Published: August 2010
Polycystic ovary syndrome (PCOS) is the most commonly encountered endocrinopathy in women of reproductive age. It has significant reproductive and nonreproductive consequences.1 Women of any ethnic background can present with PCOS. In a prospective study of 400 women of reproductive age, 4% to 4.7% of white women and 3.4% of African American women had PCOS.2 A similar rate of 4% to 6% has been found in other populations. Because patients with PCOS can present with an assortment of complaints such as menstrual disturbances, infertility, hirsutism, and acne, their point of entry into the medical system may be by way of a primary care physician, gynecologist, endocrinologist, or dermatologist. Thus, all these disciplines need to be familiar with this syndrome and its long-term consequences.
Historically, there has been a lack of consensus regarding the features that define PCOS. A meeting convened by the National Institutes of Health (NIH) in 1990 stressed three key features necessary for the diagnosis of PCOS: The patient must exhibit ovulatory dysfunction; there must be evidence of clinical hyperandrogenism or biochemical hyperandrogenemia; and other disorders must be excluded, such as nonclassic congenital adrenal hyperplasia, androgen-secreting tumors, and hyperprolactinemia or thyroid disorders.
Because 16% to 25% of the normal population has polycystic-appearing ovaries on ultrasound,3 the presence of polycystic ovaries was considered to be suggestive but not diagnostic of PCOS. However, the Rotterdam European Society of Human Reproduction/American Society for Reproductive Medicine (ESHRE/ASRM)-Sponsored PCOS Consensus Workshop Group that convened in 2003 now requires the existence of two of the following three criteria to make the diagnosis of PCOS4: oligo-ovulation or anovulation, clinical or biochemical signs of hyperandrogenism, and polycystic ovaries.
Even after these criteria have been met, other potential causes of hyperandrogenism must be excluded. To exclude congenital adrenal hyperplasia, a basal early-morning level of 17-α hydroxyprogesterone is performed in the follicular phase. Total testosterone level may be evaluated to rule out an androgen-secreting tumor. Although luteinizing hormone (LH) levels are high in most women with PCOS, the Rotterdam consensus panel considered it unnecessary to assess the LH level routinely. Free testosterone and dehydroepiandrosterone sulfate (DHEAS) levels may be checked when clinical suspicion is high for PCOS but the patient does not exhibit clinical signs of hyperandrogenism.
The basic pathophysiologic defect is unknown in PCOS; however, because this disorder tends to cluster in families, a genetic cause is believed likely. Hyperandrogenism is a common finding in first-degree relatives of women with PCOS. The rates of PCOS in mothers and sisters of patients with PCOS were 24% and 32%, respectively.5 Increased insulin resistance has been noted in mothers and sisters of women with PCOS.6 Several large research groups are actively searching for a genetic cause of this syndrome. The mode of inheritance is uncertain at this point, and the role of shared environmental factors such as diet and lifestyle in the manifestation of the disease is unknown. Loci proposed and investigated as possible PCOS genes include CYP11A, the insulin gene, and a region near the insulin-receptor gene.7 The wide range of PCOS symptoms likely plays a significant role in the inability thus far to identify a specific gene mutation.
Although the pathophysiology of PCOS is not clear, a variety of biochemical abnormalities have been described with this syndrome. Hyperinsulinemia is noted in 50% to 70% of PCOS patients. It is defined as impaired action of insulin on glucose transport and antilipolysis in adipocytes in the presence of normal insulin binding. Hyperinsulinemia causes or exacerbates hyperandrogenemia. Increased insulin levels at the ovarian level lead to increased androgen production from the ovarian thecal cells. In an in vitro model, Nestler and colleagues8 demonstrated increased testosterone production by theca cells of women with PCOS exposed to insulin compared with testosterone production from the theca cells of normal women. Also, by suppressing hepatic production of sex hormone binding globulin (SHBG), insulin increases unbound levels of testosterone. At the level of the granulosa cell, insulin amplifies the response of granulosa cells to LH. Thus, these cells undergo abnormal differentiation and premature arrest of follicular growth, and thus anovulation.
