Excerpt from
Ryan: Kistner's Gynecology & Women's Health, 7th ed.,
Copyright © 1999 Mosby, Inc.

Chapter 15 - Polycystic Ovary Syndrome and Hyperandrogenism
ANN E. TAYLOR ANDREA E. DUNAIF

KEY ISSUES
1. Most women with hyperandrogenism have either polycystic ovarian syndrome (PCOS), defined as hyperandrogenism and chronic anovulation when other disorders are excluded, or idiopathic hirsutism, defined as hyperandrogenism and regular ovulatory menstrual cycles.

2. Significant morbid conditions are associated with PCOS, including endometrial hyperplasia, insulin resistance, diabetes mellitus type 2, and probably increased risk for cardiovascular disease.

3. Diagnostic evaluation should be focused on excluding ovarian and adrenal androgen-secreting tumors and screening for associated health risks including glucose intolerance and dyslipidemia.

4. Treatments that improve insulin sensitivity and result in lower serum insulin levels are potentially new therapies for hyperandrogenism.

The most common causes of hyperandrogenism are disorders of unknown cause: polycystic ovary syndrome and idiopathic hirsutism. These disorders affect approximately 5% of premenopausal women (Knochenhauer et al, 1998). In this chapter, we discuss the differential diagnosis of hyperandrogenism, the clinical correlates of the hyperandrogenic state, a diagnostic approach to masculinized women, and therapeutic options.

PATHOPHYSIOLOGY OF HYPERANDROGENISM
Masculinization is defined as clinical evidence of excess androgen action in women; it includes acne and alopecia in addition to hirsutism (Fig. 15-1) . True virilization is defined as temporal balding, deepening of the voice, increased muscle bulk, and clitoromegaly, and is the clinical consequence of severe hyperandrogenism. Masculinization results from the effects of either increased androgen production or enhanced androgen use by target tissues. The spectrum of hyperandrogenic conditions ranges from a subtle increase in terminal body hair to true virilization. Because of genetic differences in target tissue number and sensitivity to androgens, there may be no clinical sequelae of hyperandrogenism.

Androgens are steroid hormones synthesized and secreted directly by the adrenal glands and gonads (Longcope, 1986). Potent androgens are also converted from precursors in peripheral tissues, including skin and fat cells. Androgens are defined specifically by their ability in bioassay systems to induce growth and secretion by the prostate and seminal vesicles and to bind tightly to prostatic cytosolic androgen receptors. Like insulin and growth hormone, androgens are anabolic because they cause nitrogen retention.

In humans, testosterone is the biologically important extracellular androgen. It is metabolized into biologically active products and excretory products. The biologically active metabolites include the even more potent androgen, dihydrotestosterone (DHT), formed intracellularly through the 5alpha-reduction of testosterone, and the estrogen, estradiol (E2), formed through the aromatization of testosterone.

The ovaries and the adrenals secrete the androgen prehormones androstenedione (A) and ehydroepiandrosterone (DHEA) under the control of luteinizing hormone (LH) and adrenocorticotropic hormone (ACTH), respectively. DHEA sulfate (DHEAS) is secreted almost exclusively by the adrenals. Testosterone (T) is secreted by the ovaries and the adrenals and is produced by the peripheral conversion of A and DHEA. Under normal circumstances, dihydrotestosterone (DHT) is formed entirely from the peripheral conversion of A and T. T and DHT circulate tightly bound to the high-affinity binding proteins and sex hormone-binding globulin (SHBG) and are loosely associated with albumin. Free T and that nonspecifically bound to albumin are biologically available to enter target tissues. In most androgen target tissues, T is converted to its more potent metabolite, DHT, by the 5alpha-reductase enzyme system. Both T and DHT bind to the androgen receptor, initiating androgen action. DHT is metabolized in the periphery to alpha;-androstanediol, which is then converted to 3alpha-androstanediol glucuronide and excreted. (Used with permission of A. Dunaif.)

Other androgens, such as androstenedione, dehydroepiandrosterone (DHEA), and DHEA sulfate (DHEAS), are androgenic by their conversion to testosterone or DHT and, thus, are androgenic prehormones. Androgen production results from glandular secretion and peripheral conversion of these prehormones (mostly androstenedione and DHEA). In normal women, approximately 50% of circulating testosterone is secreted equally by the ovary and the adrenal gland (Longcope, 1986). Androstenedione is also equally secreted by them (Longcope, 1986). In contrast, 50% of DHEA is secreted by the adrenal gland, 20% is secreted by the ovary, and 30% is derived from the peripheral conversion of DHEAS, which is almost completely produced by the adrenal gland (Abraham, 1976). Under normal circumstances, serum DHT is formed entirely from the peripheral conversion of androstenedione (85%) and testosterone (15%) (Ito and Horton, 1971). Thus, androgen production can be increased in abnormal states by the increased glandular secretion of the potent androgen testosterone or by the increased glandular secretion of androgen prehormones such as androstenedione, DHEA, and DHEAS.

In the ovary androgens are precursors of estrogen production, and their production is under the control of luteinizing hormone (LH). Thus, feedback control of ovarian androgens is mediated by the effects of androgen metabolites ( i.e., estrogens) on the hypothalamus and the pituitary. In the adrenal cortex, androgen production is under the control of adrenocorticotropic hormone (ACTH). The only known feedback control of adrenal androgen secretion is mediated by cortisol feedback on the hypotholamic-pituitary-adrenal axis.

Biologic availability of androgens is related to the concentration of the high-affinity, androgen-binding protein produced by the liver known as sex hormone-binding globulin (SHBG). Only free androgens and those nonspecifically bound to circulating albumin are able to enter tissues and produce biologic effects (Pardridge, 1981). SHBG has the greatest affinity for DHT, then for testosterone, and then for E2. There is minimal, if any, binding of androstenedione or DHEA to SHBG. Certain synthetic androgens and progestins ( e.g., levonorgestrel, which is contained in several oral contraceptive pills) have high affinities for SHBG and may displace endogenous steroids. Hepatic synthesis of SHBG is decreased by androgens and insulin and increased by estrogens and thyroid hormones. Consequently, testosterone and DHT availability can be enhanced by decreasing SHBG levels or by administering compounds that compete for SHBG binding sites.

The major component of the metabolic clearance rate (MCR) of androgens reflects hepatic metabolism and the renal excretion of metabolites. A smaller component of the MCR results from androgens that leave the circulation to act on target tissues. This extrasplanchnic metabolism is also known as peripheral utilization. Androgen utilization by peripheral tissues is the final determinant of target-tissue androgen action. The pilosebaceous unit consists of a pilary component and a sebaceous component and can differentiate into a terminal hair follicle or a sebaceous follicle. In the pilosebaceous unit, the chief enzyme regulating androgen utilization is 5alpha-reductase (Kutten et al, 1977; obo et al, 1983). In the pilosebaceous unit, the major 5alpha-reduced androgen DHT is more potent than testosterone. In ovarian granulosa cells, adipose tissue, muscle, and hypothalamus, the aromatase enzyme system converts testosterone to E2 and androstenedione to estrone (E1).

Dihydrotestosterone binds to specific cytosolic androgen receptors that are then translocated to the nucleus. The interaction of DHT with the nuclear androgen receptor (a member of the c-erbA oncogenic hormone receptor superfamily) initiates androgen-specific molecular and cellular responses. Testosterone may also bind directly to the androgen receptor in certain target tissues, such as skeletal muscle, without requiring 5alpha-reductase conversion to DHT. Individual variation in the masculinizing effects of androgens cannot be explained on the basis of androgen receptor levels. Although the androgen receptor is required for masculinization and its genetic absence results in a female phenotype in genetic males (testicular feminization syndrome), no differences in androgen receptor levels have been detected between hirsute women and normal women or normal men (Mowszowicz et al, 1983). Recent data suggest that in men the functional activity of the androgen receptor can be modulated by the number of polyglutamine tri-nucleotide repeats in the first exon. Hence, longer repeats are associated with decreased sperm counts (Tut et al, 1997), and shorter repeats are associated with increased risk for androgendependent prostate cancer (Irvine, 1994). Whether this androgen receptor variability modulates androgen action at the low androgen levels seen in women remains to be determined. Thus, the clinical diversity of response to androgens in women could result from variability in androgen production, androgen bioavailability, peripheral androgen utilization, or, potentially, androgen receptor sensitivity.

SIGNS AND SYMPTOMS OF HYPERANDROGENISM

Nonvirilizing

Hirsutism

Oily skin and seborrhea

Cystic acne

Diffuse alopecia

Menstrual irregularities

Infertility

Virilizing

Temporal balding

Clitoromegaly

Deepening of the voice

Skeletal muscle hypertrophy and male body habitus

Mammary atrophy

Increased libido

Clinical Features of Hyperandrogenism
Most women with clinical evidence of hyperandrogenism have increased ovarian or adrenal androgen production (Bardin and Lipsett, 1967). In most hyperandrogenic women, SHBG levels are typically low, and free testosterone levels are elevated even when total testosterone levels are normal. Occasional women with normal circulating androgen levels may have isolated increased peripheral compartment (skin and hair follicle) androgen production from increased 5alpha-reductase activity (Horton et al, 1982) (Box 15-1) .

Hirsutism
The most commonly appreciated expression of hyperan- drogenism is excess terminal hair in women. Hirsutism is defined as the transformation of vellus (fine, soft, unpigmented) to terminal (coarse, pigmented, longer) hair in androgen-dependent hair areas ( e.g., upper lip, chin, chest, back, pubis, thighs). Although the hormonal environment influences the conversion of vellus hairs to terminal hairs, the total number of hair follicles is genetically determined and is not changed by the hormonal environment. Racial and ethnic (but not sex) differences exist in total follicle number; for example, white persons have more follicles than Asians, and persons of Mediterranean descent have more follicles than those of northern European descent. The total hair follicle number influences the severity of the expression of hirsutism in hyperandrogenic women. For example, one fourth to one third of women of British and non-Scandinavian European origin may normally have terminal hair on the upper lip, periareolar area, or linea alba, whereas such hair must be considered abnormal in Asian and Scandinavian women

DIFFERENTIAL DIAGNOSIS OF MASCULINIZATION

Idiopathic hirsutism

Polycystic ovarian syndrome

Congenital adrenal hyperplasia (classical and nonclassical)

Androgen-secreting neoplasms

Factitious/iatrogenic

Hypertrichosis

Other medical conditions: Hypothyroidism, juvenile hypothyroidism, acute intermittent porphyria, malnutrition, anorexia nervosa, dermatomyositis, epidermolysis bullosa, Cornelia de Lange syndrome, postencephalitis period, onset of multiple sclerosis, nevoid hypertrichosis, and severe insulin resistance.

Medications: Minoxidil, phenytoin, corticosteroids, penicillamine, diazoxide, streptomycin, hexachlorobenzene, psoralens, and cyclosporine

(Ferriman and Gallway, 1961; Carmina et al, 1992). Unlike those areas, the upper abdomen, sternum, back, and shoulders are distinctly unusual sites for terminal hair in women and should be cause for further evaluation, even in women from genetic backgrounds with more baseline hair follicles. Finally, hirsutism must be differentiated from hypertrichosis, which is a generalized increase in vellus (lanugo in the neonate) but not terminal hair. Hypertrichosis may be associated with certain drugs, metabolic disorders, or malignancy.

Alopecia
Androgen stimulation of the pilosebaceous unit in most parts of the body produces increased hair diameter, pigmentation, rate of growth, and sebaceous gland secretion. Conversely, androgen stimulation of androgen-sensitive hair follicles on the scalp decreases the diameter and rate of hair growth, especially in the temporal area (male pattern baldness). Diffuse hair thinning can also be a sign of mild hyperandrogenism in women (Futterweit et al, 1988).

Acne
Increased production of sebum leads to acne in susceptible persons; if it is chronic, it can produce scarring. Occasionally, acne is the only sign of severe hyperandrogenism. Lucky et al (1983) found that women with acne alone had elevated free testosterone levels similar to those of women with hirsutism with or without acne.

Menstrual Irregularity and Infertility
Hyperandrogenism is often associated with chronic anovulation that results in menstrual irregularity (Hull, 1987). The anovulatory state can be associated with chronic, unopposed estrogen stimulation of the endometrium, resulting in endometrial hyperplasia, erratic heavy menstrual bleeding, and even endometrial carcinoma (Aiman et al, 1986). In addition, the reduced frequency of ovulation is associated with decreased opportunities for conception. Hyperandrogenism probably has an impact on fertility in addition to anovulation. Many studies suggest a decreased conception rate in spite of induced ovulation, and some suggest an increased risk for spontaneous miscarriage in hyperandrogenic women (Balen et al, 1993; Franks, 1995). The mechanisms for this androgen action remain to be determined.

Virilization
In more severe forms of androgen excess, other masculinizing signs develop. These are referred to as true virilization, and they include increased muscle bulk, clitoromegaly [clitoral length greater than 1 cm or diameter greater than 0.5 cm (Tagatz et al, 1979)], temporal balding, and voice deepening. Clitoromegaly and voice changes are irreversible, even if androgen levels can be reduced to normal. Profound androgen excess may also inhibit breast growth and produce an android distribution of body fat.

