Abstract
BACKGROUND: Pancreaticoduodenectomy (Whipple's procedure) is a formidable procedure when undertaken for severe pancreaticoduodenal injury. The purposes of this study were to review our experience with this procedure for trauma; to classify injury grades for both pancreatic and duodenal injuries in patients undergoing pancreaticoduodenectomy according to the American Association for the Surgery of Trauma-Organ Injury Scale for pancreatic and duodenal injury; and to validate existing indications for performance of this procedure. STUDY DESIGN: We performed a retrospective 126-month study (May 1992 to December 2002) of all patients admitted with proven complex pancreaticoduodenal injuries requiring pancreaticoduodenectomy. RESULTS: Eighteen patients were included; mean age was 32 ± 12 years (SD), mean Revised Trauma Score was 6.84 ± 2.13 (SD), and mean Injury Severity Score was 27 ± 8 (SD). There were 17 penetrating injuries (94%) and 1 blunt injury (6%). One of 18 patients had an emergency department thoracotomy and died (100% mortality); 5 of the remaining 17 patients required operating room thoracotomies, and only 1 survived (80% mortality). There was 1 AAST-OIS pancreas grade IV injury, and there were 17 pancreas grade V injuries and 18 AAST-OIS duodenum grade V injuries. Indications for pancreaticoduodenectomy were: massive uncontrollable retropancreatic hemorrhage, 13 patients (72%); massive unreconstructable injury to the head of the pancreas/main pancreatic duct and intrapancreatic portion/distal common bile duct, 18 patients (100%); and massive unreconstructable injury, 18 patients (100%). Mean estimated blood loss was 6,888 ± 7,866 mL, and overall survival was 67% (12 of 18 patients). CONCLUSIONS: Complex pancreaticoduodenal injuries requiring pancreaticoduodenectomy (Whipple's procedure) are uncommon but highly lethal; virtually all are classified as AAST-OIS grade V for both pancreas and duodenum. Current indications for performance of pancreaticoduodenectomy are valid and should be strictly applied during procedure selection.