Introduction: Pancreatic cancer has a poor prognosis. Complete surgical resection is the only therapy to offer a chance for long-term survival. While the morbidity and mortality of pancreatic resection has fallen dramatically over the last 20 years, long-term survival rates in most series have remained stagnant. The aim of this study is to identify the actual 3- and 5-year survival rate after resection of pancreatic cancer and identify clinicopathologic or operative variables that influence survival. Methods: Between 1981 and 2002, 227 consecutive patients had resection with curative intent for pancreatic adenocarcinoma. There were 188 (83%) pancreaticoduodenectomies, 27 (12%) distal resections, and 12 (5%) total pancreatectomies. Operative morbidity was 42% and 30-day mortality was 6%. Actual 3-year survival rate was 19% (N=43), and actual 5-year survival rate was 5% (N=11). We compared 21 patients (9%) [15 m, 6 f, mean age 62 + 9] who died from early tumor recurrence (< 12 mos.) with 34 patients (15%) [16 m, 18 f, mean age 61 + 11 years] who survived at least 3 years postoperatively for clinicopathologic and operative variables. Results: There were no differences in tumor size, stage, vascular invasion, operative time, or blood transfusions between groups. Multivariate analysis using logistic regression identified poor tumor differentiation [O.R. 9.4, 95% C.I. = 1.2 - 73.0, P=0.03], extent of resection [O.R. 35.5, 95%C.I. = 2.8- 449.0, P=0.006] and blood loss [O.R. 1.002, 95% C.I. =1.000-1.003, P=0.04] as significant predictors of early recurrence. Conclusion: Survival after resection with curative intent in patients with pancreatic carcinoma has an actual 19% 3-year and 5% 5-year survival in this series. Of the variables identified as independent predictors of early tumor recurrence, a positive surgical margin (R1 resection) and operative blood loss are factors which can be impacted on by surgeons.