Cognitive and Motor Impairment Severity Related to Signs of Subclinical Wernicke’s Encephalopathy in HIV Infection

Abstract

Background:

Wernicke’s encephalopathy (WE) is a neurological condition resulting from thiamine deficiency. Although commonly associated with alcoholism, non-alcoholic WE has been described in individuals with Human Immunodeficiency Virus (HIV) infection, but subclinical WE may be underdiagnosed. The current study questioned whether presence of subclinical WE signs underlies cognitive and motor deficits in HIV individuals as observed in alcoholism.

Setting:

56 HIV positive individuals (HIV+) and 53 HIV negative controls (HIV−) were assessed on 6 cognitive and motor domains: attention/working memory, production, immediate and delayed episodic memory, visuospatial abilities, and upper limb motor function.

Methods:

Based on a rating scheme by Caine et al, HIV+ individuals were categorized by subclinical WE risk factors (dietary deficiency, oculomotor abnormality, cerebellar dysfunction, and altered mental state). Performance was expressed as age- and education-corrected Z-scores standardized on controls.

Results:

Sorting by Caine criteria yielded 20 HIV+ as Caine 0 (i.e., meeting no criteria), 22 as Caine 1 (i.e., meeting one criterion), and 14 as Caine 2 (i.e., meeting two criteria). Comparison among HIV+ Caine subgroups revealed a graded effect: Caine 0 performed at control levels, Caine 1 showed mild to moderate deficits on some domains, and Caine 2 showed the most severe deficits on each domain.

Conclusion:

This graded severity pattern of performance among Caine subgroups suggests that signs of subclinical WE can partly explain the heterogeneity in HIV-related cognitive and motor impairment. This study highlights the utility of Caine criteria in identifying potential causes of HIV-related neurocognitive disorders and has implications for disease management.

INTRODUCTION

Treatment of HIV infection with highly active antiretroviral therapy (HAART) and combination antiretroviral therapy (cART) has been successful in reducing viral load and improving immune system functioning. As a consequence, the prevalence of central nervous system (CNS) opportunistic infections, acquired immunodeficiency syndrome (AIDS), and premature death in people living with HIV has decreased ^1-4. However, despite pharmacological treatment, HIV-associated neurocognitive disorder (HAND) occurs in upwards of 50% of HIV-infected patients ^5-12 and is marked by heterogeneity in the severity of cognitive impairment, ranging from asymptomatic neuropsychological impairment (ANI) to mild neurocognitive disorder (MND) and HIV-associated dementia (HAD) ^13-17.

Antiretroviral therapy (ART) has been instrumental in preventing the development of the most severe stage of HAND, i.e., HAD ¹⁸, but even in this era of improved pharmacological therapy, mild and moderate cognitive and motor dysfunctions endure in HIV-infected individuals^19,20. HIV infection is associated with deficits in selective cognitive domains including attention/working memory and executive functioning ^7,21-27, episodic memory ^26,28,29, psychomotor speed ²⁶, and visuospatial abilities ^23,30. Motor domains are also affected and include deficits in balance, gait, and manual dexterity ^31-34.

Wernicke’s encephalopathy (WE), an acute neurological condition resulting from severe thiamine depletion or deficiency, has been historically diagnosed by a clinical triad of symptoms: oculomotor abnormalities, gait and balance instability, and altered mental state ³⁵. Although this neurological syndrome is commonly associated with alcoholism, non-alcoholic Wernicke’s encephalopathy can occur in individuals with HIV infection with or without AIDS and subclinical WE could be underdiagnosed in clinical practice due to confounding HIV-related etiologies, notably infectious and neoplastic etiologies due to the immunosuppressed state ^36-39. Indeed, the diagnosis of WE is often made only postmortem and could be missed in as many as 75-80% in patients with alcoholism and AIDS ^40,41. Thiamine deficiency was observed in a large proportion of HIV-positive individuals in advanced stages of HIV infection and in clinically asymptomatic patients in the pre- and post-HAART era ^41-43. A recent review reported that genetic studies enabled determination of selective proteins linking thiamine to HIV pathology, although thiamine also affects HIV through non-genomic factors (for review ⁴⁴).

