Journal of Pediatric Psychology, Vol. 25, No. 7, 2000, pp.

503–513

Neurocognitive Functioning and Magnetic

Resonance Imaging in Children With
Sickle Cell Disease
Ronald T. Brown,1 PhD, Patricia C. Davis,2 MD, Richard Lambert,3 PhD,
Lewis Hsu,2 MD, PhD, Katharine Hopkins,2 MD, and James Eckman,2 MD
1
Medical University of South Carolina, 2 Emory University School of Medicine, and
3
University of North Carolina at Charlotte

Objective: To examine neurocognitive functioning in children classified with overt cerebral vascular acci-
dents (CVAs), silent infarcts, or without central nervous system (CNS) pathology on magnetic resonance
imaging.
Methods: Participants were 63 children and adolescents with sickle cell disease (SCD).
Results: Children with overt CVAs and silent infarcts differed from their peers without CNS pathology on
measures of attention and executive functioning.
Conclusions: We consider these deficits the result of the high frequency of frontal lobe deficits incurred by
children with SCD. Recommendations include the use of tests designed to measure attention and executive
functioning as a way of screening children with SCD for possible CNS pathology. We also suggest that fu-
ture research examine the mechanism underlying frontal lobe involvement for individuals with SCD.

Key words: sickle cell disease; neurocognitive; cerebral vascular accident; attention.

Children who suffer from sickle cell disease (SCD) quires rehabilitation (Cohen, Branch, McKie, & Ad-
are at significant risk for cognitive impairments ams, 1997; Harriman, Griffith, Hurtig, & Keehn,
(Bonner, Gustafson, Schumacher, & Thompson, 1991). Cohen and associates (Cohen et al., 1997)
1999; Brown, Armstrong, & Eckman, 1993; Brown, compared parents’ reports of children who sus-
Doepke, & Kaslow, 1993; White & DeBaun, 1998). tained CVAs in the left hemisphere with normally
For example, 7% to 17% of children with SCD developing children. Children with CVAs demon-
(HbSS type, the homozygous condition known as strated marked impairments in attention and hy-
sickle cell anemia) sustain a cerebral vascular ac- peractivity. They also showed significant deficits in
cident (CVA) (stroke), a condition that frequently intellectual functioning, academic achievement,
results in significant learning impairments and re- and visual-motor impairments. Similarly, Schatz et
al. (1999) showed that children with SCD who sus-
All correspondence should be sent to Ronald T. Brown, Department of tained anterior cerebral infarcts had deficits on neu-
Pediatrics, Medical University of South Carolina, 171 Ashley Avenue,
P.O. Box 250561, Charleston, South Carolina 29425. E-mail: brownron
rocognitive measures of attention and executive
@musc.edu. functioning. DeBaun et al. (1998) found that atten-

䉷 2000 Society of Pediatric Psychology

504 Brown et al.

tion and executive functioning screening measures The incidence of central nervous system (CNS) ab-
were the most useful in identifying children with normalities identified on MRI was 17%, with over
cerebral infarcts. 10% of the sample categorized as having a silent in-
Research further indicates that children with farct. When children with HbSS were considered
SCD but without any overt signs or symptoms of a alone, over one-fifth of the sample had abnormal
cerebral vascular accident may still be at significant findings on MRI, over 15% of which were consistent
risk for cognitive impairment. For example, Fowler with a silent infarct. Because most of the patients
and colleagues (1988) and Swift and associates with silent infarcts had HbSS type SCD, neurocogni-
(1989) reported that, compared to healthy and tive functioning was examined for these partici-
comparison sibling controls, children with SCD pants. Children with a clinical history of a CVA
without overt symptoms of a CVA still scored lower were compared to those with MRI evidence of a si-
on tests of intellectual functioning and academic lent infarct or no MRI abnormality. Measures in-
achievement. Wasserman, Wilmas, Fairclough, cluded tests of intelligence, academic achievement,
Mulhern, and Wang (1991) compared sibling con- motor speed, and parent ratings of behavioral ad-
trols to 43 children with SCD ages 8 to 16 years with justment. Children with a clinical stroke as revealed
no overt neurological dysfunction. They found on MRI showed the highest frequency of neurocog-
that the children with SCD still scored lower on nitive deficits, including impairments in intellec-
measures of intelligence, although no differences tual functioning, language abilities, visual-motor
appeared in academic achievement. In addi- and visual-spatial processing, academic achieve-
tion, children 12 years of age or younger showed ment, and sequential memory. Children with silent
language-based deficits on the Luria-Nebraska Neu- infarcts did not show problems as severe or perva-
ropsychological Battery (Golden, 1987). Brown and sive as those who demonstrated overt CVAs, al-
associates (Brown, Buchanan, et al., 1993) used a though they did perform more poorly than those
battery of neurocognitive measures to compare with no MRI abnormality on visual-motor speed,
children and adolescents with SCD but without arithmetic, and vocabulary.
CVAs or overt neurological dysfunction to nondis- We wanted to extend the research of Armstrong
eased siblings. Children with SCD performed more and associates (1996) by examining the neurocogni-
poorly than siblings on measures of sustained atten- tive functioning of children with SCD and clinical
tion. These findings are consistent with those of stroke and children with SCD and silent infarcts,
Rodgers, Clark, and Kessler (1984), who found evi- with both conditions documented on MRI scans.
dence of altered metabolism in the frontal lobe area We extended Armstrong’s original investigation by
on positron emission tomography (PET). These using additional indices of neurocognitive func-
studies indicate the likelihood of deficits in atten- tioning, including measures of attention and execu-
tion and executive functioning in patients with tive functioning and specific neuropsychological
SCD, including those with overt CVAs (DeBaun et measures of language. Also included were parent re-
al., 1998; Rodgers et al., 1984; Schatz et al., 1999) ports of behavioral adjustment and adaptive com-
and those with no documented evidence of infarcts petence. Our study was unique because it included
(Brown, Buchanan, et al., 1993; Goonan, Brown, a battery of measures associated with attention,
Buchanan, & Eckman, 1994). concentration, and executive functioning and spe-
The possibility that a significant percentage of cifically related these measures to MRI findings. In
children with SCD (11% to 20%) with no evidence addition, consistent with the Armstrong design,
of overt neurological disease on physical examina- children and adolescents in the two MRI categories
tion may have cerebral infarction detectable only (i.e., overt CVA, silent infarct) were compared to
by magnetic resonance imaging (MRI) has also been those participants with no documented MRI abnor-
investigated (Kinney et al., 1999; Pavlakis et al., mality. We hypothesized that, relative to their
1988). The most current definitive investigation of counterparts with no documented MRI abnormal-
silent cerebral infarction with neuropsychological ity, children with overt clinical strokes and those
data has been a subset of the Cooperative Study of with silent infarcts would perform less well on
Sickle Cell Disease (CCSCD) (Armstrong et al., global measures of intellectual functioning and aca-
1996). Participants in this large, multisite, collabo- demic achievement and on tasks of attention and
rative clinical research program included 194 chil- executive functioning, language, and visual-motor
dren with SCD, 135 of whom had the HbSS type. and visual-spatial processing.
Magnetic Resonance Imaging in Children 505

