First Department of Medicine (E.D.-K.), University of Athens, GR-17548 Athens, Greece; Department of Medicine (J.-P.B.), Universitaire de Sherbrooke, Sherbrooke, Canada J1H 5N4; Departments of Medicine (M.J.I., J.E.N.) and Obstetrics and Gynecology (J.E.N.), Virginia Commonwealth University, Richmond, Virginia 23298; and Hospital de Clinicas Caracas (D.J.J.), 1040 Caracas, Venezuela
Address all correspondence and requests for reprints to: John E. Nestler, M.D., Medical College of Virginia, P.O. Box 980111, Richmond, Virginia 23298-0111. E-mail: nestler{at}hsc.vcu.edu.
The polycystic ovary syndrome (PCOS) is a complex clinical entity for the clinician, a scientific challenge for the investigator, and a frustrating personal and social experience for approximately 3.55.0 million women in the United States. It is classically defined by chronic anovulation and hyperandrogenism, but it has become apparent that insulin resistance plays a prominent role in the disorder.
The discovery of the central role of insulin resistance in PCOS is important from several points of view, including the possibility that the traditional management of the disorder may have untoward long-term consequences. The aims of this report are to highlight why insulin resistance should be considered in the formulation of chronic treatment strategies of PCOS, to perhaps suggest a more judicious use of oral contraceptive pills (OCPs) in the chronic treatment of the disorder, and to issue a call for more rigorous and critical studies of the metabolic effects of both OCPs and insulin-sensitizing drugs in women with PCOS.
For many decades OCPs have been standard therapy for women with PCOS not seeking pregnancy. There have been several advantages of treatment with OCPs, including foremost protection from the development of endometrial carcinoma, regularization of menses, and amelioration of hirsutism and acne. There is no disputing the importance of these benefits. However, what has been lacking is a critical examination of whether OCPs might also exert adverse metabolic effects with long-term consequences, especially in a group of women with known insulin resistance and predisposition to type 2 diabetes and cardiovascular disease.
PCOS: a general health disorder linked to insulin resistance
Numerous studies have documented that both lean and obese women with PCOS are insulin resistant (1, 2, 3, 4). The lean woman with PCOS seems to have a form of insulin resistance that is intrinsic (and perhaps unique) to the syndrome and poorly understood (5, 6, 7). The obese woman with PCOS possesses not only the form of insulin resistance intrinsic to the syndrome, but also has an added burden of insulin resistance that is related to excess adiposity (8).
Insulin resistance is associated with an increased risk for several disorders, including type 2 diabetes, hypertension, dyslipidemia (low high-density lipoprotein cholesterol and high triglycerides), elevated plasminogen activator inhibitor type 1 (PAI-1), elevated endothelin-1, endothelial dysfunction, and heart disease. This clustering of abnormalities with insulin resistance has been termed Syndrome X, the insulin resistance syndrome, or the dysmetabolic syndrome (9, 10). The adverse potential of the dysmetabolic syndrome has recently been highlighted by the National Cholesterol Education Projects Adult Treatment Panel III guidelines, which recognize Syndrome X as a major cardiac risk factor (11), and by assignment of an ICD-9 code (277.7) for the disorder.
PCOS may rightly be considered a component of the dysmetabolic syndrome in women. Evidence that PCOS is associated with a high risk for the development of type 2 diabetes and heart disease is mounting, and optimal chronic therapy for the disorder should address these long-term general health issues.
Regarding diabetes risk, prospective clinical trials have demonstrated a 3135% prevalence of impaired glucose tolerance and 7.510.0% prevalence of type 2 diabetes mellitus in women with PCOS (12, 13). It has also been reported that the rate of conversion from impaired glucose tolerance to type 2 diabetes mellitus may be increased 5- to 10-fold in women with PCOS (12, 14). Oligomenorrhea is a highly predictive surrogate marker for PCOS. An analysis of 101,073 women in the Nurses Health Study II reported that over an 8-yr period the conversion rate to type 2 diabetes among oligomenorrheic women was approximately 2-fold greater than that for eumenorrheic women, regardless of whether the oligomenorrheic women were obese or lean, indicating that oligomenorrhea was an independent predictor of type 2 diabetes (15).
Conversely, a retrospective study of an academic diabetes clinic in Virginia revealed that 27% of the premenopausal women with type 2 diabetes had PCOS (16), and transvaginal ultrasonography of premenopausal women in the diabetes clinic at Middlesex Hospital in England revealed an 82% prevalence of anatomically polycystic ovaries (17). Collectively, the above observations indicate that women with PCOS constitute one of the groups at highest risk for the development of diabetes.
