Enrico Carmina

The Importance of Diagnosing the Polycystic Ovary Syndrome

Hyperandrogenism and insulin were linked as early as 1921, when Achard and Thiers (1) published their classic description of a bearded women with diabetes (1). Since then, researchers have realized that most women with hyperandrogenism show evidence of a disorder known as the polycystic ovary syndrome (PCOS), which is extremely common but heterogeneous. Most women with PCOS have some degree of insulin resistance, although it may be subtle. Abnormalities of insulin secretion and action have been implicated in the pathophysiology of PCOS. For this and other reasons, diagnosis of PCOS is important. The polycystic ovary syndrome, then called the SteinLeventhal syndrome, was first described in 1935. Originally, diagnosis required pathognomonic ovarian findings and the clinical triad of hirsutism, amenorrhea, and obesity (2). The next diagnostic milestone occurred 30 years later, when researchers in the late 1960s and early 1970s noted derangements in the hypothalamicpituitary axis. This focused the diagnosis on endocrine criteria, such as elevated levels of serum luteinizing hormone or ratio of luteinizing hormone to follicle-stimulating hormone (3, 4). With the advent of pelvic ultrasonography in the 1970s and 1980s (first, abdominal sonography and, later, vaginal sonography), the recognition of a characteristic polycystic ovary complicated the diagnosis. Some believed that ultrasonographic findings alone were sufficient to make a diagnosis. In this setting, it was misleading to call the disorder polycystic ovary (PCO) or polycystic ovarian disease (PCOD) because these terms completely ignored the well-accepted endocrine features. It was also quickly realized that polycystic ovaries can occur in some normal women and in women with well-defined endocrinopathies as varied as hypothalamic amenorrhea and congenital abnormal hyperplasia (5, 6). In normal ovulatory women with no other typical endocrine features, we prefer to call this finding as polycystic-appearing ovary (7) to distinguish it from the term PCO, which is often used synonymously, but incorrectly, with PCOS. The polycystic ovarian syndrome is extremely prevalent and is considered the most frequently encountered endocrinopathic condition. It has been suggested that the disorder occurs in 4% to 7% of women of reproductive age (8, 9). During the reproductive years, PCOS is associated with important reproductive morbidity, including infertility, abnormal bleeding, increased pregnancy loss, and complications of pregnancy (10). Because women with PCOS also have an increased risk for endometrial carcinoma because of long-standing unopposed estrogen stimulation, their menstrual function must be continually monitored (11). In 1990, the National Institutes of Health formed a group to investigate PCOS. No consensus was reached regarding the naming of the disorder (12). However, in women who present with hyperandrogenism and chronic anovulation, a diagnosis of PCOS is considered reasonable after other endocrine disorders (for example, congenital adrenal enzymatic deficiencies and tumors) have been ruled out. Hyperandrogenism and chronic anovulation remain the two most characteristic clinical features of the disorder and are discussed separately. Criteria for the diagnosis of PCOS are presented in the Table. Table. Criteria for Clinical Diagnosis of the Polycystic Ovary Syndrome Hyperandrogenism Hyperandrogenism is a key feature of PCOS. Although the adrenal gland may contribute, hyperandrogenism is principally ovarian in origin among women with a primary diagnosis of PCOS. In various populations around the world (13), it has been found that most women with PCOS have elevated levels of serum androgens; however, normal levels may be found in some women. For a diagnosis of PCOS, it is sufficient to have elevated serum androgen levels or a biological expression of hyperandrogenism (acne or hirsutism). Serum testosterone level is the best marker for ovarian hyperandrogenism, and dehydroepiandrosterone sulfate is the best adrenal marker. It is recommended that these analytes be measured. The measurement of free testosterone provides a higher diagnostic yield for ovarian hyperandrogenism because levels of sex-hormone binding globulin are decreased. However, clinical assays used to test this measure vary considerably, affecting its reliability. It is important to point out that hyperandrogenemia is not synonymous with hirsutism or acne. Some ethnic groups (for example, Asians) have substantial hyperandrogenism (elevated levels of testosterone and dehydroepiandrosterone sulfate) without any significant skin manifestations (13, 14). Anovulation Anovulation in PCOS is usually chronic and presents as oligomenorrhea or amenorrhea of perimenarchial onset. Nevertheless, a history of regular menses is also possible. Because the degree and chronicity of self-reported menstrual irregularity vary, it has been the most difficult aspect of the