- Email: [email protected]
Impact of More Restrictive Blood Transfusion Strategies on Clinical Outcomes: A Meta-analysis and Systematic Review
CLINICAL RESEARCH STUDY
Impact of More Restrictive Blood Transfusion Strategies on Clinical Outcomes: A Meta-analysis and Systematic Review Shelley R. Salpeter, MD,a Jacob S. Buckley,b Saurav Chatterjee, MDc a Stanford University School of Medicine, Stanford, Calif; bBrown University, Providence, RI; cSt Luke’s e Roosevelt Hospital Center, New York, NY.
ABSTRACT BACKGROUND: There is accumulating evidence that restricting blood transfusions improves outcomes, with newer trials showing greater benefit from more restrictive strategies. We systematically evaluated the impact of various transfusion triggers on clinical outcomes. METHODS: The MEDLINE database was searched from 1966 to April 2013 to find randomized trials evaluating a restrictive hemoglobin transfusion trigger of <7 g/dL, compared with a more liberal trigger. Two investigators independently extracted data from the trials. Outcomes evaluated included mortality, acute coronary syndrome, pulmonary edema, infections, rebleeding, number of patients transfused, and units of blood transfused per patient. Extracted data also included information on study setting, design, participant characteristics, and risk for bias of the included trials. A secondary analysis evaluated trials using less restrictive transfusion triggers, and a systematic review of observational studies evaluated more restrictive triggers. RESULTS: In the primary analysis, pooled results from 3 trials with 2364 participants showed that a restrictive hemoglobin transfusion trigger of <7 g/dL resulted in reduced in-hospital mortality (risk ratio [RR], 0.74; confidence interval [CI], 0.60-0.92), total mortality (RR, 0.80; CI, 0.65-0.98), rebleeding (RR, 0.64; CI, 0.45-0.90), acute coronary syndrome (RR, 0.44; CI, 0.22-0.89), pulmonary edema (RR, 0.48; CI, 0.33-0.72), and bacterial infections (RR, 0.86; CI, 0.73-1.00), compared with a more liberal strategy. The number needed to treat with a restrictive strategy to prevent 1 death was 33. Pooled data from randomized trials with less restrictive transfusion strategies showed no significant effect on outcomes. CONCLUSIONS: In patients with critical illness or bleed, restricting blood transfusions by using a hemoglobin trigger of <7 g/dL significantly reduces cardiac events, rebleeding, bacterial infections, and total mortality. A less restrictive transfusion strategy was not effective. Ó 2014 Elsevier Inc. All rights reserved. The American Journal of Medicine (2014) 127, 124-131 KEYWORDS: Clinical outcomes; Meta-analysis; Mortality; Systematic review; Transfusion SEE RELATED EDITORIAL p. 103
Red blood cell transfusions have been the standard of care for treating anemia for more than 100 years now, with little evidence that they improve clinical outcomes.1-3 By the
Funding: None. Conflict of Interest: None. Authorship: All authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to Shelley Salpeter, MD, 34 North San Mateo Drive, Suite 1, San Mateo, CA 94401. E-mail address: [email protected] 0002-9343/$ -see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjmed.2013.09.017
early 1900s, blood transfusion was considered to be “a procedure of such simple and harmless character” that no clinical indication was needed, “the mere possibility of benefitting a condition by the addition of blood being considered sufficient warrant.”1 The practice was based on the assumption that anemia is tolerated poorly and that red blood cell transfusions improve outcomes.1,4,5 Researchers did not begin to question the evidence behind the practice until the 1980s and 1990s, when the first randomized trials were performed.6-12 By that time, the practice of blood transfusion was so ingrained in our medical framework that
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
125
the approach has been to march slowly down on the transprotocol shown in Appendix Tables 1 and 2, online). We fusion trigger instead of addressing whether transfusions included trials of adults or children, including neonates, are beneficial at all. involving surgical or medical conditions. Trials that used a The standard transfusion trigger for many years had restrictive transfusion trigger more than 7 g/dL were evalbeen a hemoglobin of 10 g/dL or even higher.6,9,12,13 This uated separately as a “less restrictive” strategy. Additional arbitrary trigger has been lowered gradually to a hemoglobin searches of related articles were done to perform a syslevel of 6 to 8 g/dL because studies tematic review of the impact of showed that blood transfusions are various transfusion strategies. CLINICAL SIGNIFICANCE associated with worse outcomes in patients with anemia due to illness Pooled randomized trial data show that Data Extraction and Quality or bleeding, compared with simple blood transfusions increase in-hospital Assessment supportive measures such as hymortality, total mortality, rebleeding, 6,9,12-20 Two investigators (SS, JB) exdration. However, a liberal acute coronary syndrome, pulmonary tracted data from the trials, transfusion practice is still comedema, and bacterial infections. reconciling differences by conmon, especially for those with sensus. In addition, selected incoronary artery disease who are When a restrictive hemoglobin transvestigators were contacted for thought to benefit more from blood fusion trigger of <7 g/dL is used in additional information. Clinical transfusions.21 There have been patients with critical illness or bleed, the outcomes evaluated included inmany challenges inherent in the number needed to treat to prevent 1 hospital mortality, 30-day mortalstudy of our transfusion practices, death is 33. ity, total mortality, acute coronary such as the broad patient base syndrome, pulmonary edema, Observational data indicate that hemoincluded in the analyses, the mulbacterial infections, rebleeding, tiple indications for blood transglobin levels of 5 to 6 g/dL are well number of patients receiving any fusions, and the confounding by tolerated in normovolemic patients blood transfusion, and units of indication in observational studies. without affecting oxygen delivery. blood transfused per patient. We have found no randomized Extracted data also included inclinical trials comparing transformation on study setting, design, participant characterfusion with no transfusion. Instead, the available trials istics, and risk for bias for the included trials (detailed have compared more or less restrictive transfusion strastudy protocol shown in Appendix, online).24 tegies using different transfusion triggers. A previous meta-analysis pooled data from randomized trials that evaluated restrictive hemoglobin transfusion triggers Data Synthesis and Analysis ranging from 7 to 10 g/dL and found that restricting The results were reported as a risk ratio (RR) and risk diftransfusions significantly reduced in-hospital mortality but ference for dichotomous outcomes, for the restrictive strathad no effect on other clinical outcomes.22 We have egy compared with the liberal strategy, with the confidence chosen a different approach to evaluate the available evinterval (CI) set at 95% significance. For the amount of idence. We now update the meta-analysis through April blood transfused per patient, the results were reported as a 2013 to include a subsequent trial23 and restrict the primean difference, with 95% CIs for the restrictive compared mary analysis to those trials with a transfusion trigger of with the liberal strategy. To test for inter-study heteroge<7 g/dL. Trials that evaluated less restrictive strategies neity, the chi-square value was calculated; statistical sigwere evaluated in a separate analysis. We also provide a nificance was indicated by P < .1. The fixed-effects method systematic review of observational studies that evaluated was chosen to report the results because minimal heteroclinical outcomes related to other more restrictive transgeneity was seen in most of the analyses.25 When heterofusion strategies. geneity was noted, the random-effects method was used.26
MATERIALS AND METHODS Data Sources and Study Selection We conducted a comprehensive search of the MEDLINE database from 1966 to April 2013 using the terms blood transfusion and clinical trial, and scanned selected journals and references of identified articles. Studies of any language were included in the primary analysis if they were randomized controlled trials that evaluated a restrictive blood transfusion strategy using a transfusion trigger of <7 g/dL, compared with a more liberal strategy (detailed study
In a secondary analysis, the pooled results from trials using a less restrictive strategy were evaluated and compared with the trials in the primary analysis using the test for interaction.27 The analyses were performed using Review Manager, Version 5.2, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2012.
Role of Funding Source The investigators received no funding for the study. No sponsor had a role in any aspect of the study, including its design and conduct, data extraction and analysis, and preparation of the manuscript.
126
The American Journal of Medicine, Vol 127, No 2, February 2014
-
Potentially relevant studies identified and screened for retrieval (n = 4500)
Studies excluded: Not trials of transfusion triggers (n = 4369) Studies used for systematic review
Trials of transfusion triggers In anemia (n = 32)
Trials excluded: Not randomized (n = 2)
Randomized trials of transfusion triggers in anemia (n = 30)
Trials excluded from primary analysis: Used less-restrictive trigger (n = 19) (16 trials used in secondary analysis) Did not provide clear trigger (n = 4) Duplicate data from other trials (n = 4)
Primary analysis: Randomized controlled trials of a restrictive transfusion strategy using a hemoglobin trigger of < 7 g/dL, compared with a more liberal strategy (n = 3) Figure 1
Flow chart of trials search for meta-analysis.
