Prof R Pasquali, U.O. di Endocrinologia, Dipt Medicina Interna, Azienda Policlinico S. Orsola-Malpighi, Via Massarenti 9, 40138 Bologna, Italy. Email email@example.com
The polycystic ovary syndrome (PCOS) is one of the most common causes of infertility due to anovulation in women. The clinical features of PCOS are heterogeneous and may change throughout the lifespan, starting from adolescence to postmenopausal age. This is largely dependent on the influence of obesity and metabolic alterations, including an insulin-resistant state and the metabolic syndrome, which consistently affect most women with PCOS. Obesity does in fact have profound effects on both the pathophysiology and the clinical manifestation of PCOS, by different mechanisms leading to androgen excess and increased free androgen availability and to alterations of granulosa cell function and follicle development. Notably, simple obesity per se represents a functional hyperandrogenic state. These mechanisms involve early hormonal and metabolic factors during intrauterine life, leptin, insulin and the insulin growth factor system and, potentially, the endocannabinoid system. Compared with normal weight women with PCOS, those with obesity are characterised by a worsened hyperandrogenic and metabolic state, poorer menses and ovulatory performance and, ultimately, poorer pregnancy rates. The importance of obesity in the pathogenesis of PCOS is emphasised by the efficacy of lifestyle intervention and weight loss, not only on metabolic alterations but also on hyperandrogenism, ovulation and fertility. The increasing prevalence of obesity among adolescent and young women with PCOS may partly depend on the increasing worldwide epidemic of obesity, although this hypothesis should be supported by long-term prospective epidemiological trials. This may have great relevance in preventive medicine and offer the opportunity to expand our still limited knowledge of the genetic and environmental background favouring the development of the PCOS.
The polycystic ovary syndrome (PCOS), one of the most common causes of infertility due to anovulation, affects 4–7% of women.1 According to the National Institutes of Health, basic diagnostic criteria should be the presence of hyperandrogenism and chronic oligo-anovulation, with the exclusion of other causes of hyperandrogenism such as adult-onset congenital adrenal hyperplasia, hyperprolactinaemia and androgen-secreting neoplasms.2 A consensus conference held in Rotterdam3 agreed on the appropriateness of including ultrasound morphology of the ovaries as a further potential criteria to define the PCOS but also established that at least two of the following criteria are sufficient for the diagnosis: oligo and/or anovulation, clinical and/or biochemical signs of hyperandrogenism and polycystic ovaries at ultrasound. The pathophysiology of PCOS may have a genetic component although it can be suggested that the main factors responsible for the increasing prevalence of the PCOS are related to the influence of the environment, including dietary habits, behaviour and other still undefined factors.1 The clinical features of PCOS are heterogeneous and may change throughout the lifespan, starting from adolescence to postmenopausal age.4 This is largely dependent on the influence of obesity and metabolic alterations, including an insulin-resistant state and the metabolic syndrome, which consistently affect most women with PCOS.5 This represents an important factor in the evaluation of the PCOS throughout life and implies that the PCOS by itself may not be a hyperandrogenic disorder exclusively restricted and relevant to young and fertile-aged women but may also have some health implications later in life.
Whereas hyperandrogenism and menstrual irregularities represent the major complaints in young women with the PCOS, symptoms related to androgen excess, oligorrhoea or amenorrhoea and, particularly, infertility are the main complaints of adult women with PCOS during the reproductive age. Obesity has an important impact on the severity of these manifestations in proportion to its degree and particularly in the presence of the abdominal phenotype.5 In addition, there is consistent evidence that it renders affected women more susceptible to develop type II diabetes, with some differences in the prevalence rates between countries and, potentially, in favouring the development of cardiovascular diseases.1
This review article is particularly focused on the impact of obesity on infertility in women with PCOS by considering (i) the pathophysiological interaction between obesity and PCOS, (ii) the influence of obesity on both clinical and biochemical phenotype and, finally, (iii) the beneficial effect of weight loss on the hormonal and metabolic abnormalities, on menses, ovulation and fertility rates.
