Responses of prothrombin time and activated partial thromboplastin time to edoxaban in Japanese patients with non-valvular atrial fibrillation: characteristics of representative reagents in Japan (CVI ARO 7).

The aims of this study were to determine the distribution of plasma concentration of edoxaban (PC-Ed) with their 90% interval (on therapy range) and its correlation with anticoagulation markers in patients with non-valvular atrial fibrillation (NVAF). Consecutive 97 NVAF patients under edoxaban therapy were evaluated (60/30 mg dose, n = 48/49; men/women, n = 71/26; age, 69 years). CHADS2 score 0, 1, and ≥ 2 were 27%, 44%, and 29%, respectively. The mean (90% interval) of PC-Ed by LC-MS/MS was 194.3 (49.4-345.3) and 17.0 (4.8-40.7) ng/mL at peak (2-4 h post-dose) and trough (pre-dose), respectively. Correlation of prothrombin time (PT) with PC-Ed was higher than that of activated partial thromboplastin time (aPTT). Among 6 PT reagents, Coagupia PT-N and Simplastin Excel S (both PT reagents) showed the highest predictive capability for the upper outlier of PC-Ed at peak and trough. Among 4 aPTT reagents, only Thrombocheck APTT measured at peak had a significant predictive capability. When using PT reagents, both peak and trough sampling showed a similar predictive capability for the upper outliers of PC-Ed with a high sensitivity, but a relatively low specificity. We demonstrated the distributions of plasma concentration, PT with 6 reagents, and aPTT with 4 reagents under edoxaban therapy in Japanese patients with NVAF, showing their 90% intervals. For predicting the upper outlier of PC-Ed, PT was more sensitive compared with aPTT, whereas predicting capability for the outliers of PC-Ed was mostly similar between peak and trough samplings among PT reagents (UMIN 000032492).