A randomized controlled trial of the ketogenic diet in refractory childhood epilepsy

    loading  Checking for direct PDF access through Ovid


Non‐pharmacological treatment option in children such as ketogenic diet (KD) is under‐explored in management of drug‐resistant epilepsy. A recent Cochrane review summarizes the effectiveness of KD (4:1) to be around 58% in achieving >50% seizure reduction at 3 months.1 We read with great interest the recent randomized controlled trial (RCT) published by Lambrechts et al.2 In this trial, the authors concluded that KD was 50% effective in achieving around 50% seizure frequency reduction at 4 months in KD group as compared to care as usual (CAU) group. However, we would like to highlight our observations about this study.
Firstly, in this study, the authors intended to study the efficacy of KD in comparison with the usual care. Nonetheless, the authors overtly do not mention about the study design, if it is a superiority, equivalence, or non‐inferiority trial. Secondly, the heart of a RCT is the process of randomization and allocation concealment. However, in the study methodology, there has been no explicit mention about the technique used for generating allocation sequence or its concealment. In any anti‐epileptic drug trial, while comparing the efficacy of two drugs, the baseline seizure frequency assessment becomes very important. The authors have chosen a duration of 4 weeks for assessment of baseline seizure frequency; however, a 4 weeks duration would be insufficient to assess the mean frequency of baseline seizures as 25% of children in their cohort had a usual seizure frequency <1/per month. We suggest a baseline duration of at least 8 weeks for assessing the mean frequency of seizures in adjunctive trials.3 The authors have mentioned that the primary outcome has been analyzed by using the “intention to treat” (ITT) analysis. ITT analysis means including all patients after randomization for analysis irrespective of their compliance, group which they were randomized or any protocol deviation. It is vital to follow the dictum “once randomized, always analyzed” and avoids overestimation of study results.4 However, in the study, authors have excluded six dropouts in CAU group and three dropouts in KD group after randomization, for the assessment of primary outcome. The authors have not specified the details of diet administration, composition, and method for compliance assessment. Lastly, a mention of P‐value when describing the results of primary outcome assessment in abstract or table would have improved the statistical interpretation of the results.

Related Topics

    loading  Loading Related Articles