Articles Long-term cognitive and cardiac outcomes after prenatal

Articles
Long-term cognitive and cardiac outcomes after prenatal
exposure to chemotherapy in children aged 18 months or
older: an observational study
Frédéric Amant, Kristel Van Calsteren, Michael J Halaska, Mina Mhallem Gziri, Wei Hui, Lieven Lagae, Michèl A Willemsen, Livia Kapusta,
Ben Van Calster, Heidi Wouters, Liesbeth Heyns, Sileny N Han, Viktor Tomek, Luc Mertens, Petronella B Ottevanger
Summary
Background Chemotherapy for the treatment of maternal cancers during pregnancy has become more acceptable in
the past decade; however, the effect of prenatal exposure to chemotherapy on cardiac and neurodevelopmental
outcomes of the offspring is still uncertain. We aimed to record the general health, cardiac function, and
neurodevelopmental outcomes of children who were prenatally exposed to chemotherapy.
Methods We did an interim analysis of a multicentre observational cohort study assessing children who were prenatally
exposed to maternal cancer staging and treatment, including chemotherapy. We assessed children at birth, at age
18 months, and at age 5–6, 8–9, 11–12, 14–15, or 18 years. We did clinical neurological examinations, tests of the
general level of cognitive functioning (Bayley or intelligence quotient [IQ] test), electrocardiography and
echocardiography, and administered a questionnaire on general health and development. From age 5 years, we also
did audiometry, the Auditory Verbal Learning Test, and subtasks of the Children’s Memory Scale, and the Test of
Everyday Attention for Children, and we also completed the Child Behavior Checklist. This study is registered with
ClinicalTrials.gov, number NCT00330447.
Findings 236 cycles of chemotherapy were administered in 68 pregnancies. We assessed 70 children, born at a median
gestational age of 35·7 weeks (range 28·3–41·0; IQR 3·3; 47 women at <37 weeks), with a median follow-up period
of 22·3 months (range 16·8–211·6; IQR 54·9). Although neurocognitive outcomes were within normal ranges,
cognitive development scores were lower for children who were born preterm than for those born at full term. When
controlling for age, sex, and country, the score for IQ increased by an average 11·6 points (95% CI 6·0–17·1) for each
additional month of gestation (p<0·0001). Our measurements of the children’s behaviour, general health, hearing,
and growth corresponded with those of the general population. Cardiac dimensions and functions were within normal
ranges. We identified a severe neurodevelopmental delay in both members of one twin pregnancy.
Interpretation Fetal exposure to chemotherapy was not associated with increased CNS, cardiac or auditory morbidity,
or with impairments to general health and growth compared with the general population. However, subtle changes
in cardiac and neurocognitive measurements emphasise the need for longer follow-up. Prematurity was common
and was associated with impaired cognitive development. Therefore, iatrogenic preterm delivery should be avoided
when possible.
Funding Research Foundation-Flanders; Research Fund-K U Leuven; Agency for Innovation by Science and
Technology; Stichting tegen Kanker; Clinical Research Fund-University Hospitals Leuven; and Belgian Cancer Plan,
Ministery of Health.
Introduction
The unintended use of urethane to treat chronic myeloid
leukaemia in a pregnant woman in 1948 was one of the
first reports of chemotherapy use during pregnancy.1
Since then, more experience has been gained and
chemotherapy is now regularly given after the first
trimester if cancer treatment is needed during pregnancy.2,3 Investigators estimate that cancer complicates
one in 1000–2000 pregnancies and the incidence
increases by 2·5% per calendar year.4–6 However, the
effect of the malignancy and its treatment on fetal health
remains a serious concern.
Chemotherapy during the first trimester increases the
risk of congenital malformations, whereas the effects of
chemotherapy on the fetus beyond the first trimester could
potentially affect the development of the brain and heart.
Of concern is the potential effect on the front of the brain
(attention, memory, and executive functions), because
these are most affected by cytotoxic treatment in adults
and children.7 These cognitive functions are also the most
vulnerable in infants at risk, such as preterm neonates,
children with periventricular leucomalacia, and children
in utero exposed to toxic products such as cocaine, tobacco,
or alcohol, or to maternal emotional disturbance and
distress.8–10 A second concern is the potential cardiotoxic
effect of anthracyclines, commonly used in the treatment
of breast cancer and haematological malignancies—the
most common cancer types during pregnancy.11
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
Published Online
February 10, 2012
DOI:10.1016/S14702045(11)70363-1
See Online/Comment
DOI:10.1016/S14702045(11)70408-9
See Series Lancet 2012;
379: 558, 570, and 580
Leuven Cancer Institute,
Gynaecologic Oncology
(F Amant PhD, B Van Calster PhD,
L Heyns MSc, S N Han MD),
Department of Obstetrics and
Gynaecology
(K Van Calsteren PhD,
M Mhallem Gziri MD,
B Van Calster), and Department
of Paediatrics (L Lagae PhD,
H Wouters MSc), University
Hospitals Leuven, and Division
of Imaging and Cardiovascular
Dynamics (L Mertens PhD),
Katholieke Universiteit Leuven,
Belgium; Department of
Obstetrics and Gynaecology,
Second Medical Faculty, Charles
University, Prague, Czech
Republic (M J Halaska PhD);
Department of Paediatric
Neurology, Donders Institute
for Brain, Cognition and
Behaviour (M A Willemsen PhD),
Department of Paediatric
Cardiology (L Kapusta PhD), and
Department of Medical
Oncology (P B Ottevanger PhD),
Radboud University Nijmegen
Medical Centre, Nijmegen,
Netherlands; Edit Wolfson
Medical Centre, Holon, Israel
(L Kapusta); Paediatric
Cardiocentre, Faculty Hospital
Motol, Prague, Czech Republic
(V Tomek MD); and Division of
Cardiology, The Hospital for
Sick Children, Toronto, ON,
Canada (W Hui MD, L Mertens)
Correspondence to:
Dr Frédéric Amant, Leuven
Cancer Institute, Gynaecologic
Oncology, University Hospitals
Leuven, Katholieke Universiteit
Leuven, Herestraat 49,
Leuven 3000, Belgium
[email protected]
1
Articles
So far, solely retrospective and restricted data exist on the
long-term outcome of children exposed to chemotherapy
in utero.12–14 We aimed to assess the general health, cardiac
function, and neurodevelopmental outcomes, including
intelligence, memory, attention, and executive functions,
in children who were prenatally exposed to chemotherapy.
