Divergences in insulin resistance between the different phenotypes of the polycystic ovary syndrome

Polycystic ovary syndrome (PCOS) is a common condition that often becomes manifest during adolescence [1]. Polycystic ovary syndrome presents with both hyperandrogenic (hirsutism, acne and androgenic alopecia) and reproductive (oligoamenorrhea and impaired fertility) features. Biochemically, hyperandrogenemia is the most common manifestation [1]. Polycystic ovary syndrome is also often associated with metabolic dysfunction including insulin resistance [2]. At least, part of this association is likely to be mediated via obesity-related effects. Evidence from epidemiological and genetic studies confirms the importance of obesity as a pathogenic factor that often unmasks PCOS in women who are genetically predisposed [1,3]. Although weight gain is not an essential element in all women with PCOS (a subgroup of women with PCOS are lean), a high proportion (between 38 and 88%) of women with this condition are either overweight or obese [1]. It follows that fat mass is likely to be an important contributor toward prevalence of PCOS. Although PCOS is estimated to affect between 5 and 9% of premenopausal women [4], this figure is likely to increase in the future as the global obesity epidemic ensues.

Polycystic ovary syndrome is heterogeneous. This heterogeneity pervades many aspects of the condition, including clinical and biochemical features and ovarian morphology. This heterogeneity is also reflected by the way in which PCOS is now defined. In 2003, the internationally agreed Rotterdam diagnostic criteria for PCOS [5] superseded the prior NIH diagnostic criteria (requiring presence of both menstrual disturbance and hyperandrogenic features without reference to ovarian morphology) [6]. The Rotterdam diagnostic criteria [5] utilizes a ‘two out of three’ rule; following exclusion of other pathologies, a patient can be diagnosed with PCOS if she manifests at least two out of three cardinal features: menstrual irregularity and anovulation (inter-menstrual interval of >42 days), hyperandrogenism (raised serum testosterone, hirsutism, acne, androgenic alopecia) and polycystic ovarian (PCO) morphology on ultrasound scan [5].

The introduction of the Rotterdam diagnostic criteria for PCOS [5] has led, by definition, to the emergence of four distinct phenotypic subgroups. For clarity, we refer to these according to the abbreviations ‘P’ (presence of PCO morphology), ‘H’ (presence of hyperandrogenic features) and ‘O’ (presence of oligoamenorrhea). Two of these subgroups (‘PHO’ and ‘HO’) are equivalent to PCOS as defined by the older NIH diagnostic criteria [6]. However, the Rotterdam diagnostic criteria have also led to the generation of two extra subgroups that do not form any part of the older NIH criteria [6]: ‘PH’ (women with PCO morphology and hyperandrogenic features but normal menses) and ‘PO’ (women with PCO morphology and oligoamenorrhea but normal androgen levels). Perhaps, most controversially is the inclusion of the ‘PO’ group, given their normoandrogenemic status. This broadening of the diagnostic range of PCOS and the generation of multiple phenotypic subgroups has been a focus of much research in recent years. One important clinical question, for which we now have much published evidence to address conclusively, relates to the metabolic status (particularly insulin resistance) amongst women within each subgroup.

Regardless of diagnostic criteria employed, PCOS as a whole is associated with an adverse cardiometabolic profile [2]. It has been well established that women with PCOS have much higher rates of metabolic syndrome compared with women without this condition: in studies on white women with PCOS from the USA, prevalence of metabolic syndrome (based on National Cholesterol Education Program Adult Treatment Panel III [NCEP ATPIII] criteria) has been estimated to be between 34 and 46% [2]. Women with PCOS also have a significantly increased risk for the development of Type 2 diabetes mellitus [7]. An important contributor to this adverse cardiometabolic risk is through obesity-related effects [1]. However, it is important to emphasize that PCOS per se (and independent of any fat mass effects) is associated with adverse cardiometabolic risk [7]. Obese women with PCOS should therefore be considered to be at particular risk of metabolic disorders.

It is likely that an important factor that mediates obesity-related effects on the manifestation of PCOS and one that is central to adverse cardiometabolic profile is insulin resistance. There is a clear association between PCOS and insulin resistance, and this varies according to the population studied and the way in which insulin resistance is defined and assessed [8]. It is likely that insulin resistance plays an important pathogenic role in the development of PCOS that implicates pleiotropic effects of hyperinsulinemia on peripheral tissues, including the ovary (by acting as a ‘co-gonadotrophin’ in both steroidogenic actions [9] and in disordered follicle maturation) [10]. Furthermore, improving insulin sensitivity in women with PCOS results in improved menstrual cyclicity and fertility [1]. Given the pathogenic and therapeutic importance of insulin resistance for PCOS, it is imperative to consider the implications of the Rotterdam diagnostic criteria for PCOS on insulin resistance within the resulting phenotypic subgroups.

