The 2016 European heart failure guidelines recommend the use of implantable cardioverter-defibrillators (ICDs) to reduce the risk of sudden cardiac death (SCD) and all-cause mortality in patients with symptomatic heart failure (HF) and left ventricular ejection fraction (LVEF) of ≤35% despite optimal medical therapy (class I recommendation).1 This is a level A recommendation for patients with ischemic heart disease (IHD), as the evidence of benefit is stronger (data derived from multiple randomized clinical trials or meta-analyses2–4), whereas for non-ischemic heart disease (non-IHD), it is a level B recommendation (data derived from a single randomized trial or large non-randomized studies3–5). The studies on which these recommendations are based were published over ten years ago, and the medical therapy and cardiac resynchronization therapy (CRT) in use at that time were inferior to present-day care, meaning that the benefit of an ICD was greater at that time. Furthermore, after the publication of the European guidelines, the DANISH trial demonstrated that prophylactic ICD implantation in patients with non-IHD did not significantly reduce overall mortality, although it reduced the risk of SCD.6 As a result, the benefit of ICDs in patients with non-IHD has recently been questioned.7

The study aimed to determine whether the benefit of an ICD or cardiac resynchronization therapy-defibrillator (CRT-D) for primary prevention was different in patients with non-IHD compared to those with IHD in a real-world population.

Two hundred and eighty-one patients were studied, of whom 187 (66%) had IHD. The characteristics of the study population at the time of device implantation are shown in Table 1. Patients with IHD were older, more frequently male and with more cardiovascular risk factors and peripheral arterial disease, but there were no significant differences between the groups in terms of NYHA class, creatinine and BNP levels, LVEF and presence of AF. Patients with non-IHD more often received a CRT-D, while those with IHD more often received an ICD.

Primary endpoint

In a mean follow-up of 55±42 months, the primary composite endpoint of appropriate shock and all-cause mortality occurred in 150 patients (53%): in 105 (56%) in the IHD group and in 45 (46%) in the non-IHD group (p=0.15). Kaplan-Meier curve analysis (Figure 1) shows that the probability of appropriate shock or all-cause mortality was similar in the two study groups (log-rank p=0.10).

IHD: ischemic heart disease.'> IHD: ischemic heart disease.' title='Kaplan-Meier curve for the composite endpoint (appropriate shock and all-cause mortality). IHD: ischemic heart disease.'/>

Figure 1.

Kaplan-Meier curve for the composite endpoint (appropriate shock and all-cause mortality). IHD: ischemic heart disease.

(0.09MB).

Secondary endpoints

Twenty patients (21%) with non-IHD and 34 (18%) with IHD received appropriate shocks (p=0.64) during follow-up. The incidence of appropriate shock was 3.75 per person/year in the non-IHD group and 3.08 per person/year in the IHD group (p=0.19).

In patients who received appropriate shocks, the mean time to first shock was 3.5±2.7 years in the non-IHD group and 2.9±2.3 years in the IHD group (p=0.07). Kaplan-Meier curves for the occurrence of appropriate shock in both groups are shown in Figure 2A. The probability of appropriate shock was similar in the two groups (log-rank p=0.63).

IHD: ischemic heart disease.'> IHD: ischemic heart disease.' title='Kaplan-Meier curves for the likelihood of appropriate shock (A) and all-cause mortality (B). IHD: ischemic heart disease.'/>

Figure 2.

Kaplan-Meier curves for the likelihood of appropriate shock (A) and all-cause mortality (B). IHD: ischemic heart disease.

(0.12MB).

Multivariate regression analysis showed that LVEF (HR 1.02, 95% CI 1.01-1.04, p=0.001 for each percentage point increase) and age (HR 0.97, 95% CI 0.95-0.99, p=0.04 for each year) were independently associated with appropriate shock.

Thirty-six patients (38%) with non-IHD and 89 (47%) with IHD died during follow-up (p=0.19). The mortality rate was 7.5 deaths per 100 person/years in the non-IHD group and 9.0 deaths per 100 person/years in the IHD group (p=0.11).

