Background: Mortality rates associated with pancreatic resection for cancer have steadily decreased with time, but improvements in long-term survival are less clear. This prospective study evaluated risk factors for survival after resection for pancreatic adenocarcinoma.
Methods: Data from 366 consecutive patients recorded prospectively between November 1993 and September 2001 were analysed using univariate and multivariate models.
Results: Fifty-eight patients (15.8 per cent) underwent surgical exploration only, 97 patients (26.5 per cent) underwent palliative bypass surgery and 211 patients (57.7 per cent) resection for pancreatic adenocarcinoma. Stage I disease was present in 9.0 per cent, stage II in 18.0 per cent, stage III in 68.7 per cent and stage IV in 4.3 per cent of patients who underwent resection. Resection was curative (R0) in 75.8 per cent of patients. Procedures included pylorus-preserving Whipple resection (41.2 per cent), classical Whipple resection (37.0 per cent), left pancreatic resection (13.7 per cent) and total pancreatectomy (8.1 per cent). The in-hospital mortality and cumulative morbidity rates were 2.8 and 44.1 per cent respectively. The overall actuarial 5-year survival rate was 19.8 per cent after resection. Survival was better after curative resection (R0) (24.2 per cent) and in lymph-node negative patients (31.6 per cent). A Cox proportional hazards survival analysis indicated that curative resection was the most powerful independent predictor of long-term survival.
Conclusion: Resection for pancreatic adenocarcinoma can be performed safely. The overall survival rate is determined by the radicality of resection. Patients deemed fit for surgery who have no radiological signs of distant metastasis should undergo surgical exploration. Resection should follow if there is a reasonable likelihood that an R0 resection can be obtained.
Copyright 2004 British Journal of Surgery Society Ltd.