Differential Receptive and Expressive Language Functioning of Children with Symptomatic HIV Disease and Relation to CT Scan Brain Abnormalities

(1)

Differential

Receptive

and

Expressive

Language

Functioning

of Children

with

Symptomatic

HIV

Disease

and

Relation

to CT

Scan

Brain

Abnormalities

Pamela L. Wolters, PhD*; Pim Brouwers, PhD*; Howard A. Moss, PhD*; and Philip A. Pizzo, MD*

ABSTRACT. Objectives. To investigate the effect of

HIV disease on the receptive and expressive language of

children and the relationship between CT scan brain

abnormalities and language functioning.

Methods. Thirty-six children (mean age, 5.5 years; range, 1 through 10 years; 75% vertical transmission; 58% classified as encephalopathic) with symptomatic

HIV infection and 20 uninfected siblings (mean age,

7.8 years; range, 3 through 15 years) were administered

an age-appropriate comprehensive language test

as-sessing both receptive and expressive language

(Reynell Developmental Language Scales or Clinical

Evaluation of Language Fundamentals-Revised).

Each HIV-infected child had a CT scan of the brain as

part of the baseline evaluation, which was rated

inde-pendently and blindly by two neurologists, for

pres-ence and severity of brain abnormalities using a

semiquantitative rating system.

Results. Expressive language was significantly more impaired than receptive language in the overall sample

of HIV-infected children. The encephalopathic children

scored significantly lower than the non-encephalopathic children, however, the degree of discrepancy between

mean receptive and expressive language scores was not

significantly different between these two groups. The

uninfected sibling control group did not have a signifi-cant discrepancy between receptive and expressive

lan-guage, and they scored significantly higher than the

in-fected patient group. Greater severity of CT scan

abnormalities was significantly correlated with poorer

receptive and expressive language functioning in the

overall HIV-infected sample and a higher discrepancy

between receptive and expressive language in the

encephalopathic group.

Conclusion. Pediatric HIV disease is associated with differential receptive and expressive language function-ing in which expressive language is significantly more

impaired than receptive language. The sibling data and

CT scan correlations suggest that the observed language

impairments are associated with the direct effects of

HIV-related central nervous system disease. Pediatrics 1995;95:112-119; pediatric AIDS, HIV infection, neuropsy-chology, language, encephalopathy, CT scan.

From the *Pediatric Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, and the MedicaI Illness Counseling Center, Chevy Chase, MD.

This manuscript consists of part of the first author’s doctoral dissertation at the University of North Carolina at Chapel Hill.

Received for publication Jan 13, 1994; accepted Apr 13, 1994.

Reprint requests to (P.L.W.) Neuropsychology Group, Pediatric Branch, National Cancer Institute, National Institutes of Health, Bldg 10, 13N240, 10 Center Dr, MSC 1928, Bethesda, MD 20892-1928.

Acad-emy of Pediatrics.

ABBREVIATIONS. HIV, human immunodeficiency virus; AIDS,

acquired immune deficiency syndrome; ANOVA, analysis of

variance; CIX, Center for Disease Control; CELF-R, Clinical

Eval-uation of Language Fundamentals-Revised; CNS, central

nervous systern; CT, computed tomography; NCI, National

Cancer Institute.

Children with HIV infection are at risk for

devel-oping neurological and neuropsychological

impair-ments that are associated with the direct effect of

HIV on the central nervous system (CNS).’7 Despite

numerous studies documenting the

neurodevelop-mental deficits observed in pediatric AIDS patients,8

the effect of HIV on the language functioning of

children has not been systematically and thoroughly investigated.

Previous reports of language abnormalities in

chil-dren with HIV infection5’95 are based primarily on

general observations of behavior, tests of cognitive

development, language screening measures, or tests

that assess limited language skills. Most studies used small samples or tended to report mostly descriptive

findings. In addition, previous research has rarely

controlled for various medical and environmental

factors that may affect the cognitive development of

HIV-infected children and confound the effects of

HIV disease on language functioning.16’17

Neverthe-less, some studies suggest that children with

ad-vanced HIV-related CNS disease exhibit greater

im-pairments in their expressive language than

receptive language. 5,10,13,18,19

We previously investigated the adaptive behavior

of children with symptomatic HIV infection

mea-sured by a parent report measure and found that

expressive language was more impaired than

recep-tive language.’8 In another study,8 CT scan brain

abnormalities, rated by a semiquantitative method,

were significantly correlated with degree of

cogni-tive deficits suggesting that these impairments have

an organic basis. The purpose of the present study,

therefore, was to further investigate the extent to

which language functioning is compromised by

pe-diatric HIV disease, and more specifically, to

deter-mine whether receptive and expressive language

functions are affected differentially. Children with

symptomatic HIV infection were assessed with

age-appropriate standardized comprehensive language

tests measuring both receptive and expressive

language skills. A comparison group of uninfected

at Viet Nam:AAP Sponsored on September 1, 2020 www.aappublications.org/news

(2)

siblings of the HIV-infected patients also received a

comprehensive language assessment to evaluate the

influence of the environment on language

function-ing. Furthermore, language test scores were

cone-lated with analog ratings of concurrently obtained

brain CT scans to examine whether language

impairments are associated with structural brain

abnormalities.

