Case Presentation

A 55-year-old healthy married woman complains of reduced sexual desire and arousal for the past 10 years. She denies any dyspareunia. She underwent a hysterectomy with bilateral salpingo-oophorectomy at the age of 45 for symptomatic uterine fibroids. After the surgery, she started experiencing hot flashes, vaginal dryness, and decreased sexual desire. She was prescribed with a transdermal estradiol patch which relieved her hot flashes and vaginal dryness but did not improve her sexual symptoms. Her total and free testosterone levels are low at 12 ng/dl and 1.5 pg/ml, respectively. She had a normal mammogram within the last year. She asks whether testosterone replacement is an option to treat her sexual symptoms. How should this patient be assessed and treated?

Androgen Physiology in Postmenopausal Women

Similar to men, women also experience age-related decline in gonadal steroids [1]. At the time of natural menopause, there is sudden and permanent cessation of ovarian follicle formation and decline in estrogen production. Although serum androgen levels decline with age in women, much of this decline is between the ages of 20–40 years. Furthermore, there is no cessation of androgen production during natural menopause.

The two major sources of androgen production in women are the ovaries and the adrenal glands [2]. In women, testosterone is produced directly by the ovaries and by peripheral conversion of androstenedione and dehydroepiandrosterone (DHEA), which are synthesized by the ovaries and the adrenal glands, respectively. In young premenopausal women, the ovary is responsible for approximately 25 % of the testosterone production, while 75 % is derived from the adrenal glands. However, in postmenopausal women, the ovary becomes a major source of androgens and contributes to approximately 50 % of the total testosterone production (Fig. 10.1) [3]. Although the climacteric ovary becomes atrophic and loses capacity to synthesize estradiol, it still continues to secrete substantial amount of androgens under the stimulation of gonadotropins [2]. In fact, the steepest decline in testosterone levels occurs in the early reproductive years between the ages of 20–40 with a plateau through the menopausal transition, followed by a gradual decline with age (Fig. 10.2) [1, 4]. Interestingly, after the age of 80 years, a minor increase in serum total testosterone levels has been reported in one cross-sectional study [4]; whether this increase in older women is a true physiological phenomenon is unclear. Thus, the decline in androgens in women appears to be more of a function of aging rather than natural menopause [1, 2]. In contrast to natural menopause, surgical menopause results in a significant decline in androgen levels. Indeed epidemiologic studies show a significant decline in serum testosterone levels by more than 40 % in women undergoing bilateral oophorectomy (Fig. 10.3) [5].

Fig. 10.1
figure 1

The contribution of the ovaries and the adrenals to serum androgen levels in pre- and postmenopausal women [3]

Full size image
Fig. 10.2
figure 2

Total and free testosterone levels with age [4]

Full size image
Fig. 10.3
figure 3

Androgens after hysterectomy and oophorectomy [5]

Full size image

Female Androgen Deficiency Syndrome

The role of androgen therapy in postmenopausal women has been an area of growing scientific and public interest. In June 2001, an international expert panel met at the Princeton Consensus Conference (Princeton, NJ) and established three diagnostic criteria for the Female Androgen Deficiency Syndrome (FADS) [6]. In order to have a diagnosis of FADS, women had to have the following three criteria: (1) clinical symptoms such as diminished well-being, unexplained fatigue, sexual dysfunction, vasomotor instability, and/or decreased vaginal lubrication, (2) be adequately estrogenized (i.e., normal menstruating woman or a postmenopausal woman who is on estrogen replacement), and (3) free serum testosterone levels at or below the lowest quartile of normal range for women of reproductive age (20–40 year). Hence, it has been hypothesized that testosterone replacement might reverse symptoms of FADS in these women [7].

Specific conditions in women that may be associated with a higher risk for androgen deficiency include black race, low body mass index (BMI <18.5 kg/m2), bilateral oophorectomy, diseases of the hypothalamus, pituitary or ovaries, primary adrenal insufficiency, oral estrogen and corticosteroid use, various malignancies and their treatments, and HIV infection (Table 10.1) [2, 4]. Hence, these subsets of women have a greater likelihood of androgen deficiency and, if meeting criteria for FADS, could potentially benefit from testosterone therapy.

