Restrictive versus liberal blood transfusion in patients with coronary artery disease: a meta-analysis

<p><b>Objectives:</b> To compare clinical outcomes between restrictive versus liberal blood transfusion strategies in patients with coronary artery disease (CAD).</p> <p><b>Research design and methods:</b> A literature search from January 1966 to May 2016 was performed in PubMed, EMBASE and Cochrane Library to find trials evaluating a restrictive hemoglobin transfusion trigger of ≤8 g/dL, compared with a more liberal trigger. Two study authors independently extracted data from the trials. The primary outcome was mortality and the secondary outcome was subsequent myocardial infarction. Relative risks (RRs) with their 95% confidence intervals (CIs) were assessed.</p> <p><b>Results:</b> Six trials involving 133,058 participants were included in this study. Pooled results revealed no difference in mortality between the liberal transfusion and restrictive transfusions (RR = 1.17, 95% CI = 0.91–1.52, <i>P</i> = .22). Subgroup analysis revealed that a restrictive transfusion strategy was associated with a higher risk of in-hospital mortality (RR = 1.38, 95% CI = 1.15–1.67, <i>P</i> < .001) and 30 day mortality (RR = 1.21, 95% CI = 1.01–1.45, <i>P</i> = .03), compared with the liberal strategy. No significant difference was found between the liberal transfusion strategy and restrictive transfusion strategy in risk for subsequent myocardial infarction (RR = 1.09, 95% CI = 0.57–2.06, <i>P</i> = .80).</p> <p><b>Limitations:</b> Limitations include (1) limited number of trials, especially those evaluating myocardial infarction, (2) observed heterogeneity, (3) confounding by indication and other inherent bias may exist.</p> <p><b>Conclusion:</b> The findings suggest that restrictive blood transfusion was associated with higher in-hospital and 30 day mortality than liberal blood transfusion in CAD patients. The conclusions are mainly based on retrospective studies and should not be considered as recommendation before they are supported by randomized controlled trials.