1. health and fitness
2. disease

ORIGINAL INVESTIGATION
ONLINE FIRST
|
LESS IS MORE
Association of Blood Transfusion With
Increased Mortality in Myocardial Infarction
A Meta-analysis and Diversity-Adjusted Study Sequential Analysis
Saurav Chatterjee, MD; Jørn Wetterslev, MD, PhD; Abhishek Sharma, MD;
Edgar Lichstein, MD; Debabrata Mukherjee, MD, MS
Background: The benefit of blood transfusion in pa-
tients with myocardial infarction is controversial, and a
possibility of harm exists.
Methods: A systematic search of studies published between January 1, 1966, and March 31, 2012, was conducted using MEDLINE, EMBASE, CINAHL, Scopus,
Web of Science, and Cochrane Central Register of Controlled Trials databases. English-language studies comparing blood transfusion with no blood transfusion or a
liberal vs restricted blood transfusion strategy were
identified. Two study authors independently reviewed
729 originally identified titles and abstracts and selected 10 for analysis. Study title, follow-up period,
blood transfusion strategy, and mortality outcomes
were extracted manually from all selected studies, and
the quality of each study was assessed using the
strengthening Meta-analysis of Observational Studies in
Epidemiology checklist.
Results: Studies of blood transfusion strategy in anemia associated with myocardial infarction were abstracted, as well as all-cause mortality rates at the longest available follow-up periods for the individual studies.
Pooled effect estimates were calculated with randomeffects models. Analyses of blood transfusion in myocar-
Author Affiliations: Division of
Cardiology, Department of
Medicine, Brown University,
and Providence Veterans Affairs
Medical Center, Providence,
Rhode Island (Dr Chatterjee);
Copenhagen Trial Unit, Centre
for Clinical Intervention
Research, Rigshospitalet,
Copenhagen, Denmark
(Dr Wetterslev); Division of
Internal Medicine, Maimonides
Medical Center, New York,
New York (Drs Sharma and
Lichstein); and Division of
Cardiology, Texas Tech
University, El Paso
(Dr Mukherjee).
dial infarction revealed increased all-cause mortality associated with a strategy of blood transfusion vs no blood
transfusion during myocardial infarction (18.2% vs 10.2%)
(risk ratio, 2.91; 95% CI, 2.46-3.44; P ⬍ .001), with a
weighted absolute risk increase of 12% and a number
needed to harm of 8 (95% CI, 6-17). Multivariate metaregression revealed that blood transfusion was associated with a higher risk for mortality independent of baseline hemoglobin level, nadir hemoglobin level, and change
in hemoglobin level during the hospital stay. Blood transfusion was also significantly associated with a higher risk
for subsequent myocardial infarction (risk ratio, 2.04; 95%
CI, 1.06-3.93; P =.03).
Conclusions: Blood transfusion or a liberal blood transfusion strategy compared with no blood transfusion or
a restricted blood transfusion strategy is associated with
higher all-cause mortality rates. A practice of routine or
liberal blood transfusion in myocardial infarction should
not be encouraged but requires investigation in a large
trial with low risk for bias.
Arch Intern Med.
Published online December 24, 2012.
doi:10.1001/2013.jamainternmed.1001
T
HROMBOLYSIS, ANTICOAGUlation, and antiplatelet
drugs have revolutionized
the therapeutic approach to
acute coronary syndrome,
with significant improvements in clinical
outcomes.1-7 However, such therapy may
concomitantly increase the risk for bleeding, leading to the development of anemia during the hospital stay and to subsequent blood transfusion.8-10 Despite
advancement in reperfusion therapy, patients with low hemoglobin levels continue to have more postoperative complication and higher mortality rates,11-14 and
the presence of anemia in acute myocardial infarction has been associated with
worse prognosis.15-19
ARCH INTERN MED
PUBLISHED ONLINE DECEMBER 24, 2012
E1
In patients with significant coronary occlusion, low oxygen-carrying capacity secondary to anemia may further compromise the myocardial oxygen supply,
worsening the ischemia. By increasing the
oxygen-carrying capacity, blood transfusion might be beneficial, especially in anemia secondary to acute blood loss. However, inappropriate blood transfusion may
lead to circulatory overload and increased thrombogenicity, which can
worsen the clinical outcomes. Furthermore, elevated hemoglobin levels have
been shown to be associated with increased mortality.11,19 In post hoc analyses of a multicenter randomized controlled trial, Hébert et al20 reported no
mortality benefits of liberal blood transWWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013
Author Affil
Cardiology,
Medicine, Br
and Provide
Medical Cen
Rhode Islan
Copenhagen
for Clinical
Research, Ri
Copenhagen
Wetterslev);
Medicine, M
Center, New
(Drs Sharma
and Division
Texas Tech U
(Dr Mukher
METHODS
and for patients with a hematocrit of less than 30% (to convert
hematocrit to a proportion of 1.0, multiply by 0.01).
Data analyses were performed using commercially available software (RevMan 5.1 [Cochran IMS], TSA version 0.9 [Copenhagen Trial Unit], and STATA version 11 [StataCorp LP]).
Outcomes were assessed using random-effects models to exclude the presence of significant heterogeneity (evaluated and
quantified with the I2 statistic), and then pooled randomeffects risk ratios (RRs [95% CIs]) were calculated following
the method by DerSimonian and Laird.
The quality of the studies was assessed on the basis of elements from the strengthening Meta-analysis Of Observational
Studies in Epidemiology checklist for cohort studies. We did
not assign a threshold for study inclusion. All the studies included in the analyses met at least 15 variables in the checklist.
