Slight cognitive impairment and magnetic resonance

imaging abnormalities but normal school levels in

children treated for acute lymphoblastic leukemia
with chemotherapy only
Annette Kingma, PhD, Rieneke I. van Dommelen, MSc, Eduard L. Mooyaart, MD, PhD,
Jan T. Wilmink, MD, PhD, Betto G. Deelman, PhD, and Willem A. Kamps, MD, PhD
Presymptomatic treatment of the cen-
Objectives: To investigate persistent neuropsychologic late effects in chil- tral nervous system (CNS) is essential
dren treated for acute lymphoblastic leukemia at a young age with to the improved survival rate of children
chemotherapy only by means of serial neuropsychologic assessments with acute lymphoblastic leukemia
(NPAs), magnetic resonance imaging (MRI) of the brain, and evaluation of
ALL Acute lymphoblastic leukemia
school levels.
CNS Central nervous system
Study design: Consecutive patients (n = 17) had 2 extensive NPAs (12 psy- DCLSG Dutch Childhood Leukemia
chometric measures) after cessation of therapy. Test results were compared Study Group
MTX Methotrexate
with those of both healthy control subjects and 28 previously treated children MRI Magnetic resonance imaging
who received cranial irradiation. MRI findings were related to test scores. NPA Neuropsychologic assessment
School levels were evaluated in the patients and their healthy siblings. PP Purdue Pegboard
RAVLT Rey’s Auditory Verbal Learning Test
Results: Initial participation (n = 17) and availability of the study group after VMI Visual motor integration
8 years of follow-up were 100%. Significant group differences between pa- WISC-R Wechsler Intelligence Scale for
Children-Revised
tients who received chemotherapy and healthy control subjects were found
WPPSI Wechsler Preschool and Primary
for memory and fine-motor functioning. The 17 patients combined showed 16 Scale of Intelligence
deficits on various test measures. MRI abnormalities were seen in 6 children,
but these did not correlate with cognitive performance. No differences in (ALL).1 Unfortunately, CNS prophy-
school levels were seen when the patients who received chemotherapy were laxis has been associated with cognitive
compared with their siblings. The current nonirradiated patients demonstrat- impairment, particularly when it was
ed significantly better test results and significantly fewer learning disabilities given at a young age. Numerous studies
and MRI abnormalities than did the previously irradiated group. suggested that cranial irradiation is
Conclusion: Treatment with chemotherapy only may be associated with some probably the most important single
cognitive impairment. However, these children attained normal school levels. cause of neuropsychologic deficits in
(J Pediatr 2001;139:413-20) ALL survivors.2-4 However, CNS pro-
phylaxis with intrathecal, triple in-
trathecal, or high-dose intravenous
methotrexate (MTX) without cranial ir-
From the Department of Pediatrics, Division of Pediatric Oncology/Hematology, Department of Radiology, and De-
radiation may also have adverse cogni-
partment of Neuropsychology, University Hospital Groningen, Groningen, The Netherlands; and Department of Ra-
diology, University of Maastricht, Maastricht, The Netherlands. tive sequelae. Results of studies have
Supported by grants from the Dutch Cancer Society and the Foundation of Pediatric Oncology shown deleterious effects of combina-
Research Groningen. tion therapy with MTX on neuropsy-
Submitted for publication June 27, 2000; revisions received Nov 29, 2000, and Apr 9, 2001; ac- chologic functioning,5-10 although oth-
cepted May 3, 2001.
ers have shown no important cognitive
Reprint requests: Annette Kingma, PhD, Department of Pediatrics, Division of Pediatric On-
cology/Hematology, University Hospital Groningen, PO Box 30.001, 9700 RB, Groningen, The
deficits in children treated according to
Netherlands. various protocols with only chemother-
Copyright © 2001 by Mosby, Inc. apy.11,12 Studies relating neuroimaging
0022-3476/2001/$35.00 + 0 9/21/117066 to psychometric test results have been
doi:10.1067/mpd.2001.117066 equivocal.13-16

413
 KINGMA ET AL
We longitudinally studied cognitive
functioning, brain magnetic resonance
imaging (MRI) findings, and school
achievement in 17 children after com-
pletion of treatment for ALL at a
young age with chemotherapy only.
This study was designed to be compa-
rable to a preceding study group of 28
young children with ALL who were
treated with cranial irradiation and
MTX. At a median of 10 years after di-
agnosis, 100% of this total study popu-
lation (both the patients and their
healthy siblings) was available for
school evaluation by parent-completed
questionnaires.
We hypothesized that the patients
who received chemotherapy would
have lower test scores and inferior
school achievements when compared
with the healthy control subjects and
siblings. We also hypothesized that the
nonirradiated children would have bet-
ter test results and school performance
and fewer MRI abnormalities than the
previously irradiated patients.
METHODS
Patients and Study Design
From 1981 to 1990, 64 consecutive
children with ALL diagnosed before
the age of 7 years were eligible for the
present or the previous study if born
and raised in The Netherlands with
Dutch as their native language. Pa-
tients with initial meningeal leukemia
or preexisting conditions interfering
with normal development were exclud-
ed. Informed consent was obtained ac-
cording to institutional guidelines. The
results for patients who were treated
for relapse (n = 7), developed other
major diseases interfering with mental
development (n = 2), or died before the
second neuropsychologic assessment
(NPA) (n = 6) were omitted from the
analysis. The first NPA was scheduled
3 to 6 months after cessation of
chemotherapy; it was repeated at a me-
dian of 5 years later to optimize the
study of persistent late effects and to
minimize potential test-retest effects.
