745© 2013 David G. Wild. Published by Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/B978-0-08-097037-0.00061-0

The androgens have a wide variety of actions in both sexes at every stage of development. In women, the ovaries and adrenal glands secrete mostly weak androgens: dehydroepi-androsterone (DHEA) and androstenedione. These weak androgens are then converted to the more potent androgen, testosterone, in peripheral tissue. Testosterone is subse-quently converted to 5α-dihydrotestosterone (DHT) in many androgen-sensitive tissues, e.g. hair follicles and pros-tate. The presence of increased concentration of testoster-one and DHT in the female causes hirsutism and virilization. Hirsutism is the term for increased growth of terminal hair, usually on the face, but often including the thighs, arms, breasts, and supra-pubic areas to a varying degree. It is reported to affect between 5% and 10% of women of reproductive age. Virilization describes the appearance of male characteristics such as clitoromegaly, deepening of the voice and increased muscle mass. Hypertrichosis is a sepa-rate clinical condition not caused by androgen excess. It is excessive hair growth over and above what is the norm for age, sex, or race, not isolated to sexual areas, and can affect either sex. It may affect the whole body or isolated areas. Its cause is unknown although some forms may be inherited.

Both testosterone and DHT are required for the devel-opment of the male sex organs. Absence, or reduced syn-thesis, of these steroids or their receptors during fetal life leads to feminization. Conversely, exposure of the female fetus to high concentrations of androgens leads to viriliza-tion and masculinization.

Androgens (produced by the adrenals during the adre-narche) are required for the development of secondary sex hair (pubis, axilla) in both sexes. This is demonstrated by the development of pubic and axillary hair in agonadal men. The additional requirement for androgens to be converted to testosterone and then to DHT is shown by the lack of secondary sexual hair in the complete form of the androgen-insensitivity syndrome where there is an absence of the receptors for DHT. Asexual hair is also responsive to the action of androgens. Hence, men develop coarse terminal hair on the face, chest, and limbs; a similar picture occurring in women with increased androgen lev-els is termed hirsutism. Racial and genetic factors have to be taken into consideration when assessing excess hair growth in women. For instance, noticeable hair growth on the face of a woman of Mediterranean origin may be expected, but the same hair growth on the face of a Chi-nese woman would be regarded as abnormal.

Excessive secretion of androgens in the female leads to masculinization. If this occurs during fetal life, the child is born with ambiguous genitalia. There is enlargement of the clitoris to form a pseudo-penis and labial fusion. The extent of the masculinization is variable, but complete labial fusion with a terminal urethral opening has been reported. Masculinization of the adult female involves increase in muscle mass, clitoromegaly, deepening of the voice, development of a male escutcheon, and frontal

balding. The degree of virilization or masculinization is positively correlated with the concentration of circulating androgen.

Clinical DisordersPOLYCYSTIC OVARIAN SYNDROMEIn 1935, Stein and Leventhal described a syndrome of obe-sity, hirsutism, menstrual disturbance, and polycystic ova-ries: the Stein–Leventhal syndrome. We now know that this represents just one subgroup of the polycystic ovar-ian syndrome (PCOS), which is classically typified by the presence of enlarged cystic ovaries that are covered with a pale, thickened, glistening capsule. The cysts are numer-ous, measuring between 2 and 8 mm in diameter and arranged around the periphery of the ovary. The stroma is reported as enlarged. In a study of 1741 patients with poly-cystic ovaries, Balen et al. (1995) reported that 29.7% had a normal menstrual cycle, 47.0% had oligomenorrhea, 19.2% had amenorrhea, 2.7% had polymenorrhea, and 1.4% had menorrhagia. In addition, 66.2% had hirsutism; mild in 20.6%, moderate in 40.7%, and severe in 4.9%, while 34.7% of the patients had acne and 2.5% had acan-thosis nigricans. PCOS is therefore a spectrum of disease.

The criteria for the diagnosis of PCOS vary from practice to practice and from country to country. For example, the National Institute of Child Health and Human Develop-ment Conference of PCOS recommended that the major criteria for PCOS should include (in order of importance) hyperandrogenism and/or hyperandrogenemia, oligomen-orrhea, and the exclusion of other known disorders exhibit-ing similar abnormalities. Therefore, this is a diagnosis of exclusion where the detection of polycystic ovaries is not required. The Rotterdam consensus of 2003 proposed the inclusion of two out of three criteria, namely hyperan-drogenism (either clinical or biochemical), oligomenorrhea, and polycystic ovaries, while excluding other endocrine dis-orders. These criteria would include women with polycystic ovaries and menstrual disorder without signs of hyperandr-ogenemia. The Androgen Excess Society issued a guideline in 2006 again placing weight on the presence of hyperan-drogenism for the diagnosis of PCOS. It proposed that the diagnosis be based on the presence of hyperandrogenemia with oligomenorrhea and/or polycystic ovaries.

Anti-Müllerian hormone (AMH) is involved in follicular growth, and it has recently been shown that concentra-tions are raised in PCOS by two- to threefold. This is in line with the increased number of pre-antral and small antral follicles. AMH may therefore provide useful diag-nostic confirmation of PCOS when ultrasound evidence is unclear or not available. The concentration of AMH may also be useful as an indicator of the severity of disease as well as predicting the beneficial effects of weight loss. Increased AMH has also been noted in the prepubertal and

Hirsutism and Virilization in the FemaleMichael J. Wheeler (mjw@the2wheelers.plus.com)

C H A P T E R

9.7

746 The Immunoassay Handbook

peripubertal daughters of PCOS women indicating that ovarian abnormalities are present as early as infancy.

Testosterone concentrations are frequently above the reference interval in patients with PCOS, androstenedione concentrations occasionally so. Free testosterone concen-tration may be increased proportionately more than the total testosterone concentration because of the lowered sex hormone-binding globulin (SHBG) concentration. The concentration of SHBG has been shown to be inversely correlated with weight, but this seems to be in turn related to insulin resistance in these patients. A large number of studies have now shown that insulin appears to diminish SHBG concentration and hence increases free testosterone concentration. In most patients, the ovaries are the major source of the increased androgen secretion. Up to 10% of patients with PCOS may have increased dehydroepiandrosterone sulfate (DHEAS) concentration in the serum indicating increased androgen secretion from the adrenal. Several groups have investigated adrenal func-tion in PCOS patients, and subtle defects in adrenal ste-roidogenesis have been reported in 12–40% of these patients although no increase in adrenocorticotropic hor-mone (ACTH) secretion has been found.

