ArticlePDF Available

The role of birth weight on the causal pathway to child and adolescent ADHD symptomatology: A population-based twin differences longitudinal design

Authors:

Abstract and Figures

Background Available evidence points towards lower birth weight as a risk factor for the development of attention deficit/hyperactivity disorder (ADHD) symptoms. We probed the causal nature of this putative effect of birth weight on ADHD symptoms using the twin differences design, which accounts for genetic and shared environmental confounds. Method In a large population‐based twin sample – 3,499 monozygotic (MZ) and 6,698 dizygotic (DZ) pairs – parents, teachers or twins rated the twins’ ADHD symptoms at nine assessment waves (2–16 years). We implemented the twin differences design, which completely accounts for shared environmental and genetic confounding in MZ twins. We tested whether: (a) the lighter‐born twins had elevated ADHD symptoms compared to the heavier‐born twins, by regressing within‐pair differences of ADHD symptoms on within‐pair differences of birth weight among MZ twins; (b) the effect of birth weight on ADHD was moderated by gender, gestational age and low birth weight; (c) this effect changed with age at ADHD assessment using adapted latent growth curve models; and (d) results differed for inattention and hyperactivity/impulsivity. Results Birth weight significantly predicted ADHD symptoms from early childhood to late adolescence. The lighter‐born twin had more ADHD symptoms than the heavier‐born cotwin among MZ twins across assessment waves and raters. No moderation effect was detected. The magnitude of the effect of birth weight decreased significantly across time for hyperactivity/impulsivity, but the decrease failed to reach significance for inattention. Estimates for inattention were significantly larger than for hyperactivity/impulsivity at each time point, implying stronger effect of birth weight on inattention symptoms. Conclusions Our findings provide stringent evidence for environmental effect of lower birth weight on the causal pathway to elevated ADHD symptoms. Effect of birth weight persists across a 14‐year period from childhood into late adolescence, in particular for inattention symptoms.
Content may be subject to copyright.
The role of birth weight on the causal pathway to
child and adolescent ADHD symptomatology:
a population-based twin differences longitudinal
design
Kai Xiang Lim,
1
Chao-Yu Liu,
2,3
Tabea Schoeler,
2
Charlotte A.M. Cecil,
4
Edward D. Barker,
4
Essi Viding,
2
Corina U. Greven,
1,5,6,
* and Jean-Baptiste Pingault
1,2,
*
1
Social, Genetic, and Developmental Psychiatry Centre, Institute of Psychiatry, Psychology and Neuroscience, King’s
College London, London;
2
Department of Clinical, Educational and Health Psychology, Division of Psychology and
Language Sciences, University College London, London, UK;
3
School of Medicine, National Taiwan University, Taipei
City, Taiwan;
4
Department of Psychology, Institute of Psychiatry, Psychology & Neuroscience, King’s College London,
London, UK;
5
Department of Cognitive Neuroscience, Donders Institute for Brain, Cognition and Behaviour,
Radboud University Medical Centre, Nijmegen;
6
Karakter Child and Adolescent Psychiatry, University Center,
Nijmegen, The Netherlands
Background: Available evidence points towards lower birth weight as a risk factor for the development of attention
deficit/hyperactivity disorder (ADHD) symptoms. We probed the causal nature of this putative effect of birth weight
on ADHD symptoms using the twin differences design, which accounts for genetic and shared environmental
confounds. Method: In a large population-based twin sample 3,499 monozygotic (MZ) and 6,698 dizygotic (DZ)
pairs parents, teachers or twins rated the twins’ ADHD symptoms at nine assessment waves (216 years). We
implemented the twin differences design, which completely accounts for shared environmental and genetic
confounding in MZ twins. We tested whether: (a) the lighter-born twins had elevated ADHD symptoms compared to
the heavier-born twins, by regressing within-pair differences of ADHD symptoms on within-pair differences of birth
weight among MZ twins; (b) the effect of birth weight on ADHD was moderated by gender, gestational age and low
birth weight; (c) this effect changed with age at ADHD assessment using adapted latent growth curve models; and (d)
results differed for inattention and hyperactivity/impulsivity. Results: Birth weight significantly predicted ADHD
symptoms from early childhood to late adolescence. The lighter-born twin had more ADHD symptoms than the
heavier-born cotwin among MZ twins across assessment waves and raters. No moderation effect was detected. The
magnitude of the effect of birth weight decreased significantly across time for hyperactivity/impulsivity, but the
decrease failed to reach significance for inattention. Estimates for inattention were significantly larger than for
hyperactivity/impulsivity at each time point, implying stronger effect of birth weight on inattention symptoms.
Conclusions: Our findings provide stringent evidence for environmental effect of lower birth weight on the causal
pathway to elevated ADHD symptoms. Effect of birth weight persists across a 14-year period from childhood into late
adolescence, in particular for inattention symptoms. Keywords: Attention-deficit/hyperactivity disorder;
inattention; hyperactivity/impulsivity; birth weight; twin differences.
Introduction
Birth weight is an index of foetal growth and a
marker of the prenatal environment (Schlotz &
Phillips, 2009). Lower birth weight is associated with
a range of adverse mental health and behavioural
outcomes, such as lower intelligence quotient (IQ;
e.g. Breslau, 1995), autistic-like features (e.g. Ron-
ald, Happ!
e, Dworzynski, Bolton, & Plomin, 2010),
conduct problems (e.g. Wiles et al., 2006) and
attention-deficit/hyperactivity disorder (ADHD; e.g.
Schlotz, Jones, Godfrey, & Phillips, 2008). The
adverse outcomes associated with lower birth weight
can be viewed in light of the developmental origins of
health and disease (DOHaD) hypothesis (Barker,
1997, 1998). It proposes that changes in prenatal
environment during sensitive periods of organ
development can result in long-term structural or
physiological alterations, which subsequently
increase the risk of diseases such as psychiatric
disorders later in life (Barker, 1997, 1998).
Numerous epidemiological studies have shown
that lower birth weight is associated with ADHD
symptoms (Botting, Powls, Cooke, & Marlow, 1997;
Breslau et al., 1996; Mick, Biederman, Prince, Fis-
cher, & Faraone, 2002; Nigg & Breslau, 2007).
Despite this, the reported associations may not
reflect a direct causal relationship as they might be
confounded by genetic and shared environmental
factors. For example, ADHD and birth weight are
both heritable (76% and 40%, respectively; Claus-
son, Lichtenstein, & Cnattingius, 2000; Faraone
et al., 2005). Although previous studies found no
evidence of common genetic liability, residual genetic
confounding of the association between birth
weight and ADHD is still possible (Groen-Blokhuis,
Middeldorp, Van Beijsterveldt, & Boomsma, 2011;
*Share last authorship.
Conflict of interest statement: No conflicts declared.
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for Child and
Adolescent Mental Health.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
Journal of Child Psychology and Psychiatry **:* (2018), pp **–** doi:10.1111/jcpp.12949
PFI_12mmX178mm.pdf + eps format
Pettersson et al., 2015). In addition, shared environ-
mental factors associated with both birth weight and
ADHD, such as socioeconomic status (Blumenshine,
Egerter, Barclay, Cubbin, & Braveman, 2010; Fou-
lon et al., 2015) and prenatal smoking (Gal!
era et al.,
2011; Thapar et al., 2009) may also confound this
association. To further probe the nature of the
relationship between birth weight and ADHD symp-
toms, the twin differences design a genetically
informed design for causal inference can be imple-
mented (McGue, Osler, & Christensen, 2010; Pin-
gault et al., 2018). This design capitalizes on the
twins’ characteristics to control for shared environ-
mental and genetic confounding, partially in dizy-
gotic (DZ) twins and fully in monozygotic (MZ) twins.
Findings from this design can provide stringent
estimates of the role of birth weight on the causal
pathway to developing ADHD symptoms.
A number of twin differences studies in childhood
have demonstrated that lighter-born twins have
elevated ADHD symptoms compared to heavier-born
twins (Asbury, Dunn, & Plomin, 2006; Ficks, Lahey,
& Waldman, 2013; Groen-Blokhuis et al., 2011;
Hultman et al., 2007; Lehn et al., 2007; Pettersson
et al., 2015). However, these studies were limited in
several key ways. First, no twin differences study
used longitudinal methods to examine whether the
effect of birth weight on ADHD symptoms persisted in
the long-term (i.e. from childhood to adolescence).
The DOHaD hypothesis implies that this effect
should persist in the long-term (Barker, 1997,
1998). However, the effects of birth weight on adverse
developmental outcomes may attenuate with age. For
example, catch-up growth in childhood and adoles-
cence appears to alter the initial effects of lower birth
weight on general cognitive and psychological per-
formance in males (Lundgren, Cnattingius, Jonsson,
& Tuvemo, 2001), although this is not observed for
attention problems (Groen-Blokhuis et al., 2011). To
date, there is no longitudinal twin differences study
examining whether the effect of birth weight persists
from childhood to later stages of life.
