Evaluation of the clinical utility of a carotid bruit

BACKGROUND: Uncertainty exists over whether listening for carotid bruits as part of the clinical examination is informative in terms of predicting the presence or severity of carotid stenosis. AIM: We sought to undertake a comprehensive meta-analysis and meta-regression of all studies to date that have assessed the relationship between a carotid bruit and severity of degree of stenosis. METHODS: Electronic databases were used to identify all published studies in humans evaluating the association between bruit and stenosis published until and including October 2011. Pooled sensitivity, specificity and diagnostic odds ratio (DOR) were calculated for each stenosis group. Summary receiver operating characteristic (SROC) curve analysis was performed in studies assessing clinically relevant (i.e. >70%) stenosis. Meta-regression was performed in all studies, using random effects. RESULTS: We identified 26 studies evaluating the association between carotid bruit and stenosis, in 15 117 arteries. For clinically relevant stenosis (i.e. >70%), we found pooled sensitivity 0.53 [95% confidence interval (CI): 0.5-0.55], specificity 0.83 (95% CI: 0.82-0.84) and DOR 4.32 (95% CI: 2.78-6.66). SROC curve analysis gave an area under the curve of 0.73. Meta-regression analysis showed a (non-significant) (P = 0.067) inverse relationship between carotid bruit and stenosis. CONCLUSION: The carotid bruit is of moderate value for detecting clinically relevant carotid stenosis. It gives high specificity but low sensitivity. The likelihood of a carotid bruit does not increase at increasing degrees of stenosis.