DOI 10.1515/hmbci-2013-0014 Horm Mol Biol Clin Invest 2013; 15(3): 91–103
Thomas M. Barber*, Petra Vojtechova and Stephen Franks
The impact of hyperandrogenism in female obesity and cardiometabolic diseases associated with polycystic ovary syndrome Abstract: Polycystic ovary syndrome (PCOS) is a common condition characterized by reproductive and hyperandrogenic features and is often associated with obesity and metabolic dysfunction. Overall, women with PCOS have a substantially greater prevalence of metabolic syndrome than women from the general population. Furthermore, PCOS per se (independent of its frequent association with obesity) often confers cardiometabolic risk (including insulin resistance), and its concurrence with obesity often represents a metabolic “double-whammy” from the adverse effects of PCOS and obesity. The introduction of the Rotterdam diagnostic criteria for PCOS in 2003 has broadened the scope of this condition. The Rotterdam diagnostic criteria have also introduced two new phenotypic subgroups (including normoandrogenemic women with PCOS) that have provided novel insights into a potential role for hyperandrogenism in the development of adverse cardiometabolic risk in women with PCOS. Based on evidence from cross-sectional and interventional studies, hyperandrogenism, obesity, and cardiometabolic risk in women appear to be linked through complex and multidirectional pathways. Furthermore, data from obese women without a formal diagnosis of PCOS also suggest that these interrelationships often exist in female obesity per se (in milder forms than occurs in PCOS). Data from female-to-male transsexuals are particularly informative because these show direct effects of hyperandrogenism (induced through exogenous use of androgenic therapies) on fat distribution and cardiometabolic risk in women. A challenge for the future will be to disentangle and improve our understanding of this complex pathogenic web, thereby facilitating novel and targeted therapies for the hyperandrogenic and adverse cardiometabolic manifestations of PCOS. Keywords: cardiometabolic risk; hyperandrogenism; obesity; polycystic ovary syndrome.
*Corresponding author: Thomas M. Barber, DPhil, Clinical Sciences Research Laboratories, Division of Metabolic and Vascular Health, University of Warwick, University Hospitals Coventry and Warwickshire, Clifford Bridge Road, Coventry CV2 2DX, UK, Phone: +44-2476-968591, Fax: +44-2476-968653, E-mail: [email protected] Petra Vojtechova: Clinical Sciences Research Laboratories, Division of Metabolic and Vascular Health, Warwick Medical School, University of Warwick, University Hospitals Coventry and Warwickshire, Coventry, UK Stephen Franks: Institute of Reproductive and Developmental Biology, Imperial College, London, UK
Introduction Hyperandrogenism in women is common and is defined on the basis of clinical and/or biochemical criteria [1]. Clinical features of female hyperandrogenism include hirsutism, acne, and alopecia. Biochemical features (termed hyperandrogenemia) include elevated serum levels of androgens (increased total testosterone, androstenedione, dehydroepiandrosterone [DHEAS], and free androgen index, which is influenced by the suppression of serum sex hormone binding globulin [SHBG] level) [1]. Hyperandrogenism is one of the diagnostic features of polycystic ovary syndrome (PCOS) and is manifested in many other conditions [2]. There is also a clear association between female obesity and hyperandrogenic features, both in the context of PCOS and in women without PCOS [1]. There is evidence that female hyperandrogenism is also associated with cardiometabolic risk factors and that exposure to hyperandrogenemia (either prenatally or as an adult) may influence body fat distribution (BFD), insulin resistance, and other cardiometabolic features [3, 4]. It is likely that the mechanisms linking female hyperandrogenism with obesity and cardiometabolic dysfunction are complex and bidirectional. However, most of the studies reported in this field are based on cross-sectional data, and there is a lack of longer-term prospective studies that provide clear evidence for direct effects of serum androgens on the adipocyte and vice versa. Therefore, our current understanding of pathogenic mechanisms in this field is incomplete.
92 Barber et al.: Androgens, obesity, cardiometabolic risk in PCOS In this brief review article, we discuss the associations among PCOS, metabolic dysfunction, and cardiometabolic diseases. We then explore data that support an association between female hyperandrogenism and cardiometabolic dysfunction and data that support a role for obesity in the development of female hyperandrogenism. We discuss the possible mechanisms whereby female hyperandrogenemia may influence BFD and cardiometabolic risk and explore data relating to management of both obesity and hyperandrogenism in women and their respective effects on androgen levels and cardiometabolic risk. Finally, we provide an outlook for the future in this field and consider potential clinical implications of hyperandrogenism in the management of obese women with PCOS.
