A series of anamnestic, clinical, endoscopic and histologic parameters was evaluated in 64 patients with ulcerative colitis in remission to assess whether any of these had predictive value of relapse. All patients with quiescent disease were followed for at least 6 months. The statistical analysis was carried out by means of likelihood chi 2 test, receiver operating characteristic curve, Youden index and stepwise logistic regression. Eighteen patients out of 64 (28%) had a relapse, in a mean period of 9 months after the beginning of the study. The findings of our study suggest that only three factors were useful in selecting a subpopulation with a higher risk of recurrence: (a) a fiber-poor diet; (b) a number of previous episodes of relapse higher than 10, and (c) the presence of extraintestinal manifestations.