Polycystic ovary syndrome (PCOS) trilogy: a translational and clinical review


Stephen L. Atkin, Hull York Medical School, Hull, UK. Tel.: +44 (0) 1482 675365; Fax: +44 (0) 1482 675370; E-mail: s.l.atkin@hull.ac.uk


Polycystic ovary syndrome (PCOS) is a common disorder associated with chronic anovulatory infertility and hyperandrogenaemia with the clinical manifestation of oligomenorrhoea, hirsutism and acne. The underlying aetiology of PCOS remains elusive despite being the focus of intense basic and clinical study, while clinically these patients often present to gynaecologists, dermatologists and endocrinologists. The Society of Endocrinology recognized that a forum to forge collaborations between the basic scientists and clinicians, as well as to cross clinical disciplines, was needed and the Special Interest Group (SIG) in PCOS was formed in 2003. The mission statement of the PCOS SIG is ‘to promote the multi disciplinary collaboration between basic scientists and clinicians from all branches of medicine interested in PCOS and the metabolic syndrome’. It is hoped that this will encourage multidisciplinary networking as well as to provide an educational forum for basic and clinical trainees. In this way, the transition from fundamental scientific understanding into clinical application and practise will be facilitated.

This paper results from the PCOS SIG held at the 2007 British Endocrine Society meeting in Birmingham, where international experts addressed three important aspects of PCOS: the basic science of follicular function in PCOS, the hotly debated cardiovascular risk in PCOS and the often neglected dermatological problems associated with PCOS. This resulted in a unique contribution for the benefit of both basic and clinical scientists as well as the practising clinician. The discussions across these three different aspects of the syndrome encouraged networking between basic and clinical science, facilitating how research can be translated into clinical effect, and are summarized below.

Follicular function in polycystic ovaries

Helen Mason, Suman Rice, Asjid Qureshi, Laura Pellatt


PCOS is one of the most prevalent endocrine disorders in women, and is the focus of substantial research effort. In the USA the medical cost of its diagnosis and treatment during 2005 was around $4 billion.1 There are no equivalent data for the UK, but the prevalence of PCOS in the two countries is similar and with increasing rates of obesity the number of women diagnosed with this condition is likely to rise, with an escalation of the impact on NHS resources over the next decade. PCOS has a serious negative impact on health-related quality of life, affecting both physical and psychological well-being.2 Fortunately our understanding of PCOS is improving and there have been a number of recent important discoveries regarding the aetiology and function of the ovaries in this condition. Progress has also been made in defining the changes in the growth pattern and function of the follicles in polycystic ovaries and it is these developments that will be outlined here.

There are several distinct aspects to the follicular abnormalities associated with PCOS. The defining abnormality and the morphological feature of the condition is the increase in the number of follicles compared with normal ovaries.3 The follicles in these ovaries then over-produce androgens, leading to hyperandrogenaemia. The final aspect is the anovulation that occurs in a subgroup of women with PCOS. These abnormalities will be considered separately, although it will be seen that there may be common causative factors.

As there is still some debate about the definition of PCOS, and since our results include those women with polycystic ovaries who have ovulatory cycles, it is important to define the terminology used. Throughout this paper the definition of PCOS will essentially be that of the revised Rotterdam criteria4 except that for the in vitro studies the morphology of all of the ovaries was known and classified as normal or polycystic ovary. These studies were performed on ovaries which were microscopically dissected and morphological distinctions are more easily made during this process. Ovaries were classified into three groups according to menstrual cycle history and macroscopic morphological features at the time of dissection.5 A polycystic ovary had at least three of the following criteria; increased volume (> 9 ml), 10 or more follicles of 2–18 mm in diameter, an increase in the amount and density of stroma and thickening of the tunica. Patients with a history of anovulatory infertility and/or oligomenorrhoea or amenorrhoea and no evidence of recent corpora lutea were designated ‘anovulatory PCO’ (anovPCO) and those ovaries from women reporting regular cycles which met the above morphological criteria, but in which a dominant follicle and/or a recent corpus luteum was observed were designated ‘ovulatory PCO’(ovPCO). Normal morphology was assigned when the ovary was of normal size with soft, pliable stroma and contained not more than five follicles greater than 2 mm in diameter in a woman with regular menstrual cycles.5

The steroidogenic defect

An increase in circulating androgens of ovarian origin is one of the main features of PCOS and is present in both ovulatory and anovulatory women. Hyperandrogenaemia is responsible for some of the most distressing symptoms and the commonest presentations.2,6 Studies investigated whether increased circulating androgens were due to increased follicle number or an intrinsic defect of these follicles7 and demonstrated that androstenedione accumulation was approximately 20-fold higher per cell in the theca from polycystic ovary patients than in normal patients. Interestingly, production of 17α-hydroxyprogesterone (17OHP) and progesterone were also increased, suggesting increased activity of both P450 17α hydroxylase, 17, 20 lyase and either 3βHSD or side chain cleavage. Importantly, androgen production was increased whether the theca originated from ovulatory or anovulatory polycystic ovaries.8 This finding is important because it indicated that raised androgens could not be the primary cause of anovulation. These studies have been confirmed and greatly extended by McAllister and colleagues using long-term cultures of theca cells.9,10

The finding that increased steroid production is an intrinsic defect in PCOS led to the hypothesis that this might be the site of the genetic defect.11 Large-scale family studies however, have shown that this is not the primary defect. Surprisingly, the main link has been made to a region on chromosome 19 (19p13.2), a region which lies within an intron of the gene encoding for the protein fibrillin 3.12 The possible role of fibrillin 3, a glycoprotein component of microfibrils, in the aetiology of PCOS remains obscure, however, members of the fibrillin family are known to interact with components of the transforming growth factor beta family and many of these have well-described roles in folliculogenesis. The future of this area of research will be followed with great interest.

