OBJECTIVE To assess the effect of aging on the anticoagulant response to warfarin. METHODS Retrospective cohort study. METHODS A university hospital outpatient anticoagulation clinic. METHODS All patients (n = 530) monitored in the anticoagulation clinic over a 10-year period (1980 to 1990). The 530 study patients had a mean age of 61.5 (+/- 14.7) years (age range, 12 to 90 years). The patients were stratified into four age groups: younger than 50 years (n = 97); 50 to 59 years (n = 107); 60 to 69 years (n = 149); and 70 years or older (n = 177). METHODS For each patient, a dose-adjusted mean prothrombin time ratio was calculated by dividing the mean prothrombin time ratio by the mean daily warfarin dose. RESULTS Older patients were more likely to be female (P less than 0.001), to have more medical problems (P less than 0.001), to be taking more medications (P less than 0.001), and to weigh less than younger patients (P less than 0.001). Across age groups, there were no significant differences in the use of medications that potentiated or inhibited the anticoagulant effects of warfarin. The prothrombin time ratio, when adjusted for dose, was significantly increased in older patients (P less than 0.001). The increased anticoagulant response to warfarin seen with increasing patient age persisted even after simultaneously controlling for relevant demographic and clinical variables in a multivariate model. Other factors significantly associated with an increased sensitivity to warfarin included use of a medication with a potentiating interactive effect with warfarin, female gender, and overall medication use. Increased body weight and duration of warfarin use exceeding 6 months were found to be inversely related to anticoagulant response. CONCLUSIONS The anticoagulant response to warfarin is exaggerated with advancing age. This finding emphasizes the need for close monitoring of older patients treated with warfarin therapy.