Evaluation and Treatment of Hirsutism in Premenopausal Women (2018) PDF

CL IN IC A L P RA CT I CE G UI DE L I NE Evaluation and Treatment of Hirsutism in Premenopausal Women: An Endocrine Society* Clinical Practice Guideline...

CL IN IC A L P RA CT I CE G UI DE L I NE Evaluation and Treatment of Hirsutism in Premenopausal Women: An Endocrine Society* Clinical Practice Guideline Kathryn A. Martin,1 R. Rox Anderson,1 R. Jeffrey Chang,2 David A. Ehrmann,3 Rogerio A. Lobo,4 M. Hassan Murad,5 Michel M. Pugeat,6 and Robert L. Rosenfield3 Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 1 Massachusetts General Hospital, Boston, Massachusetts 02114; 2University of California, San Diego, La Jolla, California 92037; 3University of Chicago, Chicago, Illinois 60637; 4Columbia University, New York, New York 10032; 5Mayo Clinic Evidence-Based Practice Center, Rochester, Minnesota 55905; and 6 Hospices Civils de Lyon, Bron, France F-69677 *Co-Sponsoring Associations: Androgen Excess and Polycystic Ovary Syndrome Society and European Society of Endocrinology. Objective: To update the “Evaluation and Treatment of Hirsutism in Premenopausal Women: An Endocrine Society Clinical Practice Guideline,” published by the Endocrine Society in 2008. Participants: The participants include an Endocrine Society–appointed task force of seven medical experts and a methodologist. Evidence: This evidence-based guideline was developed using the Grading of Recommendations, Assessment, Development, and Evaluation system to describe the strength of recommendations and the quality of evidence. The task force commissioned two systematic reviews and used the best available evidence from other published systematic reviews and individual studies. Consensus Process: Group meetings, conference calls, and e-mail communications facilitated consensus development. Endocrine Society committees, members, and cosponsoring organizations reviewed and commented on preliminary drafts of the guidelines. Conclusion: We suggest testing for elevated androgen levels in all women with an abnormal hirsutism score. We suggest against testing for elevated androgen levels in eumenorrheic women with un- wanted local hair growth (i.e., in the absence of an abnormal hirsutism score). For most women with patient-important hirsutism despite cosmetic measures (shaving, plucking, waxing), we suggest starting with pharmacological therapy and adding direct hair removal methods (electrolysis, photoepilation) for those who desire additional cosmetic benefit. For women with mild hirsutism and no evidence of an endocrine disorder, we suggest either pharmacological therapy or direct hair removal methods. For pharmacological therapy, we suggest oral combined estrogen–progestin contraceptives for the majority of women, adding an antiandrogen after 6 months if the response is suboptimal. We recommend against antiandrogen monotherapy unless adequate contraception is used. We suggest against using insulin-lowering drugs. For most women who choose hair removal therapy, we suggest laser/photoepilation. (J Clin Endocrinol Metab 103: 1233–1257, 2018) ISSN Print 0021-972X ISSN Online 1945-7197 Abbreviations: CI, confidence interval; CPA, cyproterone acetate; DHEA, dehy- Printed in USA droepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; DSP, drospirenone; EE, Copyright © 2018 Endocrine Society ethinyl estradiol; FDA, Food and Drug Administration; GnRH, gonadotropin- Received 29 January 2018. Accepted 29 January 2018. releasing hormone; IPL, intense pulsed light; NCCAH, nonclassical congenital adrenal First Published Online 7 March 2018 hyperplasia; OC, oral contraceptive; PCOS, polycystic ovary syndrome; PH, paradoxical hypertrichosis; RCT, randomized controlled trial; SHBG, sex hormone–binding globulin; VTE, venous thromboembolism. doi: 10.1210/jc.2018-00241 J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 https://academic.oup.com/jcem 1233 1234 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 Summary of Recommendations 3.2. For most women with hirsutism, we suggest against antiandrogen monotherapy as initial 1.0 Diagnosis of hirsutism therapy (because of the teratogenic potential of 1.1. We suggest testing for elevated androgen levels these medications) unless these women use ade- in all women with an abnormal hirsutism score quate contraception (2 |OOO). However, for (2 |OO). In those cases where serum total tes- women who are not sexually active, have tosterone levels are normal, if sexual hair growth is undergone permanent sterilization, or who are moderate/severe or sexual hair growth is mild but using long-acting reversible contraception, we there is clinical evidence of a hyperandrogenic en- suggest using either oral contraceptives or anti- docrine disorder (such as menstrual disturbance or androgens as initial therapy (2 |OOO). The choice between these options depends on pa- Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 progression in spite of therapy), we suggest mea- suring an early morning serum total and free tes- tient preferences regarding efficacy, side effects, tosterone by a reliable specialty assay. (2 |OO) and cost. 1.2. We suggest screening hyperandrogenemic women 3.3. For most women, we do not suggest one oral for NCCAH due to 21-hydroxylase deficiency by contraceptive over another as initial therapy, as measuring early morning 17-hydroxyprogesterone all oral contraceptives appear to be equally ef- levels in the follicular phase or on a random day fective for hirsutism, and the risk of side effects is for those with amenorrhea or infrequent menses low. (2 |OO) (2 |OO). In hirsute patients with a high risk of 3.4. For women with hirsutism at higher risk for congenital adrenal hyperplasia (positive family venous thromboembolism (e.g., those who are history, member of a high-risk ethnic group), we obese or over age 39 years), we suggest initial suggest this screening even if serum total and free therapy with an oral contraceptive containing the testosterone are normal. (2 |OO) lowest effective dose of ethinyl estradiol (usu- 1.3. We suggest against testing for elevated androgen ally 20 mcg) and a low-risk progestin (Table 2). levels in eumenorrheic women with unwanted (2 |OOO) local hair growth (i.e., in the absence of an ab- 3.5. If patient-important hirsutism remains despite normal hirsutism score) because of the low like- 6 months of monotherapy with an oral contra- lihood of identifying a medical disorder that would ceptive, we suggest adding an antiandrogen. change management or outcome. (2 |OO) (2 |OO) 3.6. We do not suggest one antiandrogen over an- other (2 |OO). However, we recommend 2.0 Treatment of hirsutism in against the use of flutamide because of its po- premenopausal women tential hepatotoxicity. (1 |OO) 3.7. For all pharmacologic therapies for hirsutism, we 2.1. For most women with patient-important hirsutism suggest a trial of at least 6 months before making despite cosmetic measures, we suggest starting with changes in dose, switching to a new medication, pharmacological therapy (2 |OOO). For women or adding medication. (2 |OOO) who then desire additional cosmetic benefit, we 3.8. In patients with severe hirsutism causing emo- suggest adding direct hair removal methods. tional distress and/or in those women who have However, for women with mild hirsutism and no used oral contraceptives in the past and have not evidence of an endocrine disorder, we suggest either experienced sufficient improvement, we suggest approach. (2 |OOO) initiating combination therapy with an oral con- 2.2. For hirsute women with obesity, including those traceptive and antiandrogen (2 |OO). However, with polycystic ovary syndrome, we also rec- we suggest against combination therapy as a ommend lifestyle changes. (1 |OO) standard first-line approach. (2 |OO) 3.0 Pharmacological treatments Other drug therapies Initial therapies 3.9. We suggest against using insulin-lowering drugs 3.1. For the majority of women with hirsutism who are for the sole indication of treating hirsutism. not seeking fertility, we suggest oral contraceptives (2 |OO) as initial therapy for treating patient-important 3.10. We suggest against using gonadotropin- hirsutism. (2 |OO) releasing hormone agonists except in women doi: 10.1210/jc.2018-00241 https://academic.oup.com/jcem 1235 with severe forms of hyperandrogenemia (such other clinical evidence of hyperandrogenemia, such as as ovarian hyperthecosis) who have a subopti- progressive growth of hair in androgen-dependent areas mal response to oral contraceptives and anti- (sexual hair). We have added a recommendation to screen androgens. (2 |OOO) hyperandrogenemic women for nonclassic congenital 3.11. We suggest against the use of topical anti- adrenal hyperplasia (NCCAH) due to 21-hydroxylase androgen therapy for hirsutism. (2 |OOO) deficiency by measuring early morning 17-hydroxy- progesterone levels in the follicular phase or on a random 4.0 Direct hair removal methods day for those with amenorrhea or infrequent menses. In women with hirsutism who are at high risk for NCCAH 4.1. For women who choose hair removal therapy, (positive family history, member of a high-risk ethnic we suggest photoepilation for those whose un- group), we suggest screening even if serum total and free Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 wanted hair is auburn, brown, or black, and we testosterone are normal. suggest electrolysis for those with white or blonde hair. (2 |OO) Treatment 4.2. For women of color who choose photoepilation For treatment, we have made the