Effect of perceived and objectively-assessed frailty on outcomes in edoxaban-treated patients with atrial fibrillation: data from the ETNA-AF-Europe 4-year follow-up

Abstract

Abstract Background Frailty is common in patients with atrial fibrillation (AF). However, different definitions of frailty are used, and these may variably overlap with frailty perception among physicians. Long-term data on outcomes based on perceived versus objective frailty in patients with AF treated with a non-vitamin K antagonist oral anticoagulant are lacking. Purpose To assess the effect of perceived versus objective frailty on outcome events in patients with AF treated with edoxaban during the 4-year follow-up of the ETNA-AF-Europe registry (NCT02944019). Methods ETNA-AF-Europe is a multinational, multicentre, post-authorisation, observational study conducted in patients with AF receiving edoxaban. In this sub-analysis, patients were evaluated in terms of perceived or objective frailty status (yes or no categories). Perceived frailty was based on investigators’ own clinical binary judgement in each patient. Objective frailty was determined using a Modified Frailty Index, a simplified adaptation of the Rockwood’s Frailty Index. Patient demographics and characteristics were collected at baseline. The efficacy and safety outcomes are reported here comparing frailty status (yes versus no) among perceived or objectively classified patients. Results Among the 13,164 patients with 4 years of follow-up, the prevalence of perceived and objective frailty was 10.7% and 4.1%, respectively. Compared with the objectively frail group, those perceived as frail were older, with a lower proportion of males, had lower body weight and creatinine clearance values, yet lower CHA2DS2-VASC and modified HAS-BLED scores (Table 1). Further, patients with objective frailty tended to have more comorbidities (e.g., diabetes, heart failure, hypertension; Table 1). Those deemed frail were more likely to receive the reduced 30 mg edoxaban dose versus non-frail patients (p<0.001 for both groups), and non-recommended dosing regimens (60 or 30 mg) were more often prescribed in frail versus non-frail patients (p<0.0001 for all comparisons; Table 1). Patients perceived as frail had a significantly higher risk of all evaluated outcomes compared with non-frail patients (p-value range: <0.0001 – 0.0459; Figure 1). Similar data were seen for objectively frail patients, with the exception of no significant differences in the risk of intracranial haemorrhage (ICH, p=0.0918) or haemorrhagic stroke (p=0.1619; Figure 1). Conclusions Frailty was associated with an inappropriate dosing of edoxaban and an increased incidence of outcome events, quantitatively similar in the perceived and objective frailty groups. After 4 years of therapy, even if ICH risk was increased in frail subjects, its cumulative incidence was low.