Sociodemographic Characteristics of HIV-Associated Dementia in the South Carolina Alzheimer’s Disease Registry

Abstract

Objectives:

Studies examining the sociodemographic characteristics associated with human immunodeficiency virus (HIV)–associated dementia (HAD) are lacking, especially in the southern United States. The aim of this study was to describe the characteristics of HAD using South Carolina Alzheimer’s Disease Registry data, and compare these characteristics with other forms of dementia in South Carolina.

Methods:

Data were obtained from the population-based, South Carolina Alzheimer’s Disease Registry from 2000–2006 and 2010–2016 (N = 165,487). Crude and multivariable logistic regression models were applied to determine sociodemographic characteristics associated with HAD by time period.

Results:

Younger, Black, Other, men, and urban populations had greater odds of being diagnosed as having HAD in both time periods. For example, compared with individuals aged 85 years and older, individuals aged 18 to 34 had 97 times the odds (adjusted odds ratio 97.0; 95% confidence interval 31.6–297.8) of being diagnosed as having HAD. In 2010–2016, however, nursing facility populations had a greater odds of being diagnosed as having HAD.

Conclusions:

We found that younger populations (younger than 74 years), communities of color, men, urban populations, and nursing facility populations were more likely to have HAD. Future research should focus on the association between HAD and risk for Alzheimer’s disease.

Before antiretroviral therapy (ART), acquired immunodeficiency syndrome (AIDS) dementia complex was a common adverse health outcome of living with human immunodeficiency virus (HIV), which was first defined in 1986.¹ AIDS dementia complex is now referred to as HIV-associated neurocognitive disorders (HAND), and include asymptomatic neurocognitive impairment, minor neurocognitive disorder, and HIV-associated dementia (HAD).¹

HAD is the most severe form of HAND^1,2 and is an independent risk factor for mortality among people living with HIV.¹ HAD occurs when HIV infection crosses the blood–brain barrier and also may develop into a debilitating disease.^3,4 HAD is characterized by memory loss, difficulties with thinking and/or concentrating, speaking clearly and loss of motor skills.⁴ If left untreated HAD also can affect instrumental activities of daily living⁵ and activities of daily living.^6,7 These challenges can lower adherence to ART.⁸ Because at least 90% to 95% adherence to ART is needed to obtain viral suppression, which lowers the risk of HIV transmission,⁹ more studies are needed to determine potential ways to reduce or prevent HAD among people living with HIV.

Many illnesses and related mortality, ranging from hypertension¹⁰ to infant mortality,¹¹ disproportionately affect the southern United States. South Carolina is representative of this phenomenon with respect to HIV/AIDS. The southern United States accounts for more than half (51%) of incident HIV cases yearly in the United States.¹² In 2019, South Carolina accounted for the ninth highest HIV incidence rate in the United States, at 13.2/100,000 population.¹³ As such, studies are needed that will assess the risk factors for HAD with the aim of reducing the incidence and prevalence, especially in the southern United States.

Health disparities exist in HIV/AIDS. Health disparities are defined as the differences that exist in the diagnoses, prevalence, morbidity, and mortality of a specific outcome and the related adverse health conditions that may exist for certain populations.¹⁴ For example, Black¹⁴ and Hispanic/Latinx¹⁴ populations are disproportionately affected by HIV compared with White populations,¹⁵ although risk behaviors alone do not explain the racial disparities.¹⁶ HIV prevalence and incidence rates also are higher among men compared with women¹⁷ and among urban populations compared with rural populations,¹⁸ although some rural communities have high HIV prevalence rates.¹⁵ These disparities are seen at the national level as well as in South Carolina. In the United States in 2018, Black populations made up 13% of the total population but accounted for 42% of new HIV diagnoses.¹⁹ In the same year, men accounted for approximately 8 in 10 new HIV diagnoses, with the majority (86%) attributed to men who have sex with other men.¹⁷ In South Carolina, men accounted for 48% of the population but for 8 in 10 new HIV diagnoses from 2018–2019.²⁰ Black populations accounted for 27% of the total population but for 65% of the total population living with HIV.²⁰ In South Carolina, urban counties also have the largest HIV prevalence and incidence rates²⁰; however, individuals living in rural areas were more likely to have AIDS at or be within 1 year of an HIV diagnosis.²¹

To date, there is a lack of studies examining the sociodemographic characteristics of HAD, especially in the southern United States; therefore, the aim of this study was to determine the sociodemographic characteristics of HAD among individuals in the South Carolina Alzheimer’s Disease Registry. Findings from this study will help to determine target populations for interventions geared toward improving cognition and functional outcomes in the context of HAD.

