ICD primary prevention trials — structured written assessment

SAQ 1 (10 marks)

Practice scenario. A 62-year-old man had an anterior ST-elevation myocardial infarction (MI) 4 weeks ago, treated with primary percutaneous coronary intervention with complete revascularization on the day of the MI. His pre-discharge left ventricular ejection fraction (LVEF) was 30%. He has New York Heart Association (NYHA) class II symptoms, sinus rhythm, no documented ventricular arrhythmia and an expected survival of more than 1 year with good functional status. He asks whether he needs an implantable cardioverter-defibrillator (ICD) now.^[7][1]

1. Name two randomized trials of an ICD early after MI and give, for each, the enrolment window and the effect on overall mortality. (2)^[17][18]
2. Explain, with figures from one of these trials, why the reduction in arrhythmic deaths early after MI did not translate into lower overall mortality. (2)^[17]
3. State the 2026 European Society of Cardiology (ESC) heart failure row on a primary-prevention ICD early after MI, with its class and level. (1)^[1]
4. When should his LVEF be reassessed under the 2023 ESC acute coronary syndrome (ACS) guideline, and why? (1)^[7]
5. Which randomized trial supports an ICD for him if his LVEF remains 0.30 or less once he is past the early post-MI window, and what did it show? (2)^[14][4]
6. What did the VEST trial show about a wearable cardioverter-defibrillator (WCD) in this early period? (2)^[23]

Model answers — SAQ 1

1. DINAMIT randomized patients 6 to 40 days after MI (LVEF 0.35 or less with impaired cardiac autonomic function), open-label, to an ICD or no ICD and, over a mean of 30 months, found no difference in overall mortality (hazard ratio 1.08) (1 mark).^[17] IRIS, also randomized, 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; over a mean of 37 months overall mortality was not reduced (hazard ratio 1.04) (1 mark).^[18]
2. In DINAMIT, over a mean of 30 months, there were 12 deaths due to arrhythmia with an ICD versus 29 without, but 50 versus 29 deaths from nonarrhythmic causes (1 mark).^[17] The investigators concluded that the reduction in death due to arrhythmia was offset by an increase in death from nonarrhythmic causes; IRIS (randomized; 898 patients enrolled 5 to 31 days after MI who met criterion 1, criterion 2 or both, as defined in answer 1; ICD versus medical therapy alone; mean 37 months) also had fewer sudden (27 versus 60) but more nonsudden cardiac deaths (68 versus 39) with an ICD (1 mark).^[17][18]
3. ESC 2026: ICD implantation for primary prevention is not recommended within 40 days of an MI as implantation during this period does not improve prognosis (Class III, Level B1) (1 mark).^[1]
4. In patients with pre-discharge LVEF 40% or less, the 2023 ESC ACS guideline recommends repeat evaluation of the LVEF 6–12 weeks after an ACS (and after complete revascularization and the institution of optimal medical therapy) to assess the potential need for primary-prevention ICD implantation (Class I, Level C) (1 mark).^[7]
5. MADIT-II, which randomized 1232 patients with prior MI and LVEF 0.30 or less (3:2) to a defibrillator or conventional medical therapy without requiring electrophysiological testing; the 2025 ACC/AHA ACS guideline describes the MADIT II patients as at least 30 days after MI or 3 months after revascularization for MI (1 mark).^[14][4] Over an average of 20 months, mortality was 14.2% with the defibrillator versus 19.8% with conventional medical therapy (hazard ratio 0.69, 95% CI 0.51 to 0.93) (1 mark).^[14]
6. VEST randomized 2302 patients with acute MI and ejection fraction 35% or less, 2:1, to a WCD plus guideline-directed therapy or guideline-directed therapy alone (1 mark).^[23] Arrhythmic death at 90 days was 1.6% with the WCD versus 2.4% with guideline-directed therapy alone (relative risk 0.67, P=0.18), not significantly lower; the device was worn for a median of 18.0 hours per day (1 mark).^[23]

SAQ 2 (10 marks)

Practice scenario. A 72-year-old woman has symptomatic heart failure with reduced ejection fraction (HFrEF) of non-ischaemic aetiology, New York Heart Association (NYHA) class II symptoms and a left ventricular ejection fraction (LVEF) of 30% after 6 months of optimal foundational medical therapy (FMT). She has type 2 diabetes, a QRS duration of 100 ms and an expected survival of more than 1 year with good functional status. She has read that the DANISH trial was negative and asks whether an ICD is worthwhile.^[1][19]

1. Describe the DANISH trial population and design. (2)^[19]
2. Give the DANISH results for death from any cause and for sudden cardiac death. (2)^[19]
3. Give two explanations that ESC 2022 offers for the DANISH result. (2)^[2]
4. What does the DANISH extended follow-up show about age? (1)^[27]
5. What did EU-CERT-ICD subgroup analyses show for diabetes and for age 75 years or more, and what kind of study was it? (2)^[24]
6. State the 2026 ESC heart failure row that applies to her, with its class and level. (1)^[1]

Model answers — SAQ 2

1. DANISH was a randomized controlled trial of 1116 patients with symptomatic systolic heart failure not caused by coronary artery disease and LVEF 35% or less, assigned to an ICD (556) or usual clinical care (560) (1 mark).^[19][27] In both groups, 58% of the patients received cardiac resynchronization therapy (CRT) (1 mark).^[19]
2. After a median follow-up of 67.6 months, death from any cause occurred in 21.6% with an ICD versus 23.4% with usual clinical care (hazard ratio 0.87, P=0.28), not significantly lower with an ICD (1 mark).^[19] Sudden cardiac death occurred in 4.3% with an ICD versus 8.2% with usual clinical care (hazard ratio 0.50, P=0.005) (1 mark).^[19]
3. ESC 2022 offers the low SCD rate in the trial, excellent medical treatment (more than 90% on ACE inhibitors and beta-blockers, more than 50% on a mineralocorticoid receptor antagonist) and the high prevalence (58%) of CRT as potential explanations; any two (1 mark each).^[2]
4. In the DANISH extended follow-up (1116 patients with nonischemic HFrEF randomized to an ICD or usual clinical care), at a median of 13.2 years, the effect of the ICD on all-cause death was consistent regardless of age, but sudden cardiovascular death was reduced in patients 70 years or younger (hazard ratio 0.38) and not in those over 70 (hazard ratio 1.27) (1 mark).^[27][19]
5. EU-CERT-ICD was a prospective, investigator-initiated, controlled cohort study, not a randomized trial: it recruited 2327 patients with ischaemic or dilated cardiomyopathy and guideline indications for a prophylactic ICD, and data from 2247 were analysable (1516 in the ICD group, 731 controls without an ICD, compared using multivariable models and propensity scoring for adjustment), over a mean follow-up of 2.4 years (1 mark).^[24] Its subgroup analyses indicated no ICD benefit in diabetics (adjusted hazard ratio 0.945) or those aged 75 years or more (adjusted hazard ratio 1.063) (1 mark).^[24]
6. ESC 2026: an ICD should be considered in symptomatic HFrEF (NYHA class II/III) of non-ischaemic aetiology with LVEF ≤35% despite ≥3 months of optimal 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 mark).^[1]