Page 141

Karger_ESC London_2013

London, United Kingdom 2013 Cerebrovasc Dis 2013; 35 (suppl 3)1-854 141 5 Epidemiology of stroke 15:10 - 15:20 Changes in the Impact of Stroke Risk Factors with Age, and Implications for Stroke Epidemiology in the Elderly: the REasons for Geographic and Racial Differences in Stroke (REGARDS) Study G. Howard1, M. Banach2, B.M. Kissela3, D.C. Goff4, E.Z. Soliman5, S.E. Judd6, M.M. Saf-ford7, V.J. Howard8 University of Alabama at Birmigham, Birmingham, USA1,University of Lodz, Lodz, PO-LAND2, University of Cincinnati, Cincinnati, USA3, University of Colorado, Denver, USA4, Wake Forest University School of Medicine, Winston-Salem, USA5, University of Alabama at Birmingham, Birmingham, USA6, University of Alabama at Birmingham, Birmingham, USA7, University of Alabama at Birmingham, Birmingham, USA8 Introduction: The average age of stroke is increasing with the “graying” of the population, in-creasing the importance of assessing risk factors in the elderly. Potential changes in the impact of risk factors could be a “true” effect, or could be due to age-related changes in the correlation-al structure of risk factors and associated changes in residual confounding. Methods: REGARDS followed 27,748 stroke-free participants age 45 and over (mean = 64, interdecile range 53 to 77) for an average of 5.3 years, during which 715 incident strokes oc-curred. The association of the “Framingham” risk factors with incident stroke risk was as-sessed in age strata of 45-64 (Young), 65-74 (Middle), and 75+ (Old). The association between risk factors was assessed using Kappa statistics. Results: With the exception of atrial fibrillation, there was a monotonic decrease in the mag-nitude of the impact across the age strata; with hypertension, diabetes, smoking and LVH even becoming non-significant in the elderly (see first figure). The association between measured risk factors also weakened at older ages (see second figure, the example of the association of hypertension and diabetes). Discussion: The impact of risk factors on stroke risk substantially declined at older ages. However, forces such as the metabolic syndrome give rise to positive correlations between risk factors in the young, but these correlations may be reduced in the elderly by selective mortality of those with multiple risk factors. Residual confounding may be playing a role in the decreas-ing impact of risk factors in the elderly if: 1) if there are similar patterns of change in the cor-relation with unknown risk factors (i.e., latent risk factors exist), or 2) if there is measurement error in the known risk factors. These results suggest that the impact of stroke risk factors could be overestimated in the young, and underestimated in the elderly.


Karger_ESC London_2013
To see the actual publication please follow the link above