Risk of stroke in relation to degree of asymptomatic carotid stenosis: a population-based cohort study, systematic review, and meta-analysis

Abstract

Background: There is uncertainty around which patients with asymptomatic carotid stenosis should be offered surgical intervention. Although stroke rates were unrelated to the degree of stenosis in the medical-treatment-only groups in previous randomised trials, this could simply reflect recruitment bias and there has been no systematic analysis of a stenosis-risk association in cohort studies. We aimed to establish whether there is any association between the degree of asymptomatic stenosis and ipsilateral stroke risk in patients on contemporary medical treatment.

Methods: We did a prospective population-based study (Oxford Vascular Study; OxVasc), and a systematic review and meta-analysis. All patients in OxVasc with a recent suspected transient ischaemic attack or stroke, between April 1, 2002, and April 1, 2017, who had asymptomatic carotid stenosis were included in these analyses. We commenced contemporary medical treatment and determined ipsilateral stroke risk in this cohort by face-to-face follow-up (to Oct 1, 2020). We also did a systematic review and meta-analysis of all published studies (from Jan 1, 1980, to Oct 1, 2020) reporting ipsilateral stroke risk in patients with asymptomatic carotid stenosis. We searched MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials, and included both observational cohort studies and medical treatment groups of randomised controlled trials if the number of patients exceeded 30, ipsilateral stroke rates (or the raw data to calculate these) were provided, and were published in English.

Findings: Between April 1, 2002, and April 1, 2017, 2354 patients were consecutively enrolled in OxVasc and 2178 patients underwent carotid imaging, of whom 207 had 50-99% asymptomatic stenosis of at least one carotid bifurcation (mean age at imaging: 77·5 years [SD 10·3]; 88 [43%] women). The 5-year ipsilateral stroke risk increased with the degree of stenosis; patients with 70-99% stenosis had a significantly greater 5-year ipsilateral stroke risk than did those with 50-69% stenosis (six [14·6%; 95% CI 3·5-25·7] of 53 patients vs none of 154; p<0·0001); and patients with 80-99% stenosis had a significantly greater 5-year ipsilateral stroke risk than did those with 50-79% stenosis (five [18·3%; 7·7-29·9] of 34 patients vs one [1·0%; 0·0-2·9] of 173; p<0·0001). Of the 56 studies identified in the systematic review (comprising 13 717 patients), 23 provided data on ipsilateral stroke risk fully stratified by degree of asymptomatic stenosis (in 8419 patients). Stroke risk was linearly associated with degree of ipsilateral stenosis (p<0·0001); there was a higher risk in patients with 70-99% stenosis than in those with 50-69% stenosis (386 of 3778 patients vs 181 of 3806 patients; odds ratio [OR] 2·1 [95% CI 1·7-2·5], p<0·0001; 15 cohort studies, three trials) and a higher risk in patients with 80-99% stenosis than in those with 50-79% stenosis (77 of 727 patients vs 167 of 3272 patients; OR 2·5 [1·8-3·5], p<0·0001; 11 cohort studies). Heterogeneity in stroke risk between studies for patients with severe versus moderate stenosis (p_het<0·0001) was accounted for by highly discrepant results (p_diff<0·0001) in the randomised controlled trials of endarterectomy compared with cohort studies (trials: pooled OR 0·8 [95% CI 0·6-1·2], p_het=0·89; cohorts: 2·9 [2·3-3·7], p_het=0·54).

Interpretation: Contrary to the assumptions of current guidelines and the findings of subgroup analyses of previous randomised controlled trials, the stroke risk reported in cohort studies was highly dependent on the degree of asymptomatic carotid stenosis, suggesting that the benefit of endarterectomy might be underestimated in patients with severe stenosis. Conversely, the 5-year stroke risk was low for patients with moderate stenosis on contemporary medical treatment, calling into question any benefit from revascularisation.

Funding: NIHR Oxford Biomedical Research Centre, Wellcome Trust, Wolfson Foundation, and the British Heart Foundation.

Figure 1

Risk of recurrent vascular events by degree of asymptomatic ipsilateral carotid stenosis in 207 patients with asymptomatic carotid stenosis in the Oxford Vascular Study The exact range and cutoff points used to report the stenosis and risk data in trials and previous studies vary, with some defining severe stenosis as 70–99% and some as 80–99%. We therefore analysed both cutoff points in the Oxford Vascular Study. TIA=transient ischaemic attack. HR=hazard ratio.

Figure 2

Time trends in pooled estimates of annual ipsilateral stroke rates in medically treated patients with asymptomatic carotid stenosis, stratified by degree of baseline carotid artery stenosis, in a systematic review of studies published after 1980 Data are pooled annual ipsilateral stroke rates and 95% CIs. (A) Stroke rate distal to 50–99% stenosis in all studies (n=56). (B) Stroke rate distal to 50–69% versus 70–99% stenosis in all non-selective cohort studies reporting risk for patients with 50–69% or 70–99% stenosis (n=30). (C) Stroke rate distal to 50–99% stenosis in all non-selective cohort studies reporting risk in both patients with 50–69% stenosis and those with 70–99% stenosis (n=15). (D) Stroke rate distal to 50–69% versus 70–99% stenosis in all non-selective cohort studies reporting risk in both patients with 50–69% stenosis and those with 70–99% stenosis (n=15).

Figure 3

Time trends in statin use, current smoking prevalence, and duration of follow-up, and ipsilateral stroke rates stratified by confounding factor Data are pooled annual ipsilateral stroke rates and 95% CIs. Data are from all non-selective cohort studies (n=30) reporting stroke risk for patients with 50–69% or 70–99% stenosis. Time trends for medically treated patients with asymptomatic carotid stenosis are shown on the left.

Figure 4

Odds of ipsilateral stroke by degree of stenosis in medically treated patients with asymptomatic carotid stenosis (A) Five cohort studies reporting stroke risk for patients with 0–99% stenosis. (B) Seven cohort studies reporting stroke risk for patients with 50–99% stenosis. (C) The medical treatment group of randomised trials. OR=odds ratio.