1. health and fitness
2. disease

Intensive Care Med (2001) 27: S 1±S 2
Charles L. Sprung
Gordon R. Bernard
R. Phillip Dellinger
Introduction
)
C. L. Sprung ( ) ´ G. R. Bernard ´ R. P. Dellinger
E-mail: [email protected]
Phone: +9 72-2-6 77 71 11
Fax: +9 72-2-5 63 54 43
This supplement to Intensive Care Medicine is the result
of a sustained effort by the International Sepsis Forum
(ISF) to bring together a body of practical recommendations for the management of patients with severe sepsis
and septic shock.
The ISF is a nonprofit, nongovernmental association
whose members are healthcare professionals in critical
care and infectious disease committed to improving the
understanding and clinical management of patients
with severe sepsis. While there is still a high morbidity
and mortality associated with sepsis and its sequelae,
new data on patient management are emerging that
may ultimately significantly improve the current situation. Such findings need to be evaluated and incorporated, when appropriate, into existing treatment protocols.
Headed by a Steering Committee of international experts and opinion leaders, the ISF focuses on improving
the management of sepsis and, in particular, septic
shock by developing an international consensus on the
latest understanding of key scientific and clinical issues,
and disseminating emerging practice guidelines to researchers, intensivists and other critical care professionals worldwide. The ISF educational program comprises
three major elements: workshops on management strategies in sepsis, traveling fellowships and lecture series.
Funding for the activities of the ISF is provided through
international sponsors using unrestricted educational
grants.
The present supplement presents management
guidelines for practicing clinicians and is primarily devoted to those patients with severe sepsis, typically hospitalized patients, many of whom are in intensive care
units. Animal trials have been mentioned when relevant, but our major goal is clinical management and we
have primarily focused on human studies.
The current recommendations are based on an evidence-based methodology with categorization as previously described (Table 1). The recommendations represent the groups assessment of the evidence-based medicine literature together with clinical practice and personal experience.
The evidence-based method is devised primarily for
therapeutic trials, and some of the contributions to this
Table 1 Grading of responses to questions and levels of evidence (adapted from [1])
Grading of responses to questions
A. Supported by at least two level I investigations
B. Supported by only one level I investigation
C. Supported by level II investigations only
D. Supported by at least one level III investigation
E. Supported by level IV or V evidence
Levels of evidence
I. Large, randomized trials with clearcut results; low risk of false-positive (alpha) error or false-negative (beta) error
II. Small, randomized trials with uncertain results; moderate to high risk of false-positive (alpha) and/or false-negative (beta) error
III. Nonrandomized, contemporaneous controls
IV. Nonrandomized, historical controls and expert opinion
V. Case series, uncontrolled studies, and expert opinion
S2
issue such as definitions, epidemiology of infection and
experimental therapy are therefore not well suited to
this approach. In addition, not all treatments used in
septic patients have been the subject of randomized,
controlled trials. Therefore the level of evidence may
not be so strong, even though a particular clinical practice is well established and in routine use. Clinicians
will need to be aware of this limitation and use their
judgement in interpreting these recommendations.
The methodology used included a Medline search for
at least 10 years prior to the publication, supplemented
by a manual search of other relevant journals. Keywords
that were used are noted in each contribution to this issue. All nine ISF members or their collaborators was re-
Reference
1. Sackett DL (1989) Rules of evidence and
clinical recommendations on the use of
antithrombotic agents. Chest 95 [Suppl]:2S±4S
quired to present a first draft of their respective contributions prior to a meeting in November 1999. At the
meeting each author presented the data for the article,
which were discussed. Based on suggestions, changes
were made and contributions were revised. Each article
was then reviewed by another ISF member and then edited by one of the three editors.
We believe this supplement will be an extremely
helpful source of practical information for the clinician.
As the fields of infectious diseases, microbiology and intensive care are ever moving forward, the data and recommendations may change in the future. In the meantime, we believe this supplement will be an important
addition to every clinician's library.
Intensive Care Med (2001) 27: S 3±S 9
Definition of sepsis
Idit Matot
Charles L. Sprung
)
I. Matot ´ C. L. Sprung ( )
Department of Anesthesiology and Critical Care Medicine,
Hadassah University Medical Center,
Hebrew University of Jerusalem, Jerusalem, Israel
E-mail: [email protected]
Phone: +9 72-2-6 77 71 11
Fax: +9 72-2-5 63 54 43
History of definitions
Sepsis is the systemic inflammatory response to infection [1]. Sepsis and its sequelae represent progressive
stages of the same illness in which a systemic response
to an infection mediated by endogenous mediators may
lead to a generalized inflammatory reaction in organs
remote from the initial insult and eventually to end-organ dysfunction and/or failure [2]. Sepsis remains an important and life-threatening problem. Sepsis is the most
common cause of death in the intensive care unit [3]. Because of increasingly aggressive treatment of patients in
advanced stages of illness the incidence and mortality
from sepsis in hospital patients remains high [2, 4, 5, 6].
New efforts to improve survival have highlighted the uncertainty of the specific diagnostic criteria used to define
entry criteria for clinical trials. In the past the terms bacteremia, septicemia, sepsis, sepsis syndrome, and septic
shock were used interchangeably, which led to an imprecise understanding of sepsis and its related disorders and
to confusion in the interpretation of clinical trials [6].
A review of the literature published in the 1980s discloses remarkable disparity in the mortality of patients with
sepsis, sepsis syndrome, and septic shock [5, 7, 8, 9, 10,
11]. The conflicting results were due in part to different
definitions of the clinical entity under investigation.
Bone, Sibbald, and Sprung [9, 10, 11] listed several reasons for the lack of firm definitions at that time, including the lack of epidemiological studies to evaluate the
systemic response to infection, the lack of precise criteria for the different terms used, the early death of severely injured patients before the development of sepsis
and septic shock, and the focus on patients with Gramnegative bacteremia rather than sepsis. Two decades
ago the idea of a microbial cause of sepsis was established, and bacteremia was required for a patient to
have ªsepsisº [12]. Later, investigators recognized that
the host is not passive when invaded by an organism
but in fact secretes a large spectrum of endogenous inflammatory mediators that may also result in injury.
Moreover, it was also recognized that the same inflammatory response might result from noninfectious insults
as well [13, 14]. Further studies [7, 15] found that the
clinical response persists after eradication of the infection, and is itself associated with an increased mortality.
Although the inflammatory response is beneficial in
many patients, an exaggerated response, rather than
the invasive infection, may be the more important determinant of outcome in critically ill patients [16].
Methods
Consensus Conference definitions
To evaluate studies of sepsis for definitions a computer-based review of the literature was undertaken using Medline from 1990 until September 1999 as the primary database. The specific subject
heading keywords were definition, diagnosis, criteria, epidemiology, and classification and matching them with infection, sepsis, sepsis syndrome, septicemia, septic shock, and multiple organ failure.
A new set of definitions was proposed by the Consensus
Conference of the American College of Chest Physicians and the Society of Critical Care Medicine held in
Chicago in August 1991 [6]. These definitions included
patients in various stages of infection: bacteremia, sep-
Introduction
S4
Table 1 Clinical frequency of SIRS, sepsis, severe sepsis, and septic shock
Reference
Location
Determination
SIRS (%)
Sepsis (%)
Severe
sepsis (%)
Rangel-Frausto
et al. [19]
Pittet et al. [20]
Medical, surgical
ICUs and wards
Surgical ICU
Study period
68
26
18
Study period
93
49
16
7
Salvo et al. [21]
General ICU
Admission
52
5
2
3
Saez-Llorens et al. [22]
Pediatric ICU
Infection suspected
±
21
61
18
Proulx et al. [23]
Jones and Lowes [24]
Pediatric ICU
Medical ward
Study period
Time of blood
culture
82
55
23
16
4
5
2
3
Muckart and
Bhagwanjee [4]
Surgical ICU
(trauma)
1st 24 h
88
14
14
20
Bossink et al. [25]
Medical ward
Fever onset
95
44
±
±
sis, severe sepsis, septic shock, and multiple organ dysfunction syndrome (MODS). Recommendations from
the Consensus Conference provided both a conceptual
and practical framework for the definition of the systemic inflammatory response to infection (sepsis). It
was expected that the application of a more flexible
and easily met definition would improve early bedside
detection of sepsis, permit early intervention, standardize research protocols, and allow comparisons of results
of clinical trials of new therapies.
The conference proposed a new term, systemic inflammatory response syndrome (SIRS), to describe
widespread inflammation that occurs following a wide
variety of insults including infection, pancreatitis, trauma, burns, etc. The term SIRS validated the concept
that endogenous mediators of inflammation play an important role in ªsepsisº together with microbial factors.
The term sepsis was defined as a subset of patients with
an inflammatory response, limited to those patients
with documented infection. The systemic response included two or more abnormalities in temperature, heart
rate, respiratory rate, and white blood count. Many patients have infection without a systemic response and
are therefore not septic. SIRS, sepsis, severe sepsis, and
septic shock represent a continuum of clinical and pathophysiological severity. The process begins with an infection, with or without a systemic inflammatory response,
and may progress to a systemic response with severe sepsis (hypotension, hypoperfusion, or organ dysfunction)
or septic shock (hypotension not responsive to adequate
fluid resuscitation with hypoperfusion or organ dysfunction). It was believed that the phases of the disease process form a continuum of severity which characterize
populations at increased risk of morbidity and mortality.
The terms septicemia, sepsis syndrome, and refractory
shock were eliminated by the Consensus Conference because they were believed to be confusing and nonspecific, applying to a variety of inflammatory states.
Septic
shock (%)
4
Epidemiology
Prior to the Consensus Conference it was not possible to
compare epidemiological or outcome studies of sepsis
since the definitions were not standardized. Hypothermia, fever, tachycardia, and tachypnea or hypocapnea
were included in many studies [7, 12, 17, 18] as well as
hypotension [7, 17, 18]. Laboratory results indicative of
infection were also included in many sepsis trials. They
included abnormal white blood cell count, abnormal
neutrophil count, and thrombocytopenia [7, 12, 18].
Manifestations of organ dysfunction in septic patients
included altered mentation, hypoxemia, oliguria, coagulopathy, and increased lactate concentration [7, 17, 18].
Unfortunately, no studies had evaluated and compared
the epidemiology, morbidity, and mortality of different
definitions of sepsis.
After the Consensus Conference several studies evaluated the epidemiology of sepsis [4, 19, 20, 21, 22, 23, 24,
25]. The frequency of the problem usually decreased as
the severity increased from SIRS to sepsis to severe sepsis and to septic shock (Table 1). Unfortunately, the
population of patients studied differed and the inflammatory entity was determined at different times (admission, during the first 24 or 48 h, at fever onset, or over
the entire study period). As patients developed more
SIRS criteria (from two to four), the incidence of sepsis
increased [19, 24], as did mortality [24].
Correlation to mortality
Sepsis in its evolution to severe sepsis and septic shock
reflect increasing disease severity with increased patient
morbidity and mortality. Knaus et al. [5] found that the
presence or absence of sepsis (infection plus two of
four SIRS criteria) failed to identify subgroups with increased risk of death and concluded that while categoric
S5
Table 2 Mortality of SIRS, sepsis, severe sepsis, and septic shock
Reference
SIRS (%)
Sepsis (%)
Severe sepsis (%)
Septic shock (%)
Rangel-Frausto et al. [19]
Pittet et al. [20]
Salvo et al. [21]
Saez-Llorens et al. [22]
Jones and Lowes [24]
Muckart and Bhagwanjee [4]
Bossink et al. [25]
7
6
27
±
23
8
6
16
0
36
16
±
10
13
20
35
52
40
38
18
±
46
58
82
62
56
53
±
definitions of sepsis may be useful in selecting patients
for entry into clinical trials, they may not be useful in
characterizing individual, or perhaps group, risks. Several studies, however, have demonstrated increasing organ
failure (acute respiratory distress syndrome, disseminated intravascular coagulopathy, acute renal failure, and
shock) with worsening severity of sepsis [19, 20]. In addition, most studies [4, 19, 20, 21, 22, 24, 25] have noted
stepwise increases in mortality from SIRS to sepsis to
severe sepsis and to septic shock (Table 2).
Clinical characteristics defining the response to infection
The Consensus definition [4] of sepsis has been criticized for its restrictive and incomplete nature [26, 27].
For example, it may be very difficult to detect a respiratory rate of more than 20 breaths/min in patients already on respiratory support. Moreover, it might be difficult to diagnose organ dysfunction, for example, CNS
depression in a sedated patient [26]. The definition of
ªseptic shockº is also somewhat subjective, as hypotension that persists after ªadequate fluid resuscitationº
has not been defined and may be arbitrary. Moreover, a
blood pressure based criterion may underestimate the
number of patients with shock since shock physiologically represents the inability of oxygen delivery to support metabolic demands, which can occur in the presence of normal blood pressure [28]. Some recent epidemiological studies of sepsis have not used the consensus
definitions [29, 30, 31, 32]. Several clinical trials have
used consensus definitions [33, 34, 35] whereas others
have not [36, 37, 38, 39, 40].
Despite certain problems with the Consensus definitions they define a continuum of progressive physiological decline toward multiple organ dysfunction/failure
and mortality. Better understanding of the pathophysiology of sepsis will enable us in the future to begin therapy at an earlier stage in the disease process before the
onset of shock and multiorgan dysfunction and failure.
The Consensus Conference used specific clinical characteristics and thresholds to define sepsis, but there may
be better combinations of clinical manifestations in septic patients. Unfortunately, none has proven more use-
ful or has undergone such a rigorous evaluation of its
epidemiology, morbidity, and mortality. Although the
Consensus criteria have been useful in epidemiological
studies, they should not be the sole basis for the clinical
diagnosis of sepsis. It may be extremely difficult to diagnose sepsis in the patient who does not have the classical
findings. The symptoms and signs that should lead the
clinician to suspect sepsis are as follows:
· Clinical signs
· Fever/hypothermia
· Unexplained tachycardia
· Unexplained tachypnea
· Signs of peripheral vasodilation
· Unexplained shock
· Changes in mental status
· Invasive hemodynamic or laboratory parameters
· Low systemic vascular resistance/Increased cardiac
output
· Increased oxygen consumption
· Leukocytosis/neutropenia
· Unexplained lactic acidosis
· Unexplained alteration in renal or liver function
tests
· Thrombocytopenia/disseminated intravascular coagulation
· Increased procalcitonin
· Increased cytokines, C reactive protein
Absolute thresholds are probably less important than
combinations or severity levels. For example, sepsis
should be suspected in all patients with unexplained
shock.
Consensus vs. controversy
The Consensus Conference [5] agreed to a new set of
definitions that could readily be applied to patients in
various stages of sepsis. It was expected that application
of these definitions would improve bedside detection of
infections and permit earlier intervention before the onset of shock and multiorgan failure. Early recognition of
S6
the inflammatory response to infections should lead to
more effective therapeutic management.
The term SIRS was introduced to provide a term other than sepsis to describe a patient who looked ªsepticº
but was not infected. Some clinicians oppose the use of
the term SIRS [4, 19, 20, 26, 27, 28, 41] as it is nonspecific, very common in hospitalized patients, and can lead
physicians to become compliant with a ªdiagnosisº and
not seek a potential infection. When a patient manifests
a systemic response to infection, the terminology used is
not as important as the actions. It is agreed that sepsis,
severe sepsis, and septic shock represent a continuum
in a disease process and are correlated with increasing
organ dysfunction and failure and mortality. It is vitally
important to search continuously for a source of infection and to make a diagnosis. Treatment of infection
should eradicate the invading organism with antimicrobial therapy and surgical debridement or drainage. Early detection of infections in patients should lead to rapid
and aggressive diagnostic and therapeutic interventions
which should prevent circulatory compromise and organ dysfunction and improve survival. The exact criteria
used to define these entities may change with time as
more studies are performed.
Biological markers of infection
Early detection of sepsis is difficult because the first
signs of this disease may be minimal and are similar to
those of various noninfectious processes, and culture results are not immediately available. The availability of
laboratory tests to accurately and more rapidly identify
septic patients by the isolation of micro-organisms
from body fluid specimens would be of considerable value.
Several indicators measured in the bloodstream have
been evaluated for the diagnosis of sepsis. A prominent
and invariable component of the systemic inflammatory
response is the induction and release of cytokines and
acute-phase proteins, which rapidly increase in the serum. The principal cytokines [interleukin (IL) 1, tumor
necrosis factor (TNF) a, IL-6, IL-8, IL-10] and their soluble receptors (soluble TNF receptor) have convincingly been demonstrated to increase during sepsis [42, 43,
44, 45, 46]. Both cytokines and their soluble receptors,
however, are greatly elevated also in nonbacterial infections such as malaria [47] and in disease states such as
burn injury [48], trauma [49], pancreatitis [50], heart
failure [51], renal allograft rejection [52], and even in
such minor injury states as elective surgery [53, 54]. Critics have therefore rightly questioned the diagnostic significance of elevated cytokine concentration because of
its nonspecific character. In addition, the inflammatory
response changes with time during the course of sepsis
and a complex balance exists between cytokines and
their inhibitors and antagonists, which require that the
evaluation of inflammatory pathways take into account
the pro- and anti-inflammatory factors that affect each
pathway. Current efforts should therefore be directed
at defining the cytokine balance that exists at the onset
of sepsis, how this balance changes over time, and determining whether groups of cytokine variables (not a single cytokine) can be used to predict more accurately either the onset or the outcome of sepsis.
Recently several studies have evaluated the use of Creactive protein (CRP) and procalcitonin (PCT) in differentiating conditions that may mimic sepsis. These
studies highlight both CRP and PCT as specific markers
of bacterial sepsis [55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65,
66, 67, 68, 69]. Both are acute-phase proteins, and their
role is unknown.
It has been reported that increased CRP concentrations differentiate between viral and bacterial infections
[55], and that decreases in levels indicate resolution of
bacterial sepsis [56]. Moreover, CRP has been shown to
be superior to body temperature and white blood cell
count in identifying bacterial sepsis [57]. Other studies
have been less enthusiastic about CRP as a marker of infection. Increased levels were demonstrated in patients
with stable and unstable angina pectoris and in patients
with heart failure. Interestingly, increased preoperative
levels of CRP were predictive of postoperative infection
after cardiac surgery [58]. CRP was also markedly elevated in trauma patients, in children after head injury,
and in patients undergoing various types of surgery.
Some studies have also reported higher concentrations of PCT in patients with systemic manifestations of
infection than in noninfected patients who fulfill the
SIRS criteria [59] or patients with systemic viral and localized bacterial infection [60]. Other studies have
shown that increased levels of PCT predict infected necrosis in patients with acute pancreatitis [61, 70] or bacterial or fungal infection in severe burn injuries [62]
and can discriminate between infection and graft rejection in patients after organ transplantation [63, 64, 65,
71]. In patients having undergone cardiopulmonary bypass increases in PCT have been demonstrated only in
infected patients, with low values in patients with SIRS
alone [66]. Moreover, in injured patients [72] sustained
increases in PCT plasma levels predict sepsis and
MODS. In neonates [67] the sensitivity and specificity
of PCT concentrations in the detection of early-onset
sepsis was 92.6 % and 97.5 %, respectively, and 100 %
for both for the diagnosis of late-onset sepsis.
In a recent study of 111 infected and 79 noninfected
patients Urgate et al. [68] found a slightly higher sensitivity and specificity for CRP than PCT in the diagnosis
of infection: 71.8 % and 66.6 % for CRP and 67.6 %
and 61.3 % for PCT, respectively. However, the combination of CRP and PCT was much more specific in ascertaining the diagnosis of infection. In contrast, PCT
S7
was found to be a better marker of infection than CRP
in patients with pancreatitis [61], and in patients after
renal transplantation [63]. PCT has been found to be superior to the classical criteria of inflammation (CRP,
TNFa, IL-6, leukocyte count, and body temperature) in
identifying patients with sepsis [69, 73, 74].
How abnormal must PCT blood concentration be before it reliably predicts the presence of sepsis in the appropriate clinical setting? The answer is not known.
Based on the data reviewed, a value higher than 5 ng/
ml would be a reasonable threshold. More importantly,
PCT measurements might serve as a reasonable parameter to rule out sepsis. When blood levels are very low
or undetectable, the patient is unlikely to have sepsis as
a cause of organ dysfunction or shock.
The definition of sepsis should not depend on any
single arbitrary dichotomizing measurement. Rather,
sepsis should be defined as the systemic inflammatory
response, associated with an identified source of infection that develops under the appropriate clinical setting.
If measurement of a marker such as procalcitonin is validated to be correlated with the presence of infection
(and further studies are needed for this), a more precise
set of criteria can be proposed for the diagnosis of sepsis: (a) systemic manifestations; (b) an appropriate clinical setting (the reason for the addition of this criterion is
that the systemic manifestations are by themselves nonspecific markers of infection); (c) source of infection;
and (d) laboratory marker. In the future a biological
marker or markers of infection, alone or in combination, should be evaluated for their capability to predict/
diagnose sepsis. As discussed above, PCT, with an appropriate concentration threshold, may prove to be
such a marker. Measurements of such a marker may
have to be repeated daily as the sensitivity increases
with time after the diagnosis of infection [68]. Other
measures [75, 76] that have been proposed as markers
of sepsis are not easily determined and therefore are difficult to apply to all patients. In the meantime these
should serve as a research tool only.
Sepsis is a diverse illness, and the heterogeneity of
patient populations may account for some of the variability in the results of the published studies. Prospec-
tive epidemiological studies and clinical trials of sepsis
should include a larger number of patients, enough to
permit stratification of the population by the source of
infection (CNS, lungs, gastrointestinal, urinary, etc.)
and the pattern of immunological response (hypo- vs.
hyperactivation, and precise stage of inflammatory response). The former is a potentially important issue
since the source of infection has been found in many
studies to be statistically associated with increased mortality [77, 78]. In addition, to know whether the studied
patient populations are similar, one must also develop
two sets of appropriate tests: one that measures the severity of injury to the infected system and one that measures the overall severity of the patient's illness.
Summary and recommendations
Sepsis is the systemic inflammatory response to infection. In the past many different definitions of sepsis
were used interchangeably, which led to confusion. Today, the definitions of sepsis and its clinical manifestations are still a source of controversy. No single physiological or laboratory parameter can universally identify
sepsis. Not all patients with sepsis are equally ill. Sepsis,
severe sepsis, and septic shock constitute different gradations in the continuum of a disease process manifested by a combination of changes in vital signs, laboratory
parameters, hypoperfusion, and organ dysfunction. The
continuum of sepsis, severe sepsis, and septic shock is
correlated with increasing organ dysfunction and mortality. The source of infection and diagnosis of sepsis
must be identified as early as possible to permit early intervention with antimicrobial therapy and surgical
drainage to prevent disease progression, organ dysfunction, and mortality. Although the consensus criteria
have been useful in epidemiological studies, they should
not be the sole basis for the clinical diagnosis of sepsis.
Future prospective clinical trials may lead to a more
precise understanding of sepsis and its related disorders
and result in clearer, more universally accepted and easily interpreted definitions.
References
1. Rackow EC (1986) Clinical definition
of sepsis and septic shock. In: Sibbald
WJ, Sprung CL (eds) New horizons:
perspectives of sepsis and septic shock.
Society of Critical Care Medicine, Fullerton, pp 1±9
2. Bone RB, Grodzin CG, Balk RA (1998)
Sepsis: a new hypothesis for pathogenesis of the disease process. Chest 112:
235±243
3. Parrillo JE, Parker MM, Natanson C
(1990) Septic shock; advances in understanding of pathogenesis, cardiovascular dysfunction, and therapy. Ann Intern Med 113: 227±242
4. Muckart DJJ, Bhagwanjee S (1997)
American College of Chest Physicians/
Society of Critical Care Medicine Consensus Conference: definition of the
systemic inflammatory response syndrome and allied disorders in relation
to critically injured patients. Crit Care
Med 25: 1789±1795
S8
5. Knaus WA, Xiaolu Sun, Nystrom PO,
Wagner DP (1992) Evaluation of definitions for sepsis. Chest 101: 1656±1662
6. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, Schein
RMH, Sibbald WJ (1992) American
College of Chest Physicians/Society of
Critical Care Medicine Consensus Conference: definition for sepsis and organ
failure and guidelines for the use of innovative therapies in sepsis. Crit Care
Med 20: 864±874
7. Bone RC, Fisher CJ, Clemmer TP, Slotman GJ, Metz CA, Balk RA (1989)
The Methylprednisolone Severe Sepsis
Study Group. Sepsis syndrome: a valid
clinical entity. Crit Care Med 17:
389±393
8. Bihari DJ (1990) Septicaemia ± the clinical diagnosis. J Antimicrob Chemother
25 [Suppl C]:1±7
9. Sprung CL (1991) Definitions of sepsis
± have we reached a consensus? Crit
Care Med 19: 849±850
10. Bone RC (1991) Let's agree on terminology: definitions of sepsis. Crit Care
Med 19: 973±976
11. Bone RC, Sibbald WJ, Sprung CL
(1992) The ACCP-SCCM Consensus
Conference on sepsis and organ failure.
Chest 101: 1481±1482
12. Kreger BE, Craven DE, McCabe
(1980) Gram negative bacteremia. Reevaluation of clinical features and treatment in 612 patients. Am J Med 68:
344±355
13. Nystrom PO (1998) The systemic inflammatory response syndrome: definitions and aetiology. J Antimicrob
Chemother 41 [Suppl A]:1±7
14. Canadian Multiple Organ Failure Study
Group: Sibbald WJ, Marshall J, Christou N, Girotti M, McCormack D, Rostein O, Martin C, Meakins J (1991)
ºSepsisº ± clarity of existing terminology...or more confusion? Crit Care Med
19: 996±998
15. Jonsson B, Berglund J, Skau T, Nystrom
PO (1993) Outcome of intra-abdominal
infection in pigs depends more on host
responses than on microbiology. Eur J
Surg 159: 571±578
16. Marshall JC, Sweeney D (1990) Microbial infection and the septic response in
critical surgical illness. Sepsis, not infection, determines outcome. Arch Surg
125: 17±23
17. The Veterans Administration Systemic
Sepsis Cooperative Study Group
(1987) Effect of high-dose glucocorticoid therapy on mortality in patients
with clinical signs of systemic sepsis.
N Engl J Med 317: 659±665
18. Ziegler EJ, Fisher CH, Sprung CL, et al
(1991) Treatment of Gram-negative
bacteremia and septic shock with HA-1
A human monoclonal antibody against
endotoxin. N Engl J Med 324: 429±436
19. Rangel-Frausto MS, Pittet D, Costigan
M, Hwang T, Davis CS, Wenzel RP
(1995) The natural history of the systemic inflammatory response syndrome
(SIRS). JAMA 273: 117±123
20. Pittet D, Rangel-Frausto S, Li N, Tarara
D, Costigan M, Rempe L, Jebson P,
Wenzel RP (1995) Systemic inflammatory response syndrome, sepsis, severe
sepsis, and septic shock: incidence, morbidities and outcomes in surgical ICU
patients. Intensive Care Med 21:
302±309
21. Salvo I, de Cian W, Musicco M, Langer
M, Piadena R, Wolfler A, Montani C,
Magni E: The Sepsis Study Group
(1995) The Italian sepsis study: preliminary results on the incidence and evolution of SIRS, sepsis, severe sepsis, and
septic shock. Intensive Care Med
21:S244±S249
22. Saez-Llorens X, Vargas S, Guerra F,
Coronado L (1995) Application of new
sepsis definitions to evaluate outcome
of pediatric patients with severe systemic infections. Pediatr Infect Dis J 14:
557±561
23. Proulx F, Fayon M, Farrell CA, Lacroix
J, Gauthier M (1996) Epidemiology of
sepsis and multiple organ dysfunction
syndrome in children. Chest 109:
1033±1037
24. Jones GR, Lowes JA (1996) The systemic inflammatory response syndrome
as a predictor of bacteraemia and outcome from sepsis. Q J Med 89: 515±522
25. Bossink AWJ, Groeneveld J, Hack CE,
Thijs LG (1998) Prediction of mortality
in febrile medical patients. How useful
are systemic inflammatory response
syndrome and sepsis criteria? Chest
113: 1533±1541
26. Vincent JL (1997) Dear SIRS, I'm sorry
to say that I don't like you... Crit Care
Med 25: 372±374
27. Benjamin E, Leibowitz AB, Oropello J,
Iberti TJ (1992) Systemic hypoxic and
inflammatory syndrome: an alternative
designation for ªsepsis syndromeº. Crit
Care Med 20: 680±682
28. Opal SM (1998) The uncertain value of
the definition for SIRS. Chest 113:
1442±1443
29. Poole GV, Griswold JA, Muakkassa FF
(1993) Sepsis and infection in the intensive care unit: are they related? Am
Surg 59: 60±64
30. Kieft H, Hoepelman AIM, Zhou W, et
al (1993) The sepsis syndrome in a
Dutch university hospital. Arch Intern
Med 153: 2241±2247
31. Conboy K, Welage LS, Walawander
CY, et al (1995) Sepsis syndrome and
associated sequelae in patients at high
risk for gram-negative sepsis. Pharmacotherapy 15: 66±77
32. Sands KE, Bates DW, Lanken PN, et al
(1997) Epidemiology of sepsis syndrome in 8 academic medical centers.
JAMA 278: 234±240
33. Fisher CJ, Dhainaut J-FA, Opal SM, et
al (1994) Recombinant human interleukin 1 receptor antagonist in the treatment of patients with sepsis syndrome.
JAMA 271: 1836±1843
34. Opal SM, Fisher CJ, Dhainaut J-FA, et
al (1997) Confirmatory interleukin-1 receptor antagonist trial in severe sepsis: a
phase III, randomized, double-blind,
placebo-controlled, multicenter trial.
Crit Care Med 25: 1115±1124
35. Abraham E, Glauser MP, Butler T, et al
(1997) p55 tumor necrosis factor receptor fusion protein in the treatment of
patient with severe sepsis and septic
shock. JAMA 277: 1531±1538
36. Abraham E, Wunderink R, Silverman
H, et al (1995) Efficacy and safety of
monoclonal antibody to human tumor
necrosis factor a in patients with sepsis
syndrome. JAMA 273: 934±941
37. Cohen J, Carlet J for the INTERSEPT
study group (1996) INTERSEPT: an international, multicenter, placebo-controlled trial of monoclonal antibody to
human tumor necrosis factor a in patients with sepsis. Crit Care Med 24:
1431±1440
38. Fisher CJ, Agosti JM, Opal SM, et al
(1996) Treatment of septic shock with
the tumor necrosis factor receptor: Fc
fusion protein. N Engl J Med 334:
1697±1702
39. Bernard GR, Wheeler AP, Russell JA,
et al (1997) The effects of ibuprofen on
the physiology and survival of patients
with sepsis. N Engl J Med 336: 912±918
40. Eisele B, Lamy M, Thijs LG, et al
(1998) Antithrombin III in patients
with severe sepsis. Intensive Care Med
24: 663±672
41. Sibbald WJ, Doig G, Inman KJ (1995)
Sepsis, SIRS and infection. Intensive
Care Med 21: 299±301
42. Moscovitz H, Shofer F, Mignott H, et al
(1994) Plasma cytokine determination
in emergency department patients as a
predictor of bacteremia and infectious
disease severity. Crit Care Med 22:
1102±1107
S9
43. Damas P, Ledoux D, Nys M, et al (1992)
Cytokine serum level during severe sepsis in human. IL-6 as a marker of severity. Ann Surg 215: 356±362
44. Damas P, Reuter A, Gysen P, et al
(1989) Tumor necrosis factor and interleukin-1 serum levels during severe sepsis in humans. Crit Care Med 17:
975±978
45. Hack CE, De Groot ER, Felt-Bersma
RJF, et al (1989) Increased plasma levels of interleukin-6 in sepsis. Blood 74:
1704±1710
46. Van der Poll T, Jansen J, van Leenen D,
et al (1993) Release of soluble receptors
for tumor necrosis factor in clinical sepsis and experimental endotoxemia. J Infect Dis 168: 955±960
47. Brockhaus M (1997) Soluble TNF receptor: what is the significance? Intensive Care Med 23: 808±809
48. Nijsten MWN, de Groot ER, ten Duis
HJ, et al (1987) Serum levels of interleukin-6 and acute phase responses.
Lancet II:921±922
49. Ertel W, Keel M, Bonaccio M, Steckholzer U, et al (1995) Release of anti-inflammatory mediators after mechanical
trauma correlates with severity of injury
and clinical outcome. J Trauma 39:
879±885
50. Brivet FG, Emilie D, Galanaud P
(1999) Pro- and anti inflammatory cytokines during acute severe pancreatitis:
and early and sustained response, although unpredictable of death. Parisian
Study Group on Acute Pancreatitis.
Crit Care Med 27: 749±755
51. Testa M, Yeh M, Fanelli R, et al (1996)
Circulating levels of cytokines and their
endogenous modulators in patients with
severe congestive heart failure due to
coronary artery disease or heart failure.
J Am Coll Cardiol 28: 964±971
52. Van Oers MHJ, van der Hyeden AAPAM, Aarden LA (1988) Interleukin 6
(IL-6) in serum and urine of renal transplant recipients. Clin Exp Immunol 71:
314±319
53. Shenkin A, Fraser WD, Series J (1989)
The serum interleukin-6 response to
elective surgery. Lymphokine Res 8:
123±127
54. Bistrian BR (1999) Acute phase proteins and the systemic inflammatory response. Crit Care Med 27: 452±453
55. Shaw AC (1991) Serum C-reactive protein and neopterin concentrations in patients with viral or bacterial infection.
J Clin Pathol 44: 596±599
56. Yentis SM, Soni N, Sheldon J (1995)
C-reactive protein as an indicator of
resolution of sepsis in the intensive
care unit. Intensive Care Med 21:
602±605
57. Povoa P, Almeida E, Moreira P, et al
(1998) C-reactive protein as an indicator of sepsis. Intensive Care Med 24:
1052±1056
58. Fransen EJ, Maessen JG, Elenbaas TW,
et al (1999) Enhanced preoperative
C-reactive protein plasma levels as a
risk factor for postoperative infections
after cardiac surgery. Ann Thorac Surg
67: 134±138
59. Al-Nawas B, Krammer I, Shah PM
(1996) Procalcitonin in diagnosis of severe infection. Eur J Med Res 1:
331±333
60. Assicot M, Gendrel D, Carsin H, et al
(1993) High serum procalcitonin concentration in patients with sepsis and infection. Lancet 341: 515±518
61. Rau B, Steinbach G, Gansauge F, et al
(1997) The potential role of procalcitonin and interleukin 8 in the prediction
of infected necrosis in acute pancreatitis. Gut 41: 832±840
62. Von Heimburg D, Stieghorst W, Khorram S, et al (1998) Procalcitonin-a sepsis parameter in severe burn injuries.
Burns 24: 745±750
63. Eberhard OK, Langefeld I, Kuse ER, et
al (1998) Procalcitonin in the early
phase after renal transplantation ± will
it add to diagnostic accuracy? Clin
Transplant 12: 206±211
64. Hammer C, Reichenspurner R, Meiser
B (1998) Cytoimmunology in monitoring: the Munich experience. Transplant
Proc 30: 873±874
65. Staehler M, Hammer C, Meiser B, et al
(1997) Procalcitonin: a new marker for
differential diagnosis of acute rejection
and bacterial infection in heart transplantation. Transplant Proc 29: 584±585
66. Boeken U, Feindt P, Petzold T, et al
(1998) Diagnostic value of procalcitonin: the influence of cardiopulmonary
bypass, aprotinin, SIRS, and sepsis.
Thorac Cardiovasc Surg 46: 348±351
67. Chiesa C, Panero A, Rossi N, et al
(1998) Reliability of procalcitonin concentrations for the diagnosis of sepsis
in critically ill neonates. Clin Infect Dis
26: 664±672
68. Urgate H, Eliezer Silva, Mercan Dany,
et al (1999) Procalcitonin used as a
marker of infection in the intensive
care unit. Crit Care Med 27: 498±504
69. Oberhoffer M, Vogelsang H, Russwurm
S, et al (1999) Outcome prediction by
traditional and new markers of inflammation in patients with sepsis. Clin
Chem Lab Med 37: 363±368
70. Rau B, Steinbach G, Baumgart K, et al
(2000) The clinical value of procalcitonin in the prediction of infected necrosis
in acute pancreatitis. Intensive Care
Med 26:S159±S164
71. Kuse ER, Langefeld I, Jaeger K, et al
(2000) Procalcitonin in fever of unknown origin after liver transplantation: a variable to differentiate acute rejection from infection. Crit Care Med
28: 555±559
72. Wanner GA, Keel M, Steckholzer U, et
al (2000) Relationship between procalcitonin plasma levels and severity of injury, sepsis, organ failure, and mortality
in injured patients. Crit Care Med 28:
950±957
73. Cheval C, Timsit JF, Garrouste-Orgeas
M, et al (2000) Procalcitonin (PCT) is
useful in predicting the bacterial origin
of an acute circulatory failure in critically ill patients. Intensive Care Med
26:S153±S158
74. Oberhoffer M, Ruûbwurm S, Bredle D,
et al (2000) Discriminative power of inflammatory markers for prediction of
tumor necrosis factor-a and interleukin-6 in ICU patients with systemic inflammatory response syndrome (SIRS)
or sepsis at arbitrary time points. Intensive Care Med 26:S170±S174
75. Palombo JD, Bistrian BR (1999) Early
diagnosis of septic shock: a simple test
for a complex condition? Crit Care
Med 27: 465±466
76. Martens A, Eppink GJM, Woittiez AJJ,
et al (1999) Neutrophil function capacity to express CD10 is decreased in patients with septic shock. Crit Care Med
27: 549±553
77. Uzun O, Akalin HE, Hayran M, et al
(1992) Factors influencing prognosis in
bacteremia due to gram-negative organisms: evaluation of 448 episodes in
a Turkish university hospital. Clin Infect Dis 15: 866±873
78. Barriere SL, Lowry SF (1995) An overview of mortality risk prediction in sepsis. Crit Care Med 23: 376±393
Intensive Care Med (2001) 27: S 10±S 32
Diagnosis of infection in sepsis
Martin Llewelyn
Jonathan Cohen
)
M. Llewelyn ´ J. Cohen ( )
Department of Infectious Diseases,
Imperial College School of Medicine,
Hammersmith Hospital, London, UK
E-mail: [email protected]
Phone: +44-20-83 83 32 43
Fax: +44-20-83 83 33 94
Introduction
Infection is a sine qua non of sepsis. Sepsis can complicate infection occurring at any site, most commonly the
respiratory tract, abdomen and blood stream. More
than 90 % of cases of sepsis are caused by bacteria, and
Gram-negative and Gram-positive organisms occur
with approximately equal frequency [1]. Fungi ± in particular Candida species ± are sometimes responsible,
but a wide variety of other organisms have occasionally
been implicated [2].
There are several reasons why it is important to try
and make a microbiological diagnosis in septic patients.
First, and most important, is to ensure that effective antimicrobial therapy is given. There is good evidence to
support the intuitive belief that patients given appropriate therapy are more likely to survive than those given
inadequate or inappropriate treatment [3]. Secondly,
obtaining microbiological information will contribute
to the local epidemiological database, without which
logical prescribing is difficult, if not impossible. There
are substantial differences between intensive care units
in the microbial ecology, including the prevalence of
methicillin-resistant Staphylococcus aureus and vancomycin-resistant Enterococcus faecalis. Antimicrobial resistance patterns also vary widely, for example, penicillin resistance in Streptococcus pneumoniae, and gentamicin resistance in Enterobacteriaceae. Furthermore,
these patterns are constantly changing, and an up-todate awareness of these patterns is obviously essential
when considering empirical therapy. Finally, knowledge
of the microbial cause of sepsis may be important in
the choice of adjunctive therapies. This is not yet clinical
reality but will clearly be important if, for instance, antiendotoxin agents ever enter the clinical arena.
In this contribution we first consider the general approach to the diagnosis of infection in septic patients
and then some aspects of infections at particular sites.
We focus on the microbiological aspects of diagnosis, although where appropriate we also comment on other
modalities such as imaging. The clinical management
of these infections are discussed elsewhere.
Methods
We use where possible a systematic, evidence-based approach to
provide answers to specific questions. A computer-based review
of the literature was undertaken using Medline from 1991 until
September 1999 as the primary database. The references obtained
were searched manually for relevance. More recent clinical articles
were identified by manual search of the relevant journals. Although the literature searches were not extended further back
than 1991, key earlier papers are frequently cited. In reviewing
this field we frequently observed that whereas there is extensive
literature on microbiological aspects of infection in general, there
is a paucity of data concerning sepsis in particular. Hence we have
often been obliged to make recommendations based on the available literature and our own clinical experience, and that of others.
The search strategies used were as follows.
Bacteremia
The primary search terms bacteremia/septicemia (diagnosis, microbiology) and blood (microbiology) were combined with secondary search terms sepsis/sepsis syndrome (diagnosis, microbiology),
blood specimen collection (methods), bacteriological techniques
and diagnostic tests, routine (methods). Studies addressing exclusively pediatric populations were excluded.
S 11
Central venous catheter infection
The primary search terms catheterization, central venous (adverse
effects) and indwelling catheters (adverse effects) were combined
with secondary search terms sepsis/sepsis syndrome (diagnosis, microbiology, etiology), bacteremia/septicemia (diagnosis, microbiology, etiology), blood specimen collection (methods), bacteriological techniques and diagnostic tests, routine (methods).
Ventilator-associated pneumonia
The primary search terms artificial respiration (adverse effects)
ventilation, mechanical (adverse effects) and intensive care (methods) were combined with the secondary search terms pneumonia
(diagnosis, microbiology, etiology), sepsis/sepsis syndrome (diagnosis, microbiology, etiology), bacteremia/septicemia (diagnosis,
microbiology, etiology), bacteriological techniques and diagnostic
tests, routine (methods).
Surgical site infections and intra-abdominal sepsis
The primary search terms wound infection (diagnosis, microbiology), abscess (diagnosis, microbiology) and abdomen were combined with secondary search terms sepsis/sepsis syndrome (diagnosis, microbiology, etiology), bacteremia/septicemia (diagnosis, microbiology, etiology), bacteriological techniques, diagnostic tests,
routine (methods) and interventional radiology.
General considerations in the diagnosis of sepsis
There are three key difficulties associated with the diagnosis of infection in patients who have sepsis.
Establishing infection as the primary cause
The controversies surrounding the definition of sepsis
are discussed elsewhere (see Matot and Sprung, ªIntroductionº). Establishing that the patient has an ongoing
infection ± and therefore has sepsis rather than a noninfective cause of the systemic inflammatory response
syndrome (SIRS) ± can be extremely difficult. An important first step is a systematic consideration of possible noninfective causes. Knowledge of other pathologies that may mimic sepsis and how they apply to the
specific patient can make up for a relative paucity of
clinical information in patients who may be sedated or
critically ill. Noninfective causes of SIRS are the following:
· Tissue injury
· Surgery/trauma
· Hematoma/venous thrombosis
· Myocardial/pulmonary infarction
· Transplant rejection
· Pancreatitis
· Erythroderma
· Metabolic
· Thyroid storm
· Acute adrenal insufficiency
· Therapy related
· Blood products
· Cytokines, especially granulocyte-macrophage colony stimulating factor
· Anasthetic-related malignant hyperpyrexia, especially halothane
· Neuroleptic malignant syndrome, for example,
caused by haloperidol
· Opiates/benzodiazepines
· Malignancy
· Hypernephroma/lymphoma
· Tumor lysis syndrome
· Neurological
· Subarachnoid hemorrhage
Tissue infarction or hematoma, for example, may need
to be actively sought in a surgical or trauma patient
who develops signs of SIRS and precipitation of thyroid
storm or adrenal insufficiency should be considered in
at-risk patients following trauma or surgery.
Localizing the site of infection
This may be straightforward, but frequently it is confounded by the fact there are multiple pathological processes occurring concurrently or by the frequent use of
antibiotics which undermine microbiological diagnosis.
Occasionally the site of infection is occult, for instance,
when there is sinusitis or deep intra-abdominal infection.
Interpreting the microbiological findings
Conventional microbiology has several limitations in
hospitalized patients who may be septic:
·
·
·
·
·
Distinguishing colonization from infection
Prior antibiotic therapy
Correctly identifying unfamiliar organisms
Determining the significance of mixed culture results
Interpreting the importance of organisms normally
of low virulence
Principal among these is the fact that many organisms
isolated from nonsterile sites may represent either colonization or infection ± microbiology alone cannot answer this question. Conversely, the microbiology laboratory may report negative findings in samples from sites
that are in fact infected, either because antibiotics have
sterilized the specimen, or because special procedures
need to be carried out (e.g., immunofluorescence to detect Pneumocystis carinii).
S 12
A clinical approach
Fever is a common sign in hospitalized patients and is
often the first indication of sepsis. Practical guidelines
for the evaluation of fever on the ICU have recently
been published [4]. Focused clinical examination, guided by risk factors relevant to the individual patient, often reveal potential sources of sepsis and guide subsequent investigation. Surgical and traumatic wounds
should be exposed and examined for signs of infection.
Particular attention should be paid to vascular access
sites for signs of phlebitis or cellulitis and to pressure areas or injection sites for evidence of soft tissue infection.
Evidence of sinusitis should be sought, and fundoscopy
is invaluable in detecting candidal endophthalmitis, a
pathognomic feature of systemic fungal sepsis. Urine in
the catheter may be frankly purulent, and the presence
of diarrhea may indicate Clostridium difficile associated
colitis. The importance of repeated, complete physical
examination to detect the emergence of new signs cannot be overstated.
Nonspecific markers of infection
Traditional markers of infection such as neutrophilia
lack sufficient sensitivity among hospitalized patients
to be of value in distinguishing sepsis, although marked
neutrophilia or failure to mount a neutrophil response
may be of prognostic value. Levels of procalcitonin
(PCT) and C-reactive protein (CRP) are straightforward to assay. The evidence that these acute-phase
markers have specificity in differentiating infection
from other causes of an inflammatory response has recently been reviewed [5]. Levels of CRP and PCT are
correlated well with the degree of inflammatory response and are of particular value in monitoring response to treatment [6]. PCT may have some advantages over CRP in that it rises more quickly at the onset
of inflammation and is cleared more quickly as inflammation resolves [7]. Levels of PCT are correlated
more closely with severity of sepsis [8] and also are predictive of mortality [9]. A prospective study of ICU patients found that a CRP level of 50 mg/l or higher had a
sensitivity of 98.5 % and specificity of 75 % in identifying probable or definite sepsis [10]. De Werra et al.
[11], also in an ICU population, found PCT levels of
1.5 ng/ml or higher to have a sensitivity of 100 % and
specificity of 72 % in identifying sepsis. Such markers
therefore cannot alone differentiate sepsis from other
causes of SIRS; rather they are a part of a systematic
evaluation that includes clinical examination and directed diagnostic techniques. Daily, sequential measurement of inflammatory markers is likely to be of
more value in diagnosis of infection than single measurements [10].
Detection of circulating endotoxin might be expected to be a specific test for sepsis. Assays differ in the
sensitivity, cutoffs are not established, and the transient
nature of endotoxemia makes timing of measurements
essential. For these reasons measurement of endotoxin
levels in sepsis patients remains experimental.
Figure 1 outlines an algorithm for investigation of
suspected sepsis into which may be fed data from clinical examination and nonspecific investigations along
with the appropriate specific microbiology and imaging
investigations discussed in detail below.
Discussion: literature-based recommendations
To answer each of the following important clinical questions, a review of the literature was performed as previously described.
Bacteremia
Are there specific indications for obtaining blood for
culture?
Answer: yes; grade D.
Recommendation
Fever, chills, hypothermia, leukocytosis, left-shift of
neutrophils, neutropenia, and when infection is suspected, hypoalbuminemia, development of renal failure or
signs of hemodynamic compromise are specific indications for obtaining blood for culture. Blood cultures
should be taken as soon as possible after onset of fever
or chills.
Rationale
Blood should be obtained for culture whenever there is
reason to suspect blood stream infection, commonly
when a patient develops a new fever. In practice, as a
noninvasive, safe, and low-cost investigation, blood culture is often performed when there a few specific indications. However, a number of clinical and laboratory parameters are independently correlated with the presence of bacteremia in patients in whom infection is suspected, notably, chills, hypoalbuminemia, development
of renal failure, and a diagnosis of urinary tract infection
[12]. Other criteria are fever, hypothermia, leukocytosis,
left-shift of neutrophils, neutropenia and signs of hemodynamic compromise [13]. Ideally patients should not
be receiving parenteral antibiotics when blood cultures
are performed. While we are aware of no published
S 13
Fig. 1 Algorithm for systematic evaluation of nonneutropenic patients with suspected sepsis
data which directly address this issue, blood cultures
should be taken, where possible, immediately before a
regular dose of antibiotic so that blood levels are minimized. In suspected fungemia, therapy with antibacterial agents clearly should not impact on yield. Otherwise
the indications for performing blood culture are the
same irrespective of whether the patient is receiving antibiotics or not. In this group of patients, media containing antibiotic adsorbing substances such as BacT/Alert
FAN and BACTEC Plus/F should be used since they
are associated with increased recovery of significant
pathogens, particularly among patients on appropriate
antibiotic regimes [14].
The literature contains no clinical data relating to
timing of blood cultures with respect to timing of fever
or chills. Nevertheless, bacteria are rapidly cleared
from blood, and development of fever usually follows
an episode of bacteremia by 30±90 min. Published expert opinion is that blood cultures should be taken as
soon as possible following onset of fever [15].
Does the technique employed in obtaining blood
cultures influence the sensitivity and specificity of this
investigation?
Answer: yes; grade D.
Recommendation
Blood should be obtained by fresh venipuncture. Sites
associated with skin contamination (e.g., femoral site)
or loss of skin integrity (e.g., burns or dermatological
disease) should be avoided. Skin should be swabbed
twice with either 70 % isopropyl alcohol or with an iodine containing solution prior to venipuncture. The
blood culture stopper should also be sterilized prior to
inoculation. An adequate volume (20±60 ml) of blood
should be obtained per culture (10±30 ml per bottle). If
insufficient blood is available, only the aerobic bottle
should be inoculated. The needle used for venipuncture
should be changed prior to inoculation of blood into culture bottles.
Rationale
When a decision has been made to take blood for culture, adherence to a protocol for obtaining the specimen
results in lower contamination rates and improved yield
[13, 16]. The cost of additional investigations, treatment
and in-patient stay associated with each contaminated
blood culture has been estimated as between U. S.
$1,000 and $5,000 [17, 18]. Furthermore, with Grampositive organisms making up an increasing proportion
of significant blood culture isolates, identifying such isolates as contaminants is more difficult than ever.
Blood taken from a central venous catheter (CVC) is
significantly more likely to be contaminated by skin flora [19], as is blood taken from previously placed periph-
S 14
Table 1 Trials comparing skin
sterilization techniques
Protocol
Conclusion
Reference
10 % povidone iodine and 0.2 % chlorine
peroxide
No benefit
21
Alcohol then povidone iodine compared
with alcohol alone
No benefit
22
70 % iodophore compared with tincture
of iodine
Tincture of iodine associated with 50 %
fewer contaminated cultures
23
70 % isopropyl alcohol/70 % povidone
iodine compared with isopropyl alcohol,
10 % acetone and povidone iodine
dispenser
Isopropyl alcohol and 10 % acetone and
povidone iodine dispenser associated
with 50 % fewer contaminated cultures
24
eral cannulas, although contamination rates may not be
higher for blood obtained through a peripheral cannula
at the time of insertion, and this is acceptable as a means
of minimizing number of venipunctures [20]. It may be
possible to minimize contamination of blood obtained
through venous catheters by adherence to meticulous
sterile technique [21], but this approach should only be
used when no site for venipuncture is available.
Definitive evidence that skin disinfection reduces
blood culture contamination rates is lacking, but this
conclusion has been inferred from the findings of controlled trials which demonstrated superiority of one
skin preparation agent over another. Two other trials
have shown no benefit. One concluded that contamination occurs during laboratory specimen handling [22], a
second found unexpectedly low contamination rates in
both patient groups. The relevant trials since 1990 are
summarized in Table 1. We are aware of no reason to revise our earlier recommendations that skin should be
swabbed twice with either 70 % isopropyl or ethyl alcohol or with an iodine-containing solution [25]. The
blood culture stopper should also be sterilized prior to
inoculation of the blood sample. While we are aware of
no data that directly address this issue, it is the recommendation of previously published expert opinion [15].
Inoculation of 3 cfu into blood culture is required to
give 100 % culture positivity [26]. The concentration of
bacteremia in adult patients is frequently less than one
viable organism per milliliter [27] and may be less than
0.1 organisms per milliliter [28]. For these reasons it is
not surprising that the volume of blood inoculated into
culture is an important variable determining culture
yield [13, 29]. This effect has also been demonstrated in
clinical studies, although not specifically in sepsis patients [30, 31]. While the blood culture system employed
determines the volume of blood that may be utilized, in
adults a minimum sample size of 20 ml is required per
venipuncture (10 ml per bottle) [15, 32] while increasing
the sample volume above 30 ml is not associated with
significantly improved culture rates [13]. In infants, in
whom bacteremia is associated with higher number of
colony-forming units per milliliter, a smaller volume of
blood (0.5± ml or < 1 % of circulating blood volume)
can be used for culture [33]. Anaerobic organisms now
make up fewer than 5 % of blood culture isolates [34,
35]. Furthermore, aerobic culture bottles are more successful in culture of the overwhelming majority of organisms identified in blood [36]. Therefore, if insufficient blood is available to inoculate two culture bottles,
only the aerobic bottle should be inoculated.
Contrary to earlier reports [37, 38, 39], a recent metaanalysis has demonstrated lower contamination rates if
a needle change is performed [18]. Guidelines regarding
resheathing of needles must be strictly adhered to in order that the associated increased risk of needle-stick injury be avoided.
Is there evidence to determine how many sets of blood
cultures should be taken?
Answer: yes; grade E.
Recommendation
A minimum of two and maximum of three sets of blood
cultures should be obtained for each episode of suspected bacteremia.
Rationale
When bacteremia is associated with endocarditis, if one
culture is positive, the probability of any subsequent
culture being positive exceeds 95 %. In bacteremia associated with other sources of infection, sensitivity exceeding 99 % is reached with either two [40] or three
[41] cultures. The taking of only one culture is rarely
permissible since the rate of contamination of an individual set of blood cultures is finite, ranging from 1 %
to 4.5 % [22, 23, 24]. Interpretation of a single isolate of
a potentially contaminating organism may therefore be
exceedingly difficult.
S 15
Is there evidence that temporal separation of blood
cultures is valuable?
Answer: no; grade D.
Recommendation
In critically ill patients, in whom it may not be possible
to delay treatment, no interval is required between taking sets of blood cultures.
initions of CRB and CRS. In particular, most published
studies that appear to show that it is possible to diagnose
catheter infection without catheter removal for culture
in reality only address cases of CRB.
Can CVC infection be identified as a source of
bacteremia without resorting to catheter removal for
culture?
Answer: yes; grade C.
Rationale
Recommendations
In patients who are not critically unwell, published expert opinion has been that a 30- to 60-min interval
should be allowed between obtaining sets of blood cultures [42]. However, the only published study directly
to address the efficacy of serial versus simultaneous
blood cultures, demonstrated that drawing blood cultures simultaneously or at intervals over a 24-h period
did not effect yield [31].
When a CVC is suspected as a source of bacteremia, diagnosis of CVC infection may be made by blood culture
based techniques if (a) the patient's clinical condition
permits a potentially infected line to be left in place,
(b) treatment of CVC infection is to be attempted, or
(c) other potential sources of bacteremia are apparent.
While the acridine orange leucocyte cytospin (AOLC)
test offers the possibility of virtually immediate diagnosis, on the basis of currently available data its use should
remain experimental.
Central venous catheter infection
Definition of terms
Rationale
It is clinically important to distinguish catheter-related
bacteremia (CRB) from catheter-related sepsis (CRS).
CRB is defined by the presence of three criteria [43]:
When a patient who has a CVC in place develops bacteremia, the likelihood that the CVC is the source of the
bacteremia depends on the organism cultured. One
study of 311 patients who had CVCs found that 73 % of
bacteremias were related to CVC infection, and that if
the culture was of S. aureus, this figure rose to over
92 % [48]. Other organisms that are particularly associated with CVC infection as a source are coagulase-negative staphylococci, Corynebacterium (especially JK-1),
Bacillus species and fungi, in particular, Candida species
[49]. Two approaches to diagnosing CVC infection as a
source of bacteremia, without catheter removal for culture, have been reported in the literature: (a) culture of
CVC and peripheral blood and (b) AOLC test.
· Positive catheter culture.
· Positive peripheral blood culture taken before catheter removal.
· The same micro-organism is identified in each of the
above.
A positive catheter culture has been defined [44] as the
presence of 15 colonies or more on semiquantitative culture of the catheter tip [45] or of 10 colonies or more on
quantitative culture [46]. The various laboratory techniques used to culture CVCs is beyond the scope of this
article but have recently been reviewed [47].
CRS is defined as a positive catheter culture when
this is considered to be the source of the patient's sepsis,
but bacteremia does not occur [44]. The term catheterrelated blood stream infection is sometimes used, and
refers to cases in which peripheral cultures are positive,
but catheter tip cultures do not need to meet culture criteria as long as there is indirect evidence that the catheter is the source of infection, for example, defervescence
following catheter removal.
Unfortunately, many of the studies pertaining to diagnosis of CVC infection do not strictly adhered to def-
Culture of central venous catheter blood
This approach is based on the fact that the concentration of bacteria drawn through an infected catheter is
between 4 [50] and 30 [51] times higher than the concentration in peripheral blood drawn simultaneously. The
use of this approach in clinical diagnosis has been assessed in two ways. First, by quantitative or semiquantitive blood culture and, secondly, in continuous monitoring blood culture systems. Both methods rely on it being
possible to aspirate at least 20 ml blood through the
S 16
Table 2 Studies of continuous
blood culture monitoring (CRB
catheter-related bacteremia)
Table 3 Studies of acridine orange leukocyte cytospin
(AOLC) in adult patients
(CRB catheter-related bacteremia, CVC central venous catheter, HDU high dependency
unit)
Population
Finding
Reference
Retrospective review of 7 patients with
suspected line related infection
Time to positivity shorter for central
catheter cultures than cultures from
peripheral sites
59
Retrospective study of 11 patients in
whom CRB was diagnosed
Time to culture positivity 1±24 h earlier
for central peripheral catheter cultures
61
Retrospective analysis of 64 cancer
patients with suspected CRB
28 cases of CRB were identified; a cutoff
of +120 min had 100 % specificity and
96.4 % sensitivity
62
Prospective study of 93 cancer patients
on an ICU
28 cases of CRB were identified; a cutoff
of +120 min had 91 % specificity and
94 % sensitivity
63
Setting
Finding
Reference
100 patients with suspected CRB,
most on parenteral nutrition
35 cases of CRB, AOLC positive in 2/17;
when used in conjunction with endoluminal brushing, AOLC identified 15/18 and
was 100 % specific
4 cases of CRB, AOLC positive in 2; 10
colonized catheters, AOLC positive in 2
12 cases of catheter-related blood-stream
infection identified, AOLC negative in all
64
50 cases of CRB, AOLC was 96 %
sensitive and 92 % specific
67
55 CVC tips from ICU patients
49 patients with suspected catheterrelated blood-stream infection; mixed
ICU/HDU
128 surgical patients with suspected
CRB
catheter in question, and it is estimated that blood cannot be aspirated from 12±50 % of potentially infected
catheters [43, 52]. Furthermore, being culture based
methods, both involve a delay of up to 48 h before cultures can be expected to become positive, time when a
potentially infected catheter remains in situ.
Quantitative blood cultures
A series of studies in the 1980s demonstrated that quantitative cultures of central and peripheral blood can be
used to diagnose CRB [53, 54, 55]. In the literature since
1991 two studies have attempted to determine the ratio
of colony-forming units per milliliter that gives optimal
sensitivity and specificity. Capdevila et al. [50] found
that using a cutoff of 1:4 for the ratio of bacterial colonies in peripheral to central blood, sensitivity and specificity of 94 % and 100 %, respectively, could be
achieved. Quilici et al. [56] used a ratio of 1:8 and found
sensitivity of 92.8 % and specificity of 100 % in a prospective study of critical care patients. Published expert
opinion is that quantitative culture of central and peripheral blood samples showing a greater than 5:1 ratio
suggest CRB [57, 58]. A meta-analyis of various catheter culture techniques concluded that quantitative blood
65
66
cultures were more cost effective than any culture technique involving the catheter itself [47].
Continuous blood culture monitoring
In this approach the time taken for a blood culture to
become positive is related to the number of micro-organisms initially present. The higher the initial concentration, the faster the cutoff point is be reached to determine positivity [59]. Four clinical studies have been performed using paired central and peripheral blood samples and one of the commercially available continuous
monitoring blood culture systems (e.g., BacT/Alert)
which employs a colorimetric CO2 sensor [60]; findings
are summarized in Table 2.
Acridine orange leukocyte cytospin test
This is currently the only good candidate for a rapid diagnostic test for CRB. The test requires only 50 ml blood
to be withdrawn from a CVC and takes around 30 min
and minimal specialist laboratory expertise. We are
aware of four clinical trials in adults (Table 3). Two
have yielded positive results [64, 67], both from the
S 17
Table 4 Factors associated with risk of central venous catheter (CVC) infection (CRB catheter-related bacteremia)
References
Catheter site
Subclavian lines are associated with significantly less catheter site colonization and
infection than jugular or femoral lines although less CRB has not been demonstrated.
Difficult catheter insertions with multiple attempts are associated with higher rates
of infection.
71, 74, 75, 76
Catheter type
It has been suggested that multilumen catheters may be associated with a higher
risk of infection than single lumen catheters possibly because they are handled more
frequently. This effect is not apparent in critical care patients. Tunnelled lines are
associated with delayed and fewer incidences of CRB.
72, 74, 78, 79, 80
Antiseptic impregnated
and antibiotic coated
catheters
There are conflicting studies assessing the extent to which incorporating antimicrobial
agents into the manufacture of catheters reduces infection rates. When assessed in
critical care patients, such catheters are associated with rates of colonization, local
infection and CRB reduced by up to 70 %.
When CVCs are used for parenteral nutrition this benefit may be lost.
In an ICU setting the use of CVCs for parenteral nutrition has been shown to increase
colonization rates. In a non-ICU setting higher infection rates have been found in
CVCs used for parenteral feeding.
The risk of a central venous catheter being infected increases as duration of catheterization increases beyond 1 week.
Meta-analysis data suggest that ªguide-wireº replacement of CVC may be associated
with a higher infection rate than replacement to a new site.
The infection rate for a patient's first CVC is significantly lower than for any subsequent catheter whether replaced by ªguide-wireº or at a distant site.
81, 82, 83, 84
Trauma and burns patients at high risk, neurosurgical patients at lowest risk.
71
Catheter insertion
Catheter use
Duration of catheterization
Previous catheters
Underlying disease
same group in Leeds, while two from elsewhere [65, 66]
both produced disappointing results. None has specifically investigated sepsis patients, and only one specifically involved critical care patients [65]. Two small studies, again from the group in Leeds, have demonstrated
that culture of a wire brush passed down the lumen of a
potentially infected catheter is correlated well with subsequent culture of the catheter tip. The first looked at
115 CVCs used for parenteral nutrition in general surgical patients [68]. The second reported 95 % sensitivity
and 84 % specificity among 22 surgical patients subsequently diagnosed with CRB [69].
Can CVC infection be identified as a source of sepsis in
nonbacteremic patients without catheter removal for
culture?
Answer: no; grade E.
Recommendation
When a CVC is suspected as a source of sepsis in nonbacteremic patients, definitive diagnosis requires that
the CVC should be removed and sent for culture.
77
85, 86
74, 87
74, 88, 89
90
91
Rationale
Only a proportion (10±72 % depending on organisms involved) of infected CVCs are associated with bacteremia [70]. When critical care patients develop signs of
sepsis, even in the absence of bacteremia, replacement
of central venous catheters is frequently advised. However, using standard culture techniques, rates of proven
catheter infection range from 8.9 % to 26 % [66, 71, 72],
and therefore many more catheters are removed than
are infected [66, 69, 73]. The decision whether to remove a CVC in this setting, where blood cultures are
negative, is essentially a clinical one, and involves
weighing up risk of catheter replacement against risk of
leaving in place a potential source of infection. To help
guide this decision a range of clinical parameters have
been suggested as possible correlates of catheter infection. A number of individual associations have been
identified, summarized in Table 4, and these have been
used to make recommendations concerning CVC insertion and care [44]. Two studies have assessed the use of
such parameters in guiding clinical judgement and have
failed to show they are of value in increasing the proportion of removed CVCs that are infected [66, 92]. Clinical
criteria have been incorporated with microbiological
data and clinical response to catheter removal into a
scoring system [93] which appears more sensitive and
no less specific than the Hospital Infection Control
Practices Advisory Committee diagnostic criteria [94]
S 18
but nevertheless requires removal of the catheter for
culture.
Is infection at the catheter insertion site indicative of
CVC infection?
Answer: yes; grade C.
Recommendation
If infection is suspected at the catheter site, swabs
should be taken from the insertion site for culture. The
presence of purulence at the CVC site should prompt
catheter replacement at a distant site irrespective of culture results.
Rationale
We are unaware of any studies of insertion site skin culture as a predictor of catheter-related infection in sepsis
patients. In patients with long-term nutrition catheters,
negative site cultures had a negative predictive value
for line infection of 98 % [95] while positive culture, particularly of organisms other than coagulase-negative
staphylococci is predictive of CVC infection [96]. In patients with nontunneled CVCs, quantitative skin cultures at the time of removal for suspected infection had
a sensitivity of 75 %, a positive predictive value of
100 % and a negative predictive value of 92 % in detecting CVC infection [97].
Expert opinion in the literature is generally that inflammation and frank purulence around a catheter insertion site is a predictor of CVC infection [43, 57, 98].
Reed et al. [44] recommend that ªif the site appears to
be infected, the catheter is removed.º While local evidence of infection is predictive of systemic infection, it
is possible to find marked local signs of infection with
no evidence of systemic infection [48].
Ventilator-associated pneumonia
This section addresses specifically the diagnosis of ventilator-associated pneumonia (VAP) in sepsis patients.
The diagnoses of community-acquired pneumonia [99,
100, 101] and pneumonia in the immunocompromised
host [102] have recently been reviewed by others. The
merits of various invasive diagnostic approaches to
VAP have also been the subject of several recent editorials and reviews [103, 104, 105, 106, 107]. We therefore
specifically exclude from this discussion a review of the
extensive literature, which assesses the different modes
of invasive airway sampling. Nevertheless, the question
ªIs the lower respiratory tract the source of this patient's
sepsis?º will, particularly in the ICU frequently apply to
patients who develop sepsis while ventilated. We focus
then on the evidence to support the recommendation
of particular diagnostic tests in management of sepsis
patients in whom VAP is suspected as the source.
Can clinical parameters be used to diagnose pneumonia
as a source of sepsis in a ventilated patient?
Answer: uncertain; grade D.
Rationale
The clinical diagnosis of pneumonia in the non-ICU patient is usually based on the presence of fever, leukocytosis, purulent sputum and new radiographic infiltrates
in such patients these criteria are sensitive and specific.
In intubated patients these parameters are too nonspecific to be of diagnostic value. Purulent secretions, for
example, are almost inevitably found in patients receiving prolonged mechanical ventilation and do not specifically indicate the presence of pneumonia [108].
A range of risk factors for the development of VAP
have been identified. Cumulative incidence of VAP increases with time following intubation, but the daily increase in risk diminishes over time [109], such that rates
are approximately 3 % per day in the first week, 2 % per
day in the second and 1 % per day thereafter [109]. Other independent risk factors recently reviewed include
witnessed aspiration, neurological disease and administration of a paralyzing agent impairing airway reflexes,
presence of a nasogastric tube, enteral feeding and
drugs used to raise gastric pH [110]. One study used by
multivariate analysis of clinical parameters to generate
a scoring system for risk of developing nosocomial
pneumonia in ICU patients. In the patients studied the
scoring system had a sensitivity of 85 % and a specificity
of 66 % [111]. We are not aware of any prospective,
comparative data to assess the usefulness of such a system in clinical practice; hence the answer to the question remains ªuncertain.º
Should blood cultures be obtained in patients in whom
VAP is suspected?
Answer: yes; grade E.
Recommendation
Two sets of blood cultures should be sent in patients
with suspected VAP.
S 19
Rationale
Recommendation
Blood cultures are neither sensitive nor specific in the
diagnosis of VAP. Between 3 % and 12 % of bacteremias
which occur in ICU patients have a respiratory tract
source [112, 113], but only one-quarter of cases of VAP
are associated with bacteremia [114]. Bacteremia in patients with suspected VAP, in reality, usually arises
from outside the chest [114, 115]. Meduri et al. [115]
have demonstrated that two-thirds of patients with
nosocomial pneumonia have at least one other focus of
infection, usually urinary or CVC related. For this reason published expert opinion is that blood cultures are
an essential part of the work up of a patient with suspected VAP [116].
Pleural effusions larger than 10 mm should be aspirated.
Samples should be sent for immediate Gram and fungal
stains, culture and biochemistry including protein, lactic
dehydrogenase and glucose. Paired blood chemistry
samples should also be sent for comparison.
Are new chest radiographic infiltrates diagnostic of
pneumonia in a ventilated patient?
Answer: no; grade D.
Recommendation
A chest radiography should be performed.