The increase in follicle number

Until recently there were very few data from these ovaries regarding the precise nature of the increase in follicle number. Most of the data pertained to antral follicle counting which could be performed by pelvic ultrasound scanning, usually in one plane. The only study to address the numbers of preantral follicles was performed by Hughesden.13 He recorded similar numbers of primordial follicles in ovaries of both morphologies, but approximately twice as many of all of the growing and atretic stages of follicles in polycystic ovaries. Hughesden concluded that over-recruitment of follicles into the growing phases, with an increase in follicle turnover, was the likely cause of the appearance of polycystic ovaries. He did not address the fact that this would be likely to lead to premature menopause, a condition which is not associated with PCOS.

More recently further studies have added greatly to this area that staged and counted follicles in 5 µm sections of fixed small cortical biopsies taken from women during laparoscopy.14 Webber et al. found a more dramatic sixfold increase in the median density of preantral follicles in women with anovulatory polycystic ovaries.14 In particular, there was a decrease in the percentage of primordial follicles which was reflected in a specific increase in growing follicles of the primary stage. The necessary small size of these biopsies largely precluded the presence of secondary follicles so the fate of these primary follicles could not be determined. The authors concluded that the rise in small growing follicles could result from increased population of the ovary by germ cells during foetal life or from reduced rates of atresia of follicles prepubertally.

Erickson and colleagues counted follicles in random sections of fixed ovaries obtained from women undergoing bilateral salpingoophorectomy.15 In contrast to the previous report, this group found no change in the number of primordial follicles. Interestingly, they also found a specific increase in primary follicles in the order of fivefold. Further analysis showed that there was a ‘stockpiling’ of primary follicles in PCOS whereas non-growing primary follicles were quite rare in normal ovaries. The numbers of secondary and Graffian follicles were both twice as frequent in polycystic ovaries. These authors speculated on a number of possible causes for the mechanism; abnormal growth differentiation factor-9, which is discussed in more detail below; increased LH secretion which has been shown to slow primary follicle growth in mice,16 or as suggested by the other groups, increased androgens.

The effects of raised androgens on ovarian function have been studied for some time and cause a specific increase in the proportion of primary follicles.17,18 Data showing that prenatal androgenization of animals results in establishment of many features of PCOS add further to this argument.19–21 In order to be able to hypothesize an androgen effect on small follicles however, it is necessary to demonstrate that androgen receptors (AR) are present on granulosa cells in follicles of this stage. Although a number of studies have reported the presence of ARs in preantral follicles22–24 the precise stage at which receptors appear was unclear. We therefore undertook a study to determine the AR presence in individual preantral follicles isolated from human ovarian cortex. We did not detect AR in any primordial follicles, however, once follicles reached the primary stage, AR mRNA began to be expressed and the number of positive follicles increased exponentially at each progressive growth stage (Fig. 1).24 Interestingly, this expression preceded that of FSHR with only a small percentage of primary follicles expressing the latter. These results suggest that androgens do influence early follicle growth in polycystic ovaries.

Figure 1.

Percentage of preantral follicles positive for androgen receptor (open bars) or FSH receptor (black bar) in a series of 57 follicles. Results are arranged by follicle stage. Note linear increase in androgen receptor positive follicles with growth and comparative lack of FSH receptors.

The definitive study to determine androgen effects on human preantral follicle growth requires observation of its effects under laboratory conditions. The reason why this has not been performed is that culture of ovarian follicles and cortex is notoriously difficult, the main problems are, widespread activation of follicle growth, wholesale loss of primordial follicles and high rates of atresia, even in the best hands.25–27 It is likely that any superimposed effects of androgens on this process would be difficult to determine. In order to address this we grew small pieces of lamb ovarian cortex on the chorioallantoic membrane (CAM) of fertilized chicken eggs using a modification of the method described by Cushman et al.28,29 In this system testosterone selectively increased primary follicles reproducing the results of follicle counts in polycystic ovaries. This adds further weight to the proposal that androgens cause the selective slowing of folliculogenesis seen in PCOS at the primary stage. It remains to be determined how this change in primary follicles translates into the increase in antral follicles which is the defining feature of the syndrome.

There are a number of other factors which have been shown to have a role in initiation of follicle growth. Two of the most important are members of the bone morphogenetic protein (BMP) family. Growth differentiation factor-9 (GDF-9)30 and BMP-1531 have been shown to be expressed in developing oocytes and animals in which the genes were knocked out showed arrested and abnormal follicle growth32,33 making these factors candidates for the cause of altered follicle number in PCOS. Compared with normal ovaries, oocyte GDF-9 expression occurs later and is then maintained at a lower level in developing follicles in polycystic ovaries.34 BMP-15 was unchanged. However, alterations in GDF-9 were found in both ovulatory and anovulatory polycystic ovaries and are therefore not implicated in the mechanism of anovulation.

