Postoperative narcotic use is associated with development of clinically relevant pancreatic fistulas after distal pancreatectomy.

BACKGROUND: Various strategies to decrease postoperative pancreatic fistula after a distal pancreatectomy have proved unsuccessful. Because narcotics can cause spasm of the sphincter of Oddi and thereby increase pressure within the pancreatic duct stump, we hypothesized that increased narcotic use would be associated with increased occurrence of clinically relevant postoperative pancreatic fistula after distal pancreatectomy. METHODS: Retrospective analysis of consecutive distal pancreatectomies (2011-2016) was performed. Postoperative narcotic use was calculated in morphine equivalents. Postoperative pancreatic fistula was graded according to the International Study Group on Pancreatic Surgery. Perioperative variables were evaluated using multivariate logistic regression with clinically relevant postoperative pancreatic fistula as the dependent outcome. RESULTS: In the study, 310 distal pancreatectomies were analyzed (61% robotic, 20% open, 19% laparoscopic). Average age was 62 (53% female), and median total dose of morphine equivalents was 424 mg (interquartile range 242-768). Clinically relevant postoperative pancreatic fistula occurred in 21.6%. Clinically relevant postoperative pancreatic fistula and not clinically relevant postoperative pancreatic fistula cohorts were similar in most demographics and operative variables, but clinically relevant postoperative pancreatic fistula patients had fewer stapled transections (80 vs 90%, P=.025), less pancreatic cancers (11 vs 35%, P<.001), and greater median total morphine equivalents (577 vs 403 mg, P < .009). On univariate analysis, clinically relevant postoperative pancreatic fistula was associated with body mass index, nonstapled transection, suture ligation of the PD, a nonpancreatic cancer pathology, prophylactic octreotide, and total morphine equivalents >424 (cohort median). On multivariate analysis, only pancreatic cancer pathology was protective against a clinically relevant postoperative pancreatic fistula (odds ratio 0.24, confidence interval, 0.10-0.50, P=.001), while increasing total morphine equivalents were predictive of a clinically relevant postoperative pancreatic fistula (odds ratio 1.13, confidence interval, 1.01-1.27, P=.035) with a 13% increased risk for every approximate ≈100 mg increase in total morphine equivalents. CONCLUSION: In this retrospective analysis, postoperative narcotic use was associated with the development of clinically relevant postoperative pancreatic fistula after distal pancreatectomy. Limiting narcotic use may be one of the few available mitigating strategies against the development of a clinically relevant postoperative pancreatic fistula after distal pancreatectomy.