following revisions to treatment, we suggest using a long-wavelength, the new guideline: long pulse-duration light source such as Nd:YAG We now suggest either pharmacologic therapy or or diode laser delivered with appropriate skin direct hair removal methods as initial therapy for cooling (2 |OOO). Clinicians should warn women with mild hirsutism and no evidence of an Mediterranean and Middle Eastern women with endocrine disorder. For other women with patient- facial hirsutism about the increased risk of de- important hirsutism, we suggest starting with phar- veloping paradoxical hypertrichosis (PH) with macological therapy and adding direct hair removal photoepilation therapy. We suggest topical methods if needed. treatment or electrolysis over photoepilation with We added a recommendation that it is reasonable to these patients. (2 |OO) start with combined pharmacological therapy [oral 4.3. For women who desire more rapid response to contraceptives (OCs) and antiandrogens] in select photoepilation, we suggest adding eflornithine women with severe hirsutism that is causing distress. topical cream during treatment. (2 |OO) We added a recommendation to use lower estrogen- 4.4. For women with known hyperandrogenemia dose OCs with low-risk progestins in women at who choose hair removal therapy, we suggest higher risk for venous thromboembolism (VTE) pharmacologic therapy to minimize hair regrowth. (e.g., obese, age.39 years). For other women, our (2 |OO) approach is the same as in the original guideline: we do not suggest one OC formulation over another. Changes Since the Previous Guideline We made a stronger recommendation against the use of flutamide for hirsutism. In 2008, the Endocrine Society published the clinical We added a suggestion for electrolysis rather than practice guideline “Evaluation and Treatment of Hirsut- photoepilation in women with blond or white hair ism in Premenopausal Women.” As hirsutism is common, who choose direct hair removal methods. We also often associated with an underlying endocrine disorder, provide guidance for the use of photoepilation (and and associated with significant personal distress, treatment its potential complications) in women of color. is appropriate for most women who present with this We added a lifestyle recommendation for women problem. In this current version, we have attempted to with polycystic ovary syndrome (PCOS). address several issues, as well as incorporate insights from relevant studies published since the 2008 guideline. Im- portant modifications in this version are as follows. Method of Development of Evidence-Based Clinical Practice Guidelines Evaluation We have broadened the suggestion for determining the The Clinical Guidelines Subcommittee of the Endocrine serum total testosterone concentration to include all Society deemed the evaluation and treatment of hirsut- women with hirsutism and have broadened the suggestion ism in premenopausal women a priority area for revision for determining the serum-free testosterone concentration and appointed a task force to formulate evidence-based to include hirsute women whose serum total testosterone is recommendations. The task force followed the ap- normal in the presence of moderate to severe hirsutism or proach recommended by the Grading of Recommendations, 1236 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 Assessment, Development, and Evaluation group, an in- The Endocrine Society provided the funding for this ternational group with expertise in the development and guideline; the Task Force received no funding or re- implementation of evidence-based guidelines (1). A detailed muneration from commercial or other entities. description of the grading scheme has been published else- The Endocrine Society’s clinical practice guidelines are where (2). The task force used the best available research developed to be of assistance to endocrinologists by pro- evidence to develop the recommendations. The task force viding guidance and recommendations for particular areas also used consistent language and graphical descriptions of of practice. The guidelines should not be considered inclusive both the strength of a recommendation and the quality of of all proper approaches or methods, or exclusive of others. evidence. In terms of the strength of a recommendation, The guidelines cannot guarantee any specific outcome, nor strong recommendations use the phrase “we recommend” do they establish a standard of care. The guidelines are not and the number 1, and conditional recommendations use the intended to dictate the treatment of a particular patient. Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 phrase “we suggest” and the number 2. Cross-filled circles Treatment decisions must be made based on the independent indicate the quality of the evidence, such that OOO de- judgment of health care providers and each patient’s indi- notes very low-quality evidence; OO, low quality; vidual circumstances. O, moderate quality; and , high quality. The The Endocrine Society makes no warranty, express or task force has confidence that persons who receive care implied, regarding the guidelines and specifically ex- according to the strong recommendations will derive, on cludes any warranties of merchantability and fitness for a average, more good than harm. Conditional recommen- particular use or purpose. The Society shall not be liable dations require more careful consideration of the person’s for direct, indirect, special, incidental, or consequential circumstances, values, and preferences to determine the best damages related to the use of the information contained course of action. Linked to each recommendation is a de- in this work. scription of the evidence and the values that the task force considered in making the recommendation. In some in- Commissioned Systematic Review stances, there are remarks in which the task force offers technical suggestions for testing conditions, dosing, and The Endocrine Society Task Force commissioned two monitoring. These technical comments reflect the best systematic reviews in 2008 that were updated to support available evidence applied to a typical person being treated. the current guideline. The updated review included a Often this evidence comes from the unsystematic observa- network meta-analysis that compared the available 37 tions of the task force and their preferences; therefore, one randomized controlled trials (RCTs) of pharmacologic should consider these remarks as suggestions. therapy for hirsutism. This network meta-analysis ap- The Endocrine Society maintains a rigorous conflict- proach facilitated simultaneous comparison of multiple of-interest review process for developing clinical practice agents and allowed indirect comparison of interventions guidelines. All task force members must declare any (that have not been evaluated in head-to-head trials) potential conflicts of interest by completing a conflict-of- based on their effect on a common comparator. interest form. The Clinical Guidelines Subcommittee The goals of the systematic reviews and meta-analyses reviews all conflicts of interest before the Society’s were to: Council approves the members to participate on the task Update the analyses of the efficacy and safety of force and periodically during the development of the OCs, antiandrogens, and metformin vs placebo, and guideline. All others participating in the guideline’s de- OCs plus antiandrogens vs OCs, for the treatment of velopment must also disclose any conflicts of interest in hirsutism; and the matter under study, and most of these participants Compare the impact on hirsutism of OCs containing must be without any conflicts of interest. The Clinical antiandrogens [cyproterone acetate (CPA) and Guidelines Subcommittee and the task force have drospirenone (DSP)] vs other OCs, and OCs con- reviewed all disclosures for this guideline and resolved or taining levonorgestrel (the most androgenic pro- managed all identified conflicts of interest. gestin) vs other OCs. Conflicts of interest are defined as remuneration in any amount from commercial interests; grants; research The results of the network analysis were consistent support; consulting fees; salary; ownership interests [e.g., with the previous meta-analyses, showing that OCs, anti- stocks and stock options (excluding diversified mutual androgens, and the combination of OCs plus antiandrogens funds)]; honoraria and other payments for participation were all more effective than placebo and led to reduction in in speakers’ bureaus, advisory boards, or boards of di- hirsutism scores. The addition of antiandrogens to OCs was rectors; and all other financial benefits. Completed forms slightly more effective than OCs alone for hirsutism. Met- are available through the Endocrine Society office. formin therapy was not superior to placebo. OCs containing doi: 10.1210/jc.2018-00241 https://academic.oup.com/jcem 1237 antiandrogens were no more effective than other OCs, and Table 1. Definition of Terms OCs containing levonorgestrel were equally effective as other OCs for the treatment of hirsutism. The results of the Term Definition review serve as the evidence base for the recommendations Hirsutism Hirsutism is excessive terminal hair that about pharmacologic therapy. appears in a male pattern (excessive hair in androgen-dependent areas; i.e., sexual hair) in women. Ferriman–Gallwey The modified Ferriman–Gallwey score is Definition, Pathogenesis, and Etiology score the gold standard for evaluating of Hirsutism hirsutism. Nine body areas most