Comprehensive Gynecology ten currently recognized causes

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.


There are 10 currently recognized causes of androgen excess in women. One frequent cause is administration of androgenic medication. In addition to testosterone itself, various anabolic steroids, 19-norprogestins, and danazol have androgenic effects. Thus a careful history of medication intake is important for all women with hirsutism.

Hirsutism or virilization can also be associated with some forms of abnormal gonad development. With this cause, individuals have signs of either external sexual ambiguity or primary amenorrhea, in addition to findings of androgen excess and a Y chromosome present in the gonad. These conditions are discussed in Chapter 38 (Primary and Secondary Amenorrhea and Precocious Puberty) and will not be further described in the discussion of the differential diagnosis of androgen excess in this chapter.

Signs of androgen excess during pregnancy can be caused by increased ovarian testosterone production. This is usually caused by either a luteoma of pregnancy or hyperreactio luteinalis. The former is a unilateral or bilateral solid ovarian enlargement, whereas the latter is bilateral cystic ovarian enlargement. After pregnancy is completed, the excessive ovarian androgenic production resolves spontaneously and the androgenic signs regress.

A diagnosis of these three causes of androgen excess can usually be easily made by means of a careful history and physical examination. The remaining causes of androgen excess, together with the origin of hyperandrogenism, are listed in Table 40-5. Details of each of these causes will be described in decreasing order of their frequency.

Table 40-5 --Differential Diagnosis of Hirsutism and Virilization[*]





Abnormal gonadal or sexual development


Androgen excess in pregnancy: luteoma or hyperreactio luteinalis


Idiopathic hirsutism


Polycystic ovary syndrome

Stromal hyperthecosis

Ovarian tumors

Adrenal gland

Adrenal tumors

Cushing syndrome

Adult-onset congenital adrenal hyperplasia


Idiopathic hirsutism and polycystic ovary syndrome do not present with virilizations.

Idiopathic Hirsutism (Peripheral Disorder of Androgen Metabolism)

“Idiopathic” hirsutism is manifested by signs of hirsutism and regular menstrual cycles in conjunction with normal circulatory levels of androgens (both testosterone and DHEA-S). Because this type of disorder is frequently present in several individuals in the same family, particularly those of Mediterranean descent, it has also been called familial, or constitutional, hirsutism. Since neither ovarian nor adrenal androgen production is increased in these individuals, the cause of the androgen excess was not determined until recently, hence the term idiopathic hirsutism. This is a very common cause of hirsutism and is second in frequency only to polycystic ovary syndrome (PCOS). We have shown that about 80% of these individuals have increased levels of 3αdiol-G, indirectly indicating that the cause of hirsutism is increased 5αreductase activity (5αRA) (Fig. 40-5). Also we have directly measured the percent conversion of testosterone to DHT in genital skin as an assessment of 5αRA in the skin of women with idiopathic hirsutism. The amount of 5αRA was increased in hirsute women as compared with normal women and correlated well with both the degree of hirsutism and serum levels of 3αdiol-G. Thus idiopathic hirsutism is actually a disorder of peripheral androgen metabolism in the pilosebaceous apparatus of the skin and is possibly genetically determined although it is also possible that early exposure to androgens can “program” increased 5αRA. Antiandrogens that block peripheral testosterone action or interfere with 5αRA are effective therapeutic agents for this disorder. We have shown that hirsutism is largely a disorder of the peripheral compartment, and the most effective treatments involve peripheral blockade (described below).

Figure 40-5Serum 3αdiol-G in premenopausal nonhirsute women (Pre), hirsute women, normal men, and postmenopausal nonhirsute women (Post). The asterisks denote p < 0.05, as compared with Pre.

(Reprinted from Fertility and Sterility, 42, Paulson RJ, Serafini PC, Catalino JA, Lobo RA, Measurements of 3α,17β-androstanediol glucuronide in serum and urine and the correlation with skin 5αreductase activity, 422. Copyright 1986, with permission from The American Society for Reproductive Medicine.)

Polycystic Ovary Syndrome

Polycystic ovary syndrome was originally described in 1935 by Stein and Leventhal as a syndrome consisting of amenorrhea, hirsutism, and obesity in association with enlarged polycystic ovaries. The classic definition of PCOS includes women who are anovulatory and have irregular periods as well as hyperandrogenism (as determined by signs such as hirsutism or elevated blood levels of androgens: testosterone or DHEA-S. This should be in the absence of enzymatic disorders (such as 21-hydroxylase deficiency) or tumors.