Behavior
Behavior may be affected by hyperandrogenism. The major source of data on this question derives from adult women with classic virilizing congenital adrenal hyperplasia, most of whom underwent attempts at surgical correction of clitoromegaly with variable success rates. Hochberg et al (1987) noted various psychological changes in genotypic females with virilizing congenital adrenal hyperplasia. These included altered personality, preadolescent tomboyishness, and injured body image, despite an eventual adjustment to feminine gender identity. Increased sex drive (coital and masturbation frequency), increased satisfaction threshold, and versatility in heterosexual relationships were also described in women with "late-treated" adrenogenital syndrome (Ehrhardt et al, 1968). Whether the poor success rate of their genital correction surgery, resulting in difficult or impossible heterosexual intercourse, confounds these findings is unknown. In the common hyperandrogenic disorders in which androgen levels are only modestly elevated ( e.g., testosterone levels ~100 ng/dL), behavioral changes have not been well studied. However, quality of life is compromised in women with PCOS, particularly because of obesity (Cronin et al, 1997).

Differential Diagnosis and Pathophysiology
A woman's concern about androgenic symptoms such as hirsutism or hair loss may differ depending on her ethnic background and her interpretation of normal, which may be influenced by popular images of hairless female beauty. The physician must consider normal racial and ethnic variations in terminal hair distribution and possible causes for hypertrichosis.

Idiopathic Hirsutism
Idiopathic hirsutism (IH) is defined by the maintenance of normal ovulation, despite increased androgen effects on the skin (Lobo, 1986). Regular ovulatory menstrual cycles are reassuring, but never absolute, evidence that serious abnormal causes of hirsutism can be excluded. Women with IH have 5alpha-reductase activity that is greater than that of normal women but similar to that of normal men (Thomas and Oake, 1974; Kutten et al, 1977) (Fig. 15-3) . Although more than 90% of hirsute women have biochemical evidence of hyperandrogenemia (Wild et al, 1983), women with IH tend to have lower androgen production rates (Bardin and Lipsett, 1967). This may explain the absence of ovulatory disturbances in such women. Furthermore, PCOS usually does not develop in women with IH (Dunaif, unpublished observations). Although ultrasonography reveals polycystic ovaries in many women with IH (Adams, 1986), the diagnosis of PCOS cannot be made without some form of anovulation (see later). Recent studies by Legro et al (1998a) have found that hyperandrogenemia with regular menses is an additional familial phenotype in PCOS kindreds. Thus, it may be part of the genetic spectrum of PCOS.

Polycystic Ovary Syndrome and Ovarian Hyperthecosis
These conditions represent a morphologic spectrum of the same process. In 1935, Stein and Leventhal described a syndrome consisting of bilaterally enlarged sclerocystic and polyfollicular ovaries, menstrual dysfunction, and hirsutism (Stein and Leventhal, 1935). Geist and Gaines (1942) noted the association of clinical virilization with islands of luteinized theca cells in the ovarian stroma; this finding subsequently became known as ovarian stromal hyperthecosis.

There is general agreement in the literature that the diagnosis of PCOS requires the presence of hyperandrogenism (elevated serum androgen levels or definitive clinical evidence of excess androgen effect) and chronic anovulation (fewer than six to nine menses per year) (Zawadski and Dunaif, 1992). It is now recognized that PCOS represents the most common cause of masculinization in women with a prevalence of approximately 5% of premenopausal women (Knochenhauer et al, 1998).

Women with PCOS are chronically anovulatory, but spontaneous ovulation and conception may occasionally occur. Usually, menstrual irregularities persist from the time of menarche so that a regular pattern of menses is never established. Less often, girls can have primary amenorrhea. Hyperandrogenism may be subtle, and cystic acne may be the only sign. Hyperandrogenism may take several years to produce hirsutism in PCOS or hirtutism may be absent, depending on the 5alpha-reductase activity in the skin (McKenna et al, 1983). Acanthosis nigricans is evident on clinical examination of the skin in approximately 50% of obese women with PCOS (Dunaif et al, 1987; Stuart et al, 1986; Dunaif et al, 1991) (Fig. 15-4) . Acromegaloid features, such as acral hypertrophy, are found occasionally in women with PCOS. True virilization is uncommon in PCOS and suggests the presence of an adrenal or ovarian tumor. Clinical findings in women with PCOS are summarized:

CLINICAL FEATURES IN WOMEN WITH POLYCYSTIC OVARIAN SYNDROME

Frequent

Hirsutism

Infertility

Oligomenorrhea

Obesity

Common

Amenorrhea

Seborrhea

Acne

Acanthosis nigricans

Alopecia

Impaired glucose tolerance

Rare

True virilization

Acral hypertrophy

Endometrial carcinoma

Hypertension

Non-androgen-secreting ovarian tumors ( e.g., dermoid cysts)

In PCOS, classic ovarian morphology includes a thickened cortex, multiple subcapsular follicular cysts, hyperplasia and luteinization of the theca interna, stromal hyperplasia, and multiple immature follicles suggestive of arrested folliculogenesis (Hughesdon, 1982). Ovary size ranges from normal to substantially enlarged. On ultrasonography, these histologic findings appear as a peripheral array of at least eight small follicles (6 to 10 mm in diameter), and there is an increased amount of dense stroma (Adams criteria) (Adams et al, 1986) (Fig. 15-5) . Although most women with PCOS have the typical polycystic ovarian morphology, this ovarian morphology is not specific to PCOS. Enlarged polycystic ovaries are seen in women with adult-onset 21- hydroxylase deficiency (Hague et al, 1990) and 20% to 25% of normal controls (Polson et al, 1988). The lack of specificity of ovarian morphologic changes for which this disorder was named emphasizes the importance of defining the biochemical features of PCOS and of excluding other diseases before diagnosis.

The primary androgen-secreting cells of the ovary are thecal and stromal cells, which respond to stimulation by LH (Erickson et al, 1985). Follicles contain granulosa cells that, under normal circumstances, aromatize locally produced androgens to estrogens (primarily E2). The capacity of granulosa cells to aromatize androgens is a function of their maturity, and this is under direct follicle-stimulating hormone (FSH) control (Erickson et al, 1985).

There is considerable overlap in the reported clinical and biochemical features of hyperthecosis and PCOS (Geist and Gaines, 1942; Dunaif et al, 1985). Hyperthecosis has been defined as luteinized cells that are clustered or diffusely scattered away from the follicles in the stroma and that encroach on the hilar region. However, when ovarian sections are examined carefully by histologic analysis, islands of luteinized theca cells (stromal hyperthecosis) can be detected in most ovaries of women with PCOS (Hughesdon, 1982). More extensive stromal hyperthecosis often correlates with more severe androgen excess, producing true virilization. However, stromal hyperthecosis also can be associated with estrogen production alone and may be a cause of postmenopausal vaginal bleeding (Nagamani et al, 1988). Stromal hyperthecosis is more extensive in women with PCOS if there is substantial insulin resistance, and it has been suggested that insulin directly stimulates theca cell growth (Dunaif et al, 1985; Nagamani et al, 1988). Indeed, the extent of theca cell hyperplasia has been shown to be positively related to plasma insulin levels (Nagamani et al, 1988).

Biochemical Features.
Characteristic biochemical abnormalities of PCOS include elevated serum androgen levels, decreased SHBG levels, disordered gonadotropin release with increased LH relative to FSH secretion, and acyclic (tonic) estrogen levels in the midfollicular range of the normal menstrual cycle associated with chronic anovulation. Hyperinsulinemia secondary to insulin resistance appears to be a unique feature of PCOS and not of hyperandrogenic states in general (Dunaif et al, 1987). Biochemical abnormalities may wax and wane, and attempts to categorize this disorder by biochemical profiles are problematic.

Elevated Androgens.
Plasma levels of testosterone, biologically available testosterone (non-SHBG-bound and free), androstenedione, DHEAS, DHEA, and DHT can be elevated. Typically, testosterone, non-SHBG-bound testosterone, and androstenedione levels are mildly to moderately elevated compared with nonhirsute regularly ovulating women. If serum testosterone levels are consistently greater than 150 to 200 ng/dL when measured in an extraction and chromatography assay system or if DHEAS levels are greater than 700 mug/dL (Meldrum and Abraham, 1979; Moltz et al, 1984; Friedman et al, 1985; Derksen et al, 1994), an androgen-secreting neoplasm must be excluded. Peripheral androgen elevations fluctuate with pulsatile adrenal secretion and ovarian menstrual cyclicity so that a single determination may miss androgen excess. Occasionally repeated sampling or pooling of samples may be necessary to detect androgen excess.

By combining the results of a number of studies, it can be seen that there is increased ovarian androgen production in most women with PCOS. Under normal circumstances, more than 90% of serum DHEAS is secreted by the adrenals (Longcope, 1986); thus, its elevation in women with PCOS indicates that there is an adrenal contribution to androgen excess. However, because the adrenals also contribute substantially to circulating testosterone and androstenedione levels, both by direct secretion and by peripheral conversion of prohormones, adrenal hyperan- drogenism may occur in the absence of elevated DHEAS levels. Many women with PCOS have an adrenal component to their androgen excess (Ehrmann et al, 1995).

Sex Hormone-Binding Globulin.
Serum levels of SHBG are modestly decreased in nonobese women with PCOS and substantially decreased in obese women with PCOS. This is secondary, in part, to the direct action of androgens to decrease hepatic production of SHBG (Andersen, 1974; Winneker et al, 1989). Obesity is also associated with low SHBG levels, which may be attributed to an independent effect of insulin to decrease hepatic SHBG synthesis directly (Plymate et al, 1988; Nestler et al, 1991).

Estrogens.
Serum levels of total estrogens are elevated above the range seen in normally menstruating women. Estradiol levels are similar to those seen in the midfollicular phase of the normal menstrual cycle, whereas E1 levels are typically elevated compared with those in normal women (Rebar et al, 1976; Baird et al, 1977). Therefore, on clinical examination, women with PCOS are well estrogenized ( e.g., they have moist vaginal mucosa with rugae, and they experience vaginal bleeding after progestin challenge). Because SHBG levels are low, non-SHBG-bound, biologically available E2 levels are increased compared with those in normally ovulating women in the midfollicular phase of the menstrual cycle (Lobo et al, 1981). This may be important in the pathogenesis of the gonadotropin secretory abnormalities of PCOS (Lobo et al, 1981).

Estrogen secretion in PCOS is derived from the ovary and the extragonadal aromatization of androgens (par- ticularly androstenedione to E1). Estrogen production in PCOS is constant and is not cyclic as it is in the normal menstrual cycle. This results in a chronic, unopposed ( i.e., no progesterone) effect on the endometrium that can result in endometrial hyperplasia, dysfunctional uterine bleeding, and endometrial neoplasia if left untreated (Jackson and Docherty, 1957).

Gonadotropins.
Serum LH levels are increased in most women with PCOS as a function of increased amplitude and increased frequency of pulsatile LH secretion (Waldstreicher et al, 1985; Taylor et al, 1997). Serum FSH levels and pulse frequency are generally in the normal midfollicular range, or they may be slightly low. An increased frequency of LH pulses indicates an increased frequency of hypothalamic gonadotropin-releasing hormone (GnRH) secretion, and, therefore, a hypothalamic site of the abnormality. The increased amplitude of LH pulses may have several causes. First, women with PCOS secrete more LH in response to the same dosage of GnRH (Rebar et al, 1976) or GnRH agonists (Barnes et al, 1989), indicating increased pituitary sensitivity. Second, an increased GnRH pulse frequency has been shown to increase the LH-FSH ratio experimentally, indicating a direct effect of frequency (Spratt et al, 1987). Third, there does not appear to be an increase in the amount of GnRH secreted per pulse because LH is similarly suppressed by small dosages of a GnRH-antagonist in women with PCOS and in normal controls (Hayes et al, 1998). Finally, there does not appear to be an excess secretion of inhibin from PCOS ovaries (Lambert-Messelian et al, 1994).

Insulin Resistance.
Insulin resistance is a unique feature of PCOS and not of hyperandrogenic states in general (Dunaif et al, 1987; Dunaif et al, 1989; Robinson et al, 1993). Approximately 40% of women with PCOS have impaired glucose tolerance (31%) or frank diabetes mellitus (7.5%) (Legro et al, 1999). Although obesity and age increase this risk, glucose intolerance, including type 2 diabetes, can be seen in adolescent and nonobese PCOS women (Legro et al, 1999). PCOS is an important risk factor for the development of type 2 diabetes mellitus in women. It is estimated that approximately 10% of diabetes in premenopausal women is PCOS related (Dunaif, 1997).

Pathophysiology of Polycystic Ovary Syndrome.
There are four distinct loci of disturbed endocrine functioning in women with PCOS: ovaries, adrenal gland, hypothalamic-pituitary axis, and peripheral insulin-responsive tissues.

Ovarian Function.
Theca cells isolated from polycystic ovaries secrete markedly increased amounts of androstenedione and 17-OH progesterone basally and in response to LH (Gilling-Smith et al, 1994). Despite extensive investigation, however, no consistent specific enzymatic defect in steroidogenesis has been found in the majority of ovaries of women with PCOS. A number of steroidogenic pathways are upregulated, consistent with a diffuse hyperresponsiveness rather than a specific enzyme defect (Ehrmann et al, 1995). Many women with PCOS have exaggerated ovarian 17-OH progesterone secretion, suggesting dysregulation of the cytochrome enzyme P450c17 that controls 17 hydroxylation of pregnenolone and progesterone and 17,20 lyase activity to convert 17-OH pregnenolone to DHEA (Ehrmann et al, 1995). This is the rate-limiting enzyme for androgen biosynthesis. No mutations have been found in this enzyme in women with PCOS (Franks et al, 1997; Legro et al, 1998b).

