Semiological features of epilepsy seizures and paroxysmal non‐epileptic events in children: A systematic review and meta‐analysis

AIM: To systematically review evidence on age-specific semiological features of paediatric paroxysmal events assessed by video-electroencephalography (EEG) monitoring, and to identify key features that might help differentiate epileptic from non-epileptic events during clinical assessment of paediatric patients. METHOD: We conducted a PROSPERO-registered systematic review and meta-analysis of literature in MEDLINE, Embase, and CENTRAL up to November 2023, encompassing studies of paediatric patients aged up to 21 years with paroxysmal events recorded on video-EEG. Risk of bias was assessed using the JBI critical appraisal tools. Primary outcomes were pooled proportions and subgroup differences for 28 semiology types (95% confidence intervals), stratified by age groups (1 month-3 years, 3-12 years, 1 month-21 years). RESULTS: Forty studies involving 3123 patients were analysed. In epilepsy, arrest (p = 0.029, 1 month-3 years), hypermotor activity (p = 0.015, 3-12 years), tonic semiology (p < 0.001, 3-12 years and 1 month-21 years), and spasm (p = 0.028, 1 month-21 years) were significantly more frequent. Movements involving agonist-antagonist muscles (0 out of 271 children with epilepsy vs 45 out of 454 children without epilepsy), stereotype semiology (0 out of 249 children with epilepsy vs 89 out of 879 children without epilepsy), asynchronous motor events (0 out of 249 children with epilepsy vs 9 out of 167 children without epilepsy), tremor (1 out of 271 children with epilepsy vs 136 out of 1689 children without epilepsy), and akinetic semiology (0 out of 249 children with epilepsy vs 38 out of 688 children without epilepsy) were mostly of non-epileptic origin. INTERPRETATION: Video-EEG remains the criterion standard for definitive diagnosis. Nevertheless, certain age-specific semiological features may support differentiation and risk stratification.