Cardiovascular Risk in PCOS

In clinical practice, women with polycystic ovarian syndrome (PCOS) are often seen for immediate concerns such as infertility, menstrual irregularity, and symptoms of androgen excess. During the past two decades, however, such patients have been observed to have an increased prevalence of diabetes (1) and of risk factors for cardiovascular disease. Specifically, many women with PCOS are similar to those with metabolic cardiovascular syndrome (“Syndrome X”) (2). In addition, the chronic anovulation of PCOS implies unopposed estrogen and, therefore, an increased risk of endometrial cancer. These factors have led to a different clinical perspective about PCOS—one that recognizes the need to address the immediate issues of irregular bleeding, hirsutism, and infertility, but also emphasizes the long-term goals of preventing diabetes, heart disease, and cancer. Nearly 40% of women with PCOS will have impaired glucose tolerance or overt type 2 diabetes, a finding that is consistently seen across several geographic areas and ethnic groups (3). Moreover, women with PCOS are more likely than normally cycling women to have insulin resistance (4), central adiposity, dyslipidemia, and hypertension (5). Other markers of cardiovascular disease, such as C-reactive protein (6) and homocysteine (7), have also been found to be elevated in women with PCOS. In addition to serum markers, several measures of subclinical atherosclerosis, such as carotid intima-media thickness (IMT) and coronary artery calcium (CAC), have been studied. Women over the age of 45 yr with PCOS have been found to have greater IMT than cycling controls (8) and greater CAC (9). In this issue of JCEM, Orio et al. (10) evaluate additional subclinical measures of heart disease, namely left ventricular hypertrophy and other measures of cardiac performance as measured by echocardiography, in a case-control study of 30 young women with PCOS and 30 controls. Although the sample size is small, a major strength of the design is that cases and controls were matched by age (±2 yr) and body mass index (BMI) (±1 kg/m2). In addition to the echocardiographic assessments, PCOS cases and controls were also compared with respect to insulin and lipids. Considering the insulin and lipid results first, Orio et al. (10) found that fasting insulin concentrations and the homeostasis model assessment (HOMA) index were higher in PCOS subjects than in controls, even in comparisons stratified by BMI into normal, overweight, and obese women. This finding is consistent with many previous studies. Also consistent with previous reports were the findings that low-density lipoprotein cholesterol was higher and high-density lipoprotein cholesterol was lower in PCOS women than in controls. This adverse lipid profile was apparent in the overweight and obese BMI groups, but was not statistically significant in the normal weight group, a finding that may reflect an interaction effect between weight and PCOS on cardiovascular risk. Despite the fact that heavier women have a greater baseline level of cardiovascular risk, the incremental cardiovascular risk attributable to PCOS may be even greater as weight increases. The authors’ echocardiographic findings build upon earlier reports of alterations in the cardiac flow of patients with PCOS and are novel in demonstrating an increased left ventricular mass index (LVMi) among young women with PCOS who are asymptomatic and do not have any other risk factors for cardiovascular disease. In an effort to determine the determinants of increased LVMi, a stepwise linear regression analysis was performed on the data from PCOS patients. When LVMi was regressed on age, BMI, HOMA, fasting insulin, lipids, and blood pressure, only HOMA entered the equation as a significant independent variable. It is possible that the sample size was too small to produce reliable estimates for several independent variables that may affect LVMi. It is also possible, however, that insulin is the key mediating variable between PCOS status and cardiac dysfunction, as suggested by the authors in their discussion. In view of the mitogenic effects of insulin on myocardial tissue, there is biological rationale for such a relationship, which is a provocative observation deserving further investigation. No doubt, the authors will be encouraged by their findings to collect data on larger samples of cases and controls, and other investigators should consider undertaking their own confirmatory studies. As a complementary research design, longitudinal study of echocardiographic parameters in asymptomatic women with PCOS may shed light on the question of whether chronic PCOS leads to greater impairment of cardiac function across time. The core question, however, is whether alterations in serum markers (e.g. lipids, homocysteine, C-reactive protein) and in subclinical measures of cardiovascular disease (e.g. IMT, CAC, LVMi) ultimately lead