DATA SOURCES AND SEARCHES
DATA SYNTHESIS AND ANALYSIS
A systematic search of studies published between January 1,
1966, and March 31, 2012, was conducted using MEDLINE,
EMBASE, CINAHL, Scopus, Web of Science, and Cochrane Central Register of Controlled Trials databases. Studies were identified that evaluated mortality outcomes and compared blood
transfusion with no blood transfusion or a liberal vs restricted
blood transfusion strategy. The search terms used were transfusion, myocardial infarction, and mortality. Pertinent trials were
also searched for in clinicaltrials.gov and in proceedings from
major international cardiology meetings (American College of
Cardiology, American Heart Association, European Society of
Cardiology, and Transcatheter Cardiovascular Therapeutics).
References of original and review articles were cross-checked.
Study selection was performed by 2 of us independently (S.C.
and A.S.), with disagreement resolved by consensus among all
the authors. Citations were first reviewed at the title and abstract level. Studies that were short-listed were then retrieved
in full text.
When available, odds ratios (ORs) reported in the articles
were used to determine event rates, and unadjusted ORs (95%
CIs) were used preferentially over adjusted ORs to avoid bias
from different types of adjustments in the various studies.24 If
ORs were not reported, we calculated them using the event
and sample size frequencies. If frequencies were not given,
ORs were estimated from percentages and were rounded to
the nearest integer. If any cell had a 0 count, ORs were calculated by adding 0.5 to all cell counts from the study to avoid
division by 0. The I2 statistic was used to examine the heterogeneity of effect sizes in the overall aggregations: I2 of less
than 25% indicates low heterogeneity, and I2 exceeding 75%
indicates high heterogeneity. Publication bias was evaluated
using a combination of a funnel plot–based method, the regression test by Egger, and the trim-and-fill method to estimate the number of missing studies and to calculate a corrected OR as if these studies were present. The effect of
potential outliers was examined by comparing the pooled estimate with estimates obtained after iterations using k minus 1
findings (each study is left out, and the effect is reestimated),
where k is the number of studies. Studies were treated as statistical outliers if the k minus 1 estimate produced a 95% CI
that did not overlap with the 95% CI of the aggregated estimate. P⬍.05 was considered statistically significant.
fusion in maintaining hemoglobin levels of at least 10
mg/mL in critically ill patients with cardiovascular disease. However, because patients receiving blood transfusion tend to be older and have advanced disease and
more associated comorbid conditions, these variables may
act as confounding factors when determining mortality
and morbidity rates.21-23 Because of conflicting results reported in various investigations, the benefits and harm
of blood transfusion in myocardial infarction remain controversial. In the present meta-analysis, we systematically evaluated the potential risk-benefit of blood transfusion in patients with myocardial infarction.
STUDY SELECTION
Studies were considered suitable for inclusion if they met the
following criteria: (1) they reported the effect of blood transfusion or a liberal blood transfusion strategy (blood transfusion levels at which there was no blood transfusion in the comparator arm) on mortality, (2) they had a comparator group
with no blood transfusion or a restricted blood transfusion strategy and identified a mean hemoglobin level for both groups,
and (3) they used statistical methods to minimize confounding between the groups (matching, covariate adjustment, or propensity-based adjustment). Our search was restricted to studies in the English language. Full texts of 16 potentially relevant
articles were independently reviewed by at least 2 of us (S.C.
and A.S.) to establish eligibility according to the inclusion criteria. To avoid confounding, we excluded studies assessing the
effect of transfusion of components other than whole blood or
red blood cells.
DATA EXTRACTION AND
QUALITY ASSESSMENT
Data abstraction and study appraisal were performed by 2 of us
(S.C. and A.S.) independently, with disagreement resolved by consensus among all authors. Key study and patient characteristics
were extracted, including the outcomes of all-cause mortality and
myocardial infarction reported at the longest available follow-up
periods. Prespecified subgroup analyses were planned for patients with ST-segment elevation myocardial infarction (STEMI)
ARCH INTERN MED
STUDY SEQUENTIAL ANALYSIS
In a single study, interim analyses increase the risk for type I
error. To avoid an increase of overall type I error, monitoring
boundaries can be applied to decide whether a single study could
be terminated early because the P value is sufficiently small.
Because no reason exists why the standards for a metaanalysis should be less rigorous than those for a single study,
analogous study sequential monitoring boundaries can be applied to meta-analysis as study sequential analysis.25,26 Cumulative meta-analyses of studies are at risk for producing random errors because of few data27 and repetitive testing of
accumulating data and because the requirement for the amount
of information analogous to the sample size of a single optimally powered clinical study might not be met.25,26
The underlying assumption for study sequential analysis is
that significance testing and calculation of the 95% CIs are performed each time a new study is published. Study sequential
analysis depends on the quantification of the required amount
of information. In this context, the smaller the required amount
is, the more lenient is the trial sequential analysis and the more
lenient are the criteria for significance.25,26 A required diversity (D2)–adjusted information size was calculated, with D2 being
PUBLISHED ONLINE DECEMBER 24, 2012
E2
WWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013
STUDY QUALITY
Identification
720 Records identified through
database searching
9 Additional records identified
through other sources
729 Records retained after
duplicates removed
Screening
729 Records screened
705 Records excluded
Eligibility
24 Full-text articles assessed
for eligibility
14 Full-text articles excluded
(no comparator group,
no statistical adjustment,
CABG bleeding studies,
or no report of mortality)
Included
10 Studies included in
qualitative synthesis
10 Studies included in
quantitative synthesis
Figure 1. Search strategy according to the Meta-analysis Of Observational
Studies in Epidemiology (MOOSE) checklist.40 CABG indicates coronary
artery bypass graft.