This diagnosis does not require findings on ultrasound (US) of characteristic polycystic ovaries. For the past 15 years, this non-US-based definition has been referred to as the “NIH consensus” definition because it followed an NIH conference in 1989. However this was not a consensus conference and there was no true consensus among attendees.

Over time there was been increasing evidence that some women with all the features of PCOS may be ovulatory and have regular menstrual cycles. Also there has been renewed emphasis placed on finding polycystic ovaries on US. Accordingly, a conference in Rotterdam has come up with a new definition, which was published simultaneously in the journals Human Reproduction and Fertility and Sterility in 2004. This definition places an emphasis on finding polycystic ovaries on US as an important criterion. Other significant findings are the classic features of anovulation/menstrual irregularity and hyperandrogenism. Women with PCOS may have all three findings, but it requires any two out of the three to make a diagnosis of PCOS. Thus, hyperandrogenic women with normal ovulatory cycles and polycystic ovaries on US may be diagnosed as having PCOS. We have found that approximately 95% of women who have classic symptoms (NIH criteria) of anovulation and hyperandrogenism have polycystic ovaries on US. Therefore it may be unnecessary to “confirm” the diagnosis by US in this setting. Figures 40-6 and 40-7 show the classic appearing polycystic ovary at surgery in a sagittal section, which is nearly identical to what is seen by US in a sagittal plane. The US diagnosis of polycystic ovaries has been made on the basis of finding enlarged ovaries (>10 cm3) and the presence of 10 or more peripherally oriented cystic structures (2 to 8 mm) surrounding a dense stroma. Since the Rotterdam conference, however, ovarian size alone has been suggested to be sufficient, with ovaries larger than 10 cm3 being diagnostic, and recent evidence suggesting that this lower limit may be 7 cm3.

Figure 40-6Gross characteristics of polycystic ovaries. Bilateral enlarged ovaries with smooth and thickened capsule.

(From Yen SSC: Chronic anovulation caused by peripheral endocrine disorders. In Yen SSC, Jaffe RB [eds]: Reproductive Endocrinology, 2nd ed. Philadelphia, WB Saunders, 1986.)

Figure 40-7Sagittal section of a polycystic ovary illustrating large number of follicular cysts and thickened stroma.

It is also important to note that anywhere from 10% to 25% of the normal reproductive-age population (no symptoms or signs of PCOS) may have polycystic ovaries found on US. These ovaries have been called “polycystic-appearing ovaries” (PAO) or “polycystic ovarian morphology” (PCOM) in the literature. This isolated finding should not be confused with the diagnosis of PCOS, but may be a risk factor for other traits of PCOS (insulin resistance, cardiovascular risk factors) discussed later.

Using the classic definition of PCOS, which can also be quite heterogeneous in terms of the severity of the findings of menstrual irregularity and hyperandrogenism, approximately 3% to 7% of the reproductive-age population will have PCOS. Thus PCOS is an extremely common disorder, and its diagnosis is important because of its consequences (see section on consequences of PCOS).

Depending on the women' country of origin, 30% to 70% of women with PCOS are overweight, although it's clear that thin women may also have PCOS. All symptoms of PCOS will be worse in women who are overweight or obese. This relates most strongly to insulin resistance, which is a key factor of PCOS and will be described later. The degree of menstrual irregularity may also relate to the finding of insulin resistance.

Characteristic endocrinologic features include abnormal gonadotropin secretion caused by either increased gonadotropin-releasing hormone (GnRH) pulse amplitude or increased pituitary sensitivity to GnRH. These abnormalities result in tonically elevated levels of luteinizing hormone (LH) in about two thirds of the women with this syndrome (Fig. 40-8). After a bolus of GnRH, there is usually an exaggerated response of LH, but not of follicle-stimulating hormone (FSH) (Fig. 40-9). In addition, there are increased circulating levels of androgens pro-duced by both the ovaries and the adrenal glands (Fig. 40-10). Serum testosterone levels usually range between 0.7 to 1.2 ng/mL, and androstenedione levels are usually between 3 and 5 ng/mL. In addition, about half the women with this syndrome have elevated levels of DHEA-S, suggesting adrenal androgen involvement. Evidence also exists for adrenal hyperactivity in at least a third of women with PCOS. Although nearly all women with PCOS have elevated levels of circulating androgens, we have found that the presence or absence of hirsutism depends on whether those androgens are converted peripherally by 5αreductase to the more potent androgen DHT, as reflected by increased circulating levels of 3αdiol-G. Nonhirsute women with PCOS have elevated circulatory levels of testosterone or DHEA-S or both but not 3αdiol-G. The tonically elevated levels of LH are usually above 15 mIU/mL.