Although controversial, several lines of evidence also indicate increased cardiovascular risk in women with PCOS. PCOS is associated with increased prevalence of several cardiovascular risk factors, including hypertension (18, 19, 20) and dyslipidemia (21, 22, 23, 24). In addition, women with PCOS display surrogate markers for early atherosclerosis, such as increased PAI-1 (25, 26, 27), endothelin-1 (28), and C-reactive protein (29) concentrations. Several (30, 31, 32), but not all (33), studies suggest that PCOS is associated with endothelial dysfunction, which is linked to insulin resistance and is a risk factor for cardiovascular disease.
Consequently, it is not surprising that several studies, although again not all (34), report an increased prevalence of heart disease in PCOS (22, 30, 35, 36) or in women with the anatomic finding of polycystic ovaries (37). A powerful demonstration of the early atherogenic process in PCOS is the study by Talbott et al. (36), which revealed thickening of carotid intima media in middle-aged women with PCOS compared with age-matched normal women.
OCPs and the dysmetabolic syndrome
The above considerations indicate that PCOS is not only an infertility problem or cosmetic annoyance, but perhaps foremost a general health problem at whose crux lies insulin resistance. Considering that PCOS may affect between 3.5 and 5.0 million young women in the United States, it arguably may be the most important general health issue affecting young women. Evidence exists, albeit limited and at times less than rigorous, that some sex steroids may aggravate insulin resistance. Therefore, in evaluating the traditional long-term use of OCPs in PCOS one is compelled to ask: "How might oral contraceptives interact with the manifestations of Syndrome X in PCOS?" This is a topic that has received relatively scant attention.
OCPs, insulin resistance, and glucose intolerance
Women with PCOS are at high risk for the development of glucose intolerance and heart disease, likely related in part to their insulin resistance. OCPs have long been the mainstay for the treatment of the syndrome. However, as outlined below, some studies have shown that OCPs may themselves decrease insulin sensitivity and glucose tolerance, and these sequellae might possibly increase the risk for type 2 diabetes mellitus in women with PCOS.
For example, a review of the literature revealed only six short-term clinical trials that prospectively evaluated the metabolic effects of OCPs specifically in women with PCOS (38, 39, 40, 41, 42, 43). Only one study included a placebo control group (38), two others compared the effects of an OCP to metformin (41, 43), and all five studies assessed a limited number of women.
Two studies demonstrated significant deteriorations of insulin sensitivity, using either the hyperglycemic clamp (38) or the hyperinsulinemic-euglycemic clamp (40), in women with PCOS when a low-dose norethindrone-containing triphasic combination OCP or a high-dose OCP containing cyproterone acetate was administered for 36 months. The study by Korytkowski et al. (38) also assessed a control group of normal women, in whom OCP administration was associated with an increase in serum triglycerides, a finding consistent with the induction of insulin resistance.
In the study by Dahlgren et al. (40), the decrease in insulin sensitivity occurred despite a stable body mass index and reduction in circulating androgens. Notably, when these investigators treated eight women with PCOS with a GnRH agonist for 6 months, serum androgen concentrations also declined markedly yet insulin sensitivity improved. This suggests that the degree of insulin resistance induced by OCPs was both substantial and related directly to the pills themselves and not to suppression of ovarian function.
The remaining four studies assessed the effects of 6-month administration of a low-dose OCP on oral glucose tolerance in women with PCOS. One study administered a desogestrel-containing pill (39), and the other three studies used a pill containing cyproterone acetate that was administered to either obese (41) or nonobese (42, 43) women with PCOS.
Two of these studies (39, 41), both conducted in obese women with PCOS, demonstrated a deterioration of glucose tolerance with OCP administration, as evidenced by higher plasma glucose levels during the oral glucose tolerance test, with no change in plasma insulin concentrations. This occurred despite no change in body mass index and a marked decrease in circulating androgens. Because plasma insulin levels did not change, the deterioration in glucose tolerance was likely due to a decrease in insulin sensitivity.
The remaining two studies (42, 43), performed in nonobese women with PCOS, showed no change in glucose tolerance or insulin sensitivity after the administration of an OCP containing cyproterone acetate, suggesting that the metabolic effects of OCPs may vary with body phenotype.
No cohort studies have been performed to assess the risk of developing type 2 diabetes specifically in women with PCOS who used OCPs. Nonetheless, two Nurses Health Study cohort studies have assessed the risk of type 2 diabetes associated with the use of OCPs in a general population of healthy women (44, 45). The first cohort of the Nurses Health Study (NHS I) included healthy female nurses who were 3055 yr old in 1976 and followed them for 12 yr (45). The mean age was 58 yr at follow-up, with a total of 2265 cases. The relative risk of developing type 2 diabetes associated with past use of OCPs was 10% greater than the risk in never users. It should be noted, however, that a large proportion of OCPs used in 1976 contained high-dose estrogens.
