Back to 2011 Program

Background: Distal pancreatectomy is a simple surgical procedure with a low mortality rate, but morbidity associated with pancreatic fistula remains a problem. This study sought to identify the risk factors for clinical pancreatic fistula, as defined by the International Study Group of Pancreatic Surgery (ISGPS). Methods: The medical records of all patients who underwent distal pancreatectomy at five Japanese institutions between January 2001 and June 2009 were retrospectively reviewed. All relevant anonymized data from patients entered into electronic case report forms were synthesized into a common database Data obtained for each patient included the following: age, sex, body mass index, diabetes mellitus, American Society of Anesthesiologists classification, previous laparotomy, primary disease surgical procedures, intraoperative bleeding, operation time, hospital stay, postoperative complications, and mortality. Pancreatic fistula was assigned according to the ISGPS classification into 4 categories: no fistula, Grade A, B, and C fistula. The primary endpoint was the occurrence of clinical pancreatic fistula Grade B or C. Results: Of the 388 patients studied, 226 were male and 162 were women, with a median age of 65 years (range 13-85 years). Pancreatic ductal carcinoma was present in 169 patients (51.0 %). Management of the pancreatic remnant was varied: stapler suture was used in 224 patients, hand-sewn closure was employed in 43 patients (11.1%), and no suture with duct ligation was performed in 118 patients (30.4%). Overall incidence of any pancreatic fistula (clinical or biochemical) was 58% (n=225). Grade A, B and C pancreatic fistula occurred in 95 patients (24.5%), 128 patients (33.0%), and 2 patients (0.5%), respectively. There was no operative mortality and no in-hospital deaths, but 4 patients (1.0%) required reoperation. The distribution of grades according to the Clavien-Dindo classification was as follows: 100 patients (25.8%) grade 0; 125 (32.2%) grade 1; 156 (40.2%) grade 2; 5 (0.1%) grade 3a; 1 (0.3%) grade 3b; and 1 (0.3%) grade 4a. There was no grade 4b or 5 complications. In multivariate analysis, diabetes mellitus, method of stump closure, and duration of operation were found to be independently associated with clinical pancreatic fistula, with respective hazard ratios (95% confidence interval) of 3.55 (1.83-6.85: p<0.001), 0.32(0.199-0.53: p<0.001), and 0.32 (0.19-0.52: p<0.001).Conclusions: Pancreatic leak remains a frequent complication after distal pancreatectomy, although severe complications (≥grade III) rarely occur. Staple closure of the pancreatic remnant is associated with a significantly lower clinical fistula rate.