| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Department of Medicine, Neuroendocrine Unit, Clinical Research Center, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: Karen K. Miller, M.D., Neuroendocrine Unit, Bulfinch 457B, Massachusetts General Hospital, Boston, Massachusetts 02114. E-mail: . kkmiller{at}partners.org
Abstract
Hypopituitarism is associated with osteopenia and a reduction in lean body mass. We have recently demonstrated markedly reduced serum androgen levels in women with hypopituitarism. We hypothesized that serum androgen levels and lean body mass are important determinants of bone mineral density (BMD) in women with hypopituitarism. In addition, because IGF-I may stimulate androgen secretion in women, we investigated whether GH administration results in an increase in serum androgen levels. Sixteen women with a history of pituitary disease of adult-onset and serum GH levels less than 5 ng/ml on stimulation testing underwent BMD and body composition testing by dual-energy x-ray absorptiometry. Univariate regression analysis revealed strong correlations between androgen levels and BMD [lateral spine BMD and dehydroepiandrosterone sulfate (DHEAS) (r = 0.68, P = 0.03), total hip BMD and free T (r = 0.60, P = 0.01), Ward’s triangle BMD and DHEAS (r = 0.68, P = 0.004), Ward’s triangle BMD and free T (r = 0.54, P = 0.03), femoral neck BMD and free T (r = 0.52, P = 0.04), and femoral neck BMD and DHEAS (r = 0.51, P = 0.04)]. When adjusted for age using Z scores, correlations at the femoral neck no longer reach significance. Correlations between androgens and BMD at other sites, including anterior-posterior spine and total body, were not significant, and neither total T nor androstenedione correlated with BMD at any site. Lean body mass strongly correlated with BMD [total hip (r = 0.80, P = 0.0002), total body (r = 0.78, P = 0.0003), trochanter (r = 0.74, P = 0.001), Ward’s triangle (r = 0.56, P = 0.02), femoral neck (r = 0.53, P = 0.04), and anterior-posterior spine (r = 0.52, P = 0.04)]. In stepwise regression models, DHEAS determined 47% of the variation in Ward’s triangle BMD (R2 = 0.47, P = 0.004) and 46% of lateral spine BMD (R2 = 0.46, P = 0.03). Lean body mass determined 64% of the variation in total hip BMD (R2 = 0.64, P = 0.0002), 62% of total body (R2 = 0.62, P = 0.0003), and 55% of trochanter BMD (R2 = 0.55, P = 0.001). Subjects were then randomized to receive GH at a dose of 12.5 µg/kg per day or placebo for 12 months in a double-blind protocol. Serum androgen levels were obtained at baseline, 1, 3, 6, 9, and 12 months after initiation of GH. Androgen levels did not increase in the women receiving GH for 12 months, compared with those receiving placebo. Stimulation of androgen secretion is therefore unlikely to be a mechanism underlying the improvement in BMD, body composition, or quality of life observed with GH administration. In conclusion, androgen levels and lean body mass may be important determinants of BMD in women with hypopituitarism. It remains to be determined whether androgen replacement therapy itself or an increase in lean body mass achieved as a result of androgen administration will result in an improvement in BMD in this population.
WE HAVE RECENTLY demonstrated marked hypoandrogenemia in women with hypopituitarism of all ages (1). However, the clinical consequences of hypoandrogenemia in this population are unknown. For example, hypopituitarism is associated with a reduction in bone density and an increase in fracture rate (2, 3). Although hypogonadism has been established as one mechanism of bone loss in men with hypopituitarism (4), it is unknown whether hypoandrogenemia exerts similar effects on bone in women with hypopituitarism.
Recent data suggest that androgens may be important for the maintenance of skeletal health in women. Positive univariate correlations have been demonstrated between serum androgen levels in healthy pre- and postmenopausal women (5, 6, 7). Moreover, the addition of methyltestosterone to an estrogen replacement therapy regimen in healthy postmenopausal women results in an increase in bone formation markers, compared with estrogen administration alone (8). Two randomized, placebo-controlled studies have demonstrated that androgen administration is associated with an increase in bone mineral density (BMD) in excess of that achieved with estrogen therapy alone in healthy postmenopausal women (9, 10). In addition to stimulating bone formation directly, a secondary effect of androgens on the skeletal system may occur through maintenance of lean body mass. Androgen administration has been demonstrated to result in increased lean body mass in men with hypopituitarism (4). It has been demonstrated that lean body mass is an important determinant of BMD in healthy girls and women as well as in women with amenorrhea as a result of exercise, stress, or anorexia nervosa (11, 12, 13, 14, 15). However, whether androgens are important for the maintenance of lean body mass or whether lean body mass is important for maintenance of skeletal health in women with hypopituitarism is unknown.
