CMR scar did not identify who benefits from an ICD at LVEF 36–50%

Design

Open-label RCT, 18 sites

Population

353 adults with LVEF 36–50% and CMR-defined scar

Primary outcome

Sudden cardiac death or haemodynamically significant ventricular arrhythmia

Effect

7.8% vs 9.2%; HR 0.76 (95% CI 0.37 to 1.58)

CMR GUIDE randomised 353 adults with ischaemic or non-ischaemic cardiomyopathy, LVEF 36–50%, scar on late gadolinium enhancement and guideline-directed therapy to a primary-prevention ICD or an implantable loop recorder, at 18 sites in Australia, Germany and the UK. Median follow-up was 6.3 years; 70% had LVEF 40% or more. It was published in JAMA in August.

The primary composite of sudden cardiac death or haemodynamically significant ventricular arrhythmia occurred in 7.8% with an ICD and 9.2% with a loop recorder (HR 0.76, 95% CI 0.37 to 1.58). Sudden death alone was lower with an ICD (1.7% vs 5.8%, HR 0.26, 0.07 to 0.95), but that was a secondary component. All-cause and cardiovascular mortality were similar.

The trial does not support scar alone as a reason to implant in the mid-range ejection fraction. A benefit under 70 and possible harm over 70 is a subgroup finding and needs confirmation before it guides practice.

• Do not implant a primary-prevention ICD at LVEF 36–50% on the strength of CMR scar alone.
• Keep using CMR for diagnosis and aetiology in cardiomyopathy.
• Optimise guideline-directed therapy and reassess LVEF before any device decision.
• Discuss the uncertainty openly with younger patients who ask about an ICD.

Why it matters

It weakens the case for extending ICDs beyond LVEF 35% on the basis of scar imaging.

Don't overread it

The lower sudden-death rate with an ICD is a secondary component, not the trial's answer.

The statistics, in plain English

The primary hazard ratio of 0.76 has a wide interval (0.37 to 1.58) that includes both meaningful benefit and harm, so the trial is inconclusive rather than proof of no effect. The lower sudden-death rate is a component of the composite and was not tested as the main question. The age interaction (P = 0.01) is one of six subgroup analyses and could be chance.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.