Neuropsychiatric features in a multi-ethnic population with Alzheimer Disease and Mild Cognitive Impairment

Abstract

Background:

Alzheimer disease (AD) is more prevalent in African American (AA) and Hispanic White (HIW) compared to Non-Hispanic White (NHW) individuals. Similarly, neuropsychiatric symptoms (NPS) vary by population in AD. This is likely the result of both sociocultural and genetic ancestral differences. However, the impact of these NPS on AD in different groups is not well understood.

Methods:

Self-declared AA, HIW, and NHW individuals were ascertained as part of ongoing AD genetics studies. Participants who scored higher than 0.5 on the Clinical Dementia Rating Scale (CDR) were included. Group similarities and differences on Neuropsychiatric Inventory Questionnaire (NPI-Q) outcomes (NPI-Q total score, NPI-Q items) were evaluated using univariate ANOVAs and post hoc comparisons after controlling for sex and CDR stage.

Results:

Our sample consisted of 498 participants (26% AA; 30% HIW; 44% NHW). Overall, NPI-Q total scores differed significantly between our groups, with HIW having the highest NPI-Q total scores, and by AD stage as measured by CDR. We found no significant difference in NPI-Q total score by sex. There were six NPI-Q items with comparable prevalence in all groups and six items that significantly differed between the groups (Anxiety, Apathy, Depression, Disinhibition, Elation, and Irritability). Further, within the HIW group, differences were found between Puerto Rican and Cuban American Hispanics across several NPI-Q items. Finally, Six NPI-Q items were more prevalent in the later stages of AD including Agitation, Appetite, Hallucinations, Irritability, Motor Disturbance, and Nighttime Behavior.

Conclusions:

We identified differences in NPS among HIW, AA, and NHW individuals. Most striking was the high burden of NPS in HIW, particularly for mood and anxiety symptoms. We suggest that NPS differences may represent the impact of sociocultural influences on symptom presentation as well as potential genetic factors rooted in ancestral background. Given the complex relationship between AD and NPS it is crucial to discern the presence of NPS to ensure appropriate interventions.

Introduction

Alzheimer disease (AD) is a major public health challenge, especially among underserved diverse populations. The prevalence of late-onset AD (onset after 65 years of age) varies significantly by self-declared racial/ethnic groups, where African American (AA) and Hispanic White (HIW) are 1.5 to 2 times more likely to develop AD compared to Non-Hispanic White (NHW) individuals^1–5. In addition, AA and HIW also differ in clinical presentation, stage of disease at time of diagnosis, and dementia risk factors in relation to NHW^6–9. While race-ethnicity are overly simplistic surrogates for cultural influences on behavior and clearly not biological classifications, they often serve as proxies, albeit simplistic, for non-biologic influences (e.g., sociocultural factors and social determinants of health) and genetic ancestry^6,8–12. Non-biologic influences are especially pertinent to diverse groups and reflect a range of sociocultural factors that may impact the risk and course of AD^13–16.

The core AD phenotype is characterized by progressive cognitive and functional decline, with the presence of at least one neuropsychiatric symptom (NPS) in 80% of AD cases¹⁷. NPS are non-cognitive multidimensional behavioral disturbances (e.g. depression, anxiety, and agitation) that are commonly prodromal to AD^18–20. Further, while many of these symptoms are often overlooked or attributed to clinical manifestations of AD, they significantly affect daily living activities, often worsening prognosis and, have been shown tied to reduced lifespan^21–26. Not surprisingly, NPS are major contributors to the social and economic burden of AD, with high costs for hospitalization, residential placement, and psychopharmacologic therapy^27–29.

The assessment of NPS in AD is complicated as these symptoms are heterogenous and require clinical evaluation by trained specialists to optimize non-pharmacologic and pharmacologic approaches³⁰. Multiple screening tools have been developed to identify NPS. The widely used Neuropsychiatric Inventory Questionnaire (NPI-Q) has demonstrated clinical reliability and validity to identify NPS³¹. Studies using the NPI-Q have shown that NPS are associated with faster progression from mild cognitive impairment (MCI) to dementia³², accelerated cognitive decline³³, and increased functional decline³⁴. However, most outcome studies using the NPI-Q in AD and related dementias have been restricted to NHW individuals and have not examined its utility in diverse populations.

