Three point transfusion risk score in hepatectomy: an external validation using the American College of Surgeons – National Surgical Quality Improvement Program (ACS-NSQIP)

Abstract

BackgroundRisk of red blood cell transfusion (RBCT) in partial hepatectomy is 17–27%; strategies to reduce transfusions can be targeted in patients at increased risk. A Three Point Transfusion Risk Score (TRS) was previously developed to predict patients' risk of transfusion during and following hepatectomy. Here, it was subject to external validation using the ACS-NSQIP database. MethodsTRIPOD guidelines were followed. A validation cohort was created with the ACS-NSQIP dataset. Risk groups for RBCT were created using the TRS: anemia (hematocrit ≤36%), major liver resection (≥4 segments) and primary liver malignancy. Concordance index was used to assess the discrimination. The Hosmer–Lemeshow test for goodness of fit and calibration curves were used to assess calibration. ResultsOf 2854 hepatectomies, 18.9% received RBCT. The TRS stratified patients from low (8.5% risk of RBCT) to very high risk (40.6%) of RBCT. The concordance was 0.68 (95% CI 0.66–0.70). Hosmer–Lemeshow test and calibration curves supported good predictive performance of the model. ConclusionThe TRS adequately discriminated risk of RBCT in an external sample of patients undergoing hepatectomy. It provides a simple method to identify patients at high transfusion risk. It can be used to tailor patient blood management initiatives and reduce the use of RBCT.