The challenging role of antiandrogens in the management of polycystic ovary syndrome

Combined oral contraceptives and antiandrogenic progestins

Combined oral (estrogen-progestin) contraceptives (COCs) are recommended for the treatment of clinical hyperandrogenism and/or menstrual irregularity in women and adolescents with PCOS . Various progestin preparations are contained in COCs with distinct pharmacologic characteristics and variable androgenicity ( Table 2 ). COCs containing an antiandrogenic progestin are the drugs of choice for hirsutism, alopecia, and acne . The antiandrogenic effect of COCs is achieved by suppressing the hypothalamus-pituitary-ovarian axis, as they reduce the pulse frequency of gonadotropin hormone-releasing hormone (GnRH). This mechanism inhibits the secretion of follicle-stimulating hormone (FSH) and luteinizing hormone (LH), and subsequently, ovarian androgen production . Furthermore, COCs reduce bioavailable androgen concentrations by stimulating the hepatic production of sex hormone-binding globulin (SHBG) and slightly suppress adrenal androgen secretion . Antiandrogenic progestins ( Table 2 ) exert additional antagonizing effects on the AR and inhibit 5α-reductase activity .

However, COC administration is associated with common adverse effects, such as headaches, mood changes, nausea, breast tenderness, which are usually palliated within the first months, and an increased risk of venous thromboembolism (VTE) . In general, the incidence of VTE associated with COC use is low [i.e., 9–10 events per 10,000 women-years of exposure for COCs containing levonorgestrel (LNG)], less than the respective in gestation, which varies between different COCs . Antiandrogenic progestins may present the highest effect on hyperandrogenism, but exert a higher risk of VTE comparing with progestins with androgenicity . A meta-analysis showed that the VTE risk for COCs containing 30-μg ethinylestradiol (EE) and antiandrogenic progestins, such as cyproterone acetate (CPA) or drospirenone (DSP), was higher compared with COCs containing LNG [pooled risk ratio (RR) 1.5–2.0]. However, there was no difference for dienogest (DNG) . In general, the use of COCs containing EE 35-μg and CPA is recommended only for cases with moderate-to-severe hirsutism or acne, since they confer a higher VTE risk compared with EE 30-μg . On the other hand, antiandrogenic progestins present a more favorable metabolic profile compared with LNG [the latter increase low-density lipoprotein cholesterol (LDL-C) concentrations by 6.8 mg/dL, p < 0.001) and decrease high-density lipoprotein cholesterol (HDL-C) concentrations by 4.4 mg/dL, p < 0.001)] .

COCs containing an antiandrogenic progestin are considered the drug of choice for clinical hyperandrogenism . A meta-analysis evaluated the effect of COCs containing similar estrogen (EE 30–35 μg) and different progestins [CPA, DSP, chlormadinone acetate (CMA), or desogestrel (DSG)] on hyperandrogenism in women with PCOS. Preparations with LNG and gestodene were excluded due to insufficient data . COC use was associated with a decline in Ferriman–Gallwey (FG) score and androgenic biochemical parameters [total (tT) and free testosterone (fT), androstenedione (Δ ₄ A), and dehydroepiandrosterone sulfate (DHEAS)], as well as with an increase in SHBG concentrations . Both short-term and long-term use of all COCs decreased androgen concentrations; nevertheless, the prolonged use (6–12 months) was more effective on hirsutism and SHBG. The use of COCs containing CPA for 12 months presented the greatest impact on hirsutism . This meta-analysis confirmed that long-term treatment with COCs is needed to improve clinical hyperandrogenism (FG scores) and showed that neutral progestins could be as effective as antiandrogenic progestins in improving clinical and biochemical hyperandrogenism .

However, a network meta-analysis including hirsute women, COCs containing LNG, CPA, or DSP had similar effect on hirsutism scores . This study included women with both PCOS and non-PCOS . Moreover, a randomized-controlled trial (RCT) in women with PCOS ( n = 200) showed that COCs containing LNG had similar efficacy in clinical manifestations of hyperandrogenism after 3–6 months of therapy compared with COCs containing DSG, CPA, or DSP. However, COCs with LNG compound had a detrimental effect on lipid metabolism .