A second key limitation is that no previous twin
differences study formally tested whether birth weight
may differentially influence the development of the
two ADHD symptom dimensions, i.e. inattention and
hyperactivity/impulsivity. A singleton study showed
that adolescents with low birth weight had signifi-
cantly higher inattention problems when compared
to controls but no difference in hyperactivity was
detected (Indredavik et al., 2004). This effect might
persist until young adulthood, as another singleton
study found that 20-year-old adults born with very low
birth weight had more inattention problems but not
hyperactivity problems compared to controls (Hack
et al., 2004). In their twin differences study, Pet-
tersson et al. (2015) suggested that lower birth weight
affects both ADHD symptom dimensions in child-
hood, but whether the magnitude of this effect differs
significantly across both dimensions remains
untested. Also, most of the twin differences studies
only used parents’ ratings to assess ADHD symptoms
(Ficks et al., 2013; Groen-Blokhuis et al., 2011;
Hultman et al., 2007; Pettersson et al., 2015). This
is potentially problematic as parental reports are
prone to contrast effect, that is parents tend to
highlight differences between twins (Thapar, Hervas,
& McGuffin, 1995).
Lastly, a recent meta-analysis systematically
investigated moderators of the association between
birth weight and ADHD (Momany, Kamradt, &
Nikolas, 2017), which were not examined in previous
twin differences studies. In the current study, we
tested whether the size of this association differed
according to gender, gestational age, and whether it
differed for low birth weight (<2,500 g) versus normal
birth weight participants. Although the meta-analy-
sis reported nonsignificant results for these three
moderators, we aimed to confirm this finding within
the twin differences design. This is the first large
population-based twin differences study to examine
the differential effect of birth weight on both ADHD
symptom dimensions, assessed by multiple infor-
mants across a 14-year period. We tested whether
the relationship between birth weight and ADHD
symptoms:
1. remained after controlling for genetic and shared
environmental confounds;
2. was moderated by gender, gestational age and low
birth weight;
3. persisted from childhood to adolescence; and
4. whether differential effect of birth weight on the
two ADHD symptom dimensions could be
detected.
Methods
Participants
Participants were drawn from the Twins Early Development
Study (TEDS), a large longitudinal study of twins born in
England and Wales between 1994 and 1996, which is repre-
sentative of the UK population as shown in Table S1 (Haworth,
Davis, & Plomin, 2013). Their zygosity was determined using a
parent-rated instrument, which is 95% as accurate as DNA
markers (Price et al., 2000). Analyses were conducted after
excluding twin pairs with extreme perinatal conditions, severe
medical conditions, uncertain zygosity, unknown gender,
incomplete birth weight record and no data on ADHD symp-
toms (see footnote of Table S2 for details about exclusion
criteria). The final sample included 10,197 twin pairs (6,698
DZ pairs, 3,499 MZ pairs, 51.2% females). The number of twin
pairs included in each analysis ranged from 3,176 pairs to
7,119 pairs (see Table 1), depending on age, ADHD scale and
informants. The present study included nine waves of assess-
ments when twins were between 2 and 16 years old. Written
informed consent was obtained from all participating families.
This study was approved by the Institute of Psychiatry, King’s
College London, Ethics Committee.
Measures
Birth weight. During first contact, parents reported each
twin’s birth weight in pounds or kilograms, which were then
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for
Child and Adolescent Mental Health.
2Kai Xiang Lim et al.
converted into grams for analyses. Twins’ median age was
1.6 years when birth weight data were collected.
ADHD symptoms. ADHD symptoms were assessed using
age-appropriate parent-reported questionnaires, which are
Behar’s Preschool Behaviour Questionnaire (BPBQ; Behar &
Stringfield, 1974) at ages 2, 3 and 4 years, the hyperactivity-
inattention subscale of Strength and Difficulties Questionnaire
(SDQ; Goodman, 1997) at ages 4, 7, 9, 12 and 16 years, and the
DSM-IV-based ADHD subscale of the Conners’ Parent Rating
Scale Revised (CPRS-R; Conners, 2001) at ages 8, 12, 14 and
16 years. There were also teacher-reported questionnaires
(SDQ at ages 7, 9 and 12 years) and self-reported questionnaires
(SDQ at ages 9, 12 and 16 years). All scales are standard ADHD
scales with adequate psychometric properties (see footnote of
Table S3 for additional details on individual scales).
Statistical analyses
All analyses were conducted using R Version 3.3.1 (R Core
Team, 2016) and its Structural Equation Modelling package
Lavaan Version 0.5-22 (Rosseel, 2012).
Twin differences analyses. Using the twin differences
design, we tested whether birth weight predicted ADHD
symptoms at each wave of assessment. In order to examine
the whole spectrum of birth weight differences, twin pairs were
included irrespective of within-twin pair differences in birth
weight.
Response rates were systematically lower for teacher and
self-reports than for parental reports at the same data
collection point. We therefore created two SDQ composite
scores across ages to maximize sample size (i.e. averaging
teacher-rated SDQ measures at ages 7, 9 and 12 years and
averaging self-report SDQ measures at ages 9, 12 and 16 years).
We did not include teacher and self-reported scales available at
only one data collection point. Prior to analysis for each ADHD
scale, twins with missing ADHD data on that particular scale
were excluded and all variables were standardized.
Two types of estimates were obtained: (a) the unadjusted
phenotypic estimates from the whole twin sample and (b) the
estimates from twin differences in MZ twins (MZ estimates). To
obtain unadjusted phenotypic estimates, nonindependence of
data within twin pairswere accounted for by allowing within-twin
pair correlations (Carlin,Gurrin, Sterne, Morley, & Dwyer,2005).
MZ estimates were obtained by conducting Ordinary Least
Square (OLS) regressions through origin, regressing within-twin
pair differences in ADHD symptoms on within-twin pair differ-
ences in birth weight (Carlin et al., 2005). MZ estimates are
standardized beta coefficients from these regressions. The esti-
mates from twin differences in DZ same-gender twins were also
obtained in a similar way for comparison (see Table S4).
To account for non-normality and nonindependence, robust
95% confidence intervals (CI) were obtained using bootstrap-
ping (10,000 repetitions) for all estimates. The moderating
effects of gender, gestational age and low birth weight
(<2,500 g) on these estimates were tested (see Tables S5S7).
Unadjusted phenotypic estimates are bivariate correlations
betweenbirth weightand ADHD symptoms without controllingfor
any potential genetic and shared environmental confounds. MZ
twins share 100% of their segregated genes and shared environ-
ment. Hence, MZ estimates are more robustly adjusted, account-
ing fully for both genetic and shared environmental influences.
Modelling developmental effect. Latent growth curve
modelling was used to investigate if the change in MZ
estimates was significant across development. In each MZ
twin pair, one twin was assigned as ‘heavier twin’ and
another as ‘lighter twin’. 171 MZ twin pairs with equal birth
Table 1 Phenotypic and MZ twin difference estimates of the relationship between birth weight and ADHD symptoms
Age Scale Phenotypic estimate, b(95% CI) MZ estimate, b(95% CI) Total N(MZ)
Parents’ ratings
2BPBQ !.086 (!.106, !.065) !.126 (!.178, !.074) 5,562 (1,910)
3BPBQ !.078 (!.097, !.057) !.190 (!.241, !.143) 5,423 (1,876)
4BPBQ !.073 (!.090, !.056) !.193 (!.242, !.145) 7,119 (2,445)
4SDQ !.075 (!.091, !.058) !.193 (!.245, !.139) 7,113 (2,445)
7SDQ !.066 (!.083, !.048) !.237 (!.290, !.184) 7,011 (2,524)
8CPRS total !.052 (!.071, !.032) !.115 (!.153, !.078) 6,112 (2,177)
8CPRS H/I !.049 (!.068, !.029) !.064 (!.096, !.031) 6,110 (2,177)
8CPRS IA !.046 (!.065, !.026) !.146 (!.190, !.101) 6,109 (2,177)
9SDQ !.074 (!.101, !.048) !.169 (!.221, !.114) 3,176 (1,176)
12 SDQ !.057 (!.077, !.036) !.157 (!.203, !.114) 5,458 (1,992)
12 CPRS total !.044 (!.063, !.023) !.095 (!.133, !.056) 5,463 (1,987)
12 CPRS H/I !.033 (!.053, !.013) !.043 (!.074, !.012) 5,461 (1,987)
12 CPRS IA !.045 (!.065, !.024) !.122 (!.173, !.077) 5,463 (1,986)
14 CPRS total !.030 (!.056, !.003) !.090 (!.140, !.053) 3,194 (1,232)
14 CPRS H/I !.026 (!.051, .000) !.047 (!.084, !.010) 3,189 (1,231)
14 CPRS IA !.027 (!.055, !.001) !.108 (!.174, !.059) 3,193 (1,232)
16 SDQ !.024 (!.046, .000) !.112 (!.163, !.057) 4,699 (1,705)
16 CPRS total !.021 (!.044, .001) !.097 (!.141, !.050) 4,706 (1,708)
16 CPRS H/I !.033 (!.056, !.011) !.038 (!.082, !.001) 4,704 (1,707)
16 CPRS IA !.006 (!.029, .017) !.120 (!.173, !.065) 4,705 (1,708)
Teachers’ ratings
Mean SDQ !.010 (!.028, .009) !.067 (!.107, !.028) 7,049 (2,523)
Self-report
Mean SDQ .016 (!.002, .034) !.121 (!.177, !.065) 6,783 (2,429)
N=number of twin pairs for each analysis. H/I =Hyperactivity/impulsivity. IA =inattention. BPBQ =Behar’s Preschool
Behaviour Questionnaire. SDQ =Strength and Difficulties Questionnaire. CPRS-R =Conners’ Parent Rating Scale - Revised.