Elevated androgen levels also lead to decreased levels of SHBG. Greater unbound androgen levels are likely to produce a greater clinical response, such as hirsutism and acne. Most patients with PCOS show evidence of clinical hyperandrogenism. However, a small fraction does not, which might warrant a biochemical investigation. In such cases, measurement of free testosterone should be considered, although most direct assays for free testosterone have limited value for evaluating the hyperandrogenic woman. The methods recommended at the consensus meeting to determine free testosterone are by equilibrium dialysis, by calculation of free testosterone from measurement of SHBG and total testosterone, or by ammonium sulfate precipitation. DHEAS may be measured, because a small percentage of patients with PCOS have isolated elevations in this hormone. However, there were not enough data for the consensus meeting to support the routine measurement of androstenedione.
Another key feature of PCOS is altered gonadotropin dynamics. Several studies have shown higher LH pulse and amplitude in women with PCOS.9 Although a higher LH level drives the ovarian theca cells to produce more androgens, insufficient follicle-stimulating hormone (FSH) may be the more immediate cause of anovulation. In most women with PCOS, LH levels are elevated or the LH/FSH ratio is high; however, the mean LH pulse amplitude is attenuated in obese women with PCOS.10 Thus, the LH value or LH/FSH ratio is not helpful in establishing this diagnosis in such patients.
PCOS symptoms have a gradual onset. Although symptoms of PCOS can exist at the time of menarche, most patients do not seek help until their early to mid 20s.
Abnormal vaginal bleeding is a typical complaint that ranges from amenorrhea to oligomenorrhea to menorrhagia and metrorrhagia. Because patients with PCOS experience irregular and incomplete endometrial shedding, endometrial hyperplasia and cancer must be aggressively ruled out; thus, an endometrial biopsy should be considered. Adults with prolonged episodes of bleeding unresponsive to treatment and those with irregular bleeding and long durations of amenorrhea should undergo endometrial biopsy regardless of their age. Other studies, such as ultrasound of the uterus and endometrial lining, can assist with this decision. About 15% to 30% of women with PCOS claim to have regular periods despite documented anovulation.
Because these patients are anovulatory, they present with infertility issues. They can also have increased incidence of pregnancy loss and pregnancy complications.11,12 Spontaneous abortion occurs in one third of all pregnancies in women with PCOS, which is double the rate of normal women. After pregnancy is established, perinatal mortality is increased at least 1.5 times.13 Pregnancy complications can include gestational diabetes, pregnancy-induced hypertension, and increased rate of intrauterine fetal demise. Currently, it is unclear whether PCOS independent of obesity leads to poor obstetric outcome. In a case-control study, Haakova and colleagues compared the pregnancy outcomes of a group of women with PCOS with those of a group of healthy weight-matched women. The investigators were unable to show a higher rate of pregnancy complications in the PCOS group.14
Depending on ethnic background, skin manifestations vary. Compared with Asian women, Middle Eastern or Mediterranean women are more likely to complain of hirsutism. Other women’s primary skin complaint may be acne. Excessive growth of terminal hair is noted on the upper lip, sideburn areas, chin, and periareolar, abdominal, back, and buttock areas. Signs of virilization are lacking. In some cases, there is evidence of alopecia, which can be extremely disturbing to patients. Other dermatologic findings include acanthosis nigricans and skin tags, which may be manifestations of insulin resistance.
