Polycystic Ovary Syndrome
Ralph R. Kazer
Table Of Contents
Ralph R. Kazer, MD
In 1935, Stein and Leventhal1 described seven amenorrheic women with enlarged, polycystic ovaries who resumed cyclic menses after ovarian wedge resection. Ten years later, in a progress report in the American Journal of Obstetrics and Gynecology, Stein2 emphasized for the first time the association of hirsutism and obesity with the original clinical stigmata and so defined what came to be known as Stein-Leventhal syndrome. Subsequently the presence of enlarged polycystic ovaries alone was regarded as the sine qua non for making the diagnosis of what today is known as polycystic ovary syndrome (PCOS) or polycystic ovarian disease.
After the development of radioimmunoassay techniques for accurately measuring serum levels of peptide and steroid hormones, it became clear that women with PCOS share several distinct biochemical features. Compared with concentrations present in eumenorrheic women in the early follicular phase of the menstrual cycle, affected women frequently have elevated serum levels of luteinizing hormone (LH) and normal or low levels of follicle-stimulating hormone (FSH).3 Characteristic alterations in circulating sex-steroid concentrations also are noted.4 The androgens androstenedione, testosterone, and less frequently the adrenal androgens dehydroepiandrosterone (DHEA) and its sulfate (DHEA-S) are elevated. Women with PCOS also have elevated levels of circulating estrone (E1) in the face of early follicular phase levels of estradiol (E2).
At one time, PCOS was diagnosed on the basis of “elevated LH-to-FSH ratios” or “inverted E1-to-E2 ratios.” The discovery that many women who possess all the biochemical features of PCOS sometimes have morphologically normal ovaries, are of normal weight, and are not hirsute led to increasing confusion over how to define PCOS. No universally accepted set of clinical and biochemical criteria have been established yet to define this syndrome. It is described best as a state of chronic anovulation associated with an LH-dependent ovarian overproduction of androgens. Clinically the perimenarchal onset of symptoms has been described by several leading investigators as a prerequisite for diagnosis. The underlying pathophysiology of PCOS remains an enigma despite the unrelenting efforts of three generations of gynecologists and reproductive endocrinologists. This chapter summarizes current understanding of PCOS and provides an up-to-date diagnostic and therapeutic approach to the chronically anovulatory woman with clinical or biochemical signs of hyperandrogenism.
Grossly the ovaries of most women with PCOS are enlarged and globular, usually bilaterally, and possess smooth, glistening capsules, often described as similar to an oyster shell in appearance and color. The tunica albuginea is typically thickened, and numerous cysts several millimeters in diameter can be appreciated below the capsule on cut section, leading to the term polycystic ovary. This latter morphologic abnormality can be appreciated easily on vaginal ultrasonographic evaluation. Other chronic anovulation states, such as hypothalamic amenorrhea, may be associated with similar alterations, however, and ultrasound should not be employed as a primary diagnostic modality. Corpora lutea are seen rarely, reflecting the infrequency of ovulation. At one time, the thickened capsule was considered a mechanical barrier to ovulation, but this view has been discarded in light of the success of various medical modes of ovulation induction.
Histologically, typical findings include the presence of multiple follicles in various stages of maturation and atresia. The follicular cysts usually are lined by a thin layer of granulosa cells and are surrounded by a thickened theca interna. The latter cells are frequently luteinized. When islands of luteinized theca cells are found scattered throughout the ovarian stroma in addition to their typical perifollicular location, the term hyperthecosis sometimes is used. The clinical syndrome corresponding to this histologic finding typically reflects particularly excessive androgen production and massive obesity. It is likely that PCOS and hyperthecosis represent a spectrum of disease clinically and pathologically rather than two distinct entities.
No primary defect in steroidogenic capacity in the polycystic ovary has been documented, although women with PCOS respond acutely to gonadotropin-releasing hormone (GnRH) agonist administration with a pattern of sex-steroid release consistent with an increase in the activity of ovarian 17α-hydroxylase and C-17,20 lyase.5 This abnormality is discussed more fully later. The relative deficiency in aromatase activity that results in increased androgen production and subsequent inhibition of follicular growth can be corrected by increasing FSH stimulation in the course of various ovulation induction regimens. There is no evidence for the presence of an enzymatic defect in ovarian steroidogenesis in PCOS analogous to those seen in the various forms of congenital adrenal hyperplasia. Finally, women with PCOS do not seem to have a primary defect in inhibin secretion that might explain the characteristic abnormalities of gonadotropin secretion.6
All women with PCOS produce androgens at an increased rate compared with normal women, but the extent to which they manifest this biochemical abnormality in excessive male pattern hair growth or hirsutism is variable. The response of the hair follicle to androgen stimulation largely depends on its capacity to convert testosterone to 5α-dihydrotestosterone (DHT) locally. It is the intracellular activity of this hormone rather than that of testosterone per se that induces local androgenic activity. Testosterone is converted to DHT by the enzyme 5α-reductase. The disparity between circulating levels of androgens, in particular those of testosterone, and the appearance of hirsutism depends primarily on differences in local 5α-reductase activity. A peripheral marker for local production rates of DHT, its metabolite 3α-androstanediol glucuronide, has been identified and seems to correlate closely with the degree of hirsutism.7
Women with PCOS infrequently present with signs of overt virilization. Manifestations of markedly elevated androgen levels, such as clitoromegaly, temporal balding, or voice changes, should alert the clinician to the possible presence of an androgen-producing tumor, particularly if the development of such features took place over a short period. Women with PCOS are invariably well estrogenized, with normal breast development and copious cervical mucus on pelvic examination.
Obesity is not a universal feature of PCOS. Estimates vary widely, but perhaps half of women with PCOS significantly exceed their ideal body weight. It is unlikely that obesity per se plays a crucial role in the cause of PCOS, although some of the biochemical features that characterize obese women may reinforce indirectly the state of chronic anovulation (see later discussion).
