Population-based analysis of treatment of pancreatic cancer and Whipple resection: Department of Defense hospitals, 1989-1994

Background. The influence of hospital experience and referral patterns on the operative mortality rate of pancreaticoduodenectomy was studied in a worldwide hospital system. Methods. We analyzed computerized data on pancreatic cancer patients from U.S. Department of Defense (DOD) hospitals from 1989 to 1994. Results. Six hundred ninety-eight patients had pancreatic cancer, and 130 Whipple operations (105 for pancreatic and 25 for other cancers) were performed with an 8.5% 30-day operative mortality rate. Although most resections were done in teaching hospitals performing more than 1 Whipple procedure/per year, their results were not superior to smaller, lower volume, nonteaching hospitals. Patients transported for resection were younger than patients undergoing resection at their local DOD hospital but had similar outcomes. The operative mortality rate was higher after unusual resections and with increasing age; the tumor stage had no effect. Unresected patients undergoing combined radiation and chemotherapy had the longest survival times. Radiation therapy was associated with significantly longer survival times in patients without distant metastases, but chemotherapy was associated with a longer survival time when metastases were present. Conclusions. This mortality rate 8.5% for Whipple resections matches that from other large populations. Equivalent results were obtained in DOD teaching hospitals and smaller, community-type institutions. Because the DOD medical system minimizes financial and logistic barriers to transfer, the even distribution of DOD pancreatectomy mortality suggests that these barriers may favorably influence single institutional outcomes.