Journal of NeuroVirology, 16: 189–202, 2010

� 2010 Journal of NeuroVirology

ISSN 1355-0284 print / 1538-2443 online

DOI: 10.3109/13550284.2010.489597

Review

NeuroAIDS in Africa

Kevin Robertson,1 Jeff Liner,1 James Hakim,2 Jean-Louis Sankalé,3 Igor Grant,4 Scott Letendre,5

David Clifford,6 Amadou Gallo Diop,7 Assan Jaye,8 Georgette Kanmogne,9 Alfred Njamnshi,10

T. Dianne Langford,11 Tufa Gemechu Weyessa,12 Charles Wood,13 Mwanza Banda,14 Mina Hosseinipour,15

Ned Sacktor,16 Noeline Nakasuja,17 Paul Bangirana,17 Robert Paul,18 John Joska,19 Joseph Wong,20

Michael Boivin,21 Penny Holding,22 Betsy Kammerer,23 Annelies Van Rie,24 Prudence Ive,25 Avindra Nath,26

Kathy Lawler,27 Clement Adebamowo,28 Walter Royal III,29 and Jeymohan Joseph,30 on behalf of theNeuroAIDS in Africa Conference Participants

1University of North Carolina, Chapel Hill, North Carolina, USA; 2University of Zimbabwe, Harare, Zimbabwe;3Globomics, LLC, Decatur, Georgia, USA; 4University of California, San Diego, La Jolla, California, USA; 5University ofCalifornia, San Diego, San Diego, California, USA; 6Washington University School of Medicine, St. Louis, Missouri, USA;7University of Dakar, Dakar, Senegal; 8Medical Research Council Laboratories, Gambia Unit, Banjul, The Gambia;9University of Nebraska Medical Center, Omaha, Nebraska, USA; 10University of Yaoundé, Yaoundé, Cameroon;11Temple University School of Medicine, Philadelphia, Pennsylvania, USA; 12Addis Ababa University, Addis Ababa,Ethiopia; 13Nebraska Center for Virology, University of Nebraska–Lincoln, Lincoln, Nebraska, USA; 14UniversityTeaching Hospital, Lusaka, Zambia; 15UNC Project, Lilongwe, Lilongwe, Malawi; 16Johns Hopkins University, JohnsHopkins Bayview Medical Center, Baltimore, Maryland, USA; 17Makerere University, Kampala, Uganda; 18University ofMissouri, St. Louis, Chesterfield, Missouri, USA; 19University of Cape Town, Cape Town, South Africa; 20University ofCalifornia, San Francisco, San Francisco, California, USA; 21Michigan State University, East Lansing, Michigan, USA;22Case Western Reserve University, Cleveland, Ohio, USA; 23Children’s Hospital, Boston/Harvard Medical School, Newton,Massachusetts, USA; 24University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA; 25Clinical HIVResearch Unit, Helen Joseph Hospital, Gauteng, South Africa; 26Johns Hopkins University, Baltimore, Maryland, USA;27University of Pennsylvania, Philadelphia, Pennsylvania, USA; 28Office of Strategic Information and Research, Institute ofHuman Virology, Abuja, Nigeria; 29University of Maryland School of Medicine, Baltimore, Maryland, USA; and 30HIVPathogenesis, Neuropsychiatry and Treatment Branch, Center for Mental Health Research on AIDS, National Institute ofMental Health, National Institutes of Health, Bethesda, Maryland, USA

In July 2009, the Center for Mental Health Research on AIDS at the NationalInstitute of Mental Health organized and supported the meeting “NeuroAIDSin Africa.” This meeting was held in Cape Town, South Africa, and wasaffiliated with the 5th IAS Conference on HIV Pathogenesis, Treatment andPrevention. Presentations began with an overview of the epidemiology of HIVin sub-Saharan Africa, the molecular epidemiology of HIV, HIV-associatedneurocognitive disorders (HANDs), and HAND treatment. These introductorytalks were followed by presentations on HAND research and clinical care inBotswana, Cameroon, Ethiopia, The Gambia, Kenya, Malawi, Nigeria, Sene-gal, South Africa, Uganda, and Zambia. Topics discussed included bestpractices for assessing neurocognitive disorders, patterns of central nervoussystem (CNS) involvement in the region, subtype-associated risk for HAND,pediatric HIV assessments and neurodevelopment, HIV-associated CNSopportunistic infections and immune reconstitution syndrome, the evolvingchanges in treatment implementation, and various opportunities and strat-egies for NeuroAIDS research and capacity building in the region. Journal ofNeuroVirology (2010) 16, 189–202.

Keywords: HIV; neurodevelopment; neurology; neuropsychology; subtypes

Address correspondence to Kevin Robertson, PhD, Department of Neurology, University of North Carolina, 170 Manning Drive, ChapelHill, NC 27599-7025, USA. E-mail: kevinr@neurology.unc.edu

From the NeuroAIDS in Africa Conference, Cape Town, South Africa, July 23–24, 2009.Received 14 April 2010; accepted 21 April 2010

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

Introduction

In July 2009, the Center for Mental Health Researchon AIDS at the National Institute of Mental Healthorganized and supported the meeting “NeuroAIDS inAfrica.” This meeting was held in Cape Town, SouthAfrica, and was affiliated with the 5th IAS Confer-ence on HIV Pathogenesis, Treatment and Preven-tion. The goals of the meeting included:

1. Reviewing findings of neurocognitive impairmentdirectlyattributable toHIVatdifferentAfricansites

2. Determining the impact of comorbid factors suchas coinfections (tuberculosis [TB], malaria, syph-ilis, toxoplasmosis, cryptococcal meningitis, etc.),psychiatric illness, or nutritional deficiencies onneurocognitive function

3. Discussing relationships between neurocognitiveimpairmentandHIVdiseasevariables (currentandnadir CD4, viral load, biomarkers of HIV neuro-pathogenesis, neuroimaging, and brain pathology)

4. Evaluating differences in HIV-1 subtype-associated neuropathology

5. Evaluating the relationships between HIV-associated neurocognitive impairment and abil-ity to perform activities of daily living withindifferent African populations

In addition to these goals, conference participantsdiscussed the challenges to NeuroAIDS research inAfrica as related to infrastructure, access to neuro-logical and neuropsychological expertise, regionalHIV genetic variability, host population genetics,differences in comorbidities and coinfections, avail-ability of appropriate normative/control data for neu-ropsychological assessments, differences in languageand culture, and differences in methods used acrossstudies. Efforts were made at this meeting to furtherestablish a more standardized approach to neurocog-nitive assessments across international studies, andexpand on best practices for areas of concern such asdeveloping appropriate normal data, identifying andcategorizing levels of comorbidities, training localinvestigators, and assessing the relative impact ofneurocognitive impairment on everyday functioning,in order to improve future NeuroAIDS research inAfrica and other resource-limited settings.

Overview of epidemiology and neurologicaldisease

Epidemiology of HIV/CNS disease in sub-SaharanAfricaDr. Hakim gave an overview of the epidemiology ofHIV and central nervous system (CNS) disease insub-Saharan Africa. According to the 2008 UNAIDSReport on the Global AIDS Epidemic, approximately

22 million (67%) of the estimated 33 million adultsand children living with HIV globally, reside in sub-Saharan Africa (WHO/UNAIDS, 2008). Of the esti-mated 2.7 million adults and children newly infectedwith HIV worldwide in 2007, 1.9 million new infec-tions occurred in sub-Saharan Africa as well as1.5 million of the estimated 2 million adult and childdeaths due to AIDS. Although sub-Saharan Africa iscertainly still in dire straits, the 2008 UNAIDS reportoffered some signs of improvement. The 2008UNAIDS Report on the Global AIDS Epidemic con-firmed progress in the response to AIDS, stating thatgovernments are acting on their promises to scale updiagnosis, prognosis, and care for HIV-infected indi-viduals, that financing for HIV treatment in low-to-middle-income countries has increased 6-fold andthat additional countries (Lesotha, Namibia, Repub-lic of South Africa [RSA], and Swaziland, for exam-ple) have had improvements in the epidemic (WHO/UNAIDS, 2008). CNS opportunistic infections (OIs)such as tuberculosis (meningitis, tuberculoma), cryp-tococcal meningitis, toxoplasmosis, neurocyster-cercosis, neurosyphilis, primary CNS lymphoma(Epstein-Barr virus), progressive multifocal leuko-encephalopathy (JC virus), cytomegalovirus, as wellas stroke remain the most commonly described HIV-associated CNS diseases. Cognitive dysfunction suchas dementia and psychological perturbations un-related to OIs above has been less commonlydescribed. Dr. Hakim acknowledged problems withour understanding of HIV-associated CNS disease insub-Saharan Africa, largely due to limited data onthese conditions, in part due to the paucity of neurol-ogists managing HIV-infected patients. In addition,confidently confirming diagnoses is difficult due tocomorbid conditions, limited imaging facilities, andlimited diagnostic capabilities (microbiological,immunological, virologic, serologic, histologic, etc.).

