Are African American stroke patients receiving appropriate carotid ultrasound exams?
Background: Previous studies have suggested that carotid ultrasound (CUS) tests are less likely to be utilized in African American (AA) patients with stroke or TIA compared to whites (Stroke 1999; 30: 1350). Whether the localization of the cerebral ischemic event or the stroke subtype influences the ordering of carotid imaging tests is unclear. Objective: To determine the patterns of CUS utilization in a large cohort of AA stroke patients. Methods: The African American Antiplatelet Stroke Prevention Study database was used for this analysis. Stroke localization was evaluated and stroke subtype was also recorded. CUS utilization was analyzed for the various stroke subgroups and statistical significance was determined by chi square testing. To ensure that site variation in CUS use did not explain any observed differences, the Mantel-Haenszel test was applied using the sites as strata. Results: As of 7/99, 1087 patients were entered in the database. In 67 patients, the stroke localization was uncertain, leaving 1020 patients for analysis. Using the stroke localization and subtype information available, 479 patients were classified as carotid (CAR) territory, 278 as vertebrobasilar (VB), 139 as indeterminate (IND), and 124 as both territories (BOTH). CAR strokes were further subdivided into lacunar and nonlacunar events. There was a statistically significant difference in the CUS utilization among the five groups. CUS exams were obtained in 56% of VB patients, compared to 75 to 81% in the other groups (p<0.001). For the other groups, CUS was performed in CAR nonlacunar (75.6%), CAR lacunar (76.5%), IND (80.2%), and BOTH (80.8%). There was no difference in CUS use in patients with a lacunar vs. nonlacunar presentation (p=0.816). An adjustment stratified by center among sites with at least 20 patients produced the same statistical significance (p<0.001). Conclusions: There appears to be an underutilization of CUS exams in AA patients with carotid territory strokes, especially for patients with a nonlacunar mechanism. This may lead to decreased performance of potentially valuable stroke prevention measures such as carotid endarterectomy in symptomatic AA patients.