GH administration results in an increase in BMD and lean body mass in women with hypopituitarism (16, 17). However, it has not been established whether an increase in androgen levels contributes to the increase in bone density and other beneficial effects resulting from GH administration. IGF-I administration in vitro has been shown to increase androgen secretion by adrenal and ovarian cells (18, 19, 20, 21, 22). In vivo administration of IGF-I to females with Laron’s syndrome (primary IGF-I deficiency) results in an increase in serum androgen levels and produces signs of hyperandrogenism (23).
We therefore hypothesized that androgens and lean body mass are important determinants of BMD in women with hypopituitarism and investigated whether GH administration increases serum androgen levels.
Experimental Subjects and Methods
Patients
Sixteen women with histories of sellar disorders associated with hypopituitarism and adult-onset GH deficiency were studied. The majority of subjects with hypopituitarism had pituitary adenomas (n = 13), whereas three subjects had craniopharyngiomas. The characteristics of the pituitary disorders found in the patients, including type of tumor and number of subjects who had undergone surgery and/or radiation therapy, are shown in Table 1
. GH deficiency was confirmed by peak serum GH levels of less than 5 ng/ml on two sequential stimulation tests (clonidine and L-dopa). Patients were excluded from participation in the study if they had previously received GH or had an abnormal free T4 index, a change in glucocorticoid, thyroid hormone or gonadal steroid replacement therapy within 2 months of study entry, a history of acromegaly, current pregnancy, or fasting blood glucose more than 140 mg/dl. No subjects had histories of anorexia nervosa or anabolic steroid use. Women receiving estrogen were excluded if therapy had not been initiated at least 1 yr before study entry. Ten subjects were hypogonadal, 11 hypothyroid, and eight hypoadrenal. Seven subjects were receiving estrogen preparations, which included oral contraceptives (n = 2) and hormone replacement therapy in the form of conjugated equine estrogens (n = 4) or transdermal E2 (n = 1) plus medroxyprogesterone. Five women were experiencing spontaneous regular periods. Four subjects were amenorrheic and not receiving estrogen. Cortisol replacement included prednisone 3.5 mg once daily (n = 1), prednisone 5 mg once daily (n = 4), prednisone 5 mg each morning plus 2.5 mg each evening (n = 1), and hydrocortisone 30 mg once daily (n = 1). One subject with mild hypoadrenalism (stimulation to 13.6 µg/dl with Cortrosyn stimulation) took glucocorticoid therapy only when experiencing physical stress, which was rare.
|
Protocol. The study was approved by the Subcommittee on Human Studies of Massachusetts General Hospital, and all subjects gave written informed consent. Subjects were admitted to the General Clinical Research Center at Massachusetts General Hospital after an overnight fast for the baseline evaluation. A urine pregnancy test was performed on all subjects. Serum was obtained for hormonal assessment, and a glucose tolerance test was conducted. Height, weight, and body composition measurements were obtained. BMD was measured at the following sites in all 16 subjects: anterior-posterior (AP) lumbar spine, total hip, trochanter, Ward’s triangle, femoral neck, and total body. In addition, lateral spine bone density dual-energy x-ray absorptiometry (DXA) was performed in a subset of subjects (n = 10). Quality-of-life questionnaires were administered and cognitive function testing performed. After completion of the baseline visit, subjects were randomly assigned to receive GH (Genentech, Inc., South San Francisco, CA), 12.5 µg/kg per day sc or placebo for 12 months in a double-blind fashion. The GH dose was decreased to 6.25 µg/kg per day in four subjects who experienced side effects. Blood was obtained for hormonal evaluation at 0800 h after an overnight fast at the 3-, 6-, 9-, and 12-month visits. Repeat DXA was performed at the 6- and 12-month visits.
Hormonal assessment. Fasting 0800 h total serum T levels were measured by RIA after extraction and column chromatography (Endocrine Sciences, Inc., Calabasas Hills, CA) with an intraassay coefficient of variation of 2.9–18.8 and a sensitivity of 3 ng/dl. Free serum T levels were measured by equilibrium dialysis (Endocrine Sciences, Inc.) with an intraassay coefficient of variation of 6.6–9.4% and a sensitivity of 0.1%. Androstenedione and dehydroepiandrosterone sulfate (DHEAS) concentrations were measured by RIA (Endocrine Sciences, Inc.) with intraassay coefficients of variation of 3.0–6.1% and 3.6–6.9%, respectively. The sensitivities of the androstenedione and DHEAS assays were 10 ng/dl and 10 µg/dl, respectively. Samples from each individual were measured in duplicate and run in the same assay.
