Exploring Reasons for Differential Vulnerability and Alzheimer’s Disease Risk in Racial and Ethnic Minorities

Abstract

Background:

African American and Hispanic older adults are reported to have up to a 2-fold higher risk of Alzheimer’s disease and related disorders (ADRD), but the reasons for this increased vulnerability have not been fully explored. The Vulnerability Index (VI) was designed to identify individuals who are at risk of developing cognitive impairment in the future, capturing 12 sociodemographic variables and modifiable medical comorbidities associated with higher ADRD risk. However, a prior limitation of the VI was that the original study cohort had limited diversity. We examined the association of the VI within and between non-Hispanic White, African American, and Hispanic older adults with and without cognitive impairment and different socioeconomic strata enrolled in a community-based dementia screening study.

Objective:

To explore reasons for reported higher ADRD vulnerability in African Americans and Hispanics.

Methods:

In a cross-sectional study of 300 non-Hispanic White, African American, and Hispanic older adults with and without cognitive impairment, we studied the association between cognitive status, the VI, and socioeconomic status (SES).

Results:

When considering race/ethnicity, the presence of more vascular comorbidities drove greater vulnerability. When considering SES, vascular comorbidities played a less prominent role suggesting resources and access to care drives risk. The VI had differential effects on cognitive performance with the greatest effect in the earlier stages of impairment.

Conclusion:

Findings from this study provide a deeper understanding of the differential risk of ADRD in multicultural older adults captured by the VI and how barriers to healthcare access may increase vulnerability in racial/ethnic minorities.

Keywords

Alzheimer’s disease cognitive impairment dementia health disparities mild cognitive impairment race/ethnicity risk assessment risk factors socioeconomic status vulnerability

INTRODUCTION

Alzheimer’s disease and related dementias (ADRD) and its prodrome mild cognitive impairment (MCI) affects over 6 million Americans and more than 35 million people worldwide across all racial, ethnic, and cultural backgrounds [1]. More than 48 million adults aged 65 years and older are living in the US [2]. The Centers for Disease Control and Prevention report that African Americans have the highest prevalence of ADRD (13.8%), followed by Hispanics (12.2%) and non-Hispanic whites (10.3%) [3]. The Alzheimer’s Association reports that African Americans have a two-fold increased risk of ADRD and Hispanics a 1.5-fold increased risk of ADRD compared with non-Hispanic Whites [1]. As ethnically and racially diverse older adult populations continue to grow, ADRD poses a major and increasing public health challenge, and older adults from communities of color may be disproportionately burdened by the disease. Among Americans aged 65 and higher, Alzheimer’s disease is the fifth leading cause of death and may be particularly lethal among multicultural populations [1]. While the reasons are not fully understood, there may be unique genetic, environmental, structural, and vascular risk factors (e.g., hypertension, diabetes, obesity) in African Americans and Hispanics that place them at higher risk. Alternatively, a lack of resources and access to healthcare throughout the life course may be at root cause for these disparities in later life.

To have a better understanding of brain health and ADRD risks, we recently described 12 modifiable and non-modifiable factors associated with the risk of cognitive impairment including age, sex, race and ethnicity, education, frailty, obesity, and comorbid medical conditions to create the Vulnerability Index (VI) [4]. The VI was designed to identify individuals who are at risk of developing cognitive impairment in the future and provide actionable information for the development of prevention and treatment plans [4]. However, a prior limitation of the VI was that the study cohort had limited diversity so that its full application in multicultural cohorts of older adults could not be tested.

Although the primary risk factor for ADRD is age, race and ethnicity are also significant demographic risk factors, and estimates of ADRD among racial and ethnic minorities is understudied [5]. Racial and ethnic disparities in ADRD risk have long been recognized but most research efforts have focused on comparisons between White and African American older adults or between Non-Hispanic White and Hispanic older adults [6]. Older adults living in vulnerable communities may face additional challenges due to lack of access to care. Social determinants of health represent conditions in which individuals live, work, and age, and how these conditions effect health and health care [7]. These are distinct from medical determinants of health which represent comorbid conditions that may differentially place one group at a higher risk of disease compared with another group. In individuals from underserved communities, social determinants of health occurring in early and midlife could have a later influence on medical determinants of health supporting the need to consider both to fully understand health disparities associated with ADRD.

To address this gap in understanding, we explored the VI and each of its components within and between non-Hispanic White, African American, and Hispanic older adults with and without cognitive impairment and across different socioeconomic strata. We hypothesized that African American and Hispanic older adults who experienced a higher score on the VI will have worse scores on subjective and objective cognitive performance compared with non-Hispanic white older adults. We posited that the distribution and contribution of risk factors contained within the VI would be different across the different racial and ethnic groups and that some of the higher risk could be explained by differences in socioeconomic status.

METHODS

Participants

A total of 300 multicultural older adults completed baseline data collection between 2012 and 2015 as part of a community-based dementia screening program that examined the effects of social and medical determinants on cognition. Data collection occurred at multiple community settings such as churches, community centers, and public housing in Manhattan, Brooklyn, and Queens, New York. Inclusion criteria included 40 + years of age, and fluency in English or Spanish. Exclusion criteria were limited to closely as possible represent a “real world” community population and included: age < 40 years, non-fluency in English or Spanish, and psychiatric (i.e., schizophrenia) and neurological (i.e., non-controlled seizure disorder) conditions that could interfere with participation. Participants were divided into three groups based on race and ethnicity: 1) non-Hispanic Whites, 2) non-Hispanic African Americans, and 3) Hispanics. Studies were conducted either in English or Spanish by bilingual research staff and clinicians. All participants provided written informed consent. The study was approved by the Institutional Review Board at New York University Grossman School of Medicine.

Demographic characteristics

Participants provided demographic characteristics including age, years of education, sex, race (White, African American), and ethnicity (non-Hispanic, Hispanic), primary language (monolingual English, monolingual Spanish, bilingual English/Spanish), and preferred language for testing (English, Spanish). Participants completed a checklist of comorbid medical conditions (e.g., diabetes, stroke, hypertension) and behaviors (e.g., tobacco, alcohol, substance use) as present or absent, and if they were being treated with medications.

