Background: The purpose of the present paper was to study the incidence, presentation and management of pancreaticoenteric anastomotic (PEA) leak following pancreaticoduodenectomy (PD) and to identify risk factors associated with PEA leak.
Methods: One hundred and twenty patients underwent PD for benign and malignant pancreatic and periampullary lesions from 1989 to 1998. Prospectively collected data were analysed for incidence and outcome of PEA leak. Four clinical, three laboratory parameters, preoperative biliary drainage (PBD), perioperative octreotide use, nine intraoperative parameters, site of tumour and stage of malignant tumours were analysed by univariate and multivariate logistic regression analysis to identify factors influencing PEA leak.
Results: Pancreatic leak developed in 15 (12.5%) patients. Nine patients (60%) had a PEA leak that manifested as controlled leak through the drain. All were managed conservatively and the leak stopped after a mean duration of 17 days (range: 6-32 days). Six (40%) patients had associated intra-abdominal complications, and three (50%) died in the postoperative period. Pancreatic fistula healed in the three remaining patients after a mean duration of 18 days (range: 15-25 days). Diabetes (P = 0.02; odds ratio (OR) = 4.60; 95% confidence interval (CI): 1.23-17.18), PBD (P = 0.03; OR = 4.82; 95% CI: 1.21-19.24), sequence of reconstruction (bilioenteric anastomosis as first anastomosis; P = 0.01; OR = 6.25; 95% CI: 1.45-26.83) and duration of surgery > 8 h (P = 0.01; OR: 5.61; 95% CI: 1.54-20.39) were associated with a significantly higher incidence of PEA leak.
Conclusion: Pancreaticoenteric anastomotic leak occurred in 12% of patients undergoing PD for pancreatic and periampullary tumours. The majority of these were uncomplicated and healed with conservative treatment. Complicated leaks were associated with high mortality. Diabetes mellitus, PBD, prolonged surgery and the sequence of reconstruction were risk factors associated with an increased incidence of PEA leak.