Method behavioral ratings of internalizing and externaliz-

ing behavioral adjustment and adaptive behavior.
Participants The measures comprising the battery are described
below.
Participants were 63 children and adolescents with The Wechsler Intelligence Scale for Children-III
SCD receiving their treatment at a comprehensive (WISC-III) (Wechsler, 1991) allows for the assess-
National Institutes of Health sickle cell center. This ment of verbal (VIQ), performance (PIQ), and full-
center is located at a major university-affiliated scale intellectual functioning (FSIQ).
teaching hospital that serves primarily individuals The Woodcock-Johnson Psychoeducational Test Bat-
of lower socioeconomic status (SES). Thus, our tery: Tests of Achievement-Revised (WJ-R) (Wood-
sample is composed primarily of a fairly low SES cock & Johnson, 1990) was used to assess academic
group, and this is characteristic of other investiga- achievement in broad mathematics (calculation
tions of children and adolescents with sickle cell and applied problems subtests) and broad reading
syndromes (Hurtig, Koepke, & Park, 1989; Ievers, (letter-word identification and passage comprehen-
Brown, Lambert, Hsu, & Eckman, 1998). Approxi- sion subtests). Broad Reading and Broad Mathemat-
mately 500 children with sickle cell syndromes were ics scores were used in the analyses.
followed at this center at the time that the data were Participants were administered the Cancellation
collected, from 1993 to 1998. Participants had been A’s Task (Diller et al., 1974) that assesses capacity
referred to us by pediatric hematologists to investi- for sustained attention on a repetitive response task.
gate their cognitive, academic, and emotional func- Performance is scored for errors of omission and
tioning (Brown, Armstrong, & Eckman, 1993; commission and time to complete. In addition,
Brown, Doepke, & Kaslow, 1993; Devine, Brown, children with SCD were administered the Trail Mak-
Lambert, Donegan, & Eckman, 1999; Ievers et al., ing Test (Reitan & Wolfson, 1985) that assesses speed
1998). The children were specifically referred to the for attention, sequencing, mental flexibility, and
investigators because of possible learning problems visual search. Scoring is expressed in terms of the
and adjustment difficulties. time in seconds required for Part A and Part B of the
This investigation was approved by the Institu- test. Finally, the Freedom-from-Distractibility factor of
tional Review Board. Informed consent was obtained the WISC-III was included in this domain because
from the participants’ caregivers. For those children it is associated with sequential processing, atten-
and adolescents capable of understanding the na- tion, and automatic functioning.
ture of the study, assent was obtained. Demographic The Boston Naming Test (Kaplan & Goodglass,
information was obtained through questionnaires, 1983) assessed expressive language ability, measured
and medical information was obtained by chart re- by the ability to name pictured objects. A total nam-
views as well as an examination of the computer ing score was employed for analyses. Additionally,
database employed for the NIH sickle cell center. the Rapid Automatized Naming (RAN) (Denckla & Ru-
Caregivers were paid $30 for participation in the del, 1974) was administered. Rapid naming tasks
study. MRI examinations of the brain were con- can differentiate between individuals with a reading
ducted for clinical purposes within 3 months of the disability, those with other learning disabilities, and
neurocognitive assessment. The length of the neu- those without reading problems (Blachman, 1997;
rocognitive evaluation was approximately 3 hours. Denckla & Rudel, 1974). The total number of errors
was used for statistical analyses.
Measures The Purdue Pegboard (Tiffin, 1987) requires chil-
dren to place small round pegs into holes in a stan-
Neurocognitive Battery. Based on literature on neuro- dard board within a 30-second interval and assesses
psychological functioning in children with SCD motor speed and coordination. The raw score is de-
(Bonner et al., 1999; Brown, Armstrong, & Eckman, rived from the total number of pegs that are cor-
1993; Brown, Doepke, & Kaslow, 1993; White & rectly placed with each hand.
DeBaun, 1998), a neurocognitive battery was con- The Child Behavior Checklist (CBCL) (Achen-
structed to assess six domains of functioning: gen- bach & Edelbrock, 1991) is a 113-item measure
eral intellectual functioning, academic achieve- completed by caregivers that addresses a broad
ment, attention and executive functioning, range of internalizing (e.g., depression, anxiety) and
language, visual-spatial and motor processing, and externalizing (e.g., acting-out behaviors) behaviors
506 Brown et al.

