Effect of beginning recombinant erythropoietin treatment within the first week of life, among very-low-birth-weight neonates, on "early" and "late" erythrocyte transfusions: a meta-analysis.

OBJECTIVES

Erythrocyte transfusions to neonates can be categorized as "early" if given during the first 3 weeks of life and "late" if given thereafter. We used a meta-analysis to determine whether recombinant erythropoietin (rEpo) administration to very-low-birth-weight (VLBW, <1500 g) neonates, beginning in the first week of life, reduces either "early" or "late" transfusions.

STUDY DESIGN AND METHODS

Studies that used a randomized, placebo-controlled, double-masked design were deemed acceptable. We identified 12 acceptable, relevant, clinical trials. Additional data not provided in the publications were obtained from two of the authors.

RESULTS

The acceptable studies involved an aggregate of 561 rEpo and 529 placebo recipients. If rEpo was begun in the first week of life, the summary odds ratio (OR) for receiving any transfusion ("early" or "late") was 0.52, 95% confidence interval (CI): 0.34 to 0.79 (p=0.001). The OR for receiving an "early" transfusion was 0.54 (95% CI: 0.25 to 1.15; p=0.055), and the OR for receiving a "late" transfusion was 0.56 (95% CI: 0.37 to 0.83; p=0.036). Heterogeneity among studies was too great to estimate the effect of rEpo on the number of transfusions received or the volume of blood transfused (p<0.001 for the Q-test statistic). Subgroup analysis suggested that when rEpo is begun in the first week of life, neonates 1000 to 1500 g and >29 weeks are more likely to completely avoid transfusion than are extremely low-birth-weight (ELBW, <1000 g) neonates. No dose-response relationship was apparent between rEpo dose or iron dose and transfusion. No difference was apparent depending on whether the rEpo was given subcutaneously vs intravenously.

CONCLUSION

If rEpo is begun in the first week of life, a moderate reduction can be expected (p=0.001) in the proportion of VLBW neonates transfused. Reduction is less significant in "early" transfusion (p=0.055) than in "late" transfusion (p=0.036). Such treatment is not likely to eliminate transfusions among ELBW neonates completely.