Stephen R. Grobmyer, David E. Rivadeneira, Clayton A. Goodman, Peter Mackrell, Michael D. Lieberman, John M. Daly

Background: Pancreatic anastomotic failure has historically been regarded as one of the most feared complications after pancreaticoduodenectomy. Methods: We reviewed our recent experience (59 cases), March 1994 to December 1998, with pancreaticoduodenectomy and compared preoperative and intraoperative characteristics as well as outcomes in those patients who experienced (n = 10) versus those who did not experience a postoperative pancreatic leak (n = 49). Information was retrospectively collected from hospital records, office records, and interviews with patients. Results: The clinical leak rate in this series was 8.5%. There were no significant differences in preoperative or intraoperative characteristics comparing those with versus those without a postoperative pancreatic leak. Only 1 of 10 patients with a postoperative pancreatic leak required reoperation to manage the leak. Those with a pancreatic leak had more other postoperative complications (median 2 versus 0 complications per patient, P = 0.01) and longer hospital duration compared with those without a leak (median 13 versus 23 days, P <0.01). Overall mortality in the series was 3.4%; no mortalities occurred as a result of a pancreatic leak. Conclusions: In the 1990s pancreatic anastomotic leak remains a potentially lethal problem after pancreaticoduodenectomy. Pancreatic leakage after pancreaticoduodenectomy is associated with other postoperative complications and a longer hospital stay.