Polycystic Ovarian Syndrome, Biomarkers, and Metformin: Research, Risk, and Reality

Abstract

The potential long-term metabolic and vascular consequences of polycystic ovarian syndrome (PCOS), the commonest reproductive endocrine disorder of women, continue to attract considerable interest. There have been numerous reports of perturbances in a range of potential vascular risk parameters in this condition. It is now well appreciated that women with PCOS as a group appear more insulin resistant for a given body mass index (BMI) and, linked to this, are at increased risk of glucose intolerance. They also display slightly raised triglyceride and lower high-density lipoprotein-cholesterol levels compared with women without PCOS and, in concert with this dyslipidemic pattern, exhibit a greater preponderance of smaller denser lowdensity lipoprotein particles (1). Modest elevations in inflammatory parameters such as C-reactive protein (CRP) (2, 3), indicating subclinical inflammation and independent of degree of obesity, have also been noted in PCOS. Additionally, affected women exhibit endothelial dysfunction in both microand macrovasculature (4), perturbation of fibrinolytic pathways (5), and subclinical atherosclerosis (6). Thus, the general thrust of evidence favors an enhanced vascular risk in women with PCOS. In this issue of JCEM, Heutling et al. (7) report yet another perturbance in a novel vascular risk parameter in women with PCOS. These investigators measured circulating asymmetric dimethylarginine (ADMA) levels in 83 women with PCOS, comparing results with data from 39 controls without PCOS. ADMA is a potent endogenous inhibitor of endothelial nitric oxide synthase, and accordingly, higher ADMA levels correlate to lower NO formation and endothelial dysfunction. The authors show that ADMA levels are not only significantly higher in women with PCOS but that they also correlate to BMI and fasting insulin in affected women. Moreover, in the subgroup of women treated with metformin, ADMA levels decreased significantly independent of a reduction in BMI, with a concurrent improvement in carotid intima-media thickness (IMT).These latterresults,althoughof interest, shouldbeexamined with caution because the number of women receiving metformin wassmall (n 21),anddurationof treatmentwasshort.Moreover, there was no placebo arm, and the IMT findings have not been replicated in non-obese PCOS (8) and may be solely attributable to weight loss (9). Consequently, although of potential great interest, there is a need to put both the current study and many previous similar studies relating PCOS and cardiovascular risk factors into a clinical context. With respect to the current report, although ADMA has been linked independently to risk for incident vascular events, the prospective data are limited to a few relatively small studies (most having 150 incident vascular events), and thus confidence intervals have been large. Moreover, the majority of such studies have been conducted in individuals with existing vascular disease (10, 11). In one of the largest studies in terms of number of events (n 182), patients in the top quarter of ADMA levels at baseline had an adjusted hazard ratio of 1.70 [95% confidence intervals (CI), 1.02–2.88] for incident major cardiovascular events (10). This hazard ratio, achieved with incomplete consideration of all established coronary heart disease (CHD) risk factors with no adjustment for high-density lipoprotein-cholesterol, is in fact relatively modest and suggests ADMA levels may not necessarily enhance CHD risk prediction beyond established risk factors. On the basis of predictive studies employing receiver operating characteristics analyses or equivalent, many other novel risk factors also fail to significantly enhance risk prediction (12). Indeed, even the value of CRP for risk factor stratification is now increasingly questioned (13). Nevertheless, if we accept that many risk factors are perturbed in women with PCOS, and that vascular risk is elevated, to what extent does this translate into more frequent or earlier events, thereby providing a rationale if any for screening? The definitive answer is unfortunately lacking because there have been no adequately powered prospective epidemiological studies with the requisite baseline phenotyping. However, analysis of the contribution from the relative components of the diagnosis of PCOS is helpful in this respect. In a long-term follow-up study of 786 women diagnosed with PCOS, primarily based on ovarian wedge resection histopathology, there was no significant increased risk of death from cardiovascular-related causes after an