The aim of this study was to examine the impact of dystonia aetiology and duration, contracture, and age at deep brain stimulation (DBS) surgery on outcome in a cohort of children with medically refractory, disabling primary, secondary-static, or secondary-progressive dystonias, including neurodegeneration with brain iron accumulation (NBIA). Dystonia severity was assessed using the Burke-Fahn-Marsden Dystonia Rating Scale (BFMDRS) motor score at baseline and 6 and 12months postoperatively in a cohort of 70 consecutive children undergoing DBS between June 2005 and July 2011. Two children (3%) received unilateral DBS for hemidystonia and were excluded and five (7%) developed infections requiring part-DBS removal within 6months, leaving 63 children (90%) undergoing bilateral DBS for follow-up (34 males, 29 females; mean age at surgery for the whole group 10y 4mo, SD 4y 2mo, range 1-14y). Seventeen children were classified with primary dystonia: mean age 12years 11months, SD 4years 6months range 4years 6months to 17years 3months; 28 as having secondary-static dystonia: mean age 10years 2months, SD 4years 9months (range 3y 3mo-20y); five as having secondary-progressive dystonia: mean age 8years 11months, SD 3years 9months (range 5y 5mo-13y 1mo); and 13 as having NBIA dystonia: mean age 10years 2months, SD 3years 11months (range 1-14y). Children with primary dystonias demonstrated greater improvements in BFMDRS motor score than those in the other aetiological categories (Kruskal-Wallis test, p<0.001), which correlated negatively with dystonia duration and more strongly still against the ratio of dystonia duration normalized to age at surgery (DD/AS ratio) at 1year (Spearman's rank correlation coefficient 0.4752 and -0.599 respectively). A similar significant negative correlation was found in the secondary-static dystonia group between outcome at 1year and DD/AS ratio (-0.461). Poorer outcome in secondary dystonia coincided with the absence of a period of normal motor development in comparison with the primary dystonia group. A significant improvement in BFMDRS motor score was seen in the NBIA group at 6, but not 12months (Wilcoxon signed rank test p=0.028, p=0.85 respectively). No reduction in efficacy was seen in children with a musculoskeletal deformity at the time of surgery. Response to pallidal DBS in the treatment of dystonia declines with the proportion of life lived with dystonia in primary and secondary dystonia. Other intrinsic factors reduce the median magnitude of reduction in secondary dystonia after DBS. DBS should be offered early, preferably within 5years of onset, to maximize benefits and reduce the childhood experience of dystonia, including musculoskeletal deformity. Other multidimensional assessments are required to understand how DBS improves the lives of children with dystonia.