29
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

Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

  • Upload
    others

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 2: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

[email protected]

Page 3: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 4: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 5: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 6: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 7: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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,

Page 8: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 9: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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 --

Page 10: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 11: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 12: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 13: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 14: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 15: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 16: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 17: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 18: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 19: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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).

Page 20: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 21: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 22: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 23: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 24: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 25: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 26: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 27: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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

Page 28: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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.

Page 29: Associations between dehydroepiandrost erone (DHEA) levels, …dro.deakin.edu.au/eserv/DU:30079725/olsson-associations... · DRO Deakin Research Online, Deakin University’s Research

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