Adverse effects of red blood cell transfusions in neonates: a systematic review and meta-analysis.

BACKGROUND

Controversy exists regarding the contribution of blood transfusions to a range of adverse clinical outcomes in neonates. The aim of our systematic review was to identify the broader literature on harmful effects and associations potentially attributable to red blood cell (RBC) transfusions.

STUDY DESIGN AND METHODS

A comprehensive search of MEDLINE (PubMed) and EMBASE was undertaken. Eligible studies included both randomized controlled trials (RCTs) and nonrandomized studies examining the effects of small volume (10-20 mL/kg) RBC transfusions on neonates. Primary outcomes of interest were mortality, chronic lung disease, retinopathy of prematurity, necrotizing enterocolitis, and intraventricular hemorrhage. Two independent authors conducted a review of abstracts and then of full-text article reviews as well as data extraction and quality assessments.

RESULTS

Sixty-one studies were eligible for inclusion, including 16 (26%) randomized studies. The majority of studies were nonrandomized (n = 45; 74%), which included 32 observational studies with and 13 studies without a comparator group. There was no evidence that rates of mortality differed between restrictive and liberal strategies for transfusion (eight RCTs: risk ratio, 1.24; 95% confidence interval, 0.89-1.672, heterogeneity = 0%) or for necrotizing enterocolitis (five RCTs: risk ratio, 1.45; 95% confidence interval, 0.91-2.33; heterogeneity = 0%). A liberal strategy also was not superior to restrictive transfusion practice in the pooled randomized studies for rates of retinopathy of prematurity, chronic lung disease, or intraventricular hemorrhage.

CONCLUSIONS

Statistically significant differences in a range of harmful outcomes between neonates exposed to restrictive and liberal RBC transfusion practice were not found. However, the risks of bias identified in many studies and the lack of consistent reporting and definitions of events limits our conclusions.