The cause of anovulation

The increase in follicle number gives the classic appearance of the polycystic ovary. However many women having polycystic ovaries continue to ovulate6,35–37 and the high prevalence of this inherited condition indicates that it is not an absolute cause of infertility.38 Some additional factor or factors therefore must determine the ovulatory status of these women. As the circulating levels of FSH are within the normal range,39 the follicles are healthy and when removed from the ovary the granulosa cells respond well to FSH-stimulation, research centred on finding a locally produced inhibitor of follicle growth. Many studies have found disturbances in these important ‘fine tuners’ and amplifiers of gonadotrophin action in follicles from anovulatory women with PCOS.

Because of their ubiquitous role in folliculogenesis it was initially thought that the most likely candidate would be a member of the TGF-β family and a number of studies were performed to investigate production of inhibin, activin and follistatin. In the follicular fluid from women with normal ovaries activin A was consistently low and did not change with follicle health whereas inhibin A and B were low in small follicles and progressively increased with follicle size.40 The fact that inhibin B was consistently higher than inhibin A led to the hypothesis that this was the more important paracrine signalling factor. Comparison with levels in the follicular fluid from polycystic ovaries found that as for normal ovaries inhibin B was the predominant form but that levels were not different. This study reported lower levels of inhibin A, but the excess of inhibin B would likely mask any effect.41 The findings were then contradicted by studies demonstrating that both inhibin A and B subunit mRNA expression and protein concentrations in follicular fluid aspirated from polycystic ovaries were lower than those in size-matched follicles from normal ovaries.42–44 These data helped to explain the earlier finding that serum levels of inhibins were similar despite the increased number of follicles.44 It is not clear why the results of these studies differed and therefore the precise nature of the change in inhibin production in PCOS remains obscure, but the lack of a profound difference in intrafollicular levels suggests that they are not leading contenders for causing anovulation.

Follistatin, the activin binding protein also came under the spotlight when, in a study of candidate genes for PCOS, the strongest linkage was found to be in the region of its gene.45 Any hormone that alters activin action is of interest, in PCOS as activin not only modulates androgen production by theca and is involved in pituitary FSH secretion,46 but is also implicated in pancreatic function.47 Further studies to detect allelic variation in follistatin concluded that evidence of linkage was weak and that alterations in the follistatin gene at least were not likely to contribute to PCOS.48 Further evidence against a link comes from the measurement of follistatin levels in follicles from normal and polycystic ovaries which were found to be similar.49

Elements of the IGF system in the ovary and their binding proteins were also hypothesized to be defective in PCOS. The total amount of IGF activity was not diminished in these follicles50 and so attention turned to the numerous binding proteins which control IGF function. The main finding of these studies was that IGFBP-2 and IGFBP-4 were elevated in PCOS follicles.51,52 It is possible that this additional binding capacity would remove free active IGF from the follicle at the time that it was required for synergizing with FSH to stimulate follicle selection. Removal of this boost to aromatase might contribute to arrested follicle growth in polycystic ovary.

Anti-Müllerian hormone (AMH) is another member of the transforming growth factor β (TGF β) superfamily. In the ovary it is expressed in the granulosa cells postnatally until menopause53 and its levels are highest in small growing follicles of the preantral and antral stages and are reduced in luteinized and atretic follicles.54 Interest in AMH has been rekindled by a series of interesting publications suggesting that it could be used as a marker of ‘ovarian reserve’.55–57 Interestingly, AMH also reduces FSH-stimulated follicle growth in isolated follicle cultures.58 These data suggest an inhibitory effect of AMH on follicle growth and also that it reduces sensitivity to FSH. Our studies revealed that AMH production ceased from follicles continuing beyond selection (Fig. 2),59 reinforcing the suggestion that it has inhibitory effects on FSH-stimulated follicle growth.60 We now have preliminary data to show that AMH inhibits FSH-stimulated aromatase expression.61 Given these results, we were very interested to see that serum levels of AMH in women with PCOS were increased by twofold to threefold.62,63 Although generally considered to be due to the increase in small follicles producing AMH in these ovaries, we wished to determine whether AMH might be increased per cell and so we performed a comparative study between granulosa cells from normal and polycystic ovaries. We measured AMH in medium conditioned by equal numbers of granulosa cells from small follicles from normal, ovulatory or anovulatory polycystic ovaries. As AMH changes with follicle size, the follicles contributing to each group were size-matched. We found that whereas levels of AMH in medium from cells from ovulatory polycystic ovaries were on average four times higher than normal, those in anovulatory PCOS were 75 times higher (Fig. 3).59 These levels were well above the range seen in small follicles which are considered to be inhibitory and are of the range which we have shown to reduce FSH-stimulated aromatase expression. We suggest therefore that AMH may indeed play a significant role in causing the inhibition of follicle growth seen in anovulatory PCOS. We are currently investigating the mechanisms of the increase and of the inhibition of aromatase.

