Higher D-Dimer and Prothrombin Fragment 1+2 in Patients Taking Direct Oral Anticoagulants As Compared to Vitamin K Antagonists

Introduction: A central question regarding anticoagulant treatment of venous thromboembolism (VTE) is how long maintenance therapy should be continued. Current decision algorithms, which were developed during the era of vitamin K antagonists (VKA), are based on D-dimer measurements before and/or after discontinuation of anticoagulants. However, recent clinical data from the Apidulcis study questioned the validity of this approach in patients taking direct oral anticoagulants (DOACs). One reason for this may be generally higher D-dimer levels in patients on DOACs than in those on VKAs, for which there is growing evidence. The aim of our study was to comparatively assess D-dimer levels in a large cohort of patients on either DOAC or VKA therapy. Prothrombin fragment 1+2 (F1+2) was additionally measured in order to gain more insights into the extent of drug-induced hypocoagulation. Methods: The study population consisted of patients with VTE who received maintenance anticoagulation therapy. D-dimer and F1+2 measurements performed during visits in our outpatient thrombophilia clinic were retrospectively assessed. Samples not drawn at peak or trough post-dose intervals, or duplicate peak and trough samples were excluded from analysis. The final DOAC cohort included 419 peak and 249 trough samples from patients on rivaroxaban, 353 peak and 138 trough samples on apixaban, and 60 peak and 26 trough samples on edoxaban treatment. Patients on VKAs with an INR within the therapeutic range of 2.0 to 3.0 (n=228) served as the reference cohort. Results: In the VKA cohort, mean D-dimer and F1+2 levels were 0.32 mg/L and 41.7 pmol/L, respectively. Biomarker concentrations at peak and trough levels did not differ statistically from each other in the three DOAC subgroups. However, across all groups, peak (trough) levels of both biomarkers were significantly higher (P<.0013) than in the VKA cohort, with D-dimer levels of 0.50 (0.41) mg/L for rivaroxaban, 0.55 (0.42) mg/L for apixaban, and 0.54 (0.43) mg/L for the edoxaban subgroup. Peak (trough) F1+2 levels were 132 (147) pmol/L for rivaroxaban, 150 (162) pmol/L for apixaban, and 138 (159) pmol/L for edoxaban. Moreover, the proportion of patients with D-dimer levels above the age-adjusted cut-off was lower in the VKA cohort (9.2%) than in patients on rivaroxaban (16.8%, P=.006), apixaban (21.8%, P=4x10-5), and edoxaban (20.9%, P=.005). Of note, F1+2 levels above the upper reference limit (407 pmol/L) were not observed in the VKA and edoxaban groups, and frequencies in the rivaroxaban and apixaban groups (1.8% and 2.0%, respectively) did not differ significantly from the VKA group (0%). Conclusion: The observation of comparable D-dimer and F1+2 levels at peak and trough post-dose intervals is evidence that both biomarkers reach a steady-state level in patients taking DOACs. Thus, they can in principle be used to assess and guide anticoagulation therapy. However, higher D-dimer levels in patients on DOACs may warrant reconsideration of threshold values and new validation of clinical decision algorithms. The potential benefits of additional assessment of F1+2 should be addressed in further studies.