ICD primary prevention: the trial evidence

Overview and definitions

ESC 2026 calls the residual risk of sudden cardiac death (SCD) still high.^[1] It says this warrants primary-prevention ICD implantation in patients with HFrEF deemed at high risk of SCD.^[1] Secondary prevention is the other setting. The ESC 2026 secondary-prevention row needs expected survival >1 year with good functional status, no reversible cause and no arrhythmia <48 h after an MI.^[1] With those conditions, it recommends an ICD after recovery from a ventricular arrhythmia causing haemodynamic instability, to reduce the risk of sudden death and all-cause death (Class I, Level A).^[1]

ESC 2026 bases ICD selection in HFrEF on LVEF thresholds established in randomized controlled trials (RCTs) and on underlying aetiology.^[1] Its reason is that individuals with non-ischaemic heart failure may derive less benefit than those with ischaemic heart disease.^[1] So a row makes sense only when you know who its trials enrolled.

New York Heart Association functional classification (ESC 2026 Table 8)

^[1]

NYHA class matters because trial entry and the rows are both split by it. ESC 2026 defines HFrEF as LVEF below 50% with symptoms and/or signs of heart failure.^[1]

Classification of the trial evidence

The randomized trials fall into families by how patients were selected and when the device went in.^{[11][14][19][17]} Reading them by family explains why some show benefit and others do not.

^{[11] [12] [14] [15] [20] [21] [16] [19] [27] [2] [13] [17] [18]}

COMPANION randomized (1:2:2) 1520 patients with advanced heart failure (NYHA class III or IV) due to ischemic or nonischemic cardiomyopathy and a QRS interval of at least 120 msec to optimal pharmacologic therapy alone or with CRT by pacemaker or pacemaker-defibrillator.^[22] Its abstract does not state the follow-up duration.^[22] EU-CERT-ICD, a prospective controlled cohort study in 44 centres in 15 European countries, followed patients with ischaemic or dilated cardiomyopathy and guideline indications for a prophylactic ICD, comparing an ICD group with controls without an ICD over a mean follow-up of 2.4 years.^[24] An analysis from the PROFID pooled cohort adds 32,214 patients with LVEF 35% or less after MI (7,477 with and 24,737 without a primary-prevention ICD; propensity overlap-weighted; mean follow-up 43.7 months), analysed for temporal trends across inclusion periods from 1995 to 2020.^[32] VEST randomized (2:1) patients with acute MI and an ejection fraction of 35% or less to a wearable cardioverter-defibrillator (WCD) plus guideline-directed therapy or guideline-directed therapy alone, with arrhythmic death at 90 days as the primary outcome.^[23]

^{[11] [12] [14] [15] [20] [21] [16] [19] [13] [17] [18]}

Epidemiology and risk

^{[1] [2] [11]}

ESC 2026 notes that the risk of SCD has declined over time.^[1] ESC 2022 says early revascularization strategies and modern heart failure medication have reduced the overall risk of SCD in heart failure patients since the trials were published.^[2] DEFINITE randomized 458 patients with nonischemic dilated cardiomyopathy, LVEF below 36% and premature ventricular complexes or NSVT to standard medical therapy with or without a single-chamber ICD (mean follow-up 29.0 months).^[16] In DEFINITE, two-year mortality in the standard-therapy group was 14.1 percent, an annual rate of 7 percent.^[16]

Pathophysiology: why a shock box changes survival only in some patients

ESC 2026 states that ICDs are effective in preventing SCD and prolonging survival in patients with HFrEF at high risk of SCD.^[1] ESC 2026 states that antiarrhythmic drugs such as amiodarone and dronedarone are effective in reducing ventricular arrhythmias but do not prolong survival and may shorten it.^[1]

The second idea is competing risk. ESC 2026 states that the effect of ICD treatment depends on the competing risk of non-sudden death, both cardiovascular and non-cardiovascular.^[1] So with many comorbidities and older age, the benefit of ICD therapy may be substantially lower.^[1] ESC 2022 says the SCD risk needs to be weighed against the individual’s competing risk of a non-arrhythmic death.^[2]

The early post-MI trials show competing risk in numbers. DINAMIT was a randomized, open-label trial of an ICD versus no ICD 6 to 40 days after MI, in patients with LVEF 0.35 or less and impaired cardiac autonomic function.^[17] Over a mean of 30 months, deaths due to arrhythmia were 12 with an ICD versus 29 without.^[17] Deaths from nonarrhythmic causes were 50 versus 29 (hazard ratio 1.75).^[17] The investigators concluded that the reduction in death due to arrhythmia was offset by an increase in death from nonarrhythmic causes.^[17]

IRIS was a randomized, open-label multicentre trial that enrolled 898 patients 5 to 31 days after MI and randomly assigned them to an ICD or medical therapy alone.^[18] Its entry criteria were 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.^[18] Over a mean of 37 months, there were fewer sudden cardiac deaths with an ICD than with medical therapy alone (27 versus 60).^[18] There were more nonsudden cardiac deaths with an ICD (68 versus 39).^[18]

^{[17] [18]}

Clinical presentation: the patient the trials enrolled

Recognize the candidate by matching the patient to a trial entry profile.^[14][15][19]