in children and adolescents. Caregivers rated the se- VAs), and post hoc analyses were performed on any
verity of each of the symptoms that they observed dependent measure for which there was a signifi-
in their child. In addition, we administered the cant ANOVA.
Vineland Adaptive Behavior Scales (Sparrow, Balla, &
Cicchetti, 1984), a widely used, semistructured pa-
rental interview that assesses adaptive behavior Results
across several domains of functioning, including
communication, daily living, socialization, and mo- Of the participants, 60.3% were male (n ⫽ 38) and
tor skills. The Adaptive Behavior composite score 39.7% were female (n ⫽ 25), and the mean age was
was used in our analysis. 9.75 years (SD ⫽ 2.87, range ⫽ 6.33 to 17.00 years).
Magnetic Resonance Imaging Studies. MRI studies Grades in school ranged from kindergarten to tenth
of the brain without contrast were performed in ac- with a mean of third grade (M ⫽ 3.56, SD ⫽ 2.75).
cordance with standard practices. Children who Children were primarily in regular classrooms
were younger and those who needed assistance (63.5%; n ⫽ 40), although 34.9% (n ⫽ 22) were re-
were sedated in accordance with the guidelines out- ceiving special education resource services (i.e.,
lined by the American Academy of Pediatrics (Com- part-time resource help in certain subjects). One
mittee on Drugs, 1992). Our procedures were child was in a self-contained special education class-
similar to those of the CCSCD investigation (Arm- room (1.3%). The children and adolescents had the
strong et al., 1996) and other studies (DeBaun et al., following types of SCD: HbSS (n ⫽ 48; 76.2%), HbSC
1998; Schatz et al., 1999). Each child’s MRI was (the heterozygous condition for hemoglobin S and
reviewed independently by both a pediatric neuro- hemoglobin C) (n ⫽ 15; 23.8%). Children and
radiologist and a pediatric radiologist. Any disagree- adolescents in the two SCD groups (i.e., HbSS,
ments between the two radiologists were resolved HbSC) were compared on hemoglobin, symptoms
by consensus. MRI readings were classified without ever, and symptoms in the past year. The results of
CNS pathology as normal (no CNS pathology), cere- t tests revealed that participants with HbSS disease
bral infarction, atrophy, and cerebral infarction and (Mhemoglobin ⫽ 8.22, SD ⫽ 1.14) had lower hemo-
atrophy. Children with infarction were classified as globin than did their peers with HbSC disease
having had either clinically apparent CVAs or silent (Mhemoglobin ⫽ 11.07, SD ⫽ 1.19) (t ⫽ 12.9, p ⬍ .001).
infarction. If the records showed no history of a In addition, the results of t tests revealed that parti-
CVA and the neurologic examination was unre- cipants with HbSS disease (Msymptoms ever ⫽ 3.79, SD ⫽
markable, the children with MRI scans suggestive of 2.24) had a greater number of symptoms ever than
infarction were classified as having a silent infarct did their peers with HbSC disease (Msymptoms ever ⫽
(n ⫽ 11). Children with a documented history of a 1.90, SD ⫽ 1.48) (t ⫽ 5.59, p ⬍ .001). Finally, the
CVA that was recorded in the medical record and results of t tests indicated that children and adoles-
who also had an MRI indicating an infarct were cents with HbSS disease (M symptoms past year ⫽ 2.43,
classified as having a clinical history of a CVA (n ⫽ SD ⫽ 1.89) had a greater number of symptoms in
22). If there was no pathology shown in the medical the past year than did their counterparts with HbSC
record or the MRI, the child was classified as normal disease (M symptoms past year ⫽ 1.05, SD ⫽ 1.34) (t ⫽ 4.78,
or without CNS pathology (n ⫽ 30). p ⬍ .001). Thus, participants with HbSS disease
Data Analyses. A series of separate one-way mul- clearly had greater disease severity than did their
tivariate analyses of variance (MANOVA) was per- peers with HbSC disease.
formed for each domain of measures (i.e., measures Most caregivers in our sample were mothers
of intellectual functioning, academic achievement, (87.3%, n ⫽ 55). Other caregivers were fathers
attention and executive functioning, language, (7.9%, n ⫽ 5), grandmothers (1.6%, n ⫽ 1), and
visual-spatial and motor processing, behavioral rat- aunts (3.2%, n ⫽ 2). The marital status of caregivers
ings of adjustment), and the classification of CNS was 23.8% (n ⫽ 15) currently married, 39.7% (n ⫽
pathology shown on physical examination and MRI 25) single, 34.9% (n ⫽ 22) separated or divorced,
reading (i.e., clinical stroke or CVA, “silent” stroke, and 1.6% (n ⫽ 1) widowed. Although most caregiv-
and scan without evidence of CNS pathology) ers had graduated from high school (n ⫽ 49; 77.8%),
served as the independent variable. Thus, six MA- approximately one fourth had not completed high
NOVAs were performed; significant multivariate school and had not been granted an equivalent di-
tests were followed by analyses of variance (ANO- ploma through testing (n ⫽ 14; 22.2%). The mean
Magnetic Resonance Imaging in Children 507

Table I. Demographic Characteristics and Disease Severity

Group membership
No CVA Pathology CVA Silent Entire Sample
(n ⫽ 30) (n ⫽ 22) (n ⫽ 11) (N ⫽ 63)
M (SD) n (%) M (SD) n (%) M (SD) n (%) M (SD) n (%) F/␹ 2

