Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
DRO Deakin Research Online, Deakin University’s Research Repository Deakin University CRICOS Provider Code: 00113B
Associations between dehydroepiandrosterone (DHEA) levels, pituitary volume, and social anxiety in children
Citation: Murray, Cynthia R., Simmons, Julian G., Allen, Nicholas B., Byrne, Michelle L., Mundy, Lisa K., Seal, Marc L., Patton, George C., Olsson, Craig A. and Whittle, Sarah 2016, Associations between dehydroepiandrosterone (DHEA) levels, pituitary volume, and social anxiety in children, Psychoneuroendocrinology, vol. 64, pp. 31-39.
This is the accepted manuscript.
©2016, Elsevier
This peer reviewed accepted manuscript is made available under a Creative Commons Attribution Non-Commercial No-Derivatives 4.0 Licence.
The final version of this article, as published in volume 64 of Psychoneuroendocrinology, is available online from: https://doi.org/10.1016/j.psyneuen.2015.11.004
Downloaded from DRO: http://hdl.handle.net/10536/DRO/DU:30079725
1
Associations between dehydroepiandrosterone (DHEA) levels, pituitary volume, and
social anxiety in children.
Cynthia R. Murraya, Julian G. Simmonsa,b,c, Nicholas B. Allenb,g, Michelle L. Byrneg, Lisa K.
Mundyc,d,e, Marc L. Sealc, George C. Pattonc,d,e, Craig A. Olssonc,e,f, Sarah Whittlea,b,*.
a Melbourne Neuropsychiatry Centre, Department of Psychiatry, The University of
Melbourne and Melbourne Health, 161 Barry Street, Carlton VIC 3053, Australia.
b Melbourne School of Psychological Sciences, The University of Melbourne, Parkville VIC
3010, Australia.
c Murdoch Childrens Research Institute, 50 Flemington Road, Parkville VIC 3052, Australia.
d Centre for Adolescent Health, The Royal Children’s Hospital, 50 Flemington Road,
Parkville VIC 3052, Australia.
e Department of Paediatrics, The University of Melbourne, 50 Flemington Road, Parkville
VIC 3052, Australia.
f School of Psychology, Deakin University, 221 Burwood Highway, Burwood VIC 3125,
Australia.
g Department of Psychology, University of Oregon, 1715 Franklin Boulevard, Eugene OR
97403, USA
* Corresponding author:
Sarah Whittle
Melbourne Neuropsychiatry Centre, Department of Psychiatry, The University of Melbourne
and Melbourne Health, 161 Barry Street, Carlton VIC 3053, Australia.
+61 3 8344 1958
2
Highlights
Larger PGV is associated with higher levels of DHEA/S, independent of age.
Findings suggest early adrenarche is associated with larger PGV.
Larger PGV is associated with social anxiety symptoms in children.
PGV may mediate the association between early adrenarche and social anxiety.
Abstract
Early timing of adrenarche, associated with relatively high levels of dehydroepiandrosterone
(DHEA) and its sulphate (DHEA-S) in children, has been linked with mental health
problems, particularly anxiety. However, little is known about possible neurobiological
mechanisms underlying this association. The pituitary gland is a key component of the
hypothalamic-pituitary-adrenal (HPA) axis, the activation of which triggers the onset of
adrenarche. The purpose of this study was to examine the extent to which pituitary gland
volume mediated the relationship between levels of DHEA/DHEA-S relative to age (i.e.,
adrenarcheal timing) and symptoms of anxiety in 95 children (50 female, M age 9.50 years,
SD 0.34 years). Relatively high DHEA and DHEA-S (DHEA/S) levels were found to be
associated with larger pituitary gland volumes. There was no significant direct effect of
relative DHEA/S levels on overall symptoms of anxiety. However, results supported an
indirect link between relatively high DHEA/S levels and symptoms of social anxiety,
mediated by pituitary gland volume. No sex differences were observed for any relationship.
Our findings suggest that neurobiological mechanisms may be partly responsible for the link
between relatively early adrenarche and anxiety symptoms in children. One possible
mechanism for this finding is that an enlarged pituitary gland in children experiencing
relatively advanced adrenarche might be associated with hyper-activity/reactivity of the HPA
axis. Further research is needed to understand the role of stress in the link between
adrenarcheal timing and HPA-axis function, especially in relation to the development of
anxiety symptoms in children and adolescents.
Keywords: adrenarche, dehydroepiandrosterone, pituitary gland, social anxiety, puberty,
HPA axis.
3
1. Introduction
Individual differences in the timing of puberty (i.e., the stage of puberty of the
individual relative to their same-aged peers) can confer vulnerability to adverse mental health
outcomes, particularly mood and anxiety disorders. These disorders show a dramatic rise in
incidence during the early to mid-adolescent phase (Ge & Natsuaki, 2009; Patton & Viner,
2007). Indeed, early maturation of the hypothalamic-pituitary-gonadal (HPG) axis (i.e., early
timing of gonadarche) constitutes a significant risk factor for the development of mental
illness, particularly in girls (Ge, Conger, & Elder, 1996).
The timing of adrenarche, an early maturational phase of puberty occurring prior to
gonadarche (characterized by a marked increase in the production of the adrenal androgen
dehydroepiandrosterone [DHEA], and its sulphate DHEA-S), also predicts adverse mental
health problems in children. Girls and boys with premature adrenarche (i.e., pubarche before
age 8 in girls and 9 in boys, associated with increased DHEA/S levels) demonstrate increased
levels of anxiety and behavioral problems when compared to their later developing peers
(Dorn, Hitt, & Rotenstein, 1999; Dorn et al., 2008; Sontag-Padilla et al., 2012). The link
between premature adrenal development and increased risk for anxiety disorders is of
particular interest, given that such disorders have a high incidence in this age group (Albano,
Chorpita, & Barlow, 2003; Bernstein, Borchardt, & Perwien, 1996; Dorn et al., 1999).
It has been suggested that the action of pubertal hormones on the development of
cognitive and affective systems in the brain during sensitive periods may provide some
explanation for dysregulated affect and behavior, and associated psychopathology during
puberty (Blakemore, Burnett, & Dahl, 2010). Thus, the effect of early exposure to pubertal
hormones on developmentally immature brain systems is postulated to contribute to the link
between early pubertal timing and vulnerability to psychopathology. Indeed, we have
recently provided the first evidence that early timing of adrenarche may be associated with
alterations in brain structure and function. Specifically, we have found that relatively high
levels of DHEA in older children was associated with reduced frontal white matter volume
(Klauser et al., 2015), decreased affect-related brain activity in the cingulate cortex, and
decreased affect-related brain activity in a number of cortical and subcortical regions
(specifically in females, Whittle et al., 2015).
Investigation of the structure of the pituitary gland might be a particularly fruitful
approach to understanding the link between early adrenarche and anxiety in children. The
hypothalamic-pituitary-adrenal (HPA) axis regulates and secretes the adrenal androgens
(DHEA/S, Styne & Grumbach, 2011). The pituitary gland is an integral part of the HPA and
4
HPG axes, which become reactivated during puberty and underlies the development of
reproductive maturity (Styne & Grumbach, 2011). The volume of the pituitary gland (PGV)
has previously been utilized as an index of stress-related HPA-axis function (Axelson et al.,
1992), as it is thought to increase with the number of corticotrope cells producing ACTH,
which stimulate the release of cortisol by the adrenal cortex (Pariante et al., 2004). Indeed,
PGV has been associated with cortisol measures of HPA-axis hyper-activity and hyper-
reactivity (Kaess et al., 2013), and with early life stress (Ganella et al., 2015). PGV has also
been used as a measure of HPG axis function (Whittle et al., 2012). Importantly, the size of
the pituitary gland has also been associated with vulnerability to psychopathology
(MacMaster & Kusumakar, 2004; MacMaster et al., 2006), and particularly anxiety
symptoms and related disorders (Thomas & De Bellis, 2004). Given the role of the pituitary
gland in adrenarche, its volume is also likely to be influenced by the action of adrenal
hormones during adrenarche. The association between PGV and adrenarcheal processes,
however, has currently not been investigated. Thus, given that adrenarcheal timing has been
associated with risk for anxiety disorders (as stated above), it is plausible that PGV may play
a role in the link between adrenarcheal timing and such risk.