By conducting chart reviews of clinical histories of alcoholic patients with definitive postmortem neuropathological WE diagnosis not identified during their lifetime, Caine et al (1997) ⁴⁵ proposed operational criteria enabling the antemortem identification of subclinical WE with a high degree of specificity. Postmortem WE diagnosis required the historical in vivo presence of just two of the following criteria: (1) dietary deficiency, (2) oculomotor abnormality, (3) cerebellar dysfunction, or (4) an altered mental state or cognitive impairment. In vivo studies have identified graded effects in neuropsychological performance when alcoholic individuals were classified by number of Caine criteria met ^46,47. Those meeting two or more criteria demonstrated the most severe cognitive impairment compared with those meeting none or just one criterion. These observations raised the question whether the presence of subclinical WE signs as defined by Caine et al could contribute to the heterogeneity of cognitive and motor deficits observed in HIV-infected patients ^14-16.

The present study aimed to determine whether the heterogeneity in the pattern and severity of cognitive and motor performance deficits observed in HIV infection could be partially explained by risk factors for WE. We tested the following hypotheses: (1) categorizing the HIV+ group by Caine criteria would reveal a stepwise deficit severity, where HIV+ participants meeting two or more criteria would show the greatest and most widespread cognitive and motor deficits as similarly observed in alcoholic individuals and (2) given the prolonged life expectancy of ART-treated HIV-infected individuals, aging, disease-related clinical variables (e.g., CD4 cell count, log viral load, and length of time with HIV), comorbid infections such as hepatitis C virus, and AIDS status represent additional factors of the heterogeneity of HIV-associated neurocognitive impairment. Therefore, we hypothesized that risk factors for WE would predict cognitive and motor performance above and beyond these other clinical variables.

METHODS

Participants

Study participants included 56 HIV seropositive individuals (HIV+; 40 men and 16 women) and 53 individuals seronegative for HIV infection (HIV−; 29 men and 24 women). HIV+ individuals were included between 2008 and 2017, with 36 HIV+ participants between 2008-2013 and 20 HIV+ participants between 2014-2017. Data from these HIV− control participants were included in earlier studies that examined cognitive impairment in alcoholism related to Caine criteria ^46,47. The current study was conducted in accordance with protocols approved by the Institutional Review Boards of Stanford University and SRI International. Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.

All participants were screened using the Structured Clinical Interview for DSM-IV (SCID) ⁴⁸ and structured health questionnaires. Upon initial assessment, subjects were excluded if they had fewer than 8 years of education, a significant history of medical (e.g., epilepsy, stroke, multiple sclerosis, uncontrolled diabetes or loss of consciousness >30 minutes), psychiatric (i.e., schizophrenia or bipolar I disorder), or neurological disorder (e.g., neurodegenerative disease). All participants completed the Ohio State University Traumatic Brain Injury Identification Method-short form⁴⁹ and none of them reported signs of lifetime history of TBI. Other exclusionary criteria were recent (i.e., past month) substance abuse or dependence other than nicotine in the HIV+ group according to the DSM-IV criteria, HIV-related opportunistic infection in the HIV+ group, or any DSM-IV Axis I disorder in the HIV− control group. No HIV− control participant met DSM-IV criteria for substance abuse or dependence. Of the 56 HIV positive participants, 28 (50%) had a lifetime history of DSM-IV substance abuse/dependence. Of the 28 participants with a lifetime history of DSM-IV substance abuse/dependence, 18 (64.3%) had Cocaine Abuse/Dependence, 16 (57.1%) had Cannabis Abuse/Dependence, 9 (32.1%) had Amphetamine Abuse/Dependence, 2 (7.1%) had Opioid Abuse/Dependence, and 1 (3.6%) had Hallucinogen Abuse/Dependence. Of the 28 HIV positive participants with a lifetime history of substance abuse/dependence, all were in remission (mean remission time = 622.7 ± 498.7 weeks; median = 516 weeks; range = 9 to 1395 weeks). For nicotine, 11 (19.6 percent) of the 56 HIV+ participants were current smokers and 10 HIV+ (17.9 percent) were past smokers, whereas 1 of 53 (1.4 percent) HIV− controls were current smokers, and 2 (2.8 percent) were past smokers. In the group of 56 HIV+ subjects, 51 participants had no history of alcohol abuse or dependence, four had a remote history of alcohol abuse lasting 4 months to 6 years and one participant had a remote history of alcohol dependence lasting 2 years. Regarding DSM-IV alcohol dependence/abuse remission criteria at the time of visit, these five HIV+ individuals were in full sustained remission with length of abstinence from 5 years to 34 years for the four HIV+ subjects with alcohol abuse history and 22 years for the HIV+ individual with alcohol dependence history.

