Androgen production and therapy in women

Androgen production and therapy in women

Author
Laurence Udoff, MD
Section Editor
Robert L Barbieri, MD
William F Crowley, Jr, MD
Deputy Editor
Kathryn A Martin, MD



Last literature review version 16.2: May 2008 | This topic last updated: May 28, 2008 (More)


INTRODUCTION — All women produce some androgens, which may contribute to maintaining normal ovarian function, bone metabolism, cognition and sexual behavior. This topic will review androgen production in pre- and postmenopausal women and the effects of androgen therapy in postmenopausal women. Female sexual dysfunction, including hypoactive sexual desire disorder, is discussed elsewhere. (See "Etiology and diagnosis of sexual dysfunction in women" and see "Treatment of sexual dysfunction in women").

ANDROGEN PRODUCTION

Premenopausal women — The major androgens in the serum of normal cycling women are dehydroepiandrosterone sulfate (DHEA-S), dehydroepiandrosterone (DHEA), androstenedione (A), testosterone (T), and dihydrotestosterone (DHT) in descending order of serum concentrations [1] .

Though abundant in the circulation, DHEA-S, DHEA, and A may be considered pro-hormones requiring conversion to T or DHT to express their androgenic effects. Androgens are mainly produced in the adrenal gland, the ovary and from the peripheral conversion of pro-hormones.

• DHEA-S is produced solely by the adrenal gland at a rate of 3.5 to 20 mg per day [2] . Circulating levels are in the range of 1 to 4 mcg/mL.
• DHEA is also produced in the adrenal gland (50 percent), the ovary (20 percent) and from peripheral conversion of DHEA-S (30 percent) with total production rates of 6 to 8 mg per day [3] . Circulating levels are in the range of 1 to 10 ng/mL.
• A production is split equally between the adrenal gland and the ovary. Daily rates of production are 1.4 to 6.2 mg/day and circulating levels are in the range of 0.5 to 2 ng/mL [4,5] .
• T is synthesized in the adrenal gland (25 percent), the ovary (25 percent) and from the peripheral conversion of A (50 percent). Daily production rates are in the range of 0.1 to 0.4 mg/day and circulating levels are between 0.2 and 0.7 ng/mL with the lowest levels found during the early follicular phase followed by a 20 percent increase at midcycle [3] . Nearly all (99 percent) of circulating T is protein-bound (mainly to sex-hormone binding globulin [SHBG]) [6] . Therefore, any impact on SHBG concentration (eg, oral estrogen-mediated increase in SHBG) affects the concentration of the free/active androgen.
• Lastly, DHT is mainly a peripheral product of T conversion and has very low circulating levels [7] .

All the major androgens are metabolized and excreted into the urine almost exclusively as 17-ketosteroids.

Natural history — In a report of normal women ages 18-75 years, serum androgen concentrations (total and free testosterone, DHEA-S, and androstenedione) gradually declined in women of reproductive age [8,9] , with no further decrease after clinical menopause [10] .

Putative roles for androgens — Androgens are essential precursors for estrogen synthesis. They also play an important role in follicular development. Maintenance of a precise balance of estrogens and androgens within ovarian follicles is a requirement for successful follicular maturation [11-13] . In addition, many tissues have androgen receptors, including the central nervous system [14] and bone [15] , which has led to speculation that androgens affect their function.

It has also been proposed that androgens play a role in sexual behavior. A woman's libido is dependent upon many factors, including psychological factors [16] . As an example, in a study of 341 peri- and postmenopausal women, common menopausal symptoms, including depression, sleep disturbances, and night sweats, were associated with diminished libido [17] .

A balance of estrogen and androgen also may be necessary for normal sexual desire and responsiveness, but the results of studies on the role of androgens in sexuality in normal premenopausal women are inconclusive [9,16] . As an example, a cross-sectional study of a population-based cohort of women ages 48 to 58 years suggested that most aspects of female sexual function were not affected by age, menopausal functioning, or serum sex hormone concentrations [9] . Additionally, a study comparing women with premature ovarian failure with normal premenopausal controls did not find an important role for circulating androgens in sexual functioning [18] .

