BACKGROUND The treatment of severe ulcerative colitis (UC) flares includes measures such as hospitalization and intravenous steroids. Despite this, a quarter of patients are refractory to treatment. Given the availability of new therapeutic strategies in patients with steroid-refractory UC (cyclosporine, infliximab, apheresis, surgery) it is necessary to predict which treatment will be most effective for each patient. OBJECTIVES To determine which clinical or biological factors discriminate the lack of response to cyclosporine in steroid-refractory UC. METHODS Forty one flares of steroid-refractory UC in 35 patients treated with intravenous cyclosporine have been included. The response to cyclosporine was assessed at day 10 of treatment by using the modified Truelove and Witts disease activity score. Variables with prognostic significance were determined by a univariate analysis comparing groups with complete response and no-response, and an analysis of multiple linear regression. RESULTS Complete response was obtained in 41 flares (48%), partial response in 22%, and lack of response in 29%. The univariate analysis showed a significant difference in four predictive factors: higher age (p = 0.008), thrombocytosis (p = 0.01), disease extent (pancolitis vs. left-sided disease (p = 0.04)), and having received cyclosporine previously (p = 0.01). A multiple linear regression analysis confirmed the significance of higher age, thrombocytosis, and having received cyclosporine previously as predictive factors of poor response. CONCLUSION Higher age, thrombocytosis and previous use of cyclosporine predispose to poor response to intravenous cyclosporine in severe flares of steroid-refractory UC.