Figure 40-8Patterns of pulsatile luteinizing hormone (LH) secretion in patients with polycystic ovarian syndrome (PCO) and in control subjects with normal early follicular (EF) and midfollicular (MF) phases. The asterisks indicate significant pulses.

(From Kazer AR, Kessel B, Yen SSC: Circulating luteinizing hormone pulse frequency in women with polycystic ovary syndrome. J Clin Endocrinol Metab 65:233, 1987.)

Figure 40-9Comparison of quantitative luteinizing hormone (LH) and follicle-stimulating hormone (FSH) release in response to a single bolus of 150 mg of GnRH in patients with polycystic ovarian syndrome (PCO) and in normal women during low-estrogen (early follicular) and high-estrogen (late follicular) phases of their cycles.

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

Figure 40-10Mean (?SD) concentrations of testosterone and Δ4-androstenedione in 19 patients with polycystic ovarian syndrome (PCO) and 10 normal subjects between days 2 and 4 (D2-4) of their menstrual cycles.

(From DeVane GW, Czekala NM, Judd HL, et al: Circulating gonadotropins, estrogens, and androgens in polycystic ovarian disease. Am J Obstet Gynecol 121:496, 1975.)

Because FSH levels in women with PCOS are normal or low, an elevated LH-FSH ratio has been used to diagnose PCOS. However, we reported that among women with a clinical diagnosis of PCOS, only 70% had an elevated level of immunoreactive LH or an immunologic LH-FSH ratio greater than 3. Although almost all women with PCOS had elevated serum levels of biologically active LH (Fig. 40-11), use of LH or the LH-FSH ratio should not be part of the diagnostic evaluation of PCOS.

Figure 40-11Serum measurements of immunoreactive luteinizing hormone (LH), immunoreactive LH: follicle-stimulating hormone (FSH) ratios, and bioactive LH in control subjects (C), women with chronic anovulation (CA), and women with PCOS (PCO). Boxes represent the mean ?3 standard deviation (SD) of control levels.

(Reprinted from Fertility and Sterility, 39, Lobo RA, Kletzky OA, Campeau JD, et al, Elevated bioactive luteinizing hormone in women with the polycystic ovary syndrome, 674. Copyright 1983, with permission from The American Society for Reproductive Medicine.)

In addition to increased levels of circulatory androgens, we have found that women with PCOS have increased levels of biologically active (non-SHBG-bound) estradiol, although total circulating levels of estradiol were not increased (Fig. 40-12). The increased amount of non-SHBG-bound estradiol is caused by a decrease in SHBG levels, which is brought about by the increased levels of androgens and obesity with high insulin levels present in many of these women. Estrone is also increased because of increased peripheral (adipose) conversion of androgen. The tonically increased levels of biologically active estradiol may stimulate increased GnRH pulsatility and produce tonically elevated LH levels and anovulation. In addition, the lowered SHBG level increases the biologically active fractions of the elevated androgens in the circulation. The importance of the decreased levels of SHBG is shown schematically in Figure 40-13. This relative hyperestrogenism (elevated estrone and non-SHBG-bound estradiol), which is unopposed by progesterone because of anovulation, increases the risk of endometrial hypoplasia.

Figure 40-12Serum estrogen concentrations in 13 normal women and 22 PCOS patients (shaded areas).

(From Lobo RA, Granger L, Goebelsmann U, et al: Elevation in unbound serum estradiol as a possible mechanism for inappropriate gonadotropin secretion in women with PCO. J Clin Endocrinol Metab 52:156, 1981. Copyright 1981 by The Endocrine Society.)

Figure 40-13Scheme depicting the possible role of adrenal-derived androgen (A2, androstanediol; E2, estradiol; LH, luteinizing hormone; SHBG-BC, sex hormone-binding globulin binding capacity; T, testosterone.) In initiating androgen excess and anovulation.

(From Lobo RA, Goebelsmann U: Effect of androgen excess on inappropriate gonadotropin secretion as found in polycystic ovary syndrome. Am J Obstet Gynecol 142:394, 1982.)

About 20% of women with PCOS also have mildly elevated levels of prolactin (20 to 30 ng/mL), possibly related to increased pulsatility of GnRH, to a relative dopamine deficiency, or tonic stimulation from unopposed estrogen.

