Table 1.
Summary of published randomized controlled trials of ketogenic diet therapy efficacy for seizure management in children and adults
| Population | First author (year) |
Study design |
Subjects, n |
Diet | Comparison treatment |
Duration (mo) |
Significant study results (RR) | KDT attrition % (number; reason) |
|---|---|---|---|---|---|---|---|---|
| Children | Bergqvist (2005)18 | RCT | 48 | 24 CKD (4:1 ratio), 24–48 h fasting | 24 CKD (4:1 ratio), nonfasting | 3 | No difference in RR between groups (58 vs. 67%) | 13% (6/48; inefficacy, side effects) |
| Kossoff (2007)19 | RCT crossover | 20 | 10 MAD (10 g/d) | 10 MAD (20 g/d) | 3 (of each treatment), 6 (total study) | 60% (6/10) RR in 10 g MAD arm compared with 10% (1/10) in 20 g MAD arm @ 3 months; no difference between groups at 1 or 6 months | 20% @ 3 months (4/20; did not start or stopped diet) | |
| Seo (2007)20 | RCT | 76 | 40 CKD (4:1 ratio) | 36 KD (3:1 ratio) | 3 | 85% (34/40) RR in 4:1 KD arm compared with 72% (26/36) RR in 3:1 KD arm | 16% (12/76; intolerance, GI and other side effects) | |
| Neal (2008)21 | RCT | 145 | 37 CKD (3–4:1 ratio) 36 MCT-KD (45% fat content) | 72 usual diet | 3 | 38% (28/73) RR in diet arm compared with 6% (4/72) in control arm; also diet arm with lower mean percentage of baseline seizures | 26% (19/73; did not start, food refusal, too restrictive, increased seizures, side effects) | |
| Neal (2009)25 | RCT | 145 | 73 CKD (3–4:1 ratio) | 72 MCT-KD (40–45% fat content) | 12 | No difference in RR or mean percentage of baseline seizures between groups at 3, 6 or 12 months | 35% @ 3 months (51/145; did not start or stopped for increased seizures, food refusal, or side effects) | |
| Raju (2011)22 | RCT | 38 | 19 CKD (4:1 ratio) | 19 KD (2.5:1 ratio) | 3 | No difference in RR between groups (58 vs. 63%) | 16% (6/38; inefficacy, food refusal, nonacceptance by family) | |
| Sharma (2013)23 | RCT | 102 | 50 MAD (10 g/d) | 52 usual diet | 3 | 52% RR in MAD arm compared with 11.5% RR in control arm; also diet arm with lower mean percentage of baseline seizures | 8% (4/50; stopped diet early) | |
| El-Rashidy (2013)24 | RCT | 40 | 10 CKD (4:1 ratio) 15 MAD (10 g/d→ <10% carb) | 15 usual diet | 3, 6 | KD arm had a greater rate of change in seizure frequency than the MAD arm, and both diet groups had greater rate of change in seizure frequency than control arm | 16% @ 3 months (4/25; weight loss, intolerance) | |
| Sharma (2016)26 | RCT | 81 | 41 MAD (10 g/d) using simplified instructions | 40 usual diet | 3 | 56% (23/41) RR in diet arm compared with 8% (3/40) RR in control arm | 5% (2/41; food refusal, side effects) | |
| Kim (2016)27 | RCT | 104 | 51 CKD (4:1 ratio) | 53 MAD (10 g/d→ <10% carb) | 6 | No difference in RR between groups (39 vs. 36%); more seizure free @ 3 months on KD than on MAD if age 1 to <2 | 33% (34/104; inefficacy, intolerance, side effects) | |
| Lambrechts (2017)28 | RCT | 57 | 29 KDT (MCT-KD, CKD, or mixed) | 28 usual diet | 4 | 45% (13/29) RR in KD arm compared with 14% (4/28) RR in control arm | 21% (6/29; improved seizure, change in seizure pattern, GI side effects) | |
| Adult | Zare (2017)34 | RCT | 66 | 34 MAD (15 g/d) | 32 usual diet | 2 | 35% (12/34) RR in MAD arm compared with 0% in control arm. | 35% (12/34; stopped diet) |
| Kverneland (2018)35 | RCT | 75 | 37 MAD (15–20 g/d) | 38 usual diet | 3 | No difference in RR between groups (8 vs. 5%); more patients with 25–50% seizure reduction in diet group | 35% (13/37; did not start diet, increased seizures) | |
| McDonald (2018)36 | RCT crossover | 80 | 40 MAD (20 g/d) + ketogenic formula supplement daily | 40 MAD alone | 1 (of each treatment), 2 (total study) | No difference in RR between groups (55 vs. 52%), but trend for continued diet use in supplement arm (43 vs. 17%) | 30% (24/80; did not start /stopped diet or supplement, change in seizure medication) |
Abbreviations: carb, carbohydrate; CKD, classic ketogenic diet; d, day; g, grams; GI, gastrointestinal; KDT, ketogenic diet therapy; MAD, modified Atkins diet; MCT-KD, medium chain triglyceride-KD; NR, not reported; RCT, randomized controlled trial; RR, responder rate, defined as proportion of patients with ≥50% seizure reduction.