Molecular epidemiology of HIV in AfricaDr. Sankalé presented a review of the molecularepidemiology of HIV. HIV exhibits tremendousgenetic diversity, which has been attributed to thelack of proofreading of reverse transcriptase, rapidturnover of virions, constant recombination ofviruses, RNA polymerase II errors, and selectiveimmunologic pressures. The HIV-1 genome encodesnine open reading frames: gag, pol, env, tat, rev, nef,vpr, vif, and vpu. Major variations in these genesallow for HIV to be classified at a number of levels,beginning with distinctions between HIV-1 andHIV-2. HIV-1 consists of several major groups thatinclude group M (Major), the source of the vastmajority of HIV infections worldwide, as well asgroups O (Outlier) and N (non-M, non-O). Withingroup M, HIV can be further distinguished intoclades or subtypes, such as clades A–D, F–H, J, K,circulating recombinant forms (CRFs) A/E and A/G,many other circulating recombinant forms (whichnow number over 40) (LANL, 2010) as well as unique

NeuroAIDS in Africa 2009

190 K Robertson et al

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

recombinant forms. The distribution of HIV-1 sub-types and recombinants varies by region of the world,by country, and even within a country such asNigeria. Within Africa, CRF02_AG accounts for thebulk of infections in West Africa, subtypes A and Dare common in East Africa, and subtype C predomi-nates in Southern Africa as well as Northeastern areassuch as Ethiopia (Hemelaar et al, 2006). Differencesin disease progression and resistance profiles werealso discussed. A study of registered female sex work-ers in Dakar, Senegal, determined that HIV-2 is lessinfectious, has a much lower mother-to-child trans-mission rate, and a much longer progression time toAIDS than HIV-1. Studies investigating differences inHIV-1 subtypes found that subtype A has a longerAIDS-free survival than non-A (Kanki et al, 1999);subtype D has a shorter time to death and AIDS thansubtype A, C, or recombinant forms (Vasan et al,2006); and in subtype C, CCR5/non–syncytia-formingviruses are almost always the dominant form even inlate stages of infection (Tscherning et al, 1998). HIV-2and HIV-1 type O possess a primary resistance tonon-nucleoside reverse transcriptase inhibitors(NNRTIs) (Descamps et al, 1997; Ren et al, 2002).Naturally occurring minor mutations in the proteasesequences of many non-B subtypes such as subtype F(Poveda et al, 2008) can also confer resistance. Fewnaturally occurring polymorphisms at resistancepositions in reverse transcriptase in type M viruseshave been discovered.

There is a potential role for clade differences invarying neuropathogenicity. How much of a roleclades play in the neuropathogenicity of HIV will bedifficult to determine because treatment availabilityand practices vary from place to place and measuresof cognitive and functional performance are inher-ently subject to some degree of cultural influence.

HIV-associated neurocognitive disordersDr. Grant gave a summary of HIV-associated neuro-cognitive disorders (HANDs), which remain to becommon problems in HIV-infected patients despitethe success of combination antiretrovirals at improv-ing overall health and prolonging survival. Thenosology of HAND, revised in 2007 (Antinori et al,2007), now includes three classifications: asymptom-atic neuropsychological impairment (abnormality intwo or more cognitive abilities), mild neurocognitivedisorder (cognitive impairment with mild functionalimpairment), and HIV-associated dementia (markedcognitive impairment with marked functionalimpairment). Neuropsychological (NP) deficits inmeasures of attention, learning, verbal, motor, mem-ory, psychomotor, sensory, and abstraction are themost common among patients with global NPimpairment. HAND prevalence rates still rangefrom 30% to 50% and the cognitive profile, or areasof impairment, has not changed significantly. Onethird of patients have magnetic resonance imaging(MRI) evidence of white matter abnormality and

approximately 34% (perhaps up to 60%) havedetectable HIV RNA in the cerebrospinal fluid (CSF).

The possible mechanisms for neurotoxicity inHIV-1 infection include direct damage to neuronscaused by shed viral proteins, but not infection ofneurons themselves, as well as indirect damagecaused by neurotoxic molecules released by activatedand/or infected astrocytes, macrophages, and micro-glia. HIV infection preferentially causes damage tothe basal ganglia and white matter, but loss ofdendritic complexity and the degree and synaptoden-dritic injury are also strongly associated with pro-gressive neurocognitive impairment. Cofactors inHIV-associated neurocognitive complications suchas drug abuse (e.g., methamphetamine), coinfections(e.g., hepatitis C), aging, immune reconstitution syn-drome, neurotoxic treatments, and other backgroundfactors (e.g., head injury) tend to increase the likeli-hood and severity of cognitive impairment in HIV-infected individuals. On the contrary, patients with ahistory of good virologic control and no history ofsevere immunologic suppression have a reduced riskof neurocognitive impairment. In terms of everydayfunction, neurocognitive impairment is associatedwith higher rates of unemployment, difficulty withcomplex tasks such as driving, decreased ability toadhere to antiretroviral treatment, and increased mor-tality several years out from initial visit. Treatmentconsiderations such as starting antiretrovirals (ARVs)early and including ARVs with high CSF penetrationto reduce CSF viral load (VL) and prevent or attenuateHAND are promising, but still need further research.

Endeavors to establish HAND prevalence ratesacross countries and regions of the world have yieldedvaryingestimates, ranging from12%to56%.Empiricalquestions regarding the nature of these different rates(i.e., HIV clade differences, host responses and genet-ics, differences in cofactors such as TB, malaria,hepatitis, nutrition, etc.) are impeded by the lack ofcomparable and reliable diagnostic methods usedacross sites.

Treatment of HAND—penetration of antiretroviralsinto the CNSAs mentioned above, HAND is still a global problem.Dr. Letendre’s talk focused on the possibility that thelimited distribution of drugs into the brain could becontributing to this problem. Factors limiting drugdistribution into the CNS are based in the restrictivenature of the blood-brain barrier (tight junctions andastrocytic foot processes). As a result, diffusion ofdrugs across the blood-brain barrier is determined inpart by characteristics such as protein binding, molec-ular weight, lipophilicity, ionization, and molecularpumps (e.g., P-glycoprotein and organic anion trans-porters). Drugs within the different classes of antire-trovirals differ in their CSF pharmacokinetics. Forexample, the CSF concentration of the nucleo-side reverse transcriptase inhibitor (NRTI) abacavirreaches approximately 36% of plasma concentrations

NeuroAIDS in Africa 2009

K Robertson et al 191

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

whereas tenofovir reaches approximately 5%. The PIsdarunavir and lopinavir differ in their CSF-to-plasmaratios (1.4% versus 0.23%) as do NNRTIs such asnevirapine and efavirenz (29–63% versus 0.5%).These ratios, however, may not reflect efficacy inthe nervous system as well as metrics that estimatethe relationship between drug concentrations in thenervous system and the concentrations that inhibitHIV replication (e.g., the CSF-to-50% inhibitory con-centration ratio). When antiretrovirals improve ormaintain cognition, they probably do so by reducingHIV replication in the CNS and improving cellularimmunity. These effects are probably necessary forimproving or maintaining cognitive health but maynot always be sufficient because neurotoxins such asHIV proteins, proinflammatory cytokines, oxidativestress, and excitotoxins may persist even when HIVlevels are reduced below the quantitation limit ofcommercial assays. In addition, immune recoverymay not result in recovery of neurotrophic or neuro-protective molecules. The predominant cellularsource of HIV found in the nervous system maydetermine the effectiveness of treatment on CNScomplications. When CD4+ cell counts are higherand cognition is normal, HIV in the nervous systemmay predominantly derive from migrating cells, suchas lymphocytes. In this situation, controlling HIVin migrating cells is relatively more important andpenetration of antiretrovirals into the nervous systemmay be relatively less important. However, whenCD4+ cell counts are lower or cognition is impaired,HIV in the nervous system may derive predominan-tly from resident cells, such as macrophages andmicroglia. In this situation, controlling HIV mayrequire antiretrovirals that penetrate into the nervoussystem in therapeutic concentrations. Additionalanti-inflammatory, antioxidant, or neuroprotectivetherapies may also be warranted.

Dr. Letendre has devised a method of estimatingthe penetration and effectiveness of antiretrovirals inthe nervous system. Ranks are assigned to each drugand these ranks are summed to yield a value for agiven regimen, with higher scores indicating greaterpenetration. In most analyses to date, higher pene-tration ranks are associated with lower CSF viralloads. Other analyses have also identified that esti-mates of higher penetration are associated with betterneurocognitive performance, better mood, and bettersurvival.