Rationale
The development of a new chest radiographic infiltrate
in a ventilated patient may have many causes other
than infection [104, 117]. When assessed against diagnosis by bronchoscopy [118], final clinical diagnosis [104]
or postmortem histology [119], chest radiographic changes alone are insufficiently specific for diagnosis of
pneumonia in this group of patients. Computed tomography (CT) is more sensitive in detecting lung parenchymal changes than plain radiography and may better
demonstrate fluid collections [120]. Even so, most causes of diffuse air-space shadowing cannot be reliably differentiated on CT, which therefore adds little diagnostic
information in suspected VAP over and above plain radiography [121]. Not withstanding this lack of diagnostic
specificity, chest radiography may provide valuable information, for example, to guide invasive diagnostic approaches and detect pleural effusion.
Should thoracocentesis be performed in patients with
pleural effusions in whom VAP is suspected?
Answer: yes; grade E.
Rationale
Pleural effusions are uncommon in VAP, and empyema
develops rarely. We are not aware of any data which determine rates at which diagnostic information is gained
from analysis of pleural fluid in the context of VAP. Our
recommendation is in line with published guidelines of
the American Thoracic Society [116]. The presence of a
parapneumonic effusion larger than 10 mm warrants diagnostic thoracocentesis. Parameters suggestive of an
underlying pneumonia include: white blood cells higher
than 5 ” 109/l, more than 50 % polymorphonuclear cells,
organisms seen on Gram stain, low glucose (< 40 g/dl),
pH less than 7.3 and biochemical criteria of an exudate
(protein > 3 g/l, raised lactic dehydrogenase) [122].
Do serological tests have a role in diagnosis of VAP?
Answer: no; grade E.
Recommendation
Serology is not routinely indicated in the diagnosis of
VAP.
Rationale
With the possible exception of Legionella infection,
ªatypicalº organisms are not causes of VAP. Although
nonepidemic Legionella infections made up 22 out of
286 episodes of hospital acquired pneumonia in one
study, none developed in patients who were already
ventilated [123]. Legionella infection developing in a
ventilated patient would raise the possibility of acquisition from within the ICU. In rare cases in which Legionella infection is suspected, urinary antigen testing
provides rapid and accurate diagnosis and is now beginning to replace serology.
Should tracheal aspirates be obtained in patients in
whom VAP is suspected?
Answer: yes; grade C.
S 20
Table 5 Studies assessing diagnostic accuracy of endotracheal
aspirates (ETA) in diagnosis of
ventilator-associated pneumonia (VAP) (BAL bronchoalveolar lavage, PSB protected
specimen brush)
Protocol
Finding
Reference
12 patients with suspected VAP; compares semiquantitative culture of ETA,
PSB and BAL
Similar range and quantities of organisms
recovered by all techniques
127
52 patients with suspected VAP; compares ETA (cutoff 106 cfu/ml) and PSB
Sensitivity 82 % vs. 64 %; specificity 83 %
vs. 96 %
128
26 patients with ªdefinite VAP,º 48
ªpossible VAP,º 28 controls; compares
ETA (cutoff 105 cfu/ml) and PSB/BAL
Sensitivity 70 % vs. 60 %/57 %; specificity
72 % vs. 93 %/87 %
129
28 patients with suspected VAP; compares ETA (cutoff 105 cfu/ml) and
PSB/BAL to postmortem histology
Sensitivity 63 % vs. 57 %/47 %; specificity
75 % vs. 88 %/100 %
130
Recommendation
A sample of secretions aspirated via the endotracheal
tube should be sent for Gram stain and for bacterial
and fungal culture.
Rationale
Studies which assess different approaches to the microbiological diagnosis of VAP have been hampered by
the fact that there is no accepted diagnostic ªgold-standard.º Postmortem histology and quantitative tissue
culture (104 cfu/g tissue) are generally regarded as the
most precise techniques available but are often not
practicable for use in clinical studies. Furthermore,
characteristic histological changes in pneumonia may
be found in lung which is sterile in culture, and bacterial
counts up to 104 cfu/g lung tissue may be found in the absence of histological changes in pneumonia, due to rapid
proliferation of organisms after death (reviewed in
[124]). Published studies therefore frequently make
comparisons between different techniques or use clinical response as confirmation of the diagnosis.
The microbiology of endotracheal aspirates (ETA)
exemplifies the problem of distinguishing colonization
from infection in an ICU setting. Bacterial colonization
of the lower respiratory tract is almost universal following intubation [125]. Consequently, negative ETA cultures have powerful negative predictive value in the diagnosis of VAP. When pneumonia is present, the causal
agent is usually present in nonquantitative culture of
ETA [126, 128, 129]. True pathogens may, however, be
missed in culture if more numerous but possibly less
pathogenic organisms over-grow the plates. Gram staining is an essential component of the evaluation of respiratory tract specimens. The presence of squamous cells
(> 10 per high power field) and the absence of leukocytes (< 25 per high power field) suggests the specimen
is contaminated with saliva and unsuitable for culture.
The finding of certain organisms may influence initial
choice of antibiotics. For example, large numbers of
clustered gram positive cocci may emphasize the need
to cover S. aureus pending culture results.
Table 5 summarizes the studies of ETA in diagnosis
of VAP since 1991. What is clear from these data is that
ETA has the advantages of not only being the least invasive means of sampling the respiratory tract but also the
most sensitive. This is a crucial point in favor of use of
ETA in that the outcome of VAP is correlated with the
adequacy of the initial antibiotic regimen used. Where
the initial antibiotic regimen fails to cover pathogens
subsequently identified by microbiology, changes aimed
at broadening coverage of pathogens does not improve
outcome [131]. ETAs may be of particular value combined with clinical and radiographic parameters in a formal scoring system [132].
Should lower respiratory tract specimens be obtained
routinely for microbiology in suspected VAP?
Answer: yes; grade B.
Recommendations
Samples from the lower respiratory tract should be obtained for microbiology. No significant advantage of
one invasive diagnostic approach over another has
been consistently demonstrated. Choice of technique
depends in practice primarily on available expertise
and equipment.
Rationale
Sampling of lower respiratory tract secretions for microbiology may impact on the management of VAP in several ways. First, the range of bacteria which cause VAP,
and their susceptibility patterns, varies widely between
different hospitals [133]. For this reason knowledge of
S 21
Fig. 2 Algorithm for diagnosis
of ventilator associated pneumonia
pathogens present in an individual hospital has great importance in choice of empirical antimicrobial regimens.
Secondly, the microbiological data may alter the outcome. A number of studies on the impact of invasive diagnostic techniques on mortality of VAP have demonstrated that invasive techniques often trigger a change
in the antibiotic regimen but have failed to provide conclusive evidence of associated improvement in outcome
[131, 134, 135, 136, 137]. However, the balance of evidence is beginning to accumulate in favor of invasive
sampling. A preliminary report of a prospective randomized controlled trial comparing invasive to noninvasive strategies demonstrated that the invasive approach
is significantly better [138]. A large randomized controlled trial recently been completed in France has demonstrated reduced 14- and 28-day mortality and reduced
antibiotic use associated with invasive as oppose to noninvasive diagnosis of VAP [139]. Direct examination of
a good-quality specimen helps to guide initial empirical
therapy, and culture results allow subsequent modification of antibiotic regimen.
The information obtained by invasive sampling of the
lower respiratory tract secretions might be expected to
allow use of broad spectrum antibiotics to be restricted.
The studies cited above, however, show that in clinical
practice this is rarely the case. Only in a small minority
of cases do the changes in antibiotic regimen, which follow lower respiratory tract sampling, result in the use of
narrower spectrum drugs. This need not be so. Where appropriate cutoff values for quantitative culture are set,
sensitivity of bronchoalveolar lavage and protected specimen brush has exceeded 90 %. Two studies addressing
outcome in patients with suspected VAP who have had
antibiotics withdrawn on the basis of bronchoscopy findings showed that there was no increase in mortality associated with this strategy [140, 141].
A systematic approach to the diagnosis of VAP in
sepsis patients is described in Fig. 2. When a ventilated
S 22
patient develops sepsis, blood cultures should be drawn
and chest radiography performed. A diagnostic bronchoscopy should be performed without delay unless either facilities or expertise are not available, or the procedure is contraindicated. In such cases diagnostic endotracheal aspiration is indicated. Respiratory samples
should be sent for direct examination and the result
used to guide choice of antibiotic therapy. When culture
results become available, the antibiotic regimen may be
modified.
Should blood cultures be obtained in cases of suspected
surgical site infection or deep abdominal collection?
Answer: yes; grade E.
Recommendation
Two sets of blood cultures should be obtained.
Rationale
Surgical site infection and intra-abdominal sepsis
National Nosocomial Infections Surveillance definitions
for surgical site infection (SSI) have been in use in the
United States for a decade [142]. These are as follows
[143]:
· Superficial SSI: Occurs within 30 days, involves skin
or subcutaneous tissue of the incision, and any one
of the following; purulence, organisms cultured from
aspirate or biopsy specimen, clinical signs of local infection, diagnosis as SSI by attending clinician. Not
stitch abscess.
· Deep SSI: Occurs within 30 days, or up to 1 year if
implant in place, infection appears to relate to surgery and involves fascia and deep muscle layers and
any one of the following; purulence, dehiscence of
deep incision, abscess found at reoperation, radiology or histology, diagnosis as deep SSI by attending
clinician.
· Organ/space SSI: Occurs within 30 days, or up to
1 year if implant in place, infection appears to relate
to surgery, involves any part of the body other than
the incision, and any of the following; purulence, organisms cultured from aspirate, abscess found at reoperation, radiology or histology, diagnosis as organ/
space SSI by attending clinician.
Similar definitions have recently been adopted in Europe and the United Kingdom [144]. The literature on
microbiological diagnosis of surgical site and wound infection has recently been reviewed [145], and we know
of no more recent, relevant data. In brief, while culture
of bacteria from an aseptically collected sample of
deep fluid or tissue is diagnostic of infection, the contribution of qualitative culture of wound swabs is limited
by inevitable contamination of any open wound. Certain organisms such as b-hemolytic streptococci can be
considered as pathogenic when present at any concentration. Otherwise, on culture of tissue biopsy, growth
of more than 105 bacteria per gram of tissue is considered diagnostic of wound infection.
Superficial SSI rarely causes sepsis and uncommonly
bacteremia. Although making up fewer than 5 % of all
causes of bacteremia among hospitalized patients [113,
146], cases of deep SSI and localized intra-abdominal
sepsis are frequently associated with bacteremia. Furthermore, empirical antibiotics may need to be given
before samples from the suspected site of infection itself
are available. However, such deep infections are frequently polymicrobial, comprising fecal organisms, and
blood cultures may not identify the full range of organisms involved, particularly anaerobes.
Are there specific indications for obtaining wound
swabs or specimens of drain fluid?
Answer: yes; grade E.
Recommendations
The presence of purulence or spreading cellulitis are indications for taking wound swabs. Infection should be
suspected particularly at ªcontaminatedº or ªdirtyº surgical sites
Rationale
Certain clinical changes imply that a superficial surgical
site has become infected. Discharge of purulent fluid is
diagnostic of SSI, and spreading inflammation, in excess
of that seen in normal healing, is present. The development of these features in the first 48 h after surgery or
trauma (ªearly infectionº) suggests the presence of infection by virulent organisms such as b-hemolytic streptococci or Clostridium species. Most surgical site infections appear between the 4th and 6th postoperative
days (ªlateº) and are polymicrobial. The National Research Council wound definitions set out in 1964 continue to be of value in risk assessment of wound infection.
The category of wound is correlated well with the rate
of wound infection (Table 6).
S 23
Table 6 Wound categories and infection rates (modified from [147])
Wound category
Definition
Clean
Elective surgery, primary closure, no breach in sterile technique, no contamination
from potentially colonized body sites
Nonelective surgery, controlled opening of colonized body site, minimal breach in
sterile technique, reoperation through clean wound with-in 7 days
Clean contaminated
Infection rate
1.5 %
7.7 %
Contaminated
Nonpurulent inflammation at first surgery, major break in sterile technique or contamination from colonized body sites; penetrating trauma < 4 h old
15.2 %
Dirty
Purulent inflammation at first surgery, preoperative perforation of colonized body sites;
penetrating trauma > 4 h old
40 %
Can the contribution of anaerobic organisms to surgical
site infection be determined in routine practice?
Answer: no; grade D.
Recommendation
When contaminated or dirty abdominal wounds develop, features of wound infection, a diagnosis of anaerobic
coinfection should be assumed irrespective of whether
anaerobes are identified by routine microbiology.
Is there evidence to support the preference of particular
imaging modalities in the diagnosis of intra-abdominal
infection?
Answer: yes; grade E.
Recommendation
In most situations ultrasound is be the modality of first
choice. When ultrasound is not diagnostic, CT should
be considered.
Rationale
Rationale
Detection of anaerobic organisms in clinical specimens
is technically demanding. If anaerobic organisms are to
be cultured, specific measures may need to be taken in
obtaining samples, such as transporting pus in anaerobic
conditions. In the laboratory, culture techniques are
specialized and time consuming. For these reasons routine processing of samples in most microbiology laboratories does not include an extensive search for anaerobes.
While infections which develop in clean wounds are
frequently caused by skin flora such as S. aureus, when
contaminated or dirty wounds become infected it is possible to identify, at least one anaerobic organism in
65±94 % of samples [148]. Similarly, over 50 % of abdominal abscesses are polymicrobial, and almost 80 %
involve at least one anaerobic species [149, 150]. Consequently, while data from randomized controlled trials
are lacking, it is accepted best practice to cover anaerobic organisms when treating sepsis arising contaminated
or dirty surgical sites [151].
Gas liquid chromatography to detect bacterial shortchain fatty acids is a technique that allows rapid identification of anaerobes in a mixed culture. Although cost
currently limits the availability of gas liquid chromatography, it may in future replace culture-based methods
of anaerobe identification.
Plain radiography of the abdomen may reveal free gas
within the abdomen suggesting bowel perforation, or
demonstrate the presence of gas within an abscess, but
is only rarely yield definitive diagnostic information
[152]. In most patients further imaging, usually by ultrasound or CT, is necessary to localize a source of infection within the abdomen [153]. Ultrasound is readily
available, if necessary as a bed-side investigation. Its
limitations are that gas-filled loops of bowel, commonly
present in postoperative ileus may obscure underlying
pathology. Wounds and drains may make access to the
abdominal wall difficult. The sensitivity of ultrasound
is particularly operator dependent. CT, by comparison,
is more sensitive than ultrasound in detecting small foci
of infection, but in certain areas of the abdomen, particularly the pancreas, distinguishing an abscess from inflammation may be difficult by CT [154].
Magnetic resonance imaging is in turn more sensitive
than CT; a recent study reported 100 % sensitivity and
94 % specificity in detecting intraperitoneal abscess
[155]. In many situations, for example in ventilated patients, the use of magnetic resonance imaging is limited
in the diagnosis of sepsis by the need to keep all metallic
instruments away from the scanner's magnetic field.
S 24
Should abdominal fluid collections identified by
imaging be aspirated as a matter of routine?
Answer: yes; grade E.
Recommendation
Collections identified by radiology should, where technically possible, be aspirated and drained under radiological control, samples being sent for Gram-staining
and culture.
Rationale
Differentiation of infected material from hematoma or
inflammatory fluid is not possible on the basis of radiology alone [154]. Confirmation that infection is present
and identification particularly of any drug-resistant organisms depends on obtaining samples for microscopy
and culture [156, 157]. The pH of fluid obtained at ultrasound-guided aspiration (< 7.1) was found in one study
to be a sensitive marker (92 %) of the presence of infection [158] and has been suggested as a bed-side means of
identifying collections which require formal drainage.
Acalculous cholecystitis
Acute acalculous cholecystitis (ACC) is an infrequent
but probably underdiagnosed complication in critically
ill patients [159, 160]. It is caused by spontaneous gangrene of the gall bladder which without prompt diagnosis and treatment progresses to perforation. The cause
appears to involve infection by Clostridium perfringens
[161]. Although reported as a complication of a wide
range of critical illnesses, the majority of cases of AAC
follow trauma or biliary surgery [162, 163]. High doses
of narcotic agents may be a contributory factor [164].
Is there a standard approach to the diagnosis of
acalculous cholecystitis?
Answer: no; grade E.
Recommendations
ACC should be suspected in any sepsis patient, particularly postoperatively, when there are either signs relating to the right upper quadrant of the abdomen or obstructive liver function tests. When ACC is suspected,
ultrasound should be ordered urgently. If an initial ultrasound examination is not diagnostic, CT should be
performed. If CT is unavailable, a repeat ultrasound
should be performed after 24 h.
Rationale
Localizing right upper quadrant pain and tenderness is
often absent in sedated or ventilated patients suffering
from ACC. Diagnosis therefore requires a high index
of suspicion. The only differentiating features in a sepsis
patient may be elevation in alkaline phosphatase or gglutaryl transferase, in an at risk patient [165].
Although the ultrasound appearances of ACC (gallbladder distension, wall thickening and free fluid suggestive of perforation) are well established [166], such
changes are not diagnostic and are frequently demonstrable in critically ill patients who do not go on to develop ACC [167]. The value of a single ultrasound study
in diagnosis of ACC has been assessed in three retrospective studies. Two studies of ICU patients estimated
sensitivity at 76 % [168] and 92 %, with a specificity of
96 % [169]. More recently among 27 cases of ACC,
only half of which occurred in critically ill patients, a
sensitivity of 29 % was found [165]. Two prospective
studies looking specifically at the use of serial ultrasound examinations in patients with suspected ACC
have demonstrated that when initial diagnosis is uncertain, failure of any abnormalities to progress on followup scans has excellent negative predictive value [170,
171]. The role of CT in diagnosis of ACC has not been
thoroughly evaluated. Superior sensitivity of CT over ultrasound has been suggested by three retrospective
studies of patients with a surgical diagnosis of ACC
[165, 169, 172].
We are aware of two studies which examine the role
of laparoscopy in diagnosis and treatment of ACC. Neither study compared diagnostic accuracy with radiology.
In both the procedure was well tolerated. Laparoscopy
has the advantage over CT that it may, in some units,
be performed at the bedside and, if the diagnosis of
ACC is confirmed, can proceed directly to laparoscopic
cholecystectomy [162] or cholecystostomy [173].
Sinusitis
Since being first reported in 1974 [174], ventilator-associated sinusitis has become an increasingly well recognized cause of sepsis. It occurs usually but not exclusively in patients who have nasotracheal, as opposed to orotracheal intubation [175, 176]. The true incidence of sinusitis among critical care patients is hard to establish
since published estimates vary widely depending on the
population studied and the diagnostic techniques used.
In the only study to directly address the issue, one of
19 patients with occult sepsis on a surgical ICU had si-
S 25
nusitis as the sole focus of infection [177]. This topic has
recently been the subject of a general review [178].
Is there a standard approach to the diagnosis of
ventilator-associated sinusitis?
Answer: no; grade E.
Recommendations
Acute sinusitis should be suspected in any sepsis patient
who has either a nasotracheal tube or a fine-bore nasogastric feeding tube, or who has suffered a head injury.
When sinusitis is suspected, radiography of the maxillary sinuses should be performed to detect the presence
of fluid. When radiography does not demonstrate fluid
in the maxillary sinuses, CT should be performed. If either radiography or CT demonstrates the presence of
fluid, antral puncture should be performed to allow definitive diagnosis and therapeutic drainage before antibiotic therapy is initiated.
Rationale
If sinusitis is the source of sepsis, specific physical signs
are likely to be absent, although a mucopurulent nasal
discharge may be noted. For this reason the diagnosis
should be suspected in any patient who has a nasotracheal tube or an indwelling nasal device of any sort
(even a fine-bore nasogastric feeding tube), or who has
had a head injury.
The definitive investigation of antral sinus disease is
considered to be direct endoscopic examination [179].
In general the diagnosis is made on the basis of culture
of bacteria from purulent material obtained from the sinus cavities [180]. Because clinical evidence to support a
diagnosis of sinusitis is generally lacking in ICU patients, the first supportive evidence often comes from
radiology, either plain radiographic sinus views, ultrasound or CT. For each of these modalities there is a discrepancy between radiological diagnosis of sinusitis
(presence of fluid) and microbiological confirmation on
any subsequent aspirate. While plain radiography is of
value in diagnosis of maxillary sinusitis [181], five views
are required to achieve 88 % sensitivity [182]. Similarly,
ultrasound detected accumulation of fluid in the sinuses
of 15 of 100 patients in a consecutive prospective series
of intubated ICU patients but in only one of these patients could aspiration confirm sinusitis [183]. CT has
two advantages over sinus radiography and ultrasound.
CT is able to distinguish mucosal thickening from fluid
within the sinuses [184] and can assess the other paranasal sinuses which may, albeit less frequently, be infect-
ed in isolation [185]. The principal disadvantage is that
CT usually requires the patient to be moved from the
ICU. Although the ability of CT to detect mucosal abnormalities improves diagnostic accuracy, the discrepancy between CT diagnosis and diagnosis by antral
puncture remains significant [184, 186].
Since the discrepancy between radiological diagnosis
of sinusitis and confirmation on aspiration is unavoidable, abnormalities on radiology indicate further investigation by antral puncture to obtain fluid for culture.
Interpretation of culture results, however, requires caution. Hospitalized patients have heavy colonization of
the nose, and contamination of samples is virtually unavoidable [184]. In addition, while true infective sinusitis is frequently caused by mixed Gram-negative and
anaerobic infections, fluid obtained from antra that do
not have signs of infection quite frequently produces an
apparently significant culture result [187].
Invasive Candida infection
As a result of widespread use of broad-spectrum antibiotics and intravascular catheters, an increase in the incidence of nosocomial infection by Candida species has
occurred in both the United States and Europe [188,
189, 190, 191]. Although C. albicans remains the most
frequently isolated species, the incidence of other, potentially more drug-resistant species is also rising [192].
The importance of invasive fungal infection is further
underlined by the considerable attributable mortality,
38 % in one study [193] and 21.7 % in another [194].
Invasive Candida infections begin by colonization of
the gastrointestinal tract or skin [195]. Suppression of
indigenous intestinal bacteria allows overgrowth of
Candida in the gastrointestinal tract and mucosal adhesion. Once a critical level of colonization has been
reached translocation, across intact small bowel mucosa
may occur. In an ICU setting, where a range of physiological stresses may impair small bowel mucosal integrity, such translocation may occur at much lower concentrations. Similarly, skin colonization provides a source
for invasive disease when integrity is breached either
by intravascular catheters or burns. Finally, any disease
or drug which inhibits cellular immunity, for example,
diabetes mellitus, or corticosteroids, predisposes to invasive Candida infection [196].
Is there a standard approach to the diagnosis of invasive
candidiasis?
Answer: no; grade E.
S 26
Table 7 Examples of suggested risk factors for invasive Candida
infection and associated mortality (APACHE II Acute Physiology
and Chronic Health Evaluation II) (data summarized from: [194,
201, 202, 203, 204])
Examples of risk factors
for development of invasive
Candida infection
Examples of risk factors for
mortality associated with invasive Candida infection
Colonization with Candida
species.
Prior treatment with multiple
antibiotics
APACHE II score > 20
Delay between onset of candidemia and start of antifungal
therapy > 48 h
Total number of different
antibiotics > 2
Total number of days on
antibiotics > 14
Prior Hickman catheter
Prior hemodialysis
Recommendations
There are no data to support a policy of routine screening of hospitalized patients for Candida colonization.
However, in sepsis patients invasive fungal infection is
more likely in patients who are heavily colonized.
When sepsis develops in patients colonized by Candida
species at two or more sites, blood cultures should be
sent and lysis centrifugation performed if available. Isolates of Candida species from sterile sites should be sent
for speciation and sensitivity testing.
Rationale
Clinical features of invasive Candida infection are in
most cases nonspecific, ranging from unexplained fever
to sepsis [197]. Specific clinical manifestations are rare.
Candidal chorioretinitis occurs in fewer than 15 % of
candidemic patients [198], but when found is an absolute indication for initiation of antifungal therapy. Skin
lesions and septic arthritis occur less frequently still
[199]. Retrospective studies have demonstrated a range
of risk factors for invasive Candida infection (Table 7).
One small prospective study has suggested that after
4 days of persisting fever despite antibacterial antibiotics, antifungal agents should be started empirically
[200]. We are not aware of any studies that address a
risk stratification approach to diagnosis of fungal infection.
Development of invasive Candida infection is correlated with preceding colonization [201]. Sites at which
colonization may be detected include urine, rectum,
gastric aspirate, vascular access sites, sputum/throat
swab, wounds and surgical drains. The number of sites
has been found to be correlated with the risk of develop-
ing invasive fungal infection [202, 203] A cutoff of two
sites colonized, as an indication for beginning empirical
antifungal therapy, has a high sensitivity but low specificity, 22 % in one study [203]. This may in part be because using the number of colonized sites alone as a
measure of infection risk fails to take into account the
intensity of colonization. Using semiquantitative culture
techniques to produce a ªcorrected Candida colonization indexº; sensitivity and specificity of 100 % was
achieved in a retrospective analysis of 29 critical care
patients [203]. No evidence exists regarding how frequently samples should be taken to detect colonization
in at-risk patients. Five days would seem a reasonable
interval. At present there is no consensus on the value
or use of routine screening for Candida in hospitalized
patients, and the data are inadequate to support a general recommendation that it should be instituted.
Conventional blood culture techniques are insensitive in detecting blood-borne Candida infections. For
example, only 50 % of patients with disseminated candidiasis have positive blood cultures [196], but lysis centrifugation of blood cultures increases yield by 30±40 %.
While growth of Candida species in blood is a clear indication for initiation of antifungal agents, failure to confirm candidemia in an at-risk patient in no way disproves the diagnosis.
Candiduria in patients who have not had instrumentation of the renal tract is strongly suggestive of renal involvement in disseminated candidiasis [195]. The practical value of this is limited by the fact that the majority of
ICU patients will have been catheterized at some stage.
Furthermore, up to 50 % of patients who have disseminated candidiasis do not have candiduria [205]. In an
ICU setting the finding of candiduria in a catheterized
patient is no more significant an indicator of invasive
disease than isolation from any other single site [206].
While a single colony of Candida species isolated
from a sterile site such as blood or cerebrospinal fluid
must be regarded as significant, the greatest obstacle to
the diagnosis of invasive Candida infection by culture
from nonsterile sites is distinguishing infection from colonization. Cut-off values for quantitative diagnosis of
Candida infections are much less well established than
for bacterial infections. Diagnosis of Candida infection
by tissue biopsy is made on the basis of either quantitative culture of more than 105 organisms per gram of tissue or the presence of yeasts on microscopy pending culture results [196].
Although C. albicans continues to make up the majority of clinical isolates, the incidence of other species
is increasing. When Candida is cultured from nonsterile
sites or urine, differentiation of C. albicans from other
species using the germ-tube technique is generally sufficient. However, when Candida is identified at sterile
sites, speciation and sensitivity testing should be routine
[157]. Relative sensitivity of these species to azole anti-
S 27
fungals varies. In certain cases the use of azoles may be
possible in place of amphotericin, with its associated
toxicity (see Bochud et al., ªAntibiotics in sepsisº)
[157, 207]. In addition, knowledge of the species and resistance patterns within a particular hospital is essential
to choice of empirical antifungal agents.
A number of PCR [208] and serological [209, 210,
211] assays have shown promise in preclinical trials, but
two clinical trials in hospitalized patients have found
that such techniques are unlikely to replace culture
based diagnosis [212, 213].
References
1. Cohen J, Heumann D, Glauser MP
(1995) Do monoclonal antibodies and
anticytokines still have a future in infectious diseases? Am J Med 99:
45S±52S
2. Parrillo JE (1993) Pathogenetic mechanisms of septic shock. N Engl J Med
328: 1471±1477
3. Kreger BE, Craven DE, McCabe WR
(1980) Gram-negative bacteremia. IV.
Re-evaluation of clinical features and
treatment in 612 patients. Am J Med
68: 344±355
4. O'Grady NP, Barie PS, Bartlett JG, et
al (1998) Practice guidelines for evaluating new fever in critically ill adult patients. Task Force of the Society of
Critical Care Medicine and the Infectious Diseases Society of America.
Clin Infect Dis 26: 1042±1059
5. Carlet J (1999) Rapid diagnostic methods in the detection of sepsis. Infect
Dis Clin North Am 13: 483±494, xi
6. Yentis SM, Soni N, Sheldon J (1995) Creactive protein as an indicator of resolution of sepsis in the intensive care
unit. Intensive Care Med 21: 602±605
7. Oberhoffer M, Karzai W, Meier-Hellmann A, et al (1999) Sensitivity and
specificity of various markers of inflammation for the prediction of tumor necrosis factor-alpha and interleukin-6 in patients with sepsis. Crit Care
Med 27: 1814±1818
8. Ugarte H, Silva E, Mercan D, et al
(1999) Procalcitonin used as a marker
of infection in the intensive care unit.
Crit Care Med 27: 498±504
9. Whang KT, Steinwald PM, White JC,
et al (1998) Serum calcitonin precursors in sepsis and systemic inflammation. J Clin Endocrinol Metab 83:
3296±3301
10. Povoa P, Almeida E, Moreira P, et al
(1998) C-reactive protein as an indicator of sepsis. Intensive Care Med 24:
1052±1056
11. de Werra I, Jaccard C, Corradin SB, et
al (1997) Cytokines, nitrite/nitrate, soluble tumor necrosis factor receptors,
and procalcitonin concentrations: comparisons in patients with septic shock,
cardiogenic shock, and bacterial pneumonia. Crit Care Med 25: 607±613
12. Leibovici L, Greenshtain S, Cohen O,
et al (1991) Bacteremia in febrile patients. A clinical model for diagnosis.
Arch Intern Med 151: 1801±1806
13. Smith-Elekes S, Weinstein MP (1993)
Blood cultures. Infect Dis Clin North
Am 7: 221±234
14. McDonald LC, Fune J, Gaido LB, et al
(1996) Clinical importance of increased sensitivity of BacT/Alert FAN
aerobic and anaerobic blood culture
bottles. J Clin Microbiol 34: 2180±2184
15. Reimer LG, Wilson ML, Weinstein
MP (1997) Update on detection of
bacteremia and fungemia. Clin Microbiol Rev 10: 444±465
16. Boschman CR, Tucker LJ, Dressel
DC, et al (1995) Optimizing detection
of microbial sepsis: a comparison of
culture systems using packaged sets
with directions for blood collection.
Diagn Microbiol Infect Dis 23: 1±9
17. Bates DW, Goldman L, Lee TH (1991)
Contaminant blood cultures and resource utilization. The true consequences of false-positive results.
JAMA 265: 365±369
18. Spitalnic SJ, Woolard RH, Mermel LA
(1995) The significance of changing
needles when inoculating blood cultures: a meta-analysis. Clin Infect Dis
21: 1103±1106
19. Bryant JK, Strand CL (1987) Reliability of blood cultures collected from intravascular catheter versus venipuncture. Am J Clin Pathol 88: 113±116
20. Weinstein MP (1996) Current blood
culture methods and systems: clinical
concepts, technology, and interpretation of results. Clin Infect Dis 23: 40±46
21. Souvenir D, Anderson DE Jr, Palpant
S, et al (1998) Blood cultures positive
for coagulase-negative staphylococci:
antisepsis, pseudobacteremia, and
therapy of patients. J Clin Microbiol
36: 1923±1926
22. Shahar E, Wohl-Gottesman BS,
Shenkman L (1990) Contamination of
blood cultures during venepuncture:
fact or myth? Postgrad Med J 66:
1053±1058
23. Strand CL, Wajsbort RR, Sturmann K
(1993) Effect of iodophor vs iodine
tincture skin preparation on blood culture contamination rate. JAMA 269:
1004±1006
24. Schifman RB, Pindur A (1993) The effect of skin disinfection materials on
reducing blood culture contamination.
Am J Clin Pathol 99: 536±538
25. Lynn WA, Cohen J (eds) (1995) Microbiological requirements for studies
of sepsis. In: Sibbald WJ, Vincent J-L
(eds) Update in intensive care and
emergency medicine, vol 19. Springer,
Berlin Heidelberg New York, pp 71±85
26. Brown DR, Kutler D, Rai B, et al
(1995) Bacterial concentration and
blood volume required for a positive
blood culture. J Perinatol 15: 157±159
27. Kreger BE, Craven DE, Carling PC, et
al (1980) Gram-negative bacteremia.
III. Reassessment of etiology, epidemiology and ecology in 612 patients.
Am J Med 68: 332±343
28. Jonsson B, Nyberg A, Henning C
(1993) Theoretical aspects of bacteremia as a function of the volume of
blood cultured. APMIS 101: 595±601
29. Weinstein MP, Mirrett S, Wilson ML,
et al (1994) Controlled evaluation of 5
versus 10 milliliters of blood cultured
in aerobic BacT/Alert blood culture
bottles. J Clin Microbiol 32: 2103±2106
30. Mermel LA, Maki DG (1993) Detection of bacteremia in adults: consequences of culturing an inadequate
volume of blood. Ann Intern Med
119: 270±272
31. Li J, Plorde JJ, Carlson LG (1994) Effects of volume and periodicity on
blood cultures. J Clin Microbiol 32:
2829±2831
32. Wilson ML (1996) General principles
of specimen collection and transport.
Clin Infect Dis 22: 766±777
33. Kaditis AG, O'Marcaigh AS, Rhodes
KH, et al (1996) Yield of positive blood
cultures in pediatric oncology patients
by a new method of blood culture collection. Pediatr Infect Dis J 15: 615±620
34. Lombardi DP, Engleberg NC (1992)
Anaerobic bacteremia: incidence, patient characteristics, and clinical significance. Am J Med 92: 53±60
S 28
35. Pottumarthy S, Morris AJ (1997) Assessment of the yield of anaerobic
blood cultures. Pathology 29: 415±417
36. Murray PR, Traynor P, Hopson D
(1992) Critical assessment of blood
culture techniques: analysis of recovery of obligate and facultative anaerobes, strict aerobic bacteria, and fungi
in aerobic and anaerobic blood culture
bottles. J Clin Microbiol 30: 1462±1468
37. Krumholtz H, Cummings S, York M
(1990) Blood Culture phlebotomy:
switching needles does not prevent
contamination. Ann Intern Med 113:
290±292
38. Thamlikitkul V, Chokloikaew S, Tangtrakul T, et al (1992) Blood culture:
comparison of outcomes between
switch-needle and no-switch techniques. Am J Infect Control 20:
122±125
39. Smart D, Baggoley C, Head J, et al
(1993) Effect of needle changing and
intravenous cannula collection on
blood culture contamination rates.
Ann Emerg Med 22: 1164±1168
40. Weinstein MP, Reller LB, Murphy JR,
et al (1983) The clinical significance of
positive blood cultures: a comprehensive analysis of 500 episodes of bacteremia and fungemia in adults. I. Laboratory and epidemiologic observations. Rev Infect Dis 5: 35±53
41. Washington JA 2nd (1975) Blood cultures: principles and techniques.
Mayo Clin Proc 50: 91±98
42. Strand CL (1988) Blood cultures: consensus recommendations in 1988.
American Society for Clinical Pathologists Check Sample Continuing Education Program
43. Dobbins BM, Kite P, Wilcox MH
(1999) Diagnosis of central venous
catheter related sepsis-a critical look
inside. J Clin Pathol 52: 165±172
44. Reed CR, Sessler CN, Glauser FL, et
al (1995) Central venous catheter infections: concepts and controversies.
Intensive Care Med 21: 177±183
45. Maki DG, Jarrett F, Sarafin HW
(1977) A semiquantitative culture
method for identification of catheterrelated infection in the burn patient.
J Surg Res 22: 513±520
46. Cleri DJ, Corrado ML, Seligman SJ
(1980) Quantitative culture of intravenous catheters and other intravascular
inserts. J Infect Dis 141: 781±786
47. Siegman-Igra Y, Anglim AM, Shapiro
DE, et al (1997) Diagnosis of vascular
catheter-related bloodstream infection: a meta-analysis. J Clin Microbiol
35: 928±936
48. Gosbell IB, Duggan D, Breust M, et al
(1995) Infection associated with central venous catheters: a prospective
survey. Med J Aust 162: 210±213
49. Maki DG, Mermel LA (1998) Infections due to infusion therapy. In: Bennett JV, Brachman PS (eds) Hospital
infections. Lippincott-Raven, Philadelphia, pp 689±722
50. Capdevila JA, Planes AM, Palomar
M, et al (1992) Value of differential
quantitative blood cultures in the diagnosis of catheter-related sepsis. Eur J
Clin Microbiol Infect Dis 11: 403±407
51. Weightman NC, Simpson EM, Speller
DC, et al (1988) Bacteremia related to
indwelling central venous catheters:
prevention, diagnosis and treatment.
Eur J Clin Microbiol Infect Dis 7:
125±129
52. Sherertz RJ, Heard SO, Raad II (1997)
Diagnosis of triple-lumen catheter infection: comparison of roll plate, sonication, and flushing methodologies.
J Clin Microbiol 35: 641±646
53. Andremont A, Paulet R, Nitenberg G,
et al (1988) Value of semiquantitative
cultures of blood drawn through catheter hubs for estimating the risk of catheter tip colonization in cancer patients.
J Clin Microbiol 26: 2297±2299
54. Benezra D, Kiehn TE, Gold JW, et al
(1988) Prospective study of infections
in indwelling central venous catheters
using quantitative blood cultures. Am
J Med 85: 495±498
55. Paya C, Guerra L, Marsh M, et al
(1989) Limited usefulness of quantitative culture of blood drawn through
the device for diagnosis of intravascular-device-related bacteremia. J Clin
Microbiol 27: 1431±1433
56. Quilici N, Audibert G, Conroy MC, et
al (1997) Differential quantitative
blood cultures in the diagnosis of catheter-related sepsis in intensive care
units. Clin Infect Dis 25: 1066±1070
57. Sherertz RJ (1996) Surveillance for infections associated with vascular catheters. Infect Control Hosp Epidemiol
17: 746±752
58. Raad I (1998) Intravascular-catheterrelated infections. Lancet 351: 893±898
59. Rogers MS, Oppenheim BA (1998)
The use of continuous monitoring
blood culture systems in the diagnosis
of catheter related sepsis. J Clin Pathol
51: 635±637
60. Thorpe TC, Wilson ML, Turner JE, et
al (1990) BacT/Alert: an automated
colorimetric microbial detection system. J Clin Microbiol 28: 1608±1612
61. Bernard D, Verschraegen G, Claeys G,
et al (1996) Diagnosis of intravascular
device related sepsis by a continuous
monitoring blood culture system. Intensive Care Med 22: 829±830
62. Blot F, Schmidt E, Nitenberg G, et al
(1998) Earlier positivity of central-venous- versus peripheral-blood cultures
is highly predictive of catheter-related
sepsis. J Clin Microbiol 36: 105±109
63. Blot F (1999) Diagnosis of catheter-related-bacteremia: a prospective comparison of the time to positivity of
hub-blood versus peripheral-blood
cultures. Lancet 354: 1071±1077
64. Tighe MJ, Kite P, Thomas D, et al
(1996) Rapid diagnosis of catheter-related sepsis using the acridine orange
leukocyte cytospin test and an endoluminal brush. JPEN J Parenter Enteral
Nutr 20: 215±218
65. von Baum H, Philippi P, Geiss HK
(1998) Acridine-orange leucocyte cytospin (AOLC) test as an in-situ method for the diagnosis of central venous
catheter (CVC)-related sepsis in adult
risk patients. Zentralbl Bakteriol 287:
117±123
66. Gowardman JR, Montgomery C,
Thirlwell S, et al (1998) Central venous catheter-related bloodstream infections: an analysis of incidence and
risk factors in a cohort of 400 patients.
Intensive Care Med 24: 1034±1039
67. Kite P, Dobbins BM, Wilcox MH, et al
(1999) Rapid diagnosis of central-venous-catheter-related blood stream infection without catheter removal. Lancet 354: 1504±1507
68. Tighe MJ, Kite P, Fawley WN, et al
(1996) An endoluminal brush to detect
the infected central venous catheter in
situ: a pilot study. BMJ 313: 1528±1529
69. Kite P, Dobbins BM, Wilcox MH, et al
(1997) Evaluation of a novel endoluminal brush method for in situ diagnosis of catheter related sepsis. J Clin Pathol 50: 278±282
70. Aufwerber E, Ringertz S, Ransjo U
(1991) Routine semiquantitative cultures and central venous catheter-related bacteremia. APMIS 99: 627±630
71. McKinley S, Mackenzie A, Finfer S, et
al (1999) Incidence and predictors of
central venous catheter related infection in intensive care patients. Anaesth
Intensive Care 27: 164±169
72. Farkas JC, Liu N, Bleriot JP, et al
(1992) Single- versus triple-lumen central catheter-related sepsis: a prospective randomized study in a critically ill
population. Am J Med 93: 277±282
73. Raad II, Bodey GP (1992) Infectious
complications of indwelling vascular
catheters. Clin Infect Dis 15: 197±208
S 29
74. Gil RT, Kruse JA, Thill-Baharozian
MC, et al (1989) Triple- vs single-lumen central venous catheters. A prospective study in a critically ill population. Arch Intern Med 149: 1139±1143
75. Brun-Buisson C, Abrouk F, Legrand P,
et al (1987) Diagnosis of central venous catheter-related sepsis. Critical
level of quantitative tip cultures. Arch
Intern Med 147: 873±877
76. Pinilla JC, Ross DF, Martin T, et al
(1983) Study of the incidence of intravascular catheter infection and associated septicemia in critically ill patients.
Crit Care Med 11: 21±25
77. Sitzmann JV, Townsend TR, Siler MC,
et al (1985) Septic and technical complications of central venous catheterization. A prospective study of 200
consecutive patients. Ann Surg 202:
766±770
78. Keung YK, Watkins K, Chen SC, et al
(1994) Comparative study of infectious complications of different types
of chronic central venous access devices. Cancer 73: 2832±2837
79. Timsit JF, Sebille V, Farkas JC, et al
(1996) Effect of subcutaneous tunneling on internal jugular catheter-related sepsis in critically ill patients: a
prospective randomized multicenter
study. JAMA 276: 1416±1420
80. Randolph AG, Cook DJ, Gonzales
CA, et al (1998) Tunneling short-term
central venous catheters to prevent
catheter-related infection: a metaanalysis of randomized, controlled trials. Crit Care Med 26: 1452±1457
81. Darouiche RO, Raad II, Heard SO, et
al (1999) A comparison of two antimicrobial-impregnated central venous
catheters. Catheter Study Group.
N Engl J Med 340: 1±8
82. Maki DG, Stolz SM, Wheeler S, et al
(1997) Prevention of central venous
catheter-related bloodstream infection
by use of an antiseptic-impregnated
catheter. A randomized, controlled trial. Ann Intern Med 127: 257±266
83. Kamal GD, Pfaller MA, Rempe LE, et
al (1991) Reduced intravascular catheter infection by antibiotic bonding. A
prospective, randomized, controlled
trial. JAMA 265: 2364±2368
84. Kamal GD, Divishek D, Kumar GC,
et al (1998) Reduced intravascular
catheter-related infection by routine
use of antibiotic-bonded catheters in a
surgical intensive care unit. Diagn Microbiol Infect Dis 30: 145±152
85. Ciresi DL, Albrecht RM, Volkers PA,
et al (1996) Failure of antiseptic bonding to prevent central venous catheter-related infection and sepsis. Am
Surg 62: 641±646
86. Pemberton LB, Ross V, Cuddy P, et al
(1996) No difference in catheter sepsis
between standard and antiseptic central venous catheters. A prospective
randomized trial. Arch Surg 131:
986±989
87. Christensen ML, Hancock ML, Gattuso J, et al (1993) Parenteral nutrition
associated with increased infection
rate in children with cancer. Cancer
72: 2732±2738
88. Ena J, Cercenado E, Martinez D, et al
(1992) Cross-sectional epidemiology
of phlebitis and catheter-related infections. Infect Control Hosp Epidemiol
13: 15±20
89. Raad I, Umphrey J, Khan A, et al
(1993) The duration of placement as a
predictor of peripheral and pulmonary
arterial catheter infections. J Hosp Infect 23: 17±26
90. Cook D, Randolph A, Kernerman P, et
al (1997) Central venous catheter replacement strategies: a systematic review of the literature. Crit Care Med
25: 1417±1424
91. Badley AD, Steckelberg JM, Wollan
PC, et al (1996) Infectious rates of central venous pressure catheters: comparison between newly placed catheters and those that have been changed.
Mayo Clin Proc 71: 838±846
92. Bozzetti F, Bonfanti G, Regalia E, et al
(1991) A new approach to the diagnosis of central venous catheter sepsis.
JPEN J Parenter Enteral Nutr 15:
412±416
93. Lugauer S, Regenfus A, Boswald M, et
al (1999) A new scoring system for the
clinical diagnosis of catheter-related
infections. Infection 27:S49±53
94. Pearson ML (1996) Guideline for prevention of intravascular device-related
infections. Hospital Infection Control
Practices Advisory Committee. Infect
Control Hosp Epidemiol 17: 438±473
95. Snydman DR, Gorbea HF, Pober BR,
et al (1982) Predictive value of surveillance skin cultures in total-parenteralnutrition-related infection. Lancet 2:
1385±1388
96. Armstrong CW, Mayhall CG, Miller
KB, et al (1990) Clinical predictors of
infection of central venous catheters
used for total parenteral nutrition. Infect Control Hosp Epidemiol 11: 71±78
97. Raad II, Baba M, Bodey GP (1995)
Diagnosis of catheter-related infections: the role of surveillance and targeted quantitative skin cultures. Clin
Infect Dis 20: 593±597
98. Reimer LG (1994) Catheter-related
infections and blood cultures. Clin
Lab Med 14: 51±58
99. Brown PD, Lerner SA (1998) Community-acquired pneumonia. Lancet 352:
1295±1302
100. Finch R (1998) Community acquired
pneumonia. J R Coll Physicians
(Lond) 32: 328±332
101. Marrie TJ (1998) Community-acquired pneumonia: epidemiology, etiology, treatment. Infect Dis Clin North
Am 12: 723±740
102. Conces DJ Jr (1998) Bacterial pneumonia in immunocompromised patients. J Thorac Imaging 13: 261±270
103. Bonten MJ, Gaillard CA, Wouters EF,
et al (1994) Problems in diagnosing
nosocomial pneumonia in mechanically ventilated patients: a review. Crit
Care Med 22: 1683±1691
104. Meduri GU (1995) Diagnosis and differential diagnosis of ventilator-associated pneumonia. Clin Chest Med 16:
61±93
105. Wunderink RG (1998) Mortality and
the diagnosis of ventilator-associated
pneumonia: a new direction. Am J Respir Crit Care Med 157: 349±350
106. Bonten MJ (1999) Controversies on
diagnosis and prevention of ventilator-associated pneumonia. Diagn Microbiol Infect Dis 34: 199±204
107. Gallego M, Rello J (1999) Diagnostic
testing for ventilator-associated pneumonia. Clin Chest Med 20: 671±679
108. Griffin JJ, Meduri GU (1994) New approaches in the diagnosis of nosocommial pneumonia. Med Clin North Am
78: 1091±1105
109. Cook DJ, Walter SD, Cook RJ, et al
(1998) Incidence of and risk factors
for ventilator-associated pneumonia
in critically ill patients. Ann Intern
Med 129: 433±440
110. Cook DJ, Kollef MH (1998) Risk factors for ICU-acquired pneumonia.
JAMA 279: 1605±1606
111. Joshi N, Localio AR, Hamory BH
(1992) A predictive risk index for
nosocomial pneumonia in the intensive care unit. Am J Med 93: 135±142
112. Edgeworth JD, Treacher DF, Eykyn SJ
(1999) A 25-year study of nosocomial
bacteremia in an adult intensive care
unit. Crit Care Med 27: 1421±1428
113. Weinstein MP, Towns ML, Quartey
SM, et al (1997) The clinical significance of positive blood cultures in the
1990s: a prospective comprehensive
evaluation of the microbiology, epidemiology, and outcome of bacteremia
and fungemia in adults. Clin Infect
Dis 24: 584±602
S 30
114. Fagon JY, Chastre J, Hance AJ, et al
(1988) Detection of nosocomial lung
infection in ventilated patients. Use of
a protected specimen brush and quantitative culture techniques in 147 patients. Am Rev Respir Dis 138:
110±116
115. Meduri GU, Mauldin GL, Wunderink
RG, et al (1994) Causes of fever and
pulmonary densities in patients with
clinical manifestations of ventilatorassociated pneumonia. Chest 106:
221±235
116. Anonymous (1995) Hospital-acquired
pneumonia in adults: diagnosis, assessment of severity, initial antimicrobial
therapy, and preventive strategies. A
consensus statement, American Thoracic Society, November 1995. Am J
Respir Crit Care Med 153: 1711±1725
117. Santos E, Talusan A, Brandstetter RD
(1998) Roentgenographic mimics of
pneumonia in the critical care unit.
Crit Care Clinics 14: 91±104
118. Butler KL, Sinclair KE, Henderson
VJ, et al (1999) The chest radiograph
in critically ill surgical patients is inaccurate in predicting ventilator-associated pneumonia. Am Surg 65: 805±809
119. Wunderink RG, Woldenberg LS, Zeiss
J, et al (1992) The radiologic diagnosis
of autopsy-proven ventilator-associated pneumonia. Chest 101: 458±463
120. Peruzzi W, Garner W, Bools J, et al
(1988) Portable chest roentgenography and computed tomography in critically ill patients. Chest 93: 722±726
121. Lipchik RJ, Kuzo RS (1996) Nosocomial pneumonia. Radiol Clin North
Am 34: 47±58
122. Light RW (1977) Pleural effusions.
Med Clin North Am 61: 1339±1352
123. Carratala J, Gudiol F, Pallares R, et al
(1994) Risk factors for nosocomial Legionella pneumophila pneumonia. Am
J Respir Crit Care Med 149: 625±629
124. Brun-Buisson C (1995) Nosocomial
pneumonia during mechanical ventilation: problems with diagnostic criteria.
Thorax 50: 1128±1130
125. Cardenosa Cendrero JA, Sole-Violan
J, Bordes Benitez A, et al (1999) Role
of different routes of tracheal colonization in the development of pneumonia in patients receiving mechanical
ventilation. Chest 116: 462±470
126. Kirtland SH, Corley DE, Winterbauer
RH, et al (1997) The diagnosis of ventilator-associated pneumonia: a comparison of histologic, microbiologic,
and clinical criteria. Chest 112:
445±457
127. Middleton R, Broughton WA, Kirkpatrick MB (1992) Comparison of four
methods for assessing airway bacteriology in intubated, mechanically ventilated patients Am J Med Sci 304:
239±245
128. Marquette CH, Georges H, Wallet F,
et al (1993) Diagnostic efficiency of
endotracheal aspirates with quantitative bacterial cultures in intubated patients with suspected pneumonia.
Comparison with the protected specimen brush. Am Rev Respir Dis 148:
138±144
129. el-Ebiary M, Torres A, Gonzalez J, et
al (1993) Quantitative cultures of endotracheal aspirates for the diagnosis
of ventilator-associated pneumonia.
Am Rev Respir Dis 148: 1552±1557
130. Marquette CH, Copin MC, Wallet F, et
al (1995) Diagnostic tests for pneumonia in ventilated patients: prospective
evaluation of diagnostic accuracy using histology as a diagnostic gold standard. Am J Respir Crit Care Med 151:
1878±1888
131. Sanchez-Nieto JM, Torres A, GarciaCordoba F, et al (1998) Impact of invasive and noninvasive quantitative culture sampling on outcome of ventilator-associated pneumonia: a pilot
study. Am J Respir Crit Care Med
157: 371±376
132. Pugin J, Auckenthaler R, Mili N
(1991) Diagnosis of ventilator associated pneumonia by bacteriologic analysis and bronchoscopic and nonbronchoscopic ªblindº Bronchoalveolar lavage fluid. Am Rev Respir Dis 143:
1121±1129
133. Rello J, Sa-Borges M, Correa H, et al
(1999) Variations in etiology of ventilator-associated pneumonia across
four treatment sites: implications for
antimicrobial prescribing practices.
Am J Respir Crit Care Med 160:
608±613
134. Alvarez-Lerma F (1996) Modification
of empiric antibiotic treatment in patients with pneumonia acquired in the
intensive care unit. ICU-Acquired
Pneumonia Study Group. Intensive
Care Med 22: 387±394
135. Luna CM, Vujacich P, Niederman MS,
et al (1997) Impact of BAL data on
the therapy and outcome of ventilator-associated pneumonia. Chest 111:
676±685
136. Rello J, Gallego M, Mariscal D, et al
(1997) The value of routine microbial
investigation in ventilator-associated
pneumonia Am J Respir Crit Care
Med 156: 196±200
137. Kollef MH, Ward S (1998) The influence of mini-BAL cultures on patient
outcomes: implications for the antibiotic management of ventilator-associated pneumonia. Chest 113: 412±420
138. Perry C, Chandrasekhar A, Moorman
DW, et al (1999) Ventilator associated
pneumonia: therapy based on protected specimen brushing versus tracheal
aspirate data. Chest 116: 278S
139. Fagon JY, Chastre J, Wolff M, et al
(2000) Invasive and noninvasive strategies for management of suspected
ventilator-associated pneumonia. A
randomized trial. Ann Intern Med
132: 621±630
140. Bonten MJ, Bergmans DC, Stobberingh EE, et al (1997) Implementation
of bronchoscopic techniques in the diagnosis of ventilator-associated pneumonia to reduce antibiotic use. Am J
Respir Crit Care Med 156: 1820±1824
141. Croce MA, Fabian TC, Schurr MJ, et
al (1995) Using bronchoalveolar lavage to distinguish nosocomial pneumonia from systemic inflammatory response syndrome: a prospective analysis. J Trauma Inj Infect Crit Care 39:
1134±1139
142. Horan TC, Gaynes RP, Martone WJ,
et al (1992) CDC definitions of nosocomial surgical site infections, 1992: a
modification of CDC definitions of
surgical wound infections. Infect Control Hosp Epidemiol 13: 606±608
143. Mangram AJ, Horan TC, Pearson ML,
et al (1999) Guideline for Prevention
of Surgical Site Infection, 1999. Centers for Disease Control and Prevention (CDC) Hospital Infection Control
Practices Advisory Committee. Am J
Infect Control 27: 97±132
144. Crowe MJ, Cooke EM (1998) Review
of case definitions for nosocomial infection ± towards a consensus. Presentation by the Nosocomial Infection
Surveillance Unit (NISU) to the Hospital Infection Liaison Group, subcommittee of the Federation of Infection Societies (FIS). J Hosp Infect 39:
3±11
145. Robson MC (1997) Wound infection: a
failure of wound healing caused by an
imbalance of bacteria. Surg Clin North
Am 77: 637±650
146. Crowe M, Ispahani P, Humphreys H,
et al (1998) Bacteremia in the adult intensive care unit of a teaching hospital
in Nottingham, UK, 1985±1996. Eur J
Clin Microbiol Infect Dis 17: 377±384
147. Sawyer RG, Pruett TL (1994) Wound
infections. Surg Clin North Am 74:
519±536
S 31
148. Nichols RL, Smith JW (1993) Wound
and intraabdominal infections: microbiological considerations and approaches to treatment. Clin Infect Dis
16:S266±272
149. Brook I, Frazier EH (1998) Aerobic
and anaerobic microbiology of retroperitoneal abscesses. Clin Infect Dis
26: 938±941
150. Brook I, Frazier EH (1999) Microbiology of subphrenic abscesses: a 14-year
experience. Am Surg 65: 1049±1053
151. de Lalla F (1999) Antimicrobial chemotherapy in the control of surgical infectious complications. J Chemother
11: 440±445
152. Fry DE (1994) Noninvasive imaging
tests in the diagnosis and treatment of
intra-abdominal abscesses in the postoperative patient. Surg Clin North
Am 74: 693±709
153. Montgomery RS, Wilson SE (1996) Intraabdominal abscesses: image-guided
diagnosis and therapy. Clin Infect Dis
23: 28±36
154. McDowell RK, Dawson SL (1996)
Evaluation of the abdomen in sepsis
of unknown origin. Radiol Clin North
Am 34: 177±190
155. Noone TC, Semelka RC, Worawattanakul S, et al (1998) Intraperitoneal abscesses: diagnostic accuracy of
and appearances at MR imaging. Radiology 208: 525±528
156. McClean KL, Sheehan GJ, Harding
GK (1994) Intraabdominal infection:
a review. Clin Infect Dis 19: 100±116
157. Anonymous (1994) Management of
deep Candida infection in surgical and
intensive care unit patients. British Society for Antimicrobial Chemotherapy
Working Party. Intensive Care Med 20:
522±528
158. Wong JK, Mustard R, Sadler DJ, et al
(1999) Predicting infection in localised
intra-abdominal fluid infections: value
of pH and PO2 measurements. J Vasc
Interv Radiol 10: 421±427
159. Frazee RC, Nagorney DM, Mucha P
Jr (1989) Acute acalculous cholecystitis. Mayo Clin Proc 64: 163±167
160. Imhof M, Raunest J, Rauen U, et al
(1992) Acute acalculous cholecystitis
in severely traumatized patients: a prospective sonographic study. Surgical
Endoscopy 6: 68±71
161. Hill AG, Collins JP (1994) Acute acalculous cholecystitis. Aust N Z J Surg
64: 251±253
162. Almeida J, Sleeman D, Sosa JL, et al
(1995) Acalculous cholecystitis: the
use of diagnostic laparoscopy. J Laparoendosc Surg 5: 227±231
163. Hagino RT, Valentine RJ, Clagett GP
(1997) Acalculous cholecystitis after
aortic reconstruction. J Am Coll Surg
184: 245±248
164. Flancbaum L, Majerus TC, Cox EF
(1985) Acute posttraumatic acalculous
cholecystitis. Am J Surg 150: 252±256
165. Kalliafas S, Ziegler DW, Flancbaum L,
et al (1998) Acute acalculous cholecystitis: incidence, risk factors, diagnosis,
and outcome. Am Surg 64: 471±475
166. Beckman I, Dash N, Sefczek RJ, et al
(1985) Ultrasonographic findings in
acute acalculous cholecystitis. Gastrointest Radiol 10: 387±389
167. Molenat F, Boussuges A, Valantin V,
et al (1996) Gallbladder abnormalities
in medical ICU patients: an ultrasonographic study. Intensive Care Med 22:
356±358
168. Shapiro MJ, Luchtefeld WB, Kurzweil
S, et al (1994) Acute acalculous cholecystitis in the critically ill. Am Surg
60: 335±339
169. Mirvis SE, Vainright JR, Nelson AW,
et al (1986) The diagnosis of acute acalculous cholecystitis: a comparison
of sonography, scintigraphy, and CT.
AJR Am J Roentgenol 147: 1171±1175
170. Jeffrey RB Jr, Sommer FG (1993) Follow-up sonography in suspected acalculous cholecystitis: preliminary clinical experience. J Ultrasound Med 12:
183±187
171. Helbich TH, Mallek R, Madl C, et al
(1997) Sonomorphology of the gallbladder in critically ill patients. Value
of a scoring system and follow-up examinations. Acta Radiol 38: 129±134
172. Blankenberg F, Wirth R, Jeffrey RB Jr,
et al (1991) Computed tomography as
an adjunct to ultrasound in the diagnosis of acute acalculous cholecystitis.
Gastrointest Radiol 16: 149±153
173. Yang HK, Hodgson WJ (1996) Laparoscopic cholecystostomy for acute acalculous cholecystitis. Surg Endosc
10: 673±675
174. Arens JF, LeJeune FE Jr, Webre DR
(1974) Maxillary sinusitis, a complication of nasotracheal intubation. Anesthesiology 40: 415±416
175. Lum Cheong RS, Cornwell Eed (1992)
Suppurative sinusitis in critically ill patients: a case report and review of the
literature. J Natl Med Assoc 84:
1057±1059
176. Bach A, Boehrer H, Schmidt H, et al
(1992) Nosocomial sinusitis in ventilated patients. Nasotracheal versus orotracheal intubation. Anaesthesia 47:
335±339
177. Borman KR, Brown PM, Mezera KK,
et al (1992) Occult fever in surgical intensive care unit patients is seldom
caused by sinusitis. Am J Surg 164:
412±415
178. Westergren V, Lundblad L, Hellquist
HB, et al (1998) Ventilator-associated
sinusitis: a review. Clin Infect Dis 27:
851±864
179. Laranne JE, Penttila MA, Paakkala
TA, et al (1992) Diagnostic value of
plain radiographs in chronic maxillary
sinusitis: a comparison between radiological and endoscopic findings in
75 patients. Rhinology 30: 205±215
180. Garner JS, Jarvis WR, Emori TG, et al
(1988) CDC definitions for nosocomial infections, 1988. Am J Infect Control 16: 128±140
181. Salord F, Gaussorgues P, Marti-Flich J,
et al (1990) Nosocomial maxillary sinusitis during mechanical ventilation:
a prospective comparison of orotracheal versus the nasotracheal route
for intubation. Intensive Care Med 16:
390±393
182. Bert F, Lambert-Zechovsky N (1995)
Microbiology of nosocomial sinusitis
in intensive care unit patients. J Infect
31: 5±8
183. Kaups KL, Cohn SM, Nageris B, et al
(1995) Maxillary sinusitis in the surgical intensive care unit: a study using
bedside sinus ultrasound. Am J Otolaryngol 16: 24±28
184. Rouby JJ, Laurent P, Gosnach M, et al
(1994) Risk factors and clinical relevance of nosocomial maxillary sinusitis
in the critically ill. Am J Respir Crit
Care Med 150: 776±783
185. Bowers BL, Purdue GF, Hunt JL
(1991) Paranasal sinusitis in burn patients following nasotracheal intubation. Arch Surg 126: 1411±1412
186. Holzapfel L, Chevret S, Madinier G, et
al (1993) Influence of long-term oroor nasotracheal intubation on nosocomial maxillary sinusitis and pneumonia: results of a prospective, randomized, clinical trial. Crit Care Med 21:
1132±1138
187. Westergren V, Lundblad L, Forsum U
(1998) Ventilator-associated sinusitis:
antroscopic findings and bacteriology
when excluding contaminants. Acta
Otolaryngol (Stockh) 118: 574±580
188. Beck-Sague C, Jarvis WR (1993) Secular trends in the epidemiology of nosocomial fungal infections in the United
States, 1980±1990. National Nosocomial Infections Surveillance System.
J Infect Dis 167: 1247±1251
S 32
189. Pittet D, Wenzel RP (1995) Nosocomial bloodstream infections. Secular
trends in rates, mortality, and contribution to total hospital deaths. Arch
Intern Med 155: 1177±1184
190. Vincent JL, Bihari DJ, Suter PM, et al
(1995) The prevalence of nosocomial
infection in intensive care units in Europe. Results of the European Prevalence of Infection in Intensive Care
(EPIC) Study. EPIC International Advisory Committee. JAMA 274:
639±644
191. Henderson VJ, Hirvela ER (1996)
Emerging and reemerging microbial
threats. Nosocomial fungal infections.
Arch Surg 131: 330±337
192. Nguyen MH, Peacock JE Jr, Morris
AJ, et al (1996) The changing face of
candidemia: emergence of non-Candida albicans species and antifungal resistance. Am J Med 100: 617±623
193. Wey SB, Mori M, Pfaller MA, et al
(1988) Hospital-acquired candidemia.
The attributable mortality and excess
length of stay. Arch Intern Med 148:
2642±2645
194. Nolla-Salas J, Sitges-Serra A, LeonGil C, et al (1997) Candidemia in nonneutropenic critically ill patients: analysis of prognostic factors and assessment of systemic antifungal therapy.
Study Group of Fungal Infection in
the ICU. Intensive Care Med 23: 23±30
195. Solomkin JS (1993) Pathogenesis and
management of Candida infection syndromes in non-neutropenic patients.
New Horiz 1: 202±213
196. Dean DA, Burchard KW (1996) Fungal infection in surgical patients. Am J
Surg 171: 374±382
197. Pittet D, Garbino J (1995) Fungal infections in the critically ill. Curr Opin
Crit Care 1: 369±380
198. Donahue SP, Greven CM, Zuravleff
JJ, et al (1994) Intraocular candidiasis
in patients with candidemia. Clinical
implications derived from a prospective multicenter study. Ophthalmology
101: 1302±1309
199. Bodey GP, Luna M (1974) Skin lesions
associated with disseminated candidiasis. JAMA 229: 1466±1468
200. Fass RJ, Goff DA, Sierawski SJ (1996)
Candida infections in the surgical intensive care unit. J Antimicrob
Chemother 38: 915±916
201. Vincent JL, Anaissie E, Bruining H, et
al (1998) Epidemiology, diagnosis and
treatment of systemic Candida infection in surgical patients under intensive care. Intensive Care Med 24:
206±216
202. Wey SB, Mori M, Pfaller MA, et al
(1989) Risk factors for hospital-acquired candidemia. A matched casecontrol study. Arch Intern Med 149:
2349±2353
203. Pittet D, Monod M, Suter PM, et al
(1994) Candida colonization and subsequent infections in critically ill surgical patients. Ann Surg 220: 751±758
204. Wenzel RP (1995) Nosocomial candidemia: risk factors and attributable
mortality. Clin Infect Dis 20:
1531±1534
205. Burchard KW, Minor LB, Slotman GJ,
et al (1983) Fungal sepsis in surgical
patients. Arch Surg 118: 217±221
206. Cornwell EE 3rd, Belzberg H, offne
TV, et al (1995) The pattern of fungal
infections in critically ill surgical patients. Am Surg 61: 847±850
207. Rex JH, Bennett JE, Sugar AM, et al
(1994) A randomized trial comparing
fluconazole with amphotericin B for
the treatment of candidemia in patients without neutropenia. Candidemia
Study Group and the National Institute. N Engl J Med 331: 1325±1330
208. Shin JH, Nolte FS, Morrison CJ (1997)
Rapid identification of Candida species in blood cultures by a clinically
useful PCR method. J Clin Microbiol
35: 1454±1459
209. Sendid B, Tabouret M, Poirot JL, et al
(1999) New enzyme immunoassays
for sensitive detection of circulating
Candida albicans mannan and antimannan antibodies: useful combined
test for diagnosis of systemic candidiasis. J Clin Microbiol 37: 1510±1517
210. Walsh TJ, Merz WG, Lee JW, et al
(1995) Diagnosis and therapeutic
monitoring of invasive candidiasis by
rapid enzymatic detection of serum
D-arabinitol. Am J Med 99: 164±172
211. Pfaller MA (1992) Laboratory aids in
the diagnosis of invasive candidiasis.
Mycopathologia 120: 65±72
212. Petri MG, Konig J, Moecke HP, et al
(1997) Epidemiology of invasive mycosis in ICU patients: a prospective
multicenter study in 435 non-neutropenic patients. Paul-Ehrlich Society
for Chemotherapy, Divisions of Mycology and Pneumonia Research. Intensive Care Med 23: 317±325
213. Pallavicini F, Izzi I, Pennisi MA, et al
(1999) Evaluation of the utility of serological tests in the diagnosis of candidemia. Minerva Anesthesiol 65:
637±639
Intensive Care Med (2001) 27: S 49±S 62
Source control in the
management of sepsis
Maria F. Jimenez
John C. Marshall
)
M. F. Jimenez ´ J. C. Marshall ( )
Toronto General Hospital,
University Health Network, Toronto,
Ontario, Canada
E-mail: [email protected]
Phone: +1-4 16-3 40-52 05/52 04
Fax: +1-4 16-5 95 94 86
Introduction
Antibiotics were first used as anti-infective therapy a
mere 60 years ago. Even the administration of intravenous fluids to support the circulation is an innovation
of the twentieth century. However, the surgical therapy
of infection in the form of drainage of abscesses, removal of foreign bodies, or dØbridement of devitalized tissue
dates to antiquity. The surgical drainage of ocular abscesses was sufficiently common in ancient Mesopotamia that the fees payable to its practitioners are laid
out in the Code of Hammurabi, and the drainage of abscesses was a common practice at the temple of Asclepios in Greece in the fourth century BC. The surgical
management of wounds dates back to prehistoric times;
skulls showing evidence of trephination (craniotomy)
with healing during life have been found in both Europe
and the pre-Columbian western hemisphere.
Surgical approaches to the treatment of infection
have evolved through principle and tradition, and few
have been evaluated by randomized controlled trials.
Moreover the need to tailor therapy to circumstances
that are often unique to a particular clinical situation
makes it difficult to standardize surgical therapy sufficiently to conduct a randomized trial of surgical approaches in the management of severe sepsis and septic
shock. For example, there has been considerable controversy regarding the relative merits of open versus
closed management of the peritoneal cavity in the patient with diffuse peritonitis. Yet a randomized trial to
answer this question has proven difficult if not impossible to organize [1]. Most patients with peritonitis have
an infectious process that is at least relatively localized
within the peritoneal cavity, and open management necessitates much greater use of nursing and operating
room resources, raising important economical and ethical concerns in designing such a study. Conversely, for a
number of patients with diffuse peritonitis it is not technically possible to close the abdomen because of increased intra-abdominal pressure or concomitant abdominal wall defects. Thus while the question is of theoretical interest, there is in reality only a very small number of patients for whom the two different approaches
could reasonably be considered to be sufficiently appropriate to justify their random assignment to either therapeutic option.
The paucity of data derived from level I or level II
studies limits the development of solid evidence-based
recommendations regarding optimal approaches to
source control in patients with sepsis. The guidelines
presented here are not definitive management algorithms, and their applicability is influenced by a number
of factors best evaluated by a surgical consultant with
experience in the complex process of surgical decision
making in life-threatening sepsis.