Methods
Participants
See Online for appendix
This multicentre cohort study was started on
March 18, 2005 in three European countries and on the
basis of a collaboration between national referral centres
in Belgium (University Hospitals Leuven), the
Netherlands (Radboud University Nijmegen Medical
Centre), and the Czech Republic (Faculty Hospital Motol,
Charles University, Prague). The study is still recruiting.
Some children were enrolled and followed up prospectively; however, others were identified retrospectively (ie,
were exposed to chemotherapy before 2005), but were
followed up prospectively. We included children who
were prenatally exposed to cytotoxic drugs for cancer
treatment, but excluded children who were exposed to
low-dose chemotherapy for skin or inflammatory disorders. Further details of how children were identified
are given in the appendix. We recorded long-term toxic
effects, which we defined as those occurring after a
minimum follow-up of 18 months, secondary to in-utero
exposure to chemotherapy. A standardised study protocol
was used in the three participating centres. Written
parental permission to participate in this study was
obtained for each child. The research protocol was
approved by the institutional review board of all
participating centres.
114 children born after prenatal exposure to chemotherapy
31 retrospective (born between 1991 and 2004)
83 prospective (born aften 2004)
29 excluded
23 younger than 18 months
6 with incomplete medical records
85 eligible
15 excluded
4 lost to follow-up
1 case of sudden infant death
syndrome at 2 months
8 parents refused to participate*
2 did not speak the national language
70 included in long-term follow-up study
27 retrospective
43 prospective
Figure 1: Study profile
*Three parents did not want to be confronted with the hospital environment
because of emotional distress, one did not want to burden their child with tests,
one did not want to come to the referral hospital, and three did not want to
participate in our study.
2
Procedures
We recorded maternal disease, staging examinations,
and all treatments given in pregnancy. We organised a
standardised assessment of the general health and
development, and cardiological, cognitive, behavioural,
and neurological development of the children, including
assessments at birth, at age 18 months (the age range, at
first assessment, we allowed in this cohort was
17–29 months), and age 5–6, 8–9, 11–12, 14–15, or
18 years.15–20 We also included children from historical
datasets from the three centres and, when they entered
the study at an age between our predefined ages, their
developmental milestones were assessed by a paediatric
neurologist (LL or MAW). When these children reached
our predefined ages, we assessed them in accordance
with our protocol.
All children born after prenatal exposure to chemotherapy (both retrospectively and prospectively) were
born in hospital and assessed by a neonatologist. For the
prospective group we asked the neonatologists to
complete a datasheet (general physical examination and
neurological assessment: tonus, reflexes, active and
passive movements, eye movements). Since these are
standard examinations of neonates, this information was
available in the medical files of all patients we included
retrospectively. From age 18 months onwards we did
assessments at the national study centre. At each visit,
biometric data and a questionnaire on general health
status, school performance, and recreation and social
situation was systematically collected from the parents
(appendix). For the cognitive assessment, we developed
an age-adapted test battery for the assessment of
intelligence, verbal and non-verbal memory, attention,
working memory, and executive functions (appendix).
The Child Behavior Checklist,20 a questionnaire that
screens for behavioural and emotional problems, was
completed by the parents while their child was tested.
From age 5 years onwards we did an audiometry test once.
Tests and questionnaires were completed in the children’s
native languages. Cardiac assessment consisted of a
12-lead electrocardiograph (ECG) and a full echocardiographical assessment for structural and functional
characteristics (appendix).
Statistical analysis
We compared our findings with available norms:
national data for height, weight, head circumference, by
sex,21 and national and international reference data for
neurodevelopmental tests.15–20 Some children were
assessed twice; we ensured there were no important
differences between the two sets of results and used the
most recent data. We converted all raw scores from the
cognitive tests to standardised scores with published
normative data for the specific age group as provided by
the respective tests. For the Bayley and Wechsler
intelligence tests, 100 plus or minus 15 is thought the
normal range of index scores. For the data analysis and
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
Articles
representation of the Test of Everyday Attention for
Children, the Children’s Memory Scale, and the Auditory
Verbal Learning Test, we converted all scores to Z scores,
with a normal range defined as a Z score of 0 (SD 1). We
assessed the relation between intelligence quotient (IQ)
scores and gestational age at birth by linear (ordinary
least squares) regression. We estimated the effect size
with the ω² measure of explained variance. We added
age and sex as covariates, together with a random effect
for country.
We analysed and compared electrocardiographic
measurements with normal values in childhood and
adolescence.22 All echocardiographic measurements
were obtained in three cardiac cycles and averaged.
We compared all cardiac measurements with the
measurements in a matched control group (1:1; from
Toronto and Leuven) with age (plus or minus 12 months)
and sex as matching factors. We compared our cardiac
measurements with reference values from a historical
dataset and represented these as Z scores. To circumvent
violations of the linear regression assumptions, mainly
the assumption of normality of the residuals due to
skewed distributions for some cardiac variables, we
estimated the independent effect of chemotherapy with
median regression. The matching was accounted for by
stratifying the regression models by age (six strata) and
adding sex as a covariate.23 When both age and sex were
used as covariates, as a sensitivity analysis, we obtained
similar results. We corrected for multiple testing with
Holm’s method for all cardiac measurements. Analyses
were done with SAS (version 9.2) and Statistical Package
Number receiving electrocardiogram and
echocardiogram (cardiological
assessment; n=65)*
for Social Sciences for Windows (version 16). This
study is registered with ClinicalTrials.gov, number
NCT00330447.
Role of the funding source
The sponsors of the study had no role in the study design,
data collection, data analysis, data interpretation, or
writing of the report. The corresponding author had full
access to all the data in the study and had final
responsibility for the decision to submit for publication.
