved in previous
trials (ranging from 25% to 70%) when
research personnel were not managing patient sedation.16,22,23 Reluctance
to interrupt sedation infusions is expressed clearly in clinician surveys and
practice audits.11,12,24 Common clinical concerns include the potential for
patient discomfort, respiratory distress, patient safety, and additional
workload.13,14,25,26 These reservations
may reflect our unexpected findings of
greater opioid and benzodiazepine
doses, more bolus doses, and greater
nurse workload among patients in the
daily interruption group.
Strengths of this trial, in addition to
the multicenter pragmatic design, include a broad mix of patients and an assessment of perceived additional nursing workload associated with daily
sedation interruption. This trial also has
limitations. Blinding of caregivers was
not feasible, we did not screen for drug
withdrawal, and our results may not be
applicable to patients receiving shorteracting agents such as propofol or dex-
©2012 American Medical Association. All rights reserved.
medetomidine or to patients requiring deeper levels of sedation.
In conclusion, for critically ill patients receiving mechanical ventilation, when nurses implemented a sedation protocol that targeted light
sedation, daily sedation interruption did
not reduce the duration of mechanical
ventilation, offered no additional benefits for patients, and may have increased both sedation and analgesic use
and nurse workload.
Published Online: October 17, 2012. doi:10.1001
/jama.2012.13872
Author Affiliations: Division of Critical Care, Department of Medicine and Interdepartmental Mount Sinai Hospital and University of Toronto, Toronto,
Canada (Dr Mehta); Department of Pharmacy and
Medicine, Mount Sinai Hospital and University of Toronto, Toronto, Canada (Dr Burry); Departments of
Medicine, Clinical Epidemiology and Biostatistics,
McMaster University, Hamilton, Canada (Drs Cook and
Meade); St Joseph’s Healthcare, Hamilton, Canada (Dr
Cook); Clinical Epidemiology Program, Ottawa Hospital Research Institute and Faculty of Medicine, University of Ottawa, Ottawa, Canada (Dr Fergusson);
Mount Sinai Hospital, Toronto (Ms Steinberg); Toronto General Hospital, Division of Respirology, Interdepartmental Division of Critical Care, Faculty of
Medicine, University of Toronto (Dr Granton); Interdepartmental Division of Critical Care and Department of Medicine, University Health Network and University of Toronto (Dr Herridge); Interdepartmental
Division of Critical Care Medicine, and Division of Respirology, Department of Medicine, University Health
Network and Mount Sinai Hospital, University of Toronto (Dr Ferguson); School of Pharmacy, Northeastern University, Boston, Massachusetts (Dr Devlin); Department of Medicine, Long Beach Memorial Medical
Center, Long Beach, California (Dr Tanios); Division
of Critical Care Medicine and Center for Health Evaluation and Outcome Sciences, St. Paul’s Hospital and
University of British Columbia, Vancouver, Canada (Dr
Dodek); Departments of Medicine and Critical Care
Medicine, Sunnybrook Hospital, Toronto (Dr Fowler);
Interdepartmental Division of Critical Care Medicine
(Drs Fowler and Burns) and Institute of Health Policy
Management and Evaluation (Dr Burns), University of
Toronto; Keenan Research Centre and the Li Ka Shing
Knowledge Institute, St Michael’s Hospital, Toronto
(Dr Burns); Departments of Anesthesiology and Critical Care, University of Alberta Hospital, Edmonton,
Canada (Dr Jacka); Section of Critical Care, Department of Medicine, Faculty of Medicine, University of
Manitoba, Winnipeg, Canada (Dr Olafson); Département de Médecine, Soins Intensifs, Hôpital Maisonneuve Rosemont, Université de Montréal, Montréal,
Canada (Dr Skrobik); Department of Critical Care, University of Ottawa, Ottawa, Canada (Dr Hébert); Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa (Ms Sabri); and the
Department of Critical Care, Hamilton Health Sciences, Hamilton, Canada (Dr Meade).
Author Contributions: Dr Mehta had full access to all of
the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Mehta, Burry, Cook,
Fergusson, Steinberg, Dodek, Fowler, Burns, Skrobik,
Hébert, Meade.
Acquisition of data: Mehta, Burry, Cook, Steinberg,
Granton, Herridge, Ferguson, Devlin, Tanios, Dodek,
Fowler, Burns, Jacka, Olafson, Skrobik, Hébert, Sabri,
Meade.
JAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
1991
DAILY SEDATION INTERRUPTION IN MECHANICALLY VENTILATED CRITICALLY ILL PATIENTS
Analysis and interpretation of data: Mehta, Burry,
Cook, Fergusson, Granton, Ferguson, Devlin, Dodek,
Fowler, Burns, Hébert, Sabri, Meade.
Drafting of the manuscript: Mehta, Burry, Cook,
Fergusson, Fowler, Hébert.
Critical revision of the manuscript for important intellectual content: Fergusson, Steinberg, Granton,
Herridge, Ferguson, Devlin, Tanios, Dodek, Fowler,
Burns, Jacka, Olafson, Skrobik, Sabri, Meade.
Statistical analysis: Fergusson, Fowler, Hébert,
Sabri.
Obtained funding: Mehta, Burry, Cook, Fergusson,
Fowler, Meade.
Administrative, technical, or material support: Mehta,
Cook, Steinberg, Herridge, Tanios, Burns, Jacka,
Skrobik.
Study supervision: Mehta, Steinberg, Granton, Devlin,
Tanios, Dodek, Fowler.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of
Potential Conflicts of Interest and none were reported.
Funding/Support: Funding was provided by the Canadian Institutes of Health Research. Dr Cook is a
Canada Research Chair of the Canadian Institutes for
Health Research. Dr Burns holds a Clinician Scientist
Phase 2 Award of the Canadian Institutes for Health
Research. Dr Fowler is a Clinician Scientist of the Heart
and Stroke Foundation (Ontario).
Role of the Sponsor: The study sponsor had no role
in the design of the study; the collection, analysis, or
interpretation of the data; or the writing or approval
of the manuscript.
Online-Only Material: The eAppendix, 2 eFigures,
8 eTables, and Author Audio Interview are available
at http://www.jama.com.
Additional Contributions: We thank the Canadian
Critical Care Trials Group for their collaboration and
the ICU nurses for their dedication to patient care and
their support of this study.
interruption of sedative infusions in an adult medicalsurgical intensive care unit: randomized controlled trial.
J Adv Nurs. 2009;65(5):1054-1060.
10. Yiliaz C, Kelebek Girgin N, Ozdemir N, Kutlay O.
The effect of nursing-implemented sedation on
the duration of mechanical ventilation in the ICU.
Ulus Travma Acil Cerrahi Derg. 2010;16(6):521526.
11. Mehta S, Burry L, Fischer S, et al; Canadian Critical Care Trials Group. Canadian survey of the use of
sedatives, analgesics, and neuromuscular blocking
agents in critically ill patients. Crit Care Med. 2006;
34(2):374-380.
