Predictive factors for postoperative pancreatitis after pancreaticoduodenectomy: A single-center retrospective analysis of 1465 patients

Abstract Background Postoperative acute pancreatitis (POAP) after pancreaticoduodenectomy (PD) has been recently recognized as an independent complication that is associated with undesirable postoperative outcomes and often precedes other complications, yet predictive factors attributable to POAP after PD remain elusive. Methods The data from 1465 consecutive patients who underwent laparotomy or minimally invasive robotic PD from March 2010 to December 2018 were retrospectively reviewed. POAP was defined as an elevation of the serum amylase levels above the institution's normal upper limit (100 U/L) on postoperative day (POD) 1. Univariate and multivariate analyses were performed to investigate the predictive factors for POAP after PD and the association between POAP and clinically relevant postoperative pancreatic fistulas (CR-POPFs). Results Among the 1465 patients, 411 (28%) underwent minimally invasive robotic surgeries, and the overall POAP and CR-POPFs rates were 770 (53%) and 277 (19%), respectively. The female sex (OR 1.76), a normal bilirubin level (OR 1.55), the robotic surgery (OR 1.36), a main pancreatic duct (MPD) ≤3 mm (OR 5.69) and a high-risk nonadenocarcinoma pathology (cystic disease: OR 4.33; pNETs: OR 4.34; others: OR 2.74) were considered independent risk factors for POAP. A nondilated MPD was a predominant predictor for POAP, with 72.2% sensitivity and 71.8% specificity. POAP was also an independent predictive factor for CR-POPFs (OR 3.48). Conclusion A nondilated MPD, a high-risk pathology, the female sex, a normal bilirubin level and the robotic surgery were independent predictive factors for POAP after PD. Prevention and early treatment strategy changes can be made based on these preoperative predictive factors.