Table 10.1 Conditions associated with androgen deficiency in women
Full size table

Limited data from randomized controlled trials using transdermal testosterone use in physiologic doses in surgically menopausal women who had hypoactive sexual desire and low serum testosterone levels have shown modest improvements in several aspects of sexual function such as sexual desire, satisfaction, and frequency [8–12]. However, although the prevalent dogma is that androgens regulate libido in women, data from epidemiologic studies show that circulating androgen levels in women are only weakly associated with sexual function [13, 14]. Thus, the definition of FADS suffers from (1) lack of consensus on absolute cutoff levels for androgens that define androgen deficiency quantitatively, (2) weak associations between symptoms and androgen levels, and (3) lack of reliable and sensitive testosterone assays to measure total and free testosterone levels in the low range that is seen in women. Thus, the Princeton Consensus Conference panel’s inability to establish a precise numerical cutoff for low free testosterone levels was influenced by lack of sensitive assays to precisely detect the lower circulating testosterone concentrations seen in women [6, 15]. Previous studies have measured serum testosterone levels using commercial radioimmunoassays that had wide interassay variability and lacked accuracy and reliability in the low range. Liquid chromatography–tandem mass spectrometry and equilibrium dialysis are now widely considered the gold standard methods for measuring total and free testosterone levels, respectively, offering the highest sensitivity and specificity [16, 17]. However, these reference methods are not widely available due to challenges in methodology and expense. In addition to the lack of accuracy of testosterone assays, there is lack of normative data on serum testosterone levels in both healthy menstruating and postmenopausal women, making it further difficult to establish thresholds for defining androgen deficiency in women [7].

In spite of the limitations highlighted above, there remains an enormous public interest and media fascination with the issue of androgen replacement in women for the treatment of sexual dysfunction. Although the data from the Heart and Estrogen/Progestin Replacement Study and the Women’s Health Initiative hormone trials [18, 19] have generated concerns about the role of estrogen replacement in postmenopausal women, there has been a strong advocacy by some members of the scientific community for androgen replacement in these women [20].

Potential Benefits of Testosterone Supplementation in Postmenopausal Women

Testosterone therapy has been widely promoted in women for the treatment of sexual dysfunction and also for improving body composition, muscle performance, bone mineral density, and cognition [21]. It has been assumed that dose–response relationships for testosterone are different in women than in men and that clinically significant effects on sexual function and other health-related outcomes can be achieved in women with testosterone doses and concentrations that are substantially lower than those required to produce similar effects in men [20]. Here we summarize data from randomized trials that have evaluated the efficacy of testosterone therapy on various efficacy outcomes in women. Some trials have used a dose–response design to evaluate effects of various doses of testosterone using different formulations.

Female Sexual Function, Menopause, and Androgens

The menopausal transition has been associated with decreased sexual responsiveness independent of age [22]. Several studies have shown that women who underwent surgical menopause experience greater deterioration of sexual function compared to naturally menopausal women [23–25]. Indeed, women who have undergone bilateral oophorectomy (resulting in low serum testosterone levels) report impaired sexual function even on estrogen therapy [26, 27], suggesting that androgens play an important role in the regulation of libido in women. In 2000, the first female testosterone transdermal patch was manufactured by Procter & Gamble called Intrinsa. This patch is available in 150 ug and 300 ug doses applied to the abdomen twice weekly, providing equivalent to 50 and 100 % daily testosterone production in premenopausal women, respectively [28].