12. Patel RP, Gambrell M, Speroff T, et al. Delirium
and sedation in the intensive care unit: survey of behaviors and attitudes of 1384 healthcare professionals.
Crit Care Med. 2009;37(3):825-832.
13. Tanios MA, de Wit M, Epstein SK, Devlin JW. Perceived barriers to the use of sedation protocols and
daily sedation interruption: a multidisciplinary survey.
J Crit Care. 2009;24(1):66-73.
14. Roberts RJ, de Wit M, Epstein SK, Didomenico
D, Devlin JW. Predictors for daily interruption of sedation therapy by nurses: a prospective, multicenter
study. J Crit Care. 2010;25(4):660.e1-7.
15. Augustes R, Ho KM. Meta-analysis of randomised controlled trials on daily sedation interruption for critically ill adult patients. Anaesth Intensive
Care. 2011;39(3):401-409.
16. Mehta S, Burry L, Martinez-Motta JC, et al; Canadian Critical Care Trials Group. A randomized trial
of daily awakening in critically ill patients managed with
a sedation protocol: a pilot trial. Crit Care Med. 2008;
36(7):2092-2099.
17. Riker RR, Picard JT, Fraser GL. Prospective evaluation of the Sedation-Agitation Scale for adult critically ill patients. Crit Care Med. 1999;27(7):13251329.
18. Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond Agitation-Sedation Scale: validity and reliabil-
ity in adult intensive care unit patients. Am J Respir
Crit Care Med. 2002;166(10):1338-1344.
19. Bergeron N, Dubois M-J, Dumont M, Dial S,
Skrobik Y. Intensive Care Delirium Screening Checklist: evaluation of a new screening tool. Intensive Care
Med. 2001;27(5):859-864.
20. Patanwala AE, Duby J, Waters D, Erstad BL. Opioid conversions in acute care. Ann Pharmacother. 2007;
41(2):255-266.
21. Girard TD, Kress JP, Fuchs BD, et al. Efficacy and
safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive
care (Awakening and Breathing Controlled trial): a randomised controlled trial. Lancet. 2008;371(9607):
126-134.
22. Weisbrodt L, McKinley S, Marshall AP, Cole L,
Seppelt IM, Delaney A. Daily interruption of sedation
in patients receiving mechanical ventilation. Am J Crit
Care. 2011;20(4):e90-e98.
23. Ruokonen E, Parviainen I, Jakob SM, et al; Dexmedetomidine for Continuous Sedation Investigators.
Dexmedetomidine versus propofol/midazolam
for long-term sedation during mechanical ventilation.
Intensive Care Med. 2009;35(2):282-290.
24. Burry L, Perreault M, Williamson D, et al. A prospective evaluation of sedative, analgesic, antipsychotic and paralytic practices in Canadian mechanically ventilated adults [abstract]. Proc Am Thorac
Soc. 2009;179:A5492.
25. O’Connor M, Bucknall T, Manias E. Sedation management in Australian and New Zealand intensive care
units: doctors’ and nurses’ practices and opinions. Am
J Crit Care. 2010;19(3):285-295.
26. Burry L, Steinberg M, Kim S, et al; SLEAP Investigators and Canadian Critical Care Trials Group.
Clinicians’ perspectives on the use of a sedation protocol or daily sedative interruption in mechanically
ventilated patients enrolled in a multicenter sedation trial [abstract]. Intensive Care Med. 2011;
37:S83.
REFERENCES
1. Arroliga AC, Thompson BT, Ancukiewicz M, et al;
Acute Respiratory Distress Syndrome Network. Use of
sedatives, opioids, and neuromuscular blocking agents
in patients with acute lung injury and acute respiratory distress syndrome. Crit Care Med. 2008;36
(4):1083-1088.
2. Kollef MH, Levy NT, Ahrens TS, Schaiff R, Prentice
D, Sherman G. The use of continuous iv sedation is
associated with prolongation of mechanical ventilation.
Chest. 1998;114(2):541-548.
3. Ouimet S, Kavanagh BP, Gottfried SB, Skrobik Y.
Incidence, risk factors and consequences of
ICU delirium. Intensive Care Med. 2007;33(1):
66-73.
4. Brook AD, Ahrens TS, Schaiff R, et al. Effect of a
nursing-implemented sedation protocol on the duration of mechanical ventilation. Crit Care Med. 1999;
27(12):2609-2615.
5. Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily
interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med.
2000;342(20):1471-1477.
6. Jacobi J, Fraser GL, Coursin DB, et al; Task Force
of the American College of Critical Care Medicine
(ACCM) of the Society of Critical Care Medicine
(SCCM), American Society of Health-System Pharmacists (ASHP), American College of Chest Physicians.
Clinical practice guidelines for the sustained use of sedatives and analgesics in the critically ill adult. Crit Care
Med. 2002;30(1):119-141.
7. Bucknall TK, Manias E, Presneill JJ. A randomized
trial of protocol-directed sedation management for mechanical ventilation in an Australian intensive care unit.
Crit Care Med. 2008;36(5):1444-1450.
8. de Wit M, Gennings C, Jenvey WI, Epstein SK. Randomized trial comparing daily interruption of sedation and nursing-implemented sedation algorithm in
medical intensive care unit patients. Crit Care. 2008;
12(3):R70.
9. Anifantaki S, Prinianakis G, Vitsaksaki E, et al. Daily
1992
JAMA, November 21, 2012—Vol 308, No. 19
Corrected on November 27, 2012
©2012 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/ by a University of California - Los Angeles User on 06/19/2013
new england
journal of medicine
The
established in 1812
january 3, 2013
vol. 368 no. 1
Transfusion Strategies for Acute Upper Gastrointestinal
Bleeding
Càndid Villanueva, M.D., Alan Colomo, M.D., Alba Bosch, M.D., Mar Concepción, M.D.,
Virginia Hernandez-Gea, M.D., Carles Aracil, M.D., Isabel Graupera, M.D., María Poca, M.D.,
Cristina Alvarez-Urturi, M.D., Jordi Gordillo, M.D., Carlos Guarner-Argente, M.D., Miquel Santaló, M.D.,
Eduardo Muñiz, M.D., and Carlos Guarner, M.D.
A BS T R AC T
Background
The hemoglobin threshold for transfusion of red cells in patients with acute gastrointestinal bleeding is controversial. We compared the efficacy and safety of a restrictive transfusion strategy with those of a liberal transfusion strategy.
Methods
We enrolled 921 patients with severe acute upper gastrointestinal bleeding and randomly assigned 461 of them to a restrictive strategy (transfusion when the hemoglobin level fell below 7 g per deciliter) and 460 to a liberal strategy (transfusion
when the hemoglobin fell below 9 g per deciliter). Randomization was stratified
according to the presence or absence of liver cirrhosis.
