Abstracts

Rationale: Well adults presenting to the emergency room (ER) after isolated new-onset seizures are often discharged to the neurology outpatient service for further evaluation. A long waiting time may delay this assessment, possibly resulting in non-diagnostic EEG. The yield of EEG might be improved when performed soon after a seizure. We sought to assess the utility of EEG in these patients within an ER setting, correlating the findings with neuroimaging data and the incidence of recurrent seizures. Methods: Patients attending the ER from January 2008 to December 2009 with uncomplicated first episodes of unprovoked convulsive seizures were included. All made a complete neurological recovery and underwent 30-minute EEG using a 10-20 international montage prior to being discharged. ER physicians were blinded to the EEG findings. Correlation was made with neuroimaging (either CT or MRI) that was performed based on standard clinical ER protocols and independent of EEG results. Seizure recurrence was assessed during a 12- month period from the time of initial presentation. Results: 84 patients were included in the study (55 males, 29 females). The mean age was 31.4 years (range 16-61). 30 (35.7 %) had abnormal EEG: 13 patients with lateralized epileptiform discharges, 8 isolated focal slowing, 7 generalized spike-wave discharges and 2 diffuse slowing. The sensitivity, specificity, positive predictive value and negative predictive value of EEG in predicting abnormal neuroimaging were 67%, 60%, 35%, and 85% respectively [OR 4.8 (95% CI 1.2-19.7)] after adjusting for age and sex. In those with lateralized EEG abnormalities and positive neuroimaging, regional concordance between the two was 71.4%. The sensitivity, specificity, positive predictive value and negative predictive value of EEG in predicting 12 month seizure recurrence were 62%, 70%, 27%, 91% respectively [OR 3.6 (95% CI 1.09-11.6)] . Of 66 patients who were not commenced on anticonvulsant medication during the follow-up period, 8 had recurrent seizures, 5 of whom had abnormal initial EEGs at the time of presentation. Conclusions: The diagnostic yield of EEG in uncomplicated new onset seizures may be improved if performed acutely. It can serve as an additional tool to aid the selection of patients who will benefit from early neuroimaging beyond standard ER protocols. It can be used to stratify patients who are unlikely to have recurrent seizures. Patients with abnormal EEG who are should be closely monitored in the epilepsy clinic.