Should we Rethink the Indications for Implantable Cardioverter-Defibrillators in Non-Ischaemic Dilated Cardiomyopathy?

Heart failure (HF) represents a major cause of morbidity and health care expenditure globally,1 and increases the risk for sudden cardiac death (SCD). In the USA, about 325 000 adults die from SCD every year, accounting for half of all cardiovascular deaths.2 The implantable cardioverter-defibrillator (ICD) is currently indicated for primary SCD prophylaxis in patients with severe left ventricular systolic dysfunction, irrespective of the underlying aetiology.3, 4 The recent results of the Danish Study to Assess the Efficacy of ICDs in Patients with Non-Ischaemic Systolic Heart Failure on Mortality (DANISH)5 raised novel interest in the longstanding debate on the efficacy of ICD implantation in non-ischaemic dilated cardiomyopathy (NIDCM). In fact, the DANISH study found only a slightly lower, non-statistically significant (P = 0.28) reduction in the primary endpoint of all-cause mortality in NIDCM patients randomized to ICD implantation (21.6%) compared with the control group (23.4%) treated with optimal medical management but no ICD implantation over a median follow-up of 67.6 months (Figure 1A). These results add further evidence to the findings of previous studies with conflicting results, such as the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT)6 and the Defibrillators in Non-Ischaemic Cardiomyopathy Treatment Evaluation (DEFINITE) study.7 In order to resolve this still debated issue, in-depth analyses of these studies are required to disentangle the reasons underlying the differences observed across the results of individual studies. In fact, although the authors of the DANISH study concluded that ICD implantation is not associated with an overall survival benefit in patients with NIDCM,5 perusal of the secondary endpoints shows that the frequency of SCD was almost halved in the ICD arm (4.3% vs. 8.2%; P = 0.005) (Figure 1B). This latter finding raises doubts about the current strategy of defibrillator implantation for the primary prevention of SCD. In fact, although not statistically significant, hazard ratios (HRs) for all-cause death and cardiovascular death were 0.87 and 0.77, respectively, both in favour of ICD implantation.5 However, the benefits to be gained from ICD use should be weighed against the risks for complications and side effects. Moreover, the differential impact of ICD on different modes of death should also be taken into account in studies evaluating ICD effectiveness.8 In fact, a comprehensive evaluation of all key elements in play represents the best strategy for identifying predictors of success and failure of ICD therapy. Additionally, the results of the DANISH study5 highlight the importance of stratifying patients according to level of risk. A subgroup analysis in the DANISH study revealed a significant reduction of the primary endpoint in specific patient categories, such as in younger patients (<59 years; P = 0.02) and in those with less severe HF (NT-proBNP <1177 pg/mL; P = 0.02),5 although it should be noted that these results do not provide sufficient evidence to suggest that NT-proBNP could be efficiently used to identify patients most likely to benefit from ICD implantation. However, in line with previous observations, these findings highlight the importance of stratifying patients by level of risk, such as by using a risk score model that includes New York Heart Association (NYHA) class, age, renal function, QRS width, and concomitant atrial fibrillation.9 Similarly, studies performed in differently selected populations showed that the usefulness of ICD for primary prevention may be lower in older patients, women and in patients with multiple comorbidities.10-12 In line with these results, a very recent report including 2331 patients demonstrated that ICD use was associated with a 37% reduction in all-cause mortality (P = 0.002) after adjustment for the clinical risk profile, performed by means of the Seattle Heart Failure Model (SHFM).13 The SHFM is a comprehensive scoring system, developed and validated in large HF patient datasets,14 which has excellent prognostic efficacy and the ability to differentiate among various modes of death.15 The SHFM has been shown to predict all-cause mortality. Although the prediction of sudden death and pump failure death may simply mirror global risk, the relative proportions of deaths that represented sudden death and pump failure death, respectively, were very different among patients with different SHFM scores.16 The Seattle Proportional Risk Model (SPRM), derived from the SHFM, was able to identify patients' proportional risk for death from SCD, such as in males, younger patients and those of better NYHA functional class status.13 Of note, cardiac resynchronization therapy (CRT) itself was able to prevent life-threatening ventricular arrhythmias, most probably through its effect on left ventricular remodelling.17 However, unfortunately, no final consensus has been reached on this issue. In fact, it is still unknown whether CRT-D may be superior to standard CRT as neither previous trials18, 19 nor the present study5 directly compared the two approaches. Interestingly, in line with the results of the DANISH study,5 among patients enrolled in the HF-ACTION (Participants in Heart Failure: A Controlled Trial Investigating Outcomes of Exercise Training) study, those with an SPRM estimated proportional risk for SCD of ≤32% had no reduction in all-cause mortality with ICD implantation.13 Despite their appeal, these risk models are very complex and may be impractical for use in routine clinical practice. However, they represent powerful research tools and should be exploited to derive novel risk stratification tools that make use of as much as possible of their predictive potential in the simplest way available. The optimal selection of candidates for ICD implantation is an important clinical issue that merits attention. In fact, consideration of the mortality benefit of ICD implantation may be futile in patients with NIDCM with a low proportional risk for SCD, although these patients are exposed to all risks associated with ICD implantation. In fact, device-related infections were registered in 4.9% of patients in the ICD arm compared with 3.6% of the control group, which included patients receiving an upgrade from standard pacemaker to CRT within the study5 (Figure 1C). Interestingly, the increase in device-related infection associated with ICD implantation was much higher among patients not receiving CRT (5.1% vs. 0.8%; HR 6.35; P = 0.006). Furthermore, inappropriate shocks occurred in as many as 5.9% of patients in the ICD group (Figure 1D). Further strategies to improve risk stratification include the adoption of imaging techniques, which may be of particular help in selected subgroups. In this regard, cardiac magnetic resonance imaging provides valuable diagnostic and prognostic insights in the setting of non-ischaemic left ventricular dysfunction, in which a recent meta-analysis has shown that late gadolinium enhancement is associated with increased risk for all-cause mortality,20 and another report has highlighted a strong correlation between diffuse myocardial fibrosis detected by native T1 mapping and adverse heart failure hospitalization and SCD in NIDCM patients.21 Another promising diagnostic technique, myocardial 123I-metaiodobenzylguanidine (123I-mIBG) washout, has been shown to be an important measure of cardiac sympathetic innervation in several disease states, including HF. The use of 123I-mIBG imaging has the potential to contain or reduce the use of ICDs by identifying patients at lower risk for SCD who are likely not to derive meaningful benefit from an ICD.22 Although ICDs are cost-effective as currently used, screening would help to improve their cost-effectiveness and allow policymakers to make better use of limited health care resources.23 In conclusion, much can be learned from the results of the studies conducted so far to test the efficacy of ICD implantation for primary prevention in NIDCM patients, including the more recent DANISH trial.5-7, 24 In particular, further clinical studies should be encouraged to assess new clinical score-based risk classification systems and novel functional imaging techniques. This work was supported in part by a grant from the Italian Ministry of Education, University and Research (MIUR) (PON03PE_00009_4 OPTIMA Cardiopaths). Conflict of interest: none declared.