19

Human Immunodeficiency Virus–Associated Dementia: Review of Pathogenesis,

Prophylaxis, and Treatment Studies of Zidovudine Therapy
David M. Simpson From the Department of Neurology, Neuro-AIDS Research Program,
Department of Clinical Neurophysiology, Mount Sinai Medical Center,
New York, New York

Human immunodeficiency virus (HIV)–associated dementia (HIVD) has been reported in up to

15% of HIV-infected adult patients. Although the pathogenesis of HIVD remains unclear, HIV
probably plays an important role in the syndrome, as evidenced by the correlation between cere-
brospinal fluid (CSF) HIV load and neuropsychological functioning. Although a large number of
antiretrovirals are used to treat HIVD, zidovudine is the best studied. Zidovudine therapy has been
associated with reduced levels of HIV RNA in CSF, fewer HIV-related changes in brain tissue at
autopsy, and time-limited improvements in neurological function among AIDS and HIVD patients.
More recent studies have investigated the penetration into CSF of other antiretrovirals, including
protease inhibitors, and the clinical efficacy of abacavir in the treatment of dementia. HIV enceph-
alopathy may occur in 30%– 60% of children with AIDS and causes significant disability. Zidovudine
has been associated with improved neuropsychological functioning in children with progressive
encephalopathy, but optimum dosing levels, duration of effect, and prophylactic potential remain to
be demonstrated.

Neurological complications are among the most frequent and
detrimental of the problems that affect HIV-infected patients
[1]. Both children and adults can have such complications at
any stage of the illness [1], with 40%–70% of all HIV-infected
persons developing symptomatic neurological disorders at
some point [2]. Pharmacotherapies now available not only
allow HIV patients to live longer but also have changed the
constellation and sequence of complications most commonly
associated with AIDS. Thus, occurrence of Pneumocystis ca-
rinii pneumonia and other opportunistic infections is often
delayed, whereas neurological disorders increasingly appear as
the initial AIDS manifestation [3].
In addition to secondary infectious or neoplastic neurologi-
cal complications resulting from immunosuppression, several
primary neurological disorders may be associated with HIV-1.
Among these are myelopathy, myopathy, and distal sensory
polyneuropathy [2]. Disabling cognitive, behavioral, and motor
impairment, termed HIV-associated dementia (HIVD) in adults
and HIV encephalopathy in children, is one of the most dev-
astating and enigmatic of the primary HIV neurological com-
This article is part of a series of papers presented at a symposium entitled
“New Directions in HIV Therapy and Quality Survival” that was held on 1 July
1998 in Geneva, Switzerland, in conjunction with the 12th World AIDS Con-
ference. The symposium was supported by an unrestricted educational grant
from Ortho Biotech Inc., Raritan, New Jersey. CME supervision and accredi-
tation are being provided by Cedars-Sinai Medical Center, Los Angeles.
Reprints or correspondence: Dr. David M. Simpson, Department of Neurol-
ogy, Neuro-AIDS Research Program, Department of Clinical Neurophysiology,
Box 1052, Mount Sinai Medical Center, One Gustave L. Levy Place, New York,
New York 10029 (DSimpson@smtplink.mssm.edu).
Clinical Infectious Diseases 1999;29:19 –34
© 1999 by the Infectious Diseases Society of America. All rights reserved.
1058 – 4838/99/2901– 0004$03.00
plications. (HIVD also is sometimes called AIDS dementia
complex [ADC], multinucleate giant cell encephalitis, or HIV-
1-associated cognitive/motor complex [2]. For simplicity, all
are referred to here as HIVD.) This article summarizes the
epidemiology, clinical features, pathogenesis, and treatment of
HIVD in adults and children.
HIV Dementia (HIVD) in Adults
Epidemiology
In the United States, neurological disease is recorded for as
many as 100,000 HIV-infected patients each year [2]. Reports
suggest an incidence of HIVD of 7% per year among survivors
in the first 2 years of the disease, with 15% of the cohort
developing dementia before death [4]. The actual occurrence
may be higher than these figures indicate, because many cases
of HIVD may not be diagnosed when other life-threatening
illnesses are present [2]. In fact, some autopsy series show
neuropathologic abnormalities in as many as 80% of AIDS
patients [5]. It appears that the incidence of HIVD may have
fallen in recent years with the advent of highly active antiret-
roviral therapy, although epidemiological data are not yet
available to support this impression.
Although HIVD occasionally is an AIDS-defining illness
[2, 6], it usually develops late in the disease after the patient
has been diagnosed with other AIDS-related illnesses and
when CD41 lymphocyte counts have declined to ,200/mm3
[7]. A prospective study of 492 homosexual men with AIDS
identified risk factors associated with more rapid develop-
ment of dementia and documented median survival times for
those developing vs. those not developing HIVD [4]. Pri-
mary risk factors for rapid development of HIVD included
20 Simpson
decreased hemoglobin levels in the year before AIDS was
diagnosed, low body mass index in the 1– 6 months before
AIDS, self-reported constitutional symptoms 7–12 months
before AIDS, older age at initial AIDS diagnosis, and Ka-
posi’s sarcoma as the initial AIDS-related diagnosis. Mul-
tivariate analyses identified decreased hemoglobin level in
the 1– 6 months before the AIDS diagnosis as the most
significant predictor of dementia (relative hazard 5 0.60 for
each additional 2 g/dL [P 5 .03]). Neither demographic
characteristics, AIDS-defining illness, zidovudine use be-
fore AIDS, nor CD41 lymphocyte count before AIDS was
an important predictor in this cohort. Importantly, patients
who developed HIVD had shorter median survival times
than did those who did not develop the complication: 6.0
months among demented patients vs. 7.8 months among
nondemented patients after a second AIDS-defining illness
(P 5 .075) [4].
Clinical Features of HIVD
HIVD is largely a diagnosis of exclusion, with early signs and
symptoms that may be difficult to recognize [2]. Clinical symp-
toms fall into three categories: cognitive, behavioral, and motor [8,
9]. Cognitive symptoms include impaired short-term memory and
concentration, increased distractibility, mental slowing, and loss of
flexibility and spontaneity. Personality change, apathy, with-
drawal, irritability, and depression often characterize behavioral
changes. Motor symptoms may be evidenced by fine-motor clum-
siness or slowness, tremor, and leg weakness.
Results of neurological examination often are normal or
nearly so in patients with early HIVD [9]. In some patients,
cognitive impairment may be detected with the Luria hand
test or word reversal tests. A series of neuropsychological
tests measuring several domains of cognitive function are
being assessed by investigators within the AIDS Clinical
Trials Group (ACTG) of the U.S. National Institute of
Allergy and Infectious Diseases. There may be impairment
of smooth eye pursuits or saccadic movements, while slow
or clumsy movements may be detected by tests involving
finger tapping or alternating wrist movements. There also
may be mild hyperreflexia and difficulty with tandem gait.
As the disease progresses, major intellectual and motor
impairments become evident, and in the later stages, patients
enter a conscious but vegetative state. To grade the severity
of HIVD symptoms, a system based on functional perfor-
mance has been developed and has been used widely in
clinical trials (table 1) [10]. Scoring ranges from 0 (normal)
and 0.5 (subclinical neurocognitive symptoms) to 4 (nearly
vegetative and mute).
Although HIVD cannot be definitively diagnosed by use of
laboratory and radiographic results, studies can help rule out
alternative diagnoses and indicate whether HIVD is likely. For
example, studies of blood and CSF can be used to screen for
CID 1999;29 (July)
Table 1. Staging of HIV dementia on the basis of functional per-
formance.
Stage Clinical features
0 (normal) Normal mental and motor function
0.5 (equivocal/ Absent, minimal, or equivocal symptoms without
subclinical) impairment of work or capacity to perform
ADL; mild signs (snout response, slowed
ocular or extremity movements) may be
present; gait and strength are normal
1 (mild) Able to perform all but the more demanding
aspects of work or ADL, but with unequivocal
evidence (signs or symptoms that may include
performance on neuropsychological testing) of
functional intellectual or motor impairment;
can walk without assistance
2 (moderate) Able to perform basic activities of self-care but
cannot work or maintain the more demanding
aspects of daily life; ambulatory, but may
require a single prop
3 (severe) Major intellectual incapacity (cannot follow
news or personal events, cannot sustain
complex conversation, considerable slowing of
all output) or motor disability (cannot walk
unassisted, requiring walker or personal
support, usually with slowing and clumsiness
of arms as well)
4 (end-stage) Nearly vegetative; intellectual and social
comprehension and output are at a
rudimentary level; nearly or absolutely mute;
paraparetic or paraplegic with urinary and
fecal incontinence
NOTE. ADL 5 activities of daily living. From [10].
other infections [2]. CSF abnormalities often seen include
mild mononuclear cell pleocytosis (counts usually of ,50
cells/mm3, seen in one-fifth of patients), elevated total immu-
noglobulin fraction, increased total protein concentration (usu-
ally to ,200 mg/dL, seen in two-thirds of patients), oligoclonal
bands, and intrathecal synthesis of IgG antibody to HIV [11].
In later stages of the disease, markers of immune activation in
CSF, including b2-microglobulin, neopterin, and quinolinate,
generally are elevated [2]. Finally, detection of HIV-1 p24
antigen in CSF, which is independent of such antigen in serum,
is correlated with HIVD [12].
Radiological studies (brain CT, MRI) are needed to exclude
other infections or neoplasms and to identify cerebral atrophy
and changes in white matter. Primary neuroradiological fea-
tures of HIVD include global cerebral atrophy, proportionate
ventricular enlargement, symmetrical abnormalities of the
white matter (leukoencephalopathy), and vascular mineraliza-
tion in children [2] (figure 1). There appears to be a correlation
between the amount of cerebral atrophy seen on magnetic
imaging of the brain and severity of dementia [14]. Current
studies are investigating the utility of magnetic resonance spec-
troscopy in the diagnosis of HIVD.
CID 1999;29 (July) HIV-Associated Dementia 21

Figure 1. Neuroradiological features of HIV dementia (HIVD). A, CT scan of brain in HIVD reveals dilatation of cortical sulci (closed arrow)
and ventricular system (open arrow). B, MRI of brain (T2-weighted image) demonstrates bilateral white matter hyperintensities (termed
leukoencephalopathy; arrows). R 5 right; L 5 left. From [13] (used with permission).