Oligomenorrhea and Infertility
In their original article, Stein and Leventhal1 noted that before their detection of a relationship between the presence of enlarged polycystic ovaries and amenorrhea, other investigators had described a pattern of abnormal bleeding associated with endometrial hyperplasia presumably caused by “excess of secretion of estrogenic hormone.” Goldzieher and Axelrod,8 in an exhaustive literature review, found that about half of women with PCOS present with amenorrhea, about 30% with “functional bleeding,” and 12% with “cyclic menses.” No particular pattern of menstrual bleeding typifies women with PCOS, although a history of oligomenorrhea (fewer than four bleeding episodes per year) is probably most common. Women with PCOS may ovulate spontaneously on occasion. It is difficult to ascertain the frequency with which bleeding in women with PCOS is associated with a prior ovulation, although it is probably low. For practical reasons, essentially all studies designed to examine women with PCOS use oligomenorrhea or amenorrhea as a selection criterion. This fact and the inclusion of amenorrhea in the features of the classic Stein-Leventhal syndrome have tended to obscure the broad spectrum of bleeding patterns that may be seen in PCOS. The clinician must not rule out the diagnosis of PCOS on the basis of a history of frequent bleeding. Goldzieher and Axelrod’s review8 revealed that only three quarters of women with PCOS were infertile; this is consistent with the fact that women with PCOS spontaneously ovulate on occasion. Finally, clinical evidence of chronic anovulation is generally of perimenarchal onset.
Inappropriate Steroid Secretion
The steroids androstenedione and testosterone are secreted in significant amounts by the adrenal gland and by the ovary. Serum levels of both hormones typically are elevated in women with PCOS. In an elegant experiment, using an agonist of GnRH to effect selective inhibition of ovarian steroidogenesis in women with PCOS, Chang and colleagues9 showed suppression of testosterone and androstenedione to castrate levels. These data indicate that the increased secretion of androgens in PCOS is due to abnormally high, gonadotropin-dependent, ovarian secretion. The stromal compartment and the theca interna cells have been identified as the principal sites of androgen production in the polycystic ovary.
DHEA and DHEA-S are derived primarily from the zona reticularis of the adrenal gland in normal women. Serum DHEA-S levels are elevated in about half of women with PCOS.10 They remain elevated in the face of prolonged GnRH agonist administration and are suppressible with exogenous glucocorticoids, indicating that in PCOS as well, they are almost entirely of adrenal origin. The failure of DHEA-S levels to normalize after the long-term elimination of the ovarian contribution to sex-steroid production suggests that the abnormality is not a consequence of the abnormal steroid milieu that characterizes PCOS.9 The circulating concentrations of DHEA and DHEA-S are the manifestations of a complex interplay between production rates, clearance rates, and mutual interconversion. In men, hyperinsulinemia increases the metabolic clearance rate of DHEA-S, but this does not seem to be the case in either normal cycling women or women with PCOS.11,12,13,14,15
E2 levels are variable in PCOS but are usually in the early follicular phase range of 30 to 70 pg/mL. Individual values in the 100 to 200 pg/mL range are commonly seen, particularly in women who ovulate intermittently. Serum levels of E1 are usually elevated in PCOS. This abnormality has been traced to enhanced peripheral aromatization of androstenedione to E1 in the face of increased availability of androstenedione. Some investigators have suggested that elevated E1 levels contribute to the maintenance of the state of chronic anovulation in PCOS.3
As shown in Figure 1, progestins serve as precursors for androgen production. In one study, a group of PCOS subjects with elevated basal levels of DHEA-S had elevated basal serum levels of pregnenolone, 17OH-pregnenolone (17OH-Pe), and 17OH-progesterone (17OH-P) but normal progesterone levels. Selective suppression of the ovary with a GnRH agonist suggests that 17OH-P is primarily of ovarian origin, whereas pregnenolone and 17OH-Pe seem to be derived from the adrenal and the ovary.16 In another study that entailed frequent blood sampling over a 24-hour period, PCOS subjects with normal basal levels of DHEA had normal 24-hour integrated values for the four androgens, with the exception of 17OH-P, which was elevated.17
SEX HORMONE–BINDING GLOBULIN.
E2 and testosterone circulate bound to sex hormone–binding globulin (SHBG), a β-globulin synthesized in the liver and also known as testosterone-E2 binding globulin, and to a lesser extent to albumin. Only 2% to 4% of the steroids are normally in an unbound, or free, state. The non–SHBG-bound hormones are thought to be more “biologically free” or available, although this issue is controversial. SHBG levels are abnormally low in PCOS, and consequently the fraction of SHBG-bound E2 or testosterone also is decreased. As a result, the hyperandrogenemia that characterizes PCOS is amplified at peripheral sites, frequently resulting in oily skin, acne, and hirsutism. The decreased SHBG and increased non–SHBG-bound testosterone present in PCOS also may explain the finding of normal total circulating testosterone levels in a few affected women. Although testosterone production is increased, lower SHBG levels result in an increased metabolic clearance rate, leading to normal circulating testosterone levels. Although the decrease in SHBG levels observed in PCOS has been attributed in the past to a suppressive effect of increased circulating androgen concentrations, the increased circulating insulin levels typical of PCOS seem to be primarily responsible for the observed reduction in SHBG levels.18
Inappropriate Gonadotropin Secretion
Serum levels of LH are elevated frequently in women with PCOS. In a given woman, LH levels may vary dramatically, however, over short periods of time, and it is probably misleading to characterize individual patients with PCOS as having “high LH” or “normal LH” concentrations on the basis of a single blood sample. Because LH is secreted in a pulsatile manner from the anterior pituitary in response to the pulsatile release of GnRH from the hypothalamus, many investigators have sought to elucidate the mechanism of enhanced LH secretion in PCOS in terms of specific abnormalities of its pulsatile release. Although some have described an abnormally high pulse frequency in PCOS, the evidence supporting the hypothesis that increased pulse amplitude is the primary factor leading to overall elevation is more convincing.19,20 This abnormality is due at least in part to increased anterior pituitary sensitivity to GnRH in women with PCOS.3 Whether the amplitude of GnRH pulses also is elevated in PCOS is unknown.