It is well established that some degree of insulin resistance occurs in most women with PCOS even those of normal weight. Insulin and insulin-like growth factor-I (IGF-I) enhance ovarian androgen production by potentiating the stimulatory action of LH on ovarian androstenedione and testosterone secretion. High levels of insulin bind with the receptor for IGF-I as a result of the significant homology of the IGF-I receptor with the insulin receptor. The granulosa cells also produce IGF-I and IGF-binding proteins (IGFBP). This local production of IGF-I and IGFBP may result in paracrine control and enhancement of LH stimulation and production of androgens by the theca cells in women with PCOS. Since IGFBP levels are lower in women with PCOS, this leads to increased bioavailable IGF-I, which increases stimulation of the theca cells in combination with LH to produce higher levels of androgen production. Insulin resistance and the resultant hyperinsulinemia stimulate ovarian androgen production. It is not clear why women with PCOS have insulin resistance; the insulin resistance in PCOS is greater than in age-matched and weight-matched controls.

Dunaif and colleagues studied a group of hyperandrogenic women with and without PCOS and obesity using glucose tolerance tests and insulin levels. They found that hyperinsulinemia occurred only in the women with PCOS, whether or not they were obese, but only the obese women with PCOS had im-paired glucose tolerance. In a subsequent study they found that nonobese women with PCOS also had glucose intolerance, but the incidence was less than in the obese women. In a prospective evaluation of 254 women with PCOS who had an oral glucose tolerance test, they found that 31% had impaired glucose tolerance and 7.5% had undiagnosed diabetes. In nonobese women with PCOS, 10% had impaired glucose tolerance and 1.5% had diabetes. Norman and coworkers showed that over a mean follow-up time period of 6.2 years 9% of women with PCOS in Australia progressed to having impaired glucose tolerance and 8% became diabetic. Thus, the negative effect of obesity and PCOS on insulin resistance is additive. Fasting glucose levels are a poor predictor of diabetes in PCOS. It would appear advisable to perform an oral glucose tolerance test at the time of diagnosis in overweight women with PCOS and periodically thereafter. If abnormalities in glucose metabolism are found, appropriate interventions should be recommended.

Acanthosis nigricans (AN) has been found in about 30% of hyperandrogenic women. About half of the hyperandrogenic women who had PCOS and were obese had AN. Although it has been suggested that the presence of hyperandrogenism insulin resistance, and AN constitute a special syndrome (the HAIR-AN syndrome), most investigators believe that women with PCOS who have AN are a subgroup of those with PCOS and do not have a distinct endocrine disorder. No causal relation among PCOS, obesity, insulin resistance, and hyperandrogenism has been elucidated to date. It is clear, however, that the combination of insulin and IGF-I (which enhance AN) leads to the hyperandrogenism.

Although the ovaries of women with PCOS produce exces-sive amounts of androgen, particularly androstenedione, there is no inherent endocrinologic abnormality in the ovaries. The tonically elevated levels of LH cause the stromal tissue to produce more androstenedione and other androgens, which in turn produces premature follicular atresia. Furthermore, the ovaries are deficient in aromatase, a deficiency that results in less conversion of androstenedione to estrogen in the ovary. The polycystic ovary does not secrete increased amounts of estrone or estradiol, but the increased levels of androstenedione are peripherally converted to estrone, thereby increasing circulating estrone levels.

Whatever the cause, the endocrinologic effects of PCOS produce a cycle of events, as shown by Yen and associates (Fig. 40-14). The increased pulsatility of GnRH produces tonically elevated LH levels and increased ovarian androgen production. Peripheral conversion of androstenedione to estrone in conjunction with the decreased SHBG levels causes tonic hyperestrogenism, which increases the pituitary sensitivity to GnRH and leads to increased LH release.

Figure 40-14The interdependent event of high luteinizing hormone/follicle-stimulating hormone (LH-FSH) ratio occasioned by an increased gonadotropin-releasing hormone (GnRH) secretion as a consequence of reduced hypothalamic inhibition. This setting induces an increased ovarian androgen production by the theca cells and acyclic estrogen feedback system in maintenance of chronic anovulation in polycystic ovary syndrome (PCOS).

(Modified from Yen SSC, Chaney C, Judd HL: Functional aberrations of the hypothalamic-pituitary system in polycystic ovary syndrome: A consideration of the pathogenesis. In James VHT, Serio M, Guisti G [eds]: The Endocrine Function of the Human Ovary. New York, Academic Press, 1976.)

Some Insights into Pathophysiology

It is clear that there is a genetic predisposition to PCOS. However, it is likely that several genes are involved. A susceptibility gene for PCOS has been suggested to lie in the region of 19p 3.2 although this needs confirmation. Environmental factors are clearly involved as well, based on twin studies, where PCOS is not always concordant on a genetic basis. Maternal exposure to androgen has been shown in a monkey model to contribute to the development of PCOS.