PCOS, metabolic dysfunction, and cardiometabolic diseases PCOS is a common condition that affects between 7% and 10% of women of reproductive age [1]. Reproductive features of PCOS include oligo-amenorrhea, manifesting clinically with irregular and/or absent menses, and impaired fertility. The most common biochemical manifestation of PCOS is hyperandrogenemia (including increased serum total testosterone, androstenedione, and free androgen index), which manifests with hirsutism, acne, and alopecia [1]. It is also clear that metabolic dysfunction is much more common in women with PCOS compared with ageand body mass index (BMI)-matched women within the general population [5–10]. The metabolic syndrome is a constellation of interrelated features [11], the clinical importance of which lies with its association with a twofold increased risk of cardiovascular disease and a fivefold increased risk of developing type 2 diabetes mellitus (T2D) [12]. Using the National Cholesterol Education Program Adult Treatment Panel III criteria in a population-based study on White women from the USA, it has been estimated that metabolic syndrome affects between 34% and 46% of women with PCOS [7–10]. One of the largest studies in this field was published by Ehrmann et al. [7], who reported on the prevalence of metabolic syndrome in 394 (mostly White) women with PCOS: in those women with a BMI >27 kg/m2, 40% had metabolic syndrome. These figures compare with a prevalence of metabolic syndrome [using National Health and Nutrition Examination Survey (NHANES) III data] within the general female US population of 6% (ages 20–29 years) and 15% (ages 30–39 years). It is a reasonable assumption that the high prevalence of metabolic syndrome in women
with PCOS would translate into an elevated cardiometabolic risk in postmenopause, although this hypothesis has yet to be definitively proven through a long-term prospective cohort study in PCOS. There is a real need for such a study to be performed, particularly given that retrospective studies on PCOS in postmenopausal women are fraught with difficulties [13]. However, it is clear from a large cross-sectional study that PCOS is associated with a significantly increased risk for the development of T2D [14], and the cardiometabolic risks associated with T2D are incontrovertible. The data outlined here provide a rationale for active screening for an early and aggressive management of cardiometabolic risk factors in women with PCOS. Although metabolic dysfunction forms an important component of the clinical and biochemical manifestations of PCOS in many women with the condition, it is interesting that none of the current diagnostic criteria for PCOS incorporate any metabolic features but rather focus on the reproductive, hyperandrogenic, and ovarian morphological manifestations of the condition. There are two main sets of diagnostic criteria currently used for PCOS. These are the National Institutes of Health (NIH) criteria, proscribed in 1990 and which require the presence of both hyperandrogenism/hyperandrogenemia and menstrual disturbance (regardless of ovarian morphology) [15]. The NIH criteria were superseded by the Rotterdam criteria in 2003, which require the presence of at least two out of the following three criteria: polycystic ovarian (PCO) morphology on ultrasound scan, hyperandrogenic features (biochemical and/or clinical), and oligo-amenorrhea (intermenstrual interval >42 days) [2, 16]. PCOS has a complex pathogenesis. Based on a large Dutch twin-family study and family-based association studies, it is clear that PCOS is a genetic condition with a high heritability (0.71) [17, 18]. Recent data from genomewide association studies (GWAS) in PCOS are shedding important new insights into its complex etiology [13]. In the first of these, Chen et al. [19] demonstrated associations between PCOS and three loci located at 2p16.3, 2p21, and 9q33.3. Data from GWAS will enable a more targeted approach toward gaining a clearer understanding of the pathogenesis of PCOS. It is also clear that obesity is closely associated with PCOS, with between 38% and 88% of women with PCOS being either overweight or obese [1]. Our current understanding of PCOS pathogenesis is that this condition becomes manifest in genetically predisposed women who subsequently gain weight (through dietary means principally) [13]. However, this is clearly an oversimplification, given that PCOS can occur in a minority of lean women. Furthermore, BMI is itself heritable,
Barber et al.: Androgens, obesity, cardiometabolic risk in PCOS 93
raising the possibility (indeed likelihood based on data on variants within the fat mass and obesity-associated gene [FTO] and PCOS [20]) that at least some of the genetic predisposition toward PCOS is through variants that influence fat mass. To summarize this section, PCOS is common, and although it is defined on the basis of reproductive and hyperandrogenic features, it is also associated with substantial metabolic dysfunction that is at least in part often related to its concurrence with obesity (although PCOS per se is associated with insulin resistance and other metabolic abnormalities) [1]. The remainder of this review will focus on the interplay between hyperandrogenism and female obesity and the cardiometabolic features associated with PCOS. It has been demonstrated that, as expected, the association of PCOS with metabolic syndrome is particularly strong among the overweight and obese subgroup (with a BMI >27 kg/m2) [7, 8]. Perhaps more surprisingly though, it has been shown by Apridonidze et al. [10] that a major predictor of metabolic syndrome occurrence in women with PCOS is elevated serum free testosterone and reduced serum SHBG levels. It has also been shown that a significant inverse relationship exists between SHBG level and metabolic syndrome occurrence in women with PCOS [21]. These data suggest that hyperandrogenemia may play an important role in the development of cardiometabolic risk in women with PCOS, and this forms the focus for the next section.