No studies to date have explored the associations between timing of adrenarche, PGV,
and psychopathology during the developmentally sensitive period of late childhood. Given
independent support for the roles of both adrenarcheal timing and enlarged PGV in the
experience of anxiety, we investigated whether PGV mediated the link between timing of
adrenarche (as assessed by adrenal hormone levels) and symptoms of anxiety in children, and
whether this effect differed between females and males. We hypothesized that relatively early
adrenarche (defined here as high DHEA/S levels relative to same age peers) would be
associated with larger PGV, and that larger PGV would mediate the link between early
adrenarche and increased symptoms of anxiety. Given consistent support for sex differences
in the effects of gonadarcheal timing on psychopathology (Hyde, Mezulis, & Abramson,
2008), we hypothesized that relatively high DHEA/S levels would have differential effects on
PGV and symptoms of anxiety in females and males. However, given that sex differences in
research on adrenarche and mental health are mixed, this aim was exploratory and we did not
have specific hypotheses related to sex differences. Finally, we conducted exploratory
analyses to examine the association between DHEA/S, PGV, anxiety and cortisol, in order to
provide some comment on stress-related mechanisms that may underlie associations found
between the primary variables of interest.
5
2. Material and methods
2.1. Participants
Participants were part of a nested brain-imaging project (iCATS) in the Childhood to
Adolescence Transition Study (CATS) (Mundy et al., 2013), a longitudinal study of puberty
and health in approximately 1200 children. A detailed description of iCATS methods can be
found in Simmons et al. (Simmons et al., 2014). Briefly, 128 9-year old children participated
in iCATS, and were selected based on plotting the morning DHEA and testosterone levels
(which were highly correlated) of the participants in the CATS study, and selecting those
children falling in the lower and upper tertiles for participation in the iCATS study. The
purpose of this selection was to obtain a sample of female and male children who varied in
their exposure to adrenal androgens, taking those who were comparatively advanced in their
adrenal maturation and comparing them to a group where the hormone changes associated
with adrenarche were not yet noticeably underway. The upper and lower tertile groups were
intentionally matched for age to maximize the ability to infer the effects of relative timing of
exposure to adrenal hormones (i.e., timing of adrenarche) independent of chronological age.
Participants with a body mass index (BMI) in the upper and lower 5th percentiles were
excluded. Physical signs of pubertal development were not included in the selection process
because very few parents reported any physical signs of note in their children during the
CATS assessment, as expected at age 9. Of the 128 iCATS participants, 100 agreed to
undergo Magnetic Resonance Imaging (MRI). Ethics approval was granted by the Royal
Children’s Hospital (RCH) Human Research Ethics Committee and ratified by the University
of Melbourne Human Research Ethics Office. Informed consent was obtained from all
children and parents prior to participation.
Data from 95 participants (45 males, M age 9.6, SD 0.4 years; 50 females, M age 9.4,
SD 0.3 years) were included in analyses after exclusions based on image acquisition
problems and excessive motion (n = 5).
2.2.Procedure
Approximately 6 months (M 6.4, SD 1.8 months) after their participation in CATS,
participants (as part of iCATS) underwent a 30-minute Magnetic Resonance Imaging (MRI)
scan of the brain, assessment of salivary and hair hormones, height and weight
measurements, a brief assessment of self-reported psychological symptoms, and a parent
report of Tanner Stage.
6
2.3. Measures
2.3.1. Neuroimaging.
Approximately 6 months after recruitment to the iCATS study, participants were
scanned using a research-dedicated 3T Siemens TIM Trio scanner (Siemens, Erlangen,
Germany) at the Murdoch Childrens Research Institute (MCRI), RCH, Melbourne.
Participants completed a T1-weighted MPRAGE scan that provided high-resolution images
for structural analysis. For all scans, participants were positioned comfortably in a supine
orientation with their head located in a 32-channel head RF coil that was electrically isolated
from the participant. T1-weighted image parameters were: TR = 1900 ms, TE = 2.24 ms, flip
angle = 9°, FOV = 230mm, producing 208 0.9mm contiguous sagittal slices (voxel
dimensions = 0.9mm3).
2.3.2. Estimation of pituitary volume.
PGV was ascertained by manual tracing, which was performed by author CM on each
individual’s images using the software package ANALYZE (Mayo Clinic, Rochester, USA;
http://www.mayo.edu/research/labs/biomedical-imaging/software/analyze-software-system).
The pituitary gland was defined and quantified based on the technique described in Pariante
et al. (2004). Images were traced in the coronal plane, where the pituitary gland is best
identified (Figure 1) (Pariante et al., 2004), and the infundibular stalk was excluded from
tracing. Boundaries of the pituitary could be clearly defined by the diaphragm sellae,
superiorly; the sphenoid sinus, inferiorly; and the cavernous sinuses, bilaterally (Pariante et
al., 2004). Pituitary volume estimates were calculated by summing all voxels within the
traced region on each slice to give values in mm3. Prior to tracing, reliability was established
by comparison between CM and a researcher experienced in pituitary gland delineation.
Intraclass correlation coefficients (absolute agreement) for intra and inter-rater reliabilities,
assessed on ten independent randomly selected images, were 0.976 and 0.985, respectively.
-- insert Figure 1 about here --
2.3.3. Anxiety symptoms.
Children completed an assessment of self-reported anxiety symptoms using the
Spence Children’s Anxiety Scale (SCAS) (Spence, 1998). The SCAS is a self-report
questionnaire designed to measure the severity of anxiety symptoms in children, broadly in
line with the dimensions of anxiety disorders proposed by the DSM-IV-TR. The scale
assesses six domains of anxiety: generalized anxiety, panic/agoraphobia, social phobia,
7
separation anxiety, obsessive-compulsive disorder, and physical injury fears. The
questionnaire contains 44 items, of which 38 reflect specific symptoms of anxiety, and 6 are
positive, filler items, included to reduce negative response bias (Spence, 1998). Participants
were asked to rate the degree to which they experience each symptom on a 4-point Likert
scale involving never (0), sometimes (1), often (2), and always (3). Separate scores for each
subscale are computed by summing individual item scores for questions relating to specific
domains of anxiety. Higher total and subscale scores indicate greater symptoms of anxiety.
The SCAS has strong evidence for reliability and validity (Muris, Merckelbach, Ollendick,
King, & Bogie, 2002). In the current study, the scale showed good internal consistency (α =
.78).
2.3.4. Adrenal development (adrenarche).
Children were asked to provide two waking saliva samples; one on the morning prior
to, and the second on the morning of their MRI scan, with the help of their parent/guardian.