414
Two patients were lost to follow-up,
and 2 others had a different time
schedule. Thus, 45 patients had the 2
examinations. With the second NPA,
MRI findings were available for 40
children (2 patients refused, 2 scans
could not be evaluated because of mo-
tion artefacts, and a child with a sec-
ondary brain tumor was excluded) and
were analyzed in relation to cognitive
performance. At a median of 10 years
after diagnosis, parents of all 45 pa-
tients completed a questionnaire on
school placement and levels of the pa-
tients and their 81 siblings.
Treatment: Current and
Previous Patient Groups
The current group (chemotherapy
only) consisted of 17 consecutive pa-
tients diagnosed from 1984 to 1988 and
treated with the nationwide Dutch
Childhood Leukemia Study Group
(DCLSG) protocol ALL VI.17 CNS
prophylaxis included intrathecal MTX
and prednisolone (doses of 6-12 mg)
and administration of high-dose intra-
venous MTX (3 weekly doses of 2000
mg/m2 in 24 hours) with leucovorin
rescue. Low-dose intravenous MTX
(30 mg/m2, weekly for 5 weeks, repeat-
ed every 7 weeks) was given as part of
the maintenance treatment. Intrathecal
MTX (6-12 mg), prednisolone (6-12
mg), and cytosine arabinoside (15-30
mg) were given once in 7 weeks as part
of the consolidation treatment during
the first year.
The previous group (irradiation and
chemotherapy) consisted of 28 consec-
utive children given a diagnosis of
ALL from 1978 to 1984 and treated ac-
cording to the DCLSG protocol ALL
V.18 Until 1983, children received cra-
nial irradiation in a total dose of 25 Gy
as CNS prophylaxis (20 Gy in 3 chil-
dren between 1 and 2 years of age);
thereafter, 5 children >2 years of age
had 18 Gy. The daily dose fraction in
all patients was 1.5 Gy. Additionally,
the children received intrathecal MTX
and prednisolone (5 doses; 12.5 mg/m2
with maximum MTX dose 15 mg per
THE JOURNAL OF PEDIATRICS
SEPTEMBER 2001
dose). Low-dose intravenous MTX
(30 mg/m2) weekly for 5 weeks (re-
peated every 7 weeks) was given as
part of the maintenance treatment; the
total duration of the chemotherapy
was 24 months.
Healthy Control Subjects
The neuropsychologic test scores of
the patients were compared with test
scores of 225 healthy children of the
same ages and from the same geo-
graphic area who were randomly cho-
sen from 5 public schools with socio-
economic comparability.19 Raw scores
from all tests except IQ tests were
transformed into standard scores with
mean = 10, SD ± 3, for convenience
and easy comparison. In 2 subsets of
the 225 control children, these intelli-
gence tests were administered: the
Wechsler Preschool and Primary
Scale of Intelligence (WPPSI) (n =
34) and the Wechsler Intelligence
Scale for Children-Revised (WISC-R)
(n = 47).
Neuropsychologic Instruments
NPAs included measures of the fol-
lowing:
1. Verbal and performance intelli-
gence: Dutch adaptations of the
WPPSI for children 4 to 6 years
of age and the WISC-R for those
≥6 years of age.20,21 A full-scale
IQ test, based on 10 subtests, was
also presented but not considered
an independent measure. When
we compared the only available
norms at that time with more re-
cent Dutch norms for the Wech-
sler scales, we noted a difference
of 10 points in full-scale IQ.
Therefore, the mean IQ of the
norm group approximated the
normal mean of 100.22
2. Auditory-verbal learning and
memory: the Dutch adaptation of
Rey’s Auditory Verbal Learning
Test (RAVLT) for children ≥6
years of age.19
3. Sustained attention and speed:
dot cancellation (continuous per-
THE JOURNAL OF PEDIATRICS
VOLUME 139, NUMBER 3
formance) test for children ≥6
years of age.23 (Results of this
measure were omitted from the
first assessment because <50% of
the patients could complete this
test then.) Auditory attentional
capacity: sentence repetition
WPPSI or digit span WISC-R.
4. Beery Test of Visual Motor Inte-
gration for children ≥2 1⁄2 of age.24
5. Fine motor functioning: Purdue
Pegboard (PP) for children ≥5
years of age and finger tapping for
children ≥6 years of age.25,26 All
patients were individually evalu-
ated in a morning session by 1 of
2 psychologists.
Magnetic Resonance Imaging
Brain MRI scans were performed
with a 1.5-T magnet system (Philips Gy-
roscan; Philips Medical Systems, Eind-
hoven, The Netherlands). All scans of
patients under study and scans of an
equal number of healthy children were
independently reviewed by 2 radiolo-
gists (E.L.M. and J.T.W.) who were un-
aware of the children’s status. In case of
different opinions, the MRI scans were
reassessed until consensus was reached.