It is clear from the cycle model of Yen (see Fig. 40-14) that the syndrome may evolve from any point. Thus it was attractive to postulate that dopamine deficiency in the hypothalamus might give rise to the exaggerated LH responses in PCOS, and there are several similar hypotheses. However, it has been observed that morphologically identifiable polycystic ovaries are seen in children. This occurrence predicts puberty and other normal endocrinologic events, suggesting a central role for altered PCOM in the disorder. Furthermore not all women with isolated polycystic ovaries have PCOS as stated earlier. Thus a pathophysiologic model can be put together as follows:

An ovary is polycystic in up to a fifth of girls according to data from Bridges and colleagues. Thus the ovary transitions early in life from normal to polycystic-appearing (PAO). This influence occurs in a specific way by genetic factors, by environmental factors, or is induced by other endocrine disturbances (Fig. 40-15). The woman who develops PAO may have normal menses, normal androgen levels, and normal ovulatory function and parity. However, if subjected to various susceptibility factors or “insults” with various degrees of severity, women with PAO may develop a full-blown syndrome (PCOS). The syndrome, if full-blown, will exhibit the full extent of hyperandrogenism and anovulation, with the most extreme form of this menstrual disturbance being amenorrhea. However, in this spectrum of disorders, the androgen disturbances may also be near normal. Similarly, the menstrual disturbance may be very mild (Fig. 40-16).

Figure 40-15Pathophysiology of polycystic ovary syndrome (PCOS): The syndrome develops when one or more “insults” persist. IGFBG1, insulin-like growth factor-binding protein 1; P450c 17α, cytochrome 450.

Figure 40-16Pathophysiology of polycystic ovary syndrome (PCOS): Differences in presentation. PAO, polycystic-appearing ovaries.

This model requires that normal homeostatic factors may be able to ward off stressors or insults in some women who can go through life without PCOS but have a PAO which does not change morphologically. Alternatively, with varying degrees of success a woman's homeostatic mechanism may at any time, early or later in reproductive life, allow symptoms or PCOS to emerge with varying degrees of severity. Two of the major insults are thought to be weight gain and psychological stress. Thus, the typical teenager born with PAO may develop PCOS fairly quickly; yet a PCOS picture may develop only later in life in some women even after having children. Because of hyperandrogenism and obesity, women with PCOS have abnormal lipoprotein profiles. The increase in triglycerides and the decrease in low-density lipoprotein (LDL) cholesterol and high-density lipoprotein (HDL) may be related to the increased body weight or hyperandrogenism (Fig. 40-17). If PCOS is not treated, the endocrinologic abnormalities persist and gradually worsen until the ovary stops functioning at the menopause. Long-term sequelae of PCOS were examined in 33 women ages 40 to 59 who had undergone ovarian wedge resection 22 to 31 years previously. Compared with age-matched controls, they had a significantly greater incidence of hypertension and diabetes mellitus (Fig. 40-18). Using multiple logistic regression analysis it was predicted that women with PCOS have an increased risk of developing cardiovascular disease compared with controls. Many studies show increased cardiovascular risk factors in PCOS, including elevations in homocysteine, C-reactive pro-tein (CRP), endothelin-1, reductions in plasminogen activator inhibitor-1 (PA-1), increased coronary Ca2+, increased carotid intima media thickness, and endothelial dysfunction. However, retrospective studies by Pierpont and Wild and colleagues reported that women with PCOS do not have an increased risk of death due to cardiovascular disease, unless they were diabetic. This issue is not clear at present.

Figure 40-17Lipid and lipoprotein profiles in 13 women with polycystic ovary syndrome (PCOS) versus control group when matched for percent ideal body weight. Differences are evident in all measures (p < 0.01). HDL, high-density lipoprotein; LDL, low-density lipoprotein.

(From Wild RA, Bartholomew MJ: The influence of body weight on lipoprotein lipids in patients with polycystic ovary syndrome. Am J Obstet Gynecol 159:423, 1988.)

Figure 40-18Prevalence of hypertension (medically treated) and manifest diabetes mellitus in 33 PCOS subjects and 132 referents. The dark-shaded bars illustrate the polycystic ovary syndrome (PCOS) subjects. The light-shaded bars illustrate the referents. Statistical comparisons were made between the women with PCOS and referents. Differences were considered significant at *p = 0.05 and * * *p = 0.001.

(Reprinted from Fertility and Sterility, 57, Dahlgren E, Janson PO, Johansson S, 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, 505. Copyright 1992, with permission from The American Society for Reproductive Medicine.)

