Effects of Dehydroepiandrosterone Therapy on Pubic Hair Growth ...

ORIGINAL ARTICLE
Endocrine Care
<strong>Effects</strong> <strong>of</strong> <strong>Dehydroepiandrosterone</strong> <strong>Therapy</strong> on Pubic
Hair Growth and Psychological Well-Being in
Adolescent Girls and Young Women with Central
Adrenal Insufficiency: A Double-Blind, Randomized,
Placebo-Controlled Phase III Trial
G. Binder, S. Weber, M. Ehrismann, N. Zaiser, C. Meisner, M. B. Ranke, L. Maier,
S. A. Wudy, M. F. Hartmann, U. Heinrich, M. Bettendorf, H. G. Doerr, R. W. Pfaeffle,
E. Keller, and the South German Working Group for Pediatric Endocrinology*
Pediatric Endocrinology (G.B., S.W., M.E., M.B.R.), University-Children’s Hospital, 72076 Tuebingen, Germany; Center
for Coordination <strong>of</strong> Clinical Studies (N.Z.) and Department <strong>of</strong> Medical Biometry (C.M.), University <strong>of</strong> Tuebingen, 72076
Tuebingen, Germany; Pharmacy (L.M.), University <strong>of</strong> Ulm, 89081 Ulm, Germany; Justus Liebig University, Steroid
Research and Mass Spectrometry Unit (S.A.W., M.F.H.), Centre <strong>of</strong> Child and Adolescent Medicine, 35392 Giessen,
Germany; Pediatric Endocrinology (U.H., M.B.), University-Children’s Hospital, 69120 Heidelberg, Germany; Pediatric
Endocrinology (H.G.D.), University-Children’s Hospital, D-91054 Erlangen, Germany; and Pediatric Endocrinology
(R.W.P., E.K.), University-Children’s Hospital, 04317 Leipzig, Germany
Context and Objective: The efficacy <strong>of</strong> oral dehydroepiandrosterone (DHEA) in the treatment <strong>of</strong>
atrichia pubis and psychological distress in young females with central adrenal insufficiency is
unknown. Our study aimed to evaluate this therapy.
Design and Patients: A total <strong>of</strong> 23 young females (mean age 18 yr, range 13–25) was enrolled in
a double-blind randomized placebo-controlled trial. Inclusion criteria were ACTH deficiency plus
two or more additional pituitary deficiencies, serum DHEA less than 400 ng/ml, and pubertal stage
more than B2. Exclusion criteria were cerebral radiation with more than 30 Gy, tumor remission
less than 1 yr, amaurosis, hypothalamic obesity, psychiatric disorders, and unstable hormone medication.
Intervention: Patients were randomized to placebo (n 12) or 25 mg HPLC-purified DHEA/d (n
11) orally for 12 months after stratification into a nontumor (n 7) and a tumor group (n 16).
Main Outcome Measures: Clinical scoring <strong>of</strong> pubic hair stage was performed at 0, 6, and 12 months
(primary endpoint), and psychometrical evaluation (Symptom Check-List-90-R and the Centre for
Epidemiological Studies-Depression Scale) at 0 and 12 months (secondary endpoint). Androgen
levels and safety parameters were measured at 0, 6, and 12 months; 24-h androgen urinary excretion
rates were calculated at 0 and 12 months.
Results: In the placebo group, four patients dropped out because <strong>of</strong> recurrence <strong>of</strong> craniopharyngioma,
manifestation <strong>of</strong> type 1 diabetes, and change <strong>of</strong> residence (n 2); in the DHEA group, one
patient dropped out because <strong>of</strong> recurrent anxiety attacks. DHEA substitution resulted in normalization
<strong>of</strong> DHEA sulfate and androstanediol glucuronide morning serum levels 2 h after drug intake
(P 0.006), and <strong>of</strong> its 24 h urinary metabolite levels (P 0.0001), placebo had no effect. Morning
serum levels <strong>of</strong> androstenedione increased in the DHEA group (P 0.02) but did not normalize. The
DHEA group exhibited significant progress in pubic hair growth from Tanner stage I–III to II–V
(mean: 1.5 stages), whereas the placebo group did not (relative risk 0.138; 95% confidence
interval 0.021–0.914; P 0.0046). Importantly, eight <strong>of</strong> the 10 Symptom Check-List-90-R scores, including
those for depression, anxiety, and interpersonal sensitivity, and the global severity index improved
in the DHEA group in comparison to the placebo group (P 0.048). DHEA was well tolerated.
Conclusions: In adolescent girls with central adrenal insufficiency, daily replacement with 25 mg
DHEA orally is beneficial: atrichia pubis vanishes, and psychological well-being improves
significantly. (J Clin Endocrinol Metab 94: 1182–1190, 2009)
1182 jcem.endojournals.org J Clin Endocrinol Metab. April 2009, 94(4):1182–1190

J Clin Endocrinol Metab, April 2009, 94(4):1182–1190 jcem.endojournals.org 1183
ACTH deficiency causes a deficient supply <strong>of</strong> the adrenal hormones
cortisol and dehydroepiandrosterone (DHEA). Although
replacement <strong>of</strong> cortisol is standard in endocrine care for
children and adults, treatment <strong>of</strong> DHEA deficiency is still a matter
<strong>of</strong> discussion (1). Impaired health-related quality <strong>of</strong> life in adults
with adrenal insufficiency was reported, suggesting insufficient replacement
therapy (2). Such data are missing for childhood and
adolescence.
DHEA is a prohormone <strong>of</strong> androgens and estrogens, and by
95% <strong>of</strong> adrenal origin (3). In early childhood, DHEA production
is low until it gradually starts to increase between 6 and 10 yr <strong>of</strong>
age. This phenomenon, called adrenarche, is associated with expansion
<strong>of</strong> the adrenal zona reticularis (3). Peak DHEA serum concentration
is reached in women at about 24 yr <strong>of</strong> age, followed by
a steady decline, down to around 20% <strong>of</strong> the peak value (4).
DHEA and its sulfate ester, DHEA sulfate (DHEAS), are secreted
in very large amounts, and are the endocrine steroids with
the highest serum concentration in humans. Many peripheral
target tissues are able to convert DHEA into sex hormones (5).
While androgens in males are mainly produced in the testes independent
from DHEA production <strong>of</strong> the adrenals, androgens in
females are to more than 75% conversion products <strong>of</strong> DHEA (6).