In the last decade, there have been several clinical trials in surgically menopausal women demonstrating improvements in sexual function with physiologic transdermal testosterone replacement that increase serum testosterone levels into the mid- to high-normal range for healthy young women [8, 9, 11, 12]. The first randomized placebo-controlled trial using the patch was published in 2000 [8]. This study was a 12-week trial of 75 surgically menopausal women (ages 31–56) with impaired sexual function (based on the Brief Index of Sexual Functioning for Women Questionnaire) on 0.625 mg of oral estrogen (for at least 2 months) who were randomized to placebo, 150 ug or 300 ug testosterone patch. The mean total testosterone levels increased from 21 ng/dl to 65 ng/dl and 102 ng/dl in the 150 ug and 300 ug groups, respectively. Sexual function (measured by the Brief Index of Sexual Functioning for Women Questionnaire) only improved in the 300 ug-treated group compared to placebo; despite a large placebo response, improvements in sexual thoughts–desire, frequency, pleasure–orgasm, and well-being were seen. Specifically, there was a two- to threefold increase from baseline in percentage of women who had sexual fantasies, masturbated, or engaged in sexual intercourse at least once a week. There was no differences in androgenic (hirsutism, acne) adverse events nor metabolic profile between the groups [29]. Indeed, several follow-up well-designed randomized, controlled studies in both surgically and naturally menopausal women (most trials included women on estrogen replacement) with hypoactive sexual desire disorder (HSDD) have reported that administration of transdermal testosterone resulted in modest improvement in some domains of sexual function such as sexual desire, satisfaction, and frequency [8–12]. Davis and colleagues conducted a trial using the testosterone patch (150 or 300 ug per day vs. placebo) in 814 non-estrogen-treated postmenopausal women (natural and surgical) for 24 weeks [11]. A significant increase by an additional 2.1 satisfying sexual encounters per month was seen only in the 300 ug group versus 0.7 satisfying sexual encounters in the placebo group. There were no significant differences in the adverse side effect profile among the groups.

Despite several studies having demonstrated efficacy and short-term safety of physiologic transdermal androgen therapy for sexual dysfunction in postmenopausal women, the improvements seen have been modest and limited by large placebo effects. Furthermore, recently, two large phase III trials using a transdermal testosterone gel (LibiGel, BioSante, Inc.) failed to meet their primary endpoints of improvements in sexual function (sexual desire and total number of satisfying sexual events) in women with HSDD compared with placebo (data not published). In a recent dose–response study of intramuscular testosterone administration (ranging from physiologic to supraphysiologic doses) in 71 estrogen-treated hysterectomized postmenopausal women (ages 41–62 years) who did not have impaired sexual functioning at baseline, testosterone administration was associated with concentration-dependent improvements in sexual thoughts and desires, sexual activity scores, and arousal, but not in other domains of sexual function [30]. However, these improvements were observed only in women who were assigned to the highest testosterone dose (Fig. 10.4).

Fig. 10.4
figure 4

On-treatment total testosterone concentrations. Data represent mean and standard errors at baseline and on-treatment for each testosterone dose group [30]

Full size image

Although these clinical trials have demonstrated some efficacy of testosterone in the treatment of sexual dysfunction in postmenopausal women, long-term safety of testosterone therapy remains unclear. In 2004, the FDA voted not to approve the transdermal testosterone patch for the treatment of hypoactive sexual desire due to lack of long-term safety data. In 2006, the Endocrine Society recommended against making a diagnosis of androgen deficiency in women due to lack of a well-defined syndrome and normative data for testosterone levels and also recommended against generalized use of testosterone supplementation in women as evidence of safety in long-term studies is lacking [31]. The most updated 2014 Endocrine Society Guidelines on androgen therapy in women continue to recommend against making a diagnosis of FADS and generalized use of testosterone. They also recommended against the use of testosterone therapy in sexual dysfunction, the exception being a postmenopausal women with HSDD, and in the absence of any contraindications, for whom the guidelines suggest a 3- or 6-month trial of testosterone. However, approved testosterone preparations for women are not available in many countries including the United States. Approved testosterone products are available for women in Australia and in some European countries; clinicians in the United States are limited to off-label testosterone preparations, since no testosterone product is FDA approved for use in women.

Body Composition, Muscle Performance, Physical Function, and Androgens

In women, the menopausal transition has been associated with an accelerated loss of muscle mass and strength and decrease in physical function [32, 33]. Although the negative impact of low estrogen on bone health is well established, there is limited evidence on whether this loss of estrogen significantly influences muscle mass and physical function [34–36].