Bone density and body composition analysis. BMD, total body fat, and total lean body mass were determined by DXA using a Hologic-2000 densitometer (Hologic, Inc., Waltham, MA). The SD for measurement of bone density at the lumbar spine by the DXA technique is 0.01 g/cm2 (24). DXA has a precision error of 3% for fat and 1.4% for lean body mass (25).
Statistical analysis. Univariate regression models were constructed to determine correlations between BMD at each skeletal site with serum androgen levels and with lean body mass. Pearson correlation coefficients were calculated and are reported. Standard least squares analyses were performed to determine whether correlations remained significant when controlling for age. Stepwise regression models were constructed for each skeletal site with the following covariates in the model: DHEAS, free T, and lean body mass. The covariates were chosen in advance as important clinical variables that may affect BMD and were also based on the results of the univariate correlation analyses reported in this manuscript. An influence analysis was performed in which correlation coefficients were calculated 16 times, omitting data for one subject at a time, to determine whether correlation coefficients were dependent on any one data point. A repeated-measures ANCOVA was used, controlling for baseline, to compare the change in serum androgen levels in the subjects receiving GH with the change in subjects receiving placebo. Means reported are ± SD, except where otherwise indicated. The funding source had no role in the collection, analysis, or interpretation of the data.
Results
Baseline clinical characteristics are reported in Table 2. Subjects were 46 ± 9 yr (range 30–64 yr), with a mean body mass index (BMI) of 28.1 ± 6.8 kg/m2 (range 21.5–45.5 kg/m2). Mean serum T, DHEAS, and androstenedione levels were below the normal ranges for healthy women of reproductive age, and mean serum-free T was in the low normal range for healthy young women (Table 2).
|
Correlations between free T and BMD at all skeletal sites tested are shown in Table 3, as are correlations with DHEAS. The strongest correlations between BMD and androgen levels are demonstrated in Figs. 1 and 2. There were no significant univariate correlations between androstenedione and BMD at any site.
|
|
|
). The strongest correlations between BMD and lean body mass are demonstrated in Fig. 3. When controlling for lean body mass with a standard least squares analysis, DHEAS still correlated significantly with bone density at Ward’s triangle (P = 0.02).
|
|
Age did not correlate with BMD at any site. All significant correlations reported above remained significant after controlling for age, except for that between DHEAS and femoral neck bone density, for which the P value increased from 0.04 to 0.06. Likewise, when univariate analyses were conducted using Z scores, all significant correlations remained so, except for those between femoral neck and both free T and DHEAS (Table 3
). The results of the influence analysis performed on all statistically significant univariate correlations revealed that the correlation coefficients did not vary greatly with omission of any one data point.
Stepwise regression analysis demonstrated that DHEAS accounted for 47% of the variation in Ward’s triangle (P = 0.004) and 46% of lateral spine (P = 0.03). Lean body mass accounted for 64% of the variation in total hip (P = 0.0002), 62% of total body (P = 0.0003), and 55% of trochanter (P = 0.001). The results of the influence analyses performed on these stepwise regression models revealed that the correlation coefficients did not vary greatly with omission of any one data point.
Serum androgen levels
There was a significant correlation between lean body mass and free T (r = 0.52, P = 0.04) (Fig. 4). There was no correlation between lean body mass and total T, DHEAS or androstenedione.
|
Administration of GH did not result in an increase in serum T, free T (Fig. 5), DHEAS, or androstenedione levels at 3, 6, 9, or 12 months, compared with administration of placebo.
|
Our results are consistent with the hypothesis that serum androgen levels may be determinants of BMD in women with hypopituitarism. These findings may be particularly important in this population of women because such women have marked hypoandrogenemia, as we have recently demonstrated (1). The data presented here suggest that androgen deficiency may be one mechanism contributing to the osteopenia in this population.
Hypoandrogenemia is well established to be associated with reduced bone mass in men (4, 26, 27, 28, 29). However, the effect of androgen deficiency on bone density in women is not well established, particularly in women with hypopituitarism. A few studies have demonstrated that androgen levels and BMD correlate in healthy pre- and postmenopausal women (5, 6, 7). Androgen receptors have been identified on bone removed from women during osteotomies (30). In addition, increases in markers of bone formation have been demonstrated in postmenopausal women receiving both methyltestosterone and estrogens, compared with those receiving estrogens alone (8). Two small randomized, placebo-controlled trials, one in healthy postmenopausal women and one in surgically menopausal women, have demonstrated increases in BMD with administration of androgens plus estrogens, compared with estrogens alone over a period of 2 yr (9, 10). Therefore, androgen deficiency is probably one mechanism contributing to the osteopenia observed in some populations of women, but further data are needed to establish its importance and its specific effects.
