Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Jul 13.
Published in final edited form as: J Int Neuropsychol Soc. 1997 May;3(3):276–287.

Brain atrophy in HIV infection is more strongly associated with CDC clinical stage than with cognitive impairment

VICTORIA DI SCLAFANI 1, RD SHANE MACKAY 1,4, DIETER J MEYERHOFF 2, DAVID NORMAN 2, MICHAEL W WEINER 2,3,4, GEORGE FEIN 1,4
PMCID: PMC2709487  NIHMSID: NIHMS126116  PMID: 9161107

Abstract

HIV infection often results in MRI-detectable brain atrophy and white matter signal hyperintensities (WMSHs). Magnetic resonance images were obtained from 31 HIV+ male patients and 10 high-risk controls. Variation within the HIV+ group on neuropsychological (NP) impairment and stage of systemic disease were relatively independent, allowing examination of the relative association of MRI measures with NP impairment versus with systemic stage of disease. HIV+ patients compared to high-risk controls evidenced global atrophy, reduced caudate nuclei volume, and a trend to gray matter volume loss but no difference in white matter volume or in WMSHs. These effects were progressive with CDC clinical stage such that patients at CDC stage A had values very close to those of controls, while patients at CDC stage C had the most abnormal values. In contrast, the relationship between these MRI variables and severity of NP impairment was much less dramatic, with the mildly to moderately impaired HIV+ subjects showing MRI volume effects greater than or equal to those of the severely impaired HIV+ subjects. These results suggest that MRI-detectable brain atrophy secondary to HIV infection is not the primary substrate underlying the progressive NP impairment in HIV disease.

Keywords: HIV, MRI, Brain atrophy, Caudate, White matter signal hyperintensity

INTRODUCTION

Neurocognitive disease associated with HIV infection has been separated into two categories: (1) a more severe form, HIV-1-associated dementia complex (HADC); and (2) a less severe form, HIV-1-associated minor cognitive-motor disorder. HADC is sufficient for a diagnosis of AIDS and usually occurs in combination with other AIDS-defining illnesses (Castro et al., 1992). The less severe minor cognitive–motor disorder is not sufficient for an AIDS diagnosis, although it may be present in persons with AIDS. It is still unclear whether HADC and HIV-1-associated minor cognitive–motor disorder reflect different underlying pathophysiological mechanisms, or whether they reflect different points along the continuum of severity of a single disease process. It is also unclear whether patients with minor cognitive–motor disorder will inevitably progress to HADC (American Academy of Neurology AIDS Task Force, 1991).

While the presence of HADC in AIDS patients is an established phenomenon, the presence of less severe NP impairments in seropositive patients without symptomatic systemic disease remains controversial. The large Multicenter AIDS Cohort Study found the onset of HADC to be a fairly abrupt phenomenon, occurring only in late-stage patients. Their sample of HIV+ asymptomatic subjects (N = 727) did not differ significantly from seronegative subjects (N = 769) on any neuropsychological measure (Miller et al., 1990). In contrast, Grant et al. (1992) and Janssen et al. (1989) documented mild NP impairment in clinically symptomatic (but not late-stage) patients. Stern et al. (1991) and Bornstein et al. (1993) found mild NP impairment in clinically asymptomatic patients, and Grant et al. (1992), Levin et al. (1990), and Wilkie et al. (1990) reported progressive NP impairment in clinically asymptomatic patients. Recently, using an 8-hr NP battery, Heaton et al. (1995) documented NP impairment in 30.5% of asymptomatic HIV+ subjects (N = 249). Although the varying reports on the frequency of NP impairment in the earlier stages of HIV infection are difficult to assess (due to differences in exclusionary criteria for subjects and controls, in the extent and sensitivity of the cognitive batteries employed, and in definitions of impairment), a critical review of 57 NP studies of asymptomatic HIV+ individuals found asymptomatic HIV+ subjects to be almost 3 times more likely to be neuropsychologically impaired than their HIV− counterparts (median prevalence 35 vs. 12%, respectively; White et al., 1995).

Some of the structural effects on the brain of systemic HIV infection can be assessed in vivo using MRI. In most cases, MRI shows global brain atrophy in HIV+ subjects at some point in the course of the disease (Levin et al., 1990; Moeller & Backmund, 1990; Post et al., 1991; Pan et al., 1992; Aylward et al., 1993; Hestad et al., 1993; Paley et al., 1994). Global atrophy in patients with advanced disease is not universal. Seilhean et al. (1993), in a postmortem study of the brains of 6 patients with HIV-associated cognitive–motor complex (4 of whom met the criteria of dementia) found no difference in cerebral weight between patients and controls. Loss of brain volume has been associated with dementia (Pan et al., 1992; Hestad et al., 1993), advanced CDC stage (Paley et al., 1994), and symptomatic but not demented patients (Jernigan et al., 1993). Aylward et al. (1993) reported cerebral atrophy in 16 nondemented HIV+ patients compared to 10 HIV− controls. Half of the nondemented HIV+ group was asymptomatic. However, the HIV+ group in their study included a proportionally greater number of individuals with histories of alcohol abuse, drug abuse, head injury, psychiatric illness, and positive medical histories than the HIV− group. These preexisting conditions may have contributed to Aylward's findings of cerebral atrophy in his nondemented HIV+ group, which included asymptomatic subjects.

While there have been abundant in vivo MRI investigations of global brain atrophy in HIV+ subjects compared to HIV− controls, there have been few reports that address the differential loss or pathology of cortical gray matter, subcortical gray matter, and white matter. Jernigan et al. (1993) found that symptomatic HIV+ subjects showed significant volume loss in white matter, temporal limbic gray matter, and the caudate nuclei when compared to both low-risk (i.e., heterosexual non-drug-abusing) HIV− controls and asymptomatic HIV+ patients.

However, significant temporal–limbic and caudate nuclei volume loss were also present in high-risk HIV− controls, complicating the interpretation of the gray matter losses in the symptomatic HIV+ subjects. Several other authors have also documented atrophy of the caudate nuclei in HIV+ subjects (Pan et al., 1992; Aylward et al., 1993; Hestad et al., 1993; Jernigan et al., 1993). Most of these investigators found caudate atrophy to be significantly associated with dementia. Some studies have attempted to measure WMSHs as a measure of white matter pathology (McArthur et al., 1990; Smith et al., 1990; Sönnerborg et al., 1990; Post et al., 1991; Manji et al., 1994). Although postmortem studies show HIV infiltration and damage of the white matter, (Smith et al., 1990; Masliah et al., 1992a; Liuzzi et al., 1994) MRI studies generally do not find WMSHs to be significantly associated with HIV infection, at least until very late-stage illness.

In summary, HIV disease progression is often associated with the development of atrophic brain changes and with the development of NP impairment. Unfortunately, severity of NP impairment has almost always been confounded with the progression of systemic disease in studies that have examined MRI-detectable brain atrophy in HIV infection. In the following report, we are able to separate the association of MRI changes with severity of NP impairment from the association of MRI changes with severity of systemic disease due to a sampling design that resulted in the relative independence of NP impairment and clinical stage of systemic disease.

