Content uploaded by Kenneth J. Zucker
Author content
All content in this area was uploaded by Kenneth J. Zucker on Feb 16, 2016
Content may be subject to copyright.
ORIGINAL PAPER
Autism Spectrum Disorder Risk Factors and Autistic Traits
in Gender Dysphoric Children
Doug P. VanderLaan •Jonathan H. Leef •
Hayley Wood •S. Kathleen Hughes •
Kenneth J. Zucker
Published online: 11 December 2014
ÓSpringer Science+Business Media New York 2014
Abstract Gender dysphoria (GD) and autism spectrum
disorder (ASD) are associated. In 49 GD children (40 natal
males), we examined ASD risk factors (i.e., birth weight,
parental age, sibling sex ratio) in relation to autistic traits.
Data were gathered on autistic traits, birth weight, parents’
ages at birth, sibling sex ratio, gender nonconformity, age,
maternal depression, general behavioral and emotional
problems, and IQ. High birth weight was associated with
both high gender nonconformity and autistic traits among
GD children. Developmental processes associated with
high birth weight are, therefore, likely to underlie the GD–
ASD link either directly or indirectly. The present study is
the first to provide quantitative data bearing on possible
mechanisms that lead GD and ASD to co-occur.
Keywords Gender dysphoria Autism spectrum
disorder Birth weight Parental age Sibling sex ratio
Introduction
Gender dysphoria (GD) and autism spectrum disorder (ASD)
are rare conditions. In the fifth edition of the Diagnostic and
Statistical Manual of Mental Disorders (DSM-5) (American
Psychiatric Association 2013), GD is characterized by dis-
tress that accompanies the incongruence between one’s
experienced or expressed gender and one’s assigned gender.
ASD is characterized by persistent deficits in social com-
munication and social interaction as well as restricted and
repetitive behavior, interests, or activities. Both conditions
show low population prevalence rates with between 1 in
10,000 and 1 in 50,000 individuals exhibiting GD (Zucker
and Lawrence 2009) and between 1 in 50 and 1 in 500
individuals exhibiting ASD (Blumberg et al. 2013; Fom-
bonne 2005). Also, both conditions show biased male:
female sex ratios ranging from approximately 2:1 to 4:1
(American Psychiatric Association 2013; Blumberg et al.
2013; Fombonne 2005; Zucker and Lawrence 2009). Despite
being rare conditions characterized by distinct sets of diag-
nostic criteria, it is not uncommon for GD and ASD to co-
occur.
Most literature regarding the link between GD and ASD
consists of clinical case reports (e.g., Lande
´n and Rasmussen
1997; Parkinson 2014; Williams et al. 1996); however, some
quantitative data have begun to emerge. In one study, 6.4 %
(7 of 108) of children and 9.4 % (9 of 96) of adolescents
referred for GD were classified as having ASD based on the
Diagnostic Interview for Social and Communication Disor-
ders (de Vries et al. 2010). Similarly, a study of adults with
GD found that 5.5 % (5 of 91) showed traits consistent with
an ASD diagnosis (Pasterski et al. 2014). Both studies con-
cluded that these prevalence rates were significantly higher
than estimated population prevalence rates for ASD.
In addition, two studies hav e examined traits of ASD among
individuals with GD relative to comparison groups. Using item
responses on the Child Behavior Checklist (CBCL), Vander-
Laan et al. (2014b) examined intense/obsessional interests and
repetitive behaviors, each of which are related to DSM-5 Point
B criteria for ASD (American Psychiatric Association 2013).
D. P. VanderLaan (&)J. H. Leef H. Wood
S. K. Hughes K. J. Zucker
Gender Identity Service, Child, Youth, and Family Services,
Beamish Family Wing, Centre for Addiction and Mental Health,
80 Workman Way, 5th Floor, Toronto, ON M6J 1H4, Canada
e-mail: doug.vanderlaan@camh.ca
K. J. Zucker
Department of Psychiatry, University of Toronto, Toronto, ON,
Canada
123
J Autism Dev Disord (2015) 45:1742–1750
DOI 10.1007/s10803-014-2331-3
In a sample of 534 children clinically referred for GD, their
siblings, and the CBCL clinic-referred and non-referred stan-
dardization samples, GD children showed elevated intense/
obsessional interests relative to all other groups and elevated
repetitive behaviors relative to the sibling and non-referred
samples. In an adult sample, Jones et al. (2012) demonstrated
that female-to-male transsexuals had more traits of ASD on a
self-report measure compared to controls. These studies are,
therefore, also consistent with the notion that the prevalence of
ASD, or traits thereof, is elevated among individuals with GD.
Given this elevation, it is possible that traits of ASD
contribute toward cross-gender behavior and identity
among a subset of children clinically referred for GD. An
alternative hypothesis is that some variable related to ASD
might also underlie cross-gender behavior and identity
albeit via a different mechanism. Some research has been
in line with these hypotheses. In one study, girls with ASD
showed less female-typical play behavior, although a
similar tendency toward sex-atypical play was not found
among boys with ASD (Knickmeyer et al. 2008). In
another study, Strang et al. (2014) compared cross-sex
wishes—as reported on a single CBCL item—in a sample
of 147 children referred for ASD to a non-clinical control
sample and the CBCL non-referred standardization sample.
By maternal report, ASD children were significantly more
likely to exhibit such wishes.
Yet, even though ASD children were more likely to
show cross-gender characteristics in these studies, it is
unclear whether the play behavior or cross-sex wishes
examined were indicative of the extreme and persistent
gender nonconformity exhibited by children clinically
referred for GD. Data suggesting that cross-gender
behavior and identity are associated with early indicators of
ASD among some subset of children clinically referred for
GD would, therefore, provide a more convincing demon-
stration that these hypotheses are tenable. One viable
approach for doing so is to examine traits of ASD and
gender nonconformity in relation to ASD risk factors that
are present early in development among children who are
clinically referred for GD.