RESULTS Primary Meta-analysis Search Results. The search identified approximately 4500 studies, of which 32 were potentially relevant trials evaluating transfusion triggers (Figure 1). Of these, 3 trials met inclusion criteria for the primary analysis.12,23,28 One study provided unpublished information.23 Studies were excluded for the following reasons: Two were not randomized, 19 used a less-restrictive hemoglobin transfusion trigger of >7 g/dL, 4 did not provide a clear transfusion trigger, and 4 provided duplicate data on participants included in another trial. Of the 19 trials evaluating a
less-restrictive strategy, 16 provided data on clinical outcomes and were evaluated separately.6-11,21,29-37 Trial Characteristics. The primary analysis included 3 trials, with a total of 2364 participants followed for mean trial duration of 45 days. The characteristics of the included trials, including their risk of bias, are shown in Appendix Figure 1 (online). The mean study size was 788 participants (range, 637-889), with a mean participant age of 45.7 years (standard deviation, 16 years). Transfusion strategies were evaluated in the setting of adult critical care,12 pediatric critical care,28 and acute upper gastrointestinal bleeding.23
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
127
Table 1 Outcomes for Restrictive Transfusion Strategy, with Hemoglobin Transfusion Trigger <7 g/dL, Compared with a More Liberal Strategy Outcome Mortality Hospital mortality 30-d mortality Total mortality Cardiac events Acute coronary syndrome Pulmonary edema Other outcomes Rebleeding Bacterial infections Blood transfusions Patients exposed to blood Units transfused per patient
Trials
Patients
RR or MD
RD
No. Needed to Treat
2 3 3
1727 2364 2364
RR, 0.74 [CI, 0.60-0.92] RR, 0.81 [CI, 0.61-0.96] RR, 0.80 [CI, 0.65-0.98]
RD, 0.04 [CI, 0.04 to 0.00] RD, 0.02 [CI, 0.04 to 0.00] RD, 0.03 [CI, 0.05 to 0.00]
25 50 33
2 3
1727 2364
RR, 0.44 [CI, 0.22-0.89] RR, 0.48 [CI, 0.33-0.73]
RD, 0.02 [CI, 0.03 to 0.00] RD, 0.03CI, 0.05 to 0.01
50 33
1 3
889 2364
RR, 0.64 [CI, 0.45-0.90] RR, 0.86 [CI, 0.73-1.00]
RD, 0.06 [CI, 0.10 to 0.01] RD, 0.03 [CI, 0.06 to 0.00]
17 33
3 3
2364 2364
RR, 0.57 [CI, 0.46-0.70] MD 1.98 [CI, 3.22 to 0.74]
RD, 0.41 [CI, 0.52 to 0.29]
2
CI ¼ confidence interval; MD ¼ mean difference; RD ¼ risk difference; RR ¼ risk ratio.
Data Synthesis. The restrictive transfusion strategy was associated with a statistically significant reduction of inhospital mortality (RR, 0.74; CI, 0.60-0.92), 30-day mortality (RR, 0.77; CI, 0.61-0.96), and total mortality (RR, 0.80; CI, 0.65-0.98), compared with the liberal strategy (Table 1, Figure 2). The risk difference for total mortality was 0.03 for the restrictive strategy compared with the liberal strategy (mean duration of included trials, 45 days), with a number needed to treat of 33 to save 1 life. In addition, the restrictive strategy resulted in a reduced incidence of acute coronary syndrome (RR, 0.44; CI, 0.22-0.89), pulmonary edema (RR, 0.48; CI, 0.330.72), rebleeding (RR, 0.64; CI, 0.45-0.90), and bacterial infections (RR, 0.86; CI, 0.73-1.00), compared with the liberal strategy (Table 1, Figure 3, Appendix, online).
Figure 2
The results for bacterial infections were of nominal significance. In pooled trial data, 55% of the participants in the restrictive group received a blood transfusion, compared with 94% in the liberal group (RR, 0.57; CI, 0.46-0.70 and risk difference, 0.41; CI, 0.52 to 0.29). The restrictive strategy resulted in a significant reduction in the mean number of units transfused, with a mean difference of 1.98 units (CI, 3.22 to 0.74 units) per patient. No evidence for inter-study heterogeneity was found in any of the clinical outcomes analyzed (P > .25, for chisquare test). There was evidence for significant heterogeneity in the analyses of exposure to blood transfusions and the mean number of units transfused (P < .00001).
Restrictive transfusion strategy and mortality. CI ¼ confidence interval.
128
The American Journal of Medicine, Vol 127, No 2, February 2014
Figure 3
Restrictive transfusion strategy and cardiac events. CI ¼ confidence interval.
Secondary Meta-analysis
Systematic Review
Randomized Trials Using Less-Restrictive Transfusion Triggers. Nineteen of the trials excluded from the primary analysis evaluated hemoglobin triggers of 7.5 to 10 g/dL in the restrictive strategy group, and 16 of those provided data on clinical outcomes.6-11,21,29-37 When the data from these trials were pooled, with a total of 4572 participants, the restrictive strategy significantly reduced the exposure to blood transfusions (RR, 0.60; CI, 0.48-0.75) and amount of blood transfused (mean difference, 0.80 units; CI, 1.24 to 0.37 units), compared with a more liberal strategy, but this had no significant effect on in-hospital mortality (RR, 0.65; CI, 0.37-1.15), total mortality (RR, 1.03; CI, 0.811.31), acute coronary syndrome (RR, 1.46; CI, 0.96-2.20), pulmonary edema (RR, 1.02; CI, 0.67-1.56), rebleeding (RR, 0.68; CI, 0.34-1.34), or bacterial infections (RR, 0.80; CI, 0.64-1.02). When the pooled results for the trials from the primary analysis (hemoglobin trigger <7 g/dL) were compared with the results from the less restrictive trials (hemoglobin trigger 7.5-10 g/dL) using the test for interaction, the more restrictive strategy was associated with a greater reduction in acute coronary syndrome (P ¼ .004), pulmonary edema (P ¼ .002), and amount of blood transfused (P ¼ .01), compared with the less restrictive strategy. When all 19 trials from the primary and secondary analyses were pooled together, with a total of 6936 participants, the restrictive strategy was still associated with a significant reduction in hospital mortality (RR, 0.73; CI, 0.59-0.89), 30-day mortality (RR, 0.83; CI, 0.69-0.99), pulmonary edema (RR, 0.68; CI, 0.51-0.90), bacterial infections (RR, 0.84; CI, 0.73-0.95), and rebleeding (RR, 0.64; CI, 0.47-0.88).
Observational Studies Evaluating More Restrictive Transfusion Strategies. Recent recommendations to further restrict blood transfusions have been based on studies showing the tolerability and safety of anemia under controlled normovolemic conditions, which involve the administration of fluids, oxygen, and beta-adrenergic blockers.14,15,38-45 A meta-analysis of randomized trials in the perioperative setting showed that normovolemic hemodilution, in which blood is removed and replaced with crystalloid or colloid solution, resulted in significantly less total blood loss and allogenic blood transfusions, compared with standard care.46 A systematic review found consistent evidence that normovolemic anemia is associated with a reduction in systemic vascular resistance and an increase in cardiac output, coronary and cerebral blood flow, and synthesis of 2,3-diphosphoglycerate in red blood cells, resulting in maintenance of oxygen delivery and extraction.39 Observational studies have shown that hemoglobin levels of 5 to 6 g/dL in the setting of surgery, critical illness, and acute bleeds generally are well tolerated when standard supportive measures are given, without evidence of cardiac ischemia or decrease in oxygen extraction until the hemoglobin decreases to less than 3 to 4 g/dL.47-56
DISCUSSION Pooled data from randomized controlled trials show that restricting blood transfusions to patients whose hemoglobin decreases to less than 7 g/dL results in a significant reduction in total mortality, acute coronary syndrome, pulmonary edema, rebleeding, and bacterial infection, compared with a
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
more liberal transfusion strategy. The number needed to treat to save 1 life was 33. This strategy resulted in a 40% reduction in the number of patients receiving a blood transfusion, with an average of 2 units less per person; however, more than one half of patients still received a transfusion. In an analysis of trials that used a less restrictive strategy, with hemoglobin triggers of 7.5 to 9 g/dL, no significant reduction in morbidity or mortality was seen. With the available evidence from observational studies, an even more restrictive transfusion strategy using a hemoglobin trigger of <6 g/dL has been recommended in some settings.14,15 Observational studies have consistently shown that transfusions are associated with an increased risk for adverse events after controlling for potential confounding variables, even when using a restrictive transfusion strategy.56-63 The increased risk seems to be directly proportional to the amount of blood transfused and the length of storage of the transfused red blood cells, and may be due to an inflammatory response to the transfused blood product.58,64-66 Systematic reviews also have evaluated the effect of blood transfusions on oxygen transport variables in anemic patients in the setting of surgery, critical illness, and bleed, and found no significant improvement in oxygen delivery or use compared with supportive care, despite an increase in oxygen content.5,57,58 This inability to improve oxygen uptake in vital organs is due to the hemodynamic response to increased blood viscosity and the loss of red cell function during preservation and storage.5,57,58,67-71 Conversely, there is evidence that normovolemic anemia with hemoglobin levels of 5 to 6 g/dL is well tolerated in cardiovascular or critical illness and may have beneficial hemodynamic effects.39,42,47,51 It has been the traditional teaching that patients with cardiac ischemia should have a more liberal transfusion strategy to maintain oxygenation, but pooled observational studies on transfusion in myocardial infarction found that the rates of subsequent myocardial infarction and all-cause mortality were significantly higher in patients receiving blood transfusions compared with standard supportive measures, after adjustment for possible confounding variables.72 Multivariate meta-regression analysis of the pooled data revealed that the increased risk was independent of the baseline or nadir hemoglobin level.72 Subgroup analysis of a large critical care trial included in the primary analysis found that in patients with coronary ischemia, a transfusion trigger of <7 g/dL may be associated with improved clinical outcomes.12,73 Two small trials included in the secondary analysis evaluated a restrictive trigger of <8 g/dL in patients with symptomatic coronary artery disease; pooled results from these 2 trials showed no significant effect on cardiac events or mortality using this less restrictive strategy.21,30 This meta-analysis found that restricting transfusions using a hemoglobin trigger of <7 g/dL reduced mortality in critical illness or bleed, with a number needed to treat of 33 to save 1 life. With millions of blood transfusions given yearly over the past century, it would be hard to calculate how many deaths may have been caused by transfusions.