Blood assays confirmed self-report of HIV status via antibody testing, determined plasma viral load (log10 transformed), and CD4 T-lymphocyte count in the HIV+ group. Hepatitis C status was also tested. Twenty-eight of the 55 HIV+ participants met CDC criteria for AIDS (CD4 cell count <200 or an AIDS-defining event, e.g., coccidioidomycosis) sometime during the course of their illness. At the time of testing, only 2 of these 28 HIV+ participants had CD4 T-lymphocyte count <200. Most of the HIV+ participants (50 HIV, 90.9%) were on HAART (highly active antiretroviral therapy). In addition, 16 of the 55 (29.1%) HIV+ subjects tested positive for the hepatitis C virus (HCV). Severity of depressive symptoms was assessed with the Beck Depression Inventory-II ⁵⁰.

Demographics for the HIV+ and HIV− groups are presented in Table 1. HIV+ were older, had fewer years of education, and endorsed more depressive symptoms than HIV− controls.

Table 1.

Demographic and clinical characteristics of participants

Demographic Variables	HIV− controls	HIV+	p value
n =	53	56
BACKGROUND DEMOGRAPHICS
Sex (Men/Women)	29/24	40/16	ns
Age	47.8 ± 13.2	52.5 ± 7.9	.03 [HIV+ > HIV−]
Education (years)	15.7 ± 2.5	13.7 ± 2.5	<.001 [HIV+ < HIV−]
Hepatitis C (%)	0%	29.1%	/
Beck Depression Inventory-II (BDI-II) Score *	2.4 ± 3.5; 1.0	9.8 ± 8.4; 8.5	<.001 [HIV+ > HIV−]
HIV-RELATED DEMOGRAPHICS
HIV Approx. Age of Onset	/	35.9 ± 9.5	/
Length of time with HIV (years)	/	16.6 ± 8.2	/
Log Viral Load (n = 51)	/	2.1 ± 1.1	/
CD4 Count* (n = 54)	/	611.3 ± 286.0; 571.0	/
AIDS Status (%) (n = 55)	/	50.9%	/
HAART Medications (%) (n = 55)	/	90.9%	/

^*Mean ± SD; median

ns=not significant

Neuropsychological test measures

Six cognitive and motor domains were assessed: attention/working memory, production (ability to initiate, plan, and strategize), immediate episodic memory, delayed episodic memory, visuospatial construction, and upper limb motor speed/dexterity.

Attention/working memory

The Trail Making Test (Trails A and B) ⁵¹ required participants to sequence numbers from 1 to 25 and then sequence numbers and letters in alternating fashion (1–A–2–B–3–C…). Time to completion for each condition was the dependent measure.

Production ability

The Controlled Oral Word Association Test ⁵² required participants to say aloud as many words as they could beginning with the letters F, A and S. Each trial was 60 seconds. The total number of correct words produced across the three trials was calculated.

Ruff Figural Fluency Test ⁵³, often considered a visual analog of the Controlled Oral Word Association Test, required participants to generate as many different designs as possible by using straight lines to connect at least two of the five dots in each array. The test consists of five conditions, some with distractors and each with a 1-minute time limit. The total number of correct unique designs was summed across the five conditions.

Immediate memory

Logical Memory-I (WMS-R) ⁵⁴ required participants to recall two short narratives immediately after each story was read to them. Number of details recalled from both stories was summed (max = 50).

Rey–Osterrieth Complex Figure—immediate recall ⁵⁵ required participants to draw a complex figure from memory immediately after having copied it. The number of details recalled according to standardized scoring instructions was calculated (max = 36).

Delayed memory

Logical Memory-II (WMS-R) ⁵⁴ had participants recall the short narratives presented after a 30-minute delay.

Rey–Osterrieth Complex Figure —delayed recall ⁵⁵ required participants to draw from memory the complex figure after a 30-minute delay.

Visuospatial construction

Rey–Osterrieth Complex Figure—copy condition required participants to copy the complex figure as accurately as possible. Scoring was according to standardized criteria.