Serum concentrations of FSH are usually normal or slightly low in PCOS. Attempts to characterize abnormal pulsatile release of FSH in PCOS have been unrewarding because the relatively long serum half-life of FSH renders the identification of individual pulses impractical. Both gonadotropins circulate in multiple forms, which may vary in bioactivity. An elevated ratio of bioactive to immunoreactive LH has been reported in women with PCOS, but the importance of this finding is not clear.21
In one large series, 13% of women with PCOS were found to have prolactin levels greater than 25 ng/mL in the absence of radiographic evidence of a prolactin-secreting pituitary adenoma.22 The mean prolactin level in this group was 31 ng/mL. Because prolactin secretion by the anterior pituitary is primarily under the inhibitory control of dopamine secreted into the portal circulation, it has been suggested that the hyperprolactinemia seen in some women with PCOS implies a defect in hypothalamic dopaminergic activity, which might cause inappropriate secretion of GnRH and consequently of LH. Hyperprolactinemia also is associated with increased adrenal production of DHEA-S, but in general the elevated adrenal androgen production seen in women with PCOS does not correlate with hyperprolactinemia.
Obese and nonobese women with PCOS have increased fasting insulin levels and release increased quantities of insulin in response to a standard glucose challenge compared with weight-matched eumenorrheic women.23,24,25 This finding is subtle, and women with PCOS are generally euglycemic. Nevertheless, there is convincing evidence that they are at risk for subsequent development of type 2 diabetes mellitus. In vitro testing reveals that insulin receptor number and binding are normal; women with PCOS lack circulating antireceptor antibodies. The elimination of ovarian steroid secretion with a GnRH analogue over extended periods fails to correct the insulin resistance in women with PCOS, indicating that the defect in insulin action is not secondary to abnormal sex-steroid levels.26,27 As discussed subsequently, it seems increasingly likely that this abnormality plays a key role in the pathophysiology of PCOS.
ABNORMALITIES OF THE GROWTH HORMONE/INSULIN-LIKE GROWTH FACTOR-I AXIS.
Insulin resistance in conjunction with increased adrenal androgen secretion is seen in individuals with acromegaly and during normal puberty. Because these two states are characterized by increased activity of the growth hormone/insulin-like growth factor-I (GH/IGF-I) axis, many investigators have assessed various components of this axis in women with PCOS.
Circulating levels of GH have been reported variously to be low, normal, or high in women with PCOS when compared with those of weight-matched normal cycling women.26,27,28,29,30,31 Technically, studies looking at this question are difficult to design and execute because of the prominent role that obesity per se plays in suppressing GH secretion. In contrast, provocative testing of anterior GH release consistently has indicated decreased GH secretion in PCOS.32,33,34
Total serum IGF-I levels seem to be normal in PCOS.29 The ability of circulating IGF-I to affect target tissues is modulated by six specific binding proteins, designated IGFBPs. Most (70% to 80%) circulating IGF-I is bound to IGFBP-3, which is thought to serve as a reservoir for the peptide. Circulating concentrations of IGFBP-3 seem to be normal in PCOS.30 In contrast, IGFBP-1 levels have been reported to be significantly decreased in women with PCOS and seem to be highly (negatively) correlated with serum insulin levels.35,36 IGFBP-1 levels fluctuate rapidly in the face of changing availability of nutrients. It seems likely that this binding protein serves as a counter-regulatory hormone, interfering with the insulin-like activity of IGF-I when insulin levels are low (e.g., overnight or during fasting).
Taken together, these findings are most consistent with an increased availability of circulating IGF-I to act on target tissues. IGF-I may function in a paracrine and endocrine mode, however, and complex modes of tissue compartmentalization, mediated by the binding proteins, may be crucial in modulating IGF-I action at the cellular level. In vitro studies indicating that IGF-I can enhance LH-stimulated androgen production by ovarian stromal cells,37 in conjunction with some of the observations described earlier, suggest that abnormalities of the GH/IGF-I axis may play an important role in the pathogenesis of PCOS.
The precise sequence of events that culminates in the development of PCOS remains to be elucidated. The clinical and biochemical features of the disorder comprise important clues in this regard, and to some extent it is possible to establish causal links among them. The cardinal clinical features, as described earlier, are (1) perimenarchal onset of symptoms, (2) menstrual bleeding patterns compatible with chronic anovulation, (3) acne and hirsutism, and (4) polycystic ovary. The cardinal biochemical features are (1) ovarian hyperandrogenism, (2) adrenal hyperandrogenism, (3) inappropriate gonadotropin secretion, (4) insulin resistance.
The thecal and stromal compartments of the polycystic ovary produce excessive amounts of androgens, which generate the typical clinical manifestations of hirsutism and acne. Increased androgen levels in the ovarian microenvironment also can inhibit estrogen production by granulosa cells and interfere with ovulatory function.38 The increased ovarian androgen production in PCOS at least partly depends on LH and insulin. Selective suppression of either hormone results in decreased circulating testosterone concentrations in women with PCOS.9,39 Increased activity of IGF-I also may play a role in the induction of ovarian hyperandrogenism. Although IGF-I is not produced in the thecal and stromal compartments, both contain IGF-I receptors, and IGF-I can synergize with LH to stimulate androgen production by them.37 Finally, on a macroscopic scale, the hypertrophy of the stromal compartment, whether the result of a developmental abnormality, chronically elevated circulating levels of trophic hormones, or both, mechanically comprises the basis for the typical enlargement of the polycystic ovary and its unusually smooth surface.