Therefore, DHEA deficiency results in androgen deficiency in
females (7). Childhood onset <strong>of</strong> complete adrenal insufficiency is
regularly associated with atrichia pubis and lack <strong>of</strong> axillary hair
in females because growth <strong>of</strong> sexual hair in females is completely
dependent on the presence <strong>of</strong> the prohormone DHEA and its
conversion product dihydrotestosterone (8).
Prospective long-term studies on the consequences <strong>of</strong> chronic
DHEA deficiency with early onset do not exist. Randomized
short-term studies in adults with adrenal insufficiency suggested
a good safety pr<strong>of</strong>ile <strong>of</strong> DHEA at doses <strong>of</strong> 25–50 mg/d orally,
which were able to normalize DHEA(S) serum levels in healthy
controls after suppression <strong>of</strong> the adrenal gland (9). Positive effects
<strong>of</strong> DHEA on well-being and sexuality in adult females with
adrenal insufficiency that were shown in the first controlled
short-term study (10) were confirmed in most <strong>of</strong> the following
short-term studies (11–13) and in one long-term study (14), but
not in all studies (15). In pediatric patients there are two small
uncontrolled case studies reported on DHEAS (instead <strong>of</strong>
DHEA) therapy that suggested that this medication is able to
induce pubarche in girls with adrenal insufficiency (7, 16). Controlled
studies in late childhood or adolescence do not exist at all.
We hypothesized that medication with DHEA might be most
useful and effective in females who are severely deficient <strong>of</strong>
DHEA and at an adolescent age when DHEA(S) reaches its physiological
serum peak. Here, we report the data observed in a
randomized controlled double-blind study that included adoles-
ISSN Print 0021-972X ISSN Online 1945-7197
Printed in U.S.A.
Copyright © 2009 by The Endocrine Society
doi: 10.1210/jc.2008-1982 Received September 8, 2008. Accepted December 30, 2008.
First Published Online January 6, 2009
* For a list <strong>of</strong> members <strong>of</strong> the South German Working Group for Pediatric Endocrinology,
see Acknowledgments.
cent girls and young women at ages between 13 and 26 yr with
severe DHEA deficiency due to central adrenal insufficiency.
Subjects and Methods
Subjects
Adolescent girls and young women at ages between 13 and 26 yr with
the diagnosis <strong>of</strong> central adrenal insufficiency were enrolled in the study.
The clinicians who decided to take part in this study were asked to screen
all their patients with a chronic endocrine disorder for eligibility. This
screening was done in four pediatric endocrinology departments at university
hospital clinics by pediatric endocrinologists who are known to
have very good expertise in diagnosis and therapy <strong>of</strong> hypopituitarism.
Eligibility criteria were a pubertal breast stage B3 or more according to
Tanner, DHEAS serum levels less than 400 ng/ml in two independent
measurements (DHEA deficiency), proven central adrenal insufficiency,
presence <strong>of</strong> at least two additional pituitary hormone deficiencies, and no
planned changes in medication during the study. Exclusion criteria were
cranial irradiation with more than 30 Gy, hypothalamic obesity syndrome,
amaurosis, mental retardation, psychosis, systemic diseases like
chronic hepatopathy and heart disorders, diabetes mellitus and elevated
liver enzymes, as well as duration <strong>of</strong> remission from a cerebral tumor
shorter than 12 months.
The group consisted <strong>of</strong> seven patients with connatal and 16 with acquired
hypopituitarism after resection <strong>of</strong> a tumor <strong>of</strong> the sella region. The 16
patients included 12 cases after surgery for a craniopharyngioma and four
cases <strong>of</strong> other tumors <strong>of</strong> the sella region (in detail: two dysgerminomas, one
pituitary adenoma, and one astrocytoma). All patients and parents <strong>of</strong> patients
gave their written informed consent before enrollment.
Study design
This study was a randomized, double-blind, placebo-controlled, multicenter
trial. We decided against the cross-over design for this study to
avoid cross-over effects that were likely to occur. The study has been
registered at www.clinicaltrials.gov. The trial was conducted in compliance
with international guidelines. The protocol was reviewed by the
Ethics Committee <strong>of</strong> the Medical Faculty <strong>of</strong> Tuebingen and by the German
Federal Institute for Drugs and Medical Devices. Adherence to good
clinical practice in this multicenter trial was achieved by implementation
<strong>of</strong> a standardized protocol, an investigator’s brochure, an operational
manual with detailed instructions on how to perform each assessment,
data collection on case record forms, several investigators’ meetings before
starting the trial reviewing all procedures and assessments, an initiation
visit at each site by a monitor assistant, and ongoing monitoring
throughout the study to maintain the quality <strong>of</strong> data collected. Patients
were randomized to receive either placebo or DHEA at a dose <strong>of</strong> 25 mg
once a day in the morning orally for 12 months. Both drugs, DHEA and
placebo, were supplied as a sterile powder in identical capsules that did
not provide any information on its content to the user.
DHEA and placebo
DHEA used was produced by Synopharm GmbH (Barsbüttel, Germany)
as a highly pure (HPLC content 100%) powder. This chemical
was encapsulated and distributed by the Pharmacy <strong>of</strong> the Medical Faculty
<strong>of</strong> the University Ulm, Germany. The placebo used was D-mannitol
Abbreviations: ADIOL, 5-Androstene-3,17-diol; CES-D, Centre for Epidemiological
Studies-Depression Scale; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone
sulfate; GC, gas chromatography; SCL-90-R, Symptom Check-List-90-R.

1184 Binder et al. DHEA Replacement in Young Females J Clin Endocrinol Metab, April 2009, 94(4):1182–1190
99.5%/aerosol 0.5%. For the assessment <strong>of</strong> compliance, unused capsules
were counted at each visit.
Efficacy assessment
In the current trial, we tested the hypothesis that oral administration
<strong>of</strong> 25 mg DHEA is able to promote impaired pubic hair growth (primary
endpoint) and to decrease psychological distress (secondary endpoint 1),
both caused by central adrenal insufficiency. In addition, normalization
<strong>of</strong> adrenal androgen levels in serum 2 h after drug intake and in a 24-h
urine collection were tested (secondary endpoint 2). Efficacy measurements
were performed at baseline, and at 6 and 12 months during
therapy.
