Red blood cell transfusions in neonatal intensive care units: a nationwide observational cohort study.

Objective: To describe the use and nationwide variation of red blood cell (RBC) transfusions in neonatal intensive care units (NICUs) following the introduction of the revised national transfusion guideline in 2019.

Design And Patients: We randomly selected neonates born below 32 weeks' gestation admitted to any NICU in the Netherlands in 2020 to include in our retrospective observational cohort study.

Main Outcome Measures: Main outcome measures were the number of neonates receiving at least one transfusion, and the number of transfusions per transfused neonate.

Results: Of 762 neonates included, 34% (257/762) received at least one RBC transfusion, varying between centres from 20% (12/61) to 50% (39/77). Median phlebotomy loss during admission was 8.2 mL/kg (IQR 4.5-17.3 mL/kg), equating to 54.3% of the median transfusion volume. Of 770 transfusions, 358 (47%) were administered above the recommended threshold, and the proportion of transfusions given above the threshold varied between centres from 15% to 719%. Median transfusion dosage and mean infusion duration were 15.1 mL/kg (IQR 15.0-16.7 mL/kg) and 3.9 hours (SD 1.1 hour) and varied from 14.8 mL/kg to 18.9 mL/kg and from 2.5 hours to 5.5 hours between centres. Blood transfusion volume was positively correlated with cumulative volume of blood draws (Pearson correlation coefficient 0.84, 95% CI 0.82 to 0.86) and lower gestation.

Conclusions: Large variation in transfusion practice remains between Dutch NICUs despite a national guideline. Extreme prematurity and cumulative blood draws were associated with increased use of RBC transfusions. Benchmarking will yield leverage points to understand and potentially prevent unwarranted variation.