Dienogest (DNG)

DNG is a newer progestin used either as monotherapy or in combination with estrogens [EE or estradiol valerate (E ₂ V)] . E ₂ V/DNG was the first COC containing natural estrogen . Chemically, DNG belongs to estranes (19-nortestosterone derivatives), but it has a cyanomethyl instead of an ethinyl group at the C-17 position . As a result, DNG has fewer hepatic effects compared with other C-19 nortestosterone derivatives due to its low activity in the upregulation of mitochondrial cytochrome P450 (CYP) enzymes . Properties related to its C-19 derivative structure include its high oral bioavailability (> 90%) and short plasma half-life (~ 10 h) . DNG circulates bound to albumin by 90%, but not to other binding proteins . Regarding its receptor activities, DNG exhibits antiandrogenic activity and no glucocorticoid or estrogenic activity, in contrast to other ethylestranes. It also presents a very strong progestogenic effect on the endometrium, but no antiestrogen effect on the cardiovascular system and low gonadotropic activity . Data on DNG used alone for the treatment of endometriosis suggest a favorable safety (lipid metabolism, liver enzymes, hemostatic parameters, or thyroid gland metabolism) and tolerability profile even in a high dose (20 mg/day) . In studies conducted in women treated with COCs, the commonly reported adverse effects were comparable to other progestogens, such as headaches, nausea, mood changes, breast tenderness, and abdominal pain . In meta-analyses including COC users, DNG had a positive effect on LDL-C [− 7.7 mg/dL, p = 0.04 (data not available on other lipid parameters)] and comparable VTE risk with LNG, in contrast to the other antiandrogenic progestins [pooled RR 1.46, 95% confidence interval (CI), 0.57–5.41] .

Therefore, DNG may be considered in PCOS management due to its antiandrogenic effect, as an antagonist of the AR, and its favorable metabolic profile. In clinical practice, EE/DNG has a favorable impact on acne . The European Medicines Agency (EMA) has approved its use in the second-line treatment of moderate acne . However, there are limited clinical data concerning DNG's efficacy in the treatment of women with PCOS. In a preliminary observational study of 36 women with PCOS and mild or moderate acne, E ₂ V/DNG exerted a favorable effect on acne and SHBG concentrations after 12 months of treatment . Furthermore, this effect was also shown with EE/DNG in an RCT in women with PCOS ( n = 60), with a further decrease in fT, DHEAS, and Δ ₄ A concentrations, although to a lesser extent compared with EE/DSP, but greater compared with EE/CMA . Further studies, RCTs and meta-analyses are needed to elucidate efficacy of DNG in improving hyperandrogenism in women with PCOS compared with the other antiandrogenic progestins.

Drospirenone (DSP)

DSP is a progestin, available either as a progestogen-only birth-control pill or in combination with estrogen (EE or E ₂ ) . It is the only progestin whose chemical structure is similar to spironolactone, as it is a 17α-spironolactone derivative . DSP has a long half-life (~ 32 h) and oral bioavailability of ~ 76% . The vast majority (95–97%) of circulating DSP binds to albumin. Unlike other progestogens, it has neither binding affinity to SHBG nor corticosteroid-binding globulin (CBG) . DSP is extensively metabolized in the liver. There are two major metabolites in the plasma; M11, the open ring acid form and M14 (4,5 dihydro-drospirenone-3-sulfate), formed independently of the CYP450 and are pharmacologically inactive . DSP presents a moderate binding affinity to the progesterone receptor and a high binding affinity to the androgen and mineralocorticoid receptors (five times higher than that of aldosterone) . DSP progestogenic effect on the endometrium is relatively weak compared with other progestins . Owing to its high antimineralocorticoid property, DSP blocks aldosterone action in the kidneys and causes decreased sodium and water retention compared with other conventional progestins . As a result, DSP may additionally exert a mild antihypertensive effect .