Estimates in bold are significant. Teachers’ and self-report ratings were obtained based on the average ratings across different
ages. Note that the CI of the MZ estimate for CPRS-R H/I (16 years) is very close to zero and hence this finding should be treated
with caution.
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for
Child and Adolescent Mental Health.
Lower birth weight and ADHD symptoms 3
weight were excluded. The development of ADHD symptoms
in lighter versus heavier twins was then modelled from age 8
to age 16 years, using the CPRS-R, which provides a
comprehensive assessment of both ADHD dimensions based
on the DSM-IV, and has been used extensively in ADHD
research (Chang, Wang, & Tsai, 2016). Analyses were
conducted for the total ADHD symptoms and repeated for
the two symptom dimensions. The growth model included
linear and quadratic components to account for nonlinear
change. Missing data were handled using full information
maximum likelihood (FIML) method.
To examine the developmental trend, we investigated
whether:
1. the differences in ADHD symptoms between ‘lighter’ and
‘heavier’ twins were significant at age 8 and 16 years (i.e.
difference between the mean in lighter twins vs. the mean in
heavier twins, tested at each age); and
2. whether this difference in means between ‘lighter’ and
‘heavier’ twins was significantly larger at age 8 compared to
age 16 years (difference observed at age 8 years minus
difference at age 16 years). Computed estimates for (i) and
(ii) were bootstrapped 10,000 times to obtain 95% CI.
Differential effect on symptom dimensions. To test
whether the MZ estimates significantly differed depending on
ADHD symptom dimensions, further analyses were carried
out in MZ twins using the CPRS-R, which has nine items for
each symptom dimension. Two models were fitted for each of
the symptom dimensions assessed at each age. In the first
model, the MZ estimates for both symptom dimensions were
constrained to be the same. The second model did not
include equality constraints, thereby allowing both MZ
estimates to be different. The difference in goodness of fit
was then tested using the Satorra-Bentler Chi-square test. A
significant difference would indicate that the effect of birth
weight differs for the two dimensions.
Results
Twin differences analyses
Descriptive sample statistics regarding mean birth
weight, number of twins and gender proportion for
each ADHD scale are shown in Table S3. The mean
gestational age was 36.47 weeks for the whole sam-
ple and 36.21 weeks for MZ twins.
As shown in Table 1, most phenotypic estimates
were significant during childhood (e.g. at age 4 years,
the estimate for parent-rated SDQ was b=!.075,
95% CI: !.091, !.058), but became nonsignificant in
adolescence (e.g. at age 16 years, the estimate for
parent-rated SDQ was b=!.024, 95% CI: !.046,
.000). In contrast, despite being the most stringent
type of estimates, all MZ estimates were significant
from early childhood to adolescence with small effect
sizes ranging from b=!.038 (95% CI: !.082, !.001)
to b=!.237 (95% CI: !.290, !.184). These effect
sizes were not moderated by gender, gestational
age and low birth weight (see Tables S5S7). Comple-
mentary analyses showed that a difference in 1 kg in
birth weight corresponded to a difference of 0.42
symptom count for inattention and 0.17 symptom
count for hyperactivity/impulsivity at age 8 years
(see Table S8 for estimates at all ages). The effect of
birth weight on ADHD was not affected by twins with
very low birth weight (<1,500 g; see Table S9). Exclud-
ing the twins with extreme perinatal and severe
medical conditions also did not change the direction
and significance of the estimates (see Table S2).
Modelling developmental effect
The change in ADHD symptoms for heavier versus
lighter-born MZ twins is shown in Figure 1 for total
ADHD symptoms (see Figure S1 for inattention and
Figure S2 for hyperactivity/impulsivity).
At age 8 years, there was a significant differ-
ence of ADHD symptoms between heavier and
lighter-born twins. Lighter-born twins tended to
have higher total ADHD, inattention and hyper-
activity/impulsivity symptoms (see Table 2). At
age 16 years, lighter-born twins tended to have
higher total ADHD and inattention symptoms.
Differences between heavier and lighter-born
twins were larger at age 8 years than 16, consis-
tent with a lessening effect of birth weight over
time. This decrease in effect sizes was observed
for total and hyperactivity/impulsivity symptoms
but failed to reach significance for inattention
symptoms (see Table 2). The growth parameters
are presented in Table S10.
Differential effect on symptom dimensions
Figure 2 shows that MZ estimates for inattention
symptoms appear larger in magnitude than for
hyperactivity/impulsivity. Formal Satorra-Bentler
tests revealed that the unconstrained models fit the
data better than the constrained models from age
8 years to 16 years, suggesting that the stronger
effect of birth weight on inattention compared to
hyperactivity/impulsivity was significant and endur-
ing (see Table 3).
Discussion
In the current twin differences study, we report
four main findings. First, the epidemiological asso-
ciation observed between birth weight and ADHD
symptomatology was confirmed in stringent twin
differences analyses using MZ twins to control for
genetic and shared environmental confounding.
This points towards a plausible role of birth weight
on the causal pathway leading to the development
of ADHD symptoms. Second, this effect was not
moderated by gender, gestational age and low
birth weight. Third, the effect of birth weight on
ADHD symptoms persisted from childhood to ado-
lescence. Fourth, this effect was stronger and more
persistent for inattention compared to hyperactiv-
ity/impulsivity.
The present findings corroborate previous twin
studies as the effect of birth weight on ADHD
symptoms remained significant even in MZ analyses,
which control for all shared environmental and
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for
Child and Adolescent Mental Health.
4Kai Xiang Lim et al.
genetic confounding. Our findings add to the litera-
ture in that we used a large population-based
sample of twins, with consistent findings across
multiple ADHD scales and multiple informants. In
addition, we found that the effect of birth weight was
not moderated by gender, gestational age and low
birth weight, which confirms findings from a recent
meta-analysis (Momany et al., 2017). The small MZ
estimates found in this study are the rule rather than
exception in MZ differences analyses. Momany et al.
(2017) reported a small overall correlation between
birth weight and ADHD symptoms (r=!.15), with a
correlation of r=!.09 when considering population-
based studies only. This may reflect that the envi-
ronmental architecture underlying ADHD is just as
complex as their genetic architecture, which involves
small effects of many genetic variants (Demontis
et al., 2017).
Intriguingly, in the current study MZ estimates
were larger than phenotypic correlations between
birth weight and ADHD symptoms. This pattern of
findings replicates the results of a previous twin dif-
ferences study (Pettersson et al., 2015), suggesting
that our finding is not sample specific. Because the
twin differences design stringently controls for
genetic and shared environmental confounds, MZ
twin differences estimates are typically lower than
phenotypic estimates. Therefore, higher MZ esti-
mates may suggest a suppression effect, i.e. con-
founders hide the true effect of birth weight as a risk
factor for ADHD symptoms. Interestingly, the afore-
mentioned meta-analysis also reported a slightly
larger effect size (r=!.18) in covariate-adjusted
analyses compared to unadjusted analyses
(r=!.15, Momany et al., 2017). As the twin differ-
ences design accounts for unobserved confounders,
the level of adjustment is more stringent than in
classical epidemiological designs. It may thus not be
surprising that the differences between phenotypic
and MZ estimates appear larger in our study than
the difference between adjusted and unadjusted
estimates in the meta-analysis.
Longitudinal findings in the current study add to
current knowledge in that the effect of lower birth
weight on ADHD symptoms were persistent from age
2 to age 16 years, in particular for inattention
Figure 1 Predicted ADHD symptoms of MZ twins for CPRS-R from age 8 to 16 years
Table 2 Latent growth curve modelling parameters for total ADHD, inattention and hyperactivity/impulsivity symptoms
Predicted estimates (95% CI)
Heavier twins Lighter twins Within-twin Difference of ADHD symptoms
Total ADHD symptoms
Mean at age 8 years 10.80 (10.44, 11.18) 11.50 (11.11, 11.88) !.70 (!.91, !.48)
Mean at age 16 years 6.36 (6.04, 6.70) 6.75 (6.43, 7.10) !.40 (!.62, !.15)
Difference of means (8 vs. 16 years) –– !.30 (!.59, !.01)
Inattention symptoms
Mean at age 8 years 5.00 (4.81, 5.21) 5.44 (5.23, 5.65) !.44 (!.57, !.30)
Mean at age 16 years 3.78 (3.57, 3.99) 4.09 (3.88, 4.31) !.31 (!.48, !.15)
Difference of means (8 vs. 16 years) –– !.12 (!.31, .07)
Hyperactivity/impulsivity symptoms
Mean at age 8 years 5.79 (5.59, 6.00) 6.05 (5.84, 6.27) !.26 (!.36, !.15)
Mean at age 16 years 2.57 (2.41, 2.73) 2.65 (2.49, 2.81) !.08 (!.19, .03)
Difference of means (8 vs. 16 years) –– !.17 (!.32, !.04)
Estimates in bold are significant. Heavier MZ twins had a mean birth weight of 2,592 g, whereas lighter MZ twins had a mean birth
weight of 2,283 g.