In a sample of children clinically referred for GD, the
present study examined maternally reported gender non-
conformity and autistic traits in relation to three ASD risk
factors: birth weight, parental age at birth, and sibling sex
ratio. Regarding birth weight, deviations from an average
size at birth (i.e., small size or large size) are associated
with an elevated risk of ASD, although the precise
mechanisms responsible are unclear (Abel et al. 2013).
Advanced parental age at birth is another risk factor for
ASD (Lundstro
¨m et al. 2010; Parner et al. 2012; Sandin
et al. 2012), possibly because older parents are more
likely to have accumulated genetic mutations or encoun-
tered toxic substances that affect gene expression, thus
leading to ASD in offspring (Kondrashov 2012; Sandin
et al. 2012). A testosterone-rich prenatal environment is a
third risk factor for ASD (Knickmeyer et al. 2006) and a
high male-to-female sibling sex ratio, which is a putative
marker of elevated testosterone exposure in utero, is
associated with ASD in children (Mouridsen et al. 2010).
If any of these ASD risk factors are associated with
autistic traits and gender nonconformity among children
clinically referred for GD, then that would provide evi-
dence consistent with hypotheses arguing that ASD has a
direct or indirect association with the emergence of GD in
children.
Method
Participants
This study utilized patient data from a chart review approved
by a hospital research ethics board. Beginning in September
2009, information on autistic traits among children less than
or equal to 12 years of age were systematically gathered by
our specialty Gender Identity Service using a psychometri-
cally validated parent-report instrument. Data were col-
lected from 47 consecutive patients at the outset of clinical
assessment for GD as well as 13 outpatients who had been
assessed for GD previously and were continuing to attend
the clinic for therapeutic services. Of these 60 cases, 11
were excluded for various reasons: the required maternal
reports of all study variables were not completed in two
cases; two cases exhibited a co-occurring disorder of sex
development; one case was a twin and another was a
triplet, which may have impacted birth weight; in four
cases, the patients were adopted, in foster care, or under the
guardianship of a child protection agency and information
was not available on birth weight, biological parents’ ages
at birth, or biological sibling sex ratio; and birth weight
information was missing for one case. Of the 49 remaining
cases (40 natal males, 9 natal females) that were retained
for analysis, the mean (SD) age at the time that the study
measures were completed was 7.19 (2.71) years. Data were
collected from 12 cases (11 natal males, 1 natal female)
while they were outpatients and 37 cases (29 natal males, 8
natal females) at the time of the initial assessment. Whether
cases were outpatients or not at the time the study measures
were completed did not vary by sex based on Fisher’s exact
test, p=.42.
Measures
For all parent-report measures, the availability of maternal
reports was more consistent across cases compared to
J Autism Dev Disord (2015) 45:1742–1750 1743
123
paternal reports. As such, the current study focused on
maternal reports with the exception of one measure for one
case (noted below). Paternal reports were used to assess
inter-rater reliability wherever appropriate.
Autistic traits were assessed using the Social Respon-
siveness Scale (SRS; Constantino and Gruber 2005). The
SRS is a 65-item parent-report questionnaire with a one-
factor solution that captures social awareness, social cog-
nition, social communication, social motivation, and
repetitive behaviors. It is a well-validated instrument in
that it distinguishes between samples of children with
ASD, non-ASD clinical controls, and healthy controls. The
SRS also distinguishes among subpopulations of ASD
children, ranging from those with mild, high-functioning
ASD to severe autism. SRS Tscores C60 are defined as
being in the clinical range and indicate the presence of
autistic traits that are likely to create daily social interfer-
ence. Whether a case was in the clinical range on the SRS
based on maternal report was, therefore, the main outcome
measure in the present study. For the present sample,
maternal and paternal reports on the SRS were available for
28 cases and their Tscores showed significant agreement,
r=.62, df =26, p\.001.
Gender nonconformity was assessed using the Gender
Identity Questionnaire for Children (GIQC; Johnson et al.
2004). The GIQC is a 14-item questionnaire in which parents
rate their child’s preferences in domains such as the gender of
preferred playmates, fantasy role-playing or dress-up play,
preferred activities and toys, wishes to be the opposite sex, and
feelings about sexual anatomy. The rating scale ranges from 1
(stereotypically opposite-sex) to 5 (stereotypically same-sex),
but to ease data interpretation the present study reversed the
scores such that higher scores reflected elevated gender non-
conformity. The GIQC has a one-factor solution and, thus,
mean ratings were used to create a gender nonconformity
score. This measure provides good discriminant validity
between GD versus control children and also distinguishes the
magnitude of gender nonconformity within samples of chil-
dren clinically referred for GD. Maternal and paternal reports
on the GIQC were available for 36 cases and their scores
showed significant agreement, r=.56, df =34, p\.001.
(Note: a maternal report for the GIQC was not available for
one case and the paternal report was substituted.)
With respect to data on ASD risk factors, information on
birth weight was available from hospital birth records for 32
cases. For the remaining 17 cases, parent-reported (maternal
or paternal) birth weight was used. Parents reported birth
weight in pounds and ounces or in grams. All birth weights
were converted to kilograms to three decimal places. Hos-
pital-record and parent-reported birth weight information
were available for 25 cases and showed significant agree-
ment, r=.96, df =23, p\.001. This extremely high
correlation, which is consistent with previous studies of the
agreement between hospital records and parent-reported
birth weight (e.g., Blanchard et al. 2002; Gayle et al. 1988;
Gofin et al. 2000; O’Sullivan et al. 2000; Tomeo et al. 1999),
indicates that the parent-reported birth weight used in the
present study is reliable.
Parents’ ages were routinely recorded as part of family
background information gathering during clinical assess-
ment. Maternal and paternal ages at the time of the child’s
birth were calculated by subtracting the child’s age from
each parent’s age at the time of the assessment. Paternal
ages were missing for three cases in which the father’s
identity was unknown.