The special case of adrenal insufficiency may be an exception, as data suggests that young women with this condition may show improvement in sexuality with the addition of DHEA to their replacement regimen. (See "Dehydroepiandrosterone and its sulfate", section on Adrenal insufficiency).

Androgens may also be important for the maintenance of normal affect, cognitive functioning and of skeletal homeostasis. (See "Cognitive function" below and see "Bone metabolism" below). Regarding the latter, a significant body of evidence exists implicating a role for androgens in the maintenance of bone health. Androgens may impact bone homeostasis directly (eg, all bone cells including osteoblasts, osteoclasts, and osteocytes have androgen receptors), or indirectly by conversion to estrogen, or by their effect on local and systemic factors that control the bone cells' microenvironment [15] .

Low serum androgen concentrations may be associated with lower bone density and fracture risk:

• Serum free androgen concentrations and bone mineral density have been positively correlated in several studies [19-21] .
• In another study, postmenopausal women with a history of vertebral crush fractures had lower serum free androgen concentrations (and similar serum estrogen values) when compared to women with no fractures [19] .
• Women with hypopituitarism (who are androgen deficiency) have low bone density [22] .


Postmenopausal women — The production rate and serum concentrations of androstenedione fall by 50 percent after the menopause (show figure 1) [23,24] . This change appears to result from decreased ovarian production of the hormone with the adrenal glands becoming the major site of androstenedione production. This hypothesis is based upon the following findings in postmenopausal women:

• Administration of corticotropin (ACTH), but not human chorionic gonadotropin (hCG), raises serum androstenedione concentrations.
• Serum androstenedione concentrations fall little after oophorectomy, but markedly after the administration of dexamethasone, which suppresses the release of ACTH.
• There is a diurnal variation in serum androstenedione concentrations that parallels the diurnal variation in serum cortisol concentrations.
• The ratio of androstenedione in ovarian venous blood to peripheral venous blood is decreased [24] .


The rate of testosterone production also falls in postmenopausal women (show figure 2) [25,26] , mostly because of a decline in the peripheral production of testosterone from androstenedione [27] . Ovarian testosterone production remains relatively constant, thereby increasing the relative ovarian contribution to overall testosterone production [27] . These observations are substantiated by the larger ovarian-to-peripheral serum gradient of testosterone in postmenopausal than in premenopausal women [24] , and by the 40 to 50 percent decrease in serum testosterone concentration seen after oophorectomy in postmenopausal women [28] , a change that persists over time [29] .

In addition to this decline in ovarian androgen secretion, there is an age-related decline in the adrenal androgens dehydroepiandrosterone (DHEA) and its sulfate ester (DHEA-S). As an example, in women ages 40 to 50, serum DHEA concentrations are approximately 50 percent of the peak concentrations seen in younger women [30] . (See "Dehydroepiandrosterone and its sulfate").

Serum androstenedione and testosterone concentrations fall little with advancing age after the menopause, despite a progressive fall in serum DHEA concentrations [31] . This difference suggests that little androstenedione and testosterone are derived from DHEA in older women, and that ovarian androstenedione and testosterone production increases or their clearance decreases with age [32] .

The decline in ovarian androgen production in postmenopausal women is much less than the decline in estrogen production; as a result, the ovaries become primarily androgen-producing glands. The relatively high rate of androgen production is due to the increase in gonadotropin secretion, which stimulates steroidogenesis in ovarian hilar cells or luteinized stromal cells [32] . Ovarian stromal tissue has receptors for both follicle-stimulating hormone and luteinizing hormone [30,33] , and chorionic gonadotropin (hCG) stimulates androstenedione, estradiol and progesterone secretion by isolated ovarian cortical stromal and hilar cells [34,35] . In addition, postmenopausal women given hCG have a small increase in serum testosterone concentrations [36] and hyperplasia of their ovarian hilar cells [37] ; in comparison, their serum estrogen concentrations do not increase [38] .