Menopause is associated with an increase in fat mass and a decrease in lean body mass [37, 38]. Conversely, free serum testosterone levels are associated with increased lean mass in older women [39]. The Study of Women’s Health Across the Nation (SWAN) found higher rates of physical limitation in surgically menopausal women compared to women undergoing natural menopause [40]. Indeed, the age-related decline in serum testosterone levels in older women has been associated with frailty [41]. Thus, it is conceivable that women who have undergone surgical menopause and have low testosterone levels may be at a greater lifetime risk for physical disability. Clinical trials using various formulations and doses of testosterone and other androgens have reported some improvements in lean mass and muscle strength, although with few demonstrating improvements in physical function [42–45]. One trial reported modest increase in lean mass in women (aged 19–50 years) with hypopituitarism in response to physiologic testosterone replacement [46]. Another 16-week double-blind randomized controlled trial of 40 postmenopausal women (natural and surgical) who were randomized to either oral 1.25 mg esterified estrogen + 2.5 mg methyltestosterone versus estrogen alone showed that the combined estrogen–androgen group experienced a significant increase in lean mass and muscle strength as well as reduction in total fat mass compared to estrogen alone [42].

Testosterone therapy has also shown beneficial effects on body composition in women who are androgen deficient as a result of another medical condition. Indeed, a study of transdermal testosterone replacement (4 mg/patch) in HIV-infected women with androgen deficiency and 10 % weight loss for 6 months resulted in increases in muscle mass and strength as well as improved muscle function [47]; however, these results have not been confirmed in other studies of women with HIV [48]. In another study of relatively healthy estrogenized postmenopausal women without physical dysfunction at baseline, testosterone administration via intramuscular injections at supraphysiologic doses was associated with significant gains in lean mass, chest press power, and loaded stair climb power; these improvements were observed only at the highest dose (25 mg weekly), with nadir testosterone concentrations of 200 ng/dl [30] (Fig. 10.5). Additional studies are needed to explore the role of androgens in the regulation of body composition, muscle performance, and physical function in various subsets of women (postmenopausal with chronic diseases) with low testosterone levels.

Fig. 10.5
figure 5

Body composition and muscle performance measures. In the bar graphs on the left, data represent absolute mean (SE) changes from baseline for each treatment group. Scatter plots on the right display estimates and 95 % CI for the generalized additive model of change in outcomes as function of free testosterone levels [30]

Full size image

Cognition and Androgens

Epidemiologic studies have suggested that serum testosterone levels in women are associated with specific aspects of cognition, although data regarding the association of circulating testosterone concentrations and cognitive function are inconsistent across studies [49–52]. Testosterone is aromatized to estradiol, both in the periphery and in the brain; in addition to its direct effects via androgen receptor, some effects of testosterone administration might thus be mediated via its aromatization to estradiol [53]. For example, premenopausal women with polycystic ovary syndrome (PCOS) who have higher serum testosterone levels were noted to perform better on spatial tasks and worse on verbal tasks compared to controls [54, 55]. Similarly, endogenous testosterone levels during the menstrual cycle are positively correlated with visuospatial ability and negatively with verbal fluency in healthy women [49, 56]. Similar findings have been noted in some [50], but not in other association studies of endogenous sex hormones and cognition in healthy postmenopausal women [51, 52], and trials of testosterone administration assessing cognitive performance in postmenopausal women have yielded inconsistent results [57, 58]. A 12-week crossover study did not show improvement in cognitive function in surgically menopausal women receiving estrogen alone versus estrogen and intramuscular testosterone [57]. In short-term studies using oral testosterone, improvement in executive function (attention) was noted in one study [59]; however, other studies did not show any cognitive benefits [60, 61]. A recently conducted 24-week dose–response study of testosterone intramuscular injections in estrogen-treated postmenopausal women aged 41–62 years neither showed improvement nor worsening of cognitive performance in a number of domains of cognitive function (spatial reasoning, verbal memory, verbal fluency, and executive function) over a wide range of testosterone doses and concentrations, including doses that achieved supraphysiologic serum testosterone concentrations [62]. In a mechanistic study exploring the role of aromatization on cognition, 71 postmenopausal women who were already receiving transdermal estrogen for 12 weeks were given transdermal 0.5 % testosterone gel and were then randomized either to aromatase inhibitor (letrozole 2.5 mg) or placebo for 16 weeks [63]. Significant improvements in visual and verbal memory were seen with testosterone therapy, and the results were unaffected by aromatase inhibition. These findings suggest that testosterone may have direct cognitive effects independent of its conversion to estradiol. On the contrary, a recently conducted 24-week dose–response study of testosterone replacement in estrogen-treated postmenopausal women neither showed improvement nor worsening of cognitive performance in a number of domains of cognitive function (spatial reasoning, verbal memory, verbal fluency, and executive function) over a wide range of testosterone doses and concentrations, including doses that achieved supraphysiologic serum testosterone concentrations [62].