Luteinizing hormone (LH) concentration in PCOS may be increased above or at the top of the reference interval for the follicular phase. The follicle-stimulating hormone (FSH) concentration is usually normal. The ratio of LH:FSH has been used to indicate the presence of PCOS, and a ratio of greater than three was said to be diagnostic. Although this was a helpful diagnostic tool when radioim-munoassays were used to measure the gonadotropins, cur-rent immunometric methods for LH give lower values and this ratio is no longer valid. Although some groups have tried to reestablish a diagnostic ratio, generally, it has been found to be unhelpful, and the most useful indicator of polycystic ovaries is a LH concentration greater than 10 IU/L. Patients may present with hirsutism, or infertility due usually to oligomenorrhea or amenorrhea. The use of anti-androgens, for example, cyproterone acetate or spi-ronolactone, is the most effective treatment for hirsutism. A number of different procedures have been carried out to achieve conception. These include clomiphene citrate, conventional gonadotropin therapy, pulsatile GnRH, GnRH agonist with gonadotropin therapy, and diathermy. Although for most of these treatments, a conception rate of 50–80% has been reported, spontaneous abortion has been reported to be as high as 40% in some studies. Those obese women with PCOS who are able to lose weight will often resume normal menstrual function.

Insulin resistance is present in about 40% of patients with PCOS. It occurs both in obese and non-obese PCOS patients although obesity further exacerbates insulin resis-tance. Hyperinsulinemia is thought to play a key role in PCOS. Reported actions include increase of hyperandro-genemia, decrease in SHBG concentration, increase in LH secretion from the pituitary, and increased estrogen production by the ovaries. The key role is demonstrated by the fact that suppression of androgen concentration does not lead to normal insulin sensitivity (see FURTHER READING for reviews). Harbourne et al. (2003) reported that treat-ment with the anti-hyperglycemic drug, metformin, reduced androgen slightly but showed a greater reduction

in the Ferriman–Gallwey score when compared with Dia-nette. Patients perceived greater reduction of hirsutism with metformin treatment. In support of these findings, most studies examining the effect of metformin treatment of PCOS patients find a modest reduction in circulating free androgens.

Polycystic ovaries are frequently present in other clini-cal disorders that are associated with increased secretion of androgens from the adrenal (congenital adrenal hyperpla-sia [CAH], see below and Cushing’s syndrome, see ADRENAL CORTEX). Because about 50% of patients with PCOS are hirsute and obese, it may be difficult to decide whether the patient has Cushing’s syndrome or PCOS. Appropriate tests can quickly exclude or confirm Cushing’s syndrome.

IDIOPATHIC HIRSUTISMSome women have hirsutism without the other features of PCOS being present. In some cases, the hirsutism may be due to genetic or racial factors. For instance, it is well rec-ognized that women of Mediterranean origin are more hirsute, and dark-haired women may be more conscious of visible hair on the upper lip in particular. Social pressures can also have an important role to play. In these women, androgen concentration is normal.

However, there is a separate group of women who have a marked increase of hair on the face, which may be present on other areas of the body. In the strict definition of the term, androgen concentration is within the reference inter-val. An index of the free testosterone concentration may show this parameter to be increased, but no pathology is obvious with no ultrasound evidence of polycystic ovaries. A number of androgen indicators have been used to disclose an abnormality of androgen synthesis or secretion, and these are discussed further under the appropriate analyte. Some investigators have seen this condition as one end of the spectrum of the polycystic ovary syndrome and would so classify these patients. In 2008, the Endocrine Society published guidelines for the evaluation and diagnosis of hir-sutism. Using the Ferriman–Gallwey score, hirsutism was defined as a score of ≥8. They did not advocate measuring androgens in cases of mild hirsutism (score 8–15) because it was unlikely that a medical disorder would be revealed that would change management or outcome. In these cases, ini-tial action would be a trial of cosmetic, dermatologic, or oral contraceptives. This was also the approach in women with a score >15 but with a normal morning testosterone. These first two conditions would fall into the category of idiopathic hirsutism especially when menstrual cycles are normal. Other investigations would be carried out when the testosterone was elevated. What is unclear is what weight is put on free testosterone. These days many, if not most, laboratories would measure testosterone and SHBG in the first instance and calculate the free testosterone. Therefore, some clinicians may put more weight on the free testoster-one result rather than the total testosterone concentration. Certainly, one colleague indicated that his treatment of hir-sutism was likely to be more aggressive if the free testoster-one was elevated, i.e. using anti-androgen therapy rather than creams or oral contraceptives. Hirsutism is associated with hyperprolactinemia, acromegaly, hyperthyroidism, Cushing’s syndrome, and late-onset CAH (LOCAH). In

747CHAPTER 9.7 Hirsutism and Virilization in the Female

the initial endocrine workup of the patient, it is common to include the measurement of prolactin, 17-hydroxyproges-terone, and thyrotropin (thyroid-stimulating hormone). Other hormones such as growth hormone and cortisol would only be requested if there was a high suspicion of acromegaly or Cushing’s disease. Treatment of idiopathic hirsutism is usually influenced by the distress it causes the patient. The Endocrine Society advocates the use of creams, physical removal of hair, or oral contraceptives as the first line of treatment for idiopathic hirsutism. More aggressive treatment includes insulin lowering drugs, dexamethasone, anti-androgens (spironolactone, cyproterone acetate), and gonadotropin-releasing hormone analogs, not all advocated by the Endocrine Society. Cyproterone acetate is not avail-able in the United States.

ANDROGEN-SECRETING TUMORS OF THE OVARYAndrogen-secreting tumors of the ovary are very rare. Several different types of tumor can occur, depending on the main type of cell present. The testosterone concentra-tion is typically greater than 1.7 ng/mL (6 nmol/L). Andro-stenedione may also be elevated in concentration. Patients may present with hirsutism or with evidence of virilization. Hirsutism of sudden onset and of increasing severity should alert the clinician to the possibility of an androgen-secreting tumor.

CONGENITAL ADRENAL HYPERPLASIACAH is due to an enzyme deficiency in steroid biosynthe-sis. The consequent fall in cortisol secretion leads to increased secretion of ACTH. This, in turn, increases the stimulus of this hormone to the adrenal and hence causes hyperplasia of the adrenal cortical cells. A number of enzyme deficiencies have been identified, but about 95% of the cases are due to a deficiency of the 21-hydroxylase enzyme (see Fig. 1). The reported incidence of classical CAH is 1 in 5000 to 1 in 15,000 with variation between ethnic and racial backgrounds.

In this particular condition, there is increased secretion of 17α-hydroxyprogesterone (17-OHP), the steroid

immediately before the enzyme deficiency. At the same time, more 17-OHP is converted to the androgens, and the increased concentration of these steroids causes virilization of the female fetus. Patients are treated with exogenous glucocorticoids, but, as up to 75% of the individuals also suffer salt loss, exogenous mineralocorticoid must be given.

Diagnosis of CAH is made by measuring the steroid immediately before the deficient enzyme. Therefore, nearly all cases may be diagnosed by measuring 17-OHP in the serum or saliva of the neonate.