Although the prevalence of obesity is high in patients with PCOS, in large series the rate varies. Depending on the series, the rate of obesity in the PCOS population ranges from 38% to 87%.15 Variables that can influence this number include diagnostic criteria and geographic and environmental factors. Because obesity is associated with insulin resistance, many women with PCOS have insulin resistance, but insulin resistance in PCOS is independent of obesity. Dunaif and colleagues showed that the insulin sensitivity of nonobese women with PCOS approached that of obese controls. Obese women with PCOS had higher insulin resistance than cycling obese women in the control group.16
PCOS patients are at higher risk for the metabolic syndrome, which is a group of cardiovascular risk factors that include dyslipidemia, type 2 diabetes mellitus, hypertension, and obesity. In a study of 254 women with PCOS, the prevalence of impaired glucose tolerance was 31% compared with 10.3% in controls, and the prevalence of type 2 diabetes was 7.5 % compared with 1.5% in controls.17 At the Rotterdam PCOS consensus workshop, three of the five criteria in Table 1 were considered necessary for a diagnosis of the metabolic syndrome.
|Abdominal obesity||>88 cm|
|High-density lipoprotein cholesterol||<50 mg/dL|
|Blood pressure||≥130/≥85 mm Hg|
|Fasting and 2-h glucose levels from oral glucose tolerance test||110-126 mg/dL and/or 2-hr glucose 140-199 mg/dL|
Adapted from Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group: Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19-25.
Treatment is typically directed at the manifesting symptoms of PCOS. Several treatment options are available, but few randomized trials exist to help provide a clear treatment guideline.
In conjunction with medical therapy, focus should be directed at lifestyle modification by diet and exercise if the patient is overweight. Even small amounts of weight loss have established menstrual cyclicity, increased spontaneous ovulation and pregnancy rates, and increased sensitivity to ovulation-induction medications.18 Also, the rate of miscarriages decreased and live births increased.19 A significant impact on the metabolic consequences of PCOS is also expected with weight loss, because insulin and triglyceride levels decrease and high-density lipoprotein cholesterol increases.
Oral contraceptive pills are typically the first line of therapy for management of irregular bleeding in women with PCOS who are not interested in conception. Cyclic withdrawal of estrogen and progesterone leads to complete endometrial shedding and resolution of most abnormal bleeding. Exposure to the progestin in oral contraceptives leads to reduction in the risk of endometrial cancer and hyperplasia. In addition, the steroids cause a decrease in LH levels and a subsequent decrease in androgen production. Finally, they also increase SHBG production, and the resulting decreased free testosterone levels lead to diminished hirsutism and acne.
If the degrees of hirsutism and acne are significant, antiandrogens are used. Although they are slow to show results (a minimum of 6-9 mo), a significant change in thickness of the hair shaft and reduction of sexual hair growth can be seen. Spironolactone, up to 200 mg/day, is the most common antiandrogen used in PCOS patients in the United States. Moghetti and colleagues20 demonstrated that there is no difference in the effectiveness of spironolactone, flutamide, or finasteride. Because of their possible impact on the development of a male fetus, they are rarely used alone. Usually they are given in conjunction with oral contraceptive pills.
Ovulation can be induced with clomiphene citrate in more than 80% of anovulatory patients. This drug is typically the first line of therapy, although pregnancy rates are only about 40%. Other agents that may be used to induce follicular growth and subsequent ovulation include insulin-sensitizing agents, exogenous gonadotropins, and aromatase inhibitors.
These agents increase tissue sensitivity to insulin action. Metformin is a biguanide drug that has traditionally been used to manage frank diabetes. It is believed to increase peripheral glucose uptake, and because it does not precipitate hyperinsulinemia, it does not cause hypoglycemia. A review of the literature21 shows that on average, body mass index decreased by 4% and androgen levels decreased by 20% with metformin compared with placebo. However, this does not translate into a significant impact on hirsutism.
Metformin regularizes menstrual cyclicity, typically within 3 months of initiating the drug. If ovulation has not been induced within 3 months of initiating metformin in anovulatory infertile women, clomiphene citrate is typically added to the regimen. Several studies show higher ovulation rates in patients on metformin and clomiphene citrate than rates achieved with placebo or with clomiphene citrate alone.22,23 Currently, there is no clear consensus on whether all patients with PCOS should initially undergo a trial of metformin to induce ovulation or whether they should take metformin only after clomiphene citrate has failed to induce ovulation. A 2003 Cochrane review on the use of insulin-sensitizing agents in patients with PCOS concludes that metformin is justified as first-line therapy for ovulation induction.24
Another area of potential controversy is the use of metformin solely for restoring menstrual cyclicity. Because metformin does not consistently lead to regular periods, it cannot be recommended as the first line of therapy for this symptom; oral contraceptives are still considered the first line of therapy for cyclic endometrial shedding. However, because oral contraceptives can exacerbate insulin resistance, metformin may be considered for morbidly obese women who are at significant risk for insulin resistance.