Contrastingly, some studies have suggested that testosterone therapy may have negative effects on cognitive performance in postmenopausal women. In a study of surgically menopausal women, significant worsening of verbal memory was seen in the group receiving both testosterone and estrogen compared to estrogen alone [58]. Similarly, though in a different context, administration of supraphysiologic doses of testosterone to female-to-male transsexuals has been shown to worsen verbal fluency while improving spatial memory [64]. Long-term, adequately powered trials are needed to evaluate the cognitive effects of testosterone therapy in women. Furthermore, whether testosterone might be beneficial in women with baseline cognitive deficits remains to be investigated.

Androgens and Bone

In women, serum testosterone levels are correlated with trabecular and cortical bone mineral density (BMD), particularly in the older postmenopausal women [39, 65]. Postmenopausal women treated with combined estradiol and testosterone implants for 2 years experienced significantly greater increases in hip and lumbar spine BMD compared to those receiving estradiol implants alone [44]. Two randomized controlled trials in surgically menopausal women compared the effects of a combination of estrogen and methyltestosterone therapy with estrogen therapy alone [66, 67]. One study showed improvement only in the spine BMD in the group receiving testosterone [66], while the other showed improvements at both sites [67]. On the contrary, another study of estrogenized surgically menopausal women who were administered with oral testosterone undecanoate 40 mg daily for 24 weeks did not improve BMD at any of the studied skeletal sites [68]. In yet another small trial of 25 surgically menopausal women on estradiol implants, addition of testosterone for 16 weeks had no significant effects on serum bone markers compared to estrogen alone [69], suggesting that testosterone does not augment the positive effects of estrogen on bone. These results are consistent with studies in young women with primary ovarian insufficiency (who have profound estrogen and testosterone deficiency) where the addition of transdermal testosterone replacement to estrogen did not provide additional benefit [70]. In contrast, another study in postmenopausal women showed that short-term administration of methyltestosterone with estrogen increased markers of bone formation [71].

In summary, data from some, but not all, testosterone intervention studies demonstrate improvement in BMD in postmenopausal women; however, it is unclear whether or not testosterone itself can provide further benefits in the prevention of osteoporosis in adequately estrogenized women. Importantly, the efficacy of testosterone on fracture rates in postmenopausal women remains unknown.

Menopausal Symptoms and Androgens

Currently a combination of oral methyltestosterone and estrogen is FDA approved only for the treatment of moderate to severe menopausal vasomotor symptoms that are not improved by estrogen alone [72]. While the majority of studies have not shown a benefit of methyltestosterone over estrogen [59, 66, 71, 73, 74] in reducing postmenopausal somatic symptoms, two studies have reported greater benefit in menopausal symptoms with methyltestosterone and estrogen combination than estrogen alone [75, 76]. Sherwin et al. reported greater improvements in somatic symptoms in surgically menopausal women treated with combined testosterone and estrogen injections compared with estrogen alone [77]. A large uncontrolled study in 300 pre- and postmenopausal women with symptoms of androgen deficiency showed that continuous testosterone alone, delivered by subcutaneous pellets, was effective in relieving psychological, somatic, and urogenital symptoms, measured by a validated Menopause Rating Scale (MRS) [78]; however this study is limited by a lacking control group for comparison.

In summary, because of the small number of studies and limitation of study design, more evidence is needed to support the efficacy of the addition of testosterone to estrogen in improving menopausal symptoms in both surgically and naturally menopausal women.