Effective treatment will minimize further virilization and allow puberty to proceed at a more normal rate. There-fore, it is most important to control androgen levels. Although testosterone has been used to monitor treatment, androstenedione may be viewed as a more appropriate androgen. When androstenedione is suppressed well into the reference interval, the concentration of 17-OHP may be as much as four times the upper limit of its reference interval. If 17-OHP levels are within the reference interval, it is likely that the patient is being overtreated, and growth will be retarded as well as other side effects of glucocorti-coid excess being manifest. Therefore, it is inappropriate to monitor treatment with 17-OHP measurement alone.

To date, 127 mutations have been reported in the cyp21a2 gene. These range from complete loss of enzyme function to partial enzyme activity. The classical form of CAH has been described as an autosomal recessive gene, closely linked to the human leukocyte antigen-B (HLA-B) locus. It is now suggested that the pathophysiology is more complicated than suggested by an autosomal recessive dis-order as variations at other loci may influence steroid metabolism and steroid responsiveness. HLA typing can be carried out in families with an affected individual to assess the carrier status of relatives. Individuals who are heterozy-gous for the gene mutation also show abnormalities of ste-roid synthesis. 17-OHP may be slightly above or just within the reference interval but stimulation of steroido-genesis with ACTH results in a greater increase of 17-OHP secretion than seen in normal subjects. This abnormality may be further exposed by calculating the 17-OHP:cortisol ratio. Patients usually have no physical abnormality.

One group of patients shows no sign of abnormal steroid secretion until after puberty. These patients are said to have

FIGURE 1 Steroid pathways.

748 The Immunoassay Handbook

LOCAH. Patients have hirsutism, oligomenorrhea, and infertility, and since they also have polycystic ovaries (as do most patients with CAH), they may be wrongly diagnosed as having PCOS if appropriate biochemical tests are not carried out. 17-OHP concentrations are above the refer-ence interval for the follicular phase of the menstrual cycle and give an exaggerated response to ACTH stimulation. Patients may be treated with exogenous glucocorticoids although anti-androgens may also be required for effective treatment of hirsutism. Addition of anti-androgens may allow lower doses of glucocorticoid therapy and may even allow replacement of cortisone with prednisolone. The bal-ance is between suppressing androgen secretion and main-taining adrenal function. As androgens decrease with age it may allow decrease in steroid dosage with increasing age.

Most other cases of CAH are due to a deficiency of the 11β-hydroxylase enzyme. There is increased production of 11-deoxycortisol, the steroid produced immediately before the enzyme block (see CONGENITAL ADRENAL HYPERPLASIA), and diagnosis is made by measurement of this steroid in serum. There is also increased secretion of 11-deoxycorti-costerone and the androgens. Patients develop hyperten-sion, and affected girls have some degree of masculinization. Other enzyme deficiencies have been described, but these are very rare.

CUSHING’S SYNDROMESee also ADRENAL CORTEX.

Cushing’s syndrome results from an overproduction of cortisol. It may be the result of a pituitary adenoma (Cush-ing’s disease), an adrenal adenoma, adrenal carcinoma, or an ectopic source of ACTH. The disease is nine times more common in women. Increased steroidogenesis leads to increased secretion of androgens. The resulting hirsutism is generally mild in Cushing’s disease, but as adrenal tumors secrete greater quantities of androgens, severe hirsutism, clitoromegaly, and deepening of the voice may occur. Rarely, cortisol and androgen secretion is greatly increased when an adrenal carcinoma is present, but in the case of adrenal adenomas, DHEAS is usually below the reference interval. Treatment is appropriate for the abnormality and includes pituitary surgery, adrenalectomy, or removal of an ACTH-secreting tumor. Further treatment by pituitary irradiation or chemotherapy may be required. Careful assessment of pituitary and adrenal function is required after surgery and exogenous glucocorticoid given if necessary.

AnalytesLUTEINIZING HORMONE AND FOLLICLE-STIMULATING HORMONESee also INFERTILITY.

The following section describes the measurement of these hormones in the diagnosis of androgen disorders.

Clinical ApplicationThe secretion of LH and FSH is suppressed by very high levels of androgen. Thus, LH and FSH concentration is usually low normal or below the reference interval in

LOCAH and Cushing’s syndrome. Although the concen-tration of FSH is normal in patients with PCOS, LH con-centration is frequently above the reference range and may be up to twice the upper limit of the reference interval. A LH concentration greater than 10 IU/L with a normal FSH concentration is suggestive of polycystic ovaries for a serum sample taken day 2–5 of the menstrual cycle or in an amenorrheic woman.

Limitations � The measurement of LH and FSH is not helpful in the

diagnosis of Cushing’s syndrome, CAH, or androgen-secreting tumors.

� Although the mean ratio of LH:FSH is increased in PCOS, there is a large overlap with normal subjects, and the ratio is unhelpful in diagnosis. It has been questioned whether LH and FSH assays have a role in the diagnosis of PCOS when good ultrasound facilities are available.

Frequency of UseCommon.

ANTI-MÜLLERIAN HORMONE (AMH)See also INFERTILITY.

This following section discusses the clinical use of AMH measurement in the investigation of PCOS.

Clinical ApplicationAlthough it is uncertain whether AMH shows cyclical secre-tion or not during the menstrual cycle, intra- and inter-cycle fluctuations may be considered to be low enough to permit random blood samples for AMH measurement. All studies have found more than a twofold difference in mean AMH concentration in PCOS patients compared with nor-mal menstruating women. The increase in AMH appears to be related to the severity of the disease. The AMH concen-tration is higher in PCOS with insulin resistance compared with patients with normal insulin sensitivity. AMH is also higher in amenorrheic PCOS women compared to PCOS patients with oligomenorrhea. Since clinical specificity and sensitivity are reported as 92% and 67%, respectively, it has been suggested that AMH may be useful in the diagnosis of PCOS where ultrasound is not available or unclear.

It has been shown that in overweight PCOS patients who show menstrual improvement with weight loss, there is a significant fall in AMH concentration. Therefore, AMH measurement may be helpful is assessing therapy in PCOS patients. It is also reported that AMH measurement may be helpful in predicting outcome to clomiphene citrate therapy as well as evaluating efficacy with insulin sensitizers such as metformin.

TESTOSTERONESee also INFERTILITY.

The following section describes the use of testosterone measurement in the investigation of hirsutism and virilization.

749CHAPTER 9.7 Hirsutism and Virilization in the Female

Clinical ApplicationsA testosterone measurement is usually requested in every case of hirsutism. The primary use of this parameter is to diagnose an androgen-secreting tumor when values are typically greater than 1.7 ng/mL (6 nmol/L). The clinician should be alerted to the possibility of a tumor when hirsut-ism is of sudden onset with increasing severity.

