Objective: To evaluate changes in the epidemiological features of Clostridium difficile colitis in hospitalized patients in the United States (C difficile is a common cause of nosocomial diarrhea that has been shown to be increasing in virulence in Canada and across Europe). Design: Cohort analysis of all patients with C difficile colitis in the Nationwide Inpatient Sample. Setting: Population-based data from the Nationwide Inpatient Sample, a 20% stratified random sample of US hospital discharge abstracts from January 1, 1993, through December 31, 2003. Patients: Using standard International Classification of Diseases, Ninth Revision (ICD-9) diagnostic codes, we identified patients with C difficile colitis. We controlled for comorbid conditions by calculating the Deyo modification of the Charlson score. To determine the relationship of year of diagnosis on main outcome measures, we constructed multivariate models. Main Outcome Measures: The prevalence, case fatality, total mortality rate, and colectomy rate of C difficile colitis. Results: We found that the prevalence, case fatality, total mortality rate, and colectomy rate of C difficile colitis increased from 1993 through 2003. In our regression analysis, the year of diagnosis predicted an increase in prevalence, case fatality, total mortality rate, and colectomy rate after adjusting for potential confounders. Conclusions: The prevalence, case fatality, total mortality rate, and colectomy rate of C difficile colitis significantly increased from 1993 to 2003. These findings provide compelling evidence of the changing epidemiological features of C difficile colitis.