HAND research and clinical care in Africa

Assessment of neurological complications in HIV-2infection in West AfricaHIV-2 is uniquely prevalent in West Africa andalthough the clinical course is more indolent thanHIV-1, pathological studies of brain tissue suggestsimilar pathology to HIV-1 by time of death. Clinicalneurological manifestations of HIV-2, however, are

largely unstudied. Dr. Clifford, Dr. Jaye, and Dr. Dioppresented their aims for several cohorts underway todevelop the infrastructure for neurological evalua-tion in West Africa in order to learn more about theneurological complications of HIV-2. In addition todevelopment of infrastructure and sharing of exper-tise, the aims of the cohorts include (1) studying theprevalence and characteristics of neurocognitiveimpairment, peripheral neuropathy, and myelopathyin HIV-2-infected patients through the course ofdisease; (2) collecting preliminary neuroimagingand CSF samples from clinically characterizedHIV-2–affected patients; (3) investigating the associ-ation of aging on HIV-2 neurological manifestations;and (4) determining the spectrum of neurologicopportunistic infections associated with HIV-2. Thetwo cohorts discussed are based in Caio, The Gambia,and Dakar, Senegal.

NeuroAIDS in CameroonDr. Kanmogne andDr. Njamnshi presented the prelim-inary results of the Cameroon neuroAIDS project, acase-control study aimed at investigating the epidemi-ology of HIV-associated neurocognitive disorders inCameroon. From February 2008 to May 2009, theCameroon team enrolled 122 participants (85 HIV-seropositive cases and 37 HIV-seronegative controls).In this pilot study, both the Halstead Category Test(HCT) andWisconsin Card Sorting Test-64 (WCST-64)showedsignificant impairmentof theexecutive/frontallobe function inHIV-infected individuals, compared toseronegative controls. Data further showed increasedimpairment in HIV-seropositive individuals withadvanced immune deficiency.

As a baseline for a series of studies on HIV-associated dementia (HAD), Dr. Njamnshi and collea-gues assessed the usefulness of the recently developedInternational HIV Dementia Scale (IHDS) as ascreening tool for HAD or HIV-associated cognitiveimpairment in HIV-positive adults in Yaoundé,Cameroon. In this case-control study, HIV-positiveadults followed up in an HIV outpatient clinic werematched to HIV-negative subjects for age and sex andscreened using the IHDS. In particular, the IHDSmeantotal score, alternating hand sequence, and 4-wordrecallmeasureswere useful in distinguishing betweenHIV patients at risk for dementia and normal healthysubjects. The IHDS fingertapping subscore was notsignificantly different between the two groups. Usingthe IHDS screening tool, investigators estimated a21.1% prevalence rate of HAND, and suggested thatfuture research on HAND in Cameroon should strate-gically start with the IHDS as a screening tool.Dr. Njamnshi and colleagues also investigated riskfactors for HAND such as low hemoglobin levels, lowbodymass index (BMI),more constitutional symptomsbefore AIDS, age older than 68 at AIDS onset, intrave-nous (IV) drug use, low CD4 counts, high plasma viralloads, and antiretroviral (ARV) drug naivety. Theyfound that low CD4 count (<200/ml), advanced clinical

NeuroAIDS in Africa 2009

192 K Robertson et al

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

stage (Centers of Disease Control and Prevention [CDC]stage C), and low hemoglobin concentration (£10 g/dl)independently predicted risk of HAND in Yaoundé,Cameroon, and suggested that these factors could beused to detect HIV-infected patients at risk of develop-ing HAD, and included as criteria for eligibility tohighly active antiretroviral therapy (HAART).

Patterns of CNS involvement in the AIDSpopulation in EthiopiaUtilizing postmortem forensic pathology cases inAddisAbaba, Ethiopia, Dr. Langford andDr. Gemechuare conducting autopsies to determine the frequencyof HIV infection and neuropathological patterns, aswell as the frequency of CNS involvement ofMycobacterium tuberculosis in HIV+/� cases in thisregion. Investigators hypothesized that approximately15% of cases would be HIV+, 30% to 50% of thesecases would show evidence of CNS involvement, andthatCNS involvement inTB+caseswouldbe greater inHIV+ than in HIV� cases. The design of this experi-ment offered opportunities to characterize the naturalhistory of HIV in this population, observe HIV neuro-pathogenesis at different stages of disease, and char-acterize neuropathology patterns in HAART-naïvepatients. Half of the cases to be considered wereobtained from hospital settings and half from outsidesettings. Additionally, half of the cases within eachgroup resulted fromnatural death, andhalf from injury(accidental and intentional). Gross pathological obser-vations of 188 cases at autopsy included 112 cases(60%) suspected of chronic illness, 37 cases (20%)suspected of CNS infection, and 84 cases (45%) withsuspected pulmonary TBwith pneumonia. Histopath-ological analyses of the same 188 cases confirmed TB,encephalitis, liver disease, or other CNS-relatedpathologies in 42% of injury-related deaths. TB wasthemost common histopathological finding from bothinjury and natural cause groups, followed by liverdisease.TissuescollectedatMenelikHospital inAddisAbaba included spleen, liver, lungs, lymphnodes, andadrenal tissue from the periphery, and frontal cortex,basal ganglia (lenticular nucleus), hippocampus, andspinal cord from the nervous system. Investigatorsnoted the challenge of detecting HIV in tissue whenthe density of the virus may be low. As a result, tissuefrom the spleen and frontal cortexwere sent to TempleUniversity in Philadelphia for polymerase chain reac-tion (PCR) analyses. Further analyses are pending.

Neuropsychological assessment in HIV/AIDS andits challenges in a low-resource country (Zambia)Dr. Wood and Dr. Banda spoke about their researchin Zambia, one of the poorest countries in Africa.The HIV prevalence rate in Zambia is at 14% in apopulation estimated at 12 million. Greater than 95%of viral isolates obtained from Zambia are subtype C,and the neurological manifestations and psycholog-ical sequelae of subtype C–infected individuals inthis area are not yet known. These researchers and

their collaborators are taking measures to build thenecessary in-country capacity to conduct HIV-1 neu-ropathogenesis research to determine the impact andextent of HIV-associated neurological diseases inZambia. As part of these aims, they conducted apilot study among HIV� adults and children fromaround Lusaka to assess the utility of tools adaptedfrom Western/developed neuropsychological assess-ments. Among adult controls, scores for the IHDS,Color Trails 2, Timed Gait, and Grooved Pegboard(dominant and non-dominant hand) fell below West-ern norms, whereas scores for the Mini-Mental StateExamination and Color Trails 1 were comparable toWestern norms. These results reveal that setting/culture and level/quality of education can influenceneuropsychological performance, but that Westerndeveloped neuropsychological tests can be useful toassess HIV+ individuals in Zambia, when tests areadapted and performance compared to appropriatecontrols from the same setting.

To determine the impact and extent of HIV-associated neurological diseases in Zambia, re-searchers are conducting a prospective, 12-monthlongitudinal study, using a random sample of HIV+ and HIV� adults and children with no existingneurological or medical problems. Analysis of per-formance so far on the IHDS revealed that HIV+overall performed more poorly than HIV�, but nodifferences between HIV+ treated and untreated indi-viduals were indicated. Another interesting findingwas that HIV+ females tended to be more sensitive toIHDS motor tasks. This ongoing study will continueto analyze other adult and child assessments up to12 months, and this preliminary data will then beused to assess progression of neurological diseaseover a longer period.

Dr. Grant also spoke briefly on a pilot study inZambia to test the feasibility of using the HIV Neu-robehavioral Research Center (HNRC) Internationalneuropsychological test battery to assess HAND inZambia and lay the foundation for future neuroAIDSprojects. Due to the current lack of Zambian testnorms, US African American norms for this batterywere used, with constants added to all scaled andT-scores in order to reduce demographic effects andallow for impairment classification via a global def-icit score. Using these methods, approximately 33%of HIV+ subjects and 15% of HIV� subjects hadneurocognitive impairment. Neuropsychologicallyimpaired HIV+ subjects reported more than twicethe number of cognitive complaints than thoseconsidered normal, indicating that performancerelates to everyday life. The effects of HIV diseasewere greatest on neuropsychological domains ofsensory-information processing, verbal fluency, andexecutive functions. The results indicate that this NPtest battery shows sensitivity to HIV in Zambia whereclade C is common in addition to its sensitivity toHIV in other international settings where clade B iscommon. Use of African American norms, however,

NeuroAIDS in Africa 2009

K Robertson et al 193

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

only partly removed demographic effects (and onlyafter adjustment), and Zambian norms for properdemographic correction are clearly needed in thefuture.

