Overview and Epidemiology of PCOS
About 5% to 10% of women of reproductive age have polycystic ovary syndrome. R. Jeffrey Chang, MD, professor and director, Division of Reproductive Endocrinology, Department of Reproductive Medicine, University of California-San Diego School of Medicine, La Jolla, explained that the hallmark features of PCOS, in addition to polycystic ovaries, are excessive hair growth, anovulation, and infertility.
The syndrome usually occurs at the onset of puberty, although it can appear in women who are in the middle of their reproductive years, and it may follow a familial inheritance pattern. “This condition, with its associated endocrine changes, is likely to pose significant long-term health risks,” said Dr. Chang.
Stein and Leventhal first recognized the syndrome in 1935, when they described seven patients with the characteristic signs, primarily anovulation and infertility; most of the women were hirsute, and some were obese (Am J Obstet Gynecol 1935;29:181-191). Over time, Stein and Leventhal reported on approximately 100 patients. “However, because of the wide variety of symptoms that seemed to be associated with the condition, researchers began to question whether a ‘Stein-Leventhal syndrome’ actually existed,” said Dr. Chang.
Hirsutism and Obesity
Several studies have found that about 30% of amenorrheic and 90% of oligomenorrheic patients have PCOS. Patients with regular menses and hirsutism have a significant prevalence of polycystic-appearing ovaries, and the percentage is even higher for those women with amenorrhea or oligomenorrhea.
In a study of women who considered themselves “normal”-that is, they had regular menses and no hirsutism-only eight of 123 had polycystic-appearing ovaries on ultrasound (Lancet 1988;i:870-872). Of women with irregular menstruation, 89% had the ultrasound criteria for this condition. Although none of the women with amenorrhea had polycystic ovaries, three had multicystic ovaries.
Insulin Impairment and Diabetes
Dr. Chang stated that he and two other investigators-Dr. Andrea Dunaif (chair of this program) and Dr. Ron Barbieri¾independently concluded that PCOS patients essentially had insulin resistance and abnormal insulin catalyst (presentation at the Society for Gynecological Investigation 1983 [abstract]).
From 20% to 40% of obese patients will develop impaired glucose tolerance (IGT) and eventually type II diabetes by the end of the fourth decade; however, not all women who develop this condition are obese. Patients with PCOS commonly have hyperinsulinemia, and therefore would also be at risk for developing IGT, whether lean or obese. “Patients with PCOS make up about 10% of patients with IGT,” said Dr. Chang, and about 7% of patients with IGT progress to type II diabetes each year. Thus, up to three million women in the United States are at risk for PCOS and diabetes.
A large number of patients with noninsulin-dependent diabetes develop frank insulin-dependent diabetes, and the risk for cardiovascular disease is greater in patients with glucose metabolism abnormalities. PCOS patients, who often have abnormal lipid profiles, appear to be at risk for cardiovascular disease. “Something may be inherent in PCOS that poses serious long-term risks to the overall health of these patients,” concluded Dr. Chang.
Diagnostic Dilemmas in PCOS
Rogerio A. Lobo, MD
According to Rogerio A. Lobo, MD, chair of the Department of Obstetrics and Gynecology, Columbia Presbyterian Medical Center, New York City, a different term is needed to describe the complex of reproductive symptoms in women currently referred to as “PCOS.” Dr. Lobo prefers “hyperandrogenic chronic anovulation,” a term that encompasses the two major symptoms and avoids the distinction between the presence and absence of polycystic ovaries. Whatever it is called, PCOS is not technically a “disease,” which has certain characteristic features, but is more accurately a “syndrome,” a set of symptoms or a symptom complex.
Polycystic ovaries, defined by Adams, Paulson, and Franks as ovaries that are 9 cc or greater on sonogram, can occur in “normal” ovulatory women (Br Med J 1986;293:355-359). Studies have shown that about 16% to 25% of normal females have polycystic ovaries, which are also found in about 24% of women with hypothalamic amenorrhea, up to 50% of women with hyperprolactinemic amenorrhea, and virtually 100% of women with congenital adrenal hyperplasia.
Diagnostic Criteria: Consensus or Not?
