Background: The usual end point for defining success of dilation is subjective (relief of dysphagia). In most patents thus managed strictures recur. We asked whether an objective end point would improve outcome. Methods: After dilation to 15 mm, patients were randomized into subjective and objective groups. In subjective group patients, end point for dilation was alleviation of dysphagia; in objective group patients, passing the 12 mm barium pill test. Objective group patients who failed underwent redilation until they passed the pill or failed three times. During Part 1 of the study, patients received ranitidine, during Part 2 they received omeprazole. Results: In part 1, dysphagia was alleviated in 7 of 8 subjective group patients. Only 2 of 10 objective group patients passed the pill test and no additional patients passed after 3 sessions, although most had no dysphagia. In Part 2, 19 subjective group and 15 objective group patients were studied. End point was not achieved in 3 objective group patients. Over long-term follow-up, objective group patients had less recurrent dysphagia ( p < 0.02) and required fewer redilation sessions (p < 0.05). Overall, the pill test correlated with the presence or absence of dysphagia (p < 0.001). Predictive value of passing the pill 1 week after dilation for the absence of dysphagia was 100%, but of failing the pill test and the presence of dysphagia was only 18%. Conclusions: Achieving an objective end point reduces stricture recurrence and the need for subsequent dilation. Initial subjective improvement does not predict long-term success. (Gastrointest Endosc 1997;45:354-9.)