Subjects

Patients

METHODS

Children with symptomatic HIV infection (CIX Class P2), ages

I through 13 years, who were consecutively enrolled in

antiretro-viral research protocols at the Pediatric Branch of the National

Cancer Institute (NCI) from June 1990 to November 1991 and

spoke English as their first language, were eligible to participate in this study.

To control for extraneous factors that may confound the effects

of HN on language functioning, the following exclusionary

crite-ria were utilized. Patients could not participate in the study if their

hearing, as measured by speech reception thresholds by a

corn-plete audiological evaluation, was not within normal limits (n = 2

patients). Also excluded were children with known pre-existing

non-HIV-related conditions, such as Down’s syndrome (n = 2

patients), congenital multiple handicaps (n = 2 patients), severe

prematurity with complications (n = I patient), significant

peri-natal trauma associated with developmental delays (n = I

pa-tient), and severe intraventricular hemorrhage (n = 2 patients),

which may have compromised their functioning and confounded

the effects of HIV on the CNS. Two other patients were not

included in the study because they were too sick to participate in

the assessment.

The sample, therefore, consisted of 36 children with a mean age

of 5.5 years and a range of I .2 to 10.8 years. Twenty-seven children

were infected by vertical transmission, while nine were infected

through transfusion of blood or blood products. Twenty-one

chil-dren were classified as exhibiting clinical evidence of encephalop-athy before study entry according to a specific list of criteria (see below). Twenty-one patients were previously treated with

antiret-roviral therapy before being enrolled in this study. The mean

education of the caregivers of the patient group was 12.7 years

with a range of 9 to 18 years. Table I summarizes the sample

characteristics of the children with symptomatic HIV infection

before study entry.

Siblings

The control group consisted of 20 uninfected siblings of the

HP/-infected patients with a mean age of 7.8 years and a range of 3.5 to 15.7 years. Most of the siblings were obtained from separate

families, although in a few cases (n = 3) the family provided more

than one sibling control. The mean education of the caregivers of

the sibling group was 13.7 years with a range of 10 to 18 years. The

educational level of the caregivers who provided a sibling

(M = 13.7 years, range = 10 to 18) did not significantly differ from

the educational level of those who did not provide a sibling

control (M = 12.4 years, range = 9 to 18) (t = 1.86, ns). Sixty

percent of the siblings were female, and 85% were older than 5

years of age and administered the Clinical Evaluation of Language

Fundamentals-Revised (CELF-R). None of the siblings had

ab-normal hearing based on parent report nor any pre-existing CNS

conditions warranting exclusion and all spoke English as their

primary language.

Procedures Patient Recruitment

Children with HIV infection were referred to the NCI from

throughout the United States to participate in clinical trials of

antiretroviral therapy. Intake research nurses initially contacted

the child’s parent or legal guardian by phone after the referral to

discuss the study, medical status of the child, possible compliance issues, and their interest in being involved in research. Patients and their families that appeared to be eligible for the studies then

visited the NCI to complete initial evaluations. Parents or legal

guardians gave informed consent for their children to participate

in the research protocols, which were approved by the NCI

Insti-tutional Review Board. The patients were treated primarily on an

outpatient basis.

Uninfected siblings of the HP/-infected patients were recruited to serve as controls by asking parents if they wanted their child to participate in this study. The sibling came with the family for one visit to the NCI to complete the language evaluation.

Classification of Encephalopathy

The patients were classified before enrollment as exhibiting

HIV-associated encephalopathy by the following non-mutually

exclusive criteria: (1) evidence of deterioration in developmental

or cognitive functioning (eg, parental report of regression in the

child’s developmental milestones or significant decline in

stan-dard scores compared to previously administered tests), (2)

gen-eral level of cognitive functioning more than 2 standard deviations below the norm (eg, standard score of less than 70), and (3) general level of cognitive functioning at least I standard deviation below

the norm and evidence of brain abnormalities on CT scan, such as

cortical atrophy, calcifications, and/or white matter disease.

Al-though these criteria are consistently used at the NC! program,

they may differ somewhat in other institutions due to the lack of

a generally accepted classification system for HIV-associated

encephalopathy in children.20’1

Language Assessment

The children with HIV infection were administered an

age-appropriate comprehensive language test as part of their

pretreat-ment neuropsychological evaluation. The uninfected siblings also

were evaluated with the comprehensive language test appropriate for their age during a visit to the NCI. Since no single test assesses

the broad age range of the sample, the following two language

measures were used.

The Reynell Developmental Language Scales’ were administered to children from 1 to 5 years of age. This comprehensive language

test is organized in a developmental sequence and is comprised of

separate receptive and expressive language scales. The verbal

comprehension scale has 10 sections assessing the child’s

under-standing of a wide range of language concepts of increasing

difficulty. The expressive language scale is divided into three

main sections measuring the structure, vocabulary, and content of

the child’s expressive language. Overall receptive and expressive language z-scores are derived from these scales. For comparison

with the standard scores obtained on the CELF-R, the receptive

and expressive language z-scores were transformed into standard

scores with a mean of 100 and a standard deviation of 15.