Demographic
characteristics
Age (in months) 117.68 (34.59) 118.04 (37.87) 117.17 (29.59) 117.71 (34.44) ⬍1
Gender
Female 12 (40) 4 (18) 9 (82) 25 (40) ⬍1
Special education 10 (33) 8 (36) 5 (46) 23 (37) 2.95
Caregiver marital status
Married 9 (30) 2 (9) 4 (36) 15 (24) 6.89
Single 15 (50) 2 (9) 8 (73) 25 (40)
Separated/divorced 16 (53) 4 (18) 2 (18) 22 (35)
Widowed 0 (0) 1 (5) 0 (0) 1 (2)
Caregiver degree
High school 24 (80) 18 (82) 7 (64) 49 (78) 8.14
Technical or associate 6 (20) 1 (5) 3 (27) 10 (16)
degree
Family income
⬍ $10,000 17 (57) 13 (59) 5 (46) 35 (56) 2.03
$10,000–$19,999 9 (30) 3 (14) 1 (9) 13 (21)
$20,000–$29,999 7 (23) 2 (9) 5 (46) 12 (19)
$30,000–$40,000 1 (3) 1 (5) 0 (0) 2 (3)
⬎ $40,000 1 (3) 0 (0) 0 (0) 1 (2)
Disease severity
Symptoms ever 2.88 (2.43) 4.57 (1.73) 2.33 (2.06) 3.35 (2.31) 5.77****
Symptoms in past year 2.03 (1.90) 2.87 (1.82) 1.67 (1.37) 2.25 (1.82) 2.27
Hemoglobin 9.23 (1.85) 8.22 (.64) 8.08 (1.18) 8.72 (1.52) 4.24***
Sickle cell type
HbSS 15 (50) 22 (100) 11 (100) 48 (76) 12.40
HbSC 15 (50) 0 (0) 0 (0) 15 (24)

***p ⬍ .02.
****p ⬍ .005.

years of education was 12.56 (SD ⫽ 1.61, range ⫽ 9 avascular necrosis, osteomyelitis, dactylitis, pria-
to 16 years). Most of the families had annual in- pism, and delayed puberty. Symptoms were re-
comes of less than $10,000 (n ⫽ 35; 55.6%). The corded if they had occurred within the past year
remainder of the sample had incomes ranging (number of symptoms within the past year) or if
$10,000–$19,000 (n ⫽ 13, 20.6%), $20,000–$30,000 they had been present prior to this period of time
(n ⫽ 12; 19.1%), $31,000–$40,000 (n ⫽ 2; 3.2%), (symptoms ever). No differences were found for any
and above $40,000 (n ⫽ 1, 1.6%). of the demographic variables. As expected, how-
Table I presents the subject age and gender, care- ever, significant differences were found for symp-
givers’ marital status, education, income, and sever- toms ever, F(2, 60) ⫽ 5.77, p ⬍ .005, and the mean
ity of disease (i.e., number of symptoms within the hemoglobin within the past three clinic visits, F(2,
past year, number of symptoms ever, mean hemo- 60) ⫽ 4.24, p ⬍ .02. All post hoc comparisons for
globin of the past three visits) for each of three des- this study were performed using Tukey’s HSD proce-
ignated MRI groups. Symptoms included seizures, dure. The results of post hoc tests performed on the
headaches, auditory and ocular complications, means revealed that for symptoms ever, the group
splenic sequestration, microsplenia, splenomegaly, classified without CNS pathology differed from the
hepatomegaly, cardiomegaly, acute chest syndrome, group with overt strokes. In addition, the group
lung abnormalities, renal complications, gallstones, classified without CNS pathology differed from the
508 Brown et al.

