BACKGROUND Small esophageal varices are difficult to inject intravariceally and complete eradication of varices may not be therapeutically necessary. METHODS A prospective study designed to compare the use of different end points of sclerotherapy ie (i) total eradication (grade 0) (ii) grade I varices or (iii) grade II white (IIw) varices, by stopping variceal injection when one of these was reached. The patients were followed endoscopically and clinically at 3-4 mo intervals. METHODS Variceal recurrence and rebleeding rates. RESULTS Of 266 patients with portal hypertension (147 cirrhosis, 51 non-cirrhotic portal fibrosis, 68 extra-hepatic portal venous obstruction), 232 (87.2%) reached one of the three end points--six to grade 0, 96 to grade I and 130 to grade IIw. Among 212 patients followed up for a mean post-sclerotherapy period of 12.2 mo (range 3-48 mo), variceal recurrence rates, time intervals to recurrence and rebleed rates were comparable in patients with endpoints of grade I and grade II (white) varices (28.5% vs 38.9%, 10.8 mo vs 9.2 mo, and 7.7% vs 7.8%, respectively; all p = ns). Overall, 51 patients had interval bleed before reaching the end point--50 (98%) of them bled from grade III and IV varices while only one patient (2%) bled with grade IIw varices. CONCLUSIONS Grade I or grade IIw varices should be acceptable as the end points for endoscopic sclerotherapy of esophageal varices.