• Prior MI with LVEF 0.30 or less, no electrophysiological testing required: the MADIT-II population (randomized 3:2 to a defibrillator or conventional medical therapy; average follow-up 20 months).^[14]
• NYHA class II or III heart failure with LVEF 35% or less, ischaemic or nonischaemic: the SCD-HeFT population (randomized to placebo, amiodarone or an ICD, each with conventional therapy; median follow-up 45.5 months).^[15]
• Non-ischaemic dilated cardiomyopathy with LVEF below 36% and premature ventricular complexes or NSVT: the DEFINITE population (randomized to standard medical therapy with or without a single-chamber ICD; mean follow-up 29.0 months).^[16]
• Symptomatic systolic heart failure not caused by coronary artery disease with LVEF 35% or less, 58% with cardiac resynchronization therapy (CRT): the DANISH population (randomized to an ICD or usual clinical care; median follow-up 67.6 months).^[19]
• Coronary artery disease, LVEF 40% or less, asymptomatic unsustained VT and inducible sustained ventricular tachyarrhythmia: the MUSTT population (randomized to electrophysiologically guided antiarrhythmic therapy, including drugs and defibrillators, or no antiarrhythmic therapy; five-year estimates).^[12]

The 2017 American Heart Association/American College of Cardiology/Heart Rhythm Society (AHA/ACC/HRS) guideline describes MADIT-II as randomizing patients with LVEF 30% or less and prior MI to an ICD or not (average follow-up 20 months in the trial report).^[3][14] It reports that roughly one third of these patients had NYHA class I symptoms.^[3] It adds that a subgroup analysis supported a survival benefit of the ICD in them.^[3] It notes that patients with non-ischaemic cardiomyopathy (NICM) and class I symptoms were not included in SCD-HeFT (randomized to placebo, amiodarone or an ICD; median follow-up 45.5 months) or DANISH (randomized to an ICD or usual clinical care; median follow-up 67.6 months).^[3][15][19] Only 99 (21.6%) of the 458 DEFINITE patients (randomized to standard medical therapy with or without a single-chamber ICD; mean follow-up 29.0 months) had class I heart failure, so the 2017 guideline calls the benefit in such patients uncertain.^[3][16]

Differential diagnosis: which trial does this patient resemble?

Apply a trial result only to the kind of patient the trial enrolled.^[17][14][19] The table pairs each situation with the evidence that governs it.

Matching the patient to the trial

^{[17] [18] [1] [13] [14] [3] [4] [19] [2]}

Assessment and investigations

LVEF: the entry ticket

The 2017 AHA/ACC/HRS guideline calls LVEF a strong, independent predictor of SCD and cardiovascular mortality.^[3] It also calls LVEF a determinant of eligibility for a primary-prevention ICD.^[3] In SCD-HeFT (randomized; NYHA class II or III, LVEF 35% or less; ICD versus placebo, each with conventional therapy; median follow-up 45.5 months), it notes, ICD benefit did not depend on the modality used to assess LVEF: echocardiography, radionuclide angiography or contrast angiograms.^[3][15] Its first ischemic-heart-disease row (COR I, LOE A) is now history.^[3] That 2017 row was for LVEF 35% or less due to ischemic heart disease, NYHA class II or III HF despite GDMT, at least 40 days post-MI and at least 90 days postrevascularization, with meaningful survival of more than 1 year expected.^[3] For its ischemic-heart-disease row, the 2017 guideline bases the NYHA class II or III, LVEF ≤35% criterion on the survival benefit observed in SCD-HeFT and MADIT-II.^[3] Those randomized trials enrolled LVEF 35% or less with NYHA class II or III heart failure (SCD-HeFT, ICD versus placebo, each on conventional therapy, median 45.5 months) and prior MI with LVEF 0.30 or less (MADIT-II, defibrillator versus conventional medical therapy, average 20 months).^[15][14] On the same post-MI question, the 2025 ACC/AHA ACS guideline recommends an ICD in selected patients with LVEF 40% or less (Table 17) at least 40 days post MI and at least 90 days postrevascularization, to reduce death (COR 1, LOE A).^[4] It marks this 2025 row as adapted from the 2017 AHA/ACC/HRS guideline.^[4]

Timing of the measurement matters as much as the number. For pre-discharge LVEF 40% or less, a 2023 ESC acute coronary syndrome (ACS) row recommends repeat LVEF evaluation 6–12 weeks after an ACS, and after complete revascularization and the institution of optimal medical therapy.^[7] Its purpose is to assess the potential need for a primary-prevention ICD (ESC 2023 ACS, Class I, Level C).^[7] ESC 2022 had given a similar row for pre-discharge LVEF 40% or less: re-evaluation 6–12 weeks after MI to assess the potential need for primary-prevention ICD implantation (Class I, Level C), now dated by the 2023 row.^[2]