Figure 2.

AMH concentration in GC-conditioned medium from normal ovaries (N = 8), ovulatory (N = 9) and anovulatory (N = 6) polycystic ovaries. AMH production was significantly different between normal, ovPCO and anovPCO (P ≤ 0·001). The mean concentration in cells from anovPCO was 75 times higher than the mean for normal ovaries (Reproduced from Pellatt 2007, with permission from The Endocrine Society).

Figure 3.

(a) Effects of metformin at 10−5 and 10−7 m on E2 production by granulosa cells from follicles from naturally cycling ovaries. Metformin (m) was inhibitory from 10−5 m onwards and this dose was also inhibitory of LH and insulin-(I) stimulated E2 production. (b) Dose dependent inhibition of theca cell production of androstenedione by culture with metformin. Insulin significantly stimulated a'dione production which was again inhibited by metformin.

Finally, steroids themselves have been investigated for a role in inhibiting follicle function in PCOS. Levels of 5areduced androgens were significantly higher in follicles from PCOS and 5a-reductase activity was approximately four times higher.64 These data are interesting because 5a-reduced androgens limit the activity of aromatase by acting as a competitive inhibitor.65 It can be envisaged that this would prevent a follicle from gaining the necessary responsiveness to FSH to become selected. In this and all of the above studies it remains to be determined whether the disturbance of the paracrine factors is the cause of the arrested follicle growth or is secondary to it. That so many factors involved in paracrine control of follicle growth and selection appear to be deranged suggests that the normal balance of these factors is lost once the follicle stops growing. It is likely that this ‘chicken and egg’ question will only be answered once the genetics of the syndrome have been fully unraveled.

Effects of insulin on follicle growth

PCOS is associated with a metabolic defect and one of the main features of this is insulin resistance and compensatory hyperinsulinaemia. The glucose uptake pathway is resistant in muscle and adipose66 and has also been demonstrated to be abnormal in granulosa cells.67 Conversely, the ability of insulin to act as a gonadotrophin in ovarian tissue appears to be unaffected and it subsequently further enhances the high rate of androgen production68 and has also been implicated in causing the premature arrest of follicle growth.69,70 Improvement of insulin sensitivity either by diet or exercise consequently remains a primary goal of treatment.71 In cases where this fails, administration of insulin sensitizers such as metformin has become the norm.72,73 Results with metformin have however, been mixed and the target population who will most benefit has not been determined.

Androgen levels often fall in women with PCOS who are taking metformin, even when there has been little effect on circulating insulin. To investigate the possibility that metformin was having an effect on the ovary independently of its effects on insulin levels, we cultured granulosa and theca cells in the presence of a wide range of doses of metformin and assessed the effects on steroid production. Even at relatively low doses, metformin exerted an inhibitory effect on granulosa E2 production and on theca production of androstenedione (Fig. 4).74 Similar results were found in an ovarian theca-like cell line75 and subsequently a number of other studies have demonstrated direct effects of metformin on these cells.76–78 It is easy to envisage that inhibition of androgen production would be beneficial in PCOS, however, the fact that this drug appears to inhibit aromatase is more problematic.

Figure 4.

Left-hand panel: AMH concentration in follicular fluid from a range of individual follicles isolated from women with regular ovulatory cycles and normal ovaries. Note the loss of AMH in fluid from follicles greater than 10 mm. Right hand panel: AMH concentration (ng/ml/50 000 cells) in GC-conditioned medium from a range of follicle sizes from normal ovaries. AMH fell exponentially as the follicle size increased and again GC from follicles greater than 10 mm produced low or undetectable levels of AMH.

In conclusion, as one of the most common conditions in women of reproductive age, polycystic ovary syndrome is rightly the focus of intensive research. It is hoped that further family studies will reveal the causative gene/s and their precise roles and that this may lead to more evidence-based treatment. It is hoped that one of these genes can be linked to the cause of the ovarian changes which actually lead to the development of the polycystic morphology and that this will ultimately bring us closer to a cure.

Cardiovascular disease (CVD) and PCOS

Francesco Orio and Annamaria Colao


PCOS is a complex endocrine condition with important health implications.1,79 Several recent reports80–84 have suggested that women with PCOS have an increased cardiovascular risk (CVR) and an increased prevalence of cardiovascular disease (CVD) characterized by an impairment of cardiac structure and function, endothelial dysfunction, lipid abnormalities, chronic low-grade inflammation and cardiopulmonary impairment.81–93 All of these features are probably linked to insulin resistance that in turn may amplify androgen excess, with the combination leading to an indirect effect on CVR through an abnormal lipid profile and an increased incidence of CVR factors.86 Additionally, PCOS is associated with long-term health risks, including diabetes mellitus94 and CVD.79–81

Cardiovascular abnormalities are important long-term sequelae of PCOS that need critical and careful investigation. Recently, several early markers of subclinical CVD have been investigated81,83,93 and important CVR factors have been evaluated.90–92

This short review will focus on dyslipidaemia, heart structure, arterial morphology, endothelial function and the cardiopulmonary assessment of women with PCOS.