The vast majority of evidence suggests that the postmenopausal ovary is a major androgen-producing gland [27-38] , with the exception of one study [39] . In 10 postmenopausal women with adrenal insufficiency, women with natural and surgical menopause had undetectable serum androgen concentrations. In addition, ovarian stimulation with hCG did not increase circulating levels of androgens in the women with intact ovaries. Negligible levels of T and A were found in ovarian homogenates, and ovarian immunocytochemistry did not detect the presence of enzymes for androgen synthesis. These data have not been confirmed by other investigators.

ANDROGEN DEFICIENCY — Women with low levels of circulating androgens are said to have androgen deficiency or androgen insufficiency syndrome. However, there are no clear biochemical criteria for this syndrome; measurement of serum androgen concentrations, in particular, free testosterone, is problematic because of a lack of validated assays in the female range (much lower than the male range); there are no age-based normative data; and serum androgen concentrations do not appear to be an independent predictor of sexual function in women [40-44] .

An Endocrine Society Clinical Practice Guideline recommended against making a diagnosis of androgen deficiency because of the lack of both a well-defined clinical syndrome and age-based normative data for serum testosterone and free testosterone concentrations That said, there are a number of conditions that may represent androgen deficiency syndromes:

• Bilateral oophorectomy
• Primary adrenal insufficiency
• Hypopituitarism, particularly women with both ACTH and gonadotropin deficiency [45]


Medications including oral contraceptives and glucocorticoids may cause a relative androgen deficiency due to ovarian and adrenal androgen suppression, respectively. Oral estrogens, even at low doses (menopausal replacement) reduce serum free testosterone concentrations by increasing serum SHBG levels.

Women with anorexia nervosa have lower serum concentrations of total and free testosterone, but not DHEAS concentrations when compared to normal-weight women with hypothalamic amenorrhea or healthy controls [46] . In this report, women with anorexia nervosa who were taking oral contraceptives had the lowest concentrations of free testosterone and DHEAS.

EFFECTS OF EXOGENOUS ANDROGENS — Androgen replacement therapy has been advocated by some for postmenopausal women with decreased sexual desire associated with personal distress and with no other identifiable cause [47,48] . However, given the lack of a well-defined clinical syndrome, age-based normative data for serum testosterone concentrations, and long-term safety data for testosterone preparations, we agree with the Endocrine Society guidelines and currently do not suggest the routine use of androgen therapy in women [42] .

Exceptions to this may include women with hypopituitarism (ACTH and gonadotropin deficiency), bilateral oophorectomy and premature ovarian failure (POF). However, our ability to treat these women is limited by the lack of an approved testosterone preparation. Women with primary and secondary adrenal insufficiency are candidates for DHEA therapy. (See "Dehydroepiandrosterone and its sulfate").

Sexual function — Testosterone therapy in postmenopausal women may have a beneficial effect on sexual function in select women, but data are variable.

Replacement vs. supraphysiologic therapy — Studies of androgen therapy in women with androgen deficiency (eg, hypopituitarism, bilateral oophorectomy) should be considered separately from studies in women with sexual desire disorders who are not androgen deficient, as the former would be considered to be replacement therapy, and the latter, supraphysiologic therapy. However, in almost all trials reporting a beneficial effect of testosterone, including those considered to be "replacement" trials, serum testosterone concentrations are higher than the upper limit of normal for premenopausal women.

• Studies in heterogeneous populations - Many testosterone trials have been performed in heterogeneous populations of women (natural or surgical menopause, with normal or low libido) receiving variable types, doses, and routes of administration of estrogen and testosterone. One of these trials reported no effect of testosterone on sexual arousal [49] , while others reported an improvement in sexual function in women with normal [50,51] or low [52] libido at baseline.

A trial in naturally menopausal women diagnosed with hypoactive sexual desire disorder who were taking estrogen reported improved sexual function with the transdermal testosterone patch (dose 300 mcg/day) [53] .