The MRI scans were evaluated for 3
types of abnormalities: (1) atrophy, indi-
cated by ventricular enlargement (by
use of the Evans index)27 and/or en-
larged sulci; (2) signal abnormalities,
identified as areas of increased signal in-
tensity compared with surrounding nor-
mal white matter on T2-weighted im-
ages; and (3) calcifications, whether
absent or present. The MRI findings
were classified as normal (no abnormal-
ities), probably abnormal (slight sign of
an abnormality), or definitely abnormal
(a distinct abnormality).
School Placement and Level
The parents completed a questionnaire
for the patients and their healthy siblings
>12 years of age (when secondary edu-
cation starts in The Netherlands).
The levels of secondary education were
classified according to categories repre-
senting the 6 Dutch educational and in-
tellectual levels (1 is the lowest, 4 is the
mean level for the current population be-
tween 15 and 19 years of age, and 6 is the
highest level, those preparing for univer-
sity).28 Special education for the learn-
ing disabled comprises categories 1 and
2, representing secondary schools for
children with defective and low-normal
intelligence, respectively. Additionally,
the percentages of children referred to
special educational classes for the learn-
ing disabled during primary school were
obtained. In the Dutch educational sys-
tem, children can be temporarily placed
in special education classes but can easi-
ly change levels in the course of sec-
ondary education.
Statistics
The test performances of the current
chemotherapy group at 2 evaluations
were compared with those of the
healthy control subjects and the previ-
ously irradiated children to test the hy-
pothesis that patients receiving
chemotherapy only would have lower
test scores than the healthy control sub-
jects but higher scores than the irradiat-
ed patients. The comparisons between
groups were based on directional 1-
tailed Student t tests for unpaired
groups. Additionally, the number of
children with impaired performance
(defined as a standard score falling 1.64
SD below the mean)29 on each test was
calculated for the last assessment when
the maximum number of patients was
available.29 Then the patients who re-
ceived chemotherapy were longitudi-
nally compared with a nondirectional 2-
tailed Student t test for paired groups as
far as they completed the same tests at
both evaluations. Young age at the first
evaluation and test shift could limit the
number of longitudinal comparisons.
Significance levels are reported at P
≤ .004 (Bonferroni adjustment; a total
of 12 measures were obtained in chil-
dren ≥6 years of age) to correct for
multiple comparisons based on hy-
potheses before data collection. The
usual decisive significance levels (P ≤
.05) are also given.
KINGMA ET AL
The Fisher exact test was used to test
the hypotheses that children in the
present chemotherapy-only group (1)
would be placed more frequently in pri-
mary schools for the learning disabled
than would their siblings but less fre-
quently than would the previously irra-
diated patients and (2) would also
demonstrate fewer MRI abnormalities
than would the previous group. The hy-
pothesis that the MRI abnormalities
would correspond with lower test
scores in both patient groups was tested
with the Kruskal-Wallis test. For each
measure, the test scores were ranked for
the 3 MRI groups. The Wilcoxon
signed rank test was used to test the hy-
pothesis that the patients in both treat-
ment groups would have lower levels of
secondary education than their siblings.
We expected the children treated with
chemotherapy only to have higher
school levels than the irradiated chil-
dren; to test this, we used the Mann-
Whitney test. Analysis of variance was
applied to determine the effect of risk
factors (age at diagnosis and sex) on
neuropsychologic and MRI outcome.
All statistical analyses were carried out
with SPSS for Windows (SPSS Inc,
Chicago, Ill) or GraphPad (Ravitz Soft-
ware, San Diego, Calif) software.30,31
RESULTS
The patients receiving the present and
previous protocols did not significantly
differ from the excluded patients with re-
gard to sex, age at diagnosis, and socio-
economic status. Furthermore, no signif-
icant demographic differences were
found between the groups, except for the
younger age at the last evaluation and the
shorter follow-up period in the present
group, which was treated with a more re-
cent protocol. Characteristics of both pa-
tient groups are shown in Table I.
Neuropsychologic Evaluations
Differences are reported as not signif-
icant in case of better performance in
patients (1-tailed testing). Effect size
415
 KINGMA ET AL
(difference between group means divid-
ed by SD) is given for the last evaluation
only, when all children were ≥6 years of
age and could complete all tests.
At the first assessment, the patients
scored significantly (P ≤ .05) lower than
healthy control subjects on perfor-
mance IQ tests only. At the last evalua-
tion, the patients had significantly lower
scores on 2 of 12 (Bonferroni-corrected
P ≤ .004) or 4 of 12 (P ≤ .05) measures,
respectively (Table II). A moderate
(0.4) to large (0.8) effect size was found
for 5 comparisons. Although a deterio-
ration over time was suggested, no sig-
nificant differences were detected in the
longitudinal comparison as described in
the Methods section. According to
analysis of variance, no main effects for
age or sex could be established. Unfor-
tunately, the patient numbers were not
sufficient to study higher-order interac-
tion effects. The Figure shows impaired
functions (maximum = 12 test mea-
sures) per child at the last evaluation;
the 17 patients combined had 16 im-
pairments; 8 of the 17 patients had at
least 1 impairment.