Dyslipidaemia is the most common metabolic abnormality found in women with PCOS,95 and the atherogenic lipoprotein profile is characterized by elevated triglycerides, small low density lipoproteins (LDL) and reduced high density lipoprotein (HDL). The atherogenic LDL lipoprotein phenotype reported is an increase of LDL size and the type III or type IV LDL subclasses (LDL III, density range = 1·033–1·038 g/ml; LDL IV and LDL V in the density range = 1·038–1·050 g/ml) are a common finding in PCOS, being the second most common lipid alteration found after a decrease in HDL-cholesterol.88 Therefore, the increase of small LDLs in the ‘atherogenic lipoprotein phenotype’ may contribute to an increase in CVR, with additional qualitative changes in lipoprotein metabolism96 that may increase CVD risk in PCOS.

Elevated triglycerides, LDLs, and reduced HDL may be closely related to IR.97 Since insulin is a major positive regulator of lipoprotein lipase involved in the pathway of HDL-C production, the dyslipidaemia seen is likely to be secondary to IR, although hyperandrogenism has also been reported to affect lipoprotein levels and lipids independently of insulin levels and body weight.86

Heart morphology

Left ventricular hypertrophy (LVH) is an important predictor of CV mortality and morbidity.98 Women with PCOS, even if young and non-obese, have a significant increase in cardiac size when compared with the body mass index (BMI) of age-matched healthy women.81 In addition, PCOS patients have a significantly lower left ventricular ejection fraction (LVEF) as a measure of systolic function, and reduced early atrial mitral flow velocity as a measure of diastolic function, although all patients have normal LVEF overall.81 PCOS patients also show significantly higher levels of diastolic blood pressure (DBP) and mean blood pressure (MBP) than healthy women,81 suggesting that the increased left ventricular mass observed in women with PCOS could be caused by a chronic increase in blood pressure even in the absence of clinically defined hypertension.99 These features are also seen in young, normal weight PCOS patients, suggesting that the pathogenesis of any cardiac abnormalities in PCOS is not just BMI dependent, indeed the left ventricular mass index (LVMi) significantly correlates to both BMI and IR in women with PCOS. Hyperinsulinaemia secondary to IR is a predictor of coronary artery disease, and IR has been proposed as the key factor linking hypertension, glucose intolerance, obesity, lipid abnormalities and coronary heart disease in ‘metabolic syndrome X’.100

In a case-control study, women with PCOS had an increased isovolumetric relaxation time (IVRT), an index of early left ventricular diastolic dysfunction, and lower ejection fraction compared with weight matched healthy women,101 giving further evidence of structural cardiac changes in PCOS. There is a significant, direct relationship between plasma insulin levels and IVRT in PCOS implicating IR in the underlying aetiology. However, the existing studies suggest that IR may contribute to myocardial dysfunction in PCOS, the exact role of IR and/or hyperandrogenism on the heart structure of women with PCOS and on the increased LVM in young women with PCOS remains unclear.

Major arteries: morphology and endothelial function

Previous studies have reported an increased pulse wave velocity of the brachial artery, a measure of arterial stiffness related to increased CVR, and there is increased stiffness of both internal and external carotid arteries in women with both PCOS and polycystic ovaries (ultrasonographic polycystic ovaries alone) compared with healthy women,102 though these findings were not seen in the aorta in PCOS.85 Studies have demonstrated a significant difference in intima media thickness (IMT) between PCOS patients and healthy women,83,103 suggesting an association between PCOS and early carotid atherosclerosis; IMT is increased even in young women with PCOS.83 These young normal-weight women with PCOS were not dyslipidaemic or hypertensive, suggesting that other features were responsible for these finding such as androgen excess or IR.

One of the early signs in the development of cardiovascular lesions is endothelial injury104 that represents an early sign of atherosclerosis. There are reports of precocious anatomical and functional arterial changes in women with PCOS,83–85,102,105 and that increased IR could play a key role in the development of endothelial damage.83,85,102,104,106 Conversely, IR and BP have been shown to interact negatively with arterial structural and functional measures in overweight women with PCOS.107 Interestingly, dehydroepiandrosterone-sulphate (DHEAS) correlates inversely with arterial structure, suggesting a cardioprotective effect of higher levels of endogenous DHEAS in women with PCOS. A direct correlation between IMT and the risk of myocardial infarction and ‘stroke’ in a patient without a history of vascular disease has been shown,108 and several metabolic alterations, such as obesity, IR, hyperandrogenism,109–111 are widely accepted risk factors for increased CVD.

Cardiopulmonary assessment

It is well known that reduced functional capacity is associated with an increased risk of cardiovascular mortality.112 The maximal oxygen consumption (VO2max), is closely and directly related to insulin sensitivity113 and a strong determinant of the insulin sensitivity index.114

Testosterone levels also positively correlate with VO2max115 although little is known about how testosterone might have an effect itself or influence insulin action. However, it is well recognized that exercise improves glucose homeostasis related to an up-regulation of the expression and/or activity of proteins involved in insulin signal transduction in skeletal muscle.116

To date, only two experimental studies have demonstrated a significant reduction in VO2max in young women with PCOS, when compared with healthy women, showing an impaired cardiopulmonary pattern leading to reduced cardiopulmonary functional capacity in these patients.92,93


While surrogate markers of CVR are found in PCOS, there are no long-term prospective data available in the literature for well-characterized women with PCOS, and large-scale clinical trials evaluating end-point morbidity and mortality for CVD in PCOS subjects are lacking. The link of PCOS to primary cardiovascular events, such as stroke or myocardial infarction, still remains to be demonstrated.