ACTG (AIDS Clinical Trials Group) 5199International Neurological StudyDr. Robertson and colleagues conducted the firstlarge-scale international neurological clinical trialin resource-limited settings for HIV. The purposeof this trial was to estimate the prevalence andincidence of AIDS dementia and CNS opportunisticinfections, and study the changes in neurocognitivefunctioning at the initiation of antiretroviral treat-ment. Researchers conducted neurological and neu-ropsychological examinations at sites in Brazil,India, Malawi, Peru, South Africa, Thailand, andZimbabwe to examine the effects of three treatmentarms (3TC/ZDV/EFV, FTC/TDF/EFV, and FTC/ddI-EC/ATV). The three sections of the neurologicalexamination consisted of a face-to-face interview,an examination to observe the subject’s physicaland mental status, and a final summary section.The neuropsychological examinations included theTimed Gait Test, Finger-Tapping Test, GroovedPegboard Test, and the Verbal Fluency Test. Theseneuropsychological examinations tested for finemotor function, speed of processing, gross motorfunction, motor control, balance, executive function,speed, and attention, The neurological and neuro-psychological examinations showed a relatively highrate of peripheral neuropathy and other neurologicalabnormalities, but a low prevalence of minor cog-nitive motor disorder (MCMD) and AIDS dementia.The unexpectedly low prevalence of cognitiveimpairment may be a result of selection bias (higherfunctioning patients enrolled), which future studiesshould attempt to avoid. The results of the neuro-psychological examinations showed the expectedvariations between countries, which could be dueto differences in populations, cultures, HIV subtypesor resistance patterns, or variations in test adminis-tration. Overall performance on neuropsychologicalexaminations was similar between treatment arms.Conclusions of the study indicate there were signif-icant improvements in neuropsychological functionafter initiation of antiretroviral therapy in resource-limited settings, which could be contributed to thecontrol of HIV viral load through ARV effects and/orpractice effects. Dr. Robertson and colleagues are inthe final stages of developing a study to collectnormative comparison data at these sites to be uti-lized by future neuropsychological studies.

Depression and HIV-associated neurocognitivedeficits: Botswana pilot studiesIn the first of two pilot projects, Dr. Lawler proposedto measure the incidence of depression and dementiain Botswana, to observe the correlation between

depression and cognitive test performance, to learnif performance on a screening test correlates withother cognitive measures, and to determine whetherdepression and/or cognitive impairment caused pro-blems in daily activities. The tests used to measureDr. Lawler’s proposals were the International HIVDementia Scale (IHDS), Digit Symbol-Coding Test,WHO-UCLA Verbal Learning Test, Beck DepressionInventory–Fast Screen, Prime-MD Mood Module,and Activities of Daily Living Scale (ADLs). In thisstudy of 120 randomly selected HIV+ individuals,24% were depressed, 38% had dementia, andapproximately 10% suffered from both depressionand dementia. There was a strong correlationbetween depression measures and incidence ofdepression in Botswana was consistent with otherHIV studies in Africa. Furthermore, depression didnot affect cognitive task performance, and it was notrelated to language, nationality of examiner, age,education, or CD4 count. However, depressed indi-viduals reported more difficulty with ADLs, in par-ticular managing medications. Performance on thescreening test (IHDS) correlated with performance onother cognitive measures. This study also reportedthat dementia was not related to CD4 count or timesince HIV diagnosis, but was related to the length oftime on HAART.

The purpose of the second pilot study was todevelop culturally specific assessment tools, suchas the Botswana-specific Auditory Verbal LearningTest, and to compare depression and cognitive per-formance with well-matched groups of HIV+ andHIV� subjects. The tests used were Trails A, DigitSymbol-Coding, Botswana Auditory Verbal LearningTest, Action Fluency, Color Trails 2, Grooved Peg-board, Prime-MD (Depression, Anxiety, AlcoholModules), and Activities of Daily Living Scale.This project concluded that there was no significantdifference in depression between HIV+ and HIV�subjects (21% versus 17%), that HIV+ subjectsreported more suicidal ideation (15% versus 3%),and that there was a significant difference in anxietybetween the two groups (42% versus 22%). Further-more, HIV+ subjects were significantly impaired ontests of processing speed, verbal learning and mem-ory, fine motor speed, and executive functions. Also,subjects reported that memory was the primary rea-son they had difficulty adhering to medicationregimens.

Opportunities and challenges for NeuroAIDSresearch in NigeriaDr. Adebamowo discussed the opportunities andchallenges of NeuroAIDS research in Nigeria.Nigeria’s population of 150 million and adult HIVprevalence of 2.89% afflict the nation with the sec-ond largest population of HIV+ individuals in theworld. Additionally, Nigeria as a country has a lowpercentage of GDP (gross domestic product) spent onhealth care, a limited number of physicians and other

NeuroAIDS in Africa 2009

194 K Robertson et al

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

care providers, and as a result only approximately20% of adults have comprehensive correct knowl-edge about AIDS. In efforts to take action against therising prevalence of HIV infection, the amount offunding from the U.S. President’s Emergency Plan forAIDS Relief (PEPFAR); preventing mother to childtransmission (PMTCT), increased significantly in thepast 4 years, from around $10 million to over $1.1billion, which led to rapid improvement of HIVcounseling, screening, testing, support, and treat-ment. PEPFAR funding has also significantlyimproved the health care infrastructure throughoutthe country. The HIV-Nigeria ACTION Program nowhas 136 training centers, hub sites, ART sites, andPMTCT sites throughout the country. PEPFAR isthe largest prospective cohort of HIV-infected indi-viduals ever assembled, providing the statisticalpower to evaluate rare events such as neurologicalimpairment. In addition, PEPFAR supports substan-tial clinical, laboratory, and informatics infrastruc-ture, allowing record linkage and access to biologicalspecimens. Colocation of research at PEPFAR sitesprovides a cost effective strategy for study ofNeuroAIDS in Africa.

Studies of HIV-1–related neurocognitiveimpairment in NigeriaIn Nigeria, the prevalence, clinical features, andeffects of treatment for HIV-related neurocognitiveimpairment are unknown. Dr. Royal presented astudy in which treatment-naïve subjects in Nigeriawere evaluated for HIV-related neurocognitiveimpairment. Assessments included clinical question-naires, general clinical laboratory testing, CD4counts, viral load, and viral strain studies as wellas previously validated cognition assessments suchas the Karnofsky Performance Scale, InternationalHIV Dementia Scale, and a detailed neurocognitivebattery. Screening was critical so that the researcherscould study the sole effects of HIV-induced neuro-logical impairment rather than from other factorssuch as head trauma, psychiatric illness, drug use,learning disability, and active CNS infection. Con-clusions drawn from this study were that these indi-viduals had lower mean Karnofsky and IHDS scoresthan HIV� controls and that low IHDS scores corre-lated with worse clinical status. In a small subset ofpatients administered detailed neuropsychologicaltesting, low effect sizes were observed on the major-ity of the tests. However, several tests were found tobe potentially quite sensitive for detecting neurocog-nitive impairment in this group. Finally, analysis ofthe specific HIV subtypes that are found in Nigeriasuggested a possible association for specific strainswith the presence of cognitive impairment. Approx-imately 50% of patients in Nigeria are prescribedAZT-anchored regimens, 30% TDF-containing regi-mens, and approximately 20% receive regimenscontaining d4T. This pilot study demonstrated

feasibility for performing future studies of HIV-related neurocognitive impairment in Nigeria, whichincludes the assessment of the effects of treatment ofthe disorder.

HIV clades/subtypes

Subtype-associated differences in risk for dementiain UgandaInvestigations into subtype-associated differences inrisk for HAND are currently underway throughoutthe world, and Dr. Sacktor and Dr. Nakasujja pre-sented some of their research on this topic in Ugandawhere clades A and D are common. In a previousretrospective study of 140 HIV seroconverters inRakai, Uganda, followed over 5 years, patientswith subtype D had a faster progression to AIDSor death (Laeyendecker, 2006). Dr. Sacktor andNakasujja followed 60 antiretroviral-naïve HIV+individuals at risk for HIV-associated cognitiveimpairment in Uganda to determine if subtype D isalso associated with a greater risk of dementia com-pared to subtype A. Individuals included in thestudy completed the IHDS, Grooved Pegboard,Timed Gait, WHO-UCLA Auditory Verbal LearningTest, Symbol-Digit, Color Trails 1 and 2, and DigitSpan Forward and Backward. Using data from thephylogenetic analysis of both gp41 and gag genes,HIV+ individuals were classified as subtype A, C, D,or A/D recombinants.

Results indicate that HIV subtype D is associatedwith an increased risk of dementia compared tosubtype A. Of the HIV+ individuals with subtype D,89% were categorized with dementia (MSK 1),whereas only 24% of individuals with subtype Areceived this rating. These results provide the firstevidence in well-characterized HIV+ patients at asimilar stage of HIV disease, suggesting that HIV sub-types may have different biological properties withrespect to their capacity to cause HIV-associatedcognitive impairment.

Neuropsychological profile of patients commencingHAART in Cape Town, South Africa—preliminaryfindingsSubtype C is responsible for more than 90% of HIV-1infections in South Africa and may differ fromsubtype B in terms of protein binding sites, bindingcharacteristics, replicative capacity, and possibly inneurotoxicity and risk for cognitive impairment. Dr.Joska, Dr. Paul, and colleagues discussed a naturalmutation in the TAT protein in subtype C associatedwith reduced monocyte chemokine regulation andits relevance to the CNS. TAT is involved in themigration of monocytes into the brain via up-regu-lation of inflammatory cytokines and adhesion mole-cules. TAT also disrupts the tight junctions in theblood-brain barrier (BBB), therefore variations in the

NeuroAIDS in Africa 2009

K Robertson et al 195

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

TAT protein could lead to variations in CNSpathology.

