LAUSR

OBJECTIVES Several studies have suggested a possible relationship between atrial fibrillation (AF) and falls. However, whether the relationship depends on AF types is unclear. We investigated the relationship between sustaining AF and falls. DESIGN Single hospital-based cohort study with a follow-up of falls within 3 years after baseline. SETTING AND PARTICIPANTS A total of 14,056 patients from our cohort between February 2010 and March 2016. MEASURES Incidence of falls within 3 years by baseline cardiac rhythm was measured, and we investigated the effects of AF types on incidence of falls. RESULTS The study population was divided into younger (<75 years old; n = 11,808) and older (≥75 years old; n = 2248) groups, and then divided into 3 groups according to the baseline cardiac rhythm: sinus rhythm (SR), paroxysmal AF (PAF), and persistent AF (PeAF). There were more male patients in the PeAF group; these patients had more comorbidities both in the younger and older groups. The cumulative incidence rates of falls at 1 year in patients with SR, PAF, and PeAF were similar in the younger group (0.4%, 0.4%, and 0.6%, respectively; P = .496), whereas those were significantly different in the older group (2.3%, 2.7%, and 5.0%, respectively; P = .024). In multivariate analysis, both PAF [hazard ratio (HR) 1.179; 95% confidence interval (CI) 0.553-2.511, reference SR] and PeAF (HR 1.502; 95% CI 0.635-3.556) were not associated with falls in the younger group. In the older group, PeAF was independently associated with incidence of falls (HR 2.257; 95% CI 1.262-4.037), but PAF was not (HR 1.317; 95% CI 0.673-2.574). CONCLUSIONS/IMPLICATIONS PeAF, not PAF, was associated independently with falls in older patients, suggesting the possible effect of irregular beats on physical frailty in the older population.