Granulosa cells from polycystic ovaries secrete smaller amounts of estrogen than granulosa cells from normal ovaries. However, the addition of FSH to PCOS granulosa cells results in normal estrogen secretion (Erickson et al, 1985; Mason et al, 1994). This suggests that ovarian responsiveness is normal and that the decreased granulosa cell estrogen production characteristic of PCOS is a functional abnormality secondary to inadequate FSH stimulation.

Adrenal Function.
The adrenal glands can contribute to hyperandrogenism in PCOS as indicated by selective venous catheterization studies, elevations of serum DHEAS levels, and studies in which ovarian or adrenal steroid production has been suppressed. Indeed, the nonclassical, late-onset form of adrenal 21-hydroxylase deficiency has clinical, biochemical, and ovarian morphological features indistinguishable from PCOS (Lobo and Goebelsmann, 1980). In the absence of these definable adrenal enzyme defects, women with PCOS often have increased responses to exogenous ACTH of DHEA, 17-OH progesterone, 17- hydroxypregnenolone, and androstenedione (Luckey et al, 1986; Ehrmann et al, 1995). However, these responses do not conform to a pattern for a single adrenal enzyme defect and suggest, instead, generalized adrenal hyperresponsiveness. It is possible that this represents a primary abnormality. Alternatively, insulin has been shown to increase adrenal responsiveness to ACTH (Moghetti et al, 1996).

Function of the Hypothalamic-Pituitary Axis.
Several hypotheses have been proposed to explain the neuroendocrine defect of PCOS, including a primary hypothalamic defect, abnormal sex steroid feedback, and abnormal sensitivity to feedback. Spontaneous ovulation (Blankstein et al, 1987; Taylor et al, 1997) or progestin exposure (Christman et al, 1991; Pastor et al, 1998; Daniels and Berga, 1997) can reduce exaggerated LH secretion in women with PCOS but may not normalize it, suggesting that an underlying hypothalmic defect persists regardless of normal feedback. The antiestrogen clomiphene citrate induces a similar rise, and exogenous estrogen produces a similar fall of LH and FSH release in PCOS and in normal ovulatory women in the midfollicular phase of the menstrual cycle, suggesting that estrogen-negative feedback is intact (Rebar et al, 1976). However, estrogen administration provokes a rise (positive feedback) in LH and FSH levels in women with PCOS 24 hours earlier than it does in normal women in the midfollicular phase. In women with PCOS, there is also increased LH release during GnRH administration than there is in normal women in the midfollicular phase (Rebar et al, 1976). These findings suggest that women with PCOS are "estrogen primed" because both early positive feedback and increased pituitary sensitivity to GnRH can be produced in normal women by exogenous estrogen administration.

Elevated LH and LH-FSH ratios have been described in women with hyperandrogenism of other causes, such as androgen-secreting ovarian neoplasms (Dunaif et al, 1984b) or nonclassical congenital adrenal hyperplasia (Lobo and Goebelsmann, 1980). Androgens are able to distort gonadotropin release by their aromatization to estrogens (Dunaif, 1986). Although androgen levels in the normal male range suppress gonadotropin release in normal women (Serafini et al, 1986), androgen levels in the PCOS range do not directly alter gonadotropin release in women with PCOS or in normal women (Dunaif, 1986). Conversely, the administration of an aromatase inhibitor to women with PCOS results in increases in gonadotropin release characteristic of an antiestrogenic effect (Dunaif, et al, 1984a). However, E1 administration does not alter LH release in women with PCOS (Chang et al, 1982). This suggests that E2 is the major gonadal steroid contributing to the distortion of gonadotropin release in PCOS.

There is growing evidence that gonadotropin secretion in PCOS is also modulated by some factor related to body weight (Taylor et al, 1997; Arroyo et al, 1997). Obese patients tend to have lower mean LH levels, LH-FSH ratios, and LH pulse amplitudes, but they maintain rapid LH pulse frequencies. The inverse correlation of LH secretion with body mass index is similar with percentage body fat and fasting serum leptin levels and is slightly less with fasting insulin levels.

Biochemical-Pathology Correlation.
The ovarian pathology findings are consistent with the abnormal hormonal milieu. It is believed that the hyperplastic thecal and stromal tissues are secondary to chronic stimulation by excess LH. It has been suggested by Givens et al (1975) that the gonadotropin-responsive tumors that occasionally coexist with PCOS may be the end result of this stimulation, as are the androgen-secreting luteomas that may occur during pregnancy (see later). Hyperinsulinemia is associated with stromal hyperthecosis in PCOS (Dunaif et al, 1985). Exogenous androgen administration may also result in polycystic ovarian changes (Futterweit and Deligdisch, 1985).

Hypotheses for the Cause of Polycystic Ovary Syndrome.
It has been demonstrated that disordered gonadotropin secretion is usually necessary to perpetuate hyperandrogenism in PCOS (Chang et al, 1983) and that either ovarian or adrenal androgen excess can cause gonadotropin release identical to that of PCOS (Lobo and Goebelsmann, 1980; Dunaif et al, 1984b). Thus, a primary central abnormality leading to disrupted gonadotropin secretion or a primary ovarian or adrenal defect leading to hyperandrogenism could be postulated to cause PCOS as follows. First, it has been suggested that a vicious circle is initiated by adrenal androgen excess because of exaggerated adrenarche (Yen, 1980). This causes distorted gonadotropin release with increased LH relative to FSH secretion, resulting in arrested folliculogenesis and LH-dependent ovarian androgen production. Excess androgens are then aromatized peripherally into estrogens. Abnormal gonadal steroid production in turn perpetuates the disordered gonadotropin release. However, gonadotropin secretory abnormalities and hyperandrogenism resume promptly in women with PCOS after medical oophorectomy with long-acting GnRH analogues (Chang et al, 1983). This suggests that there is a primary persistent abnormality in the control of GnRH release or in steroidogenesis in PCOS.

Second, it has been suggested that there is a primary central nervous system alteration that results in PCOS. Indeed, adolescent girls with PCOS have disordered diurnal secretory patterns of LH, suggesting that a neuroendocrine abnormality may be involved (Zumoff et al, 1983). A number of potential neuroendocrine changes have been suggested in PCOS, including decreases in central dopaminergic tone, but these may be secondary to tonic estrogen feedback rather than primary lesions.

Third, there has been renewed interest in a primary ovarian or adrenal enzymatic defect leading to impaired folliculogenesis or increased androgen production. Primary ovarian defects that have been reported in PCOS include abnormal 11beta-hydroxylase, 3beta-hydroxysteroid dehydrogenase, 17-ketosteroid reductase, 17beta-hydroxysteroid dehydrogenase, and 17alpha-hydroxylase/17-20 lyase activities (Ehrmann et al, 1995). To date, however, there is no strong evidence for primary genetic defects in these ovarian steroidogenic enzymes in PCOS because ovaries appear to function normally in response to FSH (Erickson et al, 1985; Franks, 1995).

Fourth, there has been growing speculation that insulin or insulin-like growth factors (IGFs) play a major role in the pathogenesis of PCOS. Acute infusion of supraphysiologic amounts of insulin can directly alter gonadal steroid secretion in women with PCOS (Dunaif and Graf, 1989). Moreover, lowering insulin levels by weight loss (Kiddy et al, 1992) or the use of pharmacologic agents (Nestler et al, 1989; Dunaif et al, 1996; Nestler et al, 1998) results in decreased circulating androgen levels and the resumption of ovulation.

In summary, PCOS is characterized by masculinization and is differentiated from IH by the presence of oligoovulation or anovulation. Only by excluding other causes (Zawodski and Dunaif, 1992) can a diagnosis of PCOS be confirmed. Ultrasonographic documentation of polycystic ovaries supports the diagnosis, but it is not definitive for the disorder.

Metabolic Consequences of Polycystic Ovary Syndrome.
The association of PCOS with obesity (Evans et al, 1983), insulin resistance, and glucose intolerance raises the concern that affected women are at increased risk for coronary artery disease and other atherosclerotic conditions. Several large epidemiologic studies have linked hyperinsulinemia with an increased risk for cardiovascular events. However, none of these studies included women, and no long-term prospective studies of cardiovascular risk in women with PCOS have been conducted to address the issue.

Because it takes a long time to complete prospective cardiac event outcome studies, alternative approaches to assess cardiac risk in patients with PCOS have been conducted, including assessment of current cardiovascular risk factors such as hypertension and hyperlipidemia and retrospective studies of all women with symptoms of cardiovascular disease. Some, but not all, studies of fasting lipid levels in women with PCOS, defined by variable criteria, suggest that high-density lipoprotein levels may be reduced and triglyceride levels increased in hyperandrogenic women (Wild, 1995; Talbott et al, 1995). Some of the variability may be explained by the small numbers of patients in some studies and by the lack of weight- and age-matched controls in other studies. Our own data (Graf et al, 1990) suggest that most of the differences in lipid levels between women with PCOS and control subjects can be explained by the differences in body mass index. It remains unclear whether PCOS bestows an increased risk over what would be expected for the degree of insulin resistance and increased prevalence of impaired glucose tolerance and type 2 diabetes. Indeed, 15% of older women with a history consistent with PCOS have diabetes mellitus (Dahlgren et al, 1992b).

Hypertension has also been associated with hyperinsulinemia. However, blood pressure is clearly not increased in young patients with PCOS compared to weight- and body fat-matched controls, even though the patients with PCOS in one study were clearly insulin resistant and hyperinsulinemic (Zimmerman et al, 1992). These results raise the possibility that insulin resistance in women with PCOS, or insulin resistance in women in general, plays a different role in cardiovascular risk than it does in men. However, older women with a history of wedge resection for PCOS may have a prevalence for hypertension that is as high as 39% (Dahlgren et al, 1992b).

More recent studies have addressed other surrogate risk factors for cardiovascular disease, including the production of plasminogen activator-I (PAI-1), which is enhanced by hyperinsulinemia and is associated with a decreased fibrinolytic response to thrombosis. PAI-1 concentration and activity are elevated in women with PCOS (Andersen et al, 1995). Ehrmann et al (1997b) demonstrated reductions in PAI-1 when insulin levels were reduced by troglitazone in patients with PCOS.

To date, two retrospective studies have been reported of women undergoing cardiac catheterization for the evaluation of chest pain. Both are weakened by their small sample size, mixture of premenopausal and postmenopausal subjects, and retrospective assessment of previous hirsutism and menstrual dysfunction. However, both suggest that women who had documented coronary artery disease at the time of catheterization were more likely to have a history consistent with hyperandrogenism. In the first study, Wild et al (1990) reported that there was a significant increase in hirsutism in women with catheterization evidence of coronary artery disease. In the second study, Birdsall et al (1997) found an increase in documented coronary artery disease in women with polycystic ovarian morphology.

Thus, women with PCOS clearly have a proclivity for insulin resistance that may contribute to other risk factors for coronary artery disease and to more coronary atherosclerosis. However, data confirming this hypothesis are unavailable. Current recommendations, given these data, are to consider all cardiovascular risk factors in women with PCOS and to screen vigilantly and treat for obesity, hypertension, hyperlipidemia, and type 2 diabetes mellitus.

Androgen-Secreting Ovarian Tumors
A few ovarian tumors (less than 1%) are capable of producing masculinization by the direct secretion of androgens. These tumors are classified as sex-cord stromal tumors, steroid or lipoid cell tumors, gonadoblastomas, and tumors with functioning stroma. The category of sex-cord stromal tumors contains derivatives of the sex cords (granulosa or Sertoli cells) or stroma singly or in any combination and in various degrees of differentiation. Lipoid cell tumor refers to the morphologic features of the steroid-producing cells that make them up, such as luteinized thecoma, Leydig, and adrenal-cortical-like cells. They can be divided into three categories according their cells of origin, known or unknown. Gonadoblastomas are complex tumors composed of sex-cord, stromal, and germ-cell elements. They almost invariably arise in association with gonadal dysgenesis when a Y chromosome-containing cell line is present and produces masculinization from the excessive production of androgens by the Leydig cell elements.

Other Ovarian Tumor Hyperandrogenic Conditions.
Several tumors stimulate surrounding stroma into steroid-secreting tissue by human chorionic gonadotropin (hCG)- independent mechanisms. Despite a report of virilizing serous cystadenoma and Brenner tumor without histologic evidence of stromal hyperplasia or luteinization, it is debatable whether these types of ovarian cancer have the necessary steroidogenic enzymes for androgen secretion. Rather, it is generally accepted that neighboring stromata is mechanically stimulated to secrete androgens, as with an expanding follicle, or is trophically stimulated by paracrine or endocrine factors.

Pregnancy.
During pregnancy, certain benign ovarian lesions, most commonly luteomas and hyperreactio luteinalis, are capable of androgen hypersecretion (Shortle et al, 1987). These lesions are derived from luteinized stromal cells, present before pregnancy, that respond unusually to placental gonadotropin secretion. Maternal virilization occurs in 10% to 50% of luteomas and in 25% of hyperreactio luteinalis. Fetal masculinization is rare, however, because of the large placental capacity to aromatize androgens into estrogens.