Neuropathologic Findings
Characteristic changes in the brains of persons with HIVD
can exist singly or overlap. The most common finding is
cerebral atrophy with resultant ventricular enlargement, wid-
ened sulci, and loss of white matter (as indicated by symmet-
rical hyperintensities on T2-weighted MRI) [6] (figure 2). Neu-
ropathologic findings generally reveal the following.
HIV-1 encephalitis. Distinct pockets of inflammatory cells,
such as microglia, macrophages, and multinucleated giant cells
formed by fusion of microglia and macrophages, are found in
white and deep grey matter as well as in the cortex [15].
Although these findings are specific for the diagnosis of HIVD,
they are present in only 50% of demented patients.
HIV leukoencephalopathy. HIV leukoencephalopathy is
characterized by diffuse damage to white matter or myelin
attenuation, astrogliosis, macrophages, and multinucleated gi-
ant cells—without distinct pockets of inflammation. Myelin
pallor in young children may be difficult to distinguish because
of their incomplete myelination [15].
Diffuse poliodystrophy. Reactive astrogliosis and micro-
glial activation in cerebral gray matter may possibly be asso-
ciated with neuron loss or dendritic damage. Severe cases may
also evidence spongiform changes [15].
Vascular calcification of the basal ganglia and possibly the
centrum semiovale is often present in pediatric AIDS. In fact,
it may be found either with or without notable HIV-1 enceph-
alitis. In contrast, inflammatory changes are more common in
the spinal cord in pediatric patients, whereas vacuolar myelop-
athy is seen less often. However, corticospinal tract degenera-
tion may also be found in pediatric HIV-1 infection [15].
Theories on Pathogenesis
The pathogenesis of HIVD remains unclear, and it is not
known whether HIV plays a primary or secondary role in
prompting CNS deterioration [16, 17]. Historically, HIV anti-
gen and HIV-1 DNA have been demonstrated in the brains of
patients with HIVD. These are predominantly localized in
lymphoreticular cells, including monocytes/macrophages and
endothelial cells; it appears to be these cells, rather than neu-
rons or glial cells, that are infected [2].
In general, the severity of certain pathological findings cor-
relates with clinical state [18, 19]. There is no convincing
evidence of neuronal HIV infection [2]. b2-microglobulin and
neopterin levels in CSF and CD41 cell counts have been found
to be predictive markers for the development of HIVD [20]. A
22
Figure 2. Neuropathologic findings in HIV dementia. A, HIV en-
cephalitis is characterized by a microglial nodule inflammatory infil-
trate with multinucleate giant cells (arrow; hematoxylin-eosin stain;
bar 5 18 mm). B, HIV is located within these nodules, which are
present in subcortical structures and are often accompanied by white
matter pallor (arrow; Luxol fast blue stain for myelin; bar 5 100
mm). From [13] (used with permission).
number of indirect factors may contribute to the pathogenesis
of HIVD, including cytokines (e.g., TNF), parts of HIV itself
(gp41, gp120, Tat, Rev, Nef), and excitatory amino acids (e.g.,
quinolinic acid). Portegies et al. [12] found that although
HIV-1 antigen expression in CSF accompanied neurological
deterioration, it was not a good predictor of dementia, because
it was not detectable before such deterioration was clinically
evident. Newer data, based on assays of CSF for HIV RNA and
tests of virus load in CSF, show a better correlation between
HIV RNA levels in CSF and the severity of dementia.
McArthur et al. [21] reported that in patients with advanced
disease (CD41 cell counts of ,200 cells/mm3), virus loads in
CSF and brain were correlated, indicating that virus burden in
CSF may be a valid marker for virus burden in the brain.
Additionally, patients with dementia had significantly elevated
virus load in CSF compared with that in nondemented subjects.
Ellis et al. [22] reported a similar pattern in patients with
advanced HIV disease. The major issue concerning HIV RNA
Simpson CID 1999;29 (July)
in CSF, however, remains how well it correlates with response
to therapy.
In light of this evidence, numerous indirect mechanisms for
HIV’s effect on the brain have been suggested.
Cytokines, particularly TNF-a and interleukins, elaborated
by infiltrating macrophages and microglia, serve as cofactors
for dementia in a number of other models [23]. These cytokines
may be causative factors of both the vacuolar myelopathy and
sensory neuropathy seen in HIVD [2].
White matter pallor, rather than resulting from demyelina-
tion, may be a consequence of HIV-1 induced alterations in the
blood-brain barrier [24].
gp120, an HIV surface antigen, can cause neuronal death
in the presence of microglial cells. This antigen has been
related to the opening of calcium channels in neuronal
membranes [25].
Possible amplification of HIV’s effects by interactions be-
tween HIV-infected monocytes and astrocytes has been sug-
gested [2]. These interactions might release neurotoxic factors
and enhance glial proliferation.
Activation of N-methyl-D-aspartate (NMDA) receptors
may be the final common pathway in HIVD, as it is in other
neurodegenerative processes [26 –28]. In vitro studies show
that gp120 neurotoxicity can be blocked by NMDA antag-
onists [29].
Chemokines are proteins involved in the trafficking of cells
and their interaction with receptors. Several chemokines, in-
cluding CCR-5 and CXCR4 for neurons, provide another
mechanism by which HIV may interact with endothelial cells
and neurons [30].
Treatment
Although the pathophysiology of HIVD is still unclear, HIV
probably plays an important role in the syndrome. Thus, the
ability of therapeutic agents to penetrate the blood-brain barrier
becomes a key consideration in the choice of therapy [31].
Among the major nucleoside reverse transcriptase inhibitors,
penetration of brain tissue has been reported to be ;20% with
zidovudine, ;30% with stavudine, and ;2%– 4% with di-
danosine [32]. No studies concerning the penetration of brain
tissue by lamivudine are available; however, zidovudine is the
most lipophilic nucleoside analogue. To date, zidovudine and
lamivudine have shown the highest ratios of CSF drug concen-
tration to IC50 (the minimum drug concentration that produces
Ä50% inhibition of HIV activity), although early studies indi-
cate that indinavir and nevirapine also are promising. Insuffi-
cient data exist to evaluate the protease inhibitors and non-
nucleoside reverse transcriptase inhibitors with respect to
penetration.
Hydroxyurea (an agent used variously to treat cancer and sickle
cell disease [33]) penetrates the blood-brain barrier well, with
concentrations in brain of ;25% of concentrations in plasma [34].
CID 1999;29 (July) HIV-Associated Dementia
The addition of hydroxyurea to didanosine-containing regimens
has been shown to enhance the in vitro anti-HIV activity of
didanosine [35]. Whereas didanosine targets viral protein, hy-
droxyurea targets cellular protein, which is less vulnerable to
development of resistance. Hydroxyurea appears to lower cellular
levels of dATP, rendering DNA synthesis more susceptible to
didanosine, a dATP competitor [36].
In general, information concerning the ability of various
therapeutic agents to cross the blood-brain barrier is incom-
plete; however, in vitro blood-brain barrier permeability to
such agents recently has been reported, in descending order, as
follows: nevirapine; didanosine, stavudine, zalcitabine, and
zidovudine; indinavir; and saquinavir [37]. Abacavir also ap-
pears to have good penetration [38]. Combining drugs does not
substantially improve permeability [31]. On the basis of studies
in humans, the ordering of the ratio of drug concentration in
CSF to that in plasma for nucleoside antiretrovirals appears to
be as follows, in descending order: zidovudine, stavudine,
abacavir, didanosine, lamivudine, zalcitabine. The ratio for the
nonnucleoside reverse transcriptase inhibitor nevirapine falls
approximately midway in this continuum, while those for dela-
viridine and efavirenz fall below that for zalcitabine [31]. As a
class, protease inhibitors are highly protein-bound and do not
cross the endothelial barrier well, although indinavir, the least
protein-bound, has the greatest potential for free drug to cross
the blood-brain barrier [39]. Preliminary data suggest that
protease inhibitors, when used in combination therapies, may
improve or stabilize the neurological condition of some pa-
tients [40, 41]. Ferrando et al. [42] followed neuropsychomet-
ric performance in cohorts of patients receiving and not receiv-
ing highly active antiretroviral therapy and found that those
receiving highly active antiretroviral therapy had significantly
better performance on several measures.
Additional information concerning the movement of zidovu-
dine into CSF was provided by a study that measured levels of
zidovudine in CSF by means of an implanted spinal catheter;
the area under the curve for drug in CSF to drug in plasma was
reported to be 75% [43]. Although levels of protease inhibitors
in CSF and brain are believed to be generally quite low because
the inhibitors are highly protein-bound (60% for indinavir and
98% for saquinavir and ritonavir) [31], there is some evidence
that their use may result in regression of abnormalities in
periventricular white matter and basal ganglia signal intensity
in HIV encephalopathy [44].
Despite the fact that research has focused primarily on the
ability of antiretrovirals to cross the blood-brain barrier, it has
not been demonstrated, even for those drugs that have good
blood-brain barrier penetration and reduce levels of HIV RNA
in CSF, that any such reduction correlates with improved
neuropsychological outcome. It is reasonable at the present
time, however, to assume that this achievement is likely, be-
cause studies are ongoing. Furthermore, the use of an antiret-
roviral that does not penetrate the blood-brain barrier has been
23
found to increase the likelihood of different populations of
virus growing within the systemic tissues and CNS [45].
Although their precise mechanism(s) of action has yet to be
defined, there is evidence that antiretrovirals decrease levels of
p24 antigen in CSF, an indirect marker for virus load [46 – 48].
Royal et al. [47] found that patients with HIVD who reported
taking antiretroviral medication had p24 antigen in CSF less
frequently and at lower concentrations than did those who did
not take such drugs. A retrospective study of a consecutive
series of 196 AIDS patients in The Netherlands from 1982 to
1988 found significant declines in both HIVD and the presence
of p24 antigen in CSF among those treated with zidovudine
[46]. Heyes et al. [49] studied quinolinic acid (a potential
neurotoxin that is an NMDA receptor agonist) in CSF in 11
HIV-1-infected patients. After treatment with zidovudine for at
least 4 weeks, patients exhibited an 11-fold reduction in quino-
linic acid levels in CSF and improved neurological status, as
indicated by scores in neuropsychological tests and reduced
severity of HIVD [49].
Given the difficulties in mounting long-term trials of demen-
tia, several recent studies have focused on the ability of anti-
retroviral agents to reduce levels of HIV RNA in CSF. Foud-
raine et al. [50] reported that the combination of stavudine and
lamivudine was as effective as zidovudine plus lamivudine in
reducing HIV RNA levels in CSF. Collier et al. [51] reported
that virus was undetectable in CSF from 9 of 10 patients treated
with indinavir in addition to background therapy.
Over the past 10 years, there have been numerous studies of
the effects of antiretroviral agents on HIVD. Given zidovu-
dine’s longevity and widespread use among HIV-infected pa-
tients, it is not surprising that the most complete information
based on the results of clinical trials in humans is available for
zidovudine [48]. Preliminary studies of the efficacy of newer
antiretrovirals are, however, beginning to be published. For
example, a recent controlled study of abacavir added to “best
background antiretroviral therapy” revealed that no significant
additional neuropsychological improvement was provided by
abacavir compared with placebo [52]. These results are not
particularly surprising, because the patients included generally
had advanced disease, and many had recently had protease
inhibitors added to their treatment regimens.
The following summarizes results of studies of the effect of
zidovudine on HIVD in adults with respect to clinical efficacy,
dosage, duration of effectiveness, and prophylactic effect (table
2) [4, 46, 53– 66]. Results of autopsy studies of the use of
zidovudine for HIV encephalitis follow. Studies of combina-
tion therapies that included zidovudine generally are not re-
viewed because it is often difficult to separate the effects of the
individual agents in these reports.
Studies of clinical efficacy. In a 1987 study, Yarchoan et al.
[53] followed four patients (two with HIVD, one with HIVD
and peripheral neuropathy, and one with paraplegia) taking
zidovudine. In three cases, zidovudine at 1,500 mg/day was
24 Simpson CID 1999;29 (July)

Table 2. Studies of zidovudine in adults with HIV dementia.

