BACKGROUND: Symptomatic vertebral artery stenosis is associated with a high risk of recurrent stroke, with higher risks for intracranial than for extracranial stenosis. Vertebral artery stenosis can be treated with stenting with good technical results, but whether it results in improved clinical outcome is uncertain. We aimed to compare vertebral stenting with medical treatment for symptomatic vertebral stenosis. METHODS: We did a preplanned pooled individual patient data analysis of three completed randomised controlled trials comparing stenting with medical treatment in patients with symptomatic vertebral stenosis. The primary outcome was any fatal or non-fatal stroke. Analyses were performed for vertebral stenosis at any location and separately for extracranial and intracranial stenoses. Data from the intention-to-treat analysis were used for all studies. We estimated hazard ratios (HRs) with 95% CIs using Cox proportional-hazards regression models stratified by trial. FINDINGS: Data were from 354 individuals from three trials, including 179 patients from VIST (148 with extracranial stenosis and 31 with intracranial stenosis), 115 patients from VAST (96 with extracranial stenosis and 19 with intracranial stenosis), and 60 patients with intracranial stenosis from SAMMPRIS (no patients had extracranial stenosis). Across all trials, 168 participants (46 with intracranial stenosis and 122 with extracranial stenosis) were randomly assigned to medical treatment and 186 to stenting (64 with intracranial stenosis and 122 with extracranial stenosis). In the stenting group, the frequency of periprocedural stroke or death was higher for intracranial stenosis than for extracranial stenosis (ten (16%) of 64 patients vs one (1%) of 121 patients; p<0·0001). During 1036 person-years of follow-up, the hazard ratio (HR) for any stroke in the stenting group compared with the medical treatment group was 0·81% CI 0·45-1·44; p=0·47). For extracranial stenosis alone the HR was 0·63 (95% CI 0·27-1·46) and for intracranial stenosis alone it was 1·06 (0·46-2·42; pinteraction=0·395). INTERPRETATION: Stenting for vertebral stenosis has a much higher risk for intracranial, compared with extracranial, stenosis. This pooled analysis did not show evidence of a benefit for stroke prevention for either treatment. There was no evidence of benefit of stenting for intracranial stenosis. Stenting for extracranial stenosis might be beneficial, but further larger trials are required to determine the treatment effect in this subgroup. FUNDING: None.