Isolated Polycystic Ovaries

We have shown that normal ovulatory women with PAO or PCOM have a subtle form of ovarian hyperandrogenism (when stimulated with GnRH-A or human chorionic gonadotropin [HCG]). We have also found subtle changes in insulin sensi-tivity and altered lipoproteins in these women. There may also be some reduction in fertility in these ovulatory women. Therefore although many women with isolated PAO/PCOM may not have any problems, we view this finding as a risk for developing the consequences of PCOS.

Consequences of PCOS

The importance of diagnosing PCOS is that there are known long-term consequences of the diagnosis warranting lifelong surveillance. Figure 40-19 shows the ages at which various consequences may emerge. Although in the early reproductive years abnormal bleeding and infertility are common, later in life, concerns relate to cardiovascular disease, diabetes mellitus, and ovarian cancer. The risk of ovarian cancer has been suggested to be increased twofold, a finding seen in women with general infertility. This risk is normalized with the use of oral contraceptives (OCs).

Figure 40-19Consequences of polycystic ovary syndrome (PCOS). ca, cancer; CVD, cardiovascular disease; DM, diabetes mellitus.

Stromal Hyperthecosis

Stromal hyperthecosis is an uncommon benign ovarian disorder in which the ovaries are bilaterally enlarged to about 5 to 7 cm in diameter. Histologically there are nests of luteinized theca cells within the stroma (Fig. 40-20). The capsules of these ovaries are thick, similar to those found in PCOS, but unlike PCOS, subcapsular cysts are uncommon. The theca cells produce large amounts of testosterone as determined by retrograde ovarian vein catheterization. Like PCOS, this disorder has a gradual onset and is initially associated with anovulation or amenorrhea and hirsutism. However, unlike PCOS, with increasing age the ovaries secrete steadily increasing amounts of testosterone. Thus when women with this disorder reach the fourth decade of life, the severity of the hirsutism increases, and signs of virilization, such as temporal balding, clitoral enlargement, deepening of the voice, and decreased breast size, appear and gradually increase in severity. By this time serum testosterone levels are usually greater than 2 ng/mL, similar to levels found in ovarian and adrenal testosterone-producing tumors. However, with the latter conditions the symptoms of virilization appear and progress much more rapidly than with ovarian hyperthecosis, in which symptoms progress gradually over many years.

Figure 40-20 A, Sagittal section of typical hyperthecotic ovary illustrating small number of follicular cysts and massive amount of stromal hyperplasia. B, Islands of luteinized theca-like cells deep in stroma of ovary in hyperthecosis.

(From Wilroy RS Jr, Givens JR, Wiser WL, et al: Hyperthecosis: An inheritable form of polycystic ovarian disease. In Bergsma D [ed]: Genetic Forms of Hypogonadism. Miami, FL, Symposia Specialists for the National Foundation-March of Dimes BD:OAS XI(4):81, 1975, with permission.)

Androgen-Producing Tumors

Ovarian Neoplasms

It is possible for nearly every type of ovarian neoplasm to have stromal cells that secrete excessive amounts of testosterone and cause signs of androgen excess. Thus on rare occasions excess testosterone produced by both benign and malignant cystadenomas, Brenner's tumors, and Krukenberg's tumors has caused hirsutism or virilization or both. Certain germ cell tumors contain many testosterone-producing cells. The testosterone produced by two of these neoplasms, Sertoli-Leydig cell tumors and hilus cell tumors, nearly always causes virilization. In addition, lipoid cell (adrenal rest) tumors can produce increased amounts of testosterone or DHEA-S or both. Rarely granulosa/theca cell tumors can also produce testosterone in addition to increased levels of estradiol.

Androgen-producing ovarian tumors usually produce rapidly progressive signs of virilization. Sertoli-Leydig cell tumors usually develop during the reproductive years (second to fourth decades), and by the time they produce detectable signs of androgen excess, the tumor is nearly always (more than 85% of the time) palpable during bimanual examination. These tumors are uncommon. Less than 1% of solid ovarian neoplasms are Sertoli-Leydig cell tumors. Hilus cell tumors most often occur after menopause. They are usually small and not palpable during bimanual examination; however, the history of rapid development of signs of virilization and the presence of markedly elevated levels of testosterone (more than two and a half times the upper limits of the normal range) with normal levels of DHEA-S usually facilitate the diagnosis.