Type of study, HIVD-related

reference n Study design Duration Diagnosis Daily AZT dose outcomes Results

Studies of clinical
efficacy
Yarchoan et al. [53] 4 Case reports Unknown 2, HIVD; 1, HIVD 1,500 mg orally Cognition; 3 patients with HIVD showed improvement;
plus neuropathy; (n 5 3); 5 motor 1 patient with paraplegia did not
1, paraplegia mg/kg iv q4h skills
(n 5 1)
Schmitt et al. [54] 262 R, DB, PC Planned for 150, AIDS; 112, 1,500 mg (n 5 Affect; Significant improvement in AZT group in
trial 24 w; ARC 134); cognition; cognition and motor areas; no difference
ended at placebo (n motor in affect; study ended early because of
16 w 5 128) skills decreased mortality in AZT group
Portegies et al. [55] 40 Retrospective 1–32 mo HIVD 400–1,200 mg Cognition; 3 patients receiving AZT showed marked
(n 5 10); motor improvement for 16–32 mo; 2 patients
no AZT (n skills receiving AZT showed slight
5 20); improvement
unknown (n
5 10)
Nordic Medical 474 R, DB, MC Median, 19 AIDS (n 5 126); 400 mg (n 5 Quality of life Fewer cases of HIVD at higher doses
Research Council trial mo HIV symptoms 160); 800 (P 5 .06); no difference in quality of
[56] (n 5 248); mg (n 5 life measures
low CD41 cell 158); 1,200
count (n 5 mg (n 5
100) 156)
Sidtis et al. [57] 40 R, DB, PC, 64 w HIVD 1,000 mg (n 5 Cognitive; Averaged across tests, found significant
MC trial 12); 2,000 motor improvement in 2,000 mg group vs.
mg (n 5 skills placebo; 1,000 mg group intermediate
13); placebo
(n 5 15)
Tozzi et al. [58] 30 Open 12 mo HIVD 500 mg (n 5 Cognitive; 73% showed some improvement within 3
13); 750 mg motor mo and 83% after 6 mo; AZT dose not
(n 5 8); skills correlated with response; 8 patients had
1,000 mg (n neurological deterioration after 6–12 mo
5 9) of treatment
Brouwers et al. [59] 38 R trial 12 w CNS compromised, Simultaneous Cognitive; Improvement seen in both groups; no
21 (12, HIVD); vs. motor overall difference between groups
normal, 17 alternating skills
AZT and ddI;
alternating
doses: 600
mg of AZT,
500 mg of
ddI
Studies of duration of
effect*
Reinvang et al. [60] 11 Open 12 mo HIVD (n 5 6); 800–2,000 mg Cognitive Improvement in HIVD group after 4–6 mo
AIDS (n 5 5) of treatment; deterioration after 9–12 mo
Karlsen et al. [61] 49 Open 30.3 mo AIDS (n 5 33); AZT: 24 with Cognitive; Improvement in AIDS group receiving AZT
HIV (n 5 16) AIDS; no motor at 6 months; decline in functioning at 12
AZT: 9 with skills months in AZT group
AIDS, 16
with HIV
infection
Studies of prophylactic
effect
Portegies et al. [46] 196 Retrospective 6y AIDS and AZT (n 5 Diagnosis of Decreased incidence of HIVD paralleled
neurological 89); no AZT HIVD; p24 increased use of AZT; significantly fewer
symptoms (n 5 107) antigen in patients treated with AZT had HIVD
CSF (P , .0001) or p24 antigen in CSF
(P 5 .0002)
Volberding et al. 1,338 R, DB, PC Mean, 55 w Asymptomatic 500 mg (n 5 Development HIVD developed in 2 patients in placebo
[62] trial HIV; no HIVD 453); 1,500 of HIVD group, 1 patient in 1,500 mg group, 0
mg (n 5 patients in 500 mg group
457);
placebo (n
5 428)
CID 1999;29 (July) HIV-Associated Dementia 25

Table 2. (Continued)

Type of study, HIVD-related

reference n Study design Duration Diagnosis Daily AZT dose outcomes Results

Hamilton et al. [63] 328 R, DB, MC Mean, 27– Asymptomatic 1,500 mg early Development HIVD developed in 6 patients in late
trial 28 mo HIV; no HIVD (n 5 170) of HIVD therapy group vs. 0 in early therapy
or late (n 5 group
168)
Concorde 1,749 R, DB, PC Median, 3.3 Asymptomatic HIV 1,000 mg early Development HIVD was AIDS-defining condition in 6
Coordinating trial y (n 5 877) of HIVD patients in the early vs. 7 in the deferred
Committee [64] or deferred (n treatment groups
5 872)
McArthur et al. [4] 492 Prospective, Up to 7 y Homosexual men Varied; mean, Development AZT use before or after AIDS developed
open with AIDS ,600 mg of HIVD did not predict development of HIVD
Bacellar et al. [65] 2,641 Longitudinal 5y HIV Varied 5-y HIVD Incidence of HIVD did not change
incidence; significantly; use of any antiretroviral not
development protective against development of HIVD
of HIVD
Baldeweg et al. [66] 141; long- Natural Median, 1 y Current AZT; Median, 500 Cognitive; No effect of current AZT use on cognition;
term, 98 history asymptomatic mg; current motor fewer neurological problems in AIDS
HIV (n 5 60); AZT (n 5 skills; patients taking AZT; increased cognitive
symptomatic 67); AZT for neurological status in patients with symptomatic HIV
HIV (n 5 51); Ä1 y (n 5 or AIDS using AZT for Ä1 y compared
AIDS (n 5 37); AZT for with those using AZT for ,1 mo
30) ,1 mo (n
5 61)

NOTE. ARC 5 AIDS-related complex; AZT 5 zidovudine; DB 5 double blind; ddI 5 didanosine; HIVD 5 HIV dementia; MC 5 multicenter; PC 5 placebo
controlled; R 5 randomized. Reprinted with permission from The Annals of Pharmacotherapy [1].
* See also [55, 58].