Samples were assayed for levels of testosterone, DHEA and DHEA-S, as hormonal markers
of adrenal development (adrenarche). Hormonal assays were conducted at the MCRI, using
Salimetrics ELISA kits. Participants returned their saliva samples at the time of their RCH
visit. Samples were centrifuged and the supernatant was used in the assays. The range of
sensitivity was from 10.2 to 1000 mg/ml. The inter-assay coefficient of variation (CV) was
5.45% for DHEA, 7.53% for DHEA-S and 13.54% for testosterone, and intra-assay CVs
were 8.56% for DHEA, 9.38% for DHEA-S, and 7.32% for testosterone. The mean of the
two DHEA/S and testosterone levels in the saliva samples was used in subsequent analysis.
2.3.5. Physical maturation.
Tanner Stage was assessed via parent report of pubertal development using the Sexual
Maturity Status line drawing instrument, which has demonstrated good internal consistency,
reliability and validity (Morris & Udry, 1980).
2.3.6. BMI.
Two independent measures of height (to nearest 0.1 cm) and weight (to nearest 0.1
kg) were obtained and averaged, and used to calculate a body mass index (BMI) for each
participant.
2.3.7. Cortisol
A measure of chronic systemic cortisol was obtained from a scalp hair sample.
Cortisol was assayed in hair samples in the current study to overcome the inherent difficulties
in interpreting point measures of cortisol from saliva (i.e., point samples vary as a function of
8
circadian rhythmicity and are susceptible to confounding by unmeasured environmental
factors). Hair samples were collected from an area approximately 1cm2 on the posterior
vertex of the scalp. Hair samples were not collected where <3 cm of hair length was
available, to minimize cosmetic impact, and thus fewer males participated than females (11
and 43, respectively, in the current sample). HCC assays were conducted by Stratech
Scientific, where samples were cut down to 3cm lengths, and processed and assayed using
Salimetrics ELISA kits (sensitivity <0.007μg/dL) with 50mg samples (see Simmons et al., In
Press for further details).
2.4. Statistical Analysis
Data were entered into SPSS version 21 (IBM Corp, 2012). An alpha level of .05 was
used for all statistical tests unless stated otherwise.
2.4.1. Descriptive statistics.
Univariate comparisons between males and females, and between Tanner Stage 1 and
Tanner Stage 2/3, for mean scores on each measure were performed using independent
groups t-tests. Bivariate correlations between all continuous variables were calculated
(Pearson r). Pituitary volumes, all hormone measures (testosterone, DHEA, DHEA-S and
cortisol), BMI, and SCAS total and subscale scores were positively skewed, and were log
transformed to normalize distributions.
2.4.2. Mediation analysis.
Two initial mediation analyses were performed to test hypotheses. For the first, SCAS
total score was the outcome variable (DV), with DHEA as the independent variable (IV). For
the second, SCAS total score was the outcome variable, and DHEA-S was the independent
variable. All analyses were performed using macros developed by Preacher and Hayes
(2004). Mediation was tested by assessing the significance of the cross product of the
coefficients for the IV (DHEA/S) to mediator (pituitary volume) relation (the a path), and the
mediator to DV (anxiety symptoms) relation, controlling for the IV (the b path). The a-b
cross product tests the statistical significance of the difference between the total effect (c
path), and the direct effect (c’ path), which is the impact of the IV on the DV, adjusting for
the effect of the mediator. See Figure 2 for an illustration of a simple mediation model. Age,
BMI, and sex were used as covariates in both models.
-- insert Figure 2 about here --
9
Because mediation requires significant associations between both the IV and the
mediator (the a path), and the mediator and the DV (the b path), two further exploratory
mediation analyses were conducted after results from the bivariate correlation analyses
revealed significant associations between pituitary volume and the SCAS social anxiety
subscale. In these models, SCAS social anxiety scores were used as the outcome variable,
and DHEA/S were IVs. Age, BMI and sex were once again used as covariates in both
models. In further analyses, sex was tested as a moderator of the mediation effect of PGV on
DHEA/S and symptoms of anxiety.
In all analyses, a bootstrapping method was applied to test the significance of the
mediation effect (Preacher & Hayes, 2004). Rather than imposing the distributional
assumptions required in parametric analyses, bootstrapping generates an empirical
approximation of the sampling distribution of a statistic by repeated random resampling from
the available data, and uses this distribution to calculate p values and construct bias-corrected
and accelerated confidence intervals (for these analyses, 5000 resamples were taken, and
95% confidence intervals were used) (see Hayes, 2009, for details). In these analyses,
mediation is significant at the level indicated (p < .05) if the confidence intervals for the
indirect effect do not contain zero.
This method has advantages over Baron and Kenny’s (1986) approach, which fails to
provide a specific test of the indirect effect of the IV on the DV through a proposed mediator.
Further, this method allowed a test for significant mediation in the absence of a direct effect
of the IV on the DV, which has been argued to be a possible, if not common, scenario
(Hayes, 2009). This result is feasible because a total effect is the sum of many different paths
of both direct and indirect influence, not all of which may be included in the formal model.
In order to exclude the possibility that any associations found were due to an effect of
physical signs of maturation (potentially indicative of early gonadarche), all analyses were re-
run using Tanner stage as a covariate. Testosterone was also included in further models to
assess the specificity of findings for DHEA/S. Finally, exploratory partial correlations
between all variables of interest and cortisol (controlling for BMI and age) were performed to
aid the interpretation of the results of primary analyses.
3. Results
3.1. Characteristics of Sample
10
Descriptive statistics for all continuous variables are shown in Table 1. Seventy
percent of females, and 69% of males were in Tanner Stage 1, 24/22% of females and males
were in Tanner Stage 2, and 4/2% of females and males were in Tanner Stage 3. Independent
samples t-tests revealed that boys had higher BMIs, were older, and had lower scores on the
Physical Injury subscale of the SCAS, and had higher cortisol levels. Tanner Stage did not
differ between males and females (Χ2 = 0.22 (2), p = 0.90). PGV did not differ between the
sexes, consistent with other research in children of a similar age (MacMaster et al., 2007).
Independent samples t-tests showed that participants in Tanner Stage 1 versus Stage 2 or 3
had lower BMI (p = 0.013), lower DHEA (p = 0.021), and higher obsessive-compulsive
disorder symptoms (p = 0.045).
-- insert Table 1 about here --
3.2. Correlations
Bivariate correlations for all continuous variables are presented in Table 2. DHEA,
DHEA-S and testosterone were significantly associated with one another, as well as with
PGV and BMI. Total scores on the SCAS were not significantly associated with any of the
hormone measures, or PGV. However, a significant positive association was demonstrated
between scores on the Social Anxiety subscale of the SCAS and PGV.
-- insert Table 2 about here –
3.3. Mediation Analyses
When controlling for age, sex, and BMI, higher levels of DHEA and DHEA-S were
associated with larger pituitary volumes (paths a; DHEA: B = .040, SE = .018, t = 2.21, p =
.030; DHEA-S: B = .046, SE = .016, t = 2.97, p = .004). However, PGV was not
significantly associated with symptoms of anxiety (paths b; p’s > 0.49), nor was there a
significant mediation effect in either model (i.e., the a b cross-products were not significant,
as indicated by the 95% confidence intervals containing zero).