Socioeconomic status

Socioeconomic status (SES) was determined by two methods. First, we used a modified version of the Hollingshead two-factor index of social position [8] which combines highest education attainment across 7 categories and highest occupational attainment across 7 categories as previously described [6 , 8–10]. Information was collected about participant and head of household, if different from participant. Participant education multiplied by 4 and head of household occupation multiplied by 7 was combined to create a sum with a range of 11-77 (lower numbers are better) provided 5 SES strata. Strata I was designated the upper SES group, Strata II-III was designated the middle SES group, and Strata IV-V was designated the lower SES group. Since there were few Hispanics in the upper SES strata, we also used a median split of Hollingshead index sum scores to create two strata (upper-middle and middle-lower) for multivariate analyses. As the Hollingshead index score is assigned to each participant based on educational and occupational attainment rather than querying their actual resources, each participant was also asked a question to rate their global financial standing: “When you consider your household income from all sources, would you say you (1) are comfortable, (2) have just enough to make ends meet, or (3) do not have enough to make ends meet”.

Vulnerability index [4]

The VI is calculated as the weighted sum of twelve easily obtained sociodemographic, medical, and functional factors including age, sex, race/ethnicity, education, frailty, obesity, depression, and comorbid vascular medical conditions (diabetes, stroke, heart disease, hypertension, and hypercholesterolemia) associated with an increased risk of developing cognitive impairment (Box 1) [4]. The possible range of scores is 2-20 with higher scores corresponding to greater vulnerability. Scores higher than 8 are associated with a 17-fold increased risk of cognitive impairment. Frailty was determined by scores on the mini-Physical Performance test (mPPT) [11], an objective measurement of physical functionality, with scores < 10 rated as frail. Obesity was determined by body mass index (BMI) and dichotomized as obese (BMI > 30) or non-obese (BMI≤30). Depressive symptoms were assessed using the Hospital Anxiety and Depression Scale (HADS) [12] with the depression subscale (HADS-D) dichotomized as depressed (HADS-D>7) or non-depressed (HADS-D≤7) [12]. For the present study, the VI was calculated for all three racial/ethnic groups and the two measures of socioeconomic status.

Box 1.

Vulnerability index

Variable	Categories	Score
Age	<60	0
61-74	1
75+	3
Sex	Male	1
Female	2
Race/Ethnicity	Non-Hispanic White	1
Any other	2
Education	>16 y	0
12-16	1
<12	2
Diabetes	No	0
Yes	2
Stroke	No	0
Yes	2
Frailty	No	0
Yes	2
Heart Disease	No	0
Yes	1
Obesity	BMI≤30	0
BMI>30	1
Hypertension	No	0
Yes	1
Hypercholesterolemia	No	0
Yes	1
Depression	No	0
Yes	1
TOTAL (Range 2-20)

Cognitive performance

Montreal Cognitive Assessment (MoCA)

MoCA was used as the primary measure of objective cognitive performance. The MoCA is a 13-item assessment tool designed to screen for MCI and to examine global cognitive function [13, 14] assessing the following cognitive domains: visuospatial, executive function, naming, memory, attention, language, abstraction, and orientation. The MoCA has a maximum score of 30, with lower scores representing worse objective cognitive performance. As per MoCA instructions, an additional point was added to the total score for individuals with less than 12 years of education [13].

AD8

The patient version of the AD8 was used as the primary measure of subjective cognitive performance. The AD8 is an 8-item self-report survey used to assess memory, orientation, judgement, and everyday function [15, 16]. The AD8 yields a score from 0 to 8 depending on the number of positive responses. A score of≥2 indicates subjective cognitive impairment [15, 16].

Additional cognitive measures

In addition to MoCA and AD8, secondary measures of cognitive performance were captured including (a) the number of animals named in one minute, (b) repeating the months in reverse order, and (c) the Mini-Cog which contains a 3-word recall and clock drawing [17]. These measures were used to assist in research consensus.

Research consensus

Each participant was reviewed in consensus conference considering their age, education, scores on the MoCA and AD8 and performance on secondary cognitive measures. Individuals whose AD8 score was < 2 and MoCA score was > 24 were considered cognitively normal. Individuals whose AD8 score was ≥2 and MoCA score was ≤24 were considered cognitively impaired. Individuals who reported only subjective complaints on the AD8 but scored > 24 on the MoCA were considered cognitively intact but with subjective memory complaints. Individuals who scored ≤24 on the MoCA were considered cognitively impaired regardless of their AD8 score.

Data analysis

Analyses were conducted using IBM SPSS version 28 (Armonk, NY). General linear modeling with post-hoc least significant differences was used with Bonferroni correction to correct for multiple comparisons. We first examined differences in the VI and each of its weighted components separately by race/ethnicity and by cognitive status. We then used multivariate analyses to consider race/ethnicity and cognitive status together. We examined differences in the VI and each of its weighted components by SES using the Hollingshead strata and the participants’ self-report, and then conducted a multivariate analysis of the VI and each of its weighted components by race/ethnicity, cognitive status, and SES strata.

We then performed linear regression analyses to examine the relationship between the VI and cognitive performance on the MoCA and AD8. Of greatest interest was the degree to which a single point change in the VI is associated with changes in either an objective measure of impairment (the MoCA) or subjective reports of impairment (the AD8). A plateau effect was anticipated after a certain level of impairment, as there may be a lower limit where vulnerability would not yet influence impairment, or a higher limit where greater vulnerability no longer reflects additional impairment risk. This plateau effect was examined using Michaelis-Menten regression, a technique traditionally used to estimate enzymatic reactions using non-linear regression to determine both rate of change up to a plateau point as well as the value of that plateau [18]. To determine the single-point change up to the plateau, a linear regression was run on values preceding the plateau. A regression was also run on values after the plateau point, which would be expected to show no significant change.

RESULTS

Sample characteristics

A total of 300 community-dwelling older adults were recruited with 45.7% non-Hispanic White, 20.3% non-Hispanic Black, and 34.0% Hispanic. English was the primary language in 45.7% of the sample while 25.4% spoke only Spanish. Participants had a mean age of 68.8±9.2 years (range 40-94) and a mean education of 14.3±4.1 years (range 0-22). The sample was 64.3% female, and 43.7% were cognitively impaired at consensus. The sample reported 5.8±2.9 comorbid medical conditions (range 0-14). The mean Hollingshead Index of Social Status was 36.8±18.4 (range 11–77) supporting a wide range of SES. The patients had a mean MoCA score of 23.0±5.2 (range: 1-30) and a mean AD8 score of 1.7±1.9 (range 0-8) supporting a wide range of subjective and objective cognitive performance. The mean HADS-D score was 5.1±3.8 (range: 0–21) supporting a wide range of mood states. The mean Vulnerability Index (VI) score was 7.4±2.9 (range: 3-17).

