A Mortality Risk Prediction Score for Dilated Cardiomyopathy

Dilated cardiomyopathy (DCM) is a leading cause of pediatric heart failure and carries a substantial risk of death or heart transplantation. In low- and middle-income countries, limited access to advanced heart failure therapies and transplantation highlights the need for simple, reliable tools to identify children at highest risk. Therefore, this study aimed to identify independent predictors of mortality and develop a practical bedside risk score for mortality prediction in children with DCM. We conducted a retrospective cohort study of children with DCM treated at Dr. Sardjito Hospital, Indonesia, between 2016 and 2025. Clinical, electrocardiographic, and echocardiographic variables were evaluated for association with mortality. Independent predictors were incorporated into a point-based risk score using the Spiegelhalter-Knill-Jones method. Discriminative performance was assessed by receiver operating characteristic analysis and internal validation. Among 109 children with DCM, mortality was 33.9%. Seven independent predictors of mortality were identified: age ≥ 2 years at diagnosis, New York Heart Association/Ross functional class IV heart failure, malnutrition, requirement for vasoactive support, low-voltage electrocardiography, left ventricular end-diastolic diameter z-score > + 4, and left ventricular ejection fraction ≤ 30%. The derived score ranged from - 25 to + 28 points. A threshold score ≥ 11 accurately identified high-risk patients, achieving an area under the curve (AUC) of 0.932 (95% CI 0.886-0.977), with 73.0% sensitivity and 95.8% specificity. Internal validation demonstrated sustained performance (AUC 0.921). We developed and validated a simple, bedside mortality risk score with excellent discriminative ability for children with DCM. By relying on routinely available clinical and echocardiographic parameters, this tool enables early identification of high-risk patients and may support timely clinical decision-making and resource allocation, particularly in resource-constrained settings.