ICD primary prevention trials — viva

Stem

Practice viva. You are the cardiology registrar presenting at the device meeting. The examiner works through the randomized trials of the primary-prevention implantable cardioverter-defibrillator (ICD), why some were negative, and how the European, North American and Australian rows grew out of them.^{[14][17][19][1][4][9]}

Branch A — The trials that built the case

Examiner: Which trial made LVEF alone enough after a myocardial infarction (MI)?

• MADIT-II enrolled 1232 patients with a prior MI and a left ventricular ejection fraction (LVEF) of 0.30 or less, randomized 3:2 to an implantable defibrillator or conventional medical therapy, and did not require invasive electrophysiological testing.^[14]
• Over an average of 20 months, mortality was 14.2% with the defibrillator and 19.8% with conventional therapy (hazard ratio 0.69).^[14]
• The earlier MADIT randomized 196 patients (NYHA class I, II or III, prior MI, LVEF 0.35 or less, asymptomatic unsustained ventricular tachycardia [VT], inducible nonsuppressible ventricular tachyarrhythmia on electrophysiologic study) to an implanted defibrillator or conventional medical therapy.^[11]
• Over an average of 27 months, the MADIT hazard ratio for overall mortality with the defibrillator versus conventional medical therapy was 0.46.^[11]

Examiner: And beyond coronary disease?

• SCD-HeFT randomized 2521 patients with New York Heart Association (NYHA) class II or III heart failure and LVEF 35% or less, 52% ischaemic and 48% nonischaemic, to conventional therapy plus placebo, amiodarone or a shock-only single-lead ICD.^[15]
• Over a median of 45.5 months, amiodarone had a similar risk of death to placebo (hazard ratio 1.06) and the ICD a 23% lower risk (hazard ratio 0.77).^[15]

Branch B — The negative trials

Examiner: Why not implant straight after the infarct, when risk is highest?

• DINAMIT, an open-label randomized trial, compared ICD therapy with no ICD therapy 6 to 40 days after MI in patients with LVEF 0.35 or less and impaired cardiac autonomic function; over a mean of 30 months there was no difference in overall mortality (hazard ratio 1.08).^[17]
• In DINAMIT, arrhythmic deaths fell with an ICD (12 versus 29 without) but nonarrhythmic deaths rose (50 versus 29), and the investigators concluded that one offset the other.^[17]
• IRIS, an open-label multicentre randomized trial, enrolled 898 patients 5 to 31 days after MI who met criterion 1 (LVEF 40% or less and a heart rate of 90 beats per minute or more on the first available electrocardiogram), criterion 2 (NSVT of 150 beats per minute or more on Holter monitoring) or both criteria, and assigned them to an ICD or medical therapy alone.^[18]
• Over a mean of 37 months, IRIS overall mortality with an ICD versus medical therapy alone was not reduced (hazard ratio 1.04), with fewer sudden (27 versus 60) but more nonsudden cardiac deaths (68 versus 39).^[18]
• The 2026 European Society of Cardiology (ESC) heart failure guideline does not recommend a primary-prevention ICD within 40 days of an MI, as implantation during this period does not improve prognosis (Class III, Level B1).^[1]
• In North America, the 2025 American College of Cardiology/American Heart Association (ACC/AHA) acute coronary syndrome guideline recommends an ICD to reduce death in selected post-MI patients with LVEF 40% or less (its Table 17), at least 40 days post MI and at least 90 days postrevascularization (COR 1, LOE A).^[4]

Examiner: And at the time of bypass surgery?

• CABG Patch randomized 900 patients under 80 years with LVEF below 0.36 and an abnormal signal-averaged electrocardiogram, scheduled for elective coronary bypass surgery, to a defibrillator implanted at surgery or to control; over an average of 32 months the hazard ratio for death from any cause was 1.07.^[13]

Branch C — Non-ischaemic cardiomyopathy

Examiner: Was DANISH a negative trial?