Given the significant correlation between PGV and scores on the Social Anxiety
subscale of the SCAS (Table 2), further exploratory mediation analyses were performed
using symptoms of social anxiety as the outcome variable. When controlling for age, sex and
BMI, levels of both DHEA and DHEA-S predicted larger PGV (paths a), and larger PGV
11
was associated with increased symptoms of social anxiety (paths b). Neither the total effects
of DHEA and DHEA-S on social anxiety symptoms (paths c, i.e., not controlling for PGV),
nor the direct effects (paths c’, i.e., controlling for PGV) were significant. However, bias-
corrected 95% confidence intervals showed that larger PGV significantly partially mediated
the relationships between increased DHEA and DHEA-S and greater symptoms of social
anxiety (DHEA = CI: .004-.096, M = .036, SE = .023; DHEA-S = CI: .006-.108, M = .041,
SE = .025) (Figures 3 and 4). That is, there was an indirect path from high DHEA/S to high
social anxiety via larger PGV. Controlling for Tanner Stage did not change the pattern of
significant results for any model. See Figure S1 in Supplementary Data for scatterplots
illustrating bivariate associations between DHEA/S, PGV and social anxiety symptoms.
To evaluate the specificity of the mediation effect to social anxiety, differences in the
magnitude of the association between PGV and social versus other anxiety symptoms,
controlling for DHEA or DHEA-S (and other covariates), were tested by comparing partial
correlation coefficients using Steiger’s Z test (1-tailed; (Steiger, 1980). The association
between PGV and social anxiety symptoms was significantly different to that for all other
anxiety symptoms, with the exception of physical injury fears (DHEA: p = 0.067; DHEA-S;
p = 0.061).
-- insert Figure 3 about here --
-- insert Figure 4 about here --
3.4 Moderated Mediation Analyses
Because moderated mediation is possible in the absence of simple mediation
(Preacher & Hayes, 2004), the moderating role of sex in the initial mediation analysis was
tested, with total anxiety scores on the SCAS. There was no significant moderated mediation;
neither the interaction between DHEA and sex in the mediator model (B = -.035, SE = .036, t
= -.951, p = .344), nor the interaction between PGV and sex in the dependent variable model
(B = .205, SE = .800, t = .257, p = .798) were significant. When using DHEA-S as the
independent variable, results were also non-significant; both the interaction between DHEA-
S and sex in the mediator model (B = -.021, SE = .032, t = -.667, p = .506), and the
interaction between PGV and sex in the dependent variable model (B = .272, SE = .793, t =
.343, p = .732) were non-significant.
In subsequent moderated mediation analyses using social anxiety symptoms as the
outcome variable, sex did not significantly moderate the indirect effect of DHEA on
12
symptoms of social anxiety through PGV; neither the interaction between DHEA and sex in
the mediator model (B = -.035, SE = .036, t = -.951, p = .344), nor the interaction between
PGV and sex in the dependent variable model (B = .708, SE = .764, t = .926, p = .357) was
significant. No significant moderated mediation effect was demonstrated when using DHEA-
S as the independent variable; both the interaction between DHEA-S and sex in the mediator
model (B = -.021, SE = .032, t = -.667, p = .506), and the interaction between PGV and sex in
the dependent variable model (B = .714, SE = .761, t = .938, p = .351) were non-significant.
In further exploratory analyses, the moderating role of sex was tested in mediation
analyses with each of the other SCAS subscale scores as outcome variables. No significant
moderated mediation effects were found for any of the SCAS subscales when controlling for
BMI and age.
3.5 Effects of Testosterone
Given the strong associations between DHEA/S and testosterone, additional analyses
were performed to assess the specificity of the above findings to DHEA/S. After controlling
for testosterone levels (and other covariates, as above), the association between DHEA and
PGV weakened and was no longer significant (path a: B = .031, SE = .026, t = 1.22, p = .225;
controlling for Tanner Stage a: B = .031, SE = .026, t = 1.18, p = .240), whereas the
association between DHEA-S and PGV remained significant (path a: B = .040, SE = .017, t =
2.42, p = .018; controlling for Tanner Stage a: B = .037, SE = .017, t = 2.14, p = .035). In
models where testosterone was the sole hormonal predictor, after controlling for other
covariates, testosterone did not significantly predict PGV (path a: B = .098, SE = .052, t =
1.88, p = .064; controlling for Tanner Stage a: B = .099, SE = .054, t = 1.83, p = .071).
Combined, these results provide evidence for specificity of the effect of DHEA/S on PGV.
3.6 Associations with Cortisol
In the whole sample, hair cortisol was negatively associated with physical injury fears
(see Table 2), but was not significantly associated with any other variable. Exploratory partial
correlational analyses (controlling for age and BMI) conducted separately by sex revealed
significant and trend-level negative associations between cortisol and DHEA (r=-.70,
p=.038), PGV (r=-.52, p=.052) and social anxiety (r=-.69, p=.038) in males only.
13
4. Discussion
The precise neurobiological mechanisms linking adrenarcheal timing to symptoms of
anxiety are currently unclear. We investigated whether relatively advanced adrenarcheal
development (as indexed by high DHEA/S levels relative to age) was associated with larger
pituitary gland volumes (PGV), and whether increased PGV mediated the link between early
adrenarche and symptoms of anxiety in late childhood. The results of this unique study of
adrenarcheal timing demonstrate that children with relatively high DHEA/S levels showed
larger PGV. Whilst there was no direct effect of relative DHEA/S levels on general
symptoms of anxiety, the results did support the role of PGV in mediating the link between
relatively high DHEA/S levels and symptoms of social anxiety specifically. Interestingly, sex
did not moderate this mediation effect, nor were there any sex differences in the relationship
between DHEA/S and larger PGV.
Results provide evidence that children whose adrenarcheal development is relatively
early or mature demonstrate larger PGV than children whose adrenarcheal development is
relatively late or immature. This is the first study of its kind to investigate the association
between DHEA/S and PGV. That DHEA/S levels in children were found to be associated
with larger PGV is consistent with the suggestion that the onset of adrenarche is associated
with activation of the HPA axis, leading to increased volume and proliferation of corticotrope
cells in the pituitary gland, which ultimately stimulate the function of the adrenal cortex
(Axelson et al., 1992). These findings also suggest that the pituitary gland growth previously
attributed to activation of the HPG axis and release of sex hormones during gonadarche (Kao,
Cook, Hansen, & Simonson, 1992) likely commences prior to these processes, and is
influenced by the activation of the HPA axis.
Premature adrenarche has previously been associated with symptoms of anxiety
(Dorn et al., 1999; Dorn et al., 2008; Sontag-Padilla et al., 2012). Whilst the results did not
support a direct link between relatively high DHEA/S levels and anxiety symptoms, in
exploratory analyses, an indirect link was found whereby larger PGV partially mediated the
association between DHEA/S levels and social anxiety. Thus, the current findings may
support the notion that neurobiological factors partly underlie previously reported links
between early adrenarcheal timing and symptoms of anxiety in children. Further, these
findings are consistent with research demonstrating that larger PGV predicts increased
anxiety symptoms in early adolescence, both cross-sectionally and longitudinally (Zipursky
et al., 2010). Of note, relatively early adrenarche, as defined in our study as high DHEA/S
levels relative to age, should be differentiated from the clinical classification of premature
14
adrenarche. Our results suggest that adrenarche that is relatively early compared with peers,
but not necessarily clinically abnormal, has implications for mental health.
The exact mechanisms by which enlarged PGV might lead to increased symptoms of
social anxiety are unclear. It has been suggested that the action of pubertal hormones on the
development of the brain during sensitive periods may be responsible for dysregulated affect
and behavior during puberty (Blakemore et al., 2010). We have previously found an indirect
association between relatively high levels of adrenal hormones during adrenarche and
symptoms of mental illness via brain function (Whittle et al., 2015). As such, it is possible
that the current mediation findings suggest that relatively early exposure to adrenal hormones
during adrenarche may affect pituitary function, which in turn, influences neural vulnerability
to anxiety.