129
The main goal of blood transfusions is to increase oxygencarrying capacity, but despite increasing oxygen content, oxygen delivery is not increased.5,57,58 Another indication for transfusions has been to stabilize bleeding patients, but in fact transfusions significantly increase the risk for rebleeding.6,23 So far, there is little trial evidence that blood transfusions significantly improve oxygen delivery or clinical outcomes in any setting or with any nadir hemoglobin level.5,57,58,74
Study Limitations This meta-analysis and systematic review has several limitations. The trials in the primary meta-analysis studied disparate populations, including both pediatric and adult patients, and addressed different indications for transfusion, such as critical illness and gastrointestinal bleed. This has the potential for introducing bias in the analysis, despite the fact that we found no evidence for inter-study heterogeneity in the results. Another potential bias is that there was incomplete blinding of the participants in the individual trials because of the nature of the interventions. However, the hard clinical outcomes studied were unlikely to have been influenced by this. Another limitation of the study is that other patient populations, such as those with less lifethreatening illnesses, were not included in the primary analysis. The objective of this study was to put together the available evidence for practicing clinicians to make sense of it all, but we are left with many unanswered questions. Unfortunately, more than one half of the patients in the restrictive group still received a blood transfusion, so we cannot directly assess the effect of transfusion versus no transfusion. At present, there are no randomized trials evaluating lower transfusion triggers, such as a hemoglobin level of 6 g/dL, which is what some of the newer guidelines recommend using on the basis of observational studies.14,15 In addition, there are no randomized trials evaluating the lower transfusion triggers in acute coronary syndrome, which at present is considered an indication for the use of a more liberal trigger. Finally, there is little information on clinical situations or nadir hemoglobin levels for which transfusions are known to improve oxygen delivery and mortality.
CONCLUSIONS We have performed an updated meta-analysis of randomized trials that shows that a restrictive transfusion strategy using a hemoglobin transfusion trigger of <7 g/dL results in a significant reduction in acute coronary syndrome, pulmonary edema, rebleeding, infections, and total mortality, compared with a more liberal strategy. At present, there is no randomized trial evidence that blood transfusions improve oxygen delivery or clinical outcomes in any setting. More studies are needed to help guide clinicians in finding optimal treatment threshold and options in the setting of anemia and bleed.
130
References 1. Bernheim BM. Blood Transfusion, Hemorrhage and the Anaemias. Philadelphia and London: JB Lippincott Company; 1917. 2. Baskett TF. James Blundell: the first transfusion of human blood. Resuscitation. 2002;52:229-233. 3. Schmidt PJ, Ness PM. Hemotherapy: from bloodletting magic to transfusion medicine. Transfusion. 2006;46:166-168. 4. Rice CL, Moss GS. Blood and blood substitutes: current practice. Adv Surg. 1979;13:93-114. 5. Napolitano LM, Corwin HL. Efficacy of red blood cell transfusion in the critically ill. Crit Care Clin. 2004;20:255-268. 6. Blair SD, Janvrin SB, McCollum CN, Greenhalgh RM. Effect of early blood transfusion on gastrointestinal haemorrhage. Br J Surg. 1986;73: 783-785. 7. Hebert PC, Wells G, Marshall J, et al. Transfusion requirements in critical care. A pilot study. Canadian Critical Care Trials Group. JAMA. 1995;273:1439-1444. 8. Johnson RG, Thurer RL, Kruskall MS, et al. Comparison of two transfusion strategies after elective operations for myocardial revascularization. J Thorac Cardiovasc Surg. 1992;104:307-314. 9. Bracey AW, Radovancevic R, Riggs SA, et al. Lowering the hemoglobin threshold for transfusion in coronary artery bypass procedures: effect on patient outcome. Transfusion. 1999;39:1070-1077. 10. Bush RL, Pevec WC, Holcroft JW. A prospective, randomized trial limiting perioperative red blood cell transfusions in vascular patients. Am J Surg. 1997;174:143-148. 11. Carson JL, Terrin ML, Barton FB, et al. A pilot randomized trial comparing symptomatic vs. hemoglobin-level-driven red blood cell transfusions following hip fracture. Transfusion. 1998;38: 522-529. 12. Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med. 1999;340:409-417. 13. McIntyre L, Hebert PC, Wells G, et al. Is a restrictive transfusion strategy safe for resuscitated and critically ill trauma patients? J Trauma. 2004;57:563-568. 14. Practice guidelines for perioperative blood transfusion and adjuvant therapies. an updated report by the American Society of Anesthesiologists Task Force on Perioperative Blood Transfusion and Adjuvant Therapies. Anesthesiology. 2006;105:198-208. 15. Ferraris VA, Ferraris SP, Saha SP, et al. Perioperative blood transfusion and blood conservation in cardiac surgery: the Society of Thoracic Surgeons and The Society of Cardiovascular Anesthesiologists clinical practice guideline. Ann Thorac Surg. 2007;83: S27-S86. 16. Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med. 2010;152:101-113. 17. Ansari S, Szallasi A. Blood management by transfusion triggers: when less is more. Blood Transfus. 2012;10:28-33. 18. Jairath V, Kahan BC, Logan RF, Travis SP, Palmer KR, Murphy MF. Red blood cell transfusion practice in patients presenting with acute upper gastrointestinal bleeding: a survey of 815 UK clinicians. Transfusion. 2011;51:1940-1948. 19. Napolitano LM, Kurek S, Luchette FA, et al. Clinical practice guideline: red blood cell transfusion in adult trauma and critical care. Crit Care Med. 2009;37:3124-3157. 20. Carson JL, Grossman BJ, Kleinman S, et al. Red blood cell transfusion: a clinical practice guideline from the AABB*. Ann Intern Med. 2012;157:49-58. 21. Carson JL, Brooks MM, Abbott JD, et al. Liberal versus restrictive transfusion thresholds for patients with symptomatic coronary artery disease. Am Heart J. 2013;165:964-971.e1. 22. 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.
The American Journal of Medicine, Vol 127, No 2, February 2014 23. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368:11-21. 24. Higgins JPT, Altman DG, Sterne JAC, eds. Chapter 8: Assessing risk of bias in included studies. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. The Cochrane Collaboration, 2011. Available from www.cochranehandbook.org. 25. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959;22:719-748. 26. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177-188. 27. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ. 2003;326:219. 28. Lacroix J, Hebert PC, Hutchison JS, et al. Transfusion strategies for patients in pediatric intensive care units. N Engl J Med. 2007;356: 1609-1619. 29. Carson JL, Terrin ML, Noveck H, et al. Liberal or restrictive transfusion in high-risk patients after hip surgery. N Engl J Med. 2011;365: 2453-2462. 30. 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;108:1108-1111. 31. Foss NB, Kristensen MT, Jensen PS, Palm H, Krasheninnikoff M, Kehlet H. The effects of liberal versus restrictive transfusion thresholds on ambulation after hip fracture surgery. Transfusion. 2009;49:227-234. 32. Grover M, Talwalkar S, Casbard A, et al. Silent myocardial ischaemia and haemoglobin concentration: a randomized controlled trial of transfusion strategy in lower limb arthroplasty. Vox Sang. 2006;90: 105-112. 33. Hajjar LA, Vincent JL, Galas FR, et al. Transfusion requirements after cardiac surgery: the TRACS randomized controlled trial. JAMA. 2010;304:1559-1567. 34. Lotke PA, Barth P, Garino JP, Cook EF. Predonated autologous blood transfusions after total knee arthroplasty: immediate versus delayed administration. J Arthroplasty. 1999;14:647-650. 35. So-Osman C, Nelissen R, Te Slaa R, Coene L, Brand R, Brand A. A randomized comparison of transfusion triggers in elective orthopaedic surgery using leucocyte-depleted red blood cells. Vox Sang. 2010;98:56-64. 36. Webert KE, Cook RJ, Couban S, et al. A multicenter pilot-randomized controlled trial of the feasibility of an augmented red blood cell transfusion strategy for patients treated with induction chemotherapy for acute leukemia or stem cell transplantation. Transfusion. 2008;48:81-91. 37. Nielsen K, Dahl B, Johansson PI, Henneberg SW, Rasmussen LS. Intraoperative transfusion threshold and tissue oxygenation: a randomised trial. Transfus Med. 2012;22:418-425. 38. Shander A. Anemia in the critically ill. Crit Care Clin. 2004;20: 159-178. 39. Hebert PC, Van der Linden P, Biro G, Hu LQ. Physiologic aspects of anemia. Crit Care Clin. 2004;20:187-212. 40. Morisaki H, Sibbald WJ. Tissue oxygen delivery and the microcirculation. Crit Care Clin. 2004;20:213-223. 41. Monk TG. Acute normovolemic hemodilution. Anesthesiol Clin North America. 2005;23:271-281, vi. 42. Putney LJ. Bloodless cardiac surgery: not just possible, but preferable. Crit Care Nurs. Q 2007;30:263-270. 43. Shah DM, Prichard MN, Newell JC, Karmody AM, Scovill WA, Powers SR Jr. Increased cardiac output and oxygen transport after intraoperative isovolemic hemodilution. A study in patients with peripheral vascular disease. Arch Surg. 1980;115:597-600. 44. Spahn DR, Leone BJ, Reves JG, Pasch T. Cardiovascular and coronary physiology of acute isovolemic hemodilution: a review of nonoxygencarrying and oxygen-carrying solutions. Anesth Analg. 1994;78: 1000-1021. 45. Creteur J, Sun Q, Abid O, De Backer D, Van Der Linden P, Vincent JL. Normovolemic hemodilution improves oxygen extraction capabilities in endotoxic shock. J Appl Physiol. 2001;91:1701-1707.