METHODS

Research Participants

Thirty-one gay or bisexual HIV+ male subjects (ages 38.4 ± 6.9 years; education 15.0 ± 2.1 years) were compared with 10 HIV− male gay or bisexual controls (ages 37.4 ± 11.5 years; education 16.6 ± 1.3 years). HIV seronegativity in the controls was documented by PCR testing. Proton magnetic resonance spectroscopy data has been presented for the majority of the samples presented here (Meyerhoff et al., 1993, 1994, 1995). Participants were recruited from advertisements in the San Francisco gay/bisexual newspapers, from flyers posted throughout the gay/bisexual neighborhoods in San Francisco, and from distribution of flyers describing the study to all HIV+ individuals enrolled in a local AIDS agency program that provides meals to HIV-infected individuals with mental or physical disabilities. Participants were excluded if they reported a history of drug or alcohol abuse, a major psychiatric disorder unrelated to or predating HIV infection, a head injury with loss of consciousness, brain disease other than HIV, or opportunistic infections of the CNS. Participants were also excluded if the consulting neuroradiologist noted evidence of opportunistic infection or brain disease other than HIV on their MRI. HIV+ subjects were classified by clinical stage of systemic disease (CDC stage excluding indices of NP function) as follows: (1) Stage A (asymptomatic, primary HIV, or persistent generalized lymphadenopathy), (2) Stage B (symptomatic, but not A or C conditions), or (3) Stage C (AIDS-indicator conditions) (Castro et al., 1992). All procedures were approved by the UCSF Committee on Human Research and written consent was obtained from all subjects prior to study.

Neuropsychological Assessment

All participants underwent NP testing during which a wide range of cognitive skills (including attention, concentration, memory, verbal language, problem-solving, visuomotor and visuospatial skills, and fine motor ability) were assessed. The battery included (1) the WAIS–R Digit Symbol and Digit Span Subtests (Wechsler, 1981); (2) the Shipley Institute of Living Scale (Shipley, 1940); (3) the Rey–Osterrieth Complex Figure Test (Osterrieth & Rey, 1944); (4) the Trail Making Tests A and B, Finger Tapping, and Grip Strength (Reitan & Wolfson, 1985); (5) the Logical Memory and Visual Reproduction immediate and delayed subscales of the Wechsler Memory Scale (Russell, 1975); (6) the 15-item modification of the Fuld Object Memory Evaluation (FOME–15; Davenport et al., 1988); (7) the Short Categories Test (Wetzel & Boll, 1987); and (8) the Controlled Oral Word Association Test (COWAT; Benton & Hamsher, 1978). Each NP test was scored according to the norms published with the test, except for the Shipley Institute of Living Scale (Zachary norms, 1986), the Rey-Osterrieth Complex Figure Test (Denman norms, 1984), the Trail Making (Heaton et al. norms, 1991), and the Finger Tapping and Grip Strength tests (raw scores). We then generated a global impairment score (GIS) by first scoring severity of clinical impairment for each test on a scale from 0 to 2. A participant received a score of 0 if he scored above the 15th percentile on the test, 1 if he scored in the 5th to 15th percentile, and 2 if he scored below the 5th percentile. Points earned on all tests were then summed. A participant was classified as not impaired if his GIS was 0 to 1, mildly to moderately impaired if his GIS was 2 to 5, and severely impaired if his GIS was 6 or more. All of the HIV− control subjects had a GIS of 0 to 1 (not impaired). Table 1 shows the mean scores for each NP test for these NP impairment subgroups.

Table 1.

Mean neuropsychological test scores for impairment subgroups

HIV+ NP impairment subgroup
Controls
 No impairment
(GIS 0–1)
 Mild to Moderate impairment
(GIS 2–5)
 Severe impairment
(GIS ≥ 6)
Neuropsychological test M ± SD M ± SD M ± SD M ± SD
WAIS-R Digit Span 11.50 ± 2.32 12.33 ± 2.25 10.25 ± 1.82 7.23 ± 1.54
WAIS-R Digit-Symbol 12.50 ± 2.12 10.67 ± 1.97 10.33 ± 2.74 8.46 ± 2.60
Shipley Institute of Living Scale 112.80 ± 6.76 112.50 ± 8.69 105 ± 6.32 96.85 ± 8.85
Rey-Osterrieth Complex Figure 49.80 ± 26.38 45.50 ± 24.01 36.25 ± 32.79 11.85 ± 6.69
Trail Making Test A 77.34 ± 21.0 61.17 ± 29.53 58.87 ± 22.21 25.67 ± 20.26
Trail Making Test B 78.63 ± 17.68 74.04 ± 15.90 48.96 ± 20.80 19.72 ± 15.07
WMS Logical Memory immediate 63.50 ± 22.08 63.17 ± 23.84 43.08 ± 21.85 29.31 ± 23.86
WMS Logical Memory delayed 69.80 ± 18.48 62.17 ± 28.73 40.25 ± 24.04 27.92 ± 16.25
WMS Visual Reproduction immediate 61.0 ± 24.41 37.83 ± 25.58 28.75 ± 19.72 17.38 ± 9.10
WMS Visual Reproduction delayed 60.60 ± 28.66 44.00 ± 29.55 24.83 ± 27.44 4.69 ± 4.23
Finger Tapping dominant hand 53.44 ± 4.52 56.93 ± 6.95 50.08 ± 4.70 46.62 ± 8.67
Finger Tapping nondominant hand 50.32 ± 2.33 49.83 ± 4.56 45.79 ± 4.72 43.39 ± 7.24
Grip Strength dominant hand 45.88 ± 6.72 39.54 ± 7.97 38.31 ± 8.41 37.08 ± 8.73
Grip Strength nondominant hand 42.20 ± 6.38 36.25 ± 7.58 36.71 ± 8.51 32.35 ± 9.33
FOME-15 immediate 12.84 ± 0.87 12.72 ± 0.88 11.65 ± 1.42 10.71 ± 1.45
FOME-15 delayed 14.20 ± 1.03 14.0 ± 1.0 12.50 ± 1.78 12.38 ± 1.45
Short Categories Test 50.80 ± 23.57 58.25 ± 26.52 42.83 ± 38.09 10.0 ± 11.70
Controlled Oral Word Association Test 65.70 ± 30.40 67.0 ± 20.68 44.75 ± 29.12 44.92 ± 27.08
Open in a new tab

WAIS subtest scores are age-adjusted scaled scores; Shipley is an I.Q. score; Finger Tapping, Grip Strength, and FOME-15 are raw scores; WMS, Rey-Osterrieth (Denman norms), Trail Making (Heaton norms), Short Categories, and COWAT are percentile scores.

MRI Acquisition

MRI studies were performed on a whole body Philips 2 Tesla MRI/MRS system. A sagittal T1 weighted study (TR/TE = 450/30) was used for determination of the canthomeatal plane. This was followed by a spin-echo sequence (TR/TE 3000/30/80 ms) performed parallel to the canthomeatal plane and covering the entire brain from the cerebellum to the vertex. The spin-echo sequence yielded a set of 19- to 24-proton density (the 30 ms echo) and T2 weighted (the 80 ms echo) 5.1 mm thick (with a 0.5 mm interslice gap) axial images.