The Clinical Evaluation of Language Fundamentals-Revised

eval-uates language disorders in children ages 5 through 16 years of

TABLE 1. Sample Characteristics of the Children With Symptomatic HIV Infe ction at Study Entry

Total, (N = 36) Encephalopathy

Evidence, No Evidence,

(N=21) (N=15)

Mean age at baseline (yrs) (range) 5.5 (1.2-10.8) 4.4 (1.2-10.3) 6.9 (1 .6-10.8)

Male/female (n) 24/12 17/4 7/8

Vertical/transfusion (n) 27/9 18/3 9/6

Reynell/CELF-R(n) 19/17 14/7 5/10

Mean education of caregiver (yrs) (range) 12.7 (9-18) 12.5 (10-18) 13.0 (9-17)

(3)

age by assessing content (semantics), form (syntax and

morphol-ogy), and memory through a comprehensive battery comprised of

11 subtests. Since various language skills develop at different ages, and the test assesses language skills across a wide age range, some

subtests are more discriminating than others for identifying

ab-normal language development at a particular age. Therefore, the

CELF-R has two forms, consisting of the six core subtests (with

some different and some overlapping subtests) required to

corn-pute the receptive and expressive language standard scores (mean

= 100, S.D. = 15) for the younger and older age groups. Form I is

administered to children aged 5 to 7 years and Form 2 is for

children aged 8 to 16 years.

CT Brain Scan Assessment

Each HIV-infected child had a CT scan of the brain as part of

their baseline evaluation. Two neurologists, blinded to the name

and clinical status of the patient, independently rated the scan for presence and severity of various brain abnormalities (ventricular enlargement, cortical atrophy, white matter attenuation, interce-rebral calcifications, and other lesions) using a semiquantitative

rating system described in more detail in a previous report.24 The

severity of each abnormality was rated along a 100-mm analog

scale, from no abnormality present (0 mm) to the presence of very

severe abnormality (100 mm). Finally, each neurologist generated

a summary score, which was a rating of the overall degree of

severity of the CT scan, that took into account each individual

abnormality. Interrater reliability was very high (r = .86;

P < .0001). A mean overall CT scan abnormality score, obtained by

averaging the summary ratings of the two neurologists, was

correlated with the language measures.

Data Analysis

Analyses of variance (ANOVA) were used to compare the

receptive and expressive language scores between the various

subgroups: encephalopathic versus non-encephalopathic;

verti-cally versus transfusion infected; previously treated versus

un-treated; the two cohorts that varied by age and language test

administered (children from I to 5 years of age administered the

Reynell Scales, and those from 5 to 16 years of age administered

the CELF-R); and the HIV-infected patients versus uninfected

sibling control group. Post-hoc comparisons were calculated on

the cell means in the above analyses to further examine the

within- and between-group differences that were significant.

Additional analyses of variance with repeated measures and

the Greenhouse-Geisser correction were used to explore possible

patterns of strengths and weaknesses in the CELF-R subtest scores

for the older patients and siblings administered Form 2. Too few

cases were available to analyze the subtest scores from Form I of

the CELF-R administered to the younger subjects, and the Reynell

Scales did not yield subtest scores. Post-hoc comparisons were

then performed to examine specific within and between group

differences.

The relationship between CT scan brain abnormalities and

language functioning was tested by product moment correlational

coefficients. The mean overall CT scan abnormality scores were

correlated with the receptive language scores, expressive language scores, and the receptive-expressive difference scores (receptive

language score minus expressive language score).

RESULTS

Comparison of Receptive and Expressive Language in

Children With Symptomatic HIV Infection

Overall Sample

The mean difference between the receptive and

expressive language scores in the overall group was

12.8 ± 1.5 points.

Comparison of the Encephalopathic and

Non-encephalopathic Groups

The degree and direction of discrepancy between

the receptive and expressive language scores were

not significantly different between the

encephalo-pathic and non-encephalopathic groups (F = .06; ns).

The mean difference between the receptive and

ex-pressive language scores for the encephalopathic

children was 12.5 ± 2.3 points, while the mean

dif-ference was 13.3 ± I .7 points for the non-encephalo-pathic children. As expected, the encephalopathic

group scored significantly lower than the

non-en-cephalopathic group in overall language functioning (F = 86.6; P < .0001) as well as in both receptive (t =

8.6; P < .0001) and expressive (t = 8.0; P < .0001)

language skills. Figure 1 illustrates the lower

recep-tive and expressive language functioning of the

en-cephalopathic children as well as the similar degree

of discrepancy between the receptive and expressive

language scores in both groups.

Comparison of the Two Cohorts Differing by

Age and Language Test

Overall language functioning was not significantly

different between the two cohorts of children who

were administered the two different language tests

due to their age (F = I .6; ns). The separate effects of

age and test cannot be determined, because these

variables are confounded. As illustrated in Fig 2, the

degree and direction of discrepancy between

recep-110

100

90

U RECEPTIVE

80 EXPRESSIVE

70

60

50

ENCEPHALOPATHIC NON.EWCEPHALOPATHIC

Fig 1. Comparison of mean receptive and expressive language

scores of encephalopathic versus non-encephalopathic children

with symptomatic HIV infection.

A highly significant difference was found between

receptive and expressive language functioning in

this sample of children with symptomatic HIV

infec-tion (F = 69.4; P < .0001) in which expressive

lan-guage was significantly lower than receptive

lan-guage. Twenty out of the 36 patients (55%) had

receptive-expressive language difference scores of 12

points or greater, while this magnitude and direction

of difference occurred in only 15% of the normative

sample of the CELF-R (chi-square = 39.74, P < .001). ‘U 0 U Ca 0

0

z

I-Cl)

(4)

110

100

90

‘U 0 U

CO

80

0

z

.(

Ca

70

60

50

U RECEPTIVE

D

EXPRESSIVE

REYNELL CELFR

(Age Range : (Age Range

1 . 5 yrs) 5 . 13 yrs)

Fig 2. Comparison of mean receptive and expressive language

scores of children with symptomatic HIV infection by age group

and test administered.

five and expressive language scores were not

statistically different (F = .19; ns) in the younger

children administered the Reynell (mean difference

= 12.1 points) and older children given the CELF-R

(mean difference = 13.4 points).