EE/DSP has been used to treat moderate acne, premenstrual syndrome, premenstrual dysphoric disorder, and dysmenorrhea . However, COCs containing DSP are used off-label in women with PCOS, as DSP has a considerable antiandrogenic effect (~ 30% of CPA in in vivo rat models) by blocking testosterone binding to ARs and antigonadotropic activity . Furthermore, DSP does not bind to SHBG, and it does not increase fT . A meta-analysis showed a reduction in FG score in PCOS women with the EE/DSP (30 μg/3 mg) formulation, either after 6 [− 2.57 (− 3.94, − 1.20)] or 12 months of treatment [− 2.15 (− 3.05, − 1.24)] . Though CPA is considered a more potent antiandrogen, a meta-analysis including women with PCOS showed that the impact on clinical [FG score, body mass index (BMI)] and hormonal outcomes [(tT, SHBG, DHEAS, and free androgen index (FAI)] did not differ between DSP and CPA . No significant difference in these outcomes was found between DSP and DSG, a progestin with low androgenicity . Nonetheless, DSP exerts a more favorable effect on lipid and glucose metabolism compared with CPA, as it lowers total cholesterol (TC) concentrations and homeostatic model assessment for insulin resistance (HOMA-IR) index to a greater extent. It is also characterized by a more pronounced effect on HDL-C and LDL-C concentrations compared with DSG .

In addition, in another meta-analysis of patients with PCOS, COCs containing DSP reduced FG score after 6 months and Δ ₄ A and tT concentrations after 3 months, to a greater extent compared with COCs antiandrogenic progestin CMA . On the other hand, EE/DSP monotherapy was more efficacious compared with metformin on biochemical hyperandrogenism (tT, FAI, and SHBG), whereas metformin showed a favorable metabolic profile [BMI, TC, LDL-C, and triglycerides (TGs)]. Moreover, the combination of EE/DSP with metformin was shown to be more effective in reducing BMI and LDL-C and LH concentrations compared with the use of EE/DSP alone .

The EE (20 or 30 μg)/DSP (3 mg) preparations are well-tolerated and present mostly nonspecific, mild-to-moderate, adverse events that do not require further intervention . However, DSP may potentially increase the risk of hyperkalemia, as DSP 3 mg is equivalent to a 25-mg dose of spironolactone . Although this adverse effect has not been proven in clinical practice, except for high-risk patients, package labeling recommends potassium monitoring in the first month for women under additional medications, which predispose to hyperkalemia and against DSP use in women with renal and adrenal insufficiency . In addition, DSP prescription is contraindicated in women with hepatic disease, as DSP is extensively metabolized in the liver . Regarding metabolic parameters, a meta-analysis showed that COCs containing DSP were associated with an increase in TG (32.0 mg/dL; 95% CI, 17.5 to 44.6) and HDL-C (9.26 mg/dL, 95% CI, 4.84 to 13.7), without any effect on LDL-C (3.44 mg/dL, 95% CI -4.29 to 11.2) concentrations, as well as BMI, insulin resistance or glycemia in women with PCOS . Furthermore, COCs containing DSP present higher VTE risk compared with those containing LNG (pooled RR 1.58, 95% CI 1.12–2.14), as shown in another meta-analysis .

In conclusion, DSP is an effective antiandrogenic progestin, with a relatively favorable or neutral profile, concerning the long-term cardiovascular and metabolic aspects of PCOS and higher antimineralocorticoid capacity compared with other progestins contained in COCs . In cases of increased VTE risk or contraindication to estrogens, DSP only pill may be an option, because it does not affect hemostatic mechanisms and presents no estrogenic action .

Cyproterone acetate (CPA)

CPA is the progestin with the highest antiandrogenic activity and is used to treat various androgen-dependent conditions . It is available either as a COC (1–2 mg CPA) in combination with EE and E ₂ V or as monotherapy (at daily doses of 10 mg, 50 mg or 100 mg) or as a solution for injection (300 mg/3 mL) . After oral administration, the bioavailability of CPA is nearly 100% . It circulates bound to albumin by 93% and presents no binding affinity to SHBG and CBG . Since there is no binding of CPA to SHBG or CBG in the serum, 93% of the compound is bound to albumin . A single oral dose of 2 mg/day leads to peak serum concentrations of about 11 ng/mL, with half-lives of 2–8 h (t1/2a) and 60 h (t1/2b) . The major metabolic steps are hydroxylation and deacetylation. The metabolite 15b-hydroxy-CPA has an antiandrogenic activity similar to that of CPA, but only 10% of its progestogenic effect .

