Predictive power of first morning glucose and the ketogenic diet.

The purpose of this study was to determine if hypoglycemia or hyperglycemia predicts the response to a ketogenic diet (KD) in a cohort of children with intractable epilepsy. We evaluated whether morning blood glucose during the initial 21 days after initiation of the KD in children with IE was related to seizure reduction after 3 months of treatment. The relation between change in weight status and blood glucose was also explored. Fasting morning whole blood glucose was measured each day for the first 21 days after initiation of KD. Weight and height were obtained at baseline, day of discharge, and at 0.5 and 1 month of full KD therapy. Associations among clinical response to the KD (responder status defined as >50% reduction of seizure frequency at 3 months), hypoglycemia, hyperglycemia, style of KD initiation protocol (fasting or gradual) and weight status were evaluated. Forty-five subjects age 1-12 years were enrolled. KD responder status was not associated with low or elevated blood glucose or type of initiation style protocol. Variability in day-to-day blood glucose also did not predict response to KD. Children who had declining weight status during KD initiation were more likely to be hypoglycemic during full KD therapy. Low blood glucose during KD therapy was not necessary for clinically significant seizure reduction. Hypoglycemia was related to declining weight status irrespective of initiation style protocol. An effective KD can be provided in a manner to minimize side-effects and maximize efficacy.