Seven years ago, a National Institutes of Health (NIH) meeting of PCOS experts attempted to compile a list of criteria thought to be important in diagnosing the syndrome (Current Issues in Endocrinology and Metabolism. Boston: Blackwell Scientific Publications, 1992). However, little agreement was reached. Only about half of the participants agreed on any specific criteria. These included hyperandrogenemia, menstrual dysfunction, elevated luteinizing hormone (LH), and polycystic ovaries on ultrasound.
Dr. Lobo, who attended the NIH meeting, stated that the participants reached only a “general agreement” that they were not dealing with a single disease entity. “At most, they came up with a soft, clinical, extremely heterogeneous definition of the syndrome based on a number of different criteria, not just one, such as the appearance of the ovaries on ultrasound,” said Dr. Lobo.
A prospective, three-continent study examined how PCOS varied among Japanese, Italian, and American women (Am J Obstet Gynecol 1992;167:1807-1812). All had hyperandrogenism and chronic anovulation. The Japanese women had normal body weight, the Italian women were close to normal body weight, and the American women were overweight. The Japanese women typically were not hirsute, while the Italian and American women were. About 75% of all patients had polycystic ovaries on ultrasound. Each group had elevated LH and testosterone levels, and 80% demonstrated insulin resistance.
Another prospective study of women in whom the clinical diagnosis of PCOS was based on hyperandrogenism and chronic anovulation revealed elevated LH in only 70% to 75% (Fertil Steril 1983;39:674-678). Therefore, “requiring that an elevated LH be present to diagnose PCOS will exclude 25% of the population,” Dr. Lobo said.
A study by Bridges and colleagues showed that the emergence of polycystic ovaries starts at age six or seven and peaks at puberty, with a prevalence of around 25% (Fertil Steril 1993;60:456-460). “However, this does not mean that 25% of girls have PCOS,” said Dr. Lobo. “Polycystic ovaries in a woman who is destined to have this problem emerge very early, but this should not be confused with PCOS, a syndrome in which this symptom may be only one component, albeit an important one.”
Drawing the Line
How abnormal do menses have to be before a patient is considered anovulatory? About 30% of patients with PCOS have normal menses. But it is unusual for a woman who has the clinical syndrome, and perhaps polycystic ovaries, to have consistently normal ovulatory menses. Many women who report normal menses have anovulatory cycles.
The majority of patients with the “full-blown clinical disorder” have polycystic ovaries on ultrasound to some degree. This syndrome also encompasses a spectrum of various degrees of endocrine manifestations. “Whereas only 10% or fewer patients with PCOS have normal ovaries, virtually 100% of patients with severe adrenal androgen excess manifest polycystic ovaries,” said Dr. Lobo.
“So where do you draw the line?” Dr. Lobo asked. “How severe do the symptoms need to be before a patient can be said to have PCOS? Diagnosis is in the eyes of the beholder.” If a patient has no external manifestations of androgen excess and normal ovulatory menses, regardless of whether she has polycystic ovaries, she should not be considered to have PCOS, according to Dr. Lobo. “However, even slightly abnormal androgen levels, occasional irregular periods, and an ultrasound finding of polycystic ovaries are sufficient to diagnose PCOS,” he concluded.
Insulin Resistance and PCOS
Andrea Dunaif, MD
PCOS is an endocrine syndrome. Echoing Dr. Lobo, Andrea Dunaif, MD, director, Women’s Health, Brigham and Women’s Hospital, Boston, and program chair, said “It is unfortunate that we have named it after an ovarian morphology, but I think we’re stuck with it.”
A History of Insulin Resistance
The association between insulin resistance and hyperandrogenism was first described in 1921 by Achard and Thiers (Bull Acad Nat’l Med 1921;86:51-64). Diagnosticians again became interested in this association when Kahn and Flier reported on adolescent girls with extreme insulin resistance, diabetes mellitus, and true virilization (“type A syndrome”) (N Engl J Med 1976;294:739-745).
In 1974, Givens noted that patients with acanthosis nigricans had high insulin levels (J Clin Endocrinol Metab 1974;38:347-355), and Dr. Dunaif and associates have found that many PCOS patients, both obese and lean, have acanthosis nigricans. This raised, velvety skin lesion, which may or may not be hyperpigmented, is a marker for insulin resistance.