administered orally, while one patient received the drug via iv
infusion. Neurophysiological tests, including nerve conduction
studies and positron emission tomography, as well as tests of
cognition and motor skills, were conducted before and during
administration of zidovudine. The authors reported that the
condition of the three patients with HIVD improved while that
of the patient with paraplegia did not. Notably, paraplegia is a
symptom more commonly associated with myelopathy than
with HIVD.
A randomized, double-blind trial of zidovudine (1,500
mg/day, n 5 134) vs. placebo (n 5 128) included 150 patients
with AIDS and 112 patients with AIDS-related complex [54].
The study was planned for 24 weeks but was discontinued after
16 weeks when significantly decreased mortality in the zidovu-
dine group was detected. Among neuropsychological out-
comes, cognition (attention and memory) and motor skills were
found to be significantly improved at both weeks 8 and 16
among patients receiving zidovudine (P , .05) and especially
among AIDS patients receiving the drug. No overall significant
difference in affect was detected between patients in the
zidovudine and placebo groups. Although findings are limited
by the short duration of the study, the authors concluded that
zidovudine can reduce HIV-associated neuropsychological ab-
normalities.
Records of 536 HIV-infected patients with neurological
symptoms were retrospectively reviewed at the Academic
Medical Centre in Amsterdam between 1982 and 1992 [55].
Forty patients met U.S. Centers for Disease Control and Pre-
vention criteria for HIVD, and 39 of these were not taking
zidovudine at the time of diagnosis. Twenty of the 40 never
received zidovudine, 10 received zidovudine at unreported
stages of the disease, and 10 began receiving the drug at
dosages of 400 –1,200 mg/day after being diagnosed with
HIVD. Of these latter 10, 3 showed marked improvement with
respect to cognition and motor skills (by at least one HIVD
stage), with improvement lasting for 16, 24, and 32 months; 2
additional patients in this group showed slight improvement.
Although the authors concluded that zidovudine may improve
symptoms in patients with HIVD, the retrospective design,
small sample size, and variable dosage of zidovudine limit the
generalizability of the results.
In summary, although there are limitations in trial designs,
the results of these early studies suggest that zidovudine can
result in significant improvements in cognition and motor skills
in some patients with HIVD. The dosage of zidovudine needed
to attain these effects has been the subject of additional re-
search.
Studies of dosage. The Nordic Medical Research Council’s
HIV Therapy Group conducted a randomized, double-blind,
parallel-group, multicenter trial comparing the outcomes of
474 patients with AIDS (n 5 126), HIV-related symptoms
(n 5 248), or low CD41 cell counts (n 5 100) who received
daily doses of either 400 mg (n 5 160), 800 mg (n 5 158), or
1,200 mg (n 5 156) of zidovudine for a median of 19 months
[56]. The quality of life was evaluated monthly by both the
patient and the physician; laboratory and clinical evaluations
also were done monthly. Although HIVD was not a primary
outcome variable, fewer cases of HIVD were diagnosed among
26 Simpson
patients receiving higher dosages of zidovudine (8%, 6%, and
3% for dosages of 400 mg/day, 800 mg/day, and 1,200 mg/day,
respectively; P 5 .06). No significant differences among
groups were demonstrated in quality of life assessed by either
physicians or patients. Because the incidences of anemia and
leukopenia were dose-related, the authors concluded that
zidovudine should be limited to daily doses of between 400 mg
and 600 mg in AIDS patients, in spite of the difference in the
incidence of dementia as noted above.
The ACTG conducted a randomized, double-blind, placebo-
controlled, multicenter trial of zidovudine treatment of HIVD
[57]. Forty patients were randomly assigned to receive either
1,000 mg (n 5 12) or 2,000 mg (n 5 13) of zidovudine or
placebo (n 5 15) and were followed for an average of 64
weeks. Neuropsychological tests were done at baseline and
every 4 weeks. When the data were averaged across tests,
patients taking zidovudine were found to have significant im-
provement in symptoms at 16 weeks compared with those in
patients taking placebo (P 5 .046). With respect to pairwise
comparisons, however, only the comparison between the
groups receiving zidovudine at 2,000 mg/day and placebo was
significant, with an intermediate level of improvement noted in
the group receiving 1,000 mg/day. Patients from the placebo
group who were re-randomized to one of the two zidovudine
groups (n 5 12) showed significant improvement across neu-
ropsychological tests between weeks 16 and 32 (P , .01);
substantial loss to follow-up during the final 32 weeks of the
study limited analyses. Although the authors concluded that
zidovudine treatment benefits patients with HIVD, the study
was limited by its small sample size, high dropout rate, and
failure to demonstrate a clear optimal dose of zidovudine.
An open study of 30 consecutive patients with HIVD fo-
cused on the effect of long-term zidovudine treatment on
severity of dementia [58]. The sample was stratified by base-
line hematologic status and absolute neutrophil count, and
patients were assigned to receive one of three daily zidovudine
doses: 500 mg (n 5 13), 750 mg (n 5 8), or 1,000 mg (n 5 9).
Reversal to a less severe HIVD category, as determined by use
of Price and Brew’s staging system, was observed after 1, 3, 6,
9, and 12 months for 15, 22, 25, 19, and 14 patients, respec-
tively. Overall, 73% of patients showed some improvement
within 3 months and 83% after 6 months of treatment. Al-
though zidovudine dose was not correlated with response,
among the eight patients having neurological deterioration after
6 –12 months of drug therapy, five were receiving the lowest
dosage of zidovudine (500 mg daily). Reduced dosages of
zidovudine were administered to eight patients who developed
hematologic toxicity after an average of 4.5 months of treat-
ment and to an additional five patients who developed bone
marrow toxicity after an average of 5 months of treatment.
Despite the fact that the study was limited by its lack of a
control group and by its small sample size, the authors con-
CID 1999;29 (July)
cluded that use of zidovudine was associated with a reversal of
neurological dysfunction in patients with HIVD.
The effect of treatment with alternating vs. simultaneous
regimens of zidovudine and didanosine on the cognitive and
motor skills of patients with symptomatic HIV infections was
studied in a randomized, unblinded trial that evaluated 38
patients [59]. Twenty-one patients (12 with HIVD) exhibited
CNS symptoms. Patients in the simultaneous treatment group
received 300 mg of zidovudine and 250 mg of didanosine daily
for 12 weeks; both doses were approximately half of that
usually recommended. Remaining patients alternated between
600 mg of zidovudine daily for 3 weeks or 500 mg of di-
danosine daily for 3 weeks. Significant improvement in neu-
ropsychological function, particularly memory and attention,
was seen with both therapeutic regimens (P , .01), and pa-
tients with CNS symptoms at the beginning of the trial showed
the most improvement. The authors concluded that combina-
tion regimens of zidovudine and didanosine reduce HIV-
induced CNS symptoms, while noting that the small sample
and relatively low dosages of drugs administered limited the
generalizability of results.
In summary, although there does appear to be a dose-
response effect with zidovudine, the optimum dosage of the
drug in patients with or at risk for HIVD is not clearly dem-
onstrated by these studies. Controlled trials are needed to
evaluate the relative effects of various drug dosages, particu-
larly in the lower-dose regimens conventionally used in current
practice.
Studies of duration of effect. The duration of treatment
effect with zidovudine therapy was evaluated as part of more
comprehensive studies by Portegies et al. [55] and Tozzi et al.
[58]. In the former, a retrospective study of 40 patients with
HIVD, 5 of the 10 patients who began zidovudine treatment
after HIVD was diagnosed showed improvement in symptoms,
with 3 showing notable improvement that lasted for 16, 24, and
32 months. Similarly, Tozzi et al. observed that 8 (31%) of the
26 patients who showed some improvement in neuropsycho-
logical functioning had clinical relapses after 6 –12 months of
treatment.
Reinvang et al. [60] conducted a small, open study of 11
AIDS patients (6 with HIVD) being treated with zidovudine in
doses ranging from 800 mg to 2,000 mg daily. Overall, cog-
nitive functioning improved during the first 4 – 6 months of the
trial, although the trend was only “weakly significant.” This
improvement was not evident at the second follow-up after
9 –12 months of treatment. On the basis of these findings, the
authors expressed doubt concerning the long-term benefits of
zidovudine in the treatment of AIDS patients with HIVD. It is
notable, however, that the power of the study to detect differ-
ences was low, given the small sample and the fact that a wide
range of dosages of zidovudine was administered.
In a continuation of this study, 33 AIDS patients and 16
asymptomatic HIV-positive patients were followed for an av-
CID 1999;29 (July) HIV-Associated Dementia
erage of 30.3 months [61]. No asymptomatic patient received
zidovudine, while 24 of the 33 AIDS patients took the drug.
Zidovudine-treated AIDS patients showed significantly greater
improvement on neuropsychological tests administered 6
months after baseline than did either AIDS patients not treated
with zidovudine or asymptomatic patients who did not receive
the antiretroviral treatment (P , .05). By 12 months, however,
the mean performance of zidovudine-treated AIDS patients had
declined in all areas, and significantly so in verbal abilities
(P , .05). The authors concluded that the effect of zidovudine
appears to be transient, even though neither analyses of data for
the full follow-up period nor dosages of zidovudine were
provided.
Taken together, these studies suggest that the benefits of
zidovudine are time-limited. This tentative conclusion awaits
confirmation by the results of prospective, randomized,
placebo-controlled trials of the duration of effects of various
dosages of zidovudine as well as its interactions with other
antiretrovirals used in combination therapies.
Studies of prophylactic effect. There have been several
studies of the possible prophylactic effect of zidovudine that
used development of HIVD as the primary outcome variable. A
retrospective study of a consecutive series of AIDS patients
evaluated the incidence of HIVD and the presence of HIV-1
p24 antigen in CSF in relation to zidovudine treatment [46].
This research capitalized on the fact that zidovudine was not
available in The Netherlands before May 1987 but was used
routinely for patients with severe HIV infection after that time.
Of the 196 AIDS patients with neurological symptoms seen at
the AIDS unit at the Academic Medical Centre in Amsterdam
between 1982 and 1988, 40 (20%) had HIVD and 89 (45%)
received zidovudine. The incidence of HIVD decreased as the
use of zidovudine increased: the proportion of patients using
the drug was 26% at the end of June 1987 and 84% at the end
of 1988, while the incidence of HIVD was 21% in the last half
of 1985, 53% in the first half of 1987, 10% in the second half
of 1987, and 3% in 1988. Compared with patients who did not
take zidovudine, significantly fewer of those who took the drug
developed HIVD (38 [36%] of 107 vs. 2 [2.3%] of 89;
P , .0001) or had p24 antigen in their CSF (16 [26%] of 61 vs.
0 of 37; P 5 .0002). The authors concluded that the introduc-
tion of zidovudine in The Netherlands was accompanied by a
“striking decline” in HIVD. They speculated that zidovudine
may prevent HIVD by inhibiting reactivation of viral replica-
tion. It is important to note that although the authors state that
the same diagnostic criteria for HIVD were used throughout the
study, other parameters, such as concomitant drug therapy or
earlier detection of disease, may have changed over time and
affected the results of this retrospective study.
The ACTG conducted a randomized, double-blind, placebo-
controlled trial of the prophylactic effect of zidovudine [62].
The sample included 1,338 HIV-positive subjects with CD41
cell counts of ,500/mm3 and normal neuropsychiatric func-
27
tioning. Patients received either placebo (n 5 428) or 500 mg
(n 5 453) or 1,500 mg (n 5 457) of zidovudine daily and were
followed for an average of 55 weeks. Measures included the
drug’s ability to forestall a variety of complications, including
HIVD, in HIV-infected patients. HIVD developed in two pa-
tients in the placebo group, one in the group receiving 1,500
mg of zidovudine, and none in the group receiving 500 mg of
zidovudine. Rates of anemia and neutropenia were significantly
higher in the group receiving 1,500 mg of zidovudine than in
either the placebo group or the group receiving 500 mg of
zidovudine (P , .0001). The authors concluded that zidovu-
dine is safe and effective in this group of patients, although the
rates of development of HIVD were too low in all groups to
quantify a protective effect of zidovudine.
The effect of early vs. late administration of zidovudine has
been the focus of two studies [63, 64]. The Veterans Affairs
Cooperative Study Group on AIDS Treatment compared out-
comes in symptomatic HIV-positive patients with no HIVD
who received daily doses of 1,500 mg of zidovudine either
early (n 5 170; i.e., for the entire study period) or late
(n 5 168; i.e., after the patient’s CD41 cell count was
,200/mm3) in infection [63]. HIVD was identified in six
patients in the late-therapy group but in none of those receiving
early therapy (statistical significance not reported). Leukopenia
and anemia occurred at similar rates in both the early and late
groups (82% vs. 77% for leukopenia and 20% vs. 16% for
anemia). The authors concluded that use of zidovudine de-
serves consideration in both symptomatic and asymptomatic
HIV-positive patients with low CD41 cell counts.
In the other study of the timing of zidovudine therapy, the
Concorde Coordinating Committee randomly assigned 1,749
asymptomatic HIV-positive patients to receive 1,000 mg of
zidovudine daily early (n 5 877; i.e., from the time of ran-
domization) or to have treatment deferred (n 5 872) [64]. Of
those deferred, 418 patients eventually received zidovudine at
the onset of AIDS-related complex or AIDS or the develop-
ment of a persistently low CD41 cell count. Patients were
followed for a median of 3.3 years, during which HIVD was
the AIDS-defining illness in six patients of the early-
zidovudine group and in seven of the deferred-zidovudine
group. Anemia and neutropenia developed earlier in the early-
therapy group. The authors concluded that these results did not
support the early use of zidovudine for HIV-infected persons
who were free of symptoms.
A prospective, open study followed nearly 500 homosexual
men with AIDS for up to 7 years [4]. Zidovudine dosage was
variable but averaged ,600 mg daily. Although 15% of those
studied developed dementia before death, use of zidovudine
either before or after AIDS developed did not predict the
occurrence of HIVD. (Pre-AIDS hemoglobin level was the
most significant predictor of HIVD.) The authors suggested
that although zidovudine appears to delay the development of
AIDS, any additional protective effects against HIVD might be
28 Simpson
reduced once AIDS develops. They also noted, however, that
the average dosage of zidovudine in this study may have been
too low to provide prophylaxis.
The Multicenter AIDS Cohort Study, an ongoing, longitu-
dinal study, described trends in the incidence of HIV-related
neurological disorders in four metropolitan areas in the United
States: Baltimore-Washington, Chicago, Pittsburgh, and Los
Angeles [65]. Information concerning HIVD and sensory neu-
ropathy was analyzed for 2,641 men who were HIV-1-
seropositive. The incidence of HIVD did not change signifi-
cantly between 1988 and 1992 in this sample, and no protective
effect of antiretroviral use could be inferred. These results do
not confirm the decline in HIVD reported by Portegies et al.
[46]; however, because the major decrease in The Netherlands
occurred before 1988, the time frame of the Multicenter AIDS
Cohort Study (1988 –1992) may have resulted in its missing
any effect of zidovudine on the incidence of HIVD in these
U.S. cities.
Finally, Baldeweg et al. [66] retrospectively reviewed
records of 141 HIV-infected men who had received care for a
median of 1 year at HIV outpatient clinics. The sample was
stratified according to current (n 5 67) vs. not current (n 5 74)
zidovudine use and to whether the patient had asymptomatic
HIV (n 5 60), symptomatic HIV (n 5 51), or AIDS (n 5 30).
At a median zidovudine dose of 500 mg, lower neuropsycho-
logical functioning was associated with stage of HIV infection;
overall, no effect of current zidovudine use on neuropsycho-
logical functioning was found. Among symptomatic HIV-
infected and AIDS patients, significantly fewer of those cur-
rently taking zidovudine exhibited abnormalities detectable by
electroencephalography (P , .05). Of the 98 men who had
used zidovudine either for ,1 month (short-term group,
n 5 61) or for Ä1 year (long-term group, n 5 37), patients with
long-term zidovudine use with either symptomatic HIV infec-
tion or AIDS had significantly higher scores on tests of cog-
nition than did those in corresponding short-term groups
(P , .01). The authors concluded that long-term zidovudine
therapy may reduce neurological, neuropsychological, and
electrophysiological abnormalities, thereby decreasing the risk
for HIVD. The study is limited, however, by its natural history
design, which does not support an assessment of cause and
effect, and its exclusion of women.
Taken together, these studies suggest that zidovudine may
forestall the development of HIVD in some patients, particu-
larly when zidovudine is used before symptoms are evident
[63]. Large, prospective clinical trials that include both symp-
tomatic and asymptomatic patients treated at various stages of
illness are required to adequately address this question.
Autopsy studies. Results of autopsy studies of HIV enceph-
alitis support the efficacy of zidovudine (table 3) [18, 67–70].
Vago et al. [67] evaluated autopsy results for 202 patients who
died at the Clinic of Infectious Diseases, Sacco Hospital, in
Milan between 1984 and 1990. Overall, 82 patients had taken
CID 1999;29 (July)
zidovudine in doses of 600 –1,000 mg daily orally for 1–30
months. The incidences of multinucleated giant cells, HIV
leukoencephalopathy, and a history of severe dementia were
significantly lower in the zidovudine group (P ¶ .05). HIV
encephalitis and HIV encephalitis plus HIV leukoencephalop-
athy were found less frequently in the zidovudine group, but
the differences were not significant. The effects of zidovudine
were time- and dose-related, with the maximum effect in
patients treated for 6 –12 months and at cumulative doses of
.200 g. These results are similar to those of an autopsy study
in the United States of 56 patients who died of AIDS [18].
Among the 40 who had HIVD, significantly fewer of those
treated with antiretroviral medication for .12 months (n 5 10)
exhibited multinucleated giant cells and/or diffuse myelin pal-
lor than did those who had used the drugs for shorter time
frames (n 5 14) or not at all (n 5 16) (P ¶ .05).
Several studies evaluated the effect of taking zidovudine
until shortly before the patient’s death. Gray et al. [68]
examined the CNS of 192 patients dying of AIDS between
1982 and 1992. The prevalences of both multinucleated
giant cells and HIV encephalitis plus HIV leukoencephalop-
athy were significantly lower among patients who took
zidovudine for .3 months and until shortly before death
(n 5 72) than among patients not treated with zidovudine
(n 5 97) (P ¶ .01). The fact that the prevalence of markers
of HIV activity in the brain was intermediate for the group
who discontinued zidovudine treatment at least 1 month
before death (n 5 23) led the authors to conclude that
discontinuing zidovudine may increase the risk of HIV
encephalitis. Maehlen et al. [69] compared outcomes among
four groups of patients: those who had never taken zidovu-
dine (n 5 50) and those who took zidovudine for .2 months
and stopped treatment at ¶1 month (n 5 29) vs. at 2– 6
months (n 5 23) vs. at .6 months (n 5 26) before death.
The incidence of multinucleated giant cells was inversely
related to the length of time since zidovudine treatment was
discontinued (percentage with multinucleated giant cells of
7%, 30%, and 35% if zidovudine was discontinued at ¶1
month, 2– 6 months, and .6 months before death, respec-
tively) but was less closely associated with the presence of
either diffuse white matter damage or microglial nodules.
The authors concluded that zidovudine reduced the risk of
brain lesions only if treatment was continued until the
patient was near death. Finally, Bell et al. [70] evaluated
autopsy results for 66 patients who died of AIDS and
correlated use of zidovudine with HIV encephalitis and with
pre-death neurophysiological and neuropsychological tests.
Zidovudine use, particularly when it was used within 1 year
of death, was significantly associated with lower risks of
HIV encephalitis (P , .01) and of cognitive impairment
(P , .05). The effect on HIV encephalitis remained, even
when adjusted for the lifetime dosage of zidovudine.
CID 1999;29 (July) HIV-Associated Dementia 29

