Targeted 14-day ECG monitoring in high-risk individuals identified a higher burden of asystole >2 seconds compared to general screening (22% vs 14%, p=0.01).
RCT (n=589)
Does targeted AF screening using a risk prediction model increase the detection of clinically relevant non-AF arrhythmias compared to general screening during 14-day ECG monitoring?
14-day ECG monitoring for AF screening detects a substantial burden of other clinically relevant arrhythmias, with higher rates of asystole and ectopic beats found in machine-learning targeted high-risk individuals.
Absolute Event Rate: 22% vs 14%
p-value: p=0.01
Abstract Background/introduction Atrial fibrillation (AF) screening is recommended for high-risk individuals. During screening interventions using long-term electrocardiogram (ECG) monitoring, clinically relevant arrhythmias other than AF may be detected. However, the incidence of these arrhythmias remains unknown. Purpose Quantifying the incidence of clinically relevant arrhythmias other than AF detected during the 14-day ECG screening intervention in the CONSIDERING-AF trial. Methods The CONSIDERING-AF trial compared targeted AF screening using a machine-learning based risk prediction model (RPM) with 14-day ECG monitoring to standard care. Of its four parallell arms, two (RPM + intervention and general + intervention) underwent ECG monitoring. Per-protocol incidence of clinically relevant arrhythmias was compared between these arms, including non-sustained ventricular tachycardia (nsVT), supraventricular tachycardia (SVT), premature ventricular beats (PVC) and supraventricular beats (PAC), atrioventricular block II type II or III (AVB), and asystole 2 seconds. Results A total of 589 patients underwent ECG monitoring, 308 (52%) in the general + intervention arm and 281 (48%) in the RPM + intervention arm. Mean age was 75±5 years and 79±5 years, respectively (p0.001), and the proportion of female participants was 54% vs. 40% (p0.001). The incidence of clinically relevant arrhythmias in the general + intervention versus RPM + intervention arms was: nsVT 37% vs. 38% (p=0.86), SVT 96% vs. 97% (p=0.82), AVB 2.9% vs. 3.6% (p=0.84), and asystole 2 seconds 14% vs. 22% (p=0.01). Median PVC burden was 3% (IQR 1–23) vs. 7% (IQR 1–45) (p0.001), and median PAC burden was 13% (IQR 6–35) vs. 25% (IQR 11–76) (p0.001). In logistic regression, type II diabetes predicted high-degree AV block (adjusted OR 4.2, 95% CI 1.6–10.8). Male sex predicted nsVT (adjusted OR 1.8, 95%CI 1.3–2.5). No independent predictors remained for SVT after adjustment. Nine patients received a pacemaker, and one patient received an implantable cardioverter-defibrillator. Conclusion(s) Prolonged ECG monitoring identified a substantial burden of clinically relevant arrhythmias beyond AF, with higher burden of asystole 2 seconds, and ectopic activity among targeted high-risk individuals. Type II diabetes independently predicted high-degree atrioventricular block, and male sex predicted non-sustained ventricular tachycardia. The rate of resulting device therapy was low. Further research should determine how these findings should guide follow-up and management decisions.
Rakai et al. (Mon,) conducted a rct in High-risk for atrial fibrillation (n=589). Targeted AF screening using a risk prediction model (RPM) with 14-day ECG monitoring vs. General AF screening with 14-day ECG monitoring was evaluated on Incidence of asystole >2 seconds (p=0.01). Targeted 14-day ECG monitoring in high-risk individuals identified a higher burden of asystole >2 seconds compared to general screening (22% vs 14%, p=0.01).