The pituitary gland is an integral part of the HPA axis, whose activity also mediates
circulating levels of stress hormones, including cortisol. While there is some evidence that
HPA axis production of cortisol and DHEA/S are independent processes (for example, the
increases in DHEA/S observed during adrenarche are not accompanied by similar increases
in cortisol; Wolf & Kirschbaum, 1999), it is reasonable to speculate that DHEA/S and
cortisol release may both be associated with pituitary gland volume, and that they may
interact in some way. We conducted exploratory analyses to examine whether DHEA/S and
PGV were associated with long-term cortisol levels in scalp hair. Although no associations
were found in the whole sample, in males, higher levels of DHEA and larger PGV were both
associated with lower cortisol levels; and, lower cortisol levels were associated with higher
social anxiety symptoms. These results must be interpreted with caution, as only 11 males
had cortisol data. Nonetheless, lower levels of hair cortisol might indicate a relative
hypocortisolism, which has been associated with increased stress sensitivity (Fries, Hesse,
Hellhammer, & Hellhammer, 2005). Further, it has been suggested that hypocortisolism and
associated stress sensitivity might create a vulnerability for the future development of
psychopathology (Heim, Newport, Mletzko, Miller, & Nemeroff, 2008). Indeed, other work
supports an association between childhood anxiety and lower basal HPA axis function
(Dieleman et al., 2014; Steudte et al., 2011). It is therefore possible that, at least in males, the
relationship between increased pituitary volume and higher levels of social anxiety might
reflect dysfunction of the HPA axis. We emphasize that this interpretation is speculative as it
is based on a very small number of participants. Further research with larger numbers and
assessment of long-term basal cortisol levels, in addition to cortisol reactivity is needed to
comprehensively investigate function of the HPA axis as a potential mechanism linking
15
relatively early exposure to DHEA/S during adrenarche, PGV and anxiety. Indeed, there is
other evidence that larger pituitary volumes are related to relatively higher activity and
reactivity of the HPA axis (Kaess et al., 2013), and that hyper-reactivity of the HPA axis is
associated with social anxiety (Condren, O'Neill, Ryan, Barrett, & Thakore, 2002).
Interestingly, pituitary volume was found to partially mediate the relationship
between DHEA/S levels and symptoms of social anxiety specifically. This may reflect the
functional significance of pubertal development, which is thought to influence the
development of social brain functions specifically due to the role of puberty in preparing the
individual to enter into new types of social relationships (e.g., romantic relationships,
competition for mates, Steinberg & Morris, 2000). Indeed, previous studies have
demonstrated specific effects of pubertal development on neural measures of social emotion
processing (Goddings, Burnett Heyes, Bird, Viner, & Blakemore, 2012). The current findings
further suggest that, even at the early adrenarcheal phase of development, there may be some
specificity in the relationship between the hormonal and neural changes occurring and social
emotions.
A possible mechanistic explanation for the mediatory effect of pituitary volume on
symptoms specific to social anxiety comes from research linking HPA axis reactivity to stress
and social avoidance (Sapolsky, 1990). An alternative explanation for the findings regarding
social anxiety comes from research investigating the impact of premature adrenarche on
executive function and psychopathology. Executive function represents the coordination of
higher-order cognition and behavior, for example problem-solving, self-evaluation and
regulation, and behavioral modulation. Executive function also promotes the development of
social cognition, which may be particularly important in navigating difficult developmental
transitions like puberty (Kolb & Fantie, 2009). Indeed, research has demonstrated consistent
associations between less developed executive function and social and emotional problems
(Miller & Hinshaw, 2010; Wahlstedt, Thorell, & Bohlin, 2008). Further, research has shown
that children undergoing premature adrenarche perform more poorly on tasks of executive
function than their normally developing peers (Dorn et al., 1999). It is therefore possible that
children undergoing relatively early adrenarche may be at risk for social anxiety due to a
potentially reduced ability to manage the affective and psychosocial demands of their
environment.
Interestingly, there were no sex differences in the associations between relative
DHEA/S levels and pituitary volume, or pituitary volume and symptoms of anxiety.
Neuroimaging research into structural brain development during gonadarche has reported sex
16
differences in global and regional grey and white matter to be associated with differential
levels of sex hormones, including testosterone and estradiol (Peper et al., 2009; Peper et al.,
2008; Perrin et al., 2009). Further, research has demonstrated that the effects of sex hormones
on PGV differ between boys and girls during gonadarche (Peper et al., 2010; Whittle et al.,
2012). The adrenal androgens (DHEA/S) are considered to be weak androgens, and act as
precursors to the production of more potent estrogens and testosterone (Dorn et al., 1999).
The plasma concentration of DHEA/S during adrenarche is also similar in boys and girls
(Ilondo et al., 1982). Our results do suggest that while levels of DHEA/S influence pituitary
volume during adrenarche, observable sex differences in pituitary volume may not emerge
until gonadarche, as a result of the more potent sex hormones and their effects on the brain,
but also increasing levels of DHEA/S through adolescence.
The lack of sex differences in the associations between pituitary volume and
symptoms of anxiety may also be a result of similar DHEA/S plasma concentration in boys
and girls during adrenarche. It is also possible that pituitary volume during adrenarche is not
an early precursor that underlies later sex differences in symptoms of anxiety and related
disorders. Sex differences in these relationships might only emerge following the onset of
gonadarche and differential exposure to potent sex hormones, when an increase in female
vulnerability to anxiety symptoms and disorders become more pronounced (Slade, Johnston,
Oakley Browne, Andrews, & Whiteford, 2009). We note our preliminary findings of a male-
specific association between DHEA, PGV, social anxiety and cortisol, which are suggestive
of potential sex differences in the mechanisms underlying the observed results. Further
research with larger samples is required to investigate this possibility.
Our findings should be interpreted in light of the following limitations. First, we
employed a cross-sectional design, which reduces the ability to make assumptions about the
temporal ordering of effects. Although not part of the current theoretical model, it is plausible
that pituitary volume may in fact influence the timing of DHEA/S production. Future work
assessing baseline pituitary volumes before the onset of adrenarcheal processes would help to
clarify this. Second, there may be other factors that are important in predicting levels of
DHEA/S and cortisol, pituitary volume, and anxiety symptoms that we did not consider. For
example, early life stress may be responsible for early adrenarche (Ellis & Essex, 2007) and
larger PGV (Ganella et al., 2015), and thus might better explain the link between PGV and
social anxiety (rather than this link being driven by early exposure to adrenal hormones).
Future work should therefore incorporate measures of early life stress, as well as other factors
like socioeconomic status and birth related factors, to explore their potential influences.
17
Third, we measured adrenarcheal timing by specifically selecting individuals who were
matched for age but differed in their levels of DHEA/S. Although age was ultimately
significantly correlated with DHEA/S, it was controlled for in all analyses, which allowed
stronger conclusions to be drawn about relative timing of exposure to DHEA/S (i.e.,
adrenarcheal timing). However, further research is needed that measures changes in DHEA/S
exposure over time, to investigate relationships amongst the variables in question more
comprehensively. Fourth, relatively early adrenarche is likely to influence other brain regions
and brain networks that were not investigated in the current study. For example, specific
vulnerabilities during this developmental stage might represent the failure of HPA-axis
control by prefrontal regions, which are undergoing dramatic developmental changes during
this period (Herman et al., 2003; Sullivan & Gratton, 2002). These speculations require
further investigation with measures of brain structure and function. Fifth, the main mediation
findings of this study were based on exploratory correlational analyses; we had no specific
hypothesis about social anxiety, specifically, from the outset. As such, these results should be
interpreted with caution, and future research should aim to replicate the results in order to
corroborate the findings. Indeed, there was some evidence that the findings were not specific
to social anxiety (i.e., analyses suggested that PGV may also be associated with physical
injury fears). Finally, the speculations regarding the role of stress and the HPA-axis in the
link between pituitary volume and symptoms of anxiety require follow-up research given the
small number of male participants for whom cortisol measures were available.