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
46. Segal JB, Blasco-Colmenares E, Norris EJ, Guallar E. Preoperative acute normovolemic hemodilution: a meta-analysis. Transfusion. 2004;44:632-644. 47. Niinikoski J, Laaksonen V, Meretoja O, Jalonen J, Arstila M, Inberg MV. Oxygen transport and tissue oxygenation under moderate and extreme hemodilution during coronary bypass surgery. Ann Clin Res. 1981;13(Suppl 33):59-64. 48. Mathru M, Kleinman B, Blakeman B, Dries D, Zecca A, Rao T. Cardiovascular adjustments and gas exchange during extreme hemodilution in humans. Crit Care Med. 1991;19:700-704. 49. Fontana JL, Welborn L, Mongan PD, Sturm P, Martin G, Bunger R. Oxygen consumption and cardiovascular function in children during profound intraoperative normovolemic hemodilution. Anesth Analg. 1995;80:219-225. 50. Weiskopf RB, Viele MK, Feiner J, et al. Human cardiovascular and metabolic response to acute, severe isovolemic anemia. JAMA. 1998;279:217-221. 51. Spence RK, Costabile JP, Young GS, et al. Is hemoglobin level alone a reliable predictor of outcome in the severely anemic surgical patient? Am Surg. 1992;58:92-95. 52. van Woerkens EC, Trouwborst A, van Lanschot JJ. Profound hemodilution: what is the critical level of hemodilution at which oxygen delivery-dependent oxygen consumption starts in an anesthetized human? Anesth Analg. 1992;75:818-821. 53. Kulvatunyou N, Heard SO. Care of the injured Jehovah’s Witness patient: case report and review of the literature. J Clin Anesth. 2004;16: 548-553. 54. Shander A, Goodnough LT. Objectives and limitations of bloodless medical care. Curr Opin Hematol. 2006;13:462-470. 55. Nagarsheth NP, Sasan F. Bloodless surgery in gynecologic oncology. Mt Sinai J Med. 2009;76:589-597. 56. Kitchens CS. Are transfusions overrated? Surgical outcome of Jehovah’s Witnesses. Am J Med. 1993;94:117-119. 57. Hebert PC, McDonald BJ, Tinmouth A. Clinical consequences of anemia and red cell transfusion in the critically ill. Crit Care Clin. 2004;20:225-235. 58. Tinmouth A, Fergusson D, Yee IC, Hebert PC. Clinical consequences of red cell storage in the critically ill. Transfusion. 2006;46:2014-2027. 59. Reeves BC, Murphy GJ. Increased mortality, morbidity, and cost associated with red blood cell transfusion after cardiac surgery. Curr Opin Cardiol. 2008;23:607-612. 60. Hill GE, Frawley WH, Griffith KE, Forestner JE, Minei JP. Allogeneic blood transfusion increases the risk of postoperative bacterial infection: a meta-analysis. J Trauma. 2003;54:908-914.
131
61. Silverboard H, Aisiku I, Martin GS, Adams M, Rozycki G, Moss M. The role of acute blood transfusion in the development of acute respiratory distress syndrome in patients with severe trauma. J Trauma. 2005;59:717-723. 62. Amato A, Pescatori M. Perioperative blood transfusions for the recurrence of colorectal cancer. Cochrane Database Syst Rev. 2006: CD005033. 63. Silva Junior JM, Rezende E, Amendola CP, et al. Red blood cell transfusions worsen the outcomes even in critically ill patients undergoing a restrictive transfusion strategy. Sao Paulo Med J. 2012;130:77-83. 64. Purdy FR, Tweeddale MG, Merrick PM. Association of mortality with age of blood transfused in septic ICU patients. Can J Anaesth. 1997;44: 1256-1261. 65. Vamvakas EC, Carven JH. Transfusion and postoperative pneumonia in coronary artery bypass graft surgery: effect of the length of storage of transfused red cells. Transfusion. 1999;39:701-710. 66. Koch CG, Li L, Sessler DI, et al. Duration of red-cell storage and complications after cardiac surgery. N Engl J Med. 2008;358: 1229-1239. 67. Marik PE, Sibbald WJ. Effect of stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA. 1993;269:3024-3029. 68. Berezina TL, Zaets SB, Morgan C, et al. Influence of storage on red blood cell rheological properties. J Surg Res. 2002;102:6-12. 69. McMahon TJ, Ahearn GS, Moya MP, et al. A nitric oxide processing defect of red blood cells created by hypoxia: deficiency of S-nitrosohemoglobin in pulmonary hypertension. Proc Natl Acad Sci U S A. 2005;102:14801-14806. 70. Bennett-Guerrero E, Veldman TH, Doctor A, et al. Evolution of adverse changes in stored RBCs. Proc Natl Acad Sci U S A. 2007;104: 17063-17068. 71. Reynolds JD, Ahearn GS, Angelo M, Zhang J, Cobb F, Stamler JS. S-nitrosohemoglobin deficiency: a mechanism for loss of physiological activity in banked blood. Proc Natl Acad Sci U S A. 2007;104: 17058-17062. 72. Chatterjee S, Wetterslev J, Sharma A, Lichstein E, Mukherjee D. Association of blood transfusion with increased mortality in myocardial infarction: a meta-analysis and diversity-adjusted study sequential analysis. Arch Intern Med. 2013;173(2):132-139. 73. Hebert PC, Yetisir E, Martin C, et al. Is a low transfusion threshold safe in critically ill patients with cardiovascular diseases? Crit Care Med. 2001;29:227-234. 74. Rao SV, Jollis JG, Harrington RA, et al. Relationship of blood transfusion and clinical outcomes in patients with acute coronary syndromes. JAMA. 2004;292:1555-1562.
131.e1
APPENDIX SUPPLEMENTAL DATA Study Protocol: Types of studies: We included randomized controlled trials with a concurrent control group. Trials were included if the comparison groups were assigned on the basis of a clear transfusion trigger or threshold of 7 g/dL, described as a hemoglobin or hematocrit level. Patients in the control group were required to receive a transfusion of red blood cells at higher hemoglobin or hematocrit thresholds than the intervention group. Types of participants: We included trials of patients with surgical or medical conditions, involving adults or children, including neonates. Types of interventions: The intervention considered was the use of red blood cell transfusion thresholds (“triggers”) as a means of providing insights for rational transfusion practices. Outcomes assessed: In-hospital mortality, 30-day mortality, total mortality over the period of follow-up for individual studies, acute coronary syndrome, pulmonary edema, bacterial infections, rebleeding, number of patients receiving any blood transfusion, and units of blood transfused per patient.
The American Journal of Medicine, Vol 127, No 2, February 2014 9. (transfus* adj5 (polic*or practic* or protocol* or trigger* or threshold* or indicator* or strateg* or criteri* or standard*)).mp. 10. ((Red blood cell* or RBC) adj5 (polic*or practic* or protocol* or trigger* or threshold*or indicator* or strateg* or criteri* or standard*)).mp. 11. ((H?emoglobin or h?emocrit or HB or HCT) adj5 (polic*or practic* or protocol* or trigger* or threshold* or indicator* or strateg* or criteri* or standard*)).mp. 12. (transfus* adj5 (restrict* or liberal*)).mp. 13. ((blood or transfusion*) adj3 (management or program*)).mp. 14. 8 or 9 or 10 or 11 or 12 or 13 15. randomi?ed.ab,ti. 16. randomized controlled trial.pt. 17. controlled clinical trial.pt. 18. placebo.ab. 19. clinical trials as topic.sh. 20. randomly.ab. 21. trial.ti. 22. 15 or 16 or 17 or 18 or 19 or 20 or 21 23. (animals not (humans and animals)).sh. 24. 22 not 23 25. 24 and 14.
Search methods for identification of studies: We did not restrict our search for trials by date, language, or publication status. A comprehensive search of the MEDLINE database from 1966 to April 2013 was conducted using the following MeSH terms and strings:
Searching other resources: We contacted authors of published studies for clarification of trial methodology and data. This was provided in 1 instance.21 We searched the reference lists of relevant reviews and published articles, as well as the reference lists of all included trials for further studies.
1. *Blood Transfusion/ 2. ((Red blood cell* or RBC) adj3 (therap* or transfus*)).mp. 3. 1 or 2 4. exp Reference Standards/ 5. standards.fs. 6. methods.fs. 7. 4 or 5 or 6 8. 3 and 7
Assessment of risk of bias in included studies: We completed a “risk of bias” table for each study, incorporating a description of the study’s performance against each of the above domains as follows: “low,” “unclear” (indicating unclear or unknown risk of bias), and “high” risk of bias. One investigator (SC) studied the following domains: sequence generation; allocation concealment; blinding; incomplete outcome data; selective outcome reporting; other potential sources of bias.
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
Appendix Figure 1 confidence interval.
131.e2
Restrictive transfusion strategy and other clinical events. CI ¼
Appendix Table 1 Patient-Level Characteristics of Randomized Controlled Trials Evaluating a Liberal Transfusion Strategy (Hemoglobin Trigger of <7.0 g/dL) Compared with a More Liberal Strategy Trial Trial Characteristics
Hebert et al12
Lacroix et al28
Villanueva et al23*
Setting
Adult Intensive Care Unit
Pediatric Intensive Care Unit
Severe Acute Upper Gastrointestinal Bleeding
Transfusion strategy
Restrictive Arm
Liberal Arm
Restrictive Arm
Liberal Arm
Restrictive Arm
Liberal Arm
Transfusion trigger (g/dL) Total (N) Mean age (y) Men (%) Baseline hemoglobin (g/dL) Mean units of red-cell transfusion Mean length of storage of red cells (d [SD]) Cardiovascular disease (N [%]) Infection (N [%])
7.0 418 57.1 64 8.2 2.5 NR
10.0 420 58.1 61 8.2 2.3 NR
7.0 320 2.98 59 8.0 0.9 14.9 [11.8]
9.5 317 3.3 60 8.0 1.7 15.2 [10.6]
7.0 444 64 71 9.55 1.5 15.8 [8]
9.0 445 66 65 9.44 3.7 15.8 [8]
76 [18] 114 [27]
94 [22] 108 [26]
76 [24] 157 [49]
75 [24] 155 [49]
NR 37 [8.3]
NR 44 [9.8]
NR ¼ not reported (in primary publication, as referenced); SD ¼ standard deviation. *Additional data provided by corresponding author on request.
131.e3 Appendix Table 2
Trial
The American Journal of Medicine, Vol 127, No 2, February 2014 Risk of Bias Assessments for Included Trials Blinding of Participants and Random Sequence Allocation Researchers Concealment Generation (Selection Bias) (Performance Bias) (Selection Bias)
Hebert et al12 Low Lacroix et al28 Low Villanueva et al23 Low
Low Low Low
Low Low Low
Blinding of Selective Incomplete Outcome Other Outcome Data Reporting Assessment (Detection Bias) (Attrition Bias) (Reporting Bias) Bias Low Low Low
Low Low Low
Low Low Low
Low Low Low
Impact of More Restrictive Blood Transfusion Strategies on Clinical Outcomes: A Meta-analysis and Systematic Review Shelley R. Salpeter, MD,a Jacob S. Buckley,b Saurav Chatterjee, MDc a Stanford University School of Medicine, Stanford, Calif; bBrown University, Providence, RI; cSt Luke’s e Roosevelt Hospital Center, New York, NY.