The primary outcome measure with respect to efficacy was the proportion
<strong>of</strong> patients with a change to a more mature pubic hair stage from
baseline to month 12, as measured by Tanner stages. For clinical purposes,
Tanner has separated the pubic hair development into five stages
from stage PH1 (infantile) to stage PH5 (adult) (17) (Fig. 1). This staging
is routinely used by pediatricians. All five clinical investigators involved
are experienced pediatric endocrinologists who are very familiar with the
staging <strong>of</strong> pubic hair according to Tanner.
Secondary outcome measures were positive changes (calculated as
the difference <strong>of</strong> each score) in well-being and mood from baseline to
month 12 as measured by the German versions <strong>of</strong> the Symptom Check-
List-90-R (SCL-90-R) and <strong>of</strong> the Centre for Epidemiological Studies-
Depression Scale (CES-D). The SCL-90-R is a widely used, well-validated
psychological status symptom inventory that measures the subjectively
experienced distress by 90 given somatic and mental symptoms. The
individual tested reports the symptoms experienced during the last 7 d.
This inventory allows a multidimensional analysis and contains repeated
measurements. The German version <strong>of</strong> SCL-90-R is well established, and
reference data for schoolchildren and adolescents are available (18, 19).
The other test used, the CES-D, is one <strong>of</strong> the most frequently used and
well-validated self-report screening tool for depressive symptoms.
CES-D consists <strong>of</strong> 20 items assessing the frequency <strong>of</strong> depressive symp-
FIG. 1. Participant flow throughout the study.
toms during the preceding week. The German version <strong>of</strong> the CES-D is
well established, and reference data for schoolchildren and adolescents
are available (20, 21).
Additional secondary outcome measures were the change to normalization
<strong>of</strong> serum androgen levels (DHEAS, androstenedione, testosterone)
2 h after drug intake and <strong>of</strong> the amount <strong>of</strong> urinary excreted androgens
[DHEA, DHEA and M (16-hydroxylated downstream metabolites),
5-androstene-3,17-diol (ADIOL), C19 (total androgen metabolites)].
Serum samples were collected from the nonfasted patient in the morning
before 1000 h during each visit (0, 6, and 12 months), around 2 h after
oral administration <strong>of</strong> DHEA at visits 6 and 12 months. The samples
were collectively assayed at the end <strong>of</strong> the trial to exclude changes <strong>of</strong>
measurements due to interassay variability. Serum DHEAS levels were
measured using an automated chemiluminescence assay system (Immulite;
DPC Biermann GmbH, Bad Nauheim, Germany). Intraassay coefficients
ranged from 4.8–8.8%. Total serum testosterone was measured
by solid-phase 125J RIA in unextracted serum (Coat-A-Count total testosterone;
Diagnostic Products Corp., Los Angeles, CA). Cross-reactivity
to natural steroids was by far less than 1%, to 5-dihydrotestosterone
3%. The intraassay coefficient was 6.0% according to the manufacturer.
Androstenedione serum levels were determined by Active Androstenedione
RIA (Diagnostics Systems Laboratories, Sinsheim, Germany). Cross-reactivity
to natural androgens is less than 0.33%. The intraassay coefficient
was 5.6% according to the manufacturer. The measurements <strong>of</strong> androstanediol
glucuronide levels in serum were performed using androstanediol
glucuronide RIA DSL-6000 (Diagnostics Systems Laboratories).
Intraassay coefficients range from 4.2–8.2%. Reference values for the
hormone assays used were taken from three references (22–24). Serum
SHBG levels were determined using the Immulite 1000 SHBG assay
(Immulite). Intraassay coefficients ranged from 4.1–7.7% according to
the manufacturer.
Urinary pr<strong>of</strong>iling was performed at 0 and 12 months. Patients and
parents received instruction and written guidance to ensure compliance
in the 24 h-urine collection that was performed at home by each participant
before the visits. The urinary steroid pr<strong>of</strong>iles were determined using
quantitative data produced by gas chromatography (GC)-mass spectrometry
analysis according to the method described previously (25). In
brief, free and conjugated urinary steroids were extracted by solid-phase
extraction, and the conjugates were enzymatically hydrolyzed. Thereafter,
steroids were again recovered by solid-phase extraction and derivatized
to form methyloxime-trimethylsilyl ethers. GC was performed
using an Optima-1 fused silica column (Machery-Nagel, Dueren, Germany)
and helium as carrier gas. The GC (Agilent 6890 Series GC; Agilent
7683 Series Injector; Agilent Technologies, Inc., Palo Alto, CA) was
directly interfaced to a mass selective detector (Agilent 5973N MSD)
operated in selected ion monitoring mode.
Daily urinary excretion rates were determined for DHEA, ADIOL,
and the sum <strong>of</strong> DHEA and its 16-hydroxylated downstream metabolites
16-hydroxy-DHEA and 3,16,17-androstenetriol (DHEA
and M), reflecting major adrenarchal secretion products. Overall androgen
metabolite excretion (C19) was determined as the sum <strong>of</strong>
androsterone, etiocholanolone, ADIOL, DHEA, 16-hydroxy-
DHEA, and 3,16,17-androstenetriol.
Safety assessments
The clinical assessment performed at every visit included physical
examination with measurement <strong>of</strong> height, weight, blood pressure, serum
chemistry, and complete blood count. We especially looked for the occurrence
<strong>of</strong> acne and <strong>of</strong> hirsutism, which was quantified by the score
proposed by Ferriman and Gallwey (26). Adverse events were monitored
and recorded throughout the study according to Good Clinical Practice
guidelines.
Sample size
We estimated that the probability for a change <strong>of</strong> the pubic stage
would be 5% in the placebo group and 50% in the DHEA group, according
to the data from two small uncontrolled case studies on DHEAS

J Clin Endocrinol Metab, April 2009, 94(4):1182–1190 jcem.endojournals.org 1185
(7, 16). Based on an -error level <strong>of</strong> 5% and -error level <strong>of</strong> 20% (onesided
testing) to detect a change to a higher pubic stage, 12 patients were
required for each study group. To compensate for a dropout rate <strong>of</strong>
10–20%, we planned to include 30 patients. Actually, recruitment was
poorer than expected because some hospitals initially involved decided
not to take part in the study because <strong>of</strong> the inability to manage the
workload associated with this study. For partial compensation <strong>of</strong> this
event, the age range for inclusion was extended from 13–23 yr to 13–26
yr by the amendment No. 1 <strong>of</strong> the study protocol from April 2003.