Demographics, cognitive and physical performance, VI, and components by racial/ethnic grouping are shown in Table 1. Correcting for multiple comparisons, African American and Hispanics had lower education, more females, a lower SES, lower MoCA scores, and a higher prevalence of cognitive impairment compared to non-Hispanic Whites. Vulnerability index scores for African American and Hispanic participants were on average higher compared with non-Hispanic White participants as well as higher rates of important comorbid conditions: diabetes, obesity, and hypertension. African Americans had the highest rate of frailty.

Table 1

Participant characteristics by racial-ethnic group

Variable	Non-Hispanic White (n = 137)	African American (n = 61)	Hispanic (n = 102)	p
Age, y	68.7 (8.7)	69.3 (11.3)	68.8 (8.6)	0.909
Education, y	16.5 (2.6)	13.7 (3.3)^a	11.9 (4.6)^a,b	<0.001
Sex, % Female	48.2	75.4^a	79.4^a	<0.001
Hollingshead Index	26.2 (12.5)	40.8 (15.9)^a	48.8 (18.3)^a,b	<0.001
Cognitive Status, % impaired	21.9	57.4^a	64.7^a	<0.001
AD8	1.5 (1.7)	1.7 (1.9)	2.1 (2.1)^a	0.051
MoCA	24.9 (4.5)	21.1 (6.0)^a	21.5 (4.8)^a	<0.001
Mini-Cog	2.9 (1.2)	2.5 (1.3)	2.6 (1.2)	0.117
Animal Naming	18.9 (7.1)	16.1 (6.7)^a	16.4 (5.5)^a	0.017
Month Backwards, #errors	0.2 (0.6)	0.5 (0.8)	0.6 (0.8)^a	0.011
HADS-D	4.7 (3.7)	4.9 (3.5)	5.9 (3.9)^a	0.050
Mini-PPT	12.6 (2.3)	10.5 (3.7)^a,c	12.4 (2.2)	<0.001
Vulnerability Index	5.9 (2.3)	9.1 (3.2)^a	8.5 (2.4)^a	<0.001
Diabetes, %	5.1	26.2^a	18.6^a	<0.001
Stroke, %	10.9	13.1	5.9	0.251
Frailty, %	16.1	45.9^a,c	17.6	<0.001
Heart Disease, %	17.5	19.7	9.8	0.150
Obesity, %	19.7	38.3^a	36.3^a	0.004
Hypertension, %	35.0	59.0^a	55.9^a	<0.001
Hypercholesterolemia, %	35.0	36.1	46.1	0.195
Depression, %	24.1	23.0	29.7	0.529

Mean (SD) or %. MoCA, Montreal Cognitive Assessment; HADS-D, Hospital Anxiety and Depression Scale-Depression Subscale; PPT, Physical Performance Test. Bold p-value signifies significance for multiple comparisons using Bonferroni correction. Post-hoc comparisons: ^aDifference from Non-Hispanic White participants. ^bDifference from African American participants. ^cDifference from Hispanic participants

Vulnerability Index scores and weighted components by race and ethnicity

To better understand potential causes of increased vulnerability to ADRD in African Americans and Hispanics, we examined the VI and each of its weighted components by racial and ethnic groups in Table 2. Although the VI includes race and ethnicity as a risk variable, examining the contribution of each weighted component to the total VI can provide information as to which social and medical determinants of health may contribute to differential increased risk of cognitive impairment. There was no difference by age but there were more women in the African American and Hispanic groups. Education played an important contribution to vulnerability with non-Hispanic Whites having the highest educational attainment and Hispanics the lowest. Diabetes, obesity, and hypertension provided greater weight towards vulnerability in African Americans and Hispanics, while frailty was significantly different only in African Americans.

Table 2

Univariate analyses of vulnerability index and weighted components by race/ethnicity, cognitive status or socioeconomic status

	Race/ethnicity				Cognitive status
	White (n = 137)	Black (n = 61)	Hispanic (n = 102)	p	HC (n = 169)	CI (n = 131)	p
Age, Weighted component	1.4 (0.9)	1.5 (1.1)	1.2 (0.9)	0.409	1.2 (0.9)	1.5 (1.1)	0.016
Education, Weighted component	0.3 (0.5)	0.9 (0.7)^a	1.1 (0.8)^a,b	<0.001	0.5 (0.6)	0.9 (0.8)	<0.001
Sex, Weighted component	1.5 (0.5)	1.7 (0.4)^a	1.8 (0.4)^a	<0.001	1.6 (0.5)	0.6 (0.5)	0.947
Race/Ethnicity, Weighted component	1.0 (0.0)	2.0 (0.0)	2.0 (0.0)	n/a	1.4 (0.5)	1.8 (0.4)	<0.001
Diabetes, Weighted component	0.05 (0.2)	0.3 (0.4)	0.2 (0.4)	<0.001	0.1 (0.3)	0.2 (0.4)	0.058
Stroke, Weighted component	0.1 (0.3)	0.1 (0.3)	0.05 (0.2)	0.253	0.1 (0.3)	0.1 (0.3)	0.795
Frailty, Weighted component	0.3 (0.7)	0.9 (1.0)^a	0.3 (0.8)^b	<0.001	0.3 (0.7)	0.7 (0.9)	<0.001
Heart disease, Weighted component	0.2 (0.4)	0.2 (0.4)	0.1 (0.3)	0.151	0.2 (0.4)	0.1 (0.3)	0.022
Obesity, Weighted component	0.2 (0.4)	0.4 (0.5)	0.4 (0.5)	0.005	0.3 (0.4)	0.3 (0.5)	0.200
Hypertension, Weighted component	0.3 (0.5)	0.6 (0.5)	0.5 (0.5)	<0.001	0.4 (0.5)	0.6 (0.5)	0.004
Hypercholesterolemia, Weighted component	0.3 (0.5)	0.4 (0.5)	0.5 (0.5)	0.196	0.4 (0.5)	0.4 (0.5)	0.796
Depression, Weighted component	0.2 (0.4)	0.2 (0.4)	0.3 (0.4)	0.561	0.2 (0.4)	0.3 (0.5)	0.090
Total Vulnerability Index	5.9 (2.3)	9.1 (3.2)^a	8.5 (2.4)^a	<0.001	6.6 (2.7)	8.5 (2.9)	<0.001

Mean (SD); HC, healthy controls; CI, cognitively Impaired; n/a, not applicable. Bold p-value signifies significance for multiple comparisons using Bonferroni correction. Post-hoc comparisons: ^aDifference from non-Hispanic Whites. ^bDifference from African Americans.

