External validation of a contemporary risk stratification model in acute myocarditis: prognostic implications from the PRE-MYO cohort

Background Risk stratification in acute myocarditis is critical to guide monitoring intensity and therapeutic decision-making1. The Spanish Society of Cardiology Working Group on Myocarditis has proposed a classification which integrates clinical, functional and imaging findings to identify patients at higher risk2. However, external validation has yet to be evaluated. Purpose To externally validate the discriminative performance of the model proposed by the Spanish group in an independent contemporary myocarditis cohort (PRE-MYO). Methods Four hundred and forty-six patients with acute myocarditis from the PRE-MYO cohort were evaluated. Patients were stratified as low (n=25), intermediate (n=334) or high-risk (n=87) according to predefined criteria. Baseline characteristics were compared across groups. Kaplan-Meier analysis evaluated major adverse cardiovascular events (MACE) free survival (all-cause death, heart transplant, LV assist device implant, heart failure hospitalization, cardiogenic shock or sustained ventricular arrhythmias). Cox regression assessed association between risk stratification and MACE. Model discrimination was evaluated using Harrell’s C-statistic. Results Risk categories identified clinically distinct phenotypes. High-risk patients were significantly older (44.9±20.5, p<0.001) and exhibited a higher burden of adverse events, with significantly higher mortality (9.5%, p=0.001) compared to low and intermediate-risk patients. MACE-free survival significantly differed between groups (log-rank p<0.001), with an increasing gradient from low to high-risk. No association was observed between intermediate-risk group and MACE (HR 1.05, 95% CI 0.14–8.01, p=0.964). However, high-risk group exhibited a fourfold increased hazard (HR 4.24, 95% CI 0.55–32.39, p=0.164), although statistical significance was not reached. The model showed moderate discriminative performance (Harrell’s C statistic 0.652, 95% CI 0.55–0.755). Conclusions In the PRE-MYO cohort, the proposed risk model identified clinically distinct patient profiles and demonstrated a graded separation in event-free survival. Although independent prognostic significance was not confirmed in multivariable Cox analysis, likely due to limited event numbers and group imbalance, the consistent gradient in mortality and MACE across categories suggests that the model captures clinically meaningful risk stratification.