Table 3. Autopsy studies of zidovudine treatment among adults who died of AIDS.

Study time Dosage and duration

Reference n frame of AZT use Outcomes measured Results

Vago et al. 202 1984–1990 600–1,000 mg daily MGCs; HIVE; HIVL; HIVE MGCs and HIVL significantly lower in
[67] for 1–30 mo 1 HIVL; history of AZT group (P < .02); HIVE and HIVE
(n 5 82); no dementia 1 HIVL less frequent in AZT group
AZT (n 5 120) (NS); severe dementia significantly less
frequent in AZT group (P < .03); effects
of AZT time- and dose-related, with
maximum effect in patients treated for
6–12 mo and at cumulative doses of
.200 g
Glass et al. 56 (40 with HIVD) Not reported No antiretroviral, or MGCs; diffuse myelin Use of antiretroviral for .12 mo
[18] use for ,3 mo (n pallor; microglial nodules; significantly associated with fewer MGCs
5 16); perivascular cuffs and less diffuse myelin pallor (P < .05)
antiretroviral for
3–12 mo (n 5
14); antiretroviral
for .12 mo (n 5
10)
Gray et al. 192 1982–1992 No AZT (n 5 97); HIVE/HIVL; MGCs Prevalence of MGCs and of HIVE 1 HIVL
[68] AZT discontinued significantly lower in AZT-till-death
.1 mo before group than in no-AZT group (P < .01);
death (n 5 23); discontinuing AZT may increase risk of
AZT taken for HIVE
.3 mo and till
death (n 5 72)
Maehlen et al. 128 1983–1994 No AZT (n 5 50); MGCs; diffuse damage of Incidence of MGCs inversely related to
[69] AZT (usually 600 white matter; microglial whether AZT was taken until patient was
mg/d) for .2 mo noduli close to death; AZT reduced risk of brain
and terminated lesions only if continued until near death
¶1 mo (n 5 29),
2–6 mo (n 5 23),
or .6 mo
(n 5 26) before
death
Bell et al. 66 1990–? Never or ,6 w; .6 HIVE; cognitive impairment AZT use, particularly when used within 1 y
[70] w but stopped .1 (neurophysiological and of death, associated with lower risk of
y before death; neuropsychological tests) HIVE (P , .01); association remained
.6 w and used when adjusted for lifetime dose of AZT;
within 1 y of AZT use also associated with reduced
death (n’s not risk of cognitive impairment (P , .05)
reported)

NOTE. AZT 5 zidovudine; HIVE 5 HIV encephalitis; HIVL 5 HIV leukoencephalopathy; MGCs 5 multinucleated giant cells.