	Hollingshead index				Self-reported household means
	Upper (n = 31)	Middle (n = 169)	Lower (n = 100)	p	Comfortable (n = 167)	Make ends meet (n = 98)	Cannot meet needs (n = 35)	p
Age, Weighted component	1.3 (1.0)	1.3 (0.9)	1.3 (1.0)	0.956	1.2 (0.9)	1.3 (0.9)	0.9 (0.9)^c	0.031
Education, Weighted component	0.03 (0.2)	0.4 (0.5)^a	1.4 (0.6)^a,b	<0.001	0.5 (0.7)	0.9 (0.7)^c	1.2 (0.8)^c,d	<0.001
Sex, Weighted component	1.3 (0.5)	1.7 (0.5)^a	1.7 (0.4)^a	<0.001	1.6 (0.5)	1.7 (0.5)	1.7 (0.4)	0.227
Race/Ethnicity, Weighted component	1.1 (0.3)	1.4 (0.5)^a	1.9 (0.3)^a,b	<0.001	1.4 (0.5)	1.7 (0.4)^c	1.8 (0.4)^c	<0.001
Diabetes, Weighted component	0.1 (0.2)	0.1 (0.3)	0.2 (0.4)	0.078	0.1 (0.3)	0.1 (0.3)	0.2 (0.4)	0.274
Stroke, Weighted component	0.1 (0.3)	0.1 (0.3)	0.1 (0.2)	0.296	0.1 (0.3)	0.1 (0.3)	0.03 (0.2)	0.338
Frailty, Weighted component	0.3 (0.7)	0.3 (0.8)	0.7 (0.9)^a,b	0.010	0.4 (0.8)	0.5 (0.9)	0.6 (0.9)	0.084
Heart disease, Weighted component	0.2 (0.4)	0.2 (0.4)	0.1 (0.3)^b	0.045	0.2 (0.4)	0.1 (0.3)	0.1 (0.3)	0.297
Obesity, Weighted component	0.2 (0.4)	0.2 (0.4)	0.4 (0.5)^b	0.012	0.3 (0.4)	0.3 (0.4)	0.4 (0.5)^c,d	0.067
Hypertension, Weighted component	0.3 (0.5)	0.4 (0.5)	0.6 (0.5)^a,b	0.024	0.4 (0.5)	0.5 (0.5)	0.5 (0.5)	0.138
Hypercholesterolemia, Weighted component	0.3 (0.5)	0.4 (0.5)	0.4 (0.5)	0.732	0.4 (0.5)	0.3 (0.5)	0.3 (0.5)	0.507
Depression, Weighted component	0.1 (0.3)	0.3 (0.4)	0.3 (0.4)	0.223	0.2 (0.4)	0.3 (0.5)	0.4 (0.5)^c	0.046
Total Vulnerability Index	5.5 (2.5)	6.9 (2.6)^a	9.0 (2.9)^a,b	<0.001	6.9 (2.8)	7.9 (2.9)^c	8.4 (2.8)^c	0.002

Mean (SD). Bold p-value signifies significance for multiple comparisons using Bonferroni correction (corrected p-value<0.0038). Post-hoc comparisons: ^aDifferences from Upper SES. ^bDifferences from Middle SES. ^cDifferences from Comfortable. ^dDifferences from Make ends meet.

Vulnerability Index scores and weighted components by cognitive status

In Table 2, we examined the VI and each of its weighted components by consensus cognitive status. There was no difference by sex, but lower educational attainment was a significant contributor to vulnerability. Increased age, heart disease, frailty, and hypertension were different by cognitive status, with diabetes showing a trend. After correction for multiple comparisons, frailty and hypertension remained significant contributors.

Vulnerability Index scores and weighted components by socioeconomic status

Since socioeconomic status may play an important role in availability and access to medical care, housing quality, and nutrition, we re-examined the VI and its weighted components with two measures of SES: the Hollingshead Index and the participants self-report of available income. Using SES strata by the Hollingshead Index, VI scores increased with decreasing SES (Table 2). Participants with middle and lower SES had lower education, were more likely to be women, and less likely to be non-Hispanic White. Participants in the lowest SES strata had a higher weighted component for frailty, heart disease, obesity, and hypertension and a trend for diabetes, however these were no longer significant after correction for multiple comparisons. Using the participants self-reported available income, VI scores again increased with decreasing SES with trends for frailty, obesity, and depression. Participants with greater challenges to make ends meet had lower education and were more likely to be women, but race/ethnicity was no longer significant.

Vulnerability Index scores and weighted components by combined race, ethnicity, and cognitive status

We performed a multivariate analysis to examine the VI and each of its weighted components for cognitive status (controls versus impaired) within and between the three racial/ethnic groups (Table 3). VI scores were lowest in non-Hispanics Whites regardless of cognitive status, intermediate in Hispanics, and highest in African Americans (p < 0.001). Within group comparisons revealed significant increases in VI scores between controls and impaired individuals in African Americans (p < 0.001) and Hispanics (p < 0.001). Important social determinants contributing to increased vulnerability include lower education (p < 0.001) and female sex (p < 0.001). Important medical determinants included diabetes, obesity, heart disease, and hypertension in African Americans and Hispanics with cognitive impairment. Frailty was an important determinant in African Americans regardless of cognitive status.

Table 3

Multivariate analyses of vulnerability index and weighted components by combined race/ethnicity and cognitive status