Multiple studies suggest that there is good reason to examine the utility of the NPI-Q in diverse groups as it appears that AA and HIW individuals have different patterns of NPS compared to NHW individuals^35–38. For instance, when examining NPS in AA individuals and NHW individuals with AD, psychosis and insomnia were more prominent among AA while NHW had higher levels of anxiety and depression^39,40. Similarly, a study evaluating NPS differences between AD NHW and Hispanic individuals primarily from Central American countries showed that Hispanic individuals had a higher NPI-Q total score and a higher prevalence of specific NPS symptoms including Delusions, Hallucinations, Anxiety, Disinhibition, and Irritability compared to NHW individuals with AD⁴¹. Other studies have also demonstrated NPS differences in AD Hispanic individuals depending on the country of origin, for example the prevalence of Anxiety varies from 16% in Mexican Americans, 22% in Spaniards and 36% in Brazilians ^17,37,42,43. Overall, these results lend support to both sociocultural and genetic ancestry influences on the expression of NPS.

Understanding variations of NPS in AD in different populations is critical to enhancing evaluation and management of NPS in all groups. The objective of this study is to identify similarities and differences in NPS as a function of self-declared race-ethnicity among individuals with AD and related dementias. We hypothesize there will be population specific NPI-Q patterns.

Methods

Participants

Our study sample consisted of self-declared AA, HIW, and NHW individuals who were ascertained for ongoing studies of AD genetics. Ascertainment sites included the University of Miami, Wake Forest University, and Case Western Reserve University. Our HIW group consisted mainly of Puerto Rican (PR) (residing in the Puerto Rico or the US) and Cuban American (CA) (residing in the US) individuals as well as a small set of individuals from various Latin American and Caribbean countries. Our AA and NHW groups were composed of individuals mainly from states in the Southeastern US (NC, SC, GA, TN, and FL). All participants underwent standard clinical evaluations, during which sociodemographic, cognitive, and behavioral data were collected through direct assessment and informant report. In addition, participants were assigned consensus diagnoses using standard criteria^44–46. Individuals included in this study were selected from the parent studies based on the following criteria: (a) enrolled in a genetic study and completed study protocols; (b) age >60 years; (c) consensus diagnosis of MCI or AD; (d) Clinical Dementia Rating scale (CDR) Global score ≥ 0.5; (e) completed NPI-Q.

Neuropsychiatric Inventory Questionnaire (NPI-Q)

The NPI-Q is a 12-item questionnaire that is used to rate the presence of NPS (e.g., Hallucinations, Agitation, Depression, etc.) common to AD and dementia³¹. The NPI-Q was administered by trained clinical coordinators via an interview performed to the informant or caregiver of the individual for whom they care. Derived from the Neuropsychiatric Interview (NPI)^47,48, the NPI-Q has demonstrated association with cognitive and functional decline^33,34 as well as progression from MCI to dementia³². The NPI-Q has been validated in Spanish⁴⁹, as well as in Spanish speaking countries including Brazil⁵⁰, Mexico⁵¹, and Chile⁵². NPI-Q items are scored as absent (0) or present (1) and a total score is generated by summing positively endorsed items; higher NPI-Q total scores indicate greater NPS burden. For this study, the NPI-Q total score and individual item scores were used as dependent measures. The 12 NPI-Q items are shown in Supplementary table 1. NPI-Q items specifically reported will be italicized to differentiate them from general descriptions of NPS.

Clinical Dementia Rating scale (CDR)

The CDR is a semi-structured interview used to characterize cognitive and functional performance typical of AD and dementia⁵³. The CDR consists of six domains: Memory, Orientation, Judgment & Problem Solving, Community Affairs, Home & Hobbies, and Personal Care, which are rated individually and then synthesized to assign a Global CDR score. The Global CDR score ranges from normal (Global CDR=0) to severe dementia (Global CDR = 3). For this study, the CDR was recoded into two levels: Early stage (i.e., mild dementia; a Global CDR=0.5 or 1) and Late stage (i.e., moderate-severe dementia; Global CDR > 1). Similar to the NPI-Q, the CDR information is often provided by the information or caregiver if the individual participating in the study is cognitively unable to answers these questions.