diagnosis to define. Some women who report normal menses may be anovulatory. In a recent prospective survey of hyperandrogenic women who reported normal menses, we found that 21% experienced anovulation despite reporting regular menstrual cycles (15). It is important to note, however, that PCOS is extremely heterogeneous and that a small number of affected women may have ovulatory function. Although this contradicts our definition of PCOS, it is now accepted that a subset of women with the syndrome who have typical polycystic ovaries on ultrasonography also experience ovulation. The Ovarian Diagnosis The polycystic ovary is easily diagnosed. It is enlarged, usually greater than 9 mL with more than 8 mL peripherally oriented cystic structures (<10 mm) in a sonographic plane surrounded by an increased stromal mass (>25% of the ovarian volume) (16). However, although we and others have insisted on these strict criteria, a sonographic spectrum exists. Polycystic ovaries may sometimes be absent in women with all of the other classic clinical characteristics of PCOS. This may be related to the particular resolution of the ultrasonographic technique (for example, vaginal scans are much more sensitive than abdominal scans). We again emphasize that although the ultrasonographic diagnosis of polycystic ovaries is most common in women with PCOS, it may occur in women with other disorders and in healthy women. Ultrasonographic diagnosis alone is not sufficient to diagnose PCOS. Test for Gonadotropin-Releasing Hormone Agonists Some investigators believe that PCOS can be diagnosed by performing a test for gonadotropin-releasing hormone agonists and demonstrating an exaggerated ovarian response (17). This test is helpful because it can show that the ovary is producing more androgen in response to luteinizing hormone, but it is our view that it should be considered an adjunctive test. It is not specific enough to pinpoint the diagnosis, and it is impractical to perform. Insulin Resistance Many of the late complications of PCOS seem to be related to insulin resistance (18). Burghen and coworkers (19) first reported this observation in 1980, and we and many other groups later confirmed it (20, 21). Although severe insulin resistance, which is often associated with acanthosis nigricans (22), may be present in PCOS, most patients have only a mild form with slightly elevated fasting serum insulin levels. A ratio of fasting glucose to insulin is usually sufficient to diagnose insulin resistance (23), although this ratio is not valid in patients with overt glucose intolerance. It has been suggested that with more sophisticated techniques (for example, an intravenous glucose tolerance test or clamps), insulin resistance may be diagnosed in almost all women with PCOS. Although insulin resistance is associated with obesity, it is also found in normal-weight women with PCOS (13, 21, 24) and in women with PCOS from different ethnic groups (13). It has been suggested that anovulation is a major determinant of insulin resistance in women with PCOS. We recently confirmed this and have found that although all women with PCOS may have evidence of insulin resistance, it is more pronounced in those with chronic anovulation than in those who have ovulatory cycles. The pathogenesis of insulin resistance remains unclear. However, it has been reported that insulin resistance may be related to excessive serine phosphorylation of the insulin receptor in at least 50% of women with PCOS (25). Dysfunction of cells may also occur in women with PCOS, making them susceptible to type 2 diabetes mellitus (26). More recently, it was suggested that women with PCOS exhibit decreased action of chiroinositol, which is important for insulin signaling. Some patients have been successfully treated with d-chiroinositol (27). Impaired Glucose Tolerance and Diabetes Because of insulin resistance, all women with PCOS have increased risk for impaired glucose tolerance and overt type 2 diabetes mellitus. A recent study found that 31% of obese, reproductive-age women with PCOS had impaired glucose tolerance and that 7.5% had overt diabetes (28). In addition, 10.3% of nonobese women with PCOS had impaired glucose tolerance and 1.5% had diabetes, a rate almost three times that of the general population (28). In another study, longitudinal follow-up of women who had been treated with wedge resection showed that 16% developed type 2 diabetes mellitus by menopause (29). These and other data (30) show that women with PCOS have a high risk for diabetes and that this risk is similar in different populations and ethnic groups (28). Because of the known long-term morbidity associated with diabetes, even young women with PCOS should be followed closely for impaired glucose tolerance and diabetes. In our view, women should be screened for glucose intolerance with an oral glucose tolerance test. We suggest that this be done in obese women with PCOS before pregnancy is attempted and in all affected women with PCOS after 40

