NORMAL PITUITARY VOLUMES IN
CHILDREN AND ADOLESCENTS WITH BIPOLAR DISORDER:
A MAGNETIC RESONANCE IMAGING STUDY
Hua Hsuan Chen, B.S.,1,2Mark Nicoletti, M.S.,3Marsal Sanches, M.D.,2–4John P. Hatch, Ph.D.,3
Roberto B. Sassi, M.D.,5David Axelson, M.D.,6Paolo Brambilla, M.D.,7Matcheri S. Keshavan, M.D.,6
Neal Ryan, M.D.,6Boris Birmaher, M.D.,6and Jair C. Soares, M.D.1–3n
The volume of the pituitary gland in adults with bipolar disorder has previously
been reported to be smaller than that of healthy controls. Such abnormalities
would be consistent with the HPA dysfunction reported in this illness. We
conducted a study of children and adolescents with bipolar disorder to determine
whether size abnormalities in the pituitary gland are already present early in
illness course. Magnetic resonance imaging (MRI) morphometric analysis of the
pituitary gland was carried out in 16 DSM-IV children and adolescents with
bipolar disorder (mean age7sd¼15.573.4 years) and 21 healthy controls
(mean age7sd¼16.973.8 years). Subjects underwent a 1.5 T MRI, with 3-D
Spoiled Gradient Recalled (SPGR) acquisition. There was no statistically
significant difference between pituitary gland volumes of bipolar patients
compared to healthy controls (ANCOVA, age, gender, and ICV as covariates;
F¼1.77, df¼1,32, P¼ .19). There was a statistically significant direct
relationship between age and pituitary gland volume in both groups (r¼ .59,
df¼17, P¼.007 for healthy controls; r¼ .61, df¼12, P¼ .008 for bipolar
patients). No evidence of size abnormalities in the pituitary gland was found in
child and adolescent bipolar patients, contrary to reports involving adult bipolar
patients. This suggests that anatomical abnormalities in this structure may
develop later in illness course as a result of continued HPA dysfunction.
Depression and Anxiety 20:182–186, 2004. & 2005 Wiley-Liss, Inc.
words: magnetic resonanceimaging;bipolar disorder;pituitary;
DEPRESSION AND ANXIETY 20:182–186 (2004)
1Department of Radiology, The University of Texas Health
Science Center at San Antonio, San Antonio, Texas
2South Texas Veterans Health Care System, Audie L. Murphy
Division, San Antonio, Texas
3Division of Mood and Anxiety Disorders, Department of
Psychiatry, The University of Texas Health Science Center at
San Antonio, San Antonio, Texas
4Department of Psychiatry, Federal University of Sa ˜o Paulo,
Sa ˜o Paulo, Brazil
5Department of Psychiatry, University of Sa ˜o Paulo School of
Medicine, Sa ˜o Paulo, Brazil
6Department of Psychiatry, Institute of Psychiatry, University
of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
7Department of Pathology and Experimental and Clinical
Medicine, Section of Psychiatry, University of Udine, Udine,
Contract grant sponsor: National Institute of Mental Health; Grant
numbers: MH 01736, MH 55123, MH 30915, MH 59929.
Psychiatry, Division of Mood and Anxiety Disorders, The
University of Texas Health Science Center at San Antonio, 7703
Floyd Curl Drive, San Antonio, TX 78229.
to: Dr.JairC. Soares,Departmentof
Received for publication 15 April 2004; Revised 3 October 2004;
Accepted 12 October 2004
Published online in Wiley InterScience (www.interscience.wiley.
& & & & 2005 WILEY-LISS, INC.