Analysis

A one-way ANOVA was used to compare self-declared race-ethnicity groups on sex, CDR severity (% Late stage), age at evaluation, age at onset, and years of education. A Factorial ANOVA was used to examine the main effects of race-ethnicity (AA, HIW and NHW), sex, CDR severity (Late Stage, Early Stage) on the NPI-Q total score and individual NPI-Q items (e.g., agitation). Both estimated mean differences (EMD) and estimated proportional differences (EPD) or the differences for each factor after adjusting for the other variables in the model are reported for the Factorial ANOVA post-hoc tests. As a secondary analysis, an independent-samples t-test was used to compare our PR (n = 44) and CA (n = 77) HIW individuals on NPI-Q total score and individual NPI-Q items. Bonferroni corrections were used for post hoc tests. All analyses were conducted using IBM SPSS Statistics for Windows, Version 27 (IBM SPSS Statistics for Windows, IBM Corporation, Armonk, NY).

Results

Our sample consisted of 498 individuals (AA=131; HIW=148; NHW=219). The one-way ANOVA (See Table 1) showed that our self-declared race-ethnicity groups differed significantly in the proportion of males-females (p < 0.001), years of education (p < 0.001), and proportion of CDR Category (p = 0.017), NPI-Q total scores at early stage (p = 0.044), and NPI-Q total scores at late stage (p = 0.010). There were no statistically significant differences between group on age at exam (p = 0.111) or age at onset (p = 0.563). Post-hoc tests revealed that NHW had a lower proportion of females than AA (p = 0.003) and HIW (p = 0.007) individuals, but there was no difference in the proportion of females between the AA and HIW (p = 1.000). NHW had significantly more years of education than both AA (p <0.001) and HIW (p < 0.001), and AA had more years of education than HIW (p = 0.004). HIW had a greater proportion of individuals in the CDR Late dementia stage than AA (p = 0.015) individuals. No proportional differences in CDR stages were seen between NHW and either AA (p = 0.154) or HIW (p = 0.737) individuals. Finally, HIW had a greater NPI-Q total score than NHW at both the early (p = 0.041) and late (p = 0.013) CDR stages, while no differences were found between AA and both HIW and NHW at either the early (HIW: p = 0.752; NHW p = 0.560) or late (HIW: p = 0.078; NHW: p = 1.000) CDR stages.

Table 1.

Race-ethnicity group comparisons (ANOVA)

Variable	AA	NHW	HIW	Overall	p-value
% Female (n= 498)	71.0% a	53.4% a, c	68.9% c	62.7%	< 0.001
% CDR >1 (n= 498)	35.9% b	46.6%	52.7% b	45.6%	0.017
NPIQ-Total CDR Early (sd) (n= 270)	3.21(2.9)	2.74(2.3)c	3.69(2.5)c	3.13(2.6)	0.044
NPIQ-Total CDR Late (sd) (n= 227)	3.94(2.5)	3.88(2.9)c	5.05(2.5)c	4.30(2.7)	0.010
AAE (sd) (n= 399)	77.9 (9.4)	78.9 (9.7)	80.3 (9.3)	79.2 (9.5)	0.111
AOO (sd) (n= 361)	72.3 (9.9)	73.5 (8.9)	74.0 (9.5)	73.4 (9.2)	0.563
Education (sd) (n= 389)	12.0 (3.1) a, b	13.6 (3.5) a, c	10.2 (4.6) b, c	12.4 (3.9)	< 0.001

^a= Difference between AA and NHW p < 0.05;

^b= Difference between AA and HIW p < 0.05;

^c= Difference between NHW and HIW p < 0.05;

AA: African American; AAE: age at exam; AOO: age of onset; CDR Early: Clinical Dementia Rating scale early stage (CDR global score <1); CDR Late: Clinical Dementia Rating scale late stage (CDR global score >= 1); HIW: Hispanic White; NHW: Non-Hispanic White;NPI-Q Total: Neuropsychiatric Inventory Questionnaire total score; sd = standard deviation; n: sample size.