The NIH 1990 criteria and the Rotterdam 2003 crite-ria require testosterone to be raised in PCOS. This is also true using the Endocrine Society guidelines of 2008. Dis-crimination between patients with PCOS, idiopathic hir-sutism, and normal women is enhanced by including the measurement of SHBG. From these two parameters, an indication of the free testosterone level may be achieved. Testosterone may be normal, but if the SHBG is low, the concentration of free testosterone may be abnormally high. It is questionable whether knowing the result of any of these three parameters will change the clinician’s man-agement of the patient with isolated mild hirsutism, and the Endocrine Society does not advocate the measure-ment of androgens in such patients. However, when there is a mixed ethnic population, an indication of the circulating free testosterone concentration may help to distinguish racial or genetic causes of hirsutism from an abnormal pathology. In addition, they may be helpful in confirming suppression or compliance during treatment with suppressive therapy.

Testosterone concentrations are higher in patients with CAH, LOCAH, and Cushing’s syndrome, but the measurement of this hormone has no place in the diagno-sis of these diseases. However, the measurement of testos-terone has been used to monitor the treatment of patients with CAH.

Limitations � Total testosterone measurement is greatly influenced

by SHBG levels. The SHBG concentration is increased by estrogens and anticonvulsants and in cirrhosis of the liver and some cases of hypothyroidism. In hirsutism, SHBG is often low, resulting in an elevated concentra-tion of nonprotein-bound testosterone. This may be a result of obesity or accompanying insulin resistance in these patients. In these situations, measurement of total testosterone alone can be particularly misleading. Free testosterone concentration may be estimated by using an SHBG measurement to derive a free androgen index.

� The measurement of testosterone gives no indication of the source of increased androgen secretion in women. The testosterone concentration does not pre-dict the efficacy of any treatment instituted to treat hir-sutism, apart from when an androgen-secreting tumor is present.

� As about half the circulating testosterone in women is derived from the peripheral conversion of the weaker androgens secreted by both the adrenals and ovaries, testosterone is likely to be less sensitive than andro-stenedione in monitoring the treatment of CAH.

� It has been reported that unidentified steroid metabo-lites, which cross-react in direct assays, may occasionally be secreted in large amounts. A falsely high testosterone concentration will be achieved, and any results from a

direct assay that are unexpectedly above 1.7 ng/mL (6 nmol/L) should be confirmed by an extraction assay. In the UK, this assay is available through the Supra-Regional Assay Service.

� Testosterone exhibits diurnal variation, being highest in the early morning and falling by 25–30% to a mini-mum in the early evening. Normal males can have a testosterone concentration in the late afternoon at the bottom or just below the 9 am reference range. Testos-terone concentrations close to the lower limit of the reference range, in samples taken in the afternoon, should always be checked with a 9 am sample.

� Testosterone levels may be elevated owing to alcohol abuse, stress, or hard physical exercise of short duration. It has been shown that extended exercise such as mara-thon running lowers the testosterone concentration.

Assay TechnologySee INFERTILITY for more details.

The lack of sensitivity and poor precision of direct methods at female concentration levels have led to two approaches in the measurement of testosterone in women and in the investigation of hirsutism. One approach has been in place for a number of years and that is the use of extraction methods. Such methods are more sensitive and specific and enable the investigator to have more confi-dence in normal and elevated levels. The other approach is more recent and that is the use of tandem mass spectrom-etry (MS). There are 19 participants using tandem MS in the UK NEQAS. Not surprisingly, these methods show less variability than the automated methods. Nevertheless, most routine laboratories use automated methods, and many would send clinical cases that were difficult to inter-pret to a specialist lab using tandem MS for a more defini-tive answer.

The UK NEQAS has recently shown that the presence of norethisterone interferes in the Roche automated method for testosterone. This occurs at concentrations that might be expected in women being given the drug for heavy periods or taking oral contraceptives. Some interfer-ence may even occur in the Siemens Centaur® assay.

Frequency of UseCommon.

SEX HORMONE-BINDING GLOBULINSHBG is a β-globulin, has a molecular mass of about 52 kDa, and is secreted by the liver. Secretion of this pro-tein is stimulated by estrogens and suppressed by andro-gens. This results in a sex difference in plasma concentration. In the circulation, SHBG binds several ste-roids. It has a high avidity for testosterone and DHT (about 1.5 × 109 mol/L) but a lower avidity for estradiol (5.0 × 108 mol/L). Nevertheless, even the binding with estradiol is considerably higher than the binding of these steroids to albumin (3.7–6.4 × 104 mol/L).

An inverse correlation between SHBG and obesity and SHBG and insulin resistance has been demonstrated, situ-ations which are found in patients with PCOS.

750 The Immunoassay Handbook

Measurement of SHBG and testosterone levels enables the concentration of biologically active testosterone to be estimated.

FunctionThere is still much debate about the precise role of SHBG. It has been suggested that SHBG dampens any large fluctuations in steroid concentration, thus main-taining a fairly constant level of unbound steroid avail-able to the tissues. Considering the marked episodic secretion of testosterone, this would seem a reasonable hypothesis. However, other experiments suggest that, if this is true, it is an oversimplification of the function of SHBG. Of greatest interest are the recent studies that have shown the presence of cell membrane receptors for SHBG, as well as evidence for internalization of the pro-tein. Hence, the whole role of SHBG at the cellular level is yet to be elucidated. Two recent reviews (see FURTHER READING) deal with SHBG receptor binding, and its inter-nalization in some tissues where it can effect androgen and estrogen binding.

Reference IntervalMen: 10–60 nmol/LWomen: 47–110 nmol/L(Orion SHBG IRMA)

Clinical ApplicationThe measurement of SHBG is used in the investigation of hirsutism. When combined with a testosterone test, the free testosterone index (or free androgen index) can be cal-culated (see below). This is particularly useful when the total testosterone concentration is normal but the SHBG is low, resulting in an elevated concentration of nonpro-tein-bound testosterone. Undetectable SHBG has been reported but is very rare.

During treatment of hirsutism with the combined treat-ment of anti-androgen and estrogen, the SHBG concen-tration is usually above, often twice, the upper limit of the reference interval. Therefore, the measurement of SHBG concentration can be used to confirm or refute compliance in patients who are not responding to therapy.

Limitations 1. The concentration of SHBG is increased by estro-

gens (oral contraceptives, pregnancy) and decreased by androgens. It is also altered in a number of clini-cal conditions and by several drugs, e.g. anti-epilep-tics and barbiturates.

2. SHBG concentration is increased in hepatic cirrho-sis, thyrotoxicosis, testicular feminization and hypo-gonadism in the male.

3. SHBG concentration is decreased in myxedema, Cushing’s syndrome, CAH, and acromegaly.

4. Exogenous T3 increases the secretion of SHBG, as do most of the anticonvulsant drugs. Dexametha-sone is reported to cause a small increase in SHBG concentration.