Comparison of the Vertically and Transfusion

Infected Children

The overall language functioning of the vertically

infected children did not differ significantly from the

children infected by transfusions (F = 2.7; ns). In

addition, the degree and direction of discrepancy

be-tween receptive and expressive language was not

significantly different in the two groups (F = 1.38; ns).

Comparison of the Previously Treated and

Untreated Patients

Overall language functioning was not significantly different between the patients who received previous

antiretroviral therapy and those who had not been

previously treated (F = 3.53, ns). Furthermore, the

magnitude and direction of the discrepancy between

receptive and expressive language scores at baseline

was not significantly different between the

previ-ously treated and untreated patients (F = .30, ns).

Comparison of Receptive and Expressive Language Between HIV-Infected Patients and Uninfected Siblings

Comparison of Group Means

The uninfected sibling control group scored

signif-icantly higher than the HIV-infected patient group in

overall language functioning (F = 20.17, P < .0001) as

well as in both receptive (t = -3.02, P < .001) and

expressive language (t = -5.57; P < .0001). A

signif-icant difference was also found between receptive

and expressive language (F = 31.18; P < .0001), and

an interaction (F = 25.87, P < .0001) indicated that the

discrepancy between receptive and expressive

lan-guage scores was significantly greater in the

HIV-infected patient group than in the uninfected sibling

group. For the sibling group, the overall mean

recep-tive language score was 101.2 ± 2.5, the mean

ex-pressive language score was 100.6 ± 3.2, and the

mean difference between the receptive and

expres-sive language scores was only 0.6 points. A

compar-ison of the receptive and expressive language scores

between the patient and sibling groups are

illustrated in Fig 3.

Comparison of Matched Patient-Sibling Pairs

An additional analysis of variance was performed

by matching the HIV-infected patient with his or her

uninfected sibling who was the closest in age (n = 10

pairs) and comparing their receptive and expressive

language scores. Results confirmed the previous

analysis: A main effect showed a significant

differ-ence between receptive and expressive language

(F = 8.35; P < .05), while an interaction indicated that

the discrepancy between receptive and expressive

language was significantly different between the

pa-tients and their matched siblings (F = 7.22; P < .05).

Further analysis showed that the mean receptive and

expressive language scores were significantly

differ-ent in the patients (t = 3.97; P < .01), but not in the

siblings (t = .45, ns) of the matched pairs.

I 10

100

90

‘U 0 U

Ca

80

0 z

.(

Ca

70

60

50

PATIENTS SIBLINGS

U RECEPTIVE

EXPRESSIVE

Fig 3. Comparison of mean receptive and expressive language

scores of symptomatic HP/-infected patients versus an uninfected sibling control group.

(5)

Comparison of the CELF-R Subtests Scores Between the Older Patients and Siblings

Significant differences were found among the

mean scores on the six subtests on Form 2 of the

CELF-R (F = 5.20; P < .01). An interaction indicated

significant differences in the subtest scores between the older HIV-infected patients and older uninfected siblings (F = 3.43; P < .05). Separate post-hoc

ANO-VAs of the patient’s and sibling subtest scores

showed a significant difference between the patient’s

scores (F = 5.98; P < .01) but not between the

sib-ling’s scores (F = 2.95; ns). Post-hoc contrasts then

were calculated on the subtest means within the

patient group to determine which means were

sig-nificantly different. Within the expressive language subtests, the patient group scored significantly lower

on “Formulated Sentences” than on “Recalling

Sen-tences” (F = 5.22; P < .05). No significant differences

were found between the receptive language subtests

for the older HIV-infected patients. Post-hoc t tests

indicated that the patient group scored significantly

lower than the sibling control group on the

expres-sive language subtests “Formulated Sentences”

(t = -2.32; P < .05) and “Sentence Assembly” (t =

-3.79; P < .001). No significant differences were

found between the patients and siblings on the three

receptive language subtests or the expressive

language subtest “Recalling Sentences.”

Correlations of CT Scan Brain Abnormalities With

Measures of Language Functioning

Thirty-four children in the sample had CT scans of

the head taken before starting their treatment

proto-col (two encephalopathic children did not receive a

CT scan at NCI at baseline). According to the

neu-rologists ratings, all the encephalopathic children

ex-hibited at least one CT scan abnormality, ranging

from mild to severe atrophy, white matter

hypoden-sity, and calcifications of the basal ganglia.

Seventy-three percent of the non-encephalopathic children

had some CT scan abnormality, which tended to be

minimal cortical atrophy. No laterality in brain

ab-normalities was observed on the CT scans.

Pearson product moment correlations were

calcu-lated to determine the relationship between the mean

overall CT scan abnormality score and three

mea-sures of language functioning: receptive language

score, expressive language score, and the

receptive-expressive difference score (see Table 2). For the

overall sample, higher mean overall CT scan

abnor-mality scores were correlated with lower receptive

language scores (r = - .35; P < .05) and lower

expres-sive language scores (r = - .50; P < .01). Higher mean

overall CT scan abnormality scores also were

associ-ated with a greater difference between receptive and

expressive language scores for the encephalopathic

children (r = .52; P < .05) but not for the

non-encephalopathic group (r = - .15; ns).