	White		Black		Hispanic		p
Variable	HC	CI	HC	CI	HC	CI	Omnibus	Race/Ethnicity	CI
Age, Weighted component	1.3 (0.9)	1.5 (1.1)	1.2 (1.1)	1.6 (1.1)	0.9 (0.6)	1.4 (1.0)	0.093	0.260	0.008
Education, Weighted component	0.3 (0.5)	0.5 (0.6)	0.7 (0.6)	0.9 (0.7)	0.9 (0.8)	1.2 (0.8)	<0.001^a	<0.001	<0.001
Sex, Weighted component	1.5 (0.5)	1.3 (0.5)	1.8 (0.4)	1.7 (0.4)	1.9 (0.3)	1.7 (0.4)	<0.001^b	<0.001	0.023
Race/Ethnicity, Weighted component	1.0 (0.0)	1.0 (0.0)	2.0 (0.0)	2.0 (0.0)	2.0 (0.0)	2.0 (0.0)	n/a	n/a	n/a
Diabetes, Weighted component	0.05 (0.2)	0.03 (0.2)	0.3 (0.4)	0.2 (0.4)	0.1 (0.3)	0.2 (0.4)	0.001^b	<0.001	0.772
Stroke, Weighted component	0.1 (0.3)	0.1 (0.3)	0.1 (0.3)	0.1 (0.3)	0.05 (0.2)	0.1 (0.2)	0.714	0.310	0.861
Frailty, Weighted component	0.2 (0.6)	0.7 (0.9)	0.8 (0.9)	1.0 (1.0)	0.1 (0.5)	0.5 (0.9)	<0.001^c	<0.001	<0.001
Heart disease, Weighted component	0.2 (0.4)	0.03 (0.2)	0.3 (0.4)	0.1 (0.3)	0.1 (0.3)	0.1 (0.3)	0.039	0.164	0.042
Obesity, Weighted component	0.2 (0.4)	0.2 (0.4)	0.4 (0.5)	0.4 (0.5)	0.4 (0.5)	0.4 (0.5)	0.059^b	0.016	0.969
Hypertension, Weighted component	0.3 (0.5)	0.4 (0.5)	0.4 (0.5)	0.7 (0.5)	0.6 (0.5)	0.5 (0.5)	<0.001^b	0.015	0.059
Hypercholesterolemia, Weighted component	0.4 (0.5)	0.3 (0.5)	0.4 (0.5)	0.3 (0.5)	0.5 (0.5)	0.4 (0.5)	0.478	0.127	0.384
Depression, Weighted component	0.2 (0.4)	0.3 (0.5)	0.2 (0.4)	0.2 (0.4)	0.2 (0.4)	0.3 (0.5)	0.499	0.776	0.185
Total Vulnerability Index	5.7 (2.2)	6.3 (2.7)	8.6 (3.5)	9.6 (2.9)	7.7 (2.1)	8.9 (2.5)	<0.001^b	<0.001	0.006

Mean (SD). HC, healthy control; CI, cognitive Impairment; n/a, not applicable. Bold p-values signify significance after multiple comparisons using Bonferroni correction (corrected p-value<0.0038). Post-hoc comparisons: ^aAll groups different from each other. ^bBlack and Hispanic participants different from non-Hispanic White participants. ^cBlack participants different from non-Hispanic White and Hispanic participants.

Multivariate modeling of Vulnerability Index and weighted components by race/ethnicity, cognitive status, and socioeconomic status

To better understand independent contributions of SES and race/ethnicity to vulnerability for cognitive impairment, we conducted a multivariate analysis (Table 4). Total VI scores were significantly different by race/ethnicity and by SES with a trend by cognitive status. Non-Hispanic Whites had the lowest VI scores regardless of cognitive status or SES. Examining each of the VI weighted components revealed interesting patterns. There was a minimal effect of age with cognitive impaired individuals being older than healthy controls. Education and sex were significant components of risk by race/ethnicity, with education different by SES and cognitive status different by sex. Diabetes showed a trend by race/ethnicity and cognitive status, with obesity and hypertension also showing trends by race/ethnicity. These trends lost significance with correction for multiple comparisons. Frailty was the single strongest contributor to vulnerability in African Americans.

Table 4

Multivariate analyses of vulnerability index and weighted components by race/ethnicity, cognitive status, and socioeconomic status

Socioeconomic class	Upper-Middle (n = 159)						Middle-Lower (n = 138)						Omnibus
Cognitive status	HC (n = 113)			CI (n = 46)			HC (n = 54)			CI (n = 84)			p
Racial/ethnic group	W (n = 87)	B (n = 13)	H (n = 13)	W (n = 20)	B (n = 14)	H (n = 12)	W (n = 20)	B (n = 12)	H (n = 22)	W (n = 10)	B (n = 20)	H (n = 54)
Age, Weighted component	1.3 (0.9)	1.2 (0.8)	1.0 (0.7)	1.5 (1.2)	1.3 (0.9)	1.6 (1.1)	1.1 (0.9)	1.3 (1.3)	0.9 (0.6)	1.4 (0.8)	1.7 (1.2)	1.3 (1.0)	0.454^e
Education, Weighted component	0.1 (0.3)	0.4 (0.5)	0.1 (0.4)	0.3 (0.6)	0.3 (0.5)	0.2 (0.4)	0.8 (0.6)	1.2 (0.40	1.3 (0.6)	0.9 (0.6)	1.4 (0.5)	1.5 (0.6)	<0.001 ^a,d
Sex, Weighted component	1.5 (0.5)	1.8 (0.4)	1.9 (0.3)	1.3 (0.5)	1.9 (0.3)	1.6 (0.5)	1.5 (0.5)	1.7 (0.5)	1.8 (0.3)	1.2 (0.4)	1.6 (0.5)	1.8 (0.4)	<0.001 ^b,e
Race/Ethnicity, Weighted component	1.0 (0.0)	2.0 (0.0)	2.0 (0.0)	1.0 (0.0)	2.0 (0.0)	2.0 (0.0)	1.0 (0.0)	2.0 (0.0)	2.0 (0.0)	1.0 (0.0)	2.0 (0.0)	2.0 (0.0)	n/a
Diabetes, Weighted component	0.03 (0.2)	0.2 (0.4)	0.1 (0.3)	0.0 (0.0)	0.2 (0.4)	0.2 (0.4)	0.1 (0.4)	0.3 (0.5)	0.2 (0.4)	0.1 (0.3)	0.2 (0.4)	0.2 (0.4)	0.017^b,e
Stroke, Weighted component	0.1 (0.3)	0.0 (0.0)	0.0 (0.0)	0.1 (0.3)	0.1 (0.4)	0.1 (0.3)	0.05 (0.2)	0.2 (0.4)	0.04 (0.2)	0.1 (0.3)	0.1 (0.3)	0.05 (0.2)	0.546
Frailty, Weighted component	0.2 (0.6)	0.5 (0.9)	0.1 (0.5)	0.9 (1.0)	0.8 (1.0)	0.2 (0.6)	0.2 (0.6)	1.2 (1.0)	0.1 (0.4)	0.2 (0.6)	1.1 (1.0)	0.5 (0.9)	<0.001 ^c,e
Heart disease, Weighted component	0.2 (0.4)	0.3 (0.5)	0.1 (0.4)	0.05 (0.2)	0.1 (0.3)	0.3 (0.5)	0.2 (0.4)	0.5 (0.4)	0.04 (0.2)	0.0 (0.0)	0.1 (0.4)	0.05 (0.2)	0.044
Obesity, Weighted component	0.2 (0.4)	0.3 (0.5)	0.4 (0.5)	0.2 (0.4)	0.2 (0.4)	0.2 (0.4)	0.2 (0.4)	0.5 (0.5)	0.4 (0.5)	0.2 (0.4)	0.4 (0.5)	0.4 (0.5)	0.077^b
Hypertension, Weighted component	0.3 (0.5)	0.5 (0.5)	0.6 (0.5)	0.4 (0.5)	0.6 (0.5)	0.6 (0.5)	0.3 (0.5)	0.4 (0.5)	0.5 (0.5)	0.5 (0.5)	0.7 (0.4)	0.5 (0.5)	0.036^b
Hypercholesterolemia, Weighted component	0.3 (0.5)	0.5 (0.5)	0.5 (0.5)	0.2 (0.4)	0.3 (0.5)	0.4 (0.5)	0.5 (0.5)	0.3 (0.5)	0.4 (0.5)	0.4 (0.5)	0.3 (0.5)	0.5 (0.5)	0.586
Depression, Weighted component	0.2 (0.4)	0.2 (0.4)	0.4 (0.5)	0.4 (0.5)	0.1 (0.4)	0.2 (0.4)	0.3 (0.5)	0.2 (0.4)	0.1 (0.3)	0.2 (0.4)	0.3 (0.5)	0.3 (0.5)	0.360
Total Vulnerability Index	5.6 (2.2)	7.9 (2.8)	7.3 (1.8)	6.4 (2.9)	8.3 (2.3)	7.7 (2.3)	6.6 (2.1)	9.7 (3.9)	7.9 (2.3)	6.2 (2.2)	10.2 (3.0)	9.2 (2.5)	<0.001 ^b,d

