Real-world performance of current sudden cardiac death risk stratification tools in hypertrophic cardiomyopathy: a single-centre study

Background

Sudden cardiac death (SCD) risk stratification in hypertrophic cardiomyopathy (HCM) remains a cornerstone of disease management. However, currently available risk-prediction tools require refinement to balance the risks and benefits of implantable cardioverter-defibrillator (ICD) therapy.

Purpose

To assess SCD risk and compare the performance of contemporary stratification tools - 2023 European Society of Cardiology (ESC) and 2024 American College of Cardiology/American Heart Association (ACC/AHA) flow-charts - in a real-world HCM population.

Methods

Single-centre retrospective study of patients (pts) followed in a dedicated cardiomyopathy clinic with confirmed HCM. HCM phenocopies were excluded. Primary composite endpoint consisted of SCD, sustained ventricular arrhythmias and/or appropriate ICD therapies. Data was extracted from medical records.

Results

A total of 130 HCM pts were included; 53.1% (n=69) male, mean age 62 ± 17 years. Obstructive phenotype was present in 22.3% (n=29) and pathogenic/likely pathogenic variants were identified in 33.8% (n=44).

Based on clinician’s judgement and guideline-based risk stratification, 17.7% (n=23) received an ICD, predominantly for primary prevention of SCD (n = 19, 82.6%).

At a median follow-up of 6 (IQR 8) years, 6 pts (4.6%) reached the primary endpoint – 5 pts due to appropriate ICD shocks and 1 pt due to SCD. No inappropriate shocks in ICD-carriers were reported.

When applied at baseline data, the ESC 5-year HCM-SCD risk model was a significant predictor of events in Cox regression (HR 1.72; 95% CI 1.15–2.57; p = 0.009) and receiver operating characteristic (ROC) curve analysis yielded modest discrimination (AUC 0.623; 95% CI 0.309–0.937) – figure 1.

In contrast, the ACC/AHA risk-marker–based algorithm did not differentiate event-free survival (log-rank p = 0.453), demonstrating limited discriminatory ability in our low-event-rate cohort.

Conclusions

In our real-world HCM population, the ESC model demonstrated superior predictive performance and clinical utility compared with the ACC/AHA risk-marker approach, which showed insufficiency to detect differences in a population with predominantly low event rates. These findings underscore the need for improved risk-prediction strategies in HCM.For image description, please refer to the figure legend and surrounding text.