Alcoholic epilepsy: a unified and dynamic classification.

The interactions between alcohol and the CNS are complex and there are experimental data suggesting that chronic and acute effects are different and often opposite. An intriguing hypothesis is that repeated alcohol withdrawal seizures (AWS) may render the brain more excitable, leading to an epileptogenic state reminiscent of the 'kindling' model. To gain insight into this question we compared alcoholic patients with seizures related to episodes of AW (AWS) and patients with seizures unrelated to episodes of AW (UAWS). There were several significant differences between the AWS and UAWS groups. Age at admission for seizure was younger in the AWS groups (p < 0.005), seizure number was higher among patients with a history of seizures before admission in the UAWS group (p < 0.05). Neurologic signs were more frequent in the UAWS group (p < 0.05). Duration of intoxication was longer in the UAWS group and brain atrophy demonstrated by CT scan was more common in the UAWS group than the AWS group. Based on these findings, we propose a dynamic classification in which patients presenting seizures unrelated to any cause other than alcohol are classified in several successive stages of 'alcoholic epilepsy', the first being characterized by AWS and the last by persistent chronic seizures.