• DANISH randomized 1116 patients with symptomatic systolic heart failure not caused by coronary artery disease and LVEF 35% or less to an ICD or usual clinical care; 58% in both groups received cardiac resynchronization therapy (CRT).^[19][27]
• At a median of 67.6 months, death from any cause was not significantly lower with an ICD than with usual clinical care (hazard ratio 0.87, P=0.28), but sudden cardiac death was lower (4.3% versus 8.2%; hazard ratio 0.50).^[19]
• The 2017 American Heart Association/American College of Cardiology/Heart Rhythm Society (AHA/ACC/HRS) guideline says that, because CRT was present in 58% of both arms, the DANISH results should not be generalized to patients with non-ischaemic cardiomyopathy (NICM) who are ineligible for CRT.^[3]
• ESC 2022 reports that a meta-analysis of all six randomized primary-prevention trials in DCM with HFrEF still showed a reduction, although lower, in overall mortality with ICD therapy (relative risk 0.76); it gives no follow-up period.^[2]

Examiner: How did ESC 2026 grade it, then?

• ESC 2026 says an ICD should be considered in symptomatic heart failure with reduced ejection fraction (HFrEF), NYHA class II/III, of non-ischaemic aetiology with LVEF ≤35% despite ≥3 months of optimal foundational medical therapy (FMT), provided survival longer than 1 year with good functional status is expected, to reduce the risk of sudden death and all-cause death (Class IIa, Level B1).^[1]
• Its ischaemic row for the same symptomatic HFrEF, NYHA class II/III picture with LVEF ≤35% despite ≥3 months of optimal FMT and expected survival longer than 1 year with good functional status, unless there was an MI in the prior 40 days, is a recommendation, Class I, Level B1.^[1]
• ESC 2026 explains that selection should be based on LVEF thresholds established in RCTs and on aetiology, as individuals with non-ischaemic heart failure may derive less benefit than those with ischaemic heart disease.^[1]

Branch D — Does it still work, and for whom?

Examiner: Trials from 20 years ago — is the benefit still there?

• EU-CERT-ICD, a prospective controlled cohort that recruited 2327 patients with ischaemic or dilated cardiomyopathy and guideline ICD indications (2247 analysable; mean follow-up 2.4 years), found adjusted mortality associated with an ICD 27% lower than in controls without an ICD; its subgroup analyses indicated no benefit in diabetics or those aged 75 years or more.^[24]
• The PROFID pooled cohort of 32,214 post-MI patients with LVEF 35% or less (propensity-weighted; mean follow-up 43.7 months) found all-cause mortality hazard ratios for an ICD versus no ICD of 0.54 for patients included in 1995–2004, 0.67 in 2005–2014 and 0.89 (P=0.221) in 2015–2020 (prespecified periods by inclusion year).^[32]
• The PROFID EHRA randomized noninferiority trial is testing optimal medical therapy (OMT) without an ICD against OMT with an ICD, for all-cause mortality, in patients at least 3 months after MI with LVEF 35% or less, on OMT for at least 3 months and in NYHA class II or III.^[31]

Examiner: And the older patient?

• In the DANISH extended follow-up (randomized; nonischemic HFrEF; ICD versus usual clinical care; median 13.2 years), the ICD reduced sudden cardiovascular death in patients 70 years or younger but not in those over 70, while its effect on all-cause death was consistent regardless of age.^[27][19]
• ESC 2022 states that a simple age cut-off cannot guide the decision on ICD implantation adequately, and that the ICD indication in elderly patients should be based on a personalized assessment considering the overall condition and comorbidities.^[2]
• The ESC 2022 Recommendation Table 50 row says that in elderly patients in whom a benefit from the defibrillator is not expected due to the patient’s age and comorbidities, omission of ICD implantation for primary prevention may be considered (Class IIb, Level B).^[2]
• ESC 2026 says shared decision-making is essential when considering ICD implantation in patients with heart failure.^[1]

Examiner: What would an Australian guideline say?

• The NHFA/CSANZ 2018 heart failure guideline (National Heart Foundation of Australia and Cardiac Society of Australia and New Zealand), which predates the ESC 2022 and 2026 guidelines used here, says an ICD should be considered for primary prevention at least 1 month after MI with LVEF 30% or less, to decrease mortality (Strong recommendation FOR; high quality of evidence).^[9]
• NHFA/CSANZ 2018 says an ICD may be considered for primary prevention in HFrEF associated with dilated cardiomyopathy and LVEF 35% or less, to decrease mortality (Weak recommendation FOR; low quality of evidence).^[9]