| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Neuroendocrine Unit (K.K.M., B.M.K.B., A.S., K.P.-L., A.K.), Massachusetts General Hospital, and Department of Laboratory Medicine, Children’s Hospital (G.B., N.R.), Harvard Medical School, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: Karen K. Miller, Neuroendocrine Unit, Bulfinch 457B, Massachusetts General Hospital, Boston, Massachusetts 02114. E-mail: KKMiller{at}Partners.org.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The aim of the study was to determine the effects of physiological testosterone replacement on cardiovascular risk markers and insulin resistance in women.
Design: A 12-month, randomized, placebo-controlled study was conducted.
Setting: A General Clinical Research Center was the setting for the study.
Study Participants: A total of 51 women of reproductive age with androgen deficiency due to hypopituitarism participated.
Intervention: Study participants were randomized to physiological testosterone administration, 300 µg daily, or placebo, by patch.
Main Outcome Measures: We measured fasting glucose, fasting insulin, insulin-resistance homeostasis model of assessment (IRHOMA), quantitative insulin sensitivity check index (QUICKI), high-sensitivity C-reactive protein, vascular cell adhesion molecule (VCAM), leptin, lipoprotein (a), apolipoprotein A1, and homocysteine.
Results: At 12 months, fasting insulin and IRHOMA were significantly lower in the testosterone compared with the placebo group, and there was a trend toward a higher QUICKI level at 12 months in the testosterone compared with the placebo group. These differences were no longer significant after controlling for baseline levels. We observed no effect, either positive or negative, of testosterone administration on high-sensitivity C-reactive protein, VCAM leptin, lipoprotein (a), or apolipoprotein A1.
Conclusions: Our data suggest that physiological testosterone replacement in women with hypopituitarism for 12 months does not increase, and may improve, insulin resistance. Chronic low-dose testosterone administration does not increase markers of cardiovascular disease reflecting several different mechanistic pathways. Large, randomized, placebo-controlled, long-term prospective studies are needed to determine whether low-dose testosterone replacement affects cardiovascular risk and event rates in women.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A total of 51 women, aged 19–50 yr with hypopituitarism, defined as hypogonadism and/or adrenal insufficiency of pituitary origin, participated in the study. Baseline clinical characteristics and effects of testosterone administration on bone density and body composition, mood, sexual function, cognition, testosterone levels, and lipids [total, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol, and triglycerides] were previously reported (1). A serum-free testosterone less than the median (3.1 pg/ml) of the reference range for premenopausal women was required for participation. All patients were estrogen replete, either because they received exogenous gonadal steroid hormones or had regular menstrual periods (n = 7), for at least 1 yr before study participation. Patients were excluded from participation if they had received androgens, dehydroepiandrosterone, supraphysiological glucocorticoid therapy, or anabolic agents within the year before study enrollment. All study participants had normal free T4 levels, and if receiving levothyroxine therapy, were on stable doses. In addition, potential subjects were required to have been either GH naive or receiving stable doses of GH for at least 2 yr before study participation. Serum alanine aminotransferase and serum creatinine more than three times the upper limit of normal was also exclusion criteria. Diabetes mellitus and hyperlipidemia were not exclusion criteria. There were 18 study participants, nine in the testosterone and nine in the placebo group, who had diabetes mellitus. Six patients, three in the testosterone group and three in the placebo group, were receiving 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors. The study was approved by the institutional review boards of Partners Health Care, Inc., and written informed consent was obtained from all subjects.
Protocol
Study subjects were recruited through referrals from physicians and advertisements. Subject eligibility was determined at a screening visit. Each eligible subject returned for a baseline visit during which blood was drawn after an overnight fast for determination of serum hormones, safety laboratory values, and cardiovascular risk markers. Serum testosterone, free testosterone, alanine aminotransferase, total cholesterol, HDL cholesterol, and triglycerides were measured in real time. Serum and plasma were frozen at –80 C for measurement of all other variables. Body composition was measured using dual x-ray absorptiometry, using a Hologic QDR-4500 (Hologic Inc., Waltham, MA), with an accuracy error for body fat mass of 1.7% and for fat-free mass of 2.4% (9). Cross-sectional muscle mass was determined using single-slice quantitative computed tomography of the midthigh using 10-mm-thick images (General Electric RP High Speed Helical CT Scanner; General Electric, Milwaukee, WI) and graphical analysis software (General Electric Advantage Windows Work Station Version 2.0, General Electric). All measurements were made in duplicate. Technical factors for the scanning were as follows: 80 kVp, 70 mA, and 2-sec scan time. Abdominal adipose deposition, including cross-sectional total abdominal fat, sc fat, and intra-abdominal fat compartments, were determined in duplicate using single-slice quantitative computed tomography scans at the level of L4 using 10-mm-thick axial images with the same scanner and software as described previously.