Results
This interim analysis was done with a data cutoff at
March 1, 2011. Figure 1 shows the study profile. 32 girls
and 38 boys were exposed to chemotherapy in utero
(236 cycles) during 68 pregnancies (two twin pregnancies). Distribution of children by country was 42 in
Belgium, 20 in the Netherlands, and eight in the Czech
Republic. Maternal disease, staging examinations, and
drugs administered as supportive care are shown in the
appendix. During pregnancy, 34 women received
chemotherapy, one received chemotherapy and radiotherapy, 27 received chemotherapy and surgery, and six
received all three treatment modalities. 19 different
regimens of chemotherapy were administered (appendix)
and anthracyclines were the most common agents
(53 patients). The median cumulative dose for doxorubicin
(n=33) was 180 mg/m² (range 50–400; IQR 120), for
epirubicin (n=14) was 450 mg/m² (200–600; 240), for
idarubicin (n=3) was 72 mg/m² (36–108; 36), and for
daunorubicin (n=4) was 240 mg/m² (120–360; 60).
Number receiving paediatric examination
(n=70) plus questionnaire for parents
(n=57; general health assessment)†
Neurological examination (n=70)
Neonatal
21
70‡
Clinical neurological examination
18 months
40§
40
Clinical neurological examination and Mental Development Index of the
Bayley Scales of Infant Development15
5–6 years
15
16
Clinical neurological examination, audiometry, and age-adapted test battery¶
8–9 years
11||
13
Clinical neurological examination and age-adapted test battery**
11–12 years
2
2
Clinical neurological examination and age-adapted test battery**
14–15 years
1
1
Clinical neurological examination and age-adapted test battery**
18 years
2
2
Clinical neurological examination and age-adapted test battery††
Inclusion between
standardised ages‡‡
6§§
6
Clinical neurological examination
Only children in the long-term follow up are represented. *One child tested at age 18 months and 6 years, five at age 6 and 9 years, one at age 9 and 12 years, and one at age 15 and 18 years. †One child tested at age
18 months and 6 years, five at age 6 and 9 years, two at age 4 (between standardised ages) and 9 years, one at age 9 and 12 years, and one at age 15 and 18 years. ‡Parents were not asked to complete questionnaire
since it is not appropriate for neonates. §In six children not all echocardiographic data were obtained because of the child’s movements; in three of these children echocardiography was repeated after 1 or 2 years to
complete the data set. Two of the 40 children were tested at age 1 year, the examination was not repeated 6 months later since it was within normal ranges. ¶For this age group the age-adapted test battery comprised
the Wechsler Preschool and Primary Scale of Intelligence (one child was tested with the Snijders-Oomen non-verbal intelligence test because their native language was not the national language), subtasks from the
Children’s Memory Scale (Numbers for verbal memory; Dot Locations, Faces, and Pictures for non-verbal memory),17 and the Child Behavior Checklist (only available in Dutch and English).20 ||In a twin with an important
neurodevelopmental delay, cardiac assessment was done at 4 years and was within normal ranges. At age 9 years cardiac assessment was not done at the request of the parents. Neurological assessments were done at
ages 4 and 9 years. **For these age groups the age-adapted test battery comprised the Wechsler Intelligence Scale for Children (one child was tested with the Snijders-Oomen non-verbal intelligence test because their
native language was not the national language), the Test of Everyday Attention for Children,16 subtasks from the Children’s Memory Scale (Numbers for verbal memory; Dot Locations, Faces, and Pictures for non-verbal
memory), the Auditory Verbal Learning Test,18,19 and the Child Behavior Checklist (only available in Dutch and English). ††For this age group the age-adapted test battery comprised the Wechsler Adult Intelligence Scale,
the Auditory Verbal Learning Test, and the Child Behavior Checklist. ‡‡Ages of children in this group were 1·9, 2·1, 2·3, 3·3, 3·7, and 3·7 years (median 2·8 years; range 1·9–3·7). §§Echocardiography was not complete for
one child because of the child’s movements.
Table 1: Medical assessments by age group
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
3
Articles
A
B
100
Weight (kg)
80
60
40
20
0
C
D
200
Height (cm)
150
100
50
0
E
F
Head circumference (cm)
80
60
40
20
0
0
50
100
Age (months)
150
200
0
50
100
Age (months)
150
200
Figure 2: Biometry of children after prenatal exposure to chemotherapy
Girls’ weight (n=32; A), boys’ weight (n=38; B), girls’ height (n=31; C), boys’ height (n=30; D), girls’ head circumference (n=29; E), and boys’ head circumference (n=30;
F). For each child we used the last available measurement. The lines are the 5th, 50th, and 95th percentiles; the dots are the measurements from our study population.
For the customised weight
centile calculator see http://
www.gestation.net/
4
Median maternal age was 32·9 years (range 23·4–41·7;
IQR 6·0) and gestational age at diagnosis was 18·1 weeks
(1·7–33·1; 9·0). Median gestational age at birth was
35·7 weeks (28·3–41·0; 3·3). Seven children were born
at 28·0–31·9 weeks, nine at 32·0–33·9 weeks, 31 at
34·0–36·9 weeks, and 23 at term (≥37 weeks). Median
birthweight was 2612 g (720–3970; 955). Birthweight
was below the tenth percentile for gestational age and
sex in 14 of 70 children (20·6%, 95% CI 12·3–30·8; onetailed binomial test p=0·009). The incidence and type
of congenital malformations were similar to the general
population (appendix). Furthermore, neonatal physical
examination and echocardiographic assessments (n=21)
were within normal limits (data not shown). Neonatal
neurological examination was normal in 64 children
(91%). One child, born at 28 weeks, presented with a
contracture of the right elbow, which resolved later. In
five children a transient neonatal hypotonia was noted;
one of these children also presented with benign sleep
myoclonus.