the relative variance reduction when the meta-analysis model
is changed from a random-effects model to a fixed-effects
model.28 D2 is the percentage of the variability between trials
to the within-trial variance and constitutes the percentage of
the variability between trials to the total variance in the metaanalysis. D2 is different from the intuitively obvious adjusting
factor based on the common quantification of heterogeneity,
the inconsistency (I2 statistic), which might underestimate the
required information size.28
Study sequential analysis was performed with an intent to
maintain an overall 5% risk for type I error, which is the standard in most meta-analyses and systematic reviews, and we calculated the required information size (ie, the meta-analysis information size needed to detect or reject an intervention effect
of a 20% relative risk increase for harm, with a risk for type II
error of 10%, at a power of 90%).25,26 Study monitoring boundaries were constructed with the conventional test boundary and
using methods by O’Brien and Fleming.29
RESULTS
Our MEDLINE search returned 720 studies. After elimination of duplicate results, EMBASE, Cochrane Central
Register of Controlled Trials, and the other registries returned 9 additional studies, leaving 729 studies for evaluation. Through a review of titles and abstracts, 705 studies were rejected for relevance. The remaining 24 articles
were reviewed and assessed for satisfaction of the inclusion and exclusion criteria. Ten studies that met all criteria were included in this analysis (Figure 1).
We used the published strengthening Meta-analysis Of
Observational Studies in Epidemiology40 checklist to select the studies for this review (Figure 1). The included
studies reported statistically adjusted effect estimates for
the outcome of mortality. Among 10 studies included,
all except one (low-bias risk) were intermediate-bias risk
studies as assessed by the Newcastle-Ottawa Scale41 for
quality assessment risk evaluation of adequacy of selection, comparability, and outcomes assessment for individual studies (Table 2).
PRIMARY OUTCOME
We identified 10 studies,30-39 including 203 665 study participants, that met our inclusion and exclusion criteria.
Only one study31 was a randomized trial; the others were
observational studies. Analyses of blood transfusion in
myocardial infarction revealed increased all-cause mortality associated with a strategy of blood transfusion vs
no blood transfusion during myocardial infarction (18.2%
vs 10.2%) (RR, 2.91; 95% CI, 2.46-3.44; P ⬍ .001), with
a weighted absolute risk increase of 12% (P ⬍ .001) and
a number needed to harm of 8 (95% CI, 6-17). The risk
remained unchanged even on exclusion of the only randomized study.31 Pooled analysis using adjusted rates of
mortality instead of the actual number of events yielded
an effect largely similar to that of the primary analysis
(eFigure 1; http://www.archinternmed.com). However,
the mortality risk with blood transfusion was found to
be mitigated when restricted to studies that included patients with STEMI (RR, 2.89; 95% CI, 0.54-15.58; P = .22)
and patients with a baseline hematocrit of less than 30%
(RR, 1.72; 95% CI, 0.39-7.63; P = .47) (eFigure 2 and eFigure 3). Multivariate meta-regression were performed using
the log of the RR as the dependent variable and adjusting for the following variables as covariates: follow-up
period, history of bleeding, baseline creatinine level, baseline hemoglobin level, nadir of hemoglobin level, and
change in hemoglobin level during the hospital stay, as
well as the use of glycoprotein IIb or IIIa, thrombolytics, or antiplatelets. The meta-regression showed that
blood transfusion is associated with higher mortality after adjustment for all these variables. However, we could
not adjust for demographic variables in the multivariate
model because we did not have patient-level data. Significant heterogeneity was noted among the outcomes
(I2 = 93%) (Figure 2). A sensitivity analysis performed
by sequentially excluding one study at a time and performing a cumulative evaluation identified no single study
as the source of heterogeneity. No significant publication bias was noted among the outcomes with visual inspection of the funnel plot, the regression test by Egger
(P = .40), or with trim-and-fill adjustment (Figure 3).
STUDY SEQUENTIAL ANALYSIS
STUDY CHARACTERISTICS
Studies were fairly homogeneous for inclusion and exclusion criteria, with a few key differences. These results are summarized in Table 1.
ARCH INTERN MED
The required diversity-adjusted information size
(D2 = 97%) for the outcome of all-cause mortality was calculated based on analyzing a 10.2% proportion of events
in the no blood transfusion or a restricted blood trans-
PUBLISHED ONLINE DECEMBER 24, 2012
E3
WWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013
Table 1. Key Features of Included Studies
Source
Aronson et al,30 2008
Cooper et al,31 2011
Jani et al,32 2007
Jolicoeur et al,33 2009
Nikolsky et al,34 2009
Rao et al,35 2004
Shishehbor et al,36
2009
Singla et al,37 2007
Wu et al,38 2001
Yang et al,39 2005
No. of
Patients
Mean
Patient
Age, y
Male
Sex, %
Follow-up
Period, mo
Baseline
Hemoglobin
Level, g/dL
Baseline
Hematocrit, %
STEMI, non-STEMI
STEMI, non-STEMI
STEMI, non-STEMI
STEMI
STEMI, non-STEMI
STEMI, non-STEMI
STEMI
2326
45
4623
5188
2060
24 112
948
69
76.4
70.21
71
67.5
68.9
67
57
48
46.66
47
47.6
58.5
59
6
1
NA
3
12
1
12
11.8
...
10.47
12.9
13.1
...
13.9
...
26.9
...
...
...
39.9
...