Assay TechnologyEarly methods were based on the binding of tritiated testos-terone to SHBG after endogenous steroids had been removed from the serum with charcoal. Tritiated DHT soon replaced tritiated testosterone because its higher bind-ing affinity for SHBG meant that removal of endogenous steroids was not required. Another procedure used Con-canavalin A-Sepharose 4B to bind SHBG before addition of tritiated DHT, whereas Iqbal and Johnson used a column of Cibacron-blue Sepharose 4B layered on top of LH-20.

Although manual assays are still available all participants in the UK NEQAS use commercial assays with the Sie-mens Immulite® 2000 and the Roche methods being the most common.

International Reference Preparations are available for the calibration of assays. The first preparation was made available in 1998, but recently, this has been superseded by the 2nd International Reference Preparation, code 08/266. It is not always clear what companies have used for the calibration of their assays, but this information should be available on request.

Type of SampleSerum.

Frequency of UseCommon.

FREE TESTOSTERONEOnly about 1% of the testosterone in the blood of women remains unbound to protein; in men, about 2% remain unbound. This “free” fraction has been traditionally regarded as the biologically active portion. However, over recent years, this view has been challenged. Some workers suggest that both the free and albumin-bound fractions should be deter-mined as the biologically important portions, whereas, more controversially, the SHBG-bound part has been suggested as the fraction available to tissues, see FURTHER READING. Nevertheless, there is little doubt that the nonprotein-bound steroid correlates well with the clinical state.

Free testosterone concentration is often measured rou-tinely in the investigation of hirsutism. Because the direct measurement of free testosterone is complex and the com-mercial kits unreliable, an indication of the free testoster-one concentration can be achieved by two methods. Firstly from the ratio of testosterone and SHBG concentrations – the free testosterone or free androgen index. Secondly, a closer approximation to the free testosterone concentra-tion can be calculated (derived free testosterone) from the testosterone, SHBG, and albumin concentrations, and a variety of calculations have been published using some or all of these parameters (see ASSAY TECHNOLOGY).

Reference IntervalWomen: 2.88–14.4 pg/mL (10–50 pmol/L)Men: 38.6–243 pg/mL (134–844 pmol/L)(Determined by micro-steady-state gel filtration in the author’s laboratory.)

751CHAPTER 9.7 Hirsutism and Virilization in the Female

Clinical ApplicationAn estimate of the free testosterone concentration may influence the treatment of hirsutism in women who have a normal total testosterone concentration. Treatment may be more aggressive in women where an underlying andro-gen abnormality is revealed.

LimitationsIn general terms, the levels of free and total testosterone are positively correlated with the degree of hirsutism and virilization in women. Nevertheless, there is considerable interindividual variation; some women have increased androgen concentrations but little or no hirsutism, whereas others have normal androgen results but quite marked hirsutism.

Clinicians disagree about the importance of a free tes-tosterone measurement. Some clinicians are guided in their treatment by the result, although most argue that they treat the hirsutism, and whether the androgen con-centration is normal or slightly elevated is irrelevant.

Assay TechnologyThe percentage of free testosterone in serum can be deter-mined by equilibrium dialysis, steady-state gel filtration, ultracentrifugation, and micro-filtration (see FURTHER READING). These methods are lengthy, tedious, and require good technical skill and experience. They are unsuited to routine clinical use.

An estimate of the free testosterone concentration may be made from the total testosterone and SHBG results. There is a good correlation between the percent free tes-tosterone and the SHBG concentration (see Fig. 2).

An equation describing this correlation can be used to calculate the percent free testosterone from the SHBG result. The free testosterone concentration can then be calculated from the total testosterone result. Some labo-ratories calculate a free testosterone index or free andro-gen index from the ratio of the total testosterone and

SHBG results. Laboratories should establish their own reference intervals because of differences between methods.

It has been reported that albumin-bound testosterone diffuses readily from the circulation into tissues. It has been proposed, therefore, that the free testosterone plus the albumin-bound testosterone (the non-SHBG-bound testosterone) represent the fraction of testosterone available to target cells. The measurement of this frac-tion can be carried out using a simple ammonium sulfate precipitation of the SHBG-bound testosterone before immunoassay.

A number of mathematical models have been devised to calculate the non-SHBG-bound testosterone. Their com-plexity depends on the number of different parameters (e.g. SHBG, albumin, testosterone, cortisol-binding glob-ulin) that are included in the calculation.

A radioimmunoassay kit is available from Siemens. Studies that have examined the reliability of this kit have shown good correlations with total testosterone, free testoster-one, and the androgen index, although values about half those obtained by equilibrium dialysis have been reported. It has been suggested that rather than measuring the free testosterone concentration, the method measures a con-stant proportion of the total testosterone (see FURTHER READING). Rosner (2001) has been particularly critical of this assay.

Vermeulen et al. (1999) evaluated the methods for esti-mating free and non-SHBG-bound testosterone, com-paring results with equilibrium dialysis measurement of the free testosterone. They concluded that neither the free testosterone using DPC’s kit (now Siemens) nor the free androgen index was a reliable parameter of free tes-tosterone concentration. The free testosterone calcu-lated from the testosterone and SHBG measurement was rapid, simple, and a reliable index of bioavailable testos-terone. It was comparable to the results from equilibrium dialysis and suitable for clinical use, except in pregnancy. Calculated nonspecifically - bound testosterone reliably reflected the measured non-SHBG-bound testosterone concentration. The website of the International Society for the Study of the Aging Male provides calculation of free testosterone from the SHBG and testosterone val-ues. It assumes a normal albumin level, but this may be changed if the albumin concentration is available. The calculation is not valid for very abnormal albumin concentrations.

Type of SampleSerum.

Frequency of UseFrequent.

ANDROSTENEDIONEAndrostenedione is secreted by the testis, ovary, and adre-nal. It is a weak androgen that is converted in peripheral tissue to the more potent testosterone. Therefore, increased secretion of androstenedione gives rise to hirsut-ism through this conversion (see Fig. 3).

FIGURE 2 Percentage of free testosterone against log SHBG. The correlation between the percentage of free testosterone, measured by steady-state gel filtration, and the log SHBG concentration measured by IRMA.

752 The Immunoassay Handbook

Reference IntervalWomen: 30–330 ng/dL (1.0–11.5 nmol/L)Men: 60–310 ng/dL (2.1–10.8 nmol/L)(Quoted for the Siemens Immulite 2000 assay)

Clinical Applications � Most requests for the determination of androstenedi-

one concentration are for the investigation of hirsut-ism. However, there is disagreement among clinicians about the usefulness of this measurement. Our own investigations suggest that only 3% of hirsute women with PCOS have an increased androstenedione con-centration alone (see Fig. 4).

� Apart from salt loss, the major problem of CAH is vir-ilization. Treatment of CAH should aim at maintain-ing androgen concentration at physiological levels in order to minimize further virilization and to allow puberty to develop more normally. Although the measurement of 17-OHP, androstenedione and total testosterone have all been used to monitor the treat-ment of patients with CAH, androstenedione seems the most appropriate. Its main source in CAH is from the adrenal, whereas testosterone is from peripheral conversion of androstenedione. It is well established that when androstenedione concentration is sup-pressed to physiological levels, 17-OHP remains at concentrations up to four times the upper limit of its reference interval in men and in women for the fol-licular phase. Thus, suppression of 17-OHP into the normal range leads to overtreatment with exogenous glucocorticoid.