Information on the numbers and ages of siblings were also
routinely recorded as part of family background information
gatheringduring clinical assessment. Becausesibling sex ratio
is a putativeproxy of a testosterone-rich prenatal environment
(Mouridsen et al. 2010), only siblings who shared the same
mother with the patient were considered. Thus, full- and
maternal half-siblings were considered in the calculation of
the sibling sex ratio whereas paternal half-, step-, and foster-
siblings were ignored. The sibling sex ratio was calculated as:
(number of male siblings ?.5)/(total number of sib-
lings ?1). Thus, lower and higher values reflected female-
and male-biased sibling sex ratios, respectively. Also, pro-
bands without siblings were retained and represented as
having equal proportions of male and female siblings.
A number of control variables were also considered.
Given the relevance of age to social skills development, the
effect of the patient’s age (in years to two decimal places) at
the time the SRS was completed was evaluated. IQ scores
based on the Wechsler Intelligence Test for Children, Fourth
Edition (Wechsler 2003) were included to control for the
effect of general intelligence on social skills. (Note: IQ
testing was not completed for one case and the sample mean
IQ of 111 was imputed to retain the case without biasing
effects toward statistical significance.) To control for forms
of psychopathology other than ASD (e.g., depression,
anxiety) that might also impact social functioning, maternal
reports on the CBCL were used. The CBCL is a gold-stan-
dard measure for general behavioral and emotional problems
in children (Achenbach 1991). Maternal and paternal reports
on the CBCL were available for 38 cases and their Tscores
showed significant agreement, r=.50, df =36, p=.001.
Maternal depression has also been associated with SRS
scores (Bennett et al. 2012) and maternal depression was,
therefore, taken into account using the mothers’ self-reported
Symptom Checklist-90-Revised Tscore on the Depression
factor (Derogatis 1994).
Statistical Analyses
All statistical analyses were conducted using SPSS Version
20. Patients were divided into those who did versus did not
1744 J Autism Dev Disord (2015) 45:1742–1750
123
have clinically significant autistic traits as indicated by the
presence versus absence of a clinical range score on the SRS.
These two groups were initially compared on the control and
focal variables using independent ttests. Also, zero-order
correlations among the continuous variables were evaluated
across the entire sample. These initial analyses identified
relevant variables to consider in subsequent multiple logistic
regression analyses comparing the two groups. The regres-
sion analyses assessed whether any focal variables that dis-
tinguished the groups based on the ttests still did so after the
control variables were taken into account. The a priori pre-
diction that ASD risk factors, gender nonconformity, and
autistic traits would overlap was evaluated directly by
examining the interaction between ASD risk factors and
gender nonconformity toward the prediction of clinical-
range autistic traits. A conventional alpha value of .05 was
used for making decisions about statistical significance.
It is important to note that none of the patients in our
sample exhibited markedly low birth weight (the lowest
birth weight in the current sample was 2.27 kg). Thus, even
though both small and large size at birth are associated with
increased ASD risk (Abel et al. 2013), it was not possible
to evaluate whether both low and high birth weight were
associated with increased odds of exhibiting clinically
significant autistic traits in this sample. Rather, the analyses
presented below examined the linear effect of high birth
weight on the presence of autistic traits.
Results
Based on maternal-report SRS Tscores, 44.9 % (n=22;
17 natal males, 5 natal females) of the 49 cases were in the
clinical range. The probability of being in the clinical range
did not vary by sex based on Fisher’s exact test, p=.71.
The mean (SD) Tscore was 71.05 (9.79) for those cases in
the clinical range and 47.78 (5.06) for those in the non-
clinical range. Thus, the children in the clinical range
exhibited autistic traits ranging from mild to severe with
moderate levels on average; the non-clinical range children
showed scores that were consistent with unaffected popu-
lations (Constantino and Gruber 2005).
Table 1presents descriptive statistics and summarizes
the results of independent ttests. Patients in the clinical
range were significantly older, had significantly lower IQ
scores, had significantly more behavioral and emotional
problems based on CBCL total Tscores, had significantly
higher birth weights, and showed significantly greater
gender nonconformity. Table 2presents the zero-order
correlations among continuous variables across the entire
sample. Of note given the focus of the present study was
the significant positive correlation between birth weight
and gender nonconformity.
Based on these group and correlation effects, a series of
multiple logistic regression analyses were conducted in which
the patient’s age, sex (dummy coded as natal females =0and
natal males =1), IQ, and CBCL total Tscore were entered as
control variables. Maternal Depression scores were not con-
trolled given the lack of effects found for this variable
(Tables 1,2). Three separate models examined focal vari-
ables and their potential interaction with sex (Table 3). The
first and second models showed that high birth weight and
elevated gender nonconformity, respectively, were signifi-
cant predictors of clinical-range autistic traits even after
accounting for the control variables. The third model
showed that an interaction of high birth weight and ele-
vated gender nonconformity significantly predicted the
presence of clinical-range autistic traits. Thus, this latter
finding confirms the a priori prediction of an overlap
between high birth weight, elevated gender nonconformity,
and autistic traits.
Discussion
In a sample of children clinically referred for GD, the present
study examined autistic traits in relation to gender noncon-
formity and three variables that serve as proxies for early-life
exposure to ASD risk: high birth weight, advanced parental
age, and high sibling sex ratio. These latter two risk factors
were not associated with autistic traits or gender noncon-
formity; however, effects were found for birth weight. Even
though both groups showed mean birth weights in the normal
range, GD children with clinical-range autistic traits
weighed 372 g (11.5 %) more at birth on average, corre-
sponding to a large effect size (Cohen’s d=.76). Relatively
higher birth weight was associated with elevated gender
nonconformity and these two factors in combination were
associated with clinical-range autistic traits. Thus, relatively
speaking, higher birth weight serves as a physical marker of
ASD in statu nascendi among highly gender-nonconforming
children clinically referred for GD. The present findings
regarding birth weight were, therefore, consistent with
hypotheses suggesting that traits of ASD are associated with
the development of the extreme and persistent cross-gender
behavior and identity that is characteristic of GD in children.