Methods

Data Source and Study Population

Data were obtained from the South Carolina Alzheimer’s Disease Registry from 2010 to 2016 (HAD: n = 505; no HAD: n = 164,982). The South Carolina Alzheimer’s Disease Registry is a comprehensive dataset of South Carolina residents who have been diagnosed as having Alzheimer’s disease and related dementias (ADRD).²² The South Carolina Alzheimer’s Disease Registry has maintained a record of ADRD diagnoses since 1988 and identifies individuals with ADRD when they (or their family members) seek provider services. Because no single system identifies all newly diagnosed patients with ADRD, cases are collected from several sources such as inpatient hospitalizations, vital records, chart abstractions, Medicaid, emergency departments, and memory clinics, ensuring the capture of as many diagnoses as possible.²² The University of South Carolina institutional review board deemed the present study as exempt.

Measures

HAD was operationalized by International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10), Clinical Modification (CM) codes. The ICD-9-CM code 042 (2010–2014) and the ICD-10-CM code B20 (2015–2016) were used along with a dementia code to identify HIV-associated dementia. The dementia codes include 290.0 to 290.3, 290.8 to 290.9, 331.0, 290.4 to 290.43, 291.2, 292.82, 294.10, 294.11, and 331.82 for ICD-9-CM and F03.90 to F03.91, G30.0 to G30.9, FO1.50 to F01.51, F10.27 to F10.97, F19.97, F02.80 to F02.81, and G31.83 for ICD-10-CM. The sociodemographic risk factors included age, race, sex, rurality, and place of diagnosis. Age was categorized into six groups: 18 to 34, 35 to 44, 45 to 54, 55 to 64, 65 to 74, 75 to 84, and 85 and older. Race/ethnicity was categorized into Black, Other (Asian, American Indian, other than listed), and Hispanic versus White. Sex was categorized as male versus female. Location was categorized as urban versus rural, and place of ADRD diagnosis was categorized as living in a nursing facility versus living in the community.

Data Analysis

Descriptive statistics were used to describe the study population and by HAD status and determined HAD prevalence estimates. Crude and multivariable logistic regression models were used to determine health disparities in and sociodemographic risk factors for HAD. Adjusted models controlled for all of the sociodemographic characteristics (age, race, sex, rurality, place of diagnosis). We were interested in seeing whether younger participants, populations of color, urban populations, and participants living in a nursing facility at the time of diagnosis had higher or lower odds of HAD compared with participants aged 85 and older, White populations, rural populations, and participants living in the community. All of the analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC).

Results

The prevalence of HAD in the South Carolina Alzheimer’s Disease Registry ranged from 0.27% to 0.38% from 2001–2006, with the highest estimate in 2006 (Fig.). These differences in prevalence were statistically significant (P = 0.026). The prevalence of HAD in the South Carolina Alzheimer’s Disease Registry ranged from 0.36% to 0.49% during the period 2010–2016, with the highest estimates in 2010 and 2013 (Fig.); however, these changes were not statistically significant (P = 0.148). Table 1 shows the sociodemographic characteristics of the study sample and by HAD status in 2010–2016. The majority of the study population was White, female, living in urban areas, and living in the community when they received their diagnosis. Individuals who had a higher prevalence of HAD tended to be 18 to 34, 35 to 44, 45 to 54, and 55 to 64 years old; Black race; male; and living in the community at the time of diagnosis.

Fig.

Prevalence of human immunodeficiency virus–associated dementia in the South Carolina Alzheimer’s Disease Registry, 2000–2006 and 2010–2016. Note missing data for human immunodeficiency virus–associated dementia for 2000.