The irradiated children had lower
scores than those in the chemotherapy
group on almost all measures (P varied
from .05 to .0008) when the present
group was compared with the previous
group at the first evaluation. Only the
results of the comparison between
both groups at the second testing are
presented because more cases were
then available, and no significant
change over time could be demonstrat-
ed in either patient group. Table III
shows that the nonirradiated children
had significantly better scores than did
the irradiated patients on one measure
on the basis of Bonferroni-corrected
P < .004 or 4 measures on the basis of P
≤ .05. Differences are reported as not
significant in case of better perfor-
mance in irradiated children (1-tailed
testing). The effect size varied from
medium (0.5) to large (0.8) for 5 com-
parisons. The Figure gives the num-
bers of impaired functions per child for
the previous study group as well; the
416
THE JOURNAL OF PEDIATRICS
SEPTEMBER 2001
Figure. Number of defective performances for all test measures per individual patient treated for
ALL at a young age: present and previous study groups (DCLSG protocols VI and V). Patients with def-
initely abnormal MRI findings are indicated by 2 arrows; those with probable abnormal MRI findings,
with 1 arrow. Striped bars, Defective Beery VMI or finger tapping or Purdue Pegboard measure (ie,
fine-motor measure); dotted bars, defective WISC-R verbal or performance IQs (defective full-scale
IQs [n = 5] are omitted); gray bars, other defective functions. Defective performance is defined as an
individual score 1.64 SD below the normative mean = lowest 5% under the standard normal curve.
28 patients combined had 57 impair- and 9 abnormalities, respectively), and
ments, and 20 of the 28 had at least one atrophy was occasionally seen (2 and 5
impairment. Moreover, poor perfor- abnormalities). No significant relation-
mance on a test with an exclusively or ships were found between MRI outcome
a predominantly motor output (Beery and test scores, school placement, or ed-
VMI, finger tapping, and Purdue Peg- ucation level in either group. For exam-
board) is separate from impairment on ple, children with high IQs could have
the purely cognitive test measures to definitely abnormal MRI findings, but
distinguish treatment effects on the pe- the reverse was found as well (Figure).
ripheral motor abilities from those on Because of the low patient numbers, we
the CNS. Thus, irradiated children still thought further analysis of risk factors
had more deficits than did the nonirra- was inappropriate.
diated patients (combined 32 deficits
in 28 cases vs 7 deficits in 17 cases). School Placements and Levels
No significant differences were
Magnetic Resonance Imaging found between the present 17 patients
In the present study group, MRI scans and their 33 siblings in placement in
were obtained for 16 of the 17 children. special primary schools for the learn-
Definitely abnormal scans were seen in 3 ing disabled or the level of secondary
cases, and a probably abnormal result education. When we compared the
was seen in another 3 cases (combined placements of the 28 previously irradi-
abnormality, 38%). In the previously ir- ated children with those of their 48 sib-
radiated group, MRIs were performed lings, we noted highly significant (P ≤
in 24 of 28 patients; 11 children had a .0001) differences (25% of the patients
definitely abnormal scan, and 4 had a were in special schools vs 4% of their
probably abnormal MRI (combined ab- siblings). Moreover, the irradiated
normality, 63%). The difference between children were more frequently referred
the two study groups was highly signifi- to special primary schools than were
cant (P ≤ .0007). White-matter abnor- the present nonirradiated children
malities were the most frequent finding (25% vs 6%, P ≤ .0003). In contrast, no
for the present and previous patients (5 significant difference was found when
THE JOURNAL OF PEDIATRICS KINGMA ET AL
VOLUME 139, NUMBER 3

Table
Table I.
I. Characteristics
Characteristics of
of patients
patients treated
treated for
for ALL
ALL at
at aa young
young age:
age: Present
present (DCLSG, ALLVI)
(DCLSG, ALL VI)and
andprevious
previous(DCLSG,
(DCLSG,ALL
ALLV)V)study group
study group
Present group Previous group (cranial
(chemotherapy only) irradiation + chemotherapy)
Males/females 10/7 11/17
Right-/left-handed 14/3 25/3
Socioeconomic status
Median education of fathers* 4.8 4.5
Median age in years at diagnosis (range) 3.5 (1.10-6.6) 3.5 (1.1-6.5)
No. of patients receiving 25 Gy (or 20 Gy <2 y) — 23
No. of patients receiving 18 Gy — 5
Median age in years at first neuropsychologic evaluation (range) 5.10 (4.2-8.10) 6.4 (3.9-8.7)
Median age in years at second neuropsychologic evaluation (range) 9.3 (7.1-12.11) 11.4 (7.3-15.8)
Median age in years at MRI of the brain (range) 9.4 (7.1-12.11) 12.0 (9.7-15.4)
Median age in years at evaluation secondary education (range÷) 15.0 (13.0-17.0) 17.0 (13.0-21.0)
Median follow-up in years since diagnosis (range) 8.9 (6.6-11.9) 13.1 (8.7-16.11)
*Median education of fathers in 9 categories.29
÷In patients 13 years of age or older only.

