Introduction. - In severe attacks of ulcerative colitis (UC) treated with intravenous corticosteroids, a fulminant colitis index (FCI) greater or equal to 8 has been associated with a greater likelihood of colectomy (72 vs 16% with an FCI < 8). This retrospective study aimed to assess the accuracy of such an association in infliximab-treated patients with moderate-to-severe bouts of UC.Patients and methods. - The study was based on the medical files of 43 patients who had received at least one infusion of infliximab to treat moderate-to-severe UC (partial Mayo Clinic score). Remission and clinical response were also assessed using the partial Mayo score. The accuracy of an FCI greater or equal to 8 to predict the likelihood of colectomy was assessed by calculating the sensitivity, specificity, positive and negative predictive values, Yule's Q coefficient, Youden's index and statistical significance (Chi(2) test).Results. - After treatment with infliximab, 10 patients were in remission (23.3%), 21 (48.8%) had a clinical response, four (9.3%) had treatment failure (without, however, requiring colectomy) and eight (18.6%) had a colectomy. Calculation of the above-mentioned indicators revealed that an FCI greater or equal to 8 was not an indicator of the risk of colectomy in this patient population, and found that only an FCI greater or equal to 16 was statistically significant. However, low values for sensitivity, positive predictive value and Youden's index preclude the clinical application of this latter result.Conclusion. - In patients treated with infliximab for moderate-to-severe UC attacks, the FCI is not a predictor of colectomy. In such patients, the factors predictive of a response to treatment or likelihood of colectomy, currently acknowledged with corticosteroid treatment, need to be further assessed for infliximab treatment. (C) 2010 Published by Elsevier Masson SAS.