PCOS and Cardiovascular Risks
by Walter Futterweit, M.D., F.A.C.P., F.A.C.E.,
OBGYN.net Editorial Advisor
Women with polycystic ovarian disease (PCOS) are the largest group of women at risk for the development of cardiovascular disease (CVD). Consensus data reveal the incidence of PCOS to range between 4.5-7.5 % of reproductive aged women. The finding of morphologic evidence of polycystic ovaries on pelvic ultrasonography in 23% of apparently normal women makes the procedure helpful but not diagnostic of PCOS. A number of other endocrinopathies may mimic these findings and a history of oligomenorrhea and evidence of ovarian hyperandrogenism are key elements in defining PCOS (1).
PCOS shares some or most of the features of the Metabolic Syndrome (Syndrome X, or Insulin Resistance Syndrome) (2). Most reports indicate that nearly ½ or more of women with PCOS fulfill the criteria for the metabolic syndrome (3). They present with a clustering of metabolic and vascular abnormalities. The purpose of this review is to assess the risk factors in women with PCOS and the available data on studies which attempt to analyze these risks and relate them to the relatively few known retrospective studies and reviews of cardiovascular events (4).
The features of the Metabolic Syndrome are manifested by:
1) Dyslipidemia
2) Hypertension
3) Insulin resistance (IR) and increased tendency to type-2 diabetes mellitus
4) Obesity, particularly the presence of central obesity (increased visceral fat)
Dyslipidemia: There are inconsistent findings which may account for differences in studies of women with PCOS. These include different diagnostic criteria, ethnicity, environmental, lifestyle factors (smoking, alcohol intake, physical activity) and genetic differences. In a study of 153 obese( BMI>27 kg/m2) and 42 nonobese non-Hispanic Caucasian women with PCOS having a mean age of ~28 years, lipids were compared to BMI and waist-to hip (WHR) matched controls mostly from same ethnicity of central Pennsylvania (5). Their findings indicated that women with PCOS have an increased LDL than did controls, independent of obesity and of significance is the fact that this occurred at a mean age of 10 years earlier than a previously reported in a large cohort of women by Talbott et al (6). The latter study demonstrated lipid changes to be more common in younger PCOS pts (central obesity) but no different than age and BMI matched controls >40 years (controls also had increased BMI and LDL). Legros data found no reduction of HDL levels in PCOS women compared to control women in Central Pennsylvania who often have low HDL levels (5). Other studies have found reduced HDL (7) and increased triglyceride (TG) levels in obese and nonobese hyperinsulinemic women with PCOS (8). Obesity distinctly has a synergistic effect on TG in women with PCOS.
Endothelial Dysfunction: Altered insulin regulation of endothelial nitric oxide synthesis leads to impaired NO-dependent vasodilatation. The arterial consequences of metabolic dysregulation leading to reduced large vessel vascular compliance and altered reactivity of arterial resistance appear to be risk factors in IR. Abnormal responsiveness of the endothelial lining of the blood vessels to vasodilators, which is very important for maintaining normal perfusion is noted. An abnormal vasodilatory response correlates with long term risk for CVD. This is a direct vasodilatory response assessed by arterial cannulation looking at leg blood flow in obese PCOS women and controls demonstrating a striking decrease in endothelial dependent vasodilation in women with PCOS (9). Here is the presence of an important risk factor for CVD in young women with PCOS.
This has also been demonstrated in studies of brachial arteries of young mean aged 26-year old women with PCOS when compared to controls. PCOS patients had abnormal vascular compliance to insulin in brachial arteries suggesting impaired insulin action (i.e., IR) in vascular tissue. This links IR in vascular tissue with endothelial dysfunction, decreased arterial compliance and possible later development of CVD in PCOS (9,10).
Hypertension: No increased incidence of HBP is noted in women with PCOS after adjustment for BMI. As patients with PCOS age they have a higher incidence of HBP than matched controls (11-13). A 3-fold risk of HBP was found in 33 older WR PCOS with a mean age of 52yrs (12). Increased systolic HBP was noted in PCOS (14).There is increased daytime systolic BP in PCOS women determined by 24 hr blood pressure monitoring, albeit in the normal range persisting after adjustment for BMI, insulin sensitivity and body fat distribution when compared to matched controls (15). Of interest was the association with hyperinsulinemia in both the obese and nonobese subjects in that study. 319 elderly PCOS women (mean age 57y) had a 2 1/2-fold incidence of nonfatal cerebrovascular events despite no increased CHD mortality compared to controls (12). Consensus: Inc BP with age>controls in women with PCOS.
