In 2006 [4], update in 2015 [2], guidelines for the indications of ICD implantation were update from 2006, in the light of the publication of two large studies MADITII [9] and DINAMIT [10] in addition to one meta-analysis [11] considering all the ten studies published in the domain: MADIT I [12], CABG-Patch, MUSTT [13], MADIT II [9], CAT [14], AMIOVIRT [15], COMPANION [16], DEFINITE [17], SCD-HeFT [18], and DINAMIT [10].

A complete analysis of these data has allowed to define the indications for these devices [2–5]. Thus the patients who reaped the biggest benefit from the ICD in primary prevention are those who present a chronic left ventricular dysfunction at distance from an acute myocardial infarction or those within the context of a nonischemic dilated cardiomyopathy. Only four studies did not demonstrate the beneficial character of the ICD. The aforementioned studies, however [10, 14, 15], were carried out on a limited number of patients (CAT [14] and AMIOVIRT [15]) or in a context of recent myocardial infarction (between the 6th and the 40th day for DINAMIT [10]) or concerned implantations following coronary bypass surgery (CABG-Patch), a treatment which decreases considerably the relative risk of sudden death in the control groups.

On the other hand, the MADIT I [12] study carried out on a group of 196 patients in a context of ischemic heart disease and prior infarct with a left ventricular ejection fraction ≤35% highlighted a reduction in the annual mortality of 54% compared to the control group [12]. Coming from the same team, the MADIT II study concerning a larger sample size of 1232 patients, in a context of ischemic heart disorder with left ventricular ejection fraction ≤30% estimated at least 1 month after an infarct, reinforced these results with a reduction in annual mortality of 31% compared to the control group [9]. The MUSTT study, regarding a group of 704 patients, showed a reduction in mortality of 51% in the patients implanted in comparison to the control group which consisted of coronary patients for whom a ventricular hyper excitability could be medically treated without resort to a ICD [13]. The study COMPANION, concerning 1520 patients with ejection fraction ≤35%, 59% of whom were coronary, confirmed the advantage of the ICD with a reduction of 36% in the annual mortality when it was associated with a biventricular pacemaker compared to the patients treated medically or by a biventricular resynchronization only [17]. Regarding the group with nonischemic dilated cardiomyopathy and a left ventricular ejection fraction ≤35%, the DEFINITE study concluded a decrease in the annual mortality of 35% [17]. The largest sample size came from the SCD-HeFT study with 2521 cases with cardiac insufficiency and a left ventricular ejection fraction ≤35% of ischemic etiology for 52% of the cases [18]. In this study, the reduction in the annual mortality with regard to the control group was 23%.

Finally, a meta-analysis which ensues from the analysis of these ten randomized studies concluded a relative reduction of 25% and an absolute reduction of 7.9% in the global mortality on a 2- to 4-year follow-up of the patients with a ICD [11].

In consequence, the recommendations concerning the implantation of ICD in primary prevention of sudden cardiac death put forward are: [2–5]

In patients with cardiac failure who remains symptomatic (NYHA III or IV) under optimal medical treatment, with left ventricular dysfunction (LVEF ≤35%) and duration of the QRS >120 ms, a biventricular pacemaker is recommended in association with the ICD for cardiac resynchronization.

Secondary prevention by the ICD allowed a 27% reduction in the global mortality of the patients concerned [19]. Moreover, in the AVID study relating to 1016 patients having presented a cardiac arrest during a ventricular tachycardia or fibrillation, the ICD resulted in a decrease in mortality, respectively, of 39, 27, and 31% at 12, 24, and 36 months, in comparison to the control group treated by amiodarone alone [20]. In the CIDS study, including 658 cardiac arrest survivors due to a ventricular tachycardia or fibrillation, the ICD tended to decrease, in a not significant way, the relative risk of mortality of 33% after 5 years in comparison to the control group benefiting from an anti-arrhythmic treatment by amiodarone alone [21]. The CASH study including also a limited group of 288 patients and realized according to the same protocol as the two previous studies concluded to a nonsignificant decrease in the mortality of 23% at 9 years in comparison to the groups treated medically by amiodarone or metoprolol alone [19]. Due to the restrained sample size, these two studies were not significant. A meta-analysis merging these three studies allowed to conclude significantly in a reduction of 27% in the global mortality of the ICD group and more particularly when left ventricular ejection fraction is ≤35%. The recommendations concerning the indications ICD implantation in secondary prevention of sudden death are as follows: [ 2–5]