Non-STEMI
STEMI, non-STEMI
Non-STEMI
370
78 974
85 111
70
77.8
73
99
45.9
52.9
1
1
NA
8.91
Graded levels
...
...
...
35
Myocardial Infarction
Types
Abbreviations: Ellipsis, not known; NA, not applicable (in-hospital mortality rates reported); STEMI, ST-segment elevation myocardial infarction.
SI conversion factors: To convert hematocrit to a proportion of 1.0, multiply by 0.01; to convert hemoglobin level to grams per liter, multiply by 10.0.
Table 2. Newcastle-Ottawa Scale of Bias Risk for Individual Studies a
Adequacy of Selection
Source
Aronson et al,30
2008
Cooper et al,31
2011
Jani et al,32
2007
Jolicoeur
et al,33 2009
Nikolsky et al,34
2009
Rao et al,35
2004
Shishehbor
et al,36 2009
Singla et al,37
2007
Wu et al,38
2001
Yang et al,39
2005
Outcomes Assessment
Assessment
of Outcome
Follow-up
Period Long
Enough
for Outcome
to Occur
Adequacy
of Follow-up
Period Among
Cohorts
**
*
*
*
*
**
*
*
*
*
*
**
*
*
*
*
*
*
**
*
*
*
*
*
*
**
*
*
*
*
*
*
**
*
*
*
*
*
*
**
*
*
*
*
*
*
**
*
*
*
*
*
*
**
*
*
*
*
*
*
**
*
*
*
Representativeness
of the Exposed Cohort
Selection
of the
Nonexposed
Cohort
Ascertainment
of Exposure
Comparability
*
*
*
*
*
*
a All studies demonstrated that the outcome of interest was not present at the start of the study. Asterisks are the star ratings per the Newcastle-Ottawa Scale;
* and ** indicates the highest ratings for these categories.
fusion strategy group and evaluating for a 20% relative risk
increase with blood transfusion or a liberal blood transfusion strategy at ␣ = .05 and ␤ = .10 (90% power). The
cumulative z curves (calculated with both the conventional test boundary and the methods by O’Brien and
Fleming)29 crossed the study sequential monitoring boundary of harm, suggesting firm evidence for a 20% relative
risk increase with blood transfusion or a liberal blood transfusion strategy compared with no blood transfusion or a
restricted blood transfusion strategy (Figure 4). We also
constructed a L’Abbé plot42 and a small study regression
graph to avoid a possible error with the effect of inclusion of small studies on the composite outcome.
ARCH INTERN MED
SECONDARY OUTCOME
Blood transfusion was also significantly associated with
a higher risk for subsequent myocardial infarction (RR,
2.04; 95% CI, 1.06-3.93; P = .03). Significant heterogeneity was present for this outcome (I2 = 98%) as well. No
significant publication bias was detected (Figure 5).
COMMENT
Several important clinical findings emerged from our systematic review and meta-analysis. First and foremost, a
PUBLISHED ONLINE DECEMBER 24, 2012
E4
WWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013
No Transfusion
or Restricted
Transfusion or Liberal
Source
Aronson et al,30 2008
Cooper et al,31 2011
Jani et al,32 2007
Jolicoeur et al,33 2009
Nikolsky et al,34 2009
Rao et al,35 2004
Shishehbor et al,36 2009
Singla et al,37 2007
Wu et al,38 2001
Yang et al,39 2005
Events
54
1
150
53
11
192
39
29
1778
1463
Total
192
21
1033
204
82
2401
292
110
3690
12 724
Events
250
2
108
203
87
669
67
29
14 432
2751
20 749
Total (95% CI)
3770
Total events
2
Heterogeneity: τ = 0.05; χ 9 = 110.72 (P <.001), I 2 = 92%
Test for overall effect: z = 12.48 (P <.001)
Total
2134
24
3590
4984
1891
21 711
651
260
75 284
72 387
Weight, %
11.4
0.5
11.8
11.1
5.3
13.6
8.8
7.1
15.3
15.1
Risk Ratio,
M-H, Random
(95% CI)
2.40 (1.86-3.10)
0.57 (0.06-5.86)
4.83 (3.81-6.12)
6.38 (4.88-8.34)
2.92 (1.62-5.24)
2.60 (2.22-3.03)
1.30 (0.90-1.88)
2.36 (1.49-3.76)
2.51 (2.42-2.61)
3.03 (2.85-3.21)
182 916
100.0
2.91 (2.46-3.44)
Favors
Transfusion
Favors
Comparator
18 598
0.20
0.05
1.00
5.00
20.00
Risk Ratio, M-H, Random (95% CI)
Figure 2. Risk for all-cause mortality with liberal blood transfusion. M-H indicates Mantel-Haenszel. Diamond indicates the overall summary estimate for the
analysis (width of the diamond represents the 95% CI); boxes, the weight of individual studies in the pooled analysis; whiskers, the 95% CIs: when they are to the
left of the midline, representing a risk ratio of 1, it means that the risks of dying favor (ie, are less with) tranfusion; when they are to the right of the midline,
representing a risk ratio of 1, it means that the risks of dying favor (ie, are less with) the comparator arm; if the lines touch the midline, representing a risk ratio of
1, it means that the risks of dying are comparable for transfusion and the comparators.
400
350
Inverse SE
300
250
200
150
100
50
0
–0.10
–0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
RD
Figure 3. Trim-and-fill adjustment for assessment of publication bias.