The pituitary gland is a central component of the
neuroendocrine system. The control of its production
and release of hormones results from the influence of
hypothalamus releasing factors and feedback loops
[Wilson et al., 1998]. Dysfunctions in the hypothala-
mus-pituitary axis (HPA axis) have been implicated in
the pathophysiology of mood disorders, particularly
unipolar depression [Soares, 2003; Soares and Mann,
1997]. Several studies have focused on structures
possibly affected by these abnormalities, such as
hippocampal atrophy as a result of chronic hypercorti-
solemia [Bremner et al., 2000]. However, few neuroi-
maging studies have examined the pituitary gland
volume in patients with mood disorders. Findings of
larger pituitary glands among adult depressed patients
have been described [Krishnan et al., 1991], and the
only study to assess the pituitary gland volume in
bipolar disorder found smaller values in adult patients
compared to healthy controls [Sassi et al., 2001].
If adult bipolar patients have abnormalities in
pituitary size, it is not known whether those abnorm-
alities are already present early, which would suggest a
neurodevelopmental origin, or appear later during
illness course, which would suggest that they could
result from ongoing hypothalamus–pituitary axis dys-
function. To test this hypothesis, the study of young
bipolar patients represents an attractive approach. No
studies have analyzed the size of the pituitary in
children and adolescents with mood disorders.
We conducted this study to test the hypothesis that
children and adolescents with bipolar disorder have
smaller pituitary size compared to healthy individuals,
similar to what has been previously reported by our
group in adults with bipolar disorder [Sassi et al.,
SUBJECTS AND METHODS
The sample comprised 16 children and adolescents
with bipolar disorder (mean age7sd¼15.573.4 years;
8 males, 8 females; 12 bipolar Type I, 3 bipolar Type II,
1 bipolar NOS) and 21 healthy controls (mean
age7sd¼16.973.8 years; 12 males, 9 females). The
bipolar patients were recruited from outpatient psy-
chiatric clinics at the University of Pittsburgh, and
healthy controls were recruited from the surrounding
community through advertisements. Patients met
DSM-IV diagnostic criteria for bipolar disorder as
per the Structured Clinical Interview for DSM IV
(SCID) [Spitzer et al., 1994], for subjects 18–21 years
of age, and the Schedule for Affective Disorders and
Schizophrenia for School Age Children, Present and
Lifetime Version (K-SADS-PL) interview [Kaufman et
al., 1997], for subjects 10–17 years old. There were no
significant differences between patients and controls
regarding age (P¼.2) or gender distribution (P¼.7).
Patients did not have any other Axis I comorbid
psychiatric disorder, with the exception of attention-
deficit/hyperactivity disorder (n¼5), conduct disorder
(n¼1) and oppositional defiant disorder (n¼1). Two
patients were drug free and 14 patients were on
treatment at the point when they participated in the
study (6 were on lithium, 4 on valproate, and 4 on
lithiumþvalproate). The mean duration of the disease
among patients was 3.9 years (sd¼2.4). No subjects
(patients and healthy controls) had a lifetime history of
alcohol or drug abuse, or any past or present
neurological illness, head trauma, or major medical
illness in the 6 months preceding the MRI.
Images were obtained using a 1.5-T MRI GE Signa
image system running version 5.4.3 software (General
Electric Medical Systems, Milwaukee, WI) with 3-D
spoiled gradient recalled acquisition (SPGR) images.
Twenty-five sagittal scout images (T1 fast spin echo)
were obtained to confirm the patient’s position. We
used a midline sagittal scout view to get proton density
(PD) and T2 weighted images (fast spin echo). We then
used 3D SPGR to get 124 coronal images for the full
brain (TR¼25 ms, TE¼5 ms, thickness¼1.5 mm,
flip angle¼401, NEX¼1,matrix
Histograms for each image were obtained as
reported previously [Keshavan, et al., 1994] using a
semi-automated algorithm. The pituitary gland and
intra-cranial volumes (ICV) were traced manually by
well-trained evaluators (H.C., M.N.), with high inter-
rater reliability (ICC4.975 for ICV; ICC4.96 for
pituitary gland), using the semi-automated software
Scion Image Beta-4.0.2 for Windows (Scion Corpora-
tion, Frederick, MD).
