Brain morphology in older African Americans, Caribbean Hispanics, and whites from northern Manhattan

Abstract

Background: Aging is accompanied by a decrease in brain volume and by an increase in cerebrovascular disease.

Objective: To examine the effects of age, sex, race/ethnicity, and vascular disease history on measures of brain morphology, including relative brain volume, ventricular volume, hippocampus and entorhinal cortex volumes, and white matter hyperintensity (WMH) burden, in a large community-based cohort of racially/ethnically diverse older adults without dementia.

Design: The associations of age, sex, race/ethnicity, and self-reported vascular disease history with brain morphology were examined in a cross-sectional study using multiple linear regression analyses. Sex x race/ethnicity interactions were also considered.

Setting: The Washington Heights-Inwood Columbia Aging Project, a community-based epidemiological study of older adults from 3 racial/ethnic groups (white, Hispanic, and African American) from northern Manhattan.

Participants: Beginning in 2003, high-resolution quantitative magnetic resonance (MR) images were acquired in 769 participants without dementia.

Main outcome measures: Relative brain volume (total brain volume/intracranial volume), ventricular volume, and hippocampus and entorhinal cortex volumes were derived manually on high-resolution MR images. White matter hyperintensities were quantified semiautomatically on fluid-attenuated inversion recovery-T2-weighted MR images.

Results: Older age was associated with decreased relative brain volume and with increased ventricular and WMH volumes. Hispanic and African American participants had larger relative brain volumes and more severe WMH burden than white participants, but the associations of these variables with age were similar across racial/ethnic groups. Compared with men, women had larger relative brain volumes. Vascular disease was associated with smaller relative brain volume and with higher WMH burden, particularly among African Americans.

Conclusions: Older age and vascular disease, particularly among African Americans, are associated with increased brain atrophy and WMH burden. African American and Hispanic subjects have larger relative brain volumes and more WMH than white subjects. Racial/ethnic group differences in WMH severity seem to be partially attributable to differences in vascular disease. Future work will focus on the determinants and cognitive correlates of these differences.

Figure 1

Schematic representation of derived MRI sample.

Figure 2

The relationship among chronological age, ethnicity, and relative brain volume. While the three ethnicity groups significantly differed in relative brain volume, the relationship between age and relative brain volume was similar across groups. Note that data have been grouped into 20 bins on the y-axis and 30 bins on the x-axis; the number of participants represented by each data point is indicated by its size.

Figure 3

Relative brain volume across ethnic groups and sex. Women had significantly larger relative brain volumes than men and African American and Hispanics had significantly larger relative brain volumes than Caucasians. However, the sex by ethnicity interaction was not significant. Error bars are standard errors. Data are adjusted for vascular disease history and age.

Figure 4

The relationship among chronological age, ethnicity, and log transformed relative WMH volume. The three groups significantly differed in severity of WMH, but the relationship between chronological age and relative WMH volume was similar across groups. Note that data have been grouped into 25 bins on the y-axis and 30 bins on the x-axis; the number of participants represented by each data point is indicated by its size.

Figure 5

Interaction between ethnicity and vascular disease history on WMH burden. For illustration purposes, vascular disease history was dichotomized into no vascular disease history versus any vascular disease history. African Americans with vascular disease had a disproportionately greater degree of WMH than their Caucasian and Hispanic counterparts. Error bars represent standard errors. Data presented are adjusted for age and sex.