Central Obesity: Intra-abdominal fat is the main fat store responsible for IR. It increases free fatty acid levels and leads to hypertriglyceridemia, small dense LDL particles, and decreased HDL.
Another surrogate marker for cardiovascular disease and atherosclerosis is carotid intimal wall thickness which can also be measured with sonography and the increased thickness correlates with a long-term risk for cardiovascular disease. When we look in women with PCOS (these were studies by Evelyn Talbott, U.Pittsburgh) (16), we see if we use a cut point of plaque thickness greater than 3 , women with PCOS have significantly greater prevalence of this increased plaque thickness.
A significant number of middle aged women with PCOS have evidence of premature carotid plaques using ultrasonography compared to controls (7.2 vs, 0.7%) (16). It supports the hypothesis that PCOS patients represent the largest group of young women for possible development of early-onset CVD, with risk factors often present years before clinical onset.
Proinflammatory and Atherogenic Markers in PCOS compared to controls show evidence of decreased fibrinolytic activity, that is, higher PAI-1 levels (17), and they also have increased C-reactive protein (CRP) levels which is a marker for inflammation which correlates well in epidemiologic studies with an increased risk for CVD (18-20). Elevated CRP levels were found in both the obese and nonobese women with PCOS (20). Treatment directed to lowering CRP levels (smoking cessation, diet, aspirin, metformin, and possibly statins) should probably be more aggressive in those women with PCOS with an increased CRP (2,19).
Type 2 Diabetes Mellitus: There is a significant increase in the incidence of impaired glucose tolerance (IGT) and T2DM in women with PCOS. Studies indicated that on the initial evaluation with a 2-hour GTT 30-40% of PCOS already have IGT or T2DM (21-23). These findings also may be present in younger teens with PCOS (23) and it is suggested that the pancreatic beta-cell may be unable to compensate for the IR (24). The higher prevalence of impaired GTT in PCOS compares to an incidence of ~ 2% in a normal female population < 40 years of age. A higher prevalence of IGT and T2DM occurs in the obese women with PCOS, particularly those with a positive family history of T2DM. It is clear that oligomenorrhea is a surrogate marker for probable PCOS and may predict a 2 to 2.5 increased risk for T2DM, particularly in the presence of a family history of type 2 diabetes mellitus (25). It must be remembered that the presence of T2DM abolishes the protective effect inherent in premenopausal women without T2DM, adding a significant risk factor for CVD.
Insulin Resistance: The presence of IR is considered to be one of the important factors in the pathogenesis of PCOS. A consensus of studies indicates an incidence of at least 50%. Obesity is a major contributing factor although a number of oligomenorrheic nonobese women with the syndrome have IR as well (26). It should be noted that not all women with IR secondary to hyperinsulinemia will have PCOS (27) and this may be related to genetic differences of the ovary and pancreas. Obesity or the development of obesity often is a synergistic trigger that promotes the symptoms of PCOS associated with hyperandrogenism. Parenthetically, some nonobese women with PCOS may not demonstrate IR.
IR may be difficult to assess accurately in the absence of hyperinsulinemic-euglycemic clamp studies, particularly when the basal insulin levels are in the normal or high normal range and in the nonobese subgroup of PCOS women. Recent data demonstrate the failure of the usual means of assessing IR with methodologies that only measure fasting glucose and insulin levels including homeostasis model assessment (HOMA) and the quantitative insulin sensitivity check index (QUICKI) in assessing insulin sensitivity (28). A recent landmark study by Diamanti-Kandarakis et al (28), however, of 59 women with PCOS of varying body weights, demonstrate a lack of correlation of the HOMA and QUICKI methodologies and insulin sensitivity as determined by the euglycemic-hyperinsulinemic clamp. This study underscores the probable underestimation of published studies of IR in women with PCOS, who have a unique form of IR, and where mild IR may be present with borderline normal fasting glucose and insulin levels. Perhaps other hormonal and genetic factors, as well as ethnicity, may influence the degree of IR, and yield conflicting reports of the incidence of IR in women with PCOS.