Postmenopause
Postmenopausal women may show masculinization that results from the lowered free estrogen to free androgen ratio caused by declining estrogen secretion in the aged ovary. Androgen secretion may continue at a disproportionately high rate because of stromal stimulation by elevated postmenopausal gonadotropins. Many elderly women experience increased growth of terminal hair on the upper lip and chin, whereas pubic, axillary, and scalp hair are partially lost.

Adrenal Masculinizing Conditions
The principal disorders of the adrenal gland that can cause masculinization in women are congenital adrenal hyperplasia (CAH), Cushing syndrome with significant hyperandrogenism, and testosterone-producing or androgen precursor-producing adrenal tumors. Usually, adrenal hyperandrogenemia is suspected when elevated serum DHEAS levels are found. An isolated elevation in serum testosterone level, however, does not exclude an adrenal source. CAH and adrenal virilizing tumors must be differentiated from premature adrenarche in girls in whom the normal rise in adrenal androgen secretion occurs before 7 years of age. In premature adrenarche, there may be mild acceleration in height and bone age accompanied by precocious pubic hair development, but there is no virilization as there is with CAH and adrenal tumors.

Nonclassical Congenital Adrenal Hyperplasia.
Hyperandrogenism resulting from inherited defects in adrenal steroid biosynthesis (predominantly 21-hydroxylase deficiency) can present during adolescence or adulthood (late-onset, attenuated, or nonclassical CAH [NCCAH]) (Fig. 15-6) . In these disorders, an enzymatic defect integral to the formation of cortisol causes a slight compensatory increase in pituitary ACTH production and an increased conversion of cortisol precursors to androgens. Women with late-onset CAH may have a history of prepubertal hirsutism but rarely prepubertal virilization.

The most common form of CAH is nonclassical 21- hydroxylase deficiency, which is caused by mutations in the P450c21 gene on chromosome 6 and is inherited as an autosomal recessive trait (New and Speiser, 1986). The nonclassical form of 21-hydroxylase deficiency is not detected until puberty and is not associated with salt wasting, severe virilization, or adrenal insufficiency. This disorder is easily diagnosed by 17-hydroxyprogesterone responses after ACTH administration (see Fig. 15-6) . Unstimulated early morning 17-hydroxyprogesterone levels may also be used for diagnosis (Azziz and Zacur, 1989). The incidence of NCCAH among hirsute women ranges from 1% to 5% depending on the ethnic background, with an overall frequency of 0.3% in the general white population and approximately 3% in Jews of European origin (New and Speiser, 1986).

The clinical picture for late-onset 3beta-hydroxysteroid dehydrogenase deficiency is indistinguishable from that for PCOS--peripubertal hirsutism and menstrual irregularity (Pang et al, 1985; Zerah et al, 1994). Biochemically, baseline Delta5 steroids (DHEAS and DHEA) are more elevated than Delta4 steroids (androstenedione and testosterone), and the ACTH-stimulated 17alpha-hydroxypregnenolone-17alpha- hydroxyprogesterone ratio is elevated to a greater degree than the DHEA-androstenedione ratio. Although this steroid pattern has been reported in as many as 15% of hirsute women based on ACTH stimulation testing (Pang et al, 1985), 3beta-hydroxysteroid dehydrogenase gene cloning has demonstrated that mutations in this gene are exceedingly rare (Zerah et al, 1994). Thus, most elevations of Delta5 steroids appear to be caused by functional androgen production defects rather than by enzyme gene mutations. A late-onset form of 11beta-hydroxylase deficiency has been reported, though a consistent biochemical response has not been observed. Not all patients have increased 11-deoxycortisol-cortisol ratios.

Cushing Syndrome.
Cushing syndrome is caused by the excess, poorly modulated secretion of cortisol due to an adrenal cortisol-producing tumor, an ACTH-secreting pituitary tumor (Cushing disease), an ectopic tumor producing ACTH, or a tumor producing corticotropin-releasing hormone. The effects of chronic hypercortisolism include central obesity, hypokalemic alkalosis, hypertension, impaired glucose tolerance, muscle wasting, thinning of the stratum corneum of the skin, and osteoporosis. In addition, many tumors also secrete androgen, which can produce hirsutism, acne, seborrhea, and true virilization. Some nonandrogen-sensitive hair growth (facial lanugo hair) may occur because of the effect of excess glucocorticoids. It is estimated that approximately 75% of women with Cushing disease have hirsutism. Hypercortisolism and hyperandrogenism may impact the hypothalamic-pituitary-gonadal axis, producing ovulatory disturbances and menstrual irregularity. When significant virilization occurs in Cushing syndrome, the cause is almost always adrenal carcinoma. Adrenal carcinoma are often palpable because the tissue is steroidogenically inefficient and does not produce sufficient hormones for clinical symptoms until there is considerable tumor mass.

Because the androgens DHEA and androstenedione are derived from 17-hydroxypregnenolone by the action of the P450c17 and 3betaHSD genes, any stimulation of low-density-lipoprotein cholesterol uptake and cholesterol side-chain cleavage by ACTH produces a state of cortisol and androgen hypersecretion (Ehrmann et al, 1995). Tumors, on the other hand, are independent of ACTH stimulation and can have varying complements of steroid biosynthetic enzymes. Usually the benign adrenal adenoma that produces hypercortisolism is well differentiated and does not secrete increased androgens. Thus, DHEAS levels are typically suppressed if there are adrenocortical adenomas because of the suppression of ACTH by the elevated plasma cortisol levels. DHEAS levels may be useful in differentiating adrenocortical adenomas from other causes of Cushing syndrome (Yamaji and Ibayashi, 1969). Adrenal carcinomas often produce excess androgens and estrogens, in addition to excess cortisol, but may not be associated with elevated DHEAS levels. Because androstenedione and DHEA are really androgen prehormones, masculinizing and possible virilizing signs of Cushing syndrome result from peripheral conversion to testosterone. Occasionally, a pure testosterone-producing tumor occurs without hypercortisolism or elevated DHEAS levels (see later).

Other Androgen-Secreting Adrenal Tumors.
Androgen-secreting adrenal tumors with normal serum cortisol levels can be associated with hypertension mediated by excessive 11-deoxycorticosterone secretion. Pure testosterone-secreting tumors are rare. Approximately 20 cases have been described, usually in postmenopausal women (Gabrilove et al, 1981). Differentiating adrenal testosterone-secreting tumors from ovarian testosterone-secreting tumors is challenging. Such neoplasms may contain LH receptors; hence, hCG may stimulate testosterone secretion in both.

Because of the variable biochemical and clinical picture, 45% of 22 patients with testosterone-secreting adrenal tumors reported between 1975 and 1987 underwent ovarian exploration before the correct diagnosis was made (Mattox and Phelan, 1987). This difficulty underscores the need for initial high-resolution computed tomography (CT) or magnetic resonance imaging (MRI) to exclude an adrenal neoplasm whenever tumoral hyperandrogenemia is suspected because of an elevated serum testosterone level. Usually, pure testosterone-secreting adrenal tumors are benign, whereas adrenal masculinizing tumors whose major androgenic steroid secretory product is DHEA often are malignant.

Hyperprolactinemia
Prolactin excess has been associated with hyperandrogenism (often hirsutism) in a variety of circumstances. Prolactin may augment adrenal androgen secretion by the inhibition of 3beta-hydroxysteroid dehydrogenase activity or, less often, through selective action on the sulfation of DHEA in adrenal or extra-adrenal sites (Carter et al, 1977). However, prolactin inhibits FSH-induced ovarian aromatase, leading to intraovarian hyperandrogenemia. In hyperprolactinemic women (prolactin range, 36 to 991 ng/mL) studied by Glickman et al (1982), 40% had androgenic abnormalities of which the most common was elevated free testosterone levels. The next in frequency was depressed SHBG levels and then elevated DHEAS levels.

Modestly increased prolactin levels have been reported in as many as 40% of women with PCOS in the absence of pituitary neoplasms. However, most investigators find substantially fewer women with hyperprolactinemia and, in fact, exclude patients with elevated prolactin levels from the diagnosis of PCOS. It has been suggested that hyperprolactinemia in PCOS is related to decreased central dopaminergic tone (Paradisi et al, 1988), which, in turn, may reflect the tonic estrogenic state.

Hypothyroidism
There are many effects of thyroid disease on adrenocortical and reproductive function. Hypothyroidism is associated with an increased metabolic production rate of testosterone, diversion of testosterone metabolism from androsterone to etiocholanolone, and reduced binding activity and hepatic production of SHBG (Gordon et al, 1969). Hypothyroidism also decreases libido and causes anovulation, infertility, alopecia, and excessive and irregular menstrual bleeding. Despite derangements in androgen metabolism in women with hypothyroidism, masculinization is not significant. Conversely, masculinization may be present in juvenile hypothyroidism. A predominant feature is reversible generalized muscular hypertrophy. Furthermore, juvenile hypothyroidism is sometimes associated with precocious puberty.

Disorders of Sexual Differentiation
Hyperandrogenic symptoms, either prepubertal or postpubertal, may occur in various types of gonadal dysgenesis, among them classic Turner syndrome (streak gonads and a Turner phenotype); pure gonadal dysgenesis (aplastic or absent gonads without the recognizable Turner syndrome phenotype); true hermaphroditism (ovarian and testicular tissue); and hermaphroditism with atypical or mixed gonadal dysgenesis (streak gonad and testis). Thus, it is mandatory to exclude a Y-cell line containing gonadoblastoma in any patient with gonadal dysgenesis who has hyperandrog- enism. This can be accomplished only by laparotomy and by removal of the streak gonad because the Y-cell line may occur only there and may not be detectable by peripheral lymphocyte or skin fibroblast chromosomal analysis.

Anorexia Nervosa and Starvation
In general, weight loss is not associated with hyperandrogenism. In anorexia nervosa, however, there is an increase in lanugo hair and amenorrhea and a reduction in breast size.

Insulin Resistance
Hyperandrogenism with true virilization occurs in a number of the rare syndromes of extreme insulin resistance, such as type A insulin resistance, leprechaunism, and partial or complete lipoatrophy syndrome (Barbieri et al, 1986; Dunaif, 1997). In these conditions, insulin levels are strikingly elevated and may increase ovarian androgen production directly or by binding to the IGF-1 receptor. Moderate insulin resistance is also a feature of PCOS and can occur in association with androgen-secreting neoplasms. Thus, the presence of insulin resistance does not aid in the differential diagnosis of hyperandrogenism.

Obesity
Obesity may be both a manifestation of the hyperandrogenic state and a contributor to it. Increased adiposity has been associated with decreased hepatic SHBG synthesis and increased androgen bioavailability (Anderson, 1974). In addition, adipose cell aromatase converts androgens to estrogens, and chronically high levels of estrogen promote adipocyte replication in vitro. This supports the clinical observations of increased menstrual irregularities and of hirsutism in obese women that may be corrected with weight reduction (Rogers and Mitchell, 1952; Glass et al, 1978). However, not every obese woman has hyperandrog- enism or ovulatory disturbances (Dunaif et al, 1988). Weight reduction is an effective therapeutic modality for PCOS, but it is not possible to predict which obese women will experience improved menstrual function with weight reduction (Kiddy et al, 1992).

Iatrogenic or Factitious Masculinization
Treatment with various drugs may produce masculinizing features (Box 15-4) . Unfortunately, most virilizing side effects are irreversible, even when steroid use is discontinued. In addition to the increasing use of illicit anabolic steroids among women to enhance athletic performance, androgens may be legitimately prescribed for the treatment of menopausal signs and symptoms, breast cancer, aplastic anemia, endometriosis, cystic mastitis, and angioneurotic edema (Wilson, 1988). Hirsutism and acne are side effects of the common progestational oral contraceptive agents that contain levonorgestrel and norgestrel ( e.g., Triphasil and Lo-Ovral, respectively). Hirsutism, acne, and alopecia occasionally result from synthetic oral, intravenous, or inhaled glucocorticoids. Fetal virilization can develop if pregnant women ingest progestational or androgenic steroids.

DIAGNOSTIC STRATEGY
Whom to Evaluate
Any woman with hyperandrogenic symptoms should undergo initial evaluation for potentially abnormal causes of increased androgen production (see Box 15-2) . In addition, hyperandrogenism should be considered in women with irregular menses, amenorrhea, or infertility. The basic evaluation includes a complete history and physical examination to rule out a serious underlying cause (Fig. 15-7) .

Medical History
Important features of the history include age, rapidity of onset of masculinizing symptoms, presence of menstrual dysfunction, and family history including ethnic origin. Hirsutism that is insidious and coincides with puberty suggests a nontumoral cause ( e.g., PCOS or NCCAH), whereas rapidly progressive hirsutism with onset after puberty and significant virilizing symptoms, such as increased muscularity and voice deepening, suggests neoplasia. True virilization indicates a severe hyperandrogenic state, whereas the preservation of menstrual cyclicity suggests a more attenuated state. It is also important to determine age at menarche, menstrual and growth histories, fertility (spontaneous or induced), and libido. If these are normal and hirsutism is present, then IH is likely. If the patient is pregnant, a pregnancy luteoma should be suspected. Any history of endocrinopathy or metabolic disease is particularly important. Drug use, especially oral contraceptive pills and anabolic steroids, should be recorded. Surreptitious anabolic steroid use should be suspected in body builders and competitive athletes.