The present findings also highlight high birth weight as
an important clue for helping refine hypotheses regarding
the GD–ASD link. Elevated maternal weight prior to and/
or during pregnancy might be especially relevant because
it is associated with large size at birth (Haugen et al. 2014;
Kirchengast and Hartmann 2013) and elevated cognitive
and psychiatric problems in offspring (for review, see Van
Lieshout 2013), including ASD (Dodds et al. 2011).
Although information on maternal pre-pregnancy weight
and weight gain during pregnancy were unavailable, the
J Autism Dev Disord (2015) 45:1742–1750 1745
123
CBCL and SRS data from the present sample were con-
sistent with this literature. GD children with higher birth
weights tended to show more behavioral and emotional
problems on the CBCL as well as elevated autistic traits.
Importantly, high birth weight was associated with
autistic traits even after statistically controlling for gen-
eral behavioral and emotional problems by covarying
CBCL scores. Our results suggest, therefore, that high
birth weight has a unique association with autistic traits
that cannot simply be explained as part of a more general
pattern of psychopathology.
Little is known about the precise mechanisms that link
birth weight with ASD risk. Brain overgrowth in particular,
as opposed to large size at birth in general, may be
responsible. Infants later diagnosed with ASD show greater
rates of growth in head circumference during the first year
of life (Fukumoto et al. 2011). Furthermore, ASD has been
associated with early-life (i.e., prior to age 4) overgrowth in
several brain regions, including the amygdala, cerebellum,
cingulate cortex, prefrontal cortex, and temporal lobe (for
review, see Alley et al. 2014). The amygdala might be
particularly relevant because enlarged size and functional
abnormalities in the amygdala appear to contribute to def-
icits in social functioning (for review, see Chasson et al.
2011). Whether these brain regions contribute to GD in
children remains to be examined.
It is possible that traits of ASD and/or their neurobio-
logical underpinnings have a direct influence on the emer-
gence of cross-gender behavior and identity. To begin with,
the premise that early-onset traits of ASD can lead to a later-
emerging GD is tenable given that high birth weight sug-
gests an early predisposition toward ASD. VanderLaan
et al. (2014b) suggested that such a developmental sequence
could unfold if ASD-based intense/obsessional interests in
cross-gender activities or objects gave rise to a cross-gender
self-schema and identity. Social communication deficits
may also contribute. Robinow (2009) suggested that
neurobiological abnormalities associated with reduced
Table 1 Comparisons of
children according to whether
they were in the clinical range
for autistic traits
a
Father’s age was missing for
three cases (one non-clinical
range case and two clinical
range cases)
Non-clinical range
(n=27)
Clinical range
(n=22)
tvalue df p value
MSDMSD
Control variables
Age 6.20 2.47 8.41 2.53 -3.09 47 .003
IQ 115.93 12.82 104.55 14.41 2.92 47 .005
CBCL total Tscore 53.30 7.30 68.55 7.44 -7.21 47 \.001
Maternal Depression Tscore 56.15 8.26 59.82 10.84 -1.35 47 .185
Focal variables
Mother’s age at child’s birth 34.07 5.51 32.82 6.72 .72 47 .476
Father’s age at child’s birth
a
36.00 5.07 35.20 5.61 .51 44 .615
Birth weight (kg) 3.22 .45 3.59 .52 -2.65 47 .011
Gender nonconformity 3.34 .68 3.69 .50 -2.02 47 .049
Sibling sex ratio .52 .22 .45 .19 1.02 47 .312
Table 2 Correlations among
continuous variables across the
entire sample
*p\.05; ** p \.01;
***p \.001
12345678
1. Age –
2. IQ -.417
**
–
3. CBCL total Tscore .449
***
-.261 –
4. Maternal Depression
Tscore
-.027 .020 .324
*
–
5. Mother’s age at child’s
birth
-.204 .417
**
-.018 -.112 –
6. Father’s age at child’s
birth
-.195 .187 -.045 -.038 .689
***
–
7. Birth weight (kg) .052 .064 .307
*
.176 .034 -.041 –
8. Gender nonconformity .140 -.039 .214 .185 -.080 -.124 .344
*
–
9. Sibling sex ratio .132 .055 .001 .172 -.007 -.144 .041 .079
1746 J Autism Dev Disord (2015) 45:1742–1750
123
social functioning in ASD, such as those found for frontal
and temporal regions, might make it difficult for some
children to acquire concepts regarding gender norms. Social
communication deficits might, therefore, underlie the cog-
nitive ‘‘lag’’ that many GD children exhibit in terms of their
gender constancy development (Wallien et al. 2009; Zucker
et al. 1999). Further, Strang et al. (2014) posited that social
communication deficits limit a child’s awareness of social
cues in response to his or her gender role enactment. If such
awareness was limited, a child may not adjust his or her
behavior toward more stereotypically masculine or femi-
nine behavior. In concert, these processes may increase the
likelihood of cross-gender behavior and identity.
If ASD does influence the emergence of cross-gender
behavior and identity in this manner, then the question
arises as to why only a particular subset of children
showing traits of ASD exhibit marked gender noncon-
formity. One possibility is that traits of ASD lead to
gender nonconformity in a stochastic fashion whereby
children with ASD form intense preoccupations with
cross-gender activities or objects due to chance. Another
possibility is that factors increasing the likelihood of GD
are elevated among those ASD children who are gender
nonconforming. Cross-gender behavior and identity has a
familial (Go
´mez-Gil et al. 2010; VanderLaan et al. 2013a,
b), possibly genetic (Alanko et al. 2010; Heylens et al.
2012), component. It is also associated with excesses of
older brothers in natal males and excesses of older sisters
in natal females (Blanchard et al. 1995; Schagen et al.
2012; VanderLaan et al. 2014a; Zucker et al. 1997). With
respect to psychological correlates, GD is frequently
comorbid with internalizing problems such as depression
and anxiety (for review, see Zucker et al. 2014). Regarding
this last point, in a sample of children with neurodevel-
opmental disorders (ASD or Attention-Deficit Hyperac-
tivity Disorder), internalizing problems were indeed
elevated among those who expressed cross-sex wishes
compared to those who did not (Strang et al. 2014).