Allocation sequence
To achieve balance in respect to the medical history <strong>of</strong> the patients,
randomization was stratified into a “tumor group” defined by a medical
history <strong>of</strong> a cerebral tumor <strong>of</strong> the sella region (in remission), and a “nontumor
group” defined by connatal hypopituitarism and absence <strong>of</strong> a
history <strong>of</strong> a cerebral tumor. The allocation sequence was defined by two
computer-generated randomization lists (one for each group) that were
set up by a statistician <strong>of</strong> the Department <strong>of</strong> Medical Biometry Tuebingen,
and were given to the pharmacy at the University <strong>of</strong> Ulm. These lists
ensured that members <strong>of</strong> both groups had an equal probability to get
DHEA or placebo (1:1). Independent pharmacists (University <strong>of</strong> Ulm,
Pharmacy, Ulm, Germany) dispensed either placebo or DHEA according
to these lists whose content was not available for the researchers. The
researchers were completely blinded for the content <strong>of</strong> the capsules allocated
to each patient from the time <strong>of</strong> allocation until the end <strong>of</strong> the
trial. The code <strong>of</strong> each study drug was revealed to the researchers after
the end <strong>of</strong> the trial, when the total data set was saved and blocked against
any subsequent changes by a statistician.
Statistical analysis
The primary endpoint was the change in pubic hair stage in the treatment
group compared with the placebo group. The efficacy <strong>of</strong> the therapy
was statistically evaluated using the Cochran-Mantel-Haenszel- 2
test (27), including the estimation <strong>of</strong> the relative risk, its 95% confidence
interval, and two-sided P values. Analyses <strong>of</strong> changes in the psychological
scores and changes in serum and urinary hormone levels were performed
using the two-sided Wilcoxon test. All analyses were performed
TABLE 1. Baseline characteristics <strong>of</strong> the study participants
during the 12-month controlled study phase for the changes between
pre-therapy visit and follow-up visit at 12 months. Patients with missing
values at the 12-month visit were included in the analyses with the last
observed value before the planned time point (at the 6 month visit). All
secondary endpoints were analyzed in an explorative sense.
After collection on case record forms, all data were entered twice into
the koordobas database (www.koordobas.de) by two independent study
assistants. Koordobas is a validated s<strong>of</strong>tware for support <strong>of</strong> organization
and data management <strong>of</strong> clinical studies according to common requirements
for conduction <strong>of</strong> studies (International Conference <strong>of</strong> Harmonization/Good
Clinical Practice, Food and Drug Administration). Clean
files were submitted to the Department <strong>of</strong> Medical Biometry at the University
<strong>of</strong> Tuebingen for the final analysis, which was performed using
the statistical analyzing s<strong>of</strong>tware SAS (version 9.1.3; SAS Institute Inc.,
Cary, NC), based on the intention-to-treat population, which included
all randomized patients with the exception <strong>of</strong> those who were withdrawn
from active treatment before the first assessment <strong>of</strong> efficacy (Fig. 1).
Role <strong>of</strong> the funding source
The sponsor was the Medical Faculty <strong>of</strong> the University <strong>of</strong> Tuebingen.
The study protocol was reviewed, and the revised version was finally
selected for support by a board <strong>of</strong> full pr<strong>of</strong>essors <strong>of</strong> the Medical Faculty.
There was no contribution to conduct, data collection, data analysis, or
anything else by the sponsor.
Results
The participant flow throughout the study is shown in Fig. 1: 21 <strong>of</strong>
23 patients (91%) completed the visit at 6 month treatment, and 18
<strong>of</strong> 23 patients (78%) completed the total trial lasting 12 months.
There was one case <strong>of</strong> dropout in the DHEA group because <strong>of</strong>
recurrent anxiety attacks that occurred after a frightening ride in
an amusement park carousel. Similar psychological problems
had occurred before the trial but had not been reported to the doctor
Placebo group (n 12) DHEA group (n 11)
Age (range) 19.7 4.3 yr (13.1–25.4) 18.0 1.5 yr (15.7–21.0)
Height (range) 168.1 5.6 cm (155.8–174.8) 162.2 6.1 cm (149.7–169.1)
Weight (range) 66.6 16.4 kg (42.0–99.3) 68.6 15.6 kg (51.8–93.5)
BMI (range) 23.4 4.8 kg/m 2 (17.3–33.1) 26.0 5.3 kg/m 2 (19.4–35.0)
No. <strong>of</strong> overweight (BMI P90) 2 0
No. with obesity (BMI P97) 2 5
Systolic blood pressure (mm Hg) 113 12 115 12
Diastolic blood pressure (mm Hg) 72 6 72 7
No. with cerebral tumor 8 8
No. with post-radiation 2 4
Hydrocortisone dose (mg/m 2 d) 5.1 2.2 6.0 2.1
No. with GH deficiency 12 11
RhGH dose (g/m 2 d) 0.38 0.20 0.28 0.19
No. with TSH deficiency 12 10
L-thyroxine dose (g/m 2 d) 40.4 8.2 42.7 10.9
LH/FSH deficiency 12 11
ADH deficiency 6 7
PRL deficiency 0 1
Breast stage B (range) 4.5 0.8 (3–5) 4.6 0.7 (3–5)
No. with pubic hair stage PH1 4 5
No. with pubic hair stage PH2 7 4
No. with pubic hair stage PH3 1 1
No. with pubic hair stage PH4 0 1
Drug doses are given per day and m 2 body surface. Data are means SD, if not otherwise stated. ADH, Antidiuretic hormone; BMI, body mass index; PRL, prolactin;
RhGH, recombinant human GH.

1186 Binder et al. DHEA Replacement in Young Females J Clin Endocrinol Metab, April 2009, 94(4):1182–1190
at inclusion. In the placebo group, one withdrawal was caused
by relapse <strong>of</strong> craniopharyngioma, another due to the manifestation
<strong>of</strong> type 1 diabetes mellitus. Two patients with placebo
were lost at follow-up because <strong>of</strong> change in residence
after 6 month therapy. The remaining 18 patients completed
the trial.