Mean (SD). HC, healthy controls; CI, cognitively impaired; W, non-Hispanic White; B, African American; H, Hispanic. Bold p-value signifies significance for multiple comparisons using Bonferroni correction (corrected p-value<0.0038). Post-hoc comparisons: ^aAll racial/ethnic groups different from each other. ^bAfrican Americans and Hispanics different from non-Hispanic Whites. ^cAfrican Americans different from Hispanics and non-Hispanic Whites. ^dMiddle-Lower SES different from Upper-Middle SES. ^eCognitively impaired different from Healthy Controls.

Effect of Vulnerability Index scores on cognitive performance

Linear regressions using the VI as the independent variable and applied to all participants were significant (Fig. 1) when the dependent variable was both the MoCA (R² = 0.035, p < 0.001) and the AD8 (R² = 0.065, p < 0.001). Across the whole range of values, the MoCA was found to decrease by 0.45±0.14 points with every single point increase in the VI (p < 0.001). Additionally, the AD8 was found to increase by 0.23±0.05 points with every point increase in the VI (p < 0.001). For the MoCA, the Michaelis-Menten regression model found that there was a plateau of 20.35±.91 (p < 0.001) preceded by a decrease of 0.30±0.08 points with every point increase in the VI [18]. The post-plateau linear regression was not significant. No plateau was identified for the AD8.

Fig. 1

Effect of Vulnerability Index on Objective and Subjective Cognitive Performance. Linear regressions were used to examine the relationship between the Vulnerability Index (VI) and objective cognitive performance measured by the Montreal Cognitive Assessment (MoCA –top row) and subjective cognitive performance measured by the AD8 (bottom row). When All Participants were examined, both the MoCA (A) and the AD8 (D) had a significant linear relationship with the VI, with each point in the VI corresponding to an increase in the AD8 of 0.23±0.05 points and a decrease in the MoCA of 0.45±0.14 points. Further, the MoCA exhibited a plateau effect when approaching a value of 20.35±0.91, as determined by Michaelis-Menten regression. Examining the effects of race and ethnicity revealed that the AD8 was significantly associated with the VI in Non-White Participants (F) but not in White Participants (E). When examining the MoCA, neither Whites (B) nor Non-Whites (C) had a significant linear relationship with VI when examined separately.

Examining the effects of race and ethnicity revealed that the AD8 had a more significant association with the VI in African American and Hispanic participants (R² = 0.094, p < 0.001) than in non-Hispanic White participants (R² = 0.013, p = 0.180), with each point of the VI resulting in an AD8 increase of 0.323±0.080. Further examining the effects of race and ethnicity, African Americans showed an increase of 0.355±0.113 (R² = 0.145, p = 0.003) for every point of the VI, while Hispanics showed an increase of 0.284±0.113 (R² = 0.061, p = 0.014). No race/ethnicity effect was seen for the MoCA.

DISCUSSION

This study examined the risk of cognitive impairment using the Vulnerability Index (VI) and its weighted components by race, ethnicity, and socioeconomic status. Although the VI includes race/ethnicity as a risk factor consistent with research supporting that African Americans and Hispanics are at a 1.5-2-fold increased risk of ADRD, we were interested in exploring the reasons for this increased risk [1]. By studying demographics, SES, and comorbid medical conditions, we were able to discern important differences that could explain the higher vulnerability for ADRD in racial and ethnic minorities and provide targets for intervention.

When considering race and ethnicity, the higher burden of vascular comorbidities (diabetes, heart disease, obesity, hypertension) drives greater vulnerability [19 –21]. In African Americans, frailty was the strongest single contributor to risk. When considering SES, comorbid medical conditions played a less prominent role, suggesting the participants’ available resources and access to care drives risk in the most vulnerable groups such as racial and ethnic minorities, women, and individuals with low education [22 –25]. Lifelong challenges in securing adequate housing, nutrition, education, and job opportunities, and access to medical care may place racial and ethnic minorities at higher risk so that by mid- to late-life, the accumulation of vascular comorbidities and frailty specifically in African Americans may increase their overall risk of ADRD. Though inequities may arise on the basis of other factors, it is important to highlight SES in part because disparities based on race and ethnicity remain the most persistent threat to health and are difficult to address from biomedical perspective.

We also found that increasing vulnerability has differential effects on cognitive performance. These findings suggest that accumulation of vascular risk factors has a deleterious effect on cognitive performance, particularly in the earlier stages of impairment. This is consistent with our earlier work on the VI which showed difference in performance in healthy controls and MCI but not in participants who already have cognitive impairment past the mild stage of dementia [4]. This may be due to the fact that the risk factors captured in the VI such as frailty, obesity, depression and age are more associated with risk of developing ADRD than with progression of ADRD. For this reason, the VI should not be used by itself to stage individuals who already manifest clinical ADRD, but instead could best be used to identify individual who are risk for ADRD, or who are more likely to transition from MCI to ADRD.