Similarly, in the present study, those GD children with
clinical-range autistic traits had elevated behavioral and
emotional problems, as indicated by CBCL total Tscores.
Future research should, therefore, continue to consider
whether these factors characterize the backgrounds of
ASD children who exhibit marked gender nonconformity
as well as those of GD children who exhibit traits of
ASD.
Rather than ASD contributing to GD in a direct fashion
as described above, an alternative hypothesis is that high
birth weight is a proxy for some process that indirectly
influences both gender nonconformity and traits of ASD.
That is, high birth weight might be associated with ASD for
reasons such as those noted above while its association
with GD is due to some other circumstance(s). For exam-
ple, high birth weight is inversely associated with prenatal
testosterone exposure (Carlsen et al. 2006). In males, lower
levels of prenatal testosterone exposure are associated with
decreased gender-typical play behavior (Lamminma
¨ki
et al. 2012), possibly because low testosterone exposure
leads to less masculinization and/or defeminization in
neural regions that influence sexually dimorphic behavior.
Thus, in natal males of relatively high birth weight, GD
may be a consequence of lower prenatal testosterone
exposure.
Because high birth weight is associated with lower tes-
tosterone (Carlsen et al. 2006), and lower testosterone is
associated with greater female-typical play behavior in
females (Lamminma
¨ki et al. 2012), it seems unlikely that
the GD–ASD link in females would be similarly owing to
prenatal testosterone exposure. Instead, some alternate
explanation should be sought. For example, research has
shown male biases for both ASD (e.g., Blumberg et al.
2013; Dodds et al. 2011; Fombonne 2005) and relatively
higher birth weight (e.g., Co
ˆte
´et al. 2003). When present in
natal females, these traits may reflect a pattern of male-
typical prenatal development. Indeed, higher birth weight
in females has been associated with masculinized somatic
features such as greater anogenital distance (i.e., the dis-
tance between the anus and the fourchette) (Avidime et al.
2011). If such masculinization extends to neural regions
that underlie sexually dimorphic behavior, then that might
help explain elevated gender nonconformity among high
birth weight natal females with GD.
Table 3 Multiple logistic regression comparing clinical versus non-
clinical range cases
BSE ßtvalue pvalue
Control variables
Age .005 .022 .03 .25 .805
Sex .012 .131 .01 .10 .925
IQ -.007 .004 -.21 -1.91 .062
CBCL total Tscore .031 .005 .66 5.76 \.001
Model 1: birth weight
Birth weight .398 .174 .41 2.29 .027
Sex 9birth weight -.301 .206 -.83 -1.47 .150
Model 2: gender nonconformity
Gender nonconformity .752 .366 .93 2.06 .046
Sex 9gender
nonconformity
-.672 .377 -1.49 -1.79 .082
Model 3: birth weight 9gender nonconformity
Birth weight 9gender
nonconformity
.102 .041 .57 2.48 .017
Sex 9birth
weight 9gender
nonconformity
-.079 .046 -.64 -1.73 .091
J Autism Dev Disord (2015) 45:1742–1750 1747
123
Future Directions
It is important to note that the various hypotheses described
above to account for the GD–ASD link are not necessarily
mutually exclusive. The processes related to these
hypotheses may interact with one another. Another possi-
bility is that each of these hypotheses applies to some
subset of children who exhibit traits of ASD and GD.
Future research may help discern which is the case.
Interestingly, our data showed a marginally significant
interaction between birth weight, gender nonconformity,
and sex in the prediction of autistic traits. Specifically, this
interaction suggested that the relationship between birth
weight, gender nonconformity, and autistic traits might be
stronger among natal females. If so, then different pro-
cesses might influence the GD–ASD link in natal males
versus females. Alternatively, the GD–ASD link in natal
males and females might be underpinned by the same or
similar processes, but the magnitude of the effects of high
birth weight on gender nonconformity and autistic traits
might be greater among natal females. This speculation is
based on a small sample of nine natal females clinically
referred for GD and future research on larger samples of
such females is needed before firm conclusions can be
drawn.
In addition to examining the sexes separately, future
research regarding the GD–ASD link should consider using
dimensional metrics of autistic traits such as the SRS as
well as formal diagnostic criteria. To date, only the study
by de Vries et al. (2010) considered formal diagnostic
criteria of both conditions. In that study, many of the 16
youth referred for GD who were classified as having ASD
were only subthreshold for a diagnosis of Gender Identity
Disorder, the formal DSM-IV diagnosis that preceded GD
in DSM-5. Even though it is based on a small number of
cases, this tendency appears to be somewhat at odds with
the current study, which showed that traits of ASD were
associated with greater gender nonconformity. As such, it
is necessary to question whether the present findings
regarding high birth weight would hold in the context of
formal diagnoses. If so, then the various hypotheses
described above regarding the significance of high birth
weight for the GD–ASD link would be tenable across the
autism spectrum. If not, then it might be the case that
processes related to high birth weight contribute to an
association between gender nonconformity and autistic
traits, but alternate explanations must be sought for cases
in which patients satisfy a formal diagnosis of ASD. Such
an examination was not possible in the current study given
sample size limitations; however, future studies employing
larger samples may be better able to address these issues.
Lastly, the existing studies examining the GD–ASD link
either lacked a clinical comparison group (present study; de
Vries et al. 2010; Jones et al. 2012; Pasterski et al. 2014)or
did not employ comprehensive measures of GD (Strang
et al. 2014) or ASD (VanderLaan et al. 2014b). Further
understanding may, therefore, be gained by examining
comprehensive measures of ASD in a comparison group of
children clinically referred for reasons unrelated to GD or
ASD. Such a comparison group would help establish
whether increased prevalence of ASD is unique to the GD
population. Also, a comparison group would make it pos-
sible to evaluate whether an increased presence of ASD
risk factors such as high birth weight among GD individ-
uals accounts for any potential group differences in autistic
traits.