Phenotypic subgroups of PCOS and the importance of hyperandrogenism as a risk factor for cardiometabolic disease One of the consequences of the Rotterdam diagnostic criteria for PCOS [2, 16] is that it has led to the generation of four distinct phenotypic subgroups. In addition to the subgroup with all three diagnostic features present [“PHO”: PCO morphology (P), hyperandrogenic features (H), and oligo-amenorrhea (O)], three other phenotypic subgroups also exist, each with only two of these features present: “PH”: women with PCO morphology and hyperandrogenic features; “PO”: women with PCO morphology and oligo-amenorrhea; and “HO”: women with hyperandrogenic features and oligo-amenorrhea. Importantly, the PO subgroup has, by definition, normal androgen levels, and along with the “PH” subgroup, would not have been recognized as PCOS on the basis of the prior 1990 NIH
diagnostic criteria [15]. The study of metabolic dysfunction within each subgroup (with particular focus on the normoandrogenemic PO subgroup) will therefore provide useful insight into a potential role for hyperandrogenemia in the development of cardiometabolic risk in women with PCOS. In a large retrospective data set analysis based on 309 Europid women with PCOS, of whom 42 were in the “PO” (normoandrogenemic) phenotypic subgroup and compared with 76 Europid control women without PCOS, our group demonstrated that following adjustments for BMI and age, the women in the “PO” subgroup were indistinguishable from control women with respect to insulin sensitivity and were significantly less insulin-resistant than women in the “PHO” subgroup [22]. Furthermore, the prevalence of metabolic syndrome in women within the “PO” subgroup (7.1%) was similar to that in controls (3.9%) and significantly lower than that in women from the “PHO” subgroup (29.3%, p
Thomas M. Barber*, Petra Vojtechova and Stephen Franks
The impact of hyperandrogenism in female obesity and cardiometabolic diseases associated with polycystic ovary syndrome Abstract: Polycystic ovary syndrome (PCOS) is a common condition characterized by reproductive and hyperandrogenic features and is often associated with obesity and metabolic dysfunction. Overall, women with PCOS have a substantially greater prevalence of metabolic syndrome than women from the general population. Furthermore, PCOS per se (independent of its frequent association with obesity) often confers cardiometabolic risk (including insulin resistance), and its concurrence with obesity often represents a metabolic “double-whammy” from the adverse effects of PCOS and obesity. The introduction of the Rotterdam diagnostic criteria for PCOS in 2003 has broadened the scope of this condition. The Rotterdam diagnostic criteria have also introduced two new phenotypic subgroups (including normoandrogenemic women with PCOS) that have provided novel insights into a potential role for hyperandrogenism in the development of adverse cardiometabolic risk in women with PCOS. Based on evidence from cross-sectional and interventional studies, hyperandrogenism, obesity, and cardiometabolic risk in women appear to be linked through complex and multidirectional pathways. Furthermore, data from obese women without a formal diagnosis of PCOS also suggest that these interrelationships often exist in female obesity per se (in milder forms than occurs in PCOS). Data from female-to-male transsexuals are particularly informative because these show direct effects of hyperandrogenism (induced through exogenous use of androgenic therapies) on fat distribution and cardiometabolic risk in women. A challenge for the future will be to disentangle and improve our understanding of this complex pathogenic web, thereby facilitating novel and targeted therapies for the hyperandrogenic and adverse cardiometabolic manifestations of PCOS. Keywords: cardiometabolic risk; hyperandrogenism; obesity; polycystic ovary syndrome.
