- Wright, Clinton B;
- DeRosa, Janet T;
- Moon, Michelle P;
- Strobino, Kevin;
- DeCarli, Charles;
- Cheung, Ying Kuen;
- Assuras, Stephanie;
- Levin, Bonnie;
- Stern, Yaakov;
- Sun, Xiaoyan;
- Rundek, Tatjana;
- Elkind, Mitchell SV;
- Sacco, Ralph L
Background
Variability in dementia rates across racial and ethnic groups has been estimated at 60%. Studies suggest disparities in Caribbean Hispanic and Black populations, but community-based data are limited.Objective
Estimate the prevalence of mild cognitive impairment (MCI) and dementia in the racially and ethnically diverse community-based Northern Manhattan Study cohort and examine sociodemographic, vascular risk factor, and brain imaging correlates.Methods
Cases of MCI and dementia were adjudicated by a team of neuropsychologists and neurologists and prevalence was estimated across race/ethnic groups. Ordinal proportional odds models were used to estimate race/ethnic differences in the prevalence of MCI or dementia adjusting for sociodemographic variables (model 1), model 1 plus potentially modifiable vascular risk factors (model 2), and model 1 plus structural imaging markers of brain integrity (model 3).Results
There were 989 participants with cognitive outcome determinations (mean age 69±9 years; 68% Hispanic, 16% Black, 14% White; 62% women; mean (±SD) follow-up five (±0.6) years). Hispanic and Black participants had greater likelihood of MCI (20%) and dementia (5%) than White participants accounting for age and education differences. Hispanic participants had greater odds of MCI or dementia than both White and Black participants adjusting for sociodemographic variables, vascular risk factors, and brain imaging factors. White matter hyperintensity burden was significantly associated with greater odds of MCI or dementia (OR = 1.3, 1.1 to 1.6), but there was no significant interaction by race/ethnicity.Conclusion
In this diverse community-based cohort, cross-sectional data revealed significant race/ethnic disparities in the prevalence of MCI and dementia. Longer follow-up and incidence data are needed to further clarify these relationships.