Dermatology and PCOS

Ulrike Blume-Peytavi


Hyperandrogenism, excessive production of androgens, may have a profound effect on the women it affects, and 70–80% of patients with androgen excess demonstrate hirsutism.117,118 The most common cause of androgen excess is PCOS; most frequent cutaneous symptoms in these patients are hirsutism, acne, androgenetic alopecia (AGA) seborrhoea, obesity and acanthosis nigricans; less frequently virilization or clitoromegaly can be observed. Oligo- or amenorrhoea, secondary infertility, diabetes mellitus or insulin resistance may be systemic consequences of PCOS. Hirsutism, acne, and AGA will be discussed in more detail below.


Hirsutism is defined as the growth of excess terminal (coarse) hairs in females in a pattern typically seen in adult males,117,119 sometimes (although rare) leading to virilization (deepening of the voice, clitoro-megaly, and loss of female body shape). The Ferriman–Gallwey score.

(F-G score) is a method of evaluating and quantifying hirsutism, and was first introduced in 1961. Hair growth is rated depending on density and intensity from 0 (no growth of terminal hair) to 4 (complete and heavy cover), in nine locations, giving a maximum score of 36. The nine locations measured are the upper lip, chin, chest, upper back, lower back, upper abdomen, lower abdomen, the upper arms and the thighs,120 A score of at least 6 is needed for a diagnosis of hirsutism,121 but care should be taken to consider the amount of hair growth that is normal for the individual's ethnic background, and the F-G score is not applicable in Asian patients and may be difficult to apply in routine clinical practise. Recent publications suggest high national variations in F-G scoring for hirsutism; from a score of 3 in an American population122 to 8 in a Finish population and above 6 in German populations, indicating that there are differences depending on ethnicity.

PCOS is the most common cause of hirsutism, which in turn is the main dermatological symptom of the condition.123,124 The clinician should always keep in mind that not every increase in hair growth is hirsutism. Attention must be paid to the differential diagnosis of hypertrichosis that is independent of androgen influence and characterized by the superfluous and uniform growth of non-terminal (vellus) hair over the body, particularly in non-sexual areas and without pattern.124 This has to be considered during diagnosis as hypertrichosis may occur congenitally (or as a result of a patient's ethnicity), but it is also frequently caused by systemic drug intake such as glucocorticosteroids, cyclosporin, latanoprost, bimatoprost, interferon a, or minoxidil.

Hirsutism is usually the result of an underlying adrenal, ovarian or central endocrine abnormality. Elevated secretion of androgens, increased bioavailability of testosterone and increased sensitivity of hair follicles to androgens all contribute to the condition. It is characterized by the transformation of vellus into terminal hairs in androgen-sensitive areas125 and can present clinically with excess hair growth on the face, chest, between the breasts, and on the abdomen.103 The main androgen involved in the development of hirsutism is dihydrotestosterone (DHT, which is produced in the skin from testosterone by the enzyme 5α-reductase). Androstenedione and DHEA are also overproduced by the ovary and adrenal glands. Women with hirsutism have increased activity of 5α-reductase in their hair follicles.126,127 Besides PCOS related hirsutism, idiopathic and neoplasia associated hirsutism, and drug induced hirsutism should be excluded during diagnosis. The main drug candidates to be investigated include anabolic agents, gestagens, a gonadotrophin inhibitor (danazol), testosterone, DHEA, tibolone, and valproic acid.

Current management of hirsutism.  A careful diagnostic work-up of the patient is mandatory in this condition. A diagnostic evaluation should be made to eliminate PCOS, or any serious cause of excess hair growth, for example, androgen secreting neoplasia, and consideration should be given to possible referrals to – or from – other specialists (e.g. a dermatologist, endocrinologist, or gynaecologist).

Treatment of hirsutism aims to rectify any causal hormonal imbalance and improve the cosmetic appearance of the disorder. Treatment may be non-pharmacological intervention (weight reduction, physical or chemical procedures, laser epilation), or pharmacological intervention (central or peripheral androgen suppression, androgen blockers, insulin sensitizers, inhibitors of androgen production, or a topical enzyme inhibitor – eflornithine).

Non-pharmacological interventions.  Weight reduction in obese women decreases hirsutism, the production of ovarian androgens, and the conversion of androstenedione to testosterone.128 Physical methods of hair removal are temporary solutions to the problem of facial or body hirsutism, and most women who present with facial hirsutism will have already tried several or all of the methods.