Recruitment was started in January 2004 and ended in March
2006. All 23 patients attended the clinic at baseline after recruitment
and randomization. There were 21 patients seen at the 6
month visit, and 18 completed the trial at the 12 month visit. The
last patient had her 12-month visit in April 2007.
Study population
Baseline characteristics <strong>of</strong> the randomized study participants
are shown in Table 1. The groups were well balanced
with respect to age, weight, tumor and radiation history, presence
<strong>of</strong> additional pituitary hormone deficiencies, doses <strong>of</strong>
hormones substituted, and pubic hair stage. All 23 patients
had hypogonadotropic hypogonadism and were substituted
with sex steroids. The only clinical characteristic that was
different between the two groups was height (the placebo
group was taller), which was presumably <strong>of</strong> no clinical relevance
for this study.
The analyses <strong>of</strong> the primary and secondary outcome measures
involved all patients who took part in the 6-month visit and
FIG. 2. Chronological changes <strong>of</strong> the pubic hair stage according to Tanner. Each filled circle represents
one individual patient. Pubic hair stage V (PH5) is normal in females with an age above 13 yr.
excluded the two patients who were withdrawn from active
treatment before the first assessment <strong>of</strong> efficacy. Treatment compliance
was good; the counts <strong>of</strong> unused capsules were in good
agreement to the expected numbers. In addition, no relevant
violation <strong>of</strong> the study protocol was noted.
Primary efficacy outcome
The chronological change <strong>of</strong> the pubic hair stage in each individual
treated is shown in Fig. 2. The DHEA group exhibited
a significant progress in pubic hair growth from Tanner stage
I–III to II–V (mean: 1 1 ⁄2 stages), whereas the placebo group
did not (mean: 0stages) (P0.0046). Progress in pubic hair stage
was observed in nine <strong>of</strong> 10 patients in the DHEA group, but only
in three <strong>of</strong> 11 patients in the placebo group (relative risk: 0.138;
95% confidence interval 0.021–0.914; P 0.0046). In the
DHEA group, five adolescents reached the mature pubic hair
stages IV and V, whereas only one individual did in the placebo
group.
Secondary efficacy outcomes
The psychometric measurement <strong>of</strong> psychological distress using
the SCL-90-R at baseline and after 12 month therapy gave
different results in the DHEA group when compared with the
placebo group. We evaluated the psychological distress at baseline
and after 12 month therapy, and calculated the differences
<strong>of</strong> the scores observed. The values are
shown in Table 2. At the start, the DHEA
group scored higher than the placebo
group in all scales, but differences were not
significant. In all 10 scores, the DHEA
group improved during therapy, whereas
the placebo group worsened. These
changes were significantly different in
eight <strong>of</strong> 10 scores, including the global severity
index, which is the central integrating
score <strong>of</strong> this psychometric test. Baseline
and outcome scores <strong>of</strong> both groups
were within the normal range observed in
the reference population (19).
We used an additional psychometric instrument,
the German version <strong>of</strong> the CES-D,
to observe the presence <strong>of</strong> depressive symptoms
in each patient treated. The effects <strong>of</strong>
the treatment were evaluated by comparison
<strong>of</strong> changes in the CES-D sum score. The
treatment group improved by 2.1 points
(6.8; range 11 to 13) between baseline
and the 12-month visit (P 0.316; n 10).
In the control group, the scores decreased
[mean: 7.9, (12.0; range 38 to 5; P
0.027; n 9]. The comparison <strong>of</strong> the score
changes with the Wilcoxon test resulted in a
P value <strong>of</strong> 0.018.
The changes in serum and urinary androgens
that were additional secondary out-
come measures are shown in Fig. 3. DHEA
substitution resulted in normalization <strong>of</strong>

J Clin Endocrinol Metab, April 2009, 94(4):1182–1190 jcem.endojournals.org 1187
TABLE 2. Scores on the SCL-90-R at baseline and after 12 month treatment with DHEA or placebo
Scale Baseline 12 months P value Reference
Somatization 0.66 0.51
Placebo 0.41 0.37 0.81 0.63 0.40 0.52
DHEA 0.58 0.40 0.49 0.39 0.09 0.39 ns
Obsessive-compulsive traits 0.77 0.54
Placebo 0.43 0.41 0.61 0.63 0.18 0.38
DHEA 0.81 0.56 0.46 0.39 0.35 0.41 0.049
Interpersonal sensitivity 0.80 0.62
Placebo 0.55 0.60 0.64 0.77 0.09 0.33
DHEA 1.06 0.66 0.53 0.48 0.53 0.49 0.030
Depression 0.77 0.59
Placebo 0.48 0.49 0.80 0.83 0.32 0.64
DHEA 0.78 0.54 0.45 0.58 0.33 0.41 0.049
Anxiety 0.67 0.57
Placebo 0.38 0.30 0.66 0.48 0.28 0.35
DHEA 0.50 0.45 0.29 0.24 0.21 0.36 0.016
Hostility 0.80 0.72
Placebo 0.29 0.36 0.41 0.57 0.12 0.38
DHEA 0.62 0.39 0.50 0.48 0.12 0.45 ns
Phobic anxiety 0.25 0.33
Placebo 0.13 0.17 0.32 0.51 0.19 0.36
DHEA 0.40 0.49 0.17 0.32 0.23 0.12 0.011
Paranoid ideation 0.72 0.63
Placebo 0.26 0.33 0.61 0.75 0.35 0.45
DHEA 0.55 0.44 0.38 0.47 0.17 0.39 0.034
Psychotic tendencies 0.46 0.53
Placebo 0.20 0.28 0.46 0.47 0.26 0.31
DHEA 0.31 0.31 0.19 0.31 0.12 0.29 0.036
Global severity index 0.67 0.45
Placebo 0.38 0.35 0.65 0.57 0.27 0.35
DHEA 0.63 0.37 0.38 0.37 0.25 0.26 0.013
Reference data shown are from 139 randomly chosen German schoolgirls with an age between 16 and 17 yr (19). ns, Not significant.
DHEAS and androstanediol glucuronide morning serum levels
2 h after drug intake; placebo had no effect (P 0.006)
(Fig. 3A). Morning serum levels <strong>of</strong> androstenedione increased
significantly in the DHEA group only but did not normalize
(P 0.02). Testosterone serum levels increased only in two
individuals and normalized in only one <strong>of</strong> the DHEA group
(Fig. 3A). SHBG and IGF-I serum levels remained unchanged
(data not shown).