Methods
A computer-based literature review was undertaken using Medline
from 1985 to the present. While references were sought using specific subject heading keywords related to sepsis, i.e., sepsis, sepsis syndrome, septic shock, infection, multiple organ failure, critical care,
and intensive care units, the paucity of relevant retrievals prompted
us to repeat the search using headings specific for the questions
asked, including soft tissue infections, pancreatitis, and diverticulitis. Additional references were identified from the references cited
in these reports. Case reports and reviews were excluded, and clinical papers included only if they reported a least ten subjects and included some form of control population (often historical).
S 50
Definitions and principles
The term source control is used in preference to surgical
therapy or surgery to emphasize that measures to control the source of an infection are not limited to surgery,
but may include, for example, removal of an infected intravascular catheter or tube thoracostomy to treat an
empyema of the pleural cavity. Advances in interventional radiology have permitted access to anatomic areas previously only amenable to surgical intervention,
with the result that formal operative intervention is necessary less often than in the past. Source control measures encompass three broad types of intervention:
Drainage of abscesses, dØbridement of devitalized infected tissue and removal of colonized foreign bodies,
diversion, repair, or excision of an ongoing focus of contamination from a hollow viscus.
Drainage of abscesses
An abscess is a discrete collection of tissue debris, bacteria, and host leukocytes that has become walled off from
adjacent healthy tissues. The local host inflammatory
response activates the coagulation cascade, resulting in
the generation of fibrin; this fibrin becomes organized
to form a capsule of fibrous tissue that isolates the infectious focus from the surrounding tissues [2]. The contents of an abscess are usually liquid, because of the
presence of tissue fluid, permitting the abscess to be
drained to the exterior. This process externalizes the focus of infection: further microbial entry into host tissues
is prevented by the abscess wall, and the clinical infectious process resolves. Drainage of an abscess converts
a closed focus of infection to either a fistula if the abscess arises from perforation of an epithelially lined surface such as the gastrointestinal tract, or a sinus if it does
not.
Drainage of an abscess may occur spontaneously or
may be a result of deliberate therapeutic intervention.
The clinical consequences of such a process of spontaneous drainage depend on the site of drainage. In the
preantibiotic era, conservative management of appendiceal abscesses was an accepted practice, since the abscess generally drained into the gastrointestinal tract,
and was associated with no significant clinical sequelae.
However, a diverticular abscess may drain spontaneously into the bladder, giving rise to a colovesical fistula, or
the vagina, creating a colovaginal fistula. Therapeutic
drainage of an abscess results in the creation of a controlled sinus or fistula. Intervention should be planned
to minimize contamination of uninfected tissues adjacent to the abscess, to simplify nursing care needed during the resolution of the abscess cavity, and to minimize
the long-term morbidity of the intervention. Thus, if
percutaneous drainage of an abscess can be performed
effectively, it is generally preferable to surgical intervention, for it reduces the contamination of surrounding tissues, obviates the need for postoperative wound management and avoids the later risks of incisional hernia.
DØbridement of devitalized tissues
DØbridement ± the removal of dead tissue ± can be accomplished by surgical excision, by the use of dressings
that adhere to devitalized tissue, removing it as the
dressing is changed, or by the use of enzymatic or biological agents, including, in the past, maggots. Dead tissue is an excellent culture medium for micro-organisms,
and because it lacks a blood supply, permits microbial
proliferation in an environment that is protected from
host inflammatory cells. Thus dØbridement of infected
necrotic tissue is a critical element in source control,
eliminating a locus of unopposed microbial growth.
The benefits of excising uninfected necrotic tissue are
less clear, and must be weighed against the risks of attempted excision. In the early stages of tissue necrosis,
viable tissue is not clearly demarcated from nonviable
tissue, and an intermediate zone of hyperemic inflammatory tissue is present: bleeding during attempted excision can be considerable and is an important reason
that delayed dØbridement of infected pancreatic necrosis is associated with a better clinical outcome [3]. With
time the demarcation of viable from nonviable tissue
becomes clearer, and intervention is facilitated; however, potential morbidity of intervention must be weighed
against the risks of leaving dead, infected tissue.
A foreign body such as a colonized intravascular
catheter or a retained surgical sponge serves as a reservoir of micro-organisms that are relatively protected
from endogenous host defenses. Experimental studies
show that the presence of a foreign body significantly
reduces the inoculum of micro-organisms required for
the establishment of a focus of infection [4]. In addition,
certain organisms, notably coagulase-negative Staphylococci, form biofilms on prosthetic materials that permit
the establishment of microbial colonies, and protect the
bacteria from host immune cells [5].
Removal of a colonized foreign body is therefore a
key element of source control. Again, however, the clinician must balance risks and benefits. When the infected foreign body is a retained surgical sponge or a colonized Foley catheter, the decision to remove the device
is obvious. However, removal of a prosthetic heart
valve, a vascular graft, or even a peritoneal dialysis catheter in a patient with dialysis-dependent renal failure
entails a greater degree of risk, and justifies a trial of
noninterventional management.
S 51
Definitive management of source of contamination
When infection arises as a consequence of perforation
of a hollow viscus, ongoing contamination occurs unless
the anatomical problem is corrected. Ideally this is accomplished by surgical removal of the involved structure, for example, appendectomy for acute perforated
appendicitis. Drainage alone can serve to create a controlled fistula, and is frequently a less morbid interim solution. Finally, contamination can be reduced by proximal diversion or defunctioning of the injured viscus. A
colostomy created proximal to a colonic perforation,
for example, reduces the volume of gastrointestinal effluent passing across the injured colon and thus minimizes the leakage of bacteria into tissues surrounding
the perforation.
Indications for source control in the management
of sepsis
The potential role of source control measures should be
evaluated for all patients with severe sepsis. In many circumstances, for example, the patient with diffuse peritonitis from a perforated ulcer or clostridial myonecrosis
following a traumatic injury, the need for source control
is obvious. In others, it may be less so. The patient with
pneumonia, for example, may have a concomitant empyema; even in the absence of this complication, tracheal
suctioning, and physiotherapy to enhance clearance of
contaminated secretions are source control measures
that may be of benefit. Replacement of a colonized catheter can hasten the resolution of a urinary tract infection.
The incremental benefits of instituting source control
are largely unknown, even in the management of a number of common infectious processes. Analyses of cohorts of patients enrolled in trials of novel mediator-targeted therapies indicate that inadequate source control
is associated with a worse clinical outcome [6]; however,
the merits of surgical source control, as with those of antibiotic therapy, are largely based on expert opinion, informed by experience.
Selection of the best method of providing definitive
management of a focus of infection involves a principle
common to the application of all forms of source control
± balancing the physiological costs and risks of intervention against the benefits of more definitive control.
Timing of source control interventions
Optimal timing of intervention requires that the benefits of intervention in providing definitive control of a
focus of infection be balanced against the risks of doing
so. General anesthesia or a reduction in the intensity of
monitoring in the radiology suite result in increased
risk for the patient who is not fully resuscitated, or who
remains physiologically unstable despite resuscitation.
Surgical intervention early in the course of necrotizing
pancreatitis can eliminate a potential or actual focus of
infection but entails a significantly greater risk of lifethreatening bleeding, because tissue planes are not demarcated.
As a general principle, source control measures
should only be undertaken once the patient has been appropriately stabilized, although resuscitation should
proceed as rapidly as feasible. Rarely, source control is
part of the process of resuscitation, for example, in
emergency surgery for a ruptured mycotic aneurysm.
Moreover in septic shock associated with intestinal infarction or infections such as clostridial myonecrosis,
full resuscitation is not possible until the rapidly advancing tissue necrosis has been stopped. Nonetheless, even
in these situations, aggressive and rapid resuscitation
can reduce the risks of anesthesia. For most infections
requiring source control, timing is an urgency, but not
an emergency. The impact of the timeliness of source
control on clinical outcome was evaluated for two distinct infectious processes.
Necrotizing soft tissue infection
Does the timing of surgery alter outcome in necrotizing
fasciitis?
Answer: yes, grade E.
Recommendation
Surgical intervention in the form of dØbridement of infected, devitalized, or nonbleeding tissue should be undertaken rapidly following hemodynamic stabilization
in patients with necrotizing soft tissue infections. This
is a grade E recommendation supported by level IV
and level V evidence.
Rationale
The mortality of necrotizing fasciitis in reported series
ranges from 9 % to 30 %. A number of case series and uncontrolled studies suggest that timely intervention in the
patient with a necrotizing soft tissue infection is associated with a superior clinical outcome. Early reports suggested that delayed diagnosis and limited or inadequate
resection results in increased mortality [7, 8, 9]. Both Majestic and Alexander [10] and Miller [11] have shown that
mortality in patients with necrotizing fasciitis is less than
that in historical controls when infection is diagnosed
early, and surgical dØbridement is aggressive.
S 52
More recent retrospective cohort studies [12, 13, 14,
15, 16, 17] have confirmed the benefits of early aggressive surgical dØbridement for the treatment of patients
with necrotizing fasciitis. For example, Bilton and colleagues [12] demonstrated a mortality reduction from
38 % to 4.2 % when patients underwent radical surgical
dØbridement at the time of initial presentation, and
Freischlag and coworkers [14] reported that surgical intervention within 24 h of recognition of necrotizing soft
tissue infection resulted in 36 % mortality, compared to
70 % when the surgical procedure was delayed more
than 24 h. Various reports have suggested that improved
outcome requires surgical intervention within 3 h of admission [18, 19] or 12 h of admission [20].
Two studies evaluating clinical and laboratory predictors of survival of patients with necrotizing soft tissue
infections have suggested that surgical dØbridement
does not influence outcome [21, 22], particularly for patients with coagulopathy at the time of admission [22].
Infected pancreatic necrosis and pancreatic abscesses
Does early surgery improve outcome for patients with
infected pancreatic necrosis?
Answer: no, grade C.
Recommendation
The decision to intervene surgically in the patient with
infected pancreatic necrosis must weigh the potential
advantages of removing a source of ongoing bacterial
proliferation against the inherent morbidity of early surgery. In general, surgery should be delayed in the stable
patient to permit adequate demarcation of tissue planes.
This is a grade C recommendation supported by a single
randomized trial and expert opinion.
Rationale
Studies of the natural history of retroperitoneal infection in patients with severe acute pancreatitis demonstrate that culture positivity generally occurs during the
second or third week after the onset of pancreatitis,
and that the risk of infection is directly related to the initial clinical severity of the disease process [23]. Drainage
of infected fluid collections or dØbridement of infected
necrosis is intuitively appealing. However, the morbidity of early surgical intervention is increased because
clear biological demarcation between viable and necrotic tissues is often not present early in the course of the
disease, and significant retroperitoneal hemorrhage is a
common complication. Earlier enthusiasm for early ac-
tive surgical intervention is giving way to greater conservatism.
Delayed surgical intervention in infected pancreatic
necrosis was first suggested by Machado et al. [24], and
later supported by a randomized trial showing a reduction in complication rates and mortality in patients in
whom surgical intervention was delayed at least 2 weeks
[3]. The benefits of further intentional delay in patients
with known infected pancreatic necrosis are not established.
The discordant recommendations for the management of infected necrosis of subcutaneous or peripancreatic tissues reflect several factors, the most important
of which is the morbidity associated with intervention.
Principles of drainage of infectious foci
Surgical drainage of deep-seated foci of infection was an
innovation of the twentieth century. The classical surgical principles for the drainage of intra-abdominal abscesses were established by Ochsner and Debakey in
1938 [25] in an era when the diagnosis of intra-abdominal infection was established primarily through the clinical findings, and antibiotics were not yet available as
adjuvant anti-infective therapy. More recently improvements in diagnostic imaging and the development of
techniques for radiologically guided abscess drainage
have provided the clinician with a spectrum of therapeutic options for the management of deep space infections.
Radiographic techniques for the diagnosis of deep site
infections
Are computed tomography and ultrasonography
equally efficacious in detecting foci of intra-abdominal
infection?
Answer: yes, grade E.
Recommendation
The diagnosis of intra-abdominal infection amenable to
source control measures can generally be made by either ultrasound or computed tomography (CT). Ultrasonography has the advantage of being portable and inexpensive, but is highly operator dependent; CT is especially useful in the evaluation of the retroperitoneum.
Rationale
The presence of infection within the thoracic or abdominal cavities is usually suspected on the basis of history
S 53
and physical examination; however, radiographic documentation is almost always indicated. Not only does delineation of the process permit safer surgical approaches, in many cases deep site infections can be treated by
radiographically guided percutaneous drainage. Three
methods of radiological investigation are commonly
used to define such infections ± ultrasonography, CT,
and contrast studies. Rigorous comparative studies of
their merits are generally not available.
Abdominal ultrasound is a useful first line diagnostic
modality for intra-abdominal collections. It is inexpensive and portable, and allows confirmatory needle aspiration of fluid collections or therapeutic percutaneous
abscess drainage. The capacity to establish a diagnosis
at the bedside in the ICU obviates the risks associated
with transportation to the radiographic suite. However,
ultrasound is highly operator dependent, and adequate
visualization may be prevented by gas within the lumen
of the intestine or external dressings.
Abdominal computed tomography is the most sensitive and specific radiological study for the diagnosis of
foci of intra-abdominal infection. It provides readily interpretable information about anatomical location, localizes gas collections as being intraluminal or extraluminal, and facilitates percutaneous aspiration of cysts
and abscesses. It is especially useful in the evaluation of
the retroperitoneum, and, when performed using intravenous contrast, can differentiate viable and nonviable
tissue. However, CT is expensive and requires that the
patient be transported to the radiology suite. The accuracy of CT without contrast is reduced because it is difficult to differentiate fluid-filled bowel from an abscess.
When infection arises as a consequence of perforation of the gastrointestinal tract, studies using water-soluble contrast agents may show the site of perforation or
demonstrate distortion of normal anatomy by an adjacent abscess. Sinograms performed by the injection of
contrast solution into a therapeutic drain can document
resolution of an abscess cavity and evaluate its connection with the gastrointestinal tract. Radionuclide scans
using gallium or indium have largely given way to ultrasonography and CT.
Two retrospective studies compared the accuracy of
CT and ultrasonography for the diagnosis of intra-abdominal abscesses [26, 27]. The accuracy of ultrasound
examination ranges from 75±96 %, while CT correctly
diagnoses 71±100 % of intraperitoneal abscesses. No
statistically significant difference was demonstrated between the two imaging modalities in identifying an abdominal abscess. However, with technological improvements to both imaging modalities, the CT has generally
emerged as the preferred method for identifying intraabdominal infectious foci.
Therapeutic options in intra-abdominal infection
Is percutaneous drainage as efficacious as surgical
intervention for the treatment of intra-abdominal
abscesses?
Answer: yes, grade E.
Recommendation
The initial approach to well-defined and accessible intra-abdominal abscesses should be percutaneous drainage. Catheter drainage can also be used as a temporizing
measure to optimize the physiological and hemodynamic condition of an acutely ill patient prior to surgical exploration [28]. Laparotomy should be reserved for those
circumstances in which there are no well-defined collections, dead tissue requires dØbridement, or residual collections cannot be treated percutaneously. Surgical intervention may also be indicated to control a source of
ongoing peritoneal contamination. Rates of failure
have increased as interventional radiologists have extended the indications for percutaneous drainage. If
the clinical condition of the patient does not improve
following the initial drainage, a follow-up CT should be
performed to determine whether a residual or missed
collection is present, and surgical intervention should
be considered.
Rationale
The traditional approach to the treatment for intra-abdominal abscesses has been surgical drainage, often performed on clinical grounds alone, without definitive radiological confirmation. However, surgical drainage carries potential morbidity including bleeding, the development of fistulas, and wound infection. Improvements
in radiological imaging techniques led to new techniques for percutaneous drainage of intra-abdominal
abscesses that could be successfully employed with low
mortality [29].
Percutaneous abscess drainage was initially restricted to unilocular well-defined collections, for which a
safe drainage route could be established. Gerzof et al.
[30] initially reported a 92 % success rate for percutaneous drainage when these criteria were satisfied. Indications and access routes for percutaneous drainage have
been since expanded. Multiple, ill-defined, or complicated abscesses (for example, appendiceal, interloop,
and pelvic abscesses), collections communicating to the
gastrointestinal tract, and even abscesses whose drainage must traverse normal organs can be treated percutaneously, with a success rate of 73.6 % and 9 % mortality
[31].
S 54
Much of the literature comparing of operative and
percutaneous methods for the drainage of intra-abdominal abscesses uses historical surgical controls, without
stratification of severity of illness prior to drainage [32,
33]. There are no randomized studies comparing percutaneous and operative drainage techniques. Uncontrolled case series show that percutaneous drainage is
as effective as conventional surgery for the drainage of
intra-abdominal collections [34, 35, 36]. In a retrospective case-control study Olak and colleagues [37] stratified patients undergoing percutaneous drainage of intra-abdominal abscesses using the acute physiology
score, and matched for age, sex, diagnosis, abscess etiology, and location with a cohort of patients managed surgically. They found no differences in morbidity and mortality when patients with intra-abdominal infection were
treated surgically or percutaneously.
Are aggressive operative approaches (continuous
postoperative peritoneal lavage, open abdomen)
superior to conventional surgical treatment for intraabdominal infection?
Answer: in continuous postoperative peritoneal lavage,
uncertain, grade D; in open abdomen, no, grade D.
Recommendation
Current data support the concept that relaparotomy ªon
demand,º as indicated by worsening of the clinical status, absence of improvement, or evolving organ dysfunction is as efficacious as a more aggressive approach.
Planned relaparotomy is indicated for patients with ischemic bowel when intestinal viability is a concern
(ªsecond lookº), for patients with necrotizing pancreatitis when demarcation of necrotic tissue demarcation is
not distinct, or when bleeding precludes complete dØbridement.
Rationale
The substantial morbidity and mortality of unsuccessful
surgical management of severe intra-abdominal infection has led to the development of more aggressive surgical procedures, in particular, continuous postoperative
peritoneal lavage, and open abdomen management,
with or without planned relaparotomy. The rationale
for continuous postoperative peritoneal lavage is that
the continuous removal of bacteria and fibrin may hasten the resolution of intra-abdominal infection and reduce the risk of persistence or recurrence [38]. The technique is labor intensive and time consuming, requires
intensive care monitoring, and has been associated with
the formation of enteric fistulas. It necessitates the perioperative placement of drains in the subphrenic spaces
for infusion of dialysis solution, and of outflow catheters
in the pelvis for evacuation of the effluent. The abdomen is closed and lavage is performed with large volumes of fluid, either continuously or intermittently, for
a period of 1±5 days.
The role of postoperative peritoneal lavage remains
to be defined. A nonrandomized study by Washington
et al. [39] showed that postoperative lavage with cefamandole, erythromycin, and heparin decreases postoperative abscess formation. On the other hand, a prospective randomized control study by Hallerback et al.
[40] was unable to demonstrate benefit for postoperative lavage containing neither antibiotics nor heparin
for patients who had received broad-spectrum systemic
antibiotics. Similarly, a randomized trial of radical peritoneal dØbridement to remove fibrin from the peritoneal surface failed to show any benefit for this more aggressive form of therapy [41].
Laparostomy ± open management of the abdomen ±
has been promoted in circumstances in which multiple
reexplorations are required to control an intra-abdominal infection. The technique avoids the increased intraabdominal pressure associated with the closure of the
abdomen and facilitates reintervention. Its complications include evisceration, massive fluid losses, fistula
formation, and retraction of the abdominal wall resulting in postoperative hernias. The significant risk of
these complications led to the development of a semiopen technique, in which temporary abdominal wall closure is accomplished by the use of polypropylene and
polyglycolic acid meshes, or polytetrafluoroethylene
patches with or without zippers or adhesive sheets. This
semiopen technique facilitates reintervention without
leaving the bowel exposed, prevents evisceration,
avoids raised intra-abdominal pressures associated with
abdominal wall closure, and minimizes damage to the
abdominal wall [42].
Planned relaparotomy or staged abdominal repair
(STAR), has been advocated for the management of patients with diffuse peritonitis to prevent the development of multiple organ failure and facilitate the resolution of intra-abdominal infection [43]. STAR implies a
commitment at the first operation to perform multiple
surgical procedures at fixed intervals (24±72 h), regardless of the patient's clinical condition. In theory, planned
relaparotomy allows superior control of peritoneal contamination and earlier detection of anastomotic leaks.
The disadvantages of this method are inadvertent visceral injury and fistula formation. Increased intra-abdominal pressure resulting from abdominal wall closure
may compromise organ perfusion and pulmonary function [44].
Although favorable results with these aggressive surgical techniques have been reported worldwide [45, 46],
S 55
objective evaluation is difficult because of the heterogeneity of patients in whom the method has been employed, and the inherent difficulties associated with performing a well-designed clinical trial. A semiopen technique with planned relaparotomy was found to reduce
mortality to one third that predicted on the basis of
Acute Physiology and Chronic Health Evaluation II
scores for patients with diffuse peritonitis [47]. On the
other hand, a multicenter case-control study undertaken
by the Peritonitis Study Group of the Surgical Infection
Society Europe compared planned relaparotomy to relaparotomy on demand for the treatment of intra-abdominal infections. The two groups of patients were
comparable respect to the severity of illness, age, cause
of infection, site of origin of peritonitis, and ability to
eliminate the source of infection. No difference was
found between the groups with respect to mortality or
the need for unplanned relaparotomy. However, the incidence of anastomotic leaks, septicemia, and postoperative organ failure was increased in the patients who underwent planned relaparotomy [48].
In a collected review of 642 patients from 22 series
Schein and colleagues [42] found an overall mortality
rate of 33 % in patients treated with aggressive surgical
management compared to a range of 30±76 % in patients with intra-abdominal infection treated with conventional methods. A prospective, nonrandomized trial
comparing the closed-abdomen technique to open-abdomen technique in patients with severe peritonitis,
suggested that the major determinant of mortality in patients with intra-abdominal infection is the host response to peritoneal contamination, rather than the approach used to control bacteria in the peritoneal cavity
[1]. There currently is no clearcut evidence of the superiority of radical surgical procedures for the treatment
of severe intra-abdominal infection.
DØbridement and device removal
DØbridement ± the surgical removal of injured, necrotic,
and/or infected tissue ± is a fundamental principle of
source control, and at least in the case of necrotizing
soft tissue infections early surgical intervention is associated with a better prognosis. However, the identification of necrotic tissue may be difficult, particularly in
deep-seated infections. Moreover, as discussed earlier
for patients with pancreatic necrosis, dØbridement of
sterile necrotic tissue may not be necessary to achieve a
favorable clinical outcome.
Can the presence of tissue necrosis be ruled out by
means of nonoperative investigations?
Answer: no, grade E.
Recommendation
Although tissue necrosis can often be detected by such
characteristic radiographic findings as gas in the tissues,
or nonenhancement of tissues following administration
of intravenous contrast, there is no single test that can
exclude the presence of tissue necrosis with certainty,
and in circumstances in which necrosis may be life
threatening (for example, intestinal ischemia) it is often
necessary to establish the diagnosis operatively.
Rationale
The early diagnosis of necrotizing soft-tissue infections
may be difficult because the initial appearance of the
skin is normal, and local signs are mild, despite the presence of extensive deeper soft tissue necrosis. Local signs
of deep infection may include edema, crepitus, and cyanosis or bronzing of the skin.
Radiographic studies revealing the presence of air in
the soft tissues support the diagnosis of underlying necrosis [10]. CT with intravenous contrast can also provide information about the viability of deeper tissues.
The easy separation of the subcutaneous tissue at the
level of the fascia when probing the wound through an
incision in the skin has traditionally been considered
pathognomonic of necrotizing fasciitis, and the diagnosis of necrotizing soft-tissue infection can be confirmed
by operative exploration, with direct visualization of
whether the incised tissue bleeds. Frozen-section biopsy
early in the evolution of a suspected necrotizing soft tissue infection has been recommended in patients with
rapidly advancing infections to provide a reliable diagnosis and define the extent of dØbridement [49].
Can a vascular catheter be safely changed over a
guidewire?
Answer: yes, grade B.
Recommendation
An infected central venous catheter can be safely changed over a guidewire, provided there is not significant
local soft tissue infection at the exit site. This is a grade
B recommendation supported by level II evidence.
Rationale
The diagnosis of catheter-related infection is based on
the isolation of the same organism from the catheter
and the blood, clinical and microbiological information
S 56
ruling out another site of infection, and the presence of
signs of systemic inflammatory response. A semiquantitative technique to differentiate contamination of a
catheter from true infection was described by Maki
et al. [50], based on the demonstration of at least 15 colonies of a given organism on an agar culture roll-plate.
The presence of more than more than 103 cfu/25 cm2 or
more than 102 colonies from cultures obtained by sonication is also indicative of catheter infection [51].
Cook and colleagues [52] performed a systematic review of 12 randomized trials that addressed the relative
merits of catheter changes over a guidewire or catheter
replacement at a new site for patients with central venous catheter infections. The guidewire exchange technique was associated with fewer mechanical complications (relative risk 1.72, 95 % confidence interval
0.89±3.33) compared to replacement at a new site, but
also with a trend towards a higher rate of catheter exitsite infection (relative risk 1.52, 95 % confidence interval 0.34±6.73) and catheter-related bacteremia.
Tunneled silastic catheters such as the Hickman or
Broviac catheter have a lower rate of infection. Removal of the catheter is often required for tunnel infections
(presence erythema, induration or purulence along the
subcutaneous tract), whereas exit-site infection (signs
of infection within 2 cm of the exit site) usually resolves
with local wound care and antibiotics [53]. Tunneled hemodialysis catheters can also be safely and easily replaced over a guidewire through the preexisting subcutaneous tunnel, with infection rates comparable to de
novo catheter replacement [54].
Removal of the catheter and excision of the affected
vein leaving the wound open is the recommended approach for peripheral septic thrombophlebitis. Catheter-related septic central venous thrombosis requires
systemic antibiotics, catheter removal, and surgical excision when medical measures fail [55].
Does a policy of scheduled replacement of indwelling
central venous catheters reduce the risk of infectious
complications?
Answer: no, grade C.
Recommendation
There is no evidence that routine catheter replacement
reduces the risk of catheter-related bacteremia. Venous
catheters should be changed only as needed when evidence of infection is present (signs of inflammation,
purulent discharge at the insertion site), or when the
catheter is not working [52, 56]. This is a level C recommendation for central venous catheters, supported
by level II evidence, and a level E recommendation
for peripheral catheters, supported by level V evidence.
Rationale
A clinical trial in cardiac surgical patients evaluating the
risk of infection associated with routine guidewire exchange from a pulmonary artery to a central venous
catheter 48±72 h after the surgical procedure showed a
higher incidence of catheter-related infection (35.3 %
compared 12.5 %) in patients undergoing catheter exchange [57]. Similarly a randomized controlled trial of
routine catheter exchanges every 3 days failed to show
any reduction in rates of catheter-related infections [58].
Proximal diversion, defunctioning, and definitive therapy
An anatomical defect in the gastrointestinal tract permits continuing contamination of sterile tissues by micro-organisms found within the gut lumen. Diverting
the fecal stream or otherwise defunctioning of the gut
minimizes such contamination. Regardless of the level
of the leak proximal diversion of the gastrointestinal
tract involves two elements: drainage of the infectious
focus adjacent to the perforation and creation of a diverting stoma or ostomy proximal to the site of perforation.
For mediastinitis secondary to an intrathoracic rupture of the esophagus, exercise of this principle entails
mediastinal drainage and the creation of a cervical
esophagostomy; in the case of a perforation of the colon
secondary to cancer or diverticulitis, the principle would
dictate drainage of the abscess and the creation of a colostomy or ileostomy proximal to the site of the leak.
Application of this principle led to the classical approach to perforated diverticulitis ± a three-stage operation involving the sequential drainage of the abscess and
creation of a transverse loop colostomy, followed by a
second operation to resect the involved segment of sigmoid colon, and a final third operation to close the colostomy.
While the principle of proximal diversion remains
widely accepted as the safest and most conservative approach to complex perforations of the gastrointestinal
tract, its utility is open to evaluation. In the first place,
the approach is not applicable to many common sites
of gastrointestinal perforation. Proximal diversion of
the stomach, for example, would entail the creation of
a cervical esophagostomy, and standard therapy for perforated ulcers now consists simply of an omental patch
with or without an acid-reducing operation. Indeed a
randomized clinical trial has suggested that nonoperative management is a reasonable alternative for most
patients with perforated ulcers [59]. Secondly, drainage
S 57
Table 1 Clinical Studies Comparing Resection with Colostomy and Drainage for Perforated Sigmoid Diverticulitis
Author
No. Patients
Colostomy& Drainage
Resection
Deaths
Deaths
Complications
Comment
Complications
Kronberg, 1993 [60]
62
6/31
(19 %)
8/31
(26 %)
Smirniotis, 1992 [62]
38
4/14
(29 %)
1/24
(4 %)
Finlay, 1987 [64]
78
24 %
20 % Fistulas
21 %
0 % Fistulas
Auguste, 1985 [65]
116
10/51
20 %
Stay- 52 days
8/65
12 %
Stay- 36 days
Nagorney, 1985 [66]
121
8/31
(26 %)
and diversion, although technically simpler, leaves the
source of contamination in situ. A number of case series
have addressed the question of whether resection of the
diseased segment of colon is superior to simple drainage
and diversion, and whether anastomosis of the colon is
safe in the face of perforation and active infection.
Is resection of perforated colon preferable to simple
drainage and proximal diversion for patients with
perforated diverticulitis?
Answer: yes, grade D.
Recommendation
Definitive resection is preferable to proximal diversion
and drainage for perforated diverticulitis, and likely for
other causes of intestinal perforation, when the more
demanding procedure of resection can be performed
safely. Extension of this principle to other sites of gastrointestinal peroration such as the esophagus requires
balancing the risks of resection with the potential benefits. This is a grade D recommendation based on level
III evidence.
Rationale
Resection of a perforated segment of colon eliminates a
focus of ongoing contamination. However, the additional operative time and resulting physiological stress to a
critically ill patient may result in higher perioperative
morbidity. Moreover, definitive resection is more technically demanding; therefore the potential for significant perioperative complications is increased.
Higher rate of colostomy
closure with resection
(84 %) than drainage
(44 %)
Shorter length of disability with resection
(81 vs 148 days)
6/90
(7 %)
Medline was searched from 1985 to the present, using
as keywords, ªgastrointestinal perforation/surgery,º and
restricting the search to papers published in English. The
literature search was restricted to the past 15 years since
changes in other aspects of supportive care may have rendered earlier study conclusions less relevant. Of the 639
references identified, 73 were selected for review.
One randomized trial [60] and six retrospective case
series [61, 62, 63, 64, 65, 66] compared outcomes for patients managed by drainage and diversion or definitive
resection (Table 1). While the series were retrospective,
and the subjects randomized in only one of the trials,
consistent evidence of benefit was apparent for patients
treated with definitive resection compared to proximal
diversion and drainage. Pooled data showed a mortality
of 22.8 % (38/167) for patients treated by proximal diversion and drainage compared with 12.5 % (31/248)
for those managed with definitive resection and colostomy (odds ratio 2.06, 95 % confidence interval 1.22±3.48,
p = 0.007). It should be emphasized that intrinsic differences in patient populations, or in adjuvant care may
have created a spurious suggestion of superiority for resection. Resection also appears to be associated with
lower morbidity, principally a shorter hospital stay and
period of disability, a lower rate of postoperative fistulas, and a higher rate of colostomy reversal.
Following colonic resection for perforated diverticulitis,
is primary anastomosis a safe alternative to the
construction of a colostomy?
Answer: yes, grade D.
S 58
Table 2 Clinical Studies Comparing Primary Anastomosis to Colostomy Following Resection for Perforated Diverticulitis
Author
Umbach, 1999 [67]
Setti Carraro, 1999 [68]
No. of Patients
Primary Anastomosis
Resection and Colostomy
Deaths
Complications
Deaths
Complications
33
0
1 anastomotic
dehiscence
N/A
N/A
No control group;
mortality from literature
1±28 % for Hartmann
procedure
105
±
±
±
±
No difference in outcome; mortality related
to APACHE II
Strada, 1993 [70]
73
4/73
(5.5 %)
11/73
(15 %)
Saccomani, 1993 [71]
38
1/26
(3.8 %)
11/26
(42 %)
Smirniotis, 1992 [62]
24
1/6
(17 %)
Alanis, 1989 [73]
60
1/34
(3 %)
2/6
(33 %)
±
Recommendation
Primary anastomosis or colostomy are equally efficacious following colon resection for diverticulitis. The
choice of procedure should be dictated by other factors
such as severity of illness, presence of chronic disease,
the degree or duration of peritoneal contamination, and
the skill and experience of the surgical team. This is a
grade D recommendation based on level III evidence.
Rationale
A primary anastomosis obviates the need for a second
operative procedure to close the stoma, a procedure
with a recognized morbidity. On the other hand, the
risk of anastomotic leakage is generally held to be greater in the face of an acute inflammatory process, and the
attendant morbidity resulting from an anastomotic leak
may be formidable. Nine reports were identified that
evaluated the safety of primary anastomosis in the clinical setting of colon perforation with established inflammation [61, 62, 67, 68, 69, 70, 71, 72, 73]; four of these reported rates for concomitant controls. Six studies were
judged to be evaluable (Table 2). All studies were case
series. In no study was primary anastomosis associated
with an increased risk of mortality, and two reports suggested that the primary determinant of mortality was
the degree of physiological derangement, rather than
the specific method of therapy used. Strada and colleagues [70] reviewed the course of 73 patients injured
on the battlefield during the Afghan War, all of whom
were managed by primary anastomosis or direct suture
repair, without creation of a stoma. Despite the fact
Comment
Battlefield management
of colonic injuries during
Afghan War
3/8
(38 %)
5/8
(63 %)
0/18
(0 %)
1/18
(6 %)
4/26
(15 %)
±
1 additional death
with colostomy closure
that nearly half the patients were hypotensive at presentation, and that management was undertaken more than
12 h after injury in 445, only 4 patients died (5.5 %).
The independent morbidity and mortality associated
with colostomy closure is an additional variable to be
considered. Complications develop during one-third of
colostomy closures [74, 75, 76], although the majority
are relatively minor. In addition, colostomy closure necessitates an additional hospital stay and postoperative
convalescence.
Pooled data from the four studies that compared primary anastomosis with colostomy following colon resection for diverticulitis [62, 71, 72, 73] showed a modestly
lower mortality for patients undergoing primary anastomosis (3/69 or 4.3 % vs. 8/55 or 14.5 %, odds ratio 0.27,
95 % confidence interval 0.07±1.06, p = 0.06). Publication bias and the absence of randomization preclude
drawing firm conclusions from these data; however,
they suggest that, contrary to classical teaching, primary
anastomosis does not carry a higher mortality rate.
Dealing with diagnostic uncertainty
Does empirical or ªblindº laparotomy improve
outcome?
Answer: no, grade E.
Recommendation
Intra-abdominal infectious complications mandating
source control are almost always evident using modern
S 59
diagnostic imaging techniques. There is little if any role
for empirical laparotomy to rule out undiagnosed infection in a critically ill patient in whom radiological examination has failed to demonstrate a surgically correctable problem.
Rationale
The first descriptions of the multiple organ dysfunction
syndrome emphasized its association with uncontrolled
and/or occult infection, particularly infection within the
abdomen [77, 78, 79]. It remains a truism that the presence of remote organ dysfunction in critically ill patients
should trigger an aggressive search for an unidentified
focus of infection, especially when organ dysfunction
evolves rapidly.
The association between organ dysfunction and occult intra-abdominal infection stimulated led to the concept of ªblind laparotomy,º undertaken in the absence
of radiological evidence of infection to search for an undrained abscess [80, 81]. However, improved radiological techniques have tempered the initial enthusiasm for
this approach. Significant occult pathology is relatively
uncommon, and when it is encountered, there is no convincing evidence that a more aggressive approach alters
outcome [82, 83, 84].
With the spectrum of radiological diagnostic techniques currently available, interpreted within the specific clinical context of an individual patient, there is little
justification for the temptation to ªhave a lookº for
fear that a clinically important infection that might
benefit from source control measures is being missed.
Summary
The process of surgical decision making is based on both
general principles that are amenable to evaluation using
rigorous techniques of clinical research and the intangible element of surgical judgment that seeks to apply
those principles to the care of an individual patient.
The role of surgical judgment is inescapable, even
though it is intrinsically subjective and recalcitrant to
objective evaluation, for a host of factors modify the application of principle in each patient, and render the circumstances of a given problem sufficiently distinctive,
that evidence must be tempered with common sense.
We have tried to provide, through an evidence-based
approach to a series of questions, the rationale for the
basic principles that should guide the clinician in initiating or modifying source control, recognizing that sound
clinical judgement demands, at times, that these be set
aside. In the individual patient, evidence of clinical improvement is the most important marker of the approach selected.
Evaluation of the adequacy of source control in the
critically ill patient can be difficult. As with other modes
of anti-infective therapy, effective source control measures are expected to result in clinical improvement, reflected in:
· Resolution of clinical signs of sepsis or systemic inflammation
· Bacteriological resolution
· Evidence of reversal of the metabolic sequelae of infection, with normal progression of wound healing,
reflected in the formation of granulation tissue, and
epithelialization
· Radiographic evidence of control of an infectious focus
· Prevention of further organ dysfunction, and resolution of existing organ dysfunction
· Survival
Evaluation of the adequacy of source control may necessitate planned reoperation. The adequacy of dØbridement of necrotizing soft-tissue infections can be assessed by repeat exploration under general anesthesia,
continuing the process until there is evidence of healthy
granulation tissue throughout the wound. Planned reexploration is also indicated for patients with diffuse intestinal ischemia to ensure bowel viability.
The appropriate interventions to determine the adequacy of source control are dictated by the clinical circumstances. A residual or recurrent abscess can usually
be demonstrated by CT or ultrasound examination,
while resolution of an abscess cavity can be monitored
using sinograms. The diagnosis of persistent or evolving
tissue necrosis is guided by the clinical setting. Retroperitoneal necrosis can be detected by CT, while sigmoid ischemia following aortic aneurysmectomy can be
evaluated by sigmoidoscopy. Occasionally diagnostic
peritoneal lavage assists in establishing a diagnosis of
gut ischemia; the lavage fluid appears bloody with established ischemia. The diagnosis of an infected foreign
body requires an appropriate history and is supported
by recurrent bacteremia or by positive cultures drawn
retrograde through an indwelling vascular or peritoneal
dialysis catheter. Finally, ongoing contamination from a
breach of the gastrointestinal tract can be documented
by appropriate contrast studies.
The general principles that guide the use of source
control techniques in the management of the patient
with severe sepsis or septic shock are readily articulated.
Their implementation in practice, however, is more
complex, and does not, as a rule, lend itself to simple algorithms that are applicable in all cases. Moreover evidence-based support for these principles is weak. In the
final analysis, the elusive process of experienced surgical judgement is invaluable for all but the most straightforward problems.
S 60
References
1. Christou NV, Barie PS, Dellinger EP,
Waymack JP, Stone HH (1993) Surgical
Infection Society intra-abdominal infection study. Prospective evaluation of
management techniques and outcome.
Arch Surg 128: 193±198
2. Finlay-Jones JJ, Davies KV, Sturm LP,
Kenny PA, Hart PH (1999) Inflammatory processes in a murine model of
intra-abdominal abscess formation.
J Leukoc Biol 66: 583±587
3. Mier J, Leon EL, Castillo A, Robledo F,
Blanco R (1997) Early versus late necrosectomy in severe necrotizing pancreatitis. Am J Surg 173: 71±75
4. Dougherty SH (1988) Pathobiology of
infection in prosthetic devices. Rev Infect Dis 10: 1102±1117
5. Patrick CC, Plaunt MR, Hetherington
SV, May SM (1992) Role of the Staphylococcus epidermidis slime layer in experimental tunnel tract infections. Infect Immun 60: 1363±1367
6. Sprung CL, Finch RG, Thijs LG, Glauser MP (1996) International sepsis trial
(INTERSEPT): role and impact of a
clinical evaluation committee. Crit
Care Med 24: 1441±1447
7. Fisher JR, Conway MJ, Takeshita RT,
Sandoval MR (1979) Necrotizing fasciitis. Importance of roentgenographic
studies for soft tissue gas. JAMA 241:
803±806
8. Strasberg SM, Silver MS (1968)
Hemolytic Streptococcus gangrene: an
uncommon but frequently fatal infection in the antibiotic era. Am J Surg
115: 763±768
9. Rea WJ, Wyrick WJ (1970) Necrotizing
fasciitis. Ann Surg 172: 957±964
10. Majeski JA, Alexander JW (1983) Early diagnosis, nutritional support, and
immediate extensive debridement improve survival in necrotizing fasciitis.
Am J Surg 145: 784±787
11. Miller JD (1983) The importance of
early diagnosis and surgical treatment
of necrotizing fasciitis. Surg Gynecol
Obstet 157: 197±200
12. Bilton BD, Zibari GB, McMillan RW,
Aultman DF, Dunn G, McDonald JC
(1998) Aggressive surgical management
of necrotizing fasciitis serves to decrease mortality: a retrospective study.
Am Surg 64: 397±400
13. Voros D, Pissiotis C, Georgantas D,
Katsaragakis S, Antoniou S, Papadimitriou J (1993) Role of early and extensive surgery in the treatment of severe
necrotizing soft tissue infections. Br J
Surg 80: 1190±1191
14. Freischlag JA, Ajalat G, Busuttil RW
(1985) Treatment of necrotizing soft tissue infections. A need for a new approach. Am J Surg 149: 751±755
15. Adant JP, Bluth F, Fissette J (1998) Necrotizing fasciitis: a life-threatening infection. Acta Chir Belg 98: 102±106
16. Sjo OH, Ivarsson LE, Sudhange JO,
Mikkelsen EM (1999) Necrotizing fasciitis ± a surgical challenge. Tidsskr Nor
Laegeforen 119: 2660±2663
17. Kujath P, Eckmann C, Benecke P
(1996) Standardized treatment of necrotizing fasciitis. Zentralbl Chir 121:
35±43
18. Moss RL, Musemeche CA, Kosloske
AM (1996) Necrotizing fasciitis in children: prompt recognition and aggressive therapy improve survival. J Pediatr
Surg 31: 1142±1146
19. Kaiser RE, Cerra FB (1981) Progressive necrotizing surgical infections ± a
unified approach. J Trauma 21: 349±355
20. Sudarsky LA, Laschinger JC, Copppa
GF, Spencer FC (1987) Improved results from standardized approach in
treating patients with necrotizing fasciitis. Ann Surg 206: 661±665
21. Laor E, Palmer LS, Tolia BM, Reid RE,
Winter HI (1995) Outcome prediction
in patients with Fournier's gangrene.
J Urol 154: 89±92
22. Hsiao GH, Chang CH, Hsiao CW, Fanchiang JH, Jee SH (1998) Necrotizing
soft tissue infections. Surgical or conservative treatment? Dermatol Surg 24:
243±247
23. Beger HG, Bittner R, Block S, Büchler
M (1986) Bacterial contamination of
pancreatic necrosis. A prospective clinical study. Gastroenterology 91: 433±438
24. Machado MC, Bacchella T, Monteiro
da Cunha JE (1986) Surgical treatment
of pancreatic necrosis. Dig Dis Sci 31:
25
25. Ochsner A, DeBakey M (1938) Subphrenic abscess: collective review and
analysis of 3: 608 collected and personal
cases. Surg Gynecol Obstet 66: 426
26. Korobkin M, Callen PW, Filly RA, Hoffer PB, Shimahak RR, Kressel HY
(1978) Comparison of computed tomography, ultrasonography, and gallium-67 scanning in the evaluation of suspected abdominal abscess. Radiology
129: 89±93
27. Knochel JQ, Koehler PR, Lee TG,
Welch DM (1980) Diagnosis of abdominal abscesses wih computed tomography, ultrasound, and 111In leukocyte
scans. Radiology 137: 425±432
28. Karlson KB, Martin EC, Frankuchen
EI (1981) Non-surgical drainage of intra-abdominal and mediastinal abscesses: a report of 12 cases. Cardiovasc Intervent Radiol 4: 170±176
29. Saini S, Kellum J, O'Leary M, O'Donnell T, Tally F, Carter B, Deterling RA,
Curtis LE (1983) Improved localization
and survival in patients with intra-abdominal abscesses. Am J Surg 145:
136±142
30. Gerzof SG, Robbins AH, Johnson WC,
Birkett DH, Nabseth DC (1981) Percutaneous catheter drainage of abdominal
abscesses. A five year experience.
N Engl J Med 305: 653±657
31. Gerzof SG, Robbins AH, Johnson WC,
Birkett DH, Nasbeth DC (1985) Expanded criteria for percutaneous abscess drainage. Arch Surg 120: 227±232
32. Johnson WC, Gerzof SG, Robbins AH,
Nabseth DC (1981) Treatment of abdominal abcesses. Comparative evaluation of operative drainage versus percutaneous catheter drainage guided by
computed tomography or ultrasound.
Ann Surg 194: 510±520
33. Glass CA, Cohn Ijr (1984) Drainage of
intraabdominal abscesses. A comparison of surgical and computerized tomography guided catheter drainage.
Am J Surg 147: 315±317
34. Levison MA, Ziegler D (1991) Correlation of APACHE II score, drainage
technique, and outcome in postoperative intraabdominal abscess. Surg Gynecol Obstet 172: 89±94
35. Hemming A, Davis NL, Robins RE
(1991) Surgical versus percutaneous
drainage of intra-abdominal abscesses.
Am J Surg 161: 593±595
36. Bufalari A, Giustozzi G, Moggi L
(1996) Postoperative intraabdominal
abscesses: percutaneous versus surgical
treatment. Acta Chir Belg 96: 197±200
37. Olak JO, Christou NV, Stein LA, Casola G, Meakins JL (1986) Operative versus pecutaneous drainage of intra-abdominal abscesses. Arch Surg 121:
141±146
38. Stephen M, Loewenthal J (1979) Continuing peritoneal lavage in high-risk
peritonitis. Surgery 85: 603±606
39. Washington BC, Villalba MR, Lauter
CB, Colville J, Starnes R (1983) Cefamandole erythromycin heparin peritoneal irrigation: an adjunct to the surgical treatment of diffuse bacterial peritonitis. Surgery 94: 576±581
S 61
40. Hallerback B, Andersson C, Englund
N, Glise H, Nihlberg A, Solhaug J,
Wahlstrom B (1986) A prospective randomized study of continuous peritoneal
lavage postoperatively in the treatment
of purulent peritonitis. Surg Gynecol
Obstet 163: 433±436
41. Polk HC, Fry DE (1980) Radical peritoneal debridement for established peritonitis: the results of a prospective randomized clinical trial. Ann Surg 192:
350±355
42. Schein M, Hirshberg A, Hashmonai M
(1992) Current surgical management of
severe intraabdominal infection. Surgery 112: 489±496
43. Schein M (1991) Planned re-operations
and open management in critical intraabdominal infections. Prospective personal experience in 52 cases. World J
Surg 15: 537±545
44. Schein M, Wittmann DH, Aprahamian
CC, Condon RE (1995) The abdominal
compartment syndrome: the physiological and clinical consequences of elevated intra-abdominal pressure. J Am
Coll Surg 180: 745±753
45. Wittmann DH, Aprahamian C, Bergstein JM (1990) Etappenlavage:
avanced diffuse peritonitis managed by
planned multiple laparotomies utilizing
zippers, slide fastener, and velcro analogue for temporary abdominal closure.
World J Surg 14: 218±226
46. Mughal MM, Bancewicz J, Irving MH
(1986) Laparostomy: a technique for
the management of intractable intraabdominal sepsis. Br J Surg 73: 253±259
47. Garcia-Sabrido JL, Tallado JM, Christou NV, Polo JR, Valdecantos E (1988)
Treatment of severe intra-abdominal
sepsis and or necrotic foci by an ªopen
abdomenº approach. Arch Surg 123:
152±156
48. Hau T, Ohmann C, Wolmershauser A,
Wacha H, Yang Q (1995) Planned relaparotomy versus relaparotomy on demand in the treatment of intra-abdominal infections. Arch Surg 130:
1193±1196
49. Pruitt BA Jr (1984) Biopsy diagnosis of
surgical infections. N Engl J Med 310:
1737±1738
50. Maki DG, Weise CE, Sarafin HW
(1977) A semiquantitative culture
method for identifying intravenous
catheter-related infection. N Engl J
Med 296: 1305±1309
51. Sherertz RJ, Heard SO, Raad II (1997)
Diagnosis of triple-lumen catheter infection: comparison of roll-plate, sonication, and flushing methodologies.
J Clin Microbiol 35: 641±646
52. Cook DJ, Randolph A, Kernerman P, et
al (1997) Central venous catheter replacement strategies: a systematic review of the literature. Crit Care Med
25: 1417±1424
53. Clarke DE, Raffin TA (1990) Infectious
complications of indwelling long-term
central venous catheters. Chest 97:
966±972
54. Dusvak RJr, Haskal ZJ, Thomas-Hawkins C, Soulen MC, Baum RA, Shlansky-Goldberg RD, Cope C (1998) Replacement of failing tunneled hemodialysis catheters through pre-existing
subcutaneous tunnels: a comparison of
catheter function and infection rates
for de novo replacements and over-thewire exchanges. J Vasc Interv Radiol 9:
321±327
55. Kaufman J, Demas C, Stark G (1986)
Catheter-related septic central venous
thrombosis: current therapeutic options. West J Med 145: 200±203
56. Bregenzer T, Conen D, Sakmann P, et al
(1998) Is routine replacement of peripheral intravenous catheters necessary? Arch Intern Med 158: 151±156
57. Bach A, Bohrer H, Geis HK (1992)
Safety of a guidewire technique for replacement of pulmonary artery catheters. J Cardiothorac Vasc Anesth 6:
711±714
58. Cobb DK, High KP, Sawyer RG, et al
(1992) A controlled trial of scheduled
replacement of central venous and pulmonary artery catheters. N Engl J Med
327: 1062±1068
59. Crofts TJ, Park KG, Steele RJ, Chung
SS, Li AK (1989) A randomized trial of
nonoperative treatment for perforated
peptic ulcer. N Engl J Med 321: 970±973
60. Kronborg O (1993) Treatment of perforated sigmoid diverticulitis: a prospective randomized trial. Br J Surg 80:
505±507
61. Tucci G, Torquati A, Grande M, Stroppa I, Siaesi M, Farinon AM (1996) Major acute inflammatory complications
of diverticular disease of the colon:
planning of surgical treatment. Hepatogastroenterology 43: 839±845
62. Smirniotis V, Tsoutsos D, Fotopoulos
A, Pissiotis AC (1992) Perforated diverticulitis: a surgical dilemma. Int Surg 77:
44±47
63. Corder AP, Williams JD (1990) Optimal
operative treatment in acute septic
complications of diverticular disease.
Ann R Coll Surg Engl 72: 82±86
64. Finlay IG, Carter DC (1987) A comparison of emergency resection and staged
management in perforated diverticular
disease. Dis Colon Rectum 30: 929±933
65. Auguste L, Borrero E, Wise L (1985)
Surgical management of perforated colonic diverticulitis. Arch Surg 120:
450±452
66. Nagorney DM, Adson MA, Pemberton
JH (1985) Sigmopid diverticulitis with
perforation and generalized peritonitis.
Dis Colon Rectum 28: 71±75
67. Umbach TW, Dorazio RA (1999) Primary resection and anastomosis for perforated left colon lesions. Am Surg 65:
931±933
68. Setti Carraro PG, Magenta A, Segala
M, Ravizzini C, Nespoli A, Tiberio G
(1999) Predictive value of a pathophysiological score in the surgical treatment
of perforated diverticular disease. Chir
Ital 51: 31±36
69. Kriwanek S, Armbruster C, Beckerhinn
P, Dittrich K (1994) Prognostic factors
for survival in colonic perforation. Int J
Colorectal Dis 9: 158±162
70. Strada G, Raad L, Belloni G, Setti Carraro P (1993) Large bowel perforations
in war surgery: one-stage treatment in
a field hospital. Int J Colorectal Dis 8:
213±216
71. Saccomani GE, Santi F, Gramegna A
(1993) Primary resection with and without anastomosis for perforation of acute
diverticulitis. Acta Chir Belg 93:
169±172
72. Medina VA, Papanicolaou GK, Tadros
RR, Fielding LP (1991) Acute perforated diverticulitis: primary resection and
anastomosis? Conn Med 55: 258±261
73. Alanis A, Papanicolaou GK, Tadros
RR, Fielding LP (1989) Primary resection and anastomosis for treatment of
acute diverticulitis. Dis Colon Rectum
21: 933±939
74. Del Pino A, Cintron JR, Orsay CP,
Pearl RK, Tan A, Abcarian H (1997)
Enterostomal complications: are emergently created enterostomas at greater
risk? Am Surg 63: 653±656
75. Ghorra SG, Rzeczycki TP, Natarajan R,
Pricolo P (1999) Colostomy closure: impact of preoperative risk factors on
morbidity. Am Surg 65: 266±269
76. Belmonte C, Klas JV, Perez JJ, Wong
WD, Rothenberger DA, Goldberg SM,
Madoff RD (1996) The Hartmann procedure. First choice or last resort in diverticular disease? Arch Surg 131:
612±617
77. Polk HC, Shields CL (1977) Remote organ failure: a valid sign of occult intraabdominal infection. Surgery 81:
310±313
78. Fry DE, Pearlstein L, Fulton RL, Polk
HC (1980) Multiple system organ failure. The role of uncontrolled infection.
Arch Surg 115: 136±140
S 62
79. Bell RC, Coalson JJ, Smith JD, Johanson WG (1983) Multiple organ system
failure and infection in adult respiratory
distress syndrome. Ann Intern Med 99:
293±298
80. Ferraris VA (1983) Exploratory laparotomy for potential abdominal sepsis
in patients with multiple organ failure.
Arch Surg 118: 1130±1133
81. Hinsdale JG, Jaffe BM (1984) Re-operation for intra-abdominal sepsis. Indications and results in modern critical care
setting. Ann Surg 199: 31±36
82. Bunt TJ (1986) Non-directed relaparotomy for intraabdominal sepsis: a futile
procedure. Am Surg 52: 294±298
83. Norton LW (1985) Does drainage of intraabdominal pus reverse multiple organ failure? Am J Surg 149: 347±350
84. McCrory C, Crowley K (1997) Is repeat
laparotomy of value in patients with
suspected intra-abdominal sepsis in the
intensive care unit? Ir J Med Sci 166:
89±91
Intensive Care Med (2001) 27: S 63±S 79
Airway and lung in sepsis
Greg S. Martin
Gordon R. Bernard
)
G. S. Martin ´ G. R. Bernard ( )
Center for Lung Research, Vanderbilt University Medical Center,
Division of Allergy, Pulmonary and Critical Care Medicine,
Nashville, Tenn., USA
E-mail: [email protected]
Phone: +1-6 15-3 43 00 77
Fax: +1-6 15-3 43 74 48
Introduction
The development of respiratory dysfunction in patients
with sepsis presents a myriad of complex interactions,
which are yet to be completely understood. Similar to
sepsis itself, the respiratory dysfunction that accompanies sepsis lies on a continuum from subclinical disease
to overwhelming organ dysfunction. The most dreaded
respiratory complication of sepsis is the acute respiratory distress syndrome (ARDS) ± a severe form of acute
lung injury (ALI) at the end of the spectrum of respiratory dysfunction.
Sepsis itself encompasses an entire range of host inflammatory responses, most frequently generated in humans by an infectious source. As the criteria defining
the sepsis syndrome have become more established,
the ability to determine specific epidemiological information associated with the syndrome has become more
feasible. Recent compilations suggest a rising incidence
of sepsis, likely resulting from advancing age of the population, potent immunosuppressive medications, and increasing numbers of invasive procedures [1]. In addition, data compiled by the Centers for Disease Control
indicate that the incidence of sepsis increased more
than 100 % from 1979 to 1987, although the lack of contemporaneous standardized definitions may make this
statistic exaggerated. The sepsis syndrome remains one
of the most commonly recognized predisposing conditions for ALI, accounting for approximately 40 % of
cases [3, 4]. Pulmonary and intra-abdominal infections
are the most commonly associated sites of infection
identified in patients suffering ALI related to sepsis [5].
The development of ARDS in patients with sepsis is reported to occur in 25±42 % of patients, increasing with
persistent arterial hypotension [6].
Since its description in 1967, the defining criteria of
ARDS have varied. Most physicians include the presence of bilateral pulmonary infiltrates on frontal chest
radiography, impaired gas exchange, and the absence
of cardiac dysfunction. Many investigators believe reduced respiratory system compliance, increased extravascular lung water, or other biochemical markers of
inflammation should be included [7]. The AmericanEuropean Consensus Conference on ARDS created a
uniform definition for ALI and ARDS in 1994, outlined
in Table 1. These criteria have allowed more precise epidemiological estimates to be made, although the incidence has been reported to vary from 5 to 71 per
100000 persons in the United States [8, 9]. Imprecision
in these statistics makes quantification of the financial
burden of this disorder difficult, although rational yearly estimates approach $5 billion in the United States
alone. Broad, cooperative studies to obtain more precise estimates are underway.
The morbidity and mortality associated with ALI
and ARDS may be declining slowly, although it is widely considered to remain in excess of 40 %. Mortality is
most often due to unresolved sepsis or multisystem organ failure (MOF) as opposed to progressive respiratory failure [5]. A recent review based on data from the
period 1983±1993 in Seattle suggests that mortality rates
have declined slowly over time, particularly in young
patients with lung injury related to sepsis [10]. Several
factors have been consistently found to affect mortality
in patients with ALI, including age, severity of illness,
cause of lung injury, presence of MOF, and preexisting
comorbid conditions [11]. The degree of initial hypoxemia is not a reliable prognostic indicator, although
changes in oxygenation over the first 48 h appear to dis-
S 64
Table 1 Definitions of acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) (adapted from the American-European
Consensus Conference)
ALI
ARDS
Timing
Degree of oxygenation defect
Radiographic appearance
Hydrostatic component
Acute
Acute
PaO2/FIO2 £300
PaO2/FIO2 £200
Bilateral infiltrates
Bilateral infiltrates
PAOP < 18 mmHg
PAOP < 18 mmHg
Table 2 Calculation of the Lung Injury Score: the total number of points is divided by the number of components used
Points
0
1
2
3
4
Chest radiography
PaO2/FIO2 ratio
PEEP (cmH2O)
Cstat (ml/cmH2O)
No infiltrates
> 300
5
80
1 quadrant
225±299
6±8
60±79
2 quadrants
172±224
9±11
40±59
3 quadrants
100±174
12±14
20±39
4 quadrants
< 100
15
19
criminate eventual outcomes. In 1988 the publication of
the lung injury score (LIS, Table 2) provided a method
of grading the severity of lung injury ± a system that
has been validated prospectively for prognostic purposes. An initial LIS higher than 3.5 has been observed to
be correlated with a survival rate of 18 %, while a score
of 2.5±3.5 corresponds to a survival rate of 30 %, a score
of 1.1±2.4 with a 59 % survival, and a score below 1.1
with a 66 % rate of survival [12].
This contribution addresses the respiratory system
complications encountered in patients with sepsis, with
a focus on clinically relevant diagnostic methods and
management options for the practicing critical care physician.
Methods
A computer-based review of the literature was undertaken using
Medline from 1993 to the present as the primary database. The
specific subject heading keywords defined for each question were
combined with the following general sepsis-related subject heading
keywords: infection, sepsis, sepsis syndrome, septic shock, multiple
organ failure, critical care, intensive care. An additional manual retrieval of pertinent cited articles from the retrieved literature was
performed.
Pathophysiology
Biochemical and cellular mediators
The hallmark of ALI/ARDS is alveolar epithelial inflammation, airspace flooding with plasma proteins and
cellular debris, surfactant depletion and inactivation,
and a loss of normal endothelial reactivity [13]. This article is not intended to serve as a reference for the biochemical and cellular mediators of sepsis-induced respiratory dysfunction, although their interplay is indisputably critical to the pathophysiology common to the syn-
drome. The pathogenesis of ALI/ARDS is complex,
with a consistently observed broad activation of the
host inflammatory response. A well described pathophysiological model of ALI/ARDS is one of acute lung
inflammation mediated by neutrophils, cytokines, and
oxidant stress [14]. Bronchoalveolar lavage fluid from
patients with ALI contains increased quantities of neutrophils and their enzymes, both of which are correlated
with the severity of lung injury [15]. While it is clear that
neutrophils exert a critical role in the evolution of the
host inflammatory response, neutropenic patients can
develop ALI, thus supporting a pivotal role for other effector cells such as alveolar macrophages [16]. Both of
these cell types produce inflammatory mediators, catalyze the generation of reactive oxygen species, and encourage lipid peroxidation through arachidonic acid
metabolism pathways. Persistent plasma elevations in
the proinflammatory cytokines such as tumor necrosis
factor-a and interleukins (ILs) 1, 6, and 8 are correlated
with reduced survival, while increases in the bronchoalveolar lavage fluid anti-inflammatory cytokines such as
IL-10 are correlated directly with survival [17, 18, 19,
20].
Evidence of lipid peroxidation and oxidant stress is
uniformly present in patients with sepsis, with reported
elevations in hypoxanthine and numerous arachidonic
acid metabolites (e.g., isoprostanes) [21, 24]. Plasma thiol levels have been found to be correlated with survival
in patients with ARDS, while lipid peroxidation products are correlated with severity of disease and survival
[25]. Cytokine expression has been known to be regulated by oxidative stress mechanisms; antioxidants such as
N-acetylcysteine and tocopherol derivatives have been
shown experimentally to reduce the expression of proinflammatory cytokines [26, 27, 28]. Recent evidence exists that mechanisms independent of oxidant control
also contribute to cytokine expression [29, 30]. In addition, endogenous (e.g., transferrin and ceruloplasmin)
and exogenous substances (albumin) with the ability to
S 65
chelate iron have been shown effectively to suppress
proinflammatory cytokine production in vitro [31, 32].
Unfortunately, the endogenous antioxidants have been
shown to be either insufficient or inactive, rendering
the natural oxidant stress defense mechanisms ineffectual. As the lung functions to filter nearly the entire cardiac output, it may thus be injured as a passive participant in the systemic inflammatory cascade.
Gas exchange
The principal cause of hypoxemia associated with sepsis
is extensive right-to-left intrapulmonary shunting of
blood flow. Intrapulmonary shunting is normally limited
to less than 5 % of the total cardiac output, whereas in
ARDS it may consume more than 25 % of the total cardiac output. In ARDS, shunting is due to persistent perfusion of atelectatic and fluid-filled alveoli. Ordinarily,
compensation occurs through hypoxic pulmonary vasoconstriction to limit the amount of shunt by reducing
perfusion to poorly ventilated lung units. In states of
lung injury, however, hypoxic pulmonary vasoconstriction may be ineffective or absent, thereby increasing
the magnitude of the intrapulmonary shunt. After the
initial insult to the lung, gradients appear along a gravitational axis, in which the dependent lung is extensively
consolidated and the main source of venous admixture
[33]. Another factor affecting the ability to compensate
for intrapulmonary shunting may be differences among
patients in the mixed venous oxygen concentration of
blood perfusing the injured lung regions as a result of
differences in cardiac output or tissue oxygen consumption. Shunting of blood through nonventilated lung units
accounts for the relative refractory nature of hypoxemia
in ARDS. As a means to improve oxygenation, manipulation of airway pressure is often required to restore
ventilation to nonventilated lung units.
Lung mechanics
Decrements in lung compliance (the change in lung volume for a given change in transpulmonary pressure) related to small airway and alveolar collapse are nearly
universal in patients with ALI/ARDS. When delivered
by mechanical ventilation with no end-expiratory pressure, the static inflation pressure for typical tidal volumes of 8 ml/kg may exceed 25 cmH2O. This implies
lung compliance approaching 20 ml/cmH2O, or less
than one-fourth that of normal. To reflect the actual intrinsic elastic properties of lung tissue, compliance
should be calculated with the quantity of lung participating in gas exchange. In early ARDS the volume of
aeratable lung is reduced by alveolar edema and surfactant dysfunction. These changes account for the need
for higher inflation pressures, exclusive of any change
in the intrinsic elastic properties of lung. As such, the inflation pressure may function as an estimation of the
amount of edema and atelectasis early in the course of
ARDS. This is reflected in the concept of a ªsmallº
lung early in ARDS versus a ªstiffº lung later in the
course. Only if fibrosis develops in the later phases do
increases in inflation pressures reflect true changes in
lung compliance. In a person with normal lungs, a transpulmonary pressure of 30 cmH2O is sufficient to achieve
total lung capacity ± thus the recommended pressure
limit for mechanical ventilation adopted by the American College of Chest Physicians Consensus Conference
[34]. Interestingly, this level of airway pressure has also
been shown to induce lung injury in some animals [35].
Whereas the static inflation pressure is the best index
of transalveolar pressure during mechanical ventilation,
the mean airway pressure is the best predictor of an
overall effect on oxygenation or hemodynamics. As the
mean airway pressure increases, progressively greater
amounts of potentially recruitable lung are recruited.
Unfortunately, at the same time venous return can be
impeded and cardiac output depressed. Because volume-related alveolar overdistension is now recognized
to play a major role in airway pressure associated injury
in ARDS, the term ªvolutraumaº (instead of barotrauma) has been coined.
At times, peak airway pressures during mechanical
ventilatory support of patients with ARDS are increased out of proportion to the increase in static inflation pressures. This finding suggests an increase in airway resistance. Airway secretions, edema, mediators
that provoke bronchospasm, narrow endotracheal
tubes, etc. can all increase airway resistance. Airway resistance, as with compliance, should be normalized for
the amount of aeratable lung volume available and, although abnormal in part due to bronchoconstriction,
the extent to which airway resistance is increased in
ARDS is not completely known [36].
Work of breathing
Because these changes in mechanical properties increase the airway pressure necessary to achieve a given
tidal volume, the work of breathing (measured as the
pressure-volume product during spontaneous breaths)
is also increased in ARDS ± an effect that is multiplied
by coincident tachypnea. One cause of increased deadspace ventilation is hyperventilation of still normal or
relatively normal alveolar units, a process exaggerated
by differences in the distribution of ventilation with mechanical ventilatory support and by overinflation of normal lung units when mean airway pressure is increased
by positive end-expiratory pressure (PEEP) or other
maneuvers. Normally, the dead space-to-tidal volume
S 66
ratio (Vd/Vt) is 0.3, but in severe ARDS as much as
90 % (Vd/Vt = 0.9) of each tidal volume may fail to participate in effective gas exchange. As a consequence
minute ventilation greater than five times normal may
be necessary to maintain normal arterial CO2 concentrations.
The increases in work of breathing are multifactorial,
including hypoxemia, dead-space ventilation, and increased airflow resistance from bronchoconstriction.
Although normal work of breathing accounts for only a
small percentage of the body's overall oxygen consumption, the work of spontaneous breathing in patients with
ARDS may require nearly 50 % of the body's total oxygen consumption. To supply the energy necessary to sustain this level of work, resources (i.e., relative blood
flow) may have to be diverted from other vital organ
systems.
Patients with lung injury related to sepsis demonstrate heightened airway resistance, as least partially
mediated by bronchoconstriction, although this phenomenon has not been shown to exist specifically in sepsis (see ªLung mechanics,º above). The combination of
altered airflow and abnormal pulmonary vascular perfusion (with loss of hypoxic pulmonary vasoconstriction)
contributes to dramatic alterations in ventilation ± perfusion matching, which may lead to clinically disproportionate hypoxemia relative to radiographic changes.
Extravascular lung water
The equation described by Starling in 1896 characterizes fluid flux across a semipermeable membrane and has
been applied both experimentally and clinically to predict pulmonary edema formation in humans. The prime
factors in this equation are the hydrostatic and oncotic
gradients between the vasculature and interstitium coupled with the degree of capillary permeability. When
fluid deposition exceeds the capacity of the lung to remove such fluid (i.e., lymphatic flow), accumulation of
extravascular water occurs. Patients with sepsis demonstrate variable degrees of capillary permeability, increasing the effect of the hydrostatic pressure gradient
relative to the oncotic pressure gradient as molecules responsible for maintaining oncotic pressure may be allowed freely to cross such leaky barriers. Accumulation
of extravascular lung water and exudation of plasma
proteins into the alveolar space creates the interstitial
edema recognized as a complication of sepsis (i.e.,
ALI/ARDS).
Pulmonary hemodynamics
Increased pulmonary artery pressure is common in patients with ARDS, but pulmonary vascular resistance is
usually only mildly to moderately elevated as a consequence of increased cardiac output. The prognosis of
patients with significant elevations in pulmonary vascular resistance is worse, whether related to depressed cardiac function or worsening pulmonary hypertension.
The cause of pulmonary hypertension in ARDS is multifactorial [37]. Vasoconstriction caused by alveolar hypoxia or other vasoactive mediators such as thromboxane and endothelin and intravascular obstruction from
platelet thrombi or perivascular edema probably dominate initially. Later, sustained or worsening pulmonary
hypertension probably reflects the degree to which fibrosis is responsible for obliteration of the vascular
bed. Thus the poor prognosis associated with late pulmonary hypertension in ARDS may simply reflect the
severity of fibrosis.
