The evidence in the effectiveness of deep brain stimulation in children with medication-refractory non-degenerative monogenic dystonia is heterogeneous and long-term results are sparse. The objective is to describe long-term outcomes in a single-center cohort and compare our results with a meta-analysis cohort form literature. We performed a retrospective single-center cohort study including consecutive pediatric patients with non-degenerative genetic or idiopathic dystonia treated with globus pallidus internus deep brain stimulation at our center and a systematic review and individual-patient data meta-analysis with the same inclusion criteria. The primary outcome was the change from baseline in the Burke-Fahn-Marsden Dystonia Rating Scale-movement (BFMDRS-M) score. The clinical cohort included 25 patients with a mean study follow-up of 11.4 years. The meta-analysis cohort included 224 patients with a mean follow-up of 3 years. Overall, the BFMDRS-M mean improvements at 1 year and at last follow-up were 41% and 33% in the clinical cohort and 58.9% and 57.2% in the meta-analysis cohort, respectively. TOR1A-dystonia showed the greatest and most stable BFMDRS-M improvement in both cohorts at 1 year and at last follow-up (76.3% and 74.3% in the clinical cohort; 69.6% and 67.3% in the meta-analysis cohort), followed by SGCE-dystonia (63% and 63.9% in the meta-analysis cohort). THAP1-dystonia (70.1% and 29.8% in the clinical cohort; 52.3% and 42.0% in the meta-analysis cohort) and KMT2B-dystonia (33.3% and 41.3% in the clinical cohort; 38.0% and 26.7% in the meta-analysis cohort) showed a less pronounced or sustained response. Globus pallidus deep brain stimulation long-term treatment seems effective with a possible gene-specific differential effect. © 2024 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.