Impact of Preoperative Steroid Use on Pancreaticoduodenectomy Outcomes: Analysis of the NSQIP Database

Abstract

Introduction: While preoperative steroid use is frequently cited as a risk factor for postoperative morbidity, the impact of steroid use on the outcomes of patients undergoing pancreaticoduodenectomy has not been previously quantified. Methods: The National Surgical Quality Improvement Program (NSQIP) database from 2005-2019 was utilized. All patients who underwent pancreaticoduodenectomy were included in the analysis. Patient characteristics and unadjusted outcomes were described, and multivariate logistic regression was used to assess the impact of steroid use on postoperative outcomes. Results: A total of 49,672 patients were identified of whom 1,293 (2.6%) were taking steroids preoperatively. Complication rates were higher in patients on steroid therapy – 3.2% vs 2.2% (p=0.01) 30-day mortality, 16.1% vs 13.7% (p=0.02) 30-day readmissions, 7.7% vs 6.1% (p=0.02) 30-day reoperation rate, and 36% vs 31% (p<0.01) infectious complications. On multivariate logistic regression, steroid use was significantly associated with any complication (OR [95% CI]) 1.2 [1.1-1.3], reoperation 1.3 [1.03-1.6], postoperative infectious complications 1.2 [1.1-1.4], pancreatic fistula 1.3 [1.1-1.6] and prolonged postoperative hospitalization 1.3 [1.1-1.5]. Steroid use was not found to be a significant risk factor for 30-day mortality, readmission, or wound complications after controlling for patient characteristics. Conclusions: Preoperative steroid use has a negative impact on overall complication rates, reoperations, and infectious complications. This is important information when weighing treatment options and counseling patients preoperatively as well as treating patients postoperatively.