Abstract 4146592: Device-Recorded Physical Activity and Atrial Fibrillation Burden: A Natural History Experiment from the COVID-19 Pandemic in the TRIM-AF Clinical Trial

Introduction: Low physical activity (PA) has been identified as a risk factor for development of atrial fibrillation (AF). However, the effect of changes in PA on directly recorded AF burden has not been well studied. The COVID-19 pandemic offered an opportunity to observe whether changes in activity were correlated with changes in AF burden. To determine if reduced PA is associated with higher AF burden, we assessed daily PA and AF burden data from patients with cardiac implantable electronic devices (CIEDs) enrolled in a prospective clinical trial, Targeting Risk Interventions and Metformin for Atrial Fibrillation (TRIM-AF, NCT03603912). Methods: Daily AF burden and activity were determined from implantable cardiac devices with atrial leads. The pandemic lockdown period was analyzed for up to 1 year. Pre-pandemic periods were matched by month to pandemic periods. To test the potential confounding of aging, pre-pre-pandemic periods were matched by month to pre-pandemic periods. To reduce the confounding of study interventions, matched periods were taken on one side of the study enrollment date. For PA and AF burden, Gaussian linear mixed models and a Bayesian mixed effect model were fitted and adjusted for age, sex, and device manufacturers. A Gaussian model was used to correlate daily activity minutes and AF%. Time splines were added to adjust for non-linear time effects. Outcomes are reported as mean activity minutes and daily AF%. Results: Comparing Pandemic vs. Pre-periods (N=82 periods; 55 male, 27 female), daily activity minutes decreased by a mean of 13.16±1.06 minutes/day, and daily AF burden increased by 16% [5%-26%]. Comparing Pre vs. Pre-Pre-Pandemic periods (N=60 periods; 41 male, 19 female), mean activity decreased by 2.28±1.13 mins/day, and AF burden increased by 57% [50%-64%]. A significant negative correlation between activity and AF burden was demonstrated (coefficient -2.0, 95% CI -2.4, -1.6). A decrease in 2.0 activity minutes was associated with a 10% increase in AF burden. Conclusions: This natural history analysis of PA and AF burden demonstrated decreases in activity and increases in AF burden with time and the pandemic. Activity and AF burden were significantly negatively correlated.