MRI Analysis

Computer-assisted segmentation of the MRI into ventricular cerebrospinal fluid (CSF), sulcal CSF, gray matter, white matter, WMSHs, and caudate nuclei was performed for all subject by a single operator blind to participant identity. The undifferentiated-tissue-versus-CSF segmentation was performed on sequential slices beginning with the slice on which the red nucleus appeared and ending two slices above the last section on which the lateral ventricles were identified. The gray-versus-white-matter segmentation was performed on a subset of these slices, beginning with the fully volumed lateral ventricles and ending two slices above the last section on which the lateral ventricles were identified. Pixels that were classified by the discriminant analysis as gray matter but which must be white matter on the basis of anatomical location were manually changed to the WMSH classification. Qualitatively, most of the white matter displaying increased signal fell into a “diffuse,” rather than “focal” classification; when focal areas of increased signal were present, these areas were subsumed in the WMSH category. The head of the caudate was delineated on the addition image (i.e., the sum of the proton-density-weighted and T2-weighted images) on the slice on which the head of the caudate was most fully volumed. Figure 1 displays a representative subject's MRI data together with the segmentation results.

Figure 1.

Figure 1

Open in a new tab

A. Segmented image shows white matter signal hyperintensity in the areas below the occipital horns of the lateral ventricle. B. Segmented image shows delineation of caudate nuclei.

Preprocessing involved two steps. The skull was first stripped off the image using a modification of the algorithm developed by Lim et al. (1989). The resulting image, which was subject to all remaining analyses, was limited to the boundaries of the intracranial vault. This was followed by estimation and removal of RF-field inhomogeneity by digital filtering. The inhomogeneity-corrected images were then processed by a trained operator blind to participant identity.

On a slice-by-slice basis, the operator thresholded subtraction images (i.e., the difference image between the T2-and proton-density-weighted images) to specify conservative CSF and non-CSF samples. These samples were chosen to represent tissue that definitely belonged to the correct (i.e., CSF or non-CSF) classification and left a sizable amount of tissue unclassified. For each slice, the conservative CSF and non-CSF samples were then used as training sets for a discriminant analysis that used both the proton-density-weighted and the T2- weighted pixel intensities to classify each pixel as either CSF or non-CSF. Ventricular CSF was separated from sulcal CSF by dividing the area within the intracranial vault into two concentric subregions: CSF within the inner 55% of the intracranial vault was designated as ventricular CSF and CSF in the outer 45% of the vault was designated as sulcal CSF. Ventricular and sulcal CSF volumes aggregated over the slices analyzed were then computed as a percent of the total intracranial vault volume to correct for variation in head size.

The gray-versus-white-matter discrimination was accomplished in a manner similar to the CSF versus non-CSF discrimination. On a slice-by-slice basis, the operator thresholded the addition image to specify conservative white matter and gray matter samples. These samples were chosen to represent tissue that definitely belonged to the correct (i.e., white vs. gray) classification and left a sizable amount of tissue unclassified. For each slice, the conservative tissue samples were then used as training sets for a discriminant analysis which used both the proton-density-weighted and the T2-weighted pixel intensities to classify all non-CSF pixels as either gray or white matter. The white-matter and gray-matter volumes aggregated over the slices analyzed were then computed as a percent of total intracranial vault volume. The WMSH category was computed as a percent of total white matter. The volume of the head of the caudate was not correlated with total intracranial vault volume; therefore it was not corrected for the variation in head size. The interoperator correlations on a subset of 20 scans (both participants and controls) from this sample for percent CSF, percent white matter, percent WMSH, percent gray matter, and volume of the head of the caudate were .99, .82, .87, .66, and .83, respectively.

RESULTS

Neuropsychological Impairment versus Clinical Stage of HIV+ Participants

Table 2 displays the cross-tabulation of severity of NP impairment (GIS) versus CDC clinical status in the HIV+ participants, illustrating the relative independence of NP impairment and CDC clinical stage in this sample [x2(4) = 4.269, p = .37]. Age was not related to severity of NP impairment in the HIV+ sample; however, age was associated with clinical stage (consistent with the older subjects having been infected longer and having more advanced disease). HIV− controls were more highly educated than HIV+ participants [t(39) = 2.29, p = .03], seronegative-versus-seropositive status accounting for 12% of the variance in education. When HIV+ participants were subgrouped by level of NP impairment or by clinical stage, the variance in education accounted for by increasing severity of either of these scales was reduced to 9.4% and 4.3%, respectively.

Table 2.

Neuropsychological impairment versus clinical stage* of HIV+ participants

NP impairment group Stage A Stage B Stage C All HIV+
No cognitive impairment (GIS 0–1) n = 2 n = 2 n = 2 n = 6
age = 35.0 ± 8.2,
educ = 14.7 ± 2.8
NP% = 60.6 ± 11.6
CD4% = 32.7 ± 4.93
Mild-to-moderate cognitive impairment (GIS 2–5) n = 1 n = 4 n = 7 n = 12
age = 39.7 ± 6.7
educ = 15.7 ± 1.4
NP% = 44.3 ± 9.2
CD4% = 6.0 ± 6.7
Severe cognitive impairment (GIS ≥ 6) n = 3 n = 7 n = 3 n = 13
age = 38.7 ± 6.4
educ = 14.5 ± 2.3
NP% = 25.2 ± 5.1
CD4% = 16.2 ± 8.5
All HIV+ n = 6
age = 30.8 ± 3.7
educ = 15.2 ± 2.4
NP% = 45.9 ± 23.5
CD4% = 28.4 ± 45.2		 n = 13
age = 39.0 ± 5.7
educ = 14.2 ± 1.7
NP% = 36.7 ± 14.6
CD4% = 16.4 ± 9.5		 n = 12
age = 41.4 ± 6.7
educ 15.7 ± 2.2
NP% = 39.2 ± 12.9
CD4% = 4.5 ± 3.3
Open in a new tab
*

Centers for Disease Control clinical stages excluding indices of neuropsychological dysfunction; GIS = Global Impairment Score; all of the controls (HIV−) had a GIS of 0–1 (not impaired); NP% = average percentile score on NP tests; educ: controls (HIV−) had a mean educational level of 16.6 ± 1.3 years; all confidence limits are SDs.

In addition to the GIS, which was used to classify HIV+ subjects into impairment subgroups, a mean percentile score for all NP tests was calculated for each subject. The average NP percentile score for the control group was 65.1 ± 4.6%, the average NP percentile score for the nonimpaired HIV+ group was 60.6 ± 11.6% [t(6) = 0.91, p = .398]. Table 2 illustrates the reductions in average NP percentile score for the HIV+ participants with increasing levels of cognitive impairment as indexed by the GIS. Table 2 also illustrates the relative independence of average NP percentile score and CDC stage. Although both stage B and C patients had a lower average NP percentile scores than stage A patients, stage C patients had a higher average NP percentile scores than stage B patients, consistent with the greater percentage of participants with high GIS (severely impaired) in stage B compared to stage C subgroups. CD4% was highly associated with CDC clinical stage but, like clinical stage, relatively independent of severity of NP impairment (GIS). The large SD for the mean CD4% in stage A reflects the large range in CD4% in asymptomatic patients. In sum, in this sample, NP impairment was relatively independent of the progression of systemic HIV disease, as measured both by CDC stage and CD4%.

