SUMMARY The operation of total pancreatectomy is performed rarely. Its role in the management of patients with chronic pancreatitis remains to be elucidated. We have reviewed our series of 29 total pancreatectomies for benign disease [14 women median age 39 years; 15 men median age 34 years]. Twelve underwent standard total pancreatectomy, in 17 duodenum preserving total pancreatectomy (DPTP) was performed. There was one death (mortality 3.4%). In no patient was the total pancreatectomy the first operative procedure. The patients were compared with age and sex matched diabetic control subjects selected on a best fit basis from the diabetic clinic database. The aetliology of the pancreatitis was idiopathic nine, pancreas divisum nine, alcohol eight and other causes three. The indication for surgery was pain 27, acute pancreatitis one and cholangitis with pancreatitis one. The complications of the procedures were mainly caused by infection [wound three, chest six and central line sepsis four] and in two there was a leak from the duodenum; no patient required re-operation. The postoperative stay [standard total, median 21 days (range 13-98) DPTP median 31 days (range 17-49)] has lengthened over the period due to greater attention to analgesic, diabetic and enzyme deficiency control before discharge. In standard total pancreatectomy there were five major hypoglycaemic episodes with only two in 17 DPTP patients. The per cent ideal body weight, the insulin requirement and the HbAl compared less well in standard total pancreatectomy group compared with controls than did DPTP. With both groups large doses of enzyme replacement were required, and this proved of importance in diabetic control. Our experience with total pancreatectomy suggests that pain will be improved in over 80% of patients and that the results of surgery will improve with prolonged follow up provided attention is given to analgesic abuse, enzyme deficiency and diabetes.The role of surgery in the management of chronic pancreatitis is a subject of debate.' Most would agree that the place for surgery is clear when there is a complication of chronic pancreatitis amenable to surgical intervention. Unfortunately the problem often facing the clinician is that of a patient with chronic pain, on an increasing dosage and strength of analgesia, whose quality of life is destroyed by the effects of both pain and analgesia. There is a group of patients who fail to respond to, or who relapse after, non-surgical procedures such as nerve blocks and endoscopic sphincterotomies: of these, there are some whose symptoms are not relieved by pancreatic

The classical Whipple procedure for chronic pancreatitis has been associated with significant long term postoperative morbidity. The pylorus-preserving procedure of Longmire has reduced but not eliminated the long-term morbidity. Preservation of the whole duodenum with total pancreatectomy has been introduced for the treatment of patients with end-stage chronic pancreatitis after favourable experience with this procedure in infants for nesidioblastosis. Fourteen patients with chronic pancreatitis have had a total pancreatectomy with preservation of the duodenum and the bile duct. All patients are still alive (median follow-up 9.5 months) and none suffered major complications in the perioperative period. One patient developed a biliary stricture at 3 months, requiring biliary reconstruction. Six of the patients have returned to full-time work; nine require no analgesia. All patients require pancreatic enzyme replacement, and all patients have gained weight postoperatively. Diabetic control is satisfactory with a twice daily insulin regime. Duodenum-preserving total pancreatectomy is feasible in the adult without mortality or high morbidity; early experience suggests that preserving the duodenum improves gastrointestinal function with easier control of the diabetes.

Twenty-one patients had a concurrent splenectomy with resection of colorectal cancer between 1970 and 1988. These were matched individually with disease control patients based on age, sex, site of tumor, Dukes stage, tumor differentiation, and date of the operation. Significantly more patients in the splenectomy group (n = 11) developed postoperative infective complications than in the control group (n = 4) (McNemar test: P = 0.03). Five-year overall actuarial survival was 45 percent in the former group and 59 percent in the latter (log rank test: chi-squared = 1.07; P = 0.24). Similarly, five-year disease-free survival in 17 patients with Dukes B and C cancers who had curative resections did not differ between the groups (log rank test: chi-squared = 0.08; P > 0.25). These results suggest that splenectomy with resection of colorectal cancer increases the risk of postoperative sepsis and does not influence long-term survival. The infrequency of concurrent splenectomy at resection of colorectal cancer may not overcome Type II error.

A survey of six British centres collected data on 83 patients undergoing total pancreatectomy (TP) for chronic pancreatitis between 1977 and 1986. There were 57 men and 26 women with a median age of 38 years (range 19-61 years). Half were alcoholics and half had had previous acute pancreatitis. Besides jaundice (14 per cent) severe pain was the indication for the operation; regular opiates had been needed in 82 per cent of patients and 37 per cent were addicted to these drugs. All but 12 had had previous pancreatic or biliary surgery, with a median of two operations and a maximum of six. TP was a one-stage procedure in 32 patients, 42 had had distal resections and 9 proximal resections in the past; the pylorus was preserved in 30. Median operation time was 4 h (range 2-18 h) and median blood loss was 3 units (1-21 units). Intraoperative complications in 11 patients included haemorrhage in 9. Four deaths occurred within 30 days from bleeding (2), respiratory failure (1) and Roux-loop infarction (1). All but one of the 79 survivors required full pancreatic supplementation and 38 per cent had difficulties in endocrine control. At a median follow-up of 1.5 years (range 0.25-10 years), 57 patients (72 per cent) were pain-free and 9 (11 per cent) needed only occasional analgesia. Though 13 (17 per cent) still took regular analgesics, all were symptomatically improved. There have been 10 late deaths (13 per cent), all but one of which are attributable to the operation.

The optimum management of ingested button batteries was ascertained by postal