Prevalence of obesity in the PCOS
We are facing a worldwide public health emergency due to the increasing epidemic of obesity and related disorders.6 The price of obesity is represented by a long list of co-morbidities and social, psychological and demographic problems. Obese women are characterised by similar co-morbidities to men, particularly type II diabetes mellitus and cardiovascular diseases.7 On the other hand, they also have some specific problems, including fertility-related disorders and some hormone-dependent forms of cancer.8,9
Obesity is strongly associated with the PCOS. Although the cause of this association remains unknown, but obesity is present in at least 30% of cases, in some series, the percentage may be as high as 75%.1 Azziz et al.10 found that in a population of 400, unselected, consecutive premenopausal women the prevalence of PCOS was 6.6% and that the prevalence of overweight and obesity was 24 and 32%, respectively, with a higher proportion of black women with respect to white women. Table 1 reports the prevalence of underweight, normal weight, overweight and obesity in a large unselected cohort of consecutive women attending our outpatient endocrinology clinic. Intriguingly, differences in the prevalence of overweight or obesity among women with PCOS in different studies may be due to referral bias, although it should be remembered that women with PCOS in the USA generally have a higher body mass index (BMI) than their European counterparts.
Table 1. Prevalence of obesity in a large cohort of 320 women with PCOS attending the Division of Endocrinology, S. Orsola-Malpighi Hospital, Bologna, in the past 10 years
Waist circumference (cm)
WHR, waist-to-hip ratio.
BMI categories are defined according to the World Health Organization document.11
22.7 ± 6.0
75.5 ± 2.8
0.81 ± 0.09
25.2 ± 6.2
70.6 ± 5.2
0.76 ± 0.06
23.9 ± 5.5
83.6 ± 6.6
0.81 ± 0.07
24.9 ± 5.9
95.7 ± 6.9
0.85 ± 0.07
26.9 ± 7.5
102.2 ± 7.6
0.87 ± 0.07
27.2 ± 8.3
113.7 ± 9.1
0.92 ± 0.11
Although it is not possible to evaluate whether the present prevalence rates of PCOS and that of associated obesity is higher than in the past, due to the lack of cross-sectional and longitudinal studies, nonetheless it can be hypothesised that the increasing worldwide prevalence of obesity may play a key role in favouring the development of PCOS in susceptible individuals. Therefore, we should expect an increasing rate of problems affecting women with PCOS, with particular reference to infertility, since PCOS is one of the most common causes of infertility in women.
Pathophysiology of the obesity–PCOS association
The high prevalence of obesity in women with PCOS has profound effects on both the pathophysiology and the clinical manifestation of the disorder. Pathophysiological mechanisms by which obesity influences the expression of PCOS are complex and not completely understood.1,12,13 Undoubtedly, obesity may influence the development of hyperandrogenism. This may occur in different ways (Figure 1).
Abdominal obesity as a condition of relative hyperandrogenism
Obesity per se represents a condition of sex hormone imbalance in women. Levels of the sex hormone-binding protein (SHBG) tend to linearly decrease with increasing body fat, and this may lead to an increased fraction of free androgens delivered to target sensitive tissues.5 SHBG levels are regulated by a complex of factors, including estrogens, iodothyronines and growth hormone (GH) as stimulating agents and androgens and insulin as inhibiting factors.14 The net balance of this regulation, with the dominant role of insulin, which inhibits SHBG synthesis in the liver, may be responsible for the decrease of SHBG concentrations observed in obesity. Obesity in fact is associated with hyperinsulinaemia that compensates for the presence of insulin resistance, reduced GH levels and increased testosterone production rate, all conditions reasonably explaining lowered SHBG concentrations. This is what occurs particularly in the presence of the abdominal phenotype of obesity, which develops as a consequence of a prevalent enlargement of visceral fat depots, which are characterised by specific hormonal and metabolic activities.15,16 Women with central obesity, in fact, usually have lower SHBG concentrations in comparison with their age- and weight-matched counterparts with peripheral obesity.15 In addition, women with central obesity have higher testosterone and dihydrotestosterone production rates than those with peripheral obesity, which may exceed their metabolic clearance rates.17 Moreover, an increased production rate occurs even for androgens not bound to SHBG, such as dehydroepiandrosterone and androstenedione.18 Therefore, the abdominal phenotype of obesity can be defined as a condition of relative functional hyperandrogenic state. Due to the specific action of androgens on the morphology and metabolic activity of the visceral adipocytes, it has also been argued that this endocrine milieu may in turn play a crucial role in preferentially determining an enlargement of visceral adipose tissue, thereby causing the abdominal obesity phenotype in women. This may be relevant for PCOS, this condition being associated with a high prevalence of abdominal fatness, even in those presenting with normal (<25) BMI.