• Women post-oophorectomy — The main evidence that testosterone has an effect on sexual function comes from trials that have examined a transdermal testosterone preparation combined with exogenous estrogen in women who have undergone bilateral oophorectomy and subsequently developed hypoactive sexual desire disorder (HSDD). Although there is a modest improvement in sexual function with testosterone in these trials, serum testosterone concentrations are typically in the high normal or supranormal range for younger premenopausal women.

In one study, 300 mcg/day of transdermal testosterone improved sexual function and psychological well-being, but mean serum free testosterone concentration increased to approximately twice the mean of premenopausal women [54] . A dose of testosterone (150 mcg/day) that increased the mean serum testosterone to a value similar to that of the mean in premenopausal women did not increase sexual function or psychological well-being.

In a second, larger, multicenter trial, 532 women with hypoactive sexual desire who had undergone hysterectomy with bilateral oophorectomy received either testosterone (300 mcg/day) or placebo patch twice per week (in addition to estrogen) for 24 weeks [55] . Sexual desire and frequency of sexual activity increased more in the testosterone group compared to placebo, but only by one additional episode per 2.5 week interval in the testosterone group versus one additional episode per 5.5 week interval with placebo. Serum testosterone levels and androgenic side effects increased with transdermal therapy; however, the side effects were considered mild. Although promising, the results of this trial do not address the safety of long-term testosterone administration.

Additional, phase-III trials reported similar results on sexual desire and sexual activity [56,57] . In one study, no additional benefits were observed with a testosterone patch delivering 400 mcg/day when compared to 300 mcg/day [56] .

• Hypopituitarism — Women with hypopituitarism, in particular those with both ACTH and gonadotropin deficiency, may also benefit from testosterone therapy. In a trial of women with androgen deficiency due to hypopituitarism, treatment with 150 to 300 mcg of testosterone transdermally daily for one year improved overall sexual function, as judged by a questionnaire, by a small but statistically significant amount [58] .
• Use of testosterone without estrogen — Preliminary data from a trial in postmenopausal women (both natural and surgical) with hypoactive sexual desire disorder receiving transdermal testosterone 300 mcg (without estrogen) suggest that sexual function may also be improved in this group of patients [59] .
• Meta-analysis — In a systematic review of 23 clinical trials (with 1957 participants) of testosterone plus hormone therapy versus hormone therapy alone in peri- or postmenopausal women, a significant decrease in serum HDL concentrations was observed, there was insufficient evidence of a treatment effect in perimenopausal women, and there appeared to be an improvement in sexual function scores in postmenopausal women [60] . However, only three trials were included in the sexual function analysis. In addition, this review was not limited to women who would be considered to be truly androgen-deficient (ie post-oophorectomy). (See "Androgen deficiency" above).


Other effects

Vasomotor symptoms — Testosterone therapy may be effective for postmenopausal women who remain symptomatic (eg, hot flashes) despite estrogen or estrogen-progestin treatment [61-63] . Because androgen production declines after the menopause, it is reasonable to assume that some postmenopausal symptoms could at least in part be due to androgen deficiency. However, we do not consider persistent vasomotor flushes to be an indication for routine androgen replacement.

Cognitive function — Some data from uncontrolled studies in which postmenopausal women were treated with estrogen alone or with androgen have suggested that androgen may improve affect and cognitive functioning. This issue was also addressed in a randomized double-blind, placebo-controlled trial in which postmenopausal women were treated for two months with either estrogen, estrogen plus testosterone, or placebo [64] . Both hormone treatments were associated with better scores on a self-rating scale of anxiety and depression than placebo, and there was a trend towards better scores in the estrogen plus testosterone group as compared with the estrogen group.

In another study of the effect of hormone treatment on energy, well-being and appetite, estrogen plus testosterone was superior to estrogen alone [65] . Androgen therapy has also resulted in increased well-being, improved energy levels, and less dysphoric mood in oophorectomized women [66] .