In urine, daily excretion rates for DHEA, ADIOL, and the
sum <strong>of</strong> DHEA and its 16-hydroxylated downstream metabolites
(DHEA and M) normalized in the DHEA group, but not in the
placebo group (Fig. 3B). In addition, the overall androgen metabolite
excretion (C19) was in the high-normal range in the
DHEA group but stayed pathologically low in the placebo group
(P 0.0001 for all comparisons).
Adverse events
The proportion <strong>of</strong> patients experiencing any adverse event
was similar between the DHEA and placebo groups. Single adverse
events reported in the DHEA group were a fracture <strong>of</strong> a
finger, rhinitis, upper airway infection, and transient pain <strong>of</strong> the
ankle joint. One girl in the DHEA group developed a mild moustache,
but no other signs <strong>of</strong> hirsutism. Her Ferriman score was
normal (n 2). Acne was not observed in any patient. In the
placebo group, headache, nonorganic acral paresthesias (n 2),
mood swings, and spraining <strong>of</strong> the ankle were reported.
There was one severe adverse event in the DHEA group and
two in the placebo group: one girl suffered from recurrent anxiety
attacks that were reported after 4 month therapy. Similar
psychological problems had occurred before the trial according
to the parents. Nevertheless, because causality could not be excluded,
medication was stopped. In the placebo group, one girl
had a relapse <strong>of</strong> craniopharyngioma, and another exhibited the
first manifestation <strong>of</strong> type 1 diabetes mellitus. In both cases,
medication was stopped. Overall, the safety pr<strong>of</strong>ile <strong>of</strong> DHEA
was very good.
Discussion
Oral administration <strong>of</strong> 25 mg DHEA daily for 1 yr in a group <strong>of</strong>
female adolescents who suffered from central adrenal insufficiency
caused a clear progress <strong>of</strong> pubic hair stage by I–III stages,
whereas placebo had no effect. Psychological distress as measured
by SCL-90 scores decreased when DHEA was substituted
but increased in the placebo group. In addition, the CES-D score
that measures the presence <strong>of</strong> depressive symptoms improved in
the treatment group but decreased in the placebo group. These
effects were associated with normalization <strong>of</strong> the morning serum
levels <strong>of</strong> DHEA and androstanediol glucuronide 2 h after drug
intake, as well as normalization <strong>of</strong> the 24-h urinary excretion rates
<strong>of</strong> DHEA, DHEA and M, ADIOL, and C19-metabolites in the

1188 Binder et al. DHEA Replacement in Young Females J Clin Endocrinol Metab, April 2009, 94(4):1182–1190
FIG. 3. Chronological changes <strong>of</strong> the androgen levels in serum (A) and the daily androgen amount
excreted in the urine (B) from baseline to 12 month therapy. The boxes indicate the 50% confidence
interval, whiskers the 100% confidence interval, and lines inside the boxes the median. The grayshadowed
areas represent the 90% confidence interval <strong>of</strong> a healthy age-related female reference
population (22–25).
DHEA group, whereas these parameters remained pathologically
low in the placebo group. DHEA therapy was safe and well
tolerated.
Adolescence is the time <strong>of</strong> the only DHEA peak during the
human life span. The normal mean serum DHEAS level at this
age is at 2000 ng/ml. Therefore, DHEA deficiency is most severe
when adolescent individuals are affected by adrenal insufficiency:
our patients had serum DHEAS levels less than 200 ng/ml.
The therapeutic substitution <strong>of</strong> a severe deficiency for a hormone
should cause positive effects if the hormone is <strong>of</strong> any
relevant physiological value for the human being. DHEA is the
major source <strong>of</strong> androgens in females.
Most <strong>of</strong> the girls treated had no pubarche
or an immature stage <strong>of</strong> pubic hair development
because <strong>of</strong> a lack <strong>of</strong> androgens at
baseline. They experienced a clear
progress <strong>of</strong> sexual hair growth during the
trial.
Interestingly, these clinical changes
happened in the absence <strong>of</strong> a measurable
normalization <strong>of</strong> testosterone in the morning
serum, whereas the overall androgen
metabolite excretion in the urine (C19 metabolites)
reached a normal level. In addition,
morning serum levels <strong>of</strong> androstanediol
glucuronide normalized in the DHEA
group, which has been proposed as a marker
<strong>of</strong> peripheral androgen production and action
(28). These observations suggest that
normalization <strong>of</strong> morning serum DHEAS
and androstenedione was sufficient to increase
locally produced testosterone and dihydrotestosterone
to an extent that enabled
sexual hair growth. The same discrepancy
between normal amounts <strong>of</strong> androgen metabolites
and low-active androgens has been
observed in DHEA trials with DHEA-deficient
women or healthy postmenopausal
women in the past (10, 29).
Beside sexual hair growth in puberty,
we could confirm that DHEA and androgens
have beneficial effects on the psychological
well-being. The exact biochemical
path that is used by DHEA to cause such
effects is still obscure, and a separation
from effects caused by metabolites <strong>of</strong>
DHEA is difficult. Such effects are <strong>of</strong> most
importance in young patients having a
chronic disease.
This is the first controlled trial analyzing
replacement <strong>of</strong> DHEA in pediatric patients.
Craen et al. (16) reported on the individual
treatment results in six girls who had hypopituitarism
and low DHEAS serum levels using
15 mg DHEAS /m 2 body surface. Central
adrenal insufficiency was present in some,
but not in all, <strong>of</strong> these patients according to
the abstract published (16). Duration <strong>of</strong> treatment was between
6 and 28 months. In this preliminary report, they observed
an increase <strong>of</strong> serum DHEAS and a progress <strong>of</strong> pubic
hair development to PH2 or more in all patients. Wit et al. (8)
reported their experience with the individual treatment <strong>of</strong> four
girls with central or peripheral adrenal insufficiency (age
range 15–17.3 yr) using DHEAS at a dose <strong>of</strong> 25–50 mg orally.
Treatment duration was between 2 and 4 yr. They reported a
positive effect on pubic hair development and normalization
<strong>of</strong> DHEAS serum levels. Adverse effects <strong>of</strong> DHEAS were not
reported in these two studies.