Adrenal Tumors

Nearly all the androgen-producing adrenal tumors are adenomas or carcinomas that generate large amounts of the C19 steroids normally produced by the adrenal gland: DHEA-S, DHEA, and androstenedione. Although these tumors do not usually directly secrete testosterone, testosterone is produced by extraglandular conversion of DHEA and androstenedione. Women with these tumors usually have markedly elevated serum levels of DHEA-S (>8 μg/mL). Women with these laboratory findings and a history of rapid onset of signs of androgen excess should undergo a computerized tomography (CT) scan or magnetic resonance imaging (MRI) of the adrenal glands to confirm the diagnosis. In addition to these uncommon tumors, a few testosterone-producing adrenal adenomas have been reported. The cellular patterns of these tumors resemble those of ovarian hilus cells, and the tumors secrete large amounts of testosterone. Because adrenal adenomas also secrete DHEA-S, an adrenal adenoma is highly likely when DHEA-S levels are greater than 8 μg/mL and testosterone levels are more than 1.5 ng/mL.

Late-Onset 21-Hydroxylase Deficiency

Congenital adrenal hyperplasia (CAH) is an inherited disorder caused by an enzymatic defect (usually 21-hydroxylase [21-OHase] or less often 11β-hydroxylase), resulting in decreased cortisol biosynthesis. As a consequence, adrenocorticotropic hormone (ACTH) secretion increases and adrenal cortisol precursors produced proximal to the enzymatic block accumulate and are converted mainly to DHEA and androstenedione. These C19 steroids are in turn peripherally converted to testosterone, which produces signs of androgen excess.

Because the enzymatic defects are congenital, the classic severe form (complete block) usually becomes clinically apparent in fetal life by producing masculinization of the female external genitalia. The severe form of CAH is the most common cause of sexual ambiguity in the newborn. The more attenuated (mild) block of 21-OHase activity usually does not produce physical signs associated with increased androgen production until after puberty. Thus this condition, termed late-onset 21-hydroxylase deficiency (LOHD) or late-onset congenital adrenal hyperplasia, is associated with the development of signs of hyperandro-genism in a woman in the second or early third decade of life.

Although the incidence of classic CAH is only 1 per 14,500 live births worldwide, Speiser and coworkers, using histocompatibility locus antigen (HLA)-B genotyping of families with LOHD-affected individuals, concluded that the incidence of LOHD varied among different ethnic groups but overall was probably the most frequent autosomal genetic disorder in humans. The incidence of LOHD was estimated to be 0.1% among a diverse white population; among Yugoslavians, Hispanics, and Ashkenazi Jews, however, the incidence was 1.6%, 1.9%, and 3.7%, respectively (Fig. 40-21). Both classic CAH and LOHD are transmitted in an autosomal recessive manner at the CYP21B locus and are linked to the HLA-B locus.

Figure 40-21Relative frequencies of nonclassic 21-hydroxylase deficiency, classic 21-hydroxylase deficiency, and other autosomal recessive disorders.

(From Speiser PW, Dupont B, Rubenstein P, et al: High frequency of nonclassical steroid 21-hydroxylase deficiency. Am J Hum Genet 37:650, 1985.)

The molecular basis of the disease is complex. The gene for CYP21 is located on 6p near the HLA complex. In proximity to this gene is a nonfunctional or pseudogene (CYP21P). Depending on the population, one fourth to one fifth of indi-viduals with classic CAH have a deletion of the CYP21 locus or a rearrangement between CYP21 and CYP21P. Current molecular techniques of genotyping can pick up well over 95% of these abnormalities, with the majority of cases being one of 10 common mutations. A spectrum of mutations results in the enzymatic defects and clinical presentations shown in Table 40-6.

Table 40-6 --Genotypic Characterization of the Forms of 21-Hydroxylase Deficiency

Rights were not granted to include this data in electronic media. Please refer to the printed book.

From New MI, White PC, Pang S, et al: The adrenal hyperplasias. In Scriver CR, Beaudet AL, Sly S, Valle D (eds): Metabolic Basis of Inherited Diseases, 6th ed. New York, McGraw-Hill, 1989.

LOHD is a phenotype that is symptomatic after adolescence and does not define the genotype. Affected individuals may be homozygous for alleles, yielding mildly abnormal enzymatic activity, or compound heterozygotes with a combination of defective alleles. The so-called cryptic 21-OHase deficiency, on the other hand, represents mild or asymptomatic individuals with biochemically identified defects that, with the advent of molecular diagnostic techniques, have been redefined as belonging to several different clinical presentations.