An early report from Ethiopia suggested that HADis infrequent in clade C (Clifford et al, 2007), butstudies from India suggest that it is comparable toclade B (Gupta et al, 2007). Dr. Joska, in collaborationwith the University of Missouri, conducted a pilotstudy to determine if cognitive impairment exists inclade C HIV in South Africa and to begin to defineneuroimaging biomarkers of CNS involvement. Cog-nitive measures used included Hopkins VerbalLearning Test–R, Brief Visual Memory Test–R, StroopTest, Color Trails, Digit Symbol, Grooved Pegboard,Tapping, and Animal Fluency. Participants included52 ART-naïve HIV patients (29.6 years of age;9.7 years of education) and 25 controls (27.0 yearsof age; 11.3 years of education). HIV patients per-formed significantly worse on the measures of DigitSymbol, Color Trails 2, Animal Fluency, and StroopColor-Word interference. Analysis of diffusion tensorimaging (DTI) and fractional anisotropy also revealedsignificant differences between these groups. Thesepreliminary results provide evidence for cognitiveimpairment in HIV clade C infection in South Africa,and these investigators plan to determine if cognitivestatus is associated with a TAT defect and whetheror not relationships exist between cognitive statusand proviral DNA, inflammatory cytokines, andneuroimaging.

Variation in neurocognitive impairment based onHIV-1 subtypeA discussion of the effect of HIV-1 subtype on neu-rocognitive impairment in pediatric infection waspresented by Dr. Boivin, Dr. Wong, and Mr. Bangir-ana. These researchers sought to determine whetherHIV+ school-aged children in Uganda currently inel-igible for ART demonstrate deficits relative to HIV�children and whether the risk for progressiveencephalopathy (PE) in children differs by subtype.

Subtype, along with other clinical indicators, maybe an important consideration for when to initiateART treatment and for antiretroviral selection (e.g.,CNS penetration characteristics). One hundred andtwo HIV-infected antiretroviral therapy (ART)-naïvechildren (mean age = 8.8 years) and 104 noninfectedcontrols (mean = 8.9 years) completed the KaufmanAssessment Battery for Children, 2nd edition(KABC-2), Tests of Variables of Attention (TOVA)auditory and visual tests, and the Bruininks/Oseretsky Test for Motor Proficiency, 2nd edition(BOT-2). HIV-1 subtype was defined by the compo-sition of the env gene determined using a sequence-specific real time PCR assay (Hoelscher et al, 2002).Evaluation of performance demonstrated significantdifferences between the HIV+ and HIV� groups onthe KABC measures of memory, learning, and visual-spatial tasks, visual TOVA response time, andattention-deficit hyperactivity disorder (ADHD) score

and on the BOT’s manual dexterity, upper-limbcoordination, and speed and agility tests. Therewere no differences between HIV+ and HIV� chil-dren on auditory TOVA. A comparison of HIV+children with subtypes A and D indicated subtypeA children did more poorly than subtype D onvisual-spatial tasks and had more errors on visualTOVA. No differences were reported between HIV+children with subtype A versus subtype D on theBOT and auditory TOVA.

In summary, ART-naïve HIV+ Ugandan schoolage children have poorer neuropsychological per-formance on working memory, visual-spatial pro-cessing, ADHD index, and aspects of motordevelopment. HIV subtype A is associated withpoorer neuropsychological performance (visual spa-tial processing and TOVA impulsivity), compared tosubtype D in treatment-naïve Ugandan children. HIVsubtypes A and D may be associated with differentrisks for progressive encephalopathy (PE) in chil-dren. HIV subtype influences on encephalopathymay fundamentally differ between children andadults as subtype D was associated with a greaterrisk of dementia or HAND compared to subtype A inUgandan adults. The PE described here in childrenmay be less dependent on immune suppression butinstead may be influenced by HIV coreceptor use.According to their pathogenic model of subtypedifferences, subtype A, which is more CCR5-tropic,may be more efficient than subtype D, which isCXCR4-tropic at infecting CNS macrophage andmicroglia cells. X4 virus may be more immunolog-ically pathogenic because it infects a wider range ofcells in lymphoid tissue, but may replicate lessefficiently in the CNS.

Pediatric Neuroassessments in Africa

In 2007, an estimated 370,000 children becamenewly infected with HIV. Over 90% of theseinfections occurred in sub-Saharan Africa (WHO/UNAIDS, 2008). More than 90% of new infectionsin children occur vertically, during pregnancy, attime of birth, or through breastfeeding (WHO/UNAIDS, 2008). HIV infection in vertically infectedchildren impairs the development and growth of animmature CNS, resulting in delayed acquisition ofneurodevelopmental milestones in the majority ofchildren in the pre-ART era (Diamond et al, 1987;Epstein et al, 1986). Without treatment, approxi-mately one half of children will die by 2 years ofage (Newell et al, 2004).

Neurocognitive assessments in HIV-infected chil-dren, especially in Africa, are plagued by a long listof confounding factors, which include the socioeco-nomic status or quality of the home environment, theemotional well-being of caregivers (i.e., depression),and secondary effects of HIV such as malnutrition,

NeuroAIDS in Africa 2009

196 K Robertson et al

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

quality of education, quality of prenatal care, andtreatment with antiretroviral drugs.

Neurobehavioral development in a pediatricpopulation exposed to HIVHIV affects all domains of child functioning at leastto some extent. Motor development is the mostaffected in term of severity, early onset, and persis-tence across the age groups. Dr. Holding and collea-gues investigated the role of weight-for-age anddisease stage in poor psychomotor outcome ofHIV-infected children in Kilifi, Kenya. They used alocally developed and validated measure, the KilifiDevelopmental Inventory, to assess psychomotordevelopment in 48 children aged 6 to 35 monthsprenatally exposed to HIV (exposed and infectedodds ratios [OR] versus exposed and not infected).Investigators reported a significant main effect of HIVstatus (F(2, 38.01) = 7.89, P < 001). Children in theHIV-infected group had lower mean psychomotorscores than both exposed but uninfected childrenand controls. No significant differences wereobserved between the HIV-exposed group and thereference population. Disease stage had an inverserelationship with psychomotor outcome. Mean psy-chomotor scores by disease stage were stage 1, 0.40(1.68); stage 2, �1.01 (1.93); stage 3, �1.46 (1.57).Children in disease stages 2 and 3 differed signifi-cantly from those of the reference population.Weight-for-age had a positive relationship with psy-chomotor outcome. These findings indicate thatweight-for-age and disease stage are viable, easilymeasurable benchmarks to identify those childrenat developmental risk. These findings also imply thatnutritional intervention and other measures aimed atslowing down disease progression may delay theonset and severity of psychomotor impairment inthe paediatric HIV population in Africa.

Effects of HIV and ARV exposure on child healthand neurodevelopment, BotswanaPerinatal ARV medications have led to dramaticreductions in mother-to-child transmission (MTCT)of HIV-1. One third of HIV+ pregnant women nowreceive ARVs during pregnancy or labor, and HAARTuseduringpregnancyisincreasing.Concernsrelatedtopossible negative effects of prenatal ARV exposure onneurodevelopment have been raised, given theunknown impact of ARVs on the developing brainand issues of potential mitochondrial toxicity relatedtoNRTI exposure.A recent largewell-controlled studyshowed no cognitive or motor differences betweenARV-exposed and -unexposed HIV� children (Wil-liams et al, 2010). HIV-exposed but uninfected chil-dren have also shown poorer outcomes than HIVunexposed children (Van Rie et al, 2008). Studies ofneurological and neurodevelopmental outcomes inHIV/ARV-exposed but uninfected children haveyielded conflicting results and controlling for

confounding factors (access to care, substance abuse,socioeconomic status [SES]) is problematic.

Dr. Kammerer, Dr. Lockman, and colleagues arecurrently working on a project in Botswana to furtherinvestigate the effects of ARV and HIV exposure inchildren. Botswana has the second highest rate ofHIV in the world and infection affects all socioeco-nomic levels. In addition, 33.7% of pregnant womenare HIV+, but significant government support has ledto >90% of pregnant women being tested for HIV and>85% of HIV+ women taking ARVs. The primaryobjectives of the current study are (1) to compareneurodevelopmental outcomes following in uteroexposure to three-drug HAART (two or more NRTIsand single NRTI [zidovudine]) among HIV-unin-fected children born to HIV+ mothers (with similarCD4 cell counts); (2) to prospectively compare neu-rodevelopmental outcomes and rates of mortality;and (3) to determine the biologic and social factorsmost strongly associated with excess mortality andneurodevelopmental delay between children whoare HIV exposed but uninfected and childrenborn to HIV-uninfected mothers in a setting withHAART availability. Neurodevelopmental outcomeswill be measured by a combination of an adaptedUS-normed test (Bayley Scales) that is used in com-parable studies and tests developed and normed inAfrica (Abubakar et al, 2008). Maternal infectionstatus, mental health, physical health, social support,and nutrition status will be assessed as well as formof feeding (breastfeeding versus formula), health andnutrition, vision, hearing, and neurological status.

