Pancreatic anastomotic leakage after pancreaticoduodenectomy in 1,507 patients: a report from the Pancreatic Anastomotic Leak Study Group.

Several definitions for pancreatic leakage after pancreaticodoudenectomy exist, and the reported range of 2-50\% underscores this variation. The goal was to determine if drain data alone was predictive of a leak and validate International Study Group on Pancreatic Fistula (ISGPF) leak criteria. Participating surgeons entered de-identified data into a web-based database designed to collect Whipple-related data. Definitions used were the ISGPF definition, > or = 3 days, amylase 3x normal; and Sarr's definition, > or = 5 days, amylase 5x normal, > 30 ml. We compared how well these two definitions were at detecting a leak and its complications. There were 1,507 cases submitted from 16 international institutions. A pancreaticoduodenectomy (PPPD) was performed in 76.2\%. Drain placement occurred in 98.0\%. Using the ISGPF definition, the pancreatic leak rate was 26.7 and 14.3\% with the Sarr definition. There were more grades A and B leaks detected by the ISGPF definition. Both determined grade C leaks equally. Both definitions correlated with an increased length of stay (LOS), need for percutaneous drains, reoperation, and delayed gastric emptying (DGE). Neither was associated with an increased risk of intensive care unit (ICU) stay or 30-day mortality. The ISGPF was able to capture more patients with clinically relevant leaks than Sarr's criteria; however, the ability to detect a leak by drain data alone is imperfect.