sensitive to androgen are assigned Hirsutism is excessive terminal hair that appears in a male a score from 0 (no hair) to 4 (frankly virile), and these separate scores are Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 pattern in women (3) (Table 1). Some sexual hair growth summed to provide a hormonal is normal, but clinicians commonly diagnose hirsutism hirsutism score (Fig. 1). as a Ferriman–Gallwey score (4) above the 95th per- Local hair growth This is unwanted localized hair growth in centile for the population (Fig. 1) (5, 6). Ferriman– the absence of an abnormal hirsutism score. Gallwey total scores that define hirsutism in women of Patient-important Unwanted sexual hair growth of any reproductive age are as follows: United States and United hirsutism degree that causes sufficient distress for Kingdom black or white women, $8 (6); Mediterranean, women to seek additional treatment. Hispanic, and Middle Eastern women, $9 to 10 (6); Hyperandrogenism Hyperandrogenism (for the purposes of this guideline) is defined as clinical South American women, $6 (7); and Asian women, a features that result from increased range of $2 for Han Chinese women (6) to $7 for androgen production and/or action. Southern Chinese women (8, 9). Although widely used, Idiopathic hirsutism This is hirsutism without hyperandrogenemia or other signs or this scoring system has its limitations, which include its symptoms indicative of a hyperandrogenic subjective nature, the failure to account for a locally high endocrine disorder. score that does not raise the total score to an abnormal extent, and the lack of consideration of such androgen- (3, 15). Androgens appear to induce vellus follicles in sex- sensitive areas such as the sides of the face from the specific areas to develop into terminal hairs, which are hairline to below the ear (sideburns) and the buttocks. larger and more heavily pigmented. Hairs grow in Self-scoring can be clinically useful, but correlates only nonsynchronous cycles, and the growth (anagen) phase modestly with scoring by a trained observer (10–12). (which varies with body area) is ;4 months for facial Lower Ferriman–Gallwey scores can be clinically im- hair. Due to the long hair growth cycle, it takes portant. In one study of 633 unselected white and black ;6 months to detect the effects of hormonal therapy and women, ;70% with scores $3 and many with lower scores considered themselves to be hirsute, and most used some form of cosmetic treatment (13). It has also been shown that even minimal degrees of unwanted hair are often as- sociated with hyperandrogenemia when menstrual irregularity is present (14). Hirsutism must be distinguished from hypertrichosis—generalized ex- cessive hair growth that may be hered- itary or result from certain medications (e.g., phenytoin, cyclosporine). Hyper- trichosis is distributed in a generalized, nonsexual pattern (i.e., predominantly on forearms or lower legs) and is not caused by excess androgen (although hyperandrogenemia may aggravate it). Figure 1. Ferriman–Gallwey hirsutism scoring system (4). Each of the nine body areas most sensitive to androgen is assigned a score from 0 (no hair) to 4 (frankly virile). These separate scores are summed to provide a total hormonal hirsutism score. Generalized hirsutism (score Pathogenesis of hirsutism $8) is abnormal in the general US population, whereas locally excessive hair growth (score The growth of sexual hair is entirely ,8) is a common normal variant. The normal score is lower in some Asian populations and dependent on the presence of androgen higher in Mediterranean populations (see text). Reproduced from Hatch et al. (5). 1238 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 ;9 months for these effects to become maximal. Hir- to be of little further diagnostic utility in most, but not sutism results from an interaction between the plasma all, populations (16, 26–30) (Section 5, Androgen Test- androgens and the apparent sensitivity of the hair follicle ing Remarks). Serum total testosterone is similar to and to androgen. The sensitivity of the hair follicle is de- correlates well with serum androgenic bioactivity in young termined in part by the local metabolism of androgens, women with and without PCOS (r = 0.7 to 0.8) (31). Thus, particularly by conversion of testosterone to dihy- hirsutism cannot currently be considered synonymous drotestosterone by the enzyme 5a-reductase and sub- with “clinical evidence of hyperandrogenism” if serum sequent binding of these molecules to the androgen total and free testosterone are normal. However, the receptor. The hirsutism score does not correlate well with possibility exists that previously unsuspected circulating the androgen level (16, 17), apparently because the androgens contribute to idiopathic hirsutism (21, 30, 32) androgen-dependent pilosebaceous follicle response to (Section 5, Androgen Testing Remarks). Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 androgen varies considerably. Further workup of the eumenorrheic patient for mild hirsutism and a normal serum total testosterone is only Etiology of hirsutism clinically indicated if there is other clinical evidence The majority of hirsutism is due to androgen excess to suggest the etiology is a hyperandrogenic endocrine ($80%), and the majority of women with hirsutism (70% disorder. to 80%) have PCOS (18, 19). PCOS is defined by the As noted, among women with mild hirsutism, ap- presence of a combination of two of three symptoms or proximately half have idiopathic hirsutism. Plasma total findings: otherwise unexplained chronic hyperandrogenism, and/or free testosterone levels are elevated in the re- oligoovulation, and ultrasonographic polycystic ovarian mainder of cases of mild hirsutism and in most cases of morphology (20). Gonadotropin-dependent functional more severe hirsutism (16, 26, 27) (Section 5, Androgen ovarian hyperandrogenism is the source of the hyper- Testing Remarks). Most women with a twofold or greater androgenemia in the majority of PCOS cases (18, 20). This elevation of serum androgen levels have some degree of may be accompanied by a related mild adrenocorticotropic hirsutism or an alternative pilosebaceous response, such hormone–dependent functional adrenal hyperandrogenism, as excessive acne vulgaris, seborrhea, or female- or male- and in a minority of instances this form of adrenal hyper- pattern alopecia. androgenism may occur in isolation (20, 21). PCOS is Other causes of androgen overproduction are in- frequently associated with a metabolic syndrome that results frequent (24, 26). NCCAH, the most common of these from insulin resistance and/or central obesity and that re- disorders, is present in 4.2% of hyperandrogenic women quires considerations distinct from those for hirsutism itself. worldwide, although specific ethnic groups are at lower or Obesity may worsen or cause features of PCOS (20, 22, 23). higher risk (Section 5, Androgen Testing Remarks) (33). Many women have hirsutism without hyperandro- Androgen-secreting tumors are present in ;0.2% of genemia. We term this “idiopathic hirsutism” in hyperandrogenic women; over half are malignant (34). eumenorrheic women who have no other clinical evidence Clinicians must consider Cushing syndrome, acromegaly, suggesting PCOS or other hyperandrogenic endocrine hypothyroidism, and (rarely) hyperprolactinemia in the disorder (19), although some may have polycystic ovary differential diagnosis of hirsutism, but patients typically morphology on ultrasound and thus meet a Rotterdam will present with the features specific to these disorders. criterion for “ovulatory PCOS” (24). Idiopathic hirsut- Clinicians must consider topical androgen use by a partner ism constitutes 5% to 20% of hirsute women (19, 24). (35), exogenous androgens or anabolic steroids (36), or Available data suggest that among eumenorrheic women valproic acid when evaluating patients with hirsutism. with mild hirsutism (a Ferriman–Gallwey hirsutism score of 8 to 15 in the United States), approximately half have 1.0 Diagnosis of Hirsutism idiopathic hirsutism (16). However, the percentage of women with idiopathic hirsutism who meet Rotterdam 1.1. We suggest testing for elevated androgen levels criteria for “ovulatory PCOS” remains unclear, as studies in all women with an abnormal hirsutism score using transvaginal ultrasound and/or high-quality andro- (2 |OO). In those cases where serum total tes- gen assay technologies in this population have not yet been tosterone levels are normal, if sexual hair growth is performed. moderate/severe or sexual hair growth is mild but It is unclear whether idiopathic hirsutism is due to al- there is clinical evidence of a hyperandrogenic en- tered androgen mechanism of action within the hair follicle docrine disorder (such as menstrual disturbance or (referred to as cutaneous hyperandrogenism) or to other progression in spite of therapy), we suggest mea- alterations in hair biology (15, 17, 25). The routine assay suring an early morning serum total and free tes- of androgenic steroids other than testosterone has proven tosterone by a reliable specialty assay. (2 |OO) doi: 10.1210/jc.2018-00241 https://academic.oup.com/jcem 1239 1.2. We suggest screening hyperandrogenemic women may be found and upon whether a detected abnor- for NCCAH due to 21-hydroxylase deficiency by mality will determine the approach to treatment. Most measuring early morning 17-hydroxyprogesterone women with local hair growth and regular menses who levels in the follicular phase or on a random