A common criticism of these studies is that the testosterone given is metabolized into estrogen and therefore that the results are mainly due to an estrogen effect. However, in the last report, there was no difference in the incidence of hot flashes between the treatment and placebo groups, suggesting a mode of action for testosterone distinct from aromatization to estrogen [66] .

Bone metabolism — A direct correlation between bone density and serum androgens has been noted in postmenopausal women [20,21] . The effects of androgen therapy on bone in postmenopausal women have been examined in studies of androgen alone and androgen in combination with estrogen; numerous observations are compatible with a beneficial effect of androgen in this setting [49,67-74] :

• In a study of biochemical markers of bone resorption and formation, women receiving either estrogen or estrogen plus androgen had evidence of increased bone formation, whereas bone resorption decreased only in the women receiving estrogen alone [67] . In another report, androgen monotherapy in postmenopausal women with osteoporosis reduced markers of bone turnover (serum alkaline phosphate concentrations and urinary calcium excretion) to the same extent as estrogen [75] .
• A number of reports have shown that nandrolone increased bone mineral density at the spine and radius when compared with no treatment [68-70] .
• In studies of the effect of androgen plus estrogen, the addition of androgen to a regimen of estrogen with or without progestin had a more beneficial effect on bone density [72-74] .
• In a study of women with clear androgen deficiency due to hypopituitarism, who were taking estrogen and had normal baseline bone density, physiologic testosterone replacement (150 to 300 mcg/day transdermally) increased serum testosterone into the normal range, and increased mean hip and radius, but not spine, bone mineral density [58] .

Taken together, these results suggest that androgen alone or in combination with estrogen may protect against osteoporosis. The putative mechanisms involve a decrease in bone resorption by either direct androgenic action or conversion of androgen to estrogen, or an increase in bone formation. However, there is no strong evidence that the addition of androgen to estrogen in postmenopausal women is more beneficial than estrogen alone.

Adrenal androgen replacement — DHEA replacement therapy appears to be effective in women with adrenal androgen deficiency, including those with primary adrenal insufficiency, hypopituitarism (ACTH deficiency), and chronic glucocorticoid use (for example, women with systemic lupus erythematosus). (See "Dehydroepiandrosterone and its sulfate", section on Adrenal insufficiency and see "Overview of the therapy and prognosis of systemic lupus erythematosus in adults", section on Dehydroepiandrosterone (DHEA)).

DHEA supplementation has also been proposed as adjunctive hormone replacement therapy for aging men and women. While there is a well-known decline in serum DHEA and DHEA-S concentrations with age, the role of adrenal androgen replacement in peri- and postmenopausal women is unclear. However, DHEA supplementation in otherwise healthy peri- or postmenopausal women does not appear to have clinical benefits. (See "Dehydroepiandrosterone and its sulfate").

Risks and side effects — One concern regarding androgen replacement therapy in postmenopausal women is the possible adverse effect on cardiovascular disease risk, because androgens have been thought to be atherogenic. This supposition is based mainly on the higher rates of cardiovascular disease in men as compared with women and the higher risk in women with androgen excess (eg, polycystic ovary syndrome) [76,77] . (See "Postmenopausal hormone therapy and cardiovascular risk").

One proposed mechanism by which androgens may adversely affect the risk of cardiovascular disease is through a decline in serum high-density lipoprotein (HDL) cholesterol concentrations. As compared with normal women, women with hyperandrogenism have lower serum HDL cholesterol concentrations [78] . In postmenopausal women, however, the results are less clear. One report found no strong correlation between serum androgen and HDL cholesterol concentrations [79] , but another study found that serum testosterone concentrations were inversely correlated with serum HDL cholesterol concentrations [80] . Confounding variables that could explain the disparity in the results include the effects of diet, body weight, exercise, and heredity, as well as methodological differences such as problems associated with the standardization of methods to measure serum lipids.