Self-reported memory complaints as recorded by the AD8 were also affected by increasing VI scores without evidence of a plateau but with a greater effect in African Americans and Hispanic participants compared with non-Hispanic White participants. No plateau effect was observed for the AD8, suggesting that as subjective complaints worsen, medical co-morbidities—and thus VI scores—increase. In a prior mediation analysis [9], we demonstrated that African American and Hispanic participants offered more subjective complaints than non-Hispanic White participants, and this could be explained by lower SES and more depressive symptoms.

This study greatly expands upon findings from the original study of the VI [4] where we reported that the VI could discriminate between cognitively normal control participants and those with any form of cognitive impairment with scores of 8 or greater giving a diagnostic odds ratio of 17.1. Further, participants with high VI scores had worse cognitive, functional, and behavioral outcomes than participants with low VI scores. However, the original sample was not diverse with 93.3% of the sample being non-Hispanic White and was clinic-based. Further the original sample had a larger proportion of participants with Lewy body dementia as this was a specialty niche of the clinical practice. Here, we explored the VI in a community-based sample of racial and ethnically diverse participants with a wide range of social and medical determinants of health that more closely reflect “real-world” situations. We demonstrated the ability of the VI to explore reasons for higher ADRD risk in African American and Hispanic older adults. Knowledge of the potential reasons for increased risk coupled with consideration of social determinants of health may provide insight for the development and testing on novel interventions.

There is growing evidence that ADRD does not affect different racial and ethnic groups equally; however, the underlying medical and social determinants of these disparities are not well understood. We previously demonstrated that multiple factors such as physical functionality, depressive symptoms, and SES help explain some of the racial, ethnic, age, and sex differences in performance on cognitive testing [9]. The poorer cognitive functioning observed among older adults from racial and ethnic minority groups may reflect differences in peak cognitive reserve as measured by educational and occupational attainment [26] that could be impacted by age-related medical comorbidities associated with higher ADRD risk (e.g., diabetes, hypertension, obesity) [27]. Increased cognitive reserve could offer protection from ADRD by resistance against effects of accumulating pathology and strengthening synaptic plasticity [26, 27].

The relationship between SES, race/ethnicity, and dementia risk is complex. Individuals with lower SES have a higher risk of vascular risk factors and stroke [28]. African Americans and Hispanics experience delays in ADRD diagnosis [29], have poorer cognitive performance particularly in vocabulary and semantic memory [30], and experience more functional disability at time of diagnosis compared with non-Hispanic Whites [29]. Our findings highlight the importance of SES as a distinct contributor to ADRD risk, and disparities in health outcomes in racial and ethnic minorities likely have a differential effect on the distribution of vulnerabilities [6, 7]. African Americans and Hispanics may have higher vulnerability due to vascular risk factors as a result of differences in access to care, nutrition, housing, and other social determinants of health. Individuals in the lowest SES strata had the highest vulnerability scores and this group was much more likely to be composed of African Americans and Hispanics.

Consistent with other research, our study suggests that the higher prevalence of ADRD among African American and Hispanic adults is related to higher rates of modifiable risk factors [1, 31]. This suggests that interventions could be developed that reduce the risk of ADRD by addressing these modifiable risk factors and promoting healthy aging. Addressing inequities in these risk factors among marginalized populations and those disproportionately impacted is critical in the efforts to reduce disparities in the burden of ADRD.

Our findings should be interpreted with attention to study limitations. As this is a cross-sectional study, it is not possible to make causal inferences about the relation between VI and self-reported measured aspects of health. All data were gathered through self-report questionnaires, and thus subject to recall bias and to reply in a socially desirable fashion. Although our study population was racially and ethnically diverse, some demographic groups such as Asian Americans and American Indians, and individuals not fluent in either English or Spanish, were not included in these analyses, precluding the generalizability of our results to other populations. The participants included only older adults from New York, therefore findings could differ for individuals from different regions in the US or other countries. Genetic factors and local ancestry could also play a role in differential risk of ADRD and should be part of future investigations [35]. Strengths of the study include the recruitment of a large, diverse community-based cohort rather than being recruited from a tertiary clinic setting. Previous reports examined individuals’ risk factors and cognitive impairment during the initial screening visit [32 –34] as well as 60-day follow-up visits to assess how participants responded to the health information and referrals provided [10].

The dearth of relevant clinical and programmatic knowledge hamstrings efforts by clinicians, scientists, policymakers, and communities to reduce ADRD disparities in multicultural older adults. Inadequate knowledge and poor understanding of ADRD in different racial and ethnic groups may create a barrier to timely dementia screening, and delay ADRD diagnoses and treatment. We provide evidence that documenting root causes of differential risk distribution and vulnerability among multicultural older adults and the potential interplay between social and medical determinants of health offers the possibility to address a major public health concern. Findings from this study may provide a deeper understanding of the differential risk of ADRD in multicultural older adults captured by the VI and how barriers to healthcare access may increase vulnerability in racial and ethnic minorities.

Footnotes

ACKNOWLEDGMENTS

This study was supported by grants to JEG from the National Institute on Aging (R01 AG040211, R01 AG071514, R01 AG069765, and R01 NS101483), the Harry T. Mangurian Foundation, and the Leo and Anne Albert Charitable Trust. The funders had no role in study design, data collection, analysis, decision to publish, or preparation of the manuscript.

JEG and MJK are creators of the Vulnerability Index used in this study and the copyright is held by the University of Miami Miller School of Medicine. JEG is one of the creators of the AD8 used in this study and through Washington University School of Medicine receives licensing fees for its commercial use.

Authors’ disclosures available online ().

References

Alzheimer Association (2022) 2022 Alzheimer’s disease facts and figures. Alzheimers Dement 18, 700–789.

U.S. Census Bureau (2017) “Older Americans Month: May 2017,” Cb17-Ff.08. https://www.census.gov/newsroom/facts-for-features/2017/cb17-ff08.html. Accessed 27 September 2022

Centers for Disease Control and Prevention (2020) “What is the burden of Alzheimer’s disease in the United States?,” cdc.gov/aging/aginginfo/alzheimers.htm#burden. Accessed 27 September 2022.

Kleiman

, Galvin

(2021) The Vulnerability Index: A weighted measure of dementia and cognitive impairment risk. Alzheimers Dement (Amst) 13, e12249.