Table 1.

Distribution of sociodemographic characteristics and by HIV-associated dementia status, 2010–2016

	N (%) N = 165,487	HAD n = 505	No HAD n = 164,982	P
Age group, y				<0.001
18–34	149 (0.10)	5 (1.08)	144 (0.09)
35–44	560 (0.37)	74 (16.0)	486 (0.32)
45–54	3136 (2.06)	138 (29.7)	2998 (1.98)
55–64	16,607 (10.9)	146 (31.5)	16,461 (10.9)
65–74	31,955 (21.0)	68 (14.7)	31,887 (21.0)
75–84	54,073 (35.5)	24 (5.17)	54,049 (35.6)
≥85	45,658 (30.0)	9 (1.94)	45,649 (30.1)
Race				<0.001
Black	42,626 (27.0)	368 (74.7)	42,258 (26.9)
White	109,578 (69.5)	109 (22.1)	109,469 (69.6)
Other	4851 (3.1)	15 (3.0)	4836 (3.1)
Hispanic	660 (0.4)	1 (0.2)	659 (0.4)
Sex				<0.001
Female	103,330 (62.6)	165 (32.7)	103,165 (62.6)
Male	61,873 (37.5)	339 (67.3)	61,534 (37.4)
Location				0.313
Urban	110,773 (72.9)	356 (75.0)	110,417 (72.9)
Rural	41,197 (27.1)	119 (25.1)	41,078 (27.1)
Place of diagnosis				0.010
Nursing facility	52,870 (34.0)	134 (28.4)	52,736 (34.0)
Living in the community	102,730 (66.0)	338 (71.6)	102,392 (66.0)

Boldface type indicates statistical significance. HAD, HIV-associated dementia; HIV, human immunodeficiency virus.

After adjusting for sociodemographic characteristics, younger ages had a higher odds of having HAD compared with those 85 and older. For example, from 2010–2016, compared with individuals aged 85 and older, individuals aged 18 to 34 years had 97 times the odds (adjusted odds ratio [aOR] 97.0, 95% confidence interval [CI] 31.6–297.8), those aged 35 to 44 had 371 times the odds (aOR 371.0, 95% CI 181.5–758.4), those aged 45 to 54 had 128 times the odds (aOR 128.0, 95% CI 64.5–254.3), those aged 55 to 64 had X times the odds (aOR 27.7, 95% CI 14.0–54.8), and those aged 65 to 74 had 7 times the odds (aOR 7.24, 95% CI 3.58–14.7). Black and Other populations had 5 times the odds of having HAD compared with White populations (Black: aOR 4.91, 95% CI 3.84–6.28; Other: aOR 5.17, 95% CI 2.95–9.07). Hispanic populations had a 23% higher odds of having HAD compared with White populations, although this was not statistically significant (OR 1.23, 95% CI 0.17–8.97). Men (aOR 1.69, 95% CI 1.37–2.08) and urban populations (aOR 1.39, 95% CI 1.10–1.74) had 69% and 39% higher odds of having HAD compared with women and rural populations, respectively. Individuals who were diagnosed in a nursing facility had a 31% higher odds of having HAD compared with individuals who were living in the community (aOR 1.31, 95% CI 1.05–1.63; Table 2). Similar patterns were seen for the earlier time period, except that there was no statistically significant association between nursing facility status and HAD.

Table 2.

Association between sociodemographic characteristics and HIV-associated dementia in South Carolina