Electrophysiological study

MADIT, a randomized trial (prior MI, LVEF 0.35 or less, asymptomatic unsustained VT; defibrillator versus conventional medical therapy; average follow-up 27 months), selected patients at electrophysiological study by inducible, nonsuppressible ventricular tachyarrhythmia.^[11] MUSTT (coronary artery disease, LVEF 40% or less, asymptomatic unsustained VT; electrophysiologically guided antiarrhythmic therapy, including drugs and defibrillators, versus no antiarrhythmic therapy; five-year estimates) randomized only patients in whom sustained ventricular tachyarrhythmia was induced at electrophysiological study.^[12] The 2017 AHA/ACC/HRS guideline reports that ICD placement in MUSTT was not randomized.^[3] The treating physician chose it for VT that could not be suppressed with antiarrhythmic medication, and the average LVEF was 30%.^[3] ESC 2022 addresses CAD with LVEF 40% or less and asymptomatic NSVT.^[2] In that group, it says, inducibility by programmed electrical stimulation (PES) identifies patients who benefit from an ICD, independent from NYHA class.^[2]

Risk markers used in the early post-MI trials

DINAMIT, a randomized, open-label trial of an ICD versus no ICD 6 to 40 days after MI (mean follow-up 30 months), enrolled patients with LVEF 0.35 or less plus impaired cardiac autonomic function.^[17] DINAMIT took impaired autonomic function as depressed heart-rate variability or an elevated average 24-hour heart rate on Holter monitoring.^[17] IRIS, a randomized, open-label trial of an ICD versus medical therapy alone (mean follow-up 37 months), enrolled patients 5 to 31 days after MI.^[18] Its criterion 1 was LVEF 40% or less with a heart rate of 90 beats per minute or more on the first available electrocardiogram.^[18] Its criterion 2 was NSVT of 150 beats per minute or more on Holter monitoring, and patients could meet both criteria.^[18] In DINAMIT (randomized, open-label; 6 to 40 days after MI, LVEF 0.35 or less with impaired cardiac autonomic function), overall mortality with an ICD versus no ICD did not differ over a mean of 30 months (hazard ratio 1.08).^[17] In IRIS (randomized, open-label; 5 to 31 days after MI, on the LVEF and heart-rate or NSVT criteria), overall mortality with an ICD versus medical therapy alone was not reduced over a mean of 37 months (hazard ratio 1.04).^[18]

The landmark randomized trials

Randomized trials of a primary-prevention ICD: who, against what, for how long, and the result

^{[11] [12] [13] [14] [15] [16] [17] [18] [19] [27]}

Read the last column with its follow-up. In MADIT (randomized; prior MI, LVEF 0.35 or less, asymptomatic unsustained VT, inducible nonsuppressible ventricular tachyarrhythmia) and MADIT-II (randomized; prior MI, LVEF 0.30 or less), the ICD reduced all-cause mortality against conventional medical therapy over average follow-up of 27 and 20 months.^[11][14] In SCD-HeFT (randomized; NYHA class II or III heart failure, LVEF 35% or less), the ICD reduced all-cause mortality against placebo over a median of 45.5 months.^[15] In MUSTT (randomized; CAD, LVEF 40% or less, asymptomatic unsustained VT, inducible sustained ventricular tachyarrhythmia), electrophysiologically guided therapy (drugs or defibrillators) versus no antiarrhythmic therapy reduced the five-year primary end point of cardiac arrest or death from arrhythmia (25% versus 32%; relative risk 0.73).^[12] Its investigators concluded that guided therapy with implantable defibrillators, but not with antiarrhythmic drugs, reduces the risk of sudden death in high-risk coronary disease.^[12] In MADIT-II (randomized; prior MI, LVEF 0.30 or less; defibrillator versus conventional medical therapy; average follow-up 20 months), the effect on survival was similar in subgroups by age, sex, ejection fraction, NYHA class and QRS interval.^[14]

The non-ischaemic trials in detail

Non-ischaemic cardiomyopathy trials

^{[2] [20] [21] [16] [19] [27]}

ESC 2022 counts six RCTs of primary-prevention ICDs in dilated cardiomyopathy (DCM) with HFrEF.^[2] Five were published between 2002 and 2005 (CAT, AMIOVIRT, DEFINITE, COMPANION and SCD-HeFT), and DANISH came later.^[2] The first three enrolled only DCM patients; SCD-HeFT and COMPANION were larger and included patients with CAD and DCM.^[2] COMPANION differed from the others, it says, because it compared CRT-defibrillator, CRT-pacemaker and optimal medical therapy (OMT).^[2] The COMPANION report describes a 1:2:2 randomization of patients with NYHA class III or IV heart failure and a QRS interval of at least 120 msec; its abstract does not state the follow-up duration.^[22]

The 2017 AHA/ACC/HRS guideline reports that four prospective RCTs initially evaluated primary-prevention ICDs in NICM, and two were small studies terminated early due to a low event rate.^[3] In DEFINITE (randomized; nonischemic dilated cardiomyopathy with LVEF below 36% and premature ventricular complexes or NSVT; standard medical therapy with or without a single-chamber ICD; mean follow-up 29.0 months), it says, an ICD reduced the risk of SCD, with a trend toward reduced all-cause mortality.^[3][16] It also reports that SCD-HeFT (randomized; NYHA class II or III, LVEF 35% or less; placebo, amiodarone or an ICD, each with conventional therapy) included 792 NICM patients, with 5-year total mortality of 27% on placebo and 21% with an ICD (P=0.06).^[3][15]

^{[11] [14] [15] [13] [16] [17] [18] [19] [12]}

Management: the guideline rows built on these trials

The 2026 ESC heart failure guideline carries the ESC primary-prevention ICD rows for heart failure in its Recommendation Table 6, given here with every condition.^[1]