Potential Adverse Effects of Testosterone Supplementation in Women

Undesired Androgenic Effects

Potential adverse androgenic effects of testosterone therapy include hirsutism, acne, alopecia, clitoromegaly, and voice deepening. In the majority of randomized placebo-controlled studies of testosterone administration in postmenopausal women [9, 10, 12, 29, 30, 79], the frequency of androgenic events (mainly hirsutism and acne) was higher in the testosterone-treated groups versus placebo; however, most androgenic events were mild in nature. Furthermore, few women withdrew from these studies because of androgenic adverse events [11, 30]. Four cases of clitoral enlargement were reported in one trial of transdermal testosterone patch (one woman receiving 150 ug/day; three receiving 300 ug/day dose), but all were classified as mild by the investigators [11]. None of these women withdrew from the study and clitoromegaly either resolved or remained stable throughout the trial. However, most of the abovementioned androgenic events were collected by subjective report with only a few studies using validated assessment tools. In one dose–response study of 71 estrogen-treated postmenopausal women, androgenic adverse effects were objectively monitored using validated instruments: hair growth using the Ferriman–Gallwey scale, sebum production using Sebutape, acne using the Palatsi scale, clitoral size using a caliper scale, and voice changes using functional acoustic testing [30, 80]. In this trial, clitoral size, Palatsi score, and sebum production rate did not differ significantly between any of the testosterone dose groups (ranging from physiologic to supraphysiologic) versus placebo. However, there were small increases in Ferriman–Gallwey scores in the two highest dose groups (12.5 and 25 mg testosterone enanthate). In the same trial, testosterone administration in women was also associated with dose- and concentration-dependent reduction in average pitch in the higher dose groups [80]. Interestingly, these changes were measureable even though the participants did not report any subjective changes in voice. Early changes in acoustic parameters prior to clinical manifestation have also been reported in patients with Parkinson’s disease, in whom changes in vocal frequency can be detected a decade prior to the clinical diagnosis [81], supporting the notion that detection via functional acoustic testing can be an early marker for subsequent clinical voice changes during testosterone administration. Future clinical trials addressing safety of testosterone therapy should include both subjective and objective evaluations of androgenic side effects.

In summary, the incidence of adverse androgenic side effects of testosterone administration in postmenopausal women at physiologic doses appears to be low and infrequent, and if they do occur, they are generally tolerable.

Cardiometabolic Effects

In women with PCOS, endogenous serum testosterone levels are positively associated with fat mass, proatherogenic dyslipidemia, and insulin resistance [60, 82]. Similarly, in postmenopausal women, higher serum testosterone levels have also been associated with insulin resistance, metabolic syndrome, and coronary heart disease [39, 83]. The data from epidemiologic studies have been extrapolated to suggest that exogenous testosterone administration may also worsen metabolic outcomes in women.

Effects on Body Fat Distribution

Obesity is an important predictor of cardiovascular morbidity and mortality in postmenopausal women [84]. Higher serum-free testosterone levels have been associated with increased total fat mass in older women [39]. There is also evidence suggesting that abdominal obesity is more strongly associated with an androgenic sex hormone profile than generalized obesity in postmenopausal women [85]. Data from SWAN identify a strong positive association between bioavailable testosterone level and increased visceral fat in midlife women [86]. Similarly, women with PCOS have higher subcutaneous and visceral abdominal fat compared to age- and BMI-matched controls [87].

Contrary to epidemiologic studies, data from clinical trials evaluating the effect of exogenous testosterone therapy on central body fat distribution in postmenopausal women are lacking. In female-to-male transsexuals, long-term intramuscular testosterone injections at supraphysiologic doses result in a decrease in subcutaneous fat but increase visceral fat accumulation [88]. In a small study of postmenopausal women who had undergone hysterectomy, the use of low-dose oral methyltestosterone for 1 year showed an increase in visceral fat relative to controls, but was not associated with worsening of insulin resistance [89]. However, this increase in abdominal visceral fat was not seen in a 6-month study of physiologic transdermal testosterone therapy in women with HIV [90]. Another 24-week dose–response study in surgically menopausal women did not show any dose- or concentration-dependent changes in abdominal subcutaneous or visceral fat volumes [91].

