Sudden cardiac death in childhood hypertrophic cardiomyopathy is best predicted by a combination of electrocardiogram risk‐score and HCMRisk‐Kids score
Ingegerd Östman‐Smith, Gunnar Sjöberg, Jenny Alenius Dahlqvist, Per Larsson, Eva Fernlund
Abstract
AIM: To compare risk algorithms (HCMRisk-Kids, ECG Risk-score) in hypertrophic cardiomyopathy (HCM) without syndrome association (ns-HCM) and with Noonan-like syndromes (RAS-HCM). METHODS: A national paediatric HCM cohort (n = 151), presenting <19 years of age, mean follow-up 13.3 years, from all Swedish centres of Paediatric Cardiology (presenting 1972-2015), with 41 RAS-HCM patients (61% males), and 110 ns-HCM patients (68% familial; 65% males). The end-point was a composite of sudden cardiac death and resuscitated cardiac arrest (SCD/CA). Risk-factors were studied with Cox-hazard regression, and receiver operating characteristic curve analysis (C-statistic). RESULTS: There were 33 SCD/CA, 27/110 in ns-HCM and 6/41 in RAS-HCM (p = 0.27). In ns-HCM HCMRisk-Kids ≥6% at diagnosis had C-statistic of 0.69 for predicting SCD/CA during first 5 years of follow-up and positive predictive value (PPV) of 22%. After 7 years of age (HCMRisk-Kids7plus), C-statistic was 0.76. ECG Risk-score ≥6 at diagnosis had C-statistic 0.87 and PPV of 31%. Independent risk factors for SCD/CA were HCMRisk-Kids7plus score (p = 0.005) and ECG risk-score (p < 0.001), whereas early beta-blocker dose (p = 0.001) and myectomy (p = 0.004) reduced risk. The sum of HCMRisk-Kids7yplus and ECG Risk-score7yplus ≥14 best predicted SCD/CA within 5 years in ns-HCM with C-statistic of 0.90 [0.83-0.96], sensitivity 100% and PPV 38%. CONCLUSION: Combining the ECG Risk-score with HCMRisk-Kids improves risk stratification in ns-HCM and shows promise in RAS-HCM.