A second cohort of the Nurses Health Study (NHS II) recruited healthy nurses who were 2542 yr old in 1989 and followed them for 4 yr (44). Mean age at follow-up was only 38 yr, with a total of 170 cases. This study assessed the risk of developing type 2 diabetes associated specifically with use of low-dose OCPs. The adjusted relative risk was increased in past and current users of OCPs compared with never users (relative risk, 1.2 and 1.6, respectively). These differences did not attain statistical significance, which may have been related to the limitations of the study, including a brief period of observation, young age at follow-up, and a limited number of cases.
Whereas the above studies did not assess all available OCPs, an impossible task given the myriad of estrogen and progestin combinations, they nonetheless suggest that OCP use may decrease insulin sensitivity and/or impair glucose tolerance in some women with PCOS. Moreover, the observation that OCP use may be associated with an increased risk of type 2 diabetes in healthy women has even greater implications for women with PCOS. Women with PCOS represent a group characterized by baseline insulin resistance who are already at high risk for type 2 diabetes, and OCP use, therefore, might be expected to increase their relative risk for type 2 diabetes even more.
OCPs and cardiovascular disease
No studies exist assessing the effect of OCP use on cardiovascular outcomes specifically in women with PCOS. Nonetheless, a World Health Organization case-control study reported 4.7- and 2.9-fold increases in risk for myocardial infarction associated with low-dose OCP use in Europe and developing countries, respectively (46). The OCP-associated risk was markedly increased by the presence of hypertension or at least one risk factor for cardiovascular disease, and, as noted earlier, these are relatively common among women with PCOS.
More recently in the United States, a population-based, case-control study of 248 women between the ages of 18 and 49 yr reported a 2-fold increased risk of myocardial infarction among women who used any type of combination OCP (47).
Despite a paucity of information, evidence suggests that OCP use may increase the risk of cardiovascular disease in women. Given that PCOS is already associated with multiple risk factors for cardiovascular disease, it is reasonable to be concerned that OCP use may enhance cardiovascular risk in PCOS while bearing in mind that no studies have directly assessed this issue.
Insulin-sensitizing drugs and the dysmetabolic syndrome in PCOS
In contrast to OCPs, insulin-sensitizing drugs would be expected to reduce the risk for type 2 diabetes or heart disease in PCOS by ameliorating the putative root cause for these co-morbidities (i.e. insulin resistance). Therefore, it is appropriate to specifically ask: "Can treatment with insulin-sensitizing drugs forestall or prevent the development of type 2 diabetes or heart disease in PCOS?"
Insulin-sensitizing drugs and type 2 diabetes mellitus
Insulin resistance may be the earliest detectable abnormality in individuals who proceed to develop type 2 diabetes and is a recognized risk factor for diabetes (48, 49, 50). There are now at least two major outcome studies that suggest that interventions to improve insulin sensitivity can stave off the development of diabetes in individuals at high risk. One study, conducted in Finland, reported that improved insulin sensitivity achieved through a combination of diet and exercise reduced progression to type 2 diabetes by 58% over 4 yr in obese men with impaired glucose tolerance (51).
More recently, an National Institutes of Health study of 3234 individuals at high risk for diabetes (i.e. presence of impaired glucose tolerance and first-degree relative with diabetes or history of gestational diabetes) was terminated early (i.e. after an average follow-up period of 3 yr) because intensive lifestyle intervention consisting of diet and exercise again reduced the risk of progression to type 2 diabetes by 58%. Notably, individuals treated with the insulin-sensitizing drug metformin experienced a 31% risk reduction in the development of diabetes (52). These outcome studies strongly suggest that improving insulin sensitivity in individuals at high risk, with lifestyle changes or drug, reduces the risk for developing diabetes.
These findings are further supported by a preliminary report by Buchanan et al. (53). Two hundred thirty-five Hispanic women with recent gestational diabetes were randomized after childbirth to troglitazone (n = 114) or placebo (n = 121). After a median of 30 months, the progression to type 2 diabetes was decreased significantly by 56% in the troglitazone-treated group compared with the placebo group. The protective effect of troglitazone to prevent progression to diabetes in this high-risk group of women persisted for 8 months after discontinuation of the drug, suggesting true preservation of pancreatic ß-cell function.
Women with PCOS are an insulin-resistant group at markedly increased risk for type 2 diabetes. Therefore, it seems reasonable to presume that the demonstrated efficacy of insulin-sensitizing drugs to stave off type 2 diabetes in individuals with impaired glucose tolerance (51, 52) or history of gestational diabetes (53) should be applicable to them as well.
