Restrictive versus liberal red blood cell transfusion thresholds in very low birth weight infants: a systematic review and meta-analysis

Abstract

Aim: A systematic review was conducted to examine the effects of restrictive versus liberal red blood cell (RBC) transfusion thresholds on clinically important outcomes in very low birth weight (VLBW) infants.

Methods: Randomised controlled trials (RCTs) of varying RBC transfusion thresholds in VLBW infants were identified by searching MEDLINE, EMBASE, CINAHL, all of the Cochrane Library and other supplementary sources. Selected studies included one of the following outcomes: total number of red blood cell transfusions, donor exposure rate, cranial ultrasonographically diagnosed brain injury, retinopathy of prematurity, bronchopulmonary dysplasia, necrotising enterocolitis or death. Studies to be included were selected by two reviewers who also assessed the risk of bias of each trial. Data extraction and analyses were independently performed by two reviewers. All data were analysed using RevMan 5.

Results: Six RCTs were identified. One trial did not meet the inclusion criteria, while two had inadequate methodological quality. Pooled analysis of two trials showed that the restrictive transfusion group received a significantly lower mean number of transfusions per infant (mean difference (MD) -1.35, 95% confidence interval (CI) [-2.61, -0.09]) and donor exposure rate (MD -0.54, 95% CI [-0.93, -0.15]). No other statistically significant differences were observed.

Conclusion: Restrictive RBC transfusion thresholds in VLBW infants may be utilised without incurring clinically important increases in the risk of death or major short-term neonatal morbidities.

Keywords: premature infant; red blood cell transfusion; very low birth weight.