A prospective multicenter study of the efficacy of the ketogenic diet in treatment of 51 children with intractable seizures is reported from 6 large institutions, including Johns Hopkins, Montefiore and Boston, USA, and Halifax, Canada, and 1 private practice setting. A 4:1 ketogenic diet was introduced in hospitalized patients after 36 hours fasting, according to the Johns Hopkins protocol. Pre-diet parental estimates of seizure frequencies were compared with seizure occurrences after 3, 6, and 12 months treatment. A 50% decrease in seizure frequency or greater was observed in 54% at 3 months, in 55% at 6 months, and in 40% at 1 year. Only 5 patients were seizure-free at 1 year. The fall out rate of patients starting the diet was 12% at 3 months, 31% at 6 months, and 53% at 1 year. Reasons for discontinuance of the diet were medical intolerance due to lethargy, dehydration, vomiting, or behavioral problems in 6, the restrictive dietary protocol in 4, and lack of seizure control in 12. Seizure type (tonic-clonic in 20, myoclonic (19), atonic (14), absence (8), and partial (9)) and EEG abnormality (generalized or multifocal epileptiform) were not related to outcome. 
COMMENT. That the ketogenic diet (KD) works has been established and corroborated by numerous investigators in the past. Is it of practical value in the management of childhood epilepsy is perhaps a more pressing question, given the advent of newer antiepileptic drugs? The above multicenter study, completed primarily in university settings, corroborates once again the partial efficacy of the KD, when introduced under hospitalized supervision, and when patients are followed closely by dietary, nursing, and physician services. Antiepileptic medication must be continued in the majority of patients treated, compounding the cost of dietary services, frequently not covered by health insurance plans.
Despite close supervision, complete control of seizures may be expected in only 10% of patients and 53% will have discontinued the diet after one year. Medical intolerance, the restrictive regimen, and insufficient seizure control to justify the demands of the diet are reasons for its withdrawal. Serious side effects, not mentioned in the above report, are the subject of a paper recently published by one of the participant authors and institutions (see Ped Neur Briefs Aug 1998; 12:60) . In 5 (10%) of 52 children treated with the KD, adverse events included hypoproteinemia, lipemia, hemolytic anemia, and Fanconi’s renal tubular acidosis; two patients had severe abnormalities of liver function, one with concomitant thrombocytopenia.
The ketogenic diet may be of value as a short-term therapy in a limited group of children with refractory epilepsy or in those exhibiting serious adverse reactions to AEDs. Its use should perhaps be restricted to research institutions where studies of the mechanism of action of the diet might uncover an essential specific ingredient and further our understanding of seizure pathophysiology (see Progress in Pediatric Neurology, PNB Publ, Vol I, 1991;pp 87-88; Vol III, 1997;pp578-9). To paraphrase the conclusions of an editorial , “Those who champion alternative treatments have an obligation to study them thoroughly before promoting them for general use with patients.” “They should also be accountable for the cost-effectiveness of their methods.” The Mayo Clinic KD protocol, which avoids the necessity for pre-diet starvation and hospitalization of the patient, is in my experience more practical and cost-effective and equally successful.