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
Articles
Average group results on the different subtasks for
verbal and non-verbal memory were within normal ranges
(figure 4). Attention function for the 12 children with
available data was within the normal range. Figure 4 lists
detailed results grouped per attention function: focused
attention, sustained attention, attention flexibility, divided
attention, and response inhibition. The Child Behavior
Checklist was available for 21 children at a median age of
8·7 years (range 5·0–15·9; IQR 3·0) and the average score
was within the normal ranges (figure 4). For internalising,
A
60
Intelligence test between age 5 and 18 years
Bayley test at age 18 months
Clinical neurological examination
49
Number
40
20
10
0
5
5
x – (>2 SD)
x – (1–2 SD)
1
x ± (1 SD)
x + (1–2 SD)
x + (>2 SD)
B
140
130
120
IQ score
110
100
90
80
70
60
50
28 29
30
31 32
33
34
35 36
37
Pregnancy duration (weeks)
C
150
140
Verbal IQ
Performance IQ
Total IQ
130
120
IQ score
The median follow-up period at last examination of the
children was 22·3 months (range 16·8–211·6; IQR 54·9).
Medical assessments by age are shown in table 1.
Biometric data showed that children had normal values
for weight, height, and head circumference (figure 2).
The questionnaire on general health status and
development was completed by 57 parents (81%) at the
study visits. Medical problems reported by the parents
(appendix) are also common in the general population. We
did not identify any malignancies. General development
was reported as seriously impaired in both children of a
twin pregnancy. Apart from these children, all school-aged
children attended a normal school (n=25). Social and
leisure activities were similar to the general population
(data not shown). Children with missing questionnaires
mainly lived in the Netherlands and the French speaking
part of Belgium. Apart from their different cultural
background, we had less personal contact with these
parents since we were not the treating physicians. However,
we did not identify differences in maternal history and
children’s outcomes when compared with parents who
completed the questionnaires.
We recorded normal cognitive development in most
children. Figure 3 shows the distribution of the results for
the last cognitive evaluation (Bayley test [n=40], Wechsler
or Snijders-Oomen non-verbal intelligence test [n=26],
and clinical neurological examination between standardised
ages [n=4]). Children who scored below normal ranges
were mainly from the preterm group. A univariable linear
regression model showed that the average IQ score
increased by 11·1 points (95% CI 5·4–16·8; p=0·0003) for
each month increase in pregnancy duration (explained
variation in IQ scores was 16%; figure 3). When controlling
for age, sex, and country, the effect remained: the score
increased on average by 11·6 (95% CI 6·0–17·1, p<0·0001)
for each month increase in pregnancy duration.
Two children (one of the sets of twins) had a significant
neurodevelopmental delay, which made it impossible to
do the complete proposed in-depth cognitive test battery
(appendix). We excluded these two children from further
cognitive and behavioural assessment. In one of these
children, a cortical malformation with multiple dysmorphic characteristics was recorded (appendix).
Although a syndromal entity is probable in this child, we
were unable to diagnose a clinical syndrome. The parents
refused a brain MRI for the other twin, but this child has
no physical abnormalities.
Total IQ scores for the children were within normal
ranges (n=25; 103 [SD 14·5]). Comparing verbal and
performance IQ scores obtained in the Wechsler tests
(n=23), paired t test showed a significant difference
(verbal score 104·8 [SD 14·5] vs performance score
97·9 [14·1]; paired t test p=0·033). This observation was
driven by nine children (39%) where there was a
significant difference between verbal and performance
IQ scores (figure 3); the remaining 14 (61%) did not
show a significant difference.
110
100
90
80
70
60
50
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
Individual child
38 39 40
41
Figure 3: Cognitive
development
(A) General cognitive
development after prenatal
exposure to chemotherapy
(n=70). The distribution of the
result of the last performed
neurological evaluation for
each child (Bayley test at age
18 months [n=39], intelligence
test between age 5 and
18 years [n=24 Wechsler IQ test,
n=2 Snijders-Oomen
non-verbal intelligence test,
n=1 Bayley test], clinical
neurological examination in
between standardised ages
[n=4]). Normal range is defined
as x plus or minus 1 SD. (B)
Cognitive outcome in relation
to gestational age at birth.
IQ score was obtained from
Wechsler test (n=24),
Snijders-Oomen non-verbal
intelligence test (n=2), Bayley
test (n=40), clinical
neurological examination
whereby findings within
normal ranges were translated
into a mean score of 100 (n=4).
A linear regression model
suggested that the average IQ
score increases by 2·5
(95% CI 1·2–3·9) for each week
increase in pregnancy duration
(p=0·0003, ω²=0·16). Circles
are individual data, solid line is
predicted IQ score, and dashed
lines are the 95% CIs for the
prediction. (C) IQ scores from
Wechsler (n=23) and
Snijders-Oomen non-verbal
intelligence tests (n=2; last two
columns). Vertical lines
represent a significant
difference between verbal and
performance IQ scores. Shaded
band represents the normal
range. IQ=intelligence quotient.
5
Articles
Task or subtask
Behaviour
CBCL
Verbal memory
AVLT
Verbal memory span
CMS Numbers
Non-verbal memory
CMS Dots
CMS Faces
Tested variable
N
Median age (range; IQR) in years
Internalising
Externalising
Total problems
21
21
21
8·7 (5·0–15·9; 3·0)
8·7 (5·0–15·9; 3·0)
8·7 (5·0–15·9; 3·0)
Learning
Immediate recall
Delayed recall
12
12
12
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
Forward
Backward
Total numbers
25
25
25
8·5 (5·0–17·6; 3·0)
8·5 (5·0–17·6; 3·0)
8·5 (5·0–17·6; 3·0)
Learning
Immediate recall
Delayed recall
Immediate recall
Delayed recall
25
25
24
25
24
8·5 (5·0–17·6; 3·0)
8·5 (5·0–17·6; 3·0)
8·2 (5·0–17·6; 3·0)
8·5 (5·0–17·6; 3·0)
8·2 (5·0–17·6; 3·0)
25
8·5 (5·0–17·6; 3·0)
12
12
12
12
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
12
12
12
12
12
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
12
11
12
12
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 1·0)
9·0 (8·5–15·9; 0·6)
9·0 (8·5–15·9; 0·6)
Non-verbal memory span
CMS Pictures
Selective attention
Tea-Ch Sky Search
Correct score
Time score
Attention score
Tea-Ch Map Mission
Sustained attentions
Tea-Ch Score
Tea-Ch Code Transmission
Tea-Ch Sky Search DT*
Tea-Ch Score DT*
Tea-Ch Walk/Don’t Walk†
Attentional flexibility
Tea-Ch Creature Counting Correct score
Time score
Tea-Ch Opposite World
Same world
Opposite world
–1
0
Z score
1
Figure 4: Cognitive development and behaviour (n=25, 5–18-year-old children)
*Also tests divided attention. †Also tests response inhibition. CBCL=Child Behavior Checklist. AVLT=Auditory Verbal
Learning Test. CMS=Children’s Memory Scale. Tea-Ch=Test of Everyday Attention for Children.
externalising, and total problems, six (29%) of 21 children
with available data had an increased score (Z score >1·0).