Pathology and lung repair
The pathological hallmark of ALI/ARDS, diffuse alveolar damage, changes dynamically as ARDS evolves [38,
39]. This occurs gradually over days to weeks, depending
on the severity and resolution of the insult, and may not
resolve for months or may result in chronic fibrotic
changes along the alveolar interstitium. The changes
that develop are conveniently divided into three phases:
the early exudative phase (days 1±5), the fibroproliferative phase (days 6±10), and the fibrotic phase (after
10 days). These times are approximate, and the characteristic features in each phase often overlap. The initial
pathological abnormalities are interstitial swelling, proteinaceous alveolar edema, hemorrhage, and fibrin deposition. Basement membrane disruption and denudation, especially of alveolar epithelial cells, can be seen
with electron microscopy. After 1±2 days hyaline membranes (sloughed alveolar cellular debris mixed with fibrin) are commonly observed by light microscopy. Cellular infiltrates may be minimal or may be dominated
by neutrophils. Fibrin thrombi can be seen in some of
the alveolar capillaries and small pulmonary arteries.
Although type I alveolar epithelial cells cover 95 %
of the alveolar surface, they are terminally differentiated cells that cannot regenerate. Instead, several days after the onset of ARDS type II cells (the cell responsible
for surfactant production) proliferate and then differentiate into new type I cells to reline the alveolar walls.
After approximately 1 week most of the alveolar
edema has resolved, hyaline membranes are much less
prominent, mononuclear cells have replaced the neutrophilic infiltrate, and fibroblasts are proliferating within
the interstitium and depositing new collagen. Pulmonary fibrosis in ARDS is often referred to as ªinterstitialº because structures between airspaces appear to be
markedly widened by fibrotic material. Detailed inspection has revealed, however, that this fibrosis is often the
S 67
result of either alveolar collapse or intra-alveolar fibrosis in which the proteinaceous edema and cellular debris
of the exudative stage have been incorporated into the
alveolar wall. Actual deposition of new collagen within
the interstitial space appears to be relatively uncommon
[40]. Eventually this healing of injured tissue may result
in lung fibrosis, but the extent to which scarring develops is enormously variable. When parenchymal fibrosis
does develop, intimal fibrosis and medial hypertrophy
of pulmonary arterioles, along with complete obliteration of portions of the vascular bed, are also common.
Clinical presentation
Initial signs and symptoms
When clinical manifestations of sepsis first appear, between 28 % and 33 % of patients meet the criteria for
ARDS [41]. Few data exist regarding respiratory abnormalities prior to this time, although progression through
a clinical spectrum of dysfunction is likely the case. Patients may experience severe dyspnea, tachypnea, and
unremitting hypoxemia prior to meeting all the criteria
for ALI/ARDS. Hypoxemia results from myriad causes,
including cardiocirculatory dysfunction affecting global
oxygen delivery and shifts in the oxyhemoglobin dissociation curve. Respiratory dysfunction contributes to hypoxemia as well, with increased work of breathing. The
multifaceted increase in work of breathing is easier to
recognize than to quantify. Changes occur through increased dead space ventilation, related to ventilationperfusion mismatching, respiratory muscle dysfunction,
decreased thoracic compliance and increased airway resistance (bronchoconstriction). Both increased physiological dead-space ventilation and intrapulmonary
shunting are responsible for the tachypnea and elevated
minute ventilation required to achieve effective CO2 excretion in patients with ARDS. Patients may also experience altered mental status or extrapulmonary organ
failure confounding their respiratory dysfunction.
These physiological changes in pulmonary and cardiocirculatory function are most often radiographically
inapparent. In practice, the radiographic findings associated with sepsis vary widely [42]. During the course of
sepsis pulmonary edema may develop as a combination
of increased pulmonary vascular permeability as described earlier, increased hydrostatic pressures related
to resuscitation efforts, and/or lowered oncotic pressure
gradients from any cause. During this time bilateral infiltrates may appear on the chest radiograph without overt
evidence of fluid overload (i.e., increased vascular pedicle width or cardiothoracic ratio). When combined with
appropriate thresholds of hypoxemia, the diagnosis of
ALI or ARDS is secured. Unfortunately, standard chest
radiographs are poor predictors of the severity of oxy-
genation defect or clinical outcome. Although the classic
pulmonary parenchymal changes associated with ALI
are diffuse, bilateral, peripheral, and interstitial in nature, they may be asymmetric or even patchy and focal.
Clinical course
The natural history of ALI/ARDS tends to be dominated by the inciting event rather than the lung injury itself.
As such, treatment of the underlying cause and support
of the respiratory system remains the standard of care.
Death from refractory respiratory failure is unusual,
with the most common cause of death being from the
development of multiple organ failure or (recurrent)
sepsis [5].
In patients who resolve ARDS relatively rapidly
(over a period of 10±14 days), minute ventilation and
dead-space ventilation both decrease in tandem with
improvements in oxygenation. Given the substantial delay to peak incidence of pneumothorax, the lung appears to withstand exposure to somewhat higher forces
in the earliest phase of human ARDS without radiographically evident barotrauma [43, 44]. After this time
further improvements in oxygenation depend on whether the fibroproliferative response can restore the normal
lung architecture for gas exchange. In patients with
more severe ARDS, i.e., those in whom significant lung
fibrosis eventually develops, minute ventilatory requirements stay high even as oxygenation improves. As fibrosis develops, progressive amounts of the vascular bed
are obliterated, which contributes to the increase in
dead-space ventilation even as alveolar edema and the
intrapulmonary shunt resolve.
Prognostication
As discussed above, respiratory dysfunction related to
sepsis exists on a continuum from subclinical aberrations to overt respiratory failure. Quantifying the severity of respiratory system involvement has been of keen
interest for more than a decade. To streamline the ability to conduct research in this area, a clinical definition of
lung injury was proposed and adopted in 1994 (Table 1).
The Consensus Conference definition of ARDS emphasizes the spectrum of abnormalities present from ALI to
ARDS, using readily available clinical criteria to make
the necessary distinction. Although the exact role of respiratory failure in multiple organ failure is not clear, it
has been demonstrated that potentially injurious modes
of mechanical ventilation can produce cytokine release
in humans and end-organ damage in animal models [45].
A number of detailed models have been created in an
attempt to accurately predict the clinical outcomes in
respiratory failure and/or sepsis (Acute Physiology and
S 68
Chronic Health Evaluation, Multiple-Organ Dysfunction Syndrome, Sequential Organ Failure Assessment,
Injury Severity Score). Unfortunately, the prospective
ability to recognize those patients who will survive is
much closer to being an art than a science. Survival in
sepsis appears to be slowly improving over the past 30
or more years, with a coincident decrease in the mortality associated with ALI/ARDS [11]. For most of the first
two decades since ARDS was first reported, mortality
remained relatively constant at 60±70 %. More recent
reports, however, suggest that mortality has declined to
roughly 40 % [11]. The explanation for this apparent improvement in patient outcomes is not clear but could be
due to differences in patient populations, changes in
ventilator support strategies, greater attention to fluid
management, improved hemodynamic and nutritional
support, improved antibiotics for nosocomial infection,
corticosteroid use later in ARDS, or the potential benefits of protocol-driven patient management systems now
implemented in many ICUs.
General scoring systems provide an estimate of the
probability of mortality on admission to the ICU [46].
A specific scoring system for ARDS has been developed; however, its predictive accuracy is debated [47].
The number of acquired organ system failures is often
the most important prognostic indicator for patients requiring intensive care, including patients with ARDS.
Not surprisingly, patients developing fibrosis have a
poorer outcome than do patients in whom fibrosis does
not develop. In addition, liver failure in association
with ARDS carries a particularly poor prognosis.
More specific predictors of outcome for patients with
ARDS have been sought from measurements of various
serum and lung lavage factors. As discussed above, concentrations of proinflammatory cytokines are correlated
with outcome. The concentrations of von Willebrand
factor antigen in serum and neutrophil-activating factor
type 1, IL-8, and procollagen peptide in airspace lavage
fluid are correlated with outcomes or progressive disease in some but not all studies. Increases in unsaturated
serum acyl chain ratios appear to discriminate severity
of illness and may serve as a marker of those most at
risk of developing ALI/ARDS [48].
The integrity of the epithelial barrier in relation to
resolution of alveolar edema also appears to be a determinant of outcome in patients with ARDS [49]. Patients
who can concentrate the protein in the edema fluid during the first 12 h of illness (indicating an intact epithelial
barrier with the ability to actively transport fluid out of
the alveoli) are more likely to recover than those who
cannot do so. Similarly, the change in the PaO22/FIO2
ratio following initial treatment of ARDS can discriminate between survivors and nonsurvivors [50]. At the
present time none of these markers has been validated
as an accurate method for predicting outcome in individual patients with ARDS.
The long-term functional outlook for survivors of
ARDS is generally good [51]. Long-term abnormalities
in pulmonary function are more common if lung function is impaired for more than a few days after the onset
of ARDS. Most of the improvement in pulmonary function and perceived health occurs in the first 3 months
following an episode of ARDS. Recently more complete data concerning long-term outcomes in patients
suffering severe respiratory complications suggest a reduction in the quality of life relative to their premorbid
level of function, often attributed to objective or subjective declines in pulmonary function [52].
Management options
No therapeutic intervention has been proven effective
in reducing the incidence of respiratory failure in sepsis
or its attributable mortality. Prevention of complications is of utmost importance while general supportive
measures (e.g., antimicrobial therapy, nutrition) are undertaken. Control of the upper airway and consideration of the need for ventilatory assistance is an important first step in the management of patients with respiratory dysfunction related to sepsis. Although limited
data exist, which are somewhat conflicting, noninvasive
positive-pressure ventilation has not clearly been shown
to be effective in this clinical circumstance [53, 56]. In
addition, it is critically important to not impede the timing of other appropriate respiratory interventions, such
as institution of mechanical ventilation, regardless of
the availability and/or seeming adequacy of noninvasive
positive-pressure ventilation (NIPPV).
Given our knowledge of fluid flux in states of altered
capillary permeability (i.e., complete equalization of oncotic forces with attendant magnification of effective
hydrostatic forces predicted by Starling's equation), it
seems prudent to advocate judicious fluid resuscitation
and/or fluid restriction when possible in this condition
[57]. Recently investigators have shown that improvements in physiology and outcome occur in patients who
lose weight or whose microvascular pressures fall as a
result of diuresis or fluid restriction [58, 59]. These improvements can be produced by strategies employing
fluid restriction without any higher incidence of complications such as renal failure or hemodynamic compromise [60]. The intravenous solution of choice (i.e., crystalloid versus colloid) is still unclear despite years of detailed investigation. In hypo-oncotic patients with established lung injury, treatment with the combination of albumin and furosemide appears to improve physiology
and may reduce the duration of mechanical ventilation,
although evidence of improved outcomes requires further investigation [61].
If patients cannot adequately protect their airway,
placement of an endotracheal tube is indicated. Based
S 69
on increased rates of sinusitis, orotracheal intubation is
the preferred route [62, 63]. Mounting evidence implicates nosocomial sinusitis in the development of ventilator-associated pneumonia (VAP) ± an entity with a significant independent contribution to mortality [64, 65,
66]. Once endotracheal tube placement has occurred, institution of mechanical ventilation is almost universally
indicated due to coincident respiratory failure (i.e., severe hypoxemia and increased work of breathing). For
this reason one of the chief benefits of mechanical ventilatory support in ARDS is to reduce the patient's work
of breathing so that blood flow may be redirected to other vital organs. Commonly accepted indications for institution of mechanical ventilation from other causes of respiratory failure apply equally to this patient population,
including refractory hypoxemia (PaO2 < 60 despite high
flow oxygen), respiratory rate of more than 35 breaths/
min, and vital capacity below 15 ml/kg, among others.
Mechanical ventilation is not a therapeutic option in
patients with respiratory failure, and thus the goal is to
support the individual respiratory requirements until
the indication(s) requiring mechanical ventilation have
reversed. No mode of ventilation has been proven superior to others in terms of outcomes in patients with sepsis-related respiratory failure, although complete ventilatory support is appropriate immediately after institution of mechanical ventilation. For this reason, volumecycled ventilation using the ªassist-controlº mode (controlled mandatory ventilation) is an appropriate mode
to choose at the outset. Similar degrees of respiratory
support can probably be achieved with intermittent
mandatory ventilation or pressure-regulated volumecontrolled ventilation.
Inhaled oxygen requirements are dictated by the degree of hypoxemia, with arterial oximetric saturations
of approx. 90 % (an approximate pO2 of 60 mmHg) being desirable. To ameliorate the changes in closing volume and lung derecruitment, application of PEEP is appropriate and may provide dramatic improvements in
PaO2. Although some data suggest that levels of PEEP
should be chosen based on respiratory system pressurevolume curves, to select the level of PEEP above the
lower inflection point of the curve and thus prevent cyclic alveolar collapse, this is impractical in current clinical practice [68]. Prone positioning has been shown to
be safe and to result in improvements in oxygenation in
approximately 65 % of patients with ALI/ARDS, although no data exist to predict which patients will respond in this manner [68, 69] Those who do respond
(defined as an improvement in PaO2 > 10 % from baseline) often maintain higher oxygenation levels for up to
18 h after resuming supine positioning and are more
likely to respond to subsequent attempts at prone positioning [70].
Tidal volume should be chosen based on ideal body
weight [men = 50+2.3 (height, in.)±60; women = 45.5
+2.3 (height, in.)±60)], and should be targeted to prevent end-inspiratory plateau pressures from exceeding
30 cmH2O whenever possible. Permissive hypercapnia,
the method of allowing pCO2 to rise while reducing tidal
volume and minute ventilation to prevent alveolar overdistension or propagation of lung injury, has been shown
to be safe and effective at reducing mortality without
adverse consequences (noted increases in QS/QT and
mean pulmonary artery pressure) in small nonrandomized series [71, 72, 73]. In this era of reduced tidal volumes based on the recent National Institutes of Health
sponsored ARDS network trial, permissive hypercapnia
has become accepted as a secondary phenomenon associated with the primary goal of avoiding dangerous airway pressures [73, 74, 75]. Gradual increases in pCO2
are generally well-tolerated, particularly if significant
acidosis does not occur, although the reduction in mean
airway pressure may adversely affect indices of oxygenation [77]. In cases of severe acidosis, intravenous bicarbonate or extracorporeal removal of CO2 may be employed.
No trials have been conducted to specifically define
the most effective method of liberating septic patients
from mechanical ventilation, although it is reasonable
to presume that the literature addressing discontinuation of mechanical ventilation in other patient populations would apply equally. In patients with significant
hemodynamic instability or altered mental status, attempts at discontinuing mechanical ventilation are not
recommended. Thus a two-step method of identifying
patients ready to discontinue mechanical ventilation is
required. A daily screening test (consisting of a brief
evaluation of the resolution of the primary indication
for mechanical ventilation and adequate oxygenation
and ventilation) is the most efficient way to identify
those patients potentially capable of breathing spontaneously [78]. This screening tool is intended to identify
patients in whom the primary indication for mechanical
ventilation has reverted, using data such as frequencyto-tidal volume ratio, oxygenation (PaO2/FIO2 ratio),
maximal inspiratory pressure, maximal expiratory pressure, airway occlusion pressure, and vital capacity [79,
80]. Patients with adequate respiratory recovery according to the screening information should progress to a
simple spontaneous breathing trial to assess the true
ability of an individual patient to be liberated from mechanical ventilation. To that end, daily attempts at spontaneous breathing (through a T-piece connection or
with minimal ventilatory support such as flow-by with
PEEP of 5 cmH2O) should be offered to all hemodynamically stable patients with adequate mental status
who pass the daily respiratory screening instrument.
This simple assessment may be as short as 30 min, although roughly half of patients may fail such a trial after
that time, suggesting 60±120 min as the appropriate duration [81]. In some circumstances pressure support
S 70
may help ªbridgeº patients in the weaning process, by
slow reductions in the applied level of support to determine the ability of the patient to effectively breath spontaneously before a trial with minimal support as described above.
Experimental options
It is a well-known fact that numerous agents have challenged the unyielding morbidity and mortality of ALI/
ARDS associated with sepsis in well-designed clinical
trials, only to be added to the growing list of failed therapies. Use of systemic corticosteroids has been thoroughly tested for both prevention of lung injury as well
as treatment of early phase ALI/ARDS and found to
not be efficacious in either setting with possible increased mortality in patients with established ALI/
ARDS [82, 83, 84]. Uncontrolled or small randomized
trials have suggested benefit to intravenous corticosteroid therapy in patients with prolonged (fibroproliferative phase) ALI/ARDS (Late Steroids Rescue Study.
http://hedwig.mgh.harvard.edu/ardsnet/ards02.html.),
with a large scale trial underway to definitively answer
this important question [85]. Ketoconazole, having
demonstrated potential benefit in the prevention of sepsis-induced lung injury [86] was intensely evaluated in a
multicenter trial supported by the National Institutes
of Health sponsored ARDS network in the United
States and found to lack efficacy in treating established
ARDS [87] Intravenous prostaglandin E1 [88, 89, 90,
91] and aerosolized prostagladin I2 (prostacyclin) [92,
93] have been shown to improve pulmonary physiology
without improving outcome. Also completely tested
has been intravenous lisofylline and strategies to
achieve supranormal oxygen delivery DO2 [94, 95]. Incompletely tested strategies include antioxidants such
as N-acetyl cysteine [98], blocking of tumor necrosis factor, IL-10 therapy [97], and platelet-activating factor antagonists. There are a few experimental options worthy
of discussion based on promising clinical or preclinical
data.
Newer modes of ventilation such as pressure-controlled ventilation or airway pressure release ventilation
[98] may play a role in select patients requiring high levels of ventilatory support, although no data support improved outcomes at this time. In addition, some newer
modes of ventilation are potentially confusing to unfamiliar physicians, thus making it less than desirable to
recommend. High frequency or oscillatory ventilation
has been tested in adults with ARDS and not shown to
be of any significant benefit, although trials are ongoing
to determine whether certain patient subgroups may
benefit from such modes [99, 100, 101, 102, 103]. In addition, liquid ventilation has been shown to result in physiological improvements in small series and animal mod-
els, and phase III trials in humans are underway to evaluate the efficacy of this agent in humans with ALI/
ARDS.
Nitric oxide (either alone, or in combination with the
selective pulmonary vasoconstrictor almitrine) has been
shown to improve oxygenation without any reduction in
duration of mechanical ventilation or mortality [104,
105, 106, 107]. Extracorporeal membrane oxygenation
[108] or extracorporeal removal of CO2 [109, 110, 111]
have been shown to result in significant physiological
improvements in severely ill patients with ARDS, without clear beneficial effects on the development of organ
failure or survival. Liquid ventilation has been shown to
improve pulmonary physiology and reduce inflammation, and is currently in phase III trials to evaluate its
ability to improve outcomes [112, 113, 114]. Similarly,
surfactant therapy has been shown to improve gas exchange without directly affecting days of mechanical
ventilation or mortality [115]. Phase III investigations
are currently in progress in North America and Europe
to determine the efficacy of variations in protocolized
surfactant therapy.
A number of experimental therapies exist on the horizon for patients with sepsis-induced lung injury, including gene therapy [116, 117]. Modulation of cytokines continues to be a difficult but promising area of intervention, with a preliminary randomized trial administering the anti-inflammatory cytokine IL-10 in ALI/
ARDS patients demonstrating a trend towards reduced
organ failure.
Other relevant respiratory issues
Despite the enormous potential of future therapies we
should not ignore the simple and readily available potentially beneficial therapies. This includes the use of inhaled beta-agonists, which have been shown to reduce
inspiratory pressure and increase lung compliance by reducing airway resistance without significant benefits in
dead-space ventilation, oxygenation, or overall outcome
[118]. Elevating the head of the bed at least 30 degrees
at all times has been shown to reduce the incidence of
gastric material migrating to the trachea [119]. Hospital
beds capable of patient rotation and/or distribution of
pressure points to prevent decubitus ulceration [120].
Finally, specific enteral nutritional formulae with antioxidants and amino acid compositions designed to reduce inflammatory lipid mediators have recently been
demonstrated to improve gas exchange and reduce the
duration of mechanical ventilation and intensive care
stay in patients with ARDS [121].
S 71
Postextubation period
Recommendation
After removal of the endotracheal tube patients should
be monitored closely for signs of respiratory compromise for a period of 6±24 h depending on the cause and
severity of respiratory failure. Endotracheal intubation
may cause upper airway injuries that result in immediate or delayed airway compromise. Insertion of an endotracheal tube with an internal stylet may tear the pyriform recesses beside the larynx and result in bleeding
and hematoma formation [122]. Use of a tube that is
too large can result in vocal cord injury, edema, and hematoma and overinflation or malpositioning of the endotracheal tube cuff can cause periglottic injury and
stenosis. Prolonged intubation, coughing, or repeated
endotracheal tube placements can cause the formation
of obstructive arytenoid granulation tissue [123]. Stridor, related to upper airway injury or inflammation, occurs in 25±75 % of pediatric extubations but is rare in
adults, occurring in a small fraction of endotracheally
intubated patients [124]. Although in some circumstances this may be managed expectantly, a low level of tolerance should exist before replacement of an adequate
airway to prevent respiratory compromise. Flexible fiber-optic examination of the larynx before extubation
is often prudent in such patients.
Typically, within a few hours, patients tolerate reintroduction of oral nutrition, although this should progress through stages demonstrating adequate swallowing and airway protective reflexes. For patients with significant respiratory secretions, assistance with ªpulmonary toiletº may be required either through airway suctioning (nasotracheal or orotracheal) or chest percussion with postural drainage. A subset of patients requiring prolonged mechanical ventilation demonstrate significant respiratory muscle weakness, in which case assisted coughing and/or hyperinflation therapy (e.g., intermittent positive pressure breathing) may be of benefit.
Provide adequate supplemental oxygen to maintain an
oximetric saturation of approximately 90 % through
use of simple oxygen delivery systems (i.e., nasal cannula or face mask) if possible. For endotracheally intubated patients, use of PEEP to increase mean airway pressure may be employed to reduce concentrations of
inspired oxygen below potentially toxic thresholds
(FIO2 < 0.60).
Discussion: literature-based recommendations
To answer each of the following important clinical questions, a review of the literature was performed as previously described.
Do manipulations of airway pressure improve (a)
oxygenation or (b) outcome in patients with sepsis?
Answer: (a) yes, grade C; (b) uncertain, grade B.
Rationale
The following specific subject heading keywords were
used to answer this question: oxygen, oxygen inhalation
therapy, anoxia, anoxemia, partial pressure, and pulmonary gas exchange. The broad spectrum of respiratory
dysfunction encountered in sepsis only allows for answers in specific clinical circumstances. For instance, a
significant proportion of the 60 % of septic patients
who never develop ALI/ARDS have normal chest radiographs, and this patient population has not been intensely studied. Although the relationship between radiographic manifestations of pulmonary dysfunction
and gas exchange is poor, it is reasonable to presume
that hypoxemic patients with sepsis but without radiographic pulmonary infiltrates would respond similarly
to patients with visible interstitial edema (i.e., patients
with ALI/ARDS). It is clear that hypoxemia is modestly
correlated with prognosis in ALI/ARDS related to sepsis, and that simple methods of oxygen supplementation
raise PaO2. Raising mean airway pressure results in recruitment of additional lung units to participate in gas
exchange while maintaining the patency of units once
recruited and thus increases PaO2 [125, 126, 127]. It is
equally clear that application of PEEP, either with an
endotracheal tube in place or through a tight-fitting
face mask, has been shown to improve oxygenation in
hypoxemic patients through increases in airway pressure [128, 129, 130]. The underlying goal in providing
such therapy is to ensure adequate oxygen delivery to
critical tissue beds in states of altered microvascular
flow. Unfortunately, there is no data on which to assess
outcomes through manipulations of airway pressure.
Studies of mechanically ventilated patients with ALI/
ARDS treated with different methods of ventilation designed to achieve different inspiratory pressures have
shown differences in outcome, but attributing these improvements to the manipulation of airway pressure directly is impossible [67, 74]. Determinations made by
the most routine measure of oxygenation, pulse oximetry, are correlated well with arterial oxygen saturation
but may misrepresent arterial saturation by 7 % in patients with extremes of heart rate, cardiac index, or pulmonary arterial wedge pressure [131]. Despite this the
S 72
use of this device is recommended to monitor arterial
oxygenation in this patient population, with supplemental oxygen and PEEP administered to maintain saturation of approximately 88±90 % (approximating a PaO2
of 60 mmHg) with nontoxic concentrations of oxygen
(ideally FIO2 2 < 0.60).
Can noninvasive positive-pressure ventilation be safely
and effectively used in ALI/ARDS related to sepsis?
Answer: no, grade B.
Recommendation
Avoid the use of NIPPV in sepsis-related ALI/ARDS
patients.
Rationale
The following specific subject heading keywords were
used to answer this question: positive pressure respiration, artificial respiration, intermittent positive pressure
ventilation, adult respiratory distress syndrome. There
has been a surge of interest in applying noninvasive positive-pressure ventilation to all patients with respiratory
failure, although it appears that patients with ALI/
ARDS are more likely to fail this therapy [53]. It is clear
that NIPPV is most effective in selected patients (normal or near normal mental status without significant respiratory system secretions) with expected resolution of
respiratory failure within 72 h ± a rare situation in ALI/
ARDS [55]. Although NIPPV may avoid the use of mechanical ventilation (and its attendant risks) in a small
population of ALI/ARDS patients [54], the delay in institution of mechanical ventilation may be equally likely
to result in untoward complications in the majority of
patients.
Does (a) placement of an endotracheal tube or (b)
institution of mechanical ventilation improve outcome
in respiratory failure related to sepsis?
Answer: (a) no, grade E; (b) yes, grade E.
Recommendation
Early placement of an endotracheal tube and institution
of mechanical ventilation in patients with sepsis is appropriate based upon standard clinical criteria heralding
the onset of respiratory failure to avoid the recognized
complications associated with respiratory failure and/or
acute respiratory arrest. Indications for institution of
mechanical ventilation include severe tachypnea (respiratory rate > 40 breaths/min), muscular respiratory failure (use of accessory muscles), altered mental status,
and/or severe hypoxemia despite supplemental oxygen.
Rationale
The following specific subject heading keywords were
used to answer this question: intubation, intratracheal
intubation, respiratory insufficiency, artificial respiration, positive-pressure respiration, and adult respiratory
distress syndrome. For ethical reasons there are no randomized trials evaluating the use of endotracheal intubation in critically ill patients. It is important to recognize that placement of an endotracheal tube is not a
therapeutic maneuver. This step carries the attendant
risks of anesthesia for the procedure and subsequent
morbid events such as VAP and thus by itself does not
improve outcome in this clinical circumstance. In addition, mechanical ventilation (independently of airway
protection, etc.) has not been shown to improve outcome in patients with sepsis and respiratory failure, although this has not been studied in depth. In comparison with historical controls (i.e., the polio epidemic),
mechanical ventilation does indeed provide significant
tangible clinical benefits [132]. Alternatively, discontinuation of mechanical ventilation by removing an endotracheal tube in terminally ill patients results in more
rapid expiration than simply withholding therapy, thus
providing indirect evidence of clinical benefit from endotracheal intubation with mechanical ventilation [133].
The greatest morbidity associated with endotracheal
tube placement relates to risk of VAP, which is increased in patients with burns, trauma, central nervous
system disease, respiratory disease, cardiac disease, and
witnessed aspiration [134]. Potentially VAP may be decreased by use of orotracheal intubation, subglottic secretion drainage, kinetic hospital beds, and increased
by heated respiratory circuit humidifiers and histamine-2 receptor antagonists [135, 136]. It is well recognized that mechanical ventilation possesses the potential to initiate or propagate lung injury, and thus can be
considered an independent source of patient morbidity.
Is normalization of (a) pH or (b) pCO2 necessary in
ALI/ARDS?
Answer: (a) no, grade D; (b) no, grade D.
S 73
Recommendation
Implement permissive hypercapnia through reduced
tidal volume ventilation in mechanically ventilated
ALI/ARDS patients with high inspiratory pressures or
otherwise at risk for barotrauma/volutrauma.
Rationale
The following specific subject keyword headings were
used to answer this question: hypercapnia, artificial respiration, positive pressure respiration, adult respiratory
distress syndrome. Permissive hypercapnia, the method
of allowing pCO2 to rise while reducing tidal volume
and minute ventilation to prevent alveolar overdistension or perpetuation of lung injury has been shown to
be safe and effective at reducing mortality without adverse consequences in small nonrandomized series [75,
76, 77]. The upper limit for pCO2 has not been established, although arterial pH should be maintained at a
level higher than 7.20. Based on these data, normalization of arterial blood gas values is not considered a valuable therapeutic maneuver.
Does the use of (a) small tidal volume ventilation or (b)
pressure limited ventilation strategies affect outcome in
ALI related to sepsis?
Answer: (a) yes, grade A; (b) uncertain, grade A.
Recommendation
Mechanical ventilation of patients with ALI should be
conducted with small tidal volumes (approximately
6 ml/kg ideal body weight) with the goal to maintain
end-inspiratory plateau pressures at levels less than
30 cmH2O.
Rationale
The following specific subject heading keywords were
used to answer this question: tidal volume, lung compliance, positive pressure respiration, intrinsic positive
pressure respiration, mechanical ventilator, adult respiratory distress syndrome. Well-designed large scale randomized trials designed to alter inspiratory pressure
through variations in tidal volume have been conducted,
with varying results [137, 138, 139, 140]. It is not completely understood why the results of these well-designed trials conflict, although the intergroup differential in airway pressure is a likely contributor. In a recent
trial in the United States, absolute all-cause mortality
was reduced by 10 % in ALI patients receiving mechanical ventilation with tidal volumes of 6 ml/kg ideal body
weight [74]. This topic has also been recently reviewed
by an international expert consensus conference [141].
Does prone positioning affect (a) gas exchange or (b)
outcome in sepsis-related ALI?
Answer: (a) yes, grade C; (b) uncertain, grade C.
Recommendation
Prone positioning may be considered in patients requiring high levels of inspired oxygen (FIO2 > 0.60) in
whom positional changes are not contraindicated, and
who are cared for at facilities experienced in the management of critically ill mechanically ventilated patients.
Rationale
The following specific subject heading keywords were
used to answer this question: prone position and supine
position. Although the evidence for prone positioning
must be graded C because of the small size of randomized trials, strong data exists to confirm the physiological benefits of this intervention. Recent studies have
made it clear that prone positioning of patients with
ALI/ARDS results in improvements in oxygenation in
approximately 65 % of patients (ªrespondersº) [68, 69,
70]. In addition, the improvements in gas exchange
may persist up to 18 h, even after returning to the supine
position, and such changes in position may be accomplished safely in intensive care units accustomed to
managing critically ill mechanically ventilated patients.
Because of the limited size of trials to date, no definitive
comments can be made on the general applicability of
these maneuvers to all patient care centers or their effect on overall mortality.
Does inhaled nitric oxide affect (a) oxygenation or (b)
outcome in ALI/ARDS?
Answer: (a) yes, grade A; (b) no, grade A.
Recommendation
Restrict nitric oxide as an option for salvage therapy in
patients with life-threatening hypoxemia not responding to traditional mechanical ventilation strategies or
for evaluation in controlled clinical trials.
S 74
Rationale
The following specific subject heading keywords were
used to answer this question: oxygen, nitric oxide, inhalation administration, and pulmonary gas exchange. Inhaled nitric oxide has been studied extensively in both
preclinical models of lung injury and clinical trials of patients with ALI/ARDS. It has been shown to lower pulmonary artery pressures and improve right ventricular
function in patients with pulmonary hypertension. Inhaled nitric oxide improves oxygenation and may reduce edema formation in patients with ALI/ARDS
through effects on hydrostatic pressure. Unfortunately,
it has not been found to significantly affect mortality
[104, 105, 106]. These data support the observation that
inhaled nitric oxide consistently improves pulmonary
physiology in a large proportion of these patients, but
fails to affect outcome.
Is there a defined fluid management strategy in sepsisrelated ALI/ARDS?
Answer: uncertain, grade C.
Recommendation
Judicious use of crystalloid fluid administration should
be practiced in patients with ALI/ARDS, with colloid
solutions considered in hypo-oncotic patients with established ALI/ARDS. It is not clear if volume restriction improves outcome.
Rationale
The following specific subject heading keywords were
used to answer this question: fluid therapy, resuscitation, diuresis, intravenous infusions, hypertonic saline
solution, sodium chloride, colloids, plasma substitutes,
hetastarch, dextrans. Optimal fluid management has
been considered a critical question in patients with sepsis and ALI/ARDS since these syndromes were first described. Conflicting data exist regarding the relative
benefits of crystalloid and colloid administration in
these patient populations despite years of research.
Use of colloids in this patient population has been advocated and debated for decades, with evidence for potential benefit appearing only recently. Similarly, only recently has fluid balance been evaluated independently
with respect to its contribution to overall morbidity and
mortality [59]. Prospective, randomized trials have
been conducted which support improved clinical outcomes based on direct manipulation of fluid balance
variables in this critically ill patient population [60].
The details of the most appropriate intravenous solution
and volume of administration requires large-scale investigation.
Are corticosteroids indicated in the (a) prevention, (b)
early treatment (exudative phase), or (c) late treatment
(fibroproliferative phase) of ARDS?
Answer: (a) no, grade A; (b) no, grade A; (c) uncertain,
grade C.
Recommendation
Do not routinely administer corticosteroids to patients
at risk for, or meeting current criteria for, ALI/ARDS.
Consider intravenous methylprednisolone in patients
with persistent or refractory ARDS after actively excluding infection, pending the results of ongoing trials.
Rationale
The following specific subject heading keywords were
used to answer this question: steroids, adrenal cortex
hormones, prednisone, methylprednisolone, hydrocortisone, dexamethasone. Corticosteroids have long been
considered part of an appropriate treatment plan for patients with lung injury. There have been well-designed
trials that fail to demonstrate any significant benefit for
corticosteroids in the prevention or early treatment of
ARDS [82, 83, 84]. A recent resurgence of interest has
been generated by small trials suggesting benefit in the
subpopulation of patients failing to progress in the late
phase of ARDS (Late Steroids Rescue Study, http://
hedwig.mgh.harvard.edu/ardsnet/ards02.html). In an attempt to answer this pressing question, the National Institutes of Health have sponsored a large-scale trial in
the United States randomizing patients with ARDS for
more than 7 days to methylprednisolone therapy [85].
In conducting all of these trials, close attention was
paid to excluding infection before or during corticosteroid therapy. Until definitive trials have been completed,
a clear recommendation cannot be made regarding corticosteroid administration in patients with persistent
ALI/ARDS.
Do daily spontaneous breathing trials or weaning
protocols reduce the duration of mechanical
ventilation?
Answer: yes, grade A.
S 75
Recommendation
It is recommended that all patients requiring acceptable
levels of ventilatory support who are not overtly unstable should receive a spontaneous breathing trial on a
daily basis to determine ability to breathe unassisted.
Rationale
The following specific subject heading keywords were
used to answer this question: weaning, ventilator weaning, artificial respiration, and mechanical ventilator.
The last 10 years has seen a surge in interest in determining the optimum method of discontinuing mechanical ventilation [142, 143, 144]. Recent large-scale trials
have been conducted to demonstrate the benefits of daily trials of spontaneous breathing in reducing the duration of mechanical ventilation [78, 145, 146].
Identifying patients capable of breathing spontaneously requires a two-step process: a brief screen and a
trial of spontaneous breathing. The screening procedure
is designed to exclude patients requiring excessive levels
of mechanical ventilatory support. Thus the following
criteria may be employed to identify patients ready to
accept a trial of spontaneous breathing: FIO2 < 0.50,
PEEP 5 cmH2O, intact airway reflexes, hemodynamic
stability and adequate mental status. The definition of
spontaneous breathing trial is still a subject of debate,
with both ªT-pieceº breathing and flow-triggered ventilation with continuous positive airway pressure of
5 cmH2O currently being acceptable methods of achieving ªspontaneous breathing.º Patients who tolerate such
a breathing trial for 2 h have an approximately 85 %
success rate with complete discontinuation of mechanical ventilation.
Conclusion
Despite significant advances in both the knowledge of
sepsis-related respiratory failure and the care of critically ill patients, ALI/ARDS continues to be a complex
problem with high mortality. The recommendations
above represent the current state of knowledge for this
condition, but equally serve to highlight the vast deficiencies of knowledge that remain. To provide our patients with the best possible outcome, a continued focus
on physiological, therapeutic, and outcomes research is
necessary.
Acknowledgements This research was supported in part by the
National Institutes of Health (NHLBI HL 07123).
References
1. Hazinski MF, Iberti TJ, MacIntyre
NR, et al (1993) Epidemiology, pathophysiology and clinical presentation
of gram-negative sepsis. Am J Crit
Care 2: 224±235
2. Anonymous (1990) Current trends increase in national hospital discharge
survey rates for septicemia ± United
States, 1979±1987. MMWR Morb
Mort Wkly Rep 39: 31±34
3. Fowler AA, Hamman RF, Good JT, et
al (1983) Adult respiratory distress
syndrome: risk with common predispositions. Ann Intern Med 98: 593±597
4. Pepe PE, Potkin RT, Reus DH, Hudson LD, Carrico CJ (1982) Clinical
predictors of the adult respiratory distress syndrome. Am J Surg 144:
124±130
5. Montgomery AB, Stager MA, Carrico
CJ, Hudson LD (1985) Causes of mortality in patients with the adult respiratory distress syndrome. Am Rev Respir Dis 132: 485±489
6. Bernard GR, Artigas A, Brigham KL,
et al (1994) The American-European
consensus conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination.
Am J Respir Crit Care Med 149:
818±824
7. Kollef MH, Schuster DP (1995) The
acute respiratory distress syndrome.
N Engl J Med 332: 27±37
8. National Heart and Lung Institute,
National Institutes of Health (1972)
Respiratory distress syndromes: Task
Force on Problems, Research Approaches, Needs. US Government
Printing Office, Washington. DHEW
Publication No (NIH) 73±432, pp
165±180
9. Thomsen GE, Morris AH (1995) Incidence of the adult respiratory distress
syndrome in the State of Utah. Am J
Respir Crit Care Med 152: 965±971
10. Milberg JA, Davis DR, Steinberg KP,
Hudson LD (1995) Improved survival
of patients with acute respiratory distress syndrome (ARDS): 1983±1993.
JAMA 273: 306±309
11. Doyle RL, Szaflarski N, Modin GW,
Wiener-Kronish JP, Matthay MA
(1995) Identification of patients with
acute lung injury. Predictors of mortality. Am J Respir Crit Care Med 152:
1818±1824
12. Vasilyev S, Schaap RN, Mortensen JD
(1995) Hospital survival rates of patients with acute respiratory failure in
modern respiratory intensive care
units: an international, multi-center,
prospective survey. Chest 107:
1083±1088
13. Artigas A, Bernard GR, Carlet J, et al
(1998) The American-European Consensus Conference on ARDS. II. Ventilatory, pharmacologic, supportive
therapy, study design strategies, and issues related to recovery and remodeling. Am J Respir Crit Care Med 157:
1332±1347
14. Demling RH (1995) The modern version of the adult respiratory distress
syndrome. Annu Rev Med 46: 193±202
15. Steinberg KP, Milberg JA, Martin TR,
et al (1988) Evolution of bronchoalveolar cell populations in the adult respiratory distress syndrome. Am Rev Respir Dis 138: 720±723
S 76
16. Rinaldo JE, Christman JW (1990)
Mechanisms and mediators of the
adult respiratory distress syndrome.
Clin Chest Med 11: 621±632
17. Marks JD, Marks CB, Luce JM, et al
(1990) Plasma tumor necrosis factor
in patients with septic shock: Mortality
rate & incidence of adult respiratory
distress syndrome. Am Rev Respir
Dis 141: 94±97
18. Meduri GU, Headley S, Tolley E,
Shelby M, Stentz F, Postlethwaite A
(1995) Plasma and BAL cytokine response to corticosteroid rescue treatment in late ARDS. Chest 108
1315±1325
19. Calandra T, Baumgartner JD, Grau
GE, et al (1990) Prognostic values of
tumor necrosis factor/cachectin, interleukin-1, interferon alpha, and interferon gamma in the serum of patients
with septic shock. J Infect Dis 161:
982±987
20. Roumen RMH, Hendriks T, van der
Ven-Jongekrijg J, et al (1993) Cytokine patterns in patients after major
vascular surgery, hemorrhagic shock,
and severe blunt trauma: relation with
subsequent adult respiratory distress
syndrome and multiple organ failure.
Ann Surg 218: 769±776
21. Quinlan GJ, Lamb NJ, Tilley R, Evans
TW, Gutteridge JM (1997) Plasma hypoxanthine levels in ARDS: implications for oxidative stress, morbidity,
and mortality. Am J Respir Crit Care
Med 155: 479±484
22. Quinlan GJ, Lamb NJ, Evans TW,
Gutteridge JM (1996) Plasma fatty
acid changes and increased lipid peroxidation in patients with adult respiratory distress syndrome. Crit Care
Med 24: 241±246
23. Carpenter CT, Price PV, Christman
BW (1998) Exhaled breath condensate
isoprostanes are elevated in patients
with acute lung injury or ARDS. Chest
114: 1653±1659
24. Christman BW, Bernard GR (1993)
Antilipid mediator and antioxidant
therapy in adult respiratory distress
syndrome. New Horiz 1: 623±630
25. Quinlan GJ, Margarson MP, Mumby
S, Evans TW, Gutteridge JM (1998)
Administration of albumin to patients
with sepsis syndrome: a possible beneficial role in plasma thiol repletion.
Clin Sci (Colch) 95: 4:459±465
26. Dong W, Simeonova PP, Gallucci R, et
al (1998) Toxic metals stimulate inflammatory cytokines in hepatocytes
through oxidative stress mechanisms.
Toxicol Appl Pharmacol 151: 359±366
27. Bulger EM, Helton WS, Clinton CM,
Roque RP, Garcia I, Maier RV (1997)
Enteral vitamin E supplementation inhibits the cytokine response to endotoxin. Arch Surg 132: 1337±1341
28. Tsan M, Danis E, Del Vechhio PJ, Rosano CL (1985) Enhancement of intracellular glutathione protects endothelial cells against oxidant damage. Biochem Biophys Res Commun 127:
270±276
29. Barcellos-Hoff MA, Dix TA (1996)
Redox-mediated activation of latent
transforming growth factor-beta 1.
Mol Endocrinol 10: 1077±1083
30. Neuschwander-Tetri BA, Bellezzo JM,
Britton RS, Bacon BR, Fox ES (1996)
Thiol regulation of endotoxin-induced
release of tumor necrosis factor alpha
from isolated rat Kupffer cells. Biochem J 320: 1005±1010
31. Lin M, Rippe RA, Niemela O, Brittenham G, Tsukamoto H (1997) Role
of iron in NF-kappa B activation and
cytokine gene expression by rat hepatic macrophages. Am J Physiol
272:G1355±G1364
32. Loban A, Kime R, Powers H (1997)
Iron-binding antioxidant potential of
plasma albumin. Clin Sci (Colch) 93:
445±451
33. Gattinoni L, Pesenti A, Bombino M,
et al (1988) Relationships between
lung computed tomographic density,
gas exchange, and PEEP in acute respiratory failure. Anesthesiology 69:
824±832
34. Slutsky AS (1993) Mechanical ventilation: American College of Chest Physicians' Consensus Conference. Chest
104: 1833±1859
35. Tsuno K, Prato P, Kolobow T (1990)
Acute lung injury from mechanical
ventilation at moderately high airway
pressures. J Appl Physiol 69: 956±961
36. Wright PE, Bernard GR (1989) The
role of airflow resistance in patients
with the adult respiratory distress syndrome. Am Rev Respir Dis 139:
1169±1174
37. Jones R, Reid LM, Zapol WM, et al
(1992) Pulmonary vascular pathology:
human and experimental studies. In:
Falke KJ (ed) Lung biology in health
and disease. Philadelphia, Saunders
38. Meyrick B (1986) Pathology of the
adult respiratory distress syndrome.
Crit Care Clin 2: 405±428
39. Tomashefski JF Jr (1990) Pulmonary
pathology of the adult respiratory distress syndrome. Clin Chest Med 11:
593±620
40. McDonald JA (1991) Idiopathic pulmonary fibrosis: A paradigm for lung
injury and repair. Chest 99 [Suppl]:87±93
41. Bernard GR, Wheeler AP, Russell JA,
et al (1997) The effects of ibuprofen on
the physiology and survival of patients
with sepsis. The Ibuprofen in Sepsis
Study Group. N Engl J Med 336:
912±918
42. Wheeler AP, Carroll FE, Bernard GR
(1993) Radiographic issues in the adult
respiratory distress syndrome. New
Horiz 1: 471±477
43. Gammon RB, Shin MS, Buchalter SE
(1992) Pulmonary barotrauma in mechanical ventilation: patterns and risk
factors. Chest 102: 568±572
44. Gammon RB, Shin MS, Groves RH,
Hardin M, Hsu C, Buchalter SE
(1995) Clinical risk factors for pulmonary barotrauma: a multivariate analysis. Am J Respir Crit Care Med 152:
1235±1240
45. Ranieri VM, Suter PM, Tortorella C,
et al (1999) Effect of mechanical ventilation on inflammatory mediators in
patients with acute respiratory distress
syndrome: a randomized controlled
trial. JAMA 282: 54±61
46. Schuster DP (1992) Predicting outcome after ICU admission. The art
and science of assessing risk. Chest
102: 1861±1870
47. Murray JF, Matthay MA, Luce JM, et
al (1988) An expanded definition of
the adult respiratory distress syndrome. Am Rev Respir Dis 138:
720±723
48. Bursten SL, Federighi DA, Parsons P,
et al (1996) An increase in serum C18
unsaturated free fatty acids as a predictor of the development of acute respiratory distress syndrome. Crit Care
Med 24: 1129±1136
49. Matthay MA, Wiener-Kronish JP
(1990) Intact epithelial barrier function is critical for the resolution of alveolar edema in humans. Am Rev Respir Dis 142: 1250±1257
50. Sloane PJ, Gee MH, Gottlieb JE, et al
(1992) A multicenter registry of patients with acute respiratory distress
syndrome: physiology and outcome.
Am Rev Respir Dis 146: 419±426
51. McHugh LG, Milberg JA, Whitcomb
ME, et al (1994) Recovery of function
in survivors of the acute respiratory
distress syndrome. Am J Respir Crit
Care Med 150: 90±94
S 77
52. Davidson TA, Caldwell ES, Curtis JR,
Hudson LD, Steinberg KP (1999) Reduced quality of life in survivors of
acute respiratory distress syndrome
compared with critically ill control patients. JAMA 281: 354±360
53. Alsous F, Amoateng-Adjepong Y,
Manthous CA (1999) Noninvasive
ventilation: experience at a community teaching hospital. Intensive Care
Med 25: 458±463
54. Antonelli M, Conti G, Rocco M, et al
(1998) A comparison of noninvasive
positive-pressure ventilation and conventional mechanical ventilation in
patients with acute respiratory failure.
N Engl J Med 339: 429±435
55. Meduri GU, Turner RE, Abou-Shala
N, Wunderink R, Tolley E (1996) Noninvasive positive pressure ventilation
via face mask: first-line intervention
in patients with acute hypercapnic and
hypoxemic respiratory failure. Chest
109: 179±193
56. Patrick W, Webster K, Ludwig L, Roberts D, Wiebe P, Younes M (1996)
Noninvasive positive-pressure ventilation in acute respiratory distress without prior chronic respiratory failure.
Am J Respir Crit Care Med 153:
1005±1011
57. Schuster DP (1993) The case for and
against fluid restriction and occlusion
pressure reduction in adult respiratory
distress syndrome. New Horiz 1:
478±488
58. Simmons RS, Berdine GG, Seidenfeld
JJ, et al (1987) Fluid balance and the
adult respiratory distress syndrome.
Am Rev Respir Dis 135: 924±929
59. Humphrey H, Hall J, Sznajder I, Silverstein M, Wood L (1990) Improved
survival in ARDS patients associated
with a reduction in pulmonary capillary wedge pressure. Chest 97:
1176±1180
60. Mitchell JP, Schuller D, Calandrino
FS, Schuster DP (1992) Improved outcome based on fluid management in
critically ill patients requiring pulmonary artery catheterization. Am Rev
Respir Dis 145: 990±998
61. Martin GS, Mangialardi RJ, Wheeler
AP, Bernard GR (1999) Albumin and
diuretics in ARDS. Am J Respir Crit
Care Med 159:A376
62. Salord F, Gaussorgues P, Marti-Flich J,
et al (1990) Nosocomial maxillary sinusitis during mechanical ventilation:
a prospective comparison of orotracheal versus the nasotracheal route
for intubation. Intensive Care Med 16:
390±393
63. Holzapfel L, Chevret S, Madinier G, et
al (1993) Influence of long-term oroor nasotracheal intubation on nosocomial maxillary sinusitis and pneumonia: results of a prospective, randomized, clinical trial. Crit Care Med 21:
1132±1138
64. Holzapfel L, Chastang C, Demingeon
G, Bohe J, Piralla B, Coupry A (1999)
A randomized study assessing the systematic search for maxillary sinusitis
in nasotracheally mechanically ventilated patients. Influence of nosocomial
maxillary sinusitis on the occurrence
of ventilator-associated pneumonia.
Am J Respir Crit Care Med 159:
695±701
65. Cook D, De Jonghe B, Brochard L,
Brun-Buisson C (1998) Influence of
airway management on ventilator-associated pneumonia: evidence from
randomized trials. JAMA 279:
781±787
66. Heyland DK, Cook DJ, Griffith L,
Keenan SP, Brun-Buisson C (1999)
The attributable morbidity and mortality of ventilator-associated pneumonia in the critically ill patient. The Canadian Critical Trials Group. Am J
Respir Crit Care Med 159: 1249±1256
67. Amato MB, Barbas CS, Medeiros
DM, et al (1995) Beneficial effects of
the ªopen lung approachº with low
distending pressures in acute respiratory distress syndrome. A prospective
randomized study on mechanical ventilation. Am J Respir Crit Care Med
152: 1835±1846
68. Stocker R, Neff T, Stein S, Ecknauer
E, Trentz O, Russi E (1997) Prone positioning and low-volume pressurelimited ventilation improve survival in
patients with severe ARDS. Chest
111: 1008±1017
69. Lamm WJ, Graham MM, Albert RK
(1994) Mechanism by which prone position improves oxygenation in acute
lung injury. Am J Respir Crit Care
Med 150: 184±193
70. Jolliet P, Bulpa P, Chevrolet JC (1998)
Effects of the prone position on gas exchange and hemodynamics in severe
acute respiratory distress syndrome.
Crit Care Med 26: 1977±1985
71. Hickling KG, Walsh J, Henderson S,
Jackson R (1994) Low mortality rate
in adult respiratory distress syndrome
using low-volume, pressure-limited
ventilation with permissive hypercapnia: a prospective study. Crit Care
Med 22: 1568±1578
72. Bidani A, Tzouanakis AE, Cardenas
VJ, Zwischenberger JB (1994) Permissive hypercapnia in acute respiratory
failure. JAMA 272: 957±962
73. Thorens JB, Jolliet P, Ritz M, Chevrolet JC (1996) Effects of rapid permissive hypercapnia on hemodynamics,
gas exchange, and oxygen transport
and consumption during mechanical
ventilation for the acute respiratory
distress syndrome. Intensive Care
Med 22: 182±191
74. Anonymous (2000) Ventilation with
lower tidal volumes as compared with
traditional tidal volumes for acute
lung injury and the acute respiratory
distress syndrome. The Acute Respiratory Distress Syndrome Network.
N Engl J Med 342: 1301±1308
75. Hickling KG, Walsh J, Henderson S,
Jackson R (1994) Low mortality rate
in adult respiratory distress syndrome
using low-volume, pressure-limited
ventilation with permissive hypercapnia: a prospective study. Crit Care
Med 22: 1568±1578
76. Bidani A, Tzouanakis AE, Cardenas
VJ, Zwischenberger JB (1994) Permissive hypercapnia in acute respiratory
failure. JAMA 272: 957±962
77. Ely EW, Baker AM, Dunagan DP, et al
(1996) Effect on the duration of mechanical ventilation of identifying patients capable of breathing spontaneously. N Engl J Med 335: 1864±1869
78. Vallverdu I, Calaf N, Subirana M, Net
A, Benito S, Mancebo J (1998) Clinical characteristics, respiratory functional parameters, and outcome of a
two-hour T-piece trial in patients
weaning from mechanical ventilation.
Am J Respir Crit Care Med 158:
1855±1862
79. Chatila W, Jacob B, Guaglionone D,
Manthous CA (1996) The unassisted
respiratory rate-tidal volume ratio accurately predicts weaning outcome.
Am J Med 101: 61±67
80. Esteban A, Alia I, Tobin MJ, et al
(1999) Effect of spontaneous breathing trial duration on outcome of attempts to discontinue mechanical ventilation. Am J Respir Crit Care Med
159: 512±518
81. Bone RC, Fisher CJ Jr, Clemmer TP,
Slotman GJ, Metz CA, Balk RA
(1987) A controlled clinical trial of
high-dose methylprednisolone in the
treatment of severe sepsis and septic
shock. N Engl J Med 317: 653±658
82. Bone RC, Fisher CJ Jr, Clemmer TP,
Slotman GJ, Metz CA (1987) Early
methylprednisolone treatment for septic syndrome and the adult respiratory
distress
syndrome.
Chest
92:
1032±1036
S 78
83. Luce JM, Montgomery AB, Marks JD,
Turner J, Metz CA, Murray JF (1988)
Ineffectiveness of high-dose methylprednisolone in preventing parenchymal lung injury and improving mortality in patients with septic shock. Am
Rev Respir Dis 138: 62±68
84. Meduri GU, Headley AS, Golden E,
et al (1998) Effect of prolonged methylprednisolone therapy in unresolving acute respiratory distress syndrome: a randomized controlled trial.
JAMA 280: 159±165
85. Yu M, Tomasa G (1993) A doubleblind, prospective, randomized trial of
ketoconazole, a thromboxane synthetase inhibitor, in the prophylaxis of
the adult respiratory distress syndrome. Crit Care Med 21: 1635±1642
86. Anonymous (2000) Ketoconazole for
early treatment of acute lung injury
and acute respiratory distress syndrome: a randomized controlled trial.
The ARDS Network. JAMA 283:
1995±2002
87. Abraham E, Park YC, Covington P, et
al (1996) Liposomal prostaglandin E1
in acute respiratory distress syndrome:
a placebo-controlled, randomized,
double-blind, multicenter clinical trial.
Crit Care Med 24: 10±15
88. Bone RC, Slotman G, Maunder R, et
al (1989) Randomized double-blind,
multicenter study of prostaglandin E1
in patients with the adult respiratory
distress syndrome. Prostaglandin E1
Study Group. Chest 96: 114±119
89. Holcroft JW, Vassar MJ, Weber CJ
(1986) Prostaglandin E1 and survival
in patients with the adult respiratory
distress syndrome: a prospective trial.
Ann Surg 203: 371±378
90. Meyer J, Theilmeier G, Van Aken H,
et al (1998) Inhaled prostaglandin E1
for treatment of acute lung injury in
severe multiple organ failure. Anesth
Analg 86: 4:753±758
91. Walmrath D, Schneider T, Pilch J, et al
(1993) Aerosolized prostacyclin in
adult respiratory distress syndrome.
Lancet 342: 961±962
92. Zwissler B, Kemming G, Habler O, et
al (1996) Inhaled prostacyclin (PGI2)
versus inhaled nitric oxide in adult respiratory distress syndrome. Am J Respir Crit Care Med 154: 1671±1677
93. Hayes MA, Timmins AC, Yau EHS, et
al (1994) Elevation of systemic oxygen
delivery in the treatment of critically
ill patients. N Engl J Med 330:
1717±1722
94. Gattinoni L, Brazzi L, Pelosi P, et al
(1995) A trial of goal-oriented hemodynamic therapy in critically ill patients. N Engl J Med 333: 1025±1032
95. Bernard GR, Wheeler AP, Arons MM,
et al (1997) A trial of antioxidants
N-acetylcysteine and procysteine in
ARDS. The Antioxidant in ARDS
Study Group. Chest 112: 164±172
96. Bernard GR, Wheeler AP, Naum CC,
et al (1999) A placebo controlled randomized trial of IL-10 in acute lung injury. Chest 116: 260 S
97. Putensen C, Mutz NJ, Putensen-Himmer G, Zinserling J (1999) Spontaneous breathing during ventilatory
support improves ventilation-perfusion distributions in patients with
acute respiratory distress syndrome.
Am J Respir Crit Care Med 159:
1241±1248
98. Carlon GC, Howland WS, Ray C, et al
(1983) High-frequency jet ventilation:
a prospective randomized evaluation.
Chest 84: 551±559
99. Holzapfel L, Robert D, Perrin F, et al
(1987) Comparison of high-frequency
jet ventilation to conventional ventilation in adults with respiratory distress
syndrome. Intensive Care Med 13:
100±105
100. Macintyre NR, Follett JV, Deitz JL, et
al (1986) Jet ventilation at 100 breaths
per minute in adult respiratory failure.
Am Rev Respir Dis 134: 897±901
101. Borg UR, Stoklosa JC, Siegel JH, et al
(1989) Prospective evaluation of combined high-frequency ventilation in
post-traumatic patients with adult respiratory distress syndrome refractory
to optimized conventional ventilatory
management. Crit Care Med 17:
1129±1142
102. Hurst JM, Branson RD, Davis K, et al
(1990) Comparison of conventional
mechanical ventilation and high-frequency ventilation: a prospective randomized trial in patients with respiratory failure. Ann Surg 211: 486±491
103. Rossaint R, Gerlach H, SchmidtRuhnke H, et al (1995) Efficacy of inhaled nitric oxide in patients with severe ARDS. Chest 107: 1107±1115
104. Troncy E, Collet J-P, Shapiro S, et al
(1997) Should we treat acute respiratory distress syndrome with inhaled nitric oxide? Lancet 350: 111±112
105. Dellinger RP, Zimmerman JL, Taylor
RW, et al (1998) Effects of inhaled nitric oxide in patients with acute respiratory distress syndrome: results of a
randomized phase II trial. Inhaled Nitric Oxide in ARDS Study Group.
Crit Care Med 26: 15±23
106. Gallart L, Lu Q, Puybasset L, Umamaheswara Rao GS, Coriat P, Rouby JJ
(1998) Intravenous almitrine combined with inhaled nitric oxide for
acute respiratory distress syndrome.
The NO Almitrine Study Group. Am
J Respir Crit Care Med 1998 158:
1770±1777
107. Zapol WM, Snider MT, Hill JD, et al
(1979) Extracorporeal membrane oxygenation in severe acute respiratory
failure. JAMA 242: 2193±2196
108. Morris AH, Wallace J, Menlove RL, et
al (1994) Randomized clinical trial of
pressure controlled inverse ratio ventilation and extracorporeal CO2 removal for adult respiratory distress syndrome. Am J Respir Crit Care Med
149: 295±305
109. Brunet F, Belghith M, Mira JP, et al
(1993) Extracorporeal carbon dioxide
removal and low-frequency positivepressure ventilation. Improvement in
arterial oxygenation with reduction of
risk of pulmonary barotrauma in patients with adult respiratory distress
syndrome. Chest 104: 889±898
110. Gattinoni L, Pesenti A, Mascheroni D,
et al (1986) Low frequency positive
pressure ventilation with extracorporeal CO2 removal In severe acute respiratory failure. JAMA 256: 881±886
111. Hirschl RB, Tooley R, Parent A, Johnson K, Bartlett RH (1996) Evaluation
of gas exchange, pulmonary compliance, and lung injury during total and
partial liquid ventilation in the acute
respiratory distress syndrome. Crit
Care Med 24: 1001±1008
112. Hirschl RB, Pranikoff T, Wise C, et al
(1996) Initial experience with partial
liquid ventilation in adult patients
with the acute respiratory distress syndrome. JAMA 275: 383±389
113. Bartlett RH, Croce M, Hirschl RH, et
al (1997) A phase II randomized, controlled trial of partial liquid ventilation
(PLV) in adult patients with acute hypoxemic respiratory failure. Crit Care
Med 25 [Suppl]:A35
114. Anzueto A, Baughman RP, Guntupalli
KK, et al (1996) Aerosolized surfactant in adults with sepsis-induced
acute respiratory distress syndrome.
Exosurf Acute Respiratory Distress
Syndrome Sepsis Study Group. N Engl
J Med 334: 1417±1421
115. Brigham KL, Stecenko AA (1995)
Gene therapy in acute critical illness.
New Horiz 3: 321±329
116. Bhattacharya J (2000) Gene therapy
for pulmonary edema. Am J Respir
Cell Mol Biol 22: 640±641
S 79
117. Wright PE, Carmichael LC, Bernard
GR (1994) Effect of bronchodilators
on lung mechanics in the acute respiratory distress syndrome (ARDS) Chest
106: 1517±1523
118. Torres A, Serra-Batlles J, Ros E, et al
(1992) Pulmonary aspiration of gastric
contents in patients receiving mechanical ventilation: the effect of body position. Ann Intern Med 116: 540±543
119. Ferrell BA, Osterweil D, Christenson
P (1993) A randomized trial of lowair-loss beds for treatment of pressure
ulcers. JAMA 269: 494±497
120. Gadek JE, DeMichele SJ, Karlstad
MD, et al (1999) Effect of enteral
feeding with eicosapentaenoic acid,
gamma-linolenic acid, and antioxidants in patients with acute respiratory
distress syndrome. Enteral Nutrition
in ARDS Study Group. Crit Care
Med 27: 1409±1420
121. Laporta D, Kleiman S, Begin L, et al
(1993) Traumatic perforation of the
cervical esophagus: A complication of
endotracheal intubation. Intensive
Care Med 19: 59±60
122. Keiser GJ, Bozentka NE, Gold BD
(1991) Laryngeal granuloma: a complication of prolonged endotracheal
intubation. Anesth Prog 38: 232±234
123. Demling RH, Read T, Lind TJ, et al
(1988) Incidence and morbidity of extubation failure in surgical intensive
care patients. Crit Care Med 16:
573±577
124. Marini JJ, Ravenscraft S (1992) Mean
airway pressure: physiologic determinants and clinical importance. I. Physiologic determinants and measurements. Crit Care Med 20: 1461±1472
125. Gattinoni L, Marcolin R, Caspani ML,
et al (1985) Constant mean airway
pressure with different patterns of positive pressure breathing during the
adult respiratory distress syndrome.
Bull Eur Physiopathol Respir 21:
275±279
126. Pesenti A, Marcolin R, Prato P, et al
(1985) Mean airway pressure vs positive end-expiratory pressure during
mechanical ventilation. Crit Care
Med 13: 34±37
127. Venus B, Jacobs HK, Lim L (1979)
Treatment of the adult respiratory distress syndrome with continuous positive airway pressure. Chest 76:
257±261
128. Smith RA, Kirby RR, Gooding JM,
Civetta JM (1980) Continuous positive
airway pressure (CPAP) by face mask.
Crit Care Med 8: 483±485
129. Rasanen J, Heikkila J, Downs J, Nikki
P, Vaisanen I, Viitanen A (1985) Continuous positive airway pressure by
face mask in acute cardiogenic pulmonary edema. Am J Cardiol 55: 296±300
130. Smatlak P, Knebel AR (1998) Clinical
evaluation of noninvasive monitoring
of oxygen saturation in critically ill patients. Am J Crit Care 7: 370±373
131. Petty TL (1990) A historical perspective of mechanical ventilation. In: Tobin MJ (ed) Critical care clinics: mechanical ventilation. Saunders, Philadelphia, pp 489±504
132. Eidelman LA, Jakobson DJ, Pizov R,
Geber D, Leibovitz L, Sprung CL
(1998) Foregoing life-sustaining treatment in an Israeli ICU. Intensive Care
Med 24: 162±166
133. Cook DJ, Walter SD, Cook RJ, et al
(1998) Incidence of and risk factors
for ventilator-associated pneumonia
in critically ill patients. Ann Intern
Med 129: 433±440
134. Cook DJ, De Jonghe B, Brochard L,
Brun-Buisson C (1998) Influence of
airway management on ventilator-associated pneumonia: evidence from
randomized trials. JAMA 279:
781±787
135. Kollef MH (1999) The prevention
of ventilator-associated pneumonia.
N Engl J Med 340: 627±634
136. Amato MBP, Barbas CS, Medeiros
DM, et al (1998) Effect of a protective-ventilation strategy on mortality
in the acute respiratory distress syndrome. N Engl J Med 338: 347±354
137. Stewart TE, Meade MO, Cook DJ, et
al (1998) Evaluation of a ventilation
strategy to prevent barotrauma in patients at high risk for acute respiratory
distress syndrome. Pressure- and Volume-Limited Ventilation Strategy
Group. N Engl J Med 338: 355±361
138. Brochard L, Roudot-Thoraval F, Roupie E, et al (1998) Tidal volume reduction for prevention of ventilator-induced lung injury in acute respiratory
distress syndrome. The Multicenter
Trail Group on Tidal Volume reduction in ARDS. Am J Respir Crit Care
Med 158: 1831±1838
139. Brower RG, Shanholtz CB, Fessler
HE, et al (1999) Prospective, randomized, controlled clinical trial comparing traditional versus reduced tidal
volume ventilation in acute respiratory
distress syndrome patients. Crit Care
Med 27 8: 1661±1663
140. Anonymous (1999) International Consensus Conferences in Intensive Care
Medicine: ventilator-associated lung
injury in ARDS. Am J Respir Crit
Care Med 160: 2118±2124
141. Esteban A, Frutos F, Tobin MJ, et al
(1995) A comparison of four methods
of weaning patients from mechanical
ventilation. N Engl J Med 332:
345±350
142. Brochard L, Rauss A, Benito S, et al
(1994) Comparison of three methods
of gradual withdrawal from ventilatory
support during weaning from mechanical ventilation. Am J Respir Crit
Care Med 150: 896±903
143. Yang KL, Tobin MJ (1991) A prospective study of indexes predicting the
outcome of trials of weaning from mechanical ventilation. N Engl J Med
324: 1445±1450
144. Kollef MH, Shapiro SD, Silver P, et al
(1997) A randomized, controlled trial
of protocol-directed versus physiciandirected weaning from mechanical
ventilation. Crit Care Med 25: 567±574
145. Esteban A, Alia I, Gordo F, et al
(1997) Extubation outcome after
spontaneous breathing trials with Ttube or pressure support ventilation.
The Spanish Lung Failure Collaborative Group. Am J Respir Crit Care
Med 156 (2 Pt 1):459±465
146. Esteban A, Alia I, Tobin MJ, et al
(1999) Effect of spontaneous breathing trial duration on outcome of attempts to discontinue mechanical ventilation. Spanish Lung Failure Collaborative Group. Am J Respir Crit Care
Med 159: 512±518
Intensive Care Med (2001) 27: S 80±S 92
Jean-Louis Vincent
Hemodynamic support in septic shock
)
J.-L. Vincent ( )
Department of Intensive Care, Erasme University Hospital,
Brussels, Belgium
E-mail: [email protected]
Phone: +32-2-5 55 33 80
Fax: +32-2-5 55 45 55
Introduction
Shock, defined as an imbalance between oxygen demand and oxygen supply, results in alterations in tissue
perfusion, with reduction in delivery of oxygen and other nutrients to tissues, causing cellular, and then organ,
dysfunction. The ultimate goals of hemodynamic therapy in shock are to restore effective tissue perfusion and
to normalize cellular metabolism. In hypovolemic, cardiogenic, and obstructive shock, hypotension occurs as
the result of a decrease in cardiac output, with consequent anaerobic tissue metabolism. Septic shock, however, typically results from distributive alterations, so
that alterations in tissue perfusion result from abnormal
control of the microvasculature with abnormal distribution of a normal or increased cardiac output. Cellular alterations in sepsis also result from the important inflammatory response, with the involvement of many mediators, including nitric oxide. Hence the endpoints of therapy are much more difficult to define with certainty
than in other forms of shock in which a reduction in
blood flow is the dominant problem.
The complex nature of the pathophysiology of sepsis
has led to considerable confusion and controversy regarding optimal patient management. Nevertheless, it
is possible to develop a basic approach to the hemodynamic support of sepsis, which will almost certainly
change as our understanding of sepsis improves further.
Methods
This review of the literature enables recommendations to be established and graded according to the strength of the available evidence.
Altered tissue perfusion in septic shock
Septic shock is characterized by hypotension, which in
adults generally refers to a mean arterial pressure below
65±70 mmHg. Hypotension is usually accompanied by
signs of altered tissue perfusion, for example, oliguria,
reduced capillary refill, and altered sensorium. Some
caution is necessary in interpreting these signs in septic
patients, however, since signs of peripheral vasoconstriction may be remarkably absent; even oliguria is not
always present. The adequacy of regional perfusion is
usually assessed clinically by evaluating indices of organ
function, although none of these parameters alone has
been validated as a reliable indicator of adequate resuscitation. These parameters include: coagulation abnormalities (disseminated intravascular coagulation); altered renal function with increased blood urea nitrogen
and creatinine; central nervous system dysfunction indicated by a clouded sensorium; altered liver parenchymal
function with increased serum levels of transaminases,
lactic dehydrogenase, and bilirubin; and altered gut perfusion, manifest by ileus and malabsorption.
Mixed venous oxygen saturation
Mixed venous oxygen saturation (SvO2) can be measured in patients with a Swan-Ganz catheter in place.
SvO2 is dependent on cardiac output, oxygen demand,
hemoglobin, and arterial oxygen saturation. The normal
SvO2 value is 70±75 % in critically ill patients but can be
elevated in septic patients due to maldistribution of
S 81
blood flow. Nevertheless, it is useful to measure SvO2
because if cardiac output becomes inadequate, SvO2 decreases. Ronco and colleagues [1] studied terminally ill
patients in whom treatment was withdrawn; SvO2 decreased dramatically before oxygen consumption started to fall, indicating that oxygen extraction capabilities
are not necessarily profoundly altered even in patients
with terminal stages of the disease process. Hence
SvO2, if normal or high does not necessarily indicate
that everything is alright, while a low SvO2 should
prompt rapid intervention to increase oxygen delivery
to the tissues.