MRI Differences Between HIV− Controls and all HIV+ Subjects

Table 3 presents the MRI variable comparisons between all HIV+ participants and the HIV− controls. There were no age-related MRI changes found in the controls (whose mean age was almost the same as that of the HIV+ participants but whose age range was greater than that of the HIV+ participants); therefore, age was not used as a covariate in the analysis of MRI changes in HIV+ participants. Education was not correlated with any of the MRI measures, whether the entire sample or the HIV+ group only was considered (all rs < .18, all ps > .27). There was a 27% increase in ventricular CSF percent and a 19.5% increase in sulcal CSF percent in HIV+ participants compared to HIV− controls [t(33) = 3.0, p = .005; and t(39) = 1.98, p = .06, respectively]. Caudate volume was reduced by 8.0% in HIV+ individuals compared to HIV− controls [t(39) = 2.5, p = .018] and percent cortical gray matter evidenced a trend toward a decrease (4.9%) in HIV+ participants [t(39) = 1.73, p = .092]. No differences were detected between the groups in percent white matter [t(39) = 0.42, p = .68] or percent WMSH [t(39) = 0.18, p = .859]. Table 3 also shows the percent of variance of the MRI measures (the dependent variable) accounted for by group membership. The effect of membership in the HIV+ group versus the HIV− group for the MRI variables that demonstrated significant atrophy was between 9.1 and 13.5% of the variance of the respective MRI variable.

Table 3.

MRI variable comparisons between HIV− and HIV+ groups

Group
 Between-group comparison
HIV− (n = 10)
 HIV+ (n = 31)
 Effect Size
(% var)*
MRI variable M ± SD M ± SD t p
Ventricular CSF percent 4.8 ± 0.9 6.1 ± 1.9 10.4 2.99 .005
Sulcal CSF percent 5.1 ± 0.9 6.1 ± 1.5 9.1 1.98 .06
Gray matter percent 40.5 ± 2.6 38.6 ± 3.2 7.1 1.73 .10
White matter percent 46.2 ± 2.0 45.8 ± 2.4 0.4 0.42 .68
WMSH (% of white matter) 5.3 ± 3.1 5.1 ± 3.6 0.1 0.18 .86
Caudate volume (pixels) 306 ± 25.8 281 ± 27.8 13.5 2.46 .02
Open in a new tab
*

Percent of variance of MRI variables accounted for by group membership

MRI Differences Between HIV− Controls and HIV+ Participants Classified by Clinical Stage of HIV Disease

The linear trend analysis presented in Table 4 and illustrated in Figure 2 demonstrates that percent ventricular and sulcal volume increased and gray matter percent and total caudate volume decreased with more advanced clinical stage of HIV disease. For the purpose of this analysis, both controls and stage A were set to 1 (asymptomatic), stage B was set to 2, and stage C was set to 3. The linear trend analyses for gray-matter percent, ventricular and sulcal CSF percent, and caudate volume were all highly significant (all ps < .005). The effect of clinical stage membership accounted for 21.8 to 32.0% of the variance of the MRI variables that demonstrated significant atrophy. These effects are more than twice as large as those presented in Table 3, where the HIV+ sample was not subdivided by clinical stage. Because clinical stage and CD4% are highly associated, an analysis of covariance was performed to determine if either variable was more strongly associated with the MRI measures. Clinical stage had significant independent power beyond CD4% to predict ventricular and sulcal CSF percent and caudate volume. For example, the independent predictive ability of clinical stage beyond CD4% for sulcal CSF percent was significant [F(2,22) = 5.14, p = .015), while the independent predictive ability of CD4% beyond that of clinical stage was nonsignificant [F(1,22) = 2.37, p = .138]. The results were different for gray matter percent, where clinical stage did not have predictive power independent of CD4% [F(2,22) = 1.94, p = .168), but CD4% did have predictive power independent of clinical stage [F(1,22) = 6.36, p = .02].

Table 4.

Association of MRI variables with clinical stage*

CDC clinical stage subgroup means
 Linear trend analysis
Controls
 Stage A
 Stage B
 Stage C
 Effect size
(% var)†
MRI variable M ± SD M ± SD M ± SD M ± SD F p
Ventricular CSF percent 4.8 ± 0.9 5.5 ± 0.9 5.3 ± 1.6 7.4 ± 2.0 23.3 13.3 .0008
Sulcal CSF percent 5.1 ± 0.9 5.1 ± 0.7 5.7 ± 1.1 7.2 ± 1.6 28.8 16.8 .0002
Gray matter percent 40.5 ± 2.6 41.2 ± 2.7 38.8 ± 3.2 37.0 ± 2.6 21.8 10.7 .0023
White matter percent 46.2 ± 2.0 45.7 ± 1.9 46.6 ± 2.5 45.0 ± 2.4 1.7 0.7 .41
WMSH (% of white matter) 5.3 ± 3.1 4.3 ± 2.3 5.8 ± 5.2 4.7 ± 1.6 0.0 0.0 .96
Caudate volume (pixels) 306 ± 25.8 310 ± 24.0 281 ± 22.1 268 ± 25.8 32.0 18.0 .0001
Open in a new tab
*

Centers for Disease Control clinical stages excluding indices of neuropsychological dysfunction.

†

Percent of variance of MRI variables accounted for by increasing severity of systemic disease (stage).

Figure 2.

Figure 2

Open in a new tab

MRI measures are more strongly associated with stage of systemic disease than with cognitive impairment.

In post hoc tests, there were no significant differences in any MRI measure between the controls and stage A participants. The only significant difference between controls and stage B participants was an 8% decrease in caudate volume [t(21) = 2.51, p = .02], although the means for ventricular and sulcal CSF percent were higher and the mean for gray matter percent was lower in stage B participants than in controls (Table 2). Between controls and stage C participants, there was a 54% increase in ventricular CSF percent [t(20) = 4.04, p = .001], a 41% increase in sulcal CSF percent [t(20) = 3.60, p = .002], a 7.5% decrease in gray matter percent [t(20) = 3.11, p = .006], and a 12.4% reduction in caudate volume [t(20) = 3.43, p = .003].

MRI Differences Between HIV− Controls and HIV+ Participants Classified by Level of NP Impairment

The linear trend analysis presented in Table 4 and Figure 2 illustrates the relative lack of a monotonic relationship between MRI variables and severity of NP impairment in the HIV+ patients. The average NP percentile score for the NP impairment subgroups was used to set the distance between the groups (controls = 65.1, HIV+ non-impaired = 60.6, HIV+ mildly-to-moderately impaired = 44.3, HIV+ severely impaired = 25.2). The linear trend analysis results for NP impairment groups and MRI measures shown in Table 5 were much less dramatic than the linear trend analysis results for stage and MRI measures as shown in Table 4. For example, the linear trend analysis using NP impairment subgroups accounted for only 6.7% of the variance in ventricular CSF percent, while the linear trend analysis using clinical stage accounted for 23.3% of the variance in ventricular CSF percent. Table 5 also shows that this difference in results was due to the mildly to moderately impaired group having the largest ventricular volume. It is important to note that the mildly to moderately impaired group had the highest percentage of patients in clinical stage C. Other MRI variables such as caudate volume showed a similar pattern: Increasing severity of NP impairment accounted for only 8.7% of the variance in caudate volume, while increasing clinical stage accounted for 32.0% of the variance in caudate volume. It is interesting to note that, although all of the NP impairment groups have the same mean caudate size, the linear analysis results in a marginally significant p value for caudate volume. This is due to the large drop in mean caudate size between the control group and the NP impairment groups. This establishes a marginally significant linear relationship between the larger mean caudate volume of the controls and the smaller mean caudate volume of the NP impairment groups.