to an increased likelihood of actual cardiovascular events among women with PCOS. There is some literature to suggest otherwise, which will be addressed shortly, but first let us consider the general issue of risk factors for cardiovascular disease among women. After an extensive, systematic review of the literature on cardiovascular risk factors in women, the American College of Cardiology (11) has identified the following set of risk factors as being most important: age, smoking, diabetes, hypertension, obesity, elevated serum low-density lipoprotein cholesterol, and low serum high-density lipoprotein cholesterol. With the exception of age and smoking, which are not pertinent to the presence or absence of PCOS, all of the other factors that increase cardiovascular disease are increased in women with PCOS. Unless there is an aspect of the pathophysiology of PCOS that ameliorates the effects of hypertension, diabetes, hyperinsulinemia, or dyslipidemia on cardiovascular disease, it follows that individuals who have a greater prevalence of such conditions (i.e. women with PCOS) have a greater likelihood of developing actual cardiovascular events. That said, documentation of a clear association between PCOS and cardiovascular events or mortality is lacking. Indeed, Pierpoint et al. (12) found no increase in the rate of deaths due to circulatory disease among a large sample of women with PCOS when compared with expected death rates based on age- and sex-specific rates of mortality in the United Kingdom. This study was based on a sample of 786 women who had previously undergone ovarian wedge resection between 1930 and 1979. Although deaths due to cardiovascular disease were not increased, death due to diabetes was significantly higher than expected (six observed vs. 1.7 expected; P = 0.002). In a follow-up study (13), the same investigators sent questionnaires to 345 cohort members for the initial PCOS sample and to 1107 control women obtained from the same general practice. From the questionnaire data, it was found that the odds ratio 95% confidence interval (CI) for coronary heart disease was 1.5 (0.7–2.9) based on 15 coronary heart disease events in the PCOS group. The odds ratio for cerebrovascular disease was higher (2.8) and reached statistical significance (95% CI, 1.1–7.1) These studies are important because of their negative findings regarding deaths from circulatory disease. Interpretation of these results is constrained by several factors, however. First, the diagnosis of PCOS cases was based mainly on hospital records related to wedge resection. This is a potentially confounding factor, because surgical removal of ovarian stroma can treat PCOS as well as diagnose it. Indeed, wedge resection can correct the anovulation and metabolic changes that are seen in PCOS over long periods of time. Second, wedge resection as a method of case ascertainment may substantially underestimate the syndrome of PCOS as defined by chronic anovulation and androgen excess. Third, 23% of the women diagnosed with PCOS could not be traced. To the extent that deaths in the untraceable group are higher than in the group that was traced, there may be a downward bias in the estimates of mortality rates in the PCOS groups, although the authors did not believe that this was present. Fourth, the authors found a relative risk of 1.5 for cardiovascular disease. This was not statistically significant because of the limited number of events; however, a relative risk estimate of 1.5 may be clinically significant if such an estimate is confirmed with larger samples. The issue of sample size becomes especially important when it is recognized that the average age of the PCOS women in the first study was in the low 50s. Thus, on average, they were not yet at an age when cardiovascular events occur with any frequency. A much larger sample size was available for analysis from the Nurses Health Study, a prospective cohort study involving over 120,000 nurses in eleven states. Using menstrual cycle irregularity as a marker for PCOS in the analysis of these data, Solomon et al. (14) estimated an adjusted relative risk (95% CI) of 1.53 (1.24–1.90) for coronary heart disease. This is the same effect size as that found in the British study reported above. The study by Orio et al. (10) suggests that women with PCOS may undergo, in an asymptomatic fashion, adverse alterations in cardiac function at an early age, well before the appearance of hypertension or subclinical changes such as increased IMT or plaque. Such data gives further credence to a clinical management strategy of women with PCOS that not only addresses their immediate symptoms and concerns effectively but also undertakes appropriate counseling and surveillance regarding longer-term risks. Body mass index; Coronary artery calcium; confidence interval; homeostasis model assessment; intima-media thickness; left ventricular mass index; polycystic ovarian syndrome.