Blood lactate levels
Hyperlactatemia (> 2 mEq/l) is typically present and
may be secondary to anaerobic metabolism due to hypoperfusion. However, the interpretation of blood lactate levels in septic patients is not always straightforward. Experimental studies have not always been able
to show a reduction in high-energy phosphate levels in
animal models of sepsis [2]. The differences between
studies may be related to the severity of the septic model, with more severe sepsis being associated with depletion of ATP despite maintenance of systemic oxygen delivery and tissue oxygenation. Also, measurements of
tissue PO2 in septic patients have not demonstrated tissue hypoxia in the presence of lactic acidosis [3]. However, if inhomogeneity in blood flow distribution is a
real phenomenon, it is likely that cell hypoperfusion
also exists with ischemia/reperfusion. A number of studies have suggested that elevated lactate levels may result
from cellular metabolic failure rather than from global
hypoperfusion in sepsis. Some organs may produce
more lactate than others, in particular, the lungs in acute
lung injury or acute respiratory distress syndrome [4, 5].
Elevated lactate levels can also result from decreased
clearance by the liver. Nonetheless, the prognostic value
of raised blood lactate levels has been well established
in septic shock patients [6], particularly if the high levels
persist [7, 8]. It is also of interest to note that blood lactate levels are of greater prognostic value than oxygenderived variables [9].
the serosa and muscularis [10], resulting in mucosal hypoxia. Any further reduction in splanchnic flow has a
correspondingly greater effect on gut hypoxia. Second,
the gut may have a higher critical oxygen delivery
threshold than other organs [11]. Third, the tip of the
villus is supplied by a central arteriole and drained by
venules passing away from the tip. A countercurrent exchange mechanism operates in the villus, whereby a
base to tip PO2 gradient exists, making the tip particularly sensitive to changes in regional flow and oxygenation. Fourth, constriction of the villus arteriole occurs
during sepsis [12], rendering the villus even more sensitive to reductions in blood flow. Fifth, the capillary density at the villus tip is reduced during sepsis [13], impeding the transfer of oxygen. Finally, gut ischemia increases intestinal permeability, which may increase bacterial
translocation, a suggested trigger or ªmotorº of the sepsis response and multiple organ failure.
Gastric tonometry has been proposed as a method to
assess regional perfusion in the gut by measuring
DPCO2. Calculations of gastric intramucosal pH (pHi)
have become obsolete as bicarbonate is global, nonspecific, and measured only intermittently. For these reasons gastric mucosal PCO2 may be more accurate than
pHi because this measure is not confounded by arterial
bicarbonate. Gastric mucosal PCO2 is influenced directly by systemic arterial PCO2, and some clinicians have
proposed using the gastric-arterial PCO2 difference as
the primary tonometric variable of interest [14]. Even
this measure is not a simple measure of gastric mucosal
hypoxia because either anaerobic metabolism decreased gastric blood flow in the absence of anaerobic
metabolism, or a combination of the two can increase
gastric mucosal PCO2 [14]. An early trial suggested
that tonometry derived parameters may be useful in
guiding therapy [15], but these findings were not confirmed recently [16], and many investigators have emphasized the limited sensitivity, and especially specificity, of these measurements. Various vasoactive agents
have been shown to have divergent effects on PgCO2
and pHi that are neither consistent nor predictable [17].
In summary, each endpoint parameter should be considered in the appropriate context, and the combination
of clinical parameters (mean arterial pressure, urine
flow, skin perfusion, level of consciousness) with blood
lactate levels is most useful.
Gut tonometry
The measurement of regional perfusion as a means of
detecting inadequate tissue oxygenation has focused on
the splanchnic circulation, as the hepatosplanchnic circulation is particularly sensitive to changes in blood
flow and oxygenation for several reasons. First, under
normal conditions the gut mucosa receives the majority
of total intestinal blood flow. However, in sepsis there
is a redistribution of flow away from the mucosa toward
Fluid resuscitation in septic shock
What is the endpoint of fluid resuscitation in septic
shock?
Answer: (a) adequate tissue perfusion, grade E.
S 82
Recommendation
The goal of fluid resuscitation in septic shock is restoration of tissue perfusion and normalization of cellular
metabolism.
Does fluid resuscitation increase cardiac output in
septic shock patients?
Can the use of pulmonary artery catheter guided
therapy improve outcome from septic shock?
Answer: uncertain, grade D.
Recommendation
Answer: yes, grade C.
When central venous pressure increases, a pulmonary
artery catheter is probably required, although its role
has recently been questioned [24].
Recommendation
Rationale
Volume repletion in patients with septic shock produces
significant increases in cardiac output and systemic oxygen delivery [18, 19], and fluids alone are sometimes sufficient to reverse hypotension and restore hemodynamic stability [20].
When pulmonary artery catheter measurements of cardiac output and SvO2 are available, filling pressures
should be increased to a level associated with maximal
cardiac output. In most patients with septic shock quite
high pulmonary artery occluded pressures are required,
despite the risk of pulmonary edema.
Should fluid infusion be the initial step in the
cardiovascular support of septic shock patients?
Answer: yes, grade D.
Recommendation
Requirements for fluid infusion are not easily determined, and therefore the fluid challenge should be titrated to the clinical endpoints of blood pressure, heart
rate, and urine output. Central venous pressure is initially required to evaluate the complex relationship between intravascular blood volume and cardiac function.
It is difficult to give optimal values for cardiac filling
pressures.
Rationale
Septic shock can be associated with either absolute or
relative hypovolemia. Large fluid deficits can exist, as a
consequence of external (e.g., diarrhea, sweating) or internal (e.g., peritonitis) losses. Relative hypovolemia is
related to the maldistributive defect with vasodilation
and peripheral blood pooling. The initial phases of experimental and clinical septic shock present as a low cardiac output syndrome with low filling pressures. Failure
to appreciate the degree of underlying hypovolemia
may result in a low cardiac output. The hyperdynamic
state is apparent only after volume repletion [21, 22]. Increased blood and plasma volumes are associated with
increased cardiac output and enhanced survival from
septic shock [23].
Choice of fluid
Patients with septic shock can be successfully resuscitated with crystalloid or colloid, although the choice of fluid continues to be a matter of debate, with colloids usually preferred in Europe, and crystalloids more widely
used in North America. There are many different colloid solutions available including natural solutions (albumin, plasma protein fraction) and artificial solutions
(gelatins, dextrans and hydroxyethyl starch). The most
commonly used solutions are albumin and hetastarch.
Is resuscitation with colloids or crystalloids associated
with similar outcomes in septic shock?
Answer: uncertain, grade C.
Recommendation
Increases in cardiac output and systemic oxygen delivery are proportional to the degree of intravascular volume expansion achieved.
Rationale
When crystalloids and colloids are titrated to the same
level of filling pressure, they restore tissue perfusion to
the same degree [25], but for the same effect, two to
four times more volume of crystalloid is required than
colloid, and slightly longer infusion periods may be nec-
S 83
essary to achieve desired hemodynamic endpoints. Colloid solutions are, however, much more expensive than
crystalloid solutions.
Should albumin be avoided in resuscitation from septic
shock?
Answer: uncertain, grade C.
Albumin is a naturally occurring plasma protein accounting for approximately 80 % of the plasma colloid
osmotic pressure in normal subjects. In the presence of
peripheral edema, mobilization of extravascular volume
can be achieved by using hyperoncotic (20 or 25 %) albumin.
Hydroxyethyl starch (HES, hetastarch) is a synthetic
colloid available in a 6 % solution of normal saline.
HES molecules may also affect endothelial cell function
and reduce endothelial cell activation and injury [26],
which could account in part for the preservation of microvascular structures with reduced leakage seen with
HES in experimental sepsis models [27]. HES solutions
can decrease factor VIII activity and prolong prothrombin time, so that the total amount of starch infused
should be limited. The possibility that the long-term deposition of higher molecular weight hetastarch particles
in the reticuloendothelial system could have immunosuppressive effects has caused some concern. However,
in experimental studies, macrophage function and reticuloendothelial function were not altered by HES [28].
Gelatin solutions are cheaper but also less efficacious. They are not currently available in the United
States, although they may soon become available there.
Blood transfusion
Can one recommend a minimum hemoglobin
concentration in severe sepsis?
Answer: yes, 7±8 g/dl, grade B.
Can one recommend a minimum hemoglobin
concentration in septic shock?
Answer: uncertain, grade E.
Recommendation
The optimal hemoglobin and hematocrit for patients
with septic shock is unclear. Most experts recommend
hemoglobin levels of 9±10 gm/dl in patients with septic
shock. This degree of anemia is usually well tolerated
in most patients, even with cardiac impairment.
Rationale
Excessive tachycardia, very low SvO2, or electrocardiographic signs of myocardial ischemia may suggest the
need for higher hemoglobin levels to be maintained. A
large study by Hebert et al. [29] in a mixed group of
ICU patients showed no benefit of transfusion to a hemoglobin level of 10 vs. 7 g/dl. Indeed, this trial found
that use of the lower hemoglobin level trigger for transfusion, 7 g/dl, resulted in improved survival. When considering blood transfusion to improve oxygenation in
critically ill patients, stored blood may be less effective
than fresh blood [30].
Vasopressor therapy in septic shock
Does adrenergic support improve outcome from septic
shock?
Answer: yes, grade E.
Recommendation
When fluid challenge fails to restore an adequate arterial pressure and organ perfusion, therapy with vasopressor agents should be started. Vasopressor therapy may
also be required transiently to sustain life and maintain
perfusion in the face of life-threatening hypotension,
even when cardiac filling pressures are not elevated.
Rationale
In shock states the measurement of blood pressure using
a cuff is often unreliable and inaccurate, and patients
should have an arterial catheter in place for continuous
blood pressure monitoring. This is even more the case
in patients receiving vasopressor therapy for shock, as
restoration of an adequate blood pressure is the required endpoint and measure of effectiveness. The precise level of mean arterial pressure to aim for is, however, not entirely certain and is likely to vary among patients. In animal studies a mean arterial pressure of less
than 60 mmHg is associated with compromised autoregulation in the coronary, renal, and central nervous system vascular beds, and blood flow may be reduced.
Some patients, however, especially the elderly, may require higher blood pressures to maintain adequate perfusion. Assessment of regional and global perfusion by
a combination of the methods outlined previously is advisable.
S 84
Effects of vasopressors on renal perfusion
Although no prospective randomized studies have demonstrated a significant improvement in renal function
with vasopressors, a number of open-label clinical series
support an increase in renal perfusion pressure [31, 32,
33, 34, 35, 36, 37, 38, 39]. Excessive doses of vasopressors may shift the renal autoregulation curve to the
right, necessitating a greater perfusion pressure for a
specified renal blood flow. The precise target mean
blood pressure level depends on the premorbid blood
pressure but can be as high as 75 mmHg [31, 33, 34, 35,
36, 37, 38, 39]. However, individual levels should be
kept at the minimum needed to reestablish urine flow,
and in some patients this can be achieved with a mean
arterial pressure of 60 or 65 mmHg. Certain patients
may remain oliguric despite normalization of systemic
hemodynamic variables [32, 33, 34, 37, 39]. This may be
due to the absence of an increase in renal blood flow, a
decrease in glomerular perfusion pressure, or irreversible ischemic renal lesions.
Although in nonseptic conditions combination therapy with the use of low-dose dopamine (1±4 mg kg±1
min±1) in addition to norepinephrine in an anesthetized
dog model and healthy volunteers resulted in significantly higher renal blood flow and lower renal vascular
resistance [40, 41], such effects have not been conclusively demonstrated in septic shock, and there is no information available on the effects of such therapy on patient survival.
blood flow, increases PgCO2 production, and decreases
pHi, suggesting that the drug alters oxygen supply in
the splanchnic circulation [42]. At low doses dopamine
increases splanchnic oxygen delivery by 65 % but
splanchnic oxygen consumption by only 16 %. Despite
this, dopamine may decrease pHi, perhaps by a direct
effect on the gastric mucosal cell. The effects of dopamine on cellular oxygen supply in the gut remain incompletely defined. The effects of norepinephrine on
splanchnic circulation are hardly predictable. The combination of norepinephrine and dobutamine appears to
be more predictable and more appropriate to the goals
of septic shock therapy than the effects of epinephrine
alone.
Individual vasopressor agents
Among adrenergic agents, are dopamine or
norepinephrine the first line agents to correct
hypotension in septic shock?
Answer: yes, grade E.
Should low-dose dopamine be routinely administered
for renal protection?
Answer: no, grade D.
Dopamine
Effects of vasopressors on the splanchnic circulation
Is the combination of norepinephrine and dobutamine
superior to dopamine in the treatment of septic shock?
Answer: uncertain, grade C.
Recommendation
The effects of dopamine on cellular oxygen supply in
the gut remain incompletely defined. The effects of
norepinephrine on splanchnic circulation are hardly
predictable. The combination of norepinephrine and
dobutamine appears to be more predictable and more
appropriate to the goals of septic shock therapy than
the effects of epinephrine alone.
Rationale
Splanchnic perfusion and the integrity of the gut mucosa
may play an important role in the pathogenesis of multiple organ failure. Epinephrine decreases splanchnic
Recommendation
The hemodynamic effects of dopamine in patients with
septic shock are well established. Dopamine increases
mean arterial pressure primarily by increasing cardiac
index with minimal effects on systemic vascular resistance. The increase in cardiac index is due to an increase
in stroke volume, and to a lesser extent, to increased
heart rate [43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54]. Patients receiving dopamine at rates greater than
20 mg kg±1 min±1 show increases in right heart pressures
as well as in heart rate, and therefore doses should not
usually exceed 20 mg kg±1 min±1, at least not without adequate hemodynamic monitoring.
Rationale
Dopamine is the natural precursor of norepinephrine
and epinephrine, and possesses several dose-dependent
pharmacological effects. Generally, at doses less than
5 mg kg±1 min±1 dopamine stimulates dopaminergic DA1
receptors in the renal, mesenteric, and coronary beds,
S 85
resulting in vasodilation. Infusion of low doses of
dopamine causes an increase in glomerular filtration
rate, renal blood flow, and sodium excretion. At doses
of 5±10 mg kg±1 min±1, ª1-adrenergic effects become predominant, resulting in an increase in cardiac contractility and heart rate. Dopamine also causes the release of
norepinephrine from nerve terminals, contributing to
its cardiac effects. At higher doses (above
10 mg kg±1 min±1), ª1-adrenergic effects predominate,
leading to arterial vasoconstriction and an increase in
blood pressure.
Dopamine's effect on gastric tonometric parameters
has been evaluated with mixed results. Roukonen et al.
[51] and Meier-Hellmann et al. [54] have documented
that dopamine increases splanchnic blood flow. Neviere
and colleagues [55] reported that dopamine is associated with a reduction in gastric mucosal blood flow; there
were changes in gastric PCO2, gastric-arterial PCO2 difference, and calculated intramucosal pH. They [55] concluded that they could not determine whether the reduction in gastric mucosal blood flow was critical because there were no changes in the acid-base parameters of the patients.
Recent studies have shown that dopamine may alter
the inflammatory response in septic shock by decreasing
the release of a number of hormones, including prolactin [56].
Norepinephrine
Recommendation
Summarizing the results of studies of norepinephrine, it
can be concluded that norepinephrine markedly improves mean arterial pressure and glomerular filtration.
This is particularly true in the high output-low resistance state of many septic shock patients. After restoration of systemic hemodynamics, urine flow reappears in
most patients and renal function improves without the
use of low-dose dopamine or furosemide. This fact supports the hypothesis that renal ischemia observed during hyperdynamic septic shock is not worsened by norepinephrine infusion and even suggests that this drug
may effectively optimize renal blood flow and renal vascular resistance.
Rationale
Norepinephrine is a potent ª-adrenergic agonist with
some ª-adrenergic agonist effects. The effects of norepinephrine have been studied in a number of studies on
patients with septic shock. In open label trials norepinephrine has been shown to increase mean arterial pressure in patients with hypotension resistant to fluid resus-
citation and dopamine. The potential that norepinephrine may have negative vasoconstrictive effects on regional vascular beds such as the liver and the kidney,
with resultant regional ischemia, has meant that norepinephrine has commonly been reserved for use as a last
resort, with predictably poor results.
However, recent experience with the use of norepinephrine in patients with septic shock suggests that it
can successfully increase blood pressure without causing
the feared deterioration in organ function. Most studies
have given septic patients fluid to correct hypovolemia
before dopamine, with or without dobutamine, titrated
to doses of 7±25 mg kg±1 min±1 to achieve the target
blood pressure. Only if this regime failed was norepinephrine added [31, 32, 35, 39, 42, 51, 57, 58]. In older
studies norepinephrine was added after the use of metaraminol, methoxamine, or isoproterenol [43, 59]. A
few studies have used norepinephrine as the only adrenergic agent to correct sepsis-induced hemodynamic abnormalities [34, 37, 50, 51, 53].
In most studies the mean dose of norepinephrine was
0.2±1.3 mg kg±1 min±1, although the initial dose can be as
low as 0.01 mg kg±1 min±1 [32], and the highest reported
norepinephrine dose was 3.3 mg kg±1 min±1 [58]. Thus,
large doses of the drug can be required in some patients
with septic shock, which may be due to ª-receptor
down-regulation in sepsis [60].
Norepinephrine therapy usually causes a statistically
and clinically significant increase in mean arterial pressure due to its vasoconstrictive effects, with little change
in heart rate or cardiac output, leading to increased systemic vascular resistance. Several studies have demonstrated increases in cardiac output ranging from 10 %
to 20 % and increases in stroke volume index of
10±15 % [39, 43, 53]; other studies, however, have observed no significant changes in either cardiac output
or stroke volume index after the use of norepinephrine
in the presence of a significant increase in vascular resistance, suggesting that norepinephrine is exerting ª1receptor agonist effects [31, 32, 33, 35, 58, 61]. Obviously, since cardiac index is either increased or unchanged,
and mean arterial pressure is consistently increased,
left ventricular stroke work index is always statistically
increased with norepinephrine. With regards to pulmonary capillary wedge pressure, no clinically significant
changes are reported.
Norepinephrine should be used only to restore normal values (or values in the lower part of the normal
range) of mean arterial blood pressure and systemic vascular resistance. Higher values should be avoided during norepinephrine therapy, since elevated cardiac afterload could be deleterious in cases of severe underlying
cardiac dysfunction. Due to methodological problems
with the use of systemic vascular resistance as the sole
measurement of peripheral resistance, the use of mean
arterial pressure is more advisable, although, as men-
S 86
tioned above, the optimal target mean arterial pressure
is not known, and depends on many factors including
age and premorbid condition.
Norepinephrine is probably more effective than
dopamine at reversing hypotension in septic shock patients. Martin et al. [50] carried out a study with the most
striking findings. They prospectively randomized 32 volume-resuscitated patients with hyperdynamic sepsis
syndrome to receive either dopamine (2.5±25 mg kg±1
min±1) or norepinephrine (0.5±5.0 mg kg±1 min±1) to
achieve and maintain normal hemodynamic and oxygen
transport parameters for at least 6 h. If the goals were
not achieved with one agent, the other was added. The
groups were similar at baseline. Dopamine administration (10±25 mg kg±1 min±1) was successful in only 31 %
(5 of 16) of patients whereas norepinephrine (1.5 
1.2 mg kg±1 min±1) resulted in success in 93 % (15 of 16)
of patients (p < 0.001). Of the 11 patients who did not
respond to dopamine 10 responded when norepinephrine was added. In contrast, the one patient who did
not respond to norepinephrine failed to respond to
dopamine. The survival rate differed between the two
groups (59 % norepinephrine vs. 17 % dopamine) although the study was not statistically designed to examine this issue.
In patients with hypotension and hypovolemia, for
example, during hemorrhagic shock, norepinephrine
and other vasoconstrictor agents have severe detrimental effects on renal hemodynamics. Despite the constant
improvement in blood pressure, renal blood flow does
not increase, and renal vascular resistance continues to
rise [62]. Renal tissue oxygen tension can decrease
markedly, worsening renal ischemia [63]. Indeed, norepinephrine has been demonstrated to cause ischemia-induced acute renal failure in rats [64]. In nonseptic shock
patients norepinephrine has a marked vasoconstrictive
effect in most vascular beds, reducing blood flow to the
liver, skeletal muscle, and kidneys [65]. When used in
normotensive and hypertensive patients, norepinephrine can decrease renal blood flow and increase renal
vascular resistance [66].
However, in hyperdynamic septic shock, during
which urine flow is believed to decrease mainly as a result of lowered renal glomerular perfusion pressure,
the situation is different. Since norepinephrine has a
greater effect on efferent arteriolar resistance and increases the filtration fraction, normalization of renal
vascular resistance could effectively reestablish urine
flow. The importance of this during norepinephrine infusion was shown by Schaer et al. [40], who demonstrated that while renal vascular resistance increased during
norepinephrine infusion; renal blood flow remained stable or even slightly increased because the drug enhanced cardiac output and renal perfusion pressure.
The effects of norepinephrine on renal function in
sepsis have been evaluated in four studies. Desjars
et al. [33] studied 22 septic shock patients treated with
norepinephrine (0.5±1.5 mg kg±1 min±1) and dopamine
(2±3 mg kg±1 min±1). Serum creatinine, blood urea nitrogen, free water clearance, and fractional excretion of sodium decreased significantly, while urine output, creatinine clearance, and osmolar clearance increased significantly. In this study [33] six of seven patients considered
at risk for developing acute renal failure had improved
renal function during norepinephrine treatment, and
only one developed nonoliguric acute renal failure requiring dialysis. Martin et al. [37] studied 24 septic
shock
patients
(treated
with
norepinephrine
(1.1 mg kg±1 min±1 + dobutamine at 8±14 mg kg±1 min±1 +
dopamine at 6±17 mg kg±1 min±1). No patient received
low-dose dopamine or furosemide. Normalization of
systemic hemodynamics was followed by reestablishment of urine flow, decrease in serum creatinine, and increase in creatinine clearance. Fukuoka et al. [34] studied 15 patients with septic shock treated with norepinephrine (0.05±0.24 mg kg±1 min±1), dopamine (9 mg kg±1
min±1) and dobutamine (5 mg kg±1 min±1). Only patients
with a normal serum lactate concentration had an increase in systemic vascular resistance, and an increase
in urine flow. Creatinine clearance was not affected
(18.8+5.5 ml/min before and 20.1+6.6 ml/min after norepinephrine). Patients with elevated serum lactate concentrations had no change in vascular resistance, a decrease in creatinine clearance (32.6  6.4 to
11.9  4.9 ml/min), and required higher doses of furosemide. The authors concluded that the serum lactate
concentration may predict which patients will experience potentially adverse renal effects with norepinephrine. However, this study included only a very limited number of patients and is at variance with the findings of other studies [39, 43, 50, 53, 57] in which vascular
resistance and urine flow were increased in patients with
elevated lactate concentrations (as high as 4.8 
1.6 mmol/l [50]). Redl-Wenzel et al. [39] studied 56 patients with septic shock treated with norepinephrine
(0.1±2.0 mg kg±1 min±1) and dopamine (2.5 mg kg±1
min±1). During norepinephrine infusion creatinine clearance increased significantly from 75  37 to 102  43 ml/
min after 48 h of treatment. The authors concluded that
mean arterial pressure could be increased by norepinephrine with a positive effect on organ perfusion and
oxygenation.
The effects of norepinephrine on serum lactate concentrations have been assessed in five studies. Four
studies assessed changes in serum lactate concentrations
over a relatively short period of time, i.e., 1±3 h. Hesselvik et al. [35] reported unchanged lactate levels during
norepinephrine therapy, but the actual values were not
given. In the other three studies [51, 53, 57] mean values
of serum lactate concentrations did not change over the
1- to 3-h study period. It should be noted that initial values were not very high (1.8±2.3 mmol/l). Since blood
S 87
flow tended to significantly improve and lactic acid concentrations decreased (but not significantly) in one
study, it is unclear whether sufficient time elapsed between measurements to see a significant norepinephrine-induced change in serum lactate concentrations. In
the last study [50] initial lactate concentrations were elevated (4.8  1.6 mmol/l), and a statistically and clinically
significant decrease in lactate levels was observed at the
end of the 6-h study period. Norepinephrine thus does
not worsen, and may even improve, tissue oxygenation,
as assessed by serum lactate levels, in patients with septic shock.
Ruokonen et al. [51] measured splanchnic blood flow
and splanchnic oxygen consumption in septic shock patients receiving either norepinephrine (0.07±0.23 mg
kg±1 min±1) or dopamine (7.6±33.8 mg kg±1 min±1) to correct hypotension. With norepinephrine no overall changes in splanchnic blood flow and splanchnic oxygen consumption or extraction were noted, and in individual patients its effects on splanchnic blood flow were unpredictable (increased in three patients, decreased in two).
Dopamine caused a consistent and statistically significant increase in splanchnic blood flow. Meier-Hellman
et al. [54] studied patients changed from dobutamine to
norepinephrine. They observed a significant decrease
in hepatic venous oxygen saturation. In another group
of patients, they studied the effects of switching from
dobutamine plus norepinephrine to the latter drug
alone. They observed the previously reported changes
in hepatic venous oxygen saturation together with a decrease in splanchnic blood flow (green dye dilution technique) and in cardiac output. Splanchnic oxygen consumption remained unchanged due to a regional increase in oxygen extraction. The decrease in splanchnic
blood flow paralleled the decrease in cardiac output.
The authors concluded that as long as cardiac output is
maintained treatment with norepinephrine alone has
no negative effects on splanchnic tissue oxygenation.
This finding was confirmed by Marik and Mohedin [53]
who observed a significant increase in pHi (from
7.16  0.07 to 7.23  0.07) over 3 h of norepinephrine
treatment. During treatment with dopamine pHi decreased significantly (7.24  0.04 to 7.18  0.05).
Reinelt et al. [67] tested the hypothesis that when
dobutamine is added to norepinephrine to obtain a
20 % increase in cardiac index in septic shock patients,
splanchnic blood flow and oxygen consumption increases and hepatic metabolic activity (hepatic glucose production) improves. Splanchnic blood flow and cardiac
index increased in parallel, but there was no effect on
splanchnic oxygen consumption and hepatic glucose
production decreased. The conclusion of the authors
was that splanchnic oxygen consumption was not dependent on delivery in septic shock patients well resuscitated with norepinephrine. Levy et al. [42] studied the effects of the combination of norepinephrine and dob-
utamine on gastric tonometric variables in 30 septic
shock patients. pHi and gastric PCO2 gap were normalized within 6 h, while in epinephrine-treated patients
pHi decreased and gastric PCO2 gap increased. Changes
in the epinephrine group were only transient and were
corrected within 24 h but could potentially have caused
splanchnic ischemia. The authors concluded that the
combination of norepinephrine with dobutamine was
more predictable than epinephrine.
Clinical experience with norepinephrine in septic
shock patients suggests that this drug can successfully
increase blood pressure without causing deterioration
in cardiac index or organ function. Norepinephrine (at
doses of 0.01±3 mg kg±1 min±1), consistently improves
hemodynamic variables in the large majority of patients
with septic shock. The effects of norepinephrine on oxygen transport variables remain undefined from the
available data, but most studies find other clinical parameters of peripheral perfusion to be significantly improved. Unfortunately only one published study was
controlled [50] and a prospective, randomized clinical
trial is still required to assess whether the use of norepinephrine in septic shock patients affects mortality compared to other vasopressors. The data are sufficiently
strong to suggest that when contemplated in the treatment of septic shock patients, norepinephrine should
be used early and not merely as a last resort.
Epinephrine
In patients who fail to respond to fluid administration or
other vasopressors epinephrine can increase arterial
pressure primarily by increasing cardiac index and
stroke volume [38, 68, 69, 70]. Moran et al. [70] reported
a linear relationship between epinephrine dose and
heart rate, mean arterial pressure, cardiac index, left
ventricular stroke work index, and oxygen delivery, and
consumption. Epinephrine, however, has detrimental
effects on splanchnic blood flow and causes transient
decreases in pHi and increases in the PCO2 gap [42, 71].
Epinephrine administration has been associated with
increases in systemic and regional lactate concentrations [42, 69, 72], although the cause of these increases
is unclear. As the monitoring periods in all these studies
were short, it is unclear whether these increases are a
transient phenomenon. Other adverse effects of epinephrine include tachyarrhythmias.
In summary, epinephrine clearly increases blood
pressure in patients unresponsive to other agents. However, because of its negative effects on gastric blood flow
and blood lactate concentrations its use should be limited.
S 88
Other vasoconstricting agents
Dobutamine
Phenylephrine, a selective-1-adrenergic agonist, has
been used in septic shock patients, although there are
concerns about its potential to reduce cardiac output
and lower heart rate in these patients. Doses of phenylephrine start at 0.5 mg kg±1 min±1 and reach a maximum
dose of 5±8 mg kg±1 min±1. A few studies have evaluated
the clinical use of phenylephrine in septic shock [73, 74,
75]. Reinelt et al. [75] reported reduced splanchnic
blood flow and oxygen delivery in six septic shock patients treated with phenylephrine compared to norepinephrine.
Dobutamine is an adrenergic agonist that stimulates b1-,
b2-, and b1-adrenergic receptors. A number of studies
have investigated the effect of dobutamine on cardiac
function during sepsis or septic shock [82, 83, 84, 85,
86]. The doses utilized ranged from 2 to 28 mg kg±1 min±1.
The majority of these studies found increases in cardiac
index combined with increases in stroke volume and
heart rate.
Epinephrine
See individual vasopressor agents, above.
Inotropic therapy in septic shock
Although the cardiac index is usually maintained in the
volume resuscitated septic shock patient, cardiac function is impaired [76]. Characterized by ventricular dilatation, a decreased ejection fraction an impaired contractile response to volume loading, and a low peak systolic pressure/end-systolic volume [77, 78], the mechanism of the myocardial dysfunction is complex. Coronary blood flow is usually normal and there is no net lactate production across the coronary vascular bed, so myocardial ischemia is not implicated. Alterations in intracellular calcium homeostasis and in - -adrenergic signal
transduction may be contributory factors. Several inflammatory mediators have been shown to cause myocardial depression in various animal models, including
cytokines [79], platelet-activating factor, and nitric oxide [80].
Inotropic therapy in septic shock is thus not straightforward. Cardiac output is usually not decreased, and
multiple factors may be involved in the depressed cardiac function. In patients with decreased cardiac output
the goals of therapy are relatively clear and are aimed
at restoring normal physiology. Because of the complexity of assessment of clinical parameters in septic patients, direct measurement of cardiac output by invasive
hemodynamic monitoring is advisable, but other endpoints of global perfusion should be followed as well.
When global hypoperfusion is manifest by a decreased
SvO2, monitoring of SvO2 can be helpful to guide response to therapy. Similarly, although lactate production in sepsis is complex, a fall in blood lactate levels
during inotropic therapy is a good prognostic sign [81].
Individual inotropic agents
Is dobutamine the pharmacological agent of choice to
increase cardiac output in the treatment of septic shock?
Answer: yes, grade D.
Dopexamine
Dopexamine is a dopamine analog that stimulates b2adrenergic and dopamine 1 and 2 receptors. It is not approved for use in the United States. Several studies have
evaluated short-term infusions of dopexamine in sepsis
or septic shock and demonstrated significant improvements in cardiac index and left ventricular stroke work
index [87, 88, 89]. In addition, mesenteric perfusion, as
assessed by gastric tonometry, were improved compared
to baseline values in initial studies [88], but this has not
been confirmed in subsequent studies [90].
Phosphodiesterase inhibitors
Phosphodiesterase inhibitors alone, such as amrinone
and milrinone, have little place in the treatment of septic shock. They may be considered in combination with
adrenergic agents. One study evaluating milrinone in
pediatric patients with sepsis observed that cardiac index and right and left ventricular stroke work indices
improved significantly, with little change in heart rate
[91].
Other agents
Calcium supplementation has been proposed in the
management of myocardial dysfunction in septic shock.
However, no consistent beneficial hemodynamic effect
of calcium administration in septic patients has been reported [92], and increased mortality has been reported
in animal models [93, 94]. Digoxin has been reported
significantly to improve cardiac performance in hypodynamic septic patients [95].
S 89
The ¹supranormalª approach
Are hyperkinetic patterns associated with better
outcome in septic shock patients?
Answer: yes, grade C.
Systemic alterations occurring in sepsis including an increased oxygen demand, altered oxygen extraction, and
myocardial depression, can explain how circulatory failure may persist despite a normal or high cardiac output
(see alterations in distribution above). Hence, it may
be valuable to further increase cardiac index (to 'supranormal' values) despite the fact that it is not typically
decreased. This remains a controversial issue. Initial
studies by Shoemaker et al. [96] seemed to support this
approach, but in order to correlate with improved survival, the cardiac index needed to be greater than
4.5 l m±2 min±1, oxygen delivery greater than 600 ml
m±2 min±1, and oxygen consumption greater than 170 ml
m±2 min±1. Randomized studies [97, 98] to test this hypothesis in all critically ill patients have produced rather
negative results, with increased mortality rates in the
study by Hayes et al. [97] which sometimes involved
the administration of very high doses of dobutamine.
Moreover, it is possible that increases in cardiac index
and oxygen delivery may simply reflect higher underlying physiological reserve of the patients, associated
with an increased chance of survival. The main problem
with studying this approach in a typical ICU population
is the heterogeneity of the patients included.
What may be beneficial in certain groups of patients
could potentially be harmful in others, thus giving an
overall negative result. Tuchschmidt et al. [99] included
only patients with septic shock and obtained more positive results. However, the strategy of increasing oxygen
delivery to predetermined elevated endpoints of cardiac
index and oxygen delivery cannot be recommended routinely. Different interventions to increase oxygen delivery, such as fluid resuscitation, blood transfusion, or infusion of vasoactive agents, can have different effects
on regional perfusion. This is an area of controversy
and ongoing research; randomized controlled trials,
with clear, reproducible treatment algorithms and use
of defined measures of regional perfusion are necessary.
In the meanwhile, one should define the goals and desired endpoints of inotropic therapy in septic patients
and use these endpoints to monitor and titrate therapy.
References
1. Ronco JJ, Fenwick JC, Tweeddale MG,
Wiggs BR, Phang PT, Cooper DJ,
Cunningham KF, Russell JA, Walley
KR (1993) Identification of the critical
oxygen delivery for anaerobic metabolism in critically ill septic and nonseptic
humans. JAMA 270: 1724±1730
2. James JH, Luchette FA, McCarter FD,
Fischer JE (1999) Lactate is an unreliable indicator of tissue hypoxia in injury
or sepsis. Lancet 354: 505±508
3. Boekstegers P, Weidenhofer S, Pilz G,
Werdan K (1991) Peripheral oxygen
availability within skeletal muscle in
sepsis and septic shock: comparison to
limited infection and cardiogenic shock.
Infection 19: 317±323
4. Kellum JA, Kramer DJ, Lee K, Mankad S, Bellomo R, Pinsky MR (1997)
Release of lactate by the lung in acute
lung injury. Chest 111: 1301±1305
5. De Backer D, Creteur J, Zhang H, Norrenberg M, Vincent JL (1997) Lactate
production by the lungs in acute lung injury. Am J Respir Crit Care Med 156:
1099±1104
6. Weil MH, Afifi AA (1970) Experimental and clinical studies on lactate and
pyruvate as indicators of the severity of
acute circulatory failure (shock). Circulation 41: 989±1001
7. Vincent JL, Dufaye P, Berre J, Leeman
M, Degaute JP, Kahn RJ (1983) Serial
lactate determinations during circulatory shock. Crit Care Med 11: 449±451
8. Friedman G, Berlot G, Kahn RJ, Vincent JL (1995) Combined measurements of blood lactate concentrations
and gastric intramucosal pH in patients
with severe sepsis. Crit Care Med 23:
1184±1193
9. Bakker J, Coffernils M, Leon M, Gris P,
Vincent JL (1991) Blood lactate levels
are superior to oxygen derived variables in predicting outcome in human
septic shock. Chest 99: 956±962
10. Vallet B, Lund N, Curtis SE, Kelly D,
Cain SM (1994) Gut and muscle tissue
PO2 in endotoxemic dogs during shock
and resuscitation. J Appl Physiol 76:
793±800
11. Nelson DP, King CE, Dodd SL, Schumacker PT, Cain SM (1987) Systemic
and intestinal limits of O2 extraction in
the dog. J Appl Physiol 63: 387±394
12. Schmidt H, Secchi A, Wellmann R,
Bach A, Bohrer H, Gebhard MM, Martin E (1996) Effect of endotoxemia on
intestinal villus microcirculation in rats.
J Surg Res 61: 521±526
13. Farquhar I, Martin CM, Lam C, Potter
R, Ellis CG, Sibbald WJ (1996) Decreased capillary density in vivo in bowel mucosa of rats with normotensive
sepsis. J Surg Res 61: 190±196
14. Russell JA (1997) Gastric tonometry:
does it work? Intensive Care Med 23:
3±6
15. Gutierrez G, Palizas F, Doglio G,
Wainsztein N, Gallesio A, Pacin J, Dubin A, Schiavi E, Jorge M, Pusajo J,
Klein F, San Roman E, Dorfman B,
Shottlender J, Giniger R (1992) Gastric
intramucosal pH as a therapeutic index
of tissue oxygenation in critically ill patients. Lancet 339: 195±199
16. Gomersall CD, Joynt GM, Freebairn
RC, Hung V, Buckley TA, Oh TE
(2000) Resuscitation of critically ill patients based on the results of gastric
tonometry: a prospective, randomized,
controlled trial. Crit Care Med 28:
607±614
17. Silva E, De Backer D, Creteur J, Vincent JL (1998) Effects of vasoactive
drugs on gastric intramucosal pH. Crit
Care Med 26: 1749±1758
18. Haupt MT, Gilbert EM, Carlson RW
(1985) Fluid loading increases oxygen
consumption in septic patients with lactic acidosis. Am Rev Respir Dis 131:
912±916
S 90
19. Packman MI, Rackow LC (1983) Optimum left heart filling pressures during
fluid resuscitation of patients with hypovolemic and septic shock. Crit Care
Med 11: 165±169
20. Sugerman HJ, Diaco JF, Pollock TW,
Miller LD (1971) Physiologic management of septicemic shock in man. Surg
Forum 22: 3±5
21. Carroll G, Snyder J (1982) Hyperdynamic severe intravascular sepsis depends on fluid administration in cynomolgus monkey. Am J Physiol
243:R131±R141
22. Rackow EC, Kaufman BS, Falk JL, Astiz ME, Weil MH (1987) Hemodynamic
response to fluid repletion in patients
with septic shock: evidence for early depression of cardiac performance. Circ
Shock 22: 11±22
23. Weil MH, Nishjima H (1978) Cardiac
output in bacterial shock. Am J Med
64: 920±922
24. Connors AF, Speroff T, Dawson NV,
Thomas C, Harrell FE, Wagner D,
Desbiens N, Goldman L, Wu AW, Califf
RM, Fulkerson WJ, Vidaillet H, Broste
S, Bellamy P, Lynn J, Knaus W, SUPPORT Investigators (1996) The effectiveness of right heart catheterization
in the initial care of critically ill patients. JAMA 276: 889±897
25. Rackow EC, Falk JL, Fein IA, Siegel
JS, Packman MI, Haupt MT, Kaufman
BS, Putnam D (1983) Fluid resuscitation in circulatory shock: a comparison
of the cardiorespiratory effects of albumin, hetastarch, and saline solutions in
patients with hypovolemic and septic
shock. Crit Care Med 11: 839±850
26. Boldt J, Müller M, Heesen M, Heyn O,
Hempelmann G (1996) Influence of different volume therapies on platelet
function in the critically ill. Intensive
Care Med 22: 1075±1081
27. Morisaki H, Bloos F, Keys J, Martin C,
Neal A, Sibbald WJ (1994) Compared
with crystalloid, colloid therapy slows
progression of extrapulmonary tissue
injury in septic sheep. J Appl Physiol
77: 1507±1518
28. Schmand JF, Ayala A, Morrison MH,
Choudry IH (1995) Effects of hydroxyethyl starch after trauma-hemorrhagic
shock: Restoration of macrophage integrity and prevention of increased circulating IL-6 levels. Crit Care Med 23:
806±814
29. Hebert PC, Wells G, Blajchman MA,
Marshall J, Martin C, Pagliarello G,
Tweeddale M, Schweitzer I, Yetisir E
(1999) A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J
Med 340: 409±417
30. Marik PE, Sibbald WJ (1993) Effect of
stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA
269: 3024±3029
31. Desjars P, Pinaud M, Potel G, Tasseau
F, Touze M-D (1987) A reappraisal of
norepinephrine therapy in human septic shock. Crit Care Med 15: 134±137
32. Meadows D, Edwards JD, Wilkins RG,
Nightingale P (1988) Reversal of intractable septic shock with norepinephrine
therapy. Crit Care Med 16: 663±666
33. Desjars P, Pinaud M, Bugnon D, Tasseau F (1989) Norepinephrine therapy
has no deleterious renal effects in human septic shock. Crit Care Med 17:
426±429
34. Fukuoka T, Nishimura M, Imanaka H,
Taenaka N, Yoshiya I, Takezawa J
(1989) Effects of norepinephrine on renal function in septic patients with normal and elevated serum lactate levels.
Crit Care Med 17: 1104±1107
35. Hesselvik JF, Brodin B (1989) Low dose
norepinephrine in patients with septic
shock and oliguria: effects on afterload,
urine flow and oxygen transport. Crit
Care Med 17: 179±180
36. Bollaert PE, Bauer P, Audibert G, Lambert H, Larcan A (1990) Effects of epinephrine on hemodynamics and oxygen
metabolism in dopamine resistant
shock. Chest 98: 949±953
37. Martin C, Eon B, Saux P, Aknin P, Gouin F (1990) Renal effects of norepinephrine used to treat septic shock patients.
Crit Care Med 18: 282±285
38. Lipman J, Roux A, Kraus P (1991) Vasoconstrictor effects of adrenaline in human septic shock. Anaesth Intensive
Care 19: 61±65
39. Redl-Wenzl EM, Armbruster C, Edelmann G, Fischl E, Kolacny M, Wechsler-Fördös A, Sporn P (1993) The effects of norepinephrine on hemodynamics and renal function in severe septic shock states. Intensive Care Med 19:
151±154
40. Schaer GL, Fink MP, Parrillo JE (1985)
Norepinephrine alone versus norepinephrine plus low-dose dopamine: enhanced renal blood flow with combination pressor therapy. Crit Care Med 13:
492±496
41. Richer M, Robert S, Lebel M (1996)
Renal hemodynamics during norepinephrine and low-dose dopamine infusion in man. Crit Care Med 24:
1150±1156
42. Levy B, Bollaert PE, Charpentier C, et
al (1997) Comparison of norepinephrine and dobutamine to epinephrine
for hemodynamics, lactate metabolism,
and gastric tonometric variables in septic shock. Intensive Care Med 23:
282±287
43. Winslow EJ, Loeb HS, Rahimtoola SH,
Kamath S, Gunnar RM (1973) Hemodynamic studies and results of therapy
in 50 patients with bacteremic shock.
Am J Med 54: 421±432
44. Wilson RF, Sibbald WJ, Jaanimagi JL
(1976) Hemodynamic effects of dopamine in critically ill septic patients. J Surg
Res 20: 163±172
45. Regnier B, Rapin M, Gory G, Lemaire
F, Teisseire B, Harari A (1977) Haemodynamic effects of dopamine in septic
shock. Intensive Care Med 3: 47±53
46. Drueck C, Welch GW, Pruitt BA (1978)
Hemodynamic analysis of septic shock
in thermal injury: treatment with
dopamine. Am Surg 7: 424±427
47. Samii K, Le Gall JR, Regnier B, Gory
G, Rapin M (1978) Hemodynamic effects of dopamine in septic shock with
and without renal failure. Arch Surg
113: 1414±1416
48. Jardin F, Gurdjian F, Desfonds P, Margairaz A (1979) Effects of dopamine on
intrapulmonary shunt fraction and oxygen transport in severe sepsis with circulatory and respiratory failure. Crit
Care Med 7: 273±277
49. Regnier B, Safran D, Carlet J, Teisseire
B (1979) Comparative haemodynamic
effects of dopamine and dobutamine in
septic shock. Intensive Care Med 5:
115±120
50. Martin C, Papazian L, Perrin G, Saux P,
Gouin F (1993) Norepinephrine or
dopamine for the treatment of hyperdynamic septic shock? Chest 103:
1826±1831
51. Ruokonen E, Takala J, Kari A, Saxen
H, Mertsola J, Hansen EJ (1993) Regional blood flow and oxygen transport
in septic shock. Crit Care Med 21:
1296±1303
52. Hannemann L, Reinhart K, Grenzer O,
Meier-Hellmann A, Bredle DL (1995)
Comparison of dopamine to dobutamine and norepinephrine for oxygen
delivery and uptake in septic shock. Crit
Care Med 23: 1962±1970
53. Marik PE, Mohedin J (1994) The contrasting effects of dopamine and norepinephrine on systemic and splanchnic
oxygen utilization in hyperdynamic sepsis. JAMA 272: 1354±1357
54. Meier-Hellmann A, Bredle DL, Specht
M, Spies C, Hannemann L, Reinhart K
(1997) The effects of low-dose dopamine on splanchnic blood flow and oxygen uptake in patients with septic
shock. Intensive Care Med 23: 31±37
55. Nevi›re R, Mathieu D, Chagnon JL,
Lebleu N, Wattel F (1996) The contrasting effects of dobutamine and dopamine on gastric mucosal perfusion in septic patients. Am J Respir Crit Care
Med 154: 1684±1688
S 91
56. Bailey AR, Burchett KR (1997) Effect
of low-dose dopamine on serum concentrations of prolactin in critically ill
patients. Br J Anaesth 78: 97±99
57. Schreuder WO, Schneider AJ, Groeneveld ABJ, Thijs LG (1989) Effect of
dopamine vs norepinephrine on hemodynamics in septic shock. Chest 95:
1282±1288
58. Martin C, Saux P, Eon B, Aknin P, Gouin F (1990) Septic shock: a goal-directed
therapy using volume loading, dobutamine and/or norepinephrine. Acta
Anaesthesiol Scand 34: 413±417
59. Brown RS, Carey JS, Mohr PA, Monson DO, Shoemaker WC (1966) Comparative evaluation of sympathomimetic amines in clinical shock. Circulation
34: 260±271
60. McMillan M, Chernow B, Roth BL
(1986) Hepatic alpha 1-adrenergic receptor alteration in a rat model of
chronic sepsis. Circ Shock 19: 185±193
61. Martin C, Perrin G, Saux P, Papazian L,
Gouin F (1994) Effects of norepinephrine on right ventricular function in septic shock patients. Intensive Care Med
20: 444±447
62. Mills LC, Moyer JH (1960) The effects
of various catecholamines on specific
vascular hemodynamics in hypotensive
and normotensive subjects. Am J Cardiol 5: 652±659
63. Murakawa K, Kobayashi A (1988) Effects of vasopressors on renal tissue gas
tensions during hemorrhagic shock in
dogs. Crit Care Med 16: 789±792
64. Conger JD, Robinette JB, Guggenheim
SJ (1981) Effect of acetylcholine on the
early phase of reversible norepinephrine-induced acute renal failure. Kidney
Int 19: 399±403
65. Eckstein JW, Abboud FM (1962) Circulatory effects of sympathomimetic amines. Am Heart J 63: 119±135
66. Gombos EA, Hulet WH, Bopp P, Goldring W, Baldwin DS, Chasis H (1962)
Reactivity of renal and systemic circulations to vasoconstrictor agents in normotensive and hypertensive subjects.
J Clin Invest 41: 203±217
67. Reinelt H, Radermacher P, Fischer G,
Geisser W, Wachter U, Wiedeck H,
Georgieff M, Vogt J (1997) Effects of a
dobutamine-induced
increase
in
splanchnic blood flow on hepatic metabolic activity in patients with septic
shock. Anesthesiology 86: 818±824
68. Mackenzie SJ, Kapadia F, Nimmo GR,
Armstrong IR, Grant IS (1991) Adrenaline in treatment of septic shock: effects on hemodynamics and oxygen
transport. Intensive Care Med 17: 36±39
69. Wilson W, Lipman J, Scribante J, Kobilski S, Lee C, Krause P, Cooper J, Barr J
(1992) Septic shock: does adrenaline
have a role as a first-line inotropic
agent? Anaesth Intensive Care 20:
470±474
70. Moran JL, O'Fathartaigh MS, Peisach
AR, Chapman MJ, Leppart P (1993)
Epinephrine as an inotropic agent in
septic shock: a dose-profile analysis.
Crit Care Med 21: 70±77
71. Meier-Hellmann A, Reinhart K, Bredle
DL, Specht M, Spies CD, Hannemann
L (1997) Epinephrine impairs splanchnic perfusion in septic shock. Crit Care
Med 25: 399±404
72. Day NPJ, Phu NH, Bethell DP, Mai
NTH, Chau TTH, Hien TT, White NJ
(1996) The effects of dopamine and
adrenaline infusion on acid-base balance and systemic hemodynamics in severe infection. Lancet 348: 219±223
73. Gregory JS, Bonfiglio MF, Dasta JF,
Reilley TE, Townsend MC, Flancbaum
L (1991) Experience with phenylephrine as a component of the pharmacologic support of septic shock. Crit Care
Med 19: 1395±1400
74. Flancbaum L, Dick M, Dasta J, Sinha
R, Choban P (1997) A dose-response
study of phenylephrine in critically ill,
septic surgical patients. Eur J Clin Pharmacol 51: 461±465
75. Reinelt H, Radermacher P, Kiefer P,
Fischer G, Wachter U, Vogt J, Georgieff M (1999) Impact of exogenousadrenorecptor stimulation on hepatosplanchnic oxygen kinetics and metabolic activity in septic shock. Crit Care
Med 27: 325±331
76. Vincent JL, Gris P, Coffernils M, Leon
M, Pinsky MR, Reuse C, Kahn RJ
(1992) Myocardial depression characterizes the fatal course of septic shock.
Surgery 111: 660±667
77. Vincent JL, Reuse C, Frank RN, Contempre B, Kahn RJ (1989) Right ventricular dysfunction in septic shock: assessment by measurements of right ventricular ejection fraction using the thermodilution technique. Acta Anaesthesiol Scand 33: 34±38
78. Parker MM, McCarthy KE, Ognibene
FP, Parrillo JE (1990) Right ventricular
dysfunction and dilatation, similar to
left ventricular changes, characterize
the cardiac depression of septic shock
in humans. Chest 97: 126±131
79. Kumar A, Thota V, Dee L, Olson J, Uretz E, Parrillo JE (1996) Tumor necrosis
factor and interleukin Ib are responsible for in vitro myocardial depression
induced by human septic shock serum.
J Exp Med 183: 949±958
80. Kumar A, Brar R, Wang P, Dee L, Skorupa G, Khadour F, Schulz R, Parrillo
JE (1999) Role of nitric oxide and
cGMP in human septic serum-induced
depression of cardiac myocyte contractility. Am J Physiol 276:R265±R276
81. Bakker J, Gris P, Coffernils M, Kahn
RJ, Vincent JL (1996) Serial blood lactate levels can predict the development
of multiple organ failure following septic shock. Am J Surg 171: 221±226
82. Jardin F, Sportiche M, Bazin M, Bourokba A, Margairaz A (1981) Dobutamine: a hemodynamic evaluation in
human septic shock. Crit Care Med 9:
329±332
83. Fisher CJJ, Horowitz Z, Albertson TE
(1985) Cardiorespiratory failure in toxic
shock syndrome: effect of dobutamine.
Crit Care Med 13: 160±165
84. Tell B, Majerus TC, Flancbaum L
(1987) Dobutamine in elderly septic
shock patients refractory to dopamine.
Intensive Care Med 13: 14±18
85. Shoemaker WC, Appel PL, Kram HB,
et al (1989) Comparison of hemodynamic and oxygen transport effects of
dopamine and dobutamine in critically
ill surgical patients. Chest 96: 120±126
86. Vincent JL, Roman A, Kahn RJ (1990)
Dobutamine administration in septic
shock: addition to a standard protocol.
Crit Care Med 18: 689±693
87. Colardyn FC, Vandenbogaerde JF, Vogelaers DP, Verbeke JH (1989) Use of
dopexamine hydrochloride in patients
with septic shock. Crit Care Med 17:
999±1003
88. Smithies M, Yee TH, Jacson L, Beale R,
Bihari D (1994) Protecting the gut and
the liver in the critically ill: effects of
dopexamine. Crit Care Med 22:
789±795
89. Hannemann L, Reinhart K, MeierHellmann A, et al (1996) Dopexamine
hydrochloride in septic shock. Chest
109: 756±760
90. Meier-Hellmann A, Bredle DL, Specht
M, Hannemann L, Reinhart K (1999)
Dopexamine
increases
splanchnic
blood flow but decreases gastric mucosal pH in severe septic patients treated
with dobutamine. Crit Care Med 27:
2166±2171
91. Barton P, Garcia J, Kouatli A, Kitchen
L, Zorka A, Lindsay C, Lawless S
(1996) Hemodynamic effects of IV milrinone lactate in pediatric patients with
septic shock. Chest 109: 1302±1312
92. Woo P, Carpenter MA, Trunkey D
(1979) Ionized calcium: the effect of
septic shock in the human. J Surg Res
26: 605±610
S 92
93. Malcolm DS, Zaloga GP, Holaday JW
(1989) Calcium administration increases the mortality of endotoxic shock in
rats. Crit Care Med 17: 900±903
94. Zaloga GP, Sager A, Black KW, Prielipp R (1992) Low dose calcium administration increases mortality during septic peritonitis in rats. Circ Shock 37:
226±229
95. Nasraway SA, Rackow EC, Astiz ME,
Karras G, Weil MH (1989) Inotropic response to digoxin and dopamine in patients with severe sepsis, cardiac failure,
and systemic hypoperfusion. Chest 95:
612±615
96. Shoemaker WC, Appel PL, Kram HB,
Waxman K, Lee TS (1988) Prospective
trial of supranormal values of survivors
as therapeutic goals in high-risk surgical
patients. Chest 94: 1176±1186
97. Hayes MA, Timmins AC, Yau EH, Palazzo M, Hinds CJ, Watson D (1994) Elevation of systemic oxygen delivery in
the treatment of critically ill patients.
N Engl J Med 330: 1717±1722
98. Gattinoni L, Brazzi L, Pelosi P, Latini
R, Tognoni G, Pesenti A, Fumagalli R
(1995) A trial of goal-oriented hemodynamic therapy in critically ill patients.
N Engl J Med 333: 1025±1032
99. Tuchschmidt J, Fried J, Astiz M, Rackow E (1992) Elevation of cardiac output and oxygen delivery improves outcome in septic shock. Chest 102:
216±220
Intensive Care Med (2001) 27: S 93±S 103
J. Carlet
Immunological therapy in sepsis:
currently available
)
J. Carlet ( )
Intensive Care Unit, Fondation-Hôpital Saint-Joseph, Paris,
France
E-mail: [email protected]
Phone: +33-1 44-12 37 83
Fax: +33-1 44-12 34 15
Should corticosteroids be used during septic shock at
low doses and for a prolonged period of time?
Answer: yes, grade C.
Recommendations
Introduction
Many therapies used in our daily practice are known to
have significant effects on inflammation. These drugs
influence the activation of the inflammatory network
that occurs during severe sepsis and related syndromes
as disseminated intravascular coagulation and acute respiratory distress syndrome (ARDS). Many of these
compounds (Table 1) have already been used during experimental models of sepsis and/or human studies.
Corticosteroids should not be used in severe sepsis or
septic shock at high doses (30 mg/kg) and for a short
course (1±2 days). On the other hand, corticosteroids
may be used during ªrefractoryº septic shock but not
during severe sepsis without shock or mild shock. It
should then be used at low doses (100 mg hydrocortisone three times a day) for 5 days or more and then
with subsequent tapering of the dose according to the
hemodynamic status and the need for vasopressors.
Rationale
Methods
This contribution reviews those drugs that are available in the daily
management of severe sepsis and septic shock. A computer-based
review of the literature was undertaken using Medline from 1990
to September 1999 as the primary database. The subject heading
keywords defined for each of the compounds listed in Table 1
were combined with the following general sepsis-related subject
heading keywords: sepsis, severe sepsis, septic shock, and ARDS.
Anti-inflammatory agents
Should corticosteroids be used in the treatment of
severe sepsis or septic shock at high doses (30 mg/kg)
for a short course (one or 2 days)?
Answer: no, grade A.
An extensive literature is available for corticosteroids.
Steroids have been used for many years, and their efficacy is controversial. Numerous animal studies performed during experimental septic (endotoxic) shock or
acute lung injuries showed a very significant reduction
in both intensity of shock, acute respiratory failure and
mortality [1, 2]. They have been used at very high doses
(30 mg/kg per dose for a maximum of 24±48 h). The
ability of these high doses of corticosteroids to reduce
complement activation and to inhibit leukocyte aggregability and adherence was at that time a very logical rationale for their efficacy [3]. Very promising initial findings have been published regarding humans [12]. However, two well designed, prospective, multicenter, randomized, double-blind studies demonstrated very clearly their inability to decrease mortality [5, 8]. Some studies mention positive trends when looking at subgroups
of infections due to Gram-negative rods [5, 8, 13].
S 94
Table 1 List of therapies currently available for eventually treating severe sepsis
Therapy
References
Anti-inflammatory agents
Corticosteroids (high or low doses)
Ibuprofen
Prostaglandin E1
Pentoxifylline
Oxygen scavengers
N-Acetylcysteine
Selenium
Drugs modifying coagulation
Antithrombin III
Drugs enhancing host defenses
Immunoglobulins
Interferon-g
Granulocytes stimulating factors
Immunonutrition
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26,
27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 25a
27, 28, 29, 30, 31
32, 33, 34, 35
36, 37, 38
39, 40, 41, 42, 43, 44, 45, 46, 47
39, 40, 41, 42, 43, 44, 45
46, 47
48, 49, 50, 51, 52
48, 49, 50, 51, 52
53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70
54, 55, 56, 57, 58
59, 60, 61, 62, 63, 64
65, 66, 67, 68, 69, 70
±a
Other drugs
Growth hormone
Antibiotics
Including ketoconazole
Including polymyxin B
Taurolidine
Fresh frozen plasma
Anesthetic sedative and analgesic agents
Catecholamines
71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85
71
72, 73, 74, 75, 76, 77
73, 74
72
78
79
80
81, 82, 83, 84, 85
Hemofiltration, plasma filtration, plasma exchange
86, 87, 88, 89, 90, 91
a
See PØrez and Dellinger, ªOther supportive therapies in sepsisº
Two recent meta-analyses [13, 14] reviewing the studies confirm that corticosteroids at the dose of 30 mg/kg
(one or two doses) are ineffective [13] or even harmful
[14]. The design and the results of the nine randomized
studies are summarized in Tables 2 and 3. Similar negative results have been obtained during ARDS [15].
Pooling the results only from those patients with
Gram-negative infections, as in the meta-analysis by Lefering et al. [13], yields a rate difference of ±5.6 % [confidence interval (CI): ±21.4 to 10.1) in favor of steroids,
based on 413 patients. Those patients with Gram-positive infections (n = 306) had an overall effect of +1.8 %
(CI: ±15.8 to 18.6). Most persons stopped using steroids
when these large trials were published.
Several studies performed over the years, however,
have maintained interest in the use of corticosteroids.
Mortality was reduced using steroids during severe typhoid fever [16], and neurological sequelae were reduced during meningitis [17]. Two large double-blind case
control studies demonstrated that prolonged treatment
(10±15 days) of relatively low doses of steroids
(120±240 mg hydrocortisone) dramatically reduced
mortality during severe Pneumocystis carinii pneumonia in AIDS patients [18, 19]. In addition, Meduri et al.
[20] showed that the course of late, fibrotic ARDS was
improved by steroid use which was confirmed in a re-
cent randomized double-blind study showing a significant reduction in mortality [21]. Two small, randomized,
double-blind studies of steroids in patients with severe
and refractory septic shock recently demonstrated positive results [22, 23]. Corticosteroids were used at small
doses (100 mg hydrocortisone three times per day in
one [22] and 100 mg followed by a continuous infusion
of 0.18 mg/kg per hour in the other [23], for longer periods of time than in past studies: 5 days in one [22] and
5±10 days in the other, with tapering of the doses according to hemodynamic status and need for vasopressors.
Both studies showed a significant reversal of shock and
organ failures and a trend in reduction in mortality. Additional studies are necessary, and a French multicenter
randomized, controlled, double-blind study reported
that low dose steroids decrease mortality in patients
with septic shock [25 a].
Several factors may explain the recent positive effects of corticosteroids during sepsis [24]. These include
the treatment of ªrelativeº adrenal insufficiency [25]
and the potentiation of adrenergic receptivity [26] in addition to the anti-inflammatory effect. The lower immunosuppressive doses and a more prolonged duration of
therapy than in the initial studies could also explain discrepancies.
S 95
Table 2 Design of the nine randomized studies used in the meta-analysis (from Cronin et al. [14]) (DB double blind, M methylprednisolone, B betamethasone, D dexamethasone, H hydrocortisone)
Reference
n
Type of study
Product
Dose
Duration
Endpoints
Cooperative
Study Group [6]
Klastersky
et al. [10]
Schumer et al. [12]
194
Open
H
6d
85
Open
B
300 mg then
50 mg/d
1 mg/kg
172
DB
M
30 mg/kg
1 dose or 2
Thompson
et al. [11]
Lucas and Ledgerwood [9]
Sprung et al. [4]
60
DB
M
30 mg/kg
48
Open
D
2 mg/kg
Max. 6 doses
in 24 h
2d
59
Open
M
30 mg/kg
1 dose (or 2)
Mortality, complications
Mortality (20 d),
complications
Mortality (28 d),
complications
Mortality, complications
Mortality (14 d),
complications
Hospital mortality,
complications
3d
Bone et al. [5]
381
DB
M
30 mg/kg
1d
Mortality (14 d),
complications
Veteran Administration [8]
223
DB
M
30 mg/kg
9d
Mortality (14 d),
complications
75
DB
M
30 mg/kg”4
1d
Hospital mortality,
ARDS complications
Luce et al. [7]
Table 3 Results of the nine randomized studies used in the metaanalysis (from Cronin et al. [14])
Cooperative study group [6]
Klastersky et al. [10]
Schumer et al. [12]
Thompson et al. [11]
Lucas et al. [9]
Sprung et al. [4]
Bone et al. [5]
Veteran Administration [8]
Luce et al. [7]
n
Risk ratio
95 % CI
194
85
172
60
48
59
381
223
75
1.72
0.97
0.30
1.01
1.09
1.11
1.35
0.95
1.07
1.23±2.41
0.65±1.45
0.13±0.72
0.77±1.31
0.36±3.27
0.74±1.67
0.98±1.84
0.57±1.58
0.72±1.60
Should ibuprofen be used in the treatment of severe
sepsis and septic shock?
Answer: no, grade B.
Recommendations
Ibuprofen should not be used during severe sepsis or
septic shock. Additional studies are needed to determine whether some patients, for example, those with
hypothermia, could benefit from the drug.
Rationale
Ibuprofen is a powerful anti-inflammatory agent, acting
on the prostaglandin metabolism as a cyclo-oxygenase
inhibitor. It has been used with controversial effects in
animals during both experimental sepsis and ARDS
[27, 28]. Two small randomized, double-blind studies in
patients showed some hemodynamic effect and a normalization of pH without any significant effect upon
mortality [29, 30]. Mortality was decreased significantly
in a post hoc analysis of hypothermic patients [30]. A
large multicenter randomized, controlled, double blind
study, however, failed to demonstrate any effect upon
mortality, reversal of shock or acute respiratory failure
[31]. Ibuprofen was able to reduce the levels of prostacyclin and thromboxane and to decrease fever, tachycardia and oxygen consumption [31]. The drug was not associated with adverse affects.
Should prostaglandins be used in the treatment of
ARDS due to severe infections and sepsis?
Answer: no, grade B.
Recommendations
Prostaglandins, in particular prostaglandin E1 or liposomal prostaglandin E1 should not be used during ARDS
due to sepsis. There are no specific data allowing recommendations in severe sepsis.
S 96
Rationale
Oxygen scavengers
Several prostaglandins which have both an anti-inflammatory and a vasoactive effect have been studied including prostaglandin I2 and particularly prostaglandin
E1 [32, 33, 34, 35] during ARDS. The vast majority of
these patients had ARDS due to severe infections or
sepsis. An early, small, randomized study showed promising results [32]. However, a large multicenter, randomized, controlled, double blind study failed to show any
difference in survival [33]. An increase in oxygen delivery and oxygen consumption was noted in treated patients who survived [34]. A recent, multicenter randomized, controlled, double-blind study with liposomal
prostaglandin E1 (TLC C-53) showed that indices of oxygenation of treated ARDS patients were improved
compared with controls, but without any effect upon duration of mechanical ventilation or 28 days mortality
[35]. Again, most ARDS was due to sepsis in these two
large studies. No data are really available concerning
an overall group of patients with severe sepsis.
Several oxygen scavengers are currently available, including N-acetylcysteine (NAC), vitamin E, vitamin C,
and selenium. Vitamins E and C have been only poorly
studied in humans, and we focus on N-acetylcysteine
and selenium.
Should pentoxifylline be used in the treatment of severe
sepsis in (a) adults, (b) infants?
Answer: (a) no, grade B; (b) no, grade C.
Recommendations
Pentoxifylline should not be used in adults with severe
sepsis unless new studies show a significant effect. The
positive effect of a small study in infants should be confirmed before clinical use.
Rationale
Pentoxifylline, which has a powerful anti-inflammatory
effect including a strong inhibition of tumor necrosis
factor secretion, has been used successfully in many animal studies with the prevention of the transition from a
hyperdynamic to hypodynamic state, although no effect
upon mortality has been shown [36]. Human studies are
more scarce. A multicenter, randomized, controlled,
double-blind study during sepsis showed an increase in
PaO2/FIO2 ratio but no effect upon cytokines levels or
mortality [37]. A recent double-blind study performed
in premature infants with sepsis showed a decrease in
cytokines levels and a significant decrease in mortality
(1/40 vs. 6/38 p = 0.046) [38]. However, the size of this
study was rather small, and additional large studies are
mandatory.
Should N-acetylcysteine be used in the treatment of
severe sepsis?
Answer: no, grade C.
Recommendations
NAC should not be used in severe sepsis until new data
are available, focusing in particular on very early therapy.
Rationale
During acute lung injury an improvement in oxygenation and reduction in the required length of mechanical
ventilation was found in patients treated with NAC
compared to controls [39]. However, several randomized studies have shown no difference in mortality, gas
exchange, and development of respiratory failure in patients treated with NAC [39, 40]. Several studies have
also been performed during severe sepsis, with heterogeneous results [41, 42, 43, 44]. Depressed cardiac performance has been described in septic patients treated
with NAC [42]. A very recent multicenter, randomized,
controlled, double-blind study showed that a prolonged
infusion of NAC is unable to prevent multiple organ
failure in consecutively admitted critically ill patients
[43]. In this study treatment used more than 24 h after
the initial insult worsened the prognosis compared to
controls. Better results were obtained when the drug
was use before the insult, as during cardiac surgery
[44]. These results suggest that this compound could be
helpful when started before (or perhaps shortly after)
the insult, but possibly harmful when started too late.
Combinations of several antioxidants have also been
published, but data are too limited to allow recommendations [45].
Should selenium be used in the treatment of severe
sepsis?
Answer: no, grade C.
S 97
Fig. 1 Effect of proinflammatory cytokines. Upon coagulation cascade during sepsis leading to an activation of tissue factor, a depletion in protein C (via a decrease in thrombomodulin levels) antithrombin III and C1 inhibitor, and a decrease in fibrinolysis (via
the effect of plasminogen activator inhibitor 1)
Recommendations
Selenium should not be used for severe sepsis. Additional studies are warranted to confirm initial positive
data.
Rationale
A profound depletion in selenium levels has been demonstrated in many severe septic patients [46]. Mortality
and morbidity are far higher in patients with a very low
selenium level [46]. A recent prospective, randomized,
but nonblinded study performed in septic patients
showed that selenium replacement is able to reduce severity indexes at day 3 and reduce the need for hemodialysis but has no significant effect upon mortality (52 %
in controls and 33, 5 % in treated patients, p = 0.13)
[47]. Additional large studies are needed to confirm initial promising results.
Drugs modifying coagulation
There are complex interactions between the inflammation and coagulation systems (Fig. 1). Proinflammatory
cytokines activate coagulation cascades, in particular
via an effect upon tissue factor which is a key player in
the coagulation cascade. They can also reduce fibrinolysis and profoundly reduce the levels of protein C and of
antithrombin III which are important anticoagulant
agents. Antithrombin III inhibits several coagulation
factors of the extrinsic pathway such as factors IXa,
XIa, XIIa in addition to factors Xa, IIa, and plasmin.
Activated protein C inhibits factors Va, Vlla, and plasminogen activator inhibitor 1. The overall effect during
sepsis is a marked procoagulant balance. Conversely,
coagulation products can activate the inflammation network which creates numerous amplification loops. For
example, thrombin can induce an up-regulation of Pand E-selectin, and contact factor activation can induce
the production of bradykinin, worsening hypotension
and tissue hypoperfusion. In humans studies, both anti-
S 98
thrombin III and protein C levels are sharply decreased
[48], and mortality of septic patients is inversely correlated with the levels of those two products. This makes
the rationale for studying those types of compounds,
such as antithrombin III, protein C, and tissue factor
protein inhibitor very strong. Only antithrombin III is
currently available.
