Abstract 142: Use of Stroke Prevention Medications in Self-Reported Stroke/TIA: The REGARDS Study

Introduction: Guidelines recommend specific classes of medications for secondary stroke prevention. In this analysis we are interested in whether the use of antiplatelet, anticoagulant, anti-hypertensive, and lipid lowering medications differs by race and region among participants who self-reported stroke or TIA in a national, general population sample. Methods: Data were derived from the REasons for Geographic And Racial Differences in Stroke (REGARDS) Study, a national cohort study of 30,239 black and white participants aged 45+ enrolled between 2003 and 2007 from the 48 contiguous states, with over-sampling in the southeast region known as the stroke belt and buckle.A centralized phone interview was used for medical history and in-home evaluation for medication inventory. All medications used in the previous two weeks were recorded and subsequently coded into drug classes. The dataset was restricted to those with self-reported stroke (1930) or TIA (1114) at baseline. Chi-square tests were used to determine racial and regional (non-belt, belt, buckle) differences in the use of each medication classification, though indication or contraindication for treatment is unknown. Separate univariate logistic regression models were fit to assess the association between each medication class and race and region. Incremental models (demographics, SES factors, health behaviors, and co-morbidities) were then fitted to determine the impact of adding groups of covariates on the associations. Odds ratios and 95% confidence intervals (CI) for the association between each of race and region, and each medication class, were computed. Results: The results from the univariate and fully-adjusted models are presented in Table1. After multivariable adjustment, blacks were less likely than whites to report using antiplatelet and lipid-lowering medications, while the belt and buckle residents were more likely than non-belt residents to report antiplatelets. Table 1: Association between Medications and Race and Region, univariate and adjustment. (95% CI) Adjusted model includes demographics, SES, health behavior, and co-morbidities. Whites and non-belt region serve as reference for logistic regression models of race and region, respectively. Conclusions: These data indicate differential use of antiplatelet therapy by race and region, and lipid lowering therapy by race, in those self-reporting stroke/TIA, with differences in use of anticoagulant and antihypertensive medications accounted for by covariant adjustment. This suggests targeted interventions are needed following stroke/TIA.