Outcome of operations for upper gastrointestinal tract bleeding.

The outcome of operations for upper gastrointestinal tract bleeding during a six-year period was compared with that of the previous four years, in which indications for operation and guidelines for surgical procedures were similar. Between 1973 and 1978, 392 patients were hospitalized for hemorrhage. Endoscopy diagnosed a bleeding lesion in 92% of 234 patients studied. Of 75 patients (19%) who required operation for uncontrollable hemorrhage, 20 (27%) died and two (3%) rebled postoperatively. Most deaths (80%) were caused by esophageal variceal bleeding. Among 47 patients with nonvariceal hemorrhage, mortality was only 9%. No patient with stress ulcer bleeding was encountered. Compared with our 1969 to 1972 experience, the present study shows no improvement in overall mortality. Rebleeding was less frequent than earlier. The most significant differences in outcome were decreased mortality in alcoholic gastritis patients, no deaths from stress ulcer, and increased mortality after portosystemic shunting. Endoscopy, used frequently from 1973 to 1978, helped to improve preoperative diagnostic rates (85% vs 65%). Combined with innovations in nonoperative treatment, such as infusion of vasopressin, it did not appear to decrease the proportion of patients requiring operation.