Predictors for non-persistence to edoxaban in Asian patients with atrial fibrillation: 2-year follow-up data from the ETNA-AF programme
C C Wang, J I Choi, N Y Chan, D Piyayotai, M Clasen, R Smolnik, J Van Zyl, M Unverdorben, J GoagAbstract
Introduction
Non-persistence to anticoagulant therapy is associated with higher risks of thromboembolic events in patients with nonvalvular atrial fibrillation (NVAF). Identifying predictors of non-persistence or premature study termination would provide valuable insights for developing strategies to reduce it in the real world and in future studies.
Purpose
To identify and characterise patients who either did not reach the end of the study period or were non-persistent with edoxaban during the 2-year follow-up of Asian patients from the ETNA-AF-ASCA programme.
Methods
ETNA-AF-ASCA is a prospective, observational, noninterventional study of patients with NVAF from South Korea/Taiwan, Thailand, and Hong Kong who received edoxaban at baseline. Patients who did not reach the end of the study period (deaths were excluded) and those who were non-persistent (defined as permanent discontinuation of edoxaban, not excluding deaths) with up to 2 years of follow-up were analysed. This exploratory analysis evaluated the association between baseline characteristics and either non-persistence or premature study termination using univariate Cox regression with adjustment for CHA2DS2-VASc score.
Results
The analysis included 3299 patients (Thailand: 289; South Korea: 1812; Taiwan: 973; Hong Kong: 225). At 2 years, 17% (n=561/3299) of patients were non-persistent, and 8.5% (n=281/3299) prematurely terminated the study for reasons other than death. Median overall age was 72.0 years (interquartile range [IQR], 66.0–79.0 years), and 41.1% of patients were female. Median CHA2DS2-VASc score was 3.0 (IQR, 2.0–4.0) across all groups. After adjustment, several baseline characteristics remained significantly associated (adjusted hazard ratio [aHR; 95% confidence interval (CI)]) with a higher risk of non-persistence, including body mass index ≥30 kg/m2 vs ≥18.5 to <25 kg/m2 (1.32 [1.00–1.74]), creatinine clearance <30 mL/min vs ≥80 mL/min (2.04 [1.43–2.91]), congestive heart failure (1.40 [1.11–1.76]), chronic hepatic disease (1.74 [1.20–2.53]), perceived frailty (1.48 [1.10–2.00]), any NVAF symptoms at baseline (1.55 [1.30–1.85]), and antiarrhythmic drug (AAD) at baseline (1.26 [1.06–1.49]; Figure 1). A higher risk of premature study termination was significantly associated (aHR [95% CI]) with any NVAF symptoms at baseline (1.54 [1.21–1.98]) and almost always vs always complying with drug therapy at baseline per investigator judgement (1.39 [1.02–1.88]; Figure 2). Ongoing AAD at baseline was associated (0.67 [0.53–0.85]) with a lower risk of premature study termination.
Conclusions
In this exploratory analysis, most patients completed the study and were classified as persistent during the 2-year follow-up. After adjustment, several baseline characteristics were associated with a higher risk of non-persistence or premature study termination and may help improve treatment persistence and clinical outcomes.