ObjectiveWe synthesized evidence for effectiveness of vagus nerve stimulation (VNS) as adjuvant therapy in pediatric drug-resistant epilepsy (DRE) by obtaining pooled estimates for seizure outcomes and analyzing their determinants.MethodsMEDLINE, EMBASE, and Cochrane databases were searched up to July 2019, for original research on VNS in pediatric (≤18 years-of-age) epilepsy. The primary outcome was 50% responder rate (50%-RR), the proportion of patients with ≥50% seizure reduction, at the last reported follow-up. Other outcomes included 50%-RR and proportion of seizure-free patients at additional reported time points. A random effects meta-analysis with restricted maximum likelihood estimation was performed to obtain pooled effect estimates. Meta-regression using multiple linear models was performed to obtain determinants of seizure outcomes and sources of heterogeneity.ResultsA total of 101 studies were included. The pooled prevalence estimates for 50%-RR and seizure freedom at last follow-up (mean 2.54 years) were 56.4% (95% confidence intervals [CIs] 52.4, 60.4) and 11.6% (95% CI 9.6, 13.9) respectively. Fewer anti-seizure medications (ASMs) tried before VNS, and later age at onset of seizures were associated with better seizure outcomes following VNS implantation. An effect of sex-distribution of studies on long-term outcomes and a potential publication bias for short-term outcomes were also observed.ConclusionPooled evidence supports possible effectiveness of VNS in pediatric DRE, although complete seizure freedom is less common. Early referral (fewer trials of ASMs) may be a modifiable factor for desirable seizure outcomes with VNS from a clinical perspective.