Introduction

Polycystic ovary syndrome (PCOS) affects 6–10% of reproductive age females, making it the most common metabolic endocrine condition in this population [2]. PCOS was first officially described in 1935 by Drs. Stein and Leventhal, resulting in its eponymous name, the Stein–Leventhal syndrome. The syndrome they described were patients who presented with amenorrhea, hirsutism, polycystic-appearing ovaries, and often obesity. They described the operation known as a “wedge resection,” which was often followed by regular menses and sometimes pregnancy. On pathological examination of tissue from the ovarian “wedge resection,” they noted multiple subcapsular cysts associated with a thickening of the tunica [3].

Later, this constellation of signs and symptoms became known as PCOS, which reflected the multiple subcapsular follicular cysts. Since 1935, our understanding of this syndrome has expanded and in 2012 an evidence-based methodology workshop was sponsored by the National Institutes of Health. At this workshop one recommendation was to change the name from PCOS to more closely reflect the complex pathophysiology of this condition, which would improve accurate communication and progress in both the clinical and research arenas [4].

Nature vs. nurture

PCOS is a complex condition characterized by anovulation and androgen overproduction affecting reproductive age women. This phenotype can result in infertility due to ovulatory dysfunction, irregular menses, endometrial hyperplasia, endometrial carcinoma, iron deficiency anemia due to heavy menses, hirsutism, virilization, miscarriage, insulin resistance, glucose intolerance, gestational and type 2 diabetes mellitus, hyperlipidemia, and the potential for increased risk of cardiovascular disease. Other conditions associated with PCOS include obesity, sleep apnea, anxiety, depression, and possible reduction in quality of life.

Studies have demonstrated a strong genetic component, supported by family and twin studies, with an important environmental contribution. Recent studies, especially in the rhesus monkey and sheep, suggest that exposure to elevated androgen levels in utero may predispose a woman to the development of a polycystic ovary type of phenotype [5]. In humans, it is possible that exposure to excess androgen at any time during development prior to the onset of puberty could increase the risk of hyperandrogenic, chronic anovulation [6]. This reflects the developmental plasticity of the hypothalamic–pituitary ovarian and adrenal axes.

PCOS is often accompanied by the metabolic syndrome, with an estimated >50% of patients who are overweight or obese. With the current epidemic of obesity these numbers will most likely continue to increase. The metabolic syndrome is characterized by central obesity, insulin resistance, and increased risk of developing type 2 diabetes, dyslipidemia, and hypertension. In 1992, David J. Barker proposed the thrifty phenotype theory. This theory proposes that intrauterine growth restriction accompanied by reduced fetal growth places the individual at higher risk of obesity, hypertension, type 2 diabetes, and cardiovascular disease. This theory proposes that alterations in fetal nutritional supply result in limited fetal growth or that fetal glucocorticoid exposure as a stress response may play a critical role in the pathogenesis [7]. As the metabolic syndrome often accompanies PCOS, it is possible that understanding the effects of fetal exposures, including adrenal androgens and glucocorticoids, may help unravel the enigmatic origin of PCOS.

Diagnosis

The diagnosis of PCOS has been imprecise and has varied over time, which has resulted in difficulty for research as well as clinical management of this condition. These issues have resulted in three consensus conferences to arrive at diagnostic criteria to improve diagnosis and communication. In 2003, the Rotterdam consensus conference published their criteria, which included two out of the following three: oligo- or anovulation, clinical or biochemical evidence of androgen excess, or polycystic appearing ovaries. “Polycystic ovaries” were defined as the presence of 12 or more follicles measuring 2–9 mm in diameter or a total ovarian volume of >10 mL. There are a couple of important caveats about the ultrasound appearance of polycystic ovaries. First, this definition only requires one ovary to meet the criteria. The patient should not be on oral contraceptive pills. If there is a dominant follicle or corpus luteum cyst, then the scan needs to be repeated in a future menstrual cycle, or if there is a cystic mass on the ovary further investigation is necessary. Additionally, the endocrine phenocopies of PCOS, such as androgen-producing ovarian or adrenal tumors, exogenous androgen administration, congenital adrenal hyperplasia especially late-onset 21-hydroxylase deficiency, hyperprolactinemia, thyroid dysfunction, Cushing’s syndrome, or acromegaly, must be ruled out [8,9].