Results
A total of 225 patients assigned to the restrictive strategy (51%), as compared with
61 assigned to the liberal strategy (14%), did not receive transfusions (P<0.001). The
probability of survival at 6 weeks was higher in the restrictive-strategy group than
in the liberal-strategy group (95% vs. 91%; hazard ratio for death with restrictive
strategy, 0.55; 95% confidence interval [CI], 0.33 to 0.92; P = 0.02). Further bleeding
occurred in 10% of the patients in the restrictive-strategy group as compared with
16% of the patients in the liberal-strategy group (P = 0.01), and adverse events occurred in 40% as compared with 48% (P = 0.02). The probability of survival was
slightly higher with the restrictive strategy than with the liberal strategy in the
subgroup of patients who had bleeding associated with a peptic ulcer (hazard ratio,
0.70; 95% CI, 0.26 to 1.25) and was significantly higher in the subgroup of patients
with cirrhosis and Child–Pugh class A or B disease (hazard ratio, 0.30; 95% CI, 0.11 to
0.85), but not in those with cirrhosis and Child–Pugh class C disease (hazard ratio,
1.04; 95% CI, 0.45 to 2.37). Within the first 5 days, the portal-pressure gradient
increased significantly in patients assigned to the liberal strategy (P = 0.03) but not
in those assigned to the restrictive strategy.
From the Gastrointestinal Bleeding Unit,
Department of Gastroenterology (C.V.,
A.C., M.C., V.H.-G., C.A., I.G., M.P.,
C.A.-U., J.G., C.G.-A., C.G.), Blood and
Tissue Bank (A.B., E.M.), and the SemiCritical Unit (M.S.), Hospital de Sant
Pau, Autonomous University, and Centro
de Investigación Biomédica en Red de
Enfermedades Hepáticas y Digestivas
(C.V., A.C., I.G., C.G.) — all in Barcelona.
Address reprint requests to Dr. Villa­
nueva at Servei de Patologia Digestiva,
Hospital de la Santa Creu i Sant Pau, Mas
Casanovas, 90. 08025 Barcelona, Spain,
or at [email protected].
This article was updated on May 16, 2013,
at NEJM.org.
N Engl J Med 2013;368:11-21.
DOI: 10.1056/NEJMoa1211801
Copyright © 2013 Massachusetts Medical Society.
Conclusions
As compared with a liberal transfusion strategy, a restrictive strategy significantly
improved outcomes in patients with acute upper gastrointestinal bleeding. (Funded
by Fundació Investigació Sant Pau; ClinicalTrials.gov number, NCT00414713.)
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
11
The
n e w e ng l a n d j o u r na l
A cute upper gastrointestinal bleeding is a common emergency condition associated with high morbidity and mortality.1 It is a frequent indication for red-cell
transfusion, because acute blood loss can decrease
tissue perfusion and the delivery of oxygen to tissues. Transfusion may be lifesaving in patients
with massive exsanguinating bleeding. However,
in most cases hemorrhage is not so severe, and in
such circumstances the safest and most effective
transfusion strategy is controversial.2,3
Restricted transfusion strategies may be appropriate in some settings. Controlled trials have
shown that for critically ill patients, a restrictive
transfusion strategy is at least as effective as a
liberal strategy, while substantially reducing the
use of blood supplies.4,5 However, these studies
excluded patients with gastrointestinal bleeding.
Observational studies and small controlled trials
have suggested that transfusion may be harmful
in patients with hypovolemic anemia,6,7 even in
those with gastrointestinal bleeding.8-12 Furthermore, studies in animals suggest that transfusion can be particularly harmful in patients with
bleeding from portal hypertensive sources, since
restitution of blood volume after hemorrhage can
lead to a rebound increase in portal pressure,
which is associated with a risk of rebleeding.12-14
We performed a randomized, controlled trial
in which we assessed whether a restrictive threshold for red-cell transfusion in patients with acute
gastrointestinal bleeding was safer and more effective than a liberal transfusion strategy that was
based on the threshold recommended in guidelines at the time the study was designed.15,16
Me thods
Study Oversight
From June 2003 through December 2009, we consecutively enrolled patients with gastrointestinal
bleeding who were admitted to Hospital de la
Santa Creu i Sant Pau in Barcelona. Written informed consent was obtained from all the patients or their next of kin, and the trial was approved by the institutional ethics committee at
the hospital. The protocol, including the statistical analysis plan, is available with the full text of
this article at NEJM.org. No commercial support
was involved in the study. All the authors vouch
for the integrity and the accuracy of the analysis
12
of
m e dic i n e
and for the fidelity of the study to the protocol.
No one who is not an author contributed to the
manuscript.
Selection of Patients
Patients older than 18 years of age who had hematemesis (or bloody nasogastric aspirate), melena, or both, as confirmed by the hospital staff,
were considered for inclusion. Patients were excluded if they declined to undergo a blood transfusion. Additional exclusion criteria were massive
exsanguinating bleeding; an acute coronary syndrome, symptomatic peripheral vasculopathy,
stroke, transient ischemic attack, or transfusion
within the previous 90 days; a recent history of
trauma or surgery; lower gastrointestinal bleeding; a previous decision on the part of the attending physician that the patient should avoid specific medical therapy; and a clinical Rockall score
of 0 with a hemoglobin level higher than 12 g per
deciliter. The Rockall score is a system for assessing the risk of further bleeding or death
among patients with gastrointestinal bleeding;
scores range from 0 to 11, with a score of 2 or
lower indicating low risk and scores of 3 to 11
indicating increasingly greater risk.
Study Design
Immediately after admission, patients were randomly assigned to a restrictive transfusion strategy
or a liberal transfusion strategy. Randomization
was performed with the use of computer-generated
random numbers, with the group assignments
placed in sealed, consecutively numbered, opaque
envelopes. Randomization was stratified according to the presence or absence of liver cirrhosis
and was performed in blocks of four. Cirrhosis was
diagnosed according to clinical, biochemical, and
ultrasonographic findings.
In the restrictive-strategy group, the hemoglobin threshold for transfusion was 7 g per
deciliter, with a target range for the post-transfusion hemoglobin level of 7 to 9 g per deciliter.
In the liberal-strategy group, the hemoglobin
threshold for transfusion was 9 g per deciliter,
with a target range for the post-transfusion hemoglobin level of 9 to 11 g per deciliter. In both
groups, 1 unit of red cells was transfused initially; the hemoglobin level was assessed after
the transfusion, and an additional unit was
transfused if the hemoglobin level was below the
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Tr ansfusion Str ategies for Upper GI Bleeding
threshold value. The transfusion protocol was applied until the patient’s discharge from the hospital or death. The protocol allowed for a transfusion to be administered any time symptoms or
signs related to anemia developed, massive bleeding occurred during follow-up, or surgical intervention was required. Only prestorage leukocytereduced units of packed red cells were used for
transfusion. The volume of a unit ranged from
250 to 320 ml, with a hematocrit of approximately 60%.
Hemoglobin levels were measured after admission and again every 8 hours during the first
2 days and every day thereafter. Hemoglobin
levels were also assessed when further bleeding
was suspected.
