Can Transfusions be Eliminated in Major Abdominal Surgery? Analysis of a Vive-Year Experience of Blood Conservation in Patients Undergoing Pancreaticoduodenectomy

Abstract

Pancreaticoduodenectomy (PD) has historically required perioperative blood transfusion in 40 to 60 per cent of cases. Growing data suggest that transfusions may be deleterious in the surgical patient. We recently initiated a minimal transfusion approach to PD consisting of limited postoperative blood draws, early iron supplementation, changes in surgical technique, and elimination of hemoglobin transfusion triggers. Predictors of perioperative transfusion were analyzed in 130 consecutive patients undergoing PD by a single surgeon between 2008 and 2013, divided into two eras with 65 patients each. Patients in each era were similar with respect to age, comorbidities, American Society of Anesthesiologists class, body mass index, and diagnosis. The transfusion rate for the entire group was 22 per cent. Nonsignificant predictors of perioperative transfusion include American Society of Anesthesiologists class ≥3 (P = 0.41), vascular resections (P = 0.56), body mass index ≥30 (P = 0.72), and intraoperative blood loss (P = 0.89). Significant predictors of transfusion include PD performed in Era 1 as well as preoperative hemoglobin levels <10 g/dL. In Era 1, 38 per cent of patients required transfusion compared with 6 per cent in Era 2 (P < 0.01). Shorter length of stay and a trend toward decreased pancreatic fistulae were seen in Era 2. Transfusions can be almost completely eliminated in PD and this may contribute to improved outcomes.