Efficacy of selective splanchnic arteriography and vasopressin perfusion in diagnosis and treatment of gastrointestinal hemorrhage.

Vasopressin infusion initially controlled 80 per cent of patients bleeding from portal hypertension, and 53 per cent did not rebleed after removal of the catheter. This figure is significantly greater than the 28 per cent of patients totally controlled by esophageal tamponade (p less than 0.075). Similar rates of success were achieved by vasopressin infusion for gastric, duodenal, and colonic bleeding sites. These results suggest that visceral arterial infusion of vasopressin is the method of choice for the short-term therapeutic management of massive gastrointestinal bleeding from portal hypertension. Vasopressin infusion also appears to be a valuable means of treating patients with massive gastrointestinal bleeding secondary to shallow gastric ulcers, gastritis, Mallory-Weiss tears, colonic bleeding and "poor risk" patients with deep gastric, marginal, or duodenal ulcers when conventional medical therapy has failed. The presence of a coagulation abnormality in patients with portal hypertension significantly reduced the complete control of bleeding to only 27 per cent ( p less than 0.010) and survival rate to 14 per cent (p less than 0.050). Visceral arterial perfusion proved to be an effective means of arresting hemorrhage, but the overall improvement in hospital mortality in this group of poor risk patients remains unproved.