Study objectives Due to their inherent thrombogenicity, mechanical cardiac valves necessitate lifelong oral anticoagulation. Less intensive oral anticoagulation than recommended earlier might result in a lower incidence of bleeding complications without increasing the embolic hazard significantly. Design Comparison of three different intensities of oral anticoagulation in a prospective, randomized multicenter design. Three months after valve replacement, patients were randomly assigned to stratum A, international normalized ratio (INR) 3.0 to 4.5; stratum B, INR 2.5 to 4.0; or stratum C, INR 2.0 to 3.5. Patients Data from 2,735 patients following aortic valve replacement (AVR; n = 2,024), mitral valve replacement (MVR; n = 553), and combined AVR and MVR (n = 158) with the St. Jude Medical (SJM) valve (St. Jude Medical; St. Paul, MN) between July 1993 and May 1999 were analyzed, covering a total follow-up period of 6,801 patient-years. All complications were registered prospectively. Measurements and results Fifty-one thromboembolic events (TEs) were documented, resulting in a linearized incidence of 0.75 TEs per 100 patient-years, 22 of which were minor (0.32% per patient-year), 10 were moderate (0.15% per patient-year), and 19 were severe (0.28% per patient-year). Thromboembolism following AVR was significantly lower than after MVR (0.53% per patient-year vs 1.64% per patient-year). Patients reported 1,687 bleeding complications (24.8% per patient-year). The vast majority of bleeding complications (n = 1,509; 22.2% per patient-year) were classified as minor, 140 were classified as moderate (2.06% per patient-year), and 38 were classified as severe (0.56% per patient-year). The clinically relevant incidences of moderate and severe TEs and bleeding complications were not significantly different between the three prespecified INR strata. Conclusions The intention-to-treat analysis of the results of the German Experience With Low Intensity Anticoagulation study leads to the unexpected result that despite a sophisticated reporting system, the incidence of moderate and severe TE and bleeding complications was comparably low in all INR strata and more or less within the so-called background incidence reported for an age-related “normal” population. This study supports reexamination of the intensity of anticoagulation in patients with the SJM valve. Due to their inherent thrombogenicity, mechanical cardiac valves necessitate lifelong oral anticoagulation. Less intensive oral anticoagulation than recommended earlier might result in a lower incidence of bleeding complications without increasing the embolic hazard significantly. Comparison of three different intensities of oral anticoagulation in a prospective, randomized multicenter design. Three months after valve replacement, patients were randomly assigned to stratum A, international normalized ratio (INR) 3.0 to 4.5; stratum B, INR 2.5 to 4.0; or stratum C, INR 2.0 to 3.5. Data from 2,735 patients following aortic valve replacement (AVR; n = 2,024), mitral valve replacement (MVR; n = 553), and combined AVR and MVR (n = 158) with the St. Jude Medical (SJM) valve (St. Jude Medical; St. Paul, MN) between July 1993 and May 1999 were analyzed, covering a total follow-up period of 6,801 patient-years. All complications were registered prospectively. Fifty-one thromboembolic events (TEs) were documented, resulting in a linearized incidence of 0.75 TEs per 100 patient-years, 22 of which were minor (0.32% per patient-year), 10 were moderate (0.15% per patient-year), and 19 were severe (0.28% per patient-year). Thromboembolism following AVR was significantly lower than after MVR (0.53% per patient-year vs 1.64% per patient-year). Patients reported 1,687 bleeding complications (24.8% per patient-year). The vast majority of bleeding complications (n = 1,509; 22.2% per patient-year) were classified as minor, 140 were classified as moderate (2.06% per patient-year), and 38 were classified as severe (0.56% per patient-year). The clinically relevant incidences of moderate and severe TEs and bleeding complications were not significantly different between the three prespecified INR strata. The intention-to-treat analysis of the results of the German Experience With Low Intensity Anticoagulation study leads to the unexpected result that despite a sophisticated reporting system, the incidence of moderate and severe TE and bleeding complications was comparably low in all INR strata and more or less within the so-called background incidence reported for an age-related “normal” population. This study supports reexamination of the intensity of anticoagulation in patients with the SJM valve.