Pancreatic fistula rate after pancreatic resection. The importance of definitions

BACKGROUND: Pancreatic fistula (PF) is still regarded as a serious complication both in terms of frequency and sequelae. The incidence varies greatly in different reports because of the different definitions used. The aim of this study was to compare several definitions of PF encountered in the current literature and to demonstrate that the PF rate in the same group of patients treated in a high volume center is dependent upon the definition applied. METHODS: A Medline search of the last 10 years was performed as regards the definition of PF. A score was assigned to the reproducible definitions based upon two basic parameters: daily output (cm3) and duration of the fistula represented by the number of days between the postoperative day of onset and the duration of the complication. Four definitions were formulated and were then applied to a group of 242 patients that underwent pancreatic head or intermediate resections with pancreatico-jejunal anastomosis in our Pancreatic Unit between November 1996 and December 2000. Statistical analysis was carried out using the Yates correct chi2 test with statistical significance set at p < 0.05. RESULTS: Among 26 different definitions identified, 14 were found suitable for the applied score. We formulated four final definitions summarizing the current concepts of PF. The incidence of PF ranged between 9.9 and 28.5% according to the different definitions applied with highly statistical differences between them. CONCLUSIONS: The PF rate after pancreatic resections is strictly dependent upon the definition used. An overall general agreement for an internationally accepted definition is urgently needed to correctly compare different experiences.