Mean time to death was 3.9±2.8 years in the non-IHD group and 4.1±3.1 years in the IHD group (p=0.67). The probability of all-cause death over time was similar in the two groups (log-rank p=0.09) (Figure 2B).

Factors independently associated with increased risk of death were elevated creatinine, age, NYHA class and higher furosemide dose at the time of device implantation. The use of beta-blockers was a protective factor against death (Table 2).

Table 2.

Results of multivariate regression analysis for all-cause mortality.

	HR	95% CI	p
Creatinine	1.42	1.19-1.69	0.0001
Age	1.03	1.01-1.05	0.002
NYHA class	1.37	1.01-1.86	0.04
Furosemide dose	1.01	1.01-1.02	0.004
Use of beta-blocker	0.61	0.39-0.97	0.04

CI: confidence interval; HR: hazard ratio.

Of the 200 patients in whom the cause of death could be established (70% of deaths), 96 died of cardiovascular cause, 16 of SCD and 88 of non-cardiovascular causes. Causes of death were similar in the two study groups (Figure 3).

IHD: ischemic heart disease; SCD: sudden cardiac death.'> IHD: ischemic heart disease; SCD: sudden cardiac death.' title='Incidence of the main causes of death in the two study groups. CV: cardiovascular; IHD: ischemic heart disease; SCD: sudden cardiac death.'/>

Figure 3.

Incidence of the main causes of death in the two study groups. CV: cardiovascular; IHD: ischemic heart disease; SCD: sudden cardiac death.

(0.05MB).

Subgroup analysis

In the subgroup of patients aged <68 years (n=181), there was a higher probability of appropriate shock (Figure 4C and D). However, there were no statistically significant differences between the IHD and non-IHD groups in either older (aged ≥68 years) or younger (aged <68 years) patients (Figure 4).

IHD: ischemic heart disease; Older patients: aged ≥68 years; Younger patients: aged 68 years.'> IHD: ischemic heart disease; Older patients: aged ≥68 years; Younger patients: aged 68 years.' title='Kaplan-Meier curves for the primary and secondary endpoints by age-group. IHD: ischemic heart disease; Older patients: aged ≥68 years; Younger patients: aged 68 years.'/>

Figure 4.

Kaplan-Meier curves for the primary and secondary endpoints by age-group. IHD: ischemic heart disease; Older patients: aged ≥68 years; Younger patients: aged 68 years.

(0.22MB).

Subanalysis by type of device shows that patients receiving a CRT-D had a greater reduction in the combined endpoint of appropriate shock and all-cause mortality and in the secondary endpoint of all-cause mortality in the non-IHD group, the difference being clear from the beginning of follow-up (log-rank p=0.003 and p=0.002, respectively) (Figure 5A and E).

CRT-D: cardiac resynchronization therapy-defibrillator; ICD: implantable cardioverter-defibrillator; IHD: ischemic heart disease.'> CRT-D: cardiac resynchronization therapy-defibrillator; ICD: implantable cardioverter-defibrillator; IHD: ischemic heart disease.' title='Kaplan-Meier curves for the primary and secondary endpoints by type of defibrillator. CRT-D: cardiac resynchronization therapy-defibrillator; ICD: implantable cardioverter-defibrillator; IHD: ischemic heart disease.'/>

Figure 5.

Kaplan-Meier curves for the primary and secondary endpoints by type of defibrillator. CRT-D: cardiac resynchronization therapy-defibrillator; ICD: implantable cardioverter-defibrillator; IHD: ischemic heart disease.

(0.28MB).