Clinical Implications
There is currently no consensus regarding best practice
when treating children with GD (Zucker 2008); however,
one’s clinical formulation regarding the bases of GD in a
given child may help direct treatment approaches (Van-
derLaan and Zucker in press). This article outlined
hypotheses that may account for GD among children who
show traits of ASD and, thus, may also have implications
for therapeutic strategies. For example, in the course of
psychological therapy, one may wish to explore whether
traits of ASD such as intense/obsessional interests or social
communication deficits contribute to a child’s gender
schema and, ultimately, his or her cross-gender behavior
and identity. If so, it would be useful to evaluate the
likelihood that such ASD traits will continue to do so,
especially in light of case studies reporting the desistence
of GD among ASD youth (Robinow 2009; Parkinson
2014). Alternatively, if the presence of traits of ASD is
reflective of exposure to factors that affected prenatal
sexual differentiation of neural areas influencing sexually
dimorphic behavior, then traits of ASD may have little or
no impact on co-occurring GD. In such a circumstance,
ASD may be a treatment consideration, but not necessarily
the focus. Instead, one may focus more directly on
addressing the lack of congruence between the experienced
gender identity and the one assigned at birth.
Conclusions
The present findings identified relatively high birth weight
as an important clue that may help explain the link between
GD and ASD in children. This study is the first to inves-
tigate and identify somatic features (i.e., birth weight) that
circumscribe the set of mechanisms that might account for
this link. Specifically, a number of factors associated with
relatively high birth weight could plausibly influence both
GD and ASD either directly or indirectly. Such factors
1748 J Autism Dev Disord (2015) 45:1742–1750
123
include neuropsychological abnormalities that influence
ASD behaviors; these behaviors may then subsequently
contribute to GD. It is also possible that while high birth
weight affects brain areas related to ASD, it also reflects
exposure to low levels of prenatal testosterone in natal
males and somatic masculinization in natal females. It is,
therefore, also reasonable to speculate that these latter
processes contribute to GD among the subset of children
who exhibit both GD and traits of ASD. Detailing which of
these hypotheses, or similarly plausible hypotheses, pro-
vides an adequate explanation for the GD–ASD link has
important clinical implications in terms of informing
treatment approaches. Thus, future research is needed to
discern whether these hypotheses are indeed accurate.
Acknowledgments D.P.V. was supported by a Canadian Institutes
of Health Research Postdoctoral Fellowship. We thank two anony-
mous reviewers for comments on an earlier version of this article.
References
Abel, K. M., Dalman, C., Svensson, A. C., Susser, E., Dal, H., Idring,
S., et al. (2013). Deviance in fetal growth and risk of autism
spectrum disorder. American Journal of Psychiatry, 170,
391–398. doi:10.1176/appi.ajp.2012.12040543.
Achenbach, T. M. (1991). Manual for the child behavior checklist/4-
18 and 1991 profile. Burlington, VT: University of Vermont
Department of Psychiatry.
Alanko, K., Santtila, P., Harlaar, N., Witting, K., Varjoen, M., Jern,
P., et al. (2010). Common genetic effects of gender atypical
behavior in childhood and sexual orientation in adulthood: A
study of Finnish twins. Archives of Sexual Behavior, 39, 81–92.
doi:10.1007/s10508-008-9457-3.
Alley, C. S., Gillberg, C., & Wilson, P. (2014). Neurobiological
abnormalities in the first few years of life in individuals later
diagnosed with autism spectrum disorder: A review of recent
data. Behavioural Neurology. doi:10.1155/2014/210780.
American Psychiatric Association. (2013). Diagnostic and statistical
manual of mental disorders (5th ed.). Arlington, VA: Author.
Avidime, O. M., Avidime, S., Olorunshola, K. V., & Dikko, A. A. U.
(2011). Anogenital distance and umbilical cord testosterone level
in newborns in Zaria, Northern Nigeria. Nigerian Journal of
Physiological Sciences, 26, 23–28.
Bennett, T., Boyle, M., Georgiades, K., Georgiades, S., Thompson,
A., Duku, E., et al. (2012). Influence of reporting effects on the
association between maternal depression and child autism
spectrum disorder behaviors. Journal of Child Psychology and
Psychiatry, 53, 89–96. doi:10.1111/j.1469-7610.2011.02451.x.
Blanchard, R., Zucker, K. J., Bradley, S. J., & Hume, C. S. (1995).
Birth order and sibling sex ratio in homosexual male adolescents
and probably prehomosexual feminine boys. Developmental
Psychology, 31, 22–30. doi:10.1037/0012-1649.31.1.22.
Blanchard, R., Zucker, K. J., Cavacas, A., Allin, S., Bradley, S. J., &
Schachter, D. C. (2002). Fraternal birth order and birth weight in
probably prehomosexual feminine boys. Hormones and Behav-
ior, 41, 321–327. doi:10.1006/hbeh.2002.1765.
Blumberg, S. J., Bramlet, M. D., Kogan, M. D., Schieve, L. A., Jones,
J. R., & Lu, M. C. (2013). Changes in prevalence of parent-
reported autism spectrum disorder in school-aged U.S. children:
2007–2012. National Health Statistics Reports, 65, 1–11.
Carlsen, S. M., Jacobsen, G., & Romundstad, P. (2006). Maternal
testosterone levels during pregnancy are associated with off-
spring size at birth. European Journal of Endocrinology, 155,
365–370. doi:10.1530/eje.1.02200.
Chasson, G. S., Timpano, K. R., Greenberg, J. L., Shaw, A., Singer,
T., & Wilhelm, S. (2011). Shared social competence impairment:
Another link between the obsessive-compulsive and autism
spectrums? Clinical Psychology Review, 31, 653–662. doi:10.
1016/j.cpr.2011.02.006.
Constantino, J. N., & Gruber, C. P. (2005). Social responsiveness
scale. Los Angeles: Western Psychological Services.
Co
ˆte
´, K., Blanchard, R., & Lalumie
`re, M. L. (2003). The influence of
birth order on birth weight: Does the sex of preceding siblings
matter? Journal of Biosocial Science, 35, 455–462. doi:10.1017/
S0021932003004553.
de Vries, A. L. C., Noens, I. L., Cohen-Kettenis, P. T., van
Berckelaer-Onnes, I. A., & Doreleifers, T. A. (2010). Autism
spectrum disorders in gender dysphoric children and adolescents.