5. Conclusions
Our results provide evidence that PGV mediates the relationship between relatively
early adrenarcheal timing and symptoms of social anxiety in children. This finding suggests
that neurobiological mechanisms are partly responsible for the link between early
adrenarcheal timing and increased symptoms of social anxiety. Further research is required to
investigate the mechanisms by which an enlarged pituitary gland in children with relatively
early exposure to DHEA/S may be associated with symptoms of social anxiety, including
stress reactivity. Given that children with increased symptoms of anxiety are at risk for the
development of depressive and anxiety disorders during adolescence and adulthood
(Zipursky et al., 2010), these results may have implications for prevention and early
intervention strategies during childhood. In particular, our findings suggest that the
neurobiological vulnerabilities to adverse mental health outcomes that have been
demonstrated in pubertally-advanced individuals emerge as early as adrenarche. Therefore,
preventive strategies that target early maturing boys and girls with early symptoms of
18
anxiety, before they enter gonadarche, may be effective in preventing the onset of more
clinically-significant mental health problems.
Conflicts of Interest
For all authors: Conflicts of interest: none.
Contributors
All named authors have contributed substantially to the intellectual content of the manuscript.
SW, NBA, GP and JGS had input into the design of the study. MLB contributed to
acquisition of the data. CM and SW were responsible for data analysis, interpretation of data
and drafting the article. All authors participated in revising the manuscript critically for
important intellectual content. All authors have given final approval of the version to be
submitted.
Role of funding sources
No funding source had any input into the design or conduct of the study, or into the
drafting of the manuscript.
6. Acknowledgements
This work was supported by the Australian Research Council (Discovery Project
grant: DP120101402; Fellowship DORA DP 130101459 to CO), and the National Health and
Medial Council (NHMRC; Career Development Fellowship ID 1007716 to SW). We would
like to thank all of the families who have participated in this study, and all investigators and
staff working on the Childhood to Adolescence Transition Study (CATS) at the Murdoch
Childrens Research Institute (MCRI) for making this investigation possible. MCRI research
is supported by the Victorian Government’s Operational Infrastructure Program. We would
also like to thank the research staff who contributed to the collection of research data (Dr
Megan Dennison, Dr Rebecca Davenport-Thomas and Dr Rachel Ellis).
19
7. References
Albano, A. M., Chorpita, B. F., & Barlow, D. H. (2003). Childhood anxiety disorders. In E. J.
Mash & R. A. Barkley (Eds.), Child psychopathology (pp. 279-329). New York:
Guilford Press.
Axelson, D. A., Doraiswamy, P. M., Boyko, O. B., Rodrigo Escalona, P. R., McDonald, W.
M., Ritchie, J. C., . . . Krishnan, K. R. (1992). In vivo assessment of pituitary volume
with magnetic-resonance-imaging and systematic stereology: Relationship to
dexamethasone suppression test results in patients. Psychiatry Research, 44, 63-70.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategies, and statistical considerations. Journal
Of Personality And Social Psychology, 51, 1173-1182.
Bernstein, G. A., Borchardt, C. M., & Perwien, A. R. (1996). Anxiety disorders in children
and adolescents: A review of the past 10 years. Journal of the American Academy of
Child & Adolescent Psychiatry, 35(9), 1110-1119.
Blakemore, S.-J., Burnett, S., & Dahl, R. E. (2010). The Role of Puberty in the Developing
Adolescent Brain. Human Brain Mapping, 31, 926-933.
Condren, R. M., O'Neill, A., Ryan, M. C. M., Barrett, P., & Thakore, J. H. (2002). HPA axis
response to a psychological stressor in generalised social phobia. .
Psychoneuroendocrinology, 27(6), 693-703.
Dieleman, G. C., Huizink, A. C., Tulen, J. H., Utens, E. M., Creemers, H. E., van der Ende,
J., & Verhulst, F. C. (2014). Alterations in HPA-axis and autonomic nervous system
functioning in childhood anxiety disorders point to a chronic stress hypothesis.
Psychoneuroendocrinology.
Dorn, L. D., Hitt, S. F., & Rotenstein, D. (1999). Biopsychological and Cognitive Differences
in Children With Premature vs On-Time Adrenarche. Archives of Pediatric and
Adolescent Medicine, 153(2), 137-146.
Dorn, L. D., Rose, S. R., Rotenstein, D., Susman, E. J., Huang, B., Loucks, T. L., & Berga, S.
L. (2008). Differences in endocrine parameters and psychopathology in girls with
premature adrenarche versus on-time adrenarche. Journal of Pediatric Endocrinology
and Metabolism, 21(5), 439-448.
Ellis, B. J., & Essex, M. J. (2007). Family environments, adrenarche, and sexual maturation:
a longitudinal test of a life history model. Child development, 78(6), 1799-1817.
20
Fries, E., Hesse, J., Hellhammer, J., & Hellhammer, D. H. (2005). A new view on
hypocortisolism. Psychoneuroendocrinology, 30(10), 1010-1016.
Ganella, D. E., Allen, N. B., Simmons, J. G., Schwartz, O., Kim, J. H., Sheeber, L., &
Whittle, S. (2015). Early life stress alters pituitary growth during adolescence—A
longitudinal study. Psychoneuroendocrinology, 53, 185-194.
Ge, X., Conger, R. D., & Elder, G. H. J. (1996). Coming of age too early: Pubertal influences
on girls’ vulnerability to psychological distress. Child Development, 67, 3386-3400.
Ge, X., Conger, R. D., & Elder Jr., G. H. (2001). Pubertal transition, stressful life events, and
the emergence of gender differences in adolescent depressive symptoms. Dev
Psychol, 37, 404-417.
Ge, X., & Natsuaki, M. N. (2009). In Search of Explanations for Early Pubertal Timing
Effects on Developmental Psychopathology. Current Directions in Psychological
Science, 18(6), 327-331.
Goddings, A. L., Burnett Heyes, S., Bird, G., Viner, R. M., & Blakemore, S. J. (2012). The
relationship between puberty and social emotion processing. Developmental science,
15(6), 801-811.
Hayes, A. F. (2009). Beyond Baron and Kenny: statistical mediation analysis in the new
millennium. Commun. Monogr.(76).
Heim, C., Newport, D. J., Mletzko, T., Miller, A. H., & Nemeroff, C. B. (2008). The link
between childhood trauma and depression: insights from HPA axis studies in humans.
Psychoneuroendocrinology, 33(6), 693-710.
Herman, J. P., Figueiredo, H., Mueller, N. K., Ulrich-Lai, Y., Ostrander, M. M., Choi, D. C.,
& Cullinan, W. E. (2003). Central mechanisms of stress integration: hierarchical
circuitry controlling hypothalamo—pituitary—adrenocortical responsiveness. Front.
Neuroendocrinol, 24, 151-180.
Hyde, J. S., Mezulis, A. H., & Abramson, L. Y. (2008). The ABCs of depression: integrating
affective, biological, and cognitive models to explain the emergence of the gender
difference in depression. Psychological review, 115(2), 291.
IBM Corp. (2012). IBM SPSS Statistics for Macintosh, Version 21.0. Armonk, NY: IBM
Corp.