ABSTRACT BACKGROUND: There is accumulating evidence that restricting blood transfusions improves outcomes, with newer trials showing greater benefit from more restrictive strategies. We systematically evaluated the impact of various transfusion triggers on clinical outcomes. METHODS: The MEDLINE database was searched from 1966 to April 2013 to find randomized trials evaluating a restrictive hemoglobin transfusion trigger of <7 g/dL, compared with a more liberal trigger. Two investigators independently extracted data from the trials. Outcomes evaluated included mortality, acute coronary syndrome, pulmonary edema, infections, rebleeding, number of patients transfused, and units of blood transfused per patient. Extracted data also included information on study setting, design, participant characteristics, and risk for bias of the included trials. A secondary analysis evaluated trials using less restrictive transfusion triggers, and a systematic review of observational studies evaluated more restrictive triggers. RESULTS: In the primary analysis, pooled results from 3 trials with 2364 participants showed that a restrictive hemoglobin transfusion trigger of <7 g/dL resulted in reduced in-hospital mortality (risk ratio [RR], 0.74; confidence interval [CI], 0.60-0.92), total mortality (RR, 0.80; CI, 0.65-0.98), rebleeding (RR, 0.64; CI, 0.45-0.90), acute coronary syndrome (RR, 0.44; CI, 0.22-0.89), pulmonary edema (RR, 0.48; CI, 0.33-0.72), and bacterial infections (RR, 0.86; CI, 0.73-1.00), compared with a more liberal strategy. The number needed to treat with a restrictive strategy to prevent 1 death was 33. Pooled data from randomized trials with less restrictive transfusion strategies showed no significant effect on outcomes. CONCLUSIONS: In patients with critical illness or bleed, restricting blood transfusions by using a hemoglobin trigger of <7 g/dL significantly reduces cardiac events, rebleeding, bacterial infections, and total mortality. A less restrictive transfusion strategy was not effective. Ó 2014 Elsevier Inc. All rights reserved. The American Journal of Medicine (2014) 127, 124-131 KEYWORDS: Clinical outcomes; Meta-analysis; Mortality; Systematic review; Transfusion SEE RELATED EDITORIAL p. 103
Red blood cell transfusions have been the standard of care for treating anemia for more than 100 years now, with little evidence that they improve clinical outcomes.1-3 By the
Funding: None. Conflict of Interest: None. Authorship: All authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to Shelley Salpeter, MD, 34 North San Mateo Drive, Suite 1, San Mateo, CA 94401. E-mail address: [email protected] 0002-9343/$ -see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjmed.2013.09.017
early 1900s, blood transfusion was considered to be “a procedure of such simple and harmless character” that no clinical indication was needed, “the mere possibility of benefitting a condition by the addition of blood being considered sufficient warrant.”1 The practice was based on the assumption that anemia is tolerated poorly and that red blood cell transfusions improve outcomes.1,4,5 Researchers did not begin to question the evidence behind the practice until the 1980s and 1990s, when the first randomized trials were performed.6-12 By that time, the practice of blood transfusion was so ingrained in our medical framework that
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
125
the approach has been to march slowly down on the transprotocol shown in Appendix Tables 1 and 2, online). We fusion trigger instead of addressing whether transfusions included trials of adults or children, including neonates, are beneficial at all. involving surgical or medical conditions. Trials that used a The standard transfusion trigger for many years had restrictive transfusion trigger more than 7 g/dL were evalbeen a hemoglobin of 10 g/dL or even higher.6,9,12,13 This uated separately as a “less restrictive” strategy. Additional arbitrary trigger has been lowered gradually to a hemoglobin searches of related articles were done to perform a syslevel of 6 to 8 g/dL because studies tematic review of the impact of showed that blood transfusions are various transfusion strategies. CLINICAL SIGNIFICANCE associated with worse outcomes in patients with anemia due to illness Pooled randomized trial data show that Data Extraction and Quality or bleeding, compared with simple blood transfusions increase in-hospital Assessment supportive measures such as hymortality, total mortality, rebleeding, 6,9,12-20 Two investigators (SS, JB) exdration. However, a liberal acute coronary syndrome, pulmonary tracted data from the trials, transfusion practice is still comedema, and bacterial infections. reconciling differences by conmon, especially for those with sensus. In addition, selected incoronary artery disease who are When a restrictive hemoglobin transvestigators were contacted for thought to benefit more from blood fusion trigger of <7 g/dL is used in additional information. Clinical transfusions.21 There have been patients with critical illness or bleed, the outcomes evaluated included inmany challenges inherent in the number needed to treat to prevent 1 hospital mortality, 30-day mortalstudy of our transfusion practices, death is 33. ity, total mortality, acute coronary such as the broad patient base syndrome, pulmonary edema, Observational data indicate that hemoincluded in the analyses, the mulbacterial infections, rebleeding, tiple indications for blood transglobin levels of 5 to 6 g/dL are well number of patients receiving any fusions, and the confounding by tolerated in normovolemic patients blood transfusion, and units of indication in observational studies. without affecting oxygen delivery. blood transfused per patient. We have found no randomized Extracted data also included inclinical trials comparing transformation on study setting, design, participant characterfusion with no transfusion. Instead, the available trials istics, and risk for bias for the included trials (detailed have compared more or less restrictive transfusion strastudy protocol shown in Appendix, online).24 tegies using different transfusion triggers. A previous meta-analysis pooled data from randomized trials that evaluated restrictive hemoglobin transfusion triggers Data Synthesis and Analysis ranging from 7 to 10 g/dL and found that restricting The results were reported as a risk ratio (RR) and risk diftransfusions significantly reduced in-hospital mortality but ference for dichotomous outcomes, for the restrictive strathad no effect on other clinical outcomes.22 We have egy compared with the liberal strategy, with the confidence chosen a different approach to evaluate the available evinterval (CI) set at 95% significance. For the amount of idence. We now update the meta-analysis through April blood transfused per patient, the results were reported as a 2013 to include a subsequent trial23 and restrict the primean difference, with 95% CIs for the restrictive compared mary analysis to those trials with a transfusion trigger of with the liberal strategy. To test for inter-study heteroge<7 g/dL. Trials that evaluated less restrictive strategies neity, the chi-square value was calculated; statistical sigwere evaluated in a separate analysis. We also provide a nificance was indicated by P < .1. The fixed-effects method systematic review of observational studies that evaluated was chosen to report the results because minimal heteroclinical outcomes related to other more restrictive transgeneity was seen in most of the analyses.25 When heterofusion strategies. geneity was noted, the random-effects method was used.26
MATERIALS AND METHODS Data Sources and Study Selection We conducted a comprehensive search of the MEDLINE database from 1966 to April 2013 using the terms blood transfusion and clinical trial, and scanned selected journals and references of identified articles. Studies of any language were included in the primary analysis if they were randomized controlled trials that evaluated a restrictive blood transfusion strategy using a transfusion trigger of <7 g/dL, compared with a more liberal strategy (detailed study
In a secondary analysis, the pooled results from trials using a less restrictive strategy were evaluated and compared with the trials in the primary analysis using the test for interaction.27 The analyses were performed using Review Manager, Version 5.2, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2012.
Role of Funding Source The investigators received no funding for the study. No sponsor had a role in any aspect of the study, including its design and conduct, data extraction and analysis, and preparation of the manuscript.
126
The American Journal of Medicine, Vol 127, No 2, February 2014
-
Potentially relevant studies identified and screened for retrieval (n = 4500)
Studies excluded: Not trials of transfusion triggers (n = 4369) Studies used for systematic review
Trials of transfusion triggers In anemia (n = 32)
Trials excluded: Not randomized (n = 2)
Randomized trials of transfusion triggers in anemia (n = 30)
Trials excluded from primary analysis: Used less-restrictive trigger (n = 19) (16 trials used in secondary analysis) Did not provide clear trigger (n = 4) Duplicate data from other trials (n = 4)
Primary analysis: Randomized controlled trials of a restrictive transfusion strategy using a hemoglobin trigger of < 7 g/dL, compared with a more liberal strategy (n = 3) Figure 1
Flow chart of trials search for meta-analysis.
RESULTS Primary Meta-analysis Search Results. The search identified approximately 4500 studies, of which 32 were potentially relevant trials evaluating transfusion triggers (Figure 1). Of these, 3 trials met inclusion criteria for the primary analysis.12,23,28 One study provided unpublished information.23 Studies were excluded for the following reasons: Two were not randomized, 19 used a less-restrictive hemoglobin transfusion trigger of >7 g/dL, 4 did not provide a clear transfusion trigger, and 4 provided duplicate data on participants included in another trial. Of the 19 trials evaluating a
less-restrictive strategy, 16 provided data on clinical outcomes and were evaluated separately.6-11,21,29-37 Trial Characteristics. The primary analysis included 3 trials, with a total of 2364 participants followed for mean trial duration of 45 days. The characteristics of the included trials, including their risk of bias, are shown in Appendix Figure 1 (online). The mean study size was 788 participants (range, 637-889), with a mean participant age of 45.7 years (standard deviation, 16 years). Transfusion strategies were evaluated in the setting of adult critical care,12 pediatric critical care,28 and acute upper gastrointestinal bleeding.23
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
127
Table 1 Outcomes for Restrictive Transfusion Strategy, with Hemoglobin Transfusion Trigger <7 g/dL, Compared with a More Liberal Strategy Outcome Mortality Hospital mortality 30-d mortality Total mortality Cardiac events Acute coronary syndrome Pulmonary edema Other outcomes Rebleeding Bacterial infections Blood transfusions Patients exposed to blood Units transfused per patient
Trials
Patients
RR or MD
RD
No. Needed to Treat
2 3 3
1727 2364 2364
RR, 0.74 [CI, 0.60-0.92] RR, 0.81 [CI, 0.61-0.96] RR, 0.80 [CI, 0.65-0.98]
RD, 0.04 [CI, 0.04 to 0.00] RD, 0.02 [CI, 0.04 to 0.00] RD, 0.03 [CI, 0.05 to 0.00]
25 50 33
2 3
1727 2364
RR, 0.44 [CI, 0.22-0.89] RR, 0.48 [CI, 0.33-0.73]
RD, 0.02 [CI, 0.03 to 0.00] RD, 0.03CI, 0.05 to 0.01
50 33
1 3
889 2364
RR, 0.64 [CI, 0.45-0.90] RR, 0.86 [CI, 0.73-1.00]
RD, 0.06 [CI, 0.10 to 0.01] RD, 0.03 [CI, 0.06 to 0.00]
17 33
3 3
2364 2364
RR, 0.57 [CI, 0.46-0.70] MD 1.98 [CI, 3.22 to 0.74]
RD, 0.41 [CI, 0.52 to 0.29]
2
CI ¼ confidence interval; MD ¼ mean difference; RD ¼ risk difference; RR ¼ risk ratio.