In addition to glucose metabolism, insulin regulates protein and lipid synthesis, as well as gene transcription. In PCOS, an abnormality occurs in how the insulin receptor transmits signals, a very early abnormality compared with type II diabetes. No mutations are present; some researchers believe that a serine kinase turns off the insulin receptors.
In 1980, Burghen and colleagues were the first to show that women with PCOS were hyperinsulinemic and that this was not necessarily associated with body weight (J Clin Endocrinol Metab 1980;50:113-116). In a prospective study, Dr. Dunaif subgrouped hyperandrogenic women by ovulatory status (J Clin Endocrinol Metab 1987;65:499-507), basing the diagnosis of PCOS on hyperandrogenism and anovulation. Whether obese or lean, women with PCOS had elevated insulin responses to oral glucose. The ovulatory hyperandrogenic women did not have hyperinsulinemia.
Women with PCOS are at increased risk for developing IGT and type II diabetes mellitus, both risk factors for cardiovascular disease. “And the risk rises linearly with two-hour glucose levels,” said Dr. Dunaif. She estimated that about 10% of IGT in premenopausal women is related to PCOS, and internists often find histories of PCOS in diabetic patients.
Androgens and Insulin Resistance
Does hyperandrogenism cause hyperinsulinemia, or does hyperinsulinemia cause hyperandrogenism? Researchers looked for an answer by suppressing insulin levels. Nestler did this first using diazoxide, and subsequent researchers have used metformin, weight loss, and troglitazone, an insulin sensitizer that reduces circulating insulin levels (J Clin Endocrinol Metab 1989;68:1027). A study by Dr. Dunaif using troglitazone (discussed by Dr. Berga, below) showed increased insulin sensitivity that resulted in regular menses as early as one month later.
Despite conflicting results, studies have noted a connection between androgens and insulin resistance. For example, lowering androgen levels in PCOS patients who are not severely insulin resistant or obese improves insulin sensitivity. If virilizing doses of testosterone are given to women (as with female-to-male transsexuals), decreased insulin sensitivity results, although not of the magnitude seen in women with PCOS.
Hyperandrogenism causes some insulin resistance, and insulin resistance causes moderate hyperandrogenism, although this does not completely explain the association between insulin resistance and PCOS. Studies have shown that when insulin levels decrease, adrenal androgen, dehydroepiandrosterone sulfate (DHEAS), estrogen, and LH levels also decrease, suggesting that insulin is a “general augmentor” of steroidogenesis and LH secretions.
“The most compelling evidence indicates genetic susceptibility,” said Dr. Dunaif. Polycystic ovary morphology is transmitted within families, and hyperandrogenism with regular ovulation, as well as PCOS and insulin resistance, also are familial. About 50% of sisters of PCOS women have some form of the syndrome, either hyperandrogenism alone or associated with chronic anovulation.
One hypothesis is that an abnormal serine kinase in PCOS phosphorylates the insulin receptor and the rate-limiting enzyme for androgen synthesis, P450c17. Insulin resistance then augments hyperandrogenism, and androgens augment insulin resistance, producing the syndrome. A second hypothesis is that some women are genetically predisposed to polycystic ovaries and some to insulin resistance. When both genes are present, they develop PCOS (Endocr Rev 1997; 18:774-800).
Dr. Dunaif concluded by advising physicians to assume that a woman with PCOS is insulin resistant, noting that diagnosing insulin resistance requires very sophisticated invasive tests.
New Treatments for PCOS: Current and Future Options
Sarah L. Berga, MD
When treating patients with PCOS, physicians must consider its complex pathogenesis. “Devising an effective treatment plan takes an extraordinary effort,” said Sarah L. Berga, MD, associate professor, Division of Reproductive Endocrinology, Departments of Obstetrics, Gynecology, and Psychiatrist, University of Pittsburgh.
Determining Patients’ Needs
“Patients frequently present when they are quite young, and it is important to understand their concerns and to tailor the plan to fit their needs,” said Dr. Berga. The patient’s desires-for example, for conception, contraception, or cosmetic treatments for obesity or acanthosis nigricans-should be ascertained first.
Some patients with IGT may understand that they are at risk for diabetes, but most do not, especially if they are young. The challenge is to convey this without frightening them. “The physician needs to serve as a motivator and educator,” Dr. Berga said.
