Résumé
Background and Aims Immune checkpoint inhibitors (ICI) are associated with life-threatening myocarditis but milder presentations are increasingly recognized. The same autoimmune process that causes ICI myocarditis can manifest concurrent generalized myositis, myasthenia-like syndrome, and respiratory muscle failure. Prognostic factors for this 'cardiomyotoxicity' are lacking. The main aim of this study was to determine predictors and construct a risk score associated with negative outcomes in patients admitted for ICI myocarditis. Methods A multicentre registry collected data retrospectively from 17 countries between 2014 and 2023. A multivariable Cox regression model was used to determine risk factors for the primary composite outcome: time to severe arrhythmia, heart failure, respiratory muscle failure, and/or cardiomyotoxicity-related death. Covariates included demographics, comorbidities, cardiomuscular symptoms, diagnostics, and treatments. Time-dependent covariates were used, and missing data were imputed. A point-based prognostic risk score was derived and externally validated. Results In 748 patients (67% male, age 23-94 years), 30-day incidence of the primary composite outcome, cardiomyotoxic death, and overall death were 33%, 13%, and 17%, respectively. By multivariable analysis, the primary composite outcome was associated with active thymoma (hazard ratio [HR] 3.6, 95% confidence interval [CI] 1.7-7.7), presence of cardiomuscular symptoms (HR 2.6 [1.5-4.2]), low QRS voltage on presenting electrocardiogram (HR for <= 0.5 mV vs >1 mV 1.9 [1.1-3.1]), left ventricular ejection fraction (LVEF) < 50% (HR 1.7 [1.1-2.6]), and incremental troponin elevation (HR 1.8 [1.4-2.4], 2.9 [1.8-4.7], and 4.6 [2.3-9.3], for 20, 200, and 2000-fold above upper reference limit, respectively). A prognostic risk score developed using these parameters showed good performance; 30-day primary outcome incidence increased gradually from 4% (risk score = 0) to 81% (risk score >= 4). This risk score was externally validated in two independent French and US cohorts. This risk score was used prospectively in the external French cohort to identify low-risk patients who were managed with no immunosuppression resulting in no cardiomyotoxic events. Conclusions ICI-associated myocarditis can manifest with high morbidity and mortality. Myocarditis severity is associated with magnitude of troponin, thymoma, low QRS voltage, depressed LVEF, and cardiomuscular symptoms. A risk score incorporating these features performed well.