Correlations between Computerized Battery Testing and a Memory Questionnaire for Identification of Neurocognitive Impairment in HIV Type 1-Infected Subjects on Stable Antiretroviral Therapy

Abstract

Neurocognitive impairment (NCI) remains prevalent in the highly active antiretroviral therapy (HAART) era. Memory function is commonly affected. There is a need for a rapid, but sensitive screening tool. This study compares the Prospective and Retrospective Memory Questionnaire (PRMQ) and a computerised battery cognitive assessment to establish if the questionnaire has potential as a rapid screening tool for HIV-associated NCI. Neurologically asymptomatic patients with an undetectable HIV viral load on stable HAART were eligible to participate. Asymptomatic NCI (aNCI) was defined as a performance score more than 1SD below the normative mean in at least two domains of the computerised test. Memory impairment (MI) was defined as a t-score more than 1 SD below the normative mean using the PRMQ. Forty-five subjects participated. The mean age was 48 years (SD 11), the mean CD4 count was 546 cells/μl (SD 271), and 84% were male. Of subjects, 14/45 (24%) had NCI and 15/45 (33%) had MI. Two subjects had both types of impairment. No significant association was found between the presence of aNCI and MI (p = 0.229, r = 0.18, 95% CI −1.2, 0.23). aNCI was statistically significantly associated with younger age (p = 0.38, r = 0.31, 95% CI −0.02, 0.001). MI was statistically significantly associated with the set-shifting cognitive domain of the computerized battery (p = 0.04, r = 0.326) and time elapsed since HIV diagnosis (p = 0.035, r = 0.316). High rates of asymptomatic NCI were observed in this cohort, especially in younger individuals. The memory questionnaire did not reliably identify HIV-associated NCI other than executive function deficits and based on our data should therefore not be used as a rapid screening tool for this purpose.

Introduction

T he effects of HIV-1 infection on the central nervous system were first described over 20 years ago.¹ Early reports described a progressive dementia frequently accompanied by motor and behavioral dysfunction. In some cases severe dementia with mutism or paraplegia ensued.² Mortality was high with a median reported survival of just 6 months following a diagnosis of AIDS dementia.³

Since the introduction of widespread highly active antiretroviral therapy (HAART), the incidence of overt HIV-associated dementia (HAD) has fallen. However neurocognitive impairment (NCI) remains prevalent and HIV-associated minor cognitive and motor disorders (MCMD) are frequently observed.^4,5

Recently a reclassification of HAD has been proposed, whereby the definition requires abnormalities in at least two cognitive areas significantly affecting daily living (including attention, concentration, mental slowing, impaired memory), plus an abnormality of motor function (slowed movements, incoordination) or behavior (personality change, irritability, emotional lability).^6
–8 The American Academy of Neurology criteria for HIV-associated neurological disorders requires acquired impairment in at least two areas of cognitive function, motor function, or behavior, confirmed by neuropsychological examination.⁹ These criteria also include a category for asymptomatic individuals who have (usually mild) cognitive impairment on formal neuropsychological testing, without observed abnormalities in everyday functioning.

The effect of HAART on HIV-associated NCI remains unclear. Many studies report an apparent improvement in neurocognitive functioning following initiation of HAART.^10

–14 However, Australian AIDS notification data between 1993 and 2000 showed a proportional increase in HAD when compared to other AIDS-defining illnesses (ADIs) suggesting that HAART has less impact on HAD than other ADIs.^15,16 Furthermore, an association with a diagnosis of NCI and a low nadir CD4 count (despite good immunological restoration with therapy) suggest the disorder is, at least partly, irreversible and deficits certainly can persist.^17,18 The proportion of asymptomatic individuals stable on HAART who have NCI on neuropsychological testing and the long-term clinical implications are not currently known.^9,18 For this reason, longitudinal research studies are required and clinical vigilance remains essential.

Identifying individuals with HIV-associated NCI is crucial as the condition is associated with a reduction in the ability to perform activities of daily living and comply with medication, reductions in quality of life, difficulty in obtaining employment, and an increased mortality.^19
–21 Therefore screening tools to enable easy identification of this impairment are of great importance.

