The aim of the study was to identify factors related to the onset of pancreatic fistula and to define the characteristics of the fistula. The study group was composed of 123 patients who underwent left pancreatectomy since 1996. Pancreatic closure was accomplished by a hand-sewn technique (39 patients) or two kinds of mechanical staplers: Proximate (Ethicon Endo-Surgery, Cincinnati, OH) (46 patients) and Endo-GIA (United States Surgical, Norwalk, CT) (38 patients). Fistula was defined as output greater than 5 ml, with amylase × 5, after day 5. In case of fistula, the drain removal was scheduled at a daily output less than 5 ml. Mortality was 0%, morbidity was 48%, and pancreatic fistula rate was 34%. Fistula rate was 38% after hand-sewn closure, 26% after Proximate, and 39% after Endo-GIA (NS). None of the other factors (separate duct ligation, hand-sewn suture in addition to stapler, spleen preservation, use of pledgetted suture, sex, age, and indication for pancreatectomy) proved to be related to a reduction in the onset of fistula. All fistulas healed spontaneously. Mean fistula duration was 36 days; 92.8% of patients with fistula were discharged with drain. The policy of delayed drain removal allowed a low rate of fistula associated morbidity (16%) and of readmission (4.7%). In conclusion, fistula is an unsolved problem of left pancreatectomy. However, a careful drain management allows a good outcome in patients with fistula.
ASJC Scopus subject areas