Upper limb motor performance

Fine finger movement ⁵⁶ required participants to turn a knurled rod with their forefinger and thumb, unimanually and then bimanually. Three, 30-second trials for each condition were administered with the unimanual and bimanual scores reflecting the average number of rotations across trials.

Grooved Pegboard ⁵⁷ required participants to insert keyed pegs into a 5 × 5 grooved pegboard. The times to completion for right and left hand trials were averaged.

Caine criteria

The four Caine criteria were operationalized in this study as follows: (1) Dietary Deficiency—showing abnormal low serum levels of prealbumin assessing current nutritional status (Adult Men: <21 mg/dL, normal range 21-43 mg/dL; Adult Women: <17 mg/dL, normal range 17-34 mg/dL), (2) Oculomotor Disturbance—exhibiting nystagmus (i.e., rapid and uncontrollable eyes movements) on neurological examination, determined by the observation of saccades and smooth pursuit during frontal eye gaze and fixed head position while subjects tracked the examiner’s index finger as it moved to the extreme right, extreme left, upward, and downward; an oculomotor abnormality is scored for participants exhibiting moderate (i.e., with fast phase to left or right, looking straight ahead) or severe (i.e., with fast phase to left when looking toward the right or to the right when looking toward the left) nystagmus, (3) Cerebellar Dysfunction—scoring at least 1.5 standard deviations below the mean for age-matched and sex-matched controls on at least two of the four balance and ataxia measures with eyes open [standing heel-to-toe on a line for 60 seconds (Romberg), walking 10 steps on a line, balancing on the right leg for 30 seconds and balancing on the left leg for 30 seconds] ⁵⁸, and (4) Altered Mental State or Mild Cognitive Impairment—scoring below 137 out of 144 points on the Dementia Rating Scale ^59,60.

HIV+ participants were assessed for Caine criteria: 20 did not meet any criteria (36%, Caine 0), 22 met one criterion (39%, Caine 1), 14 met two criteria (25%, Caine 2), and no one met three or four criteria. In the HIV+ Caine 1 subgroup, one patient met the dietary deficiency criterion, one had oculomotor disturbance, 12 had cerebellar dysfunction, and 8 had mild cognitive impairment. In the HIV+ Caine 2 subgroup, one patient had dietary deficiency and mild cognitive impairment criteria, 2 had dietary deficiency and cerebellar dysfunction criteria, and 11 met cerebellar dysfunction and mild cognitive impairment criteria.

Statistical analyses

Test scores were statistically corrected for age and education and standardized on the HIV− control group [mean and standard deviation: Z = 0 ± 1], allowing direct comparisons across groups and tests. Where higher raw scores indicated worse performance, scores were multiplied by −1, so that lower Z-scores always indicated worse performance. Theoretically- derived composite scores were then calculated as the mean of all Z-scores of tests comprising each of six cognitive and motor domains: attention/working memory, production, immediate memory, delayed memory, visuospatial construction and upper limb motor function. First, we conducted initial confirmatory analyses to attest that HIV+ individuals had mild to moderate deficits on cognitive (i.e., executive functions, episodic memory and visuospatial construction) and motor (i.e., upper limb motor speed and dexterity) measures compared with HIV− healthy control participants as observed in previous studies. Two sample t-tests tested group differences on the composite scores and individual neuropsychological test scores between HIV+ and HIV− controls. Second, one-tailed nonparametric Spearman rank-order correlations were conducted within the HIV+ group to test relations between age, depression (BDI-II score), and disease-related variables (length of time with HIV infection, log plasma viral load, and CD4 T-lymphocyte count) and neuropsychological composite scores, with the following directional hypotheses: poorer cognitive performance was predicted to be related to older age, a higher current BDI-II score, longer disease duration, higher viral load, and lower CD4 T-lymphocyte count. Furthermore, nonparametric Mann-Whitney U analyses explored subgroup differences within the HIV+ group to assess whether individuals with hepatitis C or AIDS demonstrated greater cognitive impairment than those without hepatitis C or AIDS. Third, analysis of variance (ANOVA) tested group differences on the composite scores among the HIV− controls and the three Caine subgroups in the HIV+ group (i.e., HIV+ individuals with no Caine criteria, those meeting one Caine criterion, and those meeting two or more Caine criteria), with follow-up post-hoc analyses examining two group comparisons for significant omnibus results. Finally, when correlations were observed between age, depression (BDI-II score), disease-related variables and specific cognitive and motor composite scores, supplementary analyses of covariance (ANCOVAs) were conducted to control for these additional factors potentially explaining the heterogeneity of HIV-associated neurocognitive impairment.