Interestingly, the role of adipose tissue is also crucial in controlling the balance of sex hormone availability in the target nonfat tissues. In fact, adipose tissue is able to store various lipid soluble steroids, including androgens. Most of the sex hormones appear to be preferentially concentrated within the adipose tissue rather than in the blood. As a consequence, since the amount of fat tissue is greater than the intravascular space in obesity, and the steroid tissue concentration is much higher than in plasma, the steroid pool in obese subjects is greater than that found in normal weight individuals.19 In addition, fat represents a site of intensive sex hormone metabolism and interconversion, due to the presence of several steroidogenetic enzymes, such as 3β-dehydrogenase, 17β-hydroxydehydrogenase and the aromatase system.5,9,19 Obesity, particularly the abdominal phenotype, may thus add further specific mechanisms in the development of androgen excess in women with PCOS.20
Early factors during intrauterine life
Interestingly, some recent studies focused on the potential role of obesity in influencing the development of hyperandrogenism through still undefined specific factors during the intrauterine gestational period of life. In an obese woman with PCOS, the presence of obesity in her mother during pregnancy appears in fact to increase the susceptibility to develop hyperandrogenism and the PCOS phenotype of the daughter later in time.21 This finding is in line with a theory supported by longitudinal studies performed in animal models, supporting the theory that in utero androgen excess may be an important factor programming subsequent PCOS development during puberty.22 This theory has been substantiated by studies performed in nonhuman primates23 and sheep.24 Experimental data reported so far have demonstrated that androgen exposure during fetal life may favour the development of excess body fat, particularly in the visceral depots, which may be crucial in favouring a primary ovarian disordered steroidogenetic pathway, leading in turn to the development of hyperandrogenised ovaries later in life.22 Metabolic factors, particularly insulin, appear to play a crucial role in this context.
Leptin and other inappropriate signals regulating the adipose tissue
Together with this line of thinking, it is also possible that the early onset of overweight or obesity, i.e. during peripubertal age, may play a role in the development of hyperandrogenism by the intervention of multiple interrelated mechanisms, which primarily involve inappropriate signals coming from or indirectly involving the adipose tissues, together with different hormones and/or alterations of specific regulatory pathways. They include insulin, the insulin growth factor system, the opioid system, estrogens, several cytokines, particularly leptin and, possibly, the endocannabinoid system. Most of these factors have been discussed in previous recent review articles, to which the reader can refer.20,25
Leptin plays an important role in the regulation of the hypothalamic–pituitary–gonadal (HPG) axis in humans.25 In addition, current knowledge suggests that leptin is a potential candidate involved in the pathogenesis of hyperandrogenism and infertility in women with PCOS. Leptin is considered as one of the major peripheral signals that affect food intake and energy balance, and obesity is a classic condition of circulating leptin excess. The discrepancy between high leptin blood levels and its central effects represents the basis for the concept that most forms of obesity may represent a condition of leptin resistance.26,27 On the other hand, many tissues other than fat mass have been shown to express leptin and its receptors. Specifically, there is evidence that leptin acts directly on the ovaries through functional receptors detected on the surface of ovarian follicular cells, including granulosa, theca and interstitial cells.28