Table II.Results
II. Resultsof
of22neuropsychologic
neuropsychologicevaluations
evaluationsininpresent
presentpatient
patientgroup
group(n(n==17)
17)treated
treatedfor
forALL
ALLatatyoung
youngage
agewith
with
chemotherapy only (DCLSG, ALLALLVI):
VI):Comparison
Comparison(1-tailed)
(1-tailed)with
withhealthy
healthycontrol
controlsubjects
subjects
First Second
Norm evaluation evaluation
Function group patients Patients to norm patients Patients to norm Effect
and test (SD) mean (SD) t value P value 95% CI mean (SD) t value P value 95% CI size
Intelligence <6 y
WPPSI V-IQ 111.0 (10.8) 111.0 (9.7) 0.00 .50 –8.4-8.4 — — —
P-IQ 121.0 (12.0) 113.9 (9.3) 1.56 .06 –2.1-16.3 — — —
FS-IQ 117.4 (11.0) 113.6 (9.5) 0.90 .19 –4.8-12.3 — — —
Intelligence ≥6 y
WISC-R V-IQ 109.1 (15.0) 104.7 (13.2) 0.83 .21 –6.3-15.2 106.1 (15.5) 0.70 .24 –5.6-11.5 0.2
P-IQ 112.1 (15.0) 102.6 (15.0) 1.75 .04* –1.4-20.5 107.1 (10.6) 1.26 .11 –2.9-12.9 0.4
FS-IQ 111.7 (15.0) 103.6 (13.0) 1.52 .07 –2.6-18.9 107.0 (13.5) 1.14 .13 –3.6-13.0 0.3
Learning and memory
RAVLT immediate 10.0 (3.0) 9.1 (3.2) 0.83 .20 –1.2-3 9.5 (1.5) 1.10 .14 –0.4-1.3 0.2
recall
RAVLT delayed 10.0 (3.0) 9.5 (2.2) 0.42 .34 –1.8-2.7 8.5 (2.0) 2.98 .004† –0.5-2.6 0.6
recall
Attention and speed
Sentences WPPSI or 11.2 (3.7) 10.6 (3.1) 0.40 .69 –2.4-3.6 — — —
Digit span WISC-R 9.3 (3.0) 9.5 (2.5) 0.18 NS –2.5-2.1 8.6 (1.7) 1.19 .12 –0.5-1.9 0.3
BW test speed 10.0 (3.0) 10.4 (3.3) 0.49 .31 –1.9-1.2 –0.1
BW test attention 10.0 (3.0) 8.5 (2.8) 1.94 .03* –0.0-3.0 0.5
Visual Motor Integration
Beery test of VMI 10.0 (3.0) 10.0 (2.0) 0.00 .5 –1.1-1.1 9.5 (3.2) 0.78 .29 –1.1-2.0 0.2
Fine-motor functions
Finger tapping, 10.0 (3.0) 10.5 (2.2) 0.45 NS –2.6-1.6 11.4 (3.2) 1.85 NS –2.9-0.1 –0.5
pref hand
PP, pref hand 10.0 (3.0) 9.1 (2.0) 1.04 .15 –0.8-2.5 8.0 (2.3) 2.66 .004† 0.5-3.5 0.8
PP, both hands 10.0 (3.0) 9.0 (2.3) 1.19 .12 –0.7-2.7 9.0 (2.0) 1.88 .04* –0.1-2.0 0.4
PP, assembly 10.0 (3.0) 9.6 (2.6) 0.43 .34 –1.3-2.0 10.5 (1.9) 1.03 NS –1.6-0.5 –0.2
V-IQ, Verbal IQ; P-IQ, performance IQ; FS-IQ, full-scale IQ; NS, not significant; BW, Bourdon-Wiersma.
*Significance, P ≤ .05.
†Significance, P ≤ .004 (Bonferroni correction).

417
 KINGMA ET AL THE JOURNAL OF PEDIATRICS
SEPTEMBER 2001

Table III. Comparison (1-tailed) of test results at last evaluation after treatment for ALL at young age for present (DCLSG-ALL VI
with chemotherapy only) and previous study group (DCLSG, ALL V with cranial irradiation and chemotherapy)

Mean difference (in standard score)

Function and test present-previous group* t value P value 95% CI Effect size
Intelligence ≥6 y
WISC-R V-IQ 11.5 2.50 .008† 2.2-20.8 0.8
P-IQ 2.1 0.53 .30 –6.0-10.2 0.2
FS-IQ 7.7 1.72 .05† –1.3-16.7 0.5
Learning and memory
RAVLT, immediate recall 0.2 0.33 .37 –1.2-1.7 0.1
RAVLT, delayed recall –0.3 0.44 NS –2.0-1.3 –0.1
Attention and speed
Sentences, WPPSI or —
Digit span, WISC-R 1.9 2.84 .004‡ 0.5-3.2 0.7
BW test, speed –0.4 0.44 NS –2.4-1.5 –0.1
BW test, attention –2.1 1.67 NS –3.7-0.3 –0.5
Visual motor integration
Beery test of VMI 0.7 0.83 .21 –1.1-2.5 0.2
Fine motor functions
Finger tapping, pref hand 1.5 1.49 .07 –0.5-3.5 0.5
PP, pref hand –0.8 0.94 NS –2.5-0.9 –0.3
PP, both hands 0.6 0.96 .17 –0.7-1.9 0.3
PP, assembly 1.2 1.68 .05÷ –0.2-2.7 0.5
NS, Not significant.
*Positive difference: present group better than previous group; negative difference: previous group better than present group.
†Significance, P ≤ .05.
‡Significance, P ≤ .004 (Bonferroni correction).

we compared the level of secondary
education for the present and previous
groups (level 3.8 vs 3.3, respectively; P
≤ .0004). But the comparison for irra-
diated children and their siblings yield-
ed a significant difference (level 3.3 vs
4.2 respectively; P ≤ .0004).