After completion of all baseline testing, participants were randomly assigned to receive transdermal testosterone (300 µg, Intrinsa; Procter & Gamble Pharmaceuticals, Cincinnati, OH) or placebo. Randomization was stratified for GH use. Assignments were blinded to the investigators and subjects. A health care professional who was not involved in the study monitored free testosterone levels and determined dose reductions from 300 (two patches applied every 3–4 d) to 150 µg (one patch applied every 3–4 d) in subjects with serum-free testosterone levels above the upper limit of normal for women of reproductive age. Blinding of both participants and investigators was maintained at all times; specifically, a study subject receiving placebo was sham-dose reduced, i.e. instructed to decrease patch number from two to one, concurrently with each subject reducing testosterone.
Subjects returned for follow-up study visits 1, 3, 6, 9, and 12 months after baseline testing. Total testosterone, free testosterone, and SHBG were measured at all study visits, as were lipid and lipoprotein levels. All other cardiovascular risk markers were measured at baseline and 12 months. Body composition was measured at the baseline, and 6 and 12-month visits.
Laboratory methods
Total testosterone was measured by RIA after column chromatography (Esoterix Endocrinology, Calabasas Hills, CA). The sensitivity of this assay is 3 ng/dl (to convert total testosterone to nanomoles per liter, multiply nanogram per deciliter by 0.0347) and the intraassay coefficient of variation (cv) less than 8.1%. The normal range for women of reproductive age is 10–55 ng/dl, as determined by Esoterix Endocrinology. Free testosterone concentration was calculated as the product of percent-free testosterone, measured by equilibrium dialysis (Esoterix Endocrinology) and total testosterone concentration. The sensitivity of the determination of percent-free testosterone by this method is 0.1%, with an intraassay cv of 6.9%. The normal range of free testosterone for women of reproductive age is 1.1–6.3 pg/ml (to convert free testosterone to picomoles per liter, multiply pictogram per milliliter by 3.467), as determined by Esoterix Endocrinology. SHBG was measured by an in-house immunoradiometric assay (Esoterix Endocrinology), with a lower limit of detection of 10 nm, intraassay cv of 2.4–3.9% and normal female range of 40–120 nmol/liter. Apolipoprotein (apo) AI assay was performed by an immunoturbidimetric technique on the Hitachi 917 analyzer (Roche Diagnostics, Indianapolis, IN), using reagents and calibrators from Wako Chemicals USA (Richmond, VA). The interassay cvs of apoAI at concentrations of 63.7, 126.7, and 177.4 mg/dl are 1.2, 2.8, and 3.3%, respectively. The concentration of homocysteine was determined using an enzymatic assay on an Hitachi 917 analyzer (Roche Diagnostics), using reagents and calibrators from Catch, Inc. (Seattle, WA), with interassay cvs at 6.01, 13.31, and 34.76 µmol/liter of 5.3, 4.0, and 2.1%, respectively. The concentration of lipoprotein (a) [Lp(a)] was determined using a turbidimetric assay on an Hitachi 917 analyzer (Roche Diagnostics), with reagents and calibrators from Denka Seiken (Niigata, Japan). This method is the only commercial assay that is not affected by the Kringle type 2 repeats (10). The interassay cvs at Lp(a) concentrations of 17.6 and 58.1 mg/dl are 3.6 and 1.5%, respectively. Soluble vascular cell adhesion molecule (VCAM) was measured by an ELISA assay (R & D Systems, Minneapolis, MN). The assay has a sensitivity of 2.0 ng/ml, and interassay cv at concentrations of 9.8, 24.9 and 49.6 ng/ml of 10.2, 8.5, and 8.9%, respectively. Insulin and glucose were measured by previously reported methods. The insulin-resistance homeostasis model of assessment (IRHOMA) was calculated as: (I0) * (G0)/22.5, and quantitative insulin sensitivity check index (QUICKI) was calculated as: 1/[log (I0) + log (G0)], where (I0) is fasting insulin (µIU/ml), and (G0) is fasting glucose (mmol/liter).
IRHOMA has been validated as an accurate measurement of insulin resistance and QUICKI of insulin sensitivity (11, 12, 13).
