In isolated, sporadic dystonia, it has been occasionally reported that some patients might undergo symptom remission. However, the exact clinical characteristics of patients with remission remain understudied. Given the important prognostic and pathophysiological implications of dystonic remission, we here provide a systematic review of the literature and a meta-analysis to assess demographic and clinical features associated with this phenomenon. We also provide a list of operational criteria to better define dystonic remission. Using PubMed and Embase, we conducted a systematic literature search in March 2018. 626 records were screened, 31 studies comprising data of 2551 patients with predominantly reports from patients with cervical dystonia (n=1319) or blepharospasm/Meige syndrome (n=704) were included in qualitative analysis. Five studies reporting remission in cervical dystonia were eligible for meta-analysis. Complete remission was reported in 11.8% and partial remission for 4.4% of cases. Remission rates were higher in cervical dystonia than in blepharospasm/Meige (e.g. complete remission 15.4% vs. 5.8% respectively). Remission occurred on average 4.5 years after onset of dystonic symptoms. However, the majority of patients (63.8%) relapsed. Meta-analysis for cervical dystonia showed that patients with remission were significantly younger at symptom onset than patients without remission (mean difference -7.13 years 95% CI: -10.58 − -3.68, p<0.0001). Based on our findings, we propose that the degree, the conditions associated with the onset and the duration of remission are key factors to be considered in a unifying definition of dystonic remission.