Inadequate control of seizures despite maximal treatment, or refractory epilepsy, occurs in up to 30 percent of the 2.5 million persons with epilepsy in the United States. Options for treating refractory epilepsy include using the newer antiepileptic medications (lamotrigine, gabapentin, felbamate and vigabatrin), surgery (resection of an epileptiform focus) and nutritional strategies. Specifically, a ketogenic diet has been advocated—one that forces the use of ketone bodies as a fuel source. This diet contains four times more fats than carbohydrates, and protein is severely limited. About 90 percent of the calories in the diet come from fats. It is not clear what mechanism of action accounts for reports of decreased seizure activity during ketosis, but the use of such diets has been described for decades. Lefevre and Aronson reviewed the literature to determine if there was any evidence that a ketogenic diet could be useful in pediatric patients with intractable epilepsy.

A literature search required the studies to meet various inclusion criteria, including reporting the relevant clinical outcomes and including the treatment of a ketogenic diet. Reduction in seizure frequency was the main outcome measured. Studies that reported only subjective findings were excluded.

Eleven published studies and one unpublished study were analyzed. None was a controlled study. The range of patients who became seizure-free on the diet was 7 to 33 percent, the range in patients who had more than a 90 percent reduction in seizures was 22 to 56 percent and the range for more than a 50 percent reduction (still considered clinically significant) was 29 to 100 percent. Noncompliance with the diet was reported to range from 3 to 32 percent. Adverse effects included gastrointestinal symptoms (in up to 50 percent of children), kidney stones and metabolic abnormalities, although the latter two effects occurred in fewer than 5 percent of children.

Based on the available evidence, the authors conclude that a ketogenic diet may have a beneficial effect on the incidence of seizures in children with intractable epilepsy. Although the studies were not placebo-controlled, it is not likely that the results are merely caused by the placebo effect or by the spontaneous remission that sometimes occurs in childhood epilepsy. In children in whom other options have not been entirely successful at eliminating seizures, a ketogenic diet should be given some consideration.

editor's note: Clearly, the evidence in this study is not of the highest quality. However, in the subset of children who have exhausted pharmacologic options and may even have undergone surgery, a ketogenic diet may provide possible relief for refractory epilepsy. Head-to-head comparisons of the newer antiepileptic medications (which are not generally labeled for use in children, although they are fairly commonly used), surgical options, ketogenic diets and placebo are needed.—g.b.h.