Table 5.

Association of MRI variables with severity of neuropsychological impairment

NP impairment subgroup means
Controls
 No impairment
(GIS 0–1)*
 Mild-to-moderate
impairment
(GIS 2–5)
 Severe
impairment
(GIS ≥ 6)
 Linear trend analysis
MRI variable M ± SD M ± SD M ± SD M ± SD Effect size
(% var)† 		F p
Ventricular CSF percent 4.8 ± 0.9 5.6 ± 1.3 6.9 ± 1.9 5.7 ± 2.0 6.7 3.02 .091
Sulcal CSF percent 5.1 ± 0.9 5.4 ± 1.2 6.8 ± 1.7 5.9 ± 1.3 8.7 4.02 .052
Gray matter percent 40.5 ± 2.6 40.3 ± 3.3 37.7 ± 2.8 38.6 ± 3.4 9.4 4.03 .052
White matter percent 46.2 ± 2.0 45.5 ± 3.0 45.6 ± 2.0 46.1 ± 2.6 0.0 0.0 .98
WMSH (% of white matter) 5.3 ± 3.1 4.7 ± 2.3 3.8 ± 1.7 6.5 ± 4.9 0.8 0.30 .58
Caudate volume (pixels) 306 ± 25.8 282 ± 40.1 282 ± 26.2 282 ± 25.0 8.7 3.71 .062
Open in a new tab
*

GIS: Global Impairment Score.

†

Percent of variance of MRI variables accounted for by increasing severity of NP impairment.

We then compared the MRI measures of the HIV− controls to the MRI measures of each level of NP impairment of the HIV+ subjects (Table 5) to parallel the comparison of the MRI measures of the HIV− controls to each CDC clinical stage of the HIV+ subjects (Table 4). The non-impaired HIV+ group was not different from the HIV- group on any MRI measure. The mildly to moderately impaired HIV1 group showed significant increases in percent ventricular CSF [t(16) = 3.29, p = .005], and in percent sulcal CSF [t (20) = 2.82, p = .01], and significant decreases in gray matter percent [t(20) = 2.39, p = .027], and caudate volume [t(20) = 2.26, p = .035], when compared to the HIV− controls. However, the only significant difference between HIV− controls and severely impaired HIV+ subjects was a decrease in caudate volume [t(21) = 2.26, p = .034].

The effect of membership in an NP impairment subgroup for the MRI measures that showed significant atrophy either did not increase or actually decreased between the mildly to moderately impaired group and the severely impaired group. Finally, an analysis of covariance determined that there was no significant additional predictive ability obtained by adding the interaction of NP impairment and CDC clinical stage to the prediction equation.

In order to determine whether use of a composite NP score (GIS) obscured the relationship of the MRI variables with specific NP tests, we computed the correlation of each MRI variable with each NP test for the entire sample. Despite the very large number of correlations computed and the more precise ratio scaling of the NP tests versus the less precise ordinal scaling of the clinical stages (both of which favor the NP tests over the clinical stage variable for evidencing large correlations with the MRI variables), none of the NP tests were as highly correlated with the MRI measures as was clinical stage. For the most part, different NP tests achieved the highest correlations with each MRI measure. The only discernible pattern was a significant correlation for Finger Tapping and Grip Strength with most of the MRI measures; however, Finger Tapping was significantly correlated with the MRI variables only for performance with the nondominant hand, and Grip Strength was significantly correlated with MRI variables only for performance with the dominant hand.

DISCUSSION

The major finding of this study is that MRI-detectable brain atrophy is strongly associated with CDC stage of systemic HIV disease and much more weakly associated with severity of NP impairment. When HIV+ subjects were subgrouped by clinical stage, the MRI differences were directly and strongly associated with increasing severity of systemic disease. When the same comparisons were made using level of NP impairment to create the subgroups, the MRI differences were only weakly associated with increasing severity of NP impairment. Table 6 compares the percent of variance in the MRI variables accounted for by membership in the various subgroupings (effect size). For three of the four MRI variables that show moderate to large effects, simply dividing the sample into HIV− versus HIV+ is more meaningful in explaining variation among the subjects than subgrouping by increasing severity of NP impairment. In contrast, the effect size of the advancing clinical stage subgrouping is very large for ventricular CSF, sulcal CSF, gray matter, and caudate volume. Given these results, it is doubtful that the substrate of NP impairment in this sample of HIV+ subjects is MRI-detectable brain atrophy. This conclusion is illustrated by the following results: (1) although half the patients in CDC stage A were severely NP impaired, the MRI measures of those subjects were not significantly different from those of controls, and (2) the mildly to moderately NP impaired group had the largest atrophic effects and the greatest number of individuals in CDC stage C.

Table 6.

Effect size comparisons between HIV− versus HIV+, neuropsychological impairment, and clinical stage subgroupings

Membership:
HIV− versus HIV+
subgroups
 Membership:
Increasing severity of
NP impairment subgroups
 Membership:
Advancing clinical
stage subgroups
MRI variable Effect size (% var) p Effect size (% var) p Effect size*
(% var)		 p
Ventricular CSF percent 10.4 .005 6.7 .091 23.3 .0008
Sulcal CSF percent 9.1 .06 8.7 .052 28.8 .0002
Gray matter percent 7.1 .10 9.4 .052 21.8 .0023
White matter percent 0.4 .68 0.0 .98 1.7 .41
WMSH (% of white matter) 0.1 .86 0.8 .58 0.0 .96
Caudate volume (pixels) 13.5 .02 8.7 .062 32.0 .0001
Open in a new tab
*

Percent of variance of MRI variables accounted for by group membership

The mechanisms of HIV associated neuronal injury and loss are still unclear. Although opportunistic infections of the CNS remain the most common specifically identifiable source of neurologic disability (Anders et al., 1986), Masliah et al. (1992a) found moderate-to-severe direct HIV infection of the brain in two-thirds of 107 autopsies in patients with AIDS. Loss of cortical neurons and dendritic pathology (particularly spine loss) may be caused by direct HIV infection of the neuron (Wiley et al., 1991; Masliah et al., 1992b). In addition, stimulation of cytokine production by HIV envelope glycoproteins in HIV infected neurons, macrophages, microglia, and astrocytes may also injure or destroy neurons (Masliah et al., 1992b; Everall et al., 1994). Neuronal damage by direct HIV infection of the CNS is likely to present a complicated picture, in that the neurotoxic pathways are linked and circular; HIV stimulated cytokine production has been shown to lead to further expression of HIV (greater viral burden), so that further direct viral or cytokine-mediated injury to the neuron may occur (Everall et al., 1994). Furthermore, both anterograde cortical degeneration subsequent to subcortical lesions and retrograde degeneration of axons as the result of dendritic pathology may exacerbate the neuronal damage (Wiley et al., 1991). Finally, even given this neuropathology, both Weis et al. (1993) and Everall et al. (1994) failed to find any relationship between decreased neuronal density and NP impairment.