NPI-Q Total Score Factorial ANOVA

We tested the main effects of self-declared race-ethnicity, CDR, and sex on NPI-Q total score using a factorial ANOVA. This analysis revealed main effects of race-ethnicity (p < 0.001) and CDR category (p < 0.001) on the NPI-Q total score. No main effects were found for sex (p = 0.48). Post-hoc comparisons showed that HIW had a greater NPI-Q total score compared to AA (EMD = 0.93, p = 0.009) and NHW (EMD = 1.14, p < 0.001), while there was no statistically significant difference between AA and NHW (EMD = 0.20, p = 1.000). Finally, those in the CDR Late stage (i.e., more severe dementia) had a significantly higher NPI-Q score than those in the CDR Early stage (EMD = 1.11, p < 0.001).

Neuropsychiatric Inventory Questionnaire (NPI-Q) Item Level Factorial ANOVAs

We tested the main effects of self-declared race-ethnicity, CDR, and sex on each of the NPI-Q items using factorial ANOVA. Significant results are presented in Table 2.

Table 2.

Factorial ANOVA main effects for NPI-Q total score and NPI-Q items

	Sexa (F, p)	Race/Ethnicity (F, p)	CDRb (F, p)
NPI-Q Total Score	ns	7.70, p < 0.001	22.17, p < 0.001
Agitation	4.48, p = 0.04	ns	11.92, p < 0.001
Anxiety	ns	22.347, p < 0.001	ns
Apathy	ns	7.03, p < 0.001	ns
Appetite	ns	ns	12.20, p < 0.001
Delusions	ns	ns	ns
Depression	ns	22.76, p < 0.001	ns
Disinhibition	ns	6.46, p < 0.01	ns
Elation	ns	3.06, p = 0.05	ns
Hallucinations	4.54, p = 0.03	ns	33.14, p < 0.001
Irritability	5.03, p = 0.03	14.00, p < 0.001	20.74, p < 0.001
Motor Disturbance	ns	ns	14.62, p < 0.001
Nighttime Behavior	ns	ns	15.69, p < 0.001

^aReference=Female;

^bReference=Early stage;

CDR = Clinical Dementia Rating scale; F = F-statistic; p= p-value; NPI-Q = Neuropsychiatric Inventory Questionnaire; ns= not significant.

Education was not significant for the NPI-Q total score and all NPI-Q items

Race-ethnicity

In total, six NPI-Q items showed similarities and six NPI-Q items showed differences among our groups (Figure 1). We found no statistical differences in the NPI-Q items Agitation, Appetite, Delusions, Hallucinations, Motor Disturbances and Nighttime Behaviors. In contrast, we found a significant main effect of self-declared race-ethnicity on NPI-Q items including Anxiety (p < 0.001), Apathy (p < 0.001), Depression (p < 0.001), Irritability (p < 0.001), Disinhibition (p = 0.002), and Elation (p = 0.046).

Figure 1.

Frequencies of Positive Endorsement on NPI-Q Items and NPI-Q Total Scores by Race-Ethnicity.

AA: African American; HIW: Hispanic White; NHW: Non-Hispanic White; NPI-Q: Neuropsychiatric Inventory Questionnaire.

Post hoc comparisons revealed that compared to both AA and NHW individuals, HIW individuals had a higher prevalence of Anxiety (AA: EPD = 30.1%, p < 0.001; NHW: EPD = 32.2%, p < 0.001), Apathy (AA: EPD = 22.3%, p < 0.001; NHW: EPD = 15.0%, p =0.013), Depression (AA: EPD = 33.3%, p < 0.001; NHW: EPD = 30.7%, p < 0.001) and Irritability (AA: EPD = 27.8%, p < 0.001; NHW: EPD = 23.7%, p < 0.001), all symptoms common to mood disorders. AA and NHW individuals showed no statistically significant differences on the prevalence of Anxiety (p = 1.000), Apathy (p = 0.528), Depression (p = 1.000) or Irritability (p = 1.000).

AA individuals had a higher prevalence of Disinhibition, than both NHW (EPD = 17.0%, p = 0.003), and HIW individuals (EPD = 16.6%, p = 0.005). AA individuals also had a higher prevalence of Elation compared to HIW individuals (EPD = 8.3%, p = 0.04), but not NHW individuals (EPD = 5.5%, p = 0.23); HIW and NHW individuals did not differ on Elation (EPD = 2.8%, p = 1.000).

