Implications of the 2024 American Heart Association Classification of Cardiac Sarcoidosis: Characteristics and Outcome of Patients Across the Continuum of Diagnostic Likelihood

Background: There is no universal agreement on the diagnostic criteria for cardiac sarcoidosis (CS). In 2024, the American Heart Association (AHA) introduced a new diagnostic scheme featuring a quaternary likelihood stratification for CS instead of dichotomy. Its clinical implications are unknown. Methods: We analyzed 252 patients with CS and proven histology (mean age, 49 years; 61% women). Following the AHA diagnostic stratification, CS was retrospectively considered definite given sarcoid granulomas on myocardial biopsy (n=109), highly probable with both clinical and cardiac imaging findings being CS consistent (n=93), and probable with either clinical or imaging findings alone being CS consistent (n=50). The groups were compared for clinical characteristics and the composite incidence of death, transplantation, implantation of a left ventricular assist device, and ventricular tachyarrhythmia on follow-up (median, 5.0 years). Results: In AHA definite, highly probable, and probable CS: 94%, 99%, and 22% of patients (P<0.001) had high-grade atrioventricular block, ventricular tachyarrhythmia, or heart failure at presentation; median left ventricular ejection fraction was 42% (interquartile range, 30%-58%), 56% (interquartile range, 47%-62%), and 55% (interquartile range, 47%-65%) (P<0.001); cardiac troponins were elevated in 61%, 33%, and 20% (P<0.001); and the 5-year event rate (95% CI) was 44% (34%-54%), 20% (11%-29%), and 10% (2%-19%) (P<0.001), respectively. In multivariable analysis, left ventricular ejection fraction (P<0.001) and presentation with ventricular tachyarrhythmia (P<0.001) were the key prognostic factors, with the AHA classification failing to have independent predictive value (P=0.201). Conclusions: The 2024 AHA stratification of CS likelihood distinguishes groups with differences in clinical characteristics but did not have independent prognostic value in our statistical analysis.