Data Synthesis. The restrictive transfusion strategy was associated with a statistically significant reduction of inhospital mortality (RR, 0.74; CI, 0.60-0.92), 30-day mortality (RR, 0.77; CI, 0.61-0.96), and total mortality (RR, 0.80; CI, 0.65-0.98), compared with the liberal strategy (Table 1, Figure 2). The risk difference for total mortality was 0.03 for the restrictive strategy compared with the liberal strategy (mean duration of included trials, 45 days), with a number needed to treat of 33 to save 1 life. In addition, the restrictive strategy resulted in a reduced incidence of acute coronary syndrome (RR, 0.44; CI, 0.22-0.89), pulmonary edema (RR, 0.48; CI, 0.330.72), rebleeding (RR, 0.64; CI, 0.45-0.90), and bacterial infections (RR, 0.86; CI, 0.73-1.00), compared with the liberal strategy (Table 1, Figure 3, Appendix, online).
Figure 2
The results for bacterial infections were of nominal significance. In pooled trial data, 55% of the participants in the restrictive group received a blood transfusion, compared with 94% in the liberal group (RR, 0.57; CI, 0.46-0.70 and risk difference, 0.41; CI, 0.52 to 0.29). The restrictive strategy resulted in a significant reduction in the mean number of units transfused, with a mean difference of 1.98 units (CI, 3.22 to 0.74 units) per patient. No evidence for inter-study heterogeneity was found in any of the clinical outcomes analyzed (P > .25, for chisquare test). There was evidence for significant heterogeneity in the analyses of exposure to blood transfusions and the mean number of units transfused (P < .00001).
Restrictive transfusion strategy and mortality. CI ¼ confidence interval.
128
The American Journal of Medicine, Vol 127, No 2, February 2014
Figure 3
Restrictive transfusion strategy and cardiac events. CI ¼ confidence interval.
Secondary Meta-analysis
Systematic Review
Randomized Trials Using Less-Restrictive Transfusion Triggers. Nineteen of the trials excluded from the primary analysis evaluated hemoglobin triggers of 7.5 to 10 g/dL in the restrictive strategy group, and 16 of those provided data on clinical outcomes.6-11,21,29-37 When the data from these trials were pooled, with a total of 4572 participants, the restrictive strategy significantly reduced the exposure to blood transfusions (RR, 0.60; CI, 0.48-0.75) and amount of blood transfused (mean difference, 0.80 units; CI, 1.24 to 0.37 units), compared with a more liberal strategy, but this had no significant effect on in-hospital mortality (RR, 0.65; CI, 0.37-1.15), total mortality (RR, 1.03; CI, 0.811.31), acute coronary syndrome (RR, 1.46; CI, 0.96-2.20), pulmonary edema (RR, 1.02; CI, 0.67-1.56), rebleeding (RR, 0.68; CI, 0.34-1.34), or bacterial infections (RR, 0.80; CI, 0.64-1.02). When the pooled results for the trials from the primary analysis (hemoglobin trigger <7 g/dL) were compared with the results from the less restrictive trials (hemoglobin trigger 7.5-10 g/dL) using the test for interaction, the more restrictive strategy was associated with a greater reduction in acute coronary syndrome (P ¼ .004), pulmonary edema (P ¼ .002), and amount of blood transfused (P ¼ .01), compared with the less restrictive strategy. When all 19 trials from the primary and secondary analyses were pooled together, with a total of 6936 participants, the restrictive strategy was still associated with a significant reduction in hospital mortality (RR, 0.73; CI, 0.59-0.89), 30-day mortality (RR, 0.83; CI, 0.69-0.99), pulmonary edema (RR, 0.68; CI, 0.51-0.90), bacterial infections (RR, 0.84; CI, 0.73-0.95), and rebleeding (RR, 0.64; CI, 0.47-0.88).
Observational Studies Evaluating More Restrictive Transfusion Strategies. Recent recommendations to further restrict blood transfusions have been based on studies showing the tolerability and safety of anemia under controlled normovolemic conditions, which involve the administration of fluids, oxygen, and beta-adrenergic blockers.14,15,38-45 A meta-analysis of randomized trials in the perioperative setting showed that normovolemic hemodilution, in which blood is removed and replaced with crystalloid or colloid solution, resulted in significantly less total blood loss and allogenic blood transfusions, compared with standard care.46 A systematic review found consistent evidence that normovolemic anemia is associated with a reduction in systemic vascular resistance and an increase in cardiac output, coronary and cerebral blood flow, and synthesis of 2,3-diphosphoglycerate in red blood cells, resulting in maintenance of oxygen delivery and extraction.39 Observational studies have shown that hemoglobin levels of 5 to 6 g/dL in the setting of surgery, critical illness, and acute bleeds generally are well tolerated when standard supportive measures are given, without evidence of cardiac ischemia or decrease in oxygen extraction until the hemoglobin decreases to less than 3 to 4 g/dL.47-56
DISCUSSION Pooled data from randomized controlled trials show that restricting blood transfusions to patients whose hemoglobin decreases to less than 7 g/dL results in a significant reduction in total mortality, acute coronary syndrome, pulmonary edema, rebleeding, and bacterial infection, compared with a
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
more liberal transfusion strategy. The number needed to treat to save 1 life was 33. This strategy resulted in a 40% reduction in the number of patients receiving a blood transfusion, with an average of 2 units less per person; however, more than one half of patients still received a transfusion. In an analysis of trials that used a less restrictive strategy, with hemoglobin triggers of 7.5 to 9 g/dL, no significant reduction in morbidity or mortality was seen. With the available evidence from observational studies, an even more restrictive transfusion strategy using a hemoglobin trigger of <6 g/dL has been recommended in some settings.14,15 Observational studies have consistently shown that transfusions are associated with an increased risk for adverse events after controlling for potential confounding variables, even when using a restrictive transfusion strategy.56-63 The increased risk seems to be directly proportional to the amount of blood transfused and the length of storage of the transfused red blood cells, and may be due to an inflammatory response to the transfused blood product.58,64-66 Systematic reviews also have evaluated the effect of blood transfusions on oxygen transport variables in anemic patients in the setting of surgery, critical illness, and bleed, and found no significant improvement in oxygen delivery or use compared with supportive care, despite an increase in oxygen content.5,57,58 This inability to improve oxygen uptake in vital organs is due to the hemodynamic response to increased blood viscosity and the loss of red cell function during preservation and storage.5,57,58,67-71 Conversely, there is evidence that normovolemic anemia with hemoglobin levels of 5 to 6 g/dL is well tolerated in cardiovascular or critical illness and may have beneficial hemodynamic effects.39,42,47,51 It has been the traditional teaching that patients with cardiac ischemia should have a more liberal transfusion strategy to maintain oxygenation, but pooled observational studies on transfusion in myocardial infarction found that the rates of subsequent myocardial infarction and all-cause mortality were significantly higher in patients receiving blood transfusions compared with standard supportive measures, after adjustment for possible confounding variables.72 Multivariate meta-regression analysis of the pooled data revealed that the increased risk was independent of the baseline or nadir hemoglobin level.72 Subgroup analysis of a large critical care trial included in the primary analysis found that in patients with coronary ischemia, a transfusion trigger of <7 g/dL may be associated with improved clinical outcomes.12,73 Two small trials included in the secondary analysis evaluated a restrictive trigger of <8 g/dL in patients with symptomatic coronary artery disease; pooled results from these 2 trials showed no significant effect on cardiac events or mortality using this less restrictive strategy.21,30 This meta-analysis found that restricting transfusions using a hemoglobin trigger of <7 g/dL reduced mortality in critical illness or bleed, with a number needed to treat of 33 to save 1 life. With millions of blood transfusions given yearly over the past century, it would be hard to calculate how many deaths may have been caused by transfusions.