Summary HIV load in CSF concomitant with antiretroviral therapy [50,

Results of evaluations of epidemiological surveys, clinical
trials, CSF assays, and autopsy studies lead to several conclu-
sions concerning the use of antiretrovirals, particularly zidovu-
dine, in the prevention and treatment of HIVD among adults.
Although the pathophysiology of HIVD remains to be fully
explicated, the reduction in HIV or its attendant markers in
CSF probably should be an important goal of therapy. In this
regard, antiretroviral therapy has been found to be associated
with decreased levels of HIV-1 p24 antigen in CSF [46, 47,
49]. Although direct assays of HIV RNA in CSF have only
recently been used, several groups have reported reductions in
51]. The clinical relevance of HIV load in CSF remains un-
clear, although preliminary observations suggest that an in-
crease in level of HIV RNA in CSF correlates with a decline in
cognitive status [21, 22].
Postmortem studies have confirmed that patients treated with
zidovudine, especially for .1 year and until shortly before
death, have fewer HIV-related changes in brain tissue, that is,
a lower prevalence of multinucleated giant cells and of HIV
encephalitis and HIV leukoencephalopathy, than do compara-
ble patients who were not treated with zidovudine or whose
therapeutic time frames were dissimilar [18, 67–70].
30
In The Netherlands, the incidence of HIVD declined sharply
following the introduction of zidovudine [46]. In addition,
several studies directly linked the use of zidovudine with
improved neuropathologic functioning, particularly with im-
provements in cognition and motor skills, among patients with
AIDS and HIVD [46, 53–55, 65]. Such improvements, how-
ever, appear to be time-limited, generally becoming evident in
the first 4 – 6 months of treatment but reverting toward baseline
within 6 –12 months [55, 58, 60, 61]. The optimum dosage of
zidovudine has not been clearly demonstrated, although max-
imum response may occur at 1,000 –2,000 mg/day [57]. Fi-
nally, zidovudine therapy, especially among asymptomatic
HIV-infected patients and particularly at higher dosages, may
have a prophylactic effect against HIVD [63]. Future studies
must assess the effect of other antiretroviral agents as well as of
therapies aimed at indirect pathophysiological factors (e.g.,
cytokines, excitotoxins, channel abnormalities) in the prophy-
laxis and treatment of HIVD.
HIV Encephalopathy in Children
Epidemiology, Clinical Features, and Pathogenesis
Neurological disease is the initial presenting condition in
¶5% of adult patients with HIV infection but in up to 18% of
children and adolescents [19]. Overall, estimates of the inci-
dence of HIV encephalopathy among children with AIDS
range from 30% to 60% [15, 19, 71], while as many as 88% of
children with symptomatic HIV disease exhibit CNS abnor-
malities [72].
Among adults, HIV is transmitted between individuals pri-
marily by horizontal methods— by sexual contact or behaviors
related to drug use. In contrast, more than three-quarters of
cases among children are transmitted vertically (perinatally)
[19, 73]. CNS infection may occur early in fetal development,
and infection at these early stages may explain why symptoms
of HIV encephalopathy tend to develop more rapidly in chil-
dren than in adults [19, 73].
The neurological course of HIV encephalopathy in children
generally falls into one of three categories: subacute progres-
sive, plateau progressive, and static [73]. In subacute progres-
sive encephalopathy (PE), children at first develop slowly but
normally; then, social, language, and motor skills begin to be
lost. Acquired microcephaly may indicate slowed brain growth,
and HIV-1 p24 antigen may be detected in CSF. Children
demonstrating plateau PE also develop initially at a slow but
normal pace; then there is a decline in the rate of developmen-
tal progress, with little or no further acquisition of skills.
Acquired microcephaly again is present, but CSF usually is
negative for HIV-1 p24 antigen. Finally, a minority of children
(;25%) have static encephalopathy. These children are late to
acquire motor and language skills, are cognitively impaired,
and acquire skills slowly. Brain scans and CSF evaluations
generally are normal.
Simpson CID 1999;29 (July)
As with adult HIVD, the pathogenesis of childhood HIV en-
cephalopathy is uncertain. Severity of pathological findings usu-
ally correlates with clinical status, but in some children with
evidence of PE at autopsy, there is no evidence of significant
replication of HIV-1 or of inflammatory changes in the CNS [74].
Children with HIV encephalitis also demonstrate the unusual
pathological feature of calcifications and vascular mineralizations
in the brain [75]. It has been suggested that increased viral repli-
cation and emergence of drug-resistant HIV variants within the
CNS may affect the development of PE [76].
Treatment
As in adults, there is evidence that zidovudine decreases
levels of HIV-1 p24 antigen in CSF of children with PE [77,
78]. McKinney et al. [78] measured levels of p24 antigen in
CSF of 51 children between 3 months and 12 years of age.
Eighteen (35%) were positive for the antigen at baseline. After
6 months of treatment with zidovudine, eight patients who had
tested positive at baseline underwent follow-up CSF testing; all
eight showed a decrease in p24 antigen concentrations, and six
(75%) tested negative for the antigen. Sixteen of the 33 patients
who were negative at baseline had lumbar punctures after 6
months of zidovudine treatment; one then tested positive for
p24 antigen. Laverda et al. [77] performed immunologic and
virological studies on the CSF of 15 HIV-infected children (4
with PE) before and at least 6 months after treatment with 600
mg/(m2 z d) [77]. Substantial declines from baseline were ob-
served in levels of IL-6, IL-1b, and p24 antigen in CSF.
Several studies have evaluated the clinical efficacy of
zidovudine treatment of PE (table 4) [78 – 83].
Studies of efficacy. In a phase I clinical trial, Pizzo et al.
[79] evaluated changes in cognitive and adaptive functioning
after 6 months of zidovudine therapy among 13 children ages
6 months to 12 years with symptoms of AIDS. Eight of the
children had clinical evidence of encephalopathy before receiv-
ing drug therapy. Zidovudine was administered iv in dosages of
0.5, 0.9, 1.4, or 1.8 mg/(kg z h) depending on time of study
entry. Development of neutropenia was common and dose-
related. Significant improvement was found in cognitive scores
for children with and without PE at presentation at all dosage
levels (P , .01). Improvements were noted quickly, within
3– 4 weeks of initiation of therapy, at all dosage levels and
generally were steady and maintained throughout the follow-up
period. In an extension of this study, the 13 children were
reevaluated after 12 months of zidovudine therapy [80]. Gains
noted at 6 months in both cognitive functioning and adaptive
behavior were largely maintained in children with and without
PE. Although these studies were small and used an open-label
design, the authors concluded that zidovudine can benefit chil-
dren with symptomatic HIV infection, including those with
encephalopathy.
An open-label study of the efficacy of zidovudine at 180
mg/m2 orally q6h was conducted among 88 children aged 3
CID 1999;29 (July) HIV-Associated Dementia 31

Table 4. Studies of zidovudine in HIV-infected children with progressive encephalopathy.

PE-
Study Mean Diagnosis, patient related
Reference n design duration age range AZT dose outcomes Results

Pizzo et al. 13 Open 6 mo Symptomatic 0.5, 0.9, 1.4, and Cognitive; Significant improvement in cognitive
[79] HIV, 8 with 1.8 mg/(kg z h) adaptive scores for all children after 6 mo of
PE; 6 mo–12 y continuous iv behavior AZT (P , .01); improvement at
all dosages, in those with and
without PE; most adverse effects at
Ä1.4 mg/(kg z h)
Brouwers et 13 Open 1y Symptomatic 0.5, 0.9, 1.4, and Cognitive; Significantly improved cognitive and
al. [80] HIV, 8 with 1.8 mg/(kg z h) adaptive adaptive behavior scores at 6 mo
(continuation PE; 6 mo–12 y continuous iv behavior (P , .01); gain largely
of [79]) maintained at 12 mo
McKinney et 55 Open 6 mo AIDS or 180 mg/m2 p.o. Cognitive 42% had increases of Ä8 points on
al. [78] advanced HIV q6h test scores; children ,30 mo old
disease, 39 showed significant cognitive
with PE; 3 mo– improvement (P < .02) but
12 y older children did not
Nozyce et al. 54 Open 1y HIV, 4 with PE; 180 mg/m2 p.o. Cognitive No significant change in cognitive
[81] 2 mo–12 y q6h function over 12 mo
Wolters et al. 25 Open 6 mo Symptomatic Continuous iv (n Cognitive; Significantly improved (P < .01)
[82] HIV, 13 with 5 13); p.o. adaptive in all areas except motor skills
PE; 1–12 y (n 5 12); behavior; after 6 mo of treatment; no
doses not motor difference in improvement
stated skills between children with and without
PE or between routes of
administration
Brady et al. 424 R, DB 39 mo HIV with mild– 90 vs. 180 mg/ Development No significant differences between
[83] moderate m2 p.o. q6h of PE; dosage groups in development of
symptoms, no cognitive PE, cognitive functioning, or
PE; 3 mo–12 y survival

NOTE. All studies were studies of efficacy except [83], which was a study of dosage, and [79], which was both. AZT 5 zidovudine; DB 5 double-blind; PE 5
progressive encephalopathy; R 5 randomized. Reprinted with permission from The Annals of Pharmacotherapy [1].