Polycystic ovaries in hirsute women with normal menses

PURPOSE Hirsute women with normal ovulatory menstrual function are often diagnosed as having idiopathic hirsutism. We prospectively evaluated 62 hirsute ovulatory women to determine if they had a subtle form of polycystic ovary syndrome, and if they exhibited any of the metabolic abnormalities commonly associated with classic polycystic ovary syndrome. METHODS Baseline hormonal profiles, ovarian responses to gonadotropin-releasing hormone agonist, and ovarian morphology by ultrasound were compared in the hirsute women and two groups of ovulatory controls. RESULTS Among 62 women, only 8 (13%) had normal androgen levels and were considered to have idiopathic hirsutism. Twenty-four (39%) had characteristic polycystic ovaries on ultrasound, an exaggerated response of 17-hydroxyprogesterone to leuprolide, or both, suggesting ovarian hyperandrogenism and the diagnosis of mild polycystic ovary syndrome. The remaining 30 women (48%) were considered to have unspecified hyperandrogenism. Age, body weight, and androgen level were similar among the hyperandrogenic subgroups. However, when compared with both normal and overweight controls and with patients with idiopathic hirsutism, the women who had mild polycystic ovary syndrome had higher fasting insulin levels [P < 0.01, mean (+/- SD) increase of 7 +/- 3 microU/mL], lower glucose-insulin ratios (P < 0.01, mean reduction of 3 +/- 1.5), higher low-density lipoprotein cholesterol levels (P < 0.05, mean increase of 26 +/- 10 mg/dL), and lower high-density lipoprotein (HDL) cholesterol levels (P < 0.01, mean reduction of 10 +/- 4 mg/dL). Compared with patients who had unspecified hyperandrogenism, these women also had higher fasting insulin levels (P < 0.05), lower glucose-insulin ratios (P < 0.05), and lower HDL cholesterol levels (P < 0.05). CONCLUSION These data suggest that mild polycystic ovary syndrome is more common than idiopathic hirsutism, and it is also associated with subtle metabolic abnormalities.

Do hyperandrogenic women with normal menses have polycystic ovary syndrome

OBJECTIVE To determine the prevalence of polycystic ovary syndrome (PCOS) among hyperandrogenic women who report normal menses. DESIGN Prospective observational study. SETTING Academic practice in reproductive endocrinology. PATIENT(S) Fifty-eight consecutively seen new patients with hyperandrogenism who reported normal menses. INTERVENTION(S) Ovulatory status was assessed with timed serum progesterone measurements. The following tests also were carried out: vaginal ultrasound examination; measurement of the ovarian response of 17-hydroxyprogesterone (17-OHP) after the administration of leuprolide acetate, 1 mg SC; and determination of fasting serum LH, FSH, E2, 17-OHP, insulin, and androgen levels. MAIN OUTCOME MEASURE(S) Determination of ovulatory status, polycystic appearance of ovaries, and increased response of 17-OHP to leuprolide acetate. RESULT(S) Twelve (20.7%) of the hyperandrogenic women were anovulatory and met the usual criteria for the diagnosis of PCOS. The ovulatory patients had lower serum total and unbound testosterone levels. Thirty-one (53.4%) of the ovulatory women had polycystic ovaries on ultrasound examination and/or an increased 17-OHP response to leuprolide acetate, suggesting the diagnosis of PCOS despite the presence of ovulation. Considering both the anovulatory and ovulatory patients, 74% of the hyperandrogenic women studied could have PCOS. CONCLUSION(S) The data suggest that most (74%) hyperandrogenic women who report normal menses have evidence for the diagnosis of PCOS.

