A Systematic Review of Randomized Controlled Trials Investigating the Efficacy and Safety of Testosterone Therapy for Female Sexual Dysfunction in Postmenopausal Women

Channa N. Jayasena; Fatima M. Alkaabi; Curtis S. Liebers; Thomas Handley; Stephen Franks; Waljit S. Dhillo


Clin Endocrinol. 2019;90(3):391-414. 

In This Article

Nonsexual Effects of Testosterone Therapy

Testosterone therapy is anticipated to cause symptoms of androgen excess such as excess body hair (hirsutism) and acne. Accordingly, clinical trials reported that the most common adverse effects associated with testosterone therapy were skin reactions, unwanted hair growth, acne and vaginal bleeding; however, most were mild and rarely resulted in withdrawal from the study.[18,44,46,50,52,57,62,63,64,65] Furthermore, two published studies observed no significant difference in adverse effects between treatment groups.[66,67]

Breast Cancer Risk

Exogenous sex steroids may stimulate the growth of sex-hormone dependent tissues, most notably breast tissue in women,[68,69] so it is relevant to consider whether testosterone therapy increases the risk of breast cancer. The only published long-term investigation of testosterone therapy in postmenopausal women has been conducted by the product's manufacturer.[62] They performed an open-label, uncontrolled safety and tolerability trial of transdermal testosterone (daily dose, 300 μg) in over 900 women with surgical menopause and HSDD, for up to 4 years of duration. Three cases of invasive breast cancer during 4 years of TTP administration; however, the authors concluded that this was not inconsistent with background rates for women in the same age-group. The randomized controlled trial by Davis et al[48] over 24 weeks observed 4 cases of breast cancer in the testosterone group when compared with no cases in the placebo group. However, the same authors performed a follow-up study suggesting that 1 year of TTP therapy had no significant effect on digitally quantified absolute or per cent dense mammographic area in postmenopausal women when compared with placebo;[70] these data are concordant with results of a study by Hofling et al of 6 months TTP administration in 99 postmenopausal women.[71] Furthermore, Hofling et al also observed no significant increase in breast cell proliferative activity with testosterone compared to placebo.[72]

Endometrial Effects

Testosterone promotes endometrial atrophy when given without concomitant oestrogen. Davis et al[48] investigated endometrial findings in 814 women randomly assigned to receive a patch delivering 150 or 300 μg of testosterone per day or placebo. They observed that endometrial bleeding was reported more frequently on the 300 g/d dose testosterone (10.6%) compared with 150 g/d testosterone (2.7%) or placebo (2.6%). Endometrial bleeding was accompanied at the highest dose by endometrial atrophy.[48] It is therefore important to provide concurrent continuous or cyclic progestin therapy to nonhysterectomized women administered combined oestrogen and testosterone treatment as per normal practice.


Whilst many studies have suggested that testosterone administration may cause lipid disarrangements such as reduced high-density lipoprotein (HDL)[18,49,55,57,65,73,74,75,76] and an increase in low-density lipoprotein (LDL),[18,49] some indicate no change in LDL,[73,74,76] a decrease in LDL[75] or no significant change in lipid profile at all with testosterone.[77,78,79] Studies seem inconsistent regarding total cholesterol with some observing no change with testosterone compared to placebo,[73,74,78] whilst others observed a decrease.[55,57,75,76] A similar pattern of inconsistency in the literature is observed with the effects of testosterone on triglycerides; some studies observe a decrease,[55,57,75,76] whereas others observe no change with testosterone.[78] The authors of a 2011 study failed to observe any significant increase in haematocrit, or adverse changes in glycemic markers or lipid profile following up to 4 years of TTP therapy.[62] Further, long-term studies are required to investigate the potential effects of testosterone supplementation on risks of malignancy and cardiovascular disease in postmenopausal women. However, the available evidence allows us to speculate that any effects of testosterone therapy on serum lipids are likely to be minor, which is of relevance when discussing therapeutic options with patients.

Bone Mineral Density

The Women's Health Initiative Observational Study of 93 676 postmenopausal women aged 50-79 years investigated hip fracture risk in relation to the circulating level of testosterone. They observed higher circulating levels of serum SHBG is associated with an increased risk of subsequent hip fracture and high endogenous testosterone with a decreased risk, independent of each other, serum oestradiol concentration and other recognized risk factors.[80]

The reported effects of testosterone on bone of naturally postmenopausal women are heterogeneous. Elraiyah et al[18] conducted a systematic review and meta-analysis, which included 35 randomized trials, four of which reported on bone densitometry during combined testosterone and oestrogen therapy, and compared it to oestrogen alone. They concluded that testosterone had no significant effect on bone mineral density (BMD) in any tested site concordant with a study by Garnett et al in 1992.[81] However, Miller et al[82] observed BMD of the hip increased significantly with testosterone compared to HRT alone, whilst a study by Barrett-Connor et al[67] observed an increase in BMD both at the hip and lumbar spine with testosterone treatment when compared to placebo. Further, a 1995 study by Davis et al[13] on 34 postmenopausal women with oestradiol alone, or oestradiol plus testosterone implants observed BMD increased significantly for total body, at vertebrae L1-L4 and the trochanter with testosterone compared to oestradiol alone. In addition, whilst Miller et al[82] observed no difference in bone biochemical markers (Ntx, Dpd or BSAP) between treatment groups, another observed markers of bone formation (IGF-1 and P1CP) were significantly higher with testosterone compared to placebo.[83] No trials have reported long-term fracture rates during testosterone therapy in postmenopausal women. In summary, early data suggest that testosterone may have some beneficial effects of increasing BMD at specific locations in the body, but further longer-term studies are required to provide more conclusive evidence.