In one of the first published retrospective studies in this field, our own group reported data on 309 Europid women with PCOS (Rotterdam defined [5]) from the UK, of whom the majority (n = 191) belonged in the ‘PHO’ subgroup [11]. This cohort also consisted of women within the ‘PH’ subgroup with normal menses (n = 76) and ‘PO’ normoandrogenemic subgroup (n = 42). For comparison, we included Europid control women without PCOS (n = 76). Insulin resistance was confined to the ‘PHO’ subgroup, with insulin sensitivity in the ‘PH’ and ‘PO’ subgroups being equivalent to that in the control women [11]. Following adjustment for differences in BMI and age between women in the ‘PHO’ and ‘PO’ subgroups, significant differences in insulin resistance remained, suggesting that insulin resistance is at least in part inherent to the ‘PHO’ subgroup, although also contributed toward by associated obesity [11]. Prevalence of metabolic syndrome between the phenotypic subgroups followed a similar pattern to that of insulin resistance [11]. In a previously reported study on women from the UK, it was shown that those in the ‘PH’ subgroup are less insulin-resistant than those in the ‘PHO’ subgroup [12].

Other reported studies in this field have produced consistent results. Dewailly and colleagues demonstrated in a study on 406 French women with PCOS that the ‘PO’ subgroup had lower fasting insulin concentrations than women in other phenotypic subgroups [13]. In a study on 418 women with PCOS from Iceland and the USA, Welt and colleagues showed that fasting insulin concentrations are highest in the ‘PHO’ subgroup compared with that in other phenotypic subgroups [14]. In a comparison of obese women with PCOS, Broekmans and colleagues demonstrated that those within the ‘PO’ subgroup had a milder metabolic phenotype than those in the ‘PHO’ and ‘HO’ subgroups [15]. More recently, a study from Turkey showed that insulin resistance in the ‘PO’ normoandrogenemic subgroup of PCOS was similar to that in the control women and significantly lower than that in the ‘PHO’ subgroup [16]. A study from Greece demonstrated similar patterns of insulin resistance and metabolic syndrome in comparisons between ‘PHO’ + ‘HO’ subgroups and ‘PH’ + ‘PO’ subgroups [17]. In one of the largest studies to date in this field, Zhang and colleagues reported on data from 804 women from China (including 719 cases of PCOS and 85 control women) [18]. Similar patterns of insulin resistance and metabolic syndrome amongst the phenotypic subgroups were shown [18]. Finally, in a more recent study by Moghetti and colleagues on Caucasian women with PCOS and using a glucose clamp technique to assess insulin sensitivity, it was shown that women in the ‘PHO’ subgroup were insulin-resistant, those in the ‘PO’ group were insulin-sensitive and those in the ‘PH’ group had insulin sensitivity that was intermediate between that in ‘PHO’ and ‘PO’ subgroups [19]. In summary, the existing data seem very clear and consistent: insulin resistance in PCOS varies widely according to phenotypic subgroup. Women in the ‘PHO’ subgroup are typically insulin-resistant and have an adverse cardiometabolic profile, whilst those in the newer ‘PH’ and ‘PO’ subgroups are typically insulin-sensitive and metabolically equivalent to control women.

Polycystic ovary syndrome is complex and heterogeneous [3,20]. In contrast to Type 2 diabetes mellitus, which is a well-defined entity based on clear-cut biochemical diagnostic criteria, PCOS is a syndrome that is heterogeneous on many levels and therefore more problematic to define. The multidimensional nature of PCOS reflects its heterogeneity. As outlined though, insulin resistance appears to be central to PCOS in many women with this condition. We believe that there are at least three reasons why we should be concerned with insulin resistance amongst Rotterdam-defined [5] phenotypic subgroups of PCOS: to provide insight into the pathogenic role of insulin resistance and factors influencing its development; to provide guidance for healthcare professionals caring for women with PCOS regarding screening for cardiometabolic risk according to phenotypic subgroup and to provide clinically useful data regarding future cardiometabolic risk according to phenotypic subgroup. One important limitation of the reported studies to date is their cross-sectional design, providing only a snapshot view of insulin resistance within each of the phenotypic subgroups of PCOS. Migration between phenotypic subgroups over time and the longer-term implications of phenotypic history remain unanswered. To fully address this would require a long-term, prospectively designed study that follows women within each phenotypic subgroup over time. Such a study should outline ‘phenotypic journeys’ (whether women with PCOS tend to remain in just one subgroup, or whether they flit between subgroups over time and the determinants of such migrations), the longer-term implications of ‘phenotypic history’ in terms of translation into cardiometabolic events and ultimately, the morbid and mortal consequences of such events. This strategy should be a focus for future research, to gain further insight into the determinants and long-term consequences of insulin resistance in women with PCOS, particularly, how this relates temporally to the phenotypic diversity of this fascinating condition.

Acknowledgments

We acknowledge the many patients, relatives, nurses and physicians who contributed to the ascertainment of the various clinical samples reported on in this Editorial.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.