No significant relation was evident with prematurity. All
these children attended a normal school.
Clinical neurological examination (n=68) did not
show any focal or other neurological abnormalities. The
mean Bayley mental developmental index score was
96·8 (SD 14·9; n=40, at a median age of 18·2 months
[range 16·8–28·6; IQR 0·7]). Index scores for children
born at term (n=13) were 103·1 (SD 13·6) and preterm
(n=27) were 94·6 (SD 14·6).
We tested auditory function in 21 children at the
median age of 6·5 years (range 5·0–17·4; IQR 2·9). We
did not detect abnormalities in 18 children (86%; four
received cisplatin); in three children with hearing loss,
middle ear infection (n=1) and neurodevelopmental
problems (n=2) were confounding factors (appendix).
69 children aged 1–18 years underwent a non-sedated
ECG and echocardiographic assessment (two children
were tested off-protocol at 12 months). One child
refused echocardiography and we excluded four
children because it was impossible to complete the
assessment because of a lack of cooperation; so data of
65 children were available for analysis. When compared
with controls, we did not identify any statistically
significant differences by weight, height, body surface
area, and systolic blood pressure (all p values ≥0·45).
49 pregnancies (78%) of 64 had been exposed to
anthracyclines. We noted a higher heart rate in our
study participants (median 109, range 61–152) by
comparison with the control group (104, 54–139;
p=0·25); however, analysis of ECG measurements
revealed no arrhythmia or conduction abnormalities
(appendix). During the echocardiographic examination
we did not identify any structural cardiac defects.
Table 2 summarises the echocardiographic measurements of cardiac dimensions and systolic function in
children exposed to anthracyclines. All cardiac
dimensions were within the normal range for children
N*
Study children (n=50) Controls (n=50)
Shortening fraction (%)
98
35 (30 to 43)
39 (28 to 51)
–4·0 (–5·9 to –2·1)
<0·0001†
Ejection fraction (%)
98
66 (60 to 75)
71 (57 to 83)
–6·0 (–8·3 to –3·7)
<0·0001†
Left ventricle mass index (g/m²)
94
48·3 (37·4 to 65·5)
55·4 (40·0 to 82·6)
–2·53 (–7·89 to 2·82)
0·35
Left ventricle mass index (left ventricle mass per height²·⁷)
93
35·9 (19·8 to 54·5)
35·3 (20·6 to 28·8)
–0·94 (–3·79 to 1·91)
0·51
Left ventricle Tei index
94
0·33 (0·27 to 0·45)
0·32 (0·15 to 0·53)
Left ventricle end diastolic diameter (Z score)
96
0·02 (–2·28 to 3·19)
0·25 (–2·22 to 3·00)
Left ventricle posterior wall thickness (Z score)
95
0·16 (–2·32 to 1·81)
Interventricular septum thickness (Z score)
95
–1·09 (–2·50 to 0·45)
Right ventricle end diastolic diameter (Z score)
93
0·76 (–1·48 to 1·83)
Effect (95% CI)
0·014 (–0·025 to 0·052)
p value
0·47
–0·13 (–0·64 to 0·38)
0·61
–0·20 (–1·90 to 1·98)
0·35 (–0·25 to 0·94)
0·25
–0·64 (–2·10 to 1·65)
–0·47 (–0·89 to –0·05)
0·029
0·30 (–1·91 to 2·21)
0·29 (–0·19 to 0·76)
0·24
Data are median (range) unless otherwise stated. We estimated the independent effect of chemotherapy with stratified median regression with six strata for age and with sex
as covariate. *Missing data are from both the study and control groups: shortening fraction (2 control), ejection fraction (2 control), left ventricle mass index, g/m² (2 study,
4 control), left ventricle mass index, left ventricle mass per height²·⁷ (3 study, 4 control), left ventricle Tei index (2 study, 4 control), left ventricle end diastolic diameter
(2 study, 2 control), left ventricle posterior wall thickness (2 study, 3 control), interventricular septum thickness (2 study, 3 control), and right ventricle end diastolic diameter
(2 study, 5 control). †Statistically significant after correction for multiple testing with Holm’s method including the variables from the appendix.
Table 2: Echocardiographic data of children exposed to anthracyclines in utero and their matched controls
6
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
Articles
exposed to anthracyclines and for control children.
Compared with the control group, we noticed a clinically
small but statistically significant decrease for the patient
group concerning ejection fraction, fractional
shortening, and interventricular septum thickness.
However, all patients were within the normal range and
no patient had an abnormal value. Diastolic variables
were also within the normal range with some clinically
small but statistically significant differences between
both groups. Mitral valve E velocity was lower, mitral
valve A duration was shorter, and isovolumetric
relaxation time was shorter in the patient group
compared with the controls (appendix). These variables
are heart-rate dependent.
Discussion
Despite prenatal exposure to chemotherapy, radiotherapy, staging examinations, and co-medications, the
outcome for children in our study is not different from
the general population. In particular, we noted ageadequate cognitive development and normal cardiac
outcome in a cohort of children aged at least 18 months
who were prenatally exposed to chemotherapy and tested
at predefined ages. In our cohort, we confirmed the
negative prognostic effect of prematurity on cognitive
development (Bayley or IQ score).8,9 However, we cannot
exclude an additional effect of chemotherapy and other
drugs or radiation exposure on these children born
preterm. The decision to induce the delivery for nonobstetric reasons—mostly to start cancer treatment or
in a few cases because of deterioration of the maternal
health—was taken in 38 pregnancies (56%) of 68.
