Background. The aim of this study was to determine prognostic factors for survival after pancreaticoduodenectomy (PD) for carcinoma of the pancreatic head region. Methods. From 1983 to 1992. 176 patients underwent PD for ampullary carcinoma (n = 67), distal bile duct carcinoma (n = 42), or pancreatic carcinoma (n = 67). The first choice for resection was subtotal PD (n = 146), but patients with a tumor‐positive pancreatic margin or a brittle pancreatic duct underwent total PD (n = 30). Results. Hospital mortality was 4.7% after subtotal PD and 20% after total PD. Overall 5‐year survival was 31%. Survival after PD for ampullary carcinoma care. (5‐year, 50%) was significantly better (P < 0.001) than for distal bile duct carcinoma (24%) and pancreatic carcinoma (14%). Independent negative prognostic factors for survival (multivariate analysis) were involved resection margins (hazard rate ratio [HRR] 4.08), major vascular involvement (HRR 2.20), distal bile duct or pancreatic origin of carcinoma (HRR 1.93), and perioperative blood transfusion of more than 4 U (HRR 1.76). Tumor size (>2 cm), regional lymph node involvement, and a poor differentiation grade were overall negative factors in univariate analysis but not in the subgroup of ampullary carcinoma. Conclusion. Involvement of resection margins, major vascular ingrowth, site of origin of carcinoma, and perioperative blood transfusion were independent prognostic factors for survival after PD. Overall 5‐year survival was 31%, and subtotal PD is advocated for all patients with a macroscopically resectable tumor in the pancreatic head region without major vascular involvement, even for those with larger tumors or local lymph node metastasis. Care should be taken to limit the need for perioperative blood transfusions.

No abstract

Of 176 patients with carcinoma of the pancreatic head region 156 underwent standard pancreatoduodenectomy (group 2) and 20 with macroscopic suspicion of invasion of the portal vein or superior mesenteric vein (SMV) underwent pancreatoduodenectomy with partial resection of the portal vein or SMV (group 1). In 16 patients in group 1 end-to-end anastomosis was used for reconstruction of the vein. The morbidity rate in groups 1 and 2 was similar (55 versus 63 per cent). The hospital mortality rate was 15 per cent in group 1 and 7 per cent in group 2 (P = 0.22). Histological examination confirmed tumour invasion of the portal vein or SMV in ten patients in group 1. Invasion of the portal vein or SMV was significantly more frequent in patients with pancreatic cancer than in those with distal bile duct or ampullary carcinoma. Of the 20 patients in group 1 only three underwent curative resection with tumour-free margins. The median survival time after resection of the portal vein or SMV was 8 months; the 2-year survival rate was 19 per cent. Comparison of survival in group 1 with survival in subgroups of patients undergoing standard pancreatoduodenectomy, matched for all histological parameters, showed no significant difference. It is concluded that partial resection of the portal vein or SMV in patients undergoing pancreatoduodenectomy who are suspected of having tumour invasion of the portal vein or SMV does not improve either the rate of curative resection or survival.

Some evidence for the effectiveness of using feedback from real patients to improve knowledge and behavior exists; however, before implementing patient feedback into training programs, educators and policy makers should realize that the evidence for effecting actual improvement in physicians' consulting skills is rather limited.