Hair removal options include physical procedures such as shaving, tweezing, waxing/sugaring and chemical approaches such as bleaching with H2O2 or the use of depilatory creams with thioglycolates. All these options are ‘self-therapy’ performed by the women themselves. In contrast, professional therapy options are carried out by a professionally qualified or competent person and include photoepilation, electrolysis, and electro-epilation (the use of electrosurgical methods, such as electrocautery or diathermy, for destroying hair follicles).129

Photoepilation.  The most frequently used and successful technique for permanent epilation is photoepilation. This uses light focused on hair pigment, with accumulation of light energy within the pigmented hair shaft causing destruction of the hair follicle. Patients with skin types I-II (type I often burns, rarely tans. Tends to have freckles, red or fair hair, blue or green eyes. Type II usually burns, sometimes tans. Tends to have light hair, blue or brown eyes), and with dark hair, are the optimum candidates for laser treatment, since its efficacy is dependent on the hair colour and size of the hair follicle enabling the accumulation of the highest energy in the dark and thick hair shaft. There are various methods available, with differing results (Table 1).130

Table 1.  Photoepilation methods and typical results
MethodOutcome (% reduction in hair growth)
Intense pulsed light source (590–1200 nm)80% after 1 year
Normal mode Alexandrite laser (754 nm)40–80% after 6 months
Pulsed diode laser (800 nm)65–75% after 8 months
Ruby laser (694 nm)38–49% after 1 year
Long-pulsed Nd-YAG lasers (1064 nm)20–60% after 3 months, 0% after 6 months

In general, there is a 20–40% hair loss following a single treatment with photoepilation, and additional hair loss with each treatment of 20–40%. Under ideal conditions, two to seven treatments are needed for complete, permanent hair loss in patients with dark hair. However, the outcome depends on the individual hair and skin conditions, the laser device used and the experience of the therapist. There is also a synergistic effect of laser treatment with topical eflornithine, which slows hair re-growth.131

Electrolysis.  Electroysis is commonly used in hirsute PCOS patients and easily available due to the number of qualified practitioners. However, despite how widely available it is there is no single randomized controlled trial to guide its success in this patient group.

Pharmacological interventions.  Systemic and topical pharmacological intervention strategies for the management of excessive hair growth can be distinguished as follows:

Systemic interventions.  The main aim of pharmacological treatment is to block or inhibit androgen production or action. Central or peripheral androgen suppression can be achieved using three groups of drugs:

  • 1Peripheral androgen blockers (e.g. cyproterone acetate, flutamide, or spironolactone) or antiandrogens (e.g. finasteride)
  • 2Insulin-sensitizing agents (e.g. rosiglitazone, metformin)
  • 3Inhibitors of androgen production (e.g. oral contraceptives, GnRH analogues)

Three recent meta-analyses on hirsutism have suggested that each of the above (but not placebo) had a positive effect on hirsutism; however, combination randomized controlled trials are lacking in this area.132–135 It is difficult to determine which treatment is most effective as there have been no systematic comparative trials undertaken, but there does not seem to be one agent clearly superior to the others.

Topical pharmacological intervention.  Topical eflornithine 11·5% cream (Vaniqa®) may be used to inhibit hair growth. Eflornithine inhibits ornithine decarboxylase (ODC), a key enzyme of polyamine synthesis (responsible for cell proliferation, migration and differentiation). Blockage of ODC leads to apoptotic cell death and decreased hair growth.

Two studies have evaluated the efficacy and safety of eflornithine HCl 13·9% cream on excessive facial hair growth,135 Eflornithine cream, applied twice daily for 24 weeks, significantly reduced the growth of unwanted facial hair (measured by a physician's assessment and hair length/mass). Efficacy was not dependent on the method of hair removal. The treatment was generally well tolerated with some patients experiencing mild, transient burning, stinging, or tingling. The time-response profile suggested that the effect had not yet reached a plateau at week 24 and that a more pronounced response may be achieved with longer treatment periods. Recent reports have indicated that the combination of laser and eflornithine is more effective than laser alone for removing unwanted facial hair.131 The possibility that eflornithine may inhibit all hair types, including the thin and non-pigmented hairs which are not well targeted by laser treatment, requires further study.

Combination of laser with eflornithine in a double blind placebo-controlled study showed for the vehicle-laser treatment area, complete or almost complete hair removal achieved in 67·9% of patients compared with 96·4% of patients treated with laser combined with eflornithine.134

Acne vulgaris

Acne vulgaris is an androgen dependent, self-limiting inflammatory disorder of the pilosebaceous unit, which is seen in nearly 80% of all adolescents.136,137 Persistent, severe, or acne of late onset in women is suggestive of PCOS. Although the prevalence of acne in PCOS is unknown the literature reports up to 83% in PCOS vs. 19% in a control group, and most women with severe acne have PCOS.137

Acne is less prevalent in PCOS than hirsutism, and this may be explained by the difference in the expression levels and type of 5α-reductase in the sebaceous gland compared with the hair follicle. Hair follicles have higher DHT levels than the sebaceous glands and express 5a-reductase type I, whereas type II is present in the dermal papillae; DHT has been implicated in the pathogenesis of acne.127,138 However, the pathogenesis of acne is complex.

Treatment of acne.  Oral hormonal therapy (in women whose peripheral hyperandrogenism is caused by hormonal imbalance) is an excellent choice for women who need oral contraception. It should be started early in women with medium to severe grade acne, or with sseborrhoea, acne, hirsutism and alopecia (SAHA) syndrome. It is an important factor in the therapeutic management of women with or without endocrine abnormalities, and especially for women with acne tarda.139 Anti-androgens are used to antagonize the binding of testosterone and other androgens to the AR. Insulin sensitizing agents inhibit hepatic gluconeogenesis (biguanides) or increase glucose uptake in the muscle and fat (thiazolidinediones); 5α-reductase inhibitors reduce the action of the 5α-reductase enzymes type 1 and 2, and GnRH agonists suppress LH and FSH production.