day have no evidence to suggest an endocrine cause have a for those with amenorrhea or infrequent menses very low likelihood of excess androgen production. (2 |OO). In hirsute patients with a high risk Conversely, patients with hirsutism or with clinical of congenital adrenal hyperplasia (positive family features suggesting an underlying endocrine disor- history, member of a high-risk ethnic group), we der, including failure to respond to therapy over time suggest this screening even if serum total and free (Fig. 2), are more likely to have excess androgen testosterone are normal. (2 |OO) production (3). A rapid pace of development or pro- 1.3. We suggest against testing for elevated an- gression of hirsutism, progression in spite of ther- Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 drogen levels in eumenorrheic women with apy, or evidence of virilization (such as clitoromegaly, unwanted local hair growth (i.e., in the absence deepening of the voice, or increasing muscularity) of an abnormal hirsutism score) because of points to a greater likelihood of an androgen-secreting the low likelihood of identifying a medical neoplasm. However, some tumors producing only mod- disorder that would change management or erately excessive androgen have indolent presenta- outcome. (2 |OO) tions (3). Because standard assays fail to detect androgenic Evidence drugs, clinicians should be diligent in their effort Hirsutism is a clinical diagnosis. The management of to obtain a history of anabolic or androgenic steroid hirsutism is to a considerable extent independent of the use, particularly among athletes and patients who have etiology. However, hirsutism is a potential indication of endometriosis, sexual dysfunction, or partners who an underlying hyperandrogenic disorder that may require may use testosterone gel. Valproic acid is the only specific treatment and may have distinct implications for anticonvulsant medication that raises plasma testos- fertility, medical risks, and genetic counseling. There is a terone levels. wide variety of approaches specialists use to diagnose the Because of the high frequency of PCOS, clinicians disorder; however, there is uncertainty regarding the cost should check all hirsute women for evidence of anov- effectiveness, acceptability to patients, and the impact on ulation (menstrual irregularity) or more subtle ovarian outcomes of these approaches (3). dysfunction that may present as infertility (37), central The goal in assessing hirsutism is to attempt to de- obesity, abnormal carbohydrate and lipid metabolism, termine the specific etiology and to provide a baseline in acanthosis nigricans, or a family history of type 2 di- case it becomes necessary to reassess the patient. Figure 2 abetes mellitus. Clinicians can make a diagnosis of provides an approach to assessing hyperandrogenemia ovulatory PCOS in eumenorrheic women with hirsut- that depends on both determining the presence and degree ism, polycystic ovary morphology, and normal levels of of hirsutism and assessing whether there is clinical evidence testosterone (39, 40). It is unclear whether pelvic ul- of PCOS, other hyperandrogenic endocrinopathies, viril- trasonography is cost effective in the management of izing disorders, or androgenic medication use. idiopathic hirsutism (i.e., eumenorrheic hirsute women When testing for elevated androgen levels, we suggest with normal testosterone levels and no other clinical first measuring serum total testosterone levels using a re- evidence of PCOS). liable specialty assay (Fig. 2). In those cases where serum While PCOS is the most likely diagnosis in a woman total testosterone levels are normal, if sexual hair growth is with menstrual dysfunction, hirsutism, and an elevated moderate/severe or sexual hair growth is mild but there is testosterone level, clinicians need to exclude conditions clinical evidence of a hyperandrogenic endocrine disorder other than PCOS that are: sufficiently common, associ- (such as menstrual disturbance or progression in spite of ated with adverse natural histories, and treatable (e.g., therapy), we suggest measuring an early morning serum pregnancy, nonclassic congenital adrenal hyperplasia, total and free testosterone by a reliable specialty assay. ovarian or adrenal neoplasm, or other endocrinopathies). Menstrual irregularity, infertility, galactorrhea, central The most common of these is nonclassic congenital ad- obesity, acanthosis nigricans, clitoromegaly, sudden-onset renal hyperplasia due to 21-hydroxylase deficiency (as or rapid-progression hirsutism, or hirsutism progression in noted in Section 1, Etiology and Section 5, Andro- spite of therapy suggests the presence of a hyperandrogenic gen Testing Remarks). This is particularly important endocrine disorder. to detect because of its genetic implications for those The decision to test for androgen excess depends women desiring fertility (33). We suggest screening both on the pretest likelihood that an abnormal value hyperandrogenemic women for nonclassic congenital 1240 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 Figure 2. Evaluation and treatment of hirsutism in premenopausal women. Figure note: Local sexual hair growth (i.e., in the absence of an abnormal hirsutism score) that is not accompanied by clinical evidence of a hyperandrogenic endocrine disorder does not require an endocrine workup before embarking on dermatologic therapy (cosmetic or direct hair removal measures). Elevated androgen levels should be ruled out in women with hirsutism or any degree of sexual hair growth who also have clinical evidence of a hyperandrogenic endocrine disorder. Clinical evidence of menstrual irregularity, infertility, galactorrhea, signs or symptoms of hypothyroidism, Cushing syndrome, acromegaly, central obesity, acanthosis nigricans, clitoromegaly, or sudden-onset or rapid-progression hirsutism suggests the presence of a hyperandrogenic endocrine disorder. PCOS is the most common hyperandrogenic disorder associated with hirsutism. However, androgen-secreting tumors and nonclassic congenital adrenal hyperplasia are other major causes that clinicians should consider. Drugs that cause hirsutism include anabolic or androgenic steroids (a consideration in athletes, users of dietary supplements, patients with sexual dysfunction, or in patients with a partner who uses testosterone gel) and valproic acid (a consideration in patient with neurologic disorders). An accurate and specific assay, such as mass spectrometry, is the best choice for assessing serum total testosterone concentrations. Norms are standardized for early morning, when levels are the highest, and for days 4 to 10 of the menstrual cycle (see Section 5, Androgen Testing Remarks) when ovarian follicle development is the most comparable to that of women with hyperandrogenic anovulation; clinicians should interpret marginal values obtained at other times accordingly. Women with mild hirsutism, a normal total testosterone level, a pelvic ultrasound showing normal ovarian morphology (if performed), and no clinical evidence of other hyperandrogenic endocrine disorders have idiopathic hirsutism, which may be responsive to OC therapy. However, if the serum total testosterone is normal in the presence of moderate or severe hirsutism or if there is clinical evidence of PCOS or other endocrine disorder, clinicians should test serum-free testosterone levels. Assessing free testosterone levels using high-quality testosterone and SHBG or equilibrium dialysis assays with well-defined reference intervals is the single most useful, clinically sensitive marker of androgen excess in women. A simultaneous assay of early-morning 17-hydroxyprogesterone is indicated in subjects at high risk for nonclassic congenital adrenal hyperplasia (see text and Section 5, Androgen Testing Remarks). Progression of hyperandrogenism in the presence of a normal serum-free testosterone is very unusual, and clinicians should thoroughly reevaluate these patients (3). Unless fertility is an issue (37), demonstrating polycystic ovary morphology to diagnose ovulatory PCOS is unlikely to affect management. Adapted from Martin et al. (38). adrenal hyperplasia due to 21-hydroxylase deficiency by A separate Endocrine Society clinical guideline de- measuring early morning 17-hydroxyprogesterone levels scribes the approach to the diagnosis of PCOS in detail in the follicular phase or on a random day for those with (40). Different subspecialists use different strategies for amenorrhea or infrequent menses. In hirsute patients evaluating the patient with hirsutism (24, 41–43). The with a high risk of congenital adrenal hyperplasia (pos- evaluation of hyperandrogenemic women may include itive family history, member of a high-risk ethnic group), the following: pregnancy tests in patients with amenor- we suggest this screening even if serum total and free rhea; measuring dehydroepiandrosterone (DHEA) sul- testosterone are normal. fate (DHEAS) to screen for adrenal hyperandrogenism; doi: 10.1210/jc.2018-00241 https://academic.oup.com/jcem 1241 assessing for Cushing syndrome, thyroid dysfunc- 2.0 Treatment of Hirsutism in tion, acromegaly, and hyperprolactinemia if features of Premenopausal Women these conditions are present (however, all are uncom- 2.1. For most women with patient-important hirsutism mon causes of hirsutism); and pelvic ultrasonogra- despite cosmetic measures, we suggest starting phy (preferably transvaginal) to detect an ovarian with pharmacological therapy (2 |OOO). For neoplasm in women with severe or