RESULTS

Comparisons between HIV− controls and HIV+ participants on cognitive and motor performance: confirmatory analyses

Neuropsychological test composite scores

Two sample t-tests conducted on the six cognitive and motor domains revealed that the HIV+ group scored lower than the HIV− group on all age-corrected and education-corrected cognitive and motor composite scores: attention/working memory, production, immediate episodic memory, delayed episodic memory, visuospatial construction and upper limb motor function (Table 2 and Figure 1A). HIV+ subgroups divided into individuals with and those without lifetime drug history did not differ on any composite score.

Table 2.

Age- and education-corrected Z-scores for neuropsychological composite scores and individual tests (mean and standard deviation)

	HIV+	HIV− controls	t	Cohen’s d	p value
Neuropsychological composite scores a
Attention/Working Memory	−1.20 (1.90)	0.00 (0.88)	−4.16	0.81	<.001
Production	−0.70 (0.91)	0.00 (0.85)	−4.10	0.79	<.001
Immediate Episodic Memory	−0.66 (1.08)	0.00 (0.82)	−3.52	0.69	<.001
Delayed Episodic Memory	−0.68 (1.09)	0.00 (0.82)	−3.66	0.71	<.001
Visuospatial Construction	−0.95 (1.83)	0.00 (1.01)	−3.32	0.64	.001
Upper Limb Motor Function	−0.79 (1.53)	−0.03 (0.83)	−3.14	0.62	.002
Neuropsychological individual tests b
Trails Ac	−0.80 (1.40)	0.00 (1.01)	−3.15	0.66	.002
Trails Bc	−1.69 (3.05)	0.00 (1.01)	−3.82	0.74	<.001
FAS total	−0.97 (1.24)	0.00 (1.02)	−4.41	0.85	<.001
Ruff Figural Fluency total	−0.43 (0.95)	0.00 (1.02)	−2.23	0.45	.03
Logical Memory-immediate	−0.79 (1.32)	0.00 (1.01)	−3.47	0.67	<.001
Rey-Osterrieth-immediate	−0.50 (1.28)	0.00 (1.01)	−2.26	0.43	.03
Logical Memory-delayed	−0.90 (1.38)	0.00 (1.02)	−3.83	0.74	<.001
Rey-Osterrieth-delayed	−0.43 (1.18)	0.00 (1.01)	−2.04	0.39	.04
Rey-Osterrieth-copy	−0.95 (1.83)	0.00 (1.01)	−3.32	0.64	.001
Grooved Pegboardc	−1.27 (2.37)	0.00 (0.93)	−3.57	0.71	<.001
Fine Finger Movement	−0.32 (0.96)	0.00 (0.93)	−1.69	0.34	ns

ns=not significant

^aCollected on a range of 52 and 56 HIV individuals and on a range of 51 and 53 controls

^bCollected on a range of 54 and 56 HIV individuals and on a range of 52 and 53 controls

^cTest scores were inverted to have higher scores reflect better performance

Figure 1.

Neuropsychological performance: (A) Age- and education-corrected Z-scores on the cognitive and motor composites in the HIV− controls (HIV−) and HIV seropositive participants (HIV+) (mean and standard error of the mean). * Significant difference compared with HIV− (p<.05), (B) Neuropsychological performance in the HIV− group and three Caine subgroups of HIV+ patients (mean and standard error of the mean). * Significant difference compared with HIV− (p<.05). ¥ Significant difference compared with HIV+ participants who did not meet any criteria (HIV+ CAINE 0, p<.05). δ Significant difference compared with HIV+ participants at risk of WE (one criterion, HIV+ CAINE 1, p<.05). Neuropsychological data collected on a range of 18 and 20 HIV+ Caine 0, 22 HIV+ Caine 1, and a range of 12 and 14 HIV+ Caine 2.

Scores on individual neuropsychological tests

Two sample t-tests conducted on the age- and education-corrected individual test scores revealed that HIV+ scored significantly lower than HIV− on ten cognitive and motor measures: Trails A and B, FAS, Ruff Figural Fluency, Logical Memory (immediate and delayed memory), Rey-Osterrieth Complex Figure (copy, immediate, and delayed memory), and Grooved Pegboard (Table 2). Groups did not differ on Fine Finger Movement.