Ilondo, M. M., Vanderschueren-Lodeweyckx, M., Vlietinck, R., Pizarro, M., Malvaux, P.,
Eggermont, E., & Eeckels, R. (1982). Plasma androgens in children and adolescents.
Part I: control subjects. Horm Res, 16(2), 61-77.
21
Kaess, M., Simmons, J. G., Whittle, S., Jovev, M., Chanen, A. M., Yücel, M., . . . Allen, N.
B. (2013). Sex-specific prediction of hypothalamic-pituitary-adrenal axis activity by
pituitary volume during adolescence: A longitudinal study from 12 to 17 years of age.
Psychoneuroendocrinology, 38(11), 2694-2704.
Kao, S. C. S., Cook, J. S., Hansen, J. R., & Simonson, T. M. (1992). MR imaging of the
pituitary gland in central precocious puberty. Pediatric Radiology, 22, 481-484.
Klauser, P., Whittle, S., Simmons, J. G., Byrne, M. L., Mundy, L. K., Patton, G. C., . . .
Allen, N. B. (2015). Reduced frontal white matter volume in children with early onset
of adrenarche. Psychoneuroendocrinology, 52, 111-118.
Kolb, B., & Fantie, B. D. (2009). Development of the child’s brain and behavior. In R. C.R.
& E. Fletcher-Janzen (Eds.), Handbook of clinical child neuropsychology (3rd ed., pp.
19-46). New York: Springer.
MacMaster, F. P., Keshavan, M., Mirza, Y., Carrey, N., Upadhyaya, A. R., El-Sheikh, R., . . .
al., e. (2007). Development and sexual dimorphism of the pituitary gland. Life
sciences, 80(10), 940-944.
MacMaster, F. P., & Kusumakar, V. (2004). MRI study of the pituitary gland in adolescent
depression. Journal of Psychiatry Research, 38, 231-236.
MacMaster, F. P., Russell, A., Mirza, Y., Keshavan, M. S., Taormina, S. P., Bhandari, R., . . .
Rosenberg, D. R. (2006). Pituitary volume in treatment-naive pediatric major
depressive disorder. Biological Psychiatry, 60, 862-866.
Miller, M., & Hinshaw, S. P. (2010). Does childhood executive function predict adolescent
functional outcomes in girls with ADHD? Journal of Abnormal Child Psychology, 38,
315-326.
Morris, N. M., & Udry, J. R. (1980). Validation of a Self-Administered Instrument to Assess
Stage of Adolescent Development. Journal of Youth and Adolescence, 9(3), 271-280.
Mundy, L. K., Simmons, J. G., Allen, N. B., Viner, R. M., Bayer, J. K., Olds, T., . . . Patton,
G. C. (2013). Study protocol: the Childhood to Adolescence Transition Study
(CATS). BMC Pediatr, 13, 160. doi: 10.1186/1471-2431-13-160
Muris, P., Merckelbach, H., Ollendick, T., King, N., & Bogie, N. (2002). Three traditional
and three new childhood anxiety questionnaires: their reliability and validity in a
normal adolescent sample. Behaviour research and therapy, 40(7), 753-772.
Pariante, C. M., Vassilopoulou, K., Velakoulis, D., Phillips, L., Soulsby, B., Wood, S. J., . . .
Pantelis, C. (2004). Pituitary volume in psychosis. Brit. J. Psychiatr, 185, 5-10.
Patton, G. C., & Viner, R. (2007). Pubertal transitions in health. The Lancet, 369, 1130-1139.
22
Peper, J. S., Brouwer, R. M., Schnack, H. G., van Baal, G. C., van Leeuwen, M., van den
Berg, S. M., . . . Hulshoff Pol, H. E. (2009). Sex steroids and brain structure in
pubertal boys and girls. Psychoneuroendocrinology, 34, 332-342.
Peper, J. S., Brouwer, R. M., Schnack, H. G., van Baal, G. C., van Leeuwen, M., Van den
Berg, S. M., . . . Hulshoff Pol, H. E. (2008). Cerebral white matter in early puberty is
associated with luteinizing hormone concentrations. Psychoneuroendocrinology, 33,
909-915.
Peper, J. S., Brouwer, R. M., van Leeuwen, M., Schnack, H. G., Boomsma, D. I., Kahn, R.
S., & Pol, H. E. H. (2010). HPG-axis hormones during puberty: a study on the
association with hypothalamic and pituitary volumes. Psychoneuroendocrinology, 35,
133-140.
Perrin, J. S., Leonard, G., Perron, M., Pike, G. B., Pitiot, A., Richer, L., . . . Paus, T. (2009).
Sex differences in the growth of white matter during adolescence. Neuroimage, 45,
1055-1066.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect
effects in simple mediation models. Behav. Res. Methods Instrum. Comput., 36, 717-
773.
Sapolsky, R. M. (1990). Adrenocortical function, social rank, and personality among wild
baboons. Biological Psychiatry, 28, 862-878.
Simmons, J. G., Badcock, P. B., Whittle, S. L., Byrne, M. L., Mundy, L., Patton, G. C., . . .
Allen, N. B. (In Press). The lifetime experience of traumatic events is associated with
hair cortisol concentrations in community-based children.
Psychoneuroendocrinology.
Simmons, J. G., Whittle, S. L., Patton, G. C., Dudgeon, P., Olsson, C., Byrne, M. L., . . .
Allen, N. B. (2014). Study protocol: Imaging brain development in the Childhood to
Adolescence Transition Study (iCATS). BMC Pediatr, 14, 115. doi: 10.1186/1471-
2431-14-115
Slade, T., Johnston, A., Oakley Browne, M. A., Andrews, G., & Whiteford, H. (2009). 2007
National Survey of Mental Health and Wellbeing: methods and key findings.
Australasian Psychiatry, 43(7), 594-605.
Sontag-Padilla, L. M., Dorn, L. D., Tissot, A., Susman, E. J., Beers, S. R., & Rose, S. R.
(2012). Executive functioning, cortisol reactivity, and symptoms of psychopathology
in girls with premature adrenarche. Dev Psychopathol, 24(1), 211-223.
23
Spence, S. H. (1998). A measure of anxiety symptoms among children. Behaviour research
and therapy, 36(5), 545-566.
Steiger, J. H. (1980). Tests for comparing elements of a correlation matrix. Psychological
Bulletin, 87(2), 245.
Steinberg, L., & Morris, A. (2000). Adolescent development. Annual Review of
Psychology(52), 83–110.
Steudte, S., Stalder, T., Dettenborn, L., Klumbies, E., Foley, P., Beesdo-Baum, K., &
Kirschbaum, C. (2011). Decreased hair cortisol concentrations in generalised anxiety
disorder. Psychiatry Research, 186(2), 310-314.
Styne, D. M., & Grumbach, M. M. (2011). Puberty: Ontogeny, Neuroendocrinology,
Physiology, and Disorders. In H. M. Kronenberg, S. Melmed, S. P. Polonsky, & P. R.
Larsen (Eds.), Williams Textbook of Endocrinology (Twelfth ed., pp. 1054-1201):
Elsevier Saunders.
Sullivan, R. M., & Gratton, A. (2002). Prefrontal cortical regulation of hypothalamic—
pituitary—adrenal function in the rat and implications for psychopathology: side
matters. Psychoneuroendocrinology, 27, 99-114.
Thomas, L. A., & De Bellis, M. D. (2004). Pituitary volumes in pediatric maltreatment-
related posttraumatic stress disorder. Biological Psychiatry, 55(7), 752-758.