Women with PCOS respond to the acute administration of a GnRH analogue with increased secretion of 17α-OH progesterone and androstenedione, even in the face of adrenal suppression with dexamethasone.5 This response is consistent with enhanced 17α-hydroxylase and C-17,20 lyase activity in the thecal and stromal compartments of the ovary. It has been established that 17α-hydroxylase and C-17,20 lyase are mediated by a single enzyme, designated P450c17.40
There is no evidence for a defect in granulosa cell aromatase activity in the polycystic ovary, and true steroidogenic enzyme deficiencies are rare. The capacity of relatively modest pharmacologic interventions (e.g., clomiphene citrate or insulin-sensitizing agents) to induce ovulation in women with PCOS speaks against the presence of a crucial primary defect at the ovarian level.
Adrenal Gland in Polycystic Ovary Syndrome
The hyperandrogenism seen in PCOS is attributable to excessive androgen secretion by the adrenal and the ovary. The ovarian contribution generally predominates, but the adrenal component can be significant and in unusual cases can exceed the ovarian contribution.
Women with PCOS have normal secretion patterns of adrenocorticotropic hormone (ACTH) and cortisol but tend to release higher than normal amounts of 17OH-Pe and DHEA in response to an ACTH challenge (see Fig. 1).41 Circulating concentrations of DHEA-S often are modestly elevated. ACTH stimulation testing has failed to provide evidence for a classic adrenal enzyme block in most women with PCOS. Occasionally, someone with partial 21-hydroxylase (21-OH) or 3β-hydroxysteroid (3β-HSD) dehydrogenase deficiency, which can clinically mimic PCOS, is uncovered. Because the rise in the Δ5 steroids after ACTH administration is elevated compared with that of the Δ4 steroids in some women with elevated adrenal androgens, it has been proposed that a substantial fraction of women with PCOS are mildly deficient in 3β-HSD activity relative to that of 17α-hydroxylase and C-17,20 lyase.42 Pedigree analysis has failed so far to show that such a 3β-HSD deficiency is determined genetically.
It has been proposed that an as yet unidentified adrenal androgen-stimulating hormone might explain the discordant secretion patterns of DHEA-S and cortisol at the time of puberty.43 Adrenal steroidogenic patterns change at this time in a manner consistent with increased 17α-hydroxylase and C-17,20 lyase relative to that of 3β-HSD.44 As described earlier, a parallel alteration seems to be present in the polycystic ovary and is consistent with an increase in the activity of P450c17. Hypothetically, inappropriate secretion of adrenal androgen-stimulating hormone could result in the abnormalities of adrenal steroidogenesis seen in PCOS by either inducing such changes or inappropriately maintaining them into adult life (persistent adrenarche). It is possible that a common abnormality might be responsible for abnormal sex-steroid production in the ovary and the adrenal gland. In light of the fact that prolonged adrenal suppression with exogenous glucocorticoids fails to induce ovarian cyclicity reliably in PCOS, it is unlikely that an adrenal product plays a primary role in the pathogenesis of PCOS.
Abnormal Gonadotropin Secretion
Lobo and colleagues45 showed a correlation between non–SHBG-bound E2 levels and LH levels, lending further support to the view that increased LH release may reflect an estrogen-induced increase in anterior pituitary sensitivity to GnRH. It has been shown that eumenorrheic women who are chronically exposed to elevated androgen levels can develop abnormal gonadotropin secretion patterns similar to those seen in PCOS.46 Although it is likely that abnormal sex-steroid levels alter gonadotropin secretion in PCOS, the relative importance of androgens and estrogens remains to be determined. The extent to which insulin is an important modulator of gonadotropin secretion in women with PCOS is unclear. Suppression of insulin secretion by diazoxide suppresses testosterone, but not pulsatile LH secretion, in women with PCOS.39 More recent studies employing some of the newer insulin-sensitizing agents have shown significant declines in basal and GnRH-stimulated LH levels after therapy; others have not.47,48,49
The classic work of Knobil and associates,50 involving the administration of pulsatile GnRH to rhesus monkeys after the interruption of endogenous GnRH secretion, highlighted the importance of GnRH pulse frequency in determining subsequent gonadotropin secretion patterns. The patterns of pulsatile LH release across the menstrual cycle have been studied by several investigators, and the acquisition of this normative data has made possible the characterization of various reproductive endocrinopathies in terms of abnormal LH pulse amplitude and frequency.51,52 Some investigators described abnormally high LH and, by inference, GnRH pulse frequencies in women with PCOS, but. the most consistent finding is that of increased pulse amplitude, which may reflect either the well-documented increased anterior pituitary sensitivity to GnRH or “larger” pulses of GnRH itself.19,20 The latter hypothesis is difficult to test in humans. Other efforts to document primary defects in hypothalamic function that result in abnormal patterns of gonadotropin secretion in PCOS have met with relatively little success.
The abnormal patterns of gonadotropin secretion are most likely attributable to a combination of aberrant feedback by sex steroids and hyperinsulinemia. The two are interrelated because of the impact of hyperinsulinemia on SHBG levels, the direct effect of insulin on the ovary, and possibly the suppressive effect of insulin on IGFBP-1 levels (see later). The possibility that a developmental abnormality leading to persistent dysfunction of the hypothalamic/pituitary unit cannot be ruled out and animal studies involving prenatal androgen exposure in the rhesus monkey support this notion.53 Finally, it is possible that elevated serum LH concentrations are not mandatory for the development of PCOS given the fact that a substantial fraction of the women with the disorder have relatively normal values.
