Longitudinal relationships of periodic limb movements during sleep and incident atrial fibrillation

Abstract

Objectives This study aimed to examine relationship between periodic limb movements during sleep (PLMS) and incident atrial fibrillation/flutter (AF). Methods Prospective multicenter cohort (n = 2273: adjudicated AF group; n = 843: self-reported AF group) of community-dwelling men without prevalent AF were followed for an average of 8.3 years (adjudicated) and 6.5 years (self-reported). PLMS index (PLMI, <5 (ref), ≥5 to <30, ≥30) and PLM arousal index (PLMAI, <1 (ref), ≥1 to <5, ≥5) were measured by polysomnography. Incident adjudicated and self-reported AF were analyzed by Cox proportional hazards and logistic regression, respectively, and adjusted for age, clinic, race, body mass index (BMI), alcohol use, cholesterol level, cardiac medications, pacemaker, apnea–hypopnea index, renal function, and cardiac risk. The interaction of age and PLMS was examined. Results In this primarily Caucasian (89.8%) cohort of older men (mean age 76.1 ± 5.5 years) with BMI of 27.2 ± 3.7, there were 261 cases (11.5%) of adjudicated and 85 cases (10.1%) of self-reported incident AF. In the overall cohort, PLMI and PLMAI were not associated with adjudicated or self-reported AF. There was some evidence of an interaction of age and PLMI (p = 0.08, adjudicated AF) and PLMAI (p ≤ 0.06, both outcomes). Among men aged ≥76 years, the highest PLMI tertile was at increased risk of adjudicated AF (≥30 vs. <5; hazard ratio (HR) = 1.63, 1.01–2.63) and the middle PLMAI tertile predicted increased risk of both outcomes (1 to <5 vs. <1; adjudicated, HR = 1.65, 1.05–2.58; self-reported HR = 5.76, 1.76–18.84). No such associations were found in men aged <76 years. Conclusions Although PLMS do not predict AF incidence in the overall cohort, the findings suggest PLMS increases incident AF risk in the older subgroup.