DISCUSSION

This study indicates that pediatric HIV disease is

associated with differential receptive and expressive

language functioning. Both the encephalopathic and

TABLE 2. Pearson Product Moment Correlations of the Mean

Overall CT Scan Abnormality Score With Measures of Language

Functioning

Language Measure Mean Overall CT Scan

Abnormality Score

Overall Encephalopathy

Sample

(N = 34) Evidence No

(n = 19) Evidence

(n = 15)

Receptive language score

Expressive language score Rec-Exp difference score

r= r = .20 r= -.22

r= -.50 r= -.31 r= .29

r= .26 r= .52* r= -.15

*P < .05.

:1:P < .01.

non-encephalopathic groups exhibited a consistent

language pattern: Expressive language was

signifi-cantly poorer than receptive language with a similar

discrepancy between the mean receptive and

expres-sive language scores. Thus, expressive impairments

were evident in this sample of children with

symp-tomatic HIV infection regardless of HIV-related CNS

disease. Furthermore, the encephalopathic group

scored significantly lower than the

non-encephalo-pathic group on both receptive and expressive

lan-guage measures. Therefore, advanced HIV-related

CNS disease in children is associated with deficits in

both receptive and expressive language, although

expressive language skills are more severely

impaired.

Several significant correlations were noted

be-tween measures of language functioning and ratings

of CT scan brain abnormalities. In general, these

findings were consistent with data reported by other

studies8’25 in which CT scan abnormalities were

sig-nificantly correlated with neuropsychological

func-tioning in HIV-infected children. In the present

study, a higher mean overall CT scan abnormality

score was associated with lower receptive and

ex-pressive language scores in the overall sample.

Therefore, children with a greater severity of brain

abnormalities had more deficient language skills.

Furthermore, in the encephalopathic group, higher

overall CT scan abnormality scores were associated

with greater differences between receptive and

ex-pressive language scores. These correlations suggest

that the progression of CNS disease, resulting in

more severe brain abnormalities, is related to an

increasingly greater impairment in overall language

functioning, and in the more advanced stages of

encephalopathy, is particularly associated with more

impaired expressive language compared to receptive

language. Thus, children with HIV-related CNS

dis-ease may exhibit differential deterioration of some

skills compared to others.

On the other hand, in the overall and

non-en-cephalopathic groups, the comparative weakness in

expressive language was not correlated with severity

of the CT scan rating suggesting that

neuropsycho-logical changes may occur early in the disease

pro-cess before structural abnormalities are observable

on CT scans. In this regard, CT scan findings tend to

reflect gross brain abnormalites, and histological

(6)

damage to the brain may not be apparent on CT

scans. Other studies also have indicated that mild

cognitive impairments may occur in pediatric AIDS

patients before any clinical signs of HIV-related CNS

disease are present.18’2628 Thus, HIV-associated

en-cephalopathy appears to occur on a continuum

rang-ing from mild to severe CNS effects. Dichotomizing children into either encephalopathic or non-encepha-lopathic groups by a list of clinical criteria may not be

the most sensitive method for investigating the

ef-fects of HIV in the CNS, thus allowing more subtle

and differential changes in functioning to be missed.

The results of this study support the hypotheses

suggested by earlier studies that language function-ing, especially expressive language, is compromised

by HIV disease in children.5’9’15”8”9 The present

study extends previous research by administering

ob-jective and comprehensive language measures that

yield both receptive and expressive language

stan-dard scores within a single instrument, controlling

for the effect of possible confounding environmental

factors on language, correlating language test scores

with ratings of CT scan brain abnormalities, and

employing statistical methods to analyze the data

rather than simply presenting descriptive findings.

Two issues need to be considered when

interpret-ing these results: 1) the potential presence of other

extraneous factors that may have confounded the

effects of HIV on the language functioning of

chil-dren, and 2) the methodology and characteristics of

the sample used in the study. The first consideration

is the possibility that environmental, prenatal, and

medical factors often present in this population may

have had a potential negative impact on language

functioning. For example, the majority of children

with vertically acquired HIV infection tend to come

from lower socioeconomic environments,29 which is

associated with language delay3#{176}and lower usage of

verbal (expressive) language.3’ In cases where

verti-cal transmission is associated with drug abuse of the

parents, poor prenatal care, exposure in utero to

drugs or alcohol3235 and premature birth, may have

a detrimental effect on the development of the

new-born. In some children with transfusion-acquired

HIV infection, the medical condition that required

the initial transfusion, such as prematurity or

child-hood cancer,37 may have already had a detrimental

effect on their development. HIV-infected children

also are at risk for chronic otitis media that may

affect hearing and, if severe and protracted, may

interfere with language functioning.

