Objectives: Deep brain stimulation of the internal pallidum (GPi-DBS) is an established therapeutic option in treatment-refractory dystonia, and the identification of factors predicting surgical outcome is needed to optimize patient selection. Methods: In this retrospective multicenter study, GPi-DBS outcome of 8 patients with DYT6, 9 with DYT1, and 38 with isolated dystonia without known monogenic cause (non-DYT) was assessed at early (1-16 months) and late (22-92 months) follow-up using Burke-Fahn-Marsden Dystonia Rating Scale (BFMDRS) scores. Results: At early follow-up, mean reduction of dystonia severity was greater in patients with DYT1 (BFMDRS score: -60%) and non-DYT dystonia (-52%) than in patients with DYT6 dystonia (-32%; p = 0.046). Accordingly, the rate of responders was considerably lower in the latter group (57% vs >90%; p = 0.017). At late follow-up, however, GPi-DBS resulted in comparable improvement in all 3 groups (DYT6, -42%; DYT1, -44; non-DYT, -61%). Additional DBS of the same or another brain target was performed in 3 of 8 patients with DYT6 dystonia with varying results. Regardless of the genotype, patients with a shorter duration from onset of dystonia to surgery had better control of dystonia postoperatively. Conclusions: Long-term GPi-DBS is effective in patients with DYT6, DYT1, and non-DYT dystonia. However, the effect of DBS appears to be less predictable in patients with DYT6, suggesting that pre-DBS genetic testing and counseling for known dystonia gene mutations may be indicated. GPi-DBS should probably be considered earlier in the disease course. Classification of evidence: This study provides Class IV evidence that long-termGPi-DBS improves dystonia in patients with DYT1, DYT6, and non-DYT dystonia.