Background: Recent studies have suggested that the mortality rate from pancreatic resection for cancer is high in the UK compared with that in published series. A survey of specialist units was conducted to determine whether the results differed from those in general units.
Methods: The postoperative outcome following resection of pancreatic and periampullary tumours was analysed from specialist units in the UK and compared with that of other multi-institutional and large single institutional studies published recently (1900-1996).
Results: A total of 1026 resections was reported from 21 units (33 surgeons). Postoperative complications necessitated reoperation in 57 patients (6 per cent) and there were 58 deaths (6 per cent) in hospital. Pylorus-preserving resections were performed in 102 (41 per cent) of 250 patients with ampullary adenocarcinoma undergoing a major right-sided resection and in 123 (32 per cent) of 381 patients with ductal adenocarcinoma of the head of the pancreas undergoing right-sided resection (chi 2 = 4.01, 1 d.f., 2P = 0.04). The mean number of resections for pancreatic ductal adenocarcinoma was 3.41 (range 1.0-7.1) per institution per year. Combining these data with those from the nine published series from specialist units, there was a lower mortality rate compared with the results of five published general surveys (median 4.9 per cent (95 per cent confidence interval (c.i.) 3.1-8.0 per cent) versus 9.8 (2.5-23.2 per cent), 2P < 0.01) and specialist units had a higher volume caseload (median 5.5 (95 per cent c.i. 4.2-8.1) versus 0.5 (-0.2-2.0) cases per institution per year, 2P < 0.001). Postoperative mortality was related to caseload both for the UK (chi 2 = 7.17, 1 d.f., P < 0.01) and for all the data combined (chi 2 = 40.4, 1 d.f., P < 0.0001).
Conclusion: The results from specialist units in the UK compare favourably with those from specialist units outside the UK and are superior to those from non-specialist units. The mortality rate is generally lower in units with a higher caseload.