Science
Pavel Klein, Ivana Tyrlikova, Gregory C. Mathews
We review adjunctive ketogenic diet (KD) and modified Atkins diet (MAD) treatment of refractory epilepsy in adults. Only a few studies have been published, all open-label. Because of the disparate, uncontrolled nature of the studies, we analyzed all studies individually, without a meta- analysis. Across all studies, 32% of KD-treated and 29% of MAD-treated patients achieved $50% seizure reduction, including 9% and 5%, respectively, of patients with .90% seizure frequency reduction. The effect persists long term, but, unlike in children, may not outlast treatment. The 3:1 and 4:1 [fat]:[carbohydrate 1 protein] ratio KD variants and MAD are similarly effective. The anticonvulsant effect occurs quickly with both diets, within days to weeks. Side effects of both diets are benign and similar. The most serious, hyperlipidemia, reverses with treatment discontinuation. The most common, weight loss, may be advantageous in patients with obesity. Potential barriers to large-scale use of both diets in adults include low rate of diet acceptance and high rates of diet discontinuation. The eligible screened/enrolled subject ratios ranged from 2.9 to 7.2. Fifty-one percent of KD-treated and 42% of MAD-treated patients stopped the diet before study completion. Refusal to participate was due to diet restrictiveness and complexity, which may be greater for KD than MAD. However, long-term adherence is low for both diets. Most patients eventually stop the diet because of culinary and social restrictions. For treatment of refractory status epilepticus, only 14 adult cases of KD treatment have been published, providing insufficient data to allow evaluation. In summary, KD and MAD treatment show modest efficacy, although in some patients the effect is remarkable. The diets are well-tolerated, but often discontinued because of their restrictiveness. In patients willing to try dietary treatment, the effect is seen quickly, giving patients the option whether to continue the treatment.
