Effect of adherence to anticoagulation on the incidence of ischemic stroke in the patients with atrial fibrillation: A retrospective cohort study

초록

Background: Although NOAC had comparable or superior effectiveness compared to VKA in AF, outcomes in low compliance group was limited in the published literature. Objectives: The purpose of this study is to evaluate the effectiveness and safety of anticoagulation among noncompliant patients using NOAC or VKA compared to those compliant to NOAC treatment. Methods: This study was performed using the National Health Information Database from 2009 to 2018. A retrospective cohort was constructed with the patients newly diagnosed as non-valvular atrial fibrillation from 2010 to 2018. All participants were required to be between 18 to 100 years old, newly diagnosed AF from 2010.1-2018.6 without AF-related claim in the previous year, and CHA2DS2-VASc score of at least two. Patients with diagnosis codes indicating prosthetic heart valves or valvular heart disease, end stage renal disease, joint replacement surgery, ischemic stroke or major bleeding before enrollment were excluded. The primary exposure is compliance of NOAC (dabigatran, rivaroxaban, apixaban, and edoxaban) or VKA (warfarin) defined based on 80% Proportion of Days Covered (PDC) measured during 180 days after starting OAC. Study subjects were categorized into four treatment groups: VN (VKA non-compliant), VC (VKA compliant), NN (NOAC non-compliant), and NC (NOAC compliant). We followed up patients to the incidence of ischemic stroke, bleeding, switching of drug, death, 910 days after enrollment, or end of study period whichever came first. The differences in the baseline characteristics of a group from NC were calculated as standardized differences before and after the IPT weighting. The IPT weighted cumulative incidence rate of ischemic stroke was estimated considering bleeding, switching and death as competing events. Subdistributional hazard model was used to calculate the weighted hazard ratio. NNT and NNH for each outcome was presented as the reciprocal for difference of the IPT weighted cumulative incidence. Results: We identified 84,227 eligible patients who was diagnosed as AF and started to take oral anticoagulants from 2010 to 2018. IPT weighted cumulative incidence of ischemic stroke in VN, NN, and NC was 0.07005, 0.07787, and 0.05579 respectively. Weighted hazard ratios for VN and NN was 1.31 (1.26-1.37) and 1.34 (1.29-1.40). NNH by non-compliance to anticoagulation for ischemic stroke in VN and NN was 44 and 39. Conclusions: Non-compliant groups are prone to have a higher risk of stroke, regardless of OAC class.