Six hundred consecutive patients were operated on between September 1978 and October 1982 for isolated aortic (n = 298), mitral (n = 215), or multiple valve replacement (n = 87) with the St. Jude Medical bileaflet prosthesis. Mean age of the 303 female and 297 male patients was 50.7 +/- 9.6 (range 12 to 83) years. All patients were followed up prospectively; follow-up was complete and averaged 122.2 +/- 1.1 months for operative survivors. Total follow-up for aortic patients was 2904.1 patient-years, for mitral replacement 1859.5 patient-years, and for multiple valve replacement 736 patient-years. When the prothrombin times measured with different thromboplastins were converted into an international normalized ratio, four patient groups could be separated; that is, the groups comprised patients whose anticoagulation was maintained during the follow-up within an international normalized ratio corridor of 4.0 to 6.0, 3.0 to 4.5, 2.5 to 3.5, or 1.75 to 2.75. Less intensive anticoagulation in terms of the international normalized ratio values caused only a mild increase in the incidence of thromboembolic complications but a highly significant decrease in the rate of bleeding. Severe bleeding complications in the aortic valve group were highest with an international normalized ratio of 4.0 to 6.0 (1.15 per patient-year) and lowest with an international normalized ratio of 1.75 to 2.75 (0.24 per patient-year). The same held true for patients with single St. Jude Medical mitral valve replacement (2.09 per patient-year versus 0.72 per patient-year) and multiple valve replacements (4.45 per patient-year versus 1.20 per patient-year). These results suggest that the generally recommended international normalized ratio of 3.0 to 4.5 may be too high for patients with St. Jude Medical aortic valve replacement and also for patients with St. Jude Medical prostheses in the mitral position if, with respect to the thromboembolic hazard, there is not a predominating patient-related comorbidity. A large multicenter prospective randomized study is therefore proposed to establish the safe international normalized ratio levels accompanied by the lowest complication rates for both bleeding and thromboembolic events after St. Jude Medical prosthesis implantation (German experience with low intensity anticoagulation study).