129
The main goal of blood transfusions is to increase oxygencarrying capacity, but despite increasing oxygen content, oxygen delivery is not increased.5,57,58 Another indication for transfusions has been to stabilize bleeding patients, but in fact transfusions significantly increase the risk for rebleeding.6,23 So far, there is little trial evidence that blood transfusions significantly improve oxygen delivery or clinical outcomes in any setting or with any nadir hemoglobin level.5,57,58,74
Study Limitations This meta-analysis and systematic review has several limitations. The trials in the primary meta-analysis studied disparate populations, including both pediatric and adult patients, and addressed different indications for transfusion, such as critical illness and gastrointestinal bleed. This has the potential for introducing bias in the analysis, despite the fact that we found no evidence for inter-study heterogeneity in the results. Another potential bias is that there was incomplete blinding of the participants in the individual trials because of the nature of the interventions. However, the hard clinical outcomes studied were unlikely to have been influenced by this. Another limitation of the study is that other patient populations, such as those with less lifethreatening illnesses, were not included in the primary analysis. The objective of this study was to put together the available evidence for practicing clinicians to make sense of it all, but we are left with many unanswered questions. Unfortunately, more than one half of the patients in the restrictive group still received a blood transfusion, so we cannot directly assess the effect of transfusion versus no transfusion. At present, there are no randomized trials evaluating lower transfusion triggers, such as a hemoglobin level of 6 g/dL, which is what some of the newer guidelines recommend using on the basis of observational studies.14,15 In addition, there are no randomized trials evaluating the lower transfusion triggers in acute coronary syndrome, which at present is considered an indication for the use of a more liberal trigger. Finally, there is little information on clinical situations or nadir hemoglobin levels for which transfusions are known to improve oxygen delivery and mortality.
CONCLUSIONS We have performed an updated meta-analysis of randomized trials that shows that a restrictive transfusion strategy using a hemoglobin transfusion trigger of <7 g/dL results in a significant reduction in acute coronary syndrome, pulmonary edema, rebleeding, infections, and total mortality, compared with a more liberal strategy. At present, there is no randomized trial evidence that blood transfusions improve oxygen delivery or clinical outcomes in any setting. More studies are needed to help guide clinicians in finding optimal treatment threshold and options in the setting of anemia and bleed.
130
References 1. Bernheim BM. Blood Transfusion, Hemorrhage and the Anaemias. Philadelphia and London: JB Lippincott Company; 1917. 2. Baskett TF. James Blundell: the first transfusion of human blood. Resuscitation. 2002;52:229-233. 3. Schmidt PJ, Ness PM. Hemotherapy: from bloodletting magic to transfusion medicine. Transfusion. 2006;46:166-168. 4. Rice CL, Moss GS. Blood and blood substitutes: current practice. Adv Surg. 1979;13:93-114. 5. Napolitano LM, Corwin HL. Efficacy of red blood cell transfusion in the critically ill. Crit Care Clin. 2004;20:255-268. 6. Blair SD, Janvrin SB, McCollum CN, Greenhalgh RM. Effect of early blood transfusion on gastrointestinal haemorrhage. Br J Surg. 1986;73: 783-785. 7. Hebert PC, Wells G, Marshall J, et al. Transfusion requirements in critical care. A pilot study. Canadian Critical Care Trials Group. JAMA. 1995;273:1439-1444. 8. Johnson RG, Thurer RL, Kruskall MS, et al. Comparison of two transfusion strategies after elective operations for myocardial revascularization. J Thorac Cardiovasc Surg. 1992;104:307-314. 9. Bracey AW, Radovancevic R, Riggs SA, et al. Lowering the hemoglobin threshold for transfusion in coronary artery bypass procedures: effect on patient outcome. Transfusion. 1999;39:1070-1077. 10. Bush RL, Pevec WC, Holcroft JW. A prospective, randomized trial limiting perioperative red blood cell transfusions in vascular patients. Am J Surg. 1997;174:143-148. 11. Carson JL, Terrin ML, Barton FB, et al. A pilot randomized trial comparing symptomatic vs. hemoglobin-level-driven red blood cell transfusions following hip fracture. Transfusion. 1998;38: 522-529. 12. Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med. 1999;340:409-417. 13. McIntyre L, Hebert PC, Wells G, et al. Is a restrictive transfusion strategy safe for resuscitated and critically ill trauma patients? J Trauma. 2004;57:563-568. 14. Practice guidelines for perioperative blood transfusion and adjuvant therapies. an updated report by the American Society of Anesthesiologists Task Force on Perioperative Blood Transfusion and Adjuvant Therapies. Anesthesiology. 2006;105:198-208. 15. Ferraris VA, Ferraris SP, Saha SP, et al. Perioperative blood transfusion and blood conservation in cardiac surgery: the Society of Thoracic Surgeons and The Society of Cardiovascular Anesthesiologists clinical practice guideline. Ann Thorac Surg. 2007;83: S27-S86. 16. Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med. 2010;152:101-113. 17. Ansari S, Szallasi A. Blood management by transfusion triggers: when less is more. Blood Transfus. 2012;10:28-33. 18. Jairath V, Kahan BC, Logan RF, Travis SP, Palmer KR, Murphy MF. Red blood cell transfusion practice in patients presenting with acute upper gastrointestinal bleeding: a survey of 815 UK clinicians. Transfusion. 2011;51:1940-1948. 19. Napolitano LM, Kurek S, Luchette FA, et al. Clinical practice guideline: red blood cell transfusion in adult trauma and critical care. Crit Care Med. 2009;37:3124-3157. 20. Carson JL, Grossman BJ, Kleinman S, et al. Red blood cell transfusion: a clinical practice guideline from the AABB*. Ann Intern Med. 2012;157:49-58. 21. Carson JL, Brooks MM, Abbott JD, et al. Liberal versus restrictive transfusion thresholds for patients with symptomatic coronary artery disease. Am Heart J. 2013;165:964-971.e1. 22. 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.
The American Journal of Medicine, Vol 127, No 2, February 2014 23. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368:11-21. 24. Higgins JPT, Altman DG, Sterne JAC, eds. Chapter 8: Assessing risk of bias in included studies. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. The Cochrane Collaboration, 2011. Available from www.cochranehandbook.org. 25. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959;22:719-748. 26. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177-188. 27. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ. 2003;326:219. 28. Lacroix J, Hebert PC, Hutchison JS, et al. Transfusion strategies for patients in pediatric intensive care units. N Engl J Med. 2007;356: 1609-1619. 29. Carson JL, Terrin ML, Noveck H, et al. Liberal or restrictive transfusion in high-risk patients after hip surgery. N Engl J Med. 2011;365: 2453-2462. 30. 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;108:1108-1111. 31. Foss NB, Kristensen MT, Jensen PS, Palm H, Krasheninnikoff M, Kehlet H. The effects of liberal versus restrictive transfusion thresholds on ambulation after hip fracture surgery. Transfusion. 2009;49:227-234. 32. Grover M, Talwalkar S, Casbard A, et al. Silent myocardial ischaemia and haemoglobin concentration: a randomized controlled trial of transfusion strategy in lower limb arthroplasty. Vox Sang. 2006;90: 105-112. 33. Hajjar LA, Vincent JL, Galas FR, et al. Transfusion requirements after cardiac surgery: the TRACS randomized controlled trial. JAMA. 2010;304:1559-1567. 34. Lotke PA, Barth P, Garino JP, Cook EF. Predonated autologous blood transfusions after total knee arthroplasty: immediate versus delayed administration. J Arthroplasty. 1999;14:647-650. 35. So-Osman C, Nelissen R, Te Slaa R, Coene L, Brand R, Brand A. A randomized comparison of transfusion triggers in elective orthopaedic surgery using leucocyte-depleted red blood cells. Vox Sang. 2010;98:56-64. 36. Webert KE, Cook RJ, Couban S, et al. A multicenter pilot-randomized controlled trial of the feasibility of an augmented red blood cell transfusion strategy for patients treated with induction chemotherapy for acute leukemia or stem cell transplantation. Transfusion. 2008;48:81-91. 37. Nielsen K, Dahl B, Johansson PI, Henneberg SW, Rasmussen LS. Intraoperative transfusion threshold and tissue oxygenation: a randomised trial. Transfus Med. 2012;22:418-425. 38. Shander A. Anemia in the critically ill. Crit Care Clin. 2004;20: 159-178. 39. Hebert PC, Van der Linden P, Biro G, Hu LQ. Physiologic aspects of anemia. Crit Care Clin. 2004;20:187-212. 40. Morisaki H, Sibbald WJ. Tissue oxygen delivery and the microcirculation. Crit Care Clin. 2004;20:213-223. 41. Monk TG. Acute normovolemic hemodilution. Anesthesiol Clin North America. 2005;23:271-281, vi. 42. Putney LJ. Bloodless cardiac surgery: not just possible, but preferable. Crit Care Nurs. Q 2007;30:263-270. 43. Shah DM, Prichard MN, Newell JC, Karmody AM, Scovill WA, Powers SR Jr. Increased cardiac output and oxygen transport after intraoperative isovolemic hemodilution. A study in patients with peripheral vascular disease. Arch Surg. 1980;115:597-600. 44. Spahn DR, Leone BJ, Reves JG, Pasch T. Cardiovascular and coronary physiology of acute isovolemic hemodilution: a review of nonoxygencarrying and oxygen-carrying solutions. Anesth Analg. 1994;78: 1000-1021. 45. Creteur J, Sun Q, Abid O, De Backer D, Van Der Linden P, Vincent JL. Normovolemic hemodilution improves oxygen extraction capabilities in endotoxic shock. J Appl Physiol. 2001;91:1701-1707.