An attempt was made to control for some of these

variables. Patients were excluded when their birth or

medical history revealed evidence of pre-existing

CNS conditions or when an audiological evaluation

identified significant hearing loss. Other variables

did not appear to significantly influence the results

of this study. The demographic characteristics of the

sample are representative of the general population

and thus suggest that the results were not likely to be

biased by socioeconomic factors. The HIV-infected

children and their families who are treated at the

NCI form a heterogeneous sample that come from

throughout the United States, are from a range of

socioeconomic backgrounds, and have a variety of

risk factors including heterosexual and

transfusion-related transmission. Mean years of education (used

as a measure of socioeconomic status) of the primary

caregiver (most often mothers) in this sample is

com-parable to the educational level of female adults in

the United States39 and to the CELF-R

standardiza-lion sample (mean of approximately 12 years of

ed-ucation). Based on the normative data from the

CELF-R, the mean difference between receptive and

expressive language functioning is expected to be 0.23

For the HIV-infected sample of children, however,

the mean receptive-expressive language score

differ-ence was 12.8 points. Fifteen percent or less of the

standardization sample had a receptive-expressive

difference score greater than 12 points, while this

large discrepancy occurred in 55% of the sample of

HIV-infected children. Thus, a comparison of the

normative data from the CELF-R with the scores

obtained by the HIV-infected children illustrates the

significant differences in receptive and expressive

language despite the similar educational levels of the

two samples.

Evidence from this study also suggests that the

observed language impairments are most likely

as-sociated with direct effects of HIV-related CNS

dis-ease rather than by the influence of environmental

factors. First, the uninfected sibling control group,

who were raised in similar environments as the

in-fected patients, had higher overall language

func-tioning than the HIV-infected children and no

signif-icant difference between receptive and expressive

language. The sibling data, therefore, suggest that

the language deficits observed in the HIV-infected

children were primarily associated with the disease

rather than environmental influences. In addition,

measures of language functioning were significantly

correlated with CT scan abnormalities. These

corre-lations further indicate that the observed language

deficits are most likely associated with direct effects

of HIV on the CNS rather than by possible

confound-ing factors. Finally, the language functioning of the

vertically infected group was not significantly

differ-ent from the transfusion group, although the

major-ity of vertically infected children tend to live in lower socioeconomic environments.29

The second issue that needs to be addressed when

interpreting the findings of this study concerns some

aspects of the methodology. First, the sample was

not random; instead, the children were enrolled on

the study because they were medically eligible for

antiretroviral treatment. Second, all the children in

this study had developed CDC Class P2

symptom-atic HIV disease (versus asymptomatic HIV

infec-tion), thus, the results can only be generalized to

other children with similar symptoms. Third,

break-ing down the data into subgroups for comparisons

based on encephalopathy or age or for the subtest

analyses resulted in somewhat small subgroup sizes,

and thus, the power to detect differences was

re-duced. Finally, the use of two different language

tests may also be a possible limiting factor, because

the language functions measured by each test

were somewhat different although developmentally

(7)

appropriate. However, a single test does not exist

that comprehensively measures language skills of

the complete age range involved in this study, and

furthermore, the two cohorts administered the

different tests exhibited comparable language

functioning.

The neuropathogenesis of language impairments

in pediatric HIV disease is unknown. Language is a

very complex construct, comprised of many

interre-lated components, any of which are susceptible to

dysfunction.”#{176} This study yielded some new

informa-tion about language functioning and associated brain

abnormalities in HIV-infected children that may be

used to generate hypotheses regarding the

pathogen-esis of the observed language deficits. The primary

CT scan abnormality observed in the

non-encepha-lopathic children was minimal cortical atrophy,

while more severe CNS abnormalities, such as

sig-nificant atrophy, white matter hypodensity, and

cal-cifications of the basal ganglia were observed in the

encephalopathic children. The language

impair-ments progressed from weaknesses in expressive

language in the non-encephalopathic children to

se-vere impairments in both receptive and expressive

language in the encephalopathic children. The

signif-icant correlations between the brain scan

abnormal-ities and language scores provides sufficient

evi-dence to support the hypothesis that HIV-related

CNS disease contributes to the language

impair-ments observed in pediatric AIDS. The CT scan data,

however, did not suggest that language deficits were

associated with a specific pattern of abnormality:

right/left differences in the CT scan brain

abnormal-ities were not observed, and focal lesions, such as

calcifications, did not correlate with specific

lan-guage impairments in this group of children.

How-ever, calcifications of the basal ganglia have been

frequently observed in children with HIV disease

and associated with cognitive delays.8’24

Further-more, damage to the basal ganglia may be associated

with language dysfunction, particularly in the

ex-pressive modalities.41’42 Thus, a larger sample of

en-cephalopathic children with a broad range of

abnor-malities will be needed to examine the relation

between these language impairments and specific

brain pathology.

Verbal expression is comprised of both language

and speech.43 Thus, deficits in either system may

contribute to expressive language deficits. One factor that is likely to be involved, particularly in children

with encephalopathy, is motor impairments. Many

encephalopathic children exhibit mild to severe

im-pairment of motor skills” and abnormalities in

oral-motor functioning.’3 A study by Moss, Wolters,

Brouwers, Hendricks, and Pizzo (submitted for

pub-lication) also found that verbal expression was

highly correlated with motor skills in

encephalo-pathic children based on ratings of behavior in

vid-eotaped situations. Thus, motor deficits, including

those in muscle coordination and motor programing,

may affect oral-motor skills and contribute to feeding problems, articulation errors, and speech difficulties.