Characteristic	2000–2006				2010–2016
	Crude OR	95% CI	aOR	95% CI	Crude OR	95% CI	aOR	95% CI
18–34 y vs ≥85 y	405.2	190.8–860.5	196.7	91.0–425.3	175.9	58.2–531.1	97.0	31.6–297.8
35–44 y vs ≥85 y	913.3	544.1–>1000.0	426.3	250.0–726.9	771.2	384–>1000.0	371.0	181.5–758.4
45–54 y vs ≥85 y	274.9	167.7–450.7	135.2	81.5–224.1	233.1	188.7–457.8	128.0	64.5–254.3
55–64 y vs ≥85 y	59.1	36.2–96.6	32.7	19.8–53.8	44.9	22.9–88.1	27.7	14.0–54.8
65–74 y vs ≥85 y	12.6	7.56–20.9	8.08	4.84–13.5	10.8	5.39–21.6	7.24	3.58–14.7
75–84 y vs ≥85 y	2.68	1.55–4.65	1.90	1.08–3.36	2.25	1.05–4.84	1.69	0.77–3.72
Black vs White	10.1	8.49–12.1	5.98	4.87–7.34	8.75	7.06–10.8	4.91	3.84–6.28
Other vs White	5.24	3.35–8.19	6.64	4.12–10.7	3.12	1.81–5.35	5.17	2.95–9.07
Hispanic vs White	1.22	0.17–8.74	1.03	0.14–7.41	1.52	0.21–10.9	1.23	0.17–8.97
Men vs Women	3.16	2.72–3.66	1.66	1.40–1.96	3.45	2.86–4.15	1.69	1.37–2.08
Urban vs rural	0.97	0.82–1.13	1.20	1.01–1.43	1.11	0.90–1.37	1.39	1.10–1.74
Nursing facility vs living in the community	0.48	0.41–0.57	0.94	0.79–1.12	0.77	0.63–0.94	1.31	1.05–1.63

Boldface type indicates statistical significance. aOR, adjusted odds ratio; CI, confidence interval; HIV, human immunodeficiency virus; OR, odds ratio.

Discussion

To our knowledge, this is the first study examining sociodemographic characteristics associated with HAD in the southern United States. The main finding of the present study was that age, race/ethnicity, sex, rurality and location were associated with HAD. Younger ages, Black and Other populations, men, urban and nursing facility populations had a higher odds of contracting HAD compared with those 85 years of age and older, White populations, women, rural and individuals living in the community, respectively, from 2010–2016. Nursing facility status, however, was not associated with HAD in the earlier time period.

The prevalence of HAD in the South Carolina Alzheimer’s Disease Registry increased over time from 2000–2006 (statistically significant) but decreased slightly over time (from 2010–2016; not statistically significant). This finding aligns with previous research showing that the prevalence of HAD has decreased because of ART and HAND and has improved after ART intensification.²³ As such, it is possible that ART uptake has improved during the last several years, resulting in the slight decline in HAD.

Despite the decline, specific sociodemographic characteristics are associated with HAD. These findings are consistent with the health disparities that exist in HIV/AIDS. At the national level, Black populations,^14,19 men,¹⁷ and urban populations¹⁸ have higher rates of HIV compared with White, women, and rural populations. These health disparities also exist in the southern United States, and South Carolina is no exception. Male^17,20 and Black populations^19,20 continue to bear the brunt of the epidemic in the United States and in the South.

Our findings show that HAD needs to be addressed among nursing facility populations, which is a novel finding. It also is interesting that the odds ratio for place of diagnosis switched directions after adjustment for confounders. It is possible that the sociodemographic characteristics, including age, could affect the relationship between place of diagnosis and HAD.

The highest odds in the present study were seen comparing populations aged 35 to 44 years to those 85 and older for both time periods. Previous research has shown that HIV is the most prevalent cause of dementia among younger adults^3,24,25; however, research tends to focus on middle-aged and older adults with respect to HAD assessments.^26,27 Nevertheless, the present study shows that HAD is of crucial concern among younger populations. Although there are HIV intervention and prevention programs geared toward youth^28–30 and there has been increasing attention paid to older adults,^31,32 middle-aged adults are sometimes forgotten with respect to HIV prevention and intervention programs. For example, the Centers for Disease Control and Prevention has a focus on youth³³ and older adults³⁴; however, at present, individuals who fall between these age categories do not have a subgroup focus. The present study finds that adults aged 18 to 74 but especially those aged 35 to 44 have a higher odds of being diagnosed as having HAD. Another contributing factor to HAD is late diagnosis, which may result in neuronal cell death.³⁵ Research has shown that older adults (aged 50 and older) are likely to be diagnosed at later stages in the course of infection.³⁴ This finding also may extend to middle-aged adults who had a higher odds of being diagnosed as having HAD compared with adults aged 85 and older.

There also are shared pathogenic factors between Alzheimer’s disease and HAD that are notable and warrant consideration. For example, Canet et al reported the misprocessing of amyloid precursor protein and Aβ synthesis as an Alzheimer’s disease symptom.³⁶ A related component observed in patients living with HIV was an increase in cerebrospinal fluid Aβ_1–42³⁷ and amyloid plaques being found in the brain.^38,39 Another Alzheimer’s disease symptom mentioned was activation of death pathways and apoptosis.³⁶ Among patients living with HIV, there was an increase in apoptosis and contents and activities of c-Jun N-terminal kinase and extracellular signal–regulated kinase.⁴⁰ Indeed, c-Jun N-terminal kinase signaling has been shown to be a therapeutic target for Alzheimer’s disease.⁴¹ Another key symptom of Alzheimer’s disease is excitotoxicity. There are increased cerebrospinal fluid glutamate levels among patients living with HIV.⁴² These shared pathogenic factors, among others, help differentiate HAND from Alzheimer’s disease, indicating that they are complex and multilayered issues in the field of aging and neuro-HIV.⁴³

There are some limitations to be considered in the present study. The variable examined place of diagnosis of ADRD in a nursing facility or living in the community. As such, individuals could have changed their living situations after being diagnosed as having dementia. Because of the small numbers of HAD among Asian, American Indian, and Other groups, these individuals were grouped into a racial category of “Other”; therefore, we were unable to examine differences in these specific racial populations. In the dataset, Alzheimer’s disease and vascular dementia were given priority, in that if a person’s record contained multiple indicators of ADRD, then Alzheimer’s disease and vascular dementia took precedence. Therefore, it is possible that there were people who were diagnosed as having HAD, but this was not indicated in the dataset. It also is important to note that HAD is reported to be operationalized by clinician-generated diagnoses reported in the South Carolina Alzheimer’s Disease Registry. As such, there are some limitations in diagnostic information and process because this kind of diagnosis at present is not standardized in registries. Specifically, HAD is a diagnosis of exclusion of other causes of neurocognitive impairment. In addition, the level of neurocognitive impairment and associated functional status need to be documented to confirm this diagnosis. As such, the prevalence and effects may be underestimated. Future studies can use the Frascati Conference criteria.⁴⁴ We also note that diagnoses such as mild neurocognitive disorder caused by HIV and HIV-associated asymptomatic neurocognitive impairment are more prevalent than HAD; however, these health outcomes are not available in the South Carolina Alzheimer’s Disease Registry. Future research can examine the health disparities in these more prevalent outcomes. It also is likely that many of the results obtained is related more to the epidemiology of HIV infection itself than any specific etiological factors associated with dementia caused by HIV infection.

The study also had some strengths. This is the first study to examine the health disparities in HAD in a high-risk area, the southern United States. We also had a large sample size to analyze from the registry. The South Carolina Alzheimer’s Disease Registry is population based and collects cases from multiple sources such as inpatient hospitalizations, vital records, chart abstractions, Medicaid, emergency departments, and memory clinics, contributing to the comprehensiveness of the registry.

Conclusions

Even though ART has reduced the rate of HAD,⁴⁵ specific sociodemographic characteristics are associated with HAD. Younger populations, communities of color, men, urban populations, and individuals in nursing facilities had a higher odds of having HAD in 2010–2016; however, nursing facility status was not statistically significant in the earlier time period. The study suggests that there are health disparities with HAD, and further investigation is needed. Future research should focus on the association between HAD and risk for Alzheimer’s disease and younger populations for HAD assessments. HIV prevention, early HIV diagnosis and treatment, and retention in care also are ways in which HAD rates may be reduced.

Key Points.

• Younger populations had a higher odds of being diagnosed as having human immunodeficiency virus (HIV)–associated dementia (HAD) as compared with older populations.

• Black populations had greater odds of being diagnosed as having HAD as compared with White populations.

• Urban populations had greater odds of being diagnosed as having HAD as compared with rural populations.

• Nursing facility populations had greater odds of being diagnosed as having HAD as compared with populations living in the community.