Predicting Seizure Risk from Routine Electroencephalographs in Medical Intensive Care Units Using the 2HELPS2B Score

This study evaluated the utility of the 2HELPS2B score in predicting seizures from routine electroencephalographs (rEEGs). In total, 670 rEEGs obtained in a medical intensive care unit (MICU) between October 2018 and March 2023 were analyzed. More than 75% of these rEEGs were requested due to seizures and unexplained altered consciousness. Seizures occurred most frequently in patients with rEEGs characterized by brief, potentially ictal rhythmic discharges and electrographic seizures. A history of seizures was the most prevalent risk factor identified by the 2HELPS2B score. Seizures occurred in 28% of the cohort who experienced a seizure within 24 h of the rEEG and in 38% of the cohort who experienced a seizure before MICU discharge. Among the patients with suspected altered consciousness, the seizure incidence before MICU discharge (9.2%) was twice that within 24 h of the initial rEEG (4.7%). The seizure rate also increased from 12% for a 2HELPS2B score of 1 to 100% for scores ≥ 4. A score ≥ 2 was the optimal cutoff for predicting post-rEEG seizures and guiding antiseizure medication (ASM) treatment. Seizures occurred most frequently in patients whose ASMs were supplemented with new medications, and most new prescriptions for antiseizure medication were issued to patients with altered consciousness. These results demonstrate that the 2HELPS2B score can effectively predict seizures on the basis of rEEG results.