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
46. Segal JB, Blasco-Colmenares E, Norris EJ, Guallar E. Preoperative acute normovolemic hemodilution: a meta-analysis. Transfusion. 2004;44:632-644. 47. Niinikoski J, Laaksonen V, Meretoja O, Jalonen J, Arstila M, Inberg MV. Oxygen transport and tissue oxygenation under moderate and extreme hemodilution during coronary bypass surgery. Ann Clin Res. 1981;13(Suppl 33):59-64. 48. Mathru M, Kleinman B, Blakeman B, Dries D, Zecca A, Rao T. Cardiovascular adjustments and gas exchange during extreme hemodilution in humans. Crit Care Med. 1991;19:700-704. 49. Fontana JL, Welborn L, Mongan PD, Sturm P, Martin G, Bunger R. Oxygen consumption and cardiovascular function in children during profound intraoperative normovolemic hemodilution. Anesth Analg. 1995;80:219-225. 50. Weiskopf RB, Viele MK, Feiner J, et al. Human cardiovascular and metabolic response to acute, severe isovolemic anemia. JAMA. 1998;279:217-221. 51. Spence RK, Costabile JP, Young GS, et al. Is hemoglobin level alone a reliable predictor of outcome in the severely anemic surgical patient? Am Surg. 1992;58:92-95. 52. van Woerkens EC, Trouwborst A, van Lanschot JJ. Profound hemodilution: what is the critical level of hemodilution at which oxygen delivery-dependent oxygen consumption starts in an anesthetized human? Anesth Analg. 1992;75:818-821. 53. Kulvatunyou N, Heard SO. Care of the injured Jehovah’s Witness patient: case report and review of the literature. J Clin Anesth. 2004;16: 548-553. 54. Shander A, Goodnough LT. Objectives and limitations of bloodless medical care. Curr Opin Hematol. 2006;13:462-470. 55. Nagarsheth NP, Sasan F. Bloodless surgery in gynecologic oncology. Mt Sinai J Med. 2009;76:589-597. 56. Kitchens CS. Are transfusions overrated? Surgical outcome of Jehovah’s Witnesses. Am J Med. 1993;94:117-119. 57. Hebert PC, McDonald BJ, Tinmouth A. Clinical consequences of anemia and red cell transfusion in the critically ill. Crit Care Clin. 2004;20:225-235. 58. Tinmouth A, Fergusson D, Yee IC, Hebert PC. Clinical consequences of red cell storage in the critically ill. Transfusion. 2006;46:2014-2027. 59. Reeves BC, Murphy GJ. Increased mortality, morbidity, and cost associated with red blood cell transfusion after cardiac surgery. Curr Opin Cardiol. 2008;23:607-612. 60. Hill GE, Frawley WH, Griffith KE, Forestner JE, Minei JP. Allogeneic blood transfusion increases the risk of postoperative bacterial infection: a meta-analysis. J Trauma. 2003;54:908-914.
131
61. Silverboard H, Aisiku I, Martin GS, Adams M, Rozycki G, Moss M. The role of acute blood transfusion in the development of acute respiratory distress syndrome in patients with severe trauma. J Trauma. 2005;59:717-723. 62. Amato A, Pescatori M. Perioperative blood transfusions for the recurrence of colorectal cancer. Cochrane Database Syst Rev. 2006: CD005033. 63. Silva Junior JM, Rezende E, Amendola CP, et al. Red blood cell transfusions worsen the outcomes even in critically ill patients undergoing a restrictive transfusion strategy. Sao Paulo Med J. 2012;130:77-83. 64. Purdy FR, Tweeddale MG, Merrick PM. Association of mortality with age of blood transfused in septic ICU patients. Can J Anaesth. 1997;44: 1256-1261. 65. Vamvakas EC, Carven JH. Transfusion and postoperative pneumonia in coronary artery bypass graft surgery: effect of the length of storage of transfused red cells. Transfusion. 1999;39:701-710. 66. Koch CG, Li L, Sessler DI, et al. Duration of red-cell storage and complications after cardiac surgery. N Engl J Med. 2008;358: 1229-1239. 67. Marik PE, Sibbald WJ. Effect of stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA. 1993;269:3024-3029. 68. Berezina TL, Zaets SB, Morgan C, et al. Influence of storage on red blood cell rheological properties. J Surg Res. 2002;102:6-12. 69. McMahon TJ, Ahearn GS, Moya MP, et al. A nitric oxide processing defect of red blood cells created by hypoxia: deficiency of S-nitrosohemoglobin in pulmonary hypertension. Proc Natl Acad Sci U S A. 2005;102:14801-14806. 70. Bennett-Guerrero E, Veldman TH, Doctor A, et al. Evolution of adverse changes in stored RBCs. Proc Natl Acad Sci U S A. 2007;104: 17063-17068. 71. Reynolds JD, Ahearn GS, Angelo M, Zhang J, Cobb F, Stamler JS. S-nitrosohemoglobin deficiency: a mechanism for loss of physiological activity in banked blood. Proc Natl Acad Sci U S A. 2007;104: 17058-17062. 72. Chatterjee S, Wetterslev J, Sharma A, Lichstein E, Mukherjee D. Association of blood transfusion with increased mortality in myocardial infarction: a meta-analysis and diversity-adjusted study sequential analysis. Arch Intern Med. 2013;173(2):132-139. 73. Hebert PC, Yetisir E, Martin C, et al. Is a low transfusion threshold safe in critically ill patients with cardiovascular diseases? Crit Care Med. 2001;29:227-234. 74. Rao SV, Jollis JG, Harrington RA, et al. Relationship of blood transfusion and clinical outcomes in patients with acute coronary syndromes. JAMA. 2004;292:1555-1562.
131.e1
APPENDIX SUPPLEMENTAL DATA Study Protocol: Types of studies: We included randomized controlled trials with a concurrent control group. Trials were included if the comparison groups were assigned on the basis of a clear transfusion trigger or threshold of 7 g/dL, described as a hemoglobin or hematocrit level. Patients in the control group were required to receive a transfusion of red blood cells at higher hemoglobin or hematocrit thresholds than the intervention group. Types of participants: We included trials of patients with surgical or medical conditions, involving adults or children, including neonates. Types of interventions: The intervention considered was the use of red blood cell transfusion thresholds (“triggers”) as a means of providing insights for rational transfusion practices. Outcomes assessed: In-hospital mortality, 30-day mortality, total mortality over the period of follow-up for individual studies, acute coronary syndrome, pulmonary edema, bacterial infections, rebleeding, number of patients receiving any blood transfusion, and units of blood transfused per patient.
The American Journal of Medicine, Vol 127, No 2, February 2014 9. (transfus* adj5 (polic*or practic* or protocol* or trigger* or threshold* or indicator* or strateg* or criteri* or standard*)).mp. 10. ((Red blood cell* or RBC) adj5 (polic*or practic* or protocol* or trigger* or threshold*or indicator* or strateg* or criteri* or standard*)).mp. 11. ((H?emoglobin or h?emocrit or HB or HCT) adj5 (polic*or practic* or protocol* or trigger* or threshold* or indicator* or strateg* or criteri* or standard*)).mp. 12. (transfus* adj5 (restrict* or liberal*)).mp. 13. ((blood or transfusion*) adj3 (management or program*)).mp. 14. 8 or 9 or 10 or 11 or 12 or 13 15. randomi?ed.ab,ti. 16. randomized controlled trial.pt. 17. controlled clinical trial.pt. 18. placebo.ab. 19. clinical trials as topic.sh. 20. randomly.ab. 21. trial.ti. 22. 15 or 16 or 17 or 18 or 19 or 20 or 21 23. (animals not (humans and animals)).sh. 24. 22 not 23 25. 24 and 14.
Search methods for identification of studies: We did not restrict our search for trials by date, language, or publication status. A comprehensive search of the MEDLINE database from 1966 to April 2013 was conducted using the following MeSH terms and strings:
Searching other resources: We contacted authors of published studies for clarification of trial methodology and data. This was provided in 1 instance.21 We searched the reference lists of relevant reviews and published articles, as well as the reference lists of all included trials for further studies.
1. *Blood Transfusion/ 2. ((Red blood cell* or RBC) adj3 (therap* or transfus*)).mp. 3. 1 or 2 4. exp Reference Standards/ 5. standards.fs. 6. methods.fs. 7. 4 or 5 or 6 8. 3 and 7
Assessment of risk of bias in included studies: We completed a “risk of bias” table for each study, incorporating a description of the study’s performance against each of the above domains as follows: “low,” “unclear” (indicating unclear or unknown risk of bias), and “high” risk of bias. One investigator (SC) studied the following domains: sequence generation; allocation concealment; blinding; incomplete outcome data; selective outcome reporting; other potential sources of bias.
Salpeter et al
Restrictive Transfusion Strategies and Clinical Outcomes
Appendix Figure 1 confidence interval.
131.e2
Restrictive transfusion strategy and other clinical events. CI ¼
Appendix Table 1 Patient-Level Characteristics of Randomized Controlled Trials Evaluating a Liberal Transfusion Strategy (Hemoglobin Trigger of <7.0 g/dL) Compared with a More Liberal Strategy Trial Trial Characteristics
Hebert et al12
Lacroix et al28
Villanueva et al23*
Setting
Adult Intensive Care Unit
Pediatric Intensive Care Unit
Severe Acute Upper Gastrointestinal Bleeding
Transfusion strategy
Restrictive Arm
Liberal Arm
Restrictive Arm
Liberal Arm
Restrictive Arm
Liberal Arm
Transfusion trigger (g/dL) Total (N) Mean age (y) Men (%) Baseline hemoglobin (g/dL) Mean units of red-cell transfusion Mean length of storage of red cells (d [SD]) Cardiovascular disease (N [%]) Infection (N [%])
7.0 418 57.1 64 8.2 2.5 NR
10.0 420 58.1 61 8.2 2.3 NR
7.0 320 2.98 59 8.0 0.9 14.9 [11.8]
9.5 317 3.3 60 8.0 1.7 15.2 [10.6]
7.0 444 64 71 9.55 1.5 15.8 [8]
9.0 445 66 65 9.44 3.7 15.8 [8]
76 [18] 114 [27]
94 [22] 108 [26]
76 [24] 157 [49]
75 [24] 155 [49]
NR 37 [8.3]
NR 44 [9.8]
NR ¼ not reported (in primary publication, as referenced); SD ¼ standard deviation. *Additional data provided by corresponding author on request.
131.e3 Appendix Table 2
Trial
The American Journal of Medicine, Vol 127, No 2, February 2014 Risk of Bias Assessments for Included Trials Blinding of Participants and Random Sequence Allocation Researchers Concealment Generation (Selection Bias) (Performance Bias) (Selection Bias)
Hebert et al12 Low Lacroix et al28 Low Villanueva et al23 Low
Low Low Low
Low Low Low
Blinding of Selective Incomplete Outcome Other Outcome Data Reporting Assessment (Detection Bias) (Attrition Bias) (Reporting Bias) Bias Low Low Low
Low Low Low
Low Low Low
Low Low Low