Carotid artery stenting is an alternative to carotid endarterectomy for the treatment of atherosclerotic carotid artery stenosis. This review updates a previous version first published in 1997 and subsequently updated in 2004, 2007, and 2012.To assess the benefits and risks of stenting compared with endarterectomy in people with symptomatic or asymptomatic carotid stenosis.We searched the Cochrane Stroke Group Trials Register (last searched August 2018) and the following databases: CENTRAL, MEDLINE, Embase, and Science Citation Index to August 2018. We also searched ongoing trials registers (August 2018) and reference lists, and contacted researchers in the field.Randomised controlled trials (RCTs) comparing stenting with endarterectomy for symptomatic or asymptomatic atherosclerotic carotid stenosis. In addition, we included RCTs comparing carotid artery stenting with medical therapy alone.One review author selected trials for inclusion, assessed trial quality and risk of bias, and extracted data. A second review author independently validated trial selection and a third review author independently validated data extraction. We calculated treatment effects as odds ratios (OR) and 95% confidence intervals (CI), with endarterectomy as the reference group. We quantified heterogeneity using the I² statistic and used GRADE to assess the overall certainty of evidence.We included 22 trials involving 9753 participants. In participants with symptomatic carotid stenosis, compared with endarterectomy stenting was associated with a higher risk of periprocedural death or stroke (the primary safety outcome; OR 1.70, 95% CI 1.31 to 2.19; P < 0.0001, I² = 5%; 10 trials, 5396 participants; high-certainty evidence); and periprocedural death, stroke, or myocardial infarction (OR 1.43, 95% CI 1.14 to 1.80; P = 0.002, I² = 0%; 6 trials, 4861 participants; high-certainty evidence). The OR for the primary safety outcome was 1.11 (95% CI 0.74 to 1.64) in participants under 70 years old and 2.23 (95% CI 1.61 to 3.08) in participants 70 years old or more (interaction P = 0.007). There was a non-significant increase in periprocedural death or major or disabling stroke with stenting (OR 1.36, 95% CI 0.97 to 1.91; P = 0.08, I² = 0%; 7 trials, 4983 participants; high-certainty evidence). Compared with endarterectomy, stenting was associated with lower risks of myocardial infarction (OR 0.47, 95% CI 0.24 to 0.94; P = 0.03, I² = 0%), cranial nerve palsy (OR 0.09, 95% CI 0.06 to 0.16; P < 0.00001, I² = 0%), and access site haematoma (OR 0.32, 95% CI 0.15 to 0.68; P = 0.003, I² = 27%). The combination of periprocedural death or stroke or ipsilateral stroke during follow-up (the primary combined safety and efficacy outcome) favoured endarterectomy (OR 1.51, 95% CI 1.24 to 1.85; P < 0.0001, I² = 0%; 8 trials, 5080 participants; high-certainty evidence). The rate of ipsilateral stroke after the periprocedural period did not differ between treatments (OR 1.05, 95% CI 0.75 to 1.47; P = 0.77, I² = 0%). In participants with asymptomatic carotid stenosis, there was a non-significant increase in periprocedural death or stroke with stenting compared with endarterectomy (OR 1.72, 95% CI 1.00 to 2.97; P = 0.05, I² = 0%; 7 trials, 3378 participants; moderate-certainty evidence). The risk of periprocedural death or stroke or ipsilateral stroke during follow-up did not differ significantly between treatments (OR 1.27, 95% CI 0.87 to 1.84; P = 0.22, I² = 0%; 6 trials, 3315 participants; moderate-certainty evidence). Moderate or higher carotid artery restenosis (50% or greater) or occlusion during follow-up was more common after stenting (OR 2.00, 95% CI 1.12 to 3.60; P = 0.02, I² = 44%), but the difference in risk of severe restenosis was not significant (70% or greater; OR 1.26, 95% CI 0.79 to 2.00; P = 0.33, I² = 58%; low-certainty evidence).Stenting for symptomatic carotid stenosis is associated with a higher risk of periprocedural stroke or death than endarterectomy. This extra risk is mostly attributed to an increase in minor, non-disabling strokes occurring in people older than 70 years. Beyond the periprocedural period, carotid stenting is as effective in preventing recurrent stroke as endarterectomy. However, combining procedural safety and long-term efficacy in preventing recurrent stroke still favours endarterectomy. In people with asymptomatic carotid stenosis, there may be a small increase in the risk of periprocedural stroke or death with stenting compared with endarterectomy. However, CIs of treatment effects were wide and further data from randomised trials in people with asymptomatic stenosis are needed.