Endoscopic injection sclerosis in bleeding gastric varices.

Ninety-two consecutive, nonrandomized patients with bleeding varices were prospectively studied using sclerotherapy to control and prevent rebleeding. During this study, nine patients with gastric variceal bleeding were identified. A gastric variceal subset is defined and represents a 10% incidence in this series. All patients presented with indexed gastric bleeding varices that subsequently accounted for 34 bleeding sessions. The units of blood per rebleeding episode, hospital days, cost, and outcome were markedly different from the esophageal variceal groups. Initial management of indexed bleeding episodes by sclerotherapy and Sengstaken-Blakemore tubes were comparable; however, the number of rebleeding episodes was much higher. There was poor control of rebleeding with an associated higher rebleeding mortality and complications secondary to repeated sclerotherapy and Sengstaken-Blakemore tube use. In 37% of the patients, rebleeding was the direct result of gastric ulcerations at the endoscopic injection sclerotherapy site. The survival curve of this group was much lower than esophageal variceal bleeders. Endoscopic injection sclerotherapy in patients with bleeding gastric varices offers only temporary control of bleeding, and the high incidence of severe early rebleeding requires consideration of alternative methods for management or modified sclerotherapy techniques.