Within-group analyses in the HIV+ group

Relations between age and disease-related variables and composite scores in the HIV+ group

Poorer upper limb motor score correlated with older age (N=52; rho=−0.37, p=.003). Higher BDI-II score correlated with lower production composite score (N=56; rho=−0.23, p=.047. Lower CD4 cell counts correlated with lower immediate memory composite score (N=52; rho=0.25, p=.034) and lower delayed memory composite score (N=52; rho=0.27, p=.025) (Figure 2). By contrast, longer disease duration and higher viral load did not correlate with cognitive or motor composite scores in the HIV+ group.

Figure 2.

In the HIV+ group, scatterplots for significant correlations between (A) age and upper limb motor composite score, (B) depression and production composite score, (C) CD4 cell count and immediate memory composite score, (D) CD4 cell count and delayed memory composite score.

Hepatitis C, AIDS, and composite scores

The three HIV+ Caine subgroups were comparable in terms of hepatitis C and AIDS distribution (Table 3).

Table 3.

Demographic and clinical characteristics of the three HIV+ Caine subgroups

	HIV+ Caine 0 (n=20)	HIV+ Caine 1 (n=22)	HIV+ Caine 2 (n=14)	p value
Age	49.7 ± 6.3	53.3 ± 9.1	55.2 ± 7.3	ns
Education (years)	14.8 ± 1.5	13.5 ± 3.2	12.6 ± 2.1	<.05 [HIV+ Caine 2 <HIV+ Caine 0]
Beck Depression Inventory-II (BDI-II) Score *	8.5 ± 7.0; 9.0	9.5 ± 8.8; 7.5	12.2 ± 9.5; 9.0	ns
Approximate Age of HIV Onset	33.6 ± 9.0	38.5 ± 9.0	34.9 ± 10.3	ns
Disease Duration (years)	16.0 ± 8.4	14.7 ± 8.2	20.4 ± 7.1	ns
Log Viral Loada	2.0 ± 1.0	2.1 ± 1.2	2.2 ± 1.3	ns
CD4 Count*b	701.7 ± 273.4; 711.0	582.7 ± 289.4; 538.5	527.3 ± 283.7; 519.0	ns
Lifetime Alcohol Consumption (kg)*	83.8 ± 84.3; 69.5	72.0 ± 75.8; 50.0	85.9 ± 80.8; 79.5	ns
Hepatitis Cc
Yes	2	8	6	.06, Χ2=5.65
No	18	13	8
AIDSc
Yes	8	14	6	.30, Χ2=2.38
No	11	8	8
Drug History
Yes	11	10	7	.83, Χ2=.38
No	9	12	7

^*Mean ± SD; median

ns=not significant

^aCollected on 17 HIV+ Caine 0, 21 HIV+ Caine 1, and 13 HIV+ Caine 2

^bCollected on 19 HIV+ Caine 0, 22 HIV+ Caine 2, and 13 HIV+ Caine 2

^cCollected on 55 HIV+ individuals

Individuals with and without seropositive hepatitis C status were comparable in age and education. HIV+ patients seropositive for hepatitis C achieved lower scores on the attention/working memory composite score (mean±stdev: −2.28±2.51) than those without hepatitis C (mean±stdev: −0.75±1.41) (U=188.00, p=.029). Other differences related to hepatitis C were not significant.

Individuals with and without history of AIDS were comparable in age and education, as well as lifetime alcohol consumption, HIV disease duration, HIV age of onset, log viral load, CD4 count, and hepatitis C distribution. A history of AIDS was not associated with lower scores on any composite.

Comparisons between HIV− and Caine-determined HIV+ subgroups on cognitive and motor performance

The HIV+ Caine subgroups did not differ on age, Beck Depression Inventory-II (BDI-II) score, age of HIV infection onset, length of HIV infection, log viral load, or CD4 count (Table 3). Groups did differ on years of education, with Caine 2 having fewer years of education than Caine 0. Two of the four HIV+ participants with remote history of alcohol abuse were in the HIV+ Caine 0 group, one in the HIV+ Caine 1 group, and one in the HIV+ Caine 2 group. The HIV+ individual with remote history of alcohol dependence was in the HIV+ Caine 1 group. The HIV+ Caine subgroups did not differ on lifetime alcohol consumption (kg). The three HIV+ Caine subgroups were comparable in terms of drug history distribution (Table 3).

