BACKGROUND Clostridium difficile infection (CDI) is a common cause of morbidity among hospitalized patients. Multiple factors have been associated with primary CDI, but risk factors for CDI relapses are less well described. METHODS This was a retrospective cohort study of patients with CDI over a 15-month period. We compared patients with relapsing and nonrelapsing CDI, including risk factors associated with primary CDI and other variables hypothesized to be associated with relapsing CDI and 90-day mortality. Multivariable logistic regression models were created to examine risk factors for relapse and 90-day mortality. RESULTS One hundred twenty-nine consecutive patients with CDI were included; 38 (29%) had relapsing CDI. Factors associated with relapsing CDI included fluoroquinolone use (71% versus 49%, P = 0.04) and incidence of stroke (29% versus 12%, P = 0.02). In a regression model, use of a fluoroquinolone was associated with relapsing CDI (OR = 2.52, 95% CI = 1.11-5.72). Factors associated with 90-day mortality included higher Charlson comorbidity index score (4.34 +/- 1.71 versus 3.42 +/- 2.08, P = 0.02), severe CDI (58% versus 32%, P = 0.01), and the use of piperacillin/tazobactam (45% versus 23%, P = 0.03) or meropenem (10% versus 1%, P = 0.04). In the regression analysis, 90-day mortality was associated with severe CDI (OR = 1.76; 95% CI = 1.19-2.59). CONCLUSIONS Fluoroquinolone use and prior stroke are associated with an increased risk of relapsing CDI. Relapsing CDI and severe CDI are both associated with increased 90-day mortality.