Identifying variations in blood use based on hemoglobin transfusion trigger and target among hepatopancreaticobiliary surgeons.

BACKGROUND Transfusion practice among surgeons varies despite several trials supporting the restrictive use of blood products. We sought to define the variation in surgeon transfusion hemoglobin (Hb) triggers and targets among patients undergoing hepatopancreaticobiliary (HPB) procedures, as well as assess perioperative outcomes among patients receiving transfusions under a restrictive vs liberal transfusion strategy. STUDY DESIGN Using prospectively collected data, variations in transfusion Hb triggers, targets, and overall use of blood were examined among 1,554 patients undergoing an HPB procedure by 1 of 11 surgeons at Johns Hopkins Hospital between 2009 and 2013. Perioperative outcomes were compared among patients treated with a restrictive (Hb < 8 g/dL) vs liberal (Hb ≥ 8 g/dL) transfusion strategy. RESULTS Among the 1,554 patients included in the cohort, 504 (32.4%) received at least 1 transfusion of red cells. Patients who received a transfusion were older and had more medical comorbidities (both p < 0.001). Among 620 patients who met inclusion for blood use analysis, 344 (55.5%) received a transfusion using a restrictive trigger, 160 (25.8%) with a liberal trigger, and 116 (18.7%) patients had an Hb < 8 g/dL but did not receive a transfusion. The mean transfusion Hb trigger was 7.7 ± 1.1 g/dL and the mean target was 9.3 ± 1.1 g/dL. Patients transfused with a higher Hb trigger were older, male, white, and had more medical comorbidities and higher intraoperative estimated blood loss (all p < 0.05). Hemoglobin transfusion triggers varied among different surgeons (p < 0.001). Perioperative mortality, complications, and length of stay did not differ between those treated with a restrictive vs liberal transfusion strategy (p > 0.05). CONCLUSIONS Nearly 1 in 3 patients undergoing an HPB procedure received a blood transfusion in the perioperative period. Transfusion use, indication ("trigger"), and dose ("target") varied among surgeons. The use of a restrictive transfusion strategy did not affect perioperative outcomes.