Insulin ResistanceWomen with PCOS are insulin resistant.23 Basal and stimulated insulin levels commonly are elevated as a result. This feature is detectable in nonobese and in obese subjects and is distinct from the insulin resistance that results from obesity per se.24,25 It has been suggested that the insulin resistance of PCOS is a result of the abnormal steroidal milieu30; however, insulin resistance persists after prolonged suppression of ovarian sex-steroid secretion in women with PCOS.26,27 Obese women with PCOS are significantly more insulin resistant than nonobese women with the disorder but have similar circulating levels of gonadotropins and sex steroids.54,55 Many normal cycling women who happen to be obese display degrees of hyperinsulinemia comparable to lean women with PCOS. Together, these findings support the view that hyperinsulinemia alone is not solely responsible for the development of PCOS. Reducing circulating insulin concentrations in women with PCOS significantly lowers ovarian androgen output and may result in spontaneous ovulation, indicating that hyperinsulinemia may alter patterns of ovarian steroidogenesis either directly or by reducing IGFBP-1 levels.13,39,47,48,56,57 Finally, insulin has been shown to be a potent stimulator of ovarian androgen secretion in vitro.53
Elevated insulin levels seem to suppress circulating SHBG levels, rendering more E2 and testosterone biologically available.18 With respect to E2, this suppression may lead to increased LH levels via enhanced positive feedback.45 With respect to testosterone, abnormal gonadotropin release and increased clinical manifestations of hyperandrogenism might result. The mechanism of the defect remains to be elucidated but seems to be unrelated to an abnormality of insulin receptor concentration or binding, pointing toward some type of postreceptor defect. Studies indicate that this abnormality is associated with abnormal patterns of insulin secretion in first-degree relatives of women with PCOS, but attempts to identify a “PCOS gene” have been unrewarding so far.59,60,61 Women who have extreme insulin resistance are typically anovulatory and hyperandrogenic (see later discussion).
A recent important development in understanding of the mechanisms underlying PCOS is the identification of insulin resistance as a key abnormality that may contribute directly and indirectly to many of the clinical and biochemical features of the disorder. Current efforts are being directed toward understanding the mechanism of the insulin resistance at the molecular level. In particular, ongoing studies are focusing on abnormalities in the function of glucose transporter molecules that mediate insulin action at a postreceptor site. Insulin resistance is drawing increasing attention in many other contexts and may play a role in the induction of other important disorders, including cardiovascular disease and essential hypertension.62,63
The possibility that an abnormality in IGF-I activity may play a role in the development of PCOS has been proposed by several investigators.28,64,65 It has been suggested that IGF-I activity in the thecal and stromal compartments of the ovary is enhanced because of the decrease in IGFBP-I, which may result from increased insulin levels.35 In this view, the insulin resistance is still the primary defect. Alternatively, it has been proposed that the insulin resistance and the abnormalities of adrenal steroidogenesis might be the consequences of increased IGF-I activity.56 This view emphasizes the similarities between the events of normal puberty and the development of PCOS; specifically the fact that normal puberty is associated with increased activity of P450c17 and a decrease in insulin sensitivity suggests that a common mechanism might be at work. It may be that exposure of the ovary or adrenal to increased IGF-I stimulation during a crucial developmental period (e.g., antenatally or during puberty) might lead to morphologic changes or alterations in patterns of steroidogenesis that might persist throughout adult life.
Finally, at least two more recent lines of evidence support the view that the disorder may result from fetal exposure to an abnormal prenatal environment. As alluded to earlier, studies in the rhesus monkey suggest that antenatal exposure to androgens can abnormally “program” the subsequent development of the hypothalamic-pituitary-ovarian axis and permanently alter glucose homeostasis mechanisms. To the extent that this hypothesis has merit, the question naturally arises as to the source of the androgen excess. Placental aromatase activity is ordinarily more than adequate to protect a female fetus from modest levels of maternal hyperandrogenemia. This points toward the possibility that the fetus per se may be the source. The fetal ovary is not an important source of sex steroids during pregnancy, leaving the fetal adrenal as the remaining possibility. A hypothetical abnormality in fetal adrenal development or activity not only could be responsible for the ovarian and metabolic endocrine abnormalities of PCOS, but also might comprise a novel explanation for the frequently observed adrenal hyperandrogenism. A second new line of evidence, which may be related to some degree to the first, concerns the association of abnormal birth weight with an increased risk for developing PCOS.66
All women with clinical features of PCOS warrant diagnostic evaluation to rule out other pathologic states that can present with similar findings. Such disorders sometimes are referred to as PCOS-like. This principle is important for two reasons. First, appropriate therapy for a specific disorder may be withheld in the absence of a correct diagnosis. Second, it is essential to identify the occasional patient with an ovarian or adrenal neoplasm who presents with a PCOS-like picture. The clinician should be prepared to design a diagnostic evaluation based on the patient’s clinical history and presentation.
Hyperthyroidism and hypothyroidism are associated with states of chronic anovulation resulting from altered SHBG levels and other changes in sex-steroid metabolism. In the absence of other signs or symptoms of thyroid disease, it is probably unnecessary to measure the various indices of thyroid function. It may be worthwhile to obtain a thyroid-stimulating hormone level to identify the occasional patient with compensated hypothyroidism. Patients with PCOS who have normal thyroid-stimulating hormone levels should not be treated with thyroid supplementation.
The existence of a separate variant of PCOS characterized by moderately elevated prolactin levels is controversial. Any patient with documented, persistent hyperprolactinemia should have appropriate radiologic evaluation and should be treated with bromocriptine or cabergoline. Women who are amenorrheic because of hyperprolactinemia generally have low gonadotropin levels. The modest elevations in prolactin levels seen in PCOS frequently reflect the stimulatory effect of circulating estrogens on pituitary lactotropes. Because galactorrhea is not observed universally in hyperprolactinemic women, all patients complaining of oligomenorrhea or amenorrhea should have a serum prolactin determination. Ideally the blood sample is obtained late in the morning to permit the normal nocturnal elevation of prolactin levels to return to baseline; it should not be obtained immediately after a meal because food ingestion results in increased prolactin secretion.
