Refining arrhythmic risk stratification in hypertrophic cardiomyopathy through left atrial volume and strain
L Dinc Asarcikli, D Laenens, M P Lopez Santi, N Elmasry, L Caunite, T Zegkos, V Kamperidis, D Parcharidou, A Ziakas, T Li, W Kong, C H Sia, J J Bax, N Ajmone MarsanAbstract
Background
Risk stratification for sudden cardiac death (SCD) in hypertrophic cardiomyopathy (HCM) remains challenging. The ESC SCD risk score includes left atrial (LA) diameter but does not capture volumetric LA remodeling or LA functional impairment. Indexed LA volume (LAVI) and left atrial reservoir strain (LARS) may better reflect LA remodeling and dysfunction, which may precede LA dilatation. Whether adding these parameters improves arrhythmic risk prediction beyond the ESC SCD model remains unknown.
Purpose
To evaluate the incremental prognostic value of LAVI and LARS when sequentially incorporated into the 2014 ESC SCD risk model for prediction of arrhythmic events in HCM.
Methods
We included 1,122 consecutive patients with HCM from a multicentre cohort after excluding patients without follow-up or with missing data. Patients were followed for a composite arrhythmic endpoint including SCD, aborted SCD, and appropriate ICD therapy. Three Cox models were constructed. Model 1 included existing ESC SCD risk variables (Reference Model). In Model 2, LA diameter was replaced with LAVI ≥ 34 mL/m². Model 3 further incorporated LARS ≤ 23.3%. Model performance was compared using model fit, discrimination, calibration, and category-free net reclassification improvement (NRI).
Results
The mean age was 54 ± 15 years, and 55% were male. The median SCD risk score was 2.45 (IQR 1.61–3.80). During a median follow-up of 62 months (IQR 35–117), 122 patients (10.9%) experienced the composite arrhythmic endpoint, including 18 (1.6%) cases of SCD, 13 (1.2%) cases of aborted SCD, and 91 (8.1%) cases of appropriate ICD therapies. Optimal cut-off values for the composite arrhythmic endpoint were defined as LAVI ≥ 34 mL/m² according to guideline recommendations and LARS ≤ 23.3%, based on ROC analysis. In univariable analyses, the ESC SCD risk score, LVEF < 50%, LAVI ≥ 34 mL/m², LARS ≤ 23.3%, and the presence of an apical aneurysm were each significantly associated with the outcome. After multivariable adjustment, both LAVI ≥ 34 mL/m² (HR 2.98, 95% CI 1.79–4.94, p < 0.001) and LARS ≤ 23.3% (HR 1.86, 95% CI 1.11–2.63, p = 0.016) remained independently associated with arrhythmic events, together with the apical aneurysm (Table). Kaplan–Meier analysis demonstrated significantly lower event-free survival in patients with LAVI ≥ 34 mL/m² and LARS ≤ 23.3% (log-rank p < 0.001). The enhanced model showed stepwise improvement in overall fit, with chi-square increasing from 106 to 113 and 121 (Figure), along with improved discrimination, as Uno’s C-index rose from 0.816 to 0.869 and 0.874. Compared with Model 1, Model 2 improved overall NRI (0.431, p<0.001), while Model 3 provided further reclassification improvement (NRI 0.604, p<0.001) while preserving calibration.
Conclusion
Incorporation of LAVI and LARS into the ESC SCD risk model provides significant incremental improvement in arrhythmic risk stratification in patients with HCM.Figure