Effects of HIV on the neurodevelopment of younginfants in Africa: experience from MalawiDr. Van Rie discussed her work on the effects of HIVand ART on neurodevelopment of infants and chil-dren in Africa. In Kinshasa, Democratic Republic ofCongo, Van Rie and colleagues concluded that thevast majority of HIV-infected children, even youngchildren, present with severe motor and cognitivedevelopmental delay at time of diagnosis (Van Rieet al, 2008). Access to HIV care accelerated motordevelopment, but did not result in a catch up ofcognitive development. Timing of care had no effecton motor development but the cognitive develop-ment of children who accessed care early (prior todevelopment of severe immunodeficiency) tended tonormalize within 1 year of care, whereas the cogni-tive development remained delayed for those pre-senting late (Van Rie et al, 2009). These findingsunderscore the importance of early treatment ofHIV infection in children. Dr. Van Rie is currentlystudying the effects of early ART (within the first 3 to4 months of life) on neurodevelopment in Blantyre,Malawi. Preliminary data indicate that children whostart treatment early may demonstrate neurodevelop-mental delay (both motor and cognitive) at a veryearly stage in life (age 3 months) but can achieve a

NeuroAIDS in Africa 2009

K Robertson et al 197

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

motor and cognitive development similar to that ofHIV unexposed children by 6 months of age.

Overview of treatment issues

Implementing HIV treatment in Africa faces manychallenges such as a high disease burden, manycountries with poor health infrastructure, ensuringquality of care with trained staff and appropriatemonitoring tests, getting people into care beforedisease progression, dual epidemics of TB andHIV, providing affordable and convenient regimenswith favorable toxicity profiles, and sustainability offunding for ART programs. In spite of these chal-lenges many countries are making progress in gettingART to those in need. From the years 2004 to 2007,South Africa, Uganda, Malawi, and Ghana, for exam-ple, drastically increased the estimated number ofpeople receiving ART; however, only 27%, 33%,35%, and 15% of the estimated number of peopleneeding ART were receiving medication based onUNAIDS/WHO criteria.

Evolving changes of treatment implementation inAfricaDr. Ive reviewed the challenges and changes in HIVtreatment in Africa and discussed a 4-year programcalled the NSP implemented in South Africa in 2007.The NSP aims to (1) reduce the number of new HIVinfections by 50% and (2) reduce the impact of HIVand AIDS on individuals, families, communities, andsociety by expanding access to appropriate treat-ment, care, and support to 80% of all people diag-nosed with HIV. In particular, young people in theage group 15–24 are a focus of all the interventions,especially for behavior change–based prevention.The interventions needed to reach the goals of theNSP include (1) prevention; (2) treatment, care, andsupport; (3) human and legal rights; and (4) moni-toring, research, and surveillance. Training and sus-taining valuable human resources (i.e., doctors andnurses) is one of the biggest obstacles to providingquality care. In areas where doctors are scarce, gov-ernments and care providers may need to consider“task shifting,” and delegate some activities to lessqualified personnel. Training of primary health carenurses (rather than doctors) to initiate antiretroviraltreatment, lay counsellors (rather than nurses)“pricking” patients for rapid HIV tests, and laycounsellors (instead of social workers) for orphansupport activities, for example, may save time andmoney. Another pitfall of treatment programs inAfrica is “stock outs” of antiretroviral drugs in gen-eral or specific components of a drug regimen. Stockouts are problematic because they undermine themessage of adherence, increase the chance of resis-tance to first line regimens, and delay puttingpatients onto therapy.

Two important changes in HIV clinics in Africainvolve the use of d4t and viral load monitoring.From 2000 to 2004, d4t was a staple of ART in Africabecause it was effective and inexpensive. Long-termuse of the drug, however, was associated with asignificant number of adverse events (e.g., peripheralneuropathy and lactic acidosis), which led to non-adherence or necessitated a change in regimen. As aresult, d4t is now phasing out of treatment. Moni-toring viral load on the other hand is becoming morecommon. Although it was once believed to be tooexpensive, the costs of late detection of treatmentfailure and resulting high levels of drug resistancenow overshadow the costs of monitoring viral load.

HIV-associated CNS opportunistic infections andimmune reconstitution syndromeImmune reconstitution inflammatory syndrome(IRIS) is a paradoxical worsening of a patient’s clin-ical condition that is attributable to the recovery ofthe immune system after initiation of HAART. IRISoccurs in 15% to 25% of patients on HAART andin 20% to 45% of patients with OIs on HAART(Shelburne et al, 2006). Approximately 50% ofpatients that develop IRIS have been on HAARTfor around 50 days but the syndrome can appearearlier in treatment. Patients who are antiretroviralnaïve, have active or subclinical OIs at HAARTinitiation, have CD4 count <50 cells/mm3, or havea prompt decrease in viral load with HAART are atgreater risk for IRIS. Certain host genes can also makea patient more susceptible. CNS infections associatedwith IRIS include viral infections (JC virus: primarymultifocal leukoencephalopathy [PML]; varicella-zoster virus [VZV]: vasculitis; cytomegalovirus[CMV]: retinitis; Epstein-Barr virus [EBV], optic neu-ropathy and primary CNS lymphoma), bacterialinfections (mycobacterial disease: TB, leprosy), fun-gal infections (cryptococcal meningitis, candidameningitis), and parasitic infections (toxoplasmosis).Encephalitis directly attributable to HIV also occurs.

Dr. Nath presented three different cases showingthe spectrum of IRIS that clinicians encounter, aswell as a review of the histopathological correlates,potential biomarkers, and treatment options for IRIS.In the first case, a patient with cryptococcal menini-gitis on HAART and treatment for cryptococcal infec-tion for 8 months developed headache, photophobia,and nausea while their CD4 count increased from 24to 124 cells/mm3 and their viral load dropped toundetectable levels. After treatment with dexameth-asone 24 mg/day, the patient dramatically improvedand was diagnosed with IRIS with cryptococcal men-ingitis(Venkataramanaetal,2006).Inthesecondcase,apatientwithaCD4countof474andHIVviral load<200developed an episode of aphasia and seizures after10 years of HAART. In this case the dementia wasmediated by macrophages in the CNS. The patienttemporarily improved with steroids but then

NeuroAIDS in Africa 2009

198 K Robertson et al

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

encountered more cognitive andmotor issues (Venka-taramana et al, 2006). The last was a case of acuteencephalopathy with IRIS. A patient developed rightsideweakness and cognitive decline over a period of 8weeks and then started HAART. Ten weeks later, theCD4 count improved from 50 to 320 and the viral loaddropped from838,100 to4000, but thenover thecourseofanotherweekthepatientprogressivelydeclinedwithfocal and generalized seizures resulting in death withcomplications this time attributed to CD8+ T cells(Miller et al, 2004).

IRIS associated with progressive multifocal leu-koencephalopathy (PML) is characterized by CD3+ T cell and macrophage (CD68) infiltration, infectionof oligodendrocytes by JC virus, and inflammationattributed mostly to perivascular CD8+ T lympho-cytes, whereas CD4+ lymphocytes are absent. Neuro-pathological findings include multiple demyelinatingfoci in frontal, parietal, and temporal white matter(mostly subcortical areas). Potential biomarkers forIRIS include cytokines (e.g., interleukin-6 [IL-6]) andmarkers of neuronal injury (axonal and synaptic pro-teins). IRISusuallyonsetsafterseveralweeksofHAARTwhen the viral load drops and CD8+ T lymphocytesspike, thenprogresses asCD4+T lymphocytes increaseand CD8 cells decline. Whenmanaging potential CNS-associated IRIS, clinicians must first exclude effects ofdrugs, nonadherence to drugs, and progression ofunderlying disease. Then they must decide whetherto discontinue HAART and whether to give systemiccorticosteroids (Riedel et al, 2006). The treatmentoptions for IRIS are not ideal because steroids increaserisks from immune suppression and interruption ofHAART increases risk for resistance to therapy andreemergence of IRIS upon resumption of HAART.The current recommendations for use of steroids inIRIS are high-dose steroids tapered with oral steroidsfor a month (with OI prophylaxis) for catastrophicIRIS, but the same treatment for symptomatic andasymptomatic IRIS is debatable.

Discussion

Although research and treatment of HIV-associatedneurocognitive disorders has been going on for over20 years in the developed world, much less has beencarried out in resource-limited setting (RLS) such asAfrica. Participants at the conference discussedsome of the gaps in the knowledge and care ofHAND in Africa and offered several ideas to worktowards bridging the gaps. One concern was thatHIV care providers in Africa may not have the samegeneral level of awareness of HAND or neuroAIDSand a network to spread general information on thedisease may help. In addition to this need to spreadgeneral information on HAND, forming workingconsortiums and cooperative arrangements betweenresearch groups were also of great interest.