One-way ANOVAs revealed significant differences among the HIV+ Caine subgroups and HIV− controls on all six composite scores: attention/working memory [F(3,104) = 13.34, p <.001], production [F(3,103) = 14.20, p <.001], immediate memory [F(3,103) = 9.33, p <.001], delayed memory [F(3,103) = 10.90, p <.001], visuospatial construction [F(3,105) = 8.17, p <.001], and upper limb motor function [F(3,99) = 13.03, p <.001] (Figure 1B). Follow-up analyses indicated that for attention/working memory, production, immediate and delayed memory, and upper limb motor function, HIV+ Caine 2 demonstrated the most severe deficits, scoring lower than HIV+ Caine 1, HIV+ Caine 0, and HIV− controls. Further, for visuospatial construction composite score, HIV+ Caine 2 and HIV+ Caine 1 performed at the same level, but both scored lower than HIV+ Caine 0 and HIV− controls. Finally, HIV+ Caine 1 scored lower than HIV+ Caine 0 and HIV− controls on the attention/working memory, production, visuospatial construction composite scores, and lower than HIV− controls on the immediate and delayed memory, and motor composite scores.

Supplementary analyses reinforced these results when taking account confounding variables: (1) when age was included as a covariate for the analysis of the upper limb motor composite score, the significant difference endured among the HIV+ Caine subgroups [F(2,48)=7.52, p =.001], with HIV+ Caine 2 demonstrating the most severe deficits, scoring lower than HIV+ Caine 1 and HIV+ Caine 0; (2) when CD4 T-lymphocyte count was included as a covariate for the analyses of the immediate and delayed memory composite scores, the significant difference endured among the HIV+ Caine subgroups (immediate memory: [F(2,48) = 4.46, p =.02] and delayed memory [F(2,48) = 5.69, p =.006]), with HIV+ Caine 2 demonstrating the most severe deficits, scoring lower than HIV+ Caine 1 and HIV+ Caine 0; (3) when BDI-II score was included as a covariate for the analysis of the production composite score, significant difference was still observed among the HIV+ Caine subgroups [F(2,52) = 11.20, p <.001], with a graded effect where HIV+ Caine 2 demonstrated the most severe deficits, scoring lower than both HIV+ Caine 1 and HIV + Caine 0 and where HIV+ Caine 1 scored lower than HIV+ Caine 0.

DISCUSSION

Subclinical signs of WE and neurocognitive heterogeneity in HIV

Conducted on the entire HIV+ group, analyses on neurocognitive functions confirmed impairment in cognitive and motor domains previously shown to be affected by HIV infection: attention/working memory, production (i.e., verbal and visual fluency), immediate and delayed episodic memory, and motor dexterity skills ^61,62. HIV+ patients also showed deficits in visuospatial abilities assessed by Rey–Osterrieth Complex Figure copy, which contrasts with other studies reporting relative sparing of this cognitive domain ⁶³. Previous investigations report that cognitive and motor impairments are observed in upwards of 50% of HIV-infected patients despite ART treatment ^5-7.

Classification by Caine criteria revealed that 25% of the individuals with HIV in this sample met two criteria of subclinical signs of WE, consistent with the assumption that WE may be underestimated in HIV infection. This study also shows that 39% of the HIV patients met one criterion and may, therefore, be considered at risk of subclinical WE. Significant heterogeneity in cognitive deficit severity ¹³ was unmasked by subgrouping HIV+ participants. Indeed, variability in HIV-related neurocognitive performance among the total group of HIV+ individuals was partially accounted for by the number of Caine criteria met. Comparison among HIV+ Caine subgroups revealed a graded effect, with those not meeting any criterion performing at control levels, those at risk for WE showing mild-to-moderate deficits on select cognitive and motor domains, and those with subclinical signs of WE showing the most severe deficits in each domain. This graded severity pattern of performance among Caine subgroups suggests that signs of subclinical WE can partly explain the heterogeneity in pattern and severity of HIV-related cognitive and motor impairments.