Adrenal Enzyme Defects
Incomplete forms of 21-OH deficiency and mild deficiencies in 3β-ol dehydrogenase can clinically mimic PCOS. The reported incidence of such defects in several series varies but is probably less than 10% of hirsute patients. DHEA-S levels are frequently normal. The diagnosis of adrenal enzyme deficiencies can be made with certainty only with an ACTH stimulation test, although an elevated basal level of 17OH-P strongly suggests a 21-OH deficiency. Because of the cost involved, clinical judgment must dictate which patients warrant such evaluation. Women with “cryptic” 21-OH deficiency tend to be those with a strong family history of hirsutism, severe hirsutism from puberty, short stature compared with family members, or flattening of the breasts. It is not appropriate to recommend long-term glucocorticoid suppression in the absence of evidence for elevated adrenal androgen secretion.
Few patients presenting with hirsutism or menstrual dysfunction are proved to have Cushing’s syndrome, but the diagnosis always should be considered in light of the potential consequences of this disease. A 10 P.M. serum cortisol level, a 24-hour free urinary cortisol determination, or an overnight dexamethasone suppression test (8 A.M. serum cortisol after 1 mg of dexamethasone taken orally at 11 P.M. the evening before, with a value of <5 mg/dL being normal) can be used as the initial screening tests when the clinical picture suggests this possibility.
Adrenal and Ovarian Neoplasms
All patients with significant hirsutism should have a testosterone and a DHEA-S determination. Although controversy persists regarding limits above which androgen-secreting neoplasms should be suspected, testosterone levels greater than 2000 pg/mL or DHEA-S levels greater than 700 μg/dL require an evaluation to rule out such tumors. A clinical picture of rapidly progressive signs of androgen excess, particularly if signs of virilization are present, should alert the clinician to the potential presence of an adrenal or ovarian neoplasm. A DHEA-S level greater than 700 μg/dL should be evaluated with a computed tomography scan of the adrenals. Such modern radiographic tests can localize adrenal and ovarian neoplasms in most patients. Selective venous catheterization of the ovaries and adrenals is required rarely to pinpoint the source of excess androgen secretion; this should be undertaken only by an experienced radiologist. This procedure is expensive, can be accompanied by serious complications, and should be used only when clearly indicated. Rarely, granulosa-theca cell tumors of the ovary, which can secrete estrogens and androgens, can result in ovarian acyclicity when these tumors occur during the reproductive years.
Extreme Insulin Resistance and Acanthosis Nigricans
A group of women has been identified with extreme insulin resistance (in contrast to the modest degree of insulin resistance usually observed in women with PCOS), marked hyperandrogenism, and acanthosis nigricans.67,68 Acanthosis nigricans is the presence of hyperpigmented skin, most commonly in intertriginous areas such as the axillae and the back of the neck. This group of patients can be divided further into two subgroups: patients with a genetic disorder characterized by decreased affinity and quantity of insulin receptors (type A) and patients with a syndrome associated with circulating antibodies to the insulin receptor (type B). Affected women are usually amenorrheic and often obese. The presence of skin changes in this syndrome seems to be related to the degree of hyperinsulinism, but a direct cause-and-effect relationship remains to be established. In this group, the hyperandrogenic, anovulatory state has been attributed to a direct stimulatory effect of insulin on stromal and thecal tissues of the ovary.68 Marked obesity per se can induce a degree of insulin resistance sufficient to result in the appearance of acanthosis.69
The appropriate treatment of the patient with PCOS depends on whether she wishes to conceive, and ovulation induction should be attempted only if pregnancy is desired. Therapy ordinarily should be initiated with clomiphene citrate. Patients with hyperprolactinemia also may benefit from treatment with bromocriptine. It is expected that cabergoline, a long-acting dopaminergic agonist, will be a reasonable alternative when fetal safety issues are resolved. Whether euprolactinemic patients also benefit from dopamine agonist administration is controversial. Glucocorticoid administration should be reserved for women with evidence of adrenal involvement. Women with elevated DHEA-S levels seem to benefit from such treatment before clomiphene administration; women with normal levels do not. Pulsatile GnRH administration by either intravenous or subcutaneous routes has yielded disappointing results so far, although pretreatment with a GnRH agonist may improve outcomes.70 Treating PCOS patients with GnRH analogues alone for extended periods does not result in subsequent ovulatory function. Gonadotropin therapy should be used in resistant cases but only by clinicians who are familiar with its use and equipped to monitor follicular development carefully. Even in the best of hands, this therapy occasionally is associated with symptomatic ovarian enlargement (hyperstimulation).
Many investigators have reported success facilitating spontaneous and induced ovulation in insulin-resistant PCOS patients after oral administration of insulin-sensitizing agents, such as metformin, a biguanide, and troglitazone, a thiazolidinedione class medication.47,48,49,71 The latter has been associated with rare but serious hepatic toxicity and has been withdrawn from the market. The potential role of such agents in the initial therapy of anovulation in PCOS patients is currently an active area of investigation.
Bilateral ovarian wedge resection or electrosurgical destruction of ovarian tissue should be undertaken only as a last resort. It is not clear that a prolonged period of ovarian cyclicity can be expected after such procedures, and they may generate postoperative adhesions that subsequently can prevent conception. The reader is referred to other chapters of this volume for a detailed discussion of ovulation induction techniques.
The patient complaining primarily of hirsutism usually benefits from cyclic oral contraceptive administration. The combination of decreased ovarian androgen secretion and elevated SHBG levels that results from use of an oral contraceptive frequently results in reduced peripheral androgen activity and some cosmetic improvement. Spironolactone, a diuretic, may be used in selected cases at a dose of 100 to 200 mg/day. It acts by inhibiting androgen production in the ovary and by competitively blocking local androgen activity in the hair follicle. Patients treated with spironolactone should be aware that they may become ovulatory. Because the effects of this antiandrogen on any developing fetus are unknown, women taking spironolactone should use contraception if sexually active.