We traced the pituitary gland around the well
defined borders of the anterior and posterior pituitary,
excluding the infundibular stalk. The bright posterior
pituitary, one of the most common normal variations in
brain MRI [Patel and Friedman, 1997], was included in
the measurements because it is considered to represent
hyperintense signals from the neurosecretory vesicles
or intracellular lipids in the posterior pituitary cell
[Holder and Elster, 1997; Kucharczyk et al., 1989].
ICV was traced around the border of the brain, which
included CSF, dura matter, sinus, white and gray
matter. It was traced in every other slice following the
procedure described previously [Sassi et al., 2001].
Statistical analyses were conducted using SPSS
version 11.5 (SPSS, Chicago, IL) and SYSTAT version
8 (SPSS Science). Two-tailed significance levels were
set at Po.05. We carried out analysis of covariance
Research Article: Pituitary Volumes in Bipolar Children183
with age, gender, and ICVas covariates to compare the
pituitary volumes between the two groups. Results are
reported as unadjusted means and standard deviations.
Because statistical power is an important issue in the
interpretation of small studies, we calculated partial
Z2as an index of effect size. We used partial correlation
coefficients, controlling for gender and ICV, to assess
the relationship between age and pituitary volumes.
Analysis of variance (ANOVA) and t-tests were carried
out to compare subgroups of patients having different
ANCOVA, adjusting for age, gender, and ICV,
showed no statistically significant differences in the
pituitary gland volumes of bipolar patients (unadjusted
mean7sd .737.17 ml) compared to healthy controls
(.697.16 ml) (F¼1.77, df¼1/32, P¼.19, partial
Z2¼ .052). Similarly, the groups did not differ
significantly regarding ICV, including age and gender
as covariates (F¼1.75, df¼1/33, P ¼ .29, partial
Among patients, we examined the pituitary volumes
as a function of treatment status, comparing untreated
patients and those receiving lithium, valproate, or both.
No significant differences were found (F¼ .881, df¼3/
15, P¼.478). There was also no significant difference
between patients with (n¼11) and without (n¼5)
previous use of antipsychotics (t¼ .542, df¼14,
We compared the volumes of the pituitary gland in
males (.677.17 ml) and females (.777.14 ml), and
found a non-significant trend for larger pituitary
volumes among the females (t¼?1.926, df¼35,
P¼.062). To exclude the possible effect of female sex
hormones, we also compared the pituitary volumes of
males with and without bipolar disorder. No significant
differences were found (t¼?0.16, df¼14, P¼.87).
Additionally, partial correlation coefficients control-
ling for age, gender, and ICV were calculated to
examine the relationship between length of illness and
pituitary volume in bipolar patients. Length of illness
was not significantly related to pituitary volume
(r¼?.22, df¼10, P¼.49).
Last, partial correlations, controlling for ICV and
gender, showed a statistically significant positive
relationship between age and pituitary gland volume
in both groups (r¼0.59, df¼17, P¼.007 for healthy
controls; r¼ .61, df¼12, P¼.008 for bipolar patients).
Figure 1 displays unadjusted mean pituitary volume as
a function of age for bipolar and healthy subjects.
Pituitary gland abnormalities among mood disorder
patients may be interpreted in conjunction with the
several neuroendocrine abnormalities reported in uni-
polar and bipolar patients, such as elevated basal levels
of cortisol [Linkowski et al., 1987; Whalley et al.,
1989], non-suppression at the dexamethasone suppres-
sion test [Rush et al., 1997], and higher release of
cortisol by the adrenal gland when stimulated by
corticotropin [Lopez et al., 1987]. Smaller pituitary
glands among these patients suggest the possibility that
high levels of cortisol over time could induce smaller
pituitaries, due to a putative negative feedback effect.
In the present study, no statistically significant
differences in pituitary gland volume were found
between bipolar adolescents and healthy controls. This
finding is in contrast with our previous study of adults
[Sassi et al., 2001], where significantly smaller pituitary
glands were found in bipolar patients compared to
healthy controls. Besides the obvious differences in the
age of the subjects, few methodological differences
exist between the present study and Sassi et al. .
