Immediate operative repair is the most widely practised therapy for duodenal perforations and would seem to be the only conceivable course to recommend for most patients. However, most surgeons have encountered patients who refuse operations for their perforated ulcers and still recover. At operation the perforation has sometimes already been sealed off by omentum or adjacent organs and has to be reopened before it can be repaired. Indeed, the fact that non-operative treatment for perforated peptic ulcers may be successful has been recognized since 1870. In 1964 Herman Taylor reported the conservative treatment of 256 patients with perforations: only 21 patients required surgery and the overall mortality rate was 11 per cent. Seely and Campbell reported seven deaths in 139 patients treated conservatively, a mortality rate of 5 per cent. These results are better than those of surgical repair at the time and are comparable to most present day series. A more recent series from Dublin confirmed that conservative treatment can have acceptable results. The only randomized controlled comparison between non-operative treatment and surgery was performed in Hong Kong. This study indicated that an initial period of expectant treatment does not increase the mortality or morbidity in patients with perforated ulcers. More than 70 per cent of patients in the expectant treatment group recovered without an operation.

Non-operative management should only be employed if an experienced surgeon interested in this form of treatment is willing to personally supervise the patient's progress. Once the diagnosis of a perforated ulcer is made a nasogastric tube should be inserted as early as possible to empty the stomach and to reduce the leakage of gastrointestinal contents. The success of conservative treatment of perforated ulcer depends on keeping the stomach empty by nasogastric suction. Leakage from the perforation is kept to a minimum so that the omentum and surrounding organs have a chance to seal the perforation. Much attention to detail on the part of the medical and nursing staff is necessary to ensure that the tube is properly positioned in the stomach and that the stomach is properly emptied. Intravenous fluid is administered at a rate depending upon the degree of dehydration to maintain a urine output of at least 30 ml/h. In elderly patients and those in shock, central venous pressure measurements allow a more accurate assessment of fluid replacement. Broad spectrum antibiotics are also administered. The patient is carefully monitored and should be examined at least 12-hourly by the surgeon who made the initial assessment. Improvement is indicated by decrease in the pulse rate, temperature, and abdominal tenderness and by an improvement in the general well being of the patient. The surgeon must be prepared to abandon conservative management and to undertake operative intervention without delay if the patient fails to improve. The majority, however, will dramatically improve within 12 to 24 h. Oral fluids may be started when all signs of peritonitis disappear and intestinal activity returns. Normal diet is resumed within a few days. The patient should be treated with H&sub2;-receptor blocking drugs. Upper gastrointestinal endoscopy is performed 6 weeks later: evidence of healed duodenal ulceration will be seen in the great majority of patients. A decision is then made, depending on the individual circumstances, to stop medical treatment or to continue long-term maintenance H&sub2; blockade.

An upper gastrointestinal series using water-soluble contrast material is a useful adjunct for the definitive diagnosis and treatment of these patients. The demonstration of an ulcer crater without leakage in a patient with a typical history and peritonitis is reassuring and provides circumstantial evidence that the cause of peritonitis is indeed a perforated ulcer. It also suggests that spontaneous sealing has probably occurred. Contrast meals fail to detect persistent leakage in about 10 per cent of patients. One should rely on the overall clinical picture rather than the result of the contrast meal. Free leakage of contrast into the peritoneal cavity indicates that the crisis is not yet over and a high degree of vigilance needs to be maintained. Leakage shown on the contrast meal by itself do not mandate immediate surgery. In our experience perforations seal without operation in about 60 per cent of patients in whom leakage is detected on the contrast meal.

The importance of close supervision by an experienced surgeon cannot be overemphasized. The definitive diagnosis of patients with peritonitis can be difficult; some patients may have conditions such as colonic perforation, for which early operation is essential. Six to 12 h after the onset of peritonitis, the patient enters the so-called ‘stage of delusion’, which may deceive the unwary. The patient feels subjectively better, the acute pain eases, and the board-like rigidity of the abdomen disappears. In such cases there may be continued leakage. Unless an experienced surgeon is prepared to examine and supervise these patients on admission and at short intervals thereafter, it is likely that any deterioration in the patient's condition will be detected too late to be reversed. In fact, an expectant policy is much more demanding on the surgeon's time, skill, and judgement than is an immediate operation. To undertake such treatment without the close involvement of a skilled surgeon committed to conservative treatment would be foolhardy.

It was hoped that conservative treatment may improve the outlook for frail elderly patients who may be too ill to withstand general anaesthesia. Unfortunately this is not so: perforations are less likely to seal spontaneously in elderly patients. The greater omentum is atrophic in the elderly and is presumably less effective in sealing perforations. In addition, the elderly patient withstands continued intra-abdominal sepsis poorly. Early laparotomy after adequate resuscitation offers the best chance of recovery in these high risk patients.

In patients with perforated peptic ulcers, an initial period of non-operative treatment with careful observation may be safely allowed except in the elderly. The use of such an observation period can obviate the need for emergency surgery in more than 70 per cent of patients. Definitive ulcer surgery may then be reserved for patients who have frequent relapses of ulcer disease and other complications while receiving H&sub2;-receptor blocker treatment.

Crofts TJ, Park KGM, Steele RJC, Chung SCS, Li AKC. A randomized trial of nonoperative treatment for perforated peptic ulcer. N Engl J Med 1989; 320: 970–3.

Keane TE, Dillon B, Afdhal NH, McCormack CJ. Conservative management of perforated duodenal ulcer. Br J Surg 1988; 75: 583–4.

Redwood TH. Two cases of perforation of the stomach; one recovery. Lancet 1870; i: 647.

Seely SF, Campbell D. Nonoperative treatment of perforated peptic ulcer: a further report. Surg Gynecol Obstet 1956; 102: 435–6.

Taylor H. Aspiration treatment of perforated ulcer. Lancet 1951; i: 7–12.