DRUGS WHOSE USE IS ASSOCIATED WITH IATROGENIC MASCULINIZATION

Synthetic glucocorticoids

Adrenocorticotropic hormone

Metyrapone

Anabolic steroids

Levonorgestrel-containing oral contraceptive pills

Maternal use of synthetic progestational agents (fetal virilization)

Physical Examination
Hypertension, bruising, mood changes, and proximal muscle weakness should increase the index of suspicion for Cushing syndrome. Voice pitch should be noted also because a deep voice is suggestive of more severe hyperandrogenism. A masculine, feminine, Cushingoid, or Turner body habitus should be recorded. For example, if a patient has Turner syndrome, hirsutism suggests the presence of a Y-cell line gonadoblastoma. Breasts should be examined for atrophy or galactorrhea. The pelvis and abdomen should be palpated for ovarian or adrenal masses. Skin should be examined for dryness (hypothyroidism), hirsutism, striae, acne or seborrhea/oiliness, alopecia, acanthosis nigricans , or hyperpigmentation (excessive ACTH secretion). Acanthosis nigricans is a cutaneous marker for insulin resistance (Dunaif et al, 1990). Acral hypertrophy can be found in the type A syndrome of insulin resistance, acromegaly, and, occasionally, in PCOS.

External genitalia must be examined carefully for clitoral size, labial development, distribution of pubic hair, and genitourinary malformation. Bimanual examination of the uterus and adnexa is particularly important to find pelvic neoplasms and to disclose genital developmental abnormalities in patients with primary amenorrhea.

Various indices may be used to quantify hirsutism. According to the Ferriman and Gallwey (1961) scoring system, five gradings based on hair density and area are assigned for 11 anatomic regions, and a score is computed from the sum of the gradings. This semiquantitative method is suitable for the initial evaluation, but it is insensitive to changes after treatment because of the limits of visual inspection. It does not assign adequate weight to facial hirsutism because the scores reflect hair growth over the upper lip and chin but not on sideburns and cheeks. Finally, the Ferriman-Gallwey score does not account for hair removal. Precise quantitation of hirsutism is generally necessary only in a research study.

DIAGNOSTIC EVALUATION OF MASCULINIZED WOMEN

Total testosterone level

Biologically available testosterone level

DHEAS level

7 to 9 AM 17-OH progesterone level

Prolactin level

± FSH level

DHEAS, dehydroepiandrosterone sulfate; 17-OH, 17-hydroxyprogesterone; FSH, follicle stimulating hormone.

Initial Laboratory Studies
The recommended diagnostic evaluation is summarized. In more than 80% of women with suspected hyperandrogenism, levels of the total or free testosterone or DHEAS are evaluated on a single random determination (Wild et al, 1983). Obtaining other androgen levels, such as androstenedione or 3-alpha-androstanediol glucuronide, does not improve diagnostic accuracy. If hyperandrogenism is suspected, measuring circulating androgen levels is important to confirm the diagnosis and to rule out an androgensecreting tumor. If the serum testosterone level is greater than 150 to 200 ng/dL or the DHEAS level is greater than 700 mug/dL, additional studies must be performed (Friedman et al, 1985) (see below). If a patient has symptoms of Cushing syndrome, a 24-hour urine sample should be collected for urine-free cortisol testing. In addition, other conditions that could cause these clinical symptoms should be ruled out. Specifically, prolactin levels should be measured in all women, and FSH should be measured if menopause is a possibility.

Because of the high frequency of insulin resistance and glucose intolerance, an assessment of glucose tolerance should be considered in all women with PCOS (Legro et al, 1999). A diagnosis of NCCAH does not change clinical management, except for women attempting to conceive (see later). Usually, documenting abnormal adrenal enzyme function with an ACTH stimulation test is unnecessary, and measuring the early morning 17-OH progesterone level can be used to screen for NCCAH (Azziz and Zacur, 1989). A formal ACTH stimulation test should be performed if the morning 17-OH progesterone level is greater than 300 ng/dL. In women of eastern European Jewish origin, however, routine ACTH testing may be appropriate. Ovulatory status may be assessed reliably by measuring the luteal-phase progesterone level on day 21 or 22 of the menstrual cycle. Although an elevated LH-FSH ratio supports a diagnosis of PCOS, single gonadotropin determinations are confounded by pulsatile fluctuations and are generally unnecessary.

Additional Diagnostic Studies
If the initial evaluation demonstrates that a patient's serum testosterone level is greater than 150 ng/dL or the DHEAS level is greater than 700 mug/dL, additional studies are required to rule out an ovarian or adrenal tumor (Friedman et al, 1985).

Dynamic Hormonal Testing
Dynamic hormonal testing is not useful in distinguishing adrenal from ovarian hyperandrogenism because of the considerable overlap in responses. Oral contraceptive pills suppress ovarian and adrenal androgen production and may also suppress androgen-secreting neoplasms (Givens et al, 1975; Wild et al, 1982). Long-acting GnRH agonists (GnRHa) such as Lupron and Synarel selectively suppress ovarian androgen production in the absence of a tumor (Chang et al, 1983). However, androgen-secreting neoplasms have been documented to suppress their steroid production with GnRHa administration (Parr et al, 1988). Ovarian stimulation testing with hCG is nonspecific because similar degrees of stimulation of androgen production can be seen with tumorous and nontumorous causes of ovarian hyperandrogenism. In addition, adrenal virilizing tumors occasionally stimulate their androgen production with hCG (Givens et al, 1974; Gabrilove et al, 1981), whereas ovarian androgen-secreting tumors can respond to changes in ACTH (Tucci et al, 1973).If a patient plans ovulation induction therapy or pregnancy with a partner who also may be carrying CAH genes (e.g., an eastern European Jew), ACTH stimulation testing is necessary if her ethnic background, family history, or unstimulated 17-OH progesterone level raises the probability of a diagnosis of NCCAH. However, the ACTH stimulation test does not distinguish nontumoral from tumoral adrenal entities, nor does it distinguish ovarian hyperandrogenemia from adrenal hyperandrogenemia. The test requires the administration of 250 mug cosyntropin (Cortrosyn) intravenously or intramuscularly, with blood collected for cortisol (to document the receipt of ACTH) and 17-OH progesterone (to assess adrenal 21-hydroxylase function) at 0 and 60 minutes (see Fig. 15-6) . A stimulated 17-OH progesterone level greater than 1000 ng/dL is consistent with a P450c21 gene defect (New and Speiser, 1986).

Imaging Studies
Ultrasonography is an inconsistent diagnostic tool for the evaluation of hyperandrogenism. Transabdominal ovarian ultrasonography is relatively insensitive for assessing ovarian morphology, particularly in obese women. Conversely, vaginal ultrasonography is very sensitive for detecting polycystic ovaries (Fig. 15-5) . Thus, an experienced sonographer can reliably identify polycystic ovaries using this technique (Adams et al, 1986). However, as discussed in the section on ovarian morphology in PCOS, this finding is not diagnostic of the syndrome, it can occur secondary to other causes of hyperandrogenism or chronic anovulation, and it is seen in as many as 20% of normal women (Polson et al, 1988). Finally, ovarian ultrasongraphy cannot exclude the presence of small androgen-secreting neoplasms, which may not even enlarge the ovary containing them.

Most adrenal androgen-secreting neoplasms can be detected by adrenal computerized tomography with contrast (Gabrilove et al, 1981). Computerized tomographic imaging does not accurately image the ovaries. Adrenal CT or MRI must be performed before laparotomy to exclude adrenal lesions in all patients thought to have androgen-secreting neoplasms even when there is isolated testosterone elevation. Magnetic resonance imaging and newer-generation CT are equally effective in disclosing adrenal masses as small as 0.5 cm. Adrenal adenomas are rounded and well circumscribed, whereas adrenal carcinomas usually have irregular borders.

Magnetic resonance imaging is a relatively new technique, and experience with it is limited. Its sensitivity for detecting adrenal masses is similar to that of CT, and it may provide information regarding the tissue type of a mass. It also provides better resolution of pelvic tissues than does CT. Small ovarian cysts and tumoral lesions may be demonstrated using MRI, though the exact usefulness of this technique in excluding ovarian androgen-secreting neoplasms remains to be determined. In fact, instances have been reported in which MRI failed to identify tumors found at surgery. Ovarian scintigraphic localization of tumors has been reported (Mountz et al, 1988), but this procedure requires further validation and is not widely available.

Invasive Procedures
A controversial issue in the evaluation of the masculinized woman is the role of adrenal and ovarian venous sampling. This method is only considered when results from imaging procedures have been equivocal, there is a high suspicion of an adrenal or ovarian tumor, and there is local expertise in its performance. Venous catheterization can localize an androgen-producing tumor by venography and can also delineate effluent or peripheral circulation differences in androgens (Moltz et al, 1984; Wajchenberg et al, 1986). There are several disadvantages to this technique. First, it can cause complications such as bleeding, thrombosis, venous rupture, and renal exposure to contrast dye. Second, technical difficulties can develop in catheterization and sampling of the small vessels because even in expert hands, all four ovarian and adrenal veins can be sampled in only 45% of patients (Sorenson et al, 1986). Third, episodic androgen secretion often causes spurious results. Thus, we recommend that laparotomy and ovarian exploration be performed in any woman thought to have an androgen-secreting neoplasm if adrenal imaging studies exclude an adrenal mass.

Therapy for Hyperandrogenism Mechanism of Action Comments
Ovarian Suppression
Combination oral contraceptives (OC) Avoid levonorgestrel-containing agents

GnRHa Expensive, must combine with estrogen/progestin

Adrenal Suppression
Glucocorticoids Antiandrogens more effective for clinical symptoms; reserve for infertility therapy

Androgen Receptor Antagonists
Spironolactone Recommended alone or in combination with OC

Cyproterone acetate Not available in United States

Flutamide Expensive, potentially hepatotoxic, not recommended

5 alpha -Reductase Inhibitors

Finasteride Expensive, not recommended in fertile women

Androgen Biosynthesis Inhibitors
Ketaconazole Expensive, potentially hepatotoxic, not recommended

Insulin-Lowering Agents
(Efficacy for hyperandrogenic symptoms remains unproven)

Metformin Gastrointestinal side effects

Troglitazone Expensive, potentially hepatotoxic

Other
Bromocriptine Not effective

Cimetidine Not effective

THERAPEUTIC OPTIONS
Goals of Therapy
Except in the rare occurrence of an androgen-secreting tumor, therapy for hyperandrogenic women is directed at the patient's primary symptoms and personal goals (Table 15-1) . An elevated serum androgen level is not required to initiate treatment because hyperandrogenic symptoms demonstrate that there is increased androgen use by the target tissue.

In all obese women, weight reduction can substantially improve hyperandrogenism and anovulation (Kiddy et al, 1992). Medical therapy may consist of suppressing LH or ACTH release to reduce glandular secretion interfering with androgen synthesis, increasing SHBG to decreases androgen bioavailability, or decreasing tissue androgen use by inhibiting 5alpha-reductase activity or blocking androgen action with wider spectrum androgen receptor antagonists. In addition, many cosmetic therapies are available for the symptoms of hirsutism, acne, and alopecia. Nonpharmacologic techniques are particularly useful in conjunction with pharmacologic therapy.

It is essential to ensure that the patient has realistic expectations about the time course and the results of therapy. Once a hair follicle has been transformed by androgen exposure to produce a terminal hair, biochemical control of hyperandrogenism will not result in restoration of vellus hair growth. However, biochemical control will result in a slowing of the rate of hair growth and a decrease in hair diameter and color. Clinical effects on hair growth are not evident until 3 to 6 months of therapy, and maximal effects are not seen for up to 1 year. In most women it takes approximately 1 year for the effects of therapy for alopecia to result in clinically evident changes because of the cyclic nature of scalp hair growth. Conversely, improvements in acne and seborrhea can be seen within 1 to 2 months on therapy. Eradication of terminal hairs requires electrolysis or repeated laser treatments. Because these modalities are expensive, we recommend that the patient wait until biochemical control has been achieved (3 to 6 months) before adding these therapies.

Nonpharmacologic
Techniques used to manage hirsutism include depilation with shaving, tweezers, waxes, or creams; electrocoagulation with diathermy; electrolysis with direct current; repeated bleaching; and laser therapy. Male pattern baldness may be managed surgically with full-thickness, hair-bearing punch autografts obtained from resistant occipital follicles.

In addition, because obesity can unmask or aggravate hyperandrogenic syndromes and insulin resistance associated with PCOS is often observed and has significant health consequences, weight loss should be strongly encouraged from the outset in all overweight patients. Weight loss results in decreased androgen production and insulin levels and in increased SHBG levels, occasionally leading to the complete resolution of symptoms and the resumption of ovulatory menses. As little as a 7% reduction in body weight can lead to a significant decrease in androgen levels and to the resumption of ovulatory menses in obese women with PCOS (Kiddy et al, 1992).

Endometrial Protection
Women with chronic anovulation and continued estrogen production, such as women with PCOS, are at risk for endometrial neoplasia because of the unopposed estrogen effect (Jackson and Docherty, 1957). Thus, regular endometrial shedding must be induced in amenorrheic, hyperandrogenic women with intermittent oral progestin administration ( e.g., 5 mg medroxyprogesterone acetate orally for 14 days every 1 to 2 months). The exact frequency of withdrawal bleeding to prevent endometrial hyperplasia has not been established in patients with PCOS. Intermittent progesterone withdrawal must be given, in addition to agents for controlling hyperandrogenism, if the latter do not restore regular menses.

