TrialsgrambetaRCTs are beautiful

BEAGLE

2022

Artificial Intelligence-Guided Screening for Atrial Fibrillation Using Electrocardiogram During Sinus Rhythm: a Prospective Non-Randomised Interventional Trial

In 1,003 patients with stroke risk factors and no known atrial fibrillation, AI-ECG risk stratification followed by ambulatory rhythm monitoring increased detection of new atrial fibrillation compared with propensity-matched usual care.

Result

Primary endpoint: Newly diagnosed atrial fibrillation

AF detected in 7.6% (48/633) of AI-high-risk vs 1.6% (6/370) of low-risk patients (OR 4.98, 95% CI 2.11–11.75; P=0.0002); vs propensity-matched usual care, detection 10.6% vs 3.6% in the high-risk group (P<0.0001) and 2.4% vs 0.9% in the low-risk group (P=0.12) over a median 9.9 months.

Original publication
Subtopic
Atrial fibrillation
Design
Prospective, non-randomised interventional trial with propensity-matched real-world controls, USA
Relevance
Supporting: context rather than must-know.
Source
Noseworthy PA, et al. Lancet. 2022;400(10359):1206-1212.doi 10.1016/S0140-6736(22)01637-3

Discussion & critique

Showed that an AI read of a sinus-rhythm ECG can focus AF screening on the patients most likely to have it, a rationale later applied within VITAL-AF. Non-randomised: the usual-care comparison was built by propensity matching from eligible but unenrolled patients, and the endpoint is AF detection rather than stroke prevention, so whether AF found this way benefits from anticoagulation is untested.