*Corresponding author: Thomas M. Barber, DPhil, Clinical Sciences Research Laboratories, Division of Metabolic and Vascular Health, University of Warwick, University Hospitals Coventry and Warwickshire, Clifford Bridge Road, Coventry CV2 2DX, UK, Phone: +44-2476-968591, Fax: +44-2476-968653, E-mail: [email protected] Petra Vojtechova: Clinical Sciences Research Laboratories, Division of Metabolic and Vascular Health, Warwick Medical School, University of Warwick, University Hospitals Coventry and Warwickshire, Coventry, UK Stephen Franks: Institute of Reproductive and Developmental Biology, Imperial College, London, UK
Introduction Hyperandrogenism in women is common and is defined on the basis of clinical and/or biochemical criteria [1]. Clinical features of female hyperandrogenism include hirsutism, acne, and alopecia. Biochemical features (termed hyperandrogenemia) include elevated serum levels of androgens (increased total testosterone, androstenedione, dehydroepiandrosterone [DHEAS], and free androgen index, which is influenced by the suppression of serum sex hormone binding globulin [SHBG] level) [1]. Hyperandrogenism is one of the diagnostic features of polycystic ovary syndrome (PCOS) and is manifested in many other conditions [2]. There is also a clear association between female obesity and hyperandrogenic features, both in the context of PCOS and in women without PCOS [1]. There is evidence that female hyperandrogenism is also associated with cardiometabolic risk factors and that exposure to hyperandrogenemia (either prenatally or as an adult) may influence body fat distribution (BFD), insulin resistance, and other cardiometabolic features [3, 4]. It is likely that the mechanisms linking female hyperandrogenism with obesity and cardiometabolic dysfunction are complex and bidirectional. However, most of the studies reported in this field are based on cross-sectional data, and there is a lack of longer-term prospective studies that provide clear evidence for direct effects of serum androgens on the adipocyte and vice versa. Therefore, our current understanding of pathogenic mechanisms in this field is incomplete.
92 Barber et al.: Androgens, obesity, cardiometabolic risk in PCOS In this brief review article, we discuss the associations among PCOS, metabolic dysfunction, and cardiometabolic diseases. We then explore data that support an association between female hyperandrogenism and cardiometabolic dysfunction and data that support a role for obesity in the development of female hyperandrogenism. We discuss the possible mechanisms whereby female hyperandrogenemia may influence BFD and cardiometabolic risk and explore data relating to management of both obesity and hyperandrogenism in women and their respective effects on androgen levels and cardiometabolic risk. Finally, we provide an outlook for the future in this field and consider potential clinical implications of hyperandrogenism in the management of obese women with PCOS.
PCOS, metabolic dysfunction, and cardiometabolic diseases PCOS is a common condition that affects between 7% and 10% of women of reproductive age [1]. Reproductive features of PCOS include oligo-amenorrhea, manifesting clinically with irregular and/or absent menses, and impaired fertility. The most common biochemical manifestation of PCOS is hyperandrogenemia (including increased serum total testosterone, androstenedione, and free androgen index), which manifests with hirsutism, acne, and alopecia [1]. It is also clear that metabolic dysfunction is much more common in women with PCOS compared with ageand body mass index (BMI)-matched women within the general population [5–10]. The metabolic syndrome is a constellation of interrelated features [11], the clinical importance of which lies with its association with a twofold increased risk of cardiovascular disease and a fivefold increased risk of developing type 2 diabetes mellitus (T2D) [12]. Using the National Cholesterol Education Program Adult Treatment Panel III criteria in a population-based study on White women from the USA, it has been estimated that metabolic syndrome affects between 34% and 46% of women with PCOS [7–10]. One of the largest studies in this field was published by Ehrmann et al. [7], who reported on the prevalence of metabolic syndrome in 394 (mostly White) women with PCOS: in those women with a BMI >27 kg/m2, 40% had metabolic syndrome. These figures compare with a prevalence of metabolic syndrome [using National Health and Nutrition Examination Survey (NHANES) III data] within the general female US population of 6% (ages 20–29 years) and 15% (ages 30–39 years). It is a reasonable assumption that the high prevalence of metabolic syndrome in women