� A deficiency of the 17β-hydroxysteroid dehydrogenase enzyme results in reduced testosterone synthesis by the testis. This is a rare condition associated with incom-plete masculinization in the male. Nearly, all affected neonates have been assigned female status. Theoreti-cally, androstenedione secretion by the testes will be increased but practically, a human chorionic gonado-tropin (hCG) test (see INFERTILITY) is required to uncover the defect. An increased ratio of androstenedione to tes-tosterone is diagnostic.

Limitations � An androstenedione result provides little additional

information in the investigation of hirsutism. Andro-stenedione is secreted by both the ovary and the adrenal, and the measurement of this steroid does not establish the source of increased androgen secretion.

� Levels in the neonate and the child are <0.9 ng/mL (<0.3 nmol/L), which is below the sensitivity of many assays. In addition, it is difficult to establish reference intervals for neonates, and few laboratories have expe-rience at accurately diagnosing 17β-hydroxysteroid dehydrogenase deficiency. Such investigations should be referred to an expert center.

Assay TechnologyOver 90% of the participants in the UK NEQAS use com-mercial kits with about 50% using the Siemens Immulite 2000 method. Spiking experiments by the UK NEQAS and comparison with tandem MS methods indicate that commercial kits overestimate androstenedione by between 10% and 20%. There are significant differences between methods so laboratories should establish a reference inter-val appropriate for the method they are using.

Types of SamplePlasma or serum.

Frequency of UseModerate.

DEHYDROEPIANDROSTERONE SULFATEEssentially all of the DHEAS in the circulation is derived from direct secretion by the adrenal glands. Because it is sulfated, it has a long half-life, and hence lacks a circadian rhythm (see Fig. 5).

Reference IntervalThe reference interval changes with age, peaking just after puberty, and gradually falling after the third decade of life (see Fig. 6 and Table 1).

FIGURE 3 Androstenedione.

FIGURE 4 A Venn diagram showing the percentage of patients with polycystic ovarian syndrome who had one or more abnormal androgen results. In only 3.1% of patients was androstenedione the only abnormal result, whereas 20.6% had a raised testosterone level. No hormonal abnormality, 14.2%; not every test done, 4.7%.

753CHAPTER 9.7 Hirsutism and Virilization in the Female

Clinical ApplicationAbout 10% of patients with PCOS have a concentration of DHEAS above the reference interval although some studies have reported up to 42% of patients. Because this suggests that the adrenal is the major source of increased androgen secretion, exogenous glucocorticoid may be used initially to treat such patients. Response to this treatment is variable.

Patients with Cushing’s syndrome due to an adrenal adenoma frequently have a DHEAS concentration below the reference interval.

Limitations � Although DHEAS concentration may be increased in

hirsute patients, this does not exclude increased andro-gen secretion by the ovary. In fact, studies have shown that it is very rare to have increased androgen secretion from the adrenal alone.

� A normal DHEAS concentration in a patient with Cushing’s syndrome does not exclude an adrenal adenoma.

Assay TechnologyBecause the concentration of DHEAS is 1000-fold greater than the other androgens, specimens are usually diluted by at least 100-fold before assay. Therefore, assays do not suf-fer from serum effects. Iodinated DHEAS is available, and several simple non-extraction assays have been developed commercially. However, most routine clinical laboratories use automated immunoassay methods, the majority the Siemens Immulite 2000 platform. There are an increasing number of laboratories using tandem MS, currently, seven participants in the UK NEQAS.

It has been reported that some assays appear to cross-react with as yet unidentified metabolites, leading to spuri-ously high results. No specific clinical condition has been associated with these metabolites.

Type of SampleSerum.

Frequency of UseInfrequent.

17α-HYDROXYPROGESTERONEGonadal and adrenal tissues are able to synthesize 17α-hydroxyprogesterone (17-OHP) from progesterone (see Fig. 7). 17-OHP is readily converted to 11-deoxycor-tisol in the adrenal and to androstenedione in the gonads and adrenal. Levels of the steroid rise in the late follicular phase and peak at the same time as estradiol. A second increase occurs in the luteal phase that is similar to proges-terone. Thus both the follicle and the corpus luteum appear to secrete 17-OHP during the menstrual cycle.

Reference IntervalFollicular phase: 0.55–1.84 ng/mL (1.7–5.7 nmol/L)Luteal phase: 0.55–6.31 ng/mL (1.7–19.6 nmol/L)

Clinical ApplicationThe measurement of serum 17-OHP is used for the diag-nosis of CAH caused by a deficiency of the 21-hydroxylase enzyme. Affected infants have a concentration greater than 9.7 ng/mL (30 nmol/L). Heterozygote individuals also have increased 17-OHP concentrations compared with

FIGURE 5 Dehydroepiandrosterone sulfate (DHEAS).

FIGURE 6 Diagrammatic representation of the changes in DHEAS levels throughtout life.

FIGURE 7 17α-Hydroxyprogesterone.

TABLE 1 Reference Intervals for DHEAS

Approximate ranges for post-puberty15–30 years 0.7–4.5 µg/mL (1.7–11.5 µmol/L)30–40 years 1.2–4.2 µg/mL (3.1–10.8 µmol/L)40–50 years 0.8–4.0 µg/mL (2.0–10.2 µmol/L)50–60 years 0.3–2.7 µg/mL (0.8–6.9 µmol/L)>60 years 0.2–1.8 µg/mL (0.4–4.7 µmol/L)

The graph and ranges are based on data from the author’s laboratory.The conversion factor used (2.56) is based on the use of the sodium salt, which is used as a standard.

754 The Immunoassay Handbook

the normal population. Identification of heterozygotes is important in genetic counseling.

Patients with LOCAH usually have 17-OHP concen-trations above the reference interval for the follicular phase of the menstrual cycle. A synacthen test will identify those patients whose results are equivocal. In affected patients, there is an exaggerated increase in the secretion of 17-OHP (more than three times the basal level).

Limitations � 17-OHP is mildly elevated in those cases of CAH that

are due to an 11β-hydroxylase deficiency. This enzyme abnormality should always be considered when hyper-tension is present, and the concentration of 17-OHP is only moderately increased.

� Interpretation of results is complicated in the first 24h of life because of maternal steroids still present in the neonate’s circulation.

� 17-OHP should not be used to monitor the effective-ness of treatment for CAH. If 17-OHP is suppressed into the normal range, it is likely that the patient is being overtreated. Androstenedione is the most appropriate analyte for monitoring treatment for CAH.