J Clin Endocrinol Metab, April 2009, 94(4):1182–1190 jcem.endojournals.org 1189
FIG. 3. (Continued).
Instead <strong>of</strong> DHEAS, we administered DHEA, which has been
used in all recent clinical trials in adult patients with adrenal
insufficiency. In contrast to DHEAS, the pharmacology <strong>of</strong>
DHEA has been well studied in adults (9), and it is easily available.
By using the lowest dose reported to being effective in adult
females with adrenal insufficiency (12), DHEAS serum levels
were reached that were almost identical to the healthy reference
population at this adolescent age. In comparison, when adult
females with a mean/median age between 40 and 57 yr <strong>of</strong> age
were treated with 25–50 mg DHEA, DHEAS serum levels
reached were in median around 2000 ng/ml ( 5.4 mol/liter),
which is a concentration above the median for this age (10–15).
In this context it must be noted that oral
DHEA replacement once daily does not fully
restore the diurnal rhythm <strong>of</strong> serum androgen
levels present in healthy women. Arlt et
al. (9) studied the pharmacokinetics in
healthy women (mean age 23.3 yr) with
transient suppression <strong>of</strong> adrenal androgen
secretion because <strong>of</strong> dexamethasone administration.
Using this experimental setting,
they found supraphysiological concentrations
<strong>of</strong> DHEAS during the first 12 h after
oral administration <strong>of</strong> 50 mg DHEA, which
decreased to baseline values during the
second half <strong>of</strong> the day. Similar kinetics was
found for androstenedione and testosterone;
however, these androgen serum concentrations
did not reach baseline at any
time. Therefore, the normalization <strong>of</strong>
morning serum levels in our DHEA group
observed might not have been associated
with normal serum levels during the whole
day.
In this study, adolescents and young
adults with female gender only were treated.
Therefore, results cannot be generalized to
both genders. In male adolescents, androgens
produced in testes may substitute for
the missing adrenal androgens. Whether
DHEA deficiency causes a relevant estrogen
deficiency in adult males is still a matter <strong>of</strong>
debate. In pediatrics, data are missing to
support such a hypothesis. We restricted the
inclusion to those children with adrenal insufficiency
caused by ACTH deficiency.
Therefore, the rare cases <strong>of</strong> primary adrenal
insufficiency like in Addison’s disease, congenital
adrenal hypoplasia, and familial glucocorticoid
deficiency were not investigated.
To date, there are no rational arguments that
this specific group needs a different regiment
to treat for DHEA deficiency. The group <strong>of</strong>
patients studied was small and the time <strong>of</strong>
therapy limited to 12 months. Therefore,
data on any possibly occurring rare or late
adverse effect could not be detected by this
study. However, based on the data available from studies in
adults, there is no relevant concern regarding safety <strong>of</strong> the drug
DHEA at an oral dose <strong>of</strong> 25–50 mg daily.
Here, we have shown for the first time that DHEA replacement
in female adolescents and young women with central adrenal
insufficiency causes a substantial change in sexual hair
growth indicating efficacy <strong>of</strong> this drug. For the patient with the
burden <strong>of</strong> a complex hormone deficiency disorder, normalization
<strong>of</strong> the physical appearance is very important. In addition, we
observed a clear improvement <strong>of</strong> psychological distress and depression
scores that may be beneficial in the situation <strong>of</strong> adolescence,
which is very vulnerable for any psychological crisis. In the

1190 Binder et al. DHEA Replacement in Young Females J Clin Endocrinol Metab, April 2009, 94(4):1182–1190
future we need international long-term replacement studies with
larger numbers <strong>of</strong> affected girls to evaluate changes in quality <strong>of</strong>
life, body composition, and metabolic parameters.
Acknowledgments
We thank all patients and their families for their participation in the
study. We also thank Karin Weber for excellent technical assistance,
Philipp Schwarze for language editing, and Doris Guenon and Tobias
Binder for trial data collection from the case record forms. In addition,
we thank Martin Hautzinger from the Institute <strong>of</strong> Clinical Psychology
(University <strong>of</strong> Tuebingen) for advice and fruitful discussions, and Klaus
Dietz from the Department <strong>of</strong> Medical Biometry (University <strong>of</strong> Tuebingen)
for the statistical design <strong>of</strong> the study. The following members <strong>of</strong> the
South German Working Group for Pediatric Endocrinology took part in
the investigators’ meetings: Peter Heidemann, Eberhard Heinz, Beate
Karges, Ursula Kuhnle, Joachim Partsch, Hanz-Peter Schwarz, Wolfgang
Sorgo, and Martin Wabitsch.
Address all correspondence and requests for reprints to: Pr<strong>of</strong>essor Dr.
Gerhard Binder, Pediatric Endocrinology, University-Children’s Hospital,
Hoppe-Seyler-Str.1, 72076 Tuebingen, Germany. E-mail: gerhard.
binder@med.uni-tuebingen.de.
This study was exclusively funded by the Applied Clinical Research
grant <strong>of</strong> the University <strong>of</strong> Tuebingen. There was no sponsorship by
industry.
Disclosure Summary: The authors have nothing to declare.