In 2006, the Androgen Excess-PCOS Society published criteria for the diagnosis of PCOS to highlight the importance of androgen excess and ovulatory dysfunction as central and critical to this syndrome. In this consensus definition, the patient must have evidence of androgen excess, either clinically or biochemically, accompanied by ovulatory dysfunction and/or polycystic ovaries, while excluding other androgen excess disorders [10]. The recent National Institutes of Health Evidence-Based Methodology Workshop on Polycystic Ovary recommended maintaining the Rotterdam 2003 diagnostic criteria for PCOS as well as explicitly identifying the PCOS phenotype: androgen excess with ovulatory dysfunction, androgen excess with polycystic ovary morphology, ovulatory dysfunction with polycystic ovary morphology, and androgen excess with ovulatory dysfunction and polycystic-appearing ovaries [4]. The diagnosis of PCOS in adolescents is particularly challenging, with the suggestion that the diagnosis only be made when all three of the Rotterdam criteria are present: irregular menses or amenorrhea for at least 2 years after menarche, documented hyperandrogenemia, and that ultrasound should document increased ovarian volume >10 cm³ [11].

Metabolic syndrome is characterized by central obesity, hypertension, dyslipidemia, insulin resistance, and a proinflammatory and prothrombotic state, which increases the risk for developing cardiovascular disease and type 2 diabetes mellitus. The National Cholesterol Education Program’s Adult Treatment Panel III identified the following criteria, and required three of the five to make the diagnosis of metabolic syndrome in women: central obesity with a waist circumference of >88 cm or 35 inches, HDL cholesterol of <50 mg/dL, triglycerides of ≥150 mg/dL, blood pressure of ≥130/85 mmHg, and fasting glucose of ≥100 mg/dL [12].

Evaluation

The evaluation of patients for PCOS should be directed at confirming the diagnostic criteria are met, as well as ruling out PCOS phenocopies. The importance of a careful history and physical examination with supportive laboratory data needs to be emphasized. In the history and physical examination, one wants to evaluate for evidence of clinical hyperandrogenism, menstrual abnormalities, and any evidence of the metabolic syndrome (Table 3.1).

On physical examination, it is important to obtain height and weight, calculate the body mass index (BMI), measure the waist circumference, and obtain a resting blood pressure. The BMI is the mass in kilograms divided by the height in meters squared and is given in units of kg/m². Underweight is defined as a BMI <18.5, normal weight 18.5 to 25, overweight is 25 to 30, and obesity is >30 kg/m².

The skin examination can be supportive of a diagnosis of metabolic syndrome as well as PCOS by looking for evidence of hyperandrogenism, insulin resistance, and PCOS phenocopies, such as the stigmata of Cushing’s syndrome. Signs and symptoms of hyperandrogenism include acne, hirsutism, and evidence of male pattern balding. The Ferriman–Gallwey score is a common method used to determine the presence and severity of excess hair growth in women. The modified Ferriman–Gallwey score determines the presence and amount of hair growth in nine different areas including the upper lip, chin, chest, upper and lower back, upper and lower abdomen, upper arms, and thighs. Each area is given a score from 0, with no growth of terminal hair to 4, which is extensive growth of terminal hair. In Caucasian women a score > 8 is consistent with excess androgen. For other ethnic groups one should use an ethnically appropriate standardized scale [13]. The presence and rapid progression of virilizing signs should alert one to the possibility of more severe androgen excess: androgen-producing ovarian and adrenal neoplasms, cases of more severe insulin resistance, such as hyperandrogenism (HA), insulin resistance (IR) and acanthosis nigricans (AN), or HAIR-AN syndrome, or exogenous androgen exposure as with inadvertent exposure to testosterone gel [14]. Virilization is defined by the presence of increased muscle mass, clitoromegaly, lowering of pitch of the voice, hirsutism, frontal balding, and ovulatory dysfunction.