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for
Child and Adolescent Mental Health.
Lower birth weight and ADHD symptoms 5
symptoms. These findings partially support the
DoHAD hypothesis (Barker, 1997, 1998), stipulating
that prenatal environment constitutes a long-term
health risk. Different mechanisms may explain our
findings. For example, prenatal ischemia hypoxia
(i.e. insufficient nutrients and oxygen supply in
utero) is a primary pathway to lower birth weight,
and it also produces lasting changes in neurodevel-
opmental functioning, which increases risk for
ADHD (Smith, Schmidt-Kastner, McGeary, Kac-
zorowski, & Knopik, 2016). In addition to showing
persisting effect of birth weight in adolescence for
total ADHD and inattention symptoms, we also
revealed ‘catch-up’ or developmental compensation
effects, whereby early effect of birth weight decrease
over time (significantly for total ADHD and hyperac-
tivity/impulsivity symptoms).
The stronger phenotypic association between birth
weight and inattention than hyperactivity/impulsiv-
ity has been reported in previous studies (Hack et al.,
2004; Indredavik et al., 2004) and was confirmed
here using the twin differences design. We found that
the effect of birth weight on inattention symptoms
persisted into adolescence, whereas the effect on
hyperactivity/impulsivity symptoms decreased,
which further supported the differential effect of birth
weight on the two symptom dimensions.
Limitations
Our findings point towards the importance of birth
weight a marker for foetal growthin the aetiology of
long-term ADHD symptoms. However, birth weight is
a complex risk factor, and further research is needed
to identify which aspects of foetal growth might
explain the findings (e.g. by identifying restricted
development in specific brain areas). In addition,
other prenatal unmeasured nonshared environmen-
tal factors associated with foetal growth, such as
positioning in the womb, placentation, or differences
in nutritional availability could also explain the
findings (Plomin, DeFries, Knopik, & Neiderhiser,
2013). Nonshared differences within pairs after birth
may be a direct consequence of within-twin pair
differences in birth weight (e.g. additional medical
care for the lighter twin). Such postnatal differences
may therefore lie on indirect paths from birth weight
to ADHD symptoms, either buffering or accentuating
the effect of birth weight, which warrants further
investigation. Another limitation is that the current
study included retrospective parental reports of birth
weight, which may be affected by recall bias. How-
ever, retrospective parental reports of birth weight
have been shown to be reliable up to 30 years
postbirth (Catov et al., 2006; Lumey, Stein, &
Ravelli, 1994).
Conclusion
Restricted foetal development, as indexed by lower
birth weight, may play a role in the causal path-
ways leading to the development of ADHD symp-
tomatology. Whereas this effect was found to
decrease across time for total ADHD and hyperac-
tivity/impulsivity symptoms, the decrease failed to
reach significance for inattention symptoms. In
addition, birth weight was found to influence
inattention symptoms more strongly than hyperac-
tivity/impulsivity symptoms across childhood and
adolescence.
Figure 2 Absolute standardized MZ estimates with 95% CI for CPRS-R total ADHD, hyperactivity/impulsivity (Hyp/Imp) and inattention
symptoms across ages 8, 12, 14 and 16 years. All estimates are absolute values (negatively signed estimates are presented in Table 1). The
larger the estimates, the greater the effect of birth weight on ADHD symptoms
Table 3 Satorra-Bentler test results of the differential effect of
birth weight on ADHD symptom dimensions
Age Model df v
2
v
2
(df ) difference p-value
8 Unconstrained 2 1.60 17.32 (1) <.001***
Constrained 3 24.14
12 Unconstrained 2 .02 12.74 (1) <.001***
Constrained 3 14.90
14 Unconstrained 2 2.40 4.44 (1) .035*
Constrained 3 7.84
16 Unconstrained 2 .69 8.84 (1) .003**
Constrained 3 12.86
df, degrees of freedom.
*p<.05; **p<.01; ***p<.001.
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for
Child and Adolescent Mental Health.
6Kai Xiang Lim et al.
Supporting information
Additional Supporting Information may be found online
in the supporting information section at the end of the
article:
Table S1. Sample characteristics.
Table S2. Phenotypic, DZ and MZ twin difference
estimates without excluding twins with exclusion criteria.
Table S3. Descriptive statistics for birth weight and each
of the ADHD scales.
Table S4. Phenotypic, DZ and MZ twin difference
estimates of the relationship between birth weight and
ADHD symptoms.
Table S5. Moderating effect of gender.
Table S6. Moderating effect of short gestational age
(<37 weeks).
Table S7. Moderating effect of low birth weight
(<2,500 g).
Table S8. Corresponding change in symptoms for change
in 1 kg.
Table S9. Phenotypic, DZ and MZ twin difference
estimates after excluding twins with very low birth weight
(<1,500 g).
Table S10. Descriptive statistics for parameters in latent
growth curve modelling.
Figure S1. Predicted inattention levels of monozygotic
twins for Conners’ Parent Rating Scale-Revised from age
8 to 16 years.
Figure S2. Predicted hyperactivity/impulsivity levels of
monozygotic twins for Conners’ Parent Rating Scale-
Revised from age 8 to 16 years.
Acknowledgement
The Twins Early Development Study is supported by
program grant MR/M021475/1 (previously
G0901245) and G1100559 from the UK Medical
Research Council (Dr Ronald). This research is funded
by grant ES/N001273/1 from the Economic and
Social Research Council (A.M.C.) and grant MQ16IP16
from MQ: Transforming Mental Health (J-B.P.). The
authors gratefully acknowledge the ongoing contribu-
tion of the participants in the Twins Early Develop-
ment Study and their families. The authors have
declared that they have no competing or potential
conflicts of interest.
Correspondence
Jean-Baptiste Pingault, Division of Psychology and
Language Sciences, University College London, 26
Bedford Way, London, WC1H 0DS, UK; Email: j.pin-
gault@ucl.ac.uk
Key points
Using a longitudinal twin differences design, we probed the causal nature of the observed association
between birth weight and ADHD symptoms.
Lower birth weight predicted higher ADHD symptoms over a 14-year period from childhood to adolescence.
The size of the effect of birth weight decreased significantly for total ADHD and hyperactivity/impulsivity
symptoms.
A stronger and more persistent effect of birth weight was found for inattention symptoms compared to
hyperactivity/impulsivity symptoms.
Additional research to further dissect the mechanisms explaining this environmentally driven relationship
between birth weight and ADHD symptoms are required.
References
Asbury, K., Dunn, J.F., & Plomin, R. (2006). Birthweight-discor-
dance and differences in early parenting relate to monozygotic
twin differences in behaviour problems and academic achieve-
ment at age 7. Developmental Science,9,2231.
Barker, D.J.P. (1997). Maternal nutrition, fetal nutrition, and
disease in later life. Nutrition,13, 807813.
Barker, D.J.P. (1998). Mothers, babies and health in later life.
(2nd edn). Edinburgh, UK: Churchill Livingstone.
Behar, L., & Stringfield, S. (1974). A behavior rating scale for
the preschool child. Developmental Psychology,10, 601
610.
Blumenshine, P., Egerter, S., Barclay, C.J., Cubbin, C., &
Braveman, P.A. (2010). Socioeconomic disparities in adverse
birth outcomes: A systematic review. American Journal of
Preventive Medicine,39 , 263272.
Botting, N., Powls, A., Cooke, R.W., & Marlow, N. (1997).
Attention deficit hyperactivity disorders and other psychi-
atric outcomes in very low birthweight children at 12 years.
Journal of Child Psychology and Psychiatry, and Allied
Disciplines,38, 931941.
Breslau, N. (1995). Psychiatric sequelae of low birth weight.
Epidemiologic Reviews,17 , 96106.
Breslau, N., Brown, G.G., DelDotto, J.E., Kumar, S., Ezhutha-
chan, S., Andreski, P., & Hufnagle, K.G. (1996). Psychiatric
sequelae of low birth weight at 6 years of age. Journal of
Abnormal Child Psychology,24, 385400.
Carlin, J.B., Gurrin, L.C., Sterne, J.A.C., Morley, R., &
Dwyer, T. (2005). Regression models for twin studies: A
critical review. International Journal of Epidemiology,34,
10891099.
Catov, J.M., Newman, A.B., Kelsey, S.F., Roberts, J.M., Sutton-
Tyrrell, K.C., Garcia, M., ... & Ness, R.B. (2006). Accuracy
and reliability of maternal recall of infant birth weight among
older women. Annals of Epidemiology,16, 429431.
Chang, L.-Y., Wang, M.-Y., & Tsai, P.-S. (2016). Diagnostic
accuracy of rating scales for attention-deficit/hyperactivity
disorder: A meta-analysis. Pediatrics. Advanced online pub-
lication. https://doi.org/10.1542/peds.2015-2749.
Clausson, B., Lichtenstein, P., & Cnattingius, S. (2000). Genetic
influence on birthweight and gestational length determined by
studies in offspring of twins. BJOG: An International Journal of
Obstetrics and Gynaecology,107 , 375381.
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for
Child and Adolescent Mental Health.
Lower birth weight and ADHD symptoms 7
Conners, C.K. (2001). CRS-R : Conners’ rating scales-revised :
Technical manual. North Tonawanda, NY: Multi-Health
Systems.