progressive hyper- women who then desire additional cosmetic androgenism. Of note, some androgen-secreting ovar- benefit, we suggest adding direct hair removal ian tumors are too small to be detected by transvaginal methods. However, for women with mild hir- ultrasound. sutism and no evidence of an endocrine disorder, Further workup to identify the origin of androgen we suggest either approach. (2 |OOO) excess may be clinically indicated because of atypical Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 2.2. For hirsute women with obesity, including those clinical or laboratory findings and may include the fol- with PCOS, we also recommend lifestyle changes. lowing: (1) measuring serum androstenedione (the im- (1 |OO) mediate precursor for testosterone) (28) or other steroid intermediates [that have on occasion provided addi- tional information (Section 5, Androgen Testing Re- Evidence marks)]; (2) assessing the response to cosyntropin of The development of hirsutism is mostly dependent on 17-hydroxyprogesterone, DHEA, 17-hydroxypregnenolone, circulating androgen concentrations and the response of and 11-deoxycortisol, and/or genotyping to exclude the hair follicle to the local androgen milieu. Thus, there rare forms of congenital adrenal hyperplasia; (3) are two main approaches to the management of hirsutism assessing urinary corticoid metabolites by mass spec- that may be used either individually or in combination: trometry to exclude apparent cortisone reductase de- (1) pharmacologic therapies that target androgen pro- ficiency (44); (4) dexamethasone suppression testing to duction and action, and (2) direct hair removal methods suppress androgens arising from a functional adrenal (electrolysis and photoepilation). Most women also use source; (5) adrenal computed tomography, ovarian cosmetic measures (shaving, plucking, waxing) before ultrasound, or more specialized imaging studies (45) if their first medical consultation and continue to use them there is reason to suspect an androgen-secreting tumor; during pharmacotherapy. We suggest pharmacotherapy and (6) assessing the suppressive response to combined as initial therapy for most women with patient-important OC or gonadotropin-releasing hormone (GnRH) ago- hirsutism (adding direct hair removal methods later if nist (46). Lastly, transvaginal ultrasound is also helpful needed); however, some women may choose to start both if ovarian hyperthecosis is suspected; absence of folli- therapies simultaneously. cles and or the polycystic ovarian morphology supports Although experts have often made treatment recom- the diagnosis of hyperthecosis. This approach to eval- mendations based on the severity of hirsutism using uation is similar to that recommended by other groups, Ferriman–Gallwey scores [mild (score 8 to 15) or severe including the following: the American Association of (score.15)], this approach has several limitations: (1) Clinical Endocrinologists (43), the American Society for many clinicians are unfamiliar with calculating Reproductive Medicine (47), the French Endocrine Ferriman–Gallwey scores; (2) most women use cosmetic Society (48), and the Androgen Excess and PCOS So- measures before seeking consultation, making it impos- ciety (6). sible to accurately determine a Ferriman–Gallwey score; and (3) treatment decisions need to be proportionate to Values and preferences the extent that excessive hair affects patient well-being Our suggestion for testing for hyperandrogenemia in (i.e., some women with low scores may be more distressed all women with hirsutism places a relatively high value and desire more aggressive management of their hirsut- on the identification of treatable underlying hyper- ism than other women who may be less bothered, despite androgenic diseases. Our suggestion for not testing having higher hirsutism scores). We use the term patient- for hyperandrogenemia in patients with normal variant important hirsutism to refer to unwanted sexual hair unwanted hair, for whom hormonal treatment is not growth of any degree that causes sufficient distress for contemplated, places a relatively high value on avoiding women to seek additional treatment. the risk of false positives and the resulting increase in Cosmetic measures to manage hirsutism include medical tests and procedures. It places a relatively low methods that remove hair shafts from the skin surface value on the potential benefits of early detection of mild (depilation) and those that extract hairs to above the bulb hyperandrogenemia that will not affect initial man- (epilation). Shaving is a popular depilation method that agement and outcome. removes hair down to just below the surface of the skin. 1242 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 Shaving does not affect the rate or duration of the anagen therapy for treating patient-important hirsutism. phase or diameter of hair. However, it yields a blunt tip (2 |OO) rather than the tapered tip of uncut hair, which gives the 3.2. For most women with hirsutism, we suggest illusion of thicker hair. Chemical depilatory agents are against antiandrogen monotherapy as initial also commonly used to dissolve the hair. Most de- therapy (because of the teratogenic potential of pilatories contain sulfur and have an unpleasant odor. these medications) unless these women use ade- In addition, irritant dermatitis can occur. Epilation quate contraception (2 |OOO). However, for methods, such as plucking or waxing, are relatively safe women who are not sexually active, have un- and inexpensive, but cause some discomfort. These methods dergone permanent sterilization, or who are us- do not cause an increase in hair diameter. Scarring, follic- ing long-acting reversible contraception, we ulitis, and hyperpigmentation (particularly in women of suggest using either OCs or antiandrogens as Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 color) may occur. Although not a method of hair removal, initial therapy (2 |OOO). The choice between bleaching with products containing hydrogen peroxide and these options depends on patient preferences sulfates is a method for masking the presence of undesired regarding efficacy, side effects, and cost. hair, particularly facial hair. Side effects include irritation, 3.3. For most women, we do not suggest one OC over pruritus, and possible skin discoloration. another as initial therapy, as all OCs appear to be In addition to cosmetic and/or pharmacologic therapy, equally effective for hirsutism, and the risk of side lifestyle changes for obese women with PCOS may im- effects is low. (2 |OO) prove their hirsutism. In a meta-analysis of four studies 3.4. For women with hirsutism at higher risk for VTE that included 132 subjects, lifestyle changes (diet, exer- (e.g., those who are obese or over age 39 years), cise, behavioral, or combination therapy) resulted in we suggest initial therapy with an OC containing weight loss, a decrease in serum testosterone and fasting the lowest effective dose of ethinyl estradiol insulin concentrations, and a small improvement in (EE) (usually 20 mcg) and a low-risk progestin Ferriman–Gallwey scores when compared with minimal (Table 2). (2 |OOO) or no treatment—mean difference, 21.19 [95% confidence 3.5. If patient-important hirsutism remains despite interval (CI) (22.35 to 20.03)] (49). However, lifestyle 6 months of monotherapy with an OC, we suggest changes should not be considered a primary therapy for adding an antiandrogen. (2 |OO) hirsutism, as their impact is not clinically significant, par- 3.6. We do not suggest one antiandrogen over an- ticularly when compared with OCs (50). Our approach is other (2 |OO). However, we recommend consistent with the Endocrine Society clinical guidelines on against the use of flutamide because of its po- the diagnosis and treatment of PCOS (40). tential hepatotoxicity. (1 |OO) 3.7. For all pharmacologic therapies for hirsutism, we suggest a trial of at least 6 months before making 3.0 Pharmacological Treatments changes in dose, switching to a new medication, or adding medication. (2 |OOO) Initial therapies 3.8. In patients with severe hirsutism causing emo- 3.1. For the majority of women with hirsutism who tional distress and/or in those women who have are not seeking fertility, we suggest OCs as initial used OCs in the past and have not experienced Table 2. OCs and Associated VTE Risks Progestin Progestin Relative Progestin Relative Progestin Absolute Generation Androgenicity VTE Riska,b VTE Riskb,c Progestin/Dose EE Dose (mcg) 1 Medium 2.6 7 Norethindrone 0.5–1.0 mg 20, 35 2 High 2.4 6 Levonorgestrel 0.15 mg 20, 30 2–3 Low 2.5 6 Norgestimate 0.25 mg 35 3 Low 3.6 11 Gestodene 0.075 mg 20, 30 3 Low 4.3 14 Desogestrel 0.15 mg 20, 30 4 Antiandrogen 4.1 13 DSP 3 mg 20, 30 — Antiandrogen 4.3 14 CPA 2 mgd 35 a Relative risk compared with no OC use. b Vinogradova et al. (73); Stegeman et al. (57). c Extra cases VTE per 10,000 women treated with OCs per year. d OCs containing CPA are not available in the United States. doi: 10.1210/jc.2018-00241 https://academic.oup.com/jcem 1243 sufficient improvement, we suggest initiating com- There may also be a small direct effect in inhibiting 5a- bination therapy with an OC and antiandrogen reductase activity in the pilosebaceous unit. (2 |OO). However, we suggest against combi- Combination OCs carry about a threefold increased nation therapy as a standard first-line approach. risk of VTE in first-time users. VTE risk is significantly (2 |OO) but weakly related to estrogen dose and may wane with duration of estrogen