Bleeding esophagogastric varices: an endoscopic study.

Among 857 patients admitted from October 1977 to December 1984 with acute upper gastrointestinal hemorrhage, 165 (19.5%) had endoscopically proved esophageal or gastric varices. Among this group, varices were considered as the actual bleeding lesion in only 83 (9.7%). Stigmata of variceal bleeding were observed in 76 (92%) including 35 venous spurts, 12 venous oozes, 20 adherent clots, and 14 "platelet aggregates." In only seven cases, variceal rupture could only be presumed, because no other lesion was present. In the other 82 patients, a lesion other than varices was the bleeding source, mostly peptic ulcers (32 = 38%) and erosive gastritis (30 = 36%). One-third of the patients with bleeding varices required more than one endoscopy to provide evidence of variceal bleeding. The most frequent bleeding point was the cardia and a good correlation between variceal size and bleeding was observed. There was no relationship with evidence of gastroesophageal reflux.