To the Editor: The work of Saour et al. on anticoagulation in patients with prosthetic heart valves (Feb. 15 issue)1 is to be commended for its attempt to address the problem of the appropriate intensity of anticoagulation in such patients. The authors concluded that “moderate-intensity” anticoagulation with coumadin (prothrombin-time ratio, 1.3 to 1.7) is associated with no greater frequency of thromboembolic events than “high-intensity” anticoagulation (prothrombin-time ratio, 2.3 to 2.7), but with fewer hemorrhagic complications. However, the data for thromboembolic events shown in Figure 1 of their paper suggest an entirely different conclusion. If the embolic data are analyzed in.