ESC 2026 heart failure guideline, Recommendation Table 6: primary-prevention and wearable-defibrillator rows

^[1]

MADIT-II randomized patients with a prior MI and LVEF 0.30 or less to a defibrillator or conventional medical therapy, with an average follow-up of 20 months.^[14] ESC 2026 cites it behind its ischaemic-aetiology row (Class I, Level B1), reporting that 1232 patients were randomized to an ICD or control and that ICD therapy reduced all-cause death by 31%.^[1] That row, an ICD to reduce the risk of sudden death and all-cause death, covers symptomatic HFrEF (NYHA class II/III) of ischaemic aetiology, unless there was an MI in the prior 40 days.^[1] It also needs LVEF ≤35% despite ≥3 months of optimal FMT, provided survival longer than 1 year with good functional status is expected.^[1] ESC 2026 adds that a meta-analysis of RCTs demonstrated an overall survival benefit, even in non-ischaemic heart failure; it names no comparator or follow-up period there.^[1]

For post-MI patients, the 2025 ACC/AHA ACS guideline recommends an ICD to reduce death in selected patients with LVEF 40% or less (its Table 17), at least 40 days post MI and at least 90 days postrevascularization.^[4] It grades this class of recommendation (COR) 1, level of evidence (LOE) A.^[4] Its footnote says the row is adapted from the 2017 AHA/ACC/HRS ventricular arrhythmia guideline.^[4]

2025 ACC/AHA ACS Table 17: LVEF and patient characteristics defining an indication for a prophylactic ICD in ischemic heart disease

^[4]

Reading Table 17 against the trials

^{[4] [14] [15] [3] [12]}

In the MADIT-II trial report, patients with a prior MI and LVEF 0.30 or less were randomized to a defibrillator or conventional medical therapy, with an average follow-up of 20 months.^[14] The 2025 ACC/AHA ACS supportive text reports that MADIT II showed a 31% reduction in all-cause mortality (hazard ratio 0.69) for prophylactic implantation of a cardiac defibrillator.^[4] It describes those patients as having LVEF 30% or less, at least 30 days after MI or 3 months after revascularization for MI, with an anticipated life expectancy longer than 1 year.^[4] The 2017 AHA/ACC/HRS guideline describes the early window differently: it says patients within 40 days of an MI were excluded from MADIT-II.^[3]

For non-ischaemic cardiomyopathy in North America, the 2022 AHA/ACC/Heart Failure Society of America (HFSA) heart failure guideline has the row.^[5] It recommends ICD therapy for primary prevention of SCD in nonischemic DCM with LVEF 35% or less and NYHA class II or III on chronic guideline-directed medical therapy (GDMT), to reduce total mortality (COR 1, LOE A).^[5] The patient must have a reasonable expectation of meaningful survival for more than 1 year.^[5] The same 2022 row also covers ischemic heart disease at least 40 days post-MI; for post-MI patients, the later 2025 ACC/AHA ACS row adds at least 90 days after revascularization.^[5][4] The 2017 AHA/ACC/HRS guideline says selection for an ICD must be individualized, and patients with serious comorbidities associated with a survival of under 1 year are generally not considered candidates.^[3]

Other guideline rows that name a primary-prevention ICD, by guideline

^{[2] [1] [6] [7] [4] [5] [3] [9]}

Older rows given only as history (each followed by a later row on the same question)

^{[2] [1] [7] [3] [4] [5]}

Timing: the 40-day, 90-day and 3-month rules

The 2017 AHA/ACC/HRS guideline bases its 40-day wait on two facts.^[3] Such patients were excluded from MADIT-II and SCD-HeFT, and 2 other RCTs showed no survival benefit from ICDs implanted early after an acute MI.^[3] ESC 2022 reports that early routine prophylactic ICD implantation in the first 40 days after MI did not reduce mortality in post-MI patients with reduced LVEF in two randomized trials, DINAMIT and IRIS.^[2] In its trial report, DINAMIT was an open-label comparison of an ICD with no ICD 6 to 40 days after MI, in patients with LVEF 0.35 or less and impaired cardiac autonomic function (mean follow-up 30 months).^[17] IRIS, also open-label, compared an ICD with medical therapy alone in patients enrolled 5 to 31 days after MI (mean follow-up 37 months).^[18] Its entry criteria were 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.^[18]

Timing rules by source

^{[1] [4] [3] [2] [7]}

Waiting has a price. ESC 2026 states that delaying ICD implantation carries a risk of SCD of 2%–4% by 6 months and ∼7% by 1 year.^[1]

^{[1] [7] [4] [17] [18]}

Specific scenarios

Early after MI: the wearable defibrillator

VEST randomized (2:1) patients with acute MI and an ejection fraction of 35% or less to a WCD plus guideline-directed therapy or guideline-directed therapy alone, with arrhythmic death at 90 days as the primary outcome.^[23] Of 2302 participants, arrhythmic death at 90 days occurred in 1.6% with the device and 2.4% without (relative risk 0.67, P=0.18).^[23] The device was worn for a median of 18.0 hours per day.^[23] Death from any cause, a secondary outcome for which the abstract gives no separate timeframe, was 3.1% with the device versus 4.9% in controls (relative risk 0.64, uncorrected P=0.04).^[23] Of the 48 participants in the device group who died, 12 were wearing the device at the time of death.^[23]