Further bleeding was defined as hematemesis or
fresh melena associated with hemodynamic instability (systolic blood pressure of <100 mm Hg;
pulse rate of >100 beats per minute, or both) or a
fall in hemoglobin level of 2 g per deciliter or more
within a 6-hour period. Further bleeding was
considered to indicate therapeutic failure; if the
bleeding involved nonvariceal lesions, the patient
underwent repeat endoscopic therapy or emergency surgery, whereas in the case of further
variceal bleeding, transjugular intrahepatic portosystemic shunting (TIPS) was considered.
Complications were defined as any untoward
events that necessitated active therapy or prolonged hospitalization. Side effects were considered to be severe if the health or safety of the
patient was endangered.
Treatments and Follow-up
All the patients underwent emergency gastroscopy within the first 6 hours. When endoscopic
examination disclosed a nonvariceal lesion with
active arterial bleeding, a nonbleeding visible vessel, or an adherent clot, patients underwent endoscopic therapy with injection of adrenaline plus
multipolar electrocoagulation or application of endoscopic clips. Patients with peptic ulcer received
a continuous intravenous infusion of omeprazole
(80 mg per 10-hour period after an initial bolus
of 80 mg) for the first 72 hours, followed by oral
administration of omeprazole.
When portal hypertension was suspected, a
continuous intravenous infusion of somatostatin
(250 μg per hour) and prophylactic antibiotic
therapy with norfloxacin or ceftriaxone were administered at the time of admission and continued for 5 days. Bleeding esophageal varices were
also treated with band ligation or with sclerotherapy, and gastric varices with injection of cyanoacrylate. In patients with variceal bleeding,
portal pressure was measured within the first
48 hours and again 2 to 3 days later to assess
the effect of the transfusion strategy on portal
hypertension. Portal pressure was estimated with
the use of the hepatic venous pressure gradient
(HVPG), as described elsewhere.17
Outcome Measures and Definitions
The primary outcome measure was the rate of
death from any cause within the first 45 days.
Secondary outcomes included the rate of further
bleeding and the rate of in-hospital complications.
Statistical Analysis
We estimated that with 430 patients in each
group, the study would have the power to detect
a between-group difference in mortality of at
least 5 percentage points, assuming 10% mortality in the liberal-strategy group (on the basis of
results of previous trials with standard care1,3,18),
with the use of a two-tailed test and with alpha
and beta values of 0.05 and 0.2, respectively. The
statistical analysis was performed according to the
intention-to-treat principle. Standard tests were
used for comparisons of proportions and means.
Continuous variables are expressed as means and
standard deviations. Actuarial probabilities were
calculated with the use of the Kaplan–Meier
method and were compared with the use of the
log-rank test. A Cox proportional-hazards regression model was used to compare the two
transfusion-strategy groups with respect to the
primary and secondary end points, with adjustment for baseline risk factors (see the Supplementary Appendix, available at NEJM.org). The
hazard ratios and their 95% confidence intervals
were calculated. Data were censored at the time
an end-point event occurred, at the patient’s last
visit, or at the end of the 45-day follow-up period,
whichever occurred first. Prespecified subgroup
analyses were performed to assess the efficacy of
transfusion strategies according to the source of
bleeding (lesions related to portal hypertension
or peptic ulcer). All P values are two-tailed. Calculations were performed with the use of the
SPSS statistical package, version 15.0 (SPSS).
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
13
The
n e w e ng l a n d j o u r na l
m e dic i n e
Mortality
R e sult s
Study Patients
During the study period, 2372 patients were admitted to the hospital for gastrointestinal bleeding and 1610 were screened. Of these, 41 declined
to participate and 648 were excluded; among the
reasons for exclusion were exsanguinating bleeding requiring transfusion (in 39 patients) and a
low risk of rebleeding (329 patients) (Fig. 1). A
total of 921 patients underwent randomization and
32 withdrew or were withdrawn by the investigators after randomization (see Fig. 1 for details),
leaving 444 patients in the restrictive-strategy
group and 445 in the liberal-strategy group for the
intention-to-treat analysis. The baseline characteristics were similar in the two groups (Table 1).
A total of 277 patients (31%) had cirrhosis, and
the baseline characteristics of the patients in this
subgroup were similar in the two transfusionstrategy groups (Table 1). Bleeding was due to
peptic ulcer in 437 patients (49%) and to esophageal varices in 190 (21%) (Table 1).
Hemoglobin Levels and Transfusion
The hemoglobin concentration at admission was
similar in the two groups (Table 2). The lowest hemoglobin concentration within the first 24 hours
was significantly lower in the restrictive-strategy
group than in the liberal-strategy group, as was
the daily hemoglobin concentration until discharge
(P<0.001). The percentage of patients in whom
the lowest hemoglobin level was less than 7 g per
deciliter was higher in the restrictive-strategy
group than in the liberal-strategy group. The hemoglobin concentration at 45 days was similar in
the two groups.
A total of 225 patients (51%) in the restrictivestrategy group, as compared with 61 patients
(14%) in the liberal-strategy group, received no
transfusion (P<0.001). The mean (±SD) number
of units transfused was significantly lower in
the restrictive-strategy group than in the liberalstrategy group (1.5±2.3 vs. 3.7±3.8, P<0.001), and
a violation of the transfusion protocol occurred
more frequently in the restrictive-strategy group (in
39 patients [9%] vs. 15 patients [3%], P<0.001)
(Table 2). The percentage of patients who received a transfusion of fresh-frozen plasma, the
percentage of patients who received a transfusion of platelets, and the total amount of fluid
administered were similar in the two groups.
14
of
Mortality at 45 days was significantly lower in
the restrictive-strategy group than in the liberalstrategy group: 5% (23 patients) as compared with
9% (41 patients) (P = 0.02) (Fig. 2). The risk of
death was virtually unchanged after adjustment
for baseline risk factors for death (hazard ratio
with restrictive strategy, 0.55; 95% confidence
interval [CI], 0.33 to 0.92) (Table S4 in the Supplementary Appendix). Among all patients with
cirrhosis, the risk of death was slightly lower in
the restrictive-strategy group than in the liberalstrategy group (Fig. 2). In the subgroup of patients with cirrhosis and Child–Pugh class A or B
disease, the risk of death was significantly lower
among patients in the restrictive-strategy group
than among those in the liberal-strategy group,
whereas in the subgroup of patients with cirrhosis and Child–Pugh class C disease, the risk was
similar in the two groups. Among patients with
bleeding from a peptic ulcer, the risk of death was
slightly lower with the restrictive strategy than
with the liberal strategy.
Death was due to unsuccessfully controlled
bleeding in 3 patients (0.7%) in the restrictivestrategy group and in 14 patients (3.1%) in the
liberal-strategy group (P = 0.01). Death was caused
by complications of treatment in 3 patients (2 in
the liberal-strategy group and 1 in the restrictivestrategy group). In the remaining 44 patients (19
in the restrictive-strategy group and 25 in the
liberal-strategy group), hemorrhage was controlled
and death was due to associated diseases.
Further Bleeding
The rate of further bleeding was significantly
lower in the restrictive-strategy group than in the
liberal-strategy group: 10% (45 patients), as compared with 16% (71 patients) (P = 0.01) (Table 3).