The MRI parameters and the methodology for the
anatomical measurements were identical in both
studies. Our sample size was relatively small. The
effect size observed in our sample was also very small as
reflected by the effect size statistic. Only about 5% of
the total variability in pituitary volume was associated
with bipolar or control group membership. Thus, it
seems unlikely that we missed a clinically meaningful
effect due to inadequate statistical power.
Furthermore, the proportion of patients who were
receiving mood-stabilizing medication at the time of
the MRI scan was markedly higher in the present
sample than in the study of Sassi et al. . There is
evidence that the therapeutic effect of lithium and
antidepressants may result in part from actions on the
HPA axis [Bschor et al., 2002], but it is not clear if
treatment with these drugs affects pituitary size.
Nonetheless, we did not find significant differences
in the pituitary volume among patients according to
their treatment status. In addition, all but two of our
Fig. 1. Pituitary volume as a function of age for patients with
bipolar disorder and healthy controls.
Chen et al. 184
patients were euthymic, whereas 10 of 23 bipolar
patients in the study above were depressed. No studies
have analyzed pituitary size during different phases of
bipolar disorder, so it is unknown if volumetric changes
are related to affective status.
There was not a significant negative correlation
between pituitary volume and duration of illness. This
finding could be interpreted as evidence against the
suggested model of smaller pituitaries as result of
chronic hypercortisolism. Length of illness was mark-
edly short among our patients, however, and we did not
assess the endocrinological function of our sample.
Therefore, it is possible that HPA axis hyperactivity, if
present, was not active for enough time to affect the
We found a positive correlation between pituitary
gland size and age in our sample. This finding is in
accord with previous studies on the pituitary gland
volume of children and adolescents. Apparently, there
is a normal increase of the pituitary volume during
puberty, followed by a progressive decrease in its size
after adolescence [Elster et al., 1990; Lurie et al., 1990;
Takano et al., 1999]. Disruptions in this normal process
could account for larger or smaller pituitary volumes in
patients with mood disorders. The correlation was
similar among patients and controls (Fig. 1).
Finally, our data showed a trend for larger pituitary
volumes among females. These results are in accord
with previously described gender-related variations in
pituitary size during puberty, probably related to
differences in the process of maturation and hormone
secretion, which are more abrupt and dramatic in girls
[Kato et al., 2002; Takano et al., 1999]. The proportion
of males and females in the bipolar and control group
was similar, so it is unlikely that gender differences
contributed to our negative findings. Further, the
additional comparison between patients and controls,
after excluding female subjects, did not demonstrate
In conclusion, the absence of volumetric differences
in young bipolar patients suggests that volumetric
differences in the pituitary gland may appear over the
course of the illness and are not present in its initial
phases. The fact that we did not detect anatomical
abnormalities does not rule out functional abnormal-
ities, which may precede volumetric changes. There-
fore, a careful characterization of HPA function in
conjunction with anatomical MRI evaluation will be
important for future studies.
Acknowledgments. This study was supported in
part by the National Institute of Mental Health, the
Krus Endowed Chair in Psychiatry (The University of
Texas Health Science Center at San Antonio), and
CAPES Foundation (Brazil).
Bremner JD, Narayan M, Anderson ER, Staib LH, Miller HL,
Charney DS. 2000. Hippocampal volume reduction in major
depression. Am J Psychiatry 157:115–118.
Bschor T, Adli M, Baethge C, Eichmann U, Ising M, Uhr M, Modell
S, Kunzel H, Muller-Oerlinghausen B, Bauer M. 2002. Lithium
augmentation increases the ACTH and cortisol response in the
combined DEX/CRH test in unipolar major depression. Neurop-
Elster AD, Chen MY, Williams DW 3rd, Key LL. 1990. Pituitary
gland: MR imaging of physiologic hypertrophy in adolescence.
Holder CA, Elster AD. 1997. Magnetization transfer imaging of the
pituitary: Further insights into the nature of the posterior ‘‘bright
spot’’. J Comput Assist Tomogr 21:171–174.