An unexpected result of our study was that correlations with androgens appeared to be stronger in the hip than the spine, which contains the greater proportion of trabecular bone. No statistical test was performed to confirm this finding, and it may simply be an artifact caused by the small number of subjects studied. However, if confirmed, this result would not be consistent with current concepts regarding the differential effects of gonadal steroids on trabecular vs. cortical bone. Trabecular bone is more metabolically active than cortical bone and has been demonstrated to be more susceptible to the effects of gonadal steroids. Whether our finding reflects differential and unique effects of androgens to stimulate cortical bone formation in women or is an artifact deserves further study.
Another unexpected finding of this study was that DHEAS, a weak prehormone without an identified receptor of its own, appears to correlate more strongly with BMD at some sites than does T, an androgen 20 times more potent. It is unclear whether this finding reflects independent effects of DHEAS on the skeleton. Alternatively, DHEAS may be a marker of adrenal insufficiency, in which case lower DHEAS levels may reflect more profound hypopituitarism.
Also of importance, androgen levels correlated strongly with lean body mass, which was itself an important determinant of BMD in stepwise regression models. This suggests that androgens may act indirectly on the skeleton in women with hypopituitarism, mediated by effects on lean body mass. It is also possible that the correlations between androgen levels and BMD reflect the effect of BMI and/or lean body mass on bone density, rather than independent effects of androgens. Strong correlations between BMI and BMD have been demonstrated in a number of populations (31, 32). Previous studies suggest that lean body mass may be an even more important determinant of BMD than BMI in amenorrheic and eumenorrheic women (11, 12, 13, 14) as well as healthy girls (15). However, after controlling for lean body mass, correlations between DHEAS and bone density remained significant at Ward’s triangle, suggesting that androgens may exert some independent effects on the skeleton in this population. Further study is needed to confirm these findings.
In vitro models of adrenocortical and ovarian thecal cells in culture have demonstrated increased androgen secretion in response to GH and IGF-I administration (18, 19, 20, 21, 22). In vivo administration of IGF-I to girls with primary IGF-I deficiency, also known as Laron’s syndrome, results in increases in serum androgen levels and signs of hyperandrogenism (23). However, in our study, serum androgen levels were unaffected by GH administration. Therefore, the improvements in lean body mass, bone density, and quality of life observed in GH-deficient women are likely not mediated by increases in serum androgen levels. However, we cannot rule out a small effect of GH administration on androgen secretion that might have been evident if a larger sample of women had been studied.
Our data demonstrate strong correlations between androgen levels and BMD in women with hypopituitarism. Further study is needed to establish whether hypoandrogenemia is an important contributing factor to the development of osteopenia and to determine whether low-dose androgen replacement therapy will be effective in reversing bone loss in women with hypopituitarism.
Acknowledgments
We thank the staff of Massachusetts General Hospital General Clinical Research Center for dedicated patient care and the patients who participated in the study.
Footnotes
This work was supported in part by NIH Grant M01-RR-01066 and a research grant from Genentech, Inc.
Abbreviations: AP, Anterior-posterior; BMD, bone mineral density; BMI, body mass index; DHEAS, dehydroepiandrosterone sulfate; DXA, dual-energy x-ray absorptiometry.
Received August 16, 2001.
Accepted February 25, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  E. M. Gurnell, P. J. Hunt, S. E. Curran, C. L. Conway, E. M. Pullenayegum, F. A. Huppert, J. E. Compston, J. Herbert, and V. K. K. Chatterjee Long-Term DHEA Replacement in Primary Adrenal Insufficiency: A Randomized, Controlled Trial J. Clin. Endocrinol. Metab., February 1, 2008; 93(2): 400 - 409. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. K. Miller, B. M. K. Biller, C. Beauregard, J. G. Lipman, J. Jones, D. Schoenfeld, J. C. Sherman, B. Swearingen, J. Loeffler, and A. Klibanski Effects of Testosterone Replacement in Androgen-Deficient Women with Hypopituitarism: A Randomized, Double-Blind, Placebo-Controlled Study J. Clin. Endocrinol. Metab., May 1, 2006; 91(5): 1683 - 1690. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F Labrie, V Luu-The, A Belanger, S-X Lin, J Simard, G Pelletier, and C Labrie Is dehydroepiandrosterone a hormone? J. Endocrinol., November 1, 2005; 187(2): 169 - 196. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) or placebo (
) on mean free T levels over 12 months in women with hypopituitarism.