Demonstration of two distinct subsets of gastric varices. Observations during a seven-year study of endoscopic sclerotherapy.

Over a seven-year period, assessment of gastric varices was made on 225 patients receiving endoscopic sclerotherapy for variceal hemorrhage. Of 170 patients with complete data, gastric varices were observed in 26 (15.3%). Importantly, two distinct subsets of gastric varices were identified: varices distal to the gastroesophageal junction without extension into the fundus, termed "junctional varices," occurred in 11.2%, and varices that were confined only to the fundus, termed "fundal varices," occurred less frequently in 4.1%. Although rebleeding was increased in both subsets of gastric varices, junctional varices were more amenable to sclerotherapy. Patients with fundal varices (N = 7) had a significantly higher rebleeding rate, increased complications with sclerotherapy, and significantly decreased survival (P less than 0.005) when compared to patients with esophageal varices alone (N = 87) who were followed for more than three months. Cumulative survival was not significantly different (P less than 0.08) in patients with junctional varices (N = 19) when compared with patients with esophageal varices alone. We conclude that not all patients with gastric varices have a poor result with sclerotherapy. Recognition of these subsets may improve treatment strategies in patients with gastric varices.