DISCUSSION
Despite the limitation of sample size,
the current results appear to support
several hypotheses. Children treated at
a young age with chemotherapy, in-
cluding oral dexamethasone and high
cumulative doses of vincristine sulfate
(68 mg/m2), showed significantly poor-
er auditory memory and fine-motor
functioning than did the healthy con-
trol subjects. However, the hypothesis
of lower test scores across the whole
battery was not confirmed. The first
finding concerning impaired memory
corroborates evidence that cortico-
steroids as used in various treatments,
including treatment for childhood
leukemia, can affect cognitive process-
es.32-34 Our patients received oral dex-
amethasone (cumulative doses of 1444
mg/m2) instead of prednisolone during
induction and maintenance treatment;
exposure to high concentrations of
dexamethasone can have both immedi-
ate and prolonged adverse effects on
declarative memory through action on
the hippocampus. The second finding
was related to significantly poorer fine-
motor functioning, possibly reflecting
adverse effects of vincristine sulfate on
the peripheral nervous system. The im-
paired gross- and fine-motor skills as-
sociated with vincristine sulfate neu-
ropathy are widely reported in
children treated for leukemia.35
Earlier studies on children treated
with chemotherapy only yielded both
similar and contrary results. Ochs et
al7 and Mulhern et al8,9 found signifi-
cant neuropsychologic decline for sam-
ples of irradiated and nonirradiated
children, suggesting adverse effects of
MTX alone. Giralt et al10 found signif-
icantly lower IQs in children treated
with intrathecal MTX and intrathecal
arabinoside-cytosine in comparison
with their healthy siblings. Brown et
al6 concluded that modest nonverbal
impairment in children who received
chemotherapy only was associated
with the female sex. Mahoney et al36
demonstrated that general neurotoxic-
ity was more frequently associated
with intravenous MTX than with low-
dose oral MTX, but neuropsychologic
functioning was not tested as such.
Copeland et al11 and Butler et al,12 in
contrast, reported no significant cogni-
tive impairment in patients with
leukemia who were treated with a vari-
ety of chemotherapy protocols in com-
parison with patients treated for other
types of cancer.
Divergent findings are not necessarily
contradictory but may result from meth-

418
THE JOURNAL OF PEDIATRICS
VOLUME 139, NUMBER 3
odological differences.37 Patient bias by
nonrandom assignment of subjects to
study groups, significant loss of patients
in longitudinal designs, and neglecting
test shift might have introduced bias.
Our data correspond with results of
earlier studies demonstrating cognitive
impairment in irradiated patients.2,3,37
This study contributes to the existing
evidence by using a strict and equal
methodological design for the present
and previous patients who represented
consecutive nonbiased study groups
with 100% availability after 8 to 13
years of follow-up. Our irradiated pa-
tients showed more cognitive impair-
ment, more placements in schools for
the learning disabled, and more MRI
abnormalities than did the group with
chemotherapy only.
Abnormalities were still detected by
MRI in 38% of the current chemothera-
py group, but these abnormalities were
not associated with poor cognitive
achievement. We could not establish a
relationship between neuroradiologic
and neuropsychologic findings as did
Bakke et al16 and Kramer et al15, who
also observed no relationship between
the normal overall level of cognitive
functioning or IQ and the incidence of
MRI abnormalities in long-term sur-
vivors treated with MTX. In contrast,
others noticed correlations between
computed tomography or MRI abnor-
malities and measures of memory, atten-
tion, intelligence, or executive function-
ing.13-16,38 In our opinion, poor
relationships between structure and
function are not surprising, considering
the developing brain with significant but
normal variations in cognitive develop-
ment in children.
If the MRI is not sensitive enough to
explain relevant neuropsychologic late
effects of leukemia treatment, imaging
techniques, such as functional MRI,
positron emission tomography, single
photon emission tomography, and mag-
netic resonance spectroscopy, may pro-
vide a new opportunity to understand
late changes in brain functioning in
leukemia survivors. On the other hand,
Moore et al39 suggested that MRI has
become too sensitive for subtle white-
matter abnormalities to correlate with
relevant neuropsychologic impairment.
In a study of children with neurofibro-
matosis, it was indicated that the sim-
ple presence or absence (or the total
number) of white-matter abnormalities
had no relationship to the neuropsy-
chologic outcome.39 The anatomic loca-
tion was more meaningful in identify-
ing patients at risk for learning
disabilities, but the sample size in our
study was too small to detect relation-
ships between the location of MRI ab-
normalities and the cognitive deficits.
Insufficient achievement was addi-
tionally calculated per measure and per
child to judge the clinical relevance of
group findings in the present study. If
fine-motor measures were excluded, 7
of the 17 children had at least one poor
function, although the impairment was
not severe enough to affect their school
careers. Both the patients and their
healthy siblings attained the mean level
of education for the current Dutch
population of 15 to 19 years of age. Van
Dongen-Melman et al40 assessed the
prevalence of learning disorders in 36
children treated with the same
chemotherapy protocol in The Nether-
lands. At 5 years after diagnosis, 8% of
their study group was referred to spe-
cial schools for the learning disabled.
This corresponds to our finding of 6%
of the current patients. However, it
cannot be precluded that patients in
both studies needed significantly more
effort than did their siblings to reach
similar school levels and thus were suc-
cessful in overcoming minor deficits.
