PURPOSE: Most studies of oral anticoagulant-related bleeding have analyzed the incidence of adverse outcomes among patients with a variety of different conditions and without any comparison with a control group. We determined the incidence, time course, and risk factors associated with major bleeding after hospital discharge among patients with deep-vein thrombosis, and estimated the excess risk of bleeding associated with oral anticoagulant therapy.

METHODS: A total of 22,000 adults were hospitalized in California for 3 or more days with a diagnosis of deep-venous thrombosis between January 1, 1992, and September 30, 1994. We determined the risk factors associated with readmission for bleeding. We compared the incidence of readmission for bleeding with comparison cohorts of patients with pneumonia or cellulitis who were matched for age, gender, race, and length of hospital stay.

RESULTS: Of 21,250 patients with deep-venous thrombosis who were discharged without bleeding, 1.4% were readmitted for bleeding within 91 days; the rate was 2.7 times greater in the first 30 days than in the next 61 days. Risk factors for bleeding included hospitalization with gastrointestinal bleeding during the previous 18 months (relative hazard [RH] = 2.6, 95% confidence interval [CI]: 1.6 to 4.1), hospitalization with an alcohol-related diagnosis during the previous 18 months (RH = 2.6, 95% CI: 1.4 to 4.8), chronic renal disease (RH = 2.4, 95% CI: 1.4 to 4.2), female gender (RH = 1.7, 95% CI: 1.3 to 2.2), presence of a malignancy (RH = 1.6, 95% CI: 1.2 to 2.2), nonwhite race (RH = 1.6, 95% CI: 1.2 to 2.1), and age over 65 years (RH = 1.3, 95% CI: 1.0 to 1.7). Significantly more women (n = 40) had intracranial bleeding than men (n = 18, P = 0.02). In the comparison cohorts, the incidence of readmission for bleeding within 3 months of discharge was 0.7%, and the relative risk (RR) of readmission was greater in those with deep-venous thrombosis than in those with cellulitis (RR = 2.0, 95% CI: 1.6 to 2.5) or pneumonia (RR = 2.0, 95% CI: 1.7 to 2.5).

CONCLUSIONS: The incidence of rehospitalization for bleeding was greatest in the first 30 days after discharge, and was approximately twice that seen in patients hospitalized for cellulitis or pneumonia. Further studies are needed to determine why women and nonwhite patients are at increased risk for anticoagulant-related bleeding.