Impact of extremes of BMI on perioperative complications in complex pancreatic resections

Document Type

Conference Proceeding

Publication Date


Publication Title

J Am Coll Surg


INTRODUCTION: Pancreaticoduodenectomy offers the best chance for long-term survival and cure for patients with pancreatic malignancy. These complex resections are associated with a high perioperative morbidity. We hypothesized that BMI extremes (low<18, high 35-50, ultrahigh>50) would represent risk factors for perioperative complications in complex pancreatic surgery. METHODS: All patients undergoing elective complex pancreatectomy (N=24,522) from 2005 to 2014 were identified utilizing the American College of Surgeons (ACS) NSQIP database. Univariate and multivariable analyses were performed to identify the impact of BMI on perioperative outcomes. RESULTS: Patients with BMI extremes had a higher incidence of unplanned intubation (ultrahigh/high/low/normal, %: 8.3/7.0/ 5.8/4.8; p<0.001), prolonged mechanical ventilation (>48 hours) (%: 5.5/8.1/6.2/4.6; p<0.001), Clavien 4 or higher complication (%: 12.8/15.0/10.7/10.0; p<0.001) and length of hospital stay (days: 13.19/14.03/14.87/12.86; p<0.001), organ/space SSI (12.8/17.0/8.2/11.5; p<0.001) and death (%: 4.6/4.2/2.7/3.1; p<0.034). On multivariable analyses, adjusting for comorbidity status, high BMI represented an independent risk factor for reintubation (odds ratio [OR] 1.68; p<0.001), failure to wean (OR 1.91; p<0.001), Clavien 4 complication (OR 1.65; p<0.001), death (OR 1.43; p=0.002). When adjusting for comorbidity status low and ultrahigh BMI did not represent risk factors for complications. CONCLUSIONS: BMI extremes have less impact on perioperative outcomes after pancreaticoduodenectomies than previously thought when adjusting for comorbidity status. While selection bias is possible, it appears that the perioperative complication risk is not higher than the general population. This may have implications for preoperative patient counselling, treatment sequencing and preoperative nutrition optimization. In addition, better identification of patients who are at low risk of having perioperative complications in the setting of complex oncologic procedures may facilitate patient selection for future fast-track postoperative pathways.




4 Suppl 2

First Page


This document is currently not available here.