There is still controversy concerning the benefits of defibrillators in patients with non-IHD. The clinical trials that showed benefit (SCD-HeFT4 and COMPANION8) also included patients with IHD and the reported effect was based on subgroup analysis of non-IHD subjects. Trials that showed no reduction in all-cause mortality (DEFINITE,5 CAT,9 AMIOVIRT10 and DANISH6) included non-IHD patients only. The fact that the DANISH trial was considerably more recent than the others means that other treatment for HF had improved, especially in terms of the number of patients treated with aldosterone antagonists and CRT, which obviously will have affected survival. The trial, published two years ago, showed no benefit of defibrillator implantation in reducing all-cause mortality (the trial's primary endpoint), although death from arrhythmia was reduced and all-cause mortality was in fact lower in the subgroup of patients aged <68 years (HR 0.64; 95% CI 0.45-0.90, p=0.01).

In an attempt to clarify this issue, a meta-analysis11 of the DANISH, CAT, DEFINITE and SCD-HeFT trials was recently published. According to the authors, primary prevention ICDs are efficacious at reducing all-cause mortality among patients with non-IHD, supporting the current guidelines.

Haugaa et al.12 studied the impact of the DANISH trial on clinical practice among European countries and concluded that the majority of centers did not implant ICDs for primary prevention on a regular basis in patients with non-IHD, despite current guidelines. By contrast, in patients with IHD, the guidelines for primary prevention ICD were followed, and no relevant change in indications were reported. It should be noted that most centers acknowledged having changed their indications for ICD implantation in non-IHD patients following the publication of the DANISH results. This is a good illustration of how a single trial can rapidly bring about widespread changes in clinical practice.

Our results suggest that implanted defibrillators have similar benefits in both IHD and non-IHD patients with regard to the incidence of appropriate shocks and all-cause mortality, which were similar in the two groups. Unlike in the DANISH trial, this benefit was also observed in older patients, in whom there were also no differences in the study endpoints.

The probability of appropriate shock was comparable in IHD and non-IHD patients, indicating that patients with non-IHD have a similar risk of fatal or potentially fatal arrhythmic events.

The incidence of appropriate shock in patients with non-IHD in our study was higher than in the DANISH trial6 (21% vs. 12%). This is most likely due to differences in the programming of detection zones for therapeutic shock, since this programming used to be more aggressive than that used nowadays. However, the programming was the same in our two study groups.

Our results also show that time to first appropriate shock is longer in patients with non-IHD (3.5±2.7 vs. 2.9±2.3 years in the IHD group), as seen in other studies,8 which highlights the need to control risk factors for non-sudden death and thus insure that the risk of death is not higher than that of appropriate shock.

There was also no difference between our study groups in all-cause mortality, except in the CRT-D subgroup, in which mortality was lower in those with non-IHD. These patients have been shown to have a better response to CRT-D than those with IHD, probably because of differences in the myocardial substrate.13 The lower mortality we observed in patients with non-IHD receiving a CRT-D, which was not seen in those in the same group who received an ICD, may thus be explained by their better response to CRT-D and hence better survival. This is also one of the explanations put forward for the low mortality in the DANISH trial, in which 60% of patients in both arms received a CRT-D.9,10

Regarding causes of death, a large proportion was of unknown cause, but SCD accounted for only 5% of deaths, which demonstrates the value of defibrillators in reducing arrhythmic SCD.

Despite its observational, retrospective and single-center nature, this study shines a light on the value of defibrillator use in real-world populations in recent years. As it was not ethically possible to compare groups with and without ICDs, since this would deprive some patients of a therapy that has been shown to be effective by numerous randomized trials and meta-analyses, we compared patients with IHD (in whom the benefits of an ICD have been clearly demonstrated) and those with non-IHD (concerning whom there is still debate). The study has some limitations, particularly in the programming of the ICD's detection zones (which in the early years of the study period tended to be more aggressive and hence led to shocks that could have been avoided), the fact that causes of death were not known for all patients, the lack of monitoring of medical therapy during follow-up, and the relatively small number of participants.

In this population, there were no differences in appropriate shocks or all-cause mortality in patients with non-IHD compared to those with IHD. Implantation of a defibrillator (ICD or CRT-D) thus appears to be beneficial in both groups.

The authors have no conflicts of interest to declare.