Based on the available data, the ESC 2022 task force (dated on this question) did not recommend routine use of the WCD in the early post-MI phase.^[2] It added that the device may be considered in selected post-MI patients deemed to be at high risk for SCD.^[2] Its Recommendation Table 15 row, now dated, said the WCD may be considered in the early phase after MI in selected patients (Class IIb, Level B); the later ESC 2023 ACS and ESC 2026 HF rows on the WCD are given in the tables above.^[2] ESC 2026 says a wearable defibrillator may be considered in patients deemed at especially high risk of SCD or as a bridge to decision for ICD implantation.^[1] It adds that no mortality benefit has been shown in RCTs.^[1]

With a wide QRS: CRT-D or CRT-P

COMPANION randomized 1520 patients with NYHA class III or IV heart failure from ischaemic or nonischaemic cardiomyopathy and a QRS interval of at least 120 msec, 1:2:2, to optimal pharmacologic therapy alone or with CRT by pacemaker or pacemaker-defibrillator.^[22] Its abstract does not state the follow-up duration.^[22] Against drug therapy, the pacemaker reduced death from any cause, a secondary end point, by 24 percent (P=0.059).^[22] The pacemaker-defibrillator reduced it by 36 percent (P=0.003).^[22] ESC 2026 likewise reports a reduction in all-cause death with CRT-D, but not with CRT-P, in COMPANION, where each was compared with optimal pharmacologic therapy alone.^[1][22]

CRT-D or CRT-P: the ESC rows

^{[8] [2] [6]}

Chagas cardiomyopathy

CHAGASICS was an open-label randomized trial in 13 centres in Brazil.^[30] It randomized 323 patients with chronic Chagas cardiomyopathy, positive Chagas serology, a Rassi risk score of at least 10 points (intermediate to high risk) and at least 1 episode of NSVT to an ICD or amiodarone.^[30] It stopped early for administrative reasons, short of the intended 1100 patients.^[30] At a median follow-up of 3.6 years, all-cause death was 38.2% with the ICD versus 38.6% with amiodarone (hazard ratio 0.86, P=.40).^[30] SCD was lower with the ICD than with amiodarone (3.8% versus 13.9%).^[30]

Complications and pitfalls

• Device infection: in DANISH (randomized, ICD versus usual clinical care; symptomatic systolic heart failure not caused by coronary artery disease, LVEF 35% or less; median 67.6 months), 4.9% of the ICD group and 3.6% of controls (P=0.29).^[19]
• Shocks: in VEST (randomized 2:1, WCD plus guideline-directed therapy versus guideline-directed therapy alone; acute MI, ejection fraction 35% or less; primary outcome at 90 days), 1.3% of the WCD group received an appropriate shock and 0.6% an inappropriate shock.^[23]
• Lifelong commitment: ESC 2026 says an ICD commits the patient to lifelong monitoring and generator replacements.^[1]
• ESC 2026 adds that the lifelong risk of complications should be considered in the decision to implant.^[1]
• Generator change: ESC 2026 reports ICD shocks in more than 15% after replacement in patients with no appropriate shock during their first generator.^[1]

ESC 2026 says the need for an ICD should be reconsidered over time and before every generator replacement.^[1] It names SCD risk, remaining lifespan, complication risk, competing risks of death from comorbidities, frailty and patient choice.^[1] Before generator replacement, ESC 2026 says a shared decision-making process with a cardiologist experienced in cardiac implantable electronic devices (CIEDs) and heart failure should be considered, to ensure that the balance between risks and benefits still favours a CIED (Class IIa, Level C).^[1]

Prognosis: what happens in the long run

Long-term follow-up and contemporary cohorts

^{[25] [26] [15] [27] [19] [24] [32]}

In MADIT-II (randomized, ICD versus non-ICD medical therapy; ischaemic left ventricular dysfunction; post-trial mortality data for all participants, median follow-up 7.6 years), the survival gain with the ICD persisted to 8 years.^[25] In the ischaemic subgroup of SCD-HeFT (randomized; NYHA class II or III, LVEF 35% or less; ICD versus placebo), it persisted at a median of 11 years.^[15][26] The DANISH extended follow-up (randomized, ICD versus usual clinical care; median 13.2 years) covered nonischemic HFrEF with LVEF 35% or less, NYHA class II–III (class IV if CRT was planned) and elevated natriuretic peptides.^[19][27] In it, the ICD lowered the long-term rate of sudden cardiovascular death, without a significant fall in all-cause death.^[27] In the SCD-HeFT extended follow-up (randomized to an ICD, amiodarone or placebo; NYHA class II or III, LVEF 35% or less; median 11.0 years), the nonischaemic subgroup showed no long-term ICD benefit versus placebo (hazard ratio 0.97).^[15][26] In the DANISH extended follow-up (same design, population and follow-up), the ICD effect versus usual clinical care was generally consistent across other key subgroups, including CRT use at baseline, whereas age modified its effect on sudden cardiovascular death (see Older patients).^[27]