The pathogenesis of acne vulgaris is complex and includes focusing on hyperproliferation and differentiation of keratinocytes using topical retinoids. Combination therapy using retinoids plus benzoyl peroxide or acelaic acid can treat existing acne lesions faster than the individual agents alone and can also prevent the development of new lesions.140 According to the consensus on management of acne,141 the choice of treatment depends on the degree of acne. Systemic antibiotics and systemic retinoids are included in the management for either acne resistant to local therapy or when scarring occurs. The dermatologist should be integrated in the management as early as possible if the patient first presents to the endocrinologist or the gynaecologist.

During pregnancy and breast feeding, topical therapeutic agents can be used, including benzoyl peroxide (2·5–10%), erythromycin (Aknemycin®, Eryaknen®), clindamycin (Basocin Lsg, Gel®), and acelaic acid (Skinoren®).

Androgenetic alopecia (AGA)

AGA is a progressive, non-scarring loss of scalp terminal hair with miniaturization of terminal to vellus hair types in which both ovarian and adrenal androgens have been implicated.141 Normally patients with AGA have normal androgen blood levels, but may also have a genetic predisposition for the disorder, with increased enzyme activation and metabolism in the target end organ with elevated AR density in the target organ.

The current treatment for AGA is to treat the underlying hormone dysregulation (using an oral contraceptive pill with antiandrogenic activity), plus supportive topical therapy using 2% minoxidil (Regaine® or Rogaine®) for women. Anti-androgens competitively inhibit the binding of testosterone and DHT to the AR, and are the treatment of choice for androgen-dependent alopecia. A recent publication by Vexiau et al.142 has clearly indicated that systemic antiandrogens in management of AGA are only indicated if there are elevated androgen levels. In idiopathic hirsutism and AGA the first choice will be the use of 2% topical minoxidil solution.


Hirsutism, acne, and androgenetic alopecia are common symptoms of hyperandrogenism in women. Although these symptoms may also indicate other underlying diseases, the most common cause of hyperandrogenism in women of childbearing age is PCOS.

Women presenting with these symptoms should therefore be investigated for PCOS, and to exclude other potential causes of hyperandrogenism. The impact of the dermatological manifestations of this disorder should not be underestimated, and many women find these symptoms extremely debilitating, since they can affect many aspects of their lives. The treatment plan should address not only the underlying cause of the androgen excess, but also its dermatological manifestations. For patients suffering from PCOS, treatment options include reduction of androgen production and action, lifestyle modification, oral contraceptives, antiandrogens, and insulin-sensitizing agents as well as topical treatment of acne, seborrhoea and androgenetic alopecia. Local management of hirsutism should in addition include photoepilation combined with eflornithine cream.


Although PCOS is the focus of intense research, the importance of the translation and application of this research to the clinical setting cannot be underestimated. The PCOS SIG's annual meeting showcased three different aspects of the disease, with the overall aim of encouraging interdisciplinary collaboration and networking in this important area.

PCOS is one of the most common conditions in women of childbearing age, and is the most common cause of hyperandrogenism in this cohort. Dr Mason detailed the factors implicated in the cause of the ovulatory dysfunction in PCOS. It is clear that there are differences in the production of many of the paracrine signalling hormones between normal and polycystic ovaries. Although subtle changes can be observed in the inhibin/activin family it is difficult to assess the in vivo impact of such changes on follicle growth. Likewise, there are clearly alterations in the IGF/IGFBP system and in the production of various androgenic and other steroids and whereas these are not without interest, it is impossible to determine whether they reflect defective follicle function or cause it. It can however, be seen that the changes in testosterone and androstenedione are considerable and intrinsic and are likely candidates for disrupting folliculogenesis in the early stages. It is the sheer magnitude of the increase in AMH production that differentiates it from other factors and underlines the likelihood that this will prove to be an important factor in the pathogenesis of PCOS.

Professor Colao detailed the CVR factors present in PCOS. However, it was noted that there is a paucity both of long-term data for well-characterized women with PCOS, and for large-scale prospective clinical trials to determine the outcome morbidity and mortality of CVD in PCOS subjects. Furthermore, the link between PCOS and primary cardiovascular events has not yet been demonstrated.

Professor Blume-Peytavi urged that the impact of the dermatological manifestations of this disorder should not be underestimated, describing how many women find these symptoms extremely debilitating, affecting many aspects of their lives. For patients suffering from PCOS, treatment options include reduction of androgen production and action, lifestyle modification, oral contraceptives, antiandrogens, and insulin-sensitizing agents as well as topical treatment of acne, seborrhoea and AGA. Local management of hirsutism should also include photoepilation combined with eflornithine cream.


The authors gratefully acknowledge the assistance of Dr Lisa Chamberlain James in the preparation of this manuscript. The PCOS SIG meeting was sponsored by an unrestricted educational grant from Shire Pharmaceuticals.