group classified with silent infarcts ( p ⬍ .01). An vealed. None of the remaining MANOVAs was sta-
examination of the means showed that the group tistically significant.
with overt CVAs evidenced a greater frequency of Finally, the children who were diagnosed with
symptoms ever than did either the group classified either overt CVAs or silent infarcts were classified
without CNS pathology or the silent infarct group. according to the area of the brain where they had
An examination of the mean hemoglobin scores sustained damage. Areas included the frontal lobe
revealed that the group classified without CNS (n ⫽ 31), the temporal lobe (n ⫽ 17), the parietal
pathology had significantly higher mean hemo- lobe (n ⫽ 25), the occipital lobe (n ⫽ 11), the thala-
globins ( p ⬍ .03) than the group with overt CVAs. mus (n ⫽ 4), putamen (n ⫽ 10), globus pallidus (n ⫽
Thus, as expected because of the greater severity of 9), caudate nucleus (n ⫽ 14), and internal capsule
their disease, the children with overt CVAs had a (n ⫽ 10). For the entire sample, the highest fre-
higher frequency of symptoms ever and lower he- quency of damage was in the frontal lobe. In fact,
moglobin scores than children classified without 93.9% (n ⫽ 31) of the CVA and silent group had
CNS pathology. The children with silent infarcts sustained some type of frontal lobe injury. A series
had a higher frequency of symptoms ever than the of chi-square analyses was performed to determine
group classified without CNS pathology. whether there were differences among the three
The results of a one-way MANOVA on the MRI groups (without CNS pathology, silent infarcts,
attention-executive functioning factor yielded a sta- and overt CVAs). The results of the frequencies be-
tistically significant main effect, F(12, 64) ⫽ 2.13, tween the three groups revealed differences be-
p ⬍ .03. Separate ANOVAs were performed on each tween all of the areas, with the exception of the
of the dependent measures and a statistically sig- thalamus. The highest frequency of impairments
nificant effect was revealed for the Trail Making B was found for children in the overt CVA group fol-
time score, F(2, 60) ⫽ 6.62, p ⬍ .003. To determine lowed by the silent infarct group. The frequencies
the source of statistical significance between the and chi-square analyses of the various areas of im-
three groups, post hoc analyses were performed on pairment are presented in Table III. Finally, to deter-
the Trail Making B time score. The results of these mine whether the children who evidenced frontal
Tukey HSD post hoc tests indicated that the group lobe impairment as documented by MRI performed
designated without CNS pathology completed the more poorly than the children who did not evi-
test more rapidly than the group classified with si- dence frontal lobe impairments on the measures of
lent infarcts ( p ⬍ .007). The children with silent in- attention and concentration that previously had
farcts took more time to complete the task than been found to be significant, a series of t tests was
those with overt CVAs ( p ⬍ .03). See Table II for performed on the Trail Making B time score, the
these means. Cancellation A’s Omission error score, and the Can-
The results of the ANOVA performed for the cellation A’s Commission score. The results of these
Cancellation A’s Omission error score revealed a sig- analyses revealed that the children with docu-
nificant main effect, F(2, 60) ⫽ 3.72, p ⬍ .03. Post mented frontal lobe impairment performed more
hoc analyses revealed that the children classified poorly on both the Cancellation A’s Omission,
without CNS pathology differed from the children (t[48] ⫽ ⫺2.41, p ⬍ .02; MFrontal Lobe Impairment ⫽ 5.84,
with overt CVAs on the omission error score ( p ⬍ SD ⫽ 7.93, MNonfrontal Lobe Impairment ⫽ 1.72, SD ⫽ 3.23)
.01), and there was a difference that approached sta- and the Commission error scores (t[48] ⫽ ⫺2.12,
tistical significance for the children with overt CVAs p ⬍ .04; MFrontal Lobe Impairment ⫽ 5.28, SD ⫽ 6.95,
to differ from the participants classified with silent MNonfrontal Lobe Impairment ⫽ 2.08, SD ⫽ 2.98). No signifi-
infarcts ( p ⬍ .06). An examination of the means in- cant between group difference was found for the
dicates that the children classified without CNS pa- Trail Making B time score.
thology had fewer omission errors than the group
with overt CVAs. Finally, a difference that ap-
proached statistical significance was revealed for the Discussion
ANOVA performed on the Cancellation A’s Com-
mission Error score, F(2, 60) ⫽ 2.48, p ⬍ .09. No Our study examined the neurocognitive function-
other statistically significant ANOVAs for measures ing of children with SCD who had either CVAs or
of the attention-executive function domain were re- silent strokes. In addition, these two cohorts of chil-
Magnetic Resonance Imaging in Children 509

Table II. Means and Standard Deviations for Dependent Measures

Group membership
No CVA Pathology CVA Silent
(n ⫽ 30) (n ⫽ 22) (n ⫽ 11)
Measures M (SD) M (SD) M (SD) ANOVA F Post hoc

Intellectual functioning
WISC-III
Verbal IQ a 85.60 (18.65) 79.59 (15.65) 88.73 (15.67) 1.27 —
Performance IQ a 80.80 (13.80) 74.73 (16.49) 77.73 (14.25) 1.06 —
Full Scale IQ a 81.67 (16.68) 75.05 (15.53) 81.91 (14.43) 1.26 —
Academic achievement
Woodcock Johnson—Revised
Broad Reading a 83.93 (25.89) 79.11 (24.92) 84.89 (33.41) ⬍1 —
Broad Math a 87.15 (23.10) 76.89 (26.31) 87.44 (24.40) 1.08 —
Attention and executive
functioning
Cancellation As
Omission errors 1.56 (2.23) 5.30 (7.88) 1.57 (1.13) 3.72** 2⬎1; 2⬎3
Commission errors 1.17 (1.43) 6.20 (9.10) 1.14 (.69) 2.48* —
Time to completion (in sec.) 201.11 (96.62) 223.33 (110.40) 290.00 (183.96) 1.38 —
Trail Making Test
Number of errors 2.39 (4.45) 2.27 (2.92) 2.71 (3.90) ⬍1 —
Time to completion (in sec) 57.17 (27.22) 68.80 (33.94) 113.71 (52.87) 6.62**** 3⬎1; 3⬎2
FFD 91.50 (15.57) 81.47 (15.23) 87.43 (10.50) 1.90 —
Language
Boston Naming Test (total score) 27.27 (8.54) 24.50 (10.41) 25.14 (6.34) ⬍1 —
Rapid Automatized Naming Test .64 (.73) 1.63 (1.64) .86 (1.46) 3.25** 2⬎1
Visual-spatial and motor processing
Purdue
Right 10.80 (3.28) 10.00 (2.94) 9.57 (2.94) ⬍1 —
Left 10.04 (2.57) 8.41 (3.45) 9.86 (3.58) 1.52 —
Behavioral adjustment
Child Behavior Checklist
Internalizing b 57.39 (11.39) 50.81 (13.89) 60.00 (7.13) 2.58* —
Externalizing b 53.94 (12.90) 46.29 (8.70) 49.25 (12.90) 3.23** 1⬎2
Vineland Adaptive Behavior Scale 82.68 (22.79) 89.67 (20.42) 81.25 (22.46) ⬍1 —
a
Standard scores.
b
T scores.
*p ⬍ .10.
**p ⬍ .05.
****p ⬍ .005.

dren were compared to children with SCD with no frontal lobe involvement. In the area of attention,
evidence of CNS pathology as revealed on MRI. We children who showed silent strokes on the MRI also
incorporated specific assessments of neurocognitive showed similar impairments as their peers who sus-
functioning, including measures of attention and tained overt strokes. Finally, MRI scans revealed a
executive functioning, language, and adaptive com- high frequency of frontal lobe deficits in children
petence. We also measured more global indices of who sustained CVAs.
cognitive functioning such as general intelligence These data extend the findings of Armstrong
and academic achievement. Children with docu- and associates (1996), who found evidence of gen-
mented clinical strokes performed more poorly eral deficits in intellectual functioning and aca-
than their peers on tasks requiring sustained atten- demic performance in children with SCD who had
tion and effort or on tasks that were associated with CVAs and silent infarcts. Our investigation suggests
510 Brown et al.