Should antithrombin III be used in the treatment of
severe sepsis?
Answer: no, grade B.
Recommendations
Antithrombin III should not be used during severe sepsis. Countries which allow the free use of this drug in
this setting should reconsider their position.
Rationale
Antithrombin III is a drug which is widely used for septic patients in several countries. Three randomized,
small, double-blind studies were published [49, 50, 51].
Duration of disseminated intravascular coagulation
was reduced [49] as well as the number of organ failures
[51], but mortality was not different although a positive
trend was clearly noted. A meta-analysis was also performed [51] showing a 22.9 % reduction in mortality
but which did not reach statistical significance. Unfortunately a large multicenter, prospective, double-blind
study has recently been completed which showed no significant improvement in survival [52]. The complete
data have not yet been published. Other drugs such as
activated protein C and tissue factor inhibitors are not
currently available and are discussed elsewhere (see
Arndt and Abraham, ªImmunological therapy of sepsis:
experimental therapiesº).
Drugs enhancing host defenses
After the initial activation of the proinflammatory network, a profound immunodepression can occur in septic
patients [53]. This could influence outcome increasing
the risk of nosocomial infections. Several strategies
have been used to increase host defenses, including
polyvalent immunoglobulins, interferon-g, stimulating
factors for granulocytes [including granulocyte colony
stimulating factor (G-CSF)], and immunonutrition.
The latter is discussed elsewhere (see PØrez and Dellinger, ªOther supportive therapy in sepsisº).
Should intravenous immunoglobulins be used in the
treatment of severe sepsis in (a) adults or (b) neonates?
Answer: (a) no, grade C; (b) no, grade C.
Recommendations
Immunoglobulins should not be used either in adult patients or in neonates with sepsis, unless additional large
studies confirm some positive data in small-sized metaanalyses. Countries which allow a wide use of these
compounds should reconsider their position and encourage these studies.
Rationale
Intravenous immunoglobulins (IVIG) are widely used
in both infants and adults in the treatment of severe sepsis, at least in certain countries. Reports which support
their empirical use, however, are still rather weak. The
rationale is to restore immunoglobulins levels, which
may be depressed in sepsis, and to provide patients
with specific antibodies against micro-organisms. No individual well designed clinical study has been performed in adults with severe sepsis. A recent study was performed in patients with streptococcal toxic shock syndrome [54]. This was a comparative nonblinded study
performed in 21 patients which demonstrated a significantly reduced mortality (67 % vs. 34 %, p = 0.02).
Both Acute Physiology and Chronic Health Evaluation
II scores and IVIG were prognostic factors in the multivariate analysis. The odds ratio associated with IVIG
was 8.1 (95 % CI: 1.6±45). A recent meta-analysis by
the Cochrane group [55] looking at 23 studies (some of
them unpublished) on immunoglobulins, antiendotoxins, and anticytokines, extracted from the small size
studies already published, evaluated a population of
413 patients receiving polyclonal immunoglobulins.
Mortality was significantly reduced (relative risk: 0.6;
95 % CI: 0.47±0.76). Results were even more positive
when only sepsis related deaths were considered. A
large, well designed, multicenter, randomized, doubleblind study is, however, warranted before making firm
conclusions. Two prophylactic studies have been published recently [56, 57]. A study performed in cardiac
surgery patients showed no difference in the occurrence
of sepsis between polyvalent IVIG and IgM-enriched
immunoglobulin [56]. A prospective comparative study
showed that IVIG and not placebo is able to prevent
nosocomial infections after major surgery [57]. Such
prophylactic studies are needed in this field in nonsurgical critically ill patients.
In neonatal sepsis, a recent meta-analysis of 110 newborns in three studies showed that IVIG is able to re-
S 99
duce mortality significantly (odds ratio: 0.173; 95 % CI:
0.031±0.735; p = 0.007) [58]. However, the size of the
overall population was very small, and large studies are
urgently warranted. In the same meta-analysis the effect
of IVIG in the prevention of sepsis in 4933 evaluable
newborns was significant (p = 0.0193, two-tailed), although heterogeneity of the studies precluded estimation of an overall odds ratio.
Other Drugs
Interferon-g
Growth hormone should not be used in patients with
sepsis because it increases mortality.
Interferon-g has been used successfully in animals models of Gram-negative sepsis [59, 60]. Few data are available in human sepsis. The drug has been used with positive results to prevent infection during chronic granulomatous disease [61] and trauma [62, 63]. The drug, however, was unable to prevent infections in burn patients
[64]. Data are insufficient for therapy of severe sepsis
to allow recommendations.
Should growth hormone be used in the treatment of
severe sepsis?
Answer: no, grade A.
Recommendations
Rationale
Answer: no, grade C.
The administration of growth hormone could in theory
attenuate the catabolic response to injury, surgery or
sepsis. Two prospective double-blind studies with more
than 200 patients each were recently reported in critically ill patients with cardiac or abdominal surgery, multiple trauma or acute respiratory failure [71]. Mortality
was increased significantly in treated patients. The relative risk in these two pooled studies was 1.9 (95 % CI:
1.3±2.9). Length of stay and duration of mechanical ventilation were longer in treated survivors than in controls.
Recommendations
Antimicrobial compounds
G-CSF should not be used in nonneutropenic patients
with severe sepsis.
Polymixin B. Polymixin B is able to neutralize endotoxin via strong antilipid A activity [72]. Since it is very toxic, it is difficult to use intravenously in humans, although some derivates are less toxic. Extracorporeal
techniques, in which polymyxin is coated on membranes, are under investigation.
Should granulocyte colony stimulating factor be used in
the treatment of severe infections?
Rationale
G-CSF is very efficient and reduces mortality in animal
models of abdominal sepsis [65, 66]. During pneumonia
models in rats the drug has been shown to exert different effects according to the micro-organisms involved
[67]. Preliminary studies have been performed in community or hospital acquired pneumonia with controversial results [68, 69]. In patients with head trauma and receiving mechanical ventilation G-CSF prophylaxis did
not improve outcome nor lower the risk of nosocomial
pneumonia [70].
Ketoconazole. Ketaconazole, one of the new imidazoles, has a strong effect upon thromboxane synthase inhibition and has been shown to prevent ARDS in septic
patients in a small double-blind randomized study [73].
A recent study performed in 234 patients, however,
failed to demonstrate any effect upon mortality and
duration of mechanical ventilation in ARDS patients
[74]. No data are available in patients with sepsis.
Other antibiotics
Immunonutrition
See PØrez and Dellinger, ªOther supportive therapies in
sepsis.º
Some antibiotics have anti-inflammatory effects, in particular in decreasing cytokine release. Effects have
been shown for vancomycin [75] trovafloxacin [76] and
ciprofloxacin [77].
S 100
Taurolidine
Hemofiltration and plasma filtration
Taurolidine is an anti-infective agent (nonantibiotic),
used either locally, or intravenously, which has some antibacterial effect associated with an antiendotoxin effect. A randomized placebo-controlled study failed to
demonstrate any effect on outcome in sepsis [78].
Should hemofiltration be used in the treatment of
patients with severe sepsis, without renal indications?
Answer: no, grade C.
Recommendations
Other drugs currently used
Many other drugs that we use daily could have important effects upon inflammation, including heparin, fresh
frozen plasma [79], and anesthetic, sedative, and analgesic agents [80]. A recent review [80] describes the potential effects of these agents upon immunomodulation.
Catecholamines and inflammation. It is well known that
inotropic agents such as catecholamines have a significant impact upon inflammation [81]. Epinephrine inhibits tumor necrosis factor and potentiates interleukin10 leading to a significant anti-inflammatory effect [82],
via an effect upon macrophages [83]. Dopamine increases interleukin-6 release but decreases tumor necrosis
factor [84]. Recent data support the concept that the
anti-inflammatory effect of catecholamines explains
the possible beneficial effects of supranormal oxygen
delivery in critically ill surgical patients [85]. These
data do not enable clinicians to take into account the effect of catecholamines upon inflammation in deciding
which is the best to use.
Hemofiltration should not be used in patients with sepsis without renal indications unless ongoing studies provide positive results.
Rationale
Hemofiltration has been shown to decrease cytokines
levels significantly, although temporarily during severe
sepsis in humans. The technique is widely used in Europe and many authors have strong opinions [86] regarding its use, although the data are weak. A randomized, still unpublished study found no effect upon mortality [87]. Another randomized controlled study [88]
reported a 15 % (nonsignificant) increase in survival
for filtrated patients. Favorable results have been described for cardiac surgery patients [89]. Large multicenter studies are currently under way.
Plasma filtration induced a significant attenuation of
acute-phase response in a randomized, prospective
study recently performed in 22 adults with sepsis [90].
However, no difference in mortality and only a trend toward fewer organ failures were noted. Plasma exchange
has also been used in severe meningococcemia in children [91] with varying results.
References
1. Hinshaw LB, Archer LT, Beller Todd
BK, et al (1981) Survival in primates in
lethal septic shock following delayed
treatment with steroid. Circ Shock 8:
291±300
2. Hinshaw LB, Beller Todd BK, Archer
LT, et al (1979) Recovery from lethal
Escherichia coli shock in dogs. Surg Gynecol Obstet 149: 545±553
3. Hammerschmidt DE, White JG, Craddock PR, et al (1979) Corticosteroids
inhibit complement-induced granulocyte aggregation: a possible mechanism
for their efficacy in shock states. J Clin
Invest 63: 798±803
4. Sprung CL, Caralis PV, Marcial EH, et
al (1984) The effect of high-dose corticosteroids in patients with septic shock:
a prospective, controlled study. N Engl
J Med 311: 1137±1143
5. Bone RC, Fisher CJ Jr, Clemmer TP, et
al (1987) A controlled clinical trial of
high-dose methylprednisolone in the
treatment of severe sepsis and septic
shock. N Engl J Med 317: 653±658
6. Cooperative Study Group, Bennet IL,
Finland M, et al (1963) The effectiveness of hydrocortisone in the management of severe infections. JAMA 183:
462±465
7. Luce JM, Montgomery AB, Marks JD,
et al (1988) Ineffectiveness of highdose methylprednisolone in preventing
parenchymal lung injury and improving
mortality in patients with septic shock.
Am J Respir Crit Care Med 138: 62±68
8. The Veterans Administration Systemic
Sepsis Cooperative Study Group
(1987) Effect of high-dose glucocorticosteroid therapy on mortality in patients with clinical signs of systemic sepsis. N Engl J Med 317: 659±665
9. Lucas C, Ledgerwood A (1984) The
cardiopulmonary response to massive
doses of steroids in patients with septic
shock. Arch Surg 119: 537±541
10. Klastersky J, Cappel R, Debusscher L
(1971) Effectiveness of betamethasone
in management of severe infections.
N Engl J Med 284: 1248±1250
11. Thompson WL, Gurley H, Lutz B, et al
(1976) Inefficacy of corticosteroids in
shock. Clin Res 24: 258A
12. Schumer W (1976) Steroids in the treatment of clinical septic shock. Ann Surg
184: 333±341
S 101
13. Lefering R, Neugebauer EAM (1995)
Steroid controversy in sepsis and septic
shock: a meta-analysis. Crit Care Med
23: 1294±1303
14. Cronin L, Cook DJ, Carlet J, et al
(1995) Corticosteroid treatment for
sepsis: a critical appraisal and metaanalysis of the literature. Crit Care
Med 23: 1430±1439
15. Bernard GR, Luce JM, Sprung CL, et al
(1987) High-dose corticosteroids in patients with the adult respiratory distress syndrome. N Engl J Med 317:
1565±1570
16. Hoffman SL, Punjabi NH, Kumala S, et
al (1984) Reduction in mortality in
chloramphenicol treated severe typhoid
fever by high-dose dexamethasone.
N Engl J Med 310: 82±88
17. Lebel MH, Freij BJ, Syrogainnopolous
GA, et al (1988) Dexamethasone therapy for bacterial meningitis: results of
two double-blind placebo-controlled
trials. N Engl J Med 319: 964±971
18. Montaner JS, Lawson LM, Levitt N, et
al (1990) Corticosteroids prevent early
deterioration in patients with moderate
severe Pneumocystis carinii pneumonia
and the acquired immunodeficiency
syndrome. Ann Intern Med 113: 14±20
19. Gagnon S, Boota AM, Fischl MA, et al
(1990) Corticosteroids as adjunctive
therapy for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. N Engl J Med 323:
1440±1450
20. Meduri U, Chinn AJ, Leeper KV, et al
(1994) Corticosteroid rescue treatment
of progressive fibroproliferation in
late ARDS: patterns of response and
predictors of outcome. Chest 105:
1516±1527
21. Meduri GU, Headley ASH, Golden E,
et al (1998) Effect of prolonged methylprednisolone therapy in unresolving
acute respiratory distress syndrome.
JAMA 280: 159±165
22. Bollaert PE, Charpentier C, Levy S, et
al (1998) Reversal of late septic shock
with supraphysiologic doses of hydrocortisone. Crit Care Med 26: 645±650
23. Briegel J, Forst H, Haller M, et al
(1999) Stress doses of hydrocortisone
reverse hyperdynamic septic shock: a
prospective, randomized, double-blind,
single center study. Crit Care Med 27:
723±732
24. Carlet J (1999) From mega to more reasonable doses of corticosteroids: a decade to recreate hope. Crit Care Med
27: 672±674
25. Hatherill M, Tibby SM, Hilliard T, et al
(1999) Adrenal insufficiency in septic
shock. Arch Dis Child 80: 51±55
25 a. Annane D (2001) Effects of combination hydrocortisone (HC) ± fludro-cortisone (FC) on mortality in septic shock.
Crit Care Med (in press)
26. Davies AO, Lefkowitz RJ (1980) Corticosteroids induced differential regulation of beta-adrenergic receptors in circulating human polymorphonuclear
leukocytes and mononuclear leukocytes. J Clin Endocrinol Metab 51:
599±605
27. Snapper JR, Hutchison AA, Obletree
ML, Brigham KL (1983) Effects of cyclooxygenase inhibitors on the alterations of lung mechanics caused by endotoxemia in unanesthetized sheep.
J Clin Invest 72: 63±76
28. Elinger JH, Seyde WC, Longnecker DE
(1984) Methylprednisolone plus ibuprofen increase mortality in septic rats.
Circ Shock 14: 203±208
29. Haupt MT, Jastremski MS, Clemmer
TP, et al (1991) Effect of ibuprofen in
patients with severe sepsis: a randomized, double-blind, multicenter study.
Crit Care Med 19: 1339±1347
30. Arons MM, Wheeler AP, Bernard GR,
et al (1999) Effects of ibuprofen on the
physiology and survival of hypothermic
sepsis. Crit Care Med 27: 699±707
31. Bernard GR, Wheeler AP, Russell JA,
et al (1997) The effects of Ibuprofen on
the physiology and survival of patients
with sepsis. N Engl J Med 336: 912±918
32. Holcroft JW, Vassar MJ, Weber CJ
(1986) Prostaglandin E1 and survival in
patients with the adult respiratory distress syndrome. A prospective trial.
Ann Surg 203: 371±378
33. Bone RC, Slotman G, Maunder R, et al
(1989) Randomized double-blind, multicenter study of prostaglandin E1 in patients with the adult respiratory distress
syndrome. Chest 96: 114±119
34. Silverman JH, Slotman G, Bone RC, et
al (1990) Effects of prostaglandin E1
on oxygen delivery and consumption in
patients with the adult respiratory distress syndrome. Results from the prostaglandin E1 multicenter trial. The
Prostaglandin E1 Study group. Chest
98: 405±410
35. Abraham E, Baughman R, Fletcher E,
et al (1999) Liposomal prostaglandin
E1 (TLC C-53) in acute respiratory distress syndrome: a controlled, randomized, double-blind, multicenter clinical
trial. Crit Care Med 27: 1478±1485
36. Yang S, Zhou M, Koo DJ, et al (1999)
Pentoxifylline prevents the transition
from the hyperdynamic to hypodynamic response during sepsis. Am J Physiol
277: 1036±1044
37. Staubach KH, Schröder J, Stüber F, et
al (1998) Effect of pentoxifylline in severe sepsis: results of a randomized,
double-blind, placebo-controlled study.
Arch Surg 133: 94±100
38. Lauterbach R, Pawlik D, Kowalczyk D,
et al (1999) Effect of the immunomodulating agent, pentoxifylline, in the treatment of sepsis in prematurely delivered
infants: a placebo-controlled, doubleblind trial. Crit Care Med 27: 807±814
39. Suter PM, Dominghetti G, Schaller
MD, et al (1994) N-Acetylcysteine enhances recovery from acute lung injury
in man. Chest 105: 190±194
40. Jepsen S, Herlevsen P, Knudsen P, et al
(1992) antioxidant treatment with Nacetylcysteine during adult respiratory
distress syndrome: a prospective, randomized, placebo-controlled study. Crit
Care Med 20: 918±923
41. Spies CD, Reinhart K, Witt I, et al
(1994) Influence of N-acetylcysteine on
indirect indicators of tissue oxygenation
in septic shock patients: results from a
prospective, randomized, double-blind
study. Crit Care Med 22: 1746±1748
42. Peake SL, Moran JL, Leppard PI (1996)
N-Acetyl-L-cysteine depresses cardiac
performance in patients with septic
shock. Crit Care Med 24: 1302±1310
43. Molnar Z, Shearer E, Lowe D, et al
(1999) N-Acetylcysteine treatment to
prevent the progression of multisystem
organ failure: a prospective, randomized, placebo-controlled study. Crit
Care Med 27: 1100±1104
44. De Backer WA, Amsel B, Jorens PG, et
al (1996) N-Acetylcysteine pretreatment of cardiac surgery patients influences plasma neutrophil elastase and
neutrophil influx in bronchoalveolar lavage fluid. Intensive Care Med 22:
900±908
45. Ortolani O, Conti A, De Gaudio AR, et
al (2000) The effect of glutathione and
N-acethylcysteine on lipoperoxidative
damage in patients with early septic
shock. Am J Respir Crit Care Med 161:
1907±1911
46. Forceville X, Vitoux D, Gauzit RØmy,
et al (1998) Selenium, systemic immune
response syndrome, sepsis, and outcome in critically ill patients. Crit Care
Med 26: 1536±1544
47. Angstwurm MWA, Schottdorf J, Schopohl J, et al (1999) Selenium replacement in patients with severe systemic
inflammatory response syndrome improves clinical outcome. Crit Care Med
27: 1807±1813
S 102
48. Vervloet MG, Thijs LG, Hack CE
(1998) Derangements of coagulation
and fibrinolysis in critically ill patients
with sepsis and septic shock. Semin
Thromb Hemost 24: 33±44
49. Fourrier F, Chopin C, Huart JJ, et al
(1993) Double-blind, placebo-controlled trial of antithrombin III concentrates in septic shock with disseminated
intravascular coagulation. Chest 104:
882±888
50. Baudo F, Caimi TM, de Cataldo F, et al
(1998) Antithrombin III (ATIII) replacement therapy in patients with sepsis and/or postsurgical complications: a
controlled double-blind, randomized,
multicenter study. Intensive Care Med
24: 336±342
51. Eisele B, Lamy M, Thijs LG, et al
(1998) Anti-thrombin III in patients
with severe sepsis: a randomized, placebo-controlled, double-blind, multicenter trial plus a meta-analysis on all randomized placebo-controlled, doubleblind trials with anti-thrombin III in severe sepsis. Intensive Care Med 24:
663±672
52. Riess H (2000) Antithrombin in severe
sepsis. ªNewº indication of an ªoldº
drug. Intensive Care Med 26: 657±665
53. Volk HD, Reimke P, Krausch D, et al
(1996) Monocyte deactivation ± rationale for a new therapeutic strategy
in sepsis. Intensive Care Med
22:S471±S481
54. Kaul R, McGeer A, Norrby-Teglund A,
et al (1999) Intravenous immunoglobulin therapy for streptococcal toxic shock
syndrome ± a comparative observational study. Clin Infect Dis 28: 800±807
55. Alejandria MM, Lansang MA, Dans
LF, Mantaring JB (2000) Intravenous
immunoglobulins for treating sepsis
and septic shock. Cochrane database
Syst Rev (2) CD001090
56. Pilz G, Appel R, Kreuzer E, Werdan K
(1997) Comparison of early IgM-enriched immunoglobulin vs. polyvalent
IgG administration in score-identified
postcardiac surgical patients at high
risk for sepsis. Chest 111: 419±426
57. The Intravenous Immunoglobulin Collaborative Study Group (1992) Prophylactic intravenous administration of
standard immune globulin as compared
with core-lipopolysaccharide immune
globulin in patients at high-risk of postsurgical infection. N Engl J Med 327:
234±240
58. Jenson HG, Pollock BH (1997) Metaanalyses of the effectiveness of intravenous immune globulin for prevention
and treatment of neonatal sepsis. Pediatrics 99:E2
59. Livingston DH, Loder PA, Kramer SM,
et al (1994) Interferon gamma administration increases monocyte HLA-DR
antigen expression but not endogenous
interferon production. Arch Surg 129:
172±178
60. Kohler J, Heumann D, Garotta G, et al
(1993) IFN-gamma involvement in the
severity of gram-negative infections in
mice. J Immunol 151: 916±921
61. The International Chronic Granulomatous Disease Cooperative Study Group
(1999) A controlled trial of interferon
gamma to prevent infection in chronic
granulomatous disease. N Engl J Med
1324: 509±516
62. Dries DJ, Jurkovich J, Maier RV, et al
(1994) Effect of interferon gamma on
infection-related death in patients with
severe injuries. Arch Surg 129:
1031±1042
63. Polk HC Jr, Cheadle WG, Linvingston
DH, et al (1992) A randomized prospective clinical trial to determine the
efficacy of interferon-g in severely injured patients. Am J Surg 63: 191±196
64. Wasserman D, Ioannovich JD, Hinzmann RD, et al (1998) Interferon-g in
the prevention of severe burn-related
infections: a European phase III multicenter trial. Crit Care Med 26: 434±439
65. Zhang P, Bagby GH, Stoltz DA, et al
(1998) Enhancement of peritoneal leukocyte function by granulocyte colonystimulating factor in rats with abdominal sepsis. Crit Care Med 26: 315±321
66. Lundblad R, Nesland JM, Giercksky
KE (1996) Granulocyte colony-stimulating factor improves survival rate and
reduces concentrations of bacteria, endotoxin, tumor necrosis factor, and endothelin-1 in fulminant intra-abdominal
sepsis in rats. Crit Care Med 24:
820±826
67. Karzai W, Ulrich von Specht B, Parent
C, et al (1999) G-CSF during Escherichia coli versus Staphylococcus aureus
pneumonia in rats has fundamentally
different and opposite effects. Am J Respir Crit Care Med 159: 1377±1382
68. Nelson S, Farkas S, Fotheringham N, et
al (1996) Filgrastim in the treatment of
hospitalized patients with community
acquired pneumonia (CAP) (abstract).
Am J Respir Crit Care Med 151:A535
69. Wunderink RG, Leeper KV, Schein
RMH, et al (1996) Clinical response to
figrastim in pneumonia with severe sepsis. Am J Respir Crit Care Med
153:A123
70. Heard S, Fink MI, Gamelli RE, et al
(1998) The effect of prophylactic administration of recombinant human GCSF on the incidence of nosocomial infections in patients with acute traumatic
brain injury or cerebral hemorrhage.
Crit Care Med 26: 748±754
71. Takala J, Ruokonen E, Webster NR, et
al (1999) Increased mortality associated
with growth hormone treatment in critically ill adults. N Engl J Med 341:
785±792
72. Doig GS, Martin CM, Sibbald WJ
(1997) Polymixin-dextran anti entodotin pre-treatment in an ovin model of
normotensive sepsis. Crit Care Med 25:
1956±1961
73. Yu M, Tomasa G (1993) A doubleblind, prospective, randomized trial of
ketoconazole, a thromboxane synthetase inhibitor, in the prophylaxis of
the adult respiratory distress syndrome.
Crit Care Med 21: 1635±1642
74. The ARDS network (2000) Ketaconazole for early treatment of acute lung
injury and acute respiratory distress
syndrome: a randomised controlled trial. JAMA 283: 1995±2002
75. Siedlar M, Szczepanik A, Wieckiewicz
J, et al (1997) Vancomycin down-regulates lipopolysaccharide-induced tumor
necrosis factor alpha (TNF-alpha) production and TNF alpha mRNA a accumulation in human blood monotyes.
Immunopharmacology 35: 265±271
76. Khan AA, Slifer TR, Remington JS
(1998) Effect of trovafloxacin on production of cytokines by human monocytes. Antimicrob Agents Chemother
42: 1713±1717
77. Riesbeck K, Gullberg M, Forsgren A
(1994) Evidence that the antibiotic ciprofloxacin counteracts cyclosporin-dependent suppression of cytokine production. Transplantation 57: 267±272
78. Willatts SM, Radford S, Leitermann M
(1995) Effect of the antiendotoxic
agent, taurolidine, in the treatment of
sepsis syndrome: a placebo-controlled,
double-blind trial. Crit Care Med 23:
1033±1039
79. Eisenfeld L, Krause PJ, Herson VC, et
al (1992) Enhancement of neonatal
neutrophil motility (chemotaxis) with
adult fresh frozen plasma. Am J Perinatol 9: 5±8
80. Galley HF, DIMatteo MA, Webster NR
(2000) Immunomodulation by anaesthetic, sedative and analgesic agents:
does it matter? Intensive Care Med 26:
267±274
S 103
81. Severn A, Rapson NT, Hunter CA, et al
(1992) Regulation of tumor necrosis
factor production by adrenaline and badrenergic agonists. J Immunol 148:
3441±3445
82. Van Der Poll T, Coyle SM, Barbosa K,
et al (1996) Epinephrine inhibits tumor
necrosis factor-a and potentiates interleukin 10 production during human endotoxemia. J Clin Invest 97: 713±719
83. Spengler RN, Chensue SW, Giacherio
DA, et al (1994) Endogenous norephinephrine regulates tumor necrosis factor-a production from macrophages in
vitro. J Immunol 152: 3024±3031
84. Ritchie PK, Ashby M, Knight HH, et al
(1996) Dopamine increase interleukin 6
release and inhibits tumor necrosis factor release from rat adrenal zona
glomerulosa in vitro. Eur J Endocrinol
134: 610±616
85. Unsaro A, Russell JA (2000) Could
anti-inflammatory actions of catecholamines explain the possible effect of supranormal oxygen delivery in critically
ill surgical patient. Intensive Care Med
26: 299±304
86. Bellomo R, Ronco C (1999) Continuous renal replacement therapy in the intensive care unit. Intensive Care Med
25: 781±789
87. Mehta R, McDonald B, Gabbai F, et al
(1996) Continuous versus intermittent
dialysis for acute renal failure in the
ICU: results from a randomized multicenter trial. J Am Soc Nephrol 5: 1457
88. Silvester W (1998) Outcome studies of
continuous renal replacement therapy
in the intensive care. Kidney Int
53:S138±S141
89. Journois D (1999) Hemofiltration during cardiopulmonary bypass. Minerva
Anestesiol 65: 427±432
90. Reeves JH, Warwick WB, Shann F, et al
(1999) Continuous plasmafiltration in
sepsis syndrome. Crit Care Med 27:
2096±2104
91. Churchwell KB, McManus ML, Kent P,
et al (1995) Intensive blood and plasma
exchange for treatment of coagulopathy
in meningococcemia. J Clin Apheresis
10: 171±177
Intensive Care Med (2001) 27: S 93±S 103
J. Carlet
Immunological therapy in sepsis:
currently available
)
J. Carlet ( )
Intensive Care Unit, Fondation-Hôpital Saint-Joseph, Paris,
France
E-mail: [email protected]
Phone: +33-1 44-12 37 83
Fax: +33-1 44-12 34 15
Should corticosteroids be used during septic shock at
low doses and for a prolonged period of time?
Answer: yes, grade C.
Recommendations
Introduction
Many therapies used in our daily practice are known to
have significant effects on inflammation. These drugs
influence the activation of the inflammatory network
that occurs during severe sepsis and related syndromes
as disseminated intravascular coagulation and acute respiratory distress syndrome (ARDS). Many of these
compounds (Table 1) have already been used during experimental models of sepsis and/or human studies.
Corticosteroids should not be used in severe sepsis or
septic shock at high doses (30 mg/kg) and for a short
course (1±2 days). On the other hand, corticosteroids
may be used during ªrefractoryº septic shock but not
during severe sepsis without shock or mild shock. It
should then be used at low doses (100 mg hydrocortisone three times a day) for 5 days or more and then
with subsequent tapering of the dose according to the
hemodynamic status and the need for vasopressors.
Rationale
Methods
This contribution reviews those drugs that are available in the daily
management of severe sepsis and septic shock. A computer-based
review of the literature was undertaken using Medline from 1990
to September 1999 as the primary database. The subject heading
keywords defined for each of the compounds listed in Table 1
were combined with the following general sepsis-related subject
heading keywords: sepsis, severe sepsis, septic shock, and ARDS.
Anti-inflammatory agents
Should corticosteroids be used in the treatment of
severe sepsis or septic shock at high doses (30 mg/kg)
for a short course (one or 2 days)?
Answer: no, grade A.
An extensive literature is available for corticosteroids.
Steroids have been used for many years, and their efficacy is controversial. Numerous animal studies performed during experimental septic (endotoxic) shock or
acute lung injuries showed a very significant reduction
in both intensity of shock, acute respiratory failure and
mortality [1, 2]. They have been used at very high doses
(30 mg/kg per dose for a maximum of 24±48 h). The
ability of these high doses of corticosteroids to reduce
complement activation and to inhibit leukocyte aggregability and adherence was at that time a very logical rationale for their efficacy [3]. Very promising initial findings have been published regarding humans [12]. However, two well designed, prospective, multicenter, randomized, double-blind studies demonstrated very clearly their inability to decrease mortality [5, 8]. Some studies mention positive trends when looking at subgroups
of infections due to Gram-negative rods [5, 8, 13].
S 94
Table 1 List of therapies currently available for eventually treating severe sepsis
Therapy
References
Anti-inflammatory agents
Corticosteroids (high or low doses)
Ibuprofen
Prostaglandin E1
Pentoxifylline
Oxygen scavengers
N-Acetylcysteine
Selenium
Drugs modifying coagulation
Antithrombin III
Drugs enhancing host defenses
Immunoglobulins
Interferon-g
Granulocytes stimulating factors
Immunonutrition
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26,
27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 25a
27, 28, 29, 30, 31
32, 33, 34, 35
36, 37, 38
39, 40, 41, 42, 43, 44, 45, 46, 47
39, 40, 41, 42, 43, 44, 45
46, 47
48, 49, 50, 51, 52
48, 49, 50, 51, 52
53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70
54, 55, 56, 57, 58
59, 60, 61, 62, 63, 64
65, 66, 67, 68, 69, 70
±a
Other drugs
Growth hormone
Antibiotics
Including ketoconazole
Including polymyxin B
Taurolidine
Fresh frozen plasma
Anesthetic sedative and analgesic agents
Catecholamines
71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85
71
72, 73, 74, 75, 76, 77
73, 74
72
78
79
80
81, 82, 83, 84, 85
Hemofiltration, plasma filtration, plasma exchange
86, 87, 88, 89, 90, 91
a
See PØrez and Dellinger, ªOther supportive therapies in sepsisº
Two recent meta-analyses [13, 14] reviewing the studies confirm that corticosteroids at the dose of 30 mg/kg
(one or two doses) are ineffective [13] or even harmful
[14]. The design and the results of the nine randomized
studies are summarized in Tables 2 and 3. Similar negative results have been obtained during ARDS [15].
Pooling the results only from those patients with
Gram-negative infections, as in the meta-analysis by Lefering et al. [13], yields a rate difference of ±5.6 % [confidence interval (CI): ±21.4 to 10.1) in favor of steroids,
based on 413 patients. Those patients with Gram-positive infections (n = 306) had an overall effect of +1.8 %
(CI: ±15.8 to 18.6). Most persons stopped using steroids
when these large trials were published.
Several studies performed over the years, however,
have maintained interest in the use of corticosteroids.
Mortality was reduced using steroids during severe typhoid fever [16], and neurological sequelae were reduced during meningitis [17]. Two large double-blind case
control studies demonstrated that prolonged treatment
(10±15 days) of relatively low doses of steroids
(120±240 mg hydrocortisone) dramatically reduced
mortality during severe Pneumocystis carinii pneumonia in AIDS patients [18, 19]. In addition, Meduri et al.
[20] showed that the course of late, fibrotic ARDS was
improved by steroid use which was confirmed in a re-
cent randomized double-blind study showing a significant reduction in mortality [21]. Two small, randomized,
double-blind studies of steroids in patients with severe
and refractory septic shock recently demonstrated positive results [22, 23]. Corticosteroids were used at small
doses (100 mg hydrocortisone three times per day in
one [22] and 100 mg followed by a continuous infusion
of 0.18 mg/kg per hour in the other [23], for longer periods of time than in past studies: 5 days in one [22] and
5±10 days in the other, with tapering of the doses according to hemodynamic status and need for vasopressors.
Both studies showed a significant reversal of shock and
organ failures and a trend in reduction in mortality. Additional studies are necessary, and a French multicenter
randomized, controlled, double-blind study reported
that low dose steroids decrease mortality in patients
with septic shock [25 a].
Several factors may explain the recent positive effects of corticosteroids during sepsis [24]. These include
the treatment of ªrelativeº adrenal insufficiency [25]
and the potentiation of adrenergic receptivity [26] in addition to the anti-inflammatory effect. The lower immunosuppressive doses and a more prolonged duration of
therapy than in the initial studies could also explain discrepancies.
S 95
Table 2 Design of the nine randomized studies used in the meta-analysis (from Cronin et al. [14]) (DB double blind, M methylprednisolone, B betamethasone, D dexamethasone, H hydrocortisone)
Reference
n
Type of study
Product
Dose
Duration
Endpoints
Cooperative
Study Group [6]
Klastersky
et al. [10]
Schumer et al. [12]
194
Open
H
6d
85
Open
B
300 mg then
50 mg/d
1 mg/kg
172
DB
M
30 mg/kg
1 dose or 2
Thompson
et al. [11]
Lucas and Ledgerwood [9]
Sprung et al. [4]
60
DB
M
30 mg/kg
48
Open
D
2 mg/kg
Max. 6 doses
in 24 h
2d
59
Open
M
30 mg/kg
1 dose (or 2)
Mortality, complications
Mortality (20 d),
complications
Mortality (28 d),
complications
Mortality, complications
Mortality (14 d),
complications
Hospital mortality,
complications
3d
Bone et al. [5]
381
DB
M
30 mg/kg
1d
Mortality (14 d),
complications
Veteran Administration [8]
223
DB
M
30 mg/kg
9d
Mortality (14 d),
complications
75
DB
M
30 mg/kg”4
1d
Hospital mortality,
ARDS complications
Luce et al. [7]
Table 3 Results of the nine randomized studies used in the metaanalysis (from Cronin et al. [14])
Cooperative study group [6]
Klastersky et al. [10]
Schumer et al. [12]
Thompson et al. [11]
Lucas et al. [9]
Sprung et al. [4]
Bone et al. [5]
Veteran Administration [8]
Luce et al. [7]
n
Risk ratio
95 % CI
194
85
172
60
48
59
381
223
75
1.72
0.97
0.30
1.01
1.09
1.11
1.35
0.95
1.07
1.23±2.41
0.65±1.45
0.13±0.72
0.77±1.31
0.36±3.27
0.74±1.67
0.98±1.84
0.57±1.58
0.72±1.60
Should ibuprofen be used in the treatment of severe
sepsis and septic shock?
Answer: no, grade B.
Recommendations
Ibuprofen should not be used during severe sepsis or
septic shock. Additional studies are needed to determine whether some patients, for example, those with
hypothermia, could benefit from the drug.
Rationale
Ibuprofen is a powerful anti-inflammatory agent, acting
on the prostaglandin metabolism as a cyclo-oxygenase
inhibitor. It has been used with controversial effects in
animals during both experimental sepsis and ARDS
[27, 28]. Two small randomized, double-blind studies in
patients showed some hemodynamic effect and a normalization of pH without any significant effect upon
mortality [29, 30]. Mortality was decreased significantly
in a post hoc analysis of hypothermic patients [30]. A
large multicenter randomized, controlled, double blind
study, however, failed to demonstrate any effect upon
mortality, reversal of shock or acute respiratory failure
[31]. Ibuprofen was able to reduce the levels of prostacyclin and thromboxane and to decrease fever, tachycardia and oxygen consumption [31]. The drug was not associated with adverse affects.
Should prostaglandins be used in the treatment of
ARDS due to severe infections and sepsis?
Answer: no, grade B.
Recommendations
Prostaglandins, in particular prostaglandin E1 or liposomal prostaglandin E1 should not be used during ARDS
due to sepsis. There are no specific data allowing recommendations in severe sepsis.
S 96
Rationale
Oxygen scavengers
Several prostaglandins which have both an anti-inflammatory and a vasoactive effect have been studied including prostaglandin I2 and particularly prostaglandin
E1 [32, 33, 34, 35] during ARDS. The vast majority of
these patients had ARDS due to severe infections or
sepsis. An early, small, randomized study showed promising results [32]. However, a large multicenter, randomized, controlled, double blind study failed to show any
difference in survival [33]. An increase in oxygen delivery and oxygen consumption was noted in treated patients who survived [34]. A recent, multicenter randomized, controlled, double-blind study with liposomal
prostaglandin E1 (TLC C-53) showed that indices of oxygenation of treated ARDS patients were improved
compared with controls, but without any effect upon duration of mechanical ventilation or 28 days mortality
[35]. Again, most ARDS was due to sepsis in these two
large studies. No data are really available concerning
an overall group of patients with severe sepsis.
Several oxygen scavengers are currently available, including N-acetylcysteine (NAC), vitamin E, vitamin C,
and selenium. Vitamins E and C have been only poorly
studied in humans, and we focus on N-acetylcysteine
and selenium.
Should pentoxifylline be used in the treatment of severe
sepsis in (a) adults, (b) infants?
Answer: (a) no, grade B; (b) no, grade C.
Recommendations
Pentoxifylline should not be used in adults with severe
sepsis unless new studies show a significant effect. The
positive effect of a small study in infants should be confirmed before clinical use.
Rationale
Pentoxifylline, which has a powerful anti-inflammatory
effect including a strong inhibition of tumor necrosis
factor secretion, has been used successfully in many animal studies with the prevention of the transition from a
hyperdynamic to hypodynamic state, although no effect
upon mortality has been shown [36]. Human studies are
more scarce. A multicenter, randomized, controlled,
double-blind study during sepsis showed an increase in
PaO2/FIO2 ratio but no effect upon cytokines levels or
mortality [37]. A recent double-blind study performed
in premature infants with sepsis showed a decrease in
cytokines levels and a significant decrease in mortality
(1/40 vs. 6/38 p = 0.046) [38]. However, the size of this
study was rather small, and additional large studies are
mandatory.
Should N-acetylcysteine be used in the treatment of
severe sepsis?
Answer: no, grade C.
Recommendations
NAC should not be used in severe sepsis until new data
are available, focusing in particular on very early therapy.
Rationale
During acute lung injury an improvement in oxygenation and reduction in the required length of mechanical
ventilation was found in patients treated with NAC
compared to controls [39]. However, several randomized studies have shown no difference in mortality, gas
exchange, and development of respiratory failure in patients treated with NAC [39, 40]. Several studies have
also been performed during severe sepsis, with heterogeneous results [41, 42, 43, 44]. Depressed cardiac performance has been described in septic patients treated
with NAC [42]. A very recent multicenter, randomized,
controlled, double-blind study showed that a prolonged
infusion of NAC is unable to prevent multiple organ
failure in consecutively admitted critically ill patients
[43]. In this study treatment used more than 24 h after
the initial insult worsened the prognosis compared to
controls. Better results were obtained when the drug
was use before the insult, as during cardiac surgery
[44]. These results suggest that this compound could be
helpful when started before (or perhaps shortly after)
the insult, but possibly harmful when started too late.
Combinations of several antioxidants have also been
published, but data are too limited to allow recommendations [45].
Should selenium be used in the treatment of severe
sepsis?
Answer: no, grade C.
S 97
Fig. 1 Effect of proinflammatory cytokines. Upon coagulation cascade during sepsis leading to an activation of tissue factor, a depletion in protein C (via a decrease in thrombomodulin levels) antithrombin III and C1 inhibitor, and a decrease in fibrinolysis (via
the effect of plasminogen activator inhibitor 1)
Recommendations
Selenium should not be used for severe sepsis. Additional studies are warranted to confirm initial positive
data.
Rationale
A profound depletion in selenium levels has been demonstrated in many severe septic patients [46]. Mortality
and morbidity are far higher in patients with a very low
selenium level [46]. A recent prospective, randomized,
but nonblinded study performed in septic patients
showed that selenium replacement is able to reduce severity indexes at day 3 and reduce the need for hemodialysis but has no significant effect upon mortality (52 %
in controls and 33, 5 % in treated patients, p = 0.13)
[47]. Additional large studies are needed to confirm initial promising results.
Drugs modifying coagulation
There are complex interactions between the inflammation and coagulation systems (Fig. 1). Proinflammatory
cytokines activate coagulation cascades, in particular
via an effect upon tissue factor which is a key player in
the coagulation cascade. They can also reduce fibrinolysis and profoundly reduce the levels of protein C and of
antithrombin III which are important anticoagulant
agents. Antithrombin III inhibits several coagulation
factors of the extrinsic pathway such as factors IXa,
XIa, XIIa in addition to factors Xa, IIa, and plasmin.
Activated protein C inhibits factors Va, Vlla, and plasminogen activator inhibitor 1. The overall effect during
sepsis is a marked procoagulant balance. Conversely,
coagulation products can activate the inflammation network which creates numerous amplification loops. For
example, thrombin can induce an up-regulation of Pand E-selectin, and contact factor activation can induce
the production of bradykinin, worsening hypotension
and tissue hypoperfusion. In humans studies, both anti-
S 98
thrombin III and protein C levels are sharply decreased
[48], and mortality of septic patients is inversely correlated with the levels of those two products. This makes
the rationale for studying those types of compounds,
such as antithrombin III, protein C, and tissue factor
protein inhibitor very strong. Only antithrombin III is
currently available.
Should antithrombin III be used in the treatment of
severe sepsis?
Answer: no, grade B.
Recommendations
Antithrombin III should not be used during severe sepsis. Countries which allow the free use of this drug in
this setting should reconsider their position.
Rationale
Antithrombin III is a drug which is widely used for septic patients in several countries. Three randomized,
small, double-blind studies were published [49, 50, 51].
Duration of disseminated intravascular coagulation
was reduced [49] as well as the number of organ failures
[51], but mortality was not different although a positive
trend was clearly noted. A meta-analysis was also performed [51] showing a 22.9 % reduction in mortality
but which did not reach statistical significance. Unfortunately a large multicenter, prospective, double-blind
study has recently been completed which showed no significant improvement in survival [52]. The complete
data have not yet been published. Other drugs such as
activated protein C and tissue factor inhibitors are not
currently available and are discussed elsewhere (see
Arndt and Abraham, ªImmunological therapy of sepsis:
experimental therapiesº).
Drugs enhancing host defenses
After the initial activation of the proinflammatory network, a profound immunodepression can occur in septic
patients [53]. This could influence outcome increasing
the risk of nosocomial infections. Several strategies
have been used to increase host defenses, including
polyvalent immunoglobulins, interferon-g, stimulating
factors for granulocytes [including granulocyte colony
stimulating factor (G-CSF)], and immunonutrition.
The latter is discussed elsewhere (see PØrez and Dellinger, ªOther supportive therapy in sepsisº).
Should intravenous immunoglobulins be used in the
treatment of severe sepsis in (a) adults or (b) neonates?
Answer: (a) no, grade C; (b) no, grade C.
Recommendations
Immunoglobulins should not be used either in adult patients or in neonates with sepsis, unless additional large
studies confirm some positive data in small-sized metaanalyses. Countries which allow a wide use of these
compounds should reconsider their position and encourage these studies.
Rationale
Intravenous immunoglobulins (IVIG) are widely used
in both infants and adults in the treatment of severe sepsis, at least in certain countries. Reports which support
their empirical use, however, are still rather weak. The
rationale is to restore immunoglobulins levels, which
may be depressed in sepsis, and to provide patients
with specific antibodies against micro-organisms. No individual well designed clinical study has been performed in adults with severe sepsis. A recent study was performed in patients with streptococcal toxic shock syndrome [54]. This was a comparative nonblinded study
performed in 21 patients which demonstrated a significantly reduced mortality (67 % vs. 34 %, p = 0.02).
Both Acute Physiology and Chronic Health Evaluation
II scores and IVIG were prognostic factors in the multivariate analysis. The odds ratio associated with IVIG
was 8.1 (95 % CI: 1.6±45). A recent meta-analysis by
the Cochrane group [55] looking at 23 studies (some of
them unpublished) on immunoglobulins, antiendotoxins, and anticytokines, extracted from the small size
studies already published, evaluated a population of
413 patients receiving polyclonal immunoglobulins.
Mortality was significantly reduced (relative risk: 0.6;
95 % CI: 0.47±0.76). Results were even more positive
when only sepsis related deaths were considered. A
large, well designed, multicenter, randomized, doubleblind study is, however, warranted before making firm
conclusions. Two prophylactic studies have been published recently [56, 57]. A study performed in cardiac
surgery patients showed no difference in the occurrence
of sepsis between polyvalent IVIG and IgM-enriched
immunoglobulin [56]. A prospective comparative study
showed that IVIG and not placebo is able to prevent
nosocomial infections after major surgery [57]. Such
prophylactic studies are needed in this field in nonsurgical critically ill patients.
In neonatal sepsis, a recent meta-analysis of 110 newborns in three studies showed that IVIG is able to re-
S 99
duce mortality significantly (odds ratio: 0.173; 95 % CI:
0.031±0.735; p = 0.007) [58]. However, the size of the
overall population was very small, and large studies are
urgently warranted. In the same meta-analysis the effect
of IVIG in the prevention of sepsis in 4933 evaluable
newborns was significant (p = 0.0193, two-tailed), although heterogeneity of the studies precluded estimation of an overall odds ratio.
Other Drugs
Interferon-g
Growth hormone should not be used in patients with
sepsis because it increases mortality.
Interferon-g has been used successfully in animals models of Gram-negative sepsis [59, 60]. Few data are available in human sepsis. The drug has been used with positive results to prevent infection during chronic granulomatous disease [61] and trauma [62, 63]. The drug, however, was unable to prevent infections in burn patients
[64]. Data are insufficient for therapy of severe sepsis
to allow recommendations.
Should growth hormone be used in the treatment of
severe sepsis?
Answer: no, grade A.
Recommendations
Rationale
Answer: no, grade C.
The administration of growth hormone could in theory
attenuate the catabolic response to injury, surgery or
sepsis. Two prospective double-blind studies with more
than 200 patients each were recently reported in critically ill patients with cardiac or abdominal surgery, multiple trauma or acute respiratory failure [71]. Mortality
was increased significantly in treated patients. The relative risk in these two pooled studies was 1.9 (95 % CI:
1.3±2.9). Length of stay and duration of mechanical ventilation were longer in treated survivors than in controls.
Recommendations
Antimicrobial compounds
G-CSF should not be used in nonneutropenic patients
with severe sepsis.
Polymixin B. Polymixin B is able to neutralize endotoxin via strong antilipid A activity [72]. Since it is very toxic, it is difficult to use intravenously in humans, although some derivates are less toxic. Extracorporeal
techniques, in which polymyxin is coated on membranes, are under investigation.
Should granulocyte colony stimulating factor be used in
the treatment of severe infections?
Rationale
G-CSF is very efficient and reduces mortality in animal
models of abdominal sepsis [65, 66]. During pneumonia
models in rats the drug has been shown to exert different effects according to the micro-organisms involved
[67]. Preliminary studies have been performed in community or hospital acquired pneumonia with controversial results [68, 69]. In patients with head trauma and receiving mechanical ventilation G-CSF prophylaxis did
not improve outcome nor lower the risk of nosocomial
pneumonia [70].
Ketoconazole. Ketaconazole, one of the new imidazoles, has a strong effect upon thromboxane synthase inhibition and has been shown to prevent ARDS in septic
patients in a small double-blind randomized study [73].
A recent study performed in 234 patients, however,
failed to demonstrate any effect upon mortality and
duration of mechanical ventilation in ARDS patients
[74]. No data are available in patients with sepsis.
Other antibiotics
Immunonutrition
See PØrez and Dellinger, ªOther supportive therapies in
sepsis.º
Some antibiotics have anti-inflammatory effects, in particular in decreasing cytokine release. Effects have
been shown for vancomycin [75] trovafloxacin [76] and
ciprofloxacin [77].
S 100
Taurolidine
Hemofiltration and plasma filtration
Taurolidine is an anti-infective agent (nonantibiotic),
used either locally, or intravenously, which has some antibacterial effect associated with an antiendotoxin effect. A randomized placebo-controlled study failed to
demonstrate any effect on outcome in sepsis [78].
Should hemofiltration be used in the treatment of
patients with severe sepsis, without renal indications?
Answer: no, grade C.
Recommendations
Other drugs currently used
Many other drugs that we use daily could have important effects upon inflammation, including heparin, fresh
frozen plasma [79], and anesthetic, sedative, and analgesic agents [80]. A recent review [80] describes the potential effects of these agents upon immunomodulation.
Catecholamines and inflammation. It is well known that
inotropic agents such as catecholamines have a significant impact upon inflammation [81]. Epinephrine inhibits tumor necrosis factor and potentiates interleukin10 leading to a significant anti-inflammatory effect [82],
via an effect upon macrophages [83]. Dopamine increases interleukin-6 release but decreases tumor necrosis
factor [84]. Recent data support the concept that the
anti-inflammatory effect of catecholamines explains
the possible beneficial effects of supranormal oxygen
delivery in critically ill surgical patients [85]. These
data do not enable clinicians to take into account the effect of catecholamines upon inflammation in deciding
which is the best to use.
Hemofiltration should not be used in patients with sepsis without renal indications unless ongoing studies provide positive results.
Rationale
Hemofiltration has been shown to decrease cytokines
levels significantly, although temporarily during severe
sepsis in humans. The technique is widely used in Europe and many authors have strong opinions [86] regarding its use, although the data are weak. A randomized, still unpublished study found no effect upon mortality [87]. Another randomized controlled study [88]
reported a 15 % (nonsignificant) increase in survival
for filtrated patients. Favorable results have been described for cardiac surgery patients [89]. Large multicenter studies are currently under way.
Plasma filtration induced a significant attenuation of
acute-phase response in a randomized, prospective
study recently performed in 22 adults with sepsis [90].
However, no difference in mortality and only a trend toward fewer organ failures were noted. Plasma exchange
has also been used in severe meningococcemia in children [91] with varying results.
References
1. Hinshaw LB, Archer LT, Beller Todd
BK, et al (1981) Survival in primates in
lethal septic shock following delayed
treatment with steroid. Circ Shock 8:
291±300
2. Hinshaw LB, Beller Todd BK, Archer
LT, et al (1979) Recovery from lethal
Escherichia coli shock in dogs. Surg Gynecol Obstet 149: 545±553
3. Hammerschmidt DE, White JG, Craddock PR, et al (1979) Corticosteroids
inhibit complement-induced granulocyte aggregation: a possible mechanism
for their efficacy in shock states. J Clin
Invest 63: 798±803
4. Sprung CL, Caralis PV, Marcial EH, et
al (1984) The effect of high-dose corticosteroids in patients with septic shock:
a prospective, controlled study. N Engl
J Med 311: 1137±1143
5. Bone RC, Fisher CJ Jr, Clemmer TP, et
al (1987) A controlled clinical trial of
high-dose methylprednisolone in the
treatment of severe sepsis and septic
shock. N Engl J Med 317: 653±658
6. Cooperative Study Group, Bennet IL,
Finland M, et al (1963) The effectiveness of hydrocortisone in the management of severe infections. JAMA 183:
462±465
7. Luce JM, Montgomery AB, Marks JD,
et al (1988) Ineffectiveness of highdose methylprednisolone in preventing
parenchymal lung injury and improving
mortality in patients with septic shock.
Am J Respir Crit Care Med 138: 62±68
8. The Veterans Administration Systemic
Sepsis Cooperative Study Group
(1987) Effect of high-dose glucocorticosteroid therapy on mortality in patients with clinical signs of systemic sepsis. N Engl J Med 317: 659±665
9. Lucas C, Ledgerwood A (1984) The
cardiopulmonary response to massive
doses of steroids in patients with septic
shock. Arch Surg 119: 537±541
10. Klastersky J, Cappel R, Debusscher L
(1971) Effectiveness of betamethasone
in management of severe infections.
N Engl J Med 284: 1248±1250
11. Thompson WL, Gurley H, Lutz B, et al
(1976) Inefficacy of corticosteroids in
shock. Clin Res 24: 258A
12. Schumer W (1976) Steroids in the treatment of clinical septic shock. Ann Surg
184: 333±341
S 101
13. Lefering R, Neugebauer EAM (1995)
Steroid controversy in sepsis and septic
shock: a meta-analysis. Crit Care Med
23: 1294±1303
14. Cronin L, Cook DJ, Carlet J, et al
(1995) Corticosteroid treatment for
sepsis: a critical appraisal and metaanalysis of the literature. Crit Care
Med 23: 1430±1439
15. Bernard GR, Luce JM, Sprung CL, et al
(1987) High-dose corticosteroids in patients with the adult respiratory distress syndrome. N Engl J Med 317:
1565±1570
16. Hoffman SL, Punjabi NH, Kumala S, et
al (1984) Reduction in mortality in
chloramphenicol treated severe typhoid
fever by high-dose dexamethasone.
N Engl J Med 310: 82±88
17. Lebel MH, Freij BJ, Syrogainnopolous
GA, et al (1988) Dexamethasone therapy for bacterial meningitis: results of
two double-blind placebo-controlled
trials. N Engl J Med 319: 964±971
18. Montaner JS, Lawson LM, Levitt N, et
al (1990) Corticosteroids prevent early
deterioration in patients with moderate
severe Pneumocystis carinii pneumonia
and the acquired immunodeficiency
syndrome. Ann Intern Med 113: 14±20
19. Gagnon S, Boota AM, Fischl MA, et al
(1990) Corticosteroids as adjunctive
therapy for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. N Engl J Med 323:
1440±1450
20. Meduri U, Chinn AJ, Leeper KV, et al
(1994) Corticosteroid rescue treatment
of progressive fibroproliferation in
late ARDS: patterns of response and
predictors of outcome. Chest 105:
1516±1527
21. Meduri GU, Headley ASH, Golden E,
et al (1998) Effect of prolonged methylprednisolone therapy in unresolving
acute respiratory distress syndrome.
JAMA 280: 159±165
22. Bollaert PE, Charpentier C, Levy S, et
al (1998) Reversal of late septic shock
with supraphysiologic doses of hydrocortisone. Crit Care Med 26: 645±650
23. Briegel J, Forst H, Haller M, et al
(1999) Stress doses of hydrocortisone
reverse hyperdynamic septic shock: a
prospective, randomized, double-blind,
single center study. Crit Care Med 27:
723±732
24. Carlet J (1999) From mega to more reasonable doses of corticosteroids: a decade to recreate hope. Crit Care Med
27: 672±674
25. Hatherill M, Tibby SM, Hilliard T, et al
(1999) Adrenal insufficiency in septic
shock. Arch Dis Child 80: 51±55
25 a. Annane D (2001) Effects of combination hydrocortisone (HC) ± fludro-cortisone (FC) on mortality in septic shock.
Crit Care Med (in press)
26. Davies AO, Lefkowitz RJ (1980) Corticosteroids induced differential regulation of beta-adrenergic receptors in circulating human polymorphonuclear
leukocytes and mononuclear leukocytes. J Clin Endocrinol Metab 51:
599±605
27. Snapper JR, Hutchison AA, Obletree
ML, Brigham KL (1983) Effects of cyclooxygenase inhibitors on the alterations of lung mechanics caused by endotoxemia in unanesthetized sheep.
J Clin Invest 72: 63±76
28. Elinger JH, Seyde WC, Longnecker DE
(1984) Methylprednisolone plus ibuprofen increase mortality in septic rats.
Circ Shock 14: 203±208
29. Haupt MT, Jastremski MS, Clemmer
TP, et al (1991) Effect of ibuprofen in
patients with severe sepsis: a randomized, double-blind, multicenter study.
Crit Care Med 19: 1339±1347
30. Arons MM, Wheeler AP, Bernard GR,
et al (1999) Effects of ibuprofen on the
physiology and survival of hypothermic
sepsis. Crit Care Med 27: 699±707
31. Bernard GR, Wheeler AP, Russell JA,
et al (1997) The effects of Ibuprofen on
the physiology and survival of patients
with sepsis. N Engl J Med 336: 912±918
32. Holcroft JW, Vassar MJ, Weber CJ
(1986) Prostaglandin E1 and survival in
patients with the adult respiratory distress syndrome. A prospective trial.
Ann Surg 203: 371±378
33. Bone RC, Slotman G, Maunder R, et al
(1989) Randomized double-blind, multicenter study of prostaglandin E1 in patients with the adult respiratory distress
syndrome. Chest 96: 114±119
34. Silverman JH, Slotman G, Bone RC, et
al (1990) Effects of prostaglandin E1
on oxygen delivery and consumption in
patients with the adult respiratory distress syndrome. Results from the prostaglandin E1 multicenter trial. The
Prostaglandin E1 Study group. Chest
98: 405±410
35. Abraham E, Baughman R, Fletcher E,
et al (1999) Liposomal prostaglandin
E1 (TLC C-53) in acute respiratory distress syndrome: a controlled, randomized, double-blind, multicenter clinical
trial. Crit Care Med 27: 1478±1485
36. Yang S, Zhou M, Koo DJ, et al (1999)
Pentoxifylline prevents the transition
from the hyperdynamic to hypodynamic response during sepsis. Am J Physiol
277: 1036±1044
37. Staubach KH, Schröder J, Stüber F, et
al (1998) Effect of pentoxifylline in severe sepsis: results of a randomized,
double-blind, placebo-controlled study.
Arch Surg 133: 94±100
38. Lauterbach R, Pawlik D, Kowalczyk D,
et al (1999) Effect of the immunomodulating agent, pentoxifylline, in the treatment of sepsis in prematurely delivered
infants: a placebo-controlled, doubleblind trial. Crit Care Med 27: 807±814
39. Suter PM, Dominghetti G, Schaller
MD, et al (1994) N-Acetylcysteine enhances recovery from acute lung injury
in man. Chest 105: 190±194
40. Jepsen S, Herlevsen P, Knudsen P, et al
(1992) antioxidant treatment with Nacetylcysteine during adult respiratory
distress syndrome: a prospective, randomized, placebo-controlled study. Crit
Care Med 20: 918±923
41. Spies CD, Reinhart K, Witt I, et al
(1994) Influence of N-acetylcysteine on
indirect indicators of tissue oxygenation
in septic shock patients: results from a
prospective, randomized, double-blind
study. Crit Care Med 22: 1746±1748
42. Peake SL, Moran JL, Leppard PI (1996)
N-Acetyl-L-cysteine depresses cardiac
performance in patients with septic
shock. Crit Care Med 24: 1302±1310
43. Molnar Z, Shearer E, Lowe D, et al
(1999) N-Acetylcysteine treatment to
prevent the progression of multisystem
organ failure: a prospective, randomized, placebo-controlled study. Crit
Care Med 27: 1100±1104
44. De Backer WA, Amsel B, Jorens PG, et
al (1996) N-Acetylcysteine pretreatment of cardiac surgery patients influences plasma neutrophil elastase and
neutrophil influx in bronchoalveolar lavage fluid. Intensive Care Med 22:
900±908
45. Ortolani O, Conti A, De Gaudio AR, et
al (2000) The effect of glutathione and
N-acethylcysteine on lipoperoxidative
damage in patients with early septic
shock. Am J Respir Crit Care Med 161:
1907±1911
46. Forceville X, Vitoux D, Gauzit RØmy,
et al (1998) Selenium, systemic immune
response syndrome, sepsis, and outcome in critically ill patients. Crit Care
Med 26: 1536±1544
47. Angstwurm MWA, Schottdorf J, Schopohl J, et al (1999) Selenium replacement in patients with severe systemic
inflammatory response syndrome improves clinical outcome. Crit Care Med
27: 1807±1813
S 102
48. Vervloet MG, Thijs LG, Hack CE
(1998) Derangements of coagulation
and fibrinolysis in critically ill patients
with sepsis and septic shock. Semin
Thromb Hemost 24: 33±44
49. Fourrier F, Chopin C, Huart JJ, et al
(1993) Double-blind, placebo-controlled trial of antithrombin III concentrates in septic shock with disseminated
intravascular coagulation. Chest 104:
882±888
50. Baudo F, Caimi TM, de Cataldo F, et al
(1998) Antithrombin III (ATIII) replacement therapy in patients with sepsis and/or postsurgical complications: a
controlled double-blind, randomized,
multicenter study. Intensive Care Med
24: 336±342
51. Eisele B, Lamy M, Thijs LG, et al
(1998) Anti-thrombin III in patients
with severe sepsis: a randomized, placebo-controlled, double-blind, multicenter trial plus a meta-analysis on all randomized placebo-controlled, doubleblind trials with anti-thrombin III in severe sepsis. Intensive Care Med 24:
663±672
52. Riess H (2000) Antithrombin in severe
sepsis. ªNewº indication of an ªoldº
drug. Intensive Care Med 26: 657±665
53. Volk HD, Reimke P, Krausch D, et al
(1996) Monocyte deactivation ± rationale for a new therapeutic strategy
in sepsis. Intensive Care Med
22:S471±S481
54. Kaul R, McGeer A, Norrby-Teglund A,
et al (1999) Intravenous immunoglobulin therapy for streptococcal toxic shock
syndrome ± a comparative observational study. Clin Infect Dis 28: 800±807
55. Alejandria MM, Lansang MA, Dans
LF, Mantaring JB (2000) Intravenous
immunoglobulins for treating sepsis
and septic shock. Cochrane database
Syst Rev (2) CD001090
56. Pilz G, Appel R, Kreuzer E, Werdan K
(1997) Comparison of early IgM-enriched immunoglobulin vs. polyvalent
IgG administration in score-identified
postcardiac surgical patients at high
risk for sepsis. Chest 111: 419±426
57. The Intravenous Immunoglobulin Collaborative Study Group (1992) Prophylactic intravenous administration of
standard immune globulin as compared
with core-lipopolysaccharide immune
globulin in patients at high-risk of postsurgical infection. N Engl J Med 327:
234±240
58. Jenson HG, Pollock BH (1997) Metaanalyses of the effectiveness of intravenous immune globulin for prevention
and treatment of neonatal sepsis. Pediatrics 99:E2
59. Livingston DH, Loder PA, Kramer SM,
et al (1994) Interferon gamma administration increases monocyte HLA-DR
antigen expression but not endogenous
interferon production. Arch Surg 129:
172±178
60. Kohler J, Heumann D, Garotta G, et al
(1993) IFN-gamma involvement in the
severity of gram-negative infections in
mice. J Immunol 151: 916±921
61. The International Chronic Granulomatous Disease Cooperative Study Group
(1999) A controlled trial of interferon
gamma to prevent infection in chronic
granulomatous disease. N Engl J Med
1324: 509±516
62. Dries DJ, Jurkovich J, Maier RV, et al
(1994) Effect of interferon gamma on
infection-related death in patients with
severe injuries. Arch Surg 129:
1031±1042
63. Polk HC Jr, Cheadle WG, Linvingston
DH, et al (1992) A randomized prospective clinical trial to determine the
efficacy of interferon-g in severely injured patients. Am J Surg 63: 191±196
64. Wasserman D, Ioannovich JD, Hinzmann RD, et al (1998) Interferon-g in
the prevention of severe burn-related
infections: a European phase III multicenter trial. Crit Care Med 26: 434±439
65. Zhang P, Bagby GH, Stoltz DA, et al
(1998) Enhancement of peritoneal leukocyte function by granulocyte colonystimulating factor in rats with abdominal sepsis. Crit Care Med 26: 315±321
66. Lundblad R, Nesland JM, Giercksky
KE (1996) Granulocyte colony-stimulating factor improves survival rate and
reduces concentrations of bacteria, endotoxin, tumor necrosis factor, and endothelin-1 in fulminant intra-abdominal
sepsis in rats. Crit Care Med 24:
820±826
67. Karzai W, Ulrich von Specht B, Parent
C, et al (1999) G-CSF during Escherichia coli versus Staphylococcus aureus
pneumonia in rats has fundamentally
different and opposite effects. Am J Respir Crit Care Med 159: 1377±1382
68. Nelson S, Farkas S, Fotheringham N, et
al (1996) Filgrastim in the treatment of
hospitalized patients with community
acquired pneumonia (CAP) (abstract).
Am J Respir Crit Care Med 151:A535
69. Wunderink RG, Leeper KV, Schein
RMH, et al (1996) Clinical response to
figrastim in pneumonia with severe sepsis. Am J Respir Crit Care Med
153:A123
70. Heard S, Fink MI, Gamelli RE, et al
(1998) The effect of prophylactic administration of recombinant human GCSF on the incidence of nosocomial infections in patients with acute traumatic
brain injury or cerebral hemorrhage.
Crit Care Med 26: 748±754
71. Takala J, Ruokonen E, Webster NR, et
al (1999) Increased mortality associated
with growth hormone treatment in critically ill adults. N Engl J Med 341:
785±792
72. Doig GS, Martin CM, Sibbald WJ
(1997) Polymixin-dextran anti entodotin pre-treatment in an ovin model of
normotensive sepsis. Crit Care Med 25:
1956±1961
73. Yu M, Tomasa G (1993) A doubleblind, prospective, randomized trial of
ketoconazole, a thromboxane synthetase inhibitor, in the prophylaxis of
the adult respiratory distress syndrome.
Crit Care Med 21: 1635±1642
74. The ARDS network (2000) Ketaconazole for early treatment of acute lung
injury and acute respiratory distress
syndrome: a randomised controlled trial. JAMA 283: 1995±2002
75. Siedlar M, Szczepanik A, Wieckiewicz
J, et al (1997) Vancomycin down-regulates lipopolysaccharide-induced tumor
necrosis factor alpha (TNF-alpha) production and TNF alpha mRNA a accumulation in human blood monotyes.
Immunopharmacology 35: 265±271
76. Khan AA, Slifer TR, Remington JS
(1998) Effect of trovafloxacin on production of cytokines by human monocytes. Antimicrob Agents Chemother
42: 1713±1717
77. Riesbeck K, Gullberg M, Forsgren A
(1994) Evidence that the antibiotic ciprofloxacin counteracts cyclosporin-dependent suppression of cytokine production. Transplantation 57: 267±272
78. Willatts SM, Radford S, Leitermann M
(1995) Effect of the antiendotoxic
agent, taurolidine, in the treatment of
sepsis syndrome: a placebo-controlled,
double-blind trial. Crit Care Med 23:
1033±1039
79. Eisenfeld L, Krause PJ, Herson VC, et
al (1992) Enhancement of neonatal
neutrophil motility (chemotaxis) with
adult fresh frozen plasma. Am J Perinatol 9: 5±8
80. Galley HF, DIMatteo MA, Webster NR
(2000) Immunomodulation by anaesthetic, sedative and analgesic agents:
does it matter? Intensive Care Med 26:
267±274
S 103
81. Severn A, Rapson NT, Hunter CA, et al
(1992) Regulation of tumor necrosis
factor production by adrenaline and badrenergic agonists. J Immunol 148:
3441±3445
82. Van Der Poll T, Coyle SM, Barbosa K,
et al (1996) Epinephrine inhibits tumor
necrosis factor-a and potentiates interleukin 10 production during human endotoxemia. J Clin Invest 97: 713±719
83. Spengler RN, Chensue SW, Giacherio
DA, et al (1994) Endogenous norephinephrine regulates tumor necrosis factor-a production from macrophages in
vitro. J Immunol 152: 3024±3031
84. Ritchie PK, Ashby M, Knight HH, et al
(1996) Dopamine increase interleukin 6
release and inhibits tumor necrosis factor release from rat adrenal zona
glomerulosa in vitro. Eur J Endocrinol
134: 610±616
85. Unsaro A, Russell JA (2000) Could
anti-inflammatory actions of catecholamines explain the possible effect of supranormal oxygen delivery in critically
ill surgical patient. Intensive Care Med
26: 299±304
86. Bellomo R, Ronco C (1999) Continuous renal replacement therapy in the intensive care unit. Intensive Care Med
25: 781±789
87. Mehta R, McDonald B, Gabbai F, et al
(1996) Continuous versus intermittent
dialysis for acute renal failure in the
ICU: results from a randomized multicenter trial. J Am Soc Nephrol 5: 1457
88. Silvester W (1998) Outcome studies of
continuous renal replacement therapy
in the intensive care. Kidney Int
53:S138±S141
89. Journois D (1999) Hemofiltration during cardiopulmonary bypass. Minerva
Anestesiol 65: 427±432
90. Reeves JH, Warwick WB, Shann F, et al
(1999) Continuous plasmafiltration in
sepsis syndrome. Crit Care Med 27:
2096±2104
91. Churchwell KB, McManus ML, Kent P,
et al (1995) Intensive blood and plasma
exchange for treatment of coagulopathy
in meningococcemia. J Clin Apheresis
10: 171±177
Intensive Care Med (2001) 27: S 104±S 115
Immunological therapy of sepsis:
experimental therapies
Patrick Arndt
Edward Abraham
)
P. Arndt ´ E. Abraham ( )
Division of Pulmonary Sciences and Critical Care Medicine,
University of Colorado Health Sciences Center,
4200 E. Ninth Avenue, Denver, CO 80262, USA
E-mail: [email protected]
Phone: +1-3 03-3 15 70 47
Fax: +1-3 03-3 15 56 32
Introduction
An excessive inflammatory response accompanies the
initial stages of severe infection and appears to contribute to associated organ system failure and death [1, 2,
3]. A number of immunomodulatory therapies aimed
at decreasing the dysregulated inflammatory response
have been examined in patients with sepsis (Table 1).