Journal of Autism and Developmental Disorders, 40, 930–936.
doi:10.1007/s10803-010-0935-9.
Derogatis, L. R. (1994). Symptom checklist-90-R. Minneapolis:
National Computer Systems Inc.
Dodds, L., Fell, D. B., Shea, S., Armson, B. A., Allen, A. C., & Bryson,
S. (2011). The role of prenatal, obstetric and neonatal factors in
the development of autism. Journal of Autism and Developmental
Disorders, 41, 891–902. doi:10.1007/s10803-010-1114-8.
Fombonne, E. (2005). Epidemiology of autistic disorder and other
pervasive developmental disorders. Journal of Clinical Psychi-
atry, 66(Supplement 10), 3–8.
Fukumoto, A., Hashimoto, T., Mori, K., Tsuda, Y., Arisawa, K., &
Kagami, S. (2011). Head circumference and body growth in
autism spectrum disorders. Brain and Development, 33,
569–575. doi:10.1016/j.braindev.2010.09.004.
Gayle, H. D., Yip, R., Frank, M. J., Nieburg, P., & Binkin, N. J.
(1988). Validation of maternally reported birth weights among
46,637 Tennessee WIC program participants. Public Health
Reports, 103, 143–147.
Gofin, R., Neumark, Y. D., & Adler, B. (2000). Birthweight recall by
mothers of Israeli children. Public Health, 114, 161–163. doi:10.
1038/sj.ph.1900626.
Go
´mez-Gil, E., Esteva, I., Almaraz, M. C., Pasaro, E., Segovia, S., &
Guillamon, A. (2010). Familiality of gender identity disorder in
non-twin siblings. Archives of Sexual Behavior, 39, 546–552.
doi:10.1007/s10508-009-9524-4.
Haugen, M., Brantsæter, A. L., Winkvist, A., Lissner, L., Alexander,
J., Oftedal, B., et al. (2014). Associations of pre-pregnancy body
mass index and gestational weight gain with pregnancy outcome
and postpartum weight retention: A prospective observational
cohort study. BMC Pregnancy and Childbirth, 14, 201. doi:10.
1186/1471-2393-14-201.
Heylens, G., De Cuypere, G., Zucker, K. J., Schelfaut, C., Elaut, E., Vanden
Bossche, H., et al. (2012). Gender identity disorder in twins: A review
of the case report literature. Journal of Sexual Medicine, 9, 751–757.
doi:10.1111/j.1743-6109.2011.02567.x.
Johnson, L. L., Bradley,S. J., Birkenfeld-Adams, A. S., Kuksis, M. A. R.,
Maing, D. M., Mitchell, J. N., et al. (2004). A parent-report Gender
Identity Questionnaire for Children. Archives of Sexual Behavior,
33, 105–116. doi:10.1023/B:ASEB.0000014325.68094.f3.
Jones, R. M., Wheelright, S., Farrell, K., Martin, E., Green, R., Di
Ceglie, D., et al. (2012). Brief report: Female-to-male transsexual
people and autistic traits. Journal of Autism and Developmental
Disorders, 42, 301–306. doi:10.1007/s10803-011-1227-8.
Kirchengast, S., & Hartmann, B. (2013). Maternal obesity affects
newborn somatometrices and vital parameters in a gender typical
manner—Evidence for the male disadvantage hypothesis? Col-
legium Antropologicum, 4, 1057–1063.
J Autism Dev Disord (2015) 45:1742–1750 1749
123
Knickmeyer, R., Baron-Cohen, S., Fane, B. A., Wheelwright, S.,
Mathews, G. A., Conway, G. S., et al. (2006). Androgens and
autistic traits: A study of individuals with congenital adrenal
hyperplasia. Hormones and Behavior, 50, 148–153. doi:10.1016/
j.yhbeh.2006.02.006.
Knickmeyer, R. C., Wheelright, S., & Baron-Cohen, S. B. (2008). Sex-
typical play: Masculinization/defeminization in girls with an
autism spectrum condition. Journal of Autism and Developmental
Disorders, 38, 1028–1035. doi:10.1007/s10803-007-0475-0.
Kondrashov, A. (2012). The rate of mutation. Nature, 488, 467–468.
doi:10.1038/488467a.
Lamminma
¨ki, A., Hines, M., Kuiri-Ha
¨nninen, T., Kilpela
¨inen, L.,
Dunkel, L., & Sankilampi, U. (2012). Testosterone measured in
infancy predicts subsequent sex-typed behavior in boys and in
girls. Hormones and Behavior, 61, 611–616. doi:10.1016/j.
yhbeh.2012.02.013.
Lande
´n, M., & Rasmussen, P. (1997). Gender identity disorder in a
girl with autism–A case report. European Child and Adolescent
Psychiatry, 6, 170–173. doi:10.1007/BF00538990.
Lundstro
¨m, S., Haworth, C. M. A., Carlstro
¨m, E., Gillberg, C., Mill,
J., Rastam, M., et al. (2010). Trajectories leading to autism
spectrum disorders are affected by paternal age: Findings from
two nationally representative twin studies. Journal of Child
Psychology and Psychiatry, 51, 580–586. doi:10.1111/j.1469-
7610.2010.02223.x.
Mouridsen, S. E., Rich, B., & Isager, T. (2010). Sibling sex ratio of
individuals diagnosed with autism spectrum disorder as children.
Developmental Medicine and Child Neurology, 52, 289–292.
doi:10.1111/j.1469-8749.2009.03368.x.
O’Sullivan, J. J., Pearce, M. S., & Parker, L. (2000). Parental recall of
birth weight: How accurate is it? Archives of Disease in
Childhood, 82, 202–203. doi:10.1136/adc.82.3.202.
Parkinson, J. (2014). Gender dysphoria in Asperger’s syndrome: A
caution. Australasian Psychiatry, 22, 84–85. doi:10.1177/
1039856213497814.
