Prediction of transfusions in extremely low-birthweight infants in the erythropoietin era

Background: The purpose of the present paper was to detect the clinical factors most predictive of red blood cell (RBC) transfusion in extremely low-birthweight (ELBW) infants in the recombinant human erythropoietin era. Methods: Between 1995 and 2000, 66 ELBW infants were admitted to a level III neonatal intensive care unit. Fifty-four of 66 infants were eligible for enrollment in the present study. Infants were treated with erythropoietin 200 IU/kg per dose s.c. twice a week with 4-6 mg/kg per day iron supplement. Results: The mean gestational age and birthweight were 26.5 +/- 2.1 weeks and 776 +/- 134 g, respectively. Ten of 54 ELBW infants (18.5%) died during the first 21 days. Eight of 10 dead infants (80.0%) and 27 of 44 surviving infants (61.4%) received one or more RBC transfusions. The overall requirement for RBC transfusions in the surviving infants was 3.0 +/- 3.2 per infant/hospital course (range: 0-9) . There were significant differences in gestational weeks, birthweight, initial hemoglobin value, 5 min Apgar score, phlebotomy loss, phlebotomy loss/birthweight, duration of mechanical ventilation, duration of oxygen supplement, and incidence of both intraventricular hemorrhage and chronic lung disease between the transfused and non-transfused group. The predictive variables, initial hemoglobin level (odds ratio [OR] 2.61; 1 g/dL), birthweight (OR 3.00; 100 g), and gestational week (OR 1.89; 1 week), were found to be most predictive for transfusion on logistic regression analysis. Conclusions: ELBW infants are still the population at greatest risk for repeated blood transfusions after introduction of erythropoietin treatment. If labor develops, it is often impossible to extend the pregnancy period, therefore efforts should be made to increase hemoglobin level at birth.