The risk of further bleeding was significantly
lower with the restrictive strategy after adjustment for baseline risk factors for further bleeding (hazard ratio, 0.68; 95% CI, 0.47 to 0.98)
(Table S4 in the Supplementary Appendix). In addition, the length of hospital stay was shorter in
the restrictive-strategy group than in the liberalstrategy group.
In the subgroup of patients with cirrhosis, the
risk of further bleeding was lower with the restrictive transfusion strategy than with the liberal transfusion strategy among patients with
Child–Pugh class A or B disease and was similar
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Tr ansfusion Str ategies for Upper GI Bleeding
in the two groups among patients with Child–
Pugh class C disease (Table 3). Among patients
with bleeding from esophageal varices, the rate
of further bleeding was lower in the restrictivestrategy group than in the liberal-strategy group
(11% vs. 22%, P = 0.05). Rescue therapy with balloon tamponade or with transjugular intrahepatic
portosystemic shunt was required less frequently
in the restrictive-strategy group than in the liberalstrategy group.
A baseline hepatic hemodynamic study was
performed in 86 patients in the restrictive-strategy group and in 89 in the liberal-strategy
group, and it was repeated 2 to 3 days later in 74
and 77 patients, respectively, to assess changes.
Patients in the liberal-strategy group had a significant increase in the mean hepatic venous
pressure gradient between the first hemodynam­
ic study and the second (from 20.5±3.1 mm Hg
to 21.4±4.3 mm Hg, P = 0.03). There was no significant change in mean hepatic venous pressure
gradient in the restrictive-strategy group during
that interval.
Among patients with bleeding from a peptic
ulcer, there was a trend toward a lower risk of
further bleeding in the restrictive-strategy group
(Table 3). Emergency surgery to control further
bleeding was required less frequently in the
restrictive-strategy group than in the liberalstrategy group (2% vs. 6%, P = 0.04).
Adverse Events
1610 Patients were screened
648 Were excluded
962 Were eligible
41 Declined to participate
921 Underwent randomization
461 Were assigned to restrictive
strategy
460 Were assigned to liberal
strategy
17 Withdrew
444 Were included in analysis
15 Withdrew
445 Were included in analysis
Figure 1. Screening, Randomization, and Follow-up.
During the study period, 1610 patients with gastrointestinal bleeding were
screened, and 648 patients were excluded. The reasons for exclusion included massive exsanguinating bleeding requiring transfusion before randomization (39 patients) and a low risk of rebleeding (329 patients). A low
risk of rebleeding was defined as a clinical Rockall score of 0 and hemoglobin
levels higher than 12 g per deciliter. (The Rockall score is a system for assessing the risk of further bleeding or death among patients with gastrointestinal bleeding; scores range from 0 to 11, with higher scores indicating
greater risk.) Patients were also excluded if they declined blood transfusion
(14 patients); other exclusion criteria were an acute coronary syndrome
(58), symptomatic peripheral vasculopathy (12), stroke or transient ischemic
attack (7), or transfusion (10) within the previous 90 days; lower gastrointestinal bleeding (51); pregnancy (3); a recent history of trauma or surgery
(41); a decision by the attending physician that the patient should avoid
medical therapy (9); or inclusion in this study within the previous 90 days
or inclusion more than twice (75). A total of 921 patients underwent randomization, of whom 32 were withdrawn: 23 were found to be ineligible, 5 had
major protocol violations, and 4 decided to withdraw from the study.
The overall rate of complications was significantly
lower in the restrictive-strategy group than in the
liberal-strategy group (40% [179 patients] vs. 48%
[214 patients], P = 0.02), as was the rate of serious
adverse events (Table S5 in the Supplementary
Appendix). Transfusion reactions and cardiac
events, mainly pulmonary edema, occurred more
frequently in the liberal-strategy group (Table 3).
The rates of other adverse events, such as acute
kidney injury or bacterial infections, did not differ significantly between the groups (Table S5 in
a liberal transfusion strategy, in which the hemothe Supplementary Appendix).
globin threshold was 9 g per deciliter. The most
relevant finding was the improvement in survival
Discussion
rates observed with the restrictive transfusion
We found that among patients with severe acute strategy. This advantage was probably related to
upper gastrointestinal bleeding, the outcomes a better control of factors contributing to death,
were significantly improved with a restrictive such as further bleeding, the need for rescue
transfusion strategy, in which the hemoglobin therapy, and serious adverse events. All these facthreshold was 7 g per deciliter, as compared with tors were significantly reduced with the restricn engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
15
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
Table 1. Baseline Characteristics of the Patients.*
Restrictive Strategy
(N = 444)
Characteristic
Liberal Strategy
(N = 445)
P Value
In-hospital bleeding — no. (%)†
20 (5)
30 (7)
0.19
Rockall score‡
5.3±2.0
5.4±1.7
0.18
228/444 (51)
209/445 (47)
0.20
Source of bleeding — no./total no. (%)
Peptic ulcer
Location
0.95
Gastric
Duodenal
Stomal
76/228 (33)
71/209 (34)
143/228 (63)
131/209 (63)
9/228 (4)
7/209 (3)
Stigmata
0.93
Active bleeding
Visible vessel
Gastroesophageal varices
Mallory–Weiss tears
35/228 (15)
33/209 (16)
127/228 (56)
119/209 (57)
101/444 (23)
109/445 (24)
0.58
25/444 (6)
30/445 (7)
0.49
Erosive gastritis or esophagitis
38/444 (9)
29/445 (7)
0.26
Neoplasms
16/444 (4)
20/445 (4)
0.50
Other
36/444 (8)
48/445 (11)
Cirrhosis — no. (%)
Alcoholic cause — no./total no. (%)
139 (31)
138 (31)
0.94
63/139 (45)
62/138 (45)
0.49
37/139 (27)
30/138 (22)
Child–Pugh class — no./total no. (%)§
A
0.57
B
76/139 (55)
79/138 (57)
C
26/139 (19)
29/138 (21)
20.1±4.4
20.6±5.2
HVPG — mm Hg¶
0.61
Causes of bleeding — no./total no. (%)
93/139 (67)
97/138 (70)
0.60
Gastric varices
Esophageal varices
8/139 (6)
12/138 (9)
0.36
Peptic lesions
21/139 (15)
18/138 (13)
0.73
*Plus–minus values are means ±SD.
†Among patients with in-hospital bleeding, 16 (7 in the restrictive-strategy group and 9 in the liberal-strategy group)
were admitted to the intensive care unit with sepsis or for pressure support.
‡The Rockall score is a system for assessing the risk of further bleeding or death among patients with gastrointestinal
bleeding; scores range from 0 to 11, with higher scores indicating higher risk.
§ Child–Pugh class A denotes good hepatic function, class B intermediate function, and class C poor function. The mean
Model for End-Stage Liver Disease (MELD) score among patients in all Child–Pugh classes (on a scale from 6 to 40,
with higher values indicating more severe liver disease) was 11.9±7 in the restrictive-strategy group and 12.1±6 in the
liberal-strategy group (P = 0.95).