28 (74%) of these 38 patients were delivered, iatrogenically,
preterm (median gestational age 35·3 weeks; range
31·3–36·3). We suggest that the start of cancer treatment
(including chemotherapy) during pregnancy might
prevent iatrogenic prematurity and add to the preservation of long-term neurocognitive outcome.
These data should be interpreted with caution since
two children, both members of a twin pregnancy,
presented with an important neurodevelopmental delay.
The fetal cortex changes during development from a
smooth cerebral surface at 14 gestational weeks to a
complex association of sulci and gyri. Polymicrogyria is
the endpoint of different pathological processes.24 We
cannot rule out that prenatal exposure to chemotherapy
after 15 weeks of gestational age affected the developmental delay, although the syndromal presentation
makes the causal relation less probable.
Moreover, in the children older than 6 years who
did a Wechsler intelligence test (n=23), a disharmonic
intelligence profile—discrepancies between verbal and
performance IQ values—was evident in 39% (15% in the
normal population).25 Although this finding does not
directly suggest neuropsychological impairment, verbal
and performance IQ discrepancies have been associated
with several neurological disorders and learning
problems. Furthermore, children with disharmonic
intelligence profiles more often have behavioural
problems than children with harmonic intelligence
profiles.26 In our series the average results of the Child
Behavior Checklist were within normal ranges, although
in six (29%) of the 21 children the total problem score
was raised, which is higher than we expected (15% in the
general population).9 Although the general neurodevelopmental assessment was within normal ranges,
these findings show that more subtle changes in
neurodevelopment are possible. Therefore we need to
remain prudent until a larger group of children has been
assessed with a longer follow-up.
In our study, fetal exposure to chemotherapy in utero
was not associated with congenital cardiac abnormalities.
For the children exposed to anthracyclines, variables for
systolic and diastolic function were all within normal
ranges. We identified a significant difference in ejection
fraction and fractional shortening between the patients
and age-matched and sex-matched controls. However,
because values obtained in all patients in the
anthracycline group were normal, as in the report of
Aviles and colleagues,14 we do not believe this difference
is clinically relevant. Some of the diastolic variables
(isovolumetric relaxation time and mitral A duration)
were different between patients and controls, but also
these data were within the normal range and the
differences can probably be explained by the higher heart
rate in the study group. It is reassuring that cardiac
dimensions, wall thickness, and left ventricular mass
index were all within normal ranges. We discussed the
need for matching for prematurity with our cardiologists
although prematurity is not a predictor of heart
dysfunction. Furthermore, we did not identify heart
dysfunction for premature children.
We believe that three important factors contributed to
our overall reassuring outcome. First, chemotherapy was
only given after the first trimester, which is the most
vulnerable time for toxic effects.27 Chemotherapy given
during the second and third trimester does not increase
congenital malformations.11,28 Second, by contrast with
the belief that the fetal blood–brain barrier is immature
and leaky, recent data suggest that the fetal brain is well
protected.29–31 Compared with the situation in adults, fetal
cerebrospinal fluid contains high concentrations of
proteins related to specialised transcellular transfer and
needed for fetal brain development.30 Erroneously, these
high protein concentrations were once interpreted as a
consequence of a leaky blood–brain barrier. During
embryogenesis, almost all neurons are formed by
6–18 weeks of gestation, but the brain continues to
develop later in pregnancy and also postnatally, through
neuronal migration, differentiation and synaptic maturation, and myelinisation. The underlying morphological
features of the blood–brain barrier are the presence
of tight junctions, low rates of transcytosis, and the
expression of specialised influx and efflux transporters,
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
7
Articles
Panel: Research in context
Systematic review
To our knowledge, ours is the first comprehensive report on the long-term outcome of
children after prenatal exposure to chemotherapy. We searched PubMed for reports
published from 1990 to 2011, with the search terms “pregnancy”, “cancer”, “children”,
“chemotherapy”, and “outcome”. We did not limit our search by language. We also
searched review papers. We identified very limited published data on the examination
of children long term after prenatal exposure at the time of our search, indicating a
need for information in this area.
Interpretation
Our findings suggest that general health and growth, and CNS, cardiac, and auditory
function of children born to mothers treated with chemotherapy during pregnancy did
not differ from the general population. Our findings do not support a strategy of delay in
chemotherapy administration or iatrogenic preterm delivery with post-partum
chemotherapy administration to avoid harm to the fetus. However, we noted subtle
changes and underscore the need for longer follow-up in more children.
For more on Cancer in
Pregnancy see http://www.
cancerinpregnancy.org/
8
which are present early in embryological development.30
Moreover, pericytes inhibit the expression of molecules
that increase vascular permeability and CNS immunecell infiltration.29 Tight junctions are present between
cerebral endothelial cells and between choroid plexus
epithelial cells, and restrict the entry of proteins into the
brain and cerebrospinal fluid. In the immature brain
there are additional morphological barriers at the
interface between the cerebrospinal fluid and brain
tissue: strap junctions at the inner neuroependymal
surface and these and other intercellular membrane
specialisations at the outer (pia-arachnoid) surface.
These barriers disappear later in development and
are absent in adults.30 Apart from this morphological
protection, the presence of functional efflux transporters,
including P-glycoprotein, reduces brain penetration of
drugs.32 Virgintino and colleagues31 reported on the
expression of P-glycoprotein early in human fetuses
during cerebral cortex formation. At the earliest
examined stage, 12 weeks of gestation, P-glycoprotein
was detectable as diffuse cytoplasmic labelling of the
endothelial cells lining the primary cortex microvessels.