Oral Contraceptive Pills
Estrogen and progestin are more effective in combination than either agent is alone in suppressing LH-dependent hyperandrogenemia. Although at least 50 mug ethinyl estradiol (EE) or the equivalent is required to suppress LH, lower estrogen levels (approximately 35 mug EE) can increase serum SHBG and decrease ACTH-dependent adrenal DHEAS production (Wild et al, 1983). Thus, low-dose oral contraceptive pills (30 to 35 mug EE) are usually effective in controlling hirsutism and acne (see Fig. 15-8) . Contraceptive pills also induce cyclic uterine withdrawal bleeding and provide contraception.

In women in whom estrogens are contraindicated (see later), high-dose medroxyprogesterone has been used alone with some efficacy. Progesterone inhibits 5alpha-reductase activity, which decreases target-tissue androgen use. All progestins have some degree of androgenic action. Levonorgestrel has a relative androgenic activity of 9.4 (methyltestosterone, 100) and has a higher affinity for SHBG than do endogenous steroids. Thus, levonorgestrel increases biologically available potent androgens by competing for binding sites on SHBG and should be avoided in women with PCOS. Oral contraceptive pills containing progestins with acceptable progestational activity and weaker androgenic activity, such as ethynodiol diacetate (androgen activity, 0.63), norethindrone (androgen activity, 1.60), or the newer progestin norgestimate are recommended. We avoid contraceptive pills containing desogestrel because of a potential increased risk for thromboembolism.

The major risks inherent in the use of oral contraceptive pills are related primarily to their estrogen content. Risks include thromboembolic phenomena, gallbladder disease, and exacerbation of hypertension and migraine headaches. They may cause benign hepatic tumors, but this is rare. Less serious but troublesome side effects include nausea and breast tenderness, which typically resolve after the first few months of therapy, and depression and fluid retention, which may persist. Phlebitis, hypertension, and cholelithiasis are relative contraindications to oral contraceptive use. Cigarette smoking significantly increases the risk for thromboembolic phenomena during oral contraceptive use and should be discouraged vigorously. Oral contraceptive agents have been shown to worsen insulin resistance in hyperandrogenic women (Korytkowski et al, 1995; Nader et al, 1997).

In nonsmoking patients with clinically mild to moderate hyperandrogenism who are 35 years of age or younger and who do not have contraindications, oral contraceptive pills with 35 mug EE may be used as first-line therapy. Antiandrogen therapy is usually needed if hirsutism is moderate to severe or if alopecia has developed. Only approximately one third of patients respond to the point that mechanical depilation is not needed. Patients must be informed that a full clinical response may require 6 to 12 months. It is possible to maintain a cosmetic response with oral contraceptives alone after the initial therapy in combination with antiandrogens, usually spironolactone.

Glucocorticoids
Glucocorticoids are used primarily to suppress ACTHdependent adrenal androgen secretion. They are most commonly indicated for the treatment of CAH. However, recent studies demonstrate that women with NCCAH do just as well or better with antiandrogen therapy (with or without oral contraceptives) as with glucocorticoids, probably because of the difficulty of providing sufficient medication to reduce androgen production without excess glucocorticoid effects. Thus, glucocorticoids are no longer recommended as first-line therapy for NCCAH (Spritzer et al, 1990). Some studies suggest that glucocorticoids significantly improve the ovulatory response of women with NCCAH, and short-term use can be considered in women with infertility. Fertility rates may also be improved by glucocorticoid therapy in conjunction with ovulation induction in hyperandrogenic women who do not have CAH. This may be related to decreasing putative direct negative androgen effects on the uterus, follicle, or hypothalamic-pituitary axis.

The risks and side effects of glucocorticoids range from relative hypoadrenalism with attenuated adrenal stress response (even with dosages in the range of 0.5 mg a day) to iatrogenic Cushing syndrome (with dosages in the range of 0.75 mg a day). The former may be monitored and prevented by titrating the dosage to a cortisol level of more than 2 g/dL between 8 and 9 AM. However, adrenal androgen secretion is more sensitive than cortisol secretion to low-dose glucocorticoid suppression (Rittmaster et al, 1985).

Glucocorticoid replacement, usually with 0.25 to 0.5 mg oral dexamethasone at bedtime, is the specific therapy for women with NCCAH. Clinical improvement may be expected to occur any time between 3 and 12 months after the institution of therapy. Because the clinical response correlates with biochemical control of hyperandrogenemia, if androgen levels do not decrease, this treatment should be discontinued. Dexamethasone is the glucocorticoid of choice because it is virtually devoid of mineralocorticoid activity and can be given in a single bedtime dose. However, as discussed earlier, we only recommend the use of glucocorticoids for the treatment of infertility in hyperandrogenic women.

Antiandrogens
Antiandrogens act primarily by interfering with target tissue binding of androgens to their intracellular receptor. They may be used alone or in combination with oral contraceptive pills for masculinizing signs that are not responsive to a single agent. Antiandrogens can improve insulin sensitivitiy slightly in women with PCOS (Moghetti et al, 1996b).

Spironolactone
The diuretic agent spironolactone is not approved by the Food and Drug Administration (FDA) for use in hirsutism. Nevertheless, it has potent antiandrogen action in high dosages (75 to 200 mg a day) by interfering with androgen synthesis by inhibiting 17alpha-hydroxylase/17, 20-lyase activity (P450c17) and by competitively inhibiting intracellular testosterone and DHT receptor binding. Moreover, spironolactone may inhibit 5alpha-reductase activity. When spironolactone is given in high dosages (more than 75 mg daily), there is significant displacement of testosterone from SHBG; however, its overall effect is antiandrogenic. Comparative studies have shown spironolactone to have similar efficacy (alone or in combination with oral contraceptive pills) to cyproterone acetate (O'Brien et al, 1991; Erenus et al, 1996), flutamide (Erenus et al, 1994), and finasteride (Erenus et al, 1997).

Spironolactone has been used for many years and has an excellent safety profile (Cumming et al, 1982). Side effects include dysfunctional bleeding (56%) but not electrolyte disturbances. When the former occurs, the dosage should be decreased and contraceptive pills added, or the drug stopped altogether. Headache (18%), nausea (25%), and lassitude (15%) may occur, but they usually resolve after 1 to 2 months and can be lessened with the concomitant use of oral contraceptive pills (Helfer et al, 1988). The longterm safety of spironolactone is unclear in light of a reported association with breast carcinogenesis in rodents. Spironolactone should not be given during pregnancy because of the potential to interfere with the normal masculinization of a male fetus. Its teratogenic potential is particularly important because spironolactone can lead to a resumption of ovulation in women with PCOS (Evron et al, 1981). However, there is a case report of the delivery of normal male infants in a woman who was administered high-dose spironolactone during pregnancy for Bartter syndrome (Groves and Corenblum, 1995).

A starting dosage of 50 to 100 mg twice a day is recommended (Crosby and Rittmaster, 1991). If side effects do not develop, a maximal dosage of 100 mg twice daily can be used to control hyperandrogenism. If dysfunctional uterine bleeding occurs, the daily dosage should be decreased by 25 mg. A clinical response may be expected to occur approximately 3 to 6 months after therapy is started. Women with IH and PCOS often respond well to this therapy. Because of the risk for teratogenic effects and dysfunctional bleeding, spironolactone is often prescribed in combination with an oral contraceptive pill. Because the two therapies work by different mechanisms, together they are more effective at resolving symptoms than if either therapy is used alone. Generic spironolactone is available and is a cost-effective therapy for hyperandrogenism. In the United States, it is the first-line therapy for moderate to severe hyperandrogenism.

Cyproterone Acetate
The antiandrogen cyproterone acetate is highly effective in the treatment of acne and hirsutism. It also competitively inhibits intracellular androgen receptor function, but it has only 18.6% of the activity of spironolactone. The decrease in LH-dependent androgen synthesis is attributed to the progestational activity of cyproterone and the EE that is administered with it to provide contraception and further suppress LH. This drug is not available in the United States, despite its apparent safe use in Europe. Because it is a progestin, large doses can suppress the hypothalamicpituitary-adrenal axis. In approximately 10% of patients, the drug is discontinued because it causes fatigue, weight gain, and loss of libido. A clinical response in hirsutism may be expected to occur in 50% to 90% of patients roughly 2 to 3 months after cyproterone acetate is started. The relapse rate 6 months after withdrawal of the drug ranges from 40% to 80%. Acne and seborrhea improve in approximately 60% of patients. It is also teratogenic because of its anti- androgenic actions, and it is always administered in combination with EE in premenopausal women.

Flutamide
Flutamide is a pure antiandrogen that is used primarily in the medical management of prostate cancer. Preliminary clinical trials in hirsutism using 100 mg a day have been promising. Dosages of 250 mg a day have resulted in decreased hirsutism and the resumption of ovulatory menses in adolescents with PCOS (DeLeo et al, 1998). Side effects include liver toxicity, and case reports of hepatic failure leading to transplantion or death have been reported in men and in at least one woman administered flutamide for hirsutism (Wysowski et al, 1993). Flutamide is not FDA approved for the treatment of hirsutism. Given this safety profile, flutamide should be considered only for women in whom all other therapies have failed.

Other Agents
5alpha-Reductase Inhibitors
Without binding to the androgen receptor, 5alpha-reductase inhibitors block the conversion of testosterone to DHT. However, there are at least two different isoforms of the 5alpha- reductase enzyme that are differentially distributed in androgen-sensitive tissues. For example, finasteride was developed for the treatment of prostate hypertrophy, and it was not thought to affect skin 5alpha-reductase significantly. Surprisingly, recent studies demonstrated that finasteride has an efficacy similar to that of 100 mg/day spironolactone for the treatment of hirsutism (Wong et al, 1995; Erenus et al, 1997). These agents are also teratogenic because they cross the placenta and inhibit fetal 5alpha-reductase activity, resulting in ambiguous genitalia in a male fetus. These agents are not FDA approved for the treatment of hirsutism.

Ketoconazole
Technically, the antifungal agent ketoconazole is not an antiandrogen because it blocks the synthesis of andro- gens through the suppression of P450-dependent adrenal and gonadal enzymes such as P450c17. This results in a reduction in serum testosterone levels, whereas 17- hydroxyprogesterone levels increase. Ketoconazole is not FDA approved for the treatment of hirsutism. However, in a study of nine hirsute women treated with 400 to 1200 mg/day, clinical improvement was apparent at 6 months (Martikainen et al, 1988). Because serum DHEAS and cortisol levels did not change, the primary effect was on ovarian androgen synthesis. Low-dose therapy (400 mg a day) is recommended after a 1-month induction period on 1200 mg a day. Untoward effects include nausea, pruritus, and hepatocellular dysfunction. Periodic liver function testing should be performed and the drug discontinued if liver enzyme increases are detected. This drug is not used as the initial choice in the management of women with masculinization, but it may be considered in resistant patients in whom all other measures have failed.

Long-Acting Gonadotropin-Releasing Hormone Analogues
Long-acting GnRHa decrease LH secretion (and, therefore, LH-dependent androgen secretion) by producing gonado- tropin desensitization. GnRHa may be used in women in whom primary therapy with oral contraceptive pills or antiandrogens fails. When nafarelin acetate was administered to six hirsute women as a nasal spray at a dosage of 500 mug twice daily, biochemical control occurred in four of the women within 1 to 3 months (Andreyko et al, 1986). This drug also may be useful for women with IH because it decreases androgen delivery to target tissue. Indeed, it was as effective as high-dose cyproterone acetate in one study (Carmina and Lobo, 1997). An important side effect of these agents is osteoporosis because estrogen levels are lower, as are androgen levels. The major limitation is that GnRHa treatment is expensive; it can cost up to $400 per month. The FDA has not approved GnRHa for the treatment of hyperandrogenism. Osteoporosis may be minimized by supplementation with a low-dose estrogen preparation and periodic progesterone withdrawal (Rittmaster, 1995).

Cimetidine
Cimetidine is a weak androgen receptor antagonist (Eil and Edelson, 1984). A controlled clinical study has not found cimetidine to be effective in the treatment of hyperandrog- enism (Lissack et al, 1989).

Insulin-Lowering Agents
Troglitazone and metformin lower biologically available testosterone levels by approximately 25% in women with PCOS (Velazquez et al, 1994; Dunaif et al, 1996; Nestler and Jacubowicz, 1996; Ehrmann et al, 1997b). This testosterone lowering correlates with the reduction in insulin levels (Dunaif, 1996) and is not seen without them (Ehrmann, 1997a). It is predicted that this improvement in hyperandrogenemia will be associated with an improvement in hirsutism, but this remains to be documented in more prolonged trials. Resumption of ovulation has been reported in women receiving these agents alone (Dunaif, 1996; Nestler et al, 1998) or in combination with clomiphene citrate (Nestler et al, 1998). Metformin is not consistently effective at lowering insulin levels in PCOS (Crave et al, 1995; Ehrmann et al, 1997a). Only two studies have been conducted on troglitazone in PCOS (Dunaif et al, 1996; Ehrmann et al, 1997b), and to date there have been no comparative studies of efficacy. Metformin can cause lactic acidosis in persons with compromised renal function, and it commonly causes gastrointestinal side effects such as nausea and diarrhea, but these usually do not lead to discontinuation of the drug among patients with PCOS. Troglitazone is well tolerated, but it has been associated with several cases of fatal hepatic necrosis. It is unknown whether this hepatic failure can be predicted by surveillance of liver function tests. Neither of these agents is FDA approved for use in hyperandrogenism.