Table III. Frequencies of Damage as Revealed on MRI children with SCD who are suspected of having
Group membership CNS involvement. As Schatz et al. (1999) observed,
Silent
in case studies of individuals with frontal lobe injur-
CVA (n ⫽ 22) (n ⫽ 11) ies, global tests of intellectual and academic func-
Area of damage n (%) n (%) ␹2 tioning do not adequately identify frontal lobe
deficits. Although such assessments are important,
Frontal lobe 21 (95.5) 10 (90.9) 54.9
they may in fact underestimate the extent of cogni-
Temporal lobe 11 (50.0) 6 (54.6) 21.9
Parietal lobe 17 (77.3) 8 (72.7) 38.3
tive impairments for children with documented
Occipital lobe 9 (40.9) 2 (18.2) 15.7 frontal lobe involvement. Given that alternatives to
Thalamus 3 (13.6) 1 (9.1) 4.5 neuropsychological assessments require sedation
Putamen 9 (40.9) 1 (9.1) 17.4 for younger children, are quite costly, and are in-
Globus pallidus 8 (36.4) 1 (9.1) 14.9 convenient for families, the identification of neuro-
Caudate nucleus 13 (59.1) 1 (9.1) 28.4 logical impairments through cognitive screening
Internal capsule 8 (36.4) 2 (18.2) 13.5 appears to be a practical means of selecting children
There were no areas of damage in any of the 30 members of the who are candidates for more extensive medical eval-
normal group; 93.9% (n ⫽ 31/33) of all CVA and Silent groups uation.
had frontal lobe involvement.
In contrast to the CSSCD findings (Armstrong et
al., 1996) and other studies (for review see, Brown,
that children with SCD and either clinical strokes Armstrong, & Eckman, 1993; Brown, Doepke, &
or silent infarcts show attention problems, as as- Kaslow, 1993), our investigation did not provide
sessed by traditional neuropsychological measures. any evidence for statistically significant global
In fact, our findings are in accord with those of sev- deficits in mental abilities, academic achievement,
eral investigators (Brown, Buchanan, et al., 1993; or visual-motor functioning, although the children
Cohen et al., 1997; DeBaun et al., 1998; Schatz et in the CVA group performed more poorly than the
al., 1999), who have provided strong evidence that children in the other groups. The absence of evi-
cognitive measures of attention and concentration dence is surprising, and we offer three possible ex-
are the most useful indices to identify children with planations. First, the relatively low power offered by
possible cerebral infarcts. This is helpful in identi- this type of clinical study could have muted effects
fying children who should be serious candidates for that may have been significant in a larger sample.
further radiologic evaluation. In support of this no- Second, significant variability across most of the
tion, DeBaun et al. (1998) concluded that screening measures in each of the domain areas and error vari-
measures of attention and executive functioning ance may have diminished the possibility of sig-
are valuable in the identification of children and ad- nificant effects that also may have occurred with a
olescents with SCD in whom CNS involvement is larger sample. Third, the more compelling explana-
suspected. tion is related to the fact that all of the patients were
Of interest is the finding that the highest fre- specifically referred to this study because of long-
quency of functional impairments on the MRI re- standing histories of academic difficulties, delays in
vealed overt infarcts in the frontal lobe. These data cognitive development, and evidence of a clinical
are consistent with those of Rodgers et al. (1984), stroke; nearly 40% of the sample was receiving spe-
who found evidence of altered metabolism in the cial education services, including those children
frontal lobe area of adult patients with SCD as indi- designated without CNS pathology. In fact, one
cated on PET scans. Our findings are coupled with third of the group designated with no MRI pathol-
other findings of impairments in the area of atten- ogy was receiving special education services. Thus,
tion and concentration, neurocognitive processes the relatively low scores for the group designated
that traditionally have been associated with frontal without CNS pathology may have produced “floor”
lobe functioning (Boliek & Obrzut, 1997; Shue & effects that diminished the possibility of signifi-
Douglas, 1992). The data in our investigation and cance, especially on global measures of functioning
other studies (Brown, Buchanan, et al., 1993; Cohen such as intelligence and academic achievement. For
et al., 1997; DeBaun et al., 1998; Schatz et al., 1999) example, the mean FSIQ for the CVA and silent
underscore the importance of the assessment and stroke group in our study was similar to the means
identification of problems associated with atten- reported by Armstrong et al. (1996), but the mean
tion, concentration, and executive functioning in FSIQ for the children without CNS pathology in our
Magnetic Resonance Imaging in Children 511