Impaired motor functioning, however, does not

ap-pear to explain the extent and types of observed

language deficits, especially the expressive language

weakness in the non-encephalopathic children who

do not exhibit motor problems. The older

HIV-in-fected children, for example, had lower scores on

subtests involving generating sentences versus

re-peating sentences. Such a pattern suggests problems

in memory and the retrieval of information, as well

as difficulties in word-finding and syntactical

struc-ture rather than in oral-motor difficulties. Further

research is needed to study more specific areas of

language, oral-motor functioning, and memory skills

to further investigate the types and causes of

lan-guage impairments in children with symptomatic

HIV disease and to compare these deficits to children

with other developmental language disorders.

Finally, this study has implications for the clinical

management of pediatric HIV patients. First, a

com-prehensive assessment of language abilities is

impor-tant to include in a developmental evaluation of

HIV-infected children for early identification of

ab-normalities. Delays in the development of expressive

language skills in infants and young children may be

an early indicator of HIV-related CNS disease and

justify the consideration of antiretroviral treatment.

Second, the periodic assessment of language

func-tioning is useful for monitoring disease progression and assessing the effectiveness of antiretroviral

ther-apies in the recovery of impaired language skills.

Third, speech and language services are

recom-mended to provide evaluation, rehabilitation

ther-apy, and alternative communication systems to help

prevent, ameliorate, and compensate for the

lan-guage and speech deficits observed in children with

HIV disease.

ACKNOWLEDGMENTS

This research was carried out in part by Research Contract

NCI-CM-17529-41 awarded to the Medical Illness Counseling

Center, Chevy Chase, MD.

We wish to thank Lucy Civitello, MD, Charles DeCarli, MD,

Anita Pikus, PhD, Renee Smith, MS. Sharmista Bose, MS. Peggy

McCardle, PhD, Rune Simeonsson, PhD, and Eugene Tassone for

their assistance with this research.

REFERENCES

I. Brouwers P, Belman AL, Epstein LG. Central nervous system

involvement: manifestations and evaluation. In: Pizzo PA, Wilfert KM, eds. Pediatric AIDS: The Challenge ofHIV infection in Infants, Children, and

Adolescents. Baltimore, MD: Williams & Wilkins; 1990:318-335

2. Epstein LG, Goudsmit J,Paul DS, et al. Expression of human immuno-deficiency virus in cerebrospinal fluid of children with progressive

encephalopathy. Ann Neurol. 1987;21 :397-401

3. Epstein LG, Sharer LR, Goudsmit J. Neurological and neuropathological

features of human immunodeficiency virus infection in children. Ann

Neurol. 1988;23(suppl):519-S23

4. Epstein LG, Sharer LR, Joshi VV, Fojas MM, Koenigsberger MR. Oleske

JM. Progressive encephalopathy in children with acquired immune

deficiency syndrome. Ann Neurol. 1985;1 7:488-496

5. Epstein LG, Sharer LR, Oleske JM, et al. Neurologic manifestations of human innumodeficiency virus infection in children. Pediatrics. 1986; 784:678-687

6. Ho D, Pommerantz R, Kaplan J. Pathogenesis of infection with human

immunodeficiency virus. New EngI I Med. 19873175:278-286

7. Sharer LR, Epstein LG, Cho E, et al. Pathologic features of AIDS en-cephalopathy in children: evidence for LAV/HTLV-III infection of brain. Hu,n Pathol. 1986;17:271-284

8. Brouwers P, DeCarli C, Civitello L, Moss H, Wolters P. Pizzo P.

Corre-lation between CT-brain scan abnormalities and neuropsychological

(8)

function in children with symptomatic HIV disease. Arch Neurol. In

press.

9. Belman AL, Diamond G, Dickson D, et al. Pediatric acquired

immune-deficiency syndrome: neurologic syndromes. Am I Dis Child. 1988;142:

29-35

10. Condini A, Axia G, Cattelan C, et al. Development of language in

18-30-month-old HIV-1-infected but not ill children. AIDS. 1991;5:

735-739

I I. McCardle P, Nannis E, Smith R, Fischer G. Patterns of perinatal

HIV-related language deficit [Abstract WB 2021]. In: Proceedings from the VII International Conference on AIDS. Florence, Italy: 1991;2:187

12. Nozyce M, Diamond G, Belman A, et al. Neurodevelopmental

impair-ments during infancy in offspring of IVDA and HJV seropositive moth-ers [Abstractl. Pediatr Res. 1989;25:359A

13. Pressman H. Communication Disorders and Dysphagia in Pediatric AIDS. ASHA. 1992;34:45-47

14. Ultmann MH, Belman AL, Ruff HA, et al. Developmental abnormalities

in infants and children with acquired immune deficiency syndrome

(AIDS) and AIDS-Related complex. Dev Med Child Neurol. 1985;27:

563-571

15. Ultmann MH, Diamond GW, Ruff HA, et al. Developmental abnormal-ities in children with acquired immunodeficiency syndrome (AIDS): a follow up study. Int JNeurosci. 198732:661-667

16. Brouwers P, Moss H, Wolters P.Schmitt F. Developmental deficits and

behavioral change in pediatric AIDS. In: Grant I, Martin A, eds. Neuro-psychology of HIV: Current Research and New Directions. New York: Oxford University Press; 1994:310-338

17. WhittjK, Hooper SR. Tennison MB, et al. Neuropsychologic functioning of human immunodeficiency virus-infected children with hemophilia.