Age, hepatitis C, and neurocognitive heterogeneity in HIV

Older age was associated with poorer upper limb motor performance in HIV-infected individuals. Deficits in regional brain volumes of the caudate nucleus in HIV+ individuals is systematically reported in the literature and has been most consistently associated with cognitive impairment, particularly poor motor performance assessed with the Grooved Pegboard ^64-66. Older age and HIV infection have been reported to be independently correlated with smaller volumes of the caudate, but some evidence for premature aging of the caudate has been noted in HIV-infection ⁶⁷.

HIV-related cognitive disorder has also been related to comorbid conditions such as hepatitis C ^68-70. Compromised attention, memory, and psychomotor speed ^71-76, with less evidence for deficits in executive functioning ^77,78 and fine motor coordination ⁷⁹, have been reported in HCV mono-infection. A number of studies indicate that HIV/HCV co-infection results in higher level of cognitive impairment than HIV mono-infection ^80-86. A recent meta-analysis indicated that co-infected patients are more likely to be impaired in information processing speed than HIV mono-infected patients ⁷⁰. The current study consistently showed that co-infected individuals demonstrated worse performance on the attention/working memory composite score than those without HCV even if both groups were comparable with respect to age and education. In the present study, the attention/working memory composite score was calculated primarily from the mean of age- and education-corrected Z-scores from Trails A and B times. Trail making tests have been hypothesized to reflect a constellation of components processes including visual attention, sequencing and shifting, mental flexibility, working memory, and psychomotor speed ⁸⁷. Therefore, worse attention/working memory performance in HIV/HCV co-infected individuals might be explained by their more severe processing speed deficit than HIV mono-infected individuals.

Immune system restoration, memory compromise, and clinical implications

These results suggest that Caine criteria assessment could be useful in clinical settings for the detection of subclinical WE in HIV infected individuals ⁸⁸. Given the high prevalence of subclinical signs of WE in HIV+ participants and the report of thiamine deficiency in a large proportion of HIV-positive individuals in advanced stages of HIV infection as well as in clinically asymptomatic patients in the pre- and post-HAART era ^41-43, thiamine supplementation could be considered as adjunctive therapy ⁴⁴.

Similarly, the present study reported that low current CD4 cell count is associated with poor immediate and delayed episodic memory performance. Consistent relations between a low CD4 cell count in the past (i.e., CD4 nadir) and greater risk for HIV-related neurocognitive disorders have been reported in HIV+ individuals even after initiation of ART and immune recovery ^6,89. Nevertheless, the association of sustained prevalent impairment with worse current immune status (low current CD4 cell count) was also observed, suggesting that restoring immunocompetence increases the likelihood of neurocognitive recovery ⁹⁰. As suggested here, the cognitive domain of episodic memory may be particularly sensitive to a compromised immune system. Previous investigations have reported that micronutrient supplementation including thiamine can significantly improve CD4 cell counts in HIV-infected individuals on HAART ⁹¹. Thus, we speculate that thiamine supplementation is relevant for immune system restoration and prevention of memory decline in individuals with HIV infection.

Limitations and conclusion

An inherent limitation of this study is its retrospective nature restricting the choice to use prealbumin as a proxy for dietary deficiency or malnutrition. Serum proteins such as albumin and prealbumin have been widely used to determine nutritional status. Although prealbumin has been considered a more reliable indicator of acute changes in nutritional status than albumin ⁹², laboratory markers are not reliable by themselves. Rather, they should be supplemented by a comprehensive nutrition-focused physical examination including muscle wasting and subcutaneous fat loss in future research ⁹². Further, our analyses were based on a limited sample and require replication with a larger and more diverse group including HIV+ individuals meeting three or four Caine criteria. Nonetheless, the graded pattern of performance compromise among HIV+ Caine subgroups provides initial evidence that signs of subclinical WE contribute to the heterogeneity in HIV-related cognitive and motor impairment independent of other clinical variables such as age, depression level, disease-related variables (length of time with HIV infection, log plasma viral load, and CD4 T-lymphocyte count), hepatitis C, AIDS, alcohol and drug history. Even mild cognitive deficits challenge HIV disease management by compromising treatment adherence, daily life functioning, and quality of life ^93-95. We conclude that Caine criteria can aid in determining cognitive and motor impairment severity and potentially in identifying HIV+ individuals at risk for WE, promoting thiamine as a supplement treatment to mitigate further cognitive and motor decline.