Persistence and predictors for non-persistence to edoxaban therapy in patients with atrial fibrillation: 4-year follow-up data from the ETNA-AF-Europe study

Non-persistence to non-vitamin K antagonist oral anticoagulants (NOACs) is associated with increased stroke risk in atrial fibrillation (AF) patients. To date, most NOAC persistence studies have been retrospective and often did not include patients receiving edoxaban. Identifying predictors for non-persistence may be useful when considering the design of future studies and help develop strategies to reduce non-persistence in clinical practice.

To define and describe patients who did not reach the end of the study period and were non-persistent to edoxaban during the 4-year follow-up of ETNA-AF-Europe.

ETNA-AF-Europe was a prospective, observational study conducted in AF patients receiving edoxaban. In this subanalysis, patients not reaching the end of the study period and patients who were non-persistent (defined as permanent discontinuation of edoxaban) after 4 years of treatment were examined, including patients switching to other NOACs and reasons for discontinuation. Baseline predictors for not reaching the end of the study period and baseline predictors for non-persistence with edoxaban treatment during 4 years of follow-up were also examined. Persistent patients may have missed dose/interruption.

Overall, 13,164 patients were included in the analysis. During the whole study period, 14.3% died. At the end of the study, 27.3% of patients prematurely terminated the study and 71.5% were classified as completing the study. Of patients who left the study (n=3598), 30.2% were lost-to-follow-up, 6.9% withdrew consent, and 5.9% transferred to another institution. Of patients who completed the study (n=9417), 87.4% were classified as persistent within the 4 years. Only 6.2% permanently discontinued edoxaban without switching to another NOAC (Table).

Following backwards elimination, baseline characteristics associated with premature study termination (excluding death) were male sex, low body mass index, low renal function, chronic hepatic disease, long standing persistent/permanent AF, smoking, low compliance as judged by investigator, vitamin K antagonist-naïve, and no rate/rhythm control drug at baseline. Characteristics associated with non-persistence were increasing age, male sex, body weight extremes, low renal function, heart failure, vascular disease, chronic hepatic disease, alcohol use, perceived frailty, chronic obstructive pulmonary disease, smoking, current AF symptoms, and ablation (Figure). The criterion for backward elimination was p=0.05.

Most patients in ETNA-AF-Europe reached the end of the study. Less than a quarter were classified as non-persistent during the 4-year follow-up. The majority of patients who were non-persistent switched to another NOAC with a small number completely stopping anticoagulant therapy. Male sex, low renal function, chronic hepatic disease, and smoking were associated with both treatment discontinuation and non-persistence.

Termination and persistence