Assay Technology“In-house” extraction assays are common. Siemens manu-factures a direct kit that has a good correlation with the extraction assay for 17-OHP concentration in the adult. However, when using this kit, serum from neonates and children less than six months old should be extracted before the measurement of 17-OHP concentration due to the cross-reaction with other steroids present in the serum. Six of 53 participants use tandem MS in the UK NEQAS. There are no methods developed for automated platforms.

Cross-reaction with other steroids can confuse the interpretation of results when using immunoassay meth-ods. In the UK, a GC–MS is available through the Supra-Regional Assay Laboratory at the University College London Hospitals.

The investigation of CAH is usually made in neonates and prepubertal children where the collection of blood in sufficient quantity is difficult. Methods have been devel-oped to measure 17-OHP in saliva (see FURTHER READING) and from blood spots.

Types of SampleSerum, plasma, saliva, or blood spot.

Frequency of UseInfrequent.

DIHYDROTESTOSTERONEDHT is formed in peripheral tissue from testosterone by the enzyme 5α-reductase. Its biological activity, relative to testosterone, depends on the bioassay and animal species

used. However, it is generally regarded as the more potent androgen with a higher binding affinity for SHBG. Testos-terone is converted to DHT in the testes, skin, brain, sali-vary glands, lung, heart, and pectoral muscle.

Reference IntervalWomen: 377–725 pg/mL (1.3–2.5 nmol/L)Men: 116–435 pg/mL (0.4–1.5 nmol/L)(Derived by an radioimmunoassay (RIA) method after high performance liquid chromatography (HPLC) separa-tion of DHT from the other androgens.)

Clinical ApplicationSome studies suggested that the measurement of DHT might be helpful in the investigation of hirsutism. How-ever, because DHT is produced in many tissues, it is not a sensitive indicator of DHT production in the skin.

The concentration of DHT is low in both sexes, and it is difficult for most assays to achieve the required sensi-tivity. This is because it is difficult to avoid losses during the separation of DHT from testosterone and other androgens.

Assay TechnologyAntisera raised against DHT conjugates have a high cross-reaction with testosterone. Because attempts to produce a specific monoclonal antibody have been unsuccessful, tes-tosterone must be removed from patient sera before DHT can be measured.

DHT has usually been separated from testosterone by a chromatographic procedure after an initial solvent extrac-tion. These methods have the advantage of separating out all the androgens, which can then be measured by RIA. However, these procedures are usually limited to research studies.

A slightly simpler method uses potassium permanganate to oxidize the testosterone. Several companies now sell commercial kits. These include DSL (Beckman Coulter); Research Diagnostics Inc., New Jersey, USA; and IBL-Hamburg, Germany. The ELISA from IBL-Hamburg is a simple method with no oxidation or extraction steps. To achieve good precision and reliability in the extraction methods, a high degree of experience and skill is required.

Types of SampleSerum or plasma.

Frequency of UseUncommon.

ANDROSTANEDIOLS AND THEIR GLUCURONIDESDHT is rapidly converted to a variety of other compounds that include androsterone and the androstanediols, 5α-androstane-3α, 17β-diol (3α-diol) and 5α-androstane-3β, 17β-diol (3β-diol), by a number of tissues. These include the

755CHAPTER 9.7 Hirsutism and Virilization in the Female

accessory sex glands, skin, brain, salivary glands, and heart muscle. These steroids are further metabolized to their gluc-uronides and sulfates. All these steroids have been investi-gated as markers of hirsutism. The most abundant of these steroid metabolites is androsterone sulfate. However, Zwicker and Rittmaster (1993) showed that this steroid was not suitable as a marker of hirsutism. It correlated poorly with other androgens, and concentrations were not elevated in hirsute women. It was shown that the steroid is derived almost entirely from the adrenal glands. Little or no 3β-diol is formed in the human and so studies have focused on 3α-diol and its glucuronide (3α-diol G). The latter has been measured in both urine and serum although recent studies have measured 3α-diol G in serum. Many of these studies have reported that 3α-diol and 3α-diol G have a role in the investigation of hirsutism.

Reference Interval3α-Androstanediol GlucuronideAdult Males 3.4–22.0 ng/mLAdult Females

Premenopausal 0.5–5.4 ng/mLPostmenopausal 0.1–6.0 ng/mLHirsute 1.3–9.4 ng/mL

ARUP Laboratories RIA

Clinical ApplicationStudies have shown that DHT concentrations are infre-quently increased in idiopathic hirsutism. The 3α-diol was investigated as a better indicator of increased androgen activity in peripheral tissue as it is the end metabolite of androgen metabolism in the skin. Findings have been vari-able. Whereas one group of investigators reported increased 3α-diol concentrations in 93% of their patients with idiopathic hirsutism, other workers have found nor-mal results.

Interest switched to 3α-diol glucuronide (3α-diol G) when it was reported that this metabolite is predomi-nantly formed in sexual tissue and skin. Early studies showed that 3α-diol G concentration was greatly increased in hirsute women, whereas testosterone was only modestly increased. It was later shown that method-ological deficiencies in the early method caused errone-ously elevated results. More recent studies have shown that 3α-diol G is only modestly increased in hirsutism. In a detailed review Rittmaster (1991) examined studies that had looked at the origin and clinical use of 3α-diol and 3α-diol G. He concluded that the primary source of precursors for androgen conjugates appeared to be the adrenal glands and that the liver may be the major source of glucuronide conjugation. Therefore, he concluded that the clinical utility of 3α-diol G in the investigation of hirsutism was limited and androgen-dependent hir-sutism is best assessed by examining the distribution of body hair. Nevertheless there continue to be many stud-ies of PCOS and hirsutism where 3α-diol and 3α-diol G have been measured. Chen and et al. (2002) discuss these studies and their varied outcomes in a more recent review.

LimitationsResults from different studies have been inconsistent, cast-ing doubt on the usefulness of these assays. They probably add little to the information achieved from the measure-ment of other androgens. Azziz et al. (2000) concluded that the routine measurement of serum 3α-diol G is not recom-mended in the evaluation of idiopathic hirsutism or in other hirsute patients. The assays are not routinely available.

Assay TechnologyCommercial assays are available from Beckman Coulter and from ARUP Laboratories, Salt Lake City, UT 84108-1221, USA.

Type of SampleSerum.

Frequency of UseRare.

Further ReadingAzziz, R., Black, V.Y., Knochenhauer, G.A., Hines, G.A. and Boots, L.R.

Ovulation after glucocorticoid suppression of adrenal androgens in polycystic ovary syndrome is not predicted by the basal dehydroepiandrosterone sulfate level. J. Clin. Endocrinol. Metab. 84, 946–950 (1999).

Balen, A.H., Conway, G.S., Kaltsas, G., Techatraisak, K., Manning, P.J., West, C. and Jacobs, H.S. Polycystic ovary syndrome: the spectrum of the disorder in 1,741 patients. Hum. Reprod. 10, 2107–2111 (1995).

