Clinical spectrum and risk factors for severe arrhythmias in non-surgical pediatric patients: a single-center retrospective study

To analyze the clinical spectrum and identify risk factors for severe arrhythmias in children. A single-center retrospective study was conducted on 428 consecutive children (0–18 years) diagnosed with arrhythmia between January 2021 and October 2025. Patients were classified as having severe arrhythmia-related clinical presentation (severe group, n = 107) or non-severe presentation (non-severe group, n = 321) according to the presence of hemodynamic instability, syncope, heart failure, or malignant arrhythmia features. Clinical spectrum characteristics were collected. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors for severe arrhythmias. Electrocardiographic (ECG), echocardiographic, and serological parameters were compared. The severe group had higher proportions of infants/toddlers (0–3 years), underlying cardiac disease, syncope history, and class III-IV heart function ( P < 0.05). Multivariate analysis identified these as independent risk factors ( P < 0.05): infant/toddler age [2.30 (1.08–4.89)], underlying cardiac disease [12.66 (6.01–26.68)], syncope history [45.84 (12.74-164.91)], and class III-IV heart function [32.97 (13.39–81.17)]. The prediction model showed good fit (Hosmer-Lemeshow P = 0.171) and discrimination (AUC = 0.925, 95% CI : 0.893–0.958). The severe group also exhibited longer QTc, wider QRS, more third-degree atrioventricular block, higher ventricular premature burden, larger cardiac dimensions (LVEDD/LAD Z-score), higher pulmonary artery systolic pressure and myocardial injury/heart failure markers (CK-MB, cTnI, BNP, NT-proBNP), but lower heart rate variability (SDNN), left ventricular ejection fraction, and serum potassium ( P < 0.05). Infant/toddler age, underlying cardiac disease, syncope history, and advanced heart failure are independent clinical risk factors for severe pediatric arrhythmias. A model incorporating these factors demonstrates strong predictive ability for early high-risk identification.