The minimal intensity of oral anticoagulant required for antithrombotic protection in patients with a mechanical heart valve is still debatable, and that of the Westerner may not be directly applied to Thai patients. Our preliminary clinical review suggested that International Normalized Ratio (INR) 2-3 might be enough but it needs further supporting evidence. Therefore, we studied the effect of different anticoagulant intensities, expressed as INR, on the in vivo coagulation activation by measuring prothrombin fragment 1 + 2 (F1 + 2) in 116 patients with mechanical heart valve replacements. The patients had received warfarin for not less than one month with different intensities. The mean +/- S.D. of F1 + 2 level in 30 normal controls was 0.7 +/- 0.17 nmol/L. After excluding two outliers, the maximum linear correlation between INR and F1 + 2 was -0.658 (p < 0.001) when only patients whose intensities were lower than INR3 were taken into account. Adding more data from the patients having higher intensities decreased the correlation coefficient. The patients were subsequently classified by INR values in the range INR 1.1-1.9, 2-3 and 3.1-4.2. The F1 + 2 in each group was 0.6 +/- 0.30, 0.28 +/- 0.13 and 0.24 +/- 0.13 nmol/L respectively. The F1 + 2 in the first group did not differ from normal (p = 0.119) but was higher than the others (p = 0.000). The latter two groups had no difference between them (p = 0.112). Hence, from the laboratory point of view, we did not see additional benefit in the reduction of thrombin activation by the anticoagulant intensities higher than the range INR 2-3. The evidence supported that this therapeutic range might be enough for Thai patients with mechanical heart valves.