¶Portal pressure was measured with the use of the hepatic venous pressure gradient (HVPG), which is the difference
between the wedged and free hepatic venous pressures. Measurements were performed within the first 48 hours in
175 patients with variceal bleeding (86 in the restrictive-strategy group and 89 in the liberal-strategy group).
tive strategy. Our results are consistent with
those from previous observational studies and
randomized trials performed in other settings,
which have shown that a restrictive transfusion
strategy did not increase,5,19 and even de-
16
creased,4,20 the mortality observed with a liberal
transfusion strategy.
Current international guidelines recommend
decreasing the hemoglobin threshold level for
transfusion in patients with gastrointestinal
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Tr ansfusion Str ategies for Upper GI Bleeding
Table 2. Hemoglobin Levels, Transfusions, and Cointerventions.*
Restrictive Strategy
(N = 444)
Liberal Strategy
(N = 445)
At admission
9.6.±2.2
9.4±2.4
0.45
Lowest value during hospital stay
7.3±1.4
8.0±1.5
<0.001
9.2±1.2
10.1±1.0
<0.001
11.6±1.7
11.7±1.8
0.67
Variable
P Value
Hemoglobin level — g/dl
At discharge†
At day 45
Patients with lowest hemoglobin <7 g/dl — no. (%)
202 (45)
81 (18)
<0.001
Patients with lowest hemoglobin >9 g/dl — no. (%)
55 (12)
67 (15)
0.28
219 (49)
384 (86)
<0.001
671
1638
<0.001
1.5±2.3
3.7±3.8
<0.001
0
3
<0.001
Red-cell transfusion
Any — no. of patients (%)
Units transfused — no.
Total‡
Mean/patient
Median
Range
0–19
0–36
1.2±1.8
2.9±2.2
<0.001
Transfusion not adjusted to hemoglobin level —
no. of patients (%)¶
35 (8)
12 (3)
0.001
Major protocol violation — no. of patients (%)‖ ║
39 (9)
15 (3)
<0.001
15
15
During index bleeding§
Duration of storage of red cells — days**
Median
Range
0.95
1–40
1–42
Fresh-frozen plasma transfusion — no. of patients (%)††
28 (6)
41 (9)
0.13
Platelet transfusion — no. of patients (%)‡‡
12 (3)
19 (4)
0.27
5491±3448
5873±4087
0.19
86 (19)
93 (21)
0.62
Crystalloids administered within first 72 hr — ml
Receipt of colloids — no. of patients (%)
* Plus–minus values are means ±SD.
† The average difference in the daily hemoglobin level between the restrictive-strategy group and the liberal-strategy group
was 1.0±1.3 g per deciliter, from the time of admission to discharge.
‡ Included are all red-cell transfusions received from the time of admission to discharge.
§ This category refers to the units of red cells transfused before further bleeding.
¶ Transfusions were administered in 31 patients (26 in the restrictive-strategy group and 5 in the liberal-strategy group)
because of symptoms or signs (defined as tachycardia, chest pain, or signs of severe hypoxemia) in 14 patients (8 in the
restrictive-strategy group and 6 in the liberal-strategy group) because of massive bleeding, and in 2 patients (1 in each
group) because of surgery.
‖ In the restrictive-strategy group, 39 patients without signs or symptoms, massive bleeding, or surgery received a
transfusion when the hemoglobin level was higher than 7 g per deciliter. In the liberal-strategy group, 15 patients with
a hemoglobin level lower than 9 g per deciliter did not receive a transfusion.
** Red cells were stored for up to 42 days. At least 1 unit stored for more than 14 days was administered in 141 of the
219 patients in the restrictive-strategy group (64%) and 253 of the 384 patients in the liberal-strategy group (66%)
who received a transfusion.
†† Included are all patients who received a transfusion of fresh-frozen plasma from the time of admission to discharge.
‡‡ Included are all patients who received a transfusion of platelets from the time of admission to discharge.
bleeding, from 10 g per deciliter15,16 to 7 g per
deciliter.3,21 A reduction in the number of transfusions performed may have accounted for the
reduction in mortality from gastrointestinal bleed-
ing that has been observed in recent years.22,23
However, current guidelines are based on findings from trials of transfusion triggers involving
critically ill patients with normovolemic anemia
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
17
The
n e w e ng l a n d j o u r na l
of
m e dic i n e
A Survival, According to Transfusion Strategy
100
90
100
Overall Survival (%)
80
99
98
70
97
60
96
95
50
40
93
P=0.02 by log-rank test
92
30
91
20
Liberal strategy
90
0
10
0
Restrictive strategy
94
0
0
5
5
10
15
10
20
15
25
20
30
35
40
45
25
30
35
40
45
399
386
397
383
395
378
394
375
392
372
Days
No. at Risk
Restrictive strategy
Liberal strategy
444
445
429
428
412
407
404
397
401
393
B Death by 6 Weeks, According to Subgroup
Subgroup
Restrictive
Strategy
Liberal
Strategy
Hazard Ratio (95% CI)
P Value
no. of patients/total no. (%)
Overall
Patients with cirrhosis
Child–Pugh class A or B
Child–Pugh class C
Bleeding from varices
Bleeding from peptic ulcer
23/444 (5)
15/139 (11)
5/113 (4)
10/26 (38)
10/93 (11)
7/228 (3)
0.55 (0.33–0.92)
0.57 (0.30–1.08)
0.30 (0.11–0.85)
1.04 (0.45–2.37)
0.58 (0.27–1.27)
0.70 (0.26–1.25)
41/445 (9)
25/138 (18)
13/109 (12)
12/29 (41)
17/97 (18)
11/209 (5)
0.1
1.0
0.02
0.08
0.02
0.91
0.18
0.26
10.0
Restrictive Strategy Liberal Strategy
Better
Better
Figure 2. Rate of Survival, According to Subgroup.
Panel A shows the Kaplan–Meier estimates of the 6-week survival rate in the two groups. The probability of survival
was significantly higher in the restrictive-strategy group than in the liberal-strategy group. The gray arrows indicate
the day on which data from a patient were censored. The inset shows the same data on an enlarged y axis. Panel B
shows the hazard ratios, with 95% confidence intervals, for death by 6 weeks, according to prespecified subgroups.
In the subgroup of patients with Child–Pugh class A or B disease, the Model for End-Stage Liver Disease (MELD)
score (on a scale from 6 to 40, with higher values indicating more severe liver disease) was 10.3±5 in the restrictivestrategy group and 10.9±5 in the liberal-strategy group (P = 0.41). In the subgroup of patients with Child–Pugh class C
disease, the MELD score was 20.6±6 in the restrictive-strategy group and 18.1±5 in the liberal-strategy group (P = 0.11).
