Extremes of BMI are associated with a higher risk of pancreatic fistula following pancreaticoduodenectomy: an analysis using the NSQIP database

Document Type

Conference Proceeding

Publication Date


Publication Title



Background: Elevated body-mass index (BMI) is a well-described risk factor for postoperative complications. Specifically, the impact of BMI on pancreatic fistula rates following pancreaticoduodenectomy (PD) has been inconsistent. The aim of this study was to investigate pancreas-specific morbidity following pancreaticoduodenectomy for patients with extremes of BMI using a large national database. Methods: The National Surgical Quality Improvement Program database (NSQIP) was queried for patients undergoing PD between 2014–2016. BMI was classified according to the WHO classification as underweight (UW) (40). Univariate and multivariable logistic regression models were used to evaluate the effects of BMI on pancreas-specific morbidity. Stepwise selection was performed and adjustments were made for comorbidities, operative factors and pancreas-specific variables. Postoperative pancreatic fistula was classified into biochemical leak (BCL), Grade B and Grade C as per the International Study Group for Pancreatic Surgery (ISGPS) 2016 definition. P <.05 was considered statistically significant. All analyses were done in SAS 9.4 (SAS Institute, Cary, NC). Results: 10,526 patients were included in the analysis (UW n=302, NW n=3,721, OW n = 3,678, OBI n = 1,779, OBII n=653, OBIII n=363) (Table 1). On univariate analysis elevated BMI (OB I-III) was associated with pancreatic fistula development compared to normal weight patients. This difference persisted on multivariable analysis for OBI and OBIII (OBI vs. NW OR = 1.55 (1.13–2.12); p = 0.007) (OBIII vs. NW, OR = 1.86 (1.08–3.21); p = 0.026). The difference did not persist for OBII (OBII vs. NW OR = 1.53 (0.96–2.24); p = 0.08). Similarly, patients with higher BMI (OBI-III) had a lower odds of a lower grade pancreatic leak compared to NW patients. This difference persisted on multivariate ordinal logistic regression analysis for OBI and OBIII but not for OBII (OB I vs. NW OR = 0.65 (0.48–0.89); p = 0.008) (OBII vs. NW OR = 0.68 (0.43–1.08); p = 0.099) (OBIII vs. NW OR = 0.53 (0.31–0.90); p = 0.02). On multivariable logistic regression there was no statistically significant difference for 30-day mortality, readmission rates, morbidity, delayed gastric emptying among the BMI groups. Conclusion: Elevated BMI increases risk of pancreatic fistula but not mortality or general perioperative morbidity following pancreaticoduodenectomy. Extreme obesity is an independent risk factor for pancreatic fistula and higher fistula grade compared to patients with normal weight. It might be worth considering the incorporation of BMI into the pancreatic fistula risk score. [Figure presented]




Suppl 1

First Page


This document is currently not available here.