study was 81.7 (SD ⫽ 16.7). The mean FSIQ for the controls will bring the problems of these children
children without CNS pathology in the Armstrong into sharper focus. Finally, the children designated
investigation was 90, making the FSIQ for the group without CNS pathology on the MRI exams included
without CNS pathology in our investigation nearly children who also had HbSC disease who were not
10 standard score points lower than the mean for represented in the other groups (i.e., CVA group and
the group without CNS pathology in the CSSCD in- silent infarct group). Because no differences were
vestigation. Thus, the group without CNS pathol- found on any of the neurocognitive measures be-
ogy in our investigation was more similar to the tween the children having HbSS and HbSC disease
groups designated with CNS pathology and Arm- within the group designated without CNS pathol-
strong’s clinical groups. Similar trends are noted for ogy (i.e., normal group), the presence of children
the reading and mathematics achievement test with HbSC and HbSS disease in this group is not
scores, where our mean scores for the group without likely confounded.
CNS pathology were nearly two-thirds of a standard Despite these limitations, our results underscore
deviation lower than that reported by Armstrong the importance of using tests that measure atten-
et al. This may be due to the fact that the children tion and concentration in the assessment of chil-
without CNS pathology in our sample were referred dren with SCD, particularly for children who have
because of their long history of academic problems, sustained overt strokes or silent infarcts. Future re-
delays in cognitive development, as well as clinical search will need to identify the mechanisms under-
evidence of a CVA. Clearly, additional research is lying frontal lobe deficits for children who have
needed in this area, particularly clinical studies with sustained infarcts. Recently, Schatz et al. (1999) pro-
larger sample sizes. vided important preliminary data that both the spe-
The contribution of our investigation must be cific region of the brain is affected and the total
interpreted within the limitations of the design. volume of cerebral injury influence cognitive func-
Again, the small sample size may have limited tioning. Examining both the location and volume
power and diminished effects that may have oc- of cerebral infarction is a good direction for further
curred with a larger sample. In addition, significant investigation. Finally, the data in our study and the
variability within the three groups increased error findings of other investigations (Brown, Buchanan,
variance and may also have mitigated significant ef- et al., 1993; Schatz et al., 1999) provide some pre-
fects. In addition, the children participating in this liminary foundation for the management and reha-
investigation were part of a clinical sample, all of bilitation of these children, particularly in the
whom who were referred for learning problems, a educational environment. Given the pervasiveness
chronic history of school difficulties, and develop- of attention problems in children with SCD, psy-
mental delays in cognitive functioning. Thus, the chological and medical interventions designed to
group designated as “normal” may have been quite enhance attention and concentration represent a
low functioning, possibly characterized by other promising area for future research. Hopefully, such
CNS pathology not detected on traditional clinical interventions will enhance the quality of life for
MRI exams. Additional research will need to include these children as they struggle with the associated
multisite, collaborative clinical samples and larger cognitive morbidities presented by this lifelong
cohorts of children with SCD employing neuropsy- disease.
chological measures across a wide domain of func-
tioning. Assessment across a wide range of domains
can help determine more specifically the essential Acknowledgments
deficits of children who have sustained CVAs, silent
infarcts, and those who evidence no CNS pathol- This research was supported in full by a grant
ogy. In addition, any association among the reason awarded from the National Institutes of Health,
for referral, specific location of pathology as re- Heart, Lung, and Blood Branch, HLB P60-HL48-482.
vealed on the MRI, and psychological test data can- We thank Drs. Baldwin, Casey, Doepke, Ievers, and
not be addressed in this particular investigation. Tanaka, and Ms. Donegan for their help in data col-
Future research will need to examine decision- lection and Jeffrey Schatz, PhD, for assisting with
making algorithms in identifying those children an earlier draft of the manuscript.
likely to evidence specific CNS pathology on MRI
scans. In addition, using community comparison Received May 14, 1999; accepted August 14, 1999
512 Brown et al.