CPA is a highly potent progestogen (90% binding affinity for the progesterone receptor) and a moderately potent antiandrogen (6–21% binding affinity for the androgen). It also exhibits antigonadotropic, antiestrogenic, and glucocorticoid properties . The antiandrogenic effect of CPA results from its competitive inhibition of the binding of endogenous androgens to the AR and is dose-dependent . In addition, CMA blocks testosterone's conversion to the more potent androgen, 5α-dihydrotestosterone (by inhibiting 5α-reductase in the skin) and reduces androgen production in the ovary . As it presents no binding affinity to SHBG, CPA does not affect fT .

EE (35 μg)/CPA (2 mg) has been widely used in the management of hyperandrogenism in women with PCOS . In a meta-analysis which evaluated the effect of COCs in women with PCOS, both short-term (3 months) and long-term (6–12 months) use of EE/CPA was associated with a decline in FG score [− 1.22, (96% CI − 1.62 to − 0.81) and − 7.96 (95% CI −8.65 to − 7.27) after 3 and 12 months, respectively]. EE/CPA also decreases androgen (tT, fT, and Δ ₄ A) and increases SHBG concentrations at 3 and 12 months of continuous use . The greatest effect on hirsutism by EE/CPA was evident after 12 months of treatment [− 7.96 (95% CI − 8.65 to 7.27)] . EE/CPA may also improve co-existing acne and seborrhea in women with PCOS . However, after treatment cessation, it must be underlined that the antiandrogenic effect seems to vanish in a few months, indicating the need for long-term therapy .

EE/CPA use is generally well-tolerated with a safety profile similar to other COCs . However, it is contraindicated in women with a history of meningioma . From a metabolic point of view, the EE/CPA may negatively affect lipid metabolism, increasing TG concentrations . A meta-analysis of studies in premenopausal women with PCOS showed that EE/CPA increase plasma TG and HDL-C concentrations by 25.1 mg/dL (95% CI 13.8–36.4) and 6.09 mg/dL (95% CI 1.91–10.3), respectively . The effect of CPA on glucose homeostasis (fasting plasma glucose, HOMA-IR) and BMI is negligible . Nonetheless, EE/CPA use has been associated with increased VTE risk [pooled RR 2.04 (95% CI 1.55–2.49)] compared with LNG . Thus, its use is recommended only when treating moderate to severe hirsutism or acne .

Higher CPA dosages as monotherapy have also been used cyclically, in combination with a COC or estrogens, to treat hyperandrogenic states . CPA tablets are used internationally, but they are not available in the USA . The recommended dose of CPA is 50–100 mg/day for 10 days/cycle for the management of hirsutism . However, few pooled data about the use of CPA alone in treating women with PCOS are included in this context. In a Cochrane meta-analysis, no differences were found in FG score or serum androgen (tT, fT, Δ ₄ A, and DHEAS) concentrations at 6 months between EE/CPA and CPA (at daily doses of 25–100 mg) in hirsute women with PCOS . Only one study on this concept was included, with a relatively small sample size ( n = 113) . Moreover, no clinical differences on hirsutism (except from FG score at 12 months favoring CPA) were observed when CPA was compared with other antiandrogens (spironolactone, flutamide, and finasteride), as well as GnRH-analogs in hirsute women with or without PCOS . A network meta-analysis on the treatment options for hirsutism in women (PCOS and non-PCOS) showed that the combination of a COC with CPA was equally effective compared with COC monotherapy [− 0.49 (− 1.41 to 0.43)], but the certainty of comparisons was low . CPA (50–100 mg/day) may also be beneficial in treating female pattern hair loss .

Adverse effects of higher doses (> 10 mg) of CPA include liver toxicity, irregular menstrual bleeding, nausea, decreased libido and mood, liver tumors and meningiomas . Furthermore, the CPA may increase the risk of undervirilization of male fetuses at higher doses, as proved in animal models . EMA's safety committee has recommended that CPA daily doses of ≥ 10 mg should only be used for androgen-dependent conditions, when other treatment options (e.g., EE/CPA 2 mg) have failed. Furthermore, the dose should be gradually reduced to the lowest effective one, because of the increased risk for meningiomas, reported primarily at doses of ≥ 25 mg/day. In clinical practice, patients should be monitored for clinical signs and symptoms of meningioma .