IPediatr. 1993;122:52-59

18. Wolters PL, Brouwers P, Moss HA, Pizzo PA. Adaptive behavior of

children with symptomatic Hil/ infection before and after zidovudine therapy. IPediatr Psycho!. 1994;19:47-61

19. Wolters PL. The receptive and expressive language functioning of

chil-dren with Acquired Immune Deficiency Syndrome [Doctoral

dissertation; University of North Carolina at Chapel Hill 1992.] Disser-tation Abstracts International. 199353:3811B

20. American Academy of Neurology. Nomenclature and research case

definitions for neurologic manifestations of human immunodeficiency virus-type I (HP/-I) infection. Neurology. 1991;41:778-785

21. Belman AL. Acquired immunodeficiency syndrome and the child’s

central nervous system. Pediatr Clin North Am. 1992394:691-714 22. Reynell 1. Reynell Developmental Language Scales. Windsor:

NFER-Nelson; 1977

23. Semel E, Wiig E, Secord W. Clinical Evaluation ofLanguage

Fundamentals-Revised. San Antonio, TX: Psychological Corporation; 1987

24. DeCarli C, Civitello LA, Brouwers P. Pizzo PA. The prevalence of computed tomographic abnormalities of the cerebrum in 100 consecu-live children symptomatic with the human immunodeficiency virus.

Ann Neurol. 199334:198-205

25. DeCarli C, Fugate L, Falloon J, et al. Brain growth and cognitive im-provement in children with human immune deficiency virus-induced encephalopathy after six months of continuous infusion zidovudine therapy. IAcquired Immune Defic Syndr. 1991;4:585-592

26. Brouwers P, Moss H, Wolters P. et al. Effect of continuous infusion

zidovudine therapy on neuropsychologic functioning in children with

symptomatic human immurtodeficiency virus infection. I Pediatr. 1990;

117:980-985

27. Brouwers F, Moss H, Wolters P. El-Amin D, Tassone E, Pizzo P.

Neu-robehavioral typology of school-age children with symptomatic HIV

disease [Abstractl. JClin Exp Neuropsychol. 1992;14:113

28. Pizzo P, Eddy J, Falloon J, et al. Effect of continuous intravenous

infusion of zidovudine (AlT) in children with symptomatic HIV

infec-tion. New EngI I Med. 1988;31914:889-896

29. Rogers M, Thomas P, Starcher E, Noa M, Bush T, Jaffe H. Acquired immunodeficiency syndrome in children: report of the Center for Dis-ease Control National Survelliance, 1982 to 1985. Pediatrics. 1987;79:

1008-1014

30. Silva P. Epidemiology, longitudinal course, and some associated fac-tors. In: Yule W, Rutter M, eds. Lrnzguage Development and Disorders.

Philadelphia: JB Lippincott Co; 1987

31. Puckering C, Rutter M. Environmental influences on language devel-opment. In: Yule W, Rutter M, eds. Language Development and Disorders.

Philadelphia: JB Lippincott Co; 1987

32. Day N, Jasperse D, Richardson G, et al. Prenatal exposure to alcohol:

effect on infant growth and morphologic characteristics. Pediatrics. 1989;

843:536-541

33. Fulroth R, Phillips B, Durand D. Perinatal outcome of infants exposed to cocaine and/or heroin in utero. Am JDis Child. 1989;143:905-910

34. MacGregor SN, Keith LG, Chasnoff IJ, et al. Cocaine use during

pregnancy: adverse pennatal outcome. Am IObstet Gynecol. 1987;1573:

686-690

35. Naeye R, Blanc W, Leblanc W, Khatamee M. Fetal complications of

maternal heroin addiction: abnormal growth, infections, and episodes of stress. IPediatr. 1973;83:1055-1061

36. Spreen 0, Tupper D, Risser A, Tuokko H, Edgell D. Prematurity and low birth weight. In: Spreen 0, ed. Human Developmental Neuropsychol-ogy. New York: Oxford University Press; 1984

37. Poplack DG, Brouwers P. Adverse sequalae of central nervous system therapy. Clin Oncol. 1985;4:263-285

38. Klein J.Otitis media and the development of speech and language.

Pediatr Infect Dis. 1984;3:389-391

39. Census Bureau, Department of Commerce. Educational Attainment in the United States. 1987

40. Rapin I, Allen DA. Developmental language disorders: nosologic

con-siderations. In: Kirk U, ed. Neuropsychology of Linguage, Reading, and Spelling. Orlando: Academic Press; 1983:155-184

41. Brunner R, Kornhuber H, Seemuller E, Suger G, Wallesch C. Basal ganglia participation in language pathology. Brain and Language. 1982; 16:281-299

42. Wallesch C, Kornhuber H, Brunner R, Kunz T, Hollerbach B, Suger G.

Lesions of the basal ganglia, thalamus, and deep white matter: differ-ential effects on language functions. Brain and Language. 1983;20:

286-304

43. Baken RJ. Getting ready to talk: the infants acquisition of motor

capa-bility for speech. In: Kirk U, ed. Neuropsychology of Language, Reading,

and Spelling. Orlando: Academic Press; 1983:83-95

44. Aylward EH, Butz AM, Hutton N, Joyner ML, Vogelhut JW. Cognitive and motor development in infants at risk for human immunodeficiency virus. Am I Dis Child. 1992;146:218-222

Experience is what you get from not having it when you needed it most.

Anonymous

(9)

1995;95;112

Pediatrics

Pamela L. Wolters, Pim Brouwers, Howard A. Moss and Philip A. Pizzo

Symptomatic HIV Disease and Relation to CT Scan Brain Abnormalities