Monitoring Antiseizure Medication Change Using Ultra Long‐Term Electroencephalogram: A Multicenter Study

OBJECTIVE: Medication dose adjustments are common in treatment-resistant epilepsy, but lack robust evidence, and patient-reported seizure diaries have known limitations. We assessed whether ultra long-term subcutaneous electroencephalogram (sqEEG) improves detection of seizure frequency changes after antiseizure medication adjustments. METHODS: A retrospective, multicenter cohort study was carried out across 9 centers in adults with treatment-resistant epilepsy. SqEEG seizure frequency was compared with seizure diaries before and after medication dose increases and reductions, over 7-, 14-, and 30-day windows, using categorical analyses, mixed-effects models and survival analyses. RESULTS: A total of 65 participants (mean age 40.6 years; 59% female) contributed >280,000 hours of sqEEG, 9,190 electrographic seizures and 4,850 diary events, including across 160 medication changes. Categorical disagreement between modalities ranged from 36% to 50%, with significant differences at 7 and 14 days for dose reductions, and at 14 days for dose increases. sqEEG recorded 2.2- to 3.4-fold more seizures than diary across all seizure count models (all p < 0.01). Seizure count models showed no statistically significant difference between modalities in detecting treatment-related changes (interaction estimates: 5-38% larger increases with sqEEG for dose reductions, and 13-38% larger reductions with sqEEG for dose increases). SqEEG identified return to baseline seizure counts significantly earlier than diary after dose reductions at 7- and 14-day windows (median 7 vs 13 days and 10 vs 17 days, respectively; adjusted p = 0.046 and 0.042). INTERPRETATION: SqEEG detects more seizures than patient-reported diaries and identifies treatment-related changes earlier, even after modest medication adjustments, particularly after dose reductions, highlighting the potential of objective ultra long-term EEG monitoring in treatment-resistant epilepsy. ANN NEUROL 2026;100:600-612.