use. The use of OCs containing some of the recent-generation low-androgenicity progestins Evidence (desogestrel, gestodene), and androgen receptor antag- Combined oral estrogen–progestin contraceptives onists (CPA, DSP) may confer a 50% to 100% increased Most combined estrogen–progestin OCs contain the risk of VTE compared with OCs containing the second- potent, synthetic estrogen EE in combination with a generation progestin levonorgestrel, according to reviews Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 progestin. Of note, in this guideline, OCs refers only to of large-scale comparative analyses (58, 59). However, combined oral estrogen–progestin contraceptives con- the DSP risk was not found in a prospective post- taining EE and not to newer OCs that contain 17- marketing study of first-time contraceptive users (60). Of b-estradiol or estradiol valerate combined with highly note, the absolute risk is low and far less than that seen potent progestins, as the doses of estrogen in these pills during pregnancy (61). There have been concerns that the are unlikely to suppress ovarian androgens. The guideline presence of PCOS may represent an additional in- also does not refer to oral progestin-only contraceptives, dependent risk factor for VTE, but available data are which are ineffective for hirsutism. inconclusive (62, 63). There have also been concerns Most progestins are derived from 19-nortestoster- about an excess risk of VTE with OCs containing CPA, one and exhibit varying degrees of androgenicity (31, but the Pharmacovigilance Risk Assessment Committee 51). Examples of progestins with low androgenicity of the European Medicines Agency concluded in 2013 include norgestimate, desogestrel, and gestodene; those that the benefits of the drug outweighed the risks (64). with medium androgenicity include norethindrone; Increased age and obesity are additional factors associ- and those with relatively high androgenicity include ated with an increased risk of VTE. The risk in women norgestrel and levonorgestrel. CPA and DSP are over age 39 years taking OCs is approximately fourfold structurally unrelated to testosterone and function as higher than in younger women (100 vs 25 per 100,000 weak androgen receptor antagonists. Several OCs women years) (65). The risk in obese women taking OCs contain DSP, a progestin structurally related to spi- has been estimated to be 2- to 10-fold higher than ronolactone that exhibits weak antiandrogenic activ- nonobese women taking OCs (66, 67). However, the ity. In bioassays, 3 mg DSP (the dose used in OCs) was benefits outweigh the risks based on obesity alone (68) equivalent to only 9 to 10 mg spironolactone. For when using OCs for contraception. The risk-benefit ratio comparison, 100 to 200 mg spironolactone is the for women with PCOS taking OCs simply for cycle therapeutic dose for hirsutism. Also, 2 mg CPA (the control and/or androgen suppression is unclear. dose used in OCs) was equivalent to ;50 mg spi- ronolactone (52, 53). A 12-month trial comparing OCs Updated systematic review and meta-analysis containing either 3 mg DSP or 2 mg CPA showed The results of the network analysis were consistent similar reductions in hirsutism scores, suggesting that with our previous meta-analysis. Our 2008 review the efficacy is substantially related to ovarian sup- identified only one placebo-controlled, randomized trial pression (54). Although DSP is a very weak anti- (69) and a second trial that compared OCs to no therapy androgen, it is more potent than spironolactone in in women with hirsutism (70). The updated review iden- antimineralocorticoid equivalency. tified no additional trials. A combined analysis of OC therapy reduces hyperandrogenism via a number these trials associated OC therapy with a greater reduc- of mechanisms, including the following: suppression tion in hirsutism scores, with a pooled weighted mean of luteinizing hormone secretion (and therefore ovar- difference of 27.20 [95% CI (211.96 to 22.52)]. The ian androgen secretion) (55), stimulation of hepatic pro- extent to which this average reduction in hirsutism scores duction of sex hormone–binding globulin (SHBG) (thereby reflects a reduction in hirsutism-associated distress increasing androgen binding in serum and reducing serum- remains unclear. free androgen concentrations), and a slight reduction in The systematic review also compared OCs containing both adrenal androgen secretion and binding of androgens antiandrogenic progestins (CPA and DSP) vs all other to their receptor. Consequently, there is a reduction of OCs and the relatively androgenic progestin levo- testosterone production. Androgenic progestins also in- norgestrel vs all other OCs. Only four trials presented crease the metabolic clearance of testosterone (56, 57). data sufficient for meta-analysis. OCs containing 1244 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 levonorgestrel had a similar effect on hirsutism scores antiandrogens were pooled together as a class, and results compared with all other OCs. OCs containing anti- were expressed in Ferriman–Gallwey units, antiandrogens androgenic progestins (one trial using CPA and one were significantly more effective than placebo, with a pooled trial using DSP) were associated with slightly lower weighted mean difference of 27.02 [95% CI (211.51 Ferriman–Gallwey scores than other OCs, with a to 22.52)]. There was no statistically significant difference weighted mean difference of 22.86 [95% CI (24.96 to among the three antiandrogens. 20.76)], a difference that is probably not clinically im- Available antiandrogens and their dosing regimens are portant (71). shown in Table 3. Spironolactone, an aldosterone an- tagonist, exhibits dose-dependent competitive inhibition Which OCs should be used for hirsutism? of the androgen receptor as well as inhibition of 5a- For most women, we do not suggest one particular OC reductase activity (77). Although there are no rigorous Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 formulation over another for treating hirsutism. This dose-response trials to date, spironolactone’s effects are recommendation is consistent with the Endocrine Society known to be dose dependent (77). The drug is generally clinical guideline on the diagnosis and treatment of PCOS well tolerated, but should not be used if there is renal (40). There were no clinically important advantages of impairment. Spironolactone may have a dose-dependent OCs containing the antiandrogen progestins CPA or DSP association with menstrual irregularity unless the pa- in our meta-analysis. Although we were concerned that tient uses an OC concomitantly. Spironolactone use may OCs containing levonorgestrel (the most androgenic rarely result in hyperkalemia, and it may cause increased progestin) would be less effective for hirsutism, this was diuresis and occasionally postural hypotension and not observed in our meta-analysis. Whereas a potential dizziness early in treatment. OCs containing DSP have a benefit of OCs containing levonorgestrel is their lower mild mineralocorticoid effect and should not be used VTE risk, an important concern is the adverse effects of with a potassium-sparing diuretic. As with all anti- levonorgestrel on metabolic biomarkers (72). Although androgens, if a patient inadvertently uses spironolactone there are no data to suggest that the metabolic effects are during early pregnancy, there is a danger that a male associated with adverse clinical outcomes, we tend to fetus could be feminized (78) because of the exquisite avoid this progestin in women with PCOS, a population sensitivity of the fetal genitalia to exposure to maternal with metabolic concerns at baseline. As noted previously, synthetic sex hormone ingestion (79), although the for women with hirsutism at higher risk for VTE (e.g., absolute risk of this is not known. those who are obese or over age 39 years), we suggest Clinicians worldwide use CPA to treat hirsutism and initial therapy with an OC containing the lowest effective acne, but it is not available in the United States. CPA is a dose of EE (usually 20 mcg) and a low-risk progestin progestogenic compound with antiandrogen activity by (Table 2). Our approach is similar to that of the consensus virtue of its effects in inhibiting the androgen receptor and statement from the Androgen Excess and PCOS Society, to a lesser degree in inhibiting 5a-reductase activity (80). which also suggests using a low-dose OC product to It also suppresses serum gonadotropin and androgen minimize VTE risk (6). levels. In one systematic review, the OC CPA (2 mg) with Ovarian androgen suppression may be similar with 35 mcg EE was similar to antiandrogen therapy and more OCs containing different doses of EE. In a meta-analysis effective than placebo (81). of 42 studies, the suppression of serum total and free Finasteride inhibits type 2 5a-reductase activity. testosterone concentrations was similar with OCs con- Because enhanced 5a-reductase activity in hirsutism taining 20 vs 30/35 mcg EE (74). Limited data suggest that OCs containing DSP with 20 or 30 mcg EE have a similar effect on Ferriman–Gallwey scores (75). Trans- Table 3. Antiandrogens Used for the Treatment of Hirsutism dermal contraceptive patch and OCs suppressed serum androgens to a similar degree in one study, but the Antiandrogens Dosing outcome of hirsutism was not addressed (76). CPAa 50–100 mg/d on menstrual cycle days 5–15, with EE 20–35 mg on days 5–25 Antiandrogens Spironolactone 100–200 mg/d [given in divided doses (twice daily)] Our