Objective: This study was undertaken to characterize somatic symptoms and related disorders (SSD) in epilepsy.
Methods: Adults with epilepsy under active follow-up at a tertiary epilepsy center were consecutively enrolled. The diagnosis of SSD was performed by an experienced psychologist based on the structured clinical interview for Statistical Manual of Mental Disorders, 5th edition. Detailed social/demographic data, epilepsy features, psychiatric features, life quality, disability, and economic burden were collected and compared between people with SSD and those without. Bodily distress syndrome checklist, Somatic Symptom Disorder-B Criteria Scale, Patient Health Questionnaire-9, and Generalized Anxiety Disorder seven-item scale (GAD-7) were used to evaluate SSD individuals' somatic symptoms, symptomrelated psychological distress, and depressive and anxious symptoms. Quality of life and disability were assessed by Quality of Life in Epilepsy Inventory 31 (QOLIE-31) and World Health Organization Disability Assessment Schedule V.2.0 (WHO DAS 2.0). A risk prediction nomogram was generated using least absolute shrinkage and selection operator (LASSO) analysis and validated.Results: One hundred fifty of 631 participants (24%) were diagnosed with SSD.In people with SSD, the top three most common somatic symptoms were memory impairment, headache, and dizziness (85%, 80%, and 78%, respectively), and multiple systems were involved in most (82%) people with SSD. Compared with people without SSD, those with SSD had lower QOLIE-31 total scores, and higher WHO DAS 2.0 scores and disease economic burdens. LASSO analysis suggested that a history of severe traumatic brain injury, hippocampal sclerosis, low seizure worry and medication effects scores on QOLIE-31, multiple systems affected by somatic symptoms, and a high GAD-7 score were risk factors of SSD. The nomogram was validated for good accuracy in the training and testing cohorts.Significance: SSD are likely to be a common comorbidity in epilepsy and harm epilepsy prognosis. Our risk prediction nomogram was successfully developed but needs further validation in larger cohorts.