Van Honk, J., Tuiten, A., van den Hout, M., Koppeschaar, H., Thijssen, J., de Haan, E., &
Verbaten, R. (2000). Conscious and preconscious selective attention to social threat:
Different neuroendocrine response patterns. Psychoneuroendocrinology, 25, 577-591.
Wahlstedt, C., Thorell, L. B., & Bohlin, G. (2008). ADHD symptoms and executive function
impairment: Early predictors of later behavioral problems. Developmental
Neuropsychology, 33, 160-178.
Whittle, S., Simmons, J. G., Byrne, M. L., Strikwerda-Brown, C., Kerestes, R., Seal, M. L., . .
. Patton, G. C. (2015). Associations between early adrenarche, affective brain function
and mental health in children. Social cognitive and affective neuroscience, nsv014.
Whittle, S., Yucel, M., Lorenzetti, V., Byrne, M. L., Simmons, J. G., Wood, S. J., . . . Allen,
N. B. (2012). Pituitary volume mediates the relationship between pubertal timing and
depressive symptoms during adolescence. Psychoneuroendocrinology, 37, 881-891.
Wolf, O. T., & Kirschbaum, C. (1999). Actions of dehydroepiandrosterone and its sulfate in
the central nervous system: effects on cognition and emotion in animals and humans.
Brain Res Brain Res Rev, 30(3), 264-288.
24
Zipursky, A. R., Whittle, S., Yucel, M., Lorenzetti, V., Wood, S. J., Lubman, D. I., . . . Allen,
N. B. (2010). Pituitary volume prospectively predicts internalizing symptoms in
adolescence. Journal of Child Psychology and Psychiatry, 52(3), 315-323.
25
8. Figure legends
Figure 1. Traced pituitary gland on coronal slice of MR image.
Figure 2. Simple mediation model. Significant mediation is indicated if the a-b cross product
(equivalent to c - c’) is significantly greater than zero.
b
a
Mediator
DV IV
IV DV
c’
c
26
Figure 3. Mediation analysis. Path a is the total effect of DHEA on pituitary volume, path b is
the total effect of pituitary volume on symptoms of social anxiety, path c is the total effect of
DHEA on symptoms of social anxiety (i.e., not controlling for pituitary volume), and path c’
is the direct effect of DHEA on symptoms of social anxiety (i.e., controlling for pituitary
gland volume). *p <0.05
Figure 4. Mediation analysis. Path a is the total effect of DHEA-S on pituitary volume, path b
is the total effect of pituitary volume on symptoms of social anxiety, path c is the total effect
of DHEA-S on symptoms of social anxiety (i.e., not controlling for pituitary volume), and
path c’ is the direct effect of DHEA-S on symptoms of social anxiety (i.e., controlling for
pituitary gland volume).*p <0.05
b: B = .912, SE = .378,
t = 2.40, p = .018* a: B = .040, SE = .018,
t = 2.21, p = .030*
Pituitary Gland
Volume
SCAS Social Anxiety DHEA
DHEA SCAS Social Anxiety
c’: B = .011, SE = .066,
t = .164, p = .870
c: B = .047, SE = .066,
t = .717, p = .475
DHEA-S
Pituitary Gland
Volume
SCAS Social Anxiety
DHEA-S SCAS Social Anxiety
a: B = .046, SE = .016,
t = 2.97, p = .004*
b: B = .894, SE = .388,
t = 2.31, p = .024*
c’: B = .016, SE = .059,
t = .279, p = .781
c: B = .057, SE = .057,
t = 1.00, p = .319
27
9. Tables
Table 1. Descriptive statistics for all variables and tests for sex differences.
Male (n=45) Female (n=50) Total t
M (SD) M (SD) M (SD)
Testosterone
(pg/ml)
44.71 (15.18) 48.28 (23.35) 46.59 (19.89) 0.86
DHEA (pg/ml) 41.57 (30.92) 69.81 (96.98) 56.45 (74.53 1.85
DHEA-S (pg/ml) 29.16 (42.59) 48.49 (106.40) 39.44 (83.05) 1.13
Cortisol (pg/mg) 4.99 (3.54) 2.73 (2.51) 3.19 (2.86) 2.56*
PGV 273.20 (49.89) 279.80 (62.18) 276.67 (56.50) 0.57
SCAS Total 18.01 (10.22) 20.26 (12.48) 19.22 (11.47) 0.93
Separation 2.91 (2.51) 3.18 (2.60) 3.05 (2.55) 0.51
Social 2.91 (2.67) 4.04 (3.19) 3.51 (3.00) 1.85
OCD 4.34 (3.03) 3.58 (2.96) 3.94 (3.00) -1.22
Panic 1.89 (1.67) 1.82 (2.57) 1.85 (2.18) -1.46
Physical 2.45 (2.17) 3.50 (2.10) 3.01 (2.19 2.37*
GAD 3.50 (2.16) 4.16 (3.03) 3.85 (2.67) 1.20
Age 9.57 (0.37) 9.42 (0.30) 9.50 (0.34) -
2.30* BMI 17.09 (1.90) 18.04 (2.51) 17.6 (2.28) 2.08*
*p <0.05
Note. BMI = Body Mass Index; PGV = Pituitary Gland Volume; DHEA =
Dehydroepiandroesterone; DHEAS = Dehydroepiandrosterone-sulphate; SCAS Total = Total
score on the Spence Children’s Anxiety Scale; Separation = SCAS Separation Anxiety
subscale; Social = SCAS Social Anxiety subscale; OCD = SCAS Obsessive Compulsive
Disorder subscale; Panic = SCAS Panic and Agoraphobia subscale; Physical = SCAS
Physical Injury subscale; GAD = SCAS Generalized Anxiety subscale.
28
Table 2. Bivariate correlations between all variables.
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1. Sex
2. Age .23*
3. BMI -.21* .19
4. PGV -.05 .13 .19
5. DHEA -.05 .24* .38** .28**
6. DHEA-S -.03 .21* .21* .32** .56**
7. TST -.04 .14 .34** .24* .75** .44**
8. SCAS Total -.08 .07 .11 .11 .10 .15 -.01
9. Separation -.062 .025 -.00 -.07 -.03 .07 -.09 .76**
10. Social -.22* .10 .14 .29** .14 .16 .014 .82** .56**
11. OCD .13 .08 .08 .10 .03 .09 -.05 .71** .43** .53**
12. Panic .12 .13 -.03 .08 .16 .13 .09 .59** .54** .45** .32**
13. Physical -.30** .00 .20 .11 .03 .08 -.01 .58** .41** .54** .32** .17
14. GAD -.10 .02 .16 .13 .17 .17 .12 .71** .42** .55** .44** .38** .35**
15. Cortisol .34* .12 -.26 .02 -.16 .11 -.04 -.18 -.08 -.18 -.08 -.04 -.39** -.06
*Correlation significant at p <.05 (2-tailed). **Correlation significant at p <.01 (2-tailed).
Note. BMI = Body Mass Index; PGV = Pituitary Gland Volume; DHEA = Dehydroepiandroesterone; DHEAS = Dehydroepiandrosterone-
sulphate; TST = Testosterone; SCAS Total = Total score on the Spence Children’s Anxiety Scale; Separation = SCAS Separation Anxiety
subscale; Social = SCAS Social Anxiety subscale; OCD = SCAS Obsessive Compulsive Disorder subscale; Panic = SCAS Panic and
Agoraphobia subscale; Physical = SCAS Physical Injury subscale; GAD = SCAS Generalized Anxiety subscale