Objective: The purpose of this study was to assess the efficacy and safety of duloxetine, a selective inhibitor of serotonin and norepinephrine reuptake, in the treatment of stress urinary incontinence. Study Design: A double-blind, randomized, placebo-controlled study was conducted in 553 women aged 18 to 65 years with a predominant symptom of stress urinary incontinence. Subjects were randomized to placebo (n = 138 women) or duloxetine at one of three doses (20 mg/d, n = 138 women; 40 mg/d, n = 137 women; or 80 mg/d, n = 140 women). Outcome variables that were assessed after 12 weeks of treatment included incontinence episode frequency recorded in a real-time diary and answers provided to the Patient Global Impression of Improvement scale and the Incontinence Quality of Life questionnaire. Results: Duloxetine was associated with significant and dose-dependent decreases in incontinence episode frequency that paralleled improvements that were observed in the Patient Global Impression of Improvement scale and the Incontinence Quality of Life questionnaire. The median incontinence episode frequency decrease with the use of the pooled diary analysis with placebo was 41% compared with 54% for duloxetine 20 mg per day (P =.06), 59% for duloxetine 40 mg per day (P =.002), and 64% for duloxetine 80 mg per day (P <.001). One half of the subjects at the 80 mg per day dose had a ≥64% reduction in incontinence episode frequency (P <.001 vs placebo); 67% had ≥50% reduction (P =.001 vs placebo). These improvements were observed despite significant concurrent dose-dependent increases in the average voiding interval in the duloxetine groups compared with the placebo group. Similar statistically significant improvements were demonstrated in a subgroup of 163 subjects who had more severe stress urinary incontinence (≥14 incontinence episode frequency per week; 49%-64% reduction in incontinence episode frequency in the duloxetine groups compared with 30% in the placebo group). Discontinuation rates for adverse events were 5% for placebo and 9%, 12%, and 15% for duloxetine 20, 40, and 80 mg per day, respectively (P =.04). Nausea was the most common symptom that led to discontinuation. None of the adverse events that were reported were considered to be clinically severe. Conclusion: This trial provides evidence for the efficacy and safety of duloxetine as a pharmacologic agent for the treatment of stress urinary incontinence. (Am J Obstet Gynecol 2002;187:40-8.)