Glucocorticoid administration should be reserved for patients with documented elevations of adrenal androgen production. Marked suppression of testosterone levels after a 5-day course of dexamethasone (0.5 mg four times daily) may identify patients who would benefit from glucocorticoid administration even in the face of normal ACTH stimulation testing. Such patients may have unusually high adrenal P450c17 activity.72 A typical regimen would consist of a 0.25- to 0.5-mg dose of dexamethasone nightly. Particularly at the higher dose, patients should be warned about and monitored for evidence of adrenal insufficiency. A morning cortisol level of less than 2.0 μg/dL warrants a reduction in the glococorticoid dosage.
Finasteride is a 5-α reductase inhibitor that blocks conversion of testosterone to dihydrotestosterone. This agent, which originally was introduced for the treatment of prostate cancer, is similar in efficacy to spironolactone for the treatment of hirsutism and is generally well tolerated.73,74 Typical daily dosages range between 2 and 5 mg daily. Because of its potential impact on the developing external genitalia of a male fetus, effective contraception is mandatory during its use.
Cyproterone acetate is an antiandrogen used widely in Europe for the treatment of hirsutism. It is generally given with an estrogen in a cyclic fashion. It is currently unavailable for use in the United States and may never be approved for use because of its propensity to cause breast cancer in beagles and birth defects when administered to pregnant rodents.
As mentioned earlier, insulin-sensitizing agents such as metformin seem to ameliorate ovarian hyperandrogenism.13,56,57 To the extent that the abnormal folliculogenesis in PCOS is due to local overproduction of androgens, these agents ultimately may prove to be useful in treating patients complaining of hirsutism and women complaining of infertility.
For patients complaining primarily of excessive facial hair, eflornithine hydrochloride, an inhibitor of L-ornithine decarboxylase, has been introduced. This agent is applied as a topical cream. Electrolysis, shaving, and other physical modalities can be useful adjuncts to hormonal therapy, particularly for excessive facial hair.
Women with PCOS may be amenorrheic or oligomenorrheic or may have frequent, sometimes heavy, irregular bleeding, referred to as dysfunctional uterine bleeding. The fact that these patients may bleed “about once a month” should not be interpreted as reassuring. In all cases, except for the occasional patient who is ovulatory, the practitioner must actively intervene to prevent the development of endometrial hyperplasia and neoplasia, which can arise in the face of chronic unopposed endometrial stimulation by estrogen. This risk largely can be eliminated by administering oral contraceptives or periodic progestins. For affected women not using oral contraceptives, 5 to 10 mg of medroxyprogesterone acetate should be taken orally for 10 to 14 days on a monthly basis. Patients on this regimen should be advised to use a barrier form of contraception if they do not wish to conceive. Finally, the practitioner should have a low threshold for carrying out endometrial sampling in women with abnormal bleeding, particularly if they have not been provided with such therapy for an extended period of time.
Non–Insulin-Dependent Diabetes Mellitus and Cardiovascular Disease
A growing body of evidence suggests that women with PCOS, particularly women who are obese, are at increased risk for the development of glucose intolerance later in life.75 This is a natural consequence of the insulin resistance that typifies the disorder and that which is associated with obesity per se. It is reasonable to screen PCOS patients for glucose intolerance periodically, employing fasting glucose-to-insulin ratios as the initial test. A ratio of less than 4.5 is considered to be abnormal.76 The well-known association of hyperinsulinemia with hypertension and the abnormal lipid profiles frequently seen in the face of chronically elevated androgens suggest that these patients should be viewed as at higher risk for cardiovascular disease as well.
It would seem reasonable to steer PCOS patients toward the same lifestyle alterations commonly recommended for patients with non–insulin-dependent diabetes mellius: careful attention to diet, regular exercise, and an overall effort to maintain an ideal body weight. Nevertheless, it remains a frustrating clinical fact that women with PCOS have a notoriously difficult time losing weight, and so practitioner and patients should devise reasonable goals in this regard. It is likely that insulin-sensitizing agents eventually may play a significant role in the long-term management of these significant health risks. As we await longer term studies, care providers should educate their patients about these concerns and recommend appropriate, but realistic, lifestyle alterations.