months through 12 years with either AIDS or advanced HIV
disease [78]. Cognitive functioning was tested at baseline and
6 months later among 55 of these children, including 39 with
PE. Overall, 23 patients (42%) had increases of Ä8 points on
test scores; 10 had increases of Ä15 points. Although children
,30 months of age showed significant cognitive improvement
(P , .02), older children did not. Anemia or neutropenia was
reported in 61% of patients; most of these conditions were
treated by transfusion, reduced zidovudine dosage, or both. The
authors concluded that zidovudine therapy yielded some im-
provement in cognitive functioning after 6 months of use. Their
conclusion would have been strengthened had scores been
compared on tests for children with vs. without PE.
Contrasting results were obtained in a study of 54 children
ages 2 months to 12 years with HIV infection (4 with PE)
included in an open, 1-year protocol examining the effects of
oral zidovudine on neurodevelopmental functioning [81]. No
significant changes in cognitive functioning were found after
12 months of zidovudine treatment.
Wolters et al. [82] included 25 children (ages 1–12 years)
with symptomatic HIV infection (13 with PE) in an open,
6-month study of the effect of zidovudine on cognition, adap-
tive behavior, and motor skills. Twelve children received
zidovudine orally, while 13 received the drug by continuous
infusion; medication dosages were not reported. With the ex-
ception of motor skills, significant improvements were seen in
all neurodevelopmental areas measured between baseline and
final testing after 6 months of zidovudine treatment (P ¶ .01).
Although children with PE had lower scores at baseline, no
difference was found in levels of improvement between the
groups with and without PE. Route of administration also was
not found to affect the magnitude of improvement. The authors
concluded that zidovudine administered either orally or by
continuous infusion is associated with improvements in cogni-
tion and adaptive behavior in children. The generalizability of
the results is limited, however, by the small sample size and
lack of a control group.
Studies of dosage. Few studies of optimum dosage levels
of zidovudine in children have been conducted. Pizzo et al. [79]
concluded that the optimum iv zidovudine dosage ranges from
0.9 to 1.4 mg/(kg z h) on the basis of their findings that there
were no significant differences in improvement in cog-
nitive scores in children receiving either 0.5, 0.9, 1.4, or 1.8
mg/(kg z h) but that adverse effects (primarily neutropenia)
32
were most common in those receiving zidovudine in dosages of
1.4 and 1.8 mg/(kg z h). Because the research was designed to
address questions of efficacy and not of dosage, however, this
conclusion should be confirmed in additional trials.
Brady et al. [83] conducted a large (n 5 424), randomized,
double-blind clinical trial of the effects of two oral doses of
zidovudine (90 vs. 180 mg/m2 q6h) on cognitive functioning.
Children included were 3 months to 12 years of age, were
HIV-infected, and exhibited mild to moderate symptoms. No
patient had PE. Follow-up continued for an average of 39
months, with neuropsychological testing done at 6-month in-
tervals. No significant differences were found between treat-
ment groups at any time point with respect to development of
PE, cognitive functioning, or survival. Overall, 39% of patients
had at least one episode of grade 3 or greater hematologic
toxicity. There was no difference between treatment groups in
the rate of such toxicity, although fewer patients receiving
low-dose zidovudine developed neutropenia. The authors con-
cluded that in children with mild to moderate HIV disease, the
recommended oral dosage of zidovudine should be 90 mg/m2
q6h.
Summary
Fewer studies have been conducted of the efficacy and
optimal dosing of zidovudine in children than in adults. In
these young patients, duration of effect and the potential of
zidovudine use for prophylaxis for PE remain to be determined.
Overall, zidovudine appears to improve neuropsychological
functioning in children infected with HIV [78 – 80, 82]. Such
improvement has been demonstrated in children with and with-
out PE and by use of both oral and iv routes of administration.
Currently, controlled trials that follow children for a sufficient
period and use neuropsychological, neuroimaging, and virolog-
ical (in plasma and CSF) end points are needed to establish
treatment parameters for zidovudine and combination regimens
in this population.
Conclusions
HIVD in adults and PE in children are devastating compli-
cations of HIV infection. The prognosis for patients who de-
velop these conditions is uniformly poor and, as the conditions
progress, the clinical symptoms that characterize HIVD and PE
become increasingly incapacitating.
On the basis of current data, zidovudine is one of the drugs
that should be considered as part of combination regimens
designed to manage HIVD and PE. In both adults and children,
zidovudine use has been associated with decreases in p24
antigen and HIV load in CSF and with improvements in neu-
ropsychological functioning. The length of clinical response
varies among individuals but appears to decline after 6 –12
months of zidovudine use. Although dose-response relation-
Simpson CID 1999;29 (July)
ships have been demonstrated in adults, optimal dosages of
zidovudine in adults and children have not been established,
and it is not clear when in the course of the infection drug
therapy should begin. It is important to note that virtually all
patients with HIVD or PE will receive zidovudine in combi-
nation with protease inhibitors and/or nonnucleoside reverse
transcriptor inhibitors or as part of promising protocols of
highly active antiretroviral therapy [84, 85]. Although charac-
teristics of several newer antiretroviral agents and combina-
tions of agents, including their ability to cross the blood-brain
barrier, provide some promise in the treatment of HIVD, there
are minimal clinical data as yet available to guide their use.
The use of zidovudine is not without attendant problems.
Myelosuppression, with anemia and leukopenia, is a prominent
side effect [86, 87]. At least in adults, however, instances of
significant myelosuppression resulting in anemia and leukope-
nia generally do not require the discontinuation of zidovudine
therapy, because these conditions can be managed with hema-
topoietic growth factors or transfusions [87– 89].
In conclusion, a review of the neurological literature sug-
gests that maintenance treatment of patients with combinations
of antiretroviral drugs that include CNS-penetrant agents, such
as zidovudine, appears to be a strategy consonant with en-
hanced quality of life in long-term survivors. Well-designed
clinical trials are needed to refine how such therapies should be
implemented and to compare their relative efficacies and safety
profiles. It is critically important to include neuropsychological
end points and to examine a variety of CSF parameters in
primary antiretroviral trials to determine the neurological
safety and efficacy of these agents.
Acknowledgments
The author wishes to thank Jane G. Murphy, Ph.D., for editorial
assistance and Susan Morgello, M.D., for provision of photomi-
crographs.
References
1. Melton ST, Kirkwood CK, Ghaemi SN. Pharmacotherapy of HIV demen-
tia. Ann Pharmacother 1997;31:457–73.
2. Simpson DM, Berger JR. Neurologic manifestations of HIV infection.
Med Clin North Am 1996;80:1363–94.
3. Hentzen BTH, Schreij G. Patterns of morbidity and mortality in AIDS
patients on Pneumocystis carinii prophylaxis who died during hospital
admission: a report of 50 diseased patients. Neth J Med 1996;49:101–5.
4. McArthur JC, Hoover DR, Bacellar H, et al. Dementia in AIDS patients:
incidence and risk factors. Neurology 1993;43:2245–52.
5. Gray F, Gherardi R, Scaravilli F. The neuropathology of the acquired
immune deficiency syndrome (AIDS). Brain 1988;111:245– 66.
6. Berger JR, Simpson DM. Neurologic complications of AIDS. In: Scheld
WM, Whitley RJ, Durack DT, eds. Infections of the central nervous
system. 2nd ed. Philadelphia: Lippincott-Raven, 1997:255–71.
7. Janssen RS, Saykin AJ, Cannon L, et al. Neurological and neurophysio-
logical manifestations of HIV-1 infection: association with AIDS-
related complex but not asymptomatic HIV-1 infection. Ann Neurol
1989;26:592– 600.
CID 1999;29 (July) HIV-Associated Dementia
8. Navia BA, Jordan BD, Price RW. The AIDS dementia complex: I. clinical
features. Ann Neurol 1986;19:517–24.
9. Navia BA. Clinical and biologic features of the AIDS dementia complex.
Neuroimaging Clin N Am 1997;7:581–92.
10. Price RW, Brew BJ. The AIDS dementia complex. J Infect Dis 1988;158:
1079 – 83.
11. McArthur JC, Cohen BA, Farzdegan H, et al. Cerebrospinal fluid abnor-
malities in homosexual men with and without neuropsychiatric findings.
Ann Neurol 1988;23(suppl):S34 –7.
12. Portegies P, Epstein LG, Hung STA, de Gans J, Goudsmit J. Human
immunodeficiency virus type 1 antigen in cerebrospinal fluid: correla-
tion with clinical neurologic status. Arch Neurol 1989;46:261– 4.
13. Simpson DM, Tagliati M, Morgello M. Neurologic manifestations of
AIDS. In: Mildvan D, ed. Essential atlas of infectious diseases for
primary care. Philadelphia: Current Science, 1997:10.3–10.4.
14. Dal Pan GJ, McArthur JH, Aylward E, et al. Patterns of cerebral atrophy
in HIV-1–infected individuals: results of a quantitative MRI analysis.
Neurology 1992;42:2125–30.
15. Brouwers P, Belman AL, Epstein L. Central nervous system involvement:
manifestations, evaluation, and pathogenesis. In: Pizzo PA, Wilfert CM,
eds. Pediatric AIDS. Baltimore: Williams & Wilkins, 1994:433–55.
16. Glass JD, Fedor H, Wesselingh SL, McArthur JC. Immunocytochemical
quantitation of human immunodeficiency virus in the brain: correlations
with dementia. Ann Neurol 1995;38:755– 62.
17. Glass JD, Johnson RT. Human immunodeficiency virus and the brain.
Annu Rev Neurosci 1996;19:1–26.
18. Glass JD, Wesselingh SL, Selnes OA, McArthur JC. Clinical-
neuropathologic correlation of HIV-associated dementia. Neurology
1993;43:2230 –7.
19. Mintz M. Clinical comparison of adult and pediatric neuroAIDS. Adv
Neuroimmunol 1994;4:207–21.
20. Brew BJ, Dunbar N, Pemberton L, Kaldor J. Predictive markers of AIDS
dementia complex: CD4 cell count and cerebrospinal fluid concentra-
tions of b2-microglobulin and neopterin. J Infect Dis 1996;174:294 – 8.
21. McArthur JC, McClernon DR, Cronin MF, et al. Relationship between
human immunodeficiency virus–associated dementia and viral load in
cerebrospinal fluid and brain. Ann Neurol 1997;42:689 –98.
22. Ellis RJ, Hsia K, Spector SA, et al. Cerebrospinal fluid human immuno-
deficiency virus type 1 RNA levels are elevated in neurocognitively
impaired individuals with acquired immunodeficiency syndrome. Ann
Neurol 1997;42:679 – 88.
23. Griffin DE, McArthur JC, Cornblath DR. Neopterin and interferon-gamma
in serum and cerebrospinal fluid of patients with HIV-associated neu-
rologic disease. Neurology 1991;41:69 –74.
24. Petito CK, Cash KS. Blood-brain barrier abnormalities in the acquired
immunodeficiency syndrome: immunohistochemical localization of se-
rum proteins in postmortem brain. Ann Neurol 1992;32:658 – 66.
25. Dreyer EB, Kaiser PK, Offerman JT, et al. HIV-1 coat protein neurotox-
icity prevented by calcium channel antagonists. Science 1990;248:
364 –7.
26. Choi DW. Calcium-mediated neurotoxicity: relationship to specific chan-
nel types and role in ischemic damage. Trends Neurosci 1988;11:465–9.
27. Epstein LG, Gendelman HE. Human immunodeficiency virus type 1
infection of the nervous system: pathogenetic mechanisms. Ann Neurol
1993;33:429 –36.