Among the therapies currently available are oral contraceptives; gonadotropin-releasing hormone (GnRH) agonists; hormone replacement therapy; antiandrogens, such as spironolactone; ovulation-inducing agents for patients seeking fertility; electrolysis; diet; exercise; laparoscopic wedge resection; empiric use of glucocorticoids; and a new oral antidiabetic agent designed to specifically target insulin resistance, troglitazone.
Troglitazone decreases insulin resistance by targeting nuclear peroxisome proliferator activated receptors (PPARs) to reduce fatty acid and glucose output and reduce triglyceride synthesis in the liver while increasing glucose uptake in skeletal muscle-a fundamental problem for women with PCOS. It has no effect on pancreatic beta cell insulin secretion and does not cause lactic acidosis.
Metformin decreases production and uptake of glucose without causing hypoglycemia, but it does rarely cause lactic acidosis, particularly in those with impaired renal function; it also frequently causes abdominal discomfort. Studies have attempted to determine which effects of metformin are caused by weight loss and which are caused by its impact on glucose disposal and insulin dynamics.
In one study, Ehrmann and colleagues treated PCOS in 14 women with metformin, 850 mg three times a day for 12 weeks (J Clin Endocrinol and Metabol 1997;82:524-530). Body mass index (BMI) was unchanged, and insulin secretion was not improved. Free testosterone, stimulated LH, follicle-stimulating hormone (FSH), 17-hydroxy progesterone, androstenedione, DHEAS, progesterone, and estradiol levels were unchanged, suggesting that this is not a promising drug for use in PCOS women with insulin resistance.
Ehrmann also treated 13 obese, hyperandrogenic women with PCOS and IGT with 400 mg troglitazone a day for 12 weeks (J Clin Endocrinol and Metabol 1997;82:2108-2116). Fasting and two-hour glucose levels declined significantly, with reduced hemoglobin A-1C, improved insulin sensitivity, decreased androgenic hormones, and stimulated 17-hydroxyprogesterone, without any change in gonadotropins.
In a three-month, double-blind, randomized study by Dr. Dunaif and colleagues, 25 obese, hyperandrogenic women with PCOS were given troglitazone 200 mg or 400 mg a day (J Clin Endocrinol and Metabol 1996;81:3299-3306). Within a month, BMI was unchanged; insulin sensitivity increased; and free testosterone, estradiol, and estrone decreased. The improvement in insulin sensitivity resulted in regular menses as early as one month later. In those given 400 mg per day, androstenedione and LH decreased and sex hormone-binding globulin (SHBG) increased; FSH levels were unchanged, but two women ovulated.
Uses for Troglitazone
Troglitazone, used either alone or with other therapies, can ameliorate insulin resistance and cardiovascular risk; may reduce hirsutism, acne, and acanthosis nigricans; and may potentiate ovulation and eumenorrhea.
Because it is new, criteria need to be developed for its use. For example, should troglitazone be given for insulin resistance alone or for known hyperglycemia? What are the long-term side effects in women anticipating childbearing? And how should women with PCOS be treated?
In women with PCOS, troglitazone may reduce triglyceride levels and blood pressure, and therefore cardiovascular risk, as demonstrated in a study by Dr. Berga and colleagues (Arterioscler Thromb Vasc Biol 1995;15:821-826). However, because oral contraceptives raise triglyceride levels, the results of concomitant troglitazone use are hard to predict.
A study by Fulghesu and associates demonstrated that women at greatest risk for ovarian hyperstimulation syndrome (OHSS) have high insulin levels, which have been linked to OHSS in women with PCOS (J Clin Endocrinol and Metab 1997:82:644-648). If exercise and dietary methods fail, many physicians recommend pretreatment with an GnRH agonist or oral contraceptive. Now, they also have the choice of troglitazone.
Questions remain about troglitazone’s use to facilitate ovulation. Although conversion of excess androgens to excess estrogens may suppress FSH, it is unknown to what extent troglitazone use would help.
Dr. Berga concluded: “The response to troglitazone will clearly allow us to dissect out more about the interacting causes of this disorder. It may become a primary therapy to restore ovulation.”