To our knowledge, there are no other studies in the neurological literature that compare the association of atrophic MRI changes with severity of systemic disease versus the association of atrophic MRI changes with severity of NP impairment. Because NP impairment and clinical stage of systemic disease are confounded in most studies of HIV-associated brain atrophy, it has been very difficult to ascertain whether the relationship of morphological brain changes is to NP impairment, to stage of systemic disease, or to both. For example, Hestad et al. (1993), in a study of global and caudate nuclei atrophy, found “only mild to moderate correlations between MRI ratios and NP tests in most cases” (p. 117). Hestad, in the absence of any test of the relationship of MRI variables and clinical stage (or any variable, such as CD4%, that is usually highly associated with clinical stage), concluded that “the results indicate a clear relationship between morphological changes and performance on many of the neuropsychological tests” (p. 117). In a much larger sample of 154 HIV+ individuals versus 44 HIV− controls, Heaton et al. (1995) also found only “modest” associations between NP impairment and quantitative measures of global brain atrophy (p. 247). In their study, clinical stage was compared to global NP impairment and found to be highly related. However, the authors concluded that the “results of MR brain imaging provide additional independent evidence that NP impairment was due to HIV-related brain disease,” ignoring the possibility that the MRI results may have reflected the effects of clinical stage rather than the NP impairment. We also found modest associations between the level of NP impairment and MRI measures of global and caudate nuclei atrophy and a very strong association between NP impairment and clinical stage. However, since our investigation was based on a sampling design that provided relative independence of clinical stage of disease and NP impairment, we were able to conclude that MRI-detectable brain atrophy in HIV infection is primarily associated with clinical stage of disease rather than with NP performance.

Our findings of global and caudate nuclei atrophy in HIV+ subjects compared to high-risk controls is consistent with the findings of other investigators (Levin et al., 1990; Moeller & Backmund, 1990; Post et al., 1991; Pan et al., 1992; Aylward et al., 1993; Hestad et al., 1993; Paley et al., 1994). The volume decrease in the caudate nuclei is also consistent with histopathological studies that show the greatest burden of HIV in the deep gray matter (Masliah et al., 1992a). There was a trend toward a difference between HIV+ participants and controls in cortical gray matter percent and a significant difference in cortical gray matter percent between controls and stage C participants. The loss of cortical gray matter also agrees with postmortem reports of neuronal loss in relatively late-stage HIV infection (Ketzler et al., 1990; Everall et al., 1991; Wiley et al., 1991).

We found no difference in percent white matter between HIV+ subjects and HIV− controls, nor any association with the clinical stage or NP subgroups. However, Jernigan et al. (1993) found a significant decrease in white matter volumes in symptomatic HIV+ subjects relative to asymptomatic HIV+ subjects, high-risk controls and low-risk controls. In addition, the postmortem literature shows clear evidence of HIV associated damage to the white matter (Smith et al., 1990; Masliah et al., 1992a; Liuzzi et al., 1994). We cannot account for the difference in our white matter findings. We do note that direct HIV invasion of the tissue and reactive gliosis are probable and opposing factors in atrophy. Masliah et al. (1992a) found that most white matter has minimal to moderate viral burden but moderate to severe gliosis. Studies of small samples (such as ours) may have difficulty in detecting group differences because of variation in both white matter loss and in degree of white matter gliosis.

We also found no difference in percent WMSH between HIV+ subjects and HIV− controls, nor any association with the stage or NP subgroups. This finding has been replicated in the aforementioned study by Heaton et al. (1995) of 154 HIV+ individuals versus 44 HIV− controls. Either MRI-detectable white matter pathology in HIV infection is an uncommon and/or very late-stage phenomenon (none of our patients were institutionalized), or standard MRI lacks the sensitivity necessary to detect the milder forms of white matter abnormality present in earlier stages of disease. Since autopsy shows less HIV in the cortical gray matter than in the white matter, it is unlikely that the gray matter destruction in our HIV+ sample would be present without preceding or at least concurrent white matter damage. In fact, given that the highest level of HIV is in the deep gray matter, the next highest level in the white matter and the least virus present in the cortical gray matter, it has been hypothesized that direct HIV infection of the brain begins in the subcortical gray matter and proceeds through the white matter to the cortex (Masliah et al., 1992a). This hypothesis would rule out white matter damage as an uncommon or solely late-stage phenomena. It is more likely that mild-to-moderate white matter damage is beyond the resolution of standard MRI. In a review of the postmortem studies of AIDS, Anders et al. (1986:552) notes that there are “many reports” of “patchy, ill-defined areas of demyelination and gliosis”; “often (these areas) are inapparent on cut slices of brain and are noted only on histologic sections, where they may, however, be numerous in a given case.” Newer methods of measuring WMSHs, such as magnetization transfer imaging, may be much more useful than standard MRI in characterizing changes at the histological level, since magnetization transfer effects are theoretically caused by macromolecular changes which precede even the cellular changes associated with subsequent disease development (Lexa et al., 1994; Loevner et al., 1995; Santyr & Mulkern, 1995; Wong et al., 1995).

Although we found MRI-detectable global and caudate nuclei atrophy in HIV infection to be more strongly associated with CDC clinical stage than with cognitive impairment, study of a larger sample of unimpaired subjects in CDC clinical stage C would make our findings more definitive. It is also possible that NP dysfunction may correlate more strongly with measures of specific atrophy in regions other than the caudate. We also cannot rule out a disconnection syndrome between subcortical gray structures and the cortex due to white matter damage; this hypothesis must be tested with a more precise methodology than standard MRI. Finally, other measures of in vivo cerebral function, such as perfusion, electrophysiological, and metabolic studies may advance our understanding of the substrate underlying the progressive NP impairment in HIV disease.

ACKNOWLEDGMENTS

Supported by grants from NIMH, No. MHAZ 5 401 MH45680 (G.F.); NIAAA, No. R01 AA08968 (G.F.); by Department of Veterans' Affairs (DVA) General Medical Research Funds, by a DVA Career Research Scientist Award (G.F.), and by a DVA Psychiatry Research Training Award (S.M.). We acknowledge Nancy Poole, M. S., for her invaluable help in subject recruitment and assessment.

