Health disparities between communities with greater and lesser advantages are a global concern. In the USA, self-identified race as African American (AA) is consistently associated with mild cognitive impairment (MCI) and dementia, compared to Americans of European descent. In a prospective population-based study, we sought to confirm this association and investigate potential explanatory factors.

The Monongahela-Youghiogheny Healthy Aging Team (MYHAT) and Seniors Project 15104 (SP15104) studies recruited adults aged 65+ years from a group of small towns of lower socioeconomic status in the US. MYHAT recruited by age-stratified random sampling from the voter registration list for all towns; SP15104 recruited by intensive community engagement from three towns with populations that are 60% AA. Based on the Clinical Dementia Rating (CDR), MCI was defined as CDR=0.5 and dementia as CDR > 1. Using Cox proportional hazard models, we modeled time to incident CDR > 0.5 from baseline as a function of race (AA vs. all other), other demographics, and several other covariates at baseline.

The sample of 2120 individuals was 8% AA, and 62% female, with median age of 73y, and median educational level of partial college. During follow up of up to 14.5 years, 499 participants developed new-onset MCI/dementia (CDR >0.5). Cox models revealed that being AA was significantly associated with incident CDR > 0.5 (HR=1.45. 95% CI:1.01,2.10). Inclusion of age, sex, and education in the model increased the HR for race to 1.63 (1.1, 2.3). Adding number of regularly taken prescription drugs (reflecting overall morbidity), depression symptoms, preceding year alcohol consumption, and number of visits to emergency or urgent care together reduced the HR to 1.4 (0.96, 2.0), no longer statistically significant

In this population-based cohort study, self-identified African Americans had an about 40% elevated risk of developing MCI/dementia. Adjusting for demographics, the significant association between race and incident MCI/dementia was attenuated by variables reflecting depression, greater general morbidity, and lesser access to regular health services. These variables possibly reflect downstream effects of historic discrimination, but couldstill be modifiable risk factors for MCI/dementia. Addressing them could potentially mitigate ethno-racial disparities in cognitive impairment.