Effects on Insulin Resistance

Despite the association between hyperandrogenemia and insulin resistance in PCOS women, most studies in postmenopausal women have not found these associations [92]. Whether insulin resistance precedes or is a consequence of hyperandrogenism still remains unclear. Holmang et al. demonstrated that in an oophorectomized rat model, supraphysiologic doses of testosterone induce insulin resistance [93], suggesting that these effects are dose and concentration dependent. In one euglycemic–hyperinsulinemic clamp study of postmenopausal women with insulin resistance and high free serum testosterone levels, therapy with metformin for 12 weeks resulted in greater insulin sensitivity than those receiving androgen suppression therapy with leuprolide, suggesting that it is insulin resistance that leads to hyperandrogenism [94]. In contrast, another small trial in postmenopausal women, a combination therapy with oral testosterone undecanoate and estradiol, reported a reduction in insulin sensitivity compared to estradiol alone [95], suggesting that it is hyperandrogenism that might lead to insulin resistance. Conversely, in a trial of hysterectomized postmenopausal women, short-term testosterone administration for 24 weeks did not result in significant changes in fasting glucose, fasting insulin, and insulin resistance, even at doses that achieved supraphysiologic serum testosterone concentrations [91]. Similarly, pivotal trials of transdermal testosterone replacement in postmenopausal women also did not demonstrate any significant effects on fasting insulin and glucose [8–12]. Although there appears to be no evidence of harm of testosterone therapy on metabolic parameters, understanding the mechanisms by which androgens influence insulin action and glucose metabolism needs further study.

Lipid Profile

Some of the concerns regarding long-term testosterone therapy in women is the potential to lower HDL and increase LDL cholesterol [92]. Davis et al. found no significant changes in serum lipid profile in naturally and surgically menopausal women who were receiving physiologic doses of transdermal testosterone patch without concurrent estrogen for 6 months [11]. In another study, combined administration of estrogen and testosterone undecanoate did not significantly change HDL, but a reduction in total and LDL cholesterol was seen [95]. Treatment with DHEA (an adrenal androgen) in postmenopausal women significantly lowers HDL particles [96]. Postmenopausal obese women administered nandrolone decanoate (a non-aromatizable androgen) results in significant reduction in HDL and increases in LDL cholesterol [97]. Thus, it is possible that adverse effects of androgens on plasma lipid profile are limited to supraphysiologic doses of androgens and/or non-aromatizable androgen formulations. In a randomized placebo-controlled 24-week dose–response trial in postmenopausal women, total cholesterol, LDL cholesterol, and triglycerides did not change significantly even in women receiving supraphysiologic doses of testosterone injections [30]. On the contrary, HDL levels decreased in all testosterone dose groups compared to placebo; however, these changes were not statistically significant. In contrast, Basaria and colleagues showed that addition of low-dose oral methyltestosterone to oral estrogen in postmenopausal women (surgical or natural) for 16 weeks significantly lowered plasma viscosity (an established risk factor for cardiovascular disease) [98] as well as lipoproteins (total cholesterol, HDL, and triglycerides) [99]. This lowering of plasma viscosity was achieved despite an increase in fibrinogen levels possibly due to significant lowering of lipoproteins.

Overall, the evidence from the transdermal testosterone patch studies have consistently demonstrated no significant changes in lipid profile [10, 12, 29]; however, these trials were primarily designed to test the efficacy of testosterone on sexual function rather than metabolic side effects. Meta-analyses of studies that have investigated effects of combined oral testosterone with estrogen versus estrogen alone on lipid profiles show that testosterone has favorable effects in reducing triglycerides; however, unfavorable effects of decreased HDL and increased LDL cholesterol have also been reported [21].

Vasculature

Studies investigating relationship between testosterone and indices of atherosclerosis have yielded conflicting results. Some studies have shown that postmenopausal women with low testosterone levels have impaired endothelial function [100], while women with PCOS with high testosterone levels have evidence of carotid atherosclerosis and endothelial dysfunction [101]. These findings suggest that the effect that both low and high levels of serum testosterone levels compromise arterial function and the optimum levels are somewhere in between. This concept is supported in animal studies, where administration of testosterone at physiologic doses to androgen-deficient female rats resulted in an improvement in vasodilation [102]. On the contrary, supraphysiologic testosterone administration in female cynomolgus monkeys fed an atherogenic diet resulted in worsening of coronary atherosclerosis [103]. Long-term clinical trials in women are needed to establish the effects of exogenous use of testosterone on atherosclerosis and cardiovascular outcomes.

