Increased Prevalence of Elevated D-Dimer Levels in Patients on Direct Oral Anticoagulants: Results of a Large Retrospective Study

Abstract

Elevated D-dimer levels during anticoagulant therapy with vitamin K antagonists (VKA) are associated with an increased risk of thrombosis. It has been hypothesized that elevated D-dimer levels in patients receiving direct oral anticoagulants (DOACs) also indicate an increased risk of thrombosis recurrence, but data on the distribution of D-dimer levels in patients with VTE on DOACs are sparse. In the present study we retrospectively analyzed D-dimer levels in patients taking DOACs after first or recurrent venous thrombosis (n = 1,716, 1,126 thereof rivaroxaban, 481 apixaban, 62 edoxaban, and 47 dabigatran). Patients on VKA (n = 402) served as control group. Thrombotic events in the study population were categorized into distal deep venous thrombosis (DVT, n = 552 patients), distal DVT with pulmonary embolism (PE, n = 166), proximal DVT (n = 685), proximal DVT with PE (n = 462), PE without DVT (n = 522), DVT of the upper extremity (n = 78), cerebral venous sinus thrombosis (CVST, n = 48), and other venous thrombosis (n = 74). In VKA users a median D-dimer level of 0.20 mg/l was observed. In patients on DOACs D-dimer levels were significantly higher, with 0.26 mg/l for rivaroxaban, 0.31 mg/l for apixaban (P < 10−16 each), 0.24 mg/l for edoxaban (P = 2 × 10−5), and 0.25 mg/l for dabigatran (P = 4 × 10−4). These differences in comparison to patients on VKA treatment could not be explained by the patients' age, sex, body mass index, and type of thrombosis as these characteristics did not differ significantly between cohorts. Moreover, the prevalence of D-dimer levels above age-adjusted cut-offs [≥0.50 mg/l in ≤50-year-old patients, ≥(age × 0.01) mg/l in >50-year-old patients] was higher in patients on rivaroxaban (13.9%, RR 1.74, 95% CI 1.21–2.50), apixaban (17.0%, RR 2.14, 95% CI 1.45–3.15) and dabigatran (23.4%, RR 2.94, 95% CI 1.59–5.44) than in patients on VKA (8.0%). In patients on edoxaban D-dimer levels above the reference range were observed in 14.5%, but no statistical significance was reached in comparison to the VKA cohort. In conclusion, the obtained data suggest, that the type of oral anticoagulant should be considered in the clinical assessment of D-dimer levels in thrombosis patients. Further studies are warranted to evaluate a potential association between elevated D-dimer levels and thrombosis risk in patients on DOACs.