Diagnostic Criteria & Epidemiology
There are three main sets of diagnostic criteria for PCOS currently in use for both clinical and research purposes. The 1990 NIH diagnostic criteria require the presence of hyperandrogenism (defined biochemically and/or clinically) and menstrual disturbance (without reference to ovarian morphology). The 2003 Rotterdam diagnostic criteria for PCOS are based on the presence of at least two of the following three cardinal features: polycystic ovarian morphology on ultrasound scan, hyperandrogenic features and oligo-amenorrhea (defined as an intermenstrual interval >42 days).[9,10] More recently in 2009, the Androgen Excess Society published diagnostic criteria for PCOS, which state that PCOS should be defined by the presence of hyperandrogenism (defined on either clinical and/or biochemical grounds) and ovarian dysfunction (defined by oligo-anovulation and/or polycystic ovaries), with the exclusion of other related disorders.
The Rotterdam criteria for PCOS have broadened the diagnostic range of PCOS.[9,10] It has been demonstrated by our group that metabolic heterogeneity exists between Rotterdam-defined phenotypic subgroups,[9,10] with metabolic dysfunction (including insulin resistance) being restricted to the subgroup with both oligo-amenorrhea and hyperandrogenic features. Owing to a lack of long-term prospective studies in this field, the long-term effect of such metabolic heterogeneity on the future development of T2D and cardiovascular events is not clear, and should be a focus for future research.
A detailed discussion of the pathogenesis of PCOS is beyond the scope of this article (reviewed in[2,13]). An important aspect of PCOS pathogenesis is genetic predisposition, based on data from twin and family-based association studies.[14,15] Data from >1300 monozygotic twins and >1800 dizygotic twins/singleton sisters of twins from The Netherlands Twin Register have demonstrated a high heritability of PCOS with a monozygotic twin correlation of 0.71. Genetic data on metabolic traits from male relatives of women with PCOS provide further evidence to support a genetic predisposition to PCOS.[16–18] Although the actual gene variants that influence development of PCOS remain unknown, important candidates are those involved in ovarian and adrenal steroidogenesis, given that hyperandrogenemia is an important biochemical feature of PCOS. It is likely that hyperinsulinemia (resulting from insulin resistance/obesity) augments steroidogenesis through its action as a co-gonadotrophin on ovarian theca cells.
The pathophysiology of PCOS is complex and factors such as chronic inflammation and increased oxidant stress are likely to be implicated, with increased levels of advanced glycation end products having been demonstrated in women with PCOS. Endocrine disruptors such as bisphenol A may also be implicated in pathogenesis. Weight gain is also often an important pathogenic factor, with PCOS usually becoming clinically manifest in women with a presumable genetic predisposition for PCOS who subsequently gain weight. Therefore, environmental (particularly dietary) factors are important. However, BMI is also influenced by genetic factors such as the fat mass and obesity-associated protein (FTO), and obesity is a highly heritable condition. Therefore, the weight gain responsible for the manifestation of PCOS in many women with this condition is itself influenced by genetic factors. Recent genetic evidence from our own group has demonstrated that variants within FTO are associated with the development of PCOS, at least in part through effects on BMI. This was the first published evidence to genetically corroborate a link between obesity and PCOS.
It is likely that PCOS is an oligogenic condition with a handful of gene variants, each with relatively small effect sizes implicated in the genetic predisposition to the development of PCOS. The existing literature on the genetics of PCOS (mainly based on candidate-gene studies) is generally disappointing. The first genome-wide association study (GWAS) in PCOS, published recently by Chen and colleagues, demonstrated associations between PCOS and three loci: 2p16.3, 2p21 and 9q33.3. Future GWAS on large numbers of subjects are likely to identify further genome variants implicated in PCOS susceptibility. PCOS is a disease with an ovarian origin and it is likely that gene variants that influence pathways such as those involved in androgen biosynthesis and function, insulin resistance and proinflammatory mechanisms are implicated in susceptibility to the development of PCOS. In this review, we focus on variants relevant to T2D pathogenesis such as those that influence weight gain and insulin secretion.
Links Between Obesity, Insulin Resistance & PCOS
It is clear that obesity and PCOS are closely linked. Between 30 and 70% of women with PCOS are obese, and modest weight loss of just 5% can significantly improve many of the clinical and biochemical features of PCOS. Furthermore, PCOS often manifests following weight gain. In one study, PCOS was shown to be highly prevalent among women with morbid obesity (PCOS present in 17 of the 36 morbidly obese premenopausal women included in the study). Features of PCOS (e.g., hirsutism, testosterone levels, insulin resistance, menstrual cyclicity and ovulation) showed marked improvements, and PCOS frequently resolved after substantial weight loss induced by bariatric surgery. These observations support the view that obesity is one of the major factors implicated in the development of PCOS.
The link between obesity and PCOS is likely to be mediated by a variety of mechanisms, although the effects of insulin resistance and compensatory hyperinsulinemia seem to be of particular importance. Insulin resistance, as with many other features of the condition, manifests heterogeneity amongst women with PCOS. Our own group have demonstrated that insulin resistance and other dysmetabolic features appear to be confined to the subgroup of women with PCOS who have both hyperandrogenism and oligo-amenorrhea. Despite this heterogeneity, the majority of women with PCOS (between 50 and 90%) have been reported to be insulin resistant,[29,30] and improved insulin sensitivity is associated with improvements of many clinical features of PCOS. However, there is also substantial data available to support the notion that many women with PCOS, especially lean women with the condition, are not insulin resistant.[32,33]
The mechanism by which hyperinsulinemia is implicated in the development of PCOS includes the co-gonadotrophic effects of insulin (i.e., augmentation of gonadotrophin action by insulin) on ovarian and adrenal steroidogenesis, reflected by an association between levels of insulin and testosterone. Other mechanisms include the inhibitory effects of insulin on ovarian folliculogenesis and on the synthesis of sex hormone-binding globulin within the liver. The development of PCOS and T2D share many common features, including a prerequisite for a genetic predisposition (although the commonality of the genetic variants implicated is incompletely understood), coupled with subsequent weight gain and associated insulin resistance. The links between PCOS and T2D are discussed in more detail in the next section.
Women's Health. 2012;8(2):147-154. © 2012 Future Medicine Ltd.