References

Achenbach, T. M., & Edelbrock, C. S. (1991). Manual for the Diller, L., Ben-Yishay, Y., Gerstman, L. J., Goodkin, R., Gor-
Child Behavior Checklist and Child Behavior Profile. Bur- don, W., & Weinberg, J. (1974). Studies in cognition and
lington: University of Vermont, Department of Psy- rehabilitation in hemiplegia (Rehabilitation Monograph
chiatry. No. 50). New York: New York University Medical Cen-
Armstrong, F. D., Thompson, R. J., Wang, W., Zimmer- ter Institute of Rehabilitation Medicine.
man, R., Pegelow, C., Miller, S., Moser, F., Bello, J., Hur- Fowler, M. G., Whitt, J. K., Nash, K. B., Atkinson, S., Wells,
tig, A., & Vass, K. (1996). Cognitive functioning and R. J., & McMillan, C. (1988). Neuropsychologic and
brain magnetic resonance imaging in children with academic functioning of children with sickle cell ane-
sickle cell disease. Pediatrics, 97, 864–870. mia. Developmental and Behavioral Pediatrics, 9,
Blachman, B. A. (Ed.). (1997). Foundations of reading acqui- 213–220.
sitions and dyslexia: Implications for early intervention. Golden, C. J. (1987). Luria-Nebraska Neuropsychological Bat-
Mahwah, NJ: Lawrence Erlbaum. tery: Children’s revision manual. Los Angeles: Western
Boliek, C. A., & Obrzut, J. E. (1997). Neuropsychological Psychological Services.
aspects of attention deficit/hyperactivity disorder. In Goonan, B. T., Goonan, L. J., Brown, R. T., Buchanan, I., &
C. R. Reynolds & E. Fletcher-Janzen (Eds.), Handbook Eckman, J. R. (1994). Sustained attention and inhibi-
of clinical child neuropsychology (2nd ed., pp. 619–633). tory control in children with sickle cell syndrome. Ar-
New York: Plenum. chives of Clinical Neuropsychology, 9, 89–104.
Bonner, M. J., Gustafson, K. E., Schumacher, E., & Thomp- Harriman, L. M. F., Griffith, E. R., Hurtig, A. L., & Keehn,
son, R. J., Jr. (1999). The impact of sickle cell disease M. T. (1991). Functional outcomes of children with
on cognitive functioning and learning. School Psychol- sickle cell disease affected by stroke. Archives of Physi-
ogy Review, 28, 182–193. cal Medicine and Rehabilitation, 72, 498–502.
Brown, R. T., Armstrong, F. D., & Eckman, J. R. (1993). Hurtig, A. L., Koepke, D., & Park, K. B. (1989). Relation be-
Neurocognitive aspects of pediatric sickle cell disease. tween severity of chronic illness and adjustment in
Journal of Learning Disabilities, 26, 33–45. children and adolescents with sickle cell disease. Jour-
Brown, R. T., Buchanan, I., Doepke, K., Eckman, J. R., nal of Pediatric Psychology, 14, 117–132.
Baldwin, K., Goonan, B., & Schoenherr, S. (1993). Ievers, C. E., Brown, R. T., Lambert, R. G., Hsu, L., & Eck-
Cognitive and academic functioning in children with man, J. R. (1998). Family functioning and social sup-
sickle cell disease. Journal of Clinical Child Psychology, port in the adaptation of caregivers of children with
22, 207–218. sickle cell syndromes. Journal of Pediatric Psychology,
Brown, R. T., Doepke, K. J., & Kaslow, N.J. (1993). Risk- 23, 377–388.
resistance-adaptation model for pediatric chronic ill- Kaplan, E. F., & Goodglass, H. (1983). The Boston Naming
ness: Sickle cell syndrome as an example. Clinical Psy- Test. 2nd ed. Philadelphia: Lea & Febiger.
chology Review, 13, 119–132. Kinney, T. R., Sleeper, L. A., Wang, W. C., Zimmerman,
Cohen, M. J., Branch, W. B., McKie, V. C., & Adams, R. J. R. A., Pegelow, C. H., Ohene-Frempong, K., Wethers,
(1997). Neuropsychological impairment in children D. L., Bello, J. A., Vichinsky, E. P., Moser, F. G., Gal-
with sickle cell anemia and cerebrovascular accidents. lagher, D. M., DeBaun, M. R., Platt, O. S., & Miller,
Clinical Pediatrics, 33, 517–524. S. T. (1999). Silent cerebral infarcts in sickle cell ane-
Committee on Drugs. (1992). Guidelines for monitoring mia: A risk factor analysis. Pediatrics, 103, 640–645.
and management of pediatric patients during and Pavlakis, S. G., Bello, J., Prohovnik, I., Sutton, M., Ince, C.,
after sedation for diagnostic and therapeutic proce- Mohr, J. P., Piomelli, S., Hilal, S., & DeVivo, D.C.
dures. Pediatrics, 93, 1110–1115. (1988). Brain infarction in sickle cell anemia: Mag-
DeBaun, M. R., Schatz, J., Siegel, M. J., Koby, M., Craft, S., netic resonance imaging correlates. Annals in Neurol-
Resar, L., Chu, J. Y., Launius, G., Dadash-Zadeh, M., ogy, 23, 125–130.
Lee, R. B., & Noetzel, M. (1998). Cognitive screening Reitan, R. M., & Wolfson, D. (1985). The Halstead-Reitan
examinations for silent cerebral infarcts in sickle cell Neuropsychological Test Battery. Tucson, AZ: Neuropsy-
disease. Neurology, 50, 1678–1682. chology Press.
Denckla, M. B., & Rudel, R. G. (1974). Rapid automatized Rodgers, G. P., Clark, C. M., & Kessler, R. M. (1984). Re-
naming of pictured objects, colors, letters, and num- gional alterations in brain metabolism in neurologi-
bers by normal children. Cortex, 10, 186–202. cally normal sickle cell patients. Journal of the American
Devine, D., Brown, R. T., Lambert, R., Donegan, J., & Eck- Medical Association, 256, 1692–1700.
man, J. R. (1999). Predictors of psychosocial and Schatz, J., Craft, S., Kobny, M., Siegel, M. J., Resar, L., Lee,
cognitive adaptation in children with sickle cell syn- R., Chu, Y. F., Launius, G., Dadash-Zadeh, M., & De-
dromes. Journal of Clinical Psychology in Medical Set- Baun, M. R. (1999). Neuropsychologic deficits in chil-
tings, 5, 295–313. dren with sickle cell disease and cerebral infarction:
Magnetic Resonance Imaging in Children 513

The role of lesion location and volume. Child Neuro- Wasserman, A. L., Wilmas, J. A., Fairclough, D. L., Mul-
psychology, 5, 92–103. hern, R. K., & Wang, W. (1991). Subtle neuropsycho-
Shue, K. L., & Douglas, V. I. (1992). Attention deficit hy- logical deficits in children with sickle cell disease.
peractivity disorder and the frontal lobe syndrome. American Journal of Pediatric Hematology and Oncology,
Brain and Cognition, 20, 104–124. 13, 14–20.
Sparrow, S. S., Balla, D. A., & Cicchetti, D. V. (1984). Vine- Wechsler, D. (1991). Wechsler Intelligence Scale for Children–
land Adaptive Behavior Scales. Circle Pines, MN: Ameri- III. San Antonio, TX: Psychological Corporation.
can Guidance Service. White, D. A., & DeBaun, M. (1998). Cognitive and behav-
Swift, A. V., Cohen, M. J., Hynd, G. W., Wisenbaker, J. M., ioral function in children with sickle cell disease: A
McKie, K. M., Makari, G., & McKie, V. (1989). Neuro- review and discussion of methodological issues. Jour-
psychological impairment in children with sickle cell nal of Pediatric Hematology/Oncology, 20, 458–462.
anemia. Pediatrics, 84, 1077–1085. Woodcock, R. W., & Johnson, M. B. (1990). Woodcock-
Tiffin, J. (1987). Purdue Pegboard examiner’s manual. Chi- Johnson Revised Tests of Achievement. Allen, TX: DLM
cago: Science Research Associates. Teaching Resources.