Delineation of critical factors in the treatment of pancreatic trauma.

An analysis of 100 patients sustaining multiple injury and pancreatic trauma was completed. Sixteen patients with penetrating injury died within the first 24 hours, 14 of whom died intraoperatively from major hepatic and/or retroperitoneal venous injury. Eighty-four patients survived long enough to permit evaluation of treatment. There was no statistically significant relationship between mode (p = 0.3) or anatomic area (p = 0.5) of injury and death. However, death was more common in the presence of duct injury (p less than 0.0001). Thirty-nine patients were determined to have duct injury and 45 did not. These two groups were equivalent, with the exception of a higher incidence of concomitant bowel injury (p less than 0.05) in those with duct violation. Combined sump and Penrose drainage was found to be adequate treatment of both proximal and distal nonductal injury with no significant difference in mortality or morbidity rates (p = 0.5). Resection of distal ductal injuries as opposed to drainage alone resulted in significantly lower morbidity and mortality rates (p less than 0.05), comparable to those of drained nonductal injuries. No conclusions could be made relevant to proximal duct injuries, except that drainage alone is inadequate. Seventeen (20 percent) of the 84 patients evaluated died. Pancreatic related mortality rate was 17 percent (14 patients). Two of 23 patients with blunt injury (9 percent) and 12 of 61 patients with penetrating injury (20 percent died). Gram-negative sepsis (82 percent) was the most common cause of death (p less than 0.01), and sepsis was correlated with the presence of pancreatic duct (p less than 0.0001) and bowel (p less than 0.001) injury.