The primary objective of DILEMMA study is to assess whether the "heart failure optimal therapy alone (HFOT)" strategy is non inferior to the "HFOT+ICD" strategy in terms of overall survival 48 months after randomization, in patients ≥ 70 years with an ICD indication for primary prevention of SCD whether there is an indication for cardiac resynchronization therapy or not.

Rationale Although not recognized by guidelines, there is no available data demonstrating the benefit of Implantable Cardioverter Defibrillator (ICD) for primary prevention strategy of Sudden Cardiac Death (SCD) in elderly. Nevertheless, ICD are currently implanted in this population by extending the results obtained in randomized trials involving younger subjects to the elderly. Finally, if the absence of implantation in the elderly was not inferior to the implantation of such a device, the non-implantation would avoid the device-related complications and decrease the health costs.

Main objective The primary objective of DILEMMA study is to assess whether the "heart failure optimal therapy alone (HFOT)" strategy is non inferior to the "HFOT+ICD" strategy in terms of overall survival 48 months after randomization, in patients ≥ 70 years with an ICD indication for primary prevention of SCD whether there is an indication for cardiac resynchronization therapy or not.

Design This is a 2-arm parallel non-inferiority, randomized, open label, multicenter trial. 730 patients will be included over 4 years. Follow up will last 4 years.

Inclusion Criteria:

• ≥70 years old,
• Left ventricular ejection fraction ≤ 35%, assessed by echocardiography, single-photon emission computed tomography and radionuclide ventriculography and/or cardiac magnetic resonance (CMR) (assessed at inclusion or within the 6 weeks prior to inclusion).
• NYHA class II or III
• Heart failure HFOT ≥ 3 months (Possible exceptions : possibility of including patients who do not tolerate the maximum dose or who do not tolerate all four therapeutic classes, as well as the possibility of not waiting for three months when one of the classes is discontinued or when dosage adjustments are made due to poor tolerance ...)
• Providing informed consent
• Affiliated to a French Health Insurance system.

Exclusion Criteria:

• Enrolled in or planning to enroll in a conflicting interventional trial (trial evaluating the interest of ICD or modifying HFOT outside the last ESC Guidelines)
• Prior unstable sustained ventricular arrhythmia requiring external cardioversion
• Myocardial infarction within the 40 days
• Coronary artery intervention (catheter or surgical) within 90 days
• History of syncope in the previous 6 months
• Advanced cerebrovascular disease (cerebrovascular disease with functional repercussions or effect on the patient's autonomy)
• Cognitive impairment leading to the incapacity of consent
• Any disease other than cardiac disease (e.g. cancer, uremia, liver failure), associated with a likelihood of survival less than 1 year.
• Patient under tutorship, curatorship, or legal safeguard
• Persons deprived of their liberty by judicial or administrative decision (prisoner)

Individual participant data (IPD) that underlie results in publication could be shared. IPD detailed in the protocol of a planned metaanalysis could be shared

Two years after the last publication

Data sharing must be accepted by the sponsor and the PI based on a scientific project and scientific involvement of the PI team. Collaboration will be fostered.

Data sharing must respect the agreements made with funders.

Teams wishing obtain IPD must meet the sponsor and PI team to present scientific (and commercial) purpose, IPD needed, format of data transmission, and timeframe. Technical feasability and financial support will be discussed before mandatory contractual agreement.

Processing of shared data must comply with European General Data Protection Regulation (GDPR).

Three experts in cardiac electrophysiology and ICD will be in charge of uniformly approving, while blinded to the study group, ICD tracings, cardiovascular events, including specific causes of death, in order to validate the primary and secondary endpoints.