The published literature from 1985 to 2000 was reviewed in preparing this summary of the major approaches to modifying immunological pathways that
have been examined in clinical trials, using the key
words of sepsis, septic shock, endotoxemia, cytokines,
immunology, coagulation, tumor necrosis factor (TNF)
a, interleukin (IL) 1, and human.
Pathophysiology
Although microbial products, such as endotoxins, have
been targets for therapy in sepsis, a fundamental concept is that the constellation of abnormalities in these
patients is not due to the direct effect of the infectious
agent or its products but rather results from the patient's
own inflammatory response to infection. Initially, activation of complement was considered causal, particularly the fifth component of complement, which is a potent
neutrophil activator and produces a capillary hyperpermeability syndrome [4, 5]. The release of platelet-activating factor (PAF) was also thought to be responsible
for hypotension and organ system dysfunction in sepsis,
particularly since PAF is a potent hypotensive agent [6,
7, 8, 9, 10]. Using specific inhibitors of PAF, animals given lethal bacterial toxins survive [8, 9]. Similar results
were obtained when cyclo-oxygenase inhibitors were
administered to animals lethally challenged with endotoxin or bacteria, implicating cyclo-oxygenase products
as a contributing cause to septic shock [11, 12, 13].
In the absence of infection, high doses of TNF administered to animals induced circulatory collapse and
organ necrosis which were similar to those observed in
humans with septic shock [13]. Similar results were observed with high doses of IL-1 [13]. Injecting a combination of low doses of IL-1 plus TNF revealed synergistic
effects in inducing shock [13]. Neutralizing TNF activity
with antibodies [14, 15, 16] or soluble receptor fusion
constructs [17], or blocking IL-1 receptors [18, 19] was
effective in preventing death in animal models of lethal
bacteremia or endotoxemia. The findings in animal
models were confirmed when humans were injected
with either IL-1 or TNF as cancer chemotherapy [20,
21, 22], the most impressive physiological consequence
of which was the fall in blood pressure. The hypotension
was dose-dependent and, despite a short plasma halflife of less than 10 min, the biological consequences
could be observed for days. The logical clinical conclusion from these data was that reducing the biological effects of systemic TNF or IL-1 would reduce the risk of
dying from septic shock since these cytokines appeared
to be essential for the manifestation of the disease.
The biological basis for the development of a shocklike state after systemic IL-1 or TNF has been established at the molecular level. Both cytokines activate
the transcription of genes that increase the production
of small, potent, proinflammatory mediator molecules.
For example, IL-1 and TNF increase gene expression
and synthesis for phospholipase A2 (PLA2) leading to
increased PAF synthesis [23]. Similarly, an increase in
cyclo-oxygenase type II (COX-2) by IL-1 or TNF results
S 105
Table 1 Immunomodulatory
therapies examined in sepsis
Agent
Number of patients enrolled in each trial of the agent
Antiendotoxin antibodies
HA-1A
E5
543 [58], 2199 [61]
488 [60], 847 [59], 1102 [118]
Interleukin-1 receptor antagonist
99 [71], 893 [72], 696 [113]
Bradykinin antagonist
Anti-TNF-a monoclonal antibodies
Murine anti-TNF
Murine anti-TNF-a Fab2'fragments
251 [3], 504 [112]
971 [6], 564 [7], 1879 [10]
122 [76], 39 [3], 446 [3]
TNF receptor fusion proteins
p75 TNF receptor fusion protein
p55 TNF receptor fusion protein
141 [78]
498 [8], 1340 [79]
Platelet-activating factor antagonists
Ibuprofen (cyclo-oxygenase inhibitor)
262 [114], 668 [115]
29 [116], 30 [113], 455 [117]
in elevated levels of prostaglandin E2 [24, 25, 26] Nitric
oxide (NO) is a potent vasodilator and thought to be
primarily responsible for the hypotension and myocardial suppression in septic shock [27]. IL-1, TNF, and interferon-g, particularly the combination of the three, activate gene expression and synthesis of inducible NO
synthase (iNOS) [13, 28, 29, 30, 31].
Compared to IL-1 and TNF, no other cytokine has
been shown to reproduce the dramatic hypotension
and pathophysiological parameters of septic shock in
animals or in humans. This does not mean that other cytokines do not participate in the pathogenesis of the
event. For example, animals treated with neutralizing
antibodies to interferon-g are protected against lethal
endotoxemia [29]. Neutralizing antibodies to IL-8 in
models of inflammation reduce neutrophil infiltration
in the lung, kidneys, and myocardium [32]. Other cytokines, both proinflammatory, such as IL-12 [33] and
macrophage-inducing factor [34], and anti-inflammatory, such as IL-10 [35, 36, 37], as well as other mediators
of inflammation such as adhesion molecules [38, 39, 40,
41], also appear to have a role in modulating cellular
function in models of endotoxemia or bacteremia.
Antiendotoxin therapies
Antiendotoxin therapies, both nonspecific, such as intravenous immunoglobulins, and specific, such as human antiserum to heat killed Escherichia coli J5 and
murine (E5) and humanized (HA1A) antibodies directed against the lipid A component of endotoxin, have
been investigated in large populations of adult patients
with presumed sepsis [42, 43, 44, 45, 46, 47, 48, 49, 50,
51]. Although initial results were encouraging, particularly in patients with Gram-negative sepsis, larger clinical trials showed no benefit.
In part, the lack of effect of the monoclonal E5 and
HA1A antibodies may have been due to the inability
of these antibodies to effectively block the effects of endotoxin. In particular, when cultured macrophages are
first incubated with either E5 or HA1A, and endotoxin
added, there is no decrease in the release of either IL-1
or TNF by these cells [52]. These findings indicate that
neither of these antibodies is able to prevent endotoxin
from activating cells to produce proinflammatory cytokines. Given these in vitro results, it is in retrospect not
surprising that clinical trials with these antibodies were
also negative.
Bactericidal/permeability increasing (BPI) protein is
a member of a group of naturally occurring proteins
that bind to the lipid A portion of endotoxin [53]. BPI
is produced in neutrophils and stored in their primary
granules. In in vitro studies BPI effectively binds to lipopolysaccharide, prevents the LPS-induced inflammatory response from occurring (i.e., release of TNF-a
and IL-6 from cells), and significantly reduces Gramnegative bacteria viability [54, 55]. In animal models of
septic shock the administration of BPI significantly improved hemodynamics, lowered endotoxin levels, decreased markers of inflammation (serum levels of
TNF-a and IL-6), and improved survival when given
up to 30 min after an infusion of LPS or E. coli [56, 57].
In a small, phase I/II, open-labeled clinical trial in patients with severe meningococcal sepsis, BPI given by
continuous infusion improved survival compared to
that predicted by the Glasgow Meningococcal Prognostics Septicemia Score (96 % compared to < 70 %) [58].
Unfortunately, these results were unable to be confirmed in a larger, double-blind, randomized phase III
study of BPI in the therapy of pediatric patients with severe meningococcal disease.
There are several important issues that may limit the
utility of any antiendotoxin therapy, no matter how potent. First, high levels of endotoxin are associated primarily with Gram-negative bacterial infections, and ideally patients with a high probability of such infections
would therefore be targeted for antiendotoxin thera-
S 106
pies. However, it is very difficult to identify such patients since the clinical presentation of Gram-negative
infection is often not significantly different from that of
Gram-positive, or culture-negative sepsis. Second, endotoxin often initiates inflammatory cascades leading
to organ system dysfunction in septic patients. By the
time these patients are first seen and antiendotoxin therapy considered, cellular dysfunction may be established
and not amenable to correction by such therapy, unless
endotoxin continues to be present and to drive such
pathophysiological sequelae. Animal models investigating BPI and other antiendotoxin therapies have generally used pretreatment strategies. It is unclear whether
benefit can be shown in preclinical models or in the clinical setting when such therapies are administered several hours to days after the onset of endotoxemia or
Gram-negative sepsis.
Interleukin-1 receptor antagonist
The interleukin-1 receptor antagonist (IL-1ra) is a naturally occurring inhibitor of IL-1 which competitively
binds to the IL-1 receptor [18, 19]. In preclinical studies,
involving primarily rabbits and mice, infusion of IL-1ra
starting before or shortly after the onset of endotoxemia
or bacteremia improve survival [18, 19]. Interestingly, in
baboon studies [59, 60] IL-1ra therapy appears to have a
minimal effect in blunting endotoxemia-induced increases in IL-6, TNF-a, or circulating levels of TNF receptors, suggesting that in primates IL-1 does not have
a central role in the sepsis-induced inflammatory response.
Although a small (n = 99) nonblinded phase II trial
[61] suggested that IL-1ra can improve survival in septic
patients, two large subsequent phase III studies [62, 63]
were unable to demonstrate similar efficacy. In part,
the inability of IL-1ra to provide benefit to septic patients may be because IL-1 does not occupy a central,
pivotal role in perpetuating the inflammatory response
and producing organ system dysfunction in human sepsis, as was suggested by the baboon endotoxemia studies
[59, 60] reviewed above. Additionally, the failure of IL1ra to reduce mortality may reflect a timing issue, generic in sepsis studies. In particular, even if IL-1 is important in initiating a proinflammatory response, its role
may be minimal by the time the patients are recognized
and entered into a clinical trial. Indeed, very few septic
patients have elevated plasma levels of IL-1, and therefore the number of patients who truly had increased
IL-1 tissue expression at the time of enrollment into
the IL-1ra studies is unknown. Such information is of
obvious importance because therapies inhibiting IL-1
would be expected to have a beneficial effect only in patients with increased expression of this cytokine.
Anti-tumor necrosis factor therapies
The two major approaches taken to neutralizing TNF
have involved either monoclonal anti-TNF antibodies
or fusion protein constructs in which the extramembrane portion of the p55 (type I) or p75 (type II) TNF
receptor is joined to the Fc fragment of a human IgG1
antibody. Pretreatment of endotoxemic or bacteremic
animals with anti-TNF-a antibodies or TNF receptor fusion protein constructs results in clear improvements in
survival and amelioration of organ system dysfunction
[14, 15, 16, 17]. In some models of Gram-negative or
Gram-positive bacteremia, administration of anti-TNFa antibodies at the time of initiation of the bacteremic
insult or even shortly thereafter (i.e., within the first
hour) is still associated with a significant survival benefit
[64]. However, the use of such antibodies at later time
points in endotoxemic or bacteremic models does not
appear to be associated with any clear benefits.
Although several small studies [65, 66] have suggested that anti-TNF-a antibody therapy improved certain
physiological parameters, such as cardiac output, in septic patients, they were too small to detect any survival
benefit. The initial study powered to examine day 28
survival with such therapy was the North American Sepsis Trial (NORASEPT I), which examined a murine
IgG1 monoclonal antibody in the treatment of severe
sepsis and septic shock [67]. A total of 994 patients
were enrolled, of whom approximately one-half were
in shock at the time of randomization. Overall, there
was no statistically significant benefit associated with
anti-TNF therapy. However, in the prospectively defined subgroup of patients with septic shock a statistically significant reduction in mortality was present during
the first 2 weeks after administration of monoclonal
anti-TNF-a antibody compared to placebo. At day 28
after anti-TNF-a therapy the reduction in mortality
among septic shock patients was 17 % compared to
those receiving placebo. By contrast, no benefit was
found with anti-TNF-a therapy in patients not in shock
at study entry.
In the NORASEPT I shock patients the beneficial
effect of anti-TNF-a antibodies on survival appeared
within the first 24 h after enrollment; the greatest separation between the survival curves for placebo and
anti-TNF antibody-treated patients occurred during
this time [67]. Approximately 60 % of the placebo
deaths occurred within the first 3 days of the study.
Treatment with 7.5 mg/kg monoclonal anti-TNF-a antibodies was associated with a 49 % reduction in mortality
versus placebo at day 3 after study enrollment.
A second study, the International Sepsis Trial (INTERSEPT), using the same murine monoclonal antiTNF-a antibody as NORASEPT I, was undertaken in
14 primarily European countries [68]. Although the INTERSEPT study initially enrolled septic patients with
S 107
and without shock, after the results of NORASEPT I
were available, only shock patients were entered into
INTERSEPT. A total of 564 patients, of whom 420
were in septic shock, were enrolled. Day 28 mortality
was reduced by 14.5 % in patients who received 3 mg/
kg monoclonal anti-TNF-a antibody, with no reduction
in mortality found in those receiving 15 mg/kg. There
was no evidence of early survival benefit (i.e., within
the first 3 days after anti-TNF antibody infusion), a similar finding to that seen in NORASEPT I. Additionally,
whereas 60 % of the placebo deaths among patients in
shock occurred within the first 3 study days in NORASEPT I, fewer than 45 % of placebo deaths occurred
within this period in the INTERSEPT study.
A recently completed study (NORASEPT II) enrolled 1900 patients with septic shock and examined
the potential utility of 7.5 mg/kg the murine monoclonal
anti-TNF-a antibody [69]. No improvement in survival
was found in the actively treated group, all-cause mortality at day 28 being 40.3 % in monoclonal anti-TNF-a
antibody treated patients compared to 42.8 % in those
receiving placebo. Although the Acute Physiology and
Chronic Health Evaluation II scores, day 28 mortality
rates, sex ratio, and percentage of patients with one or
more organ failures present at baseline were similar in
NORASEPT I and NORASEPT II, there did appear
to be substantial differences in patient survival patterns.
Whereas more than 60 % of the deaths in the placebo
arm of NORASEPT I occurred in the first 3 days after
study entry, the mean time to death in the placebo group
of NORASEPT II was delayed, averaging 6.8 days.
These differences in survival may reflect improvements
between the two studies in the supportive care provided
to patients with septic shock, resulting in better survival
from the initial hypotensive episode and associated immediate complications, a period in which proinflammatory cytokine release, including that for TNF-a, may be
greatest. If advances in management have permitted
critically ill septic patients to better survive the initial
state of accelerated cytokine expression, this would diminish the efficacy of therapies aimed at modulating
the early proinflammatory response.
An additional concern in interpreting the NORASEPT II data revolves about the efficacy of the antiTNF antibody used. Even though only a minority of patients had detectable levels of circulating TNF-a at
baseline and posttreatment time points, review of posttreatment plasma TNF-a levels showed continued presence of circulating TNF-a in the antibody treated group.
Therefore a question remains as to the ability of the
anti-TNF antibody employed in the doses used in NORASEPT I, INTERSEPT, and NORASEPT II, to actually block cytokine activity.
The utility of administering F(ab')2 fragments of a
murine IgG3 monoclonal antibody to TNF-a has been
examined in patients with severe sepsis or septic shock
[70, 71]. There were 122 patients entered in the initial
clinical trial, and no increase in survival from sepsis for
the patients receiving anti-TNF treatment was detected
overall [70]. However, a retrospective stratification of
patients according to their plasma interleukin-6 (IL-6)
concentrations suggested beneficial effects for the drug
in patients (n = 37) with baseline levels greater than
1000 pg/ml. In patients with IL-6 levels greater than
1000 pg/ml, mortality decreased from 80 % in the placebo group to 35 % in patients who received the highest
dose (1 mg/kg) of the anti-TNF-a therapy. Two larger
unpublished studies in Europe and North America
have shown an aggregate reduction in mortality of approximately 3.5 % for patients receiving anti-TNF-a antibody fragments. These results were consistent with
those found in NORASEPT II [69], INTERSEPT [68],
and the p55 TNF receptor fusion protein [72] studies
which found no relationship between IL-6 levels and response to anti-TNF therapy.
Three clinical studies have reported results using soluble TNF receptor constructs as anti-TNF agents. In the
first of these clinical trials [73] the molecule used consisted of the extramembrane components of the human
type II (p75) receptor joined to the Fc portion of a human IgG1 antibody molecule [17]. Patients (n = 141)
with septic shock, with or without associated organ system dysfunction, were entered into the study. A significant dose-dependent increase in mortality was found in
patients treated with this p75 soluble TNF receptor construct, with mortality rising from 30 % in the placebo
group to 53 % in the patients treated with the highest
dose (1.5 mg/kg) of the anti-TNF compound.
The enhanced mortality associated with treatment
with the p75 TNF receptor molecule may be related to
the extremely high doses used in the study. Although
potency estimates are difficult to quantitate, soluble
TNF receptor fusion proteins appear to inactivate
TNF-a more than 50 times as effectively as the monoclonal antibodies [72], and therefore therapy with a
dose of 1.5 mg/kg of the p75 TNF receptor fusion protein would be expected to completely neutralize TNF-a
for a prolonged period, especially given the long halflife of the compound (> 60 h). TNF-a is an essential
component of normal inflammatory responses, and prolonged neutralization of its activity may have potent immunosuppressive effects leading to increased mortality.
Two clinical trials have examined the role of a p55
TNF receptor fusion protein construct in septic patients.
In the initial 498 patient study, separate randomization
lists were used for patients with severe sepsis with or
without early shock, shock and for those with refractory
septic shock [72]. The doses of the p55 TNF receptor
complex used in this study (0.008, 0.042, and 0.08 mg/
kg) were substantially lower than those administered in
the p75 TNF receptor complex clinical trial [73]. Therapy with 0.08 mg/kg of the p55 TNF receptor fusion pro-
S 108
tein complex, but not other doses, was associated with a
36 % reduction (p = 0.07) in day 28 mortality in the prospectively defined patient group with severe sepsis with
or without early septic shock. By contrast, no beneficial
effects were apparent with any dose of the p55 receptor
complex in patients with refractory septic shock.
Because of the apparent benefit of the p55 TNF receptor fusion protein in severe sepsis with or without
early septic shock, a 1340 patient, phase III study was
undertaken in this patient population [74]. No improvement in day 28 all cause mortality or in surrogate endpoints, such as organ failure scores, was found in patients treated with the p55 TNF receptor construct compared to placebo. Notably, the p55 TNF receptor fusion
protein used in this phase III study was from a different
batch than that in the phase II study, with differences in
glycosylation, and had slightly lower TNF neutralizing
ability, and therefore higher doses (approximately
0.125 mg/kg) were used. It is unknown what role if any
these alterations in molecular structure or binding potency played in the different outcomes between the
phase II and phase III clinical trials.
Phospholipase A2 inhibition
Nonpancreatic PLA2 (sPLA2) is released into the systemic circulation after endotoxin exposure and reproduces the hemodynamic profile of severe sepsis when
administered intravenously to animals [75]. sPLA2 catalyzes the hydrolysis of membrane phospholipids resulting in the production of PAF as well as other lysophospholipids [75]. PLA2 levels are increased in the serum
of humans with severe sepsis, without significant increases in patients hospitalized with other diagnoses or
trauma patients. Levels are correlated with hypotensive
episodes and survival in patients with severe sepsis [75,
77, 78, 79, 80]. These results suggest that inhibition of
sPLA2 would be of benefit in patients with severe sepsis.
Inhibitors of sPLA2 activity decrease PAF and leukotriene levels and improve survival in murine models of
septic shock [76]. A human phase II trial of the use of a
sPLA2 inhibitor in severe sepsis has recently been completed and demonstrated no benefit in all enrolled patients. Results of subgroup analyses should be available
in the near future.
Nitric oxide inhibition
NO, previously called endothelium-derived relaxing
factor, is synthesized by endothelial cells by way of
NOS, can be released into the systemic circulation, and
then can function to regulate blood flow to tissues [81].
Exposure to endotoxin increases endothelial secretion
of NO [82, 83], predominately by upregulating the in-
ducible form of NOS (iNOS) [84, 85]. Specific inhibitors
of iNOS, in particular NG-methyl-L-arginine (LNAME), have reduced the hypotensive response in animals infused with endotoxin, but have not improved
their survival [85, 86]. The adverse effects of iNOS inhibitors, including a decrease in cardiac output and an
increase in pulmonary artery pressure, have resulted in
questions regarding their potential benefit in humans
with severe sepsis [81, 87]. Additionally, continuous infusion of L-NAME resulted in an increase in mortality
in an animal model of septic shock [87].
Two small, open-label studies of L-NAME infusion
have reported reproducible and sustained increases in
mean arterial pressure in humans with severe sepsis
[81, 88]. However, both studies were associated with increases in pulmonary artery pressures and a fall in cardiac output requiring an increase in dobutamine infusion
in one study [81, 88]. The fall in cardiac output was
greatest in the high-dose group (20 mg/kg per hour)
and was associated with electrocardiographic evidence
of cardiac ischemia in 27 % of the patients in this group
[88]. A recent phase III clinical trial utilizing continuous
L-NAME infusion in patients with severe sepsis was discontinued due to an increase in adverse effects, including statistically significantly increased mortality in the
L-NAME group [89].
Anticoagulation
The coagulation system is activated in animal models
and in humans with severe sepsis as evidenced by the
presence of intravascular thrombi in vessels on tissue
specimens and the frequent occurrence of disseminated
intravascular coagulation (DIC) [90, 91]. The activation
of the coagulation system is associated with decreased
levels of fibrinogen, increased levels of activated factor
X, and increased levels of tissue factor and its inhibitor,
tissue factor pathway inhibitor (TFPI) [90]. Additionally, activation of coagulation in the setting of severe infection appears to potentiate proinflammatory responses, primarily through the activation of endothelial cells,
which then produce inflammatory mediators, including
cytokines such as TNF-a. Several studies have investigated modulation of the coagulation system including
the administration of TFPI, anti-Xa, and activated protein C (APC).
TFPI is a naturally occurring, circulating protein
which can inhibit the procoagulant effects of tissue factor by binding to factor Xa and then to the tissue factor±VIIa complex [90, 92]. Circulating levels of TFPI increase after exposure to endotoxin [90, 92]. A phase II
study investigating the infusion of TFPI in patients with
severe sepsis was recently completed with a trend toward clinical benefit [93]. A large international phase
III study of TFPI in severe sepsis is presently underway.
S 109
Blockade of factor Xa in the extrinsic coagulation
pathway has been investigated in animal models of severe sepsis. In animals administered endotoxin, dansyl
glutamyl-glycyl-arginyl chloromethyl ketone-treatedXa, a factor Xa inhibitor, prevented endotoxemia induced DIC but did not affect survival [94]. Early human
trials are underway investigating the potential benefit of
factor Xa inhibition in severe sepsis.
APC is an anticoagulant formed when protein C is
cleaved by thrombin [95, 96]. In children and adults
with severe meningococcemia and/or purpura fulminans, serum APC levels are depressed and are correlated
with clinical outcome [97, 98]. Several small nonrandomized trials utilizing infusions of protein C
(50±100 IU/kg every 6 h) have been performed in adults
and children with severe meningococcal disease and
purpura fulminans [96, 99, 100, 101]. Although these
studies were not powered to detect a difference in mortality, protein C infusion normalized sera protein C levels, increased fibrinogen levels and was accompanied
by resolution of DIC [96, 99, 100, 101]. A phase II trial
of APC in sepsis showed a nonsignificant trend towards
improved survival in APC-treated patients [95]. A large
phase III study of APC in sepsis was recently stopped
after approximately 1500 patients were enrolled because of efficacy associated with APC therapy [101 a].
Such positive results with APC suggest that infusions of
APC will become part of the standard therapy in severely ill septic patients.
Definitions and entry criteria
Entry criteria for sepsis trials have been designed primarily to include patients with clinical evidence of infection associated with the recent development of organ
system dysfunction believed to be due to this infectious
process [102, 103]. Because of the perceived need to enroll patients early in their clinical course, positive microbiological cultures have not been required. Indeed, in a
number of recent studies only a minority of patients
had positive blood cultures, with bacteremia present in
about 30 % [63, 67, 68, 69].
The definitions of sepsis and septic shock used in
most clinical trials did not include consideration of the
length of time that the infective process had been present, nor of its anatomic site. Such a classification may
be particularly important since animal studies have
shown differing patterns of response to anticytokine
therapies, such as anti-TNF-a monoclonal antibodies,
for intra-abdominal infections compared to bacteremias, and for rapidly initiated infectious processes,
such as acute bacteremia, compared to more slowly developing infections, such as peritonitis [104]. No microbiological classifications were used prospectively in clinical trials of immunomodulatory agents even though re-
sponses to anti-inflammatory therapies may differ between Gram-positive and Gram-negative infections,
and also may be more effective in patients with documented infections [73].
Even though recent clinical trials have targeted specific mediators, including endotoxin, IL-1, or TNF-a,
which were postulated to have a pivotal role in the inflammatory cascade leading to organ system failure
and death in sepsis, the actual presence of excessive levels of these mediators was not required for entry into
the study. Rather, clinical criteria, such as the presence
of one or more organ system failures, with or without
shock, were used. Although elevation in the mediator
of interest was postulated to accompany organ system
dysfunction in septic patients, data gathered from the
IL-1ra and anti-TNF clinical trials suggested that such
a correlation was present in only a minority of cases.
For example, in the North American Sepsis Trial (NORASEPT II) of a murine monoclonal antibody for septic shock only 40 % of patients had detectable circulating TNF-a levels at the time of enrollment [69]. In contrast, decreased levels of protein C were present in approximately 90 % of patients with signs of severe infection with associated organ failures, indicating that the
Consensus Conference criteria were highly useful in
identifying patients with deficiency of protein C, who
would be an appropriate target population for APC
therapy.
The lack of a requirement to recruit demonstrably infected patients with elevated plasma levels of endotoxin
or of the cytokine of interest may therefore have adversely affected outcome in trials examining immunomodulatory agents in sepsis. For example, improved efficacy of therapies, such as monoclonal anti-TNF-a antibodies, was found when blinded data safety monitoring
committees eliminated patients without clear evidence
of infection [105]. In the NORASEPT II trial of murine
monoclonal anti-TNF-a antibodies there was an 18 %
relative reduction in day 28 mortality compared to placebo in patients with detectable circulating TNF-a levels at the time of enrollment. By contrast, no effect of
such anti-TNF therapy was seen in patients without elevated circulating TNF-a concentrations.
Future directions
Although a meta-analysis combining most of the clinical
trials using anti-inflammatory agents has suggested that
benefit in survival could be achieved with such therapies, the magnitude of such an effect was small [3]. By
contrast, therapies directed against specific proinflammatory cytokines such as TNF-a and IL-1 have produced remarkable clinical response in diseases such as
rheumatoid arthritis and Crohn's disease [106, 107,
108].
S 110
The failure of immunomodulatory therapies to improve outcome in sepsis raises several questions about
drug development strategies and the design of clinical
trials in this area. First, there are reasons to believe
that some of the agents tested were insufficiently potent
to block the mediator of interest, or had properties directed at mediators which were not of central importance in determining clinical outcome. It is therefore
not surprising that clinical studies with these agents
were negative. Second, most of the immunomodulatory
therapies used in the treatment of patients with sepsis
showed impressive efficacy in animal models. Their subsequent failure in clinical trials raises concerns about the
relevance of preclinical experimental models. Specifically, differences between the biochemical and immunological responses of patients with a clinical diagnosis of
sepsis and animals with known bacterial infections or
endotoxemia may explain the divergence of results between experimental and clinical studies. Third, the fact
that agents such as anti-TNF-a antibodies are clinically
effective in the setting of rheumatoid arthritis and
Crohn's disease raises questions about the heterogeneous nature of patients with sepsis. Overly broad definitions for sepsis may have diluted out any effect that
such therapies could have in more clinically limited patient populations. Thus, whereas rheumatoid arthritis
or Crohn's disease are well described entities with specific immunological, radiological, and pathophysiological diagnostic criteria, the defining features of sepsis
are based mainly on organ system dysfunction developing in the appropriate clinical setting.
Almost all clinical trials of immunomodulatory therapies have enrolled patients on the basis of clinical evidence of infection with associated organ system dysfunction, rather than requiring evidence that the immunological abnormality of interest is present. Such inclusion criteria may permit the entry of excessively heterogeneous patient populations and thereby prevent detection of beneficial effects that might be apparent in more
homogeneous subgroups. Positive results associated
with anti-TNF therapy in preventing Jarisch-Herxheimer reactions [109] suggest that there are specific
groups of infected patients who do respond to immunomodulatory agents.
Classification of infected patients based on demonstrable abnormalities in immunological or biochemical
pathways may permit inclusion into therapeutic studies
at earlier points in their clinical course, before organ
dysfunction develops and becomes irreversible. In this
case, patients with clinical evidence of infection and alteration in plasma levels of the mediator of interest,
such as TNF-a, would be eligible for study enrollment.
Such a classification scheme, in addition to allowing appropriate targeting of patient populations who may
benefit from a specific therapeutic agent, would also
mean that the development of clinically relevant organ
system dysfunction can be used as an endpoint, rather
than as entry criteria for clinical trials.
There are substantial differences in the intracellular
signaling cascades initiated by Gram-negative and
Gram-positive infections [110], as well as by the site of
infection. It would therefore appear reasonable to limit
the nature and site of the infection in patients enrolled
in future trials of immunomodulatory agents. For example, an immunomodulatory therapy would initially be
investigated only in patients with evidence of meningococcemia or pneumococcal pneumonia.
The use of mechanistic definitions to define patient
populations who are at risk for infection-initiated organ
system dysfunction lacks the reductionist simplicity of
the Consensus Conference definitions of sepsis [101,
102], which generate large numbers of patients presenting with similar constellations of clinical findings. However, mechanistic definitions should provide more homogeneous groups of patients with activation of similar
immunological or biochemical pathways, at earlier stages in their clinical course and who should respond to interruption of immunological or other cascades. Although the approach taken in almost all clinical trials
has been to block completely the effects of the mediator
of interest, it may be more appropriate to modulate,
rather than to ablate biochemical responses [111]. Improved techniques to monitor immunological markers
of inflammatory and host defense responses will be important in assessing the effects of future therapies on
central mechanisms contributing to organ dysfunction
in sepsis.
The extent to which such a mechanistic approach is
beneficial will depend upon how important the mediators generated are in contributing to the subsequent
clinical course, particularly the development of organ
system dysfunction and mortality. Using more stringent
definitions in clinical trials would of course restrict the
numbers of patients eligible for a specific therapy until
subsequent studies examine larger at risk populations.
However, unless benefit can first be shown in a rigorously defined patient population in which there is specific
evidence of mediator activation, there is little reason to
anticipate that efficacy will be achieved in larger, more
heterogeneous patient groups.
S 111
References
1. Abraham E. Therapies for sepsis
(1997) Emerging therapies for sepsis
and septic shock. West J Med 166:
195±200
2. Parrillo JE, Parker MM, Natanson C,
Suffredini AF, Danner RL, Cunnion
RE, Ognibene FP (1990) Septic shock
in humans. Ann Intern Med 113:
227±242
3. Zeni F, Freeman B, Natanson C (1997)
Anti-inflammatory therapies to treat
sepsis and septic shock: a reassessment. Crit Care Med 25: 1095±1100
4. Mollnes TE, Redl H, Hogasen K,
Bengtsson A, Garred P, Speilberg L,
Lea T, Oppermann M, Gotze O, Schlag G (1993) Complement activation
in septic baboons detected by neoepitope-specific assays for C3b/iC3b/C3c,
C5a and the terminal C5b-9 complement complex (TCC). Clin Exp Immunol 91: 295±300
5. Granowitz EV, Porat R, Gelfand JA,
Wolff SM, Dinarello CA (1994) Administration of low-dose endotoxin to
healthy humans increases C5a binding
to circulating neutrophils. J Infect Dis
169: 480±482
6. Coughlan AF, Hau H, Dunlop LC,
Berndt MC, Hancock WW (1994) Pselectin and platelet-activating factor
mediate initial endotoxin-induced
neutropenia. J Exp Med 179: 329±334
7. Dhainaut JF, Tenaillon A, Le Tulzo Y,
Schlemmer B, Solet JP, Wolff M, Holzapfel L, Zeni F, Dreyfuss D, Mira JP,
et al (1994) Platelet-activating factor
receptor antagonist BN 52021 in the
treatment of severe sepsis: a randomized,
double-blind,
placebo-controlled, multicenter clinical trial. BN
52021 Sepsis Study Group. Crit Care
Med 22: 1720±1728
8. Koltai M, Hosford D, Braquet PG
(1993) Platelet-activating factor in
septic shock. New Horiz 1: 87±95
9. Kruse-Elliott KT, Albert DH, Summers JB, Carter GW, Zimmerman JJ,
Grossman JE (1996) Attenuation of
endotoxin-induced pathophysiology
by a new potent PAF receptor antagonist. Shock 5: 265±273
10. Ishii S, Nagase T, Tashiro F, Ikuta K,
Sato S, Waga I, Kume K, Miyazaki J,
Shimizu T (1997) Bronchial hyperreactivity, increased endotoxin lethality
and melanocytic tumorigenesis in
transgenic mice overexpressing platelet-activating factor receptor. EMBO
J 16: 133±142
11. Coran AG, Drongowski RA, Paik JJ,
Remick DG (1992) Ibuprofen intervention in canine septic shock: reduction of pathophysiology without decreased cytokines. J Surg Res 53:
272±279
12. Mansilla-Rosello A, Ferron-Orihuela
JA, Ruiz-Cabello F, Garrote-Lara D,
Fernandez-Mondejar E, DelgadoCarrasco ML (1997) Differential effects of IL-1b and ibuprofen after endotoxic challenge in mice. J Surg Res
67: 199±204
13. Okusawa S, Gelfand JA, Ikejima T,
Connolly RJ, Dinarello CA (1988) Interleukin 1 induces a shock-like state
in rabbits. Synergism with tumor necrosis factor and the effect of cyclooxygenase inhibition. J Clin Invest 81:
1162±1172
14. Tracey KJ, Fong Y, Hesse DG, Manogue KR, Lee AT, Kuo GC, et al
(1987) Anti-cachectin/TNF monoclonal antibodies prevent septic shock
during lethal bacteremia. Nature 330:
662±664
15. Fong Y, Tracey KJ, Moldawer LLL,
et, et al (1989) Antibodies to cachectin/TNF reduce interleukin-1 and interleukin-6 appearance during lethal
bacteremia. J Exp Med 170:
1627±1633
16. Hinshaw LB, Tekamp-Olson P, Chang
AC, et al (1990) Survival of primates
in LD 100 septic shock following therapy with antibody to tumor necrosis factor (TNF). Circ Shock 30: 279±292
17. Van Zee KJ, Moldawer LL, Oldenburg HAS, Thompson WZ, Stackpole
SA, Montegut WJ, Rogy MA, Meschter C, Gallati H, Schiller CD, et al
(1996) Protection against lethal Escherichia coli bacteremia in baboons
(Papio anubis) by pretreatment with a
55-kDa TNF receptor (CD120a)-Ig fusion protein, Ro 45-2081. J Immunol
156: 2221±2230
18. Ohlsson K, Bjork P, Bergenfeldt M,
Hageman R, Thompson RC (1990) Interleukin-1 receptor antagonist reduces mortality from endotoxin shock.
Nature 348: 550±552
19. Wakabayashi G, Gelfand JA, Burke
JF, Thompson RC, Dinarello CA
(1991) A specific receptor antagonist
for interleukin-1 prevents Escherichia
coli-induced shock. FASEB J 5:
338±343
20. Chapman PB, Lester TJ, Casper ES, et
al (1987) Clinical pharmacology of recombinant human tumor necrosis factor in patients with advanced cancer.
J Clin Oncol 5: 1942±1951
21. Smith JW, Urba WJ, Curti BD, et al
(1992) The toxic and hematologic effects of interleukin-1 alpha administered in a phase I trial to patients with
advanced malignancies. J Clin Oncol
10: 1141±1152
22. Van der Poll T, Buller HR, ten Cate
HT (1990) Activation of coagulation
after administration of TNF to normal
subjects. N Engl J Med 322: 1622±1627
23. Endo S, Inada K, Nakae H, et al (1995)
Plasma levels of type II phospholipase
A2 and cytokines in patients with sepsis. Res Commun Mol Pathol Pharmacol 90: 413±421
24. Diaz A, Chepenik KP, Korn JH, Reginato AM, Jimenez SA (1998) Differential regulation of cyclooxygenases 1
and 2 by interleukin-1b, tumor necrosis factor-a, and transforming growth
factor-b1 in human lung fibroblasts.
Exp Cell Res 241: 222±229
25. Newton R, Stevens DA, Hart LA,
Lindsay M, Adcock IM, Barnes PJ
(1997) Superinduction of COX-2
mRNA by cycloheximide and interleukin-1b involves increased transcription and correlates with increased NFk B and JNK activation. FEBS Lett
418: 135±138
26. Guan Z, Baier LD, Morrison AR
(1997) p38 mitogen-activated protein
kinase down-regulates nitric oxide
and up-regulates prostaglandin E2
biosynthesis stimulated by interleukin-1b. J Biol Chem 272: 8083±8089
27. Moncada S, Palmer RMJ, Higgs EA
(1991) Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 43: 109±142
28. Dinarello CA (1996) Biological basis
for interleukin-1 in disease. Blood 87:
2095±2147
29. Heremans H, van Damme J, Dillen C,
Dikman R, Billiau A (1990) Interferon-g, a mediator of lethal lipopolysaccharide-induced Shwartzman-like
shock in mice. J Exp Med 171:
1853±1861
30. Huang S, Hendriks W, Althage A, et al
(1993) Immune response in mice that
lack the interferon-gamma receptor.
Science 259: 1742±1745
S 112
31. Car BD, Eng VM, Schnyder B, Ozmen
L, Huang S, Gallay P, Heumann D,
Aguet M, Ryffel B (1994) Interferon
gamma receptor deficient mice are resistant to endotoxic shock. J Exp Med
179: 1437±1444
32. Herada A, Mukaida N, Matsushima K
(1996) Interleukin-8 as a novel target
for intervention therapy in acute inflammatory diseases. Mol Med Today
2: 482±489
33. Zisman DA, Kunkel SL, Strieter RM,
Gauldie J, Tsai WC, Bramson J, Wilkowski JM, Bucknell KA, Standiford
TJ (1997) Anti-interleukin-12 therapy
protects mice in lethal endotoxemia
but impairs bacterial clearance in murine Escherichia coli peritoneal sepsis.
Shock 8: 349±356
34. Calandra T, Bernhagen J, Metz CN,
Spiegel LA, Bacher M, Donnelly T,
Cerami A, Bucala R (1995) MIF as a
glucocorticoid-induced modulator of
cytokine production. Nature 377:
68±71
35. Standiford TJ, Strieter RM, Lukacs
NW, Kunkel SL (1995) Neutralization
of IL-10 increases lethality in endotoxemia. J Immunol 155: 2222±2229
36. Howard M, Muchamuel T, Andrade S,
Menon S (1993) Interleukin 10 protects mice from lethal endotoxemia.
J Exp Med 177: 1205±1208
37. Gerard C, Bruyns C, Marchant A,
Abramowicz D, Vandenabeele P,
Delvaux A, Fiers W, Goldman M,
Velu T (1993) Interleukin 10 reduces
the release of tumor necrosis factor
and prevents lethality in experimental
endotoxemia. J Exp Med 177: 547±550
38. Xu H, Gonzalo JA, St Pierre Y, Williams IR, Kupper TS, Cotran RS,
Springer TA, Gutierrez-Ramos JC
(1994) Leukocytosis and resistance to
septic shock in intercellular adhesion
molecule 1-deficient mice. J Exp Med
180: 95±109
39. Sessler CN, Windsor AC, Schwartz M,
Watson L, Fisher BJ, Sugerman HJ,
Fowler AA (1995) Circulating ICAM1 is increased in septic shock. Am J
Respir Crit Care Med 151: 1420±1427
40. Watanabe S, Mukaida N, Ikeda N,
Akiyama M, Harada A, Nakanishi I,
Nariuchi H, Watanabe Y, Matsushima
K (1995) Prevention of endotoxin
shock by an antibody against leukocyte integrin beta 2 through inhibiting
production and action of TNF. Int Immunol 7: 1037±1046
41. Kayal S, Jais JP, Aguini N, Chaudiere
J, Labrousse J (1998) Elevated circulating E-selectin, intracellular adhesion molecule 1, and von Willebrand
factor in patients with severe infection.
Am J Respir Crit Care Med 157:
776±784
42. Nydegger UE (1997) Sepsis and polyspecific intravenous immunoglobulins.
J Clin Apheresis 12: 93±99
43. Werdan K, Pilz G (1996) Supplemental immune globulins in sepsis: a critical appraisal. Clin Exp Immunol 104
[Suppl 1]:83±90
44. Ziegler EJ, McCutchan JA, Fierer J,
Glauser MP, Sadoff JC, Douglas H,
Braude AI (1982) Treatment of gramnegative bacteremia and shock with
human antiserum to a mutant Escherichia coli. N Engl J Med 307: 1225±1230
45. Zanetti G, Glauser MP, Baumgartner J
(1991) Use of immunoglobulins in prevention and treatment of infection in
critically ill patients: review and critique. Rev Infect Dis 13: 971±972
46. Calandra T, Glauser MP, Schellekens
J, Verhoef J (1988) Treatment of
gram-negative septic shock with human IgG antibody to Escherichia coli
J5: a prospective, double-blind, randomized trial. J Infect Dis 158:
312±319
47. J5 Study group (1992) Treatment of severe infectious purpura in children
with human plasma from donors immunized with Escherichia coli J5: a
prospective double-blind study. J Infect Dis 165: 695±701
48. The HA-1A Sepsis Study Group
(1991) Treatment of gram-negative
bacteremia and septic shock with HA1A human monoclonal antibody
against endotoxin: a randomized, double-blind placebo-controlled trial.
N Engl J Med 324: 429±436
49. Bone RC, Balk RA, Fein AM, Perl
TM, Wenzel RP, Reines HD, Quenzer
RW, Iberti TJ, MacIntyre N, Schein
RM (1995) A second large controlled
clinical study of E5, a monoclonal antibody to endotoxin: results of a prospective, multicenter, randomized,
controlled trial. Crit Care Med 23:
994±1006
50. Greenman RL, Schein RM, Martin
MA, et al (1991) A controlled clinical
trial of E5 murine monoclonal IgM antibody to endotoxin in the treatment of
gram-negative sepsis. JAMA 266:
1097±1102
51. McCloskey RV, Straube RC, Sanders
C, Smith SM, Smith CR (1994) Treatment of septic shock with human monoclonal antibody HA-1A. A randomized, double-blind, placebo-controlled
trial. CHESS Trial Study Group. Ann
Intern Med 121: 1±5
52. Warren HS, Amato SF, Fitting C,
Black KM, Loiselle PM, Pasternack
MS, Cavaillon JM (1993) Assessment
of ability of murine and human antilipid A monoclonal antibodies to bind
and neutralize lipopolysaccharide.
J Exp Med 177: 89±97
53. Tobias PS, Mathison JC, and Ulevitch
R (1988) A family of lipopolysaccharide binding proteins involved in responses to gram-negative sepsis.
J Biol Chem 263: 13479±13481
54. Marra M, Wilde C, Griffith J, Snable J,
Scott R (1990) Bactericidal/permeability-increasing protein has endotoxinneutralizing activity. J Immunol 144:
662±666
55. Weiss J, Elsbach P, Shu C, Castillo J,
Grinna L, Horwitz A, Theofan G
(1992) Human bactericidal/permeability-increasing protein and a recombinant NH2-terminal fragment cause
killing of serum-resistant gram-negative bacteria in whole blood and inhibit tumor necrosis factor release induced by the bacteria. J Clin Invest
90: 1122±1130
56. Jin H, Yang R, Marsters S, Ashkenazi
A, Bunting S, Marra MN, Scott RW,
Baker JB (1995) Protection against endotoxic shock by bactericidal/permeability-increasing protein in rats.
J Clin Invest 95: 1947±1952
57. Rogy M, Oldenburg HS, Calvano SE,
Montegut WJ, Stackpole SA, Van Zee
KJ, Marra M, Scott R, Seilhammer JJ,
Moldawer LL, Lowry SF (1994) The
role of bactericidal/permeability-increasing protein in the treatment of
primate bacteremia and septic shock.
J Clin Immunol 14: 120±133
58. Giroir BP, Quint PA, Barton P, Kirsch
EA, Kitchen L, Goldstein B, Nelson
BJ, Wedel NJ, Carroll SF, Scannon PJ
(1997) Preliminary evaluation of recombinant amino-terminal fragment
of human bactericidal/permeability-increasing protein in children with severe meningococcal sepsis. Lancet
350: 1439±1443
S 113
59. Hawes AS, Fischer E, Marano MA,
Van Zee KJ, Rock CS, Lowry SF, Calvano SE, Moldawer LL (1993) Comparison of peripheral blood leukocyte
kinetics after live Escherichia coli, endotoxin, or interleukin-1 alpha administration. Studies using a novel interleukin-1 receptor antagonist. Ann
Surg 218: 79±90
60. Fischer E, Marano MA, Van Zee KJ,
Rock CS, Hawes AS, Thompson WA,
DeForge L, Kenney JS, Remick DG,
Bloedow DC, et al (1992) Interleukin1 receptor blockade improves survival
and hemodynamic performance in Escherichia coli septic shock, but fails to
alter host responses to sublethal endotoxemia. J Clin Invest 89: 1551±1557
61. Fisher CJ Jr, Slotman GJ, Opal SM,
Pribble JP, Bone RC Emmanuel G,
Ng D, Bloedow DC, Catalano MA
(1994) Initial evaluation of human recombinant interleukin-1 receptor antagonist in the treatment of sepsis syndrome: a randomized, open-label, placebo-controlled multicenter trial. The
IL-1RA Sepsis Syndrome Study
Group. Crit Care Med 22: 12±21
62. Fisher CJ Jr, Dhainaut JF, Opal SM,
Pribble JP, Balk RA, Slotman GJ,
Iberti TJ, Rackow EC, Shapiro MJ,
Greenman RL, et al (1994) Recombinant human interleukin 1 receptor antagonist in the treatment of patients
with sepsis syndrome. Results from a
randomized, double-blind, placebocontrolled trial. Phase III rhIL-1ra
Sepsis Syndrome Study Group.
JAMA 271: 1836±1843
63. Opal SM, Fisher CJ, Pribble JP, Dhainaut J-F, VincentJ-L, Brase R, Lowry
SF, Sadoff JC, Slotman GJ, Levy H,
Balk RA, Shelly MP, LaBrecque JF,
Lookabaugh J, Donovan H, Baughman R, Norman J, DeMaria E, Matzel
K, Abraham E, Seneff M, and the Interleukin-1 Receptor Antagonist Sepsis Investigator Group (1997) The confirmatory interleukin-1 receptor antagonist trial in severe sepsis: a phase
III randomized, double-blind, placebo-controlled, multicenter trial. Crit
Care Med 25: 1115±1124
64. Hinshaw LB, Emerson TE Jr, Taylor
FB Jr, Chang ACK, Duerr M, Peer
GT, Flournoy DJ, White GL, Kosanake SD, Murray CK, Xu R, Passey
RB, Fournel MA (1992) Lethal S. aureus shock in primates: prevention of
death with anti-TNF antibody. J Trauma 33: 568±573
65. Vincent JL, Bakker J, Marecaux G,
Schandene L, Kahn RJ, Dupont E
(1992) Administration of anti-TNF antibody improves left ventricular function in septic shock patients. Results
of a pilot study. Chest 101: 810±815
66. Dhainaut J-F A, Vincent J-L, Richard
C, Lejeune P, Martin C, Fierobe L,
Stephens S, Ney UM, Sopwith M
(1995) CDP571, a humanized antibody
to human tumor necrosis factor-a:
safety, pharmacokinetics, immune response, and influence of the antibody
on cytokine concentrations in patients
with septic shock. Crit Care Med 23:
1461±1469
67. Abraham E, Wunderink R, Silverman
H, et al (1995) Monoclonal antibody
to human tumor necrosis factor alpha
(TNF MAb): efficacy and safety in patients with the sepsis syndrome.
JAMA 273: 934±941
68. Cohen J, Carlet J (1996) INTERSEPT:
an international, multicenter, placebocontrolled trial of monoclonal antibody to human tumor necrosis factora in patients with sepsis. Crit Care
Med 24: 1431±1440
69. Abraham E, Anzueto A, Gutierrez G,
Tessler S, San Pedro G, Wunderink R,
Dal Nogare A, Nasraway S, Berman
S, Cooney R, Levy H, Baughman R,
Rumbak M, Light RB, Poole L, Allred
R, Constant J, Pennington J, Porter S,
for the NORASEPT II Study Group
(1998) Monoclonal antibody to human
tumor necrosis factor alpha (TNF
Mab) in the treatment of patients with
septic shock: a multi-center, placebo
controlled, randomized, double-blind
clinical trial. Lancet 351: 929±933
70. Reinhart K, Wiegand-Lohnert C,
Grimminger F, et al (1996) Assessment of the safety and efficacy of the
monoclonal anti-tumor necrosis factor
antibody-fragment, MAK 195F, in patients with sepsis and septic shock: a
multicenter, randomized, placebo-controlled, dose-ranging study. Crit Care
Med 24: 733±742
71. Panacek E (2000) Results of the Monarcs Study. American Thoracic Society
International Conference, ªThe Heart
in Sepsisº, May 2000
72. Abraham E, Glauser MP, Butler T, et
al (1997) p55 tumor necrosis factor receptor fusion protein in the treatment
of patients with severe sepsis and septic shock. JAMA 277: 1531±1534
73. Fisher CJ Jr, Agosti JM, Opal SM, et al
(1996) Treatment of septic shock with
the tumor necrosis factor receptor Fc
fusion protein. N Engl J Med 334:
1697±1702
74. Abraham E, Reynaert M, Lew D, Pingleton S, Butler I, Dugernier I, Margolis B, Kudsk K, Zimmerli W, Anderson
P, Laterre PF, Lesslauer W, Cooper P,
Burdeska A, Passe S, Modi M, Leighton A, Salgo M, Van der Auwera P,
for the Tenefuse Study Group (2001)
Lenercept (p55-Tumor Necrosis Factor Receptor Fusion Protein, Ro 452081, Tenefuse) patients with severe
sepsis or early septic shock. A randomized double-blind placebo-controlled
multicenter phase III trial with
1342 patients. Crit Care Med (in press)
75. Vadas P (1984) Elevated plasma phospholipase A2 levels: correlation with
the hemodynamic and pulmonary
changes in gram-negative septic shock.
J Lab Clin Med 104: 873±881
76. Marshall LA, Hall RH, Winkler JD,
Badger A, Bolognese B, Roshak A,
Flamberg PL, Sung CM, ShabotFletcher M, Adams JL, Mayer RJ
(1995) SB 203347, an inhibitor of 14
kDa phospholipase A2, alters human
neutrophil arachidonic acid release
and metabolism and prolongs survival
in murine endotoxin shock. J Pharmacol Exp Ther 274: 1254±1262
77. Vadas P, Pruzanski W, Stefanski E,
Sternby B, Mustaqrd R, Bohnen J,
Fraser I, Farewell V, Bombardier C
(1988) Pathogenesis of hypotension in
septic shock: correlation of circulating
phospholipase A2 levels with circulatory collapse. Crit Care Med 16: 1±7
78. Endo S, Inada K, Nakae H, Takakuwa
T, Yamada Y, Suzuki T, Taniguchi S,
Yoshida M, Ogawa M, Teraoka H
(1995) Plasma levels of type II phospholipase A2 and cytokine in patients
with sepsis. Res Commun Mol Pathol
Pharmacol 90: 413±421
79. Sorenson J, Kald B, Tagesson C, Lindahl M (1994) Platelet-activating factor and phospholipase A2 in patients
with septic shock and trauma. Intensive Care Med 20: 555±561
80. Nyman KM, Uhl W, Forsstrom J,
Buchler M, Beger HG, Nevalainen TJ
(1996) Serum phospholipase A2 in patients with multiple organ failure.
J Surg Res 60: 7±14
81. Avontuur JA, Tutein-Nolthenius RP,
Bodegom JW, Bruining H (1998) Prolonged inhibition of nitric oxide synthesis in severe septic shock: a clinical
study. Crit Care Med 26: 660±667
82. Salvemini D, Korbut R, Anggard E,
Vane J (1990) Immediate release of a
nitric oxide-like factor from bovine
aortic endothelial cells by Eschericha
coli lipopolysaccharide. Proc Natl
Acad Sci USA 87: 2593±2597
S 114
83. Evans T, Carpenter A, Kinderman H
(1993) Evidence of increased nitric oxide production in patients with the sepsis syndrome. Circ Shock 41: 77±81
84. Moncada S, Palmer RM, Higgs EA
(1991) Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 43: 109±142
85. Kilbourn RG, Jubran A, Gross SS,
Griffith OW, Levi R, Adams J, Lodato
RF (1990) Reversal of endotoxin-mediated shock by NG-methyl-L-arginine, an inhibitor of nitric oxide synthesis. Biochem Biophys Res Commun
172: 1132±1138
86. Kilbourn RG, Gross SS, Jubran A,
Adams J, Griffith OW, Levi R, Lodato
RF (1990) NG-methyl-L-arginine inhibits tumor necrosis factor-induced
hypotension: implications for the involvement of nitric oxide. Proc Natl
Acad Sci USA 87: 3629±3632
87. Cobb JP, Natanson C, Hoffman WD,
Lodato RF, Banks S, Koev CA, Solomon MA, Elin RJ, Hosseini JM, Danner RL (1992) NW-amino-L-arginine,
an inhibitor of nitric oxide synthase,
raises vascular resistance but increases
mortality rates in awake canines challenged with endotoxin. J Exp Med
176: 1175±1182
88. Grover R, Zaccardelli D, Colice G,
Guntupalli K, Watson D, Vincent JL
(1999) An open-label dose escalation
study of the nitric oxide synthase inhibitor, NG-methyl-L-arginine hydrochloride (546C88), in patients with
septic shock. Crit Care Med 27:
913±922
89. Nasraway SA (1999) Sepsis research:
we must change course. Crit Care
Med 27: 427±430
90. Camerota A, Creasey A, Patla V, Larkin V, Fink M (1998) Delayed treatment with recombinant human tissue
factor pathway inhibitor improves survival in rabbits with gram-negative
peritonitis. J Infect Dis 177: 668±676
91. Park T, Creasey A, Wright S (1997)
Tissue factor pathway inhibitor blocks
cellular effects of endotoxin by binding to endotoxin and interfering with
transfer to CD14. Blood 89: 4268±4274
92. Creasey AA, Change AK, Feigen L,
Wun TC, Taylor FB, Hinshaw LB
(1993) Tissue factor pathway inhibitor
reduces mortality from Escherichia
coli septic shock. J Clin Invest 91:
2850±2860
93. Abraham E (2000) New therapies in
sepsis. American Thoracic Society International Conference, ªNew insights
into Acute Lung Injuryº, May 2000
94. Taylor FB, Change AC, Peer GT,
Mather T, Blick K, Catlett R, Lockhart
MS, Esmon CT (1991) DEGR-factor
Xa blocks disseminated intravascular
coagulation initiated by Escherichia
coli without preventing shock or organ
damage. Blood 78: 364±368
95. Abraham E (2000) Coagulation abnormalities in acute lung injury and sepsis.
Am J Respir Cell Mol Biol 22:
401±404
96. Smith OP, White B, Vaughan D, Rafferty M, Claffey L, Lyons B, Casey W
(1997) Use of protein-C concentrate,
heparin, and haemodiafiltration in
meningococcus-induced purpura fulminans. Lancet 350: 1590±1593
97. Fourrier F, Lestavel P, Chopin C, Marey A, Goudemand J, Rime A, Mangalaboy J (1990) Meningococcemia
and purpura fulminans in adults: acute
deficiencies of proteins C and S and
early treatment with antithrombin III
concentrates. Intensive Care Med 16:
121±124
98. Leclerc F, Hazelzet J, Jude B, Hofhuis
W, Hue V, Martinot A, Van der Voort
E (1992) Protein C and S deficiency in
severe infectious purpura of children:
a collaborative study of 40 cases. Intensive Care Med 18: 202±205
99. Rivard G, David M, Farrell C,
Schwarz H (1995) Treatment of purpura fulminans in meningococcemia with
protein C concentrate. J Pediatr 126:
646±652
100. Kreuz W, Veldman A, Escuriola-Ettingshausen C, Schneider W, Beeg T
(1998) Protein-C concentrate for meningococcal purpura fulminans. Lancet 351: 986±987
101. Rintala E, Seppala OP, Kotilainen P,
Pettila V, Rasi V (1998) Protein C in
the treatment of coagulopathy in meningococcal disease. Crit Care Med
26: 965±968
101 a. Bernard GR, la Rosa SP et al (2001)
The efficacy and safety of recombinant
human activated protein C for the
treatment of patients with severe sepsis. Crit Care med (in press)
102. Bone RC, Balk RA, Cerra FB, et al
(1992) ACCP/SCCM consensus conference: definitions for sepsis and organ failure and guidelines for the use
of innovative therapies in sepsis. Chest
101: 1644±1655
103. Bone RC, Fisher Jr CJ, Clemmer TP,
et al (1989) Sepsis syndrome: a valid
clinical entity. Crit Care Med 17:
389±393
104. Zanetti G, Heumann D, Gerain J, et al
(1992) Cytokine production after intravenous or peritoneal gram-negative
bacterial challenge in mice. J Immunol
148: 1890±1897
105. Sprung CL, Finch RG, Thijs LG,
Glauser MP (1996) International sepsis trial (INTERSEPT): role and impact of a clinical evaluation committee. Crit Care Med 24: 1441±1447
106. Targan SR, Hanauer SB, van Deventer
SJ, Mayer L, Present DH, Braakman
T, DeWoody KL, Schaible TF, Rutgeerts PJ (1997) A short-term study of
chimeric monoclonal antibody cA2 to
tumor necrosis factor alpha for
Crohn's disease. N Engl J Med 337:
1029±1035
107. Lorenz HM, Antoni C, Valerius T,
Repp R, Grunke M, Schwedtner N,
Nusslein H, Woody J, Kalden JR,
Manger B (1996) In vivo blockade of
TNF-alpha by intravenous infusion of
a chimeric monoclonal TNF-alpha antibody in patients with rheumatoid arthritis. Short term cellular and molecular effects. J Immunol 156: 1646±1653
108. Elliott MJ, Maini RN, Feldmann M,
Long-Fox A, Charles P, Bijl H, Woody
JN (1994) Repeated therapy with
monoclonal antibody to tumor necrosis factor alpha (cA2) in patients with
rheumatoid arthritis. Lancet 344:
1125±1127
109. Fekade D, Knox K, Hussein K, Melka
A, Lalloo DG, Coxon RE, Warrell
DA (1996) Prevention of JarischHerxheimer reactions by treatment
with antibodies against tumor necrosis
factor alpha. N Engl J Med 335:
311±315
110. Bone RC (1994) Gram-positive organisms and sepsis. Arch Intern Med 154:
26±34
111. Kopf M, Baumann H, Freer G, et al
(1994) Impaired immune and acute
phase responses in interleukin-6-deficient mice. Nature 368: 339±342
112. Fein AM, Bernard GR, Criner GJ, et
al (1997) Treatment of severe systemic
inflammatory response syndrome and
sepsis with a novel bradykinin antagonist, Deltibant (CP-0127) JAMA 277:
482±487
113. Natanson C, Esposito CJ, Banks SM
(1998) The sirens' songs of confirmatory sepsis trials: selection bias and
sampling error. Crit Care Med 26:
1927±1931
S 115
114. Dhainaut JF, Tenaillon A, Le Tulzo Y,
et al (1994) Platelet-activating factor
receptor antagonist BN 52021 in the
treatment of severe sepsis: a randomized double-blind, placebo-controlled,
multicenter clinical trial. Crit Care
Med 22: 1720±1728
115. Dhainaut JF, Tenaillon A, Hemmer M,
et al (1998) Confirmatory platelet-activating factor receptor antagonist trial
in patients with severe Gram-negative
bacterial sepsis: a phase III, randomized double-blind, placebo-controlled,
multicenter clinical trial. Crit Care
Med 26: 1963±1971
116. Bernard GR, Reines HD, Halushka
PV (1991) Prostacyclin and thromboxane A2 formation is increased in human sepsis syndrome: effects of cyclooxygenase inhibition. Am Rev Respir Dis 144: 1095±1101
117. Bernard GR, Wheeler AP, Russell JA,
et al (1997) The effects of ibuprofen on
the physiology and survival of patients
with sepsis. N Engl J Med 336: 912±918
118. Angus DC, Birmingham MC, Balk
RA, et al (2000) E5 murine monoclonal antiendotoxin antibody in
gram-negative sepsis: a randomized
controlled
trial.
JAMA
283:
1723±1730
Intensive Care Med (2001) 27: S 116±S 127
Other supportive therapies in sepsis
Javier PØrez
R. Phillip Dellinger
)
J. PØrez ´ R. P. Dellinger ( )
Division of Cardiovascular and Critical Care Medicine,
Rush-Presbyterian-St. Luke's Medical Center, Chicago, Ill., USA
E-mail: [email protected]
Phone: +1-3 12-9 42 33 30
Fax: +1-3 12-9 42 63 59
Introduction
Because sepsis is associated with multisystem organ failure, there are many other supportive therapies used to
treat these patients that do not directly relate to the sepsis process. Although most of these have been studied in
randomized controlled clinical trials in hospitalized patients, few have been tested specifically in sepsis. Thus
research and the literature support some, while others
are extrapolations from other ill populations.
Deep vein thrombosis (DVT) prophylaxis, nutritional support, and stress ulcer prophylaxis (SUP) are important adjunctive considerations in the management
of sepsis. Consumption coagulopathy makes the septic
patient at risk for development of venous clots. Nutritional support, especially enteral, is recognized as important in supporting the critically ill septic patient
who is unable to eat. In addition, sepsis and the associated organ dysfunction put the patient at increased risk
for development of stress ulcers.
Methods
We performed a comprehensive Medline literature search from
January 1966 to February 2000. The following terms were independently searched: DVT, deep vein thrombosis, thrombophlebitis,
venous thrombosis, thromboembolic disease, pulmonary embolism, anticoagulation, warfarin, heparin, low-molecular weigh heparin, DVT prophylaxis, mechanical compression devices, external
pneumatic compression, nutritional support, parenteral nutrition,
total parenteral nutrition, enteral nutrition, immunoenhancing
diets, immunomodulating diets, stress ulcers, gastrointestinal
bleeding, SUP, gastrointestinal bleeding prophylaxis, gastrointestinal bleeding prevention, antacids, histamine antagonists, sucralfate, antiulcer agents, omeprazole, and proton-pump inhibitors.
Each one of those terms was searched and crossed with the following: critical care, intensive care, infection, systemic inflammatory response syndrome, sepsis, severe sepsis, sepsis syndrome,
septic shock, and multiple organ dysfunction syndrome.
Deep vein thrombosis prophylaxis in sepsis
The use of DVT prophylaxis in higher risk postoperative patients has been universally accepted since the early 1970s when it was found to reduce the risk of thromboembolic phenomena in this group [1]. Many subsequent trials have continued to emphasize the value of
DVT prophylaxis in most postoperative patients [2, 3,
4, 5]. In addition to postoperative patients, subcutaneous heparin has also proven efficacious in reducing the
risk of thromboembolism among myocardial infarction
[6, 7, 8, 9] and ischemic stroke patients [10, 11, 12].
Only a few studies of venous thromboembolism prophylaxis have been carried out on general medical wards,
and medical intensive care units. In those studies, patients treated with subcutaneous heparin [13, 14, 15, 16,
17] or with low molecular weight heparin (LMWH)
[18, 19] reduced the risk of thromboembolic events.
Does DVT prophylaxis improve clinical outcome in
patients with sepsis?
Answer: yes, grade A.
Recommendations
Considering the frequent occurrence of independent
risk factors for DVT in septic patients and the high per-
S 117
Table 1 DVT prophylaxis studies performed in general populations of the acutely ill: percentage of sepsis/infected patients
Reference
Design, methods
Setting
n
Infection,
sepsis
n
%
Pingleton et al. [13]
Prospective/historical
controls, V/Q angio, autopsy
Respiratory
care unit
188
53
28
Reduction in incidence of
pulmonary embolism
Cade [14]
Prospective, double-blind
placebo-control, I-fibrinogen
scan
ICU/medical
ward
119/131
ND
±
Reduction in incidence of
DVT (29 % vs. 13 %)
Halkin et al. [15]
Randomized prospective
control, no data
Medical ward
1358
138
10
Reduction in mortality
(10.9 % vs. 7.8 %)
Belch et al. [16]
Prospective randomized,
control, I-fibrinogen scan
ICU
100
52
52
Reduction in incidence of
DVT (26 % vs. 4 %)
Gardlund et al. [17]
Prospective randomized,
no data
Medical ward
11693
1610
14
Minor thromboembolic
events reduced
Samana et al. [18]
Placebo-control, doubleblind, randomized, venography, ultrasound,
V/Q angio, CAT, autopsy
Medical ward/
ICU
1102
584
53
Reduction in incidence of
DVT (14.9 % vs. 5.5 %)
Dahan et al. [19]
Placebo-control randomized,
I-fibrinogen
Medical ward
270
11
4
Reduction in incidence of
DVT (9 % to 3 %)
centage of sepsis/infected patients included in studies
that have demonstrated efficacy of DVT prophylaxis in
general, septic patients should be treated with DVT
prophylaxis. Even though there is not a randomized
study that establishes the impact of DVT prophylaxis
on morbidity and mortality specifically in septic patients, the significant number of septic patients included
in the populations of patients enrolled in other prospective randomized trials supports that the use of DVT prophylaxis reduces morbidity and mortality in septic patients. Moreover, septic patients, especially those with
severe sepsis and multiple organ failure, have less cardiopulmonary reserve, and the impact of a minor thromboembolic event in this group of patients could be very
compromising.
Rationale
Patients in the intensive care unit are at high risk of development of thromboembolic phenomena [14, 20, 21].
Septic patients as described above are expected to be
in the intensive care unit (ICU) and to be part of the
population at risk. No definitive study restricted to the
incidence of DVT in septic patients has been carried
out. The significance of DVT prophylaxis on morbidity
and mortality in septic patients needs to be implied
based on the analysis of proportion of the septic patients
included in the studies of the acutely ill patient in general (Table 1). Pingleton et al. [13], observed a reduction
in the incidence of pulmonary embolism in patients admitted to the respiratory intensive care unit. Cade [14]
Results
found a reduction in the risk of thromboembolic events
from 29 % to 13 % among patients admitted to ICU
and treated with subcutaneous heparin. In the latter
study, using a control group consisting of patients admitted to the medical ward and coronary care unit, a significantly higher incidence of thromboembolic events was
found among patients admitted to the ICU. Halkin and
coworkers [15], in a randomized prospective study of
patients admitted to medical wards, compared treatment with low-dose unfractionated heparin to patients
who did not receive any treatment. They found a significant reduction in mortality in heparin-treated patients
(7.8 % vs. 10.9 %). Belch et al. [16], in a study carried
out in medical patients admitted to the intensive care
unit, found a significantly reduced incidence of thromboembolic events (4 % vs. 26 %) in the group treated
with unfractionated heparin.
Mortality was not addressed in the study. Gardlund
et al. [17], found a significant reduction in minor embolic events in patients admitted to the hospital with infectious disease diagnoses who were treated with subcutaneous heparin versus those not treated, although there
was no difference in mortality or major thromboembolic
events. In a recently published trial [18] 1102 patients
received either LMWH (in two different doses) or placebo. Although the patients included in this trial were
not admitted to the ICU, many of them suffered from
complicated conditions. Patients receiving 40 mg enoxaparin had a significant reduction in the incidence of
thromboembolic phenomena (5.5 % vs. 14.9 %). No significant difference was found in mortality among any of
the groups, but a trend toward decreased mortality in
S 118
patients receiving 40 mg enoxaparin was reported. Dahan et al. [19] compared medical patients using treatment with LMWH versus placebo in a double-blind,
placebo-controlled randomized trial. LMWH reduced
the incidence of thromboembolic phenomena (9.1 %
vs. 3 %). Hirsch and colleagues [20] studied 100 patients
admitted to ICU and found the incidence of DVT to be
33 %. There was an association with increased mortality
in patients suffering DVT (although it is not possible to
determine whether death was caused by DVT or was a
consequence of the deteriorated state of those patients).
Although it is difficult to demonstrate mortality benefit
from DVT prophylaxis unless either a very large study
or patients at very high risk are studied, many argue
that demonstrating a decrease in DVT without increase
in bleeding complications implies that mortality benefit
could be demonstrated if higher powered studies were
performed.
Septic patients, especially those admitted to the ICU,
frequently have one or more risk factors for thromboembolic phenomena. These have been widely described
in postoperative, medical and critically ill patients [14,
22, 23, 24.]. These factors are: age (> 40 years), history
of venous thromboembolism, malignancy, bed rest
(> 5 days), major surgery, congestive heart failure, fracture (pelvic, hip or leg), estrogen replacement, stroke,
myocardial infarction, multiple trauma, and hypercoagulable states. The concurrence of two or more factors increases the risk of thromboembolic events [23, 24]. Other risk factors frequently present in septic patients include use of central venous catheters [20, 25, 26, 27],
use of neuromuscular blockade, use of deep sedation
[28], and presence of coagulopathy [29].