Recent in vitro data indicate that leptin may exert a direct inhibitory effect on ovarian function, by inhibiting both human granulosa and thecal cell steroidogenesis, probably by antagonising stimulatory factors, such as insulin growth factor-1, transforming growth factor-β, insulin and LH.28 Moreover, high leptin concentrations in the ovary may interfere with the development of a dominant follicle and oocyte maturation, as demonstrated by in vitro and in vivo studies.29 Finally, exogenous infusion of leptin has been shown to significantly decrease ovulation rate in the female rat.29 Taking into account the involvement of leptin in the control of ovulation and reproduction, there has been much interest focused on leptin levels in women with PCOS, since most women with PCOS are obese. To date, higher circulating levels of leptin than those expected in relation to BMI, or normal concentrations of leptin, have been reported.5,8,25 Whether high leptin levels in the peripheral circulation and/or in ovarian tissues may play a role in determining anovulation in obese women with PCOS is, however, at present unknown. It could be argued that obesity-associated hyperleptinaemia may represent an additional factor involved not only in the development of insulin resistance but also in the impairment of ovarian function, particularly in women with PCOS. This topic has been extensively discussed in a recent review to which the reader should refer for further information.25
Potential role of the endocannabinoid system
The historical description of the evolutionary concept and biological basis of the endocannabinoid system has been reviewed in recent articles to which the reader should refer for a deeper understanding.30 The identification of the endocannabinoids, the enzymatic machinery for their synthesis and degradation and the specific cannabinoid receptors type 1 (CB1) and type 2 (CB2) is the result of the scientific research performed in the last decade. Endocannabinoids are derivatives of arachidonic acid and they belong to a family composed of an increasing number of compounds, the most studied so far being arachidonoyl ethanolamide, named anandamide (AEA) (8), and 2-arachidonoyl glycerol (2-AG).31,32 An interesting aspect of the endocannabinoid system is its activity ‘on demand’, meaning that the system can be activated with a closely regulated spatial and temporal selectivity: only ‘when’ and ‘where’ it is needed.33 This property poses an important distinction that should be taken into account when the physiological functions of the endocannabinoid system are compared with the pharmacological actions of exogenous cannabinoid receptor compounds which lack such selectivity.
Astonishing amount of data have been accumulated, providing increasingly deeper insights regarding the biological roles of the endocannabinoid system. In general, it can be summarised that the endocannabinoid system is involved in different physiological functions, many of which are related to the stress recovery systems and to the maintenance of homeostatic balance.34 Among other functions, the endocannabinoid system is involved in neuroprotection, modulating nociception, regulating motor activity and controlling certain phases of memory processing. In addition, the endocannabinoid system contributes to modulate the immune and inflammatory responses. It also influences the cardiovascular and the respiratory systems by controlling heart rate, blood pressure and bronchial functions. Finally, yet importantly, endocannabinoids are known to exert important antiproliferative actions.35
Endocannabinoids are deeply involved in the dynamic and homeostatic regulation of feeding and energy metabolism. They act as orexigenic compounds, by contrast, antagonists of the CB1 receptors may reduce food intake.35 The regulation of metabolic processes includes both central and peripheral actions at the hypothalamic levels and in the adipose tissue and the liver, respectively. The peripheral action of the endocannabinoids has been documented in animal models treated with a selective CB1 antagonist, rimonabant.36–38 While endogenous cannabinoids AEA and 2-AG promote lipogenesis, CB1 antagonists have been found to increase energy expenditure through activation of futile cycles, enhance lipolysis and stimulate glucose metabolism, by increasing glucose transpoter-4 recruitment and activation.36–38 By these mechanisms, the CB1 antagonist rimonabant has been proposed as a new pharmacological treatment for tackling obesity.39,40 Indeed, it has been clearly documented that the endocannabinoid system may be overactivated in dietary-induced and genetic animal models, therefore favouring increased adipocyte enlargement.35 Interestingly, however, the endocannabinoid system has been found to regulate multiple endocrine functions, including the HPG axis.
