Original article Variation in triggers and use of perioperative blood transfusion in major gastrointestinal surgery A. Ejaz 1,3 , G. Spolverato 1 , Y. Kim 1 , S. M. Frank 2 and T. M. Pawlik 1 Departments of 1 Surgery and 2 Anesthesiology/Critical Care Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland, and 3 Department of Surgery, University of Illinois Hospital and Health Sciences System, Chicago, Illinois, USA Correspondence to: Professor T. M. Pawlik, Department of Surgery, Johns Hopkins Hospital, 600 North Wolfe Street, Blalock 688, Baltimore, Maryland 21287, USA (e-mail: tpawlik1@jhmi.edu) Background: The decision to perform intraoperative blood transfusion is subject to a variety of clinical and laboratory factors. This study examined variation in haemoglobin (Hb) triggers and overall utilization of intraoperative blood transfusion, as well the impact of transfusion on perioperative outcomes. Methods: The study included all patients who underwent pancreatic, hepatic or colorectal resection between 2010 and 2013 at Johns Hopkins Hospital, Baltimore, Maryland. Data on Hb levels that triggered an intraoperative or postoperative transfusion and overall perioperative blood utilization were obtained and analysed. Results: Intraoperative transfusion was employed in 437 (156 per cent) of the 2806 patients identiied. Older patients (odds ratio (OR) 168), patients with multiple co-morbidities (Charlson co-morbidity score 4 or above; OR 166) and those with a lower preoperative Hb level (OR 495) were at increased risk of intraoperative blood transfusion (all P < 0001). The Hb level employed to trigger transfusion varied by sex, race and service (all P < 0001). A total of 105 patients (240 per cent of patients transfused) had an intraoperative transfusion with a liberal Hb trigger (10g/dl or more); the majority of these patients (78; 743 per cent) did not require any additional postoperative transfusion. Patients who received an intraoperative transfusion were at greater risk of perioperative complications (OR 155; P = 0002), although patients transfused with a restrictive Hb trigger (less than 10g/dl) showed no increased risk of perioperative morbidity compared with those transfused with a liberal Hb trigger (OR 122; P = 0514). Conclusion: Use of perioperative blood transfusion varies among surgeons and type of operation. Nearly one in four patients received a blood transfusion with a liberal intraoperative transfusion Hb trigger of 10g/dl or more. Intraoperative blood transfusion was associated with higher risk of perioperative morbidity. Paper accepted 18 June 2014 Published online in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.9617 Introduction The decision to transfuse blood in the perioperative period is subject to a variety of clinical and laboratory factors 1 . Several studies have found an association between blood transfusion and worse short- and long-term postopera- tive outcomes 2 7 . As such, several randomized clinical trials 8 12 have been conducted to evaluate the impact of the restrictive use of blood transfusion, and have found equiva- lent or superior outcomes compared with the liberal use of blood transfusion. Based largely on the results from these trials, numerous national guidelines 13 16 have been issued that provide evidenced-based recommendations against the use of blood transfusion in most situations in patients with a haemoglobin (Hb) level above 8 g/dl. Despite this, several studies 17 19 have shown that transfusion practice varies signiicantly between and within institutions. These studies have, however, focused largely on blood utilization dur- ing the perioperative period rather than the intraoperative management of blood transfusion. Thus, appropriate tim- ing and use of blood transfusion in the operating room are less well deined. Particularly in operations with high surgical blood loss, intraoperative blood transfusion is given to maintain haemodynamic stability, and to provide adequate oxygen delivery and maintain organ perfusion. The decision to transfuse is often made jointly by the operating surgeon © 2014 BJS Society Ltd BJS Published by John Wiley & Sons Ltd