A controlled randomized multicenter trial of pancreatogastrostomy or pancreatojejunostomy after pancreatoduodenectomy

Abstract

Background Only 2 large (more than 100 patients) prospective trials comparing pancreatogastrostomy (PG) with pancreatojejunostomy (PJ) after pancreatoduodenectomy (PD) have been reported until now. One nonrandomized study showed that there were less pancreatic and digestive tract fistula with PG, whereas the other, a randomized trial from a single high-volume center, found no significant differences between the two techniques. Methods Single blind, controlled randomized, multicenter trial. The main endpoint was intra-abdominal complications (IACs). Results Of 149 randomized patients, 81 underwent PG and 68 PJ. No significant difference was found between the two groups concerning pre- or intraoperative patient characteristics. The rate of patients with one or more IACs was 34% in each group. Twenty-seven patients sustained a pancreatoenteric fistula (18%), 13 in PG (16%; 95% confidence interval [CI] 8–24%) and 14 in PJ (20%; 95% CI 10.5–29.5%). No statistically significant difference was found between the 2 groups concerning the mortality rate (11% overall), the rate of reoperations and/or postoperative interventional radiology drainages (23%), or the length of hospital stay (median 20.5 days). Univariate analysis found the following risk factors: (1) age ≥70 years old, (2) extrapancreatic disease, (3) normal consistency of pancreas, (4) diameter of main pancreatic duct <3 mm, (5) duration of operation >6 hours, and (6) a center effect. Significantly more IAC, pancreatoenteric fistula, and deaths occurred in one center (that included the most patients) ( P = .05), but there were significantly more high-risk patients in this center (normal pancreas consistency, extrapancreatic pathology, small pancreatic duct, higher transfusion requirements, and duration of operation >6 hours) compared with the other centers. In multivariate analysis, the center effect disappeared. Independent risk factors included duration of operation >6 hours for IAC and for pancreatoenteric fistula ( P = .01), extrapancreatic disease for pancreatoenteric fistulas ( P < .04), and age ≥70 years for mortality ( P < .02). Conclusions The type of pancreatoenteric anastomosis (PJ or PG) after PD does not significantly influence the rate of patients with one or more IAC and/or pancreatic fistula or the severity of complications.