The skin is also a good site for evidence of insulin resistance. Patients with insulin resistance may have skin tags, acanthosis nigricans, and varying degrees of lipodystrophy. Skin tags or acrochordons are benign lesions which may be sessile or pedunculated, smooth or irregular, and their color may vary from skin-color to dark. Acanthosis nigricans is a darkening, thickening of the skin usually in intertriginous zones like the nape of the neck, axilla, and medial thigh. Lipodystrophy is a metabolic abnormality in fat metabolism, which results in loss of subcutaneous fat and this can be a localized or more generalized phenomenon.

Tanner staging of the breasts can help determine whether there was normal puberty with adequate levels of estrogen to induce breast development. Tanner staging of pubic and axillary hair can also determine whether there is an adult pattern and extent of hair distribution. The pelvic examination can provide additional information, including evidence of clitoromegaly or an ovarian mass, which would suggest an androgen-producing ovarian neoplasm. The pelvic examination can give evidence of adequate estrogenization. In a female with PCOS and unopposed estrogen production, one would expect to discover normal external genitalia with normal vaginal rugae and thick stratified squamous, vaginal epithelium. In the presence of unopposed estrogen there will be abundant, clear cervical mucus with spinnbarkeit. Under progesterone’s influence, the cervical mucus becomes scant, thick, and tacky. In the absence of estrogen, the vaginal walls shrink and become flat-surfaced and pale pink in color. In the presence of estrogen, vaginal pH is usually less than 4.5 because of glycogen metabolism by vaginal bacteria, while in the absence of estrogen the pH is usually >4.5 due to loss of this lactic acid production. A careful physical examination should be able to determine whether the patient has been exposed to estrogen during puberty (Tanner stages) and whether there is current estrogen production by the ovaries.

The physical examination must also evaluate for PCOS mimics: including thyroid dysfunction, galactorrhea, late-onset or nonclassic 21-hydroxylase deficiency (congenital adrenal hyperplasia), acromegaly, and Cushing’s syndrome.

Laboratory evaluation

After a thorough history and physical examination, the physician will have developed a differential diagnosis, which helps to focus appropriate laboratory testing in the patient with suspected PCOS. Many of these patients will present with absent or irregular menses. The initial evaluation of the patient with absent or irregular menses includes a pregnancy test, thyroid function tests, and a serum prolactin. This is accompanied by evaluation of the patient’s estrogen status. As noted above, physical examination can often give insight into whether estrogen is present or not. One laboratory test that can help determine estrogen status is the Vaginal Maturation Index. The Vaginal Maturation Index is performed by obtaining a scraping from the lateral vaginal wall, and microscopically determining the percentage of basal, para-basal, and superficial epithelial cells. Under the influence of estrogen, there will be normal maturation of the vagina and one will find more superficial cells, while in the absence of estrogen there will be primarily basal and para-basal cells.

The most common method for determining estrogen status is using the patient as her own in vivo biological assay, the progesterone challenge, in which progesterone is administered and she is observed for the onset of a withdrawal bleed. In the patient with chronic, hyperandrogenic anovulation (PCOS), the administration of progesterone is expected to induce a withdrawal bleed. If the patient does not have a withdrawal bleed, then the differential diagnosis includes primary ovarian insufficiency, müllerian agenesis and other absent or obstructive abnormalities of the uterus and outflow tract, hypogonadotropic hypogonadism, pregnancy, congenital adrenal hyperplasia with elevated progestin levels (17-hydroxyprogesterone), and occasionally with markedly elevated testosterone with its accompanying hypoestrogenism.

The presence of a progestin withdrawal bleed in this setting is consistent with normogonadotropic anovulation (WHO group II) [15]. In this case, further testing is aimed at discovering the etiology of this normogonadotropic, anovulatory state. Thyroid function tests are obtained in the initial part of the evaluation, because hypothyroidism can cause ovarian dysfunction, but rarely hyperandrogenism.