^{[25] [15] [26] [27]}

Special populations

Older patients

MADIT-II randomized patients with a prior MI and LVEF 0.30 or less to a defibrillator or conventional medical therapy (average follow-up 20 months).^[14] ESC 2022 reports that in MADIT-II patients with ischaemic cardiomyopathy, a risk scheme of 5 clinical factors, including age over 70 years, predicted absence of long-term ICD benefit.^[2] It gives no follow-up period for that analysis.^[2] ESC 2022 reports a recent analysis of ICD patients enrolled in four MADIT studies, for which it gives no follow-up period: age below 75 years predicted VT/VF and age 75 years or more predicted non-arrhythmic mortality.^[2] In the DANISH extended follow-up (randomized, ICD versus usual clinical care; nonischemic HFrEF, LVEF 35% or less; median 13.2 years), the effect on all-cause death was consistent regardless of age.^[19][27] Sudden cardiovascular death fell in patients 70 years or younger (hazard ratio 0.38) but not in those over 70 (hazard ratio 1.27).^[27] Subgroup analyses of EU-CERT-ICD (prospective controlled cohort that recruited 2327 patients with ischaemic or dilated cardiomyopathy and guideline ICD indications, 2247 analysable; mean follow-up 2.4 years) indicated no ICD benefit at 75 years or more (adjusted hazard ratio 1.063; P for interaction 0.0902).^[24] ESC 2022 adds that a high comorbidity index is associated with less survival gain in primary- or secondary-prevention ICD patients.^[2] ESC 2022 concludes 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] Its Recommendation Table 50 has the row: 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 (ESC 2022, Class IIb, Level B).^[2]

Diabetes, kidney disease and women

ESC 2022 cites evidence that patients with end-stage renal disease, with diabetes, and elderly patients benefit less or not at all from a primary-prevention ICD.^[2] Subgroup analyses of the same EU-CERT-ICD cohort (ischaemic or dilated cardiomyopathy with guideline ICD indications; 2247 analysable patients; mean follow-up 2.4 years) indicated no ICD benefit in diabetics (adjusted hazard ratio 0.945; P for interaction 0.0887).^[24] ESC 2022 notes that women have been underrepresented in all primary-prevention trials and that data suggest they may benefit less.^[2]

Pregnancy and peripartum cardiomyopathy

ESC 2025 states that an ICD is indicated in pregnant women at high risk of SCD.^[10] It says implantation is not associated with a higher risk of major complications and can be performed safely, particularly after 8 weeks of gestation.^[10] ESC 2025 says a WCD may be considered in women with peripartum cardiomyopathy (PPCM) and LVEF below 35% (Class IIb, Level C).^[10] ESC 2022 explains that in PPCM, thresholds for early ICD implantation are higher because of a high rate of spontaneous recovery after delivery.^[2]

Evidence, guidelines and controversies

The non-ischaemic question after DANISH

The DANISH investigators state that earlier evidence for prophylactic ICDs in systolic heart failure not due to coronary artery disease rested primarily on subgroup analyses.^[19] DANISH randomized patients with symptomatic systolic heart failure not caused by coronary artery disease and LVEF 35% or less to an ICD or usual clinical care.^[19] Over a median follow-up of 67.6 months, it found no significant reduction in death from any cause (hazard ratio 0.87).^[19] ESC 2022 offers three potential explanations for that result.^[2] The first two potential explanations are the low SCD rate (4.3% in the ICD group and 8.2% in the control group) and excellent medical treatment (more than 90% on ACE inhibitors and beta-blockers, more than 50% on a mineralocorticoid receptor antagonist).^[2] The third is the high prevalence (58%) of CRT.^[2] The 2017 AHA/ACC/HRS guideline adds that, with CRT in 58% of both arms, DANISH should not be generalized to NICM patients ineligible for CRT.^[3] The 2022 AHA/ACC/HFSA guideline warns that subgroup analyses of the main ICD and CRT trials it cites were not their primary endpoints and should be interpreted with caution.^[5]

Pooled analyses of the primary-prevention ICD trials

^{[2] [28] [3] [29]}

NHFA/CSANZ 2018 (dated) reports a substudy of DANISH (randomized, ICD versus usual clinical care in symptomatic systolic heart failure not caused by coronary artery disease, LVEF 35% or less) with the prespecified endpoint of age; it gives no follow-up period for the substudy.^[9][19] It identified a significant reduction in mortality in patients aged under 70 years.^[9] ESC 2022 reports a further analysis of DANISH (randomized, ICD versus usual clinical care; nonischaemic symptomatic systolic heart failure, LVEF 35% or less) linking ICDs with reduced all-cause mortality at 70 years or younger (hazard ratio 0.70), but not over 70.^[2][19] It gives no follow-up period for that analysis.^[2] ESC 2022 concludes that DANISH emphasizes the need for further refining the indication in contemporary DCM patients, and that age and comorbidity need to be considered.^[2]

Is the ischaemic benefit still there?

