Red blood cell (RBC) transfusion is common during infant cardiac surgery. A previous report of pediatric heart transplant recipients showed that increased RBC transfusion volume was independently associated with increased length of intensive care unit stay. It is unclear whether transfusion to infants as a subgroup carries similar risks. This study investigated relationships between intraoperative RBC transfusion during heart transplantation and postoperative length of stay (LOS), morbidity, and mortality in infants.
Retrospective analysis of medical records from infants <1 year old undergoing primary heart transplantation at Loma Linda University Medical Center from 1985 to 2012 was conducted. Exclusion criteria included preoperative exchange transfusion or extracorporeal membrane oxygenation. Data sought included patient characteristics; intraoperative RBC transfusion volume and cardiopulmonary bypass details; and postoperative vasoactive support, ventilator support, morbidity, LOS, and 30-day mortality. The relationship of RBC transfusion volume (mL/kg) to these postoperative variables was assessed by univariate analysis. Multiple regression analysis of postoperative LOS included variables that were independent predictors of LOS or associated with ≥10% change in the β-estimate for RBC effect.
Data from 307 infants showed that most (66.8%) had single-ventricle physiology. Median age at transplant was 50 days, weight 3.95 kg, and intraoperative transfusion volume 109 mL/kg. Transfusion volume was inversely related to age and weight. Median postoperative LOS was 18.2 days. Univariate linear regression analysis of transfused volume showed no relationship to log-transformed postoperative LOS (F(1,305) = 0.00; P = 0.960; R2 = 0.000; β-coefficient = 0.004; 95% confidence interval = −0.1542 to 0.1623). Transfused volume was not related to 30-day mortality (difference −0.162; −0.048 to 0.371 mL/kg; P = 0.112) or to postoperative ventilator support (R2 = 0.047), but was greater in patients who required reoperation (difference −0.246; −0.494 to −0.025; P = 0.004). Multiple regression analysis for all patients revealed age, preoperative ventilator support, prolonged postoperative ventilatory or vasoactive support, transplant year, and 30-day mortality, but not major adverse events, to be significant confounding variables. Adjusting for these variables, transfused volume was not associated with prolonged postoperative LOS.
In contrast to a prior report, we found no correlation between intraoperative RBC transfusion and postoperative LOS when studying only infants. Infants have maturing organ systems, less physiologic reserve, and increased surgical blood loss (evaluated as mL/kg) during cardiac surgery than their larger, older counterparts, distinguishing them from the general pediatric population. These differences require additional studies to determine the outcome impact of transfusion strategies in the infant subgroup.