Kato K, Saeki N, Yamaura A. 2002. Morphological changes on MR
imaging of the normal pituitary gland related to age and sex: Main
emphasis on pubescent females. J Clin Neurosci 9:53–56.
Kaufman J, Birmaher B, Brent D, Rao U, Flynn C, Moreci P,
Williamson D, Ryan N. 1997. Schedule for Affective Disorders
and Schizophrenia for School-Age Children-Present and Lifetime
Version (K-SADS-PL): Initial reliability and validity data. J Am
Acad Child Adolesc Psychiatry 36:980–988.
Keshavan MS, Beckwith C, Bagwell W, Pettegrew JW, Krishnan KR.
1994. An objective method for edge detection in MRI morpho-
metry. European Psychiatry 9:205–207.
Krishnan KR, Doraiswamy PM, Lurie SN. 1991. Pituitary size in
depression. J Clin Endocrinol Metab 72:256–259.
Kucharczyk J, Kucharczyk W, Berry I, de Groot J, Kelly W, Norman
D, Newton TH. 1989. Histochemical characterization and
functional significance of the hyperintense signal on MR images
of the posterior pituitary. AJR Am J Roentgenol 152:153–157.
Linkowski P, Mendlewicz J, Kerkhofs M, Leclercq R, Golstein J,
Brasseur M, Copinschi G, Van Cauter E. 1987. 24-hour profiles of
adrenocorticotropin, cortisol, and growth hormone in major
depressive illness: Effect of antidepressant treatment. J Clin
Endocrinol Metab 65:141–152.
Lopez JF, Kathol RG, Jaeckle RS, Meller W. 1987. The HPA axis
response to insulin hypoglycemia in depression. Biol Psychiatry
TABLE 1. Subject characteristics
Age, mean7sd (yr)
Age range (yr)
Bipolar NOS, Bipolar not otherwise specified.
Research Article: Pituitary Volumes in Bipolar Children185
Lurie SN, Doraiswamy PM, Husain MM, Boyko OB, Ellinwood EH Download full-text
Jr, Figiel GS, Krishnan KR. 1990. In vivo assessment of pituitary
gland volume with magnetic resonance imaging: The effect of age.
J Clin Endocrinol Metab 71:505–508.
Patel VH, Friedman L. 1997. MRI of the brain: Normal anatomy and
normal variants. Philadelphia: W.B. Saunders.
Rush AJ, Giles DE, Schlesser MA, Orsulak PJ, Weissenburger JE,
Fulton CL, Fairchild CJ, Roffwarg HP. 1997. Dexamethasone
response, thyrotropin-releasing hormone stimulation, rapid eye
movement latency, and subtypes of depression. Biol Psychiatry
Sassi RB, Nicoletti M, Brambilla P, Harenski K, Mallinger AG, Frank
E, Kupfer DJ, Keshavan MS, Soares JC. 2001. Decreased pituitary
volume in patients with bipolar disorder. Biol Psychiatry 50:271–280.
Soares JC, Mann JJ. 1997. The anatomy of mood disordersFreview
of structural neuroimaging studies. Biol Psychiatry 41:86–106.
Soares JC. 2003. Contributions from brain imaging to the elucidation
of pathophysiology of bipolar disorder. Int J Neuropsychophar-
Spitzer RL, Williams JBW, Gibbon M, Williams JBW. 1994.
Structured Clinical Interview for DSM-IV (SCID-IV). Washing-
ton, DC: American Psychiatric Press.
Takano K, Utsunomiya H, Ono H, Ohfu M, Okazaki M. 1999.
Normal development of the pituitary gland: Assessment with
three-dimensional MR volumetry. AJNR Am J Neuroradiol
Whalley LJ, Christie JE, Blackwood DH, Bennie J, Dick H,
Blackburn IM, Fink G. 1989. Disturbed endocrine function in
the psychoses. I: Disordered homeostasis or disease process? Br J
Wilson JD, Foster JW, Kronemberg HM, Larsen PR. 1998. Willians
textbook of endocrinology. Philadelphia: W.B. Saunders.
Chen et al.186