Abstract Introduction Loop recorders have provided the opportunity to systematically screen individuals for arrhythmias, and in doing so provide new appreciated mechanistic insights to arrhythmia pathophysiology. Atrial fibrillation (AF) is a complex arrhythmia with a multitude of precipitants and substrates. Previous studies looking at the circadian rhythm of AF, have required patient symptoms to define onset. All studies on the circadian rhythm of AF suggest that AF does not occur randomly. It has been suggested that the circadian variation in autonomic function may provide an underlying substrate for AF initiation. This is the largest study to utilise loop recorder detected AF to explore the circadian pattern of AF. Purpose We sought to ascertain how the detection of incident AF varied throughout the day, in patients monitored with a loop recorder Methods This was a single-centre retrospective observational study. Consecutive adult patients referred for ILR implantation between March 2009 and November 2019, without prevalent AF or atrial flutter, were included. Demographic and anthropometric data was identified from the individuals medical notes. The time of incident AF episodes was identified from loop recorder reports. Results A total of 1168 AF episodes of any duration were identified by the ILR, of which onset time was available for 1118 episodes. Onset of AF episodes in our data for the entire population demonstrated a monophasic distribution (figure 1a). The OR for developing AF between 13.00-14.59 compared to 03.00-04.59 was 4.89 (95% CI 3.47-6.90). Amongst the ESUS population time of onset distribution showed multiple daytime peaks (figure 1b). The OR for developing AF between 10.00-11.59 compared to 03.00-04.59 was 4.39 (95% CI 2.68-7.20). Amongst the non- ESUS population time of onset distribution (figure 1c) shows a monophasic pattern with the peak at time blocks 12.00-14.59. The OR for developing AF between 13.00-14.59 compared to 02.00-03.59 was 6.75 (95% CI 4.17-10.93). Conclusion(s) Our study demonstrated that there is a diurnal variation for incident AF detection using a loop recorder. There is a general propensity for episodes to occur in the afternoon, with a nadir at night. This is in contrast to previous reports which have suggested that the nocturnal and morning peaks are more apparent. Moreover, different cohorts appear to show different temporal patterns, with ESUS patients having multiple time peaks, whereas the non-stroke cohort had a solitary peak period. Understanding the diurnal variation of AF may help to shed more light on the complex pathogenesis of AF, and whether different AF phenotypes are associated with stroke risk. Moreover, the finding of a day-time propensity for AF, adds weight to the utilisation of consumer facing non-invasive monitoring tools during the day for AF identification in high-risk populations.
Read full abstract