A Prediction Rule to Identify Children and Young Adults at Low Risk for Myocarditis

(1) To derive a prediction rule for pediatric myocarditis that distinguishes low-risk patients for whom additional work-up, including venipuncture or cardiac imaging, may be avoided, (2) to assess the test characteristics of troponin in our study population.

This retrospective case-control study included all patients who presented to a pediatric emergency department between 2010 and 2021 and underwent troponin testing for suspected myocarditis. Myocarditis cases (identified using American Heart Association criteria) and controls were to approximate a 1:2 ratio. Logistic regression with forward selection was used to derive a prediction rule for myocarditis. As the goal was to derive a rule for low-risk children, in whom venipuncture would be unnecessary, laboratory results were analyzed separately.

We identified 93 case patients and 202 control patients. The final prediction rule included chest pain [adjusted odds ratio (aOR): 3.5, 95% CI: 1.8 to 7.0], reported or measured fever (aOR: 1.7, 95% CI: 1.0 to 3.1,) and atrioventricular conduction delays or ST segment changes (aOR: 2.6, 95% CI: 1.4 to 4.7). Sensitivity, calculated as the proportion of cases with at least one of the 3 predictors was 99% (95% CI: 0.94 to 0.99), and specificity was 14% (95% CI: 0.09 to 0.20). With at least 2 predictors, sensitivity was 60% (95% CI: 0.50 to 0.71) and specificity was 72% (95% CI: 0.65 to 0.78).

The prediction rule developed can help identify children at low risk for myocarditis and, therefore, avoid troponin testing and/or further evaluation including cardiology consult or cardiac imaging. Specificity was insufficient to rule in myocarditis without additional investigation.