Aims: To evaluate the efficacy and safety of salvage therapy, and to identify risk factors of operative complications among hospitalized ulcerative colitis (UC) patients. Patients and Methods: We evaluated 88 UC patients hospitalized at our center between April 2010 and November 2012. We compared characteristics of corticosteroid-refractory patients treated with calcineurin inhibitor and those with infliximab as second-line therapy. Furthermore, we compared the characteristics of operative and nonoperative patients. The association between perioperative treatments and complications was also investigated. Results: Calcineurin inhibitor and infliximab were used in 42 and 22 patients, respectively. We found no difference in the clinical background between them. Efficacy rates were 67 and 50%, respectively. Eight out of 10 nonresponders of each treatment were treated with the other drug as third-line therapy. The efficacy rates of calcineurin inhibitor and infliximab as the third-line therapy were 75 and 50%, respectively. Operative patients had more severe disease (87.5 vs. 31%, p < 0.01), higher Lichtiger score (14.1 vs. 11.5, p < 0.01), higher Rachmilewitz endoscopic index (10.5 vs. 8.4, p < 0.01), higher C-reactive protein (7.6 vs. 4.0, p = 0.015) and lower serum albumin (3.1 vs. 3.6, p = 0.014) than nonoperative patients. Complications were observed in 7 out of 16 (44 %) operative patients. Postoperative complications were not increased even when patients were treated with second- or third-line therapy. However, the complication rate in corticosteroid users was 54.5 (6/11) and 20% (1/5) in nonusers. Conclusions: Third-line salvage therapy is effective and tolerable in carefully selected UC patients. Perioperative use of corticosteroids may lead to more adverse outcomes.