REFERENCES

  1. American Academy of Neurology AIDS Task Force Nomenclature and research case definitions for neurologic manifestations of human immunodeficiency virus-type 1 (HIV-1) infection. Neurology. 1991;41:778–785. doi: 10.1212/wnl.41.6.778. [DOI] [PubMed] [Google Scholar]
  2. Anders KH, Guerra WF, Tomiyasu U, Verity MA, Vinters HV. The neuropathology of AIDS: UCLA experience and review. American Journal of Pathology. 1986;124:537–558. [PMC free article] [PubMed] [Google Scholar]
  3. Aylward EH, Henderer JD, McArthur JC, Brettschneider PD, Harris GJ, Barta PE, Pearlson GD. Reduced basal ganglia volume in HIV-1-associated dementia: Results from quantitative neuroimaging. Neurology. 1993;43:2099–2104. doi: 10.1212/wnl.43.10.2099. [DOI] [PubMed] [Google Scholar]
  4. Benton AL, Hamsher K.deS. Multilingual Aphasia Examination: Manual of instructions. University of Iowa; Iowa City, IA: 1978. [Google Scholar]
  5. Bornstein RA, Nasrallah HA, Para MF, Whitacre CC, Rosenberger P, Fass RJ. Neuropsychological performance in symptomatic and asymptomatic HIV infection. AIDS. 1993;7:519–524. doi: 10.1097/00002030-199304000-00011. [DOI] [PubMed] [Google Scholar]
  6. Castro KG, Ward J, Slutsker L, Buehler J, Jaffe H, Berkelman R, Curran J. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. U.S. Department of Health and Human Services, Centers for Disease Control; Atlanta, GA: 1992. [Google Scholar]
  7. Dal Pan GJ, McArthur JH, Aylward E, Selnes OA, Nance-Sproson TE, Kumar AJ, Mellits ED, McArthur JC. Patterns of cerebral atrophy in HIV-1-infected individuals: Results of a quantitative MRI analysis. Neurology. 1992;42:2125–2130. doi: 10.1212/wnl.42.11.2125. [DOI] [PubMed] [Google Scholar]
  8. Davenport L, Brown FF, Fein G, Van Dyke C. A fifteen-item modification of the Fuld Object-Memory Evaluation: Preliminary data from healthy middle-aged adults. Archives of Clinical Neuropsychology. 1988;3:345–349. [PubMed] [Google Scholar]
  9. Denman SB. Denman Neuropsychology Memory Scale. Author; Charleston, SC: 1984. [Google Scholar]
  10. Everall IP, Glass JD, McArthur J, Spargo E, Lantos P. Neuronal density in the superior frontal and temporal gyri does not correlate with degree of human immunodeficiency virus-associated dementia. Acta Neuropathologica. 1994;88:538–544. doi: 10.1007/BF00296490. [DOI] [PubMed] [Google Scholar]
  11. Everall IP, Luthert PJ, Lantos PL. Neuronal loss in the frontal cortex in HIV infection. Lancet. 1991;337:1119–1121. doi: 10.1016/0140-6736(91)92786-2. [DOI] [PubMed] [Google Scholar]
  12. Grant I, Heaton RK, Atkinson JH, Wiley CA, Kirson D, Velin R, Chandler J, McCutchan JA, the HIV Neurobehavioral Research Center HIV-1 associated neurocognitive disorder. Clinical Neuropharmacology. 1992;15(Suppl 1 Part A):364–365. doi: 10.1097/00002826-199201001-00189. [DOI] [PubMed] [Google Scholar]
  13. Heaton RK, Grant I, Butters N, White DA, Kirson D, Atkinson JH, McCutchan JA, Taylor MJ, Kelly MD, Ellis RJ, Wolfson T, Velin R, Marcotte TD, Hesselink JR, Jernigan TL, Chandler J, Wallace M, Abramson I, the HIV Neurobehavioral Research Group The HNRC 500—Neuropsychology of HIV infection at different stages of disease. Journal of the International Neuropsychological Society. 1995;1:231–251. doi: 10.1017/s1355617700000230. [DOI] [PubMed] [Google Scholar]
  14. Heaton RK, Grant I, Matthews CG. Comprehensive norms for an expanded Halstead-Reitan Battery: Demographic corrections, research findings, and clinical applications. Psychological Assessment Resources, Inc; Odessa, FL: 1991. [Google Scholar]
  15. Hestad K, McArthur JH, Dal Pan GJ, Selnes OA, Nance-Sproson TE, Aylward E, Mathews VP, McArthur JC. Regional brain atrophy in HIV-1 infection: Association with specific neuropsychological test performance. Acta Neurologica Scandinavica. 1993;88:112–118. doi: 10.1111/j.1600-0404.1993.tb04201.x. [DOI] [PubMed] [Google Scholar]
  16. Janssen RS, Saykin AJ, Cannon L, Campbel J, Pinksy P, Hessol NA, O'Malley P, Lifson AR, Doll LS, Rutherford GW, Kaplan JE. Neurological and neuropsychological manifestations of HIV-1 infection: Association with AIDS-related complex but not asymptomatic HIV-1 infection. Annals of Neurology. 1989;26:592–600. doi: 10.1002/ana.410260503. [DOI] [PubMed] [Google Scholar]
  17. Jernigan T, Archibald S, Hesselink J, Atkinson JH, Velin RA, McCuthan JA, Chandler J, Grant I, the HIV Neurobehavioral Research Center Magnetic resonance imaging morphometric analysis of cerebral volume loss in human immunodeficiency virus infection. Archives of Neurology. 1993;50:250–255. doi: 10.1001/archneur.1993.00540030016007. [DOI] [PubMed] [Google Scholar]
  18. Ketzler S, Weis S, Haug H, Budka H. Loss of neurons in the frontal cortex in AIDS brains. Acta Neuropathologica. 1990;80:90–94. doi: 10.1007/BF00294228. [DOI] [PubMed] [Google Scholar]
  19. Levin HS, Williams DH, Borucki MJ, Hillman GR, Williams JB, Guinto FC, Jr., Amparo EG, Crow WN, Pollard RB. Magnetic resonance imaging and neuropsychological findings in human immunodeficiency virus infection. Journal of Acquired Immune Deficiency Syndromes. 1990;3:757–762. [PubMed] [Google Scholar]
  20. Lexa F, Grossman R, Rosenquist A. MR of wallerian degeneration in the feline visual system: Characterization by magnetization transfer rate with histopathologic correlation. American Journal of Neuroradiology. 1994;15:201–212. [PMC free article] [PubMed] [Google Scholar]
  21. Lim K, Pfefferbaum A. Segmentation of MR brain images into cerebrospinal fluid spaces, white and grey matter. Journal of Computer Assisted Tomography. 1989;13:588–593. doi: 10.1097/00004728-198907000-00006. [DOI] [PubMed] [Google Scholar]
  22. Liuzzi GM, Mastroianni CM, Fanelli M, Massetti AP, Vullo V, Delia S, Riccio P. Myelin degrading activity in the CSF of HIV-1-infected patients with neurological diseases. NeuroReport. 1994;6:157–160. doi: 10.1097/00001756-199412300-00040. [DOI] [PubMed] [Google Scholar]
  23. Loevner L, Grossman R, Cohen J, Lexa FJ, Kessler D, Kolson DL. Microscopic disease in normal-appearing white matter on conventional MR images in patients with multiple sclerosis: Assessment with magnetization-transfer measurements. Radiology. 1995;196:511–515. doi: 10.1148/radiology.196.2.7617869. [DOI] [PubMed] [Google Scholar]
  24. Manji H, Connolly S, McAllister R, Valentine AR, Kendall BE, Fell MC, Durrance P, Thompson AT, Newman S, Weller IVD, Harrison MJG. Serial MRI of the brain in asymptomatic patients infected with HIV: Results from the UCMSM/Medical Research Council Neurology Cohort. Journal of Neurology, Neurosurgery and Psychiatry. 1994;57:144–149. doi: 10.1136/jnnp.57.2.144. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Masliah E, Achim CL, Ge N, DeTeresa R, Terry RD, Wiley CA. Spectrum of human immunodeficiency virus-associated neocortical damage. Annals of Neurology. 1992a;32:321–329. doi: 10.1002/ana.410320304. [DOI] [PubMed] [Google Scholar]