Moreover, our chemotherapy group
was slightly younger and in lower
grades than were the irradiated chil-
dren when their levels of secondary ed-
ucation were evaluated.
Although encouraging to parents,
patients, and pediatric oncologists, our
results support the need for careful
and continued follow-up of children
treated with this and other chemother-
apy protocols. Slight impairments in
KINGMA ET AL
the domains of memory and attention
when young may add to cognitive dis-
abilities later in life. Neuropsychologic
follow-up is therefore continued in our
institution, and each child’s perfor-
mance is analyzed to identify those
who need educational intervention.
Rehabilitation of children who have
treatment-related cognitive or fine-
motor deficits is essential to further im-
prove their quality of life.
We thank Mrs R. J. Schilperoort-Otte for
her valuable assistance in preparing this
manuscript.
REFERENCES
1. Pinkel D. Lessons from 20 years of cu-
rative therapy of childhood acute
leukemia. Br J Cancer 1992;65:148-53.
2. Roman DD, Sperduto PW. Neuropsy-
chological effects of cranial radiation:
current knowledge and future direc-
tions. Int J Radiat Oncol Biol Phys
1995;31:983-98.
3. Stebhens JA, Kaleita TA, Noll RB.
CNS prophylaxis of childhood
leukemia: what are the long term neu-
ropsychological, neurological and be-
havioral effects? Neuropsychol Rev
1991;2:147-77.
4. Madan-Swain A, Brown RT. Cognitive
and psychosocial sequelae for children
with acute lymphoblastic leukemia and
their families. Clin Psychol Rev
1991;11:267-94.
5. Brown RT, Sawyer MB, Antoniou G,
Toogood I, Rice M, Thompson N, et
al. A 3-year follow-up of the intellectu-
al and academic functioning of chil-
dren receiving central nervous system
prophylactic chemotherapy for
leukemia. J Dev Behav Pediatr
1996;17:392-8.
6. Brown RT, Madan-Swain A, Walco
GA, Cherrick I, Ievers CE, Conte PM,
et al. Cognitive and academic late effects
among children previously treated for
acute lymphocytic leukemia receiving
chemotherapy as CNS prophylaxis. J
Pediatr Psychol 1998;23:333-40.
7. Ochs J, Mulhern RK, Fairclough D,
Parvey L, Whitaker J, Ch’ien L, et al.
Comparison of neuropsychologic func-
tioning and clinical indicators of neu-
rotoxicity in longterm survivors of
childhood leukemia given cranial irra-
diation or parenteral methotrexate: a
prospective study. J Clin Oncol
1991;1:145-51.
419
 KINGMA ET AL
8. Mulhern RK, Fairclough D, Ochs J.
A prospective comparison of neu-
ropsychologic performance of children
surviving leukemia who received 18-
Gy, 24-Gy, or no cranial irradiation. J
Clin Oncol 1991;9:1348-56.
9. Mulhern RK, Wasserman AL, Fair-
clough D, Ochs J. Memory function in
disease-free survivors of childhood
acute lymphocytic leukemia given
CNS prophylaxis with or without
1,800 cGy cranial irradiation. J Clin
Oncol 1988;6:315-20.
10. Giralt J, Ortega JJ, Olive T, Verges
R, Forio J, Salvador L. Long-term
neuropsychologic sequelae of child-
hood leukemia: comparison of two
CNS prophylaxis regimens. Int J Ra-
diat Oncol Biol Phys 1992;24:49-53.
11. Copeland DR, Moore BD, Francis
DJ, Jaffe N, Culbert SJ. Neuropsy-
chologic effects of chemotherapy on
children with cancer: a longitudinal
study. J Clin Oncol 1996;14:2826-35.
12. Butler RW, Hill JM, Steinherz PG,
Meyers PA, Finlay JL. Neuropsycho-
logic effects of cranial irradiation, in-
trathecal methotrexate, and systemic
methotrexate in childhood cancer. J
Clin Oncol 1994;12:2621-9.
13. Brouwers P, Poplack D. Memory and
learning sequelae in long-term survivors
of acute lymphoblastic leukemia: associ-
ation with attention deficitis. Am J Pe-
diatr Hematol Oncol 1990;12:174-81.
14. Hertzberg H, Huk WJ, Ueberall MA,
Langer T, Meier W, Dopfer R, et al.
CNS late effects after ALL therapy in
childhood. Part I: Neuroradiological
findings in long-term survivors of
childhood ALL—an evaluation of the
interferences between morphology and
neuropsychological performance. The
German Late Effects Working Group.
Med Pediatr Oncol 1997;28:387-400.
15. Kramer JH, Norman D, Brant-Za-
wadski M. Absence of white matter
changes on magnetic resonance imag-
ing in children treated with CNS pro-
phylaxis therapy for leukemia. Cancer
1988;61:928-30.
16. Bakke SJ, Fossen A, Storm-Math-
iesen J, Lie SC. Long-term cerebral ef-
fects of CNS chemotherapy in children
with acute lymphoblastic leukemia.
Am J Pediatr Hematol Oncol
1993;10:267-70.
