European Heart Journal | 2019
Defibrillator therapy for non-ischaemic cardiomyopathy: are sharks lurking beneath the rippling waters?
Abstract
Twenty-eight years ago, William C Roberts, long-time editor of the American Journal of Cardiology, posed this simple yet surprisingly provocative question to clinical researchers. His first suggestion was to keep the study population as homogeneous as possible. This month’s European Heart Journal features an article that highlights the challenges of this seemingly unassuming question. Implantable cardioverter defibrillators (ICDs) have been a mainstay of the treatment and prevention of sudden cardiac arrest (SCA) for a generation of patients with heart disease. Initially limited to those who had suffered a cardiac arrest, ICDs are now mostly prescribed as primary prevention of SCA for those with cardiomyopathies due to ischaemic heart disease and other aetiologies. In spite of conflicting findings in clinical trials and meta-analyses for patients with non-ischaemic cardiomyopathy (NICM) due to left ventricular (LV) systolic dysfunction, current guidelines in both Europe and the United States support ICD therapy for NICM patients as a Class I recommendation. In 2016, the randomized Danish Study to Assess the Efficacy of ICDs in Patients with Non-ischemic Systolic Heart Failure on Mortality (DANISH) trial reported that ICD therapy provided no improvement in overall or cardiovascular mortality over conventional medical therapy for NICM patients. However, patients with an ICD had a 50% relative risk reduction in death due to SCA, though absolute reduction in both groups was low. This has led some to suggest that current guidelines should be modified to reflect these new findings, while others are reluctant for such a change because some patients who would otherwise benefit from an ICD could be denied a potentially life-saving therapy. Knowing the inherent risks of SCA in this population, many of us can attest to our angst when we recommend against an ICD for a NICM patient. These same clinicians are also aware of the patient’s long journey with an ICD: discomfort, infection, mechanical failures, recalls, or the anxiety and depression with inappropriate shocks. We live in a healthcare world where our clinical decisions to benefit an individual are weighed against a collective economic impact on large healthcare systems. Although guidelines were never designed to become mandates, national regulators and insurance payers have idolized guidelines to drive the appropriateness of certain treatments. Our efforts to maximize the efficiencies of standardized care by treating patients as if they were all the same are thwarted when we repeatedly discover the fact that patients are not the same. Just as ripples on the water’s surface cannot tell us the type of fish swimming underneath, carrying a NICM diagnosis may not be enough to determine risk, and thus the appropriateness of prescribing or withholding ICD therapy. In the current issue of the European Heart Journal, Gutman and colleagues examined whether the presence or absence of LV scarring as seen by cardiac magnetic resonance (CMR) imaging could predict mortality in NICM patients with or without an ICD. The investigators leveraged the variability of practice patterns due to the selfdescribed ‘less prescriptive’ ICD therapy recommendations in Australia. In this study, 452 patients with NICM underwent CMR scoring for LV scar and were matched between those with or without an ICD. As in the DANISH trial, there was no significant difference in 3 year overall survival: 81% for those without an ICD and 89% for those with an ICD. However, the presence of LV scar as seen by CMR was strongly associated with higher overall mortality in both patients with and without an ICD. In patients with an LV scar, 3 year survival was 61% for those without an ICD and 86% in those with an ICD. The