At 18 weeks of gestation, a punctate P-glycoprotein
staining pattern was detected on cortex and subcortical
vessels and on their side branches. At 22 weeks of
gestation, P-glycoprotein staining was linear and
concentrated on endothelial cell membranes. Third, the
placenta filters cytotoxic drugs and shields a proportion
from the fetus. Data on human beings are anecdotal, but
transfer rates in pregnant baboons vary and range from
0 to 57% of the maternal serum concentrations.33,34
The limitations of our study include our small sample
size, a short follow-up period, and a lack of a direct
comparison with identically assessed children, born at
the same gestational age but without prenatal exposure
to chemotherapy. To tackle the latter, we have now started
to build a control group: children matched for the
gestational age at birth, which will be followed up in
accordance with the same protocol as the children
exposed to chemotherapy. Because of the short follow-up
period, our study does not allow us to investigate for
secondary malignancies. In children, the use of low-dose
etoposide is associated with secondary leukaemia, but
genetic susceptibility to neoplasia might be a confounding factor.35 So far, only one case of secondary neoplasm
after transplacental exposure to chemotherapy has been
described. One member of a twin exposed to DNAdamaging cyclophosphamide and prednisone for acute
lymphocytic leukaemia was born with congenital
malformations and was diagnosed with papillary thyroid
cancer at age 11 years and neuroblastoma at age 14 years.36
Given the unknown effect on secondary malignancies,
where possible, preference to microtubule-targeting
agents and especially taxanes should be given.2,37 Longterm follow-up is thus mandatory but sensitive to survivor
bias. Nevertheless, all children should be followed up
and reasons for exclusion, loss of follow-up, or refusal for
participation need to be documented properly.
We show that children who were prenatally exposed to
chemotherapy do as well as other children. Based on our
data, the long-term effect of prenatal exposure to
chemotherapy seems to outweigh the maternal need for
treatment during pregnancy, especially when we consider
confounding factors such as prematurity and maternal
stress.8–10 Although we cannot exclude a role of chemotherapy in the poor outcome of a twin and a higher
incidence of disharmonic intelligence profiles, our
findings suggest that chemotherapy could be used during
pregnancy, if needed. The decision to administer
chemotherapy should follow the same guidelines as in
non-pregnant patients.3 In practice, it is possible to
administer chemotherapy from 14 weeks gestational age
onwards with specific attention to prenatal care.3 To allow
the bone marrow to recover and to minimise the risk of
maternal and fetal sepsis and haemorrhage, delivery
should be planned at least 3 weeks after the last cycle of
chemotherapy, and chemotherapy should not be given
after 35 weeks since spontaneous labour becomes more
probable.3 Furthermore, neonates, especially preterm
babies, have restricted capacity to metabolise and
eliminate drugs because of liver and renal immaturity.
The delay of delivery after chemotherapy will allow fetal
drug excretion via the placenta.3 Given the negative
prognostic effect of prematurity on cognitive development,
preterm birth should be avoided, if possible. Only time
will inform us of the full consequences, including fertility
and secondary malignancies (especially if DNA damaging
drugs are used), of fetal exposure to chemotherapy.
Therefore, we continue our international collaborative
initiative, Cancer in Pregnancy, and strive to follow up a
greater number of children over a longer period (panel).
Contributors
FA thought of the concept of the study. FA and KVC searched the
published work. FA, MJH, and PBO were the national study
coordinators. FA, KVC, MMG, LH, SNH, MJH, and PBO identified
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
Articles
children and organised examinations in the national study centres; they
collected all clinical data and study results. MMG, LK, and VT did the
cardiac assessments. FA, MMG, LM, WH, and LK did the cardiac data
analysis and interpretation. FA, KVC, MJH, LL, MAW, and HW did the
cognitive tests, cognitive data analysis, and interpretation. KVC and
BVC did the statistical analyses. FA and KVC wrote the first draft of the
report. All authors approved the final version of the report.
Conflicts of interest
We declare that we have no conflicts of interest.
Acknowledgments
This report was presented as a late breaking abstract during the
European Multidisciplinary Cancer Congress in Stockholm, Sweden
(Sept 26, 2011). FA is senior clinical researcher and BVC postdoctoral
researcher for the Research Foundation-Flanders; this research is
supported by Research Foundation-Flanders Project G 0358.06,
“Stichting tegen kanker Project SCIE2006-17”, Research Fund-K.U.Leuven
(OT/07/053), Flemish Government, Agency for Innovation by Science
and Technology Project TBM070706-IOTA3, Clinical Research Fund-UZ
Gasthuisberg, and Belgian Cancer Plan, Ministry of Health NKP 29 038.
We thank the parents and children for their participation in our study,
and all colleagues who contributed to this study. We thank Javier Ganame
and Daisy Thijs (Leuven), Bob Rijk and Imke Tomasouw-Janssen
(Nijmegen) for echocardiographic assessment of the children. We thank
Ilse De Croock, Griet De Mulder, and Caroline Sterken (Leuven);
Anja Vinck (Nijmegen); and Jitka Zackova (Prague) for cognitive
evaluation of the children. We thank Anna Jansen from Vrije Universiteit
Brussel for mutational analysis in two patients. We thank
Liesbet Van Eycken from the Belgian Cancer Registry and Frank De Smet
from the National Alliance of Christian Sickness Funds for helping with
data extraction. We thank Marieke Taal, Diane Wolput, and
Marie-Astrid Van Hoorick for their administrative support. The full list of
investigators is in the appendix.
References
1
Creskoff A, Fitz-Hugh TJ, Frost J. Urethane therapy in leukemia.
Blood 1948; 3: 896–910.
2
Amant F, Van Calsteren K, Halaska MJ, et al. Gynecologic cancers
in pregnancy: guidelines of an international consensus meeting.
Int J Gynecol Cancer 2009; 19 (suppl 1): S1–12.
3
Amant F, Loibl S, Neven P, Van Calsteren K. Breast cancer in
pregnancy. Lancet 2012; 379: 570–79.
4
Nieminen U, Remes N. Malignancy during pregnancy.
Acta Obstet Gynecol Scand 1970; 49: 315–19.
5
Potter JF, Schoeneman M. Metastasis of maternal cancer to the
placenta and fetus. Cancer 1970; 25: 380–88.
6
Stensheim H, Moller B, van Dijk T, Fossa SD. Cause-specific
survival for women diagnosed with cancer during pregnancy or
lactation: a registry-based cohort study. J Clin Oncol 2009; 27: 45–51.
7
Mennes M, Stiers P, Vandenbussche E, et al. Attention and
information processing in survivors of childhood acute
lymphoblastic leukemia treated with chemotherapy only.
