Evidence table for systematic reviews
Study reference Design Inclusion Criteria Quality Results Comments

Levy, 2012

Cochrane review

Studies of ketogenic diets and similar diets for people with epilepsy. Trials using adequate methods of concealment as well as quasi-randomised trials were considered.

Search: adequate

 

Selection: adequate

 

Quality assessment of individual studies: adequate

Four RCT’s which generated five publications were analysed.  Also six prospective and five retrospective studies were included. These are not described here.

 

Berqvist 2005

Randomised trial of fasting versus gradual introduction of ketogenic diet. Both arms contained 24 children. At 3 months the fasting group showed 58% >50% reduction and 21% seizure free and the gradual group 67% >50% reduction and 21% seizure free. Differences were not significant, but evidence suggests higher metabolic hypoglycaemia for fast onset.

 

Kossof 2007

Randomised crossover study comparing two modified Atkins diets (10g vs 20g carbohydrate daily) in 20 children. Percentage >50% reduction for 10g diet and 60% for 20g diet at 3-month. 20g diet was better tolerated.

 

Neal 2008

RCT of immediate versus (3-month)delayed introduction of ketogenic diet. 145 children included, 103 completed trial. After 3 months initial treatment  showed 38% >50% reduction and 1 seizure free compared to 6% >50% reduction and 0 seizure free in the control group. No differences in adverse effects were found.

 

Neal 2009

RCT comparing classical diet with Medium Chained Triglyceride (MCT) to 145 children. At 3-months both groups showed 32% >50% reduction and 1 patient seizure-free. Evidence suggests no difference in adverse effects

 

Seo 2007

RCT comparing lipid contents of two ketogenic diets (3:1 vs. 4:1) to 76 children. The 4:1 diet was more effective (3-month: 85% >50% reduction, 55% seizure-free), however children maintained response when switched to 3:1 (3-month: 72,1% >50% reduction, 30.5% seizure-free). The 3:1 diet was more tolerable.

Martin et al., 2016

Cochrane review

All RCTs or quasi-RCTs (using adequate methods of allocation concealment) of KD and other dietary interventions for people with epilepsy.

Search: adequate

Selection: adequate

Quality assessment of individual studies: adequate

In total, 7 RCT’s were included. Four of the identified studies were included in previous versions of this review (see above in the study of Levy (2012) (Bergqvist 2005; Kossoff 2007; Neal 2008; Seo 2007). Therefore three new RCTs were included in this update:

Raju, 2011

Randomized, non-blinded, open-label, parallel controlled trial of childrend 6 mnths- 5 years olld with refracttory epilepsy, comparing a 4:1 (n=19) and a 2.5:1 ratio KD (n=19).

-Seizure reduction:

No significant difference in seizure reduction between a 4 : 1

KD and a 2.5 : 1 KD was found.

-Seizure freedom:

26% (5/19) of participants following a 4 : 1 KD and 21% (4/19) of participants following a 2.5 : 1 KD was seizure free at three months.

->50% seizure reduction after three months: 58% (11/19) in the 4 : 1 KD group and 63% (12/19) in the 2.5 : 1 KD group; however, there was no significant difference.

-Adverse effects:

the main adverse effects were gastrointestinal symptoms, including vomiting, constipation and diarrhea. Weight loss did affect more participants (3/19) in the 4 : 1 ratio KD group than in the 2.5 : 1 (1/19) ratio KD group.

-

El-Rashidy, 2013

Single-centre RCT of 40 patients 12-36 mnths old to compare 2 different dietary interventions and a control group: MAD and classic ketogenic liquid diet.

-Adverse effects:

the main adverse effects were gastrointestinal symptoms, including vomiting, constipation and diarrhea. Constipation affected 15.4% of participants in the MAD group and 25% of participants in the classic group, but no significance was reported.

 -

Sharma, 2013

An open-label, parallel-group, RCT of children aged two to 14 years with refractory epilepsy comparing the MAD (n=50) to a control group (n=52). (details are described above in this table). 

The studies included in this review were limited by small numbers of participants and they only included children; therefore, the quality of the evidence was low.