Implantable cardioverter defibrillator outcome: beyond ejection fraction?

Over three decades of clinical experience have brought evidence that implantable cardioverter defibrillators (ICDs) should be considered as ‘gold standard’ therapy in prevention of sudden cardiac death. Implantable cardioverter defibrillators are currently indicated for secondary prevention in patients after cardiac arrest and life-threatening ventricular tachyarrhythmias or for primary prevention in patients considered at high risk of sudden arrhythmic death. The vast majority of ICDs for primary prevention are implanted in patients with significantly reduced left ventricular ejection fraction (LVEF) due to ischaemic or non-ischaemic cardiomyopathy. Although current eligibility for primary prevention ICD implantation is based merely on poor LVEF, we still experience an ongoing debate on how to identify patients who will benefit from ICD implantation. Furthermore, according to current guidelines, patients qualified for device implantation should have reasonable expectation of survival with a ‘good functional status’ for at least 1 year. Therefore, attempts are made to identify not only the ‘appropriate ICD candidate’, but also the one who may be ‘too sick’ to benefit from ICD therapy. The amount of benefit from an ICD is not uniform across the population eligible for ICD implantation, and may be limited due to early death from other than arrhythmic causes. Lack of survival benefit observed within the first year after implantation may be attributed to a higher risk of non-sudden death. Data from the MADIT II (Multicenter Automatic Defibrillator Implantation Trial), SCD-HeFT (Sudden Cardiac Death in heart Failure Trial), and DINAMIT (Defibrillation in Acute Myocardial Infarction Trial) studies demonstrated no separation of the Kaplan–Meier curves within the first 10–18 months after ICD implantation, suggesting that early mortality remains unaffected by ICD therapy. This may be explained by an increase of non-sudden death in the ICD arm of the trials. It should also be emphasized that recent decades brought significant changes in the mode of death among heart failure patients. As shown by Cubbon et al. who compared historical (1993–1995 UK-Heart Study-1) and contemporary (2006–2009 UK-Heart-2) cohorts the mortality rate during the first year declined from 12.5 to 7.8%, with sudden death contributing less to overall mortality and increasing the rate of non-cardiovascular deaths (27 vs. 14%). Older age, advanced heart failure, and other coexisting comorbidities, especially renal dysfunction, are the most common factors related to an unfavourable outcome in ICD recipients due to the competing risk of increased non-arrythmic mortality. Taking into account the number of comorbidities contributing to the increased mortality rate, it is unlikely that only one parameter will be suitable to identify patients at the highest risk. Therefore, efforts have been made to define a useful clinical model, based on a combination of risk parameters to predict sudden and nonsudden death. A number of approaches have been proposed to identify patients who either will or will not benefit from ICD therapy (Table 1). The most commonly used models in clinical practice are the Seattle Heart Failure Model (SHFM) and Chalson Comorbidity Index (CCI). The SHF multivariate prediction model for mortality in heart failure patients based on routinely collected clinical covariates was validated in both ICD as well as cardiac resynchronization therapy defibrillator (CRT-D) populations. In a study by Levy et al. 2487 patients from the SCD-HeFT trial were divided into five groups with a presumed increasing risk of all-cause mortality, based on a modified SHFM. While relative risk reduction from ICD therapy in the lowest risk quintile was 54%, patients at the highest quintile of predicted risk gained no significant benefit from ICD implantation. Modified CCI, originally based on 17 different categories of comorbidities, was validated in ICD recipients, and showed over three-fold higher mortality in patients with .3 non-cardiac comorbidities adjusted for age, gender, and heart failure. In patients with CRT-D devices, age-adjusted CCI ≥ 5 points were related with a nearly four-fold higher mortality.