Pediatr Blood Cancer 2005; 44: 478–86.
8
Lohaugen GC, Gramstad A, Evensen KA, et al. Cognitive profile in
young adults born preterm at very low birthweight.
Dev Med Child Neurol 2010; 52: 1133–38.
9
van Baar AL, Vermaas J, Knots E, de Kleine MJ, Soons P.
Functioning at school age of moderately preterm children born at
32 to 36 weeks’ gestational age. Pediatrics 2009; 124: 251–57.
10 Van den Bergh BR, Van Calster B, Smits T, Van Huffel S, Lagae L.
Antenatal maternal anxiety is related to HPA-axis dysregulation
and self-reported depressive symptoms in adolescence:
a prospective study on the fetal origins of depressed mood.
Neuropsychopharmacology 2008; 33: 536–45.
11 Van Calsteren K, Heyns L, De Smet F, et al. Cancer during
pregnancy: an analysis of 215 patients emphasizing the obstetrical
and the neonatal outcomes. J Clin Oncol 2010; 28: 683–89.
12 Aviles A, Neri N. Hematological malignancies and pregnancy:
a final report of 84 children who received chemotherapy in utero.
Clin Lymphoma 2001; 2: 173–77.
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Hahn KM, Johnson PH, Gordon N, et al. Treatment of pregnant
breast cancer patients and outcomes of children exposed to
chemotherapy in utero. Cancer 2006; 107: 1219–26.
Aviles A, Neri N, Nambo MJ. Long-term evaluation of cardiac
function in children who received anthracyclines during
pregnancy. Ann Oncol 2006; 17: 286–88.
van der Meulen BF, Ruiter SAJ, Spelberg HCL, Smrkovský M.
BSID-II-NL: Bayley Scales of Infant Development, 2nd edn—
Nederlandse versie. Lisse: Swets Test Publishers, 2002.
Manly T, Robertson I, Anderson V, Nimmo-Smit I. TEA-Ch: test of
everyday attention for children—Nederlandstalige vertaling.
Amsterdam: Harcourt Test Publishers, 2007.
Cohen MJ. Children’s memory scale. Paris: Les Editions du Centre
de Psychologie Appliquée, 1997.
Forrester G, Geffen G. Performance measures of 7- to 15-year-old
children on the auditory verbal learning test. Clin Neuropsychol
1991; 5: 345–59.
Rey A. L’examen clinique en psychologie. Paris: Presses
Universitaires de France, 1964.
Achenbach T. Child behavior checklist/4-18, Nederlandstalige
vertaling. Rotterdam: ASEBA, Erasmus MC-Sophia
Kinderziekenhuis, Afdeling Kinder-en jeugdpsychiatrie, 1991.
Roelants M, Hauspie R. Groeicurven vlaanderen 2004. Brussel:
Laboratorium voor Antropogenetica, Vrije Universiteit Brussel, 2004.
Dickinson DF. The normal ECG in childhood and adolescence.
Heart 2005; 91: 1626–30.
Koenker R. Quantile regression. New York: Cambridge University
Press, 2005.
Verrotti A, Spalice A, Ursitti F, et al. New trends in neuronal
migration disorders. Eur J Paediatr Neurol 2010; 14: 1–12.
Kaufman AS, Lichtenberger EO. Essentials of WISC-III and
WPPSI-R assessment. New York, NY: John Wiley and Sons, 2000.
Isen J. A meta-analytic assessment of Wechsler’s P>V sign in
antisocial populations. Clin Psychol Rev 2010; 30: 423–35.
Cardonick E, Iacobucci A. Use of chemotherapy during human
pregnancy. Lancet Oncol 2004; 5: 283–91.
Cardonick E, Usmani A, Ghaffar S. Perinatal outcomes of a
pregnancy complicated by cancer, including neonatal follow-up
after in utero exposure to chemotherapy: results of an international
registry. Am J Clin Oncol 2010; 33: 221–28.
Daneman R, Zhou L, Kebede AA, Barres BA. Pericytes are required
for blood–brain barrier integrity during embryogenesis. Nature 2010;
468: 562–66.
Saunders NR, Knott GW, Dziegielewska KM. Barriers in the
immature brain. Cell Mol Neurobiol 2000; 20: 29–40.
Virgintino D, Errede M, Girolamo F, et al. Fetal blood–brain barrier
P-glycoprotein contributes to brain protection during human
development. J Neuropathol Exp Neurol 2008; 67: 50–61.
Schinkel AH, Wagenaar E, Mol CA, van Deemter L. P-glycoprotein
in the blood-brain barrier of mice influences the brain penetration
and pharmacological activity of many drugs. J Clin Invest 1996;
97: 2517–24.
Van Calsteren K, Verbesselt R, Devlieger R, et al. Transplacental
transfer of paclitaxel, docetaxel, carboplatin, and trastuzumab in a
baboon model. Int J Gynecol Cancer 2010; 20: 1456–64.
Van Calsteren K, Verbesselt R, Beijnen J, et al. Transplacental transfer
of anthracyclines, vinblastine, and 4-hydroxy-cyclophosphamide in a
baboon model. Gynecol Oncol 2010; 119: 594–600.
Le Deley MC, Vassal G, Taibi A, Shamsaldin A, Leblanc T,
Hartmann O. High cumulative rate of secondary leukemia after
continuous etoposide treatment for solid tumors in children and
young adults. Pediatr Blood Cancer 2005; 45: 25–31.
Reynoso EE, Shepherd FA, Messner HA, Farquharson HA,
Garvey MB, Baker MA. Acute leukemia during pregnancy: the
Toronto Leukemia Study Group experience with long-term
follow-up of children exposed in utero to chemotherapeutic agents.
J Clin Oncol 1987; 5: 1098–106.
Mir O, Berveiller P, Ropert S, et al. Emerging therapeutic options
for breast cancer chemotherapy during pregnancy. Ann Oncol 2008;
19: 607–13.
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1
9
Articles
10
www.thelancet.com/oncology Published online February 10, 2012 DOI:10.1016/S1470-2045(11)70363-1