Clinical Dementia Rating (CDR) scale

There were significant main effects of CDR category on several NPI-Q items (Figure 2). Individuals in the CDR Late stage had significantly more Agitation (EPD = 14.9%, p < 0.001), Appetite (EPD = 14.5%, p < 0.001), Hallucinations (EPD = 15.9%, p < 0.001), Irritability (EPD = 19.1%, p < 0.001), Motor Disturbance (EPD = 15.8%, p < 0.001), and Nighttime Behavior (EPD = 16.1%, p < 0.001). NPI-Q items related to mood changes like Anxiety, Apathy, Depression, Disinhibition, and Elation did not differ by CDR stage.

Figure 2.

Frequencies of Positive Endorsement on NPI-Q Items and NPI-Q Total Scores by CDR Category.

CDR: Clinical Dementia Rating scale global score; NPI-Q Neuropsychiatric Inventory Questionnaire.

Sex

We found a significant main effect of sex on select NPI-Q items (Supplementary Figure 1). Males had a higher prevalence of Agitation (EPD = 9.5%, p = 0.04) and Irritability (EPD = 9.8%, p = 0.03), while females had a higher prevalence of Hallucinations (EPD = 6.1%, p =0.03).

HIW NPI-Q differences by country of origin

We evaluated NPS differences between our main HIW groups, PR and CA individuals (Figure 3). The results revealed that NPI-Q total scores were similar between PR and CA individuals (t = −1.32, p = 0.19), mean difference (MD) = 0.60). However, CA individuals had a greater prevalence of Anxiety (t = −3.62, p < 0.001, MD = 32.5%) and Irritability (t = −2.56, p = 0.01, MD = 22.1%), while PR individuals had a higher prevalence of Appetite (t = 3.38, p = 0.001, MD = 28.9%) and Motor Disturbance (t = 2.45, p = 0.02, MD = 22.1%). No differences between PR and CA individuals were found for the items Agitation (t = −0.76, p = 0.45, MD = 6.8%), Apathy (t = −1.47, p = 0.15, MD = 13.6%), Delusions (t = −1.33, p = 0.19, MD = 10.4%), Depression (t = −1.91, p =0.06, MD = 17.9%), Disinhibition (t = −0.81, p = 0.42, MD = 5.8%), Elation (t = 0.37, p = 0.72, MD = 1.6%), Hallucinations (t = 1.27, p = 0.21, MD = 8.1%), or Nighttime Behavior (t = −1.17, p = 0.25, MD = 9.7%).

Figure 3.

Frequencies of Positive Endorsement on NPI-Q Items and NPI-Q Total Scores for Puerto Rican and Cuban American individuals.

CA: Cuban; NPI-Q Neuropsychiatric Inventory Questionnaire; PR: Puerto Rican.

Discussion

The primary aim of this study was to examine similarities and differences in the prevalence of neuropsychiatric symptoms (NPS) among individuals with AD or MCI from different self-declared race-ethnicity groups using the NPI-Q. Our results demonstrated that while the prevalence of several NPS are comparable among race-ethnicity groups and AD stages, our HIW individuals had a significantly higher prevalence of select NPS including Anxiety, Apathy, Depression, and Irritability. The results also revealed that AA individuals had a higher prevalence of Disinhibition compared to both NHW and HIW individuals, and a higher prevalence of Elation compared to HIW individuals. As expected, those individuals in the later stages of dementia had more overall NPS, and a higher prevalence of Agitation, Appetite, Delusions, Hallucinations, Irritability, Motor Disturbance, and Nighttime Behavior. Finally, males had a higher prevalence of Agitation and Irritability, while females had a higher prevalence of Hallucinations.