Breast Cancer

Androgen receptors (AR) are widely expressed in the majority of breast tumors, for which modulation of AR signaling can be either inhibitory or stimulatory [104]. Thus, in the breast, testosterone may exert its effects directly via AR or indirectly by its aromatization to estradiol. Historically, the oncogenic role of AR has been described in ER−/AR+ subtypes that have been demonstrated to have similar signaling to ER+ breast cancers [105]. On the contrary, preclinical studies in animal models and breast cancer cell lines have demonstrated that androgens may also have apoptotic and antiproliferative effects that protect against the oncogenic effects of estrogen on breast tissue [106]. In fact, androgens have been shown to improve response rates in combination with tamoxifen in the treatment of advanced ER+ breast cancer [107, 108]. However, this has not been studied in large randomized controlled trials and is not standard clinical practice, and similar to the safety concerns with estrogen therapy in postmenopausal women, there have been concerns that exogenous testosterone therapy may potentially increase the risk for breast cancer in postmenopausal women. Interestingly, hyperandrogenic women with PCOS who are exposed to a prolonged unopposed (by progesterone) estrogen state have not been shown to have increased risk for breast cancer [105, 106]. However, combined oral estrogen plus progesterone therapy has been shown to increase risk for breast cancer in postmenopausal women [18, 109]. In contrast, in estrogen-depleted postmenopausal women, higher endogenous testosterone levels have been shown to be associated with greater breast cancer risk [106]. Thus, there is potential for exogenous testosterone therapy to have either androgenic or indirect estrogenic effects (via aromatization) on breast tissue, potentially increasing breast cancer risk. In one testosterone patch study in postmenopausal women not taking estrogen, breast cancer was diagnosed in three women in the testosterone-treated groups after 4–12 months of randomization; however, the investigators reported that one of these women in retrospect already had symptoms suggestive of breast cancer prior to randomization [11]. In contrast, in one prospective observational study of pre- and postmenopausal women not on estrogen, treatment with subcutaneous testosterone implants alone or combined with aromatase inhibitors was associated with reduced incidence of breast cancer over 5 years compared with that of age-matched historical controls [110], suggesting that the effects of testosterone on the breast may be protective. Furthermore, emerging evidence suggests that selective androgen receptor modulators may even have a beneficial role in specific subtypes of breast cancer. While AR signaling may have an oncogenic role in ER-breast cancers, development of AR-directed therapies may have positive anti-oncogenic roles in ER+ breast cancers.

In summary, the safety data on exogenous testosterone use and risk for breast cancer in healthy women is inconclusive and needs further study in larger adequately powered randomized placebo-controlled trials.

Summary: Back to the Patient

Our patient is a healthy surgically menopausal women with unequivocally low testosterone levels complaining of sexual dysfunction. The Endocrine Society recommends against generalized use of testosterone by women for sexual dysfunction (except for a specific diagnosis of HSDD). Testosterone therapy would be indicated if she was appropriately diagnosed with HSDD as clinical trial evidence supports efficacy and short-term safety of testosterone therapy for this condition. In the United States (US), there is no FDA-approved testosterone product available for the treatment of female sexual dysfunction due to inadequate long-term safety data. In countries where a testosterone formulation is approved for women, a 3–6-month trial of testosterone therapy in women with properly diagnosed HSDD who do not have contraindications could be considered [31]. Long-term use (beyond 6 months) of testosterone therapy in women is not recommended until long-term efficacy and safety data are available. Despite the lack of an approved testosterone product for the treatment of HSDD, testosterone is commonly being prescribed off-label by practitioners, often as a custom-compounded 1 % topical cream or reduced doses of testosterone gels (e.g., Testim, Androgel) approved for men. However, compounded products are not under strict government regulation and are subject to significant variability in terms of quality, purity, and bioavailability. Whereas women using testosterone formulations approved for men are at increased risk for significant overdosing, the challenges and difficulties in standardizing the dose to achieve physiologic testosterone levels in women can potentially lead to accidental overuse and abuse. Therefore, treatment of women with testosterone products formulated for men or those manufactured by compounding pharmacies is not recommended due to lack of efficacy and safety data.