Statistical analysis
JMP statistical discoveries software (version 4.0.2; SAS Institute, Inc., Cary, NC) was used for statistical analysis. The study power was 80% at a two-sided 
level to detect a 75% SD unit difference in means between the study groups at 12 months assuming no difference between baseline means. All variables were tested for normality using the Shapiro-Wilk test. All variables not normally distributed were log transformed. Clinical characteristics and cardiovascular risk markers were compared using ANOVA. Analysis of covariance and multivariate analyses were used to control for baseline values. Total cholesterol, LDL, HDL, and triglycerides were measured at multiple time points (baseline, and 1, 3, 6, 9, and 12 months after randomization). Therefore, for these variables, a mixed model was used. Variables were transformed [log (y + 1)] toward normality. The change from baseline was analyzed using a random intercepts model, with a fixed treatment effect and a random intercept with baseline value as a single covariate. This analysis averaged the difference between the on-study value and the baseline, and used this as a measure of response to therapy. Undetectable hormone levels were assigned values just below the lower limit of detection. Statistical significance was defined as a two-tailed P value < 0.05. Baseline clinical characteristics are reported as mean ± SD. All other results are recorded as mean ± SEM.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Clinical characteristics of the study subjects are shown in Table 1
. Table 1 also compares baseline characteristics of those patients randomized to receive testosterone with those randomized to receive placebo. There were no significant differences between the groups in the parameters tested. Mean free testosterone was below the lower limit of the normal range for women of reproductive age.
|
As previously reported, 55% of participants had free testosterone levels below the lower limit of the reference range, and mean free testosterone levels increased into the normal range during testosterone administration in women receiving testosterone (from a mean 0.6 ± 0.9 to 2.7 ± 2.1 pg/ml at 12 months; normal range 1.1–6.3 pg/ml), and did not increase in women receiving placebo (from a mean 0.5 ± 0.6 to 0.3 ± 0.3 pg/ml at 12 months). Total testosterone levels increased from a mean of 8.6 ± 8.2 to 36.3 ± 22.5 ng/dl at 12 months in the testosterone group and from 7.1 ± 6.1 to 6.2 ± 5.3 ng/dl at 12 months in the placebo group (normal range 10–55 ng/dl). Neither estradiol nor IGF-I increased in women receiving testosterone compared with placebo (1).
Cardiovascular risk markers
Cardiovascular risk markers at baseline and 12 months after testosterone/placebo administration are shown in Table 2. At baseline, there were no significant differences between the testosterone and placebo groups. There was a trend toward a lower baseline fasting insulin (P = 0.10) and IRHOMA (P = 0.09), and a higher Lp(a) (P = 0.06) in the group that received testosterone. At 12 months, fasting insulin and IRHOMA were significantly lower in the testosterone compared with the placebo group, and there was a trend toward a higher QUICKI level at 12 months in the testosterone compared with the placebo group. These differences were no longer significant after controlling for baseline levels. There was a trend toward a higher mean homocysteine in the testosterone compared with the placebo group at 12 months; there was no longer a trend after controlling for baseline levels. There was no significant difference at 12 months in any other cardiovascular risk markers between the testosterone and placebo groups. This included high-sensitivity C-reactive protein (hsCRP), which remained without change after controlling for estrogen/progestin use. There were no significant associations found between changes in cardiovascular risk markers and changes in testosterone, free testosterone, change in fat-free mass, change in fat mass, change in muscle mass, or change in abdominal fat mass over 12 months. Significant inverse associations were observed between androgens (both total testosterone and free testosterone) and IRHOMA and fasting insulin, and significant positive associations were observed between androgens and QUICKI at 12 months (Table 3 and Fig. 1).
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Cross-sectional studies in women with PCOS have raised the question of whether hyperandrogenemia may contribute to the elevation in cardiovascular risk markers, including higher prevalence of insulin resistance, observed in this population (7). However, studies of women with PCOS are confounded by a high rate of obesity, and despite many cross-sectional studies demonstrating an increased rate of diabetes mellitus, hypertension, and other cardiovascular risk factors in women with PCOS, the only two studies investigating actual rates of cardiovascular events in PCOS have not shown increases (15, 16). Another model of the effects of androgen administration on cardiovascular risk is that of hypogonadal men, in whom data largely demonstrate beneficial changes in body composition and cardiovascular risk markers (8). However, testosterone doses in male studies are 10–20 times the replacement doses appropriate for women, and gender-specific effects of androgen replacement are plausible. Data in animals are conflicting, with some suggestion of differential gender effects, i.e. deleterious effects in female, but not in male, animals (17, 18, 19, 20, 21). However, such data are far from definitive, with contrasting effects on different cardiovascular markers within the same studies (19, 20). The design of our study (randomized, placebo-controlled) and the population studied (women with severe androgen deficiency due to hypopituitarism) provide a model in which the specific effect of testosterone administration can be determined with few confounders. In our study, measures of insulin resistance and sensitivity did not worsen with testosterone administration, and there was a suggestion of an improvement. Our results are similar to those observed in HIV-positive women in a recent study, in which insulin resistance decreased with testosterone administration, using the same testosterone preparation and dose administered in this study (22). However, in that study it was not clear whether the change observed was due to a true decrease in the group that received testosterone or to an increase in the placebo group. It should be noted that as in our study, insulin resistance did not worsen in HIV-positive women, which is also consistent with the changes in body composition we observed, specifically increases in fat-free and muscle mass. Although not definitive, our data are reassuring with regard to the effects of testosterone replacement on insulin resistance in women with androgen deficiency.