The observation of changes in reproductive functions by using cannabis derivatives strongly supports the suggestion that they exert potent negative effects on reproduction in both sexes and this finding has been confirmed in various species.41 The primary negative effects are ascribed to a hypothalamic action, although some of these downregulating influences may be mediated directly at the level of the pituitary and the ovary.
By suppressing the secretory pulse of luteinizing hormone (LH),41 cannabinoids have been shown to downregulate blood LH levels,42 but administration of gonadotrophins or gonadotrophin-releasing hormone (GnRH) can restore ovulation or LH release, respectively, even in the presence of high levels of tetrahydrocannabinol.43 Importantly, tolerance to the antireproductive effect seems to develop after a few chronic treatment cycles.44 The common notion is that cannabinoids indirectly modify GnRH secretion by negatively modulating the activity of neurotransmitters known to facilitate GnRH secretion, such as norepinephrine and glutamate, and by stimulating those modulators known to downregulate GnRH secretion, such as dopamine, Gamma aminobutyric acid, opioids and corticotpropin releasing hormone.45
While cannabinoids are able to modulate the HPG axis, it is not yet known how, where and under what circumstances the endocannabinoids are produced to do so. It has been demonstrated that AEA fluctuates during the ovarian cycle in both the hypothalamus and pituitary,46 thus influencing hormonal secretion and sexual behaviour through CB1 receptor activation.47 Furthermore, considerable endocannabinoid production was found in the ovary, in particular at the time of ovulation, making it possible to hypothesise that the endocannabinoids may help to regulate follicular maturation and development of the ovary.48 However, an excess of cannabinoids may conversely impair regular ovulation, not only acting at hypothalamic level but also directly affecting ovarian granulosa layers.49
The endocannabinoid system also displays an important role during early pregnancy and in modulating embryo–uterine interactions, which are probably mediated by the fatty acid amide hydroxylase (FAAH) enzyme system, which plays a key role in the regulation of the endocannabinoid metabolism and degradation.50 An additional part of the scientific work in this area is that the FAAH is under the strict regulation of several hormones, such as progesterone, leptin and follicle-stimulating hormone (FSH), very well known modulators of fertility.51
Intriguingly, there are therefore several possibilities that specific, still undefined dysfunctions of the endocannabinoid system may play a role in the pathophysiology of obesity-related PCOS, with particular reference to ovulatory alteration and infertility. On the other hand, as reported above, it should be pointed out that overactivity of the system has not been documented in humans so far, but only in experimental animals. In addition, whether the PCOS may be a potential target for CB1 antagonist therapy is not at present known, although clinical trials will probably be carried out in the near future. Rimonabant and related upcoming CB1 receptor antagonist drugs may therefore be proposed not only for tackling obesity but also for dealing with the variety of metabolic alterations related to the pathological fat increase in abdominal depots. In this sense, the PCOS represents a unique model in which CB1 receptor antagonists may exhibit a powerful dual effect. In addition, there are theoretical possibilities that the endocannabinoids are involved in PCOS-related infertility, as preliminary data in men suggest.52
Metabolic factors: insulin
As reported above, obesity, particularly the abdominal phenotype, is a condition of insulin resistance and compensatory hyperinsulinaemia. Contrary to what occurs in the classic target tissues (i.e. muscle, liver and adipose tissue) of insulin action, that become resistant to insulin, the ovaries remain responsive to insulin throughout the interaction with its own receptor. A large number of in vitro studies have demonstrated that excess insulin is capable of stimulating steroidogenesis and excessive androgen production by the ovarian theca cell system.13 In women with PCOS, excess insulin can therefore participate in determining increased ovarian androgen synthesis (particularly androstenedione and testosterone) and blood levels. In vivo, numerous studies have subsequently demonstrated that both acute and chronic hyperinsulinaemia can stimulate testosterone production and that suppression of insulin levels can conversely decrease blood androgen concentrations (reviewed by Gambineri et al.5 and Poretsky et al.13). The excess of local ovarian androgen production induced by high circulating insulin may also cause premature follicular atresia and thus favour anovulation.13 Insulin resistance and hyperinsulinaemia, which develop together with the obesity state, may theoretically play a primary role in favouring hyperandrogenism in overweight or obese women susceptible to developing PCOS, particularly during pubertal age. In addition, numerous clinical studies have clearly demonstrated that obesity tends to amplify both hyperandrogenism and insulin resistance in the presence of PCOS, as reported in the following paragraph.