Ovarian Surgery
Ovarian wedge resection results in transient decreases in androgen levels (Katz et al, 1978). Thus, this procedure is used only as a last resort in infertile women and not for control of hyperandrogenism. Oophorectomy may be considered to manage women with tumors or for those in whom pharmacologic therapy has failed or is contraindicated.

Bromocriptine
Placebo-controlled trials have shown that bromocriptine is no more effective than placebo in controlling hyperandrogenism (Buvat et al, 1986).

Thyroid Hormone Therapy
Thyroid hormones should be given only to women with documented hypothyroidism. The attendant risk for complication limits the use of empiric thyroid hormone therapy designed to promote weight loss or to increase SHBG.

SUMMARY
Hyperandrogenic disorders are common problems in women and are usually secondary to PCOS or IH. These conditions are straightforward to diagnose, and rare serious disorders can be easily excluded. Most of the clinical symptoms can be controlled with therapy. It is important to recognize that women with PCOS are at increased risk for glucose intolerance and cardiovascular disease.

REFERENCES

Abraham GE: Ovarian and adrenal contribution to peripheral androgens during the menstrual cycle. J Clin Endocrinol Metab 1976; 39:340-346.

Adams J, Polson DW, Franks S: Prevalence of polycystic ovaries in women with anovulation and idiopathic hirsutism. BMJ 1986; 293: 355-359.

Aiman J, Fomey JP, Parker CR: Secretion of androgens and estrogens by normal and neoplastic ovaries in postmenopausal women. Obstet Gynecol 1986; 68:1-5.

Andersen P, Seljeflot I, Abdelnoor M, et al: Increased insulin sensitivity and fibrinolytic capacity after dietary intervention in obese women with polycystic ovary syndrome. Metabolism 1995; 44:611-616.

Anderson DC: Sex hormone-binding globulin. Clin Endocrinol (Oxf) 1974; 3:69-96.

Andreyko JL, Monroe SE, Jaffe RB: Treatment of hirsutism with a gonadotropin-releasing hormone agonist (Nafarelin). J Clin Endocrinol Metab 1986; 63:854-859.

Arroyo A, Laughlin GA, Morales AJ, Yen SSC: Inappropriate gonadotropin secretion in polycystic ovary syndrome: Influence of adiposity. J Clin Endocrinol Metab 1997; 82: 3728-3733.

Azziz R, Zacur HA: 21-Hydroxylase deficiency in female hyperandrogenism: Screening and diagnosis. J Clin Endocrinol Metab 1989; 69:577-584.

Baird DT, Corker CS, Davidson DW, et al: Pituitary-ovarian relationships in polycystic ovary syndrome. J Clin Endocrinol Metab 1977; 45:798-809.

Balen A, Tan S, MacDougall J, Jacobs H: Miscarriage rates following in-vitro fertilization are increased in women with polycystic ovaries and reduced by pituitary desensitization with buserelin. Hum Reprod 1993; 8: 959-964.

Barbieri RL, Makris A, Randall RW, et al: Insulin stimulates androgen accumulation in incubations of ovarian stroma obtained from women with hyperandrogenism. J Clin Endocrinol Metab 1986; 62:904-910.

Bardin CW, Lipsett M: Testosterone and androstenedione production rates in normal women and in idiopathic hirsutism and polycystic ovary syndrome. J Clin Invest 1967; 46:891-902.

Barnes RB, Rosenfield RL, Burstein S, Ehrmann DA: Pituitary-ovarian responses to Nafarelin testing in the polycystic ovary syndrome. N Engl J Med 1989; 320:559-565.

Birdsall MA, Farquhar CM, White HD: Association between polycystic ovaries and extent of coronary artery disease in women having cardiac catheterization. Ann Intern Med 1997; 126:32-35.

Blankstein J, Rabinovici J, Goldenberg M, et al: Changing pituitary reactivity to follicle-stimulating hormone and luteinizing hormonereleasing hormone after induced ovulatory cycles and after anovulation in patients with polycystic ovarian disease. J Clin Endocrinol Metab 1987; 65:1164-1167.

Buvat J, Buvat-Herbaut M, Marrolin G, et al: A double-blind controlled study of the hormonal and clinical effects of bromocriptine in the polycystic ovary syndrome. J Clin Endocrinol Metab 1986; 63:119-124.

Carmina E, Koyama T, Chang L, et al: Does ethnicity influence the prevalence of adrenal hyperandrogenism and insulin resistance in polycystic ovary syndrome? Am J Obstet Gynecol 1992; 167: 1807-1812.

Carmina E, Lobo RA: Gonadotrophin-releasing hormone agonist therapy for hirsutism is as effective as high dose cyproterone acetate but results in a longer remission. Hum Reprod 1997; 12:663-666.

Carter JN, Tyson JE, Wayne GL, et al: Adrenocortical function in hyperprolactinemic women. J Clin Endocrinol Metab 1977; 45: 973-980.

Chang RJ, Lauger LR, Meldrum DR, et al: Steroid secretion in polycystic ovarian disease after ovarian suppression by a long-acting gonadotropin releasing hormone agonist. J Clin Endocrinol Metab 1983; 56:897-903.

Chang RJ, Mandel FP, Lu JKH, et al: Enhanced disparity of gonadotropin secretion by estrone in women with polycystic ovarian disease. J Clin Endocrinol Metab 1982; 54:490-494.

Christman GM, Randolph JF, Kelch RP, et al: Reduction of gonadotropin-releasing hormone pulse frequency is associated with subsequent selective follicle-stimulating hormone secretion in women with polycystic ovarian disease. J Clin Endocrinol Metab 1991; 72:1278-1285.

Crave J, Fimbel S, Lejeune H, et al: Effects of diet and metformin administration on sex hormone- binding globulin, androgens, and insulin in hirsute and obese women. J Clin Endocrinol Metab 1995; 80:2057-2062.

Cronin L, Guyatt G, Griffith L, et al: Development of a health-related quality of life questionnaire for patients with polycystic ovary syndrome. J Clin Endocrinol Metab 1998; 83:1976-1987.

Crosby PDA, Rittmaster RS: Predictors of clinical response in hirsute women treated with spironolactone. Fertil Steril 1991; 55:1076-1081.

Cumming DC, Yang JC, Rebar RW, et al: Treatment of hirsutism with spironolactone. JAMA 1982; 247:1295-1298.

Dahlgren E, Janson PO, Johansson S, et al: Polycystic ovary syndrome and risk for myocardial infarction. Acta Obstet Gynecol Scand 1992a; 71:599-603.

Dahlgren E, Johansson S, Lindstedt G, et al: Women with polycystic ovary syndrome wedge resected in 1956 to 1965: A long-term follow-up focusing on natural history and circulating hormones. Fertil Steril 1992b; 57:505-513.

Daniels TL, Berga SL: Resistance of gonadotropin releasing hormone drive to sex steroid-induced suppression in hyperandrogenic anovulation. J Clin Endocrinol Metab 1997; 82: 4179-4183.

DeLeo V, Lanzetta D, D'Antona D, et al: Hormonal effects of flutamide in young women with polycystic ovary syndrome. J Clin Endocrinol Metab 1998; 83:99-102.

Derksen J, Nagesser SK, Meinders AE, et al: Identification of virilizing adrenal tumors in hirsute women. N Engl J Med 1994; 331: 968-973.

Dunaif A: Do androgens directly regulate gonadotropin secretion in the polycystic ovary syndrome? J Clin Endocrinol Metab 1986; 63:215-221.

Dunaif A: Insulin resistance and the polycystic ovary syndrome: Mechanisms and implications for pathogenesis. Endocr Rev 1997; 18:774-800.

Dunaif A, Graf M: Insulin administration alters gonadal steroid metabolism independent of changes in gonadotropin secretion in insulin-resistant women with the polycystic ovary syndrome. J Clin Invest 1989b; 83:23-29.

Dunaif A, Graf M, Mandeli J, et al: Characterization of groups of hyperandrogenic women with acanthosis nigricans, impaired glucose tolerance, and/or hyperinsulinemia. J Clin Endocrinol Metab 1987; 65:499-507.

Dunaif A, Green G, Futterweit W, et al: Suppression of hyperandrogenism does not improve peripheral or hepatic insulin resistance in the polycystic ovary syndrome. J Clin Endocrinol Metab 1990; 70:699-704.

Dunaif A, Green G, Phelps RG, et al: Acanthosis nigricans, insulin action, and hyperinsulinemia: Clinical, histological, and biochemical findings. J Clin Endocrinol Metab 1991; 73:590-595.

Dunaif A, Hoffman AR, Scully RE, et al: Clinical, biochemical, and ovarian morphologic features in women with acanthosis nigricans and masculinization. Obstet Gynecol 1985; 66:545-552.

Dunaif A, Longcope C, Canick J, et al: The effects of the aromatase inhibitor Deltal-testolactone on gonadotropin release and steroid metabolism in polycystic ovarian disease. J Clin Endocrinol Metab 1984a; 60:773-780.

Dunaif A, Mandeli J, Fluhr H, et al: The impact of obesity and chronic hyperinsulinemia on gonadotropin release and gonadal steroid secretion in the polycystic ovary syndrome. J Clin Endocrinol Metab 1988; 66:131-139.

Dunaif A, Scott D, Finegood D, et al: The insulin sensitizing agent troglitazone: A novel therapy for the polycystic ovary syndrome. J Clin Endocrinol Metab 1996; 81:3299-3306.

Dunaif A, Scully RE, Andersen RN, et al: The effects of continuous androgen secretion on the hypothalamic-pituitary axis in women: Evidence from a luteinized thecoma of the ovary. J Clin Endocrinol Metab 1984b; 59:389-393.

Dunaif A, Segal KR, Futterweit W, et al: Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes 1989a; 38:1165-1174.

Ehrhardt AA, Evers K, Money J: Influence of androgen and some aspects of sexually dimorphic behavior in women with the late-treated adrenogenital syndrome. Johns Hopkins Med J 1968; 123:115-122.

Ehrmann DA, Barnes RB, Rosenfield RL: Polycystic ovary syndrome as a form of functional ovarian hyperandrogenism due to dysregulation of androgen secretion. Endocr Rev 1995; 16:322-353.

Ehrmann DA, Cavaghan MK, Imperial J, et al: Effects of metformin on insulin secretion, insulin action, and ovarian steroidogenesis in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1997a; 82:524-530.

Ehrmann DA, Schneider DJ, Sobel BE, et al: Troglitazone improves defects in insulin action, insulin secretion, ovarian steroidogenesis, and fibrinolysis in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1997b; 82:2108-2116.

Eil C, Edelson SK: The use of human skin fibroblasts to obtain potency estimates of drug binding to androgen receptors. J Clin Endocrinol Metab 1984; 59:51-55.

Erenus M, Gurbuz O, Durmosoglu F, et al: Comparison of the efficacy of spironolactone versus flutamide in the treatment of hirsutism. Fertil Steril 1994; 61:613-616.

Erenus M, Yucelten D, Durmosoglu F, et al: Comparison of finasteride versus spironolactone in the treatment of idiopathic hirsutism. Fertil Steril 1997; 68:1000-1003.

Erenus M, Yucelten D, Gurbuz O, et al: Comparison of spironolactone-oral contraceptive versus cyproterone acetate-estrogen regimens in the treatment of hirsutism. Fertil Steril 1996; 66:216-219.

Erickson GF, Magoffin DA, Dyer CA, et al: The ovarian androgen producing cells: A review of structure/function relationships. Endocr Rev 1985; 6:371-399.

Evans DJ, Hoffman RG, Kalkhoff RK, et al: Relationship of androgenic activity to body fat topography, fat cell morphology, and metabolic aberrations in premenopausal women. J Clin Endocrinol Metab 1983; 57:304-310.

Evron S, Shapiro G, Diamant YZ: Induction of ovulation with spironolactone (Aldactone) in anovulatory oligomenorrheic and hyperandrogenic women. Fertil Steril 1981; 36:468-471.

Ferriman D, Gallwey JD: Clinical assessment of body hair growth in women. J Clin Endocrinol Metab 1961; 21:1440-1447.

Franks S: Polycystic ovary syndrome: A changing perspective. N Engl J Med 1995; 333:853-861.

Franks S, Gharani N, Waterworth D, et al: The genetic basis of polycystic ovary syndrome. Hum Reprod 1997; 12:2641-2648.

Friedman CI, Schmidt GE, Kim MH, Powell J: Serum testosterone concentrations in the evaluation of androgen-producing tumors. Am J Obstet Gynecol 1985; 153:44-49.

Futterweit W, Deligdisch L: Histopathological effects of exogenously administered testosterone in 19 female to male transsexuals. J Clin Endocrinol Metab 1985; 62:16-21.

Futterweit W, Dunaif A, Yeh H-C, et al: The prevalence of hyperandrogenism in 109 consecutive female patients with diffuse alopecia. J Am Acad Dermatol 1988; 19:831-836.