Participants discussed establishing these arrange-ments in order to share resources on tissue samples,neuropsychological data and tools, and neuroimag-ing amongst other needs and considered modelingthe consortium on the Institutional DevelopmentAward (IDeA) program, which extends NIHresources to foster health-related research in ruraland medically undeserved areas of the US. Possiblearrangements could include collaboration buildingbetween medical schools and institutions in theUnited States/Europe and RLS as well as site-to-site (or south-to-south) sharing of expertise. Ifresearch sites with adequate expertise can be iden-tified, the field of HAND research in Africa couldbenefit from networks developed on the continent,which would have less overwhelming differencesbetween sites. Before these efforts can move forward,however, a concrete set of training needs andresources needs to be established.

Establishing a set of neurological screens, neuro-psychological examinations, and functional assess-ments that can be used across sites, however, is aproblem on its own. Even within Africa, differencesin culture, language, education, and access toresources exist, which could affect performance,interpretation, and validity of many instruments.

The issue of adapting, validating, and gatheringnormative data for neuropsychological assessmentsoccupied much of the discussion section. Is it pos-sible to develop and validate a set of tests that can beused across cultures? Most participants agreed thatmany instruments developed in Western setting canbe used in RLS, if properly adapted for educationlevel and culture of the setting; however, they alsoestablished that developing one set of instruments tobe used across sites is not the best way to screen forneurocognitive impairment. Instead, developing a setof domains to be tested, in place of a strict set of tests,may be useful. It may not be possible to test eachdomain in every setting due to restraints on time,staff, and multiple issues with language proficiencyand culture. Starting with a breadth of assessmentsand working to develop a set of domains for eachsetting might be better than creating many tests fromscratch or using the same tests across settings. To laythe foundation for HIV neurological studies in RLS,the ACTG 5199 International Neurological Studyselected a brief battery of neuropsychological testssensitive to HIV-associated changes and also gearedtowards limiting cultural influences. In addition,researchers at these sites will soon begin a normativecontrol study with a slightly expanded battery, aim-ing to enroll approximately 2400 participants, inorder to address the lack of normative data in RLS.

The appropriateness of activities of daily livingscales and other functional assessment also cameinto question. In RLS, where most people are unem-ployed and don’t drive cars, and where work anddaily life are less complex, thus many existing instru-ments may not be useful to determine functional

NeuroAIDS in Africa 2009

K Robertson et al 199

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

impairment. The Bolton questionnaire developed inUganda represents an improvement in this area, butdeveloping other culturally appropriate functionalassessments, possibly with help from communitymembers, presents an ideal research project.

Other problems that occur with neuropsycholog-ical instruments and neurological examinationsinvolve standardized administration, scoring, anddefinitions of clinical outcomes. Cultural differencesin administration and interpretation of instructionsexist, so quality control to measure performance ofthose giving assessments should be in place, andmethods for monitoring quality should be publishedwith research findings. Experimenter drift (awayfrom standardized assessment) is common in bothdeveloped and developing settings, and must bemanaged. Primary investigators should considerobservingmultiple testing sessions every fewmonths,even if they don’t know the language. Reviewing CRFsand rescoring neuropsychological tests randomly canalso help catch mistakes and reveal systematic errors.Periodically reviewing training DVDs that modelappropriate testing techniques and blinding peopleadministering assessments as much as possible tothe disease status of the individual may also benefitstandardadministration.Regardingclinicaloutcomes,depression, functional impairment, and cognitiveimpairment are defined in different ways by differentstudies, which create issues with comparing perfor-mance and health across populations. The “Frascatipaper” in 2007 provided an update of the researchnosology for HAND and laid out a set of suggestedcognitive domains and definitions for cognitive andfunctional impairment (Antinori et al, 2007). The key,however, may not be a set of definitions, but rather forresearchers to be very specific when presenting find-ings as to what definitions were used to define cogni-tive/functional/behavioral impairment, and to pointout differences in those definitions from other studies.

There are many challenges to conducting neuro-psychological assessments in pediatric populationsin Africa. Quality of home environment, physicaland mental well-being of caregivers, gestationaleffects (MTCT and prenatal treatment), and malnu-trition may all impact performance. Participantsdiscussed different strategies for evaluating neuro-cognition in children, which included screening ver-sus in-depth assessments, home- or village-basedevaluations when clinic access is difficult, use ofcomputers and cutting-edge developments, as wellas strategies for standardizing scores. One proposalwas to use reference groups of children based on age,gender, and environment instead of large normativedata sets. Developing assessments appropriate tocultural contexts that are sustainable (do not involvepublished/copyrighted materials) and then deter-mining appropriate developmental milestones forassessments in resource-limited settings were alsodiscussed. Neuropsychological “interventions” toboost positive neuroplasticity, such as games for

learning and development in school-age HIV+ chil-dren, may compliment ART and help amelioratesome of the neurocognitive effects of HIV.

In terms of neuropathogenesis, several areas wereconsidered. As older, better penetrating ARVs aregradually replaced by newer drugs, some unintendedloss of neurocognitive benefits may occur. Theeffects of HIV-2 dual infection on neurocognition,aswell as the other opportunistic infections and cofac-tors such as crystal methamphetamine abuse in SouthAfrica, were also of concern. Linking systemic immu-nological activation and symptoms to CNS immuneactivation and clarifying the clade- or subtype-relatedmechanisms for disease need further attention. Forexample, in adults, clade D demonstrates increasedneuropathogenesis over cladeA,whereas theoppositeeffect was observed in children. Studies investigatingclade C have reported less motor effects than otherclades. These and other studies need to be replicatedto determine if clade-specific neurocognitive profilesexist. Studies such as these are difficult to interpret,however, because evenwithin clades, different motifsand protein products exist. The differences in patho-genesis may not be due to clades per se, but specificproteinmotifswithin a genome that cause toxicity. Dr.Nath demonstrated that even a single amino acid shiftcan make a difference. If serine is substituted for cys-teine, theTATprotein is less toxicbecause thecysteineon TAT hooks in with cysteine on a receptor to open achannel. Post-translational modifications of proteinssuch as TATwithinmammalian cellsmake them1000times more toxic than TAT produced from Esherichiacoli. Ultimately determining the mechanismsbehind clade-related differences in neurocognitiveimpairment will be very complicated and require alot of sequencing data in addition to neurocognitiveassessments and analyses of other variables.

Conclusions

NeuroAIDS research in Africa faces many challenges,but also offers many opportunities. Because sub-Saharan Africa was the “ground zero” for the HIVpandemic, the majority of HIV-infected individualsreside there. If the effects of culture, language, andeducation can be teased out, this large pool of pro-spective patient populations presents opportunities toinvestigate how differences in viral and host geneticsinfluence clinical presentation of neurocognitive dis-ease. In addition, collections of postmortem tissuesamples acquired in the region could be valuableresources for further characterizing the neuropatho-genesis of HIV and comorbid conditions. The popu-lation of the region is generally resource poor and athigh risk of disease, so access to treatment andcounseling provided by research is very meaningful.Knowledge gained from HAND research in the regioncould significantly improve HAND treatmentthroughout the world.

NeuroAIDS in Africa 2009

200 K Robertson et al

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

Complete list of conference participants

Speakers and chairpersons

Clement Adebamowo (Abuja, Nigeria)Mwanza Banda (Lusaka, Zambia)Paul Bangirana (Kampala, Uganda)Michael Boivin (East Lansing, USA)David Clifford (St. Louis, USA)Amadou Gallo Diop (Dakar, Senegal)Graham Fieggen (Cape Town, South Africa)Igor Grant (San Diego, USA)James Hakim (Harare, Zimbabwe)Jeannine Heckmann (Cape Town, South Africa)Penny Holding (Mombasa, Kenya)Mina Hosseinipour (Lilongwe, Malawi)Prudence Ive (Gauteng, South Africa)Assan Jaye (Banjul, The Gambia)Jeymohan Joseph (Bethesda, USA)John Joska (Cape Town, South Africa)Betsy Kammerer (Newton, USA)Georgette Kanmogne (Omaha, USA)T. Dianne Langford (Philadelphia, USA)Kathy Lawler (Philadelphia, USA)Scott Letendre (San Diego, USA)Jeanne McDermott (Bethesda, USA)Marc Mendelson (Cape Town, South Africa)Noeline Nakasujja (Kampala, Uganda)Avindra Nath (Baltimore, USA)Alfred Njamnshi (Yaoundé, Cameroon)Robert Paul (Chesterfield, USA)Kevin Robertson (Chapel Hill, USA)Walter Royal III (Baltimore, USA)Ned Sacktor (Baltimore, USA)Jean-Louis Sankalé (Decatur, USA)Soraya Seedat (Cape Town, South Africa)Annelies Van Rie (Chapel Hill, USA)Tufa Gemechu Weyessa (Addis Ababa, Ethiopia)Joseph Wong (San Francisco, USA)Charles Wood (Lincoln, USA)