The more androgenic progestins (eg, norethindrone, levonorgestrel), when given as the progestin component of estrogen-progestin replacement therapy in postmenopausal women, tend to blunt the estrogen-related rise or even produce a treatment-related decline in serum HDL cholesterol concentrations [81] . However, the addition of testosterone to estrogen replacement therapy has produced conflicting results: it caused a decline in serum HDL cholesterol concentrations in studies using oral estrogen [52,74,82,83] , but not in others using non-oral routes of administration [54,62,84,85] .

One study suggests that testosterone administration might decrease cardiovascular risk in postmenopausal women on hormone replacement therapy (HRT) [86] . In 33 postmenopausal women on HRT compared with 15 controls, both endothelial-dependent and -independent brachial artery vasodilatation was improved by the addition of a testosterone implant (50 mg) for six weeks. Additionally, it has been shown in a group of sixty postmenopausal women that endogenous testosterone levels are positively correlated with brachial artery vasodilation [87] .

Other potential side effects of adding androgen to estrogen in postmenopausal women are acne, hirsutism, deepening of the voice, and clitoromegaly. Among women given testosterone implants, 15 to 20 percent had slight increase in downy facial hair after several years, but acne, voice changes and clitoromegaly were very rare [88] . Women given oral methyltestosterone doses (2.5 mg daily) may become mildly hirsute [74] . Androgen replacement therapy does not affect body weight or blood pressure [63,89] .

A possible association between testosterone administration and breast cancer risk has been reported. However, data are limited. (See "Postmenopausal hormone therapy and the risk of breast cancer", section on Effect of testosterone).

In a review of available testosterone clinical trials, hirsutism and acne appeared to be the major adverse reactions (both were dose- and duration-related and generally reversible). Virilization was rare, and oral, but not parental or transdermal, testosterone was associated with a decrease in serum HDL concentrations, which could have a negative impact on cardiovascular risk. However, there were no adverse effects on blood pressure, vascular reactivity, blood viscosity, hemoglogin concentration, coagulation factors or insulin sensitivity. All available trials are limited by their short duration (≤ two years) and the co-administration of estrogen or estrogen-progestin therapy [90] .

INFORMATION FOR PATIENTS — Educational materials on this topic are available for patients. (See "Patient information: Sexual problems in women" and see "Patient information: Postmenopausal hormone therapy" and see "Patient information: Alternatives to postmenopausal hormone therapy"). We encourage you to print or e-mail these topic reviews, or to refer patients to our public web site, www.uptodate.com/patients, which includes these and other topics.

SUMMARY — The use of androgens as an addition to estrogen or estrogen-progestin therapy for postmenopausal women is controversial.

• The clearest indication for androgen replacement therapy is in patients with symptomatic androgen insufficiency in association with hypopituitarism, adrenal insufficiency, premature ovarian failure, or in women who have undergone bilateral oophorectomy.
• Androgens may significantly improve sexual functioning in select postmenopausal women. (See "Sexual function" above).
• The data that androgens significantly improve cognitive function and affect are not compelling. (See "Cognitive function" above).
• Androgen may have a beneficial effect on bone, which may simply be due to the additional estrogen formed from the administered androgen; studies with nonaromatizable androgens could resolve this question. (See "Bone metabolism" above).
• Serum HDL cholesterol concentrations decline slightly in postmenopausal women receiving oral testosterone therapy, but it is not known if the change substantially affects overall cardiovascular risk. (See "Risks and side effects" above).
• Cosmetic side effects such as hirsutism and acne are usually mild and are well tolerated and irreversible virilizing changes are rare.

In conclusion, for highly selected postmenopausal women, especially those who have undergone bilateral oophorectomy, estrogen replacement alone may not be adequate therapy. Based upon the finding that endogenous androgen production declines after spontaneous menopause as well as oophorectomy, it may be logical to offer some women androgen therapy in conjunction with estrogen therapy. The practitioner and patient should realize, however that many questions remain to be answered regarding the efficacy and safety of this therapy. Most androgen preparations that have been used for this indication are not approved the United States Food and Drug Administration. (See "Treatment of sexual dysfunction in women").