Demontis, D., Walters, R.K., Martin, J., Mattheisen, M., Als,
T.D., Agerbo, E., ... & Neale, B.M. (2017). Discovery of the
first genome-wide significant risk loci for ADHD. BioRxiv,
Available from: http://biorxiv.org/content/early/2017/06/
03/145581.abstract [last accessed 25 June 2018].
Faraone, S.V., Perlis, R.H., Doyle, A.E., Smoller, J.W., Goral-
nick, J.J., Holmgren, M.A., & Sklar, P. (2005). Molecular
genetics of attention-deficit/hyperactivity disorder. Biologi-
cal Psychiatry,57 , 13131323.
Ficks, C.A., Lahey, B.B., & Waldman, I.D. (2013). Does low
birth weight share common genetic or environmental risk
with childhood disruptive disorders? Journal of Abnormal
Psychology,122, 842853.
Foulon, S., Pingault, J.B., Larroque, B., Melchior, M., Falis-
sard, B., & C^
ot!
e, S.M. (2015). Developmental predictors of
inattention-hyperactivity from pregnancy to early childhood.
PLoS ONE,10,112.
Gal!
era, C., C^
ot!
e, S.M., Bouvard, M.P., Pingault, J.B., Melchior,
M., Michel, G., ...& Tremblay, R.E. (2011). Early risk factors
for hyperactivity-impulsivity and inattention trajectories
from age 17 months to 8 years. Archives of General Psychi-
atry,68, 12671275.
Goodman, R. (1997). The strengths and difficulties question-
naire: A research note. Journal of Child Psychology and
Psychiatry,38, 581586.
Groen-Blokhuis, M.M., Middeldorp, C.M., Van Beijsterveldt,
C.E.M., & Boomsma, D.I. (2011). Evidence for a causal
association of low birth weight and attention problems.
Journal of the American Academy of Child and Adolescent
Psychiatry,50, 12471254. e2.
Hack, M., Youngstrom, E.A., Cartar, L., Schluchter, M., Taylor,
H.G., Flannery, D.J., ... & Borawski, E. (2004). Behavioral
outcomes and evidence of psychopathology among very
low birth weight infants at age 20 years. Pediatrics,114,
932940.
Haworth, C.M.A., Davis, O.S.P., & Plomin, R. (2013). Twins
Early Development Study (TEDS): A genetically sensitive
investigation of cognitive and behavioral development from
childhood to young adulthood. Twin Research and Human
Genetics: The Official Journal of the International Society for
Twin Studies,16, 117125.
Hultman, C.M., Torr"
ang, A., Tuvblad, C., Cnattingius, S.,
Larsson, J.-O., & Lichtenstein, P. (2007). Birth weight and
attention-deficit/hyperactivity symptoms in childhood and
early adolescence. Journal of the American Academy of Child
and Adolescent Psychiatry,46, 370377.
Indredavik, M.S., Vik, T., Heyerdahl, S., Kulseng, S., Fayers,
P., & Brubakk, A.-M. (2004). Psychiatric symptoms and
disorders in adolescents with low birth weight. Archives of
Disease in Childhood, Fetal and Neonatal Edition,89 , F445
F450.
Lehn, H., Derks, E.M., Hudziak, J.J., Heutink, P., van
Beijsterveldt, T.C.E.M., & Boomsma, D.I. (2007). Attention
problems and attention-deficit/hyperactivity disorder in
discordant and concordant monozygotic twins: Evidence of
environmental mediators. Journal of the American Academy
of Child and Adolescent Psychiatry,46, 8391.
Lumey, L.H., Stein, A.D., & Ravelli, C.J. (1994). Maternal recall
of birthweights of adult children: Validation by hospital and
well baby clinic records. International Journal of Epidemiol-
ogy,23, 10061012.
Lundgren, E.M., Cnattingius, S., Jonsson, B., & Tuvemo, T.
(2001). Intellectual and psychological performance in males
born small for gestational age with and without catch-up
growth. Pediatric Research,50, 9196.
McGue, M., Osler, M., & Christensen, K. (2010). Causal
inference and observational research: The utility of twins.
Perspectives on Psychological Science: A Journal of the
Association for Psychological Science,5, 546556.
Mick, E., Biederman, J., Prince, J., Fischer, M.J., & Faraone,
S.V. (2002). Impact of low birth weight on attention-deficit
hyperactivity disorder. Journal of Developmental and Behav-
ioral Pediatrics,23, 1622.
Momany, A.M., Kamradt, J.M., & Nikolas, M.A. (2017). A meta-
analysis of the association between birth weight and atten-
tion deficit hyperactivity disorder. Journal of Abnormal Child
Psychology. Advanced online publication. https://doi.org/
10.1007/s10802-017-0371-9.
Nigg, J.T., & Breslau, N. (2007). Prenatal smoking exposure,
low birth weight, and disruptive behavior disorders. Journal
of the American Academy of Child and Adolescent Psychia-
try,46, 362369.
Pettersson, E., Sj
olander, A., Almqvist, C., Anckars
ater, H.,
D’Onofrio, B.M., Lichtenstein, P., & Larsson, H. (2015). Birth
weight as an independent predictor of ADHD symptoms: A
within-twin pair analysis. Journal of Child Psychology and
Psychiatry, and Allied Disciplines,56, 453459.
Pingault, J.-B., O’Reilly, P.F., Schoeler, T., Ploubidis, G.B.,
Rijsdijk, F., & Dudbridge, F. (2018). Using genetic data to
strengthen causal inference in observational research.
Nature Review Genetics. Advanced online publication.
https://doi.org/10.1038/s41576-018-0020-3.
Plomin, R., DeFries, J.C., Knopik, V.S., & Neiderhiser, J.
(2013). Behavioral genetics. New York: Worth.
Price, T.S., Freeman, B., Craig, I., Petrill, S.A., Ebersole, L., &
Plomin, R. (2000). Infant zygosity can be assigned by parental
report questionnaire data. Twin Research,3, 129133.
R Core Team (2016). R: A languageand environment for statistical
computing. Vienna, Austria: R Foundation for Statistical
Computing. Available from: https://www.r-project.org/
Ronald, A., Happ!
e, F., Dworzynski, K., Bolton, P., & Plomin, R.
(2010). Exploring the relation between prenatal and neona-
tal complications and later autistic-like features in a repre-
sentative community sample of twins. Child Development,
81, 166182.
Rosseel, Y. (2012). lavaan: An R Package for Structural
Equation Modeling. Journal of Statistical Software,48,136.
Schlotz, W., Jones, A., Godfrey, K.M., & Phillips, D.I.W. (2008).
Effortful control mediates associations of fetal growth with
hyperactivity and behavioural problems in 7- to 9-year-old
children. Journal of Child Psychology and Psychiatry and
Allied Disciplines,49 , 12281236.
Schlotz, W., & Phillips, D.I.W. (2009). Fetal origins of mental
health: Evidence and mechanisms. Brain, Behavior, and
Immunity,23, 905916.
Smith, T.F., Schmidt-Kastner, R., McGeary, J.E., Kaczorowski,
J.A., & Knopik, V.S. (2016). Pre- and perinatal ischemia-
hypoxia, the ischemia-hypoxia response pathway, and
ADHD risk. Behavior Genetics,46, 467477.
Thapar, A., Hervas, A., & McGuffin, P. (1995). Childhood
hyperactivity scores are highly heritable and show sibling
competition effects: Twin study evidence. Behavior Genetics,
25, 537544.
Thapar, A., Rice, F., Hay, D., Boivin, J., Langley, K., Van
Den Bree, M., ... & Harold, G. (2009). Prenatal smoking
might not cause attention-deficit/hyperactivity disorder:
Evidence from a novel design. Biological Psychiatry,66,
722727.
Wiles, N.J., Peters, T.J., Heron, J., Gunnell, D., Emond, A., &
Lewis, G. (2006). Fetal growth and childhood behavioral
problems: Results from the ALSPAC cohort. American Jour-
nal of Epidemiology,163, 829837.
Accepted for publication: 8 June 2018
©2018 The Authors. Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for
Child and Adolescent Mental Health.
8Kai Xiang Lim et al.
... In support of a stable and persistent association, a meta-analysis showed that age did not moderate the association [26] and crosssectional studies suggest attention deficits linked with low BW and SGA persist until adulthood [5,24,36,37]. Further, twin studies show that the lower BW twin continues to report higher ADHD symptoms up to the mid-teens [16,23]. In support of a weakening association over time, Hultman et al. [16] found the continuous association between BW and ADHD symptoms among twins was significant at 8-9 years but not at 13-14 years, and Lim, et al. [23] found BW had a decreasing association with ADHD symptoms from ages 8 to 16. ...
... Further, twin studies show that the lower BW twin continues to report higher ADHD symptoms up to the mid-teens [16,23]. In support of a weakening association over time, Hultman et al. [16] found the continuous association between BW and ADHD symptoms among twins was significant at 8-9 years but not at 13-14 years, and Lim, et al. [23] found BW had a decreasing association with ADHD symptoms from ages 8 to 16. The potential age-dependency of this association therefore requires further exploration. ...