Is there any pharmacological method for DVT
prophylaxis preferred in septic patients?
Answer: no, grade A.
Recommendations
Septic patients who do not have a contraindication to
heparin use should receive prophylaxis with either lowdose unfractionated heparin (5,000 U either two or
three times daily) or LMWH (at recommended doses;
grade A). For those septic patients who have an absolute contraindication for heparin use (i.e., thrombocytopenia, severe coagulopathy, active bleeding, recent intracerebral hemorrhage), the use of a mechanical prophylactic device is advised since this method has proven
to be effective in postsurgical patients and therefore
would likely work in septic patients (grade E).
Rationale
Unfractionated subcutaneous heparin (UH) is widely
used for the prevention of DVT among postoperative
patients and in medical high-risk patients. UH is inexpensive and has been demonstrated in critically ill medical and surgical patients to be safe and to be associated
with minor bleeding complications (bruising, hematoma
at the site of injections) and rarely with heparin-induced
thrombocytopenia [13, 14, 24, 30, 31].
Although smaller studies have suggested that
LMWH may be either as effective as UH with less bleeding complications or more effective with the same bleeding complications [31, 32, 33] in the treatment of thromboembolic disease, larger studies have not demonstrated
statistically significant differences between the drugs.
However, LMWH has been demonstrated to be more effective than UH in several high-risk populations for prophylaxis of DVT [34]. Enoxaparin has been demonstrated to be safe and efficacious in treatment of thromboembolic disease in medical patients with minimal adverse
events [18]. Two other randomized studies [35, 36, 37]
in acutely ill medical patients compared LMWH and
UH and showed equal effectiveness in the prevention
of DVT. Each hospital should assess which form of heparin is most cost effective at that institution. Both are effective in presenting DVT in at-risk patients.
Special considerations: patients with sepsis-induced
coagulopathy
Sepsis is frequently associated with hemostatic defects
[29, 40]. The sepsis milieu may include consumptive coagulopathy and liver dysfunction leading to predisposition for both clotting and bleeding. Thrombocytopenia
is frequently present in septic patients. In the setting of
active hemorrhage or in septic patients with significant
abnormality in clotting function we recommend the use
of mechanical leg compression devices as a preferred alternative to heparin. Intermittent pneumatic compression devices applied to legs have been demonstrated to
be efficacious in postoperative patients [38, 39], and the
use of these devices is recommended in septic patients
with contraindication to the use of heparin (grade E).
Nutrition in sepsis
Septic patients are characterized by having increased
energy expenditure and enhanced catabolism [41, 42].
The need to provide adequate nutritional support to
septic patients is thus generally accepted as part of standard care in the ICU. However, many issues regarding
nutrition to septic patients remain controversial. This
controversy is enhanced by the fact that most nutritional
S 119
data available come from studies performed in trauma
or postsurgical patients, as opposed to a population of
septic patients alone.
Does institution of nutritional support improve clinical
outcome of patients with sepsis?
body mass index and mortality in critically ill patients
[55] whereas another placebo-controlled study demonstrated no difference in clinical outcome between patients receiving enteral nutrition and those receiving intravenous crystalloid [56].
Answer: yes, grade E
Are there any nutritional routes or formulations
preferred for patients with sepsis?
Recommendations
Answer: yes, grades C, E, B (based on different populations)
Based on the assumption that sepsis produces a hypercatabolic state and leads to protein-energy malnutrition,
and given that protein loss is associated with poor outcome, nutritional support in septic patients is recommended. The correlation of nutritional support with
outcome in septic patients comes from data extrapolated from studies performed in perioperative patients
and from expert opinion that allow us to establish this
recommendation. Many important questions remain regarding what kind of nutrition and when in the course of
sepsis should nutrition begin.
Rationale
Nutritional status has been closely related with outcome
of critically ill patients. Malnutrition has been associated with increased morbidity and longer hospital stays
[43]. Mullen et al. [44] in 1980 demonstrated a reduction
in perioperative complications in surgical patients with
the use of adequate nutritional support. Scientific evidence supports the important role of nutritional status
in the outcome of septic and other critically ill patients
[45, 46, 47]. Decreased gastrointestinal mucosal permeability [48], improved healing function [49], and lower
infection rates [47] have been attributed to the use of
enteral feeding in critically ill patients.
The activation of the inflammatory cascade in sepsis
alters the body's metabolism. Patients with sepsis have
elevated energy requirements, net catabolism, and rapid
loss of lean mass [50, 51]. For this reason the use of nutritional support has been axiomatically accepted. The
ability of nutrition to alter the clinical outcome of critically ill patients, however, is controversial [52, 53]. Studies have identified enteral nutrition as a major factor in
maintaining normal gut mucosal function [48, 54], both
in humans and in animals. Thus the use of enteral formulas would be expected to maintain mucosal integrity
in the critically ill septic patient. Most of the studies investigating metabolic changes and effects of nutrition
have been carried out in postoperative patients and
have provided conflicting conclusions. For example,
one study demonstrated a direct relationship between
Recommendations
Enteral nutrition is the preferred method of nutritional
support in the catabolic critically ill patient in general,
inclusive of the septic patient (grade C). For those patients who cannot tolerate enteral nutrition for a prolonged time or when contraindications do not allow its
use (mesenteric ischemia, mechanical bowel obstruction), parenteral nutritional support should be used
(grade E). Immune-enhancing formulas may be better
than other enteral formulations in critically ill patients,
but effects on ultimate outcome (i.e., survival) remain
to be demonstrated in large randomized trials (grade B).
Rationale
Although controversy exists, most authorities advocate
the use of enteral nutrition in critically ill patients [57,
58]. Several studies have compared enteral and parenteral nutrition in critically ill patients, most of them perioperative. Cerra and coworkers [53] compared standard
nutrition with total parenteral nutrition in septic patients. No difference was found in clinical outcome.
However, enteral nutrition has proven superior to
parenteral nutrition in reduction in stress ulcers [59],
gut protection [48], and costs [52, 60]. In addition, catheter placement and indwelling catheters have been associated with increased complications [61, 62]. A recent
meta-analysis [47] found an increased rate of complications and mortality in ICU medical patients receiving
parenteral nutrition when compared with those receiving enteral feeding. The advantage of enteral nutrition
versus total parenteral nutrition in some high-risk
groups has been demonstrated [63, 64].
Recent studies have also examined the potential advantage of enriched mixtures of enteral feeding formulas compared with standard formulas [50, 65, 66, 67,
68]. Bower et al. [50], published a prospective randomized clinical trial in septic patients comparing standard
enteral feeding versus an immunomodulatory formula
that contained arginine, nucleotide, and fish oil.
S 120
Although mortality was not modified, a significant
reduction in length of stay and infections was noted in
the immunomodulatory formula group. Galbµn and coworkers [67] concluded a benefit of immune-modulatory diets in septic patients from their study which revealed a decrease in mortality from 32 % to 19 %, and in infection from 20 % to 7 %. Atkinson and colleagues [68]
published a controlled double-blind clinical trial involving medical and surgical ICU patients. They compared
different formulations of enteral nutrition. The use of
immunomodulatory formula reduced mechanical ventilation time, ICU stay, hospital length of stay and duration of systemic inflammatory response syndrome. A recent study by Gadek et al. [69] in patients with acute respiratory distress syndrome, including a proportion of
septic patients, resulted in significant differences in outcome in those patients who received an immunomodulatory diet.
Are there any preferred range of calories and/or
proportion of elements in nutritional support in sepsis?
Answer: yes, grade E
Recommendations
The following are specific recommendations for septic
patients, according to the guidelines established by the
American College of Chest Physicians [58] and American Society of Parenteral and Enteral Nutrition [70]
consensus conferences:
· Daily caloric intake: 25±30 kcal/kg usual body weight
· Protein: 1.3±2.0 g/kg per day
· Glucose: 30±70 % of total nonprotein calories, to
maintain serum glucose level below 225 mg/dl
· Lipids: 15±30 % of total nonprotein calories. w6Polyunsaturated fatty acid should be reduced in septic patients, maintaining that level which avoids deficiency of essential fatty acids (7 % of total calories ±
generally 1 g/kg per day).
No specific recommendations are offered for use of medium-chain triglycerides, branched-chain amino acids,
or specific microelements added to the nutritional formulas. The use of any of these strategies, although supported in concept, does not have enough investigational
evidence to determine any clinical benefit in outcome of
septic patients.
Rationale
No randomized clinical trial has addressed optimal total
caloric requirements or the amount of fat and protein
needed in the diet of septic patients. Much of our knowledge regarding these issues derives from studies carried
out in patients with trauma, burns, and surgery, who, as
in the case of septic patients, are frequently hypercatabolic. Despite a lack of clinical outcome evidence from
randomized trials, expert panels have offered recommendations for general critically ill patients and for septic patients as well.
In 1993 the American Society of Parenteral and Enteral Nutrition used an evidence-based approach to
publish practice guidelines for nutritional support in
the ICU [70]. Although the guidelines do not address
specific recommendations for septic patients, they provide a grade B recommendation for total caloric requirements in critically ill patients. In presenting the results of a more recent conference the authors emphasize the use of branched-chain amino acids in the composition of enteral formulas although the existing data
did not allow establishing specific recommendations
[71]
The American College of Chest Physicians (ACCP)
in 1997 published a consensus statement of nutrition
guidelines in ICU patients [58]. Specific recommendations on caloric requirements in septic patients as well
as proportion of nutrients in formulations were offered.
Since then these recommendations have found agreement by most experts, but large gaps remain in our scientific basis for recommending enteral feeding in the
short-term critically ill patient [72, 73, 74].
Stress ulcer prophylaxis in sepsis
The use of SUP to prevent upper gastrointestinal bleeding in critically ill patients has become a routine in the
ICU. However, there are controversial points in this
practice: (a) SUP has not demonstrated a benefit in
mortality [23]; (b) there are many definitions of upper
gastrointestinal bleeding in critically ill patients that
could be responsible for the heterogeneity in results in
several controlled studies [75, 76]; (c) the use of SUP
has been implicated in the development of ventilator-associated pneumonia although the impact of this complication on mortality and morbidity has not been established [77, 78]; (d) only specific subgroups of patients in
the ICU are likely to benefit from SUP [79].
Comparing the various studies is made difficult by
the varied criteria used for diagnosing stress ulcer
bleeding. The use of microscopic bleeding (either orthotoluidine or guaiac in nasogastric aspirate or feces)
as a marker of stress ulcer bleeding entails several problems that have already been identified: (a) guaiac is
S 121
Table 2 Proportion of septic patients in different studies of SUP (R randomized trial, P placebo, C control, SU stress ulcer, MV mechanical ventilation)
Study
Trial
n
Septic
patients (%)
Sepsis definition
Summary of results
Cook et al. [79]
Cohort
2252
1.6
Fever-hypothermia, leukocytosis/leukopenia, + blood
culture
Risk factors for SU bleeding:
prolonged MV and coagulopathy
Schuster et al. [82]
Cohort
179
7.8
Not listed
Risk factors for SU bleeding:
coagulopathy, hypotension and
MV
Zandstra and
Stoutenbeek [83]
Cohort
167
40
Severe bacterial infection
Pinilla et al. [84]
R-C
259
3.8
2 criteria of: fever;
WBC > 15,000, shift to
the left, + culture
Minimal SU bleeding episodes;
prolonged MV identified as a
risk factor
No difference between patient
treated with antacids and control
Peura and Johnson [85]
R/P-C
39
15
Not listed
Cimetidine superior to placebo
in preventing SU; fewer transfusions required in treated group
Groll et al. [86]
R/P-C
221
30±15a
Not listed
No significant differences between placebo and cimetidine
Basso et al. [87]
R/C
168
22
Foci of infection or septicemia and fever, leukocytosis, elevated sed rate and
culture +
Cimetidine and antacid decreased the risk of SU bleeding
compared to placebo
Ben-Menachem et al. [88]
R/C
300
21
Not listed
No differences between cimetidine and sucralfate vs. control
Borrero et al. [89]
R
155
30
Not listed
No differences between sucralfate and antacids
Bressalier et al. [90]
R
74
23
Systemic infection with +
cultures or hypotension
Sucralfate advantages vs. antacids (both in safety and effectiveness)
Cook et al. [91]
R
1200
6.5
Not listed
Ranitidine offers better protection than sucralfate; no differences in ventilator-associated
pneumonia
Poleski and Spanier [92]
R
37
45
Blood culture with evidence
of infection (fever, leukocytosis)
Cimetidine and antacids equally
effective
Stothert et al. [93]
R
123
28
Culture and clinical evidence; sepsis confirmed at
autopsy or surgery
Antacids and cimetidine equally
effective
a
Referred to 30 % of septic patients in the placebo group and 5 % in the cimetidine group
nonspecific [80]; (b) cimetidine may produce false-positive results in gastric aspirates [81]; (c) the clinical relevance of microscopic bleeding is usually minimal, and
a minority of cases progress toward overt or clinically
significant bleeding. The use of overt bleeding (hematemesis, gross blood, or coffee ground material in nasogastric aspirates, hematochezia, or melena) or clinically
important bleeding (associated with a decrease in systolic blood pressure > 20 mmHg, orthostatic changes,
decrease in hemoglobin > 2 g/dl, transfusion of at least
2 U blood in 24 h caused by the bleeding episode, or
the need of surgical intervention) seems more reason-
able when evaluating the impact of stress ulcers in morbidity and mortality and the efficacy of the prophylactic
measures.
Although there are no specific studies of SUP in septic patients, many randomized trials have been carried
out in critically ill patients that include some number of
septic patients. Unfortunately, only few studies do allow
identification of the precise number of septic patients
enrolled (Table 2 lists the proportion of septic patients
in prospective studies). Furthermore, it is possible to
compare the frequency of occurrence of stress ulcer
bleeding in septic patients with that of patients at higher
S 122
risks, because many of the risk factors for development
of stress ulcer bleeding are common in septic patients.
Does SUP improve clinical outcome in patients with
sepsis?
Answer: yes, grade C
Recommendations
No randomized trial has evaluated the effect of SUP on
clinical outcome in septic patients. Examination of successful clinical trials of SUP does not allow precise identification of patients with diagnosis of sepsis. Therefore
no definitive data exist in septic patients on the effectiveness of SUP in diminishing episodes of overt or clinically significant bleeding. The clinical utility of SUP as
it affects clinical outcome in septic patients is therefore
not clear. Septic patients have been assumed to have an
increased risk for SUP since they have multiple risk factors known to increase the risk of stress ulcer bleeding.
Since data do support SUP as being efficacious in preventing upper gastrointestinal bleeding in populations
of critically ill patients, which would be expected to contain large proportions of septic patients, the use of SUP
is recommended in this group (see below).
Rationale
The use of SUP has become accepted practice in the
great majority of ICUs. Early studies associated sepsis
with stress ulcer bleeding and with an increased risk of
mortality in critically ill patients [94]. The initial study
by Skillman et al. [94] retrospectively reported a mortality of 87 % in patients admitted to the ICU (medical and
surgical) who developed stress ulcer related gastrointestinal bleeding. The use of SUP has become accepted
practice in the great majority of ICUs. However, recent
studies report significantly less mortality related to
stress ulcer bleeding [79, 83, 91]. Schuster et al. [82] reported a 14 % incidence of bleeding in patients admitted
to a respiratory intensive care unit. Although the mortality was significantly higher among patients who bled
(64 % vs. 9 %), death was related to bleeding only in 3
of the 25 patients who bled. Other studies [85, 95] comparing histamine receptor antagonists or antacids versus
placebo report similar results. Moreover, several authors believe that the modernization of anesthesia and
ventilation techniques and, in general, the improvement
in the management of critically ill patients have decreased the incidence of stress ulcers and therefore prophylaxis is not warranted [79, 80, 81, 82, 83]. Lacroix and
colleagues [96] in a meta-analysis observed a range of
overt bleeding from 1.6 % to 52.8 % of in control groups
and from 0 to 23.1 % in antacids groups. The conclusion
of the study was that cimetidine and antacids are effective in preventing stress ulcer bleeding (33 % and 43 %
better than control, respectively).
Collectively these studies support the assertion that
patients who develop bleeding from stress ulcers require
more transfusions. However, no difference in clinical
outcome has been noted. Patients with stress ulcer
bleeding who do not receive SUP often show two factors: coagulopathy and liver failure. A recent meta-analysis [97] reporting risk reduction for bleeding in critically ill patients with antacids, sucralfate, or histamine-2
receptor antagonists could not establish any impact on
clinical outcome compared with control groups. BenMenachem et al. [88], in a randomized single-blind, control trial, reported that the incidence of bleeding did not
differ among three groups of 100 patients (control, sucralfate, and cimetidine). The mortality and hospital
length of stay did not vary with prophylaxis.
Is there any specific subgroup of septic patients who
should receive SUP?
Answer: yes, grades A, C
Recommendations
Although no large randomized trial has addressed septic
patients alone, abundant data exist regarding subgroups
of septic patients with prolonged mechanical ventilation, hypotension, and coagulopathy. For these patients
the use of SUP is recommended (grade A). For other
septic patients in whom these factors are not present
SUP is recommended based on several small randomized trials in which SUP has proven efficacious in preventing bleeding and therefore reducing morbidity in
critically ill patients (grade C).
Rationale
Cook et al. [79] in a prospective study found an increased risk of stress ulcer bleeding in patients with prolonged mechanical ventilation (> 48 h) and those with
coagulopathy. The low number of septic patients in this
study does not allow the determination of the true impact of sepsis as an independent risk factor for the development of stress ulcer bleeding. Schuster et al. [82]
found increased risk of bleeding associated with coagulopathy, prolonged mechanical ventilation, and sepsis.
The authors of this study did not perform a multivariate
analysis that would help to determine the true impact of
sepsis as a single variable risk factor for stress ulcer
S 123
bleeding. Coagulopathy, frequently found in severe sepsis, has been classically associated with increased incidence of bleeding [98]. Risk factors for stress ulcer
bleeding have been demonstrated to be additive [95,
96, 97, 98, 99]. A score has been offered to predict the
risk of SU bleeding [100].
Are some methods to be preferred over others in the
prevention of stress ulcers in patients with sepsis?
Answer: uncertain, grade B.
Recommendations
Several trials have confirmed the efficacy of antacids,
sucralfate, or histamine-2 receptor antagonists in preventing stress ulcer bleeding. Since the data are conflicting, no single one can be determined as preferable. General recommendations should be based on the individual experience in the use of one or another, the availability, or cost-analysis in individual centers. In septic patients with risk factors the use of enteral nutrition following the preventive strategies currently available
may be beneficial for preventing stress ulcer bleeding.
less in patients treated with ranitidine, without increased associated pneumonia. These findings have
been corroborated in other studies [105, 106]. A metaanalysis published in 1996 found sucralfate to be associated with a trend toward a lower incidence of pneumonia compared with both antacid and histamine-2 receptor antagonists. In this meta-analysis sucralfate was associated with less mortality. A recently published cost
effectiveness analysis [107] pointed out the high costs involved in SUP. In 1999 a national survey in the United
States [108] found a wide variation in the forms of SUP
among intensivists. The costs of prophylaxis in low-risk
patients were considered by these authors as prohibitive. In this survey the authors called for the creation of
hospital-based algorithms, based on individualization
of cost and care issues at the institution as it applies to
patients with higher risks of bleeding.
There are also data supporting the use of enteral nutrition as SUP [59]. The beneficial effect of enteral feeding has been demonstrated with distal enteral nutrition
rather than gastric. Patient's position, type of tube
(orogastric versus nasogastric, small-bore versus largebore), and continuous versus intermittent delivery, are
factors implicated by the findings of various studies
that could modify stress ulcer bleeding in patients receiving enteral nutrition [109].
Rationale
Summary
There are many studies comparing the efficacy of histamine-2 receptor antagonists, antacids, and sucralfate in
the prevention of stress ulcer bleeding [85, 86, 87, 88,
90, 100, 101, 102]. Cook et al. [103] in a meta-analysis
of SUP studies found histamine-2 receptor antagonists
more effective than antacids in controlling overt bleeding. No data about nosocomial pneumonia were presented. There was no difference in mortality between
the three methods, and no difference was found when
compared with no prophylaxis. Similar results have
been reported by two other meta-analysis [96, 97]. A
controversy related to SUP stems from the ability of
both antacids and histamine-2 receptor antagonists to
raise the gastric pH, which may be associated with increase in gastric bacterial colonization. Increased bacterial presence in the gastrointestinal tract can lead to an
increase in pneumonia if it is a route that leads to pharyngeal colonization. This area is controversial, and although some studies have demonstrated an increase in
ventilator-associated pneumonia with the use of histamine-2 blockers and antacids, these data have not been
validated in all clinical trials [77, 91, 103]. Furthermore,
prospective studies suggest that gastric colonization is
not a frequent route to pharyngeal colonization [104].
In a recently published Canadian trial [91] the risk of
bleeding (in 1200 patients studied) was significantly
Patients who survive the circulatory and organ deficits
in sepsis may still fall victim to complications such as
pulmonary embolism and stress ulcer bleeding. Although there is no clearcut evidence to quantitate the
impact of such complications on mortality, the anticipated impact is grave when considering the compromised
physiological reserve of these patients. For this reason
it is important to institute effective prophylaxis to minimize the impact. In addition, catabolism associated
with sepsis likely influences the recovery of patients
with sepsis and moreover can compromise the response
of the immune system against an infectious insult. Early
and adequate nutritional support therefore appears important. There is much controversy and lack of prospective research regarding effect of supportive therapies on
outcome in patients with severe sepsis. This research is
needed.
S 124
References
1. Kakkar VV, Corrigan TP, Fossard DP
and the International Multicentre Trial (1975) Prevention of fatal postoperative pulmonary embolism by low doses of heparin: an international multicentre trial. Lancet II:45±51
2. Upchurch GR, Demling RH, Davies J,
Gates JD, Knox JB (1995) Efficacy of
subcutaneous heparin in prevention of
venous thromboembolic events in
trauma patients. Am Surg 61: 749±755
3. Fauno P, Suomalainen O, Rehnberg V,
Hansen TB, Kroner K, Soimakallio S,
Nielsen E (1994) Prophylaxis for the
prevention of venous thromboembolism after total knee arthroplasthy.
J Bone Joint Surg Am 76: 1814±1818
4. Amstutz HC, Friscia DA, Dorey F,
Carney BT (1989) Warfarin prophylaxis to prevent mortality from pulmonary embolism after total hip replacement. J Bone Surg Am 71: 321±326
5. Fisher CG, Blachut PA, Salvian AJ,
O'Brien PJ (1995) Effectiveness of
pneumatic leg compression devices
for the prevention of thromboembolic
disease in orthopaedic trauma patients: a prospective, randomized
study of compression alone versus no
prophylaxis. J Orthop Trauma 9: 1±7
6. Medical Research Council (1969) Assessment of short-term anticoagulant
administration after cardiac infarction.
BMJ 1: 335±e42
7. Drapkin A, Merskey C (1972) Anticoagulant therapy after acute myocardial
infarction. JAMA 222: 541±548
8. Warlow C, Beattie AG, Terry G, Ogston D, Kenmure ACF, Douglas AS
(1973) A double-blind trial of low doses of subcutaneous heparin in the prevention of deep vein thrombosis after
myocardial infarction. Lancet II:
934±936
9. Kierkegaard A, Norgren L (1993)
Graduated compression stockings in
the prevention of deep vein thrombosis in patients with acute myocardial
infarction. Eur Heart J 14: 1365±1368
10. Prins MH, Gelsena R, Sing AK, van
Heerde LR, den Ottolander GJ
(1989) Deep vein thrombosis prophylaxis with a low molecular weight heparin in stroke patients. Haemostasis
19: 245±250
11. McCarthy ST, Turner JJ, Roberston D,
Hawkey CJ, Macey DJ (1977) Low
dose heparin as prophylaxis against
deep vein thrombosis after acute
stroke. Lancet II:800±801
12. Turpie AG, Levine MN, Hirsh J, Carter CJ, Jay RM, Powers PJ, Andrew
M, Magnani HN, Hull RD, Gent M
(1987) A double blind randomized trial of ORG 10172 low molecular weight
heparinoid in the prevention of deep
vein thrombosis in thrombotic stroke.
Lancet I:523±526
13. Pingleton SK, Bone RC, Pingleton
WW, Ruth WE (1981) Prevention of
pulmonary emboli in a respiratory intensive care unit. Chest 79: 647±650
14. Cade JF (1982) High risk of the critically ill for venous thromboembolism.
Crit Care Med 10: 448±450
15. Halkin H, Goldberg J, Modan M, Modan B (1982) Reduction in mortality
in general medical in-patients by lowdose heparin prophylaxis. Ann Intern
Med 96: 561±565
16. Belch JJ, Lowe DO, Ward AG, Forbes
CD, Prentice CRM (1981) Prevention
of deep vein thrombosis in medical patients by low-dose heparin. Scott Med
J 26: 115±117
17. Gardlund B and the Heparin Prophylaxis Study Group (1996) Randomized, controlled trial of low-dose heparin for prevention of fatal pulmonary
embolism in patients with infectious
diseases. Lancet 347: 1357±1361
18. Samama MM, Cohen AT, Darmon JY,
Desjardins L, Eldor A, Janbon C,
Leizorovicz A, Nguyen H, Olsson
CG, Turpie AG, Weisslinger N (1999)
A comparison of enoxaparin with placebo for the prevention of venous
thromboembolism in acutely ill medical patients. N Engl J Med 341:
793±800
19. Dahan R, Houlbert D, Caulin C, Cuzin C, Viltart C, Woler M, Segrestaa
JM (1986) Prevention of deep vein
thrombosis in elderly medical in-patients by a low molecular weight heparin: a randomized double-blind trial.
Haemostasis 16: 159±164
20. Hirsch DR, Ingenito EP, Goldhaber
SZ (1995) Prevalence of deep venous
thrombosis among patients in medical
intensive care. JAMA 274: 335±337
21. Gallus AS, Hirsh J, Tuttle RJ, Trebilcock R, O'Brien S, Carroll JJ, Minden
JH, Hudecki S (1973) Small subcutaneous doses of heparin in prevention
of venous thrombosis. N Engl J Med
288: 545±551
22. Thromboembolic Risk Factors Consensus Group (1992) Risk of and prophylaxis for venous thromboembolism
in hospital patients BMJ 305: 567±574
23. Saint S, Matthay M (1998) Risk reduction in the intensive care unit. Am J
Med 105: 515±523
24. Anderson FA, Wheeler HB (1995) Venous thromboembolism. Risk factors
and prophylaxis. Clin Chest Med 16:
235±251
25. Wheeler HB, Anderson FA Jr, Cardullo PA, Patwardhan NA, Jian-Ming
L, Cutler BS (1982) Suspected deep
vein thrombosis: management by impedance plethysmography. Arch Surg
117: 1206±1209
26. Randolph AG, Cook DJ, Gonzales
CA, Andrew M (1998) Benefit of heparin in central venous and pulmonary
artery catheters. A meta-analysis of
randomized controlled trials. Chest
113: 165±171
27. Borow M, Crowley JG (1985) Evaluation of central venous catheter
thrombogenicity. Acta Anaesthesiol
Scand 81 [Suppl]:59S±64S
28. Durbec O, Viviand X, Potie F (1997)
Lower extremity deep vein thrombosis: a prospective, randomized, controlled trial in comatose or sedated patients undergoing femoral vein catheterization. Crit Care Med 25:
1982±1985
29. Carvalho AC, Freeman NJ (1994)
How coagulation defects alter outcome in sepsis: survival may depend
on reversing procoagulant conditions.
J Crit Illness 9: 51±75
30. Clagett GP, Reisch JS (1988) Prevention of venous thromboembolism in
general surgical patients. Results of
meta-analysis. Ann Surg 208: 227±240
31. Levine MN, Raskob G, Landfield S,
Hirsh J (1995) Hemorrhagic complications of anticoagulant treatment.
Chest 108 [Suppl]:276S±290S
32. Warkentin TE, Levine MN, Hirsh J
(1995) Heparin induced thrombocytopenia in patients treated with low
molecular weight heparin or unfractionated heparin. N Engl J Med 332:
1330±1335
33. Hirsh J, Raschke R, Warkentin, Dalen
JE, Poller L (1995) Heparin: mechanism of action, pharmacokinetics, dosing considerations, monitoring, efficacy, and safety. Chest 108 [Suppl]:258S±275S
34. Green D, Hirsh J, Heit J, Prins M, Davidson B, Lensing AWA (1994) Low
molecular weight heparin: a critical
analysis of clinical trials. Pharmacol
Rev 46: 89±121
S 125
35. Bergman JF, Neuhart E (1988) A multicenter randomized, double-blind
study of enoxaparin compared with
unfractionated heparin in the prevention of venous thromboembolic disease in elderly in-patients bedridden
for an acute medical illness. Thromb
Haemost 60: 407±410
36. Haremberg J, Kallenbach B, Martin
CE, Dempfle E, Zimmermann R,
Kübler W, Heene DL (1990) Randomized controlled study of heparin and
low molecular weight heparin for prevention of deep-vein thrombosis in
medical patients. Thromb Res 59:
639±650
37. Simmonneau G, Sors H, Charbonnier
B, Page Y, Laaban JP, Azarian R, Laurent M, Hirsch JL, Ferrari E, Bosson
JL, Mottier D, Beau B (1997) A comparison of Low-molecular-weight heparin with unfractionated heparin for
acute pulmonary embolism. N Engl J
Med 337: 663±669
38. Anglen JO, Goss K, Edwards J (1998)
Foot pump prophylaxis for deep vein
thrombosis: the rate of effective usage
in trauma patients. Am J Orthop 27:
580±582
39. Black PM, Crowell RM, Abbott WM
(1986) External pneumatic calf compression reduces deep thrombosis in
patients with ruptured intracranial aneurysms. Neurosurgery 18: 25±28
40. Mammen EF (1994) Coagulation defects in liver disease. Med Clin North
Am 78: 545±553
41. Cerra FB, Siegel JH, Coleman B
(1980) Septic autocanibalism: a failure
of exogenous nutritional support. Ann
Surg 192: 570±580
42. Bartlett RH, Dechert RR, Mault JR,
Ferguson SK, Kaiser AM, Erlandson
EE (1982) Measurement of metabolism in multiple organ failure. Surgery
92: 771±779
43. Dempsey DT, Mullen JL, Buzby GP
(1988) The link between nutritional
status and clinical outcome: can nutritional intervention modify it? Am J
Clin Nutr 47: 351±356
44. Mullen JL, Busby GP, Matthews DC,
Smale BF, Rosato EF (1980) Reduction of operative morbidity and mortality by combined pre-operative and
post-operative nutritional support.
Ann Surg 192: 604±613
45. Young B, Ott L, Twyman D, Norton J,
Rapp R, Tibbs P, Haack D, Brivins B,
Dempsey R (1987) The effect of nutritional support on outcome from severe
head injury. J Neurosurg 67: 668±676
46. Border JR, Hasset J, LaDuca J, Seibel
R, Steinberg S, Mills B, Losi P, Border
D (1987) The gut origin septic states
in blunt multiple trauma in the ICU.
Ann Surg 206: 427±448
47. Moore FA, Feliciano DV, Andrassy RJ
(1992) Early enteral feeding compared
with parenteral, reduces septic complications: the results of a meta-analysis.
Ann Surg 216: 172±183
48. Hadfield RJ, Sinclair DG, Houldsworth PE, Evans TW (1995) Effects
of enteral and parenteral nutrition on
gut mucosal permeability in the critically ill. Am J Respir Crit Care Med
152: 1545±1548
49. Schroeder D, Gillanders L, Mahr K,
Hill GL (1991) Effects of immediate
postoperative enteral nutrition on
body composition, muscle function,
and wound healing. JPEN J Parenter
Enteral Nutr 15: 376±383
50. Bower RH, Cerra FB, Bershadsky B,
Licari JJ, Hoyt DB, Jensen GL, Van
Buren CT, Rothkopf MM, Daly JM,
Adelsberg BR (1995) Early enteral administration of a formula (impact registered trademark) supplemented with
arginine, nucleotides, and fish oil in intensive care unit patients: results of a
multicenter, prospective, randomized,
clinical trial. Crit Care Med 23:
436±449
51. Müller TF, Müller A, Bachem MG,
Lange H (1995) Immediate metabolic
effects of different nutritional regimens in critically ill medical patients.
Intensive Care Med 21: 561±566
52. Wheeler A, Bernard GR (1999) Current concepts: treating patients with
severe sepsis. N Engl J Med 340:
207±214
53. Cerra FB, McPherson JP, Konstantinides FN, Konstantinides NN, Teasley
KM (1988) Enteral nutrition does not
prevent multiple organ failure syndrome (MOFS) after sepsis. Surgery
104: 727±733
54. Steiner M, Burges HR, Freeman LS,
Gray SJ (1968) Effect of starvation on
the tissue composition of intestine in
the rat. Am J Physiol 215: 75±77
55. Galanos AN, Pieper CF, Kussin PS,
Winchell MT, Fulkerson WJ, Harrell
FE Jr, Teno JM, Layde P, Connors AF
Jr, Phillips RS, Wenger NS (1997) Relationship of body mass index to subsequent mortality among seriously ill
hospitalized patients. Crit Care Med
25: 1962±1968
56. Heslin M, Latkany L, Leung D,
Brooks AD, Hochwald SN, Pisters
PW, Shike M, Brennan MF (1997) A
prospective randomized trial of early
enteral feeding after resection of upper gastrointestinal malignancy. Ann
Surg 226: 567±580
57. Zaloga GP (1999) Early enteral nutritional support improves outcome: hypothesis or fact? Crit Care Med 27:
259±261
58. American College of Chest Physicians
Consensus Statement. Applied nutrition in ICU patients (1997) Chest 111:
769±778
59. Pingleton SK, Hadzima SK (1983) Enteral alimentation and gastrointestinal
bleeding in mechanically ventilated
patients. Crit Care Med 11: 13±16
60. Frost P, Bihari D (1997) The route of
nutritional support in the critically ill:
physiological and economical considerations. Nutrition 13 [Suppl]:58S±63S
61. Orr ME, Ryder MA (1993) Vascular
access devices: perspective on designs,
complications and management. Nutr
Clin Pract 8: 145±152
62. Richet H, Hubert B, Nitemberg G,
Andremont A, Buu-Hoi A, Ourbak P,
Galicier C, Veron M, Boisivon A, Bouvier AM (1990) Prospective multicenter study of vascular-catheter-related
complications and risk factors for positive central-catheter cultures in intensive care unit patients. J Clin Microbiol 28: 2520±2525
63. Heyland DK, Cook DJ, Guyatt GH
(1993) Enteral nutrition in the critically ill patient: a critical review of the
evidence. Intensive Care Med 19:
435±442
64. Braga M, Gianotti L, Vignalli A, Cestari A, Bisagni P, Di Carlo V (1998)
Artificial nutrition after major abdominal surgery: impact of route of administration and composition of the diet.
Crit Care Med 26: 24±30
65. Moore FA, Moore EE, Kudsk KA,
Brown RO, Bower RH, Koruda MJ,
Baker CC, Barbul A (1994) Clinical
benefit of an immune-enhancing diet
for early postinjury feeding. J Trauma
37: 607±615
66. Cerra FB, Lehmann S, Konstantinides
N, Dzik J, Fish J, Konstantinides F, LiCari JJ, Holman RT (1991) Improve in
immune function in ICU patients by
enteral nutrition supplemented with
arginine, RNA and menhaden oil is independent of nitrogen balance. Nutrition 7: 193±199
S 126
67. Galbµn C, Montejo JC, Mesejo A,
Marco P, Celaya S, Sanchez-Segura J,
Farre M, Bryg DJ (2000) An immuneenhancing enteral diet reduces mortality rate and episodes of bacteremia in
septic intensive care unit patients. Crit
Care Med 28: 643±648
68. Atkinson S, Sieffert E, Bihari D (1998)
A prospective, randomized, doubleblind, controlled clinical trial of enteral immunonutrition in the critically ill.
Crit Care Med 26: 1164±1172
69. Gadek JE, DeMichele SJ, Karlstad
MD, Pacht ER, Donahoe M, Albertson TE, Van Hoozen C, Wenneberg
AK, Nelson JL, Noursalehi M (1999)
Effect of enteral feeding with eicosapentaenoic acid, gamma-linolenic acid
and antioxidants in patients with acute
respiratory distress syndrome. Crit
Care Med 27: 1409±1420
70. American Society of Parenteral and
Enteral Nutrition Board of Directors
(1993) Guidelines for the use of parenteral and enteral nutrition in adult and
paediatric patients. JPEN J Parenter
Enteral Nutr 17: 1SA±26SA
71. Klein S, Kinney J, Jeejeebhoy K, Alpers D, Hellerstein M, Murray M,
Twomey P (1997) Nutrition support in
clinical practice: review of published
data and recommendations for future
research directions. JPEN J Parenter
Enteral Nutr 21: 133±156
72. Wojnar MM, Hawkins WG, Lang CH
(1995) Nutritional support of the septic patient. Crit Care Clin 11: 717±733
73. Preiser JC, BerrØ J, Carpentier Y, Jolliet P, Pichard C, Van Gossum A, Vincent JL (1999) Management of nutrition in European intensive care units:
results of a questionnaire. Intensive
Care Med 25: 95±101
74. Mizock BA, Troglia S (1997) Nutritional support of the hospitalized patient. Dis a month 43: 351±426
75. Fisher RL, Pipkin GA, Wood JR
(1995) Stress-related mucosal disease.
Crit Care Clin 11: 323±345
76. Tryba M (1999) Research on stress ulcer prophylaxis: wrong questions,
wrong answers? Crit Care Med 27:
16±17
77. Tryba M (1987) Risk of acute stress
bleeding and nosocomial pneumonia
in ventilated intensive care unit patients: sucralfate verus antacids. Am J
Med 83: 117±124
78. Garvey BM, McCambley JA, Tuxen
DV (1989) Effects of gastric alkalinization on bacterial colonization in
critically ill patients. Crit Care Med
17: 211±216
79. Cook DJ, Fuller HD, Guyatt GH,
Marshall JC, Leasa D, Hall R, Winton
TL, Rutledge F, Todd TJ, Roy P
(1994) Risk factors for gastrointestinal
bleeding in critically ill patients.
N Engl J Med 330 377±381
80. Layne EA, Mellow MH, Lipman TO
(1981) Insensitivity of guaiac slide test
for detection of blood in gastric juice.
Ann Intern Med 94: 774±776
81. Schentag JJ (1980) False positive
ªhemoccultº reaction with cimetidine.
N Engl J Med 303: 110
82. Schuster DP, Rowley H, Feinstein S,
McGue MK, Zuckerman GR (1984)
Prospective evaluation of the risk of
upper gastrointestinal bleeding after
admission to a medical intensive care
unit. Am J Med 76: 623±629
83. Zandstra DF, Stoutenbeek CP (1994)
The virtual absence of stress-ulceration related bleeding in ICU patients
receiving prolonged mechanical ventilation without any prophylaxis: a prospective cohort study. Intensive Care
Med 20: 335±340
84. Pinilla JC, Oleniuk FH, Reed D, Malik B, Laverty WH (1985) Does antacid prophylaxis prevent upper gastrointestinal bleeding in critically ill patients? Crit Care Med 13: 646±650
85. Peura DA, Johnson LF (1985) Cimetidine for prevention and treatment of
gastroduodenal mucosal lesions in patients in an Intensive care unit. Ann
Intern Med 103: 173±177
86. Groll A, Simon JB, Wigle RD, Taguchi K, Todd RJ, Depew WT (1988) Cimetidine prophylaxis for gastrointestinal bleeding in an intensive care unit.
Gut 27: 135±140
87. Basso N, Bagarani M, Materia A, Fiorani S, Lunardi P, Speranza V (1981)
Cimetidine and antacid prophylaxis of
acute upper gastrointestinal bleeding
in high risk patients. Am J Surg 141:
339±342
88. Ben-Menachem T, Fogel R, Patel RV,
Touchette M, Zarowitz BJ, Hadzijahic
N, Divine G, Verter J, Bresalier RS
(1994) Prophylaxis for stress-related
gastric hemorrhage in the medical intensive care unit: a randomized, controlled, single-blind study. Ann Intern
Med 121: 568±575
89. Borrero E, Bank S, Margolis I, Schulman ND, Chardavoyne R (1985) Comparison of antacid and sucralfate in the
prevention of gastrointestinal bleeding
in patients who are critically ill. Am J
Med 79: 62±64
90. Bresalier RS, Grendell JH, Cello JP,
Meyer AA (1987) Sucralfate versus titrated antacid for the prevention of
acute stress-related gastrointestinal
hemorrhage in critically ill patients.
Am J Med 83: 110±116
91. Cook D, Guyatt G, Marshall J, Leasa
D, Fuller H, Hall R, Peters S, Rutledge F, Griffith L, McLellan A,
Wood G, Kirby A (1998) A comparison of sucralfate and ranitidine for the
prevention of upper gastrointestinal
bleeding in patients requiring mechanical ventilation. N Engl J Med 338:
791±797
92. Poleski MH, Spanier AH (1986) Cimetidine versus antacids in the prevention of stress erosions in critically
ill patients. Am J Gastroenterol 81:
107±111
93. Stothert JC, Simonowitz DA, Dellinger EP, Farley M, Edwards WA, Blair
AD, Cutler R, Carrico CJ (1980) Randomized prospective evaluation of cimetidine and antacid control of gastric
pH in the critically ill. Ann Surg 192:
169±174
94. Skillman JJ, Bushnell LS, Goldman H,
Silen W (1969) Respiratory failure, hypotension, sepsis and jaundice: a clinical syndrome associated with lethal
hemorrhage from acute stress ulceration of the stomach. Am J Surg 117:
523±530
95. Hastings PR, Skillman JJ, Bushnell L,
Sillen W (1978) Antacid titration in
the prevention of acute gastrointestinal bleeding. N Engl J Med 298:
1041±1045
96. Lacroix J, Infante-Rivard C, Jenicek
M, Gauthier M (1989) Prophylaxis of
upper gastrointestinal bleeding in intensive care units: a meta-analysis.
Crit Care Med 17: 862±886
97. Cook DJ, Reeve BK, Guyatt GH,
Heyland DK, Griffith LE, Buckingham L, Tryba M (1996) Stress ulcer
prophylaxis in critically ill patients: resolving discordant meta-analysis.
JAMA 275: 308±314
98. MacDougall BRD, Bailey RJ, Williams R (1977) H-2 receptor antagonists and antacids in the prevention of
acute gastrointestinal hemorrhage in
fulminant hepatic failure. Lancet
I:617±619
99. Metz CA, Livingston DH, Smith JS,
Larson GM, Wilson TH (1993) Impact
of multiple risk factors and ranitidine
prophylaxis on the development of
stress-related upper gastrointestinal
bleeding: a prospective, multicenter,
double-blind randomized trial. Crit
Care Med 21: 1844±1849
S 127
100. Tryba M, Zevounou F, Torok M, Zenz
M (1985) Prevention of acute stress
bleeding with sucralfate, antacids or
cimetidine. Am J Med 79 [Suppl
2C]:55±60
101. Luk GD, Summer WR, Messersmith
JF (1982) Cimetidine and antacid in
prophylaxis of acute gastrointestinal
bleeding: a randomized, double blind,
controlled study (abstract). Gastroenterology 82: 1121
102. Priebe HJ, Skillman JJ, Bushnell LS,
Long PC, Silen W (1980) Antacid versus cimetidine in preventing acute gastrointestinal bleeding. N Engl J Med
302: 426±430
103. Cook DJ, Witt LG, Cook RJ, Guyatt
GH (1991) Stress ulcer prophylaxis in
the critically ill: a meta-analysis. Am J
Med 91: 519±527
104. Bonten MJM, Gaillard CA, Van der
Geest S, van Tiel FH, Beysens AJ,
Smeets HG, Stobberingh EE (1995)
The role of intragastric acidity and
stress ulcus prophylaxis on colonization and infection in mechanically ventilated ICU patients. Am J Respir Crit
Care Med 152: 1825±1834
105. Prod'hom G, Leuenberger P, Koerfer
J, Blum A, Chiolero R, Schaller MD,
Perret C, Spinnler O, Blondel J, Siegrist H (1994) Nosocomial pneumonia
in mechanically ventilated patients receiving antacid, ranitidine or sucralfate for stress ulcer: a randomized controlled trial. Ann Intern Med 120:
653±662
106. Hanisch EW, Encke A, Naujoks F, et
al (1998) A randomized, double-blind
trial for stress ulcer prophylaxis shows
no evidence of increased pneumonia.
Am J Surg 176: 453±457
107. Ben-Menachem T, McCarthy BD, Fogel R, Schiffman RM, Patel RV, Zarowitz BJ, Nerenz DR, Bresalier RS
(1996) Prophylaxis for stress-related
gastrointestinal hemorrhage: a cost effectiveness analysis. Crit Care Med
24: 338±345
108. Lam NP, Le PDT, Crawford SY (1999)
National survey of stress ulcer prophylaxis. Crit Care Med 27: 98±103
109. American Thoracic Society (1995)
Hospital-acquired
pneumonia
in
adults: diagnosis, assessment of severity, initial antimicrobial therapy and
preventive strategies. Am J Respir
Crit Care Med 153: 1711±1725
Intensive Care Med (2001) 27: S 128±S 134
Summary of recommendations
Definition of sepsis (I. Matot, C. L. Sprung)
· In the past many different definitions of sepsis were
used interchangeably, which led to confusion.
· Sepsis is the systemic inflammatory response to infection.
· No single physiological or laboratory parameter can
universally identify sepsis.
· Not all patients with sepsis are equally ill. Sepsis, severe sepsis, and septic shock constitute different gradations in the continuum of a disease process manifested by a combination of changes in vital signs, laboratory parameters, hypoperfusion, and organ dysfunction.
· The continuum of sepsis, severe sepsis, and septic
shock is correlated with increasing organ dysfunction
and mortality.
· The source of infection and diagnosis of sepsis must
be identified as early as possible to permit early intervention with antimicrobial therapy and surgical
drainage to prevent disease progression, organ dysfunction, and mortality.
Diagnosis of infection in sepsis (M. Llewelyn, J. Cohen)
Bacteremia
· Fever, chills, hypothermia, leukocytosis, left-shift of
neutrophils, neutropenia, and when infection is suspected, hypoalbuminemia, development of renal failure or signs of hemodynamic compromise are specific indications for obtaining blood for culture.
· Blood cultures should be taken as soon as possible after onset of fever or chills.
· Blood should be obtained by fresh venipuncture.
Sites associated with skin contamination (e.g., femoral site) or loss of skin integrity (e.g., burns or dermatological disease) should be avoided.
· Skin should be swabbed twice with either 70 % isopropyl alcohol or with an iodine containing solution
prior to venipuncture. The blood culture stopper
should also be sterilized prior to inoculation.
· An adequate volume (20±60 ml) of blood should be
obtained per culture (10±30 ml per bottle)
· If insufficient blood is available, only the aerobic bottle should be inoculated.
· The needle used for venipuncture should be changed
prior to inoculation of blood into culture bottles.
· A minimum of two and a maximum of three sets of
blood cultures should be obtained for each episode
of suspected bacteremia.
· In critically ill patients in whom it may not be possible to delay treatment, no interval is required between taking sets of blood cultures.
Central venous catheter infections
· When a central venous catheter (CVC) is suspected
as a source of bacteremia, diagnosis of CVC infection
may be made by blood culture based techniques if (a)
the patients clinical condition permits a potentially
infected line to be left in place, (b) treatment of
CVC infection is to be attempted, or (c) other potential sources of bacteremia are apparent.
· While the acridine orange leukocyte cytospin test offers the possibility of virtually immediate diagnosis,
on the basis of currently available data, its use should
remain experimental.
· When a CVC is suspected as a source of sepsis in nonbacteremic patients, definitive diagnosis requires
that the CVC should be removed and sent for culture.
· If infection is suspected at the catheter site, swabs
should be taken from the insertion site for culture.
· The presence of purulence at the CVC site should
prompt catheter replacement at a distant site irrespective of culture results.
S 129
Ventilator-associated pneumonia
Acute cholecystitis
· Ventilator-associated pneumonia (VAP) should be
considered as a source of sepsis in any ventilated patient particularly in the first week following intubation, following aspiration, when a nasogastric/entral
feeding tube is in place, or when drugs have been given to raise gastric pH.
· Investigation of suspected VAP should include the
taking of two sets of blood cultures and a chest radiograph.
· Pleural effusions larger than 10 mm should be aspirated. Samples should be sent for immediate Gram
and fungal stains, culture and biochemistry including
protein, lactic dehydrogenase and glucose. Paired
blood chemistry samples should also be sent for comparison.
· Serology is not routinely indicated in the diagnosis of
VAP.
· A sample of secretions aspirated via the endotracheal tube should be sent for Gram stain and for bacterial and fungal culture.
· Bronchoscopy should be performed unless contraindicated or unavailable.
· No significant advantage of one invasive diagnostic
approach over another has been consistently demonstrated. Choice of technique depends in practice primarily on available expertise and equipment.
· Acute acalculous cholecystitis should be suspected in
any sepsis patient, particularly postoperatively, when
there are either signs relating to the right upper
quadrant of the abdomen or obstructive liver function tests.
· When acute acalculous cholecystitis is suspected, ultrasound should be ordered urgently.
· If an initial ultrasound examination is not diagnostic,
computed tomography should be ordered.
· If computed tomography is unavailable, a repeat ultrasound should be performed after 24 h.
Surgical site infection and intra-abdominal sepsis
· Blood cultures should be sent when investigating suspected surgical site infection or deep abdominal infection.
· The presence of purulence or spreading cellulitis are
indications for taking wound swabs.
· Infection should be suspected particularly at ºcontaminatedº or ªdirtyº surgical sites.
· When contaminated or dirty abdominal wounds develop features of wound infection, a diagnosis of
anaerobic coinfection should be assumed irrespective of whether anaerobes are identified by routine
microbiology.
· In most situations ultrasound is be the modality of
first choice. When ultrasound is not diagnostic, computed tomography should be considered.
· Collections identified by radiology should, where
technically possible, be aspirated and drained under
radiological control, samples being sent for Gramstaining and culture.
Sinusitis
· Acute sinusitis should be suspected in any sepsis patient who has either a nasotracheal tube or a finebore nasogastric feeding tube, or who has suffered a
head injury.
· When sinusitis is suspected, radiography of the maxillary sinuses should be performed to detect the presence of fluid.
· When radiography does not demonstrate fluid in the
maxillary sinuses, computed tomography should be
performed.
· If either radiography or computed tomography demonstrates the presence of fluid, antral puncture
should be performed to allow definitive diagnosis
and therapeutic drainage before antibiotic therapy is
initiated.
Invasive Candida infection
· There are no data to support a policy of routine
screening of hospitalized patients for Candida colonization. However, in sepsis patients, invasive fungal
infection is more likely in patients who are heavily
colonized.
· When sepsis develops in patients colonized by Candida species at two or more sites, blood cultures should
be sent and lysis centrifugation performed if available.
· Isolates of Candida species from sterile sites should
be sent for specificity and sensitivity testing.
Antibiotics in sepsis (P.-Y. Bochud, M. P. Glauser,
T. Calandra)
· Retrospective studies have shown that early administration of appropriate antibiotics reduces the mortality in patients with bloodstream infections caused by
Gram-negative bacteria.
S 130
· By analogy with the observations made in patients
with Gram-negative sepsis and despite the lack of
substantial clinical data in the literature, it is likely
that appropriate antibiotic therapy reduces the morbidity and the mortality of Gram-positive sepsis.
· Antifungal therapy is recommended for patients with
candidemia. Whether early treatment is associated
with better outcome is unknown, and additional studies are needed to evaluate this question.
· Monotherapy with carbapenem antibiotics is as effective as combination therapy with a b-lactam and
an aminoglycoside for the empirical treatment of
nonneutropenic patients with severe sepsis.
· Monotherapy with third- or fourth-generation cephalosporins is as effective as combination therapy
with a beta-lactam and an aminoglycoside for the
empirical treatment of nonneutropenic patients with
severe sepsis.
· Extended-spectrum carboxypenicillins or ureidopenicillins combined with b-lactamase inhibitors have
been shown to be effective for the treatment of suspected infections in febrile, neutropenic cancer patients, and in patients with peritonitis or nosocomial
pneumonia. However, similar studies have not yet
been carried out in patients with severe sepsis or
shock.
· Monotherapy with aztreonam appears to be as effective as combination of a b-lactam and an aminoglycoside for the treatment of patients with documented
Gram-negative sepsis. The fact that aztreonam lacks
any appreciable activity against Gram-positive or
anaerobic bacteria precludes its use as empirical single-agent therapy in patients with severe sepsis.
· Fluoroquinolones have been shown to be highly effective for the treatment of documented Gram-negative bloodstream infections. However, data are lacking to support their use as single-agent treatment of
severe sepsis, especially as first-generation fluoroquinolones display suboptimal activities against Grampositive bacteria.
· Third- and fourth-generation cephalosporins and
carbapenem antibiotics are equally effective as empirical therapy in patients with severe sepsis.
· The indiscriminate use of glycopeptide antibiotics
(i.e., vancomycin or teicoplanin) for presumed
Gram-positive infections in patients with severe sepsis and septic shock should be avoided. However, glycopeptides are appropriate in severely ill patients
with catheter-related infections or in centers in which
methicillin-resistant staphylococci predominate. The
possible clinical benefit associated with the empirical
use of glycopeptides should be weighed against the
risks of selecting resistant organisms and of increased
toxicity. Most cases require additional Gram-negative coverage, at least until microbiological results
are available.
· Antifungal agents, such as fluconazole, should not be
used on a routine basis as empirical therapy in patients with severe sepsis and septic shock.
· Fluconazole is as effective as and less toxic than amphotericin B for the treatment of candidemia in nonneutropenic patients. However, if the patient is unstable or has been treated previously with fluconazole, it might be prudent to begin therapy with amphotericin B while waiting for the identification of
the Candida species and for the results of susceptibility testing. Whether 5-fluorocytosine should be combined with amphotericin B in unstable patients is debatable.
Hemodynamic support in septic shock (J.-L. Vincent)
· The goal of fluid resuscitation in septic shock is restoration of tissue perfusion and normalization of cellular metabolism.
· Volume repletion in patients with septic shock produces significant increases in cardiac output and systemic oxygen delivery, and fluids alone are sometimes sufficient to reverse hypotension and restore
hemodynamic stability.
· Requirements for fluid infusion are not easily determined, and therefore that the fluid challenge should
be titrated to the clinical endpoints of blood pressure,
heart rate, and urine output. Central venous pressure
is initially required to evaluate the complex relationship between intravascular blood volume and cardiac
function. It is difficult to give optimal values for cardiac filling pressures.
· When central venous pressure increases, a pulmonary artery catheter is probably required, although
its role has recently been questioned.
· Increases in cardiac output and systemic oxygen delivery are proportional to the degree of intravascular
volume expansion achieved.
· The optimal hemoglobin and hematocrit for patients
with septic shock is unclear. Most experts recommend hemoglobin levels of 9±10 gm/dl in patients
with septic shock. This degree of anemia is usually
well tolerated in most patients, even with cardiac impairment.
· When fluid challenge fails to restore an adequate arterial pressure and organ perfusion, therapy with vasopressor agents should be started. Vasopressor therapy may also be required transiently to sustain life
and maintain perfusion in the face of life-threatening
hypotension, even when cardiac filling pressures are
not elevated.
· The effects of dopamine on cellular oxygen supply in
the gut remain incompletely defined. The effects of
norepinephrine on splanchnic circulation are hardly
predictable. The combination of norepinephrine and
S 131
dobutamine appears to be more predictable and
more appropriate to the goals of septic shock therapy
than the effects of epinephrine alone.
· The hemodynamic effects of dopamine in patients
with septic shock are well established. Dopamine increases mean arterial pressure primarily by increasing cardiac index with minimal effects on systemic
vascular resistance. The increase in cardiac index is
due to an increase in stroke volume, and to a lesser
extent, to increased heart rate. Patients receiving
dopamine at rates greater than 20 mg kg±1 min±1
show increases in right heart pressures as well as in
heart rate, and therefore doses should not usually exceed 20 mg kg±1 min±1, at least not without adequate
hemodynamic monitoring.
· Norepinephrine markedly improves mean arterial
pressure and glomerular filtration. This is particularly true in the high output-low resistance state of
many septic shock patients. After restoration of systemic hemodynamics, urine flow reappears in most
patients and renal function improves without the use
of low-dose dopamine or furosemide. This fact supports the hypothesis that renal ischemia observed
during hyperdynamic septic shock is not worsened
by norepinephrine infusion and even suggests that
this drug may effectively optimize renal blood flow
and renal vascular resistance.
· Dobutamine is an adrenergic agonist that stimulates
b1, b2, and b1 adrenergic receptors. A number of studies have investigated the effect of dobutamine on cardiac function during sepsis or septic shock. The doses
utilized ranged from 2 to 28 mg kg±1 min±1. The majority of these studies found increases in cardiac index
combined with increases in stroke volume and heart
rate.
Source control in the management of sepsis
(M. F. Jimenez, J. C. Marshall)
· Surgical intervention in the form of dØbridement of
infected, devitalized, or nonbleeding tissue should
be undertaken rapidly following hemodynamic stabilization in patients with necrotizing soft tissue infections. This is a grade E recommendation supported
by level IV and level V evidence.
· The decision to intervene surgically in the patient
with infected pancreatic necrosis must weigh the potential advantages of removing a source of ongoing
bacterial proliferation against the inherent morbidity
of early surgery. In general, surgery should be delayed in the stable patient to permit adequate demarcation of tissue planes. This is a grade C recommendation supported by a single randomized trial and expert opinion.
· The diagnosis of intra-abdominal infection amenable
to source control measures can generally be made by
either ultrasound or computed tomography. Ultrasonography has the advantage of being portable and
inexpensive, but is highly operator dependent; computed tomography is especially useful in the evaluation of the retroperitoneum.
· The initial approach to well-defined and accessible
intra-abdominal abscesses should be percutaneous
drainage. Catheter drainage can also be used as a
temporizing measure to optimize the physiological
and hemodynamic condition of an acutely ill patient
prior to surgical exploration. Laparotomy should be
reserved for those circumstances in which there are
no well-defined collections, dead tissue requires
dØbridement, or residual collections cannot be treated percutaneously. Surgical intervention may also be
indicated to control a source of ongoing peritoneal
contamination. Rates of failure have increased as interventional radiologists have extended the indications for percutaneous drainage. If the clinical condition of the patient does not improve following the
initial drainage, follow-up computed tomography
should be performed to determine whether a residual
or missed collection is present, and surgical intervention should be considered.
· Current data support the concept that relaparotomy
ªon demand,º as indicated by worsening of the clinical status, absence of improvement, or evolving organ dysfunction is as efficacious as a more aggressive
approach. Planned relaparotomy is indicated for patients with ischemic bowel when intestinal viability
is a concern (ªsecond lookº), for patients with necrotizing pancreatitis when demarcation of necrotic tissue demarcation is not distinct, or when bleeding precludes complete dØbridement.
· Although tissue necrosis can often be detected by
such characteristic radiographic findings as gas in
the tissues, or nonenhancement of tissues following
administration of intravenous contrast, there is no
single test that can exclude the presence of tissue necrosis with certainty, and in circumstances in which
necrosis may be life threatening (for example, intestinal ischemia), it is often necessary to establish the diagnosis operatively.
· An infected central venous catheter can be safely
changed over a guidewire, provided there is not significant local soft tissue infection at the exit site.
This is a grade B recommendation supported by level
II evidence.
· There is no evidence that routine catheter replacement reduces the risk of catheter-related bacteremia.
Venous catheters should be changed only as needed
when evidence of infection is present (signs of inflammation, purulent discharge at the insertion site),
or when the catheter is not working. This is a level C
S 132
recommendation for central venous catheters, supported by level II evidence, and a level E recommendation for peripheral catheters, supported by level V
evidence.
· Definitive resection is preferable to proximal diversion and drainage for perforated diverticulitis, and
likely for other causes of intestinal perforation,
when the more demanding procedure of resection
can be performed safely. Extension of this principle
to other sites of gastrointestinal peroration such as
the esophagus requires balancing the risks of resection with the potential benefits. This is a grade D recommendation based on level III evidence.
· Primary anastomosis or colostomy are equally efficacious following colon resection for diverticulitis. The
choice of procedure should be dictated by other factors such as severity of illness, presence of chronic
disease, the degree or duration of peritoneal contamination, and the skill and experience of the surgical
team. This is a grade D recommendation based on
level III evidence.
· Intra-abdominal infectious complications mandating
source control are almost always evident using modern diagnostic imaging techniques. There is little if
any role for empirical laparotomy to rule out undiagnosed infection in a critically ill patient in whom radiological examination has failed to demonstrate a
surgically correctable problem.
· Mechanical ventilation of patients with ALI should
be conducted with small tidal volumes (approximately 6 ml/kg ideal body weight) with the goal to maintain end-inspiratory plateau pressures at levels less
than 30 cmH2O.
· Prone positioning may be considered in patients requiring high levels of inspired oxygen (FIO2 > 0.60)
in whom positional changes are not contraindicated,
and who are cared for at facilities experienced in the
management of critically ill mechanically ventilated
patients.
· Restrict nitric oxide as an option for salvage therapy
in patients with life-threatening hypoxemia not responding to traditional mechanical ventilation strategies.
· Judicious use of crystalloid fluid administration
should be practiced in patients with ALI/ARDS,
with colloid solutions considered in hypo-oncotic patients with established ALI/ARDS. It is not clear
whether or not volume restriction improves outcome.
· Do not routinely administer corticosteroids to patients at risk for, or meeting criteria for, ALI/ARDS.
Consider intravenous methylprednisolone in patients
with persistent or refractory ARDS after actively excluding infection.
· All patients requiring acceptable levels of ventilatory
support who are not overtly unstable should receive a
spontaneous breathing trial on a daily basis to ascertain their ability to breathe unassisted.
Airway and lung in sepsis (G. S. Martin, G. R. Bernard)
· Provide adequate supplemental oxygen to maintain
an oximetric saturation of 90 % through use of simple
oxygen delivery systems (i.e., nasal cannula or face
mask), if possible. In endotracheally intubated patients, use of positive end-expiratory pressure to increase mean airway pressure may be employed to reduce concentrations of inspired oxygen below potentially toxic thresholds (FIO2 < 0.60).
· Avoid the use of noninvasive positive-pressure ventilation in patients with sepsis-related acute lung injury (ALI)/acute respiratory distress syndrome
(ARDS).
· Use early placement of an endotracheal tube and institution of mechanical ventilation in patients with
sepsis. Indications for institution of mechanical ventilation include severe tachypnea (respiratory rate
> 40 bpm), muscular respiratory failure (use of accessory muscles), altered mental status, and severe hypoxemia despite supplemental oxygen.
· In mechanically ventilated ALI/ARDS patients with
high inspiratory pressures or otherwise at risk for
barotrauma or volutrauma, implement permissive
hypercapnia through reduced tidal volume ventilation.
Immunological therapy in sepsis: currently available
(J. Carlet)
· Corticosteroids should not be used in severe sepsis or
septic shock at high doses (30 mg/kg), and for a short
course (1±2 days). On the other hand, corticosteroids
may be used during ªrefractoryº septic shock but not
during severe sepsis without shock or mild shock.
They should then be used at low doses (100 mg hydrocortisone three times a day) for 5 days or more
(up to 10 days) and then with subsequent tapering of
the dose according to the hemodynamic status. The
results of a large trial will be available shortly and
must be considered before definite recommendations can be made.
· Ibuprofen should not be used during severe sepsis
and septic shock. Additional studies are needed to
determine whether some patients, for example, those
with hypothermia, could benefit from the drug.
· Prostaglandins, in particular prostaglandin E1 or liposomal prostaglandin E1 should not be used during
acute respiratory distress syndrome due to sepsis.
There are no specific data allowing recommendations in severe sepsis.
S 133
· Pentoxifyilline should not be used in adults with severe sepsis unless new studies show a significant effect. The positive effect of a small study in infants
should be confirmed before clinical use.
· N-Acetylcysteine should not be used in severe sepsis
until new data are available, focusing in particular
on very early therapy.
· Selenium should not be used for severe sepsis. Additional studies are warranted to confirm initial positive data.
· Antithrombin III should not be used during severe
sepsis. Countries which allow the free use of this
drug in this setting should reconsider their position.
· Immunoglobulins should not be used either in adult
patients or in neonates with sepsis, unless additional
large studies confirm some positive data in smallsized meta-analyses. Countries which allow a wide
use of these compounds should reconsider their position and encourage these studies.
· Granulocyte colony stimulating factor should not be
used in nonneutropenic patients with severe sepsis.
· Growth hormone should not be used in patients with
sepsis because it increases mortality.
· Hemofiltration should not be used in patients with
sepsis, without renal indications unless ongoing studies provide positive data.
Other supportive therapies in sepsis
(J. PØrez, R. P. Dellinger)
· Considering the frequent occurrence of independent
risk factors for Deep vein thrombosis (DVT) in septic patients and the high percentage of sepsis/infected
patients included in studies that have demonstrated
efficacy of DVT prophylaxis in general, septic patients should be treated with DVT prophylaxis.
Even though there is not a randomized study that establishes the impact of DVT prophylaxis on morbidity and mortality specifically in septic patients, the significant number of septic patients included in the
populations of patients enrolled in other prospective
randomized trials supports that the use of DVT prophylaxis reduces morbidity and mortality in septic
patients. Moreover, septic patients, especially those
with severe sepsis and multiple organ failure, have
less cardiopulmonary reserve, and the impact of a minor thromboembolic event in this group of patients
could be very compromising.
· Septic patients who do not have a contraindication to
heparin use should receive prophylaxis with either
low-dose unfractionated heparin (5,000 U either two
or three times daily) or low molecular weight heparin
(at recommended doses; grade A).
· For those septic patients who have an absolute contraindication for heparin use (i.e., thrombocytopenia,
severe coagulopathy, active bleeding, recent intracerebral hemorrhage), the use of a mechanical prophylactic device is advised since this method has proven
to be effective in postsurgical patients and therefore
would likely work in septic patients (grade E).
· Based on the assumption that sepsis produces a hypercatabolic state and leads to protein-energy malnutrition, and given that protein loss is associated
with poor outcome, nutritional support in septic patients is recommended. The correlation of nutritional
support with outcome in septic patients comes from
data extrapolated from studies performed in perioperative patients and from expert opinion that allow
us to establish this recommendation. Many important questions remain regarding what kind of nutrition and when in the course of sepsis should nutrition
begin.
· Enteral nutrition is the preferred method of nutritional support in the catabolic critically ill patient in
general, inclusive of the septic patient (grade C).
For those patients who cannot tolerate enteral nutrition for a prolonged time or when contraindications
do not allow its use (mesenteric ischemia, mechanical
bowel obstruction), parenteral nutritional support
should be used (grade E). Immune-enhancing formulas may be better than other enteral formulations in
critically ill patients, but effects on ultimate outcome
(i.e., survival) remain to be demonstrated in large
randomized trials (grade B).
· The following are specific recommendations for septic patients, according to the guidelines established
by the American College of Chest Physicians and
American Society of Parenteral and Enteral Nutrition consensus conferences:
· Daily caloric intake: 25±30 kcal/kg usual body
weight
· Protein: 1.3±2.0 g/kg per day
· Glucose: 30±70 % of total nonprotein calories, to
maintain serum glucose level below 225 mg/dl
· Lipids: 15±30 % of total nonprotein calories. w 6
Polyunsaturated fatty acid should be reduced in septic patients, maintaining that level which avoids deficiency of essential fatty acids (7 % of total calories ±
generally 1 g/kg per day).
· No specific recommendations are offered for use of
medium-chain triglycerides, branched-chain amino
acids, or specific microelements added to the nutritional formulas. The use of any of these strategies, although supported in concept, does not have enough
investigational evidence to determine any clinical
benefit in outcome of septic patients.
· No randomized trial has evaluated the effect of stress
ulcer prophylaxis (SUP) on clinical outcome in septic
patients. Examination of successful clinical trials of
S 134
SUP does not allow precise identification of patients
with diagnosis of sepsis. Therefore no definitive data
exist in septic patients on the effectiveness of SUP
in diminishing episodes of overt or clinically significant bleeding. The clinical utility of SUP as it affects
clinical outcome in septic patients is therefore not
clear. Septic patients have been assumed to have an
increased risk for SUP since they have multiple risk
factors known to increase the risk of stress ulcer
bleeding. Since data do support SUP as being efficacious in preventing upper gastrointestinal bleeding
in populations of critically ill patients, which would
be expected to contain large proportions of septic patients, the use of SUP is recommended in this group
(see below).
· Although no large randomized trial has addressed
septic patients alone, abundant data exist regarding
subgroups of septic patients with prolonged mechan-
ical ventilation, hypotension, and coagulopathy. For
these patients the use of SUP is recommended (grade
A). For other septic patients in whom these factors
are not present, SUP is recommended based on several small randomized trials in which SUP has proven
efficacious in preventing bleeding and therefore reducing morbidity in critically ill patients (grade C).
· Several trials have confirmed the efficacy of antacids,
sucralfate or histamine-2 receptor antagonists in the
prevention of stress ulcer bleeding. Since the data
are conflicting, no single one can be determined as
preferable. General recommendations should be
based on the individual experience in the use of one
or another, the availability, or cost-analysis in individual centers. In septic patients with risk factors,
the use of enteral nutrition following the preventive
strategies currently available may be beneficial for
preventing stress ulcer bleeding.