A review of the literature indicates that use of metformin in women with PCOS decreased fasting insulin levels, low-density lipoprotein, and blood pressure.24 However, because of the lack of long-term data, it is unknown whether these changes translate into a decreased likelihood of cardiovascular disease and diabetes. Hence other pharmacologic interventions may be necessary, as directed by an internist, to address the specific metabolic abnormality. Finally, metformin cannot be advocated as a replacement for exercise and diet modification. The Diabetes Prevention Program Research Group demonstrated that although metformin significantly reduces the frequency of type 2 diabetes in nondiabetic patients who have high serum glucose concentrations, exercise has a greater impact in reducing the incidence of diabetes.25
Metformin is used in doses of 1500 to 2000 mg/day. Gastrointestinal symptoms are its main side effects and include nausea, vomiting, diarrhea, and flatulence. These symptoms are typically transient. Gradual increase of the dose can decrease these side effects. This medication is contraindicated in women with hepatic or renal impairment and in those with conditions that increase the risk of lactic acidosis.
Laparoscopic ovarian drilling is the modern version of the ovarian wedge resection. The premise is destruction of stromal tissue leading to decreased androgen production and subsequent change in the hormonal profile. Testosterone levels decrease and LH levels seem to drop after ovarian drilling. Hirsutism has been noted to improve for up to 9 years. Ovulatory rates of up to 74% and pregnancy rates of 50% to 70% have been reported.26 Because these results are similar to those noted with metformin, it may be more prudent to proceed with metformin therapy initially in anovulatory women not at risk for tubal factor infertility. Laparoscopy carries the inherent risks associated with surgery. In addition, this procedure may be associated with subsequent periadnexal adhesions.
Awareness has developed of some of the long-term implications of PCOS that extend well into the postmenopausal years. The physician is responsible not only for offering medical treatment options but also for educating the patient regarding the possible long-term consequences of the syndrome.
Women with PCOS have chronic anovulation. The endometrial lining is exposed to unopposed estrogen for long durations. These patients are believed to be at increased risk for endometrial cancer. However, there is minimal epidemiologic evidence to support this theory.27
There is clear evidence that women with PCOS have an increased likelihood of impaired glucose tolerance and are at increased risk for type 2 diabetes mellitus.17 Women with PCOS who have a family history or are obese are at even greater risk for developing diabetes. Unfortunately, fasting glucose concentration is a poor predictor of diabetes in this population. However, impaired glucose tolerance is known to be a risk factor for the development of frank diabetes. Because the prevalence of impaired glucose tolerance and diabetes is high in obese PCOS patients,7 the Rotterdam PCOS consensus group advises performing a screening 2-hour oral glucose tolerance test in obese women.
Cardiovascular risk factors are increased in patients with PCOS when compared with weight-matched controls.28 High-density lipoprotein cholesterol levels are lower and triglyceride levels and very-low- density lipoprotein cholesterol are higher.29 Several studies on women with PCOS seem to demonstrate an increase in blood pressure. However, blood pressure is affected by a variety of factors that include weight, genetics, and stress. Thus, it is not clear whether lean insulin-resistant women with PCOS are also at higher risk for developing hypertension. An increase in coronary artery calcification has also been reported in women with PCOS.30
It has been very difficult to sort out the individual effects of PCOS and obesity on cardiovascular risk factors, because obesity can negatively affect blood pressure, the lipid profile, and insulin resistance. All these factors are believed to lead to a higher incidence of cardiac events in women with PCOS. However, despite these risk factors, there is no consistent epidemiologic evidence that these patients in fact do have an increased incidence of coronary heart disease.31,32