Systematic review finds low-certainty evidence against endovascular therapy superiority in vertebral artery stenosis

Rationale Vertebral artery stenosis (VAS), which refers to the narrowing of the vertebral artery, is a significant cause of posterior circulation ischaemic stroke. Medical treatment involves managing risk factors and using medications, while endovascular treatment typically consists of percutaneous transluminal angioplasty, which may be performed with or without stenting. Despite these available options, optimal management of people with symptomatic vertebral artery stenosis has not yet been established. Objectives To assess the benefits and harm of percutaneous transluminal angioplasty, with or without stenting, plus medical treatment (MT), compared with MT alone, in people with episodes of vertebrobasilar ischaemia due to vertebral artery stenosis. Search methods We searched MEDLINE, Embase, BIOSIS, and two other indexes on the Web of Science, China Biological Medicine Database, Chinese Science and Technique Journals Database, China National Knowledge Infrastructure and Wanfang Data, as well as ClinicalTrials.gov trials register and the World Health Organisation (WHO) International Clinical Trials Registry Platform to 9 Dec 2025. Eligibility criteria We included all randomised controlled trials (RCTs) that compared endovascular therapy (ET) plus MT with MT alone in treating people aged 18 years or over with symptomatic VAS. We included all types of ET modalities, such as angioplasty alone, balloon‐mounted stenting, and angioplasty followed by the placement of a self‐expanding stent. The MT regimen encompassed risk‐factor control, antiplatelet therapy, lipid‐lowering therapy, and individualised management for patients with hypertension or diabetes. Outcomes Our outcomes were 30‐day post‐randomisation death or stroke; fatal or non‐fatal stroke in the territory of the treated vertebral artery from 30 days after randomisation to the end of follow‐up; stroke (ischaemic or haemorrhagic) during the entire follow‐up period; death during the entire follow‐up period; stroke or transient ischaemic attack (TIA) during the entire follow‐up period; and ≥ 50% restenosis of the treated vertebral artery documented by conventional cerebral angiography. All strokes and TIAs were newly diagnosed events. We evaluated all important outcomes during the entire follow‐up period. Risk of bias We assessed risk of bias in RCTs using version 1 of the Cochrane tool (RoB 1). Synthesis methods Two review authors independently screened studies, extracted data, and assessed risk of bias. For dichotomous outcomes, we calculated risk ratios with 95% confidence intervals and pooled results using fixed‐effect meta‐analysis when studies were sufficiently similar; otherwise, we provided a narrative synthesis. We used GRADE methods to assess the certainty of evidence and summarised key outcomes in a summary of findings table. Included studies We included a total of four multicentre RCTs with 429 participants who had symptomatic vertebral artery stenosis, all comparing endovascular treatment plus medical treatment versus medical treatment alone in adults with recent posterior circulation transient ischaemic attack or non‐disabling ischaemic stroke. Synthesis of results We included four RCTs with 429 participants who had symptomatic vertebral artery stenosis, with a mean age of 63.4 years. Three of the four RCTs (VAST, VIST, and SAMMPRIS) were stopped early, and all trials had a high risk of performance bias because blinding of the endovascular therapy was not feasible, which limits the precision of treatment‐effect estimates. Overall, the certainty of the evidence is low, mainly due to the high risk of performance bias, early termination of three of the four trials, small sample sizes, and imprecision of the effect estimates.There was no clear evidence of a difference in 30‐day post‐randomisation deaths/strokes between ET plus MT and MT alone (risk ratio (RR) 2.02, 95% confidence interval (CI) 0.73 to 5.55; 4 studies, 429 participants; low‐certainty evidence). There was no clear evidence of a difference between ET plus MT and MT alone in fatal/non‐fatal strokes in the territory of the treated vertebral artery stenosis after 30 days post‐randomisation to completion of follow‐up (RR 0.54, 95% CI 0.29 to 1.01; 4 studies, 429 participants; low‐certainty evidence); in ischaemic or haemorrhagic stroke during the entire follow‐up period (RR 0.76, 95% CI 0.46 to 1.26; 4 studies, 429 participants; low‐certainty evidence); or in death during the entire follow‐up period (RR 0.83, 95% CI 0.41 to 1.66; 4 studies, 429 participants; low‐certainty evidence). Low‐certainty evidence suggests that ET plus MT may make little or no difference to stroke or transient ischaemic attack during follow‐up (RR 0.65, 95% CI 0.39 to 1.06; 2 studies, 234 participants; low‐certainty evidence). Of the prespecified outcomes, no included study reported restenosis or good functional outcome, so these outcomes could not be analysed. Authors' conclusions Based on four RCTs including 429 participants, there may be little to no difference between endovascular treatment plus medical treatment and medical treatment alone in 30‐day post‐randomisation death/stroke, fatal/non‐fatal stroke in the territory of the treated vertebral artery stenosis after 30 days post‐randomisation to completion of follow‐up, stroke (ischaemic or haemorrhagic) during the entire follow‐up period, death during the entire follow‐up period, and stroke or transient ischaemic attack during the entire follow‐up period. Overall, the certainty of the evidence is low. We found no reliable evidence that endovascular treatment plus medical treatment is superior to medical treatment alone in preventing stroke or death. Confidence intervals are wide and compatible with either modest benefit or modest harm from the addition of endovascular treatment. Funding This work was supported by the National Natural Science Foundation of China (grant number: 82301468 and 82501574), the Beijing Nova Program (grant number: 20230484336), the Beijing Hospitals Authority's Ascent Plan (grant number: DFL20220702), the Xuanwu Hospital Talent Seed Program (grant number: YC20250107), and the Beijing Hospitals Authority Clinical Medicine Development of Special Funding Support (grant number: ZLRK202320). Registration The protocol to this review has not been published. The original review can be accessed as: DOI: 10.1002/14651858.CD013692.pub2. PICOs PICOs Population Intervention Comparison Outcome