Purpose: To determine the incidence of major bleeding in outpatients treated with warfarin and to identify predictive factors known at the start of therapy. Patients and Methods: The records of 565 patients starting outpatient therapy with warfarin upon discharge from a university hospital were reviewed. Follow-up information was obtained for 562 patients (99.5%). Bleeding was classified as major or minor using explicit criteria. The cumulative incidence of bleeding was estimated by means of survival analysis. Independent risk factors for major bleeding were identified using Cox regression analysis in 375 randomly chosen patients; they were tested in the remaining 187 patients. Results: Major bleeding occurred in 65 patients (12%) and was fatal in 10 patients (2%). The cumulative incidences of major bleeding at one, 12, and 48 months were 3%, 11%, and 22%, respectively. The monthly risk of major bleeding decreased over time, from 3% during the first month of outpatient therapy to 0.3% per month after the first year of therapy. Five independent risk factors for major bleeding-age 65 years or greater, history of stroke, history of gastrointestinal bleeding, a serious comorbid condition (recent myocardial infarction, renal insufficiency, or severe anemia), atrial fibrillation-predicted major bleeding in the testing group; the cumulative incidence of major bleeding at 48 months was 2% in 57 low-risk patients, 17% in 110 middle-risk patients, and 63% in 20 high-risk patients. Conclusions: These findings provide a quantitative basis for evaluating the risk of major bleeding in individual patients at the start of outpatient therapy with warfarin. Whether the risk of bleeding can be reduced in high-risk patients without reducing the benefit of therapy remains to be determined. © 1989 Reed Publishing USA.