6. Buckler HM, MacLachlan RI, MacLachlan VB, et al: Serum inhibin levels in polycystic ovary syndrome: Basal levels and response to luteinizing hormone-releasing hormone agonist and exogenous gonadotropin administration. J Clin Endocrinol Metab 66:798, 1988
9. Chang RJ, Laufer LR, Meldrum DR, et al: Steroid secretion in polycystic ovarian disease after ovarian suppression by a long-acting gonadotropin releasing hormone agonist. J Clin Endocrinol Metab 56:897, 1983
12. Nestler JE, Singh R, Matt DW, et al: Suppression of serum insulin by diazoxide does not alter serum testosterone or sex hormone-binding globulin levels in healthy nonobese women. Am J Obstet Gynecol 163:1243, 1990
13. Nestler JE, Jakubowicz DJ: Decreases in ovarian cytochrome P450c17a activity and serum free testosterone after reduction in insulin secretion in women with polycystic ovary syndrome. N Engl J Med 335:617, 1996
14. Velazquez EM, Mendoza S, Hamer T, et al: Metformin therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin resistance, hyperandrogenemia, and systolic blood pressure, while facilitating normal menses and pregnancy. Metabolism 43:647, 1994
18. Nestler JE, Powers LP, Matt DW, et al: A direct effect of hyperinsulinemia on serum sex hormone-binding globulin levels in obese women with the polycystic ovary syndrome. J Clin Endocrinol Metab 72:83, 1991
19. Burger CW, Korsen T, Van Kessel H, et al: Pulsatile luteinizing hormone patterns in the follicular phase of the menstrual cycle, polycystic ovarian disease (PCOD) and non-PCOD secondary amenorrhea. J Clin Endocrinol Metab 61:1126, 1985
25. Dunaif A, Segal KR, Futterweit W, et al: Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes 38:1165, 1989
26. Geffner ME, Kaplan SA, Bersch N, et al: Persistence of insulin resistance in polycystic ovarian disease after inhibition of ovarian sex steroid secretion. Fertil Steril 45:327, 1986
27. Dunaif A, Green G, Futterweit W, et al: Suppression of hyperandrogenism does not improve peripheral or hepatic insulin resistance in the polycystic ovary syndrome. J Clin Endocrinol Metab 70:699, 1990
30. Morales AJ, Laughlin GA, Butzow T, et al: Insulin, somatotropic and luteinizing hormone axes in lean and obese women with polycystic ovary syndrome: Common and distinct features. J Clin Endocrinol Metab 81:2854, 1996
31. Prelevic GM, Wurzburger MI, Balint-Peric L, Ginsburg J: Twenty-four hour serum growth hormone, insulin, C-peptide and blood glucose profiles and serum insulin-like growth factor-I concentrations in women with polycystic ovaries. Horm Res 37:125, 1992
37. Cara JF, Rosenfield RL: Insulin-like growth factor I and insulin potentiate luteinizing hormone-induced androgen synthesis by rat ovarian thecal-interstitial cells. Endocrinology 123:733, 1988
39. Nestler JE, Barlascini CO, Matt DW, et al: Suppression of serum insulin by diazoxide reduces serum testosterone levels in obese women with polycystic ovary syndrome. J Clin Endocrinol Metab 68:1027, 1989
42. Pang S, Lerner AJ, Stoner E, et al: Late-onset adrenal steroid 3b-hydroxysteroid dehydrogenase deficiency: I. A cause of hirsutism in pubertal and postpubertal women J Clin Endocrinol Metab 60:428, 1985
44. Schiebinger RJ, Albertson BD, Cassorla FG, et al: The developmental changes in plasma adrenal androgens during infancy and adrenarche are associated with changing activities of adrenal microsomal 17-hydroxylase and 17, 20 desmolase. J Clin Invest 67:1177, 1981
45. Lobo RA, Granger L, Goebelsmann U, et al: Elevations in unbound serum estradiol as a possible mechanism for inappropriate gonadotropin secretion in women with PCO. J Clin Endocrinol Metab 52:16, 1981
46. Dunaif A, Scully RE, Andersen RN, et al: The effects of continuous androgen secretion on the hypothalamic-pituitary axis in woman: Evidence from a luteinized thecoma of the ovary. J Clin Endocrinol Metab 59:389, 1984
47. Velazquez EM, Mendoza SG, Hamer T, et al: Metformin therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin resistance, hyperandrogenemia, and systolic blood pressure, while facilitating normal menses and pregnancy. Metabolism 43:647, 1994
49. Vandermolen DT, Ratts VS, Evans WS, et al: Metformin increases the ovulatory rate and pregnancy rate from clomiphene citrate in patients with polycystic ovary syndrome who are resistant to clomiphene citrate alone. Fertil Steril 75:310, 2001
51. Filicori M, Santoro N, Merriam GR, et al: Characterization of the physiological pattern of episodic gonadotropin secretion throughout the human menstrual cycle. J Clin Endocrinol Metab 62:1136, 1986
52. Reame N, Sauder SE, Kelch RP, et al: Pulsatile gonadotropin secretion during the human menstrual cycle: Evidence for altered frequency of gonadotropin-releasing hormone secretion. J Clin Endocrinol Metab 59:328, 1984
55. Dunaif A, Mandeli J, Fluhr H, Dobrjansky A: The impact of obesity and chronic hyperinsulinemia on gonadotropin release and gonadal steroid secretion in the polycystic ovary syndrome. J Clin Endocrinol Metab 66:131, 1988
56. Dunaif A, Scott D, Finegood D, et al: The insulin-sensitizing agent troglitazone improves metabolic and reproductive abnormalities in the polycystic ovary syndrome. J Clin Endocrinol Metab 81:3299, 1996
57. Nestler JE, Jakubowicz DJ: Decreases in ovarian cytochrome P450c17 alpha activity and free testosterone after reduction of insulin secretion in polycystic ovary syndrome. N Engl J Med 335:617, 1996
61. Colilla S, Cox NJ, Ehrmann DA: Heritability of insulin secretion and insulin action in women with polycystic ovary syndrome and their first degree relatives. J Clin Endocrinol Metab 86:2027, 2001
63. DeFronzo RA, Ferrannini E: Insulin resistance: A multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabetes Care 14:173, 1991
67. Kahn CR, Flier JS, Bar RS, et al: The syndromes of insulin resistance and acanthosis nigricans. N Engl J Med 294:739, 1976
69. Flier JS, Eastman RC, Minaker KL, et al: Acanthosis nigricans in obese women with hyperandrogenism: Characterization of an insulin resistant state distinct from the type A and B syndromes. Diabetes 34:101, 1985
74. Wong IL, Morris RS, Chang L, et al: A prospective randomized trial comparing finasteride to spironolactone in the treatment of hirsute women. J Clin Endocrinol Metab 80:233, 1995
75. Legro RS, Kunselman AR, Dodson WC, et al: Prevalence and predictors of risk for type 2 diabetes mellitus and impaire glucose tolerance in polycystic ovary syndrome: A prospective, controlled study in 254 affected women. J Clin Endocrinol Metab 84:165, 1999
76. Legro RS, Finegood D, Dunaif A: A fasting glucose to insulin ratio is a useful measure of insulin sensitivity in women with polycystic ovary syndrome. J Clin Endocrinol Metab 83:2694, 1998