Many diagnostic and screening tools to detect subjects with NCI have been investigated. Traditional neuropsychological assessments, although the gold standard, take several hours to perform and require a highly trained staff. Computerized battery tests are an alternative. Such batteries have shown high sensitivity in a variety of subtle cognitive deficits.^22
–24 One such tool (CogState) has demonstrated a high positive predictive value of 81% for detecting HIV-associated NCI when compared to formal neuropsychological assessment in individuals with advanced HIV disease (mean CD4 cell count 338 cells/μl) and with AIDS dementia (CD4 cell count 406 cells/μl).²⁵ This tool therefore has the potential for more widespread screening. Despite this, such computerized tests have limitations including the time required to complete testing and the need for full supervision throughout assessment, plus the purchase of equipment and software. There is a need for a more rapid, but sensitive screening tool.

Memory function is commonly affected in mild HIV-associated NCI.²⁶ This type of cognitive deficit may therefore represent a potential area for screening subjects. Typical memory problems experienced by individuals with HIV-associated NCI include working memory and encoding difficulties affecting learning efficiency, rather than the accelerated forgetting of information more typical of cortical dementias such as Alzheimer's disease.²⁷ It has also been demonstrated that prospective memory, that is episodic memory for remembering future intentions, can be particularly affected in HIV-associated NCI.²⁸ The Prospective and Retrospective Memory Questionnaire (PRMQ) is a short questionnaire by which an individual self-rates different types of memory error. Prospective memory deficits have previously been reported to adversely affect adherence to HIV medicines.²⁹ To our knowledge the PRMQ has not yet been evaluated as a screening tool for HIV-associated NCI, yet if memory (particularly prospective) is commonly affected in the condition, there exists a rationale for examining this area further.

The aims of this study, therefore, were to establish the rate of NCI in asymptomatic HIV-infected individuals on stable HAART using a computerized battery cognitive test and to compare results of the computerized battery test with the PRMQ memory questionnaire to establish if the latter has potential as a rapid screening tool for HIV-associated NCI.

Materials and Methods

Ethical approval was attained via the National Research Ethics Service, including provisions of Site Specific Assessment (SSA) and local Research and Development (R&D) approval. The study took place during April and May 2008 at the HIV Outpatient Clinic, St Mary's Hospital, London.

Subjects were eligible to participate if they were proficient in English, had HIV-1 infection, and were receiving continuous HAART with an undetectable plasma HIV RNA level for a minimum of 3 months prior to study entry. Exclusion criteria included any active neurological disease, hepatitis C infection, dementia, current use of recreational drugs, and alcohol abuse.

Participants provided informed consent. Demographic information was collected from the medical records including age, sex, nadir CD4 count, time since HIV diagnosis (years), current CD4 count, HIV viral load (VL), and current HAART. Current HAART was also scored for central nervous system penetration, using the Cerebrospinal fluid (CSF) Penetration Effectiveness (CPE) Rank Score devised by Letendre et al. ³⁰ This ranking system is based on available data of antiretroviral (ARV) chemical properties, their concentration in the CSF and/or their effectiveness in the CNS in clinical studies. A significant association between lower CPE ranks (less CNS penetrating ARVs) and higher CSF viral loads, even after adjustment for possible confounding factors, has been demonstrated.²⁹

Participants then performed two neuropsychological tests.

Cogstate test computerized battery (Melbourne, Australia)

This cognitive test is designed to measure performance in a number of cognitive domains. Software is installed on a computer and subjects complete the test while seated at a computer and are instructed how to input responses using two buttons on the keyboard. All subjects complete a practice session for familiarization under supervision by a staff member (approximately 20 min) and then an observed assessment of the same length. The test consists of eight tasks in card-game format. Reaction speeds and accuracy scores (including error rates) are recorded for each task (Table 1).

Table 1.

