Anand Vaishakh, Gulati Sheffali, Agarwala Anuja, Kamila Gautam, Mahesan Aakash, Sondhi Vishal, Gupta Kanak L, Chakrabarty Biswaroop, Jauhari Prashant, Panda Prateek Kumar, Pandey Ravindra Mohan
Centre of Excellence & Advanced Research for Childhood Neurodevelopmental Disorders, Child Neurology Division, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India.
Department of Pediatrics, Senior Dietician, All India Institute of Medical Sciences, New Delhi, India.
Epilepsia. 2025 May;66(5):1550-1559. doi: 10.1111/epi.18292. Epub 2025 Jan 30.
The ketogenic diet has been the mainstay of treatment of drug-resistant epilepsy (DRE). No comparative trials have been conducted to assess the efficacy of the two less strict ketogenic diets: modified Atkins diet (MAD) and low glycemic index treatment (LGIT). This study assesses the non-inferiority of LGIT compared with MAD.
This was an open-label randomized non-inferiority trial. Children with DRE were randomized to receive either MAD or LGIT as an add-on to anti-seizure medications. The primary endpoint was percentage seizure reduction at the end of 24 weeks of therapy compared to the baseline. The non-inferiority margin of -15% was predefined to calculate the sample size.
Ninety-one children were enrolled and randomized to receive either MAD (n = 45) or LGIT (n = 46). Intention-to-treat analysis done at the end of 24 weeks of therapy showed a mean (±standard deviation [SD]) percentage seizure reduction of 60.7% (±41.3) in the MAD sub-group and 57% (±39.4) in the LGIT sub-group (p = 0.664). The absolute difference between the means of percentage seizure reduction was -3.7 (-20.5 to 13.2) and crossed the non-inferiority margin. Ten children in the MAD group and nine children in the LGIT group did not complete 24 weeks of therapy. Adverse effects were comparable between the arms (MAD, 66.6%; LGIT, 50%), although serious adverse effects were higher in the MAD arm. The most common adverse effect was decreased acceptance (24.2%) followed by decreased satiety (9.9%), vomiting (9.9%), weight loss (5.5%), constipation (5.5%), and diarrhea (3.3%). Dyslipidemia was more commonly seen in the MAD group (MAD, six; LGIT, one). One death in the LGIT arm was unrelated to therapy. Although there was no statistically significant difference in improvement in cognition, behavior, and quality of life scales, improvement was noted from baseline scores.
LGIT may be non-inferior to MAD in the treatment of children with DRE with the advantage of increased acceptance and fewer adverse effects.
生酮饮食一直是耐药性癫痫（DRE）治疗的主要方法。尚未进行比较试验来评估两种限制没那么严格的生酮饮食：改良阿特金斯饮食（MAD）和低血糖指数疗法（LGIT）的疗效。本研究评估LGIT与MAD相比的非劣效性。
这是一项开放标签随机非劣效性试验。DRE患儿被随机分配接受MAD或LGIT作为抗癫痫药物的附加治疗。主要终点是治疗24周结束时与基线相比癫痫发作减少的百分比。预先设定-15%的非劣效性界值来计算样本量。
91名儿童入组并随机接受MAD（n = 45）或LGIT（n = 46）治疗。在治疗24周结束时进行的意向性分析显示，MAD亚组癫痫发作减少的平均（±标准差[SD]）百分比为60.7%（±41.3），LGIT亚组为57%（±39.）（p = 0.664）。癫痫发作减少百分比均值的绝对差异为-3.7（-20.5至13.2），超过了非劣效性界值。MAD组有10名儿童和LGIT组有9名儿童未完成24周治疗。两组的不良反应相当（MAD组为66.6%；LGIT组为50%），尽管MAD组严重不良反应更高。最常见的不良反应是接受度降低（24.2%），其次是饱腹感降低（9.9%）、呕吐（9.9%）、体重减轻（5.5%）、便秘（5.5%）和腹泻（3.3%）。血脂异常在MAD组更常见（MAD组6例；LGIT组1例）。LGIT组有1例死亡与治疗无关。虽然在认知、行为和生活质量量表的改善方面没有统计学上的显著差异，但从基线评分来看有改善。
在治疗DRE儿童方面，LGIT可能不劣于MAD，具有接受度更高和不良反应更少的优势。
