Background and Purpose: Despite successful use of Ketogenic Diets (KD) in refractory epilepsy, its efficacy in children has been still limited. The aim of this meta-analysis was to summarize the scientific evidence of KD, especially the clinical efficacy of KD in children with refractory epilepsy. Methods: Electronic searches of PubMed, EMBASE, and the Cochrane Library were conducted to compare the efficacy of KD in children with refractory epilepsy in randomized controlled trails (RCTs) with that incare as usual (CAU); the included studies were reviewed. Meta-analyses were performed using RevMan 5.3 software. The proportion of 50%, 90% reduction in seizures frequency and free seizures were assessed by risk ratio (RR). The percentage of reduction in seizure frequency was evaluated by weight mean difference (WMD) and the 95% confidence interval (95% CI) was reported in all results. In addition, constipation and vomiting were assessed by risk ratio (RR), while other AEs were reported narratively. Results: 3 RCTs were included, with a total sample size of 295 children. Compared with the CAU group, the numbers of 50% reduction in seizures frequency (RR 4.65, 95% CI 2.75-7.85, P < 0.00001), 90% reduction in seizures frequency (RR 3.42; 95% CI: 1.61-7.29, P=0.001) and the free seizures (RR 3.12; 95% CI 0.91-10.70, P=0.07) were all statistically significantly higher in the KD group, while the percentage of reduction in seizures frequency was lower in the KD group (MD 46.41, 95% CI -62.03--30.78, P < 0.00001). Gastrointestinal symptoms were the most common adverse effects (AEs), while the cases of constipation (RR 40.65, 95% CI 5.73-288.21, P < 0.0002) and vomiting (RR 18.1, 95% CI 2.42-133.98, P=0.005) were all statistically significantly higherin the KD group than that inthe CAU group. Conclusion: The results indicate that a KD is a kind of promising complementary therapy for children with refractory epilepsy, while stricter monitor measures should be implemented for the potential AEs.