Prolactin is an important hormone in human reproduction and abnormally elevated levels can disturb ovulation, often inducing hypogonadotropic hypogonadism. There are a number of situations which can elevate prolactin levels, including stress, diet, chronic chest stimulation, drugs, and a prolactin-secreting, pituitary adenoma. As prolactin is often elevated by stress and diet it is important to draw the level in the fasting, relaxed state. In a patient discovered to have an elevated prolactin, repeating the blood draw in the fasting state, after 30 minutes of rest following placement of an intravenous heparin lock to allow for an atraumatic blood draw, will assist in determining whether the elevation is due to stress or diet. One must also remember the high-dose hook effect and macroprolactinemia as causes of spurious results. The high-dose hook effect can occur in a one-step, immunometric, sandwich assay when there is a large amount of analyte present in the unknown sample. In this situation there is such a large amount of analyte that one molecule binds to the labeled antibody and a different molecule binds to the solid phase antibody, which results in no sandwich forming and no signal. This high-dose hook effect can be detected by a two-step sandwich assay, in which there is a washing step prior to addition of the second antibody. In this case, one would detect saturation of the assay, where all of the initial antibodies are bound by analyte, then any free analyte is washed off, followed by signal due to formation of multiple sandwiches in the assay. The high-dose hook effect can also be detected by diluting the unknown specimen and noting that the results are not linear in relation to the dilution. Macroprolactinemia occurs when prolactin molecules are bound together by endogenous antibodies to form big prolactin and big-big prolactin. These macrocomplexes still have antigenic determinants, so they can be detected in the immunoassay, but have limited biological activity. Such macrocomplexes can be precipitated by addition of polyethylene glycol to the sample, and if >50% of the prolactin activity precipitates out, it is consistent with macroprolactinemia.

The most important circulating androgen is testosterone. The adrenal androgens, dehydroepiandrosterone sulfate (DHEA-S) and androstenedione, primarily function as pre-hormones, which have to undergo conversion to testosterone prior to biological activity. Testosterone is highly protein bound with only 1–2% in a free state and 60–70% tightly bound to the high-affinity sex hormone-binding globulin, and the rest loosely associated with albumin. In adult females, most testosterone assays can differentiate between the normal range and hyperandrogenism. However, these same assays are inaccurate in quantifying the degree of hyperandrogenism. In women, free testosterone levels, including the free androgen index (FAI), correlate well with the degree of hyperandrogenism. One should suspect an androgen-producing ovarian tumor in women with markedly elevated testosterone levels (defined as >2.5 to 3 standard deviations above the upper limit of the normal range or a concentration of total testosterone >200 ng/dL in a premenopausal woman), especially if present in women with new-onset, rapidly progressing hirsutism or virilization [16]. Most androgen-producing ovarian tumors are palpable on pelvic examination. Pelvic ultrasound remains the gold standard for imaging the ovaries for the presence of an ovarian neoplasm. In androgen-producing adrenal tumors, there are usually marked elevations in DHEA-S (8 µg/mL) and other adrenal androgen precursors [17].

There are some important points in evaluating the patient with hyperandrogenism. First, one needs to have a high index of suspicion for an androgen-producing neoplasm in the patient with new-onset, rapidly progressing hirsutism or virilization. Second, androgen-producing neoplasms may episodically secrete androgens, so multiple sampling times may be necessary to detect abnormally high androgens. Third, other precursor hormones may be secreted by the neoplasm and be reflected in elevated circulating levels. An example is androstenedione, which may be markedly elevated in androgen-producing neoplasms. Lastly, some androgen-producing tumors will present with lower levels of androgens, which is especially important to remember in evaluating the postmenopausal patient who presents with new-onset, rapidly progressing hirsutism or virilization. On the other hand, there are patients with PCOS and hyperthecosis with markedly elevated androgens, and no evidence of an androgen-producing neoplasm.