systematic review identified seven trials of anti- Finasteride 2.5–5 mg/d androgens, as follows: three of finasteride, two of flutamide, Flutamideb 250–500 mg/d (high dose) and two of spironolactone. In analyses of individual anti- 62.5 to #250 mg/d (low dose) androgens compared with placebo, spironolactone 100 mg/d, a Not available in the United States; also prescribed as an OC (2 mg CPA + finasteride 2.5 to 5 mg/d, and flutamide 500 mg/d, each 35 mcg EE). showed a significant reduction in hirsutism scores. When all b Flutamide not recommended because of hepatotoxicity. doi: 10.1210/jc.2018-00241 https://academic.oup.com/jcem 1245 probably involves both type 1 and 2 5a-reductase en- Addition of antiandrogens to OCs zymes, only a partial inhibitory effect may be expected If hirsutism has not improved despite 6 or more with finasteride. One review of available trials reported months of monotherapy with an OC, we suggest adding that finasteride reduced hirsutism scores by 30% to 60% an antiandrogen. Our 2008 updated systematic reviews and reduced hair shaft diameters as well (82). This effect identified five RCTs of antiandrogens combined with was found to be similar to that with the use of other OCs vs OCs alone. The addition of antiandrogen therapy antiandrogens with no major adverse effects. Although to OCs was slightly more effective for hirsutism than OC 5 mg finasteride is the most commonly used dose, some therapy alone (five trials) and was associated with in- data suggest that 7.5 mg is more effective (83), and that cremental reduction of hirsutism scores—weighted mean doses of 2.5 and 5 mg appear to be equally effective (84). difference, 21.73 [95% CI (23.32 to 20.13)]. Our systematic reviews also demonstrated a significant Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 reduction in hirsutism scores with finasteride compared OCs vs antiandrogens with placebo (85). Dutasteride has been approved for the In the only RCT comparing an OC to an anti- treatment of men with prostate cancer, and it inhibits both androgen (finasteride), the OC contained a low-dose type 1 and 2 isoenzymes. Although this would seem- antiandrogen (2 mg CPA) (99). After 9 months of ingly be an attractive option for the treatment of hir- treatment, there was no significant difference in hir- sutism, there are no clinical data to support its use at sutism score between the finasteride group and the this time. group receiving this OC. Flutamide is a “pure” antiandrogen with a dose- response inhibition of the androgen receptor (86). Glucocorticoid therapy Whereas the most frequently used dose in RCTs is 500 Clinicians administer glucocorticoids long-term to mg/d, some experts have suggested equal efficacy with suppress adrenal androgens in women with classic con- 250 and 500 mg/d (87). Retrospective studies, trials of genital adrenal hyperplasia due to 21-hydroxylase de- combination therapy (low-dose flutamide with other ficiency. In these patients, glucocorticoids help prevent or drugs) (88, 89), and nonrandomized studies of low-dose manage hirsutism, and they are effective for maintaining flutamide (as low as 62.5 mg) (90) suggest that flutamide normal ovulatory cycles. In women with the nonclassic doses of 62.5 to 250 mg may be effective for hirsutism form of 21-hydroxylase deficiency, glucocorticoids are (91), but there is no evidence from RCTs of low-dose effective for ovulation induction, but their role in the flutamide monotherapy vs placebo to support this. management of hirsutism is less clear. The major concern with flutamide is its propensity for In patients with pure adrenal hyperandrogenism, even hepatic toxicity. This is not trivial, as numerous studies in those who are very sensitive to glucocorticoids, sup- have associated flutamide with liver failure and even death pressing adrenal androgens results in only minor im- (92–94). Although some studies have reported that low provements in hirsutism, although these patients can doses of flutamide are not hepatotoxic (88, 95, 96), others achieve prolonged remission after therapy withdrawal have raised important concerns. A 10-year surveillance (100, 101). study of 203 women receiving flutamide at doses of 62.5 or 125 mg identified 22 (11%) who experienced elevated Women with NCCAH serum concentrations of alanine aminotransferase and/or Our approach to treating hirsutism in women with aspartate aminotransferase (97). In a retrospective study of NCCAH is the same as for women with PCOS. We suggest 414 women with hirsutism receiving low-dose flutamide starting with an OC and adding an antiandrogen after alone or with OCs, 6% stopped therapy due to elevated 6 months if necessary. Clinicians can administer an anti- transaminases (all were taking 125 to 250 mg and all androgen as initial therapy if the woman is not pursuing occurred in the first year of therapy) (91). Lastly, one pregnancy and has a reliable form of contraception. We center reported a series of seven women who developed only suggest glucocorticoids for the management of hepatoxicity while taking flutamide (150 to 250 mg/d) for hirsutism in women who have a suboptimal response to acne or hirsutism; five required urgent liver trans- OCs and/or antiandrogens, or cannot tolerate them. For plantation and four of five survived (98). hirsutism, we use prednisone 4 to 6 mg daily or dexa- In our 2008 guideline, we suggested against standard methasone 0.25 mg/d. Clinicians should counsel women dose flutamide (.250 mg). Based upon emerging evi- that are considering pregnancy about the teratogenic dence of hepatotoxicity, unproven efficacy for hirsutism, risks of antiandrogens. For ovulation induction, we and the availability of alternative antiandrogens, we suggest glucocorticoid therapy. We typically start with also recommend against the use of low-dose flutamide prednisone 5 mg daily (102). If ovulation has not oc- (#250 mg). curred, the dose can be increased to 7.5 mg. Clomiphene 1246 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 citrate is then added if ovulation does not occur Other drug therapies with prednisone alone. We do not suggest dexametha- 3.9. We suggest against using insulin-lowering drugs sone because it is not inactivated by placental 11- for the sole indication of treating hirsutism. b-hydroxysteroid dehydrogenase type 2 (i.e., fetal (2 |OO) exposure occurs). 3.10. We suggest against using GnRH agonists except In women with adrenal hyperandrogenism, although in women with severe forms of hyperandro- glucocorticoids may improve hirsutism, both OCs and genemia (such as ovarian hyperthecosis) who antiandrogens may be more effective. In a study of have a suboptimal response to OCs and anti- women with hirsutism of adrenal origin or enzyme de- androgens. (2 |OOO) ficiency randomized to receive an OC (CPA plus EE) or 3.11. We suggest against the use of topical anti- dexamethasone (103), serum DHEA and DHEAS con- Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 androgen therapy for hirsutism. (2 |OOO) centrations decreased in the dexamethasone group, but not the OC group. However, more women in the OC group experienced a significant reduction in hirsutism (10 Evidence of 15 patients; 66%) than in the dexamethasone group (4 Insulin-lowering drugs—updated systematic review of 13 patients; 31%). and meta-analysis In a trial of women with NCCAH receiving CPA or Reducing insulin levels pharmacologically attenuates hydrocortisone (104), CPA-treated patients experi- hyperandrogenemia. Metformin, an insulin-lowering enced a significantly greater decrease in hirsutism drug, has been used for a number of indications in scores (54%) after 1 year than hydrocortisone-treated women with PCOS, including hirsutism. In our 2008 women (26%); in contrast, androgen levels normalized meta-analysis of eight RCTs, metformin was no more only in the hydrocortisone-treated subgroup, sug- effective than placebo for hirsutism treatment (105), gesting that half of the cutaneous expression of and we suggested against its use (38). Similar results hyperandrogenism is dependent on the peripheral re- were seen in our updated systematic review of nine ceptivity to androgens. trials; in a pooled analysis, metformin was no more effective than placebo for lowering hirsutism scores. Adverse effects associated with Other insulin sensitizers, troglitazone and rosiglita- glucocorticoid therapy zone, had no significant effect on hirsutism. Our results Slight overdosing can occur even at recommended are consistent with other meta-analyses of metformin doses and is independent of daily or alternate-day ad- therapy for hirsutism (106). ministration. Slight overdosing may be associated with side effects, such as adrenal atrophy, increased blood GnRH agonists pressure, weight gain, Cushingoid striae (particularly with Uncontrolled trials of GnRH agonist therapy in dexamethasone), and decreased bone mineral density. women with ovarian hyperandrogenism have reported DHEAS levels indicate the degree of adrenal suppression; significant reductions in luteinizing hormone, ovarian the target level is ;70 mcg/dL (25). androgens, and Ferriman–Gallwey scores (107–110). When compared with OC therapy, GnRH agonist Values and preferences therapy alone seems to have similar benefit for reducing Our recommendation not to use flutamide for the hirsutism scores (111–113). GnRH agonist with low- routine management of hirsutism places a high value on dose estrogen–progestin add-back was more effective avoiding potential hepatotoxicity and medication costs in for hirsutism than an OC in two