Insulin-sensitizing drugs and heart disease
Evidence for a possible cardioprotective effect of insulin-sensitizing drugs in PCOS is primarily indirect and circumstantial. Insulin-sensitizing drugs have been reported to exert salutary effects on multiple cardiovascular risk factors in PCOS. Specifically, some, but not all (54), studies have reported that these drugs improve the cardiovascular risk profile by decreasing serum triglycerides (55, 56, 57), reducing blood pressure (55, 56, 57), and decreasing serum PAI-1 (25, 26) and endothelin-1 (28) concentrations in women with PCOS. These effects have been variable from study to study, but certainly no aggravation of cardiovascular risk factors has been reported. Thus, insulin-sensitizing drugs do not seem to be harmful from a cardiovascular risk perspective and may be beneficial.
Extrapolating from another insulin-resistant disorder, namely type 2 diabetes, it is notable that the United Kingdom Prospective Diabetes Study reported a reduced incidence of myocardial infarction in obese type 2 diabetic patients assigned to receive initial treatment with metformin monotherapy (58). Therefore, for one of the insulin-sensitizing drugs, namely metformin, there is a long-term outcome study that suggests it may decrease cardiovascular risk. This is bolstered by a recent population-based mortality study from Canada that also reported a cardioprotective action of metformin in diabetic patients (59).
Prevention of endometrial cancer in PCOS: can insulin-sensitizing drugs help?
Finally, another important consideration of long-term therapy in PCOS is the prevention of endometrial cancer. Chronic anovulation is associated with an increased risk for endometrial cancer (60), and there seems to be an increased prevalence of endometrial cancer among women with PCOS, including young women with the disorder (61, 62). Notably, increased risk for endometrial cancer was also reported in women with increased serum levels of insulin (63) and lower serum levels of SHBG (64), both prominent features of women with PCOS and of insulin resistance. Whether insulin resistance itself plays a role in the development of endometrial cancer is unknown, although studies by Nagamani et al. (65, 66) provide evidence supporting this possibility. Notably, there are no prospective studies of the incidence of endometrial cancer in PCOS.
Although OCPs in the general population have been associated with a significant reduction in endometrial cancer risk (50% decrease; Refs. 67, 68, 69), the frequency of induced cycles that prevents endometrial cancer specifically in women with PCOS is unknown. Standard practice to prevent endometrial cancer in women with PCOS is to induce a withdrawal bleed with a progestin on an every 1- to 3-month basis, but no studies justifying this practice in premenopausal women have been reported.
While studies with insulin-sensitizing drugs are limited, they suggest that administration of these drugs is associated with substantially increased frequency of ovulation, followed by menstrual bleeding, resulting in at least six ovulatory menses per year in 5585% of treated women (70, 71, 72). Although no prospective study focusing on prevention of endometrial cancer in PCOS by insulin-sensitizing drug has been conducted, this frequency of ovulation would be consistent with the current standard of care for the prevention of endometrial cancer in women with PCOS.
Furthermore, the theoretical possibility exists that ovulation induced by an insulin-sensitizing drug might be more successful in reducing endometrial cancer risk than a withdrawal bleed induced by exogenous progestins, because only the former may be associated with a multitude of physiological ovulation-induced hormonal events (e.g. local changes in endometrial growth factors) that may attenuate endometrial cancer risk.
Summary
Recognition that insulin resistance plays a major role in the pathophysiology of PCOS has led to the novel application of insulin-sensitizing drugs as a treatment option for the disorder. Less appreciated, however, is the high risk for type 2 diabetes and perhaps cardiovascular disease associated with insulin resistance, and the important implications this may have for chronic treatment of PCOS.
OCPs are the traditional therapy for the chronic treatment of PCOS, and they are recognized to exert a number of beneficial effects. Their use cannot be dismissed lightly. Nonetheless, limited evidence raises the issue that OCPs may aggravate insulin resistance and exert other untoward metabolic actions that possibly enhance the long-term risk for diabetes and heart disease. This important clinical issue has received relatively scant attention from clinical investigators and remains unsettled. This report is as much a call for additional studies, critical and controlled, as for judicious caution.
In the interim, it is prudent at the time of treatment to assess an individual woman with PCOS for risk factors for diabetes (e.g. positive family history, history of gestational diabetes, high-risk ethnic group) and heart disease (e.g. dyslipidemia, positive family history, elevated homocystine). This may help determine the relative risk to benefit ratio of OCP therapy and allow the clinician to tailor therapy to the individual.
For clinical investigators, the challenge now is to critically explore the immediate and long-term metabolic effects of both OCPs and insulin-sensitizing drugs in PCOS, so that clinicians may treat this prevalent disorder with confidence.
Acknowledgments
Footnotes
Abbreviations: OCP, Oral contraceptive pill; PAI-1, plasminogen activator inhibitor type 1; PCOS, polycystic ovary syndrome.
Received October 3, 2002.
Accepted February 10, 2003.
References