Clinical utility of EEG in alcohol-related seizures.

OBJECTIVE: To study if electroencephalogram (EEG) can discriminate between alcohol-related seizures (ARS) and seizures unrelated to alcohol use.
MATERIAL AND METHODS: Alcohol-related seizures was defined as a seizure in a patient with score > or = 8 in the Alcohol Use Disorders Identification Test (AUDIT). Twenty-seven patients with ARS (22 without epilepsy: ARSwE), 21 AUDIT-negative epileptic patients with seizures (ES), and 30 other AUDIT negative patients with seizures (OS) were studied. Thirty-seven epilepsy outpatients and 79 sciatica inpatients were controls.
RESULTS: Epileptiform and slow activity were less frequent in the ARSwE than in the ES group. Alpha amplitude was lower in the ARSwE than the other groups. Photoparoxysmal activity was not observed. EEG was associated with a larger negative predictive value (78% probability of non-ARS if EEG was abnormal) than a positive predictive value (55% probability of ARS if EEG was normal).
CONCLUSION: A definitely abnormal EEG suggests epilepsy or symptomatic seizures unrelated to alcohol. The predictive value of a normal EEG is limited, but the typical post-ictal finding in ARS is nevertheless a normal low-amplitude EEG record.