trials—one by pho- women with a relatively benign disorder and a relatively tographic hair density (114) and one by Ferriman– lower value on foregoing a potentially useful intervention. Gallwey scores (115). Because GnRH agonists alone The suggestion not to offer glucocorticoid therapy as first- result in severe hypoestrogenism and eventual bone loss line therapy to hirsute women with NCCAH places a (116), clinicians prescribe low doses of estrogen or es- relatively higher value on avoiding the potential for ad- trogen plus progestin (in women with a uterus) as add- verse effects of glucocorticoids and a relatively lower value back therapy (117, 118). on the potential benefits of suppressing endogenous an- Although GnRH agonist therapy is more effective than drogens and inducing a more prolonged remission of placebo or no therapy for hirsutism, it appears to have no hirsutism and hyperandrogenism after therapy with- advantages when compared with other available agents drawal (100, 101). Our approach does recognize the (such as OCs and antiandrogens). In addition, GnRH importance of glucocorticoid therapy for ovulation in- agonist therapy is expensive, requires injections, and, duction in NCCAH. unless clinicians add some form of estrogen, results in doi: 10.1210/jc.2018-00241 https://academic.oup.com/jcem 1247 severe estrogen deficiency with menopausal symptoms Evidence (such as hot flashes and bone loss). We therefore suggest Temporary methods of hair removal against using GnRH agonists except in women with (cosmetic methods) severe forms of hyperandrogenemia (such as ovarian Depilation is the removal of the hair shaft from the skin hyperthecosis) who have a suboptimal response to OCs surface, for example by shaving. Depilation in humans has and antiandrogens. no known biological effect on the hair follicle, producing Topical antiandrogens no change in hair growth, hair diameter, or hair color. Creams with antiandrogens appear to have lim- Shaving yields a sharply cut hair tip, which upon regrowth ited efficacy, with one study of a cream containing feels coarse and gives the illusion of thicker hair compared canrenone (the active metabolite of spironolactone) with a naturally tapered hair tip. Plucking, waxing, or Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 reporting both benefit and no benefit (119). Similarly, mechanical devices that extract hairs are relatively safe trials of finasteride have yielded inconsistent results, and inexpensive, but cause some discomfort. Scarring, with local applications of preparations with 0.25% folliculitis, and (particularly in women of color) hyper- showing benefit (120) and 0.5% showing no benefit pigmentation may occur. (121). We therefore suggest against their use. Note that Chemical depilatory agents dissolve the hair. Most are eflornithine (which has been approved as a topical thioglycolates, which disrupt disulfide bonds in the hair. treatment) is not an antiandrogen (see Topical Treat- Side effects include emission of a sulfurous odor and ment below). irritant dermatitis (especially on the face), which may be followed by hyperpigmentation. Values and preferences Although not a method of hair removal, bleaching Our suggestion against the use of GnRH agonists with products containing hydrogen peroxide and sulfates for the routine management of hirsutism places a high is a method for masking the appearance of pigmented value on avoiding an expensive, inconvenient therapy hair. Side effects include irritation, pruritus, and possible that requires the addition of estrogen (with or with- skin discoloration. out progestin) to avoid side effects and bone loss and a relatively lower value on foregoing a potentially useful Permanent methods of hair reduction: electrolysis intervention. and photoepilation The Food and Drug Administration (FDA) has 4.0 Direct Hair Removal Methods approved a large number of photoepilation devices [laser 4.1. For women who choose hair removal therapy, we and intense pulsed light (IPL)] for permanent hair re- suggest photoepilation for those whose un- duction. They define permanent hair reduction as wanted hair is auburn, brown, or black, and we attaining at least a 30% reduction of terminal hairs and suggest electrolysis for those with white or blonde sustaining this reduction for a period longer than the hair. (2 |OO) complete growth cycle of hair follicles (4 to 12 months, 4.2. For women of color who choose photoepilation depending on body site). Photoepilation is a method treatment, we suggest using a long-wavelength, capable of rapidly treating large areas; it requires the long pulse-duration light source such as Nd: presence of pigmented, terminal hair. Electrolysis is YAG or diode laser delivered with appropri- generally limited to small treatment areas, and it does not ate skin cooling (2 |OOO). Clinicians should depend on hair pigmentation. warn Mediterranean and Middle Eastern women with facial hirsutism about the increased risk Electrolysis of developing PH with photoepilation therapy. Despite being available as a hair reduction method We often suggest topical treatment or electrol- for well over a century, prospective clinical trials have ysis over photoepilation with these patients. rarely studied electrolysis. Electrical current is passed (2 |OOO) through a fine wire electrode, which is manually inserted 4.3. For women who desire more rapid response to sequentially into individual hair follicles. The galvanic photoepilation, we suggest adding eflornithine electrolysis technique uses direct current, causing elec- topical cream during treatment. (2 |OO) trochemical reactions that locally release toxic products 4.4. For women with known hyperandrogenemia within the hair follicle. The thermolysis technique uses a who choose hair removal therapy, we suggest higher level of alternating current to produce heat in the pharmacologic therapy to minimize hair regrowth. hair follicle immediately surrounding the wire electrode. (2 |OO) Some claim a combination of these (“The Blend”) is more 1248 Martin et al Guidelines on Hirsutism J Clin Endocrinol Metab, April 2018, 103(4):1233–1257 effective (122). Electrolysis is generally regarded as ef- 60 times faster (30 seconds vs 30 minutes). Six months fective for permanent hair reduction. In one small com- following the initial treatment, there was a 74% re- parative study, electrolysis was more effective than duction in terminal hair count after laser and 35% after plucking for permanent reduction of axillary hair (123). electrolysis. The thermolysis and blend techniques are painful; topical anesthetic applications prior to treatment can reduce this Efficacy of photoepilation discomfort (124). All FDA-approved photoepilation devices have met the FDA hair removal criteria after a single treatment in at Photoepilation least one prospective study. This includes most of the Permanent hair reduction by photoepilation appeared commercially available photoepilation lasers and many ,20 years ago (125) and is now the third most prevalent IPL devices. Complete or nearly complete alopecia occurs Downloaded from https://academic.oup.com/jcem/article/103/4/1233/4924418 by guest on 18 November 2024 nonsurgical aesthetic procedure in the United States after for 4 to 6 weeks after each photoepilation treatment, botulinum toxin and hyaluronic acid injections, with followed by gradual regrowth of terminal hair that is ;890,000 procedures performed during 2012 (126). typically reduced in number compared with baseline. Photoepilation uses pulses of light absorbed by mel- A meta-analysis of 24 prospective trials published anin in the hair shaft and follicle to cause selective between 1998 and 2003 found that hair reduction at least photothermolysis of pigmented terminal hair follicles 6 months after the last treatment averaged 57.5%, (127): it selectively injures pigmented tissues based upon 54.0%, 52.8%, and 42.3% for diode, alexandrite, ruby, wavelength, pulse duration, and fluence (energy applied and Nd:YAG lasers, respectively. Although diode had the per area of skin surface). Photoepilation sources include highest percentage reduction rate, the differences among four types of laser (ruby, alexandrite, diode, and Nd: all four lasers were not statistically significant (131). An YAG) and various IPL sources emitting specific wave- earlier systematic review of 11 RCTs involving 444 lengths between 500 and 1200 nm that the melanin people reported a similar 50% reduction in hair growth absorbs. Pulse durations of milliseconds permit heat to for up to 6 months with alexandrite and diode lasers diffuse from the pigmented hair shaft into the sur- (132). The review did not perform a meta-analysis for rounding epithelium of a terminal hair follicle (128). IPL, ruby, or Nd:YAG lasers because of heterogeneous Clinicians can adjust fluence and pulse duration interventions and outcome measures. Prospective, con- according to a particular patient’s hair and skin type. trolled studies with objective quantitative endpoints that Effective and safe treatment requires producing irre- compared lasers or IPL devices for photoepilation gen- versible thermal damage to hair follicles, and not to the erally support these conclusions. Efficacy increases with surrounding skin. Some photoepilation devices are able the number of treatments (133), but rarely achieves to rapidly treat very large areas (e.g., the lower face, neck, 100% hair removal (134). In a retrospective report chest, and both axillae) within 20 minutes. The FDA has on.2000 consecutive patients treate