Background/Aims: Several earlier studies on factors predicting the long-term outcome of ulcerative colitis only encompassed treatment failure for one severe episode, or suffered from a lack of multivariate analyses. We aimed to identify factors assessable at diagnosis or after the first induction therapy which predicted relapse or later colectomy in patients with mild to severe ulcerative colitis. Methods: Clinical parameters (age, sex, disease extent, and disease activity at diagnosis) and laboratory data (hemoglobin, albumin, C-reactive protein, and erythrocyte sedimentation rate at diagnosis and 4 weeks after the first induction therapy) were evaluated in 296 patients (median follow-up 87 months). Factors predicting relapse and later colectomy were sought using the Cox proportional hazard model. Results: The presence of moderate or severe disease at diagnosis were significant predictors of relapse [adjusted hazard ratio (95% CI) 2.07 (1.48–2.89) and 1.70 (1.06–2.72), respectively] and later colectomy [3.40 (1.09–10.54) and 6.77 (1.92–23.86)]. After the first induction therapy, hemoglobin and albumin were associated with relapse [0.87 (0.76–0.99) and 0.58 (0.41–0.83)] and later colectomy [0.60 (0.47–0.77) and 0.11 (0.06–0.22)]. Conclusion: Relapse and later colectomy were associated with (1) disease activity at diagnosis and (2) lower levels of hemoglobin and albumin after the first induction therapy.