While dose and gender may be important determinants of effects on inflammation, our findings in women are consistent with published results in men demonstrating no increase in hsCRP levels with testosterone administration (23). An investigation of the effects of testosterone administration on hsCRP in women is important because hsCRP has been established as a more important predictor of cardiovascular events than LDL cholesterol (14). Our findings are also consistent with a published study in HIV-positive women (22).
Our study demonstrated no change in VCAM, Lp(a), leptin, or apoA1. There have been no previous studies investigating the effects of testosterone replacement on adhesion molecules, Lp(a), apoA1, or leptin in women with androgen deficiency. All of these markers have been shown to be independent risk factors for cardiovascular disease. In vitro studies have demonstrated increases in expression of the adhesion molecule, VCAM-1, on human umbilical vein endothelial cells of male origin, which are blocked by androgen receptor antagonism (25, 26), and decreases in VCAM-1 expression on human umbilical vein endothelial cells of female origin, in response to androgen administration (27, 28). Although Lp(a), a lipoprotein that is an independent risk factor for coronary events (29, 30), is primarily genetically determined (31), levels have consistently decreased in men after testosterone administration in open-label studies (32), including when administered in concert with an aromatase inhibitor (33), suggesting an effect of testosterone independent from conversion to estradiol. ApoA1 levels also decline with testosterone administration (34) in men. The lack of change in these markers in our study is reassuring in that it likely excludes a large deleterious effect; it also makes a major beneficial effect unlikely. Larger studies will be necessary to detect small effects, either positive or negative.
Our data suggest a trend toward higher homocysteine levels at 12 months in women receiving testosterone compared with placebo, though it should be noted that this largely reflected a decrease in the placebo group, and the trend was lost after controlling for baseline levels. Male replacement doses of testosterone have resulted in increases in homocysteine levels in men with Klinefelter’s syndrome (35) and idiopathic hypogonadotropic hypogonadism (36), as well as female to male transsexuals (37), though a study of testosterone administration to healthy male weightlifters did not demonstrate an increase in homocysteine levels (38). Further studies are needed to determine whether lower dose testosterone appropriate for use in women will elevate homocysteine levels, and, if so, whether such elevations pose a threat to cardiovascular health.
Many studies in women have shown no deleterious effects on HDL, LDL, or triglycerides with low-dose transdermal testosterone replacement (1, 2, 3, 4, 5, 39, 40, 41). This is in contrast to oral androgens (42, 43) and pre-androgens, e.g. dehydroepiandrosterone (44), administration, which results in HDL reductions. This is also in contrast to the effects of male replacement-dose im testosterone esters in men, which have been shown in some studies to decrease HDL (24).
We investigated the effects of testosterone replacement on markers reflecting a number of different mechanisms underlying increased cardiovascular risk. Our data suggest that low-dose testosterone replacement does not increase, and may improve, insulin resistance, despite 8-fold increases in mean serum-free testosterone levels (1). Moreover, we observed no effect on hsCRP, leptin, Lp(a), apoA1, or VCAM. Large, randomized, placebo-controlled, long-term prospective studies are needed to determine whether low-dose testosterone replacement affects cardiovascular event rates in women.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Summary: K.K.M. and A.K. receive study medications from and have previously consulted for Procter & Gamble Pharmaceuticals. No other authors have anything to declare.
First Published Online April 10, 2007
Abbreviations: apo, Apolipoprotein; cv, coefficient of variation; HDL, high-density lipoprotein; hsCRP, high-sensitivity C-reactive protein; IRHOMA insulin-resistance homeostasis model of assessment; LDL, low-density lipoprotein; Lp(a), lipoprotein (a); PCOS, polycystic ovarian syndrome; QUICKI, quantitative insulin sensitivity check index; VCAM, vascular cell adhesion molecule.
Received January 25, 2007.
Accepted April 4, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. Baylis Sexual Dimorphism of the Aging Kidney: Role of Nitric Oxide Deficiency Physiology, June 1, 2008; 23(3): 142 - 150. [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 |
