Article Text

Download PDFPDF

Increasing evidence that risk scores underperform in predicting sudden death in hypertrophic cardiomyopathy
  1. Martin Maron,
  2. Ethan Rowin,
  3. Barry J Maron
  1. Hypertrophic Cardiomyopathy Center Division of Cardiology, Tufts Medical Center, Boston, Massachusetts, USA
  1. Correspondence to Dr Martin Maron, Tufts Medical Center, Boston, MA 02116, USA; mmaron{at}

Statistics from

Hypertrophic cardiomyopathy (HCM) remains the most common cause of sudden cardiac death (SCD) in young people, underscoring the need for accurate selection of patients for primary prevention implantable cardioverter-defibrillator (ICD) therapy.1 Since 2014, the European Society of Cardiology (ESC) has promoted a quantitative risk score, composed of seven disease-related features that can be imputed on a smart phone (ESC Risk Score App) to predict sudden death events over 5 years.2 The ESC risk score ascribes management recommendations for primary prevention ICDs based on whether a patient falls into one of three risk categories: low (<4%, ICD not indicated), intermediate (4%–6%, ICD could be considered) or high (≥6%, ICD should be considered). Of note, the ESC risk score excludes children and adolescent patients with HCM, a subgroup traditionally considered at particularly high risk for SCD. Prior investigations of this risk score have included only a small number of Asian patients with HCM2 3 and therefore the primary aim of the present Heart study by Choi et al,4 was to determine the efficacy of the HCM Risk-SCD calculator in a cohort of Korean patients with HCM.

In this observational cohort study, 730 Korean patients with HCM were followed for 4288 person-years for the primary endpoint of SCD events. Over the follow-up, 11 patients died suddenly and six experienced appropriate ICD shocks for ventricular tachyarrythmias, including seven patients with low ESC risk score, three with intermediate scores and six with high-risk scores. Of particular importance was the observation that of the 11 patients who died suddenly due to HCM, the majority (n=7; 64%) had low ESC risk scores that clearly excluded an ICD recommendation.4 Choi et al 4 appropriately conclude that the ESC risk calculator underperforms in detecting Asian patients with HCM who are at increased risk for sudden death events and based on low risk scores would not be protected from sudden death by ICD therapy.

Therefore, these important data from Korea are in fact representative of the broad experience of the ESC risk score and not unique to this specific ethnic group. For example, in a recent large prospective single-centre experience from the USA (Tufts HCM Institute),5 aimed at assessing the most effect risk stratification strategy for patients with HCM, the ESC risk score was associated with low sensitivity (34%) for identifying high-risk patients with HCM, with low ESC risk scores in almost 50% of patients with HCM who experienced potentially life-saving appropriate ICD interventions terminating ventricular tachycardia (VT)/ventricular fibrillation (VF). Furthermore, in a recent meta-analysis of nine studies (not including the recent Tufts experience) comprising close to 10 000 patients with HCM from diverse regions and ethnicities internationally, the ESC risk score was associated with sensitivities as low as 41% for predicting individual patients with HCM who suffered a SCD event.6 Therefore, the totality of these broad and expansive experiences demonstrate that the ESC risk score does not reliably identify an important and substantial subset of at-risk patients with HCM who likely would be excluded from an ICD implant recommendation and remain at risk for potentially life-threatening SCD events.5–8

Therefore, a pertinent clinical question raised by these data is what risk stratification strategy is most reliable for identifying patients with HCM at greatest risk for sudden death? This is highly relevant given that there are now two markedly distinct strategies for risk stratifying patients with HCM for ICD therapy, the new ESC HCM Risk-SCD calculator and the well established but non-mathematical approach of the American College of Cardiology/American Heart Association (ACC/AHA) guidelines,1 5 which emploes the use of ≥1 conventional risk markers considered major within the clinical profile of the patient. Since no risk stratification strategy can provide absolute perfection in predicting future risk in such a heterogeneous disease as HCM, the answer to this critical question rests with which strategy protects more patients from sudden death at the cost of some degree of overtreatment with ICDs.

To this end, the recently published Tufts HCM Center experience provides important clarity into this issue, with a reported sensitivity for the ACC/AHA risk factor strategy that was very high (95% with intention to treat), resulting in identification of nearly all at-risk patients with HCM.5 Achieving such high sensitivity (compared with 34% for the ESC risk score) was possible in part due to the opportunity for incorporating novel high-risk cardiac magnetic resonance markers such as extensive late gadolinium enhancement, systolic dysfunction (ejection fraction <50%) and left ventricular apical aneurysms, which accounted for a significant proportion (~25%) of appropriate ICD therapies but are excluded from the ESC risk score algorithm.5 In addition, a recent investigation in close to 1500 patients with HCM provided additional support to extensive LGE as a powerful independent risk marker, strengthening SCD prediction when included in current risk model.9 These observations underscore an important distinguishing strength of the ACC/AHA strategy, its inherent flexibility to incorporate emerging and novel markers which enhance risk prediction, a limitation of the more rigid ESC risk score.

On the other hand, as well noted by Choi et al 5 as well as other investigators,8 9 the ESC risk score is associated with relatively high specificity, implying that by using this prediction strategy, the number of ICD implants could be reduced in low-risk patients and ICD overuse decreased. However, the number of ICDs needed to appropriately treat one patient for VT/VF is similar between the ESC and ACC/AHA strategies and does not different from randomised clinical trials of ICDs in non-HCM diseases.5 Therefore, the cost of preventing SCD events in patients with HCM may be only a modest number of ICDs placed in low-risk patients and not different from that of ischaemic heart disease.

Another underappreciated issue related to the two HCM risk stratification strategies involves the capability to integrate a narrative of shared decision-making, between cardiologist and patient, so important in a heterogeneous disease such as HCM. In this regard, quantitative mathematically based risk scores are inherently inflexible to incorporating physician judgement in making management decisions for individual patients with HCM, particularly since in the case of the ESC risk score concrete ICD management recommendations are assigned to risk estimates. Indeed, relying solely on statistical modelling to derive management recommendations can provide an illusion of certainty and a false oversimplification of accuracy and precision. On the other hand, the opportunity to apply physician judgement and medical reasoning to complex medical decision-making is a particular strength associated with the more accommodating ACC/AHA risk marker strategy.

In conclusion, identifying patients at the highest risk for sudden death deserving lifesaving ICD therapy remains a critical management priority in this disease. In this regard, these data from Choi et al,4 extend our experience with risk stratification by underscoring the principle that the new ESC risk score underperforms in identifying at-risk patients with HCM, throughout the broad disease spectrum. It is our view, now substantiated by considerable data, that the highly sensitive ACC/AHA major risk factor method of using ≥1 of the established individual risk factors considered major within the clinical profile of the patient with HCM, and associated with a greater measure of shared decision-making and opportunity for physician judgement, remains the preferred strategy for risk stratifying individual patients with HCM for primary prevention device therapy.



  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles