The Veterans Aging Cohort Study Index is Not Associated with HIV-Associated Neurocognitive Disorders in Uganda

Abstract

The Veterans Aging Cohort Study (VACS) Index has been associated with HIV-associated neurocognitive disorder (HAND) in some populations but has not been studied in sub-Saharan Africa. We investigated whether the VACS Index is associated with HAND in a rural population in Rakai, Uganda. HIV-infected (HIV+) adults on antiretroviral therapy underwent a neurocognitive battery for determination of HAND stage using Frascati criteria. VACS component scores were recorded for all participants. Out of 156 study participants, HAND stages were: 49% normal cognition, 15% asymptomatic neurocognitive impairment, 31% minor neurocognitive disorder, and 7% HIV-associated dementia. There was no significant association between VACS index and any HAND stage. In this first study of the VACS Index in sub-Saharan Africa, we found no association between VACS Index score and HAND.

INTRODUCTION

The widespread use of antiretroviral therapy (ART) has led to general improvement in HIV associated morbidity and mortality (Hammer et al., 2002; Hogg et al., 1998). However, the same has not been seen in HIV-associated neurocognitive disorders (HAND), which remain a prevalent comorbidity (Heaton et al., 2010). Evidence from two prospective studies in the post-ART era, the CNS HIV Antiretroviral Therapy Effects Research (CHARTER) study and the Multicenter AIDS Cohort study, shows the prevalence of HAND has remained persistently high even in virally suppressed persons living with HIV (PLWH) on ART (Heaton et al., 2010; Sacktor et al., 2016). Resource-poor settings such as sub-Saharan Africa (SSA) carry a disproportionate burden of HIV disease, yet diagnosis of HAND in these settings is limited by lack of trained personnel and resources for assessments at routine clinic visits (Kwasa et al., 2012). Utilizing routinely collected blood samples to identify patients at highest risk of of HAND could be used to identify and prioritize such patients for further neuropsychological testing.

The Veterans Aging Cohort Study (VACS) index, which was developed to capture the interacting effects of HIV disease, comorbidities, substance use and treatment toxicity to predict important clinical outcomes, may be a useful tool for this purpose (Justice et al., 2012). The VACS Index takes into account traditional HIV markers (CD4 cell count, plasma HIV RNA), age, hemoglobin level, renal function status based on the calculated glomerular filtration rate (eGFR) and the Fibrosis-4 (FIB-4) score, as well as chronic hepatitis C infection status (Shah, Lydecker, & Murray, 2009; Vallet-Pichard et al., 2007). The FIB-4 score is a non-invasive index that uses data readily available in primary care settings including age, platelet count and liver enzyme measures, to identify patients at with liver disease at highest risk of liver fibrosis who require advanced evaluation (Hudson, Sheron, Rowe, & Hirschfield, 2017). Thus, the VACS index takes into account injury to multiple organ systems in the setting of chronic HIV infection. (Tate et al., 2013). While initially developed to predict all-cause mortality amongst HIV+ veterans in the United States (US), it also has been shown to have utility in predicting cause-specific mortality, including in myocardial infarction and pneumonia. (Barakat et al., 2015; Salinas et al., 2016). The VACS index has been associated with neurocognitive impairment in HIV+ individuals in the US and Canada, but this association varies with ethnicity (M. J. Marquine et al., 2016; María J. Marquine et al., 2014; Rourke, 2015). The VACS Index has not previously been investigated in SSA, so its utility in predicting clinical outcomes in these ethnically diverse populations is unknown. We hypothesized the VACS Index would be associated with HAND in a population of PLWH receiving treatment with ART in rural Uganda.

METHODS

Study Participants.

A subset of 156 HIV+ adults on ART from the Rakai Neurology Study, which has been previously described (Sacktor et al., 2019), underwent laboratory investigations for determination of the VACS Index and were included in this analysis. Briefly, all participants underwent a detailed neurological history, physical examination, the Uganda neurocognitive battery (Sacktor et al., 2019), functional status assessment, and peripheral blood draw for CD4 cell count, plasma HIV viral load, complete blood count and comprehensive metabolic panel.

Standard Protocols, Approvals, Registrations and Patient Consents:

Written informed consent was obtained from all study participants. This study was approved by the Western Institutional Review Board, the Uganda Virus Research Institute Research and Ethics Committee, and the Uganda National Council for Science and Technology.

Statistical Analyses:

HAND stage (e.g., asymptomatic neurocognitive impairment (ANI), minor neurocognitive disorder (MND), and HIV-associated dementia (HAD)) was determined using normative neurocognitive data locally derived from 400 HIV-uninfected age- and sex-matched adults from Rakai District and the Frascati criteria as previously described. (Sacktor et al., 2019) HAND stage was then categorized as follows: (1) Any HAND stage (e.g. ANI, MND, and HAD) versus normal cognition; (2) symptomatic HAND (MND and HAD) versus ANI and normal cognition; and (3) HAD versus no dementia (normal, ANI or MND). The VACS Index was calculated as previously described,(Justice et al., 2012), but all participants were assumed to be hepatitis C virus uninfected as prior work has shown nearly absent rates of hepatitis C virus infection in Rakai District (Mullis et al., 2013). Mean VACS Index score and the mean score of each VACS index component were compared across all HAND stages using analysis of variance (ANOVA) and between dichotomous HAND groupings using t-tests.

RESULTS

Study participants were 49% male, had a mean age of 37 years (standard deviation (SD) ±8) and a mean education of 6 years (SD ±3) (Table 1). All participants were on ART with a median CD4 count of 357 (IQR 240 – 473), while 79% had an undetectable viral load. Following neurocognitive evaluation, 49% had normal cognition, 24% had ANI, 31% had MND, and 5% had HAD.

Table 1:

Demographic characteristics of the study cohort.

Demographics	n=156
Age (years) [mean (SD)]	37 (9)
Male [n (%)]	76 (49%)
Education (years) [mean (SD)]	6 (3)
Taking ART [n (%)]	156 (100%)
CD4 count [median (IQR)]	357 [240, 473]
Undetectable Viral Load [n (%)]	123 (79%)
Depression (CESD >=16) [n (%)]	9 (6%)
HAND Diagnosis (%)
Normal	76 (49%)
ANI	24 (15%)
MND	49 (31%)
HAD	7 (5%)

Abbreviations: ANI: Asymptomatic neurocognitive impairment; ART: antiretroviral therapy; CESD: Center for Epidemiologic Studies – Depression Scale; HAD: HIV-associated dementia; HAND: HIV-associated neurocognitive disorder; MND: minor neurocognitive disorder

There was no significant association between VACS index score and HAND stage. There was also no significant difference in VACS index score when comparing any HAND stage to normal cognition, symptomatic HAND to ANI/normal cognition, or HAD to those without dementia (Table 2). When comparing scores of each VACS index component across groups, the only significant differences were seen in age (participants with HAND were older) and FIB-4 score (participants with symptomatic HAND had higher scores than those with ANI and normal cognition) (Table 3).

Table 2:

Comparison of VACS Index scores by HAND stage. Of note, higher VACS Index scores indicate worse performance.

	VACS Index [mean (SD)]	p-value
Overall Cohort	16.4 (12.4)	---
HAND Diagnoses
Normal (n=76)	15.9 (11.3)	0.19
ANI (n=24)	13.2 (9.5)
MND (n=49)	18.1 (14.2)
HAD (7)	21 (16.6)
Any HAND
Normal (n=76)	15.9 (11.3)	0.62
ANI/MND/HAD (n=80)	16.9 (13.3)
Symptomatic HAND
Normal/ANI (n=100)	15.2 (10.9)	0.15
MND/HAD (n=56)	18.5 (14.5)
Dementia
Normal/ANI/MND (n=149)	16.2 (12.2)	0.32
HAD (n=7)	21 (16.6)

Abbreviations: ANI: asymptomatic neurocognitive disorder; HAD: HIV-associated dementia; HAND: HIV-associated neurocognitive disorder; MND: minor neurocognitive disorder; VACS: Veterans Aging Cohort Study

Table 3:

VACS Index component scores by HAND stage. Of note, higher scores indicate worse performance.

	Age Score [mean (SD)]	p	CD4 Score [mean (SD)]	p	Viral Load Score [mean (SD)]	p	Hemoglobin Score [mean (SD)]	p	FIB4 Score [mean (SD)]	p	GFR Score [mean (SD)]	p
Overall Cohort	1.0 (3.3)	---	6.0 (5.2)	---	1.0 (3.3)	---	7.1 (8.3)	---	1.3 (2.5)	---	0 (0.0)	---
HAND Diagnoses
Normal (n=76)	0.5 (2.3)	0.007	5.8 (4.9)	0.75	0.8(3.0)	0.38	7.7 (8.2)	0.71	1.0 (2.3)	0.03	0 (0)	----
ANI (n=24)	0 (0)		5.6 (3.3)		1.5 (3.6)		5.2(7.0)		1.0 (2.3)		0 (0)
MND (n=49)	2.2 (4.7)		5.8(5.6)		0.9 (3.4)		7.4 (9.3)		1.8 (2.8)		0 (0)
HAD (7)	1.7 (4.5)		10.4 (8.4)		2 (5.3)		4.3 (5.3)		2.6 (3.2)		0 (0)
Any HAND
Normal (n=76)	0.5 (2.4)	0.051	5.8 (4.9)	0.72	0.8 (3.0)	0.56	7.7(8.2)	0.34	1.0 (2.3)	0.1	0 (0)	----
ANI/MND/HAD (n=80)	1.5 (4.0)		6.1 (5.4)		1.1 (3.6)		6.5 (8.4)		1.7(2.7)		0 (0)
Symptomatic HAND
Normal/ANI (n=100)	0.4 (2.1)	0.008	5.8 (4.6)	0.53	1.0 (3.1)	0.97	7.1 (8.0)	0.95	1.0 (2.3)	0.03	0 (0)	----
MND/HAD (n=56)	2.1 (4.6)		6.4 (6.1)		1.0 (3.6)		7.0 (8.9)		1.9 (2.8)		0 (0)
Dementia
Normal/ANI/MND (n=149)	1.0 (3.3)	0.56	5.8 (4.9)	0.19	0.9 (3.2)	0.62	7.2 (8.4)	0.36	1.3 (2.5)	0.19	0 (0)	----
HAD (n=7)	1.7 (4.5)		10.4 (8.4)		2.0 (5.3)		4.3 (5.3)		2.6 (3.2)		0 (0)

