RESUMO
OBJECTIVE: Identification of patients who will benefit from carotid endarterectomy is not entirely effective, primarily utilising degree of carotid stenosis. This study aimed at determining if microembolic signals (MES) detected by transcranial Doppler ultrasound (TCD) can provide clinically useful information regarding stroke risk in patients with carotid atherosclerosis. METHODS: A meta-analysis of prospective studies was performed. Three analyses were proposed investigating MES detection as a predictor of: stroke or TIA, stroke alone, and stroke or TIA but with an increased positivity threshold. Subgroup analysis was used to compare pre-operative (symptomatic or asymptomatic) patients and peri- or post-operative patients. RESULTS: Twenty-eight studies reported data regarding both MES status and neurological outcome. Of these, 22 papers reported data on stroke and TIA as an outcome, 19 on stroke alone, and eight on stroke and TIA with increased positivity threshold. At the median pre-test probability of 3.0%, the post-test probabilities of a stroke after a positive and negative TCD were 7.1% (95% CI 5-10.1) and 1.2% (95% CI 0.6-2.5), respectively. In addition, the sensitivities and specificities of each outcome showed that increasing the threshold for positivity to 10 MES per hour would make TCD a more clinically useful tool in peri- and post-operative patients. CONCLUSION: TCD provides clinically useful information about stroke risk for patients with carotid disease and is technically feasible in most patients. However, the generally weak level of evidence constituting this review means definitive recommendations cannot be made.