Description of CogState Computerized Test Battery

Task	Cognitive domain assessed	Task instruction	Primary outcome measure (unit)	Optimal test performance
Associate learning	Episodic, nonverbal learning	“Does the card pair match?”	Accuracy (arcsine proportion correct)	Higher score
Detection	Psychomotor function (speed)	“Has the card turned over?”	Speed (log₁₀ ms)	Lower score
Identification	Visual attention	“Is the card red?”	Speed (log₁₀ ms)	Lower score
Congruent reaction time	Attention	“Are the cards the same color?”	Speed (log₁₀ ms)	Lower score
Monitoring	Divided attention	“Has a card touched a white line?”	Speed (log₁₀ ms)	Lower score
One-card learning	Visual learning and memory	“Have you seen this card before in this task?”	Accuracy (arcsine proportion correct)	Higher score
One-back learning	Working memory	“Is the previous card the same?”	Accuracy (arcsine proportion correct)	Higher score
Set-shifting tasks	Executive function	“Is this a target card?”	Accuracy (total number of errors)	Lower score

The Prospective & Retrospective Memory Questionnaire (PRMQ)

This is a 16-item questionnaire by which individuals self-rate how often they make memory errors on a five-point scale. Scores can range between 16 (no memory errors) and 80 (multiple memory errors). The questionnaire can also be analyzed in sections relating to retrospective, prospective, self-cued, environmentally cued, and short- and long-term memory. It takes 10 min to complete, has high internal consistency, and normative population data exist.³¹

Data analysis

For each cognitive task a single primary outcome measure was selected to minimize experiment error rates. Each of these is drawn from a distribution that is either distributed normally or that can be corrected to normal through the use of appropriate mathematical transformation. Reaction times were log transformed and accuracy scores were arcsine root transformed. Scores in six tasks were compared to the age-stratified population data provided by the CogState manufacturer (n = 879 healthy adults for tasks 1–5, n = 154 healthy adults for task 6).³² Participants' scores were then ranked according to distance from normative mean and SD.

Total PRMQ scores were calculated. Scores were then transformed to t-scores using published normative population data (n = 551 adults) and ranked according to the mean and SD.³¹

Univariate linear regression modeling was used to investigate the association between computerized battery results and PRMQ score. p-values of <0.05 were considered statistically significant. SPSS (Chicago, IL) version 15.0 was used for analysis.

Definition of neurocognitive impairment and memory impairment (MI)

Asymptomatic neurocognitive impairment (aNCI) was defined as a score more than 1 SD below the normal population mean in at least two tasks using the computerized battery test. Memory impairment (MI) was defined as a t-score of more than 1 SD below the normative mean using the PRMQ.

Additional analysis

Demographic factors [including patient age, current CD4 count, nadir CD4 count, elapsed time since HIV diagnosis, whether receiving a nonnucleoside reverse transcriptase inhibitor (NNRTI) or boosted protease inhibitor (PI)-containing HAART regimen, and CSF Penetration Effectiveness Rank Score] were investigated for association with aNCI and MI using linear regression.

Results

Forty-five HIV-positive patients took part in the study and completed both assessments. All had been taking continuous HAART with a plasma HIV VL < 50 copies/ml for a minimum of 3 months. Of these, 20/45 (44%) were currently receiving an NNRTI-containing HAART regimen and 25/45 (56%) were receiving a boosted PI-containing HAART regimen. The mean age was 48 years (SD 11), the mean CD4 count was 546 cells/μl (SD 271), and 84% were male. Patient demographics are shown in Table 2.

Table 2.

Patient Demographics and PRMQ Results

Parameter	Mean (SD) unless otherwise stated
Number of participants	45
Age (years)	48 (11)
Male, n (%)	38 (84)
CD4⁺ count (cells/μl)	546 (271)
Nadir CD4⁺ count (cells/μl)	180 (129)
HIV RNA undetectable for >3 months, n (%)	45 (100)
Years since HIV diagnosis	12.2 (6)
Line of antiretroviral therapy, n (%)
First line therapy	23 (51)
Second line therapy	17 (38)
Third line or later therapy	5 (11)
Average CNS penetration effectiveness rank of ARVs (individual drug score 0, 0.5, or 1)	1.76 (0.68)

Asymptomatic neurocognitive impairment (aNCI)

Of participants, 14/45 (24%) had aNCI when scores were compared to population data. aNCI was statistically significantly associated with younger age of participant (p = 0.038, r = 0.31, 95% CI −0.02, 0.001). aNCI was present in 54, 27, 36, and 9% of subjects in ascending interquartile age groups (p-value for trend = 0.215). No statistically significant associations were observed between the presence of aNCI and current or nadir CD4 count, time since HIV diagnosis, CSF Penetration Effectiveness Rank Score, or type of HAART (NNRTI versus boosted PI) (p > 0.16 for all observations).