Matthews

, Xu

, Gaglioti

, Holt

, Croft

, Mack

, McGuire

(2019) Racial and ethnic estimates of Alzheimer’s disease and related dementias in the United States (2015-2060) in adults aged≥65 years. Alzheimers Demen 15, 17–24.

Galvin

, Chrisphonte

, Chang

(2021) Medical and social determinants of brain health and dementia in a multicultural community cohort of older adults. J Alzheimers Dis 84, 1563–1576.

Majoka

, Schimming

(2021) Effect of social determinants of health on cognition and risk of Alzheimer disease and related dementias. Clin Ther 43, 922–929.

Hollingshead

, Redlich

(1958) Social class and mental illness: A community study. Am J Public Health 97, 1756–1757.

Tolea

, Chrisphonte

, Galvin

(2020) The effect of sociodemographics, physical function, and mood on dementia screening in a multicultural cohort. Clin Interv Aging 15, 2249–2263.

10.

Galvin

, Tolea

, Chrisphonte

(2020) What older adults do with the results of dementia screening programs. PLoS One 15, e0235534.

11.

Wilkins

, Roe

, Morris

(2010) A brief clinical tool to assess physical function: The mini-physical performance test. Arch Gerontol Geriatr 50, 96–100.

12.

Zigmond

, Snaith

(1983) The hospital anxiety and depression scale. Acta Psychiatr Scand 67, 361–370.

13.

Nasreddine

, Phillips

, Bedirian

, Charbonneau

, Whitehead

, Collins

, Cummings

, Chertkow

(2005) The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc 53, 695–699.

14.

Koski

, Xie

, Finch

(2009) Measuring cognition in a geriatric outpatient clinic: Rasch analysis of the Montreal Cognitive Assessment. J Geriatr Psychiatry Neurol 22, 151–160.

15.

Galvin

, Roe

, Powlishta

, Coates

, Muich

, Grant

, Miller

, Storandt

, Morris

(2005) The AD8: A brief informant interview to detect dementia. Neurology 65, 559–564.

16.

Galvin

, Roe

, Coats

, Morris

(2007) Patient’s rating of cognitive ability: Using the AD8, a brief informant interview, as a self-rating tool to detect dementia. Arch Neurol 64, 725–730.

17.

Borson

, Scanlan

, Chen

, Ganguli

(2003) The Mini-Cog as a screen for dementia: Validation in a population-based sample. J Am Geriatr Soc 51, 1451–1454.

18.

Raaijmakers

JGW

(1987) Statistical analysis of the Michaelis-Menten Equation. Biometrics 43, 793–803.

19.

Bahrami

, Kronmal

, Bluemke

, Olson

, Shea

, Liu

, Burke

, Lima

(2008) Differences in the incidence of congestive heart failure by ethnicity: The multi-ethnic study of atherosclerosis. Arch Intern Med 168, 2138–2145.

20.

Ferris

, Robertson

, Fabunmi

, Mosca

(2005) American Heart Association and American Stroke Association national survey of stroke risk awareness among women. Circulation 111, 1321–1326.

21.

Graham

(2015) Disparities in cardiovascular disease risk in the United States. Curr Cardiol Rev 11, 238–245.

22.

Farmer

, Ferraro

(2005) Are racial disparities in health conditional on socioeconomic status? Soc Sci Med 60, 191–204.

23.

Kaufman

, Cooper

, McGee

(1997) Socioeconomic status and health in blacks and whites: The problem of residual confounding and the resiliency of race. Epidemiology 8, 621–628.

24.

Sudano

, Baker

(2006) Explaining US racial/ethnic disparities in health declines and mortality in late middle age: The roles of socioeconomic status, health behaviors, and health insurance. Soc Sci Med 62, 909–922.

25.

Williams

, Mohammed

, Leavell

, Collins

(2010) Race, socioeconomic status, and health: Complexities, ongoing challenges, and research opportunities. Ann N Y Acad Sci 1186, 69–101.

26.

Manly

, Schupf

, Tang

, Stern

(2005) Cognitive decline and literacy among ethnically diverse elders. J Geriatr Psychiatry Neurol 18, 213–217.

27.

Scarmeas

, Stern

(2004) Cognitive reserve: Implications for diagnosis and prevention of Alzheimer’s disease. Curr Neurol Neurosci Rep 4, 374–380.

28.

Maalouf

, Fearon

, Lipa

, Chow-Johnson

, Tayeh

, Lipa

(2021) Neurologic Complications of poverty: Theassociations between poverty as a social determinant of health andadverse neurologic outcomes.. Curr Neurol Neurosci Rep 21, 29.

29.

Lin

, Daly

, Olchanski

, Cohen

, Neurman

, Faul

, Filit

, Freund

(2021) Dementia diagnosis disparities by race and ethnicity. Med Care 59, 679–686.

30.

Hernandez Saucedo

, Whitmer

, Glymour

, DeCarli

, Mayeda

, Gilsanz

, Miles

, Bhulani

, Tomaszewski Farias

, Olichney

, Mungas

(2022) Measuring cognitive health in ethnically diverse older adults. J Gerontol B Psychol Sci Soc Sci 77, 261–271.

31.

Matthews

, Hughes

(2001) Mental health service use by African American women: Exploration of subpopulation differences. Cultur Divers Ethnic Minor Psychol 7, 75–87.

32.

Tolea

, Chrisphonte

, Galvin

(2018) Sarcopenic obesity and cognitive performance. Clin Interv Aging 13, 1111–1119.

33.

Tolea

, Galvin

(2016) The relationship between mobility dysfunction staging and global cognitive performance. Alzheimer Dis Assoc Disord 30, 230–236.

34.

Tolea

, Heo

, Chrisphonte

, Galvin

(2021) A modified CAIDE risk score as a screening tool for cognitive impairment in older adults. J Alzheimers Dis 82, 1755–1768.

35.

Griswold

, Celis

, Bussies

, Rajabli

, Whitehead

, Hamilton-Nelson

, Beecham

, Dykxhoorn

, Nuytemans

, Wang

, Gardner

, Dorfsman

, Bigio

, Mesulam

, Weintraub

, Geula

, Gearing

, McGrath-Martinez

, Dalgard

, Scott

, Haines

, Pericak-Vance

, Young

, Vance

(2021) Increased APOE ɛ4 expression is associated with the difference in Alzheimer’s disease risk from diverse ancestral backgrounds. Alzheimers Dement 17, 1179–1188.