RD indicates risk difference.
significant mortality risk is demonstrated with a policy
of liberal blood transfusion in patients with myocardial
infarction, especially in those patients without STEMI or
with a hematocrit of less than 30%, bringing to light a
real possibility of harm with the practice of routine blood
transfusion in patients with myocardial infarction. The
mortality outcome was associated with statistically significant heterogeneity, likely associated with concomitant therapies, heterogeneity in the patient population,
and clinical settings (which could not be adjusted for in
our analysis because of the unavailability of patientlevel data), but might also be due to the precision of the
large observational studies, conferring an artificially high
I2 statistic. Also notable was the fact that the results remained unchanged with meta-regression adjusting for certain variables, including the follow-up period, a history
of bleeding, baseline hemoglobin level, nadir of hemoglobin level, and change in hemoglobin level during the
hospital stay, indicating a significant risk for blood transfusion over and above that associated with bleeding (and
anemia) in myocardial infarction.43 The single randomARCH INTERN MED
ized trial included in our analysis, the Conservative Versus Liberal Red Cell Transfusion in Acute Myocardial Infarction Trial, 31 had a small sample size and was
underpowered to make any relevant conclusions regarding clinically important intervention effects. We attempted to overcome the risk for increased random error due to sparse data and repetitive testing in our analysis
by constructing study sequential monitoring boundaries, and our results indicated that random error was not
the greatest risk in our meta-analysis. However, confounding by indication may be an important problem because the patients liberally transfused may also be the patients with the most serious disease and at the greatest
risk for mortality as estimated by prognostic factors at
baseline independent of the interventions subsequently
used. We also drew a regression plot to avoid overestimating the effects of small studies on the overall outcomes. Our 95% CIs were narrow, with a large sample
size, indicating at least a possibility of real harm with liberal and indiscriminate blood transfusion practices.
Of even greater concern was our finding of a significantly greater risk for myocardial reinfarction with
blood transfusion. This finding seems to conform to recent findings of detrimental effects on platelet aggregation with blood transfusion.44 Overall, our findings are
consistent with recent recommendations by the AABB
(formerly the American Association of Blood Banks)45
and in a prior Cochrane review46 for a restrictive blood
transfusion policy in critically ill patients. Our analysis
attempts to address the lacuna in the knowledge about
blood transfusion practices among patients with acute
coronary syndromes as expressed in the aforementioned guidelines by providing an updated meta-analysis on the topic in the absence of an adequately powered
randomized trial.
We also found that the risks for blood transfusion became mitigated during our subgroup analyses in patients with STEMI and in patients with a baseline hematocrit of less than 30%. This suggests a future direction
of further research in identifying specific subgroups that
PUBLISHED ONLINE DECEMBER 24, 2012
E5
WWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013
9
Cumulative
z score
Favors Transfusion in Myocardial Infarction
8
7
6
5
4
3
2
1
0
No. of patients
(linear scaled)
203 665
–1
Favors Not Specified
–2
–3
–4
–5
–6
–7
z curve
–8
–9
Figure 4. Study sequential analysis of observational studies on all-cause mortality with a 20% relative risk increase (RRI), number needed to harm of 50, control
event proportion of 10%, ␣ = .05, ␤ = .10, diversity (D 2) of 97%, and required information size of 320 310. The analysis uses a random-effects model with a type I
error risk of 5% and a power of 90%. The blue line is the cumulative z score from the cumulative random-effects meta-analyses, and each black box indicates the
addition of data from a new study; the red lines are the study sequential monitoring boundaries calculated according to the O’Brien-Fleming ␣-spending function
and stopping rule. The red vertical line indicates the diversity-adjusted information size of 320 310 patients based on an a priori 20% RRI corresponding to a
number needed to harm of 50. There seems to be evidence for a 20% RRI even when the cumulative meta-analysis is adjusted for sparse data and multiple testing
on accumulated data, disregarding possible bias.
Transfusion or Liberal
Source
Cooper et al,31 2011
Jani et al,32 2007
Jolicoeur et al,33 2009
Nikolsky et al,34 2009
Rao et al,35 2004
Singla et al,37 2007
Yang et al,39 2005
Events
1
35
16
6
604
8
242
Total
21
1033
204
82
2401
110
12 724
No Transfusion
or Restricted
Events
0
102
128
44
1771
9
1447
16 575
Total (95% CI)
912
Total events
2
Heterogeneity: τ = 0.62; χ 9 = 246.84 (P <.001), I 2 = 98%
Test for overall effect: z = 2.14 (P = .03)
Total
24
3590
4984
1975
21 711
260
72 387
Weight, %
3.5
17.0
16.3
14.0
18.0
13.3
17.9
Risk Ratio
M-H, Random
(95% CI)
3.41 (0.15-79.47)
1.19 (0.82-1.74)
3.05 (1.85-5.04)
3.28 (1.44-7.49)
3.08 (2.84-3.35)
2.10 (0.83-5.30)
0.95 (0.83-1.09)
104 931
100.0
2.04 (1.06-3.93)
Favors
Liberal Tx/Tx
Favors
Restricted Tx/No
3501
0.01
0.1
1.0
10
100
Risk Ratio, M-H, Random (95% CI)
Figure 5. Risk for myocardial infarction with liberal blood transfusion (Tx). M-H indicates Mantel-Haenszel. Diamond indicates the overall summary estimate for
the analysis (width of the diamond represents the 95% CI); boxes, the weight of individual studies in the pooled analysis; whiskers, the 95% CIs: when they are to
the left of the midline, representing a risk ratio of 1, it means that the risks of myocardial infarction (ie, are less with) liberal Tx/Tx; when they are to the right of the
midline, representing a risk ratio of 1, it means that the risks of dying favor (ie, are less with) the comparator arm; if the lines touch the midline, representing a risk
ratio of 1, it means that the risks of a myocardial infarction are comparable for the 2 arms.