Conditions that can mimic PCOS besides androgen-producing neoplasms include late-onset congenital adrenal hyperplasia due to 21-hydroxylase deficiency, Cushing’s syndrome, and acromegaly. Late-onset, or nonclassic, 21-hydroxylase deficiency is due to a partial deficiency in 21-hydroxylation resulting in hyperandrogenism in female patients, and can be tested for by drawing an early morning serum 17-hydroxyprogesterone level. A follicular phase, early morning, basal 17-hydroxyprogesterone level of <3 ng/mL is normal. If the 17-hydroxyprogesterone level is ≥ 3 ng/mL then further testing with an ACTH stimulation test is indicated. In the ACTH stimulation test, ACTH 250 µg is given intravenously and blood drawn for 17-hydroxyprogesterone levels at 0 and 60 minutes. These blood levels are then compared to nomograms developed by Dr. New to determine the existence and degree of 21-hydroxylase blockade [18]. A common cause of an elevated 17-hydroxyprogesterone level is the inadvertent drawing of the sample during the luteal phase, as this hormone is made by the corpus luteum.

Cushing’s syndrome can be screened for with a 24-hour urine free cortisol or by administering an overnight dexamethasone suppression test. The 24-hour urine free cortisol is an excellent screening tool, but is sometimes difficult to obtain and may be incomplete. Determination of 24-hour urine creatinine can help determine whether the collection is a complete specimen. For the overnight dexamethasone suppression test, dexamethasone 1 mg is given at 11 p.m. with a serum cortisol drawn the next morning at 8 a.m. In a normal individual, one expects the a.m. cortisol level to be <1.8 µg/dL. An elevated 24-hour urine free cortisol and elevated cortisol after overnight dexamethasone suppression can be due to Cushing’s or pseudo-Cushing’s syndrome, which requires further differentiation. Pseudo-Cushing’s syndrome is a heterogeneous group of conditions including obesity, depression, anorexia nervosa, alcohol withdrawal, and poorly controlled diabetes. One method that may discern between Cushing’s and pseudo-Cushing’s syndrome is the maintenance of diurnal cortisol’s variation in the latter condition. In the patient with clinical signs and symptoms as well as supporting laboratory evidence for Cushing’s syndrome, further testing is necessary to confirm and determine the source of excess cortisol [19].

Acromegaly is an often insidious, slowly progressing condition with common signs and symptoms making it difficult to diagnose until it has significantly progressed. Symptoms include a coarsening of facial features, enlarged hands and feet, excessive perspiration, muscle weakness often accompanied by fatigue, acrochordon, snoring and sleep apnea, deepening of the voice, headaches, visual changes, and irregular menstruation in women. These patients have an increased frequency of insulin resistance and type 2 diabetes mellitus [20]. Insulin-like growth factor-1 (IGF-1) is often used as a screening test for the presence of acromegaly [21]. In acromegaly, the IGF-1 level is often elevated, but it can also be raised in pregnancy. IGF-1 levels decline with age and in patients with poorly controlled disease of the liver and kidney, and diabetes. When a patient has an elevated IGF-1 level, a more provocative test is performed to confirm the diagnosis of acromegaly. The glucose tolerance test, in which 75–100 g of glucose is given to the patient, should suppress growth hormone secretion to less than 1 ng/mL in a healthy individual. In patients with acromegaly the growth hormone will not suppress during the oral glucose tolerance test [22].

In the evaluation of irregular menses, follicle-stimulating hormone (FSH) and luteinizing hormone (LH) levels will often be drawn. FSH levels are often drawn to rule out primary ovarian insufficiency. In PCOS, one may find an inverted LH:FSH ratio frequently being >2:1 or 3:1 in about 50% of patients. In the past, the inverted LH/FSH ratio was often used as one of the diagnostic markers of PCOS [23–25].