Real Events Studies
Dahlgren et al have calculated via a risk model analysis that patients with PCOS had a 4 to 7-fold higher risk of myocardial infarction compared to age matched controls (29). In a study of 143 women undergoing cardiac catheterization, aged 60 years or younger, PCOS was detected in 42% of women on pelvic ultrasonography. Patients with PCOS exhibited coronary artery segments with >50% stenosis with significantly greater clinical heart disease than women with normal ovaries on ultrasonography (30).
A Mayo Clinic study of coronary calcifications with electron-beam CT (EBCT) by Christian et al of 36 non-diabetic women with PCOS 30-45 year old revealed a 3-fold higher level of coronary artery calcification (CAC) than population controls (31). When compared to obese controls there was a 2-fold increase in CAC in PCOS subjects. A correlation was found between CAC score and BMI, visceral adiposity, and elevated levels of serum triglycerides in PCOS women.
A prospective Nurses Health Study of over 101,000 women linked menstrual irregularity, not only to an increased 2-2.5 fold risk of diabetes mellitus (25), but to an increased risk of mortality due to fatal coronary heart disease. In a prospective Nurses Health Study of 82,439 nurses, an average follow-up of 14 years was performed (32). A history of oligomenorrhea was noted in the women studied, and a 50% increased incidence of fatal and nonfatal instances of coronary heart disease (CHD) was found in this subgroup of women. This study indicated that oligomenorrhea is a good surrogate marker for the potential development of cardiovascular disease (CVD).
Pierpoint et al studied a total of 786 women diagnosed with PCOS (mostly by having had a wedge resection of the ovaries) in the UK between 1930 and 1979 whose records were traced from hospital records and who were followed for an average of 30 years (33). Standardized mortality rates (SMRs) were calculated to compare death rates of these women with the national rates. There were 59 deaths prior to the age of 79 years: 15 deaths from circulatory disease, 13 of who were from ischemic heart disease, and 2 from other circulatory diseases. Diabetes mellitus was also found to be more commonly mentioned in analysis of data and contributing to mortality in the PCOS group. Breast cancer was the commonest cause of death. The SMR for women with PCOS was 0.83 compared to the national average of 0.90. They concluded that women with PCOS do not have a higher than average mortality from circulatory disease, despite the numerous risk factors for CVD. They speculated that the endocrine profile in these women may protect them against CVD.
In the retrospective UK study by Wild et al, 319 PCOS women (mean age 57 yr, with a range of 38-98 yr, most of whom had a tissue diagnosis by wedge resection) there was no significantly higher incidence of coronary heart disease than women in the general population (12). However, a 2.5-fold higher incidence of T2DM and nonfatal cerebrovascular events was noted than age-matched controls. It should be noted that in the UK study only had 26% of patients demonstrated obesity (mean BMI was 26 kg/m2), while the incidence in the USA is higher, and may range between 50-60%.
Concluding Remarks:
It is clear that many factors make any study of PCOS difficult, particularly the UK retrospective studies by Pierpoint and Wild et al.
1) Lack of adequate diagnostic criteria. The heterogeneity of these criteria is confusing.
2) The presence or absence of measurements of insulin resistance, central fat distribution, and co-existing morbidities has not been addressed.
3) Many patients are not included in view of the fact that they may not have been followed adequately, and a number may not have been included for a variety of other reasons including change of tenure or lack of follow-up response in view of change of address, or other factors including illness or mortality. This may underestimate the incidence of women with PCOS and the presence of CVD.
4) Changes in the phenotype of women with PCOS occur with advancing age, and as is well known, most women with PCOS have reduced hyperandrogenism and more regular menses after the age of 40 years. Some studies which are stated above included a relatively small sample size.
5) The presence or absence of anovulation is not stated in some studies. Women with regular menses and those with an absence of visceral adiposity may dilute the incidence of CVD. There is as yet an absence of a study of the menopausal phenotype in PCOS (4).
6) The UK studies mostly include women who have had a wedge resection of the ovaries. Thus therapeutic intervention may possibly have contributed to the results found by the investigators.
One must conclude that at the present time one can only state the women with PCOS have a risk profile which appears to be predictive of a higher incidence of CVD. The available data on actual cardiovascular events may be suggestive, but does not definitively support the prediction. A prospective multicenter study of a large number of aged women with PCOS is sorely needed to answer this important question.