Development and Validation of a Model to Predict Secondary Arrhythmia in Patients With Epilepsy

Compared with healthy individuals, epilepsy patients are more prone to arrhythmias, which may contribute to poor prognosis. To enable early identification of this risk, we developed a clinical prognostic prediction model to assess the risk of arrhythmia comorbidity in epilepsy patients, thereby facilitating timely clinical intervention to improve patient outcomes.

We retrospectively collected clinical data from epilepsy patients treated at the Affiliated Hospital of Qingdao University between January 2022 and February 2025, including gender, age, medical history, antiseizure medications, electrocardiograms and electroencephalograms. A total of 495 eligible patients were enrolled and randomly divided into development and validation datasets at a 7:3 ratio. Variable selection was performed using LASSO regression with a penalty term, and the selected variables were incorporated into the construction of a logistic regression model. The area under the receiver operating characteristic curve (AUC) and its 95% confidence interval were used to preliminarily evaluate the model's discriminative ability, while cross-validation and bootstrapping were employed to assess its generalizability. Calibration curves and the Brier score were utilized to evaluate the model's calibration, and decision curve analysis was plotted to analyze the net clinical benefit.

The C-indices for the development and validation datasets were 0.737 (95% CI 0.675-0.799) and 0.790 (95% CI: 0.707-0.884), respectively, with an overall C-index of 0.752 (95% CI: 0.701-0.804). The corresponding sensitivity and specificity were 74.6% and 68.1%, respectively. Finally, a nomogram was constructed for the visual presentation of the predictive model.

Our predictive model can accurately assess the risk of arrhythmia comorbidity in epilepsy patients, assisting clinicians in early intervention to improve prognosis.