may accrue a real benefit from blood transfusion, overcoming its detrimental influence.47,48
Our study had several methodological limitations. All
the studies but one in our meta-analysis were observational, diverse study designs and patient characteristics
made interpretation of aggregated estimates challenging, and causality could not be inferred. In addition, despite our efforts at adjusting for different variables, the
ARCH INTERN MED
relevance and reliability of the results were limited. For
more definitive conclusions, randomized designs by blood
transfusion at different hemoglobin levels and hematocrits and relevant outcomes are needed. For these analyses, we had only summary estimates and were unable to
adjust for important patient-level covariates. Red blood
cell transfusion may be a surrogate for other variables that
could have favorable or adverse effects on outcomes (eg,
PUBLISHED ONLINE DECEMBER 24, 2012
E6
WWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013
baseline risk, infection, adenosine diphosphate, and
soluble CD40 ligand).49,50 However, these variables were
not reported in the studies used for this analysis and could
not be considered. We also restricted our search to English-language sources.
Our study method also had several strengths. The magnitude and consistency of the observed effects for blood
transfusion in myocardial infarction make the likelihood of random error affecting this observation unlikely. Moreover, we rigorously controlled for publication bias and used random-effects models, which are
generally better suited when studies are gathered only
from the published literature. We also evaluated for the
validity of our mortality findings by constructing study
sequential monitoring boundaries and inferred that our
results indicated a firm evidence of harm with a practice
of liberal blood transfusion in myocardial infarction.
In conclusion, this meta-analysis provides evidence
that rates of all-cause mortality and subsequent myocardial infarction are significantly higher in patients with
acute myocardial infarction receiving blood transfusion. Additional outcomes data are needed from randomized clinical trials that investigate important outcomes with adequate sample size and with low risk for
bias.
Accepted for Publication: August 20, 2012.
Published Online: December 24, 2012. doi:10.1001/2013
.jamainternmed.1001
Correspondence: Saurav Chatterjee, MD, Division of Cardiology, Department of Medicine, Brown University, and
Providence Veterans Affairs Medical Center, 40 Roger Williams Green, Providence, RI 02904 (sauravchatterjeemd
@gmail.com).
Author Contributions: Dr Chatterjee had full access to
all the data in the study and takes responsibility for the
integrity of the data and the accuracy of the data analysis. Study concept and design: Chatterjee, Wetterslev, and
Lichstein. Acquisition of data: Chatterjee and Sharma.
Analysis and interpretation of data: Chatterjee, Wetterslev, Lichstein, and Mukherjee. Drafting of the manuscript: Chatterjee, Wetterslev, and Sharma. Critical revision of the manuscript for important intellectual content:
Chatterjee, Wetterslev, Lichstein, and Mukherjee. Statistical analysis: Chatterjee and Wetterslev. Administrative, technical, and material support: Chatterjee and Lichstein. Study supervision: Wetterslev, Lichstein, and
Mukherjee.
Conflict of Interest Disclosures: None reported.
Online-Only Material: The eFigures are available at http:
//www.archinternmed.com.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
REFERENCES
16.
1. Anderson JL, Adams CD, Antman EM, et al; American College of Cardiology; American Heart Association Task Force on Practice Guidelines (Writing Committee to
Revise the 2002 Guidelines for the Management of Patients With Unstable Angina/
Non ST-Elevation Myocardial Infarction); American College of Emergency
Physicians; Society for Cardiovascular Angiography and Interventions; Society
of Thoracic Surgeons; American Association of Cardiovascular and Pulmonary
Rehabilitation; Society for Academic Emergency Medicine. ACC/AHA 2007 guidelines for the management of patients with unstable angina/non ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart
ARCH INTERN MED
17.
18.
19.
Association Task Force on Practice Guidelines (Writing Committee to Revise the
2002 Guidelines for the Management of Patients With Unstable Angina/Non STElevation Myocardial Infarction): developed in collaboration with the American
College of Emergency Physicians, the Society for Cardiovascular Angiography
and Interventions, and the Society of Thoracic Surgeons: endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation and the Society for Academic Emergency Medicine. Circulation. 2007;116(7):e148-e304. http:
//circ.ahajournals.org/content/116/7/e148.long. Accessed October 5, 2012.
Wright RS, Anderson JL, Adams CD, et al. 2011 ACCF/AHA focused update of
the Guidelines for the Management of Patients with Unstable Angina/Non-STElevation Myocardial Infarction (updating the 2007 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on
Practice Guidelines developed in collaboration with the American College of Emergency Physicians, Society for Cardiovascular Angiography and Interventions, and
Society of Thoracic Surgeons [published correction appears in Lancet.
2002;359(9323):2120]. J Am Coll Cardiol. 2011;57(19):1920-1959.
Eikelboom JW, Anand SS, Malmberg K, Weitz JI, Ginsberg JS, Yusuf S. Unfractionated heparin and low-molecular-weight heparin in acute coronary syndrome
without ST elevation: a meta-analysis. Lancet. 2000;355(9219):1936-1942.
Boersma E, Harrington RA, Moliterno DJ, et al. Platelet glycoprotein IIb/IIIa inhibitors in acute coronary syndromes: a meta-analysis of all major randomised
clinical trials. Lancet. 2002;359(9302):189-198.
Steinhubl SR, Berger PB, Mann JT III, et al; CREDO Investigators. Early and sustained dual oral antiplatelet therapy following percutaneous coronary intervention: a randomized controlled trial [published correction appears in JAMA.