Bentley-Lewis, R., Koruda, K. and Seely, E.W. The metabolic syndrome in women. Nat. Rev. Endocrinol. Metab. 3, 696–704 (2007).

Broekmans, F.J., Visser, J.A., Laven, J.S.E., Broer, S.L., Themmen, A.P.N. and Fauser, B.C. Anti-Müllerian hormone and ovarian dysfunction. Trends Endocrinol. Metab. 19, 340–347 (2008).

Cardozo, E., Pavonne, M.E. and Hirshfield-Cytron, J.E. Metabolic syndrome and oocyte quality. Trends Endocrinol. Metab. 22, 103–109 (2011).

Chen, W., Thiboutot, D. and Zouboulis, C.C. Cutaneous androgen metabolism: Basic research and clinical perspectives. J. Invest. Dermat. 119, 992–1007 (2002).

Ekins, R. Measurement of free hormones in blood. Endocr. Rev. 11, 5–46 (1990).Goudas, V.T. and Dumesic, D.A. Polycystic ovary syndrome. Endocrinol. Metab.

Clin. North Am. 26, 893–912 (1997).Fears, T.R., Ziegler, R.G., Donaldson, J.L., Falk, R.T., Hoover, R.N., Stanczyk,

F.Z., Vaught, J.B. and Gail, M.H. Reproducibility studies and interlaboratory concordance for androgen assays in female plasma. Cancer Epidem. Biomarkers Prevent. 9, 403–412 (2000).

Goodarzi, M.O., Dumesic, D.A., Chazenbalk, G. and Aziz, R. Polycystic ovary syndrome: etiology, pathogenesis and diagnosis. Nat. Rev. Endocrinol. 7, 219–231 (2011).

Gower, D.B. Analysis of androgens and their derivatives. In: Steroid Analysis (ed Makin H.L.J. and Gower, D.B.), 457–558 (Springer, London, 2010).

Harbourne, L., Fleming, R., Lyall, H., Sattar, N. and Norman, J. Metformin or antiandrogen in the treatment of hirsutism in polycystic ovary syndrome. J. Clin. Endocrinol. Metab. 88, 4116–4123 (2003).

Harrison, S., Somani, N. and Bergfeld, W.F. Update on the management of hirsut-ism. Cleve. Clin. J. Med. 77, 388–398 (2010).

Ibanez, L., Potau, N. and Carrascosa, A. Insulin resistance, premature adrenarche, and a risk factor of the polycystic ovary syndrome (PCOS). Trends Endocrinol. Metab. 9, 72–77 (1998).

Jacobs, H.S. (ed), Polycystic Ovary Syndrome. Bailliere’s Clinical Endocrinology and Metabolism, vol. 10 (2). (Bailliere Tindall, London, 1996).

La Marca, A., Broekmans, F.J., Volpe, A., Fauser, B.C. and Macklon, N.S. Anti-Müllerian hormone (AMH): what do we still need to know? Hum. Reprod. 24, 2264–2275 (2009).

Martin, K.A., Chang, R.J., Ehrmann, D.A., Ibanez, L., Lobo, R.A., Rosenfield, R.L., Shapiro, J., Montori, V.M. and Swiglo, B.A. Evaluation and treatment of hirsutism in premenopausal women: an Endocrine Society clinical practice guideline. J. Clin. Endo. Metab. 93, 1105–1120 (2008).

New, M.L. and Speiser, P.W. Genetics of adrenal steroid 21-hydroxylase defi-ciency. Endocr. Rev. 7, 331–349 (1986).

756 The Immunoassay Handbook

Norman, R.J., Dewailly, D., Legro, R.S. and Hickey, T.E. Polycystic ovary syn-drome. Lancet 370, 685–697 (2007).

Pasquali, R., Stener-Victorin, E., Yildiz, B.O., Duleba, A.J., Hoeger, K., Mason, H., Homborg, R., Hickey, T., Franks, S., Tapanainen, J.S., Balen, A., Abbott, D.H., Diamani-Kandarakis, E. and Legro, R.S. PCOS forum: research in polycystic ovary syndrome today and tomorrow. Clin. Endocrinol. 74, 424–433 (2011).

Pucci, E. and Petraglia, F. Treatment of androgen excess in females: Yesterday, today and tomorrow. Gynaecol. Endocrinol. 11, 411–433 (1997).

Rittmaster, R.S. Androgen conjugates: Physiology and clinical significance. Endocr. Rev. 14, 121–132 (1993).

Rittmaster, R.S. Medical treatment of androgen-dependent hirsutism: clinical review. J. Clin. Endocrinol. Metab. 80, 2559–2563 (1995).

Rosner, W. An extraordinary inaccurate assay for free testosterone is still with us. J. Clin. Endocrinol. Metab. (Lett.) 86, 2903 (2001).

Rosner, W., Auchus, R.J., Azziz, R., Sluss, P.M. and Raff, H. Position statement: Utility, limitations, and pitfalls in measuring testosterone: An Endocrine Society position statement. J. Clin. Endocrinol. Metab. 92, 405–413 (2007).

Rosner, W., Hryb, D.J., Kahn, S.M., Nakhla, A.M. and Romas, N.A. Interactions of sex-hormone binding globulin with target cells. Mol. Cell. Endocrinol. 316, 79–85 (2010).

Speiser, P.W. and White, P.C. Congenital adrenal hyperplasia due to steroid 21-hydroxylase deficiency. Clin. Endocrinol. 48, 411–417 (1998).

Stewart, P.M., Penn, R., Holder, R., Parton, A., Ratcliffe, J. and London, D.R. The hypothalamo-pituitary-adrenal axis across the normal menstrual cycle and in polycystic ovary syndrome. Clin. Endocrinol. 38, 387–391 (1993).

Vermeulen, A., Verdonck, L. and Kaufman, J.M. A critical evaluation of simple methods for the estimation of free testosterone in serum. J. Clin Endocrinol. Metab. 84, 3666–3672 (1999).

Wallace, A.M. Analytical support for the detection and treatment of congenital adrenal hyperplasia. Ann. Clin. Biochem. 32, 9–27 (1995).

Wheeler, M.J. The determination of bio-available testosterone. Ann. Clin. Biochem. 32, 345–357 (1995).

Winters, S.J., Kelley, D.E. and Goodpaster, B. The analog free testosterone assay. Are the results useful? Clin. Chem. 44, 2178–2182 (1998).

Witchel, S.F. and Azziz, R. Congential adrenal hyperplasia. J. Pediatr. Adolesc. Gynecol. 24, 116–126 (2011).

Wood, P. Salivary steroids—research or routine? Ann. Clin. Biochem. 46, 183–196 (2009).

Zwicker, H. and Rittmaster, R.S. Androsterone sulphate: physiology and clinical significance in hirsute women. J. Clin. Endocrinol. Metab. 76, 112–116 (1993).