RESUMO
The attractiveness of direct oral anticoagulants (DOACs) over vitamin K antagonists, in addition to a better benefit-risk ratio, stems from the fact that no therapeutic drug monitoring is deemed necessary. This has been recently mitigated by the fact that increased dabigatran (D) plasma levels have been associated with hemorrhages, and is currently under scrutiny of the European Medicines Agency. We aimed to evaluate, in real conditions of use, whether patients with out-of-range DOAC blood concentrations (too high or too low) were associated with bleeding or thrombosis. Patients treated with D or rivaroxaban (R) were prospectively included in a hospital cohort. D and R plasma levels were measured by high-pressure liquid chromatography-tandem mass spectrometry-at the physician's demand. We defined concentration range as "expected" within the 95% confidence interval of the mean concentration obtained from pivotal trials, and "out of range" when concentrations were outside of that interval. A blind assessment of concentrations versus occurrence of bleeding or thrombosis was performed by means of univariate and multivariate analysis. Three hundred and twenty-two patients (mean age 78.5 years ± 13.1), treated with D or R were included consecutively. They had a mean CHA2DS2-VASc at 4.4 ± 1.7 and a mean HAS-BLED score at 1.7 ± 0.9. Irrespective of the DOAC prescribed, patients presenting with out-of-range concentrations had significantly more bleeding or thrombosis than patients with expected concentrations (P < 0.001). Patients with bleeding were more prone to have concentrations beyond the 95 percentile (N = 62, P < 0.001), whereas patients with thrombosis were more likely to have concentrations below the fifth percentile (N = 26, P < 0.05). The main risks associated with hemorrhages were abnormal concentrations, a high HAS-BLED score, the patient's age, and the creatinine blood level. For thrombosis, a concentration below the fifth percentile was the only risk factor that was significant in our cohort. While D and R under current recommendation have a better benefit-risk ratio than warfarin, their safe usage could be further optimized by some degree of therapeutic monitoring.