... Generalized mixed models with a gamma distribution modelled the observed distribution of SDQ scores rather than dichotomizing scales to simplify analyses or violating assumptions of normality. Limitations of this study include its reliance on parent reports of both BW and child behaviour (the latter of which may be inaccurate for older adolescents) and that the SDQ does not differentiate between inattention and hyperactivity symptoms, which may have different relationships with birth weight [23]. Finally, the flexibility of our models was somewhat limited by only acknowledging linear time within interactions (i.e., time × sex, time × BW) despite the inclusion of a time 2 main effect. ...
Article
Full-text available
Low birth weight for one’s gestational age is associated with higher rates of child psychopathology, however, most studies assess psychopathology cross-sectionally. The effect of such foetal growth restriction appears to be strongest for attention problems in childhood, although adult studies have found associations with a range of outcomes, from depression to psychosis. We explore how associations between foetal growth and psychopathology change across age, and whether they vary by sex. We used a large nationally representative cohort of children from Ireland (N ~ 8000). Parents completed the Strengths and Difficulties Questionnaire (SDQ) at 3 time points (age 9, 13 and 17). Outcomes included a total problems scale and subscales measuring attention/hyperactivity, peer, conduct and emotional problems. Foetal growth had significant associations with all problem scales, even after controlling for sex, socioeconomic factors and parental mental health. The magnitude of these effects was small but relatively stable across ages 9–17. In males, foetal growth had the strongest associations with attention/hyperactivity and peer problems, whereas females showed more widespread associations with all four subscales. There was a trend for the association between foetal growth and emotional problems to increase with advancing age, approaching the borderline-abnormal threshold by age 17. Reduced foetal growth predicted persistently higher scores on all measured aspects of child and adolescent psychopathology. Associations with child attention/hyperactivity may generalize to a wider array of adult psychopathologies via adolescent-onset emotional problems. Future studies should explore potential age-dependent effects of foetal growth into the early 20s.
... Lower birth weights, even among those born at full term, are linked with an elevated risk of ADHD and academic difficulties (Class et al., 2014;Groen-Blokhuis et al., 2011;Hultman et al., 2007;Momany et al., 2018). Within-twin and within-sibling studies suggest that the association may be independent from genetics and other familial factors (Class et al., 2014;Ficks et al., 2013;Groen-Blokhuis et al., 2011;Hultman et al., 2007;Lim et al., 2018;Pettersson et al., 2015). Experimentally-induced fetal growth restriction in animals has also been shown to influence neurodevelopment and behavior (Lauritz et al., 2012;Meyer et al., 2014). ...
Article
Full-text available
The association between restricted fetal growth and symptoms of attention deficit/hyperactivity disorder (ADHD) in childhood is well-replicated and robust. However, fetal growth is determined by many prenatal factors and associations with mental health may be confounded by familial and social context. In this study, we sought to quantify the relative contributions of prenatal factors and familial confounds to the association between fetal growth and ADHD symptoms. Two independent cohorts were analyzed, the Adolescent Brain Cognitive Development study (ABCD; United States) and the Growing Up in Ireland (GUI) study. ADHD symptoms were measured by the Child Behavior Checklist (ABCD) and the Strengths & Difficulties questionnaire (GUI) at age 9–10. Using sequential regression models, we assessed the change-in-association between fetal growth and ADHD symptoms after controlling for sex, familial factors (socioeconomic/demographic factors & family psychiatric history) and prenatal factors (pregnancy complications & maternal substance-use during pregnancy). Converging findings from cohorts suggested that over a quarter of the association between fetal growth and ADHD symptoms is attributable to familial confounds. The degree to which the association was explained by prenatal factors differed by cohort—pregnancy complications explained a larger proportion of the effect in ABCD (7.9%) than GUI (2.7%), and maternal substance-use explained a larger proportion of the effect in GUI (22.7%) compared to ABCD (4.8%). Different explanations of the fetal growth-ADHD association across cohorts suggests cohort-specific, and potentially nationally-specific, risk factors for fetal growth and related neurodevelopmental outcomes. The evidence suggests early prevention of ADHD in Ireland should focus on minimizing maternal smoking during pregnancy. In the US, prevention and treatment of pregnancy complications are highlighted as viable targets for intervention.
... However, this model had a low resolution (i.e., the confidence intervals were extremely wide e.g., -37 to 477 for a causal path estimate of 0.66). We therefore opted to further constrain rE to 0 as this is commonly done in MZ twin difference models (e.g., Kohler et al. 2011;Lim et al. 2018) to improve resolution. All ACE components of the exposures and AE components of the outcomes in the final models were significant 4) PRS for insomnia, derived from a GWAS for insomnia symptoms in the UK Biobank (Hammerschlag et al. 2017). ...
Article
Full-text available
Previous genetically informed studies have uncovered likely causal relationships between mental health problems and self-harm but resulting causal estimates may be biased due to unmediated pleiotropy. By fitting Mendelian Randomization - Direction of Causation (MR-DoC) models that explicitly model pleiotropy, we investigated the effect of four quantitatively measured mental health problems - major depressive disorder (MDD), schizophrenia, attention-deficit hyperactivity disorder (ADHD), and insomnia, on non-suicidal self-harm (NSSH) and suicidal self-harm (SSH), separately. We used data of 12,723 twins (56.6% females) in the Twins Early Development Study. Besides substantial pleiotropy, we found effects from child-rated depressive symptoms to both NSSH ( β = 0.194, 95% CIs: 0.131, 0.257) and SSH ( β = 0.210, 95% CIs: 0.125, 0.295). Similarly, effects flowed from parent-rated depressive symptoms to NSSH ( β = 0.092, 95% CIs: 0.004, 0.181) and SSH ( β = 0.165, 95% CIs: 0.051, 0.281). We did not find evidence of aetiological difference between NSSH and SSH.
... Recently, studies employing advanced causal inference methods have helped to better separate associations due to confounding versus likely causal drivers. For example, studies have shown that for factors such as maternal smoking, effects appear to reflect mainly genetic or familial confounding, while for others, such as birth weight, effects are supported by causally informative designs [27,30]. However, even when causal effects are supported, risk factors tend to show small effects and be associated with many other outcomes besides ADHD. ...
Article
Full-text available
Attention-deficit/hyperactivity disorder (ADHD) is a common and debilitating neurodevelopmental disorder influenced by both genetic and environmental factors, typically identified in the school-age years but hypothesized to have developmental origins beginning in utero. To improve current strategies for prediction, prevention and treatment, a central challenge is to delineate how, at a molecular level, genetic and environmental influences jointly shape ADHD risk, phenotypic presentation, and developmental course. Epigenetic processes that regulate gene expression, such as DNA methylation, have emerged as a promising molecular system in the search for both biomarkers and mechanisms to address this challenge. In this Current Opinion, we discuss the relevance of epigenetics (specifically DNA methylation) for ADHD research and clinical practice, starting with the current state of knowledge, what challenges we have yet to overcome, and what the future may hold in terms of methylation-based applications for personalized medicine in ADHD. We conclude that the field of epigenetics and ADHD is promising but is still in its infancy, and the potential for transformative translational applications remains a distant goal. Nevertheless, rapid methodological advances, together with the rise of collaborative science and increased availability of high-quality, longitudinal data make this a thriving research area that in future may contribute to the development of new tools for improved prediction, management, and treatment of ADHD.
... We did observe that hypoxaemia during pregnancy is not the only factor that may lead to LBW, but it also produces lifelong adverse effects on neurodevelopmental function, which in turn increases the risk of ADHD. In addition to birth weight, different growth environments may promote the ability of children to adapt to different states, with subsequent influences on body indexes [43][44][45]. ...
Article
Full-text available
Background: Evidence of the association between children born with marginally low birth weight (MLBW) and obesity-related outcomes was controversial, and our study aimed to examine the role of attention deficit hyperactivity disorder (ADHD) and/or abnormal eating in these associations. Methods: A retrospective cohort study consisting of 677 Chinese children was conducted. Obesity-related outcomes (BMI, waist circumference, skin fold thickness, body fat, blood pressure, lipids and blood glucose), behaviour problems (ADHD and eating behaviour) and birth weight were collected. Mediation analyses were used to explore whether ADHD and/or abnormal eating was an intermediary factor in the MLBW-OB relationship. Results: The children with MLBW tended to have higher skin fold thickness, triglycerides, fasting blood glucose, waistline, body fat and abdominal obesity risks. Birth weight was negatively related to obesity-related outcomes, and the associations were mediated, partially, by increased risk of ADHD or abnormal eating behaviour after adjustment for BMI z-score. Furthermore, lower birth weight predicted higher waist circumference indirectly through emotional overeating caused by ADHD (β: -0.10; 95% CI: -0.19, -0.01). Conclusion: Our study suggests the hypothetical role of ADHD and abnormal eating as underlying mechanisms in the association between MBLW and obesity-related outcomes, which provides novel scientific evidence for interventions of childhood development.
... Furthermore, pre-and perinatal environmental influences have been related to ADHD such as low birth weight (Franz et al., 2018;Lim et al., 2018), smoking during pregnancy (Holz et al., 2014), low family income, and prematurity (Sciberras et al., 2017). Nevertheless, these findings are rather correlative in nature due to a lack of experimental studies with longitudinal designs in the research field, not (yet) allowing for conclusions on causality in these relationships (for details, see Thapar et al., 2013). ...
