Abstract
BACKGROUND: Red blood cell (RBC) transfusions are used frequently in pediatric patients admitted to the intensive care unit (ICU) after cardiac surgery. To improve data-driven transfusion decision-making in the ICU, we conducted a retrospective analysis to assess the effect of RBC transfusion on cerebral and somatic regional oxygenation (rSO2). METHODS: We evaluated post- versus pre-RBC transfusion cerebral rSO2 and somatic rSO2 in all consecutive pediatric patients (age >28 days to <18 years) who underwent biventricular cardiac surgery at a single center between July 2016 and April 2020. RESULTS: The final data set included 263 RBC postoperative transfusion events in 75 patients who underwent 83 surgeries. The median pretransfusion hemoglobin was 10.6 g/dL (25th-75th percentile, 9.3-11.6). The median pretransfusion cerebral and somatic rSO2 were 63% (54-71) and 69% (55-80), which increased by a median of 3 percentage points (-2 to 6) and 2 percentage points (-3 to 6), respectively, after transfusion. After adjusting for pretransfusion hemoglobin, change in hemoglobin posttransfusion versus pretransfusion, and potential confounders (age, sex, and STAT surgical mortality risk score), the posttransfusion versus pretransfusion change in cerebral or somatic rSO2 was not statistically significant. Pretransfusion cerebral rSO2 (crSO2) was ≤50%, a previously described threshold for increased risk for unfavorable neurological outcome, for 22 of 138 (16%) transfusion events with complete pre- and post-crSO2 data. Sixteen of these 22 (73%) transfusions resulted in a posttransfusion crSO2 >50%. When restricting analysis to the first (index) transfusion after arrival to the ICU from the operating room (administered at a median of 1.15 postoperative days [25th-75th percentile, 0.84-1.93]), between-patient pretransfusion hemoglobin was not associated with pretransfusion crSO2 but within-patient posttransfusion versus pretransfusion hemoglobin difference was significantly associated with posttransfusion versus pretransfusion crSO2 difference (mean posttransfusion versus pretransfusion crSO2 difference, 2.54; 95% confidence interval, 0.50-4.48). CONCLUSIONS: In this study, neither cerebral nor somatic rSO2 increased significantly post- versus pre-RBC transfusion in pediatric cardiac surgery patients admitted to the ICU after biventricular repairs. However, almost three-quarters of transfusions administered when pretransfusion crSO2 was below the critical threshold of 50% resulted in a posttransfusion crSO2 >50%. In addition, the significant within-patient change in crSO2 in relation to the change in posttransfusion versus pretransfusion hemoglobin in the immediate postoperative period suggests that a personalized approach to transfusion following within-patient trends of crSO2 rather than absolute between-patient values may be an important focus for future research.