Clinical characteristics and outcomes of Japanese atrial fibrillation patients with poor medication adherence: A sub-analysis of the GENERAL study.

BACKGROUND

Oral anticoagulation therapy is essential for preventing stroke in patients with atrial fibrillation (AF). However, poor anticoagulant adherence may hamper medication safety and effective prevention of stroke.

METHODS

GENERAL is a prospective cohort study of AF patients taking rivaroxaban prescribed by general practitioners in Japan. In this study, anticoagulant adherence was calculated as the proportion of days covered (PDC), and patients were retrospectively divided into two groups: good adherence (PDC ≥80 %) and poor adherence (<80 %).

RESULTS

Of 5680 patients in the GENERAL study, the poor adherence group consisted of 223 patients (3.9 %). Baseline clinical characteristics were almost comparable regarding age (PDC ≥80 % vs. <80 %: 73.9 vs. 74.0 years, p = 0.92) and sex (male 64.6 % vs. 66.8 %, p = 0.52). The PDC <80 % group more often had various co-morbidities, and had significantly higher CHADS2 (2.14 vs. 2.28, p = 0.04) and CHA2DS2-VASc scores (3.12 vs. 3.31, p = 0.045). There was no significant difference in HAS-BLED score (1.41 vs. 1.47, p = 0.39). During 2-year follow-up, the incidences of stroke or systemic embolism (1.14 vs. 3.56 % per patient-year, p < 0.01), major bleeding (0.59 vs. 1.78 % per patient-year, p < 0.01), and net clinical outcome (the composite of stroke, systemic embolism, major bleeding, or death) (3.49 vs. 7.78 % per patient-year, p < 0.01) were significantly higher in the poor adherence group; however, there was no significant difference in all-cause (1.89 vs. 2.73 % per patient-year, p = 0.23) and cardiovascular mortality (0.86 vs. 1.49 % per patient-year, p = 0.18). Multivariate analysis revealed that the poor adherence group was independently associated with stroke or systemic embolism (adjusted hazard ratio 3.12, 95 % confidence interval 1.79-5.47), major bleeding (2.87, 1.31-6.34), and net clinical outcome, (2.02, 1.39-2.93), but not with all-cause (1.18, 0.64-2.17) or cardiovascular death (1.39, 0.60-2.93).

CONCLUSIONS

Poor anticoagulant adherence, as measured by PDC <80 %, was associated with higher incidence of stroke or systemic embolism and major bleeding in the GENERAL study.