Abstracts

RATIONALE: People with seizures often complain of memory problems. We performed a pilot study of donepezil (Aricept?) in treatment of memory impairment associated with epilepsy. METHODS: The study was open-label, with a 3 month baseline, 1 month of donepezil 5 mg, then 2 months of donepezil 10 mg daily. 18 patients completed clinically, and 12 completed neuropsychological testing. All patients had partial seizures with or without secondary generalization, and most were on 1-2 AEDs. Patients took tests assessing verbal memory and attention. The Buschke Selective Reminding Test (STR) assesses rote verbal learning and memory, by sequential presentations of 12 unrelated words over 12 trials. Alternate forms were used at baseline and follow-up to control for practice effects. RESULTS: The total number of words recalled across learning trials was higher while on Aricept (p=0.04). Consistency of recall over repeated trials, a measure of long-term storage, was also improved on Aricept (p=0.05). Delayed recall at 30' was not significantly different (p=0.11), nor was random long-term retrieval (p=0.76), which would indicate fluctuation in the ability to retrieve words from long term storage. Aricept did not facilitate improvement in other cognitive domains, including basic attention, visual sequencing, mental flexibility, or psychomotor speed. Quality of life scores showed no significant changes. Mean 3-month seizure frequency at baseline was 2.70 4.60, and during Aricept 3.06 4.52 (n=18, p=0.19, two-tailed paired t-test). Generalized tonic-clonic seizures averaged 0.24 0.97 in baseline vs. 0.82 2.40 on Aricept (p=0.09). However, two patients experienced tonic-clonic seizures while on Aricept, with no such seizures in the prior year. Side effects included diarrhea in four patients, and stomach cramps, insomnia, depression and blurred vision in one patient each. CONCLUSIONS: This open-label pilot study suggests that donepezil may improve memory in patients with epilepsy. Treatment was associated with a non-significant trend towards more convulsions. A randomized, controlled trial is warranted, but with close monitoring for possible increases of seizures.