Cardiovascular Effects

Exogenous testosterone increases haematocrit, so it is important to consider if the risk of thromboembolic and cardiovascular disease is altered during therapy.[84] Some recent studies have shown an increased risk of cardiovascular events in men treated with testosterone.[85,86,87] However, other studies data did not provide any evidence of an association between testosterone and cardiovascular events.[88,89,90] Testosterone-containing medicines are licensed in the European Union (EU) for the treatment of male hypogonadism. There is limited experience on the safety and efficacy of the use of these medicines in patients over 65 years of age, and the use of testosterone to boost these levels in healthy older men is not authorized in the EU. In the United States (US), the US Food and Drug Administration (FDA) cautions that prescription testosterone are approved only for men who have low testosterone levels caused by certain medical conditions. The safety and benefit of these medications have not been established for the treatment of low testosterone levels due to ageing.[91,92] There is no specific evidence about the cardiovascular effect in women treated with testosterone. The randomized control trials comparing testosterone therapy in women with placebo have not observed any significant differences in event rates for any cardiovascular disease outcomes, including venous thromboembolic events in short-term trials. Furthermore, adverse cardio-metabolic changes have not been frequent during short-term observations (12-24 months) in women treated with testosterone.[28] Data suggest that the testosterone transdermal patch improves exercise tolerance, muscle strength and insulin resistance without side effects in elderly female patients with stable chronic heart failure. In a double-blind, randomized, placebo-controlled study in women with heart failure, testosterone therapy (the TTP releasing 300 μg/d) was associated with significant functional improvements assessed by peak oxygen consumption, distance walked over the 6-minute walking test, muscle strength and insulin resistance compared with placebo. Moreover, a systematic review of testosterone therapy on cardiovascular outcomes in postmenopausal women was conducted by Spoletini et al 2014; they observed a favourable effect of testosterone therapy in postmenopausal women, such as high-density lipoprotein cholesterol, total cholesterol, body fat mass and triglycerides.[93,94,95,96,97] In addition, a study by Kocoska-Maras et al in 2009[98] observed testosterone counteracts the oestrogen induced rise in high-sensitivity C-Reactive Protein (hs-CRP) but had no effects on other inflammatory markers of cardiovascular disease, such as Interleukin-6, Tumour Necrosis Factor-Alpha and Homocysteine. In contrast, a 2014 study[99] observed no significant difference in hs-CRP levels between testosterone and placebo treatments. Studies appear to agree that testosterone has no significant effect on systolic and diastolic blood pressure compared to placebo.[83,99] Testosterone has also been observed to significantly increase fibrinogen levels with testosterone compared to placebo.[76] In summary, despite many studies being conducted on this topic, the long-term cardiovascular consequences of testosterone therapy remain unclear.[28]

Anthropometric Measurements

Studies seem to agree that testosterone has no significant effect on BMI compared to placebo,[73,83] but are inconclusive when it comes to body composition. A 2006 study by Zang et al on 63 postmenopausal women[77] observed no significant difference in body fat with testosterone compared to other treatment groups, similar with a study by Davis et al,[79] whilst a 2005 study[83] observed lean body mass significantly increased, consistent with another study[57] indicating significant increases in lean body mass of the legs, arms and trunk. A study by Duarte et al[73] observed visceral fat significantly increased with testosterone by 11% vs placebo, in direct contrast with a 2014 study stating no difference in both abdominal and visceral fat volumes between treatment groups.[99] One study observed free fat mass increased[100] when given testosterone vs placebo, but others observed no change in fat mass[73,77] with one indicating a decrease in % fat tissue with testosterone.[57] Overall, whilst studies agree that testosterone has no effect on BMI, there is no consensus on its effects on fat mass in the absence of long-term large scale trials.

Cognitive Function

The effects of steroid hormones on human cognition are of considerable interest, particularly during the menopause. A study by Krug et al on 12 postmenopausal women observed testosterone significantly increased divergent thinking, (fantasies and fluency of speech) compared to placebo,[101] but no other aspects of cognitive function were affected. This is in direct contrast to many other studies indicating no change in verbal fluency, or any other measures of cognitive function.[102,103,104] A 2014 study observed testosterone therapy had no effect on cognition after 12 weeks of treatment compared to placebo, but at 26 weeks, the Cogstate International Shopping List Test (ISLT) score significantly increased by 1.57 units vs placebo.[104] One study observed immediate verbal memory to be impaired with testosterone, but no other memory functions were affected.[105] A 2002 study by Wisniewski et al on 26 naturally and surgically postmenopausal women discovered testosterone significantly increased building memory score compared to placebo, despite no other aspects of cognitive function being affected (cube comparison, shape memory or identical pictures).[106] In summary, studies published to date have observed a variety of effects on cognition following testosterone therapy; it is therefore difficult to draw conclusions on the predicted effects of testosterone in postmenopausal women.

Physical Symptoms

A 2002 study by Dobs and colleagues gave esterified oestrogens with or without methyltestosterone to 40 naturally and surgically menopausal women and observed testosterone significantly improved both somatic and vasomotor symptoms compared to oestrogen alone.[57] However, the majority of studies seem to agree testosterone therapy has no significant beneficial effect on suppressing hot flash frequency and severity.[58,64,67,107] A further study observed no significant improvement in somatic or psychological symptoms with testosterone compared to placebo[58] whilst another observed no significant difference in sweating and vaginal dryness between treatment groups.[67] In summary, most studies suggest that testosterone does not affect menopausal hot flushes and it remains to be determined whether somatic symptoms or vaginal dryness is altered during testosterone therapy.