270 A Randomized Controlled Trial of Low vs High Hemoglobin Thresholds for Transfusion in Extremely Low-Birth-Weight Infants

Abstract

Background: Infants of extremely low-birth-weight (ELBW, or <1000g) are frequently exposed to multiple transfusions of packed red blood cells. Our objective was to evaluate the effects of using low vs. high levels (thresholds) of blood hemoglobin concentration for transfusion of ELBW infants on mortality or severe morbidity until discharge.Methods: We selected infants <1000 g birth-weight and <48 hr of age, excluding infants with established blood disease, shock or sepsis. We randomized 451 (223 low, 228 high) infants to transfusion threshold algorithms of low vs. high hemoglobin. The algorithms differed by 10–20 g/l and were adjusted for age and need for respiratory support (low from 115 to 75 g/l, high from 135 to 85 g/l). These were maintained until discharge. Additional transfusions were permitted for shock, sepsis or surgery.Results: Maternal and infant prognostic factors were similar between the groups. Median (IQR) birth-weight was 770 (670–887) g and gestational age 26 (25–27) wk. Transfusion and Hematologic Outcomes: Comparing infants in the low vs. high threshold group, the time to first transfusion was delayed by a median of two days, fewer infants were exposed to any transfusion (194 (87%) vs. 214 (94%), p<0.01), and fewer transfusions were given (means 4.8 vs. 5.5, p=0.09). Hemoglobin levels fell faster from admission to discharge (from 164 to 101g/l vs. from 165 to 111g/l; mean difference after first week was 11 g/l, p<0.0001). Clinical Outcomes: There were no significant differences between low and high threshold groups respectively in mortality (22% vs. 18%, OR 1.3 95% c.i. 0.8–2.2), bronchopulmonary dysplasia as oxygen supplementation at 36 wk (59% vs. 56%, OR 1.2 c.i. 0.7–1.8), retinopathy of prematurity of stage 3 or worse(18% vs. 17%, OR 1.2 c.i.0.7–2.0) or periventricular leukomalacia/ventriculomegaly by ultrasound (19% vs. 21% OR 0.8 c.i. 0.5–1.4). Rates of combined death or severe morbidity were not significantly different (74% vs. 70% (OR 1.3 c.i. 0.8–2.0). Overall, 28% of all infants (or 35% of survivors) survived with no severe morbidity to discharge.Conclusion: Using a regimen of low rather than high hemoglobin transfusion thresholds results in a reduced exposure to blood transfusion for infants of extremely low birth weight with postnatal hemoglobin levels approximately 11 g/l lower. There is no difference in mortality or severe morbidity in these very high-risk infants managed with low vs. high hemoglobin thresholds.