Abbreviations: ANI: asymptomatic neurocognitive disorder; HAD: HIV-associated dementia; HAND: HIV-associated neurocognitive disorder; MND: minor neurocognitive disorder; VACS: Veterans Aging Cohort Study

DISCUSSION

Among HIV+ adults in a rural population in Uganda, we found no significant association between the VACS index and HAND, suggesting it has limited utility to predict patients at risk for neurocognitive impairment in this region. Of note, overall VACS Index scores were relatively low in this population, and several component scores, in particular the eGFR, were very low in this population.

Many factors may have impacted performance of the VACS Index in this population. For example, lifestyle factors, including mode of HIV acquisition (heterosexual in Rakai versus intravenous drug use or homosexual transmission in many Western cohorts), and the low prevalence of vascular risk factors in Rakai are notable differences between our study cohort and the VACS Index validation cohorts. Ethnicity may also be playing a role. Whereas Marquine et.. al demonstrated the association of the VACS index with HAND among populations in the US and Canada, outcomes were less certain among Hispanic sub-groups, which first raised the possibility that the association between HAND and VACS Index varied by ethnicity (M. J. Marquine et al., 2016). Of note, the VACS Index was associated with HAND amongst non-Hispanic black cohorts in the US, suggesting that VACS Index performance may differ between black Africans and non-Hispanic blacks in the US.

Other considerations include the role of other communicable and non-communicable comorbidities in the development of HAND amongst HIV+ populations in SSA that vary from those of HIV+ populations in Western cohorts. For instance, whereas the VACS index was primarily validated in populations with higher rates of hepatitis C virus infection, this differs significantly in our study population where there is instead a higher prevalence of hepatitis B virus (4.3%) (Pirillo M.F. et al., 2007) and virtually no hepatitis C infection (Mullis, et al. 2013). Other co-morbidities and factors unique to sub-Saharan Africa such as parasitic infections, exposure to endemic tuberculosis, differing diets, and genetic factors, may be driving the ongoing high rates of cognitive impairment amongst HIV+ individuals in this region and are not captured by the VACS Index. Further research is needed to investigate predictive factors in this population and develop clinical scales that identify individuals at high risk of neurocognitive impairment in this setting.

The significant association seen in the analysis of VACS components between older age and higher FIB-4 scores amongst individuals with symptomatic HAND compared to those with normal cognition and ANI likely reflects the known independent effect of age on cognitive functioning and the negative impact of sustained inflammation in patients with high FIB-4 scores (María J. Marquine et al., 2018). In particular, our findings were similar to other studies which have found an association between liver dysfunction and HAND. In a study of 1479 participants in the Women’s Interagency Health Study (WIHS), 89% of who were on ART, significant liver fibrosis was found to be associated with worse neuropsychological testing performance, and these associations were independent of both HIV and hepatitis C infection (Valcour et al., 2016). In a study involving 413 HIV+ adults at two centers in Italy, FIB-4 Index was also associated with a higher risk of cognitive impairment (Ciccarelli et al., 2014).

However, other components of the VACS Index which have been shown to be associated with HAND stage in other populations, including hemoglobin (Kallianpur et al., 2016) and renal function, were not significant in this cohort. For example, lower eGFR was shown to be the strongest predictor of neurocognitive decline among 191 HIV_ individuals in the longitudinal component of the CHARTER Study, 98% of whom were virally suppressed (Yuen et al., 2017). In addition, elevated Cystatin C levels, a biomarker of eGFR, have also been associated with cognitive decline in PLWH (Cantres-Rosario et al., 2013; Sakoda et al., 2017). The possible mechanism for the correlation between eGFR and future cognitive decline is unclear but could reflect vascular pathology in both the brain and the kidney. However, PLWH in this study did not have renal impairment, so there was no identifiable association between eGFR and HAND.

The major limitations of this study are its relatively small sample size and its cross-sectional design. Thus, larger longitudinal studies with longer follow up may better define the utility of the VACS index, if any, in predicting HAND in SSA.

CONCLUSION

To the best of our knowledge, this is the first study of the association of the VACS index with HAND in SSA. The lack of association in this study suggests it may have limited utility, but larger longitudinal studies in other regions of SSA are warranted. Further studies are needed to identify factors affecting neurocognitive outcomes in HIV+ populations in SSA and to create relevant clinical scales to identify individuals at highest risk of HAND that can then be widely implemented in these clinical settings.