Cardiac Magnetic Resonance GUIDEd Management of Mild-moderate Left Ventricular Systolic Dysfunction

doi:10.1001/jama.2026.17078

CMR GUIDE

Patient / Population Intervention / Exposure Comparison Outcome

In 353 patients with LVEF 36–50 % and scar on cardiac MRI, a primary-prevention ICD did not reduce sudden cardiac death or haemodynamically significant ventricular arrhythmia compared with an implantable loop recorder.

N

353patients

Design

Prospective, multicentre, open-label RCT, 1:1 (ICD vs implantable loop recorder), web-based randomisation stratified by site after core-lab LGE-CMR confirmation; recruitment mid-2015 to mid-2022, follow-up at least 3 years; Australia, UK and Europe

Endpoint

Time to composite of sudden cardiac death or haemodynamically significant ventricular arrhythmia (ventricular arrhythmia producing syncope or hypotension, SBP <90 mmHg)

Relevance

2Important — one of several pillars.

Result7.8% vs 9.2% (HR 0.76, 95% CI 0.37-1.58); no significant difference. Sudden cardiac death alone 1.7% vs 5.8% (HR 0.26, 95% CI 0.07-0.95). Below 70 years a 72% relative reduction in the primary endpoint, none above.

Selvanayagam JB, et al. JAMA. 2026; published online 28 August. 10.1001/jama.2026.17078

Discussion & critique

Scar on CMR does not by itself earn a defibrillator at an ejection fraction of 36-50%: the primary endpoint is neutral and the interval spans a halving and a 58% increase. The isolated reduction in sudden cardiac death and the benefit below 70 years are the interesting part and the fragile part — few events, a secondary endpoint and a subgroup. The comparator was an implantable loop recorder, so the control arm was monitored rather than treated, and recruitment ran from 2015 to 2022 while medical therapy changed underneath the trial.