Circulating leptin levels during ovulation induction : relation to adiposity and ovarian morphology

OBJECTIVE To assess serum leptin levels based on body habitus and ovarian morphology during controlled ovarian hyperstimulation. DESIGN Prospective analysis. SETTING University IVF program. PATIENT(S) Women undergoing IVF-ET were divided into two groups, obese ovulatory women (n = 6; mean (+/-SD) body mass index, 30.1 +/- 0.6 kg/m(2)) and lean ovulatory women (n = 20); mean (+/- SD) body mass index 22.0 +/- 0.2 kg/m(2)). Lean women were categorized further according to whether they had polycystic-appearing ovaries (n = 8) or normal-appearing ovaries (n = 12). INTERVENTION(S) Controlled ovarian hyperstimulation and IVF. MAIN OUTCOME MEASURE(S) Serum estradiol, testosterone, and leptin. RESULT(S) Mean (+/- SD) leptin levels were significantly higher before and after GnRH agonist down-regulation in obese women (41.7 +/- 5.2 pg/mL and 36.1 +/- 5.8 pg/mL, respectively) compared with lean women (8.4 +/- 1.0 pg/mL and 6.9 +/- 1.1 pg/mL, respectively). Mean (+/- SD) leptin levels increased significantly in both groups (54.5 +/- 5.1 pg/mL and 11.7 +/- 1.2 pg/mL, respectively), and the mean (+/-SD) percentage increase was similar (55% +/- 18% and 54.8% +/- 17%, respectively). Mean (+/-SD) leptin levels were similar in women with polycystic-appearing and normal-appearing ovaries before controlled ovarian hyperstimulation, but increased significantly in women with polycystic-appearing ovaries afterward (14.7 +/- 1.8 pg/mL and 9.3 +/- 1.0 pg/mL, respectively). CONCLUSION(S) Significant increases in leptin levels occur during controlled ovarian hyperstimulation, suggesting that leptin plays a role in follicular growth and maturation. The exaggerated response in women with polycystic-appearing ovaries reflects either a greater number of recruited follicles or a predisposition of adipocytes to leptin production.

Is the inappropriate gonadotropin secretion of patients with polycystic ovary syndrome similar to that of patients with adult-onset congenital adrenal hyperplasia?

OBJECTIVE To assess gonadotropin alterations in adult-onset congenital adrenal hyperplasia (CAH) and to compare these findings with those of patients with polycystic ovary syndrome (PCOS) in an effort to better understand the pathophysiology of these abnormalities. DESIGN Prospective study of 9 newly diagnosed patients with CAH, 10 with PCOS, and 10 ovulatory controls. INTERVENTIONS Baseline measurements of serum androgens, progestins, estradiol (E2), estrone (E1), unbound E2, luteinizing hormone (LH), and follicle-stimulating hormone (FSH). Serum LH and FSH were measured after intravenous gonadotropin-releasing hormone (GnRH) and in 15-minute blood samples for 6 hours to determine LH pulsatility. RESULTS Serum androgens were elevated but comparable in the two patient groups. Serum LH was also elevated (P less than 0.05) but was higher in PCOS than CAH. Serum LH:FSH ratios were similar as were the responses to GnRH. Serum E1 was elevated only in PCOS, but unbound E2 was elevated to the same degree in both PCOS and CAH (P less than 0.05). Patients with PCOS had a decreased LH interpulse interval compared with controls and CAH (P less than 0.05), but LH pulse amplitude was increased in both PCOS and CAH (P less than 0.05). Serum E2 and unbound E2 correlated significantly with LH (P less than 0.05), LH responses to GnRH as well as to LH pulse amplitude in CAH (P less than 0.05). The LH interpulse interval did not correlate with estrogen in any group. None of the LH parameters correlated with serum progestin levels in CAH. CONCLUSIONS The gonadotropin abnormalities of CAH appear to be intermediate between those of controls and PCOS. Although elevated estrogen may explain these abnormalities in CAH, additional factors may be operative in PCOS.