28. Lipton SA. Models of neuronal injury in AIDS: another role for the
NMDA receptor? Trends Neurosci 1992;15:75– 80.
29. Lipton SA, Sucher NJ, Kaiser PK, Dreyer EB. Synergistic effects of the
HIV coat protein and NMDA receptor-mediated neurotoxicity. Neuron
1991;7:111– 8.
30. Sanders VJ, Pittman CA, White MG, Wang G, Wiley CA, Achim CL.
Chemokines and receptors in HIV encephalitis. AIDS 1998;12:1021– 6.
31. Clifford DB, Simpson D. Targeting HIV therapy for the brain. HIV
Advances in Research and Treatment 1998;8:10 –7.
33
32. Enting RH, Hoetelmans RMW, Lange JMA, Burger DM, Beijnen JH,
Portegies P. Antiretroviral drugs and the central nervous system. AIDS
1998;12:1941–55.
33. Gwilt PR, Tracewell WG. Pharmacokinetics and pharmacodynamics of
hydroxyurea. Clin Pharmacokinet 1998;34:347–58.
34. Beckloff GL, Lerner HJ, Frost D, Russo-Alesi FM, Gitomer S. Hydroxy-
urea (NSC-32065) in biologic fluids: dose-concentration relationship.
Cancer Chemother Rep 1965;48:57– 8.
35. Federici ME, Lupo S, Cahn P, et al. Hydroxyurea in combination regimens
for the treatment of antiretroviral naive, HIV-infected adults [abstract
no. 287/12235]. In: Conference record of the 12th World AIDS Con-
ference. Geneva: Marathon Multimedia, 1998.
36. Lori F, Malykh AG, Foli A, et al. Combination of a drug targeting the cell
with a drug targeting the virus controls human immunodeficiency virus
type 1 resistance. AIDS Res Hum Retroviruses 1997;13:1403–9.
37. Glynn SL, Yazdanian M. In vitro blood-brain barrier permeability of
nevirapine compared to other HIV antiretroviral agents. J Pharm Sci
1998;87:306 –10.
38. Ravitch JR, Jarrett JL, Robertson White H, et al. CNS penetration of the
antiretroviral 1592 in human and animal models [abstract no. 636]. In:
Program and abstracts of the 5th Conference on Retroviruses and
Opportunistic Infections (Chicago). Alexandria, VA: Foundation for
Retroviruses and Human Health, 1998.
39. Stahle L, Martin C, Svensson JO, Sonnerborg A. Indinavir in cerebrospinal
fluid of HIV-1-infected patients [letter]. Lancet 1997;350:1823.
40. Miralles P, Padilla B, Berenguer J, et al. Therapy of AIDS-associated
progressive multifocal leuko-encephalopathy with protease inhibitors
combination regimens [abstract no. 464]. In: Program and abstracts of
the 5th Conference on Retroviruses and Opportunistic Infections (Chi-
cago). Alexandria, VA: Foundation for Retroviruses and Human Health,
1998.
41. Borleffs JC, CHEESE Study Team. First comparative study of saquinavir
soft gel capsules vs indinavir as part of triple therapy regimen
(CHEESE) [abstract no. 387b]. In: Program and abstracts of the 5th
Conference on Retroviruses and Opportunistic Infections (Chicago).
Alexandria, VA: Foundation for Retroviruses and Human Health, 1998.
42. Ferrando S, van Gorp W, McElhiney M, Goggin K, Sewell M, Rabkin J.
Highly active antiretroviral treatment in HIV infection: benefits for
neuropsychological function. AIDS 1998;12:F65–70.
43. Rolinski B, Bogner JR, Sadri I, Wintergest N, Goebel FD. Absorption and
elimination kinetics of ZDV in the CSF of HIV-1 infected patients. J
Acquir Immune Defic Syndr Hum Retrovirol 1997;15:192–7.
44. Filippi CG, Sze G, Farber SJ, Shahmanesh M, Selwyn PA. Regression of
HIV encephalopathy and basal ganglia signal intensity abnormality at
MR imaging in patients with AIDS after the initiation of protease
inhibitor therapy. Radiology 1998;206:491– 8.
45. Wong JK, Ignacio CC, Torriani F, Havlor D, Fotch NJS, Richman DD. In
vivo compartmentalization of HIV: evidence from the examination of
pol sequences from autopsy tissue. J Virol 1997;71:2059 –71.
46. Portegies P, de Gans J, Lange JMA, et al. Declining incidence of AIDS
dementia complex after introduction of zidovudine treatment. BMJ
1989;299:819 –21.
47. Royal W III, Selnes OA, Concha M, Nance-Sproson TE, McArthur JC.
Cerebrospinal fluid human immunodeficiency virus type 1 (HIV-1) p24
antigen levels in HIV-1-related dementia. Ann Neurol 1994;36:32–9.
48. Portegies P. Review of antiretroviral therapy in the prevention of HIV-
related AIDS dementia complex (ADC). Drugs 1995;49(suppl 1):25–31.
49. Heyes MP, Brew BJ, Martin A, et al. Quinolinic acid in cerebrospinal fluid
and serum in HIV-1 infection: relationship to clinical and neurological
status. Ann Neurol 1991;29:202–9.
50. Foudraine NA, Hoetelmans RMW, Lange JMA, et al. Cerebrospinal fluid
HIV-1 RNA and drug concentrations after treatment with lamivudine
plus zidovudine or stavudine. Lancet 1998;351:1547–50.
34 Simpson
51. Collier AC, Marra C, Coombs, RW, et al. Cerebrospinal fluid indinavir and
HIV RNA levels in patients on chronic indinavir therapy [abstract]. Clin
Infect Dis 1997;25:359.
52. Brew BJ, Brown SJ, Catalan J, et al. Safety and efficacy of abacavir (ABC,
1592) in AIDS dementia complex (study CNAB 3001) [abstract no.
561/32192]. In: Conference record of the 12th World AIDS Conference.
Geneva: Marathon Multimedia, 1998.
53. Yarchoan R, Brouwers P, Spitzer AR, et al. Response of human-
immunodeficiency-virus-associated neurological disease to 39-azido-39-
deoxythymidine. Lancet 1987;1:132–5.
54. Schmitt FA, Bigley JW, McKinnis R, et al. Neuropsychological outcome
of zidovudine (AZT) treatment of patients with AIDS and AIDS-related
complex. N Engl J Med 1988;319:1573– 8.
55. Portegies P, Enting RH, de Gans J, et al. Presentation and course of AIDS
dementia complex: 10 years of follow-up in Amsterdam, The Nether-
lands. AIDS 1993;7:669 –75.
56. Nordic Medical Research Council’s HIV Therapy Group. Double blind
dose-response study of zidovudine in AIDS and advanced HIV infec-
tion. BMJ 1992;304:13–7.
57. Sidtis JJ, Gatsonis C, Price RW, et al. Zidovudine treatment of the AIDS
dementia complex: results of a placebo-controlled trial. Ann Neurol
1993;33:343–9.
58. Tozzi V, Narciso P, Galgani S, et al. Effects of zidovudine in 30 patients
with mild to end-stage AIDS dementia complex. AIDS 1993;7:683–92.
59. Brouwers P, Hendricks M, Lietzau JA, et al. Effect of combination therapy
with zidovudine and didanosine on neuropsychological functioning in
patients with symptomatic HIV disease: a comparison of simultaneous
and alternating regimens. AIDS 1997;11:59 – 66.
60. Reinvang I, Frøland SS, Karlsen NR, Lundervold AJ. Only temporary
improvement in impaired neuropsychological function in AIDS patients
treated with zidovudine. AIDS 1991;5:228 –9.
61. Karlsen NR, Reinvang I, Frasoland SS. A follow-up study of neuropsy-
chological functioning in AIDS-patients: prognostic significance and
effect of zidovudine therapy. Acta Neurol Scand 1995;91:215–21.
62. Volberding PA, Lagakos SW, Koch MA, et al. Zidovudine in asymptom-
atic human immunodeficiency virus infection: a controlled trial in
persons with fewer than 500 CD4-positive cells per cubic millimeter.
N Engl J Med 1990;322:941–9.
63. Hamilton JD, Hartigan PM, Simberkoff MS, et al. A controlled trial of
early versus late treatment with zidovudine in symptomatic human
immunodeficiency virus infection: results of the Veterans Affairs Co-
operative Study. N Engl J Med 1992;326:437– 43.
64. Concorde Coordinating Committee. Concorde: MRC/ANRS randomised
double-blind controlled trial of immediate and deferred zidovudine in
symptom-free HIV infection. Lancet 1994;343:871– 81.
65. Bacellar H, Muñoz A, Miller EN, et al. Temporal trends in the incidence
of HIV-1-related neurologic diseases: Multicenter AIDS Cohort Study,
1985–1992. Neurology 1994;44:1892–900.
66. Baldeweg T, Catalan J, Lovett E, Gruzelier J, Riccio M, Hawkins D.
Long-term zidovudine reduces neurocognitive deficits in HIV-1 infec-
tion. AIDS 1995;9:589 –96.
67. Vago L, Castagna A, Lazzarin A, Trabattoni G, Cinque P, Costanzi G.
Reduced frequency of HIV-induced brain lesions in AIDS patients
treated with zidovudine. J Acquir Immune Def Syndr 1993;6:42–5.
68. Gray F, Bélec L, Keohane C, et al. Zidovudine therapy and HIV enceph-
alitis: a 10-year neuropathological survey. AIDS 1994;8:489 –93.
69. Maehlen J, Dunlop O, Liestøl K, Dobloug JH, Goplen AK, Torvik A.
Changing incidence of HIV-induced brain lesions in Oslo, 1983–1994:
effects of zidovudine treatment. AIDS 1995;9:1165–9.
70. Bell JE, Donaldson YK, Lowrie S, et al. Influence of risk group and
zidovudine therapy on the development of HIV encephalitis and cogni-
tive impairment in AIDS patients. AIDS 1996;10:493–9.
CID 1999;29 (July)
71. Epstein LG, Sharer LR, Oleske JM, et al. Neurologic manifestations of
human immunodeficiency virus infection in children. Pediatrics 1986;
78:678 – 87.
72. Belman AL. HIV-1 associated CNS disease in infants and children. Res
Publ Assoc Res Nerv Ment Dis 1994;72:289 –309.
73. Civitello LA. Neurologic complications of HIV infection in children.
Pediatr Neurosurg 1991/2;17:104 –12.
74. Vazeux R, Lacroix-Ciaudo C, Blanche S. Low levels of human immuno-
deficiency virus replication in the brain tissue of children with severe
acquired immunodeficiency syndrome encephalopathy. Am J Pathol
1992;140:137– 44.
75. Epstein LG, Sharer LR, Goudsmit J. Neurological and neuropathological
features of HIV in children. Ann Neurol 1988;23(suppl):S19 –23.
76. Sei S, Stewart SK, Farley M, et al. Evaluation of human immunodeficiency
virus (HIV) type 1 RNA levels in cerebrospinal fluid and viral resistance
to zidovudine in children with HIV encephalopathy. J Infect Dis 1996;
174:1200 – 6.
77. Laverda AM, Gallo P, DeRossi A, et al. Cerebrospinal fluid analysis in
HIV-1-infected children: immunological and virological findings before
and after AZT therapy. Acta Paediatr 1994;83:1038 – 42.
78. McKinney RE, Maha MA, Connor EM, et al. A multicenter trial of oral
zidovudine in children with advanced human immunodeficiency virus
disease. N Engl J Med 1991;324:1018 –25.
79. Pizzo PA, Eddy J, Falloon J, et al. Effect of continuous intravenous
infusion of zidovudine (AZT) in children with symptomatic HIV infec-
tion. N Engl J Med 1988;319:889 –96.
80. Brouwers P, Moss H, Wolters P, et al. Effect of continuous-infusion
zidovudine therapy on neuropsychologic functioning in children with
symptomatic human immunodeficiency virus infection. J Pediatr 1990;
117:980 –5.
81. Nozyce M, Hoberman M, Arpadi S, et al. A 12-month study of the effects
of oral zidovudine on neurodevelopmental functioning in a cohort of
vertically HIV-infected inner-city children. AIDS 1994;8:635–9.
82. Wolters PL, Brouwers P, Moss HA, Pizzo PA. Adaptive behavior of
children with symptomatic HIV infection before and after zidovudine
therapy. J Pediatr Psychol 1994;19:47– 61.
83. Brady MT, McGrath N, Brouwers P, et al. Randomized study of the
tolerance and efficacy of high- versus low-dose zidovudine in human
immunodeficiency virus–infected children with mild to moderate symp-
toms. J Infect Dis 1996;173:1097–106.
84. d’Arminio MA, Testa L, Adorni F, et al. Clinical outcome and predictive
factors of failure of highly active antiretroviral therapy in antiretroviral-
experienced patients in advanced stages of HIV-1 infection. AIDS
1998;12:1631–7.
85. Bisset LR, Cone RW, Huber W, et al. Highly active antiretroviral therapy
during early HIV infection reverses T-cell activation and maturation
abnormalities. AIDS 1998;12:2115–23.
86. Coyle TE. Hematologic complications of human immunodeficiency virus
infection and the acquired immunodeficiency syndrome. Med Clin
North Am 1997;81:449 –70.
87. Groopman JE, Feder D. Hematopoietic growth factors in AIDS. Semin
Oncol 1992;19:408 –14.
88. Aboulafia DM. Use of hematopoietic hormones for bone marrow defects
in AIDS. Oncology 1997;11:1827–50.
89. Sullivan PS, Hanson DL, Chu SY, Jones JL, Ward JW, Adult-Adolescent
Spectrum of Disease Group. Epidemiology of anemia in human immu-
nodeficiency virus (HIV)–infected persons: results from the multistate
Adult and Adolescent Spectrum of HIV Disease Surveillance Project.
Blood 1998;91:301– 8.