Parner, E. T., Baron-Cohen, S., Lauritsen, M. B., Jørgensen, M.,
Schieve, L. A., Yeargin-Allsopp, M., et al. (2012). Parental age
and autism spectrum disorders. Annals of Epidemiology, 22,
143–150. doi:10.1016/j.annepidem.2011.12.006.
Pasterski, V., Gilligan, L., & Curtis, R. (2014). Traits of autism
spectrum disorders in adults with gender dysphoria. Archives of
Sexual Behavior, 43, 387–393. doi:10.1007/s10508-013-0154-5.
Robinow, O. (2009). Paraphilia and transgenderism: A connection
with Asperger’s disorder? Sexual and Relationship Therapy, 24,
143–151. doi:10.1080/14681990902951358.
Sandin, S., Hultman, C. M., Kolevzon, A., Gross, R., MacCabe, J. H.,
& Reichenberg, A. (2012). Advancing maternal age is associated
with increasing risk for autism: A review and meta-analysis.
Journal of the American Academy of Child and Adolescent
Psychiatry, 51, 477–486. doi:10.1016/j.jaac.2012.02.018.
Schagen, S. E. E., Delemarre-van de Waal, H. A., Blanchard, R., &
Cohen-Kettenis, P. T. (2012). Sibling sex ratio and birth order in
early-onset gender dysphoric adolescents. Archives of Sexual
Behavior, 41, 541–549. doi:10.1007/s10508-011-9777-6.
Strang, J. F., Kenworthy, L., Dominska, A., Sokoloff, J., Kenealy, L.
E., Berl, M., et al. (2014). Increased gender variance in autism
spectrum disorders and attention deficit hyperactivity disorder.
Archives of Sexual Behavior. doi:10.1007/s10508-014-0285-3.
Tomeo, C. A., Rich-Edwards, J. W., Michels, K. B., Berkey, C. S.,
Hunter, D. J., Frazier, A. L., et al. (1999). Reproducibility and
validity of maternal recall of pregnancy-related events. Epide-
miology, 10, 774–777.
Van Lieshout, R. J. (2013). Role of maternal adiposity prior to and
during pregnancy and cognitive and psychiatric problems in
offspring. Nutrition Reviews, 71, s95–s101. doi:10.1111/nure.
12059.
VanderLaan, D. P., Blanchard, R., Wood, H., & Zucker, K. J. (2014a).
Birth order and sibling sex ratio of children and adolescents
referred to a gender identity service. PLoS One, 9, e90257.
doi:10.1371/journal.pone.0090257.
VanderLaan, D. P., Forrester, D. L., Petterson, L. J., & Vasey, P. L.
(2013a). The prevalence of fa’afafine relatives among Samoan
gynephilic men and fa’afafine.Archives of Sexual Behavior, 42,
353–359. doi:10.1007/s10508-012-0015-7.
VanderLaan, D. P., Postema, L., Wood, H., Singh, D., Fantus, S.,
Hyun, J., et al. (2014b). Do children with gender dysphoria have
intense/obsessional interests? Journal of Sex Research,. doi:10.
1080/00224499.2013.860073.
VanderLaan, D. P., Vokey, J. R., & Vasey, P. L. (2013b). Is
transgendered male androphilia familial in non-Western popu-
lations? The case of a Samoan village. Archives of Sexual
Behavior, 42, 361–370. doi:10.1007/s10508-012-0037-1.
VanderLaan, D. P., & Zucker, K. J. (in press). Gender dysphoria in
children and adolescents. In M. H. Ebert, J. F. Leckman, & I.
Petrakis (Eds.), Current diagnosis and treatment psychiatry 3rd
edition. California: McGraw-Hill.
Wallien, M. S. C., Quilty, L. C., Steensma, T. D., Singh, D., Lambert,
S. L., Leroux, A., et al. (2009). Cross-national replication of the
Gender Identity Interview for Children. Journal of Personality
Assessment, 91, 545–552. doi:10.1080/00223890903228463.
Wechsler, D. (2003). WISC-IV technical and interpretive manual. San
Antonio, TX: Psychological Corporation.
Williams, P. G., Allard, A. M., & Sears, L. (1996). Case study: Cross-
gender preoccupations with two male children with autism.
Journal of Autism and Developmental Disorders, 26, 635–642.
doi:10.1007/BF02172352.
Zucker, K. J. (2008). Children with gender identity disorder: Is there a
best practice? [Enfants avec troubles de l’identite
´sexue
´e: y-a-t-il
une pratique la meilleure?]. Neuropsychiatrie de l’Enfance et de
l’Adolescence, 56, 358–364. doi:10.1016/j.neurenf.2008.06.003.
Zucker, K. J., Bradley, S. J., Kuksis, M., Pecore, K., Birkenfeld-
Adams, A., Doering, R. W., et al. (1999). Gender constancy
judgements in children with gender identity disorder: Evidence
for a developmental lag. Archives of Sexual Behavior, 28,
475–502. doi:10.1023/A:1018713115866.
Zucker, K. J., Green, R., Coates, S., Zuger, B., Cohen-Kettenis, P. T.,
Zecca, G. M., et al. (1997). Sibling sex ratio of boys with gender
identity disorder. Journal of Child Psychology and Psychiatry, 38,
543–551. doi:10.1111/j.1469-7610.1997.tb01541.x.
Zucker, K. J., & Lawrence, A. (2009). Epidemiology of gender
identity disorder: Recommendations for the Standards of Care of
the World Professional Association for Transgender Health.
International Journal of Transgenderism, 11, 8–18. doi:10.1080/
15532730902799946.
Zucker, K. J., Wood, H., & VanderLaan, D. P. (2014). Models of
psychopathology in children and adolescents with gender
dysphoria. In B. P. C. Kreukels, T. D. Steensma, & A. L. C.
de Vries (Eds.), Gender dysphoria and disorders of sex
development: Progress in care and knowledge (pp. 171–192).
New York: Springer.
1750 J Autism Dev Disord (2015) 45:1742–1750
123