Memory impairment (MI)

Of individuals, 15/45 (33%) were identified with MI when compared to population data using the questionnaire. MI was significantly associated with time elapsed since HIV diagnosis (p = 0.035, r = 0.316, 95% CI 0.002, 0.048). No other statistically significant clinical associations were seen (p >0.17 for all observations).

Association of computerized battery scores with questionnaire scores

Two individuals were identified with impairment by both tests. No significant association was found between the presence of MI and NCI (p = 0.229, r = 0.18, 95% CI −1.2, 0.23). However, statistically significant associations were observed between the questionnaire score and the set-shifting cognitive domain of the computer test (a measure of executive function) (p = 0.040, r = 0.326, 95% CI 0.02, 0.89). Other associations are shown in Table 3.

Table 3.

Association between Computerized Battery Task Scores and Questionnaire Score

		95% confidence intervals
Computer task	p-value	Lower bound	Upper bound	r-value
Associate learning	0.980	−19.4	18.96	0.004
Detection	0.994	−26.99	27.19	0.001
Identification	0.329	−20.44	59.73	0.149
Congruent reaction time	0.331	−18.87	54.84	0.148
Monitoring	0.952	−27.07	25.51	0.009
One-card learning	0.909	−21.35	19.05	0.018
One-back learning	0.670	−10.93	16.83	0.065
Set-shifting	0.040	0.022	0.89	0.326

Discussion

In our study, a high proportion of HIV-infected adults with undetectable plasma HIV replication on stable antiretroviral therapy had aNCI when examined using a validated computerized battery assessment. The clinical significance of this neurological deficit has not yet been fully elucidated. Earlier studies suggest that clinical progress can be variable, with some individuals exhibiting later cognitive decline and others remaining stable or having improved scores on longitudinal cognitive assessments.^9,19 In the absence of established copathologies, such as concurrent neurological disease or hepatitis C virus, the factors that cause ongoing neurocognitive decline in some individuals, despite adherence to HAART, remain unclear. Poorer CNS penetration of some antiretroviral agents may allow ongoing replication of HIV within the CNS compartment resulting in cerebral inflammation and damage.³⁰ Alternative explanations include direct damage to cerebrovascular endothelium from antiretroviral agents and disruption of the blood–brain barrier permeability.³³

Interestingly, NCI was associated with younger age in this cohort. Possible reasons for this finding include the fact that the younger, less mature brain may be more susceptible to the damaging effects of HIV virus than in older patients, a theory supported by the observation of more frequent and fulminant CNS inflammation and disease in HIV-infected children than adults.³⁴ Alternatively, these age-related differences observed may reflect differing socioeconomic and education levels between our cohort and normative controls (for whom detailed demographic data are not available). Such marked age-related differences were not, however, observed for the memory questionnaire, and if such dramatic bias exists, it is reasonable to expect it not to be limited to one cognitive assessment.

No statistically significant associations between the PRMQ score and computer assessment were observed and therefore the questionnaire should not be used as a screening tool for HIV-associated NCI. An association between PRMQ score and the set-shifting task of the computerized test was observed (p = 0.04), suggesting the questionnaire is able to capture part of the executive function deterioration in HIV-associated NCI, but not the decline in other cognitive domains. Executive function, which enables an individual to plan, sequence, initiate, and sustain behavior with the ability to incorporate feedback and make adjustments, can be caused by direct HIV damage to frontostriatal regions.³⁵ Prospective memory is a cognitive function also largely dependent on prefrontostriatal circuits and involves forming, monitoring, and executing future intentions. Deficits in prospective memory that correlate with validated measures of executive function have previously been demonstrated in HIV-positive individuals.²⁸

This study has several limitations. A relatively small sample size was recruited. A computerized battery test was utilized as the gold standard to define NCI. Previous studies have described an inferior sensitivity for detecting HIV-associated NCI when utilizing a computer battery compared to formal neuropsychometric testing, raising the possibility for misclassification of subjects.^22
–24 The population examined in this study was in better health and had higher mean CD4⁺ lymphocyte counts than subjects participating in the validation study of the CogState battery for the diagnosis of HIV-associated NCI.²⁵ Finally, the PRMQ is a subjective self-assessment and therefore prone to innate bias in the event of an individual with poor insight. For this reason proxy-rated (e.g., spouse or relative) data have been published and differences in self-score and proxy-rated score were observed.³⁶ It was beyond the scope of this study for a proxy to also complete the PRMQ.

