The quest for perfection: the contribution of the electrocardiogram to the prevention of sudden death in hypertrophic cardiomyopathy

Haghjoo et al.1 report a retrospective electrocardiographic study investigating the prognostic value of the 12-lead electrocardiogram (ECG) in 173 patients with hypertrophic cardiomyopathy (HCM). During a follow-up period of 50 months, 6% of the study group experienced the composite endpoint of sudden cardiac death (SCD), aborted SCD, or appropriate internal cardioverter defibrillator (ICD) shock. Univariable analysis confirmed the association of conventional risk markers such as non-sustained ventricular tachycardia, severe left ventricular (LV) hypertrophy, and a history of unexplained syncope with the composite endpoint, but also identified ST-segment depression in the high lateral leads (I and aVL) as a predictor of SCD. When the relative contribution of each predictor to the composite endpoint was examined with multivariable analysis, only ST-segment depression in the high lateral leads and syncope emerged as independent predictors. In spite of its still powerful reputation as a disease typified by premature death, most contemporary survival studies indicate that HCM is associated with a low annual risk of SCD. Nevertheless, it is clear that the susceptibility to ventricular arrhythmia varies substantially between individuals and so risk stratification remains central to clinical management. Patients with HCM develop ventricular arrhythmia for many reasons. Myocardial hypertrophy causes dispersion of repolarization and refractoriness making the myocardium vulnerable to triggered arrhythmias while the accompanying disarray and fibrosis create areas of conduction block predisposing to re-entry arrhythmias. In addition, abnormal ion fluxes during repolarization of cardiomyocyte cell membranes cause after-depolarizations and triggered activity. This arrhythmogenic substrate is further influenced by myocardial ischaemia, caused by intimal and medial smooth muscle cell hyperplasia of the small intramural coronary arteries, LV outflow tract obstruction, diastolic dysfunction, and abnormal autonomic function. 2 The current approach to risk stratification for SCD relies on an assessment of a small number of routinely collected clinical parameters. International guidelines rank these risk factors into ‘major’ (family history of SCD, abnormal blood pressure response to exercise, maximal wall thickness ≥30 mm, non-sustained ventricular tachycardia, and unexplained syncope) and ‘possible’ [myocardial ischaemia, high-risk mutations, competitive exercise, late gadolinium enhancement in cardiac magnetic resonance (CMR) imaging, and LV outflow tract obstruction]. 3 However, with the exception of aborted SCD, there is very little evidence to suggest that any one risk factor is more predictive than another. 3 , 4 Despite their modest predictive value for SCD when used in isolation, the aggregation of these risk factors in an individual patient reflects the severity of the underlying myocardial disease and thus the risk for SCD. Patients with two or more risk factors have a 72% 6-year survival with a 4–5% annual risk of SCD and are thus candidates for ICD implantation. 3 , 4 In contrast to their weak positive predictive value, these risk factors have a very high negative predictive value, and patients with no risk factors for SCD have a good prognosis and do not required aggressive primary prevention as the mortality rate is <1% per year. 3 , 4 However, there is less certainty with the management of patients with a single risk factor. On the one hand, 5% of SCD occurs in patients with a single risk factor. On the other hand, 25% of HCM patients have a single major risk factor, the majority of whom will not die suddenly with a 6-year survival rate of 93%. 2 , 4 The clinical challenge is to identify which patients with a single risk factors are at greatest risk for SCD. One approach to reducing this uncertainty is to seek additional risk markers. Programmed electrical stimulation, fractionation of paced right ventricular elecrtograms, late gadolinium enhancement on CMR, and genotyping have all been suggested but the results are contradictory or preliminary. 4 The approach by Haghjoo et al.1 is attractive because of the simplicity and widespread availability of electrocardiography. A number of previous studies have examined the relationship between ECG patterns and prognosis in HCM. In a retrospective analysis of the clinical course of 254 patients with HCM followed over a mean period of 6 years, McKenna et al.5 found that incomplete bundle branch block at the time of diagnosis was significantly commoner in patients that died, but was not predictive of SCD. Many other ECG variables including repolarization abnormalities, Q-waves, atrioventricular block, and intraventricular block have been examined but none is consistently associated with SCD. 6 , 7 QT dispersion (the interlead difference of the maximum and minimum QTc in a 12-lead ECG) caused by regional myocardial repolarization abnormalities is increased in HCM patients. Despite being a marker of electrical instability, QT dispersion has been inconsistently associated with SCD. 8 Several mechanisms cause ST-segment depression. Myocardial ischaemia and microvascular dysfunction are the common features of HCM and have been associated with adverse clinical outcomes. 9 However, ST-segment changes on exercise testing are unreliable markers of myocardial ischaemia as detected by myocardial perfusion imaging and studies of myocardial lactate metabolism. 10 ST-segment depression is also a feature of LV hypertrophy. As LV hypertrophy is also predictive of SCD risk, this could explain the association of the high lateral ST-segment depression and outcomes. Against this is the fact that in the multivariate analysis, only ST-segment depression in the high lateral leads was significant in its association with SCD. However, Haghjoo et al.1 did not examine the extent and distribution of the LV hypertrophy which is likely to explain the association of SCD with ST-segment depression in the high lateral leads. 11 The authors of this study 1 describe a novel predictor of SCD in HCM. However, the small size of the study population, its retrospective nature, and the low positive predictive value (21%) of ST-segment depression in the high lateral leads for SCD are major limitations. A larger prospective study will be required to confirm the prognostic significance of ST-segment depression in the high lateral leads and elucidate the underlying pathophysiology. In the meantime, the current risk algorithm should be used as the basis for an individualized assessment in every patient. 3 For example, the prognostic significance of all major risk factors is much greater in younger individuals. Similarly, the presence of other clinical features such as LV outflow tract obstruction, myocardial ischaemia, and extensive myocardial scar detected by cardiovascular magnetic resonance imaging can tip the balance in favour of aggressive management in a borderline case. Most importantly, the life saving potential of ICDs should always be balanced against the risks and potential complications of device therapy that accrue over the life-time of an individual patient. Conflict of interest: none declared.