New and associates have proposed a schema for identifying and classifying the clinical spectrum of disease shown in Table 40-6. Since there are three possible manifestations of CYP21Y alleles (normal, mildly defective, or severely defective), there are six possible genotypes representing three clinical phenotypes (asymptomatic, LOHD, and classic CAH). Individuals with LOHD may be compound heterozygotes, with one mildly and one severely defective allele, or homozygous, with two mildly defective alleles. Although biochemical differences in the hormonal response to ACTH have been shown between these two genotypes, their phenotypes are similar. Carriers can be identified among family members who are heterozygous with one normal allele. These individuals have normal basal 17-hydroxyprogesterone (17-OHP) levels, a mild degree of hirsutism, if present, and smaller increases of 17-OHP after ACTH stimulation, usually between 3.5 and 10 ng/mL.

LOHD is also usually associated with menstrual irregularity. It has been hypothesized that the mechanism for anovulation is similar to that which occurs with PCOS. The increased levels of androgen lower SHBG levels, thus increasing the amount of biologically active circulating estradiol. The increased estradiol stimulates tonic LH release, which increases ovarian androgen production and locally inhibits follicular growth and ovulation. Thus women with this disorder present with postpubertal onset of hirsutism and oligomenorrhea or amenorrhea, similar to women with PCOS. However, women with LOHD, unlike those with PCOS, commonly have a history of prepubertal accelerated growth (ages 6 to 8 years) with later decreased growth and a short ultimate height. A history of this growth pattern, a family history of postpubertal onset of hirsutism, and findings of mild virilization are indicators of the presence of CAH.

To differentiate LOHD from PCOS, measurement of basal (early-morning) serum 17-OHP levels should be performed. If basal levels of 17-OHP are greater than 8 ng/mL, the diagnosis of LOHD is established. If 17-OHP is above normal (2.5 to 3.3 ng/mL) but less than 8 ng/mL, an ACTH stimulation test should be performed. A baseline 17-OHP should be measured and 0.25 mg of synthetic ACTH infused as a single bolus. One hour later another serum sample of 17-OHP should be measured. If the level of 17-OHP increases more than 10 ng/mL, the diagnosis of LOHD is established (Fig. 40-22). Individuals with LOHD should be treated with continuous corticosteroids to arrest the signs of androgenicity and restore ovulatory menstrual cycles.

Figure 40-22Means and ranges of 17αhydroxyprogesterone levels before and after cosyntropin administered intramuscularly in normal subjects, suspected heterozygotes, patients with late-onset congenital adrenal hyperplasia (CAH), and one patient with adrenal carcinoma.

(From Baskin HJ: Screening for late-onset congenital adrenal hyperplasia in hirsutism or amenorrhea. Arch Intern Med 147:847, 1987.)

Cushing's Syndrome

Excessive adrenal production of glucocorticoids due to increased ACTH secretion (Cushing's disease) or adrenal tumors produces the signs and symptoms of Cushing's syndrome. These findings include hirsutism and menstrual irregularity in addition to the classic findings of central obesity, dorsal neck fat pads, abdominal striae, and muscle wasting and weakness. The latter catabolic effect of glucocorticoid excess differs from the anabolic effects of testosterone excess, but some women with PCOS may have other clinical findings that are similar to those found with Cushing's syndrome. In such instances Cushing's syndrome can be easily excluded by performing an overnight dexamethasone suppression test. To perform this test, 1 mg of dexamethasone is ingested at 11 pm, and plasma cortisol is measured the following morning at 8 am (Fig. 40-23). If the cortisol level is less than 5 μg/100 mL, Cushing's syndrome is ruled out. If the cortisol level fails to suppress to this degree, the diagnosis of Cushing's syndrome is not established. It is necessary to perform a complete dexamethasone suppression test (Liddle's test) or measurement of urinary free cortisol and plasma ACTH to determine whether Cushing's syndrome exists.

Figure 40-23Outline of overnight dexamethasone suppression test.

(From Goebelsmann U, Lobo RA: Androgen excess. In Mishell DR Jr, Davajan V, Lobo RA [eds]: Infertility, Contraception and Reproductive Endocrinology, 2nd ed. Cambridge, MA, Blackwell Scientific, 1986.)

Depression and other conditions can cause failure to suppress with the dexamethasone screening test just described. Accordingly many endocrinologists prefer to depend on measurements of 24-hour urinary free cortisol. A creatinine is also measured to gauge completeness of collection. Values above 100 μg/24 hr are abnormal, and values greater than 240 μg are virtually diagnostic of Cushing's syndrome. Cushing's syndrome may result from a pituitary tumor producing ACTH (Cushing's disease), from an ectopic tumor in the body, from adrenal neoplasms or hyperplasia. Various algorithms have been developed for this differential diagnosis.