In summary, high rates of aNCI were observed in this clinic cohort, especially in younger individuals. The memory questionnaire did not reliably identify HIV-associated NCI and based on our data should not be used as a rapid screening tool for this purpose.

Footnotes

Acknowledgments

Computerized cognitive battery assessments were performed using equipment from Cogstate^TM Limited (Melbourne, Australia).

Disclosure Statement

No competing financial interests exist.

References

, Rota

, Schooley

. et al. Isolation of HTLV-III from cerebrospinal fluid and neural tissues of patients with neurologic syndromes related to the acquired immunodeficiency syndrome. N Engl J Med, 1985; 313,24:1493–1497.

Navia

, Jordon

, Price

. The AIDS dementia complex: I. Clinical features. Ann Neurol, 1986; 19,6:517–524.

McArthur

, Hoover

et al. Dementia in AIDS patients: Incidence and risk factors. Multicenter AIDS Cohort Study. Neurology, 1993; 43,11:2245–2252.

Bhaskaran

, Mussini

. et al. Changes in the incidence and predictors of human immunodeficiency virus-associated dementia in the era of highly active antiretroviral therapy. Ann Neurol, 2008; 63,2:213–221.

McArthur

, McDermott

, McClernon

. et al. Attenuated central nervous system infection in advanced HIV/AIDS with combination antiretroviral therapy. Arch Neurol, 2004; 61,11:1687–1696.

Cherner

, Cysique

, Heaton

. et al. Neuropathologic confirmation of definitional criteria for human immunodeficiency virus–associated neurocognitive disorders. J NeuroVirol, 2007; 13:23–28.

Heaton

, Grant

et al. Clinical confirmation of the American Academy of Neurology algorithm for HIV-1-associated cognitive/motor disorder. American Academy of Neurology. The Dana Consortium on Therapy for HIV Dementia and Related Cognitive Disorders. Neurology, 1996; 47:1247–1253.

Janssen

, Cornblath

, Epstein

. et al. Nomenclature and research case definitions for neurological manifestations of human immunodeficiency virus type-1 (HIV-1) infection. Neurology, 1991; 41:778–785.

Antinori

, Arendt

, Becker

. et al. Updated research nosology for HIV-associated neurocognitive disorders. Neurology, 2007; 69,18:1789–1799.

10.

Tozzi

, Balestra

, Galgani

, Narciso

, Sampaolesi

, Antinori

. et al. Changes in neurocognitive performance in a cohort of patients treated with HAART for 3 years. J Acquir Immune Defic Syndr, 2001; 28,1:19–27.

11.

Robertson

, Robertson

, Ford

, Watson

, Fiscus

, Harp

. et al. Highly active antiretroviral therapy improves neurocognitive functioning. J Acquir Immune Defic Syndr, 2004; 36,1:562–566.

12.

Cohen

, Boland

, Paul

, Tashima

, Schoenbaum

. et al. Neurocognitive performance enhanced by highly active antiretroviral therapy in HIV-infected women. AIDS, 2001; 15,3:341–345.

13.

Tozzi

, Balestra

, Galgani

, Narciso

, Ferri

, Sebastiani

. et al. Positive and sustained effects of highly active antiretroviral therapy on HIV-1-associated neurocognitive impairment. AIDS, 1999; 13,14:1889–1897.

14.

Robertson

, Smurzynski

, Parsons

. et al. The prevalence and incidence of neurocognitive impairment in the HAART era. AIDS, 2007; 21,14:1915–1921.

15.

Dore

, Corrella

, Lia

. et al. Changes to AIDS dementia complex in the era of highly active antiretroviral therapy. AIDS, 1999; 13:1249–1253

16.