2003;289(8):987]. JAMA. 2002;288(19):2411-2420.
Antithrombotic Trialists’ Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002;324(7329):71-86.
Yusuf S, Zhao F, Mehta SR, Chrolavicius S, Tognoni G, Fox KK; Clopidogrel in
Unstable Angina to Prevent Recurrent Events Trial Investigators. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without
ST-segment elevation. N Engl J Med. 2001;345(7):494-502.
Moscucci M, Fox KA, Cannon CP, et al. Predictors of major bleeding in acute
coronary syndromes: the Global Registry of Acute Coronary Events (GRACE).
Eur Heart J. 2003;24(20):1815-1823.
Alexander KP, Newby LK, Cannon CP, et al; American Heart Association Council
on Clinical Cardiology; Society of Geriatric Cardiology. Acute coronary care in
the elderly, part I: non-ST-segment-elevation acute coronary syndromes: a scientific statement for healthcare professionals from the American Heart Association Council on Clinical Cardiology: in collaboration with the Society of Geriatric
Cardiology. Circulation. 2007;115(19):2549-2569.
Segev A, Strauss BH, Tan M, Constance C, Langer A, Goodman SG; Canadian
Acute Coronary Syndromes Registries Investigators. Predictors and 1-year outcome of major bleeding in patients with non-ST-elevation acute coronary syndromes: insights from the Canadian Acute Coronary Syndrome Registries. Am
Heart J. 2005;150(4):690-694.
Reinecke H, Trey T, Wellmann J, et al. Haemoglobin-related mortality in patients
undergoing percutaneous coronary interventions. Eur Heart J. 2003;24(23):
2142-2150.
McKechnie RS, Smith D, Montoye C, et al; Blue Cross Blue Shield of Michigan
Cardiovascular Consortium (BMC2). Prognostic implication of anemia on inhospital outcomes after percutaneous coronary intervention. Circulation. 2004;
110(3):271-277.
Lee PC, Kini AS, Ahsan C, Fisher E, Sharma SK. Anemia is an independent predictor of mortality after percutaneous coronary intervention. J Am Coll Cardiol.
2004;44(3):541-546.
Carson JL, Duff A, Poses RM, et al. Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet. 1996;348(9034):1055-1060.
Arant CB, Wessel TR, Olson MB, et al; National Heart, Lung, and Blood Institute
Women’s Ischemia Syndrome Evaluation Study. Hemoglobin level is an independent predictor for adverse cardiovascular outcomes in women undergoing
evaluation for chest pain: results from the National Heart, Lung, and Blood Institute Women’s Ischemia Syndrome Evaluation Study. J Am Coll Cardiol. 2004;
43(11):2009-2014.
Nikolsky E, Mehran R, Aymong ED, et al. Impact of anemia on outcomes of patients undergoing percutaneous coronary interventions. Am J Cardiol. 2004;
94(8):1023-1027.
Vis MM, Sjauw KD, van der Schaaf RJ, et al. Prognostic value of admission hemoglobin levels in ST-segment elevation myocardial infarction patients presenting with cardiogenic shock. Am J Cardiol. 2007;99(9):1201-1202.
Eikelboom JW, Mehta SR, Anand SS, Xie C, Fox KA, Yusuf S. Adverse impact of
bleeding on prognosis in patients with acute coronary syndromes. Circulation.
2006;114(8):774-782.
Sabatine MS, Morrow DA, Giugliano RP, et al. Association of hemoglobin levels
PUBLISHED ONLINE DECEMBER 24, 2012
E7
WWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
with clinical outcomes in acute coronary syndromes. Circulation. 2005;111
(16):2042-2049.
Hébert PC, Yetisir E, Martin C, et al; Transfusion Requirements in Critical Care
Investigators for the Canadian Critical Care Trials Group. Is a low transfusion
threshold safe in critically ill patients with cardiovascular diseases? Crit Care Med.
2001;29(2):227-234.
Engoren MC, Habib RH, Zacharias A, Schwann TA, Riordan CJ, Durham SJ.
Effect of blood transfusion on long-term survival after cardiac operation. Ann
Thorac Surg. 2002;74(4):1180-1186.
Koch CG, Li L, Duncan AI, et al. Morbidity and mortality risk associated with red
blood cell and blood-component transfusion in isolated coronary artery bypass
grafting. Crit Care Med. 2006;34(6):1608-1616.
Koch CG, Khandwala F, Li L, Estafanous FG, Loop FD, Blackstone EH. Persistent
effect of red cell transfusion on health-related quality of life after cardiac surgery.
Ann Thorac Surg. 2006;82(1):13-20.
Deeks JJ, Dinnes J, D’Amico R, et al; International Stroke Trial Collaborative Group;
European Carotid Surgery Trial Collaborative Group. Evaluating nonrandomised intervention studies. Health Technol Assess. 2003;7(27):iii-x, 1-173.
Brok J, Thorlund K, Gluud C, Wetterslev J. Trial sequential analysis reveals insufficient information size and potentially false positive results in many
meta-analyses. J Clin Epidemiol. 2008;61(8):763-769.
Wetterslev J, Thorlund K, Brok J, Gluud C. Trial sequential analysis may establish when firm evidence is reached in cumulative meta-analysis. J Clin Epidemiol.
2008;61(1):64-75.
Thorlund K, Imberger G, Walsh M, et al. The number of patients and events required to limit the risk of overestimation of intervention effects in metaanalysis—a simulation study. PLoS One. 2011;6(10):e25491. http://www.ncbi
.nlm.nih.gov/pmc/articles/PMC3196500/. Accessed October 6, 2012.