— trials from which patients with acute bleeding have been excluded.4,5 Transfusion requirements may be different for patients with acute
hemorrhage due to factors such as hemodynamic instability or rapid onset of anemia to
extremely low hemoglobin levels. The current
study addressed the effects of transfusion in this
setting. Previous observational studies and small
18
controlled trials have supported the use of a restrictive transfusion strategy for patients with
gastrointestinal bleeding.8-11 Our results, which
are consistent with the results from those studies, showed that a restrictive strategy significantly
reduced the rates of factors related to therapeutic
failure such as further bleeding and the need for
rescue therapy, as well as reducing the length of
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Tr ansfusion Str ategies for Upper GI Bleeding
Table 3. Study Outcomes.*
Restrictive Strategy
(N = 444)
Liberal Strategy
(N = 445)
Hazard Ratio with
Restrictive Strategy
(95% CI)
P Value
23 (5)
41 (9)
0.55 (0.33–0.92)
0.02
Overall
45/444 (10)
71/445 (16)
0.62 (0.43–0.91)
0.01
Patients with cirrhosis
16/139 (12)
31/138 (22)
0.49 (0.27–0.90)
0.02
12/113 (11)
23/109 (21)
0.53 (0.27–0.94)
0.04
4/26 (15)
8/29 (28)
0.58 (0.15–1.95)
0.33
10/93 (11)
21/97 (22)
0.50 (0.23–0.99)
0.05
Outcome
Death from any cause within 45 days — no. (%)
Further bleeding — no. of patients/total no. (%)
Child–Pugh class A or B
Child–Pugh class C
Bleeding from esophageal varices
Rescue therapies
Balloon tamponade
3/139 (2)
11/138 (8)
0.03
TIPS
6/139 (4)
15/138 (11)
0.04
23/228 (10)
33/209 (16)
20/228 (9)
26/209 (12)
4/228 (2)
12/209 (6)
0.04
9.6±8.7
11.5±12.8
0.01
179 (40)
214 (48)
14 (3)
38 (9)
0.35 (0.19–0.65)
0.001
12 (3)
16 (4)
0.74 (0.35–1.59)
0.56
Patients with bleeding from peptic ulcer
0.63 (0.37–1.07)
0.09
Rescue therapies
Second endoscopic therapy
Emergency surgery
No. of days in hospital
0.21
Adverse events — no. (%)†
Any‡
Transfusion reactions
Fever
0.73 (0.56–0.95)
0.02
Transfusion-associated circulatory overload
2 (<1)
16 (4)
0.06 (0.01–0.45)
0.001
Allergic reactions
1 (<1)
6 (1)
0.16 (0.02–1.37)
0.12
Cardiac complications§
49 (11)
70 (16)
0.64 (0.43–0.97)
0.04
8 (2)
13 (3)
0.61 (0.25–0.49)
0.27
12 (3)
21 (5)
0.56 (0.27–1.12)
0.07
Acute coronary syndrome¶
Pulmonary edema
Pulmonary complications
48 (11)
53 (12)
0.89 (0.59–1.36)
0.67
Acute kidney injury
78 (18)
97 (22)
0.78 (0.56–1.08)
0.13
3 (1)
6 (1)
0.49 (0.12–2.01)
0.33
119 (27)
135 (30)
0.87 (0.63–1.21)
0.41
Stroke or transient ischemic attack
Bacterial infections
*Plus–minus values are means ±SD. TIPS denotes transjugular intrahepatic portosystemic shunt.
†Patients may have had more than one type of adverse event.
‡Included are all patients who had at least one adverse event during the study period.
§ This category includes patients with acute coronary syndrome, pulmonary edema, or arrhythmias.
¶Unstable angina developed in 13 patients (8 in the restrictive-strategy group and 5 in the liberal-strategy group), and myocardial infarction
occurred in 8 patients (all in the liberal-strategy group).
stay in the hospital. These harmful effects of transfusion may be related to an impairment of hemostasis. Transfusion may counteract the splanchnic
vasoconstrictive response caused by hypovolemia,
inducing an increase in splanchnic blood flow
and pressure that may impair the formation of
clots.24,25 Transfusion may also induce abnormalities in coagulation properties.8,10
Concerns about transfusion have been raised
primarily with respect to patients who have cirrhosis with portal hypertension. Experimental
studies have shown that restitution of blood
volume can induce rebound increases in portal
pressure that may precipitate portal hypertensiverelated bleeding.12-14 Clinical studies have also
shown that transfusion increases portal pressure
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
19
The
n e w e ng l a n d j o u r na l
during acute variceal bleeding, an increase that
may be prevented with somatostatin.17 In keeping with these observations, we found that the
beneficial effect of a restrictive transfusion strategy with respect to further bleeding was observed mainly in patients with portal hypertension. We also observed that despite treatment with
somatostatin, patients in the liberal-strategy
group had a significant increase in portal pressure during acute variceal bleeding that was not
observed in patients in the restrictive-strategy
group. This may have accounted for the higher
rate of further bleeding with the liberal strategy.
We found a reduction in the rate of complications with the restrictive transfusion strategy. This
finding is consistent with results from a previous
trial involving critically ill adults.4 However, conflicting results have been shown in other settings.5,19 Several factors, such as coexisting conditions or age, may account for this discrepancy.
Cardiac complications, particularly pulmonary
edema, occurred more frequently with the liberal
transfusion strategy, both in the current study
and in the trial that involved critically ill adults.4
The higher level of cardiac complications may
indicate a higher risk of circulatory overload associated with a liberal transfusion strategy. Other
effects of transfusion, such as transfusion-related
immunomodulation,26 may increase the risk of
complications or death. These are unlikely to have
occurred in the current study given the similar
incidence of bacterial infections in the two groups
and the universal use of prestorage leukocytereduced red cells. Adverse outcomes have also
been associated with long storage time of transfused blood.27 In our study, the storage time was
similar in the two groups. However, the median
duration of storage was 15 days, and storage lesions become apparent after about 14 days.28
Therefore, the fact that there were more transfusions of blood with these long storage times in
the liberal-strategy group may have contributed
to the worse outcome. Further research is needed to determine whether the use of newer blood
may influence the results with respect to the transfusion strategy. We found that a restrictive transfusion strategy significantly decreased the number of units transfused and the percentage of
patients who received no transfusions — findings that were also seen in previous trials.4,5,19
The goal of red-cell transfusions is to improve
20
of
m e dic i n e
the delivery of oxygen to tissues. The safest and
most effective transfusion strategy depends not
only on the hemoglobin trigger level but also on
factors such as coexisting conditions, age, and
hemodynamic status.1,3 Consequently, we allowed
transfusions to be performed at the discretion of
attending physicians when symptoms related to
anemia developed, when massive bleeding occurred, or when surgical intervention was required. Transfusions that were not adjusted to the
hemoglobin level and violations of the transfusion
protocol occurred more often in the restrictivestrategy group than in the liberal-strategy group.
However, both these deviations from the protocol occurred in less than 10% of cases.
Our trial has several limitations. First, the results cannot be generalized to all patients with
acute gastrointestinal bleeding. Patients with a low
risk of rebleeding were not included in this study.
However, these patients are less likely to require
a transfusion. Patients with massive exsanguinating hemorrhage were also excluded from this
trial because red-cell transfusion may be lifesaving for them. However, only a minority of eligible
patients were excluded for this reason. Second,
because we compared two transfusion strategies,
the study was not blinded, and this may have
introduced a bias. It is unlikely that bias was introduced, however, owing to the objective definition of the primary outcome and the use of a
randomized design with concealed assignments.