Anti-müllerian hormone (AMH) is a glycoprotein of the transforming growth factor beta (TGF-β) family made by the granulosa cells of the primary, secondary, preantral, and small antral follicles, which are still gonadotropin-releasing hormone (GnRH) independent. The concentration of AMH has been correlated with ovarian reserve. Recent studies suggest that elevated levels of AMH may be predictive of PCOS [26].

Pelvic ultrasound, preferably with a vaginal probe transducer, is important in hyperandrogenic patients to exclude ovarian neoplasms as well as determine the presence of polycystic appearing ovaries. The Rotterdam consensus conference defined a polycystic ovary as one having ≥12 follicles of 2 to 9 mm in diameter or an ovarian volume of >10 cm³. The ovarian volume is calculated using the formula for a prolate ellipse: volume = 0.5 × length × width × thickness. These standards require that only one ovary has to meet the criteria, that the patient should be scanned in the follicular phase, and not on oral contraceptive pills (OCPs). It is also important to use state-of-the-art equipment [27,28].

Should women with PCOS be screened with an oral glucose tolerance test and lipid panel? In women with PCOS, as many as 40% will have evidence of glucose intolerance and by the fourth decade 10% will have type 2 diabetes mellitus [29]. Some recommend that we should only screen those at high risk for developing diabetes: increased BMI, history of gestational diabetes, family history of diabetes, or membership in a high-risk ethnic group [30]. The oral glucose tolerance test is a 75 g glucose load, followed by blood draws in the fasting state and 2 hours after ingestion of the glucose. Normal levels are fasting plasma glucose less than 100 mg/dL and 2-hour glucose level below 140 mg/dL. Fasting glucose levels between 100 and 125 mg/dL are diagnosed as having impaired fasting glycemia, while 2-hour levels between 140 and 200 mg/dL are diagnosed as having impaired glucose tolerance. Diabetes is diagnosed when the fasting level is ≥126 mg/dL and 2-hour glucose levels are ≥200 mg/dL [31]. A fasting lipid panel can also be helpful in diagnosing the metabolic syndrome (HDL cholesterol of <50 mg/dL, triglycerides of ≥150 mg/dL) [11].

Women with PCOS and ovulatory dysfunction have a higher frequency of developing endometrial cancer. Endometrial cancer is the most common gynecologic cancer in the USA, with increasing frequency with advancing age, but it has recently been seen increasingly in younger women [30]. In those 13 to 18 years of age, the development of endometrial cancer is rare, but one should consider further evaluation with an endometrial biopsy in the patient with prolonged abnormal bleeding, obesity, no other etiology for the bleeding, and failed medical treatment. For those between the ages of 19 and 39, endometrial cancer is a relatively unusual diagnosis, but consider an endometrial biopsy if there has been a prolonged period of unopposed estrogen, failed medical therapy, or a family history of endometrial cancer or applicable cancer syndrome. If the endometrial biopsy is non-diagnostic, further testing with hysteroscopy and possible dilatation and curettage may be necessary. In women over the age of 40, especially those over age 45, an endometrial biopsy should be performed for abnormal uterine bleeding [32–34].

Sleep apnea is a common complaint in PCOS and does not appear to be explained solely by obesity, but may be more correlated with insulin resistance. Common signs and symptoms of sleep apnea include daytime sleepiness, long pauses in breathing, loud and chronic snoring, as well as choking or gasping during sleep. Other minor symptoms include waking with a headache, dry or sore throat, insomnia, waking up feeling short of breath, and moodiness, irritability or depression [35,36]. If the patient has symptoms suggestive of sleep apnea, then a formal sleep study with polysomnography can assist in making the diagnosis.

Recent studies have demonstrated an increased incidence of depression and anxiety in women with PCOS. Deeks et al. showed that in women with PCOS, 29% met criteria for depression, especially if infertile, and 57% met the criteria for anxiety. Of these women, few were diagnosed and treated for these conditions, highlighting the importance of screening for anxiety and depression [37]. To evaluate their psychological condition, the Beck Depression Inventory or the Hospital Anxiety and Depression Scale can be administered to patients with PCOS [38,39].