Hypogonadism and Hormone Replacement Therapy on Bone Mass of Adult Women with Thalassemia Major

We studied bone mass and metabolism in 30 adult women (age 28.5 ± 1.3) with thalassemia major (TM) and evaluated whether prolonged hormone replacement therapy (HRT) was able to optimize bone accrual. TM patients had reduced bone mass, increased bone turnover and lower serum gonadotropin and estradiol levels compared with 10 normal women of similar age. A significant correlation was found between bone mass and sex hormone levels. Six TM patients with normal ovarian function had normal bone turnover markers and modestly low bone mass (lumbar spine −1.29 ± 0.31; femoral neck −0.60±0.21; Z-score). The other 24 TM women were hypogonadic and had significantly lower bone mass for age (lumbar spine −2.35 ± 0.2, femoral neck −1.83 ± 0.2) and increased bone turnover relative to eugonadal women. Of the hypogonadal patients, 13 had taken HRT since age 15 ± 1 years, but their bone mass and turnover markers were not different than untreated hypogonadal patients. In conclusion, while hypogonadism negatively affects bone mass acquisition in adult TM women, HRT at the standard replacement doses is not sufficient to secure optimal bone accrual.

Reassessment of adrenal androgen secretion in women with polycystic ovary syndrome

Objective To reevaluate the clinical significance of elevations of adrenal androgens in polycystic ovary syndrome (PCOS). Methods Thirty women with PCOS and ten ovulatory controls were evaluated. Serum dehydroepiandrosterone (DHEA) sulfate and 11β-hydroxyandrostenedione were measured before and after 3 and 6 months of GnRH agonist (GnRH-A) therapy. All controls and 15 women with PCOS received intravenous ACTH before and after GnRH-A therapy Results Twenty-one (70%) of the women with PCOS had elevations of DHEA sulfate, and 16 (53%) had elevations in 11β-hydroxyandrostenedione. Only two women with PCOS had normal values of both adrenal androgens. After GnRH-A therapy, only 11 subjects (37%) had elevated values of DHEA sulfate. Four of 16 women had reductions in 11β-hydroxyandrostenedione. Only those with elevated baseline DHEA sulfate levels had reductions after GnRH-A therapy. The reduction of DHEA sulfate with GnRH-A correlated with the reduction in androstenedione. Of the subjects who had reductions in DHEA sulfate with GnRH-A therapy, there was a blunted response of DHEA to ACTH after treatment. Conclusion Our findings suggest that the ovary may influence the prevalence and magnitude of adrenal androgen excess in PCOS.

Ovarian suppression reduces clinical and endocrine expression of late-onset congenital adrenal hyperplasia due to 21-hydroxylase deficiency.