In summary, we found that a restrictive transfusion strategy, as compared with a liberal transfusion strategy, improved the outcomes among
patients with acute upper gastrointestinal bleeding. The risk of further bleeding, the need for
rescue therapy, and the rate of complications
were all significantly reduced, and the rate of
survival was increased, with the restrictive transfusion strategy. Our results suggest that in patients with acute gastrointestinal bleeding, a
strategy of not performing transfusion until the
hemoglobin concentration falls below 7 g per
deciliter is a safe and effective approach.
Supported in part by the Fundació Investigació Sant Pau.
Dr. Guarner reports receiving consulting fees from Sequana
Medical. No other potential conflict of interest relevant to this
article was reported.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
We thank the nursing and medical staffs from the Semi-Critical Unit at the Hospital de la Santa Creu i Sant Pau for their
cooperation in this study.
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Tr ansfusion Str ategies for Upper GI Bleeding
References
1. Gralnek IM, Barkun AN, Bardou M.
Management of acute bleeding from a peptic ulcer. N Engl J Med 2008;359:928-37.
2. Barkun A, Bardou M, Marshall JK.
Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med
2003;139:843-57.
3. Barkun AN, Bardou M, Kuipers EJ, et
al. International consensus recommendations on the management of patients with
nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2010;152:101-13.
4. Hébert PC, Wells G, Blajchman MA, et
al. A multicenter, randomized, controlled
clinical trial of transfusion requirements
in critical care. N Engl J Med 1999;340:40917. [Erratum, N Engl J Med 1999;340:1056.]
5. Lacroix J, Hébert PC, Hutchison JS, et
al. Transfusion strategies for patients in
pediatric intensive care units. N Engl J
Med 2007;356:1609-19.
6. Malone DL, Dunne J, Tracy JK, Putnam AT, Scalea TM, Napolitano LM.
Blood transfusion, independent of shock
severity, is associated with worse outcome
in trauma. J Trauma 2003;54:898-905.
7. Robinson WP III, Ahn J, Stiffler A, et
al. Blood transfusion is an independent
predictor of increased mortality in nonoperatively managed blunt hepatic and
splenic injuries. J Trauma 2005;58:437-44.
8. Blair SD, Janvrin SB, McCollum CN,
Greenhalgh RM. Effect of early blood
transfusion on gastrointestinal haemorrhage. Br J Surg 1986;73:783-5.
9. Villarejo F, Rizzolo M, Lopéz E, Domeniconi G, Arto G, Apezteguia C. Acute
anemia in high digestive hemorrhage:
margins of security for their handling
without transfusion of red globules. Acta
Gastroenterol Latinoam 1999;29:261-70.
(In Spanish.)
10. Hearnshaw SA, Logan A, Palmer KR,
Card TR, Travis SPL, Murphy MF. Out-
comes following early red blood cell transfusion in acute upper gastrointestinal
bleeding. Aliment Pharmacol Ther 2010;
32:215-24.
11. Halland M, Young M, Fitzgerald MN,
Inder K, Duggan JM, Duggan A. Characteristics and outcomes of upper gastrointestinal hemorrhage in a tertiary referral
hospital. Dig Dis Sci 2010;55:3430-5.
12. Kravetz D, Sikuler E, Groszmann RJ.
Splanchnic and systemic hemodynamics
in portal hypertensive rats during hemorrhage and blood volume restitution. Gastroenterology 1986;90:1232-40.
13. McCormick PA, Jenkins SA, McIntyre
N, Burroughs AK. Why portal hypertensive varices bleed and bleed: a hypothesis.
Gut 1995;36:100-3.
14. Castañeda B, Morales J, Lionetti R, et
al. Effects of blood volume restitution following a portal hypertensive–related
bleeding in anesthetized cirrhotic rats.
Hepatology 2001;33:821-5.
15. British Society of Gastroenterology
Endoscopy Committee. Non-variceal upper gastrointestinal haemorrhage: guidelines. Gut 2002;51:Suppl 4:iv1-iv6.
16. de Franchis R. Updating consensus in
portal hypertension: report of the Baveno
III consensus workshop on definitions,
methodology and therapeutic strategies
in portal hypertension. J Hepatol 2000;33:
846-52.
17. Villanueva C, Ortiz J, Miñana J, et al.
Somatostatin treatment and risk stratification by continuous portal pressure
monitoring during acute variceal bleeding. Gastroenterology 2001;121:110-7.
18. Jairath V, Hearnshaw S, Brunskill SJ,
et al. Red cell transfusion for the management of upper gastrointestinal haemorrhage. Cochrane Database Syst Rev 2010;
9:CD006613.
19. Carson JL, Terrin ML, Noveck H, et al.
Liberal or restrictive transfusion in high-
risk patients after hip surgery. N Engl J
Med 2011;365:2453-62.
20. Marik PE, Corwin HL. Efficacy of red
cell transfusion in the critically ill: a systematic review of the literature. Crit Care
Med 2008;36:2667-74.
21. de Franchis R. Revising consensus in
portal hypertension: report of the Baveno
V consensus workshop on methodology
of diagnosis and therapy in portal hypertension. J Hepatol 2010;53:762-8.
22. Crooks C, Card TIM, West J. Reductions in 28-day mortality following hospital admission for upper gastrointestinal
hemorrhage. Gastroenterology 2011;141:
62-70.
23. British Society of Gastroenterology.
UK upper GI bleeding audit (http://www
.bsg.org.uk/clinical/general/uk-upper-gibleeding-audit.html).
24. Roberts I, Evans P, Bunn F, Kwan I,
Crowhurst E. Is the normalisation of
blood pressure in bleeding trauma patients harmful? Lancet 2001;357:385-7.
25. Duggan JM. Transfusion in gastrointestinal haemorrhage — if, when and how
much? Aliment Pharmacol Ther 2001;15:
1109-13.
26. Vamvakas EC, Blajchman MA.
Transfusion-related immunomodulation
(TRIM): an update. Blood Rev 2007;21:
327-48.
27. Koch CG, Li L, Sessler DI, et al. Duration of red-cell storage and complications
after cardiac surgery. N Engl J Med 2008;
358:1229-39.
28. Kanias T, Gladwin MT. Nitric oxide,
hemolysis, and the red blood cell storage
lesion: interactions between transfusion,
donor, and recipient. Transfusion 2012;
52:1388-92.
Copyright © 2013 Massachusetts Medical Society.
nejm clinical practice center
Explore a new page designed specifically for practicing clinicians,
the NEJM Clinical Practice Center, at www.NEJM.org/clinical-practice-center.
Find practice-changing research, reviews from our Clinical Practice series,
a curated collection of clinical cases, and interactive features
designed to hone your diagnostic skills.
n engl j med 368;1 nejm.org january 3, 2013
The New England Journal of Medicine
Downloaded from nejm.org at UC SHARED JOURNAL COLLECTION on June 18, 2013. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
21