Use of the ketogenic diet in drug resistant epilepsy syndromes during early infancy: Differences between 3: 1 and 4: 1 formula, a pilot study

The ketogenic diet (KD) is a high fat – low carbohydrate and sufficient in protein diet and has been established in the treatment of drug resistant epilepsy. The aim of this study was to assess the efficacy and safety of the KD in very young infants and toddlers, as KD formula are now readily available also below 1 year of age.

Included were prospectively all infants and toddlers with epilepsy who started on the KD at the Department of Paediatrics and Adolescent Health, Medical University of Vienna between July 2006 and September 2010. Data on efficacy and safety were analysed after three, six, and twelve months using Student's t-test and chi-square.

28 individuals were included, mean age at start was 1.68 years (SD±1.60). The mean number of antiepileptic drugs before starting the KD was 3.43 (SD±5.04, min.0, max. 10). 12 infants were started on a 4: 1 formula adjusted to a 3: 1 adding rapeseed oil and carbohydrates and 16 infants on a readily available 3: 1 KD-formula. 57,14% (n=16) were responders, i.e. showing a seizure reduction of ≥50%, 14,4%(4) became seizure-free, the remaining 42,85% (n=12) were non-responders. Infants who were on the 3:1 formula were significantly younger, showed a significant shorter treatment lag before starting, but had a trend in experiencing more short-term side effects. There were no differences for efficacy in both groups (43.8% vs.56.8%). The younger children were significantly more impaired on the neurological exam (Chi-quadrat 0.021). Carbohydrates per kg/day were significantly higher in the adjusted group. We noticed in a high number of young infants a good response in difficult-to-treat epilepsy, and hypothesize that this may be due to a high presence of patients with infantile spasm. There were no differences between available formulas in efficacy and safety. Moreover, only few side effects with no need to interrupt the KD were observed. Study limitations are the relatively small sample size and that outcome results were based on clinical neurological performance only. We have to increase our sample size prospectively.