Thesis
The current thesis addresses neurobiological characteristics associated with Attention-Deficit/Hyperactivity Disorder and the transdiagnostic symptom dimension of impulsivity. A special focus was set on developmental effects in brain-behavior relationships for a more detailed understanding of the relevance of identified neuromarker-candidates within distinct maturational stages. Further, the current work took into account the highly relevant topic of electroencephalographic data quality as an essential prerequisite for validly translating neurophysiological study results into clinical practice. Due to its ease of administration, its high tolerability, and the rather low costs, electroencephalography was the focus of most of the current work. Therefore, three meta-analytical or empirical studies were conducted and reviewed within this thesis: first, our meta-analysis on event-related potential differences between patients with Attention-Deficit/Hyperactivity Disorder (n=1576) and healthy controls (n=1794) in childhood, adolescence, and young adulthood was presented. We identified relevant medium to large effect sizes between patients diagnosed with Attention-Deficit/Hyperactivity Disorder and healthy controls (−0.32<d<−0.57), mainly regarding later cognitive event-related potentials (P300, Contingent Negative Variation, and error-related positivity), indicating deficits in higher-order cognitive functioning (study 1). Second, results on electroencephalographic data quality were reported from our ESCAlife trial exploring Attention-Deficit/Hyperactivity Disorder patients in childhood (n=184), adolescent (n=39), and young adult (n=57) age compared to a small sample of healthy controls in school-age (n=25). We were able to show that participant-related characteristics, especially age and symptoms of hyperactivity/impulsivity, affect electroencephalographic data quality subsequently impacting on results obtained from spectral power analyses (study 2). And finally, we introduced our analyses on the large population-based IMAGEN-cohort of healthy adolescents (n=2034) assessed longitudinally into young adulthood (n=1383). We found that measures of different facets of impulsivity are related to brain activity in the pre-supplementary motor area and inferior frontal gyrus during inhibitory control during adolescence and in the ventral striatum during reward anticipation in young adulthood with distinct effects for different age groups. Associations between brain activity and impulsivity change substantially from adolescence to young adulthood, especially due to maturational changes on a neural level (study 3). Across all studies relevant developmental effects were identified. The studies presented here indicate that a variety of neurobiological characteristics and processes can be related to Attention-Deficit/Hyperactivity Disorder and impulsivity from either a categorical or a dimensional perspective and possibly represent promising biomarker-candidates. However, the current findings are in line with previous literature highlighting that no single biomarker might be sufficient to characterize aspects of healthy as well as deviant, clinically relevant human behavior. Future large-scale longitudinal studies using multidimensional assessment methods are needed to disentangle effects and further prove sensitivity and specificity of already identified neuromarker-candidates, thereby combining categorical and dimensional approaches. Additionally, data quality should be in the focus of future work to ensure a valid translation of neuroscientific study results into clinical practice. Specifically, developmental effects need to be explicitly taken into account. Further, future studies should address the predictive value of the neuromarker-candidates identified and prove their effectiveness as targets of neuromodulation-treatment within a personalized-medicine framework.
Article
DNA methylation (DNAm) is one of the most frequently studied epigenetic mechanisms facilitating the interplay of genomic and environmental factors, which can contribute to externalizing behaviours and related psychiatric disorders. Previous epigenome-wide association studies (EWAS) for externalizing behaviours have been limited in sample size, and, therefore, candidate genes and biomarkers with robust evidence are still lacking. We 1) performed a systematic literature review of EWAS of attention-deficit/hyperactivity disorder (ADHD)- and aggression-related behaviours conducted in peripheral tissue and cord blood and 2) combined the most strongly associated DNAm sites observed in individual studies (p<10⁻³) to identify candidate genes and biological systems for ADHD and aggressive behaviours. We observed enrichment for neuronal processes and neuronal cell marker genes for ADHD. Astrocyte and granulocytes cell markers among genes annotated to DNAm sites were relevant for both ADHD and aggression-related behaviours. Only 1% of the most significant epigenetic findings for ADHD/ADHD symptoms were likely to be directly explained by genetic factors involved in ADHD. Finally, we discuss how the field would greatly benefit from larger sample sizes and harmonization of assessment instruments.
Article
Randomized control trials are considered the pinnacle for causal inference. In many cases, however, randomization of participants in social work research studies is not feasible or ethical. This paper introduces the co-twin control design study as an alternative quasi-experimental design to provide evidence of causal mechanisms when randomization is not possible. This method maximizes the genetic and environmental sameness between twins who are discordant on an “exposure” to provide strong counterfactuals as approximations of causal effects. We describe how the co-twin control design can be used to infer causality and in what type of situations the design might be useful for social work researchers. Finally, we give advantages and limitations to the design, list a set of Twin Registries with data available after application, and provide an example code for data analysis.
Conference Paper
Background Attention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder characterised by age-inappropriate, disruptive and pervasive manifestations of inattention and/or hyperactivity/impulsivity. ADHD symptoms typically emerge in childhood and persist into later stages of life. ADHD also frequently co-occurs with a number of psychiatric disorders and medical conditions, thereby bringing a tremendous burden to affected individuals as well as society. In addition to symptom severity and chronicity, the development of ADHD also plays a determinant role in disease outcomes. However, few studies have systematically investigated different predictive factors and underlying aetiologies associated with the development of ADHD. Aims This thesis aims to examine patterns, influences and genetic underpinnings of the development of ADHD from childhood to adolescence. The first study investigates childhood factors that differentiate late-onset ADHD from childhood-onset ADHD and differences in adolescent outcomes. The second study examines genetic and environmental contributions underlying the effects of the development of inattention on academic performance. The third and the fourth studies investigate the developmental relationships between ADHD and BMI through triangulation of evidence from longitudinal statistical analyses and genetically informed causal inference approaches. Methods All of the studies adopt a development-sensitive design using data from the “Twin Early Development Study” (TEDS), a longitudinal cohort in the UK. A pluralistic statistical approach is employed for different study objectives. To strengthen causal inference, this thesis compares and contrasts findings from longitudinal statistical approaches and different genetically informed methods under a triangulation framework. Results Findings of this thesis suggest that 1) late-onset ADHD is more likely to be found in males and children who exhibit increased conduct problems and experience more childhood family adversity. Moreover, low socioeconomic status specifically predicts de novo late-onset ADHD, while additional factors predict subthreshold late-onset ADHD; 2) both the baseline level and the developmental course of inattention influence academic performance. Genetic contributions to the development of inattention also affect academic performance; 3) longitudinal statistical analyses identify unidirectional effects from ADHD symptoms to subsequent BMI, while genetic methods suggest a bidirectional causal relationship. Triangulation of evidence shows that multiple sources of confounding are involved in the relationships between ADHD and BMI, including unmeasured confounding and dynastic effects. Conclusions This thesis identifies specific childhood risk factors and genetic underpinnings associated with different developmental patterns of ADHD. Influences of the developmental course of ADHD on psychological and functional outcomes can be attributable to direct causal relationships, genetic and environmental confounding, or a combination of both. Altogether, these findings contribute to a more complete and systematic understanding of different developmental aspects of ADHD. To disentangle aetiological pathways between the development of ADHD and associated conditions, a pluralistic statistical approach to triangulate evidence regarding causal mechanisms is necessary.
Article
Background Low birth weight is associated with adult mental health, cognitive and socioeconomic problems. However, the causal nature of these associations remains difficult to establish owing to confounding. Aims To estimate the contribution of birth weight to adult mental health, cognitive and socioeconomic outcomes using two-sample Mendelian randomisation, an instrumental variable approach strengthening causal inference. Method We used 48 independent single-nucleotide polymorphisms as genetic instruments for birth weight (genome-wide association studies’ total sample: n = 264 498) and considered mental health (attention-deficit hyperactivity disorder (ADHD), autism spectrum disorder, bipolar disorder, major depressive disorder, obsessive–compulsive disorder, post-traumatic stress disorder (PTSD), schizophrenia, suicide attempt), cognitive (intelligence) and socioeconomic (educational attainment, income, social deprivation) outcomes. Results We found evidence for a contribution of birth weight to ADHD (OR for 1 s.d. unit decrease (~464 g) in birth weight, 1.29; 95% CI 1.03–1.62), PTSD (OR = 1.69; 95% CI 1.06–2.71) and suicide attempt (OR = 1.39; 95% CI 1.05–1.84), as well as for intelligence (β = −0.07; 95% CI −0.13 to −0.02) and socioeconomic outcomes, i.e. educational attainment (β = −0.05; 95% CI −0.09 to −0.01), income (β = −0.08; 95% CI −0.15 to −0.02) and social deprivation (β = 0.08; 95% CI 0.03–0.13). However, no evidence was found for a contribution of birth weight to the other examined mental health outcomes. Results were consistent across a wide range of sensitivity analyses. Conclusions These findings support the hypothesis that birth weight could be an important element on the causal pathway to mental health, cognitive and socioeconomic outcomes.