Our results showing more Anxiety and mood symptoms among HIW individuals compared to AA and NHW individuals, are not surprising. Several studies have demonstrated that older HIW individuals with dementia (primarily Mexican Americans) have a higher prevalence of mood problems (as measured by the NPI) compared to older NHW individuals^36,41,54. Further, anxiety and depression symptoms (as measured by the Spielberger State Trait Anxiety Inventory), appear to be greater among healthy older Hispanic adults as well suggesting that such symptoms may predate dementia^55–57. Finally, our secondary analyses suggest that the role of country of origin may be an important factor in understanding NPS among HIW. Specifically, we found that the prevalence of Anxiety and Irritability were greater in CA vs PR individuals (with trends suggesting a higher prevalence of Apathy and Depression in CA as well). At a minimum, these results confirm that HIW are not uniform in the occurrence of NPS. However, they also point to the potential contributions of sociocultural differences⁵⁸ between these groups as well as differences in genetic ancestry (i.e., different proportions of African, European, and Amerindian ancestry)^59–61.

Our AA and NHW had a similar total number of neuropsychiatric symptoms, however there were differences in select symptoms. AA individuals had a higher prevalence of Disinhibition compared to both NHW and HIW individuals, and a higher prevalence of Elation compared to HIW individuals. However, unlike some other studies we did not find differences between AA and NHW individuals on Anxiety, Depression, or psychosis related symptoms^39,40. We believe that the lack of differences may be related to measurement issues. This is especially pertinent when considering the lower prevalence of Anxiety and mood symptoms in AA individuals. An alternative explanation may be that there are reporting differences for select NPS (e.g., Depression) due to stigma or a focus on somatic symptoms that are common to depression^62–63. All these results, highlight the importance of further study in this area as clinical decision making is guided by measures such as the NPI-Q and could lead to underutilization of well-established pharmacologic therapies.

Conceptualizing NPS in AD is a complex undertaking. Neuropathological studies have demonstrated that NPS might be the result of molecular and functional alteration in specific brain regions and not just a behavioral consequence of AD⁶⁴. While neuropathological studies highlight potential biologic underpinnings it does not take into account non-biological differences between multi-ethnic groups. We suggest that sociocultural differences may exert a strong influence on the measurement of NPS and diagnostic endpoints^65–66 Thus, awareness of sociocultural influences on measurement and uptake of treatment is critical to ensure that these services are delivered. At the same time, the role of genetic risk and its variation among groups with different ancestral backgrounds may also contribute to NPS endpoints. Ancestry informed genetic studies of NPS across different populations can identify potential ancestry specific targets for future interventions and also provide insight on ancestral influences in genes associated to individual NPS such as Anxiety and Depression. Finally, our results have implications for how NPS are measured and ultimately treated, with special consideration on the impact that caregiver or informant distress is present at the time NPS are being reported.

Limitations

First, while the NPI-Q is a well-validated measure of neuropsychiatric symptoms common to AD, it is brief and may overestimate the prevalence of these symptoms as opposed to clinical examination and formal diagnosis⁶⁷. Second, NPI-Q results may be influenced by culturally related communication issues pertinent to NPS (e.g., understanding what constitutes the specific symptoms)⁶⁶. This limitation can be addressed in future studies in which individuals from various groups are asked to describe their understanding of NPS and its expression. Third, while country of origin may be a proxy for socio-cultural practices, it does not consider acculturation among individuals who no longer live in their country of origin, future analyses will examine differences in those still living in their home country (e.g., PR, CA), vs those who have immigrated to a new country. Finally, language and socio-cultural language barriers may contribute to the differences observed. While all individuals were examined in their preferred language (native or acquired), the language of administration and socio-culture language interpretation of NPI-Q items could influence response patterns from informant and caregivers and is worth investigating.

Conclusions

Consistent with prior studies, our results point to self-declared race-ethnicity group differences in the prevalence of NPS among individuals with AD. The burden of these symptoms on top of the cognitive and functional impairments of AD or MCI is exacerbated by poor understanding of these phenomena and how to best treat them. Future efforts to optimize the treatment of neuropsychiatric symptoms in AD or MCI requires careful consideration of both biologic and non-biologic factors that disproportionately affect diverse populations. Even with improvements in dissecting the biology of these symptoms in diverse populations, translational benefit will hinge on addressing both biological and non-biological influences. Our task now is to further explore the way such information can be used to better inform culturally attuned interventions—particularly given evidence that effective and timely treatment of neuropsychiatric symptoms in AD or MCI can have broad effects on cognition and behavior as well as caregiver stress.