Participants

Jing Bao (Bethesda, USA)Kathleen Bateman (Cape Town,South Africa)Mariana Cherner (San Diego, USA)Ronald Collman (Philadelphia, USA)Marc Combrinck (Cape Town, South Africa)Luminita Ene (Bucharest, Romania)Edwin Escobar G. (Miranda, Venezuela)Nathan Geffen (Cape Town, South Africa)Karl Goodkin (Los Angeles, USA)Hetta Gouse (Cape Town, South Africa)W. David Hardy (Los Angeles, USA)Jacqueline Hoare (Cape Town, South Africa)Jennifer Jardine (Cape Town, South Africa)Cecilia Kanyama (Cape Town, South Africa)Johnstone Kumwenda (Blantyre, Malawi)Barbara Laughton (Tygerberg, South Africa)Jeff Liner (Chapel Hill, USA)Suzaan Marais (Cape Town, South Africa)Ernesta M. Meintjes (Cape Town,South Africa)Rosie Mngqibisa (Durban, South Africa)Agnes Moses (Lilongwe, Malawi)Daphne Radebe (Gauteng, South Africa)Ian Sanne (Johannesburg, South Africa)Fabienne Shumbusho (Kigali, Rwanda)Lawrence Lee Soon-U (Singapore)Stephen Ssebulime (Kampala, Uganda)Kevin Stoloff (Cape Town, South Africa)Helen Struthers (Diepkloof, South Africa)Innocent Turate (Kigali, Rwanda)Victor G. Valcour (San Francisco, USA)Moleen Waison (Harare, Zimbabwe)

Declaration of interest: The authors report no con-flicts of interest. The authors alone are responsiblefor the content and writing of the paper.

References

Abubakar A, Holding P, van Baar A, Newton CR, van deVijver FJ (2008). Monitoring psychomotor developmentin a resource-limited setting: an evaluation of the KilifiDevelopmental Inventory. Ann Trop Paediatr 28:217–226.

Antinori A, Arendt G, Becker JT, Brew BJ, Byrd DA, ChernerM, Clifford DB, Cinque P, Epstein LG, Goodkin K, GisslenM, Grant I, Heaton RK, Joseph J, Marder K, Marra CM,McArthur JC, Nunn M, Price RW, Pulliam L, RobertsonKR, Sacktor N, Valcour V, Wojna VE (2007). Updatedresearch nosology for HIV-associated neurocognitive dis-orders. Neurology 69: 1789–1799.

Clifford DB, Mitike MT, Mekonnen Y, Zhang J, Zenebe G,Melaku Z, Zewde A, Gessesse N, Wolday D, Messele T,Teshome M, Evans S (2007). Neurological evaluation ofuntreated human immunodeficiency virus infectedadults in Ethiopia. J NeuroVirol 13: 67–72.

Descamps D, Collin G, Letourneur F, Apetrei C, Damond F,Loussert-Ajaka I, Simon F, Saragosti S, Brun-Vezinet F(1997). Susceptibility of human immunodeficiency virustype 1 group O isolates to antiretroviral agents: in vitrophenotypic andgenotypic analyses. J Virol71: 8893–8898.

Diamond GW, Kaufman J, Belman AL, Cohen L, Cohen HJ,Rubinstein A (1987). Characterization of cognitive func-tioning in a subgroup of children with congenital HIVinfection. Arch Clin Neuropsychol 2: 245–256.

NeuroAIDS in Africa 2009

K Robertson et al 201

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.

Epstein LG, Sharer LR, Oleske JM, Connor EM, Goudsmit J,Bagdon L, Robert-Guroff M, Koenigsberger MR (1986).Neurologic manifestations of human immunodeficiencyvirus infection in children. Pediatrics 78: 678–687.

Gupta JD, Satishchandra P, Gopukumar K, Wilkie F,Waldrop-Valverde D, Ellis R, Ownby R, SubbakrishnaDK, Desai A, Kamat A, Ravi V, Rao BS, Satish KS,Kumar M (2007). Neuropsychological deficits in humanimmunodeficiency virus type 1 clade C-seropositiveadults from South India. J NeuroVirol 13: 195–202.

Hemelaar J, Gouws E, Ghys PD, Osmanov S (2006). Globaland regional distribution of HIV-1 genetic subtypes andrecombinants in 2004. AIDS 20: W13–W23.

Hoelscher M, Dowling WE, Sanders-Buell E, Carr JK, HarrisME, Thomschke A, Robb ML, Birx DL, McCutchan FE(2002). Detection of HIV-1 subtypes, recombinants, anddual infections in east Africa by a multi-region hybrid-ization assay. AIDS 16: 2055–2064.

Kanki PJ, Hamel DJ, Sankale JL, Hsieh C, Thior I, Barin F,Woodcock SA, Gueye-Ndiaye A, Zhang E, Montano M,Siby T, Marlink R, I ND, Essex ME, S MB (1999). Humanimmunodeficiency virus type 1 subtypes differ in dis-ease progression. J Infect Dis 179: 68–73.

Laeyendecker O, Li X, Arroyo M, McCutchan F, Gray R,Wawer M, Serwadda D, Nalugoda F, Kigozi G, Quinn T,and Rakai Health Science Program (2006). 13th Con-ference on Retroviruses and Opportunistic Infections;February 5–February 8, 2006; Denver, CO.

LANL (2010). The circulating recombinant forms (CRFs).In: HIV Sequence Database. Los Alamos, NM: LosAlamos National Laboratory.

Miller RF, Isaacson PG, Hall-Craggs M, Lucas S, Gray F,Scaravilli F, An SF (2004). Cerebral CD8+ lymphocyto-sis in HIV-1 infected patients with immune restorationinduced by HAART. Acta Neuropathol 108: 17–23.

Newell ML, Coovadia H, Cortina-Borja M, Rollins N,Gaillard P, Dabis F (2004). Mortality of infected anduninfected infants born to HIV-infected mothers inAfrica: a pooled analysis. Lancet 364: 1236–1243.

Poveda E, de Mendoza C, Parkin N, Choe S, Garcia-GascoP, Corral A, Soriano V (2008). Evidence for differentsusceptibility to tipranavir and darunavir in patients

infected with distinct HIV-1 subtypes. AIDS 22:611–616.

Ren J, Bird LE, Chamberlain PP, Stewart-Jones GB, StuartDI, Stammers DK (2002). Structure of HIV-2 reversetranscriptase at 2.35-A resolution and the mechanismof resistance to non-nucleoside inhibitors. Proc NatlAcad Sci USA 99: 14410–14415.

Riedel DJ, Pardo CA, McArthur J, Nath A (2006). Therapyinsight: CNS manifestations of HIV-associated immunereconstitution inflammatory syndrome. Nat Clin PractNeurol 2: 557–565.

Shelburne SA, Montes M, Hamill RJ (2006). Immune recon-stitution inflammatory syndrome: more answers, morequestions. J Antimicrob Chemother 57: 167–170.

Tscherning C, Alaeus A, Fredriksson R, Bjorndal A, Deng H,Littman DR, Fenyo EM, Albert J (1998). Differences inchemokine coreceptor usage between genetic subtypesof HIV-1. Virology 241: 181–188.

Van Rie A, Dow A, Mupuala A, Stewart P (2009). Neuro-developmental trajectory of HIV-infected childrenaccessing care in Kinshasa, Democratic Republic ofCongo. J Acquir Immune Defic Syndr 52: 636–642.

Van Rie A, Mupuala A, Dow A (2008). Impact of the HIV/AIDS epidemic on the neurodevelopment of preschool-aged children in Kinshasa, Democratic Republic of theCongo. Pediatrics 122: e123–e128.

Vasan A, Renjifo B, Hertzmark E, Chaplin B, Msamanga G,Essex M, Fawzi W, Hunter D (2006). Different rates ofdisease progression of HIV type 1 infection in Tanzaniabased on infecting subtype. Clin Infect Dis 42: 843–852.

Venkataramana A, Pardo CA, McArthur JC, Kerr DA, IraniDN, Griffin JW, Burger P, Reich DS, Calabresi PA, Nath A(2006). Immune reconstitution inflammatory syndromein the CNS of HIV-infected patients. Neurology 67:383–388.

WHO/UNAIDS (2008). Report on the Global AIDS Epide-mic. Epub: http://www.unaids.org/en/KnowledgeCentre/HIVData/GlobalReport/2008/2008_Global_report.asp

Williams PL, Marino M, Malee K, Brogly S, Hughes MD,Mofenson LM (2010). Neurodevelopment and in uteroantiretroviral exposure of HIV-exposed uninfectedinfants. Pediatrics 125: e250–e260.

This paper was first published online on Early Online on 25 May 2010.

NeuroAIDS in Africa 2009

202 K Robertson et al

J N

euro

viro

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Nat

iona

l Ins

titut

es o

f H

ealth

Lib

rary

on

09/0

2/11

For

pers

onal

use

onl

y.