Article
Full-text available
https://rdcu.be/VT5g Causal inference is essential across the biomedical, behavioural and social sciences.By progressing from confounded statistical associations to evidence of causal relationships, causal inference can reveal complex pathways underlying traits and diseases and help to prioritize targets for intervention. Recent progress in genetic epidemiology - including statistical innovation, massive genotyped data sets and novel computational tools for deep data mining - has fostered the intense development of methods exploiting genetic data and relatedness to strengthen causal inference in observational research. In this Review, we describe how such genetically informed methods differ in their rationale, applicability and inherent limitations and outline how they should be integrated in the future to offer a rich causal inference toolbox.
Article
Full-text available
This review focuses on how measured pre- and perinatal environmental and (epi)genetic risk factors are interrelated and potentially influence one, of many, common developmental pathway towards ADHD. Consistent with the Developmental Origins of Health and Disease hypothesis, lower birth weight is associated with increased ADHD risk. Prenatal ischemia-hypoxia (insufficient blood and oxygen supply in utero) is a primary pathway to lower birth weight and produces neurodevelopmental risk for ADHD. To promote tissue survival in the context of ischemia-hypoxia, ischemia-hypoxia response (IHR) pathway gene expression is altered in the developing brain and peripheral tissues. Although altered IHR gene expression is adaptive in the context of ischemia-hypoxia, lasting IHR epigenetic modifications may lead to increased ADHD risk. Taken together, IHR genetic vulnerability to ischemia-hypoxia and IHR epigenetic alterations following prenatal ischemia-hypoxia may result in neurodevelopmental vulnerability for ADHD. Limitations of the extant literature and future directions for genetically-informed research are discussed.
Article
Full-text available
Objective: The objective of the study was to characterize the developmental sequence of pre- and postnatal risk factors for inattention-hyperactivity symptoms in preschoolers. Materials and methods: Longitudinal data came from a French population based birth cohort study (EDEN; N = 1311 mother-child pairs followed from the pregnancy onwards). Inattention-hyperactivity symptoms were assessed with the Strengths and Difficulties Questionnaire when participating children were 3 years of age. Potential risk factors were classified in four domains (fetal exposures and child somatic characteristics, child temperament, child neurodevelopmental status, psychosocial environment) and four periods (before pregnancy, prenatal/birth, infancy, toddlerhood). Their role as potential moderator or mediator was tested with path analysis to determine the developmental sequence. Results: A low family socioeconomic status before pregnancy was the main environmental risk factor for inattention-hyperactivity symptoms at 3 years, and its effect occurred via two pathways. The first was a risk pathway, where lower SES was associated with higher maternal depression and anxiety during pregnancy; then to higher maternal and child distress and dysregulation in infancy; and in turn to higher levels of inattention-hyperactivity at 3 years. The second was a protective pathway, where higher SES was associated with longer duration of breastfeeding during infancy; then to better child neurodevelopmental status in toddlerhood; and in turn to lower levels of inattention-hyperactivity at 3 years. Discussion: This study identified psychosocial factors at several developmental periods that represent potential targets for preventing the emergence of inattention-hyperactivity symptoms in early childhood.
Article
Full-text available
Structural equation modeling (SEM) is a vast field and widely used by many applied researchers in the social and behavioral sciences. Over the years, many software pack-ages for structural equation modeling have been developed, both free and commercial. However, perhaps the best state-of-the-art software packages in this field are still closed-source and/or commercial. The R package lavaan has been developed to provide applied researchers, teachers, and statisticians, a free, fully open-source, but commercial-quality package for latent variable modeling. This paper explains the aims behind the develop-ment of the package, gives an overview of its most important features, and provides some examples to illustrate how lavaan works in practice.
Article
A large body of work has investigated the association between birth weight and ADHD and has resulted in mixed findings with regard to the direction and magnitude of this association. Despite the vast amount of research on this topic, a comprehensive and systematic quantification of the association between birth weight and ADHD has yet to be undertaken. A meta-analysis of 88 unique studies (N = 4,645,482) was conducted to quantify the overall effect size of birth weight on ADHD symptoms. Several variables were examined as moderators that may contribute to systematic variation in effect sizes. Overall, birth weight was found to have a small, but significant, association with ADHD symptoms such that individuals born at lower birth weights manifested greater symptoms of ADHD (r = −0.15). Sample type, mean birth weight of the sample, geographic region, the informant of ADHD symptoms, ADHD symptom measurement method, and race were all found to contribute significantly to heterogeneity in effect sizes. Notably, several early life risk factors previously found to be associated with both ADHD and birth weight, gestational age and prenatal smoking exposure, were not found to contribute to heterogeneity in effect sizes. The findings of the current analyses align with the growing recognition that early life adversity contributes to neurodevelopmental difficulties, and the findings highlight the importance of a better understanding of the mechanisms underlying the association between early life risk factors and adverse neurodevelopmental sequela, such as that observed in ADHD.
Article
Context: The Child Behavior Checklist-Attention Problem (CBCL-AP) scale and Conners Rating Scale-Revised (CRS-R) are commonly used behavioral rating scales for diagnosing attention-deficit/hyperactivity disorder (ADHD) in children and adolescents. Objective: To evaluate and compare the diagnostic performance of CBCL-AP and CRS-R in diagnosing ADHD in children and adolescents. Data sources: PubMed, Ovid Medline, and other relevant electronic databases were searched for articles published up to May 2015. Study selection: We included studies evaluating the diagnostic performance of either CBCL-AP scale or CRS-R for diagnosing ADHD in pediatric populations in comparison with a defined reference standard. Data extraction: Bivariate random effects models were used for pooling and comparing diagnostic performance. Results: We identified and evaluated 14 and 11 articles on CBCL-AP and CRS-R, respectively. The results revealed pooled sensitivities of 0.77, 0.75, 0.72, and 0.83 and pooled specificities of 0.73, 0.75, 0.84, and 0.84 for CBCL-AP, Conners Parent Rating Scale-Revised, Conners Teacher Rating Scale-Revised, and Conners Abbreviated Symptom Questionnaire (ASQ), respectively. No difference was observed in the diagnostic performance of the various scales. Study location, age of participants, and percentage of female participants explained the heterogeneity in the specificity of the CBCL-AP. Conclusions: CBCL-AP and CRS-R both yielded moderate sensitivity and specificity in diagnosing ADHD. According to the comparable diagnostic performance of all examined scales, ASQ may be the most effective diagnostic tool in assessing ADHD because of its brevity and high diagnostic accuracy. CBCL is recommended for more comprehensive assessments.
Article
Background Studies have found an association between low birth weight and ADHD, but the nature of this relation is unclear. First, it is uncertain whether birth weight is associated with both of the ADHD dimensions, inattentiveness and hyperactivity-impulsivity. Second, it remains uncertain whether the association between birth weight and ADHD symptom severity is confounded by familial factors.Method Parents of all Swedish 9- and 12-year-old twins born between 1992 and 2000 were interviewed for DSM-IV inattentive and hyperactive-impulsive ADHD symptoms by the Autism – Tics, AD/HD and other Comorbidities (A-TAC) inventory (N = 21,775 twins). Birth weight was collected prospectively through the Medical Birth Registry. We used a within-twin pair design to control for genetic and shared environmental factors.ResultsReduced birth weight was significantly associated with a mean increase in total ADHD (β = −.42; 95% CI: −.53, −.30), inattentive (β = −.26; 95% CI: −.33, −.19), and hyperactive-impulsive (β = −.16; 95% CI: −.22, −.10) symptom severity. These results imply that a change of one kilogram of birth weight corresponded to parents rating their child nearly one unit higher (going from “no” to “yes, to some extent” on a given symptom) on the total ADHD scale. These associations remained within pairs of MZ and DZ twins, and were also present when restricting the analyses to full term births.Conclusions There is an independent association between low birth weight and all forms of ADHD symptoms, even after controlling for all environmental and genetic confounds shared within twin pairs. These results indicate that fetal growth restriction (as reflected in birth weight differences within twin pairs) and/or the environmental factors which influence it is in the casual pathway leading to ADHD.
Article
Although advances in neonatal care over the past century have resulted in increased rates of survival among at-risk births, including infants with low birth weight, we have much to learn about psychological outcomes in this population. In particular, despite growing evidence that low birth weight may be associated with an increased risk for Attention-Deficit/Hyperactive Disorder (ADHD) symptoms in childhood, few studies have examined birth weight as a risk factor for disruptive disorders that commonly co-occur with ADHD. In addition, the etiology of the relation between birth weight and these disorders is unknown. The current investigation aimed to better understand these associations in the context of potentially confounding genetic and environmental influences by examining phenotypic associations between birth weight and disruptive disorder symptoms both between families and within families in two independent twin samples (Sample 1: N = 1,676 individuals; Sample 2: N = 4,038 individuals). We found negative associations between birth weight and inattentive, hyperactive-impulsive, and broad externalizing symptoms in both samples. Nonetheless, the overall magnitude of these associations was very small, contributing to less than 1% of the variance in these symptom dimensions. Within-family associations between birth weight and disruptive disorder symptoms did not differ for monozygotic and dizygotic twin pairs, suggesting that nonshared environmental influences rather than common genetic influences are responsible for these associations. Overall, the consistent albeit weak associations between birth weight and disruptive disorder symptoms suggest that low birth weight may not represent a major risk factor in the development of these symptoms. (PsycINFO Database Record (c) 2013 APA, all rights reserved).