Optimizing Transfusion-Related Postoperative Outcomes in Craniosynostosis Repair

BACKGROUND: As cranial vault reconstruction for craniosynostosis is associated with significant blood loss and transfusions, managing intraoperative and postoperative hematologic status is a significant challenge for both plastic surgeons and anesthesiologists. Factors contributing to these challenges include young patient age with low total blood volume (TBV), as well as the difficulty of quantifying intraoperative estimated blood loss (EBL) in real time. However, optimizing intraoperative transfusion management for these cases is critical: blood products are independently associated with an increased risk of overall mortality, postoperative complications, multiorgan failure, and prolonged intensive care unit (ICU) stays. This study aims to evaluate how intraoperative fluid management, including blood transfusion, affects incidence of postoperative complications, and respiratory morbidity. METHODS: We conducted a retrospective review of prospectively collected data from October 2012 to November 2019 using the Pediatric Craniofacial Surgery Perioperative Registry at Johns Hopkins Hospital. Pediatric patients (<18 years) undergoing open craniosynostosis repair were included. Endoscopic strip craniectomies were excluded. Outcomes of interest included postoperative complication incidence, intraoperative and postoperative respiratory complications, and hospital length of stay (LOS). RESULTS: Sixty-one patients were included with a median age of 1.2 years (SD = 3.3); 36% were female, 54% Caucasian, and median ASA score was 2. Mean ICU and total hospital LOS were 3.3 and 6 days, respectively. Intraoperatively, mean EBL was 494 ml (SD = 403) and mean EBL/TBV was 0.55 (SD = 0.42). Patients were given an average of 1,412-ml crystalloid fluids for a mean crystalloid/EBL ratio of 5.1:1. On average, 646-ml blood products were given (mean 75% TBV). When controlling for ASA, odds of any postoperative complication were increased over 14-fold by intraoperatively transfusing >85% TBV in blood products compared with <85% (P = 0.028). Increasing %TBV transfused was significantly associated with increased incidence of intraoperative or postoperative respiratory complications, with an odds ratio of 5.2 (P = 0.049). Total and ICU LOS were increased as intraoperative %TBV transfused increased, although these findings did not reach significance (P = 0.08, 0.09). A higher difference between the highest intraoperative and preoperative hemoglobin values was also associated with 1.75 increased odds of postoperative complication (P = 0.03). Increasing the intraoperative crystalloid:EBL ratio was significantly associated with incidence of any postoperative complication (P = 0.04), with a ratio of ≥7:1 associated with an odds ratio of 13.07 (P = 0.05). Age, gender, surgeon, and specific procedure were not significantly associated with outcome in univariate analysis; crystalloid:blood ratios were not significantly associated with complication rates. Intraoperative %TBV transfused was not associated with postoperative transfusion requirements. CONCLUSION: Postoperative morbidity may be optimized by utilizing transfusion and crystalloid thresholds. As transfusing >85% TBV was associated with increased postoperative complications, we advocate for adopting practices that may either decrease transfusion need to below this threshold (eg, antifibrinolytic therapy, bloodless surgical technique) or provide alternative methods to minimize external transfusion (eg, using cell saver). Additionally, maintaining a crystalloid:EBL ratio of <7:1 may also prevent postoperative complications; colloid replacement may be considered if volume needs exceed this ratio. We aim to conduct further interinstitutional research to identify additional optimal infusion and transfusion practices in this patient population.