Screening for Atrial Fibrillation in Older Adults

By Michael H. Crawford, MD, Editor

SYNOPSIS: A two-week ambulatory electrocardiogram monitor in a large group of individuals 70 years of age or older with no history of atrial fibrillation (AF) showed a very low incidence of AF (4.4%), almost all of which was paroxysmal. In less than 2% of the subjects did it represent ≥ 2% of the monitoring time. However, some patients had hours of AF, raising a concern for thromboembolic risk.

SOURCE: Singer DE, Atlas SJ, Go AS, et al. Atrial fibrillation burden on a 14-day ECG monitor: Findings from the GUARD-AF trial screening arm. JACC Clinical Electrophysiol. 2024; Sep 1. doi: 10.1016/j.jacep.2024.08.010. [Online ahead of print].

The burden of atrial fibrillation (AF), defined as the frequency of episodes, their distribution in time, and duration, has been associated with an increased risk of thromboembolic (TE) events. Thus, efforts to detect and classify AF in individuals at higher risk may be of value to reduce future TE events, such as stroke, by preventive treatment. Hence, the reducinG stroke by screening for UndiAgnosed Atrial Fibrillation in elderly inDividuals (GUARD-AF) randomized trial is of interest.

Individuals age 70 years or older with no prior history of AF were recruited from 149 primary care sites in the United States and randomized to usual care or AF screening with a single-lead patch monitor that stores up to 14 days of data (Zio^®XT). AF episodes had to last at least 30 seconds to be counted. Study outcomes were whether AF (or atrial flutter) were detected and, if so, the percentage of the total recording time when AF was detected, the duration of the longest episode, and the total number of episodes. Also, the CHA₂DS₂-VASc score was calculated once at baseline.

GUARD-AF enrolled 11,905 subjects from 2020-2021 with a mean age of 75 years, 57% women, and 89% white. The median CHA₂DS₂-VASc score was 3. The median ECG monitor wear time was 14 days. AF was detected in 4.4% of the subjects; of these, 0.5% had continuous AF and 4% had paroxysmal (pAF). The median percent time in AF was 0.46%, or 75 minutes, and the median longest episode was 38 minutes. The upper quartile threshold of 2.5% time in AF corresponded to 7.6 hours.

In multivariate analysis that included the variables in the CHA₂DS₂-VASc score, only age older than 80 years was associated with the detection of AF (odds ratio [OR], 1.46; 95% confidence interval [CI], 1.06-2.02). The authors concluded that there was a very low burden of AF in older individuals, but one-quarter of those with AF detected had multiple hours of AF, raising concern about TE risk.

Commentary

Despite being cut short by COVID-19, this is the largest trial of detecting subclinical AF to date. Also, the patients were recruited from primary care offices and randomized to monitoring or no monitoring. In addition, the majority of the patients were in the oral anticoagulant-recommended CHA₂DS₂-VASc score range. There will be a second publication after the long-term outcome phase of the study is completed.

The treatment of device-detected, or so-called subclinical AF, is less clear compared to continuous AF or clinically detected pAF. Prior studies of device-detected AF have suggested that the percent time in AF and the length of the longest episode are associated with TE events. In GUARD-AF, these indices of AF burden are skewed toward low magnitude. For example, 60% had AF for < 1.0% of the time and the median longest AF episode was 38 minutes (range, 2 minutes to 245 minutes). However, the upper quartile of the longest episodes was four hours or longer. Also, the median number of AF episodes detected was three, with the upper quartile being ≥ 19.

In those with pAF, the correlation coefficient between percent time in AF with the length of the longest episode was 0.94, but the correlation of percent time in AF with the number of AF episodes was less impressive (0.54). The correlation between the number of episodes with the length of the longest episode was even weaker (0.33). This suggests that 24- to 48-hour monitoring may not be adequate to detect the longest episode of AF.

Another strength of the study was the use of the Zio^®XT patch monitor, which is commonly used clinically. There were weaknesses, though. The only clinical variables included were those in the CHA₂DS₂-VASc score. Also, the length of monitoring undoubtedly will influence the amount of AF detected. Other devices, such as implanted loop recorders and pacemaker/defibrillators, can provide much longer monitoring. However, these devices are not usually found in the average older individual, and comorbidities certainly affect the risk of AF. The long-term outcomes from this study when completed will be of interest.

Michael H. Crawford, MD, is a Professor of Medicine and Consulting Cardiologist, University of California Health, San Francisco.