Background and Purpose The prothrombin time (expressed as the international normalized ratio [INR]) is the standard method of monitoring warfarin therapy in patients with atrial fibrillation. Prothrombin activation fragment F1.2 provides an index of in vivo thrombin generation and might provide a better index of the effective intensity of anticoagulation. We examined the relationship between F1.2 and INR in patients with atrial fibrillation. Methods We measured INR and F1.2 levels in 846 patients with atrial fibrillation participating in the Stroke Prevention in Atrial Fibrillation III study. Two hundred nineteen (26%) were taking aspirin alone, 326 (39%) were taking adjusted-dose warfarin, and 301 (36%) were taking a low fixed dose of warfarin (1 to 3 mg) plus aspirin (combination therapy). F1.2 levels were measured with an enzyme-linked immunosorbent assay. Results Patients receiving adjusted-dose warfarin or combination therapy had significantly higher INR and significantly lower F1.2 values than those on aspirin alone ( P ≤.0001 for each of the four comparisons). F1.2 values (nanomolar) were inversely correlated with INR (F1.2=−0.1+2.3[1/INR]; R ² =.37; P <.0001; simple linear regression). However, significant variability remained. Among patients receiving warfarin, older patients had higher F1.2 values than younger patients after adjustment for INR intensity ( P <.001) in the model. There was no difference in the relationship between F1.2 and INR between men and women. Conclusions Increasing intensity of anticoagulation, as measured by the INR, is associated with decreasing thrombin generation as measured by the F1.2 level, but significant variability exists in this relationship. Older anticoagulated patients have higher F1.2 values than younger patients at equivalent INR values. The clinical significance of these differences is not clear. F1.2 measurement might provide information regarding anticoagulation intensity in addition to that reflected by the INR.