Comparison of the native prothrombin antigen and the prothrombin time for monitoring oral anticoagulant therapy

We have measured the fully carboxylated (native) prothrombin antigen and the undercarboxylated (abnormal) prothrombin antigen in patients treated with sodium warfarin using specific immunoassays to evaluate a new approach for monitoring oral anticoagulant therapy. Plasma and serum samples (391) were assayed for the prothrombin time, native prothrombin antigen, and abnormal prothrombin antigen. The results were correlated with the presence of bleeding or thromboembolic complications at the time of phlebotomy. The native prothrombin antigen correlated with the occurrence of complications in 95% of samples. Of 13 samples from patients with bleeding complications, 13/13 (100%) had a native prothrombin of 12 micrograms/mL or lower. Of seven samples from patients with thromboembolic complications, 6/7 (86%) had a native prothrombin of 24 micrograms/mL or greater. By comparison, a prothrombin time index of 1.5 to 2.5, 1.5 to 2.2, 1.5 to 2.0, or 1.3 to 1.8 identified 6/20 (30%), 9/20 (45%), 11/20 (55%), or 12/20 (60%) patients at risk, respectively. Although the prothrombin time index did correlate with the presence of bleeding complications, the native prothrombin antigen correlated closely with the presence of bleeding and thromboembolic complications. According to these results, the native prothrombin antigen, maintained in a range of 12 to 24 micrograms/mL by regular adjustment of the warfarin dosage, may be associated with a reduced risk of complications due to excessive or insufficient warfarin therapy. On the basis of these preliminary data, we recommend that the native prothrombin antigen be considered to monitor warfarin therapy.