Wetterslev J, Thorlund K, Brok J, Gluud C. Estimating required information size
by quantifying diversity in random-effects model meta-analyses. BMC Med
Res Methodol. 2009;9:86. http://www.ncbi.nlm.nih.gov/pmc/articles
/PMC2809074/. Accessed October 5, 2012.
O’Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics.
1979;35(3):549-556.
Aronson D, Dann EJ, Bonstein L, et al. Impact of red blood cell transfusion on
clinical outcomes in patients with acute myocardial infarction. Am J Cardiol. 2008;
102(2):115-119.
Cooper HA, Rao SV, Greenberg MD, et al. Conservative versus liberal red cell
transfusion in acute myocardial infarction (the CRIT Randomized Pilot Study).
Am J Cardiol. 2011;108(8):1108-1111.
Jani SM, Smith DE, Share D, et al. Blood transfusion and in-hospital outcomes
in anemic patients with myocardial infarction undergoing percutaneous coronary intervention. Clin Cardiol. 2007;30(10)(suppl 2):II49-II56.
Jolicoeur EM, O’Neill WW, Hellkamp A, et al; APEX-AMI Investigators. Transfusion and mortality in patients with ST-segment elevation myocardial infarction
treated with primary percutaneous coronary intervention. Eur Heart J. 2009;
30(21):2575-2583.
Nikolsky E, Mehran R, Sadeghi HM, et al. Prognostic impact of blood transfusion after primary angioplasty for acute myocardial infarction: analysis from the
CADILLAC (Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications) Trial. JACC Cardiovasc Interv. 2009;2(7):624-632.
ARCH INTERN MED
35. Rao SV, Jollis JG, Harrington RA, et al. Relationship of blood transfusion and
clinical outcomes in patients with acute coronary syndromes. JAMA. 2004;
292(13):1555-1562.
36. Shishehbor MH, Filby SJ, Chhatriwalla AK, et al. Impact of drug-eluting versus
bare-metal stents on mortality in patients with anemia. JACC Cardiovasc Interv.
2009;2(4):329-336.
37. Singla I, Zahid M, Good CB, Macioce A, Sonel AF. Impact of blood transfusions
in patients presenting with anemia and suspected acute coronary syndrome. Am
J Cardiol. 2007;99(8):1119-1121.
38. Wu WC, Rathore SS, Wang Y, Radford MJ, Krumholz HM. Blood transfusion in
elderly patients with acute myocardial infarction. N Engl J Med. 2001;345(17):
1230-1236.
39. Yang X, Alexander KP, Chen AY, et al; CRUSADE Investigators. The implications
of blood transfusions for patients with non-ST-segment elevation acute coronary syndromes: results from the CRUSADE National Quality Improvement
Initiative. J Am Coll Cardiol. 2005;46(8):1490-1495.
40. Stroup DF, Berlin JA, Morton SC, et al; Meta-analysis of Observational Studies
in Epidemiology (MOOSE) Group. Meta-analysis of observational studies in epidemiology: a proposal for reporting: Meta-analysis Of Observational Studies in
Epidemiology (MOOSE) group. JAMA. 2000;283(15):2008-2012.
41. Wells GA, Shea B, O’Connell D, et al; Ottawa Hospital Research Institute. The
Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses: case-control studies. http://www.ohri.ca/programs/clinical
_epidemiology/oxford.asp. Accessed March 2012.
42. Song F. Exploring heterogeneity in meta-analysis: is the L’Abbé plot useful? J Clin
Epidemiol. 1999;52(8):725-730.
43. Kumbhani DJ, Bhatt DL. Platelet activation: yet another strike against routine
TRANSFUSION. Eur Heart J. 2010;31(22):2712-2714.
44. Silvain J, Pena A, Cayla G, et al. Impact of red blood cell transfusion on platelet
activation and aggregation in healthy volunteers: results of the TRANSFUSION
study. Eur Heart J. 2010;31(22):2816-2821.
45. Carson JL, Grossman BJ, Kleinman S, et al; Clinical Transfusion Medicine Committee of the AABB. Red blood cell transfusion: a clinical practice guideline from
the AABB. Ann Intern Med. 2012;157(1):49-58.
46. Carson JL, Carless PA, Hebert PC. Transfusion thresholds and other strategies
for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev.
2012;4:CD002042.
47. Valente S, Lazzeri C, Chiostri M, et al. The impact of blood transfusion on short
and long term prognosis in STEMI patients treated with primary percutaneous
coronary intervention: a single center–experience. Int J Cardiol. 2012;157(2):
281-283.
48. Alexander KP, Chen AY, Wang TY, et al; CRUSADE Investigators. Transfusion
practice and outcomes in non-ST-segment elevation acute coronary syndromes.
Am Heart J. 2008;155(6):1047-1053.
49. Twomley KM, Rao SV, Becker RC. Proinflammatory, immunomodulating, and
prothrombotic properties of anemia and red blood cell transfusions. J Thromb
Thrombolysis. 2006;21(2):167-174.
50. Urbich C, Dernbach E, Aicher A, Zeiher AM, Dimmeler S. CD40 ligand inhibits
endothelial cell migration by increasing production of endothelial reactive oxygen species. Circulation. 2002;106(8):981-986.
PUBLISHED ONLINE DECEMBER 24, 2012
E8
WWW.ARCHINTERNMED.COM
©2012 American Medical Association. All rights reserved.
Downloaded From: http://archinte.jamanetwork.com/ by a University Of North Carolina - Chapel Hill User on 01/23/2013