Dore

, McDonalda

, Corrella

. et al. Marked improvement in survival following AIDS dementia complex in the era of highly active antiretroviral therapy. AIDS, 2003; 17,10:1539–1545.

17.

Munoz-Moreno

, Fumaz

, Ferrer

. et al. Nadir CD4 cell count predicts neurocognitive impairment in HIV-infected patients. AIDS Res Hum Retroviruses, 2008; 24,10:1301–1307.

18.

Cole

, Margolick

, Cox

. et al. Longitudinally preserved psychomotor performance in long-term asymptomatic HIV-infected individuals. Neurology, 2007; 69,24:2213–2220.

19.

Tozzi

, Balestra

, Bellagamba

. et al. Persistence of neuropsychologic deficits despite long-term highly active antiretroviral therapy in patients with HIV-related neurocognitive impairment prevalence and risk factors. J Acquir Immune Defic Syndr, 2007; 45:174–182.

20.

Albert

, Weber

, Todak

. An observed performance test of medication management ability in HIV: Relation to neuropsychological status and adherence outcomes. AIDS Behav, 1999; 3:121–128.

21.

Farinpour

, Miller

, Satz

et al.

Psychosocial risk factors of HIV morbidity, mortality: Findings from the Multicenter AIDS Cohort Study (MACS)

J Clin Exp Neuropsychol, 2003; 25,5:654–670.

22.

Collie

, Darby

, Maruff

. Computerised cognitive assessment of athletes with sports related head injury. Br J Sports Med, 2001; 35:297–302.

23.

Collie

, Darby

, Falleti

et al. Determining the extent of cognitive change after coronary surgery: A review of statistical procedures. Ann Thorac Surg, 2002; 73:2005–2011.

24.

Mollica

, Maruff

. et al. Repeated assessment of cognition in children and the measurement of performance change. Child Neuropsychol, 2005; 11:303–310.

25.

Cysique

, Maruff

, Darby

, Brew

. The assessment of cognitive function in advanced HIV-1 infection and AIDS dementia complex using a new computerised cognitive test battery. Arch Clin Neuropsychol, 2006; 21,2:185–194.

26.

Wilkie

, Goodkin

, van Zuilen

, Tyll

. et al. Cognitive effects of HIV-1 infection. CNS Spectr, 2000; 5,5:33–51.

27.

Stout

, Salmon

et al. Decline in working memory associated with HIV infection. HNRC Group. Psychol Med, 1995; 25,6:1221–1232.

28.

Carey

, Woods

. et al. Prospective memory in HIV-1 infection. J Clin Exp Neuropsychol, 2006; 28,4:536–548.

29.

Woods

S P

, Iudicello

, Moran

. et al. HIV-associated prospective memory impairment increases risk of dependence in everyday functioning. Neuropsychology, 2008; 22,1:110–117.

30.

Letendre

, Marquie-Beck

. et al. Validation of the CNS Penetration-Effectiveness rank for quantifying antiretroviral penetration into the central nervous system. Arch Neurol, 2008; 65,1:65–70.

31.

Crawford

, Smith

, Maylor

et al. The Prospective, Retrospective Memory Questionnaire (PRMQ): Normative data and latent structure in a large non-clinical sample. Memory, 2003; 11,3:261–275.

32.

Maruff

. Validity of the CogState computerized test battery: Sensitivity to mild traumatic brain injury, mild cognitive impairment, schizophrenia, and AIDS dementia complex. Arch Clin Neuropsychol(in press).

33.

Langford

. et al. Changing patterns in the neuropathogenesis of HIV during the HAART era. Brain Pathol, 2003; 13,2:195–210.

34.

Sharer

, Cho

. Neuropathology of HIV infection: Adults versus children. Prog AIDS Pathol, 1989; 1:131–141.

35.

Melrose

, Tinaz

, Castelo

et al. Compromised fronto-striatal functioning in HIV: An fMRI investigation of semantic event sequencing. Behav Brain Res, 2008; 188,2:337–347.

36.

Crawford

, Henry

, Ward

et al. The Prospective, Retrospective Memory Questionnaire (PRMQ): Latent structure, normative data and discrepancy analysis for proxy-ratings. Br J Clin Psychol, 2006; 45:83–104.