ESC 2022 states that, although total mortality has been reduced, the relative reduction by the ICD is a consistent 27%.^[2] It says this has been corroborated in two prospective registry studies: EU-CERT-ICD (2327 European patients enrolled 2014–2018) and the SwedeHF registry (2610 Swedish patients enrolled 2000–2016).^[2] It gives no comparator or follow-up for either registry.^[2] In its own report, EU-CERT-ICD (prospective controlled cohort; ischaemic or dilated cardiomyopathy with guideline ICD indications; ICD group versus controls without an ICD; mean follow-up 2.4 years) found 27% lower adjusted mortality with an ICD (hazard ratio 0.731).^[24] The later PROFID pooled cohort, of patients with LVEF 35% or less after MI (ICD versus no ICD; propensity-weighted; mean follow-up 43.7 months), found the estimated all-cause mortality reduction smaller in more recent periods of inclusion (hazard ratio 0.89, P=0.221, in 2015–2020).^[32] The PROFID EHRA trial is a randomized, open-label, blinded-outcome noninferiority trial.^[31] It plans to recruit approximately 3,595 patients at least 3 months after MI, with LVEF 35% or less, at least 3 months of OMT and NYHA class II or III.^[31] It asks whether OMT without an ICD is not inferior to OMT with an ICD for all-cause mortality.^[31]

Regional differences

Regional rows compared (every class carries its body and year)

^{[1] [4] [9] [5]}

Australia and New Zealand

The National Heart Foundation of Australia and Cardiac Society of Australia and New Zealand (NHFA/CSANZ) 2018 guideline is the Australian heart failure guideline in the register.^[9] It is older than the ESC 2022 to 2026, 2022 AHA/ACC/HFSA and 2025 ACC/AHA rows it is compared with, so its rows are given as dated.^{[9][2][6][1][5][4]} The three NHFA/CSANZ 2018 primary-prevention rows sit in the table "Other guideline rows that name a primary-prevention ICD, by guideline", with their GRADE strength and quality of evidence.^[9] NHFA/CSANZ 2018 says ICD benefits must be balanced against living with a device, implant complications and psychological sequelae of inappropriate shocks for atrial tachyarrhythmias or lead fracture.^[9]

Guidelines checked

Guidelines checked for this topic (census 2026-10-09); a row called current, or the later one, is so among these guidelines

^{[1] [2] [6] [7] [8] [10] [4] [5] [3] [9]}

Exam pearls

• MADIT-II (randomized 3:2 to an ICD or conventional therapy; prior MI, LVEF 0.30 or less; average follow-up 20 months) did not require invasive electrophysiological testing for risk stratification.^[14]
• SCD-HeFT (randomized to conventional therapy plus placebo, amiodarone or an ICD; NYHA class II or III, LVEF 35% or less; median 45.5 months) used a conservatively programmed, shock-only, single-lead ICD.^[15]
• In SCD-HeFT (randomized; NYHA class II or III, LVEF 35% or less), over a median of 45.5 months, amiodarone had no favourable effect on survival against placebo.^[15]
• In SCD-HeFT (randomized; NYHA class II or III, LVEF 35% or less), the ICD gave an absolute fall in mortality of 7.2 percentage points versus placebo after five years in the overall population.^[15]
• DINAMIT (randomized, ICD versus no ICD; 6 to 40 days after MI, LVEF 0.35 or less and impaired cardiac autonomic function; mean 30 months): with an ICD, fewer arrhythmic deaths, offset by more nonarrhythmic deaths, and no difference in overall mortality.^[17]
• IRIS (randomized, ICD versus medical therapy alone; enrolled 5 to 31 days after MI, with either LVEF 40% or less and a heart rate of 90 beats per minute or more on the first available electrocardiogram, or NSVT of 150 beats per minute or more on Holter, or both; mean 37 months): with an ICD, fewer sudden (27 versus 60) but more nonsudden cardiac deaths (68 versus 39), and overall mortality not reduced.^[18]
• In the SCD-HeFT extended follow-up (randomized; NYHA class II or III, LVEF 35% or less; intention-to-treat), at a median of 11 years ICD benefit versus placebo persisted overall (hazard ratio 0.87), was attenuated after 6 years and was most evident in ischaemic heart failure and in NYHA class II at enrolment.^[26][15]
• ESC 2026 grades the ICD, given to reduce the risk of sudden death and all-cause death, Class I, Level B1 for ischaemic and Class IIa, Level B1 for non-ischaemic aetiology.^[1]
• Both ESC 2026 rows need symptomatic HFrEF, NYHA class II/III, LVEF 35% or less despite at least 3 months of optimal FMT, and expected survival longer than 1 year with good functional status; the ischaemic row also excludes an MI in the prior 40 days.^[1]