Gabrilove JL, Seman AT, Sabet R, et al: Virilizing adrenal adenoma with studies on the steroid content of the adrenal venous effluent and a review of the literature. Endocr Rev 1981; 2:462-470.

Geist SH, Gaines JA: Diffuse luteinization of the ovaries associated with the masculinization syndrome. Am J Obstet Gynecol 1942; 43:975-983.

Gilling-Smith C, Willis DS, Beard RW, et al: Hypersecretion of androstenedione by isolated thecal cells from polycystic ovaries. J Clin Endocrinol Metab 1994; 79:1158-1165.

Givens JR, Andersen RN, Wiser WL, et al: A gonadotropin-responsive adrenocortical adenoma. J Clin Endocrinol Metab 1974; 38: 126-133.

Givens JR, Andersen RN, Wiser WL, et al: A testosterone secreting, gonadotropin responsive pure thecoma and polycystic ovarian disease. J Clin Endocrinol Metab 1975; 41:845-853.

Glass AR, Dahms WT, Abraham G, et al: Secondary amenorrhea in obesity: Etiologic role of weight-related androgen excess. Fertil Steril 1978; 30(2):243-244.

Glickman SP, Rosenfeild RL, Bergenstal RM, et al: Multiple androgenic abnormalities, including elevated free testosterone, in hyperprolactinemic women. J Clin Endocrinol Metab 1982; 55:251-257.

Gordon GG, Southern AL, Tochimoto S, et al: Effect of hyperthyroidism and hypothyroidism on the metabolism of testosterone and androstenedione in man. J Clin Endocrinol 1969; 29(2):164-170.

Graf M, Brown V, Richards C, et al: The independent effects of hyperandrogenemia, hyperinsulinemia, and obesity on lipid and lipoprotein profiles in women. Clin Endocrinol (Oxf) 1990; 33:119-131.

Groves TD, Corenblum B: Spironolactone therapy during human pregnancy. Am J Obstet Gynecol 1995; 172:1655.

Hague WM, Adams J, Rodda C, et al: The prevalence of polycystic ovaries in patients with congenital adrenal hyperplasia and their close relatives. Clin Endocrinol 1990; 33: 501-510.

Hammerstein J, Meckies J, Leo-Rossberg I, et al: Use of cyproterone acetate in the treatment of acne, hirsutism and virilism. J Steroid Biochem 1975; 6:827-836.

Hayes FJ, Taylor AE, Martin KA, et al: Use of a gonadotropin-releasing hormone antagonist as a physiologic probe in polycystic ovary syndrome: Assessment of neuroendocrine and androgendynamics. J Clin Endocrinol Metab 1998; 83: 2342-2349.

Helfer EL, Miller JL, Rose LI: Side-effects of spironolactone therapy in the hirsute woman. J Clin Endocrinol Metab 1988; 66:208.

Hochberg Z, Gardos M, Benderly A: Psychosexual outcome of assigned females and males with 46,XX virilizing congenital adrenal hyperplasia. Eur J Pediatr 1987; 146:497-499.

Horton R, Hawks D, Lobo R: 3-Alpha, 17-beta-androstanediol glucuronide in plasma--a marker of androgen action in idiopathic hirsutism. J Clin Invest 1982; 69:1203-1206.

Hughesdon PE: Morphology and morphogenesis of the Stein-Leventhal ovary and of so-called 'hyperthecosis.' Obstet Gynecol 1982; 37: 59-77.

Hull MGR: Epidemiology of infertility and polycystic ovarian disease: Endocrinological and demographic studies. Gynecol Endocrinol 1987; 1:235-245.

Irvine RA, Yu MC, Ross RK, Coetzee GA: The CAG and GGC micro-satellites of the androgen receptor gene are in linkage disequlibrium in men with prostate cancer. Cancer Res 1994;54:2861-2864.

Ito T, Horton R: The source of plasma dihydrotestosterone in man. J Clin Invest 1971; 50:162-1627.

Jackson RL, Docherty MD: The Stein-Leventhal syndrome analysis of 45 cases with special reference to association with endometrial carcinoma. Am J Obstet Gynecol 1957; 73:161-173.

Katz M, Carr PJ, Cohen BM, et al: Hormonal effects of wedge resection of polycystic ovaries. Obstet Gynecol 1978; 51:437-444.

Kiddy DS, Hamilton-Fairley D, Bush A, et al: Improvement in endocrine and ovarian function during dietary treatment of obese women with polycystic ovary syndrome. Clin Endocrinol (Oxf) 1992; 36:105-111.

Knochenhauer ES, Key TJ, Kahsar-Miller M, et al: Prevalence of the polycystic ovary syndrome in unselected black and white women of the southeastern United States: A prospective study. J Clin Endocrinol Metab 1998; 83:3078-3082.

Korytkowski MT, Mokan M, Horwitz MJ, et al: Metabolic effects of oral contraceptives in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1995; 80:3327-3334.

Kuttenn F, Mowszowicz I, Schaison G, et al: Androgen production and skin metabolism in hirsutism. J Endocrinol 1977; 75:83-91.

Lambert-Messerlian GM, Hall JE, Sluss PE, et al: Relatively low levels of dimeric inhibin circulate in men and women with polycystic ovary syndrome using a specific two-site enzyme-linked immunosorbent assay. J Clin Endocrinol Metab 1994; 79:45-50.

Legro RS, Driscoll D, Strauss JF, et al: Evidence for a genetic basis for hyperandrogenemia in polycystic ovary syndrome. Proc Natl Acad Sci USA 1998a; 95:14956-14960.

Legro RS, Kunselman AR, Dodson WC, et al: Prevalence and predictors of risk for type 2 diabetes mellitus and impaired glucose tolerance in polycystic ovary syndrome: A prospective, controlled study in 254 affected women. J Clin Endocrinol Metab 1999;84:165-169.

Legro RS, Spielman R, Urbanek M, et al: Phenotype and genotype in polycystic ovary syndrome. Recent Prog Horm Res 1998b; 53:217-256.

Lissak A, Sorokin Y, Calderon I, et al: Treatment of hirsutism with cimetidine: A prospective randomized controlled trial. Fertil Steril 1989; 51:247-250.

Lobo R: 'Idiopathic hirsutism'--fact or fiction? Semin Reprod Endocrinol 1986; 4:179-187.

Lobo RA, Goebelsmann U: Adult manifestation of congenital adrenal hyperplasia due to incomplete 21-hydroxylase deficiency mimicking polycystic ovarian disease. Am J Obstet Gynecol 1980; 138:720-726.

Lobo RA, Goebelsmann U, Horton R: Evidence for the importance of peripheral tissue events in the development of hirsutism in polycystic ovary syndrome. J Clin Endocrinol Metab 1983; 57:393-397.

Lobo RA, Grange L, Goebelsmann U, et al: Elevations in unbound serum estradiol as a possible mechanism for inappropriate gonadotropin secretion in women with PCOS. J Clin Endocrinol Metab 1981; 52:156-158.

Longcope C: Adrenal and gonadal androgen secretion in normal females. J Clin Endocrinol Metab 1986; 15:213-228.

Lucky AW, McGuire J, Rosenfield RL, et al: Plasma androgens in women with acne vulgaris. J Invest Dermatol 1983; 81:70-74.

Lucky AW, Rosenfeld RL, McGuire J, et al: Adrenal androgen hyperresponsiveness to adrenocorticotropin in women with acne and/or hirsutism: Adrenal enzyme defects and exaggerated adrenarche. J Clin Endocrinol Metab 1986; 62: 840-848.

Martikainen H, Heikkinen J, Ruokonen A, et al: Hormonal and clinical effects of ketoconazole in hirsute women. J Clin Endocrinol Metab 1988; 66:987-991.

Mason HD, Willis DS, Beard RW, et al: Estradiol production by granulosa cells of normal polycystic ovaries: Relationship to menstrual cycle history and concentrations of gonadotropins and sex steroids in follicular fluid. J Clin Endocrinol Metab 1994; 79:1355-1360.

Mattox JH, Phelan S: The evaluation of adult females with testosterone producing neoplasms of the adrenal cortex. Surg Gynecol Obstet 1987; 164:98-101.

McKenna TJ, Moore A, Magee F, et al: Amenorrhea with cryptic hyperandrogenemia. J Clin Endocrinol Metab 1983; 56:893-896.

Mechanick J, Dunaif A: Masculinization: A clinical approach to the diagnosis and treatment of hyperandrogenic women. In Mazzaferri E, editor: Advances in endocrinology and metabolism, St.Louis, Mosby Year Book.

Meldrum DR, Abraham GE: Peripheral and ovarian nevous concentrations of various steroid hormones in virilizing ovarian tumors. Obstet Gynecol 1979; 53: 36-43.

Moghetti P, Castello R, Negri C, et al: Insulin infusion amplifies 17alpha-hydroxycorticosteroid intermediate response to adrenocorticotropin in hyperandrogenic women: Apparent relative impairment of 17,20-lyase activity. J Clin Endocrinol Metab 1996a; 81:881-886.

Moghetti P, Tosi F, Castello R, et al: The insulin resistance in women with hyperandrogenism is partially reversed by antiandrogen treatment: Evidence that androgens impair insulin action in women. J Clin Endocrinol Metab 1996b; 81:952-960.

Moltz L, Schwartz U, Sorensen R, et al: Ovarian and adrenal vein steroids in patients with nonneoplastic hyperandrogenism: Selective cath- eterization flndings. Fertil Steril 1984; 42: 69-75.

Mountz JM, Gross MD, Shapiro B, et al: Scintigraphic localization of ovarian dysfunction. J Nucl Med 1988; 29: 1644-1650.

Mowszowicz I, Melinatou E, Doukani A, et al: Androgen binding capacity and 5-alphareductase activity in pubic skin fibroblasts from hirsute patients. J Clin Endocrinol Metab 1983; 56:1209-1213.

Nader S, Riad-Gabriel MG, Saad MF: The effect of a desogestrel-containing oral contraceptive on glucose tolerance and leptin concentrations in hyperandrogenic women. J Clin Endocrinol Metab 1997; 82: 3074-3077.

Nagamani M, Hannigan EV, Kinh TV, et al: Hyperinsulinemia and stromal luteinization of the ovaries in postmenopausal women with endometrial cancer. J Clin Endocrinol Metab 1988; 67:144-153.

Nestler JE, Barlascini CO, Matt DW, et al: Suppression of serum insulin by diazoxide reduces serum testosterone levels in obese women with polycystic ovary syndrome. J Clin Endocrinol Metab 1989; 68:1027-1032.

Nestler JE, Jacubowicz DJ, Evans WS, et al: Effects of metformin on spontaneous and clomiphene- induced ovulation in the polycystic ovary syndrome. N Engl J Med 1998; 338: 1876-1880.

Nestler JE, Jacubowicz DJ: Decreases in ovarian cytochrome P450c17 activity and serum free testosterone after reduction of insulin secretion in polycystic ovary syndrome. N Engl J Med 1996; 335: 617-623.

Nestler JE, Powers LP, Matt DW, et al: A direct effect of hyperinsulinemia on serum sex hormone- binding globulin levels in obese women with polycystic ovary syndrome. J Clin Endocrinol Metab 1991; 72: 83-89.

New Ml, Speiser PW: Genetics of adrenal steroid 21-hydroxylase deficiency. Endocr Rev 1986; 7:331-349.

O'Brien RC, Cooper ME, Murray RML, et al: Comparison of sequential cyproterone acetate/estrogen versus spironolactone/oral contraceptive in the treatment of hirsutism. J Clin Endocrinol Metab 1991; 72:1008-1013.

Pang S, Lerner A, Stoner E, et al: Late-onset adrenal steroid 3-beta-hydroxysteroid dehydrogenase deficiency, I: A cause of hirsutism in pubertal and postpubertal women. J Clin Endocrinol Metab 1985; 60(3):428-439.

Paradisi R, Grossi G, Venturoli S, et al: Evidence for a hypothalamic alteration of catecholamine metabolism in polycystic ovarian syndrome. Clin Endocrinol (Oxf) 1988; 29:317-326.

Pardridge WM: Transport of protein-bound hormones into tissues in vivo. Endocr Rev 1981; 2: 103-123.

Parr JH, Abraham RR, Seed M, et al: The treatment of a hyperandrogenic and virilizing state in an elderly female with a synthetic LHRH agonist. J Endocrinol Invest 1988; 11:433-436.

Pastor CL, Griffin-Korf ML, Aloi JA, et al: Polycystic ovary syndrome: Evidence for reduced sensitivity of the gonadotropin-releasing hormone pulse generator to inhibition by estradiol and progesterone. J Clin Endocrinol Metab 1998; 83:582-590.

Plymate SR, Matej LA, Jones RE, et al: Inhibition of sex hormone-binding globulin production in the human hepatoma (Hep G2) cell line by insulin and prolactin. J Clin Endocrinol Metab 1988; 67:460-464.

Polson DW, Wadsworth J, Adams J, Frank S: Polycystic ovaries: A common finding in normal women. Lancet 1988; 1:870-872.

Rebar R, Judd HL, Yen SSC, et al: Characterization of the inappropriate gonadotropin secretion in polycystic ovary syndrome. J Clin Invest 1976; 57:1320-1329.

Rittmaster R, Loriaux DL, Culter GB Jr: Sensitivity of cortisol and adrenal androgens to dexamethasone s