with PCOS would translate into an elevated cardiometabolic risk in postmenopause, although this hypothesis has yet to be definitively proven through a long-term prospective cohort study in PCOS. There is a real need for such a study to be performed, particularly given that retrospective studies on PCOS in postmenopausal women are fraught with difficulties [13]. However, it is clear from a large cross-sectional study that PCOS is associated with a significantly increased risk for the development of T2D [14], and the cardiometabolic risks associated with T2D are incontrovertible. The data outlined here provide a rationale for active screening for an early and aggressive management of cardiometabolic risk factors in women with PCOS. Although metabolic dysfunction forms an important component of the clinical and biochemical manifestations of PCOS in many women with the condition, it is interesting that none of the current diagnostic criteria for PCOS incorporate any metabolic features but rather focus on the reproductive, hyperandrogenic, and ovarian morphological manifestations of the condition. There are two main sets of diagnostic criteria currently used for PCOS. These are the National Institutes of Health (NIH) criteria, proscribed in 1990 and which require the presence of both hyperandrogenism/hyperandrogenemia and menstrual disturbance (regardless of ovarian morphology) [15]. The NIH criteria were superseded by the Rotterdam criteria in 2003, which require the presence of at least two out of the following three criteria: polycystic ovarian (PCO) morphology on ultrasound scan, hyperandrogenic features (biochemical and/or clinical), and oligo-amenorrhea (intermenstrual interval >42 days) [2, 16]. PCOS has a complex pathogenesis. Based on a large Dutch twin-family study and family-based association studies, it is clear that PCOS is a genetic condition with a high heritability (0.71) [17, 18]. Recent data from genomewide association studies (GWAS) in PCOS are shedding important new insights into its complex etiology [13]. In the first of these, Chen et al. [19] demonstrated associations between PCOS and three loci located at 2p16.3, 2p21, and 9q33.3. Data from GWAS will enable a more targeted approach toward gaining a clearer understanding of the pathogenesis of PCOS. It is also clear that obesity is closely associated with PCOS, with between 38% and 88% of women with PCOS being either overweight or obese [1]. Our current understanding of PCOS pathogenesis is that this condition becomes manifest in genetically predisposed women who subsequently gain weight (through dietary means principally) [13]. However, this is clearly an oversimplification, given that PCOS can occur in a minority of lean women. Furthermore, BMI is itself heritable,
Barber et al.: Androgens, obesity, cardiometabolic risk in PCOS 93
raising the possibility (indeed likelihood based on data on variants within the fat mass and obesity-associated gene [FTO] and PCOS [20]) that at least some of the genetic predisposition toward PCOS is through variants that influence fat mass. To summarize this section, PCOS is common, and although it is defined on the basis of reproductive and hyperandrogenic features, it is also associated with substantial metabolic dysfunction that is at least in part often related to its concurrence with obesity (although PCOS per se is associated with insulin resistance and other metabolic abnormalities) [1]. The remainder of this review will focus on the interplay between hyperandrogenism and female obesity and the cardiometabolic features associated with PCOS. It has been demonstrated that, as expected, the association of PCOS with metabolic syndrome is particularly strong among the overweight and obese subgroup (with a BMI >27 kg/m2) [7, 8]. Perhaps more surprisingly though, it has been shown by Apridonidze et al. [10] that a major predictor of metabolic syndrome occurrence in women with PCOS is elevated serum free testosterone and reduced serum SHBG levels. It has also been shown that a significant inverse relationship exists between SHBG level and metabolic syndrome occurrence in women with PCOS [21]. These data suggest that hyperandrogenemia may play an important role in the development of cardiometabolic risk in women with PCOS, and this forms the focus for the next section.
Phenotypic subgroups of PCOS and the importance of hyperandrogenism as a risk factor for cardiometabolic disease One of the consequences of the Rotterdam diagnostic criteria for PCOS [2, 16] is that it has led to the generation of four distinct phenotypic subgroups. In addition to the subgroup with all three diagnostic features present [“PHO”: PCO morphology (P), hyperandrogenic features (H), and oligo-amenorrhea (O)], three other phenotypic subgroups also exist, each with only two of these features present: “PH”: women with PCO morphology and hyperandrogenic features; “PO”: women with PCO morphology and oligo-amenorrhea; and “HO”: women with hyperandrogenic features and oligo-amenorrhea. Importantly, the PO subgroup has, by definition, normal androgen levels, and along with the “PH” subgroup, would not have been recognized as PCOS on the basis of the prior 1990 NIH
diagnostic criteria [15]. The study of metabolic dysfunction within each subgroup (with particular focus on the normoandrogenemic PO subgroup) will therefore provide useful insight into a potential role for hyperandrogenemia in the development of cardiometabolic risk in women with PCOS. In a large retrospective data set analysis based on 309 Europid women with PCOS, of whom 42 were in the “PO” (normoandrogenemic) phenotypic subgroup and compared with 76 Europid control women without PCOS, our group demonstrated that following adjustments for BMI and age, the women in the “PO” subgroup were indistinguishable from control women with respect to insulin sensitivity and were significantly less insulin-resistant than women in the “PHO” subgroup [22]. Furthermore, the prevalence of metabolic syndrome in women within the “PO” subgroup (7.1%) was similar to that in controls (3.9%) and significantly lower than that in women from the “PHO” subgroup (29.3%, p