References32ShowHide

1. [1]Køber L, et al. 2026 ESC Guidelines for the management of heart failure. Eur Heart J, 2026.PMID 42661420
2. [2]Zeppenfeld K, et al. 2022 ESC Guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death. Eur Heart J, 2022.PMID 36017572
3. [3]Al-Khatib SM, et al. 2017 AHA/ACC/HRS Guideline for Management of Patients With Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. Circulation, 2018.PMID 29084731
4. [4]Rao SV, et al. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients With Acute Coronary Syndromes: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol, 2025.PMID 40013746
5. [5]Heidenreich PA, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation, 2022.PMID 35363499
6. [6]Arbelo E, et al. 2023 ESC Guidelines for the management of cardiomyopathies. Eur Heart J, 2023.PMID 37622657
7. [7]Byrne RA, et al. 2023 ESC Guidelines for the management of acute coronary syndromes. Eur Heart J, 2023.PMID 37622654
8. [8]Glikson M, et al. 2021 ESC Guidelines on cardiac pacing and cardiac resynchronization therapy. Eur Heart J, 2021.PMID 34455430
9. [9]Atherton JJ, et al. National Heart Foundation of Australia and Cardiac Society of Australia and New Zealand: Guidelines for the Prevention, Detection, and Management of Heart Failure in Australia 2018. Heart Lung Circ, 2018.PMID 30077227
10. [10]De Backer J, et al. 2025 ESC Guidelines for the management of cardiovascular disease and pregnancy. Eur Heart J, 2025.PMID 40878294
11. [11]Moss AJ, et al. Improved survival with an implanted defibrillator in patients with coronary disease at high risk for ventricular arrhythmia. Multicenter Automatic Defibrillator Implantation Trial Investigators. N Engl J Med, 1996.PMID 8960472
12. [12]Buxton AE, et al. A randomized study of the prevention of sudden death in patients with coronary artery disease. Multicenter Unsustained Tachycardia Trial Investigators. N Engl J Med, 1999.PMID 10601507
13. [13]Bigger JT Jr, et al. Prophylactic use of implanted cardiac defibrillators in patients at high risk for ventricular arrhythmias after coronary-artery bypass graft surgery. Coronary Artery Bypass Graft (CABG) Patch Trial Investigators. N Engl J Med, 1997.PMID 9371853
14. [14]Moss AJ, et al. Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced ejection fraction. N Engl J Med, 2002.PMID 11907286
15. [15]Bardy GH, et al. Amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. N Engl J Med, 2005.PMID 15659722
16. [16]Kadish A, et al. Prophylactic defibrillator implantation in patients with nonischemic dilated cardiomyopathy. N Engl J Med, 2004.PMID 15152060
17. [17]Hohnloser SH, et al. Prophylactic use of an implantable cardioverter-defibrillator after acute myocardial infarction. N Engl J Med, 2004.PMID 15590950
18. [18]Steinbeck G, et al. Defibrillator implantation early after myocardial infarction. N Engl J Med, 2009.PMID 19812399
19. [19]Køber L, et al. Defibrillator Implantation in Patients with Nonischemic Systolic Heart Failure. N Engl J Med, 2016.PMID 27571011
20. [20]Bänsch D, et al. Primary prevention of sudden cardiac death in idiopathic dilated cardiomyopathy: the Cardiomyopathy Trial (CAT). Circulation, 2002.PMID 11914254
21. [21]Strickberger SA, et al. Amiodarone versus implantable cardioverter-defibrillator:randomized trial in patients with nonischemic dilated cardiomyopathy and asymptomatic nonsustained ventricular tachycardia--AMIOVIRT. J Am Coll Cardiol, 2003.PMID 12767651
22. [22]Bristow MR, et al. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med, 2004.PMID 15152059
23. [23]Olgin JE, et al. Wearable Cardioverter-Defibrillator after Myocardial Infarction. N Engl J Med, 2018.PMID 30280654
24. [24]Zabel M, et al. Clinical effectiveness of primary prevention implantable cardioverter-defibrillators: results of the EU-CERT-ICD controlled multicentre cohort study. Eur Heart J, 2020.PMID 32372094
25. [25]Goldenberg I, et al. Long-term benefit of primary prevention with an implantable cardioverter-defibrillator: an extended 8-year follow-up study of the Multicenter Automatic Defibrillator Implantation Trial II. Circulation, 2010.PMID 20837894
26. [26]Poole JE, et al. Long-Term Outcomes of Implantable Cardioverter-Defibrillator Therapy in the SCD-HeFT. J Am Coll Cardiol, 2020.PMID 32703511
27. [27]Butt JH, et al. Long-Term Effect of ICDs in Nonischemic Heart Failure With Reduced Ejection Fraction: Extended Follow-Up Analysis of DANISH. J Am Coll Cardiol, 2025.PMID 41123523
28. [28]Al-Khatib SM, et al. Primary Prevention Implantable Cardioverter Defibrillators in Patients With Nonischemic Cardiomyopathy: A Meta-analysis. JAMA Cardiol, 2017.PMID 28355432
29. [29]Theuns DA, et al. Effectiveness of prophylactic implantation of cardioverter-defibrillators without cardiac resynchronization therapy in patients with ischaemic or non-ischaemic heart disease: a systematic review and meta-analysis. Europace, 2010.PMID 20974768
30. [30]Martinelli-Filho M, et al. Amiodarone or Implantable Cardioverter-Defibrillator in Chagas Cardiomyopathy: The CHAGASICS Randomized Clinical Trial. JAMA Cardiol, 2024.PMID 39356542
31. [31]Dagres N, et al. PRevention of sudden cardiac death aFter myocardial infarction by defibrillator implantation: Design and rationale of the PROFID EHRA randomized clinical trial. Am Heart J, 2026.PMID 40774643
32. [32]Sepehri Shamloo A, et al. Mortality reduction with implanted defibrillator for primary prevention of sudden death after Myocardial Infarction: temporal trends in the PROFID study. Eur Heart J, 2026.PMID 42663237