17. Veerman AJ, Hählen K, Kamps WA,
Van Leeuwen EF, De Vaan GA, Solbu
G, et al. High cure rate with a moder-
ately intensive treatment regimen in
420
non-high risk childhood acute lym-
phoblastic leukemia. Results of proto-
col ALL VI from the Dutch Childhood
Leukemia Study Group. J Clin Oncol
1996;3:911-8.
18. Van der Does-van den Berg A, van
Wering ER, Suciu S, Solbu G, Van’t
Veer MB, Rammeloo JA, et al. Effec-
tiveness of rubidomycin in induction
therapy with vincristine, prednisone,
and L-asparaginase for standard risk
childhood acute lymphocytic leukemia:
results of a Dutch phase III study
(ALL V). Am J Pediatr Hematol
Oncol 1989;11:125-33.
19. Van der Burg W, Kingma A. Perfor-
mance of 225 Dutch School Children
on Rey’s Auditory Verbal Learning
Test (AVLT): Parallel Test-Retest Reli-
abilities with an Interval of 3 Months:
Normative Data. Arch Clin Neuropsy-
chol 1999;14:545-59.
20. Wechsler D. Manual Wechsler
Preschool and Primary Scale of Intelli-
gence. New York: The Psychological
Corporation; 1967.
21. Wechsler D. Wechsler Intelligence
Scale for Children–Revised. New York:
The Psychological Corporation; 1974.
22. Bos H, De Sonneville LMJ. De Ned-
erlandse WISC-R. Kind Adolesc
1989;10:72-9.
23. Kamphuis GH. Een bijdrage tot de
geschiedenis van de Bourdontest. Ned
Tijdschr Psychol 1962;17:247-50.
24. Beery KE. Revised administration,
scoring and searching manual for the
development test of visual-motor inte-
gration. Cleveland: Modern Curricu-
lum Press; 1982.
25. Gardner RA, Broman M. The Purdue
Pegboard: normative data on 1334
schoolchildren. J Clin Child Psychol
1979;8:156-62.
26. Lezak MD. Neuropsychological as-
sessment. New York: Oxford Univer-
sity Press; 1995.
27. Evans WA. An encephalographic ratio
for estimating ventricular enlargement
and cerebral atrophy. Arch Neurol
Psychiatry 1942;47:931-47.
28. CBS. Enquete beroepsbevolking 1995
Centraal Bureau voor de Statistiek.
Heerlen: CBS, Div. Sociaal-Economische
Statistieken; 1996.
29. Runyon A, Haber B. Fundamentals of
behavior statistics. Reading (PA): Ad-
dison-Wesley; 1967.
30. Norusis MJ. SPSS for Windows. Base
system user’s guide. Release 6.0.
Chicago: SPSS Inc; 1993.
THE JOURNAL OF PEDIATRICS
SEPTEMBER 2001
31. Ravitz S. Graphpad Instat Tm. San
Diego: Ravitz Software Inc; 1991.
32. Squire LR. Memory and the hip-
pocampus: a synthesis from findings
with rats, monkeys and humans. Psy-
chol Rev 1992;99:195-231.
33. Keenan PA, Jacobson MW, Soleymani
RM, Mayes MD, Stress ME, Yaldoo
DT. The effect on memory of chronic
prednisone treatment in patients with
systemic disease. Neurology
1996;47:1396-402.
34. Waber DP, Carpentieri SC, Klar N,
Silverman LB, Schwenn M, Hurwitz
CA, et al. Cognitive sequelae in chil-
dren treated for acute lymphoblastic
leukemia with dexamethasone or pred-
nisone. J Pediatr Hematol Oncol
2000;22:206-13.
35. Reinders-Messelink HA, Schoemaker
MM, Hofte M, Göeken LHN, Kingma
A, Van den Briel MM, et al. Fine
motor and handwriting problems after
treatment for childhood acute lym-
phoblastic leukemia. Med Pediatr
Oncol 1996;27:551-5.
36. Mahoney DH Jr, Shuster J, Nitschke
R, Lauer SJ, Winick N, Steuber CP, et
al. Intermediate-dose intravenous
methotrexate with intravenous mer-
captopurine is superior to repetitive
low-dose oral methotrexate with intra-
venous mercaptopurine for children
with lower-risk B-lineage acute lym-
phoblastic leukemia: a Pediatric On-
cology Group phase III trial. J Clin
Oncol 1998;16:246-54.
37. Butler RW, Copeland DR. Neuropsy-
chological effects of central nervous sys-
tem prophylactic treatment in childhood
leukemia: methodological considera-
tions. J Pediatr Psychol 1993;18:319-38.
38. Chessels JH, Cox TCS, Kendall B, Ca-
vanough NPC, Jannoun L, Richards S.
Neurotoxicity in lymphoblastic leukemia:
comparison of oral and intra-muscular
methotrexate and 2 doses of radiation.
Arch Dis Child 1990;65:416-22.
39. Moore BD, Slopis JM, Schomer D,
Jackson EF, Levy BM. Neuropsycho-
logical significance of areas of high sig-
nal intensity on brain MRIs of children
with neurofibromatosis. Neurology
1996;46:1660-8.
40. Van Dongen-Melman J, De Groot A,
Van Dongen JJM, Verhulst FC,
Hählen K. Cranial irradiation is the
major cause of learning problems in chil-
dren treated for leukemia and lym-
phoma: a comparative study. Leukemia
1997;11:1197-2000.