OBJECTIVE To determine the effectiveness of GnRH-agonist (GnRH-a) treatment in women with late onset congenital adrenal hyperplasia. DESIGN Prospective assessment of GnRH-a treatment in six women with documented late-on-set congenital adrenal hyperplasia who were not preselected. Comparisons were made to previous responses in the same patients receiving dexamethasone. Eight age- and weight-matched ovulatory women served as controls. SETTING Academic medical center. INTERVENTION Baseline blood determinations before and after i.v. ACTH, before and after 6 months of GnRH-a treatment. Estrogen and progestin replacement was begun in all women after the 3rd month of treatment. MAIN OUTCOME MEASURES Serum 17-hydroxyprogesterone (17-OHP), gonadotropin, and androgen levels before and after GnRH-a treatment. Responses of 17-OHP and androgens to ACTH assessment of hirsutism using a modified Ferriman-Gallwey score. RESULTS Gonadotropins, estrogen, androgen, and 17-OHP were suppressed with GnRH-a treatment. Levels were similar before and after estrogen and progestin replacement. Responses of 17-OHP after ACTH were blunted but still were elevated compared with responses in controls. Ferriman-Gallwey scores decreased significantly (-8 +/- 1; mean +/- SE). This response was greater than that observed previously with 6 months of dexamethasone (-2 +/- 0.3). CONCLUSIONS Suppression of the ovary with GnRH-a treatment was beneficial in these patients with late-onset congenital adrenal hyperplasia. An ovarian influence on the clinical and biochemical findings of the disorder is suggested.

The Addition of Dexamethasone to Antiandrogen Therapy for Hirsutism Prolongs the Duration of Remission

OBJECTIVE To determine whether the addition of dexamethasone to antiandrogen therapy prolongs the duration of remission in women with hirsutism. DESIGN Follow-up study of patients treated with one of four regimens: spironolactone (100 mg/d) for 1 year, dexamethasone (0.37 mg/d) for 1 year, dexamethasone (0.37 mg/d) plus spironolactone (100 mg/d) for 1 year, or dexamethasone (0.37 mg/d) plus spironolactone (100 mg/d) for 2 years. SETTING Academic medical practice in reproductive endocrinology. PATIENT(S) Fifty-four women with hirsutism and hyperandrogenism. INTERVENTION(S) Ferriman-Gallwey-Lorenzo scores were obtained and serum levels of testosterone, unbound testosterone, and dehydroepiandrosterone sulfate were measured before therapy, every 6 months during therapy, and for 1 year after the withdrawal of therapy. MAIN OUTCOME MEASURE(S) Hirsutism scores and serum indices of hormonal changes were monitored. RESULT(S) Ferriman-Gallwey-Lorenzo scores and androgen levels remained low 1 year after the withdrawal of therapy in patients who were treated with dexamethasone, either alone or in combination with spironolactone. In patients who were treated with spironolactone alone, hirsutism scores had returned to baseline values after 1 year. CONCLUSION(S) The addition of an agent that suppresses androgen levels may be useful to prolong the duration of remission of hirsutism in women with hyperandrogenism who are treated with antiandrogens.

Serum androsterone conjugates differentiate between acne and hirsutism in hyperandrogenic women

OBJECTIVE To determine if among hyperandrogenic women acne may be differentiated from hirsutism by markers of peripheral androgen metabolism. DESIGN Prospective outpatient study of 36 hyperandrogenic women and controls divided into groups based on the presence or absence of significant hirsutism and the presence or absence of moderate to severe acne. Serum levels of adrenal and ovarian derived androgens were elevated but similar in all patient groups. INTERVENTIONS Measurement of serum androgens including metabolites of 5 alpha-reductase activity: 3 alpha-androstanediol glucuronide and sulfate and androsterone (A) glucuronide and sulfate. RESULTS 3 alpha-androstanediol glucuronide and sulfate were elevated in all groups (P less than 0.05) and could differentiate between hirsute and nonhirsute patients but were similar in patients with and without acne. Serum A glucuronide and sulfate were only significantly elevated in patients with acne (P less than 0.01) and were higher than levels in controls and hirsute patients without acne. Ratios of precursor androgens to A glucuronide and sulfate were significantly higher in patients with acne compared with patients without acne (P less than 0.05). CONCLUSIONS Altered peripheral metabolism in acne may favor the formation of A conjugates, which may help differentiate acne from hirsutism among hyperandrogenic women.