Predictive value of EGSYS score in the differential diagnosis of cardiac syncope and neurally mediated syncope in children

BACKGROUND AND OBJECTIVE: Syncope is a common emergency with diverse etiologies in children. Among these, cardiac syncope (CS) is associated with high mortality and is usually difficult to diagnose. However, there is still no validated clinical prediction model to distinguish CS from other forms of pediatric syncope. The Evaluation of Guidelines in Syncope Study (EGSYS) score was designed to identify CS in adults and has been validated in several studies. In this study, we aimed to assess the ability of the EGSYS score in predicting CS in children. METHODS: In this retrospective study, we calculated and analyzed the EGSYS scores of 332 children hospitalized for syncope between January 2009 and December 2021. Among them, 281 were diagnosed with neurally mediated syncope (NMS) through the head-up tilt test, and 51 were diagnosed with CS using electrocardiography (ECG), echocardiography (ECHO), coronary computed tomography angiography (CTA), myocardial enzymes and genetic screening. The receiver operating characteristic (ROC) curve and Hosmer-Lemeshow test were used to evaluate the predictive value of the EGSYS score system. RESULTS: The median scores of 51 children with CS and 281 children with NMS were 4 [interquartile range (IQR): 3-5] and −1 (IQR: -2-1), respectively. The area under the ROC curve (AUC) was 0.922 [95% confidence interval (CI): 0.892-0.952; P < 0.001], indicating that the EGSYS score system has good discrimination. The best cutoff point was ≥3, with a sensitivity and specificity of 84.3% and 87.9%, respectively. The Hosmer-Lemeshow test demonstrated satisfactory calibration (χ²=1.468, P > 0.05) of the score, indicating a good fit of the model. CONCLUSION: The EGSYS score appeared to be sensitive for differentiating CS from NMS in children. It might be used as an additional diagnostic tool to aid pediatricians in accurately identifying children with CS in the clinical practice.