  26. Masliah E, Ge N, Morey M, DeTeresa R, Terry RD, Wiley CA. Cortical dendritic pathology in human immunodeficiency virus encephalitis. Laboratory Investigation. 1992b;66:285–291. [PubMed] [Google Scholar]
  27. McArthur J, Kumar A, Johnson D, Selnes OA, Becker JT, Herman C, Cohen BA, Saah A, the Multicenter AIDS Cohort Incidental white matter hyperintensities on magnetic resonance imaging in HIV-1 infection. Journal of Acquired Immune Deficiency Syndromes. 1990;3:252–259. [PubMed] [Google Scholar]
  28. Meyerhoff DJ, MacKay S, Bachman L, Poole N, Weiner NW, Fein G. Reduced brain n-acetylaspartate in cognitively impaired human immunodeficiency virus seropositive individuals suggests neuronal loss: In vivo1H magnetic resonance spectroscopic imaging. Neurology. 1993;43:509–515. doi: 10.1212/wnl.43.3_part_1.509. [DOI] [PubMed] [Google Scholar]
  29. Meyerhoff DJ, MacKay S, Poole N, Dillon WP, Weiner MW, Fein G. N-acetylaspartate reductions measured by 1H MRSI in cognitively impaired HIV-seropositive individuals: Replication and extension. Magnetic Resonance Imaging. 1994;12:653–659. doi: 10.1016/0730-725x(94)92460-0. [DOI] [PubMed] [Google Scholar]
  30. Meyerhoff DJ, MacKay S, Sappey-Marinier D, Deicken R, Calabrese G, Dillon WP, Weiner MW, Fein G. Effects of chronic alcohol abuse and HIV infection on brain phosphorus metabolites. Alcoholism: Clinical and Experimental Research. 1995;19:685–692. doi: 10.1111/j.1530-0277.1995.tb01568.x. [DOI] [PubMed] [Google Scholar]
  31. Miller EN, Selnes OA, McArthur JC, Satz P, Becker JT, Cohen BA, Sheridan K, Machado AM, Van Gorp WG, Visscher B. Neuropsychological performance in HIV-1-infected homosexual men: The multicenter AIDS Cohort Study (MACS) Neurology. 1990;40:197–203. doi: 10.1212/wnl.40.2.197. [DOI] [PubMed] [Google Scholar]
  32. Moeller AA, Backmund HC. Ventricle brain ratio in the clinical course of HIV infection. Acta Neurologica Scandiiavica. 1990;81:512–515. doi: 10.1111/j.1600-0404.1990.tb01010.x. [DOI] [PubMed] [Google Scholar]
  33. Osterrieth P, Rey A. Le test de cope d'une figure complexe [The complex-figure test] Archives de Psychologie. 1944;30:205–221. [Google Scholar]
  34. Paley MNJ, Chong WK, Wilkinson ID, Shepherd J, Clews A, Sweeney B, Hall-Craggs MA, Kendall BE, Newman SP, Harrison MJG. Cerebrospinal fluid-intracranial volume ratio measurements in patients with HIV infection: CLASS image analysis technique. Radiology. 1994;190:879–886. doi: 10.1148/radiology.190.3.8115644. [DOI] [PubMed] [Google Scholar]
  35. Post MJD, Berger JR, Quencer RM. Asymptomatic and neurologically symptomatic HIV-seropositive individuals: Prospective evaluation with cranial MR imaging. Radiology. 1991;178:131–139. doi: 10.1148/radiology.178.1.1984291. [DOI] [PubMed] [Google Scholar]
  36. Reitan RM, Wolfson D. The Halstead-Reitan Neuro-psychological Test Battery: Theory and interpretation. Neuropsychology Press; Tucson, AZ: 1985. [Google Scholar]
  37. Russell EW. A multiple scoring method of assessment of complex memory factors. Journal of Consulting and Clinical Psychology. 1975;43:800–809. [Google Scholar]
  38. Santyr G, Mulkern R. Magnetization transfer in MR imaging. Journal of Magnetic Resonance Imaging. 1995;5:121–124. doi: 10.1002/jmri.1880050121. [DOI] [PubMed] [Google Scholar]
  39. Seilhean D, Duyckaerts C, Vazeux R, Bolgert F, Brunet P, Katlama C, Gentilini M, Hauw JJ. HIV-1-associated cognitive/motor complex: Absence of neuronal loss in the cerebral neocortex. Neurology. 1993;43:1492–1499. doi: 10.1212/wnl.43.8.1492. [DOI] [PubMed] [Google Scholar]
  40. Shipley WC. A self-administering scale for measuring intellectual impairment and deterioration. Journal of Psychology. 1940;9:371–377. [Google Scholar]
  41. Smith TW, DeGirolami U, Henin D, Bolgert F, Hauw JJ. Human immunodeficiency virus (HIV) leukoencephalopathy and the microcirculation. Journal of Neuropathology and Experimental Neurology. 1990;49:359–370. doi: 10.1097/00005072-199007000-00001. [DOI] [PubMed] [Google Scholar]
  42. Sönnerborg A, Sääf J, Alexius B, Strannegård Ö, Wahlund LO, Wetterberg L. Quantitative detection of brain aberrations in human immunodeficiency virus type 1-infected individuals by magnetic resonance imaging. Journal of Infectious Diseases. 1990;162:1245–1251. doi: 10.1093/infdis/162.6.1245. [DOI] [PubMed] [Google Scholar]
  43. Stern Y, Marder K, Bell K, Chen J, Dooneief G, Goldstein S, Mindry D, Richards M, Sano M, Williams J, Gorman J, Ehrhardt A, Mayeux R. Multidisciplinary baseline assessment of homosexual men with and without human immunodeficiency virus infection. Archives of General Psychiatry. 1991;48:131–138. doi: 10.1001/archpsyc.1991.01810260039006. [DOI] [PubMed] [Google Scholar]
  44. Wechsler D. Wechsler Adult Intelligence Scale-Revised. The Psychological Corporation; New York: 1981. [Google Scholar]
  45. Weis S, Haug H, Budka H. Neuronal damage in the cerebral cortex of AIDS brains: A morphometric study. Acta Neuropathologica. 1993;85:185–189. doi: 10.1007/BF00227766. [DOI] [PubMed] [Google Scholar]
  46. Wetzel L, Boll T. Short Category Test, Booklet Format. Western Psychological Services; Los Angeles: 1987. [Google Scholar]
  47. White DA, Heaton RK, Monsch AU, the HIV Neurobehavioral Research Group Neuropsychological studies of asymptomatic human immunodeficiency virus-type-1 infected individuals. Journal of the International Neuropsychological Society. 1995;1:304–315. doi: 10.1017/s1355617700000308. [DOI] [PubMed] [Google Scholar]
  48. Wiley CA, Masliah E, Morey M, Lemere C, DeTeresa R, Grafe M, Hansen L, Terry R. Neocortical damage during HIV infection. Annals of Neurology. 1991;29:651–657. doi: 10.1002/ana.410290613. [DOI] [PubMed] [Google Scholar]
  49. Wilkie FL, Eisendorfer C, Morgan R, Loewenstein D, Szapocznik J. Cognition in early human immunodeficiency virus infection. Archives of Neurology. 1990;47:443–440. doi: 10.1001/archneur.1990.00530040085022. [DOI] [PubMed] [Google Scholar]
  50. Wong KT, Grossman RI, Boorstein J, Lexa FJ, McGowan JC. Magnetization transfer imaging of periventricular hyperintense white matter in the elderly. American Journal of Neuroradiology. 1995;16:253–258. [PMC free article] [PubMed] [Google Scholar]
  51. Zachary RA. Shipley Institute of Living Scale Revised manual. Western Psychology Services; Los Angeles: 1986. [Google Scholar]

RESOURCES