References
1. Achermann JC, Siverman BL 2001 <strong>Dehydroepiandrosterone</strong> replacement for
patients with adrenal insufficiency. Lancet 357:1381–1382
2. Løvås K, Loge JH, Husebye ES 2002 Subjective health status in Norwegian
patients with Addison’s disease. Clin Endocrinol (Oxf) 56:581–588
3. Abraham GE 1994 Ovarian and adrenal contribution to peripheral androgens
during the menstrual cycle. J Clin Endocrinol Metab 39:340–346
4. Ibanez L, Dimartino-Nardi J, Potau N, Saenger P 2000 Premature adrenarche—normal
variant or forerunner <strong>of</strong> adult disease? Endocr Rev 21:671–
696
5. Sulcova J, Hill M, Hampl R, Starka L 1997 Age and sex related differences in
serum levels <strong>of</strong> unconjugated dehydroepiandrosterone and its sulphate in normal
subjects. J Endocrinol 154:57–62
6. Labrie F, Luu-The V, Belanger A, Lin S-X, Simard J, Pelletier G, Labrie C 2005
Is dehydroepiandrosterone a hormone? J Endocrinol 187:169–196
7. Young J, Couzinet B, Nahoul K, Brailly S, Chanson P, Baulieu EE, Schaison
G 1997 Panhypopituitarism as a model to study the metabolism <strong>of</strong> dehydroepiandrosterone
(DHEA) in humans. J Clin Endocrinol Metab 82:2578–2585
8. Wit JM, Langenhorst VJ, Jansen M, Oostdijk W, Van Doorn J 2001 <strong>Dehydroepiandrosterone</strong>-sulfate
(DHEAS) treatment for atrichia pubis. Horm Res
56:134–139
9. Arlt W, Justl HG, Callies F, Reincke M, Hübler D, Oettel M, Ernst M, Schulte
HM, Allolio B 1998 Oral dehydroepiandrosterone for adrenal androgen replacement:
pharmacokinetics and peripheral conversion to androgens and estrogens
in young healthy females after dexamethasone suppression. J Clin
Endocrinol Metab 83:1928–1934
10. Arlt W, Callies F, van Vlijmen JC, Koehler I, Reincke M, Bidlingmaier M,
Huebler D, Oettel M, Ernst M, Schulte HM, Allolio B 1999 Dehydroepi-
androsterone replacement in women with adrenal insufficiency. N Engl J Med
341:1013–1020
11. Hunt PJ, Gurnell EM, Huppert FA, Richards C, Prevost AT, Wass JA, Herbert
J, Chatterjee VK 2000 Improvement in mood and fatigue after dehydroepiandrosterone
replacement in Addison’s disease in a randomized, double blind
trial. J Clin Endocrinol Metab 85:4650–4656
12. Johannsson G, Burman P, Wirén L, Engström BE, Nilsson AG, Ottosson M,
Jonsson B, Bengtsson BA, Karlsson FA 2002 Low dose dehydroepiandrosterone
affects behavior in hypopituitary androgen-deficient women: a placebocontrolled
trial. J Clin Endocrinol Metab 87:2046–2052
13. van Thiel SW, Romijn JA, Pereira AM, Biermasz NR, Roelfsema F, van Hemert
A, Ballieux B, Smit JW 2005 <strong>Effects</strong> <strong>of</strong> dehydroepiandrostenedione, superimposed
on growth hormone substitution, on quality <strong>of</strong> life and insulin-like
growth factor I in patients with secondary adrenal insufficiency: a randomized,
placebo-controlled, cross-over trial. J Clin Endocrinol Metab 90:3295–3303
14. Gurnell EM, Hunt PJ, Curran SE, Conway CL, Pullenayegum EM, Huppert
FA, Compston JE, Herbert J, Chatterjee VK 2008 Long-term DHEA replacement
in primary adrenal insufficiency: a randomized, controlled trial. J Clin
Endocrinol Metab 93:400–409
15. Løvås K, Gebre-Medhin G, Trovik TS, Fougner KJ, Uhlving S, Nedrebø BG,
Myking OL, Kämpe O, Husebye ES 2003 Replacement <strong>of</strong> dehydroepiandrosterone
in adrenal failure: no benefit for subjective health status and sexuality
in a 9-month, randomized, parallel group clinical trial. J Clin Endocrinol
Metab 88:1112–1118
16. Craen M, Du Caju MVL, Bourguignon JP, Ernould C, Malvaux P,
Vandeweghe M, Wolter R, Vanderschueren-Lodeweyckx M 1982 <strong>Effects</strong> <strong>of</strong>
oral treatment with DHEAS in girls with hypopituitarism. Pediatr Res 16:888
(Abstract)
17. Marshall W, Tanner J 1969 Variations in pattern <strong>of</strong> pubertal changes in girls.
Arch Dis Child 44:291–303
18. Ge X, Lorenz FO, Conger RD, Elder GH, Simons RL 1994 Trajectories <strong>of</strong>
stressful life events and depressive symptoms during adolescence. Dev Psychol
30:467–483
19. Essau CA, Groen G, Conradt J, Turbanisch U, Petermann F 2001 Validität und
Reliabilität der SCL-90-R: Ergebnisse der Bremer Jugendstudie. ZDDP 22:
139–152
20. Radl<strong>of</strong>f LS 1977 The CES-D scale: a self-report depression scale for research
in the general population. Appl Psychol Measure 1:385–401
21. Steinhausen HC, Metzke CW 2000 [The General Depression Scale in diagnosis
<strong>of</strong> adolescents]. Prax Kinderpsychol Kinderpsychiatr 49:419–434 (German)
22. Emlinger MW, Kühnel W, Ranke MB 2002 Reference ranges for serum concentrations
<strong>of</strong> lutropin (LH), follitropin (FSH), estradiol (E2), prolactin, progesterone,
sex hormone-binding globulin (SHBG), dehydroepiandrosterone
sulfate (DHEAS), cortisol and ferritin in neonates, children and young adults.
Clin Chem Lab Med 40:1151–1160
23. Ankarberg C, Norjavaara E 1999 Diurnal rhythm <strong>of</strong> testosterone secretion
before and throughout puberty in healthy girls: correlation with 17-estradiol
and dehydroepiandrosterone sulfate. J Clin Endocrinol Metab 84:975–984
24. Rittner HL, Lee PD, Blum WF, Doerr HG, Steiss J, Kreuder J, Rascher W, Kiess
W 1997 Developmental patterns <strong>of</strong> serum 3 -androstanediol glucuronide.
J Endocrinol Invest 20:245–250
25. Remer T, Boye KR, Hartmann MF, Wudy SA 2005 Urinary markers <strong>of</strong> adrenarche:
reference values in healthy subjects, aged 3–18 years. J Clin Endocrinol
Metab 90:2015–2221
26. Ferriman D, Gallwey JD 1961 Clinical assessment <strong>of</strong> body hair growth in
women. J Clin Endocrinol Metab 21:1440–1447
27. Kleinbaum DG, Kupper LL, Morgenstern H 1982 Epidemiologic research:
principles and quantitative methods. Belmont, CA: Lifetime Learning Publications.
Sections 17.4, 17.5
28. Rittmaster RS 1993 Androgen conjugates: physiology and clinical significance.
Endocr Rev 14:121–132
29. Labrie F, Bélanger A, Cusan L, Candas B 1997 Physiological changes in dehydroepiandrosterone
are not reflected by serum levels <strong>of</strong> active androgens and
estrogens but <strong>of</strong> their metabolites: intracrinology. J Clin Endocrinol Metab
82:2403–2409