The phenotype of PCOS in relation to the presence of obesity
Various studies have evaluated the impact of obesity on the hyperandrogenic state in women with PCOS. They have uniformly demonstrated that obese women with PCOS are characterised by significantly lower SHBG plasma levels and worse hyperandrogemia in comparison with their normal weight counterparts.5 In addition, a negative correlation has been reported between body fat mass and circulating androgens.5 It has also been repeatedly reported that a higher proportion of obese women with PCOS complain of hirsutism and menstrual disorders in comparison with normal weight women.5 Therefore, there is consistent evidence that the increase in body weight may favour a more severe hyperandrogenism in women with PCOS.
Women with PCOS are characterised by a high prevalence of several metabolic abnormalities that are strongly influenced by the presence of obesity. Adequate confirmation of the role of obesity in determining hyperinsulinaemia and insulin resistance in women with PCOS derives from studies comparing groups of normal weight and obese women with PCOS. Both fasting and glucose-stimulated insulin concentrations were significantly higher in obese than in nonobese PCOS subgroups (reviewed by Gambineri et al.5 and Dunaif12). Accordingly, studies examining insulin sensitivity by using different methods such as the euglycaemic–hyperinsulinaemic clamp technique, the frequently sampled intravenous glucose tolerance test (FSIVGTT) and the intravenous insulin test have further demonstrated that obese women with PCOS have significantly lower insulin sensitivity than in their nonobese counterparts with PCOS and, therefore, a more severe insulin-resistant state (reviewed by Gambineri et al.5).
The percentage of women affected by PCOS and obesity who present with glucose intolerance is rather high, ranging from 20 to 49%,12 which is substantially above the prevalence rates reported in premenopausal women in population-based studies. In contrast, glucose intolerance in normal weight women with PCOS is uncommon.12,53 Collectively, this suggests that obesity per se plays an important role in altering the insulin–glucose system in PCOS.
In addition, several studies have identified defects of insulin secretion in obese women with PCOS.12 Using the FSIVGTT, Dunaif and Finegood53 reported that obese women with PCOS mount an inadequate insulin secretory response to compensate for the peripheral insulin-resistant state, suggesting relative β-cell dysfunction. However, regardless of alterations in insulin secretion, in a 10-year follow-up study, we found that both fasting and glucose-stimulated insulin and C-peptide levels tended to increase spontaneously and significantly in women with PCOS, suggesting a worsened insulin-resistant state over time.54 In the same study, we also found that several women developed impaired glucose tolerance. Longitudinal data are therefore warranted to investigate which factors, namely progressive insulin resistance and/or subtle alterations of insulin secretion, can predict the well-documented susceptibility of obese women with PCOS towards the development of type II diabetes.
Although PCOS per se may be associated with alterations of both lipid and lipoprotein metabolism, the coexistence of obesity usually leads to a more atherogenetic lipoprotein pattern. A greater reduction of high-density lipoproteins (HDLs) together with a higher increase of both triglycerides and total cholesterol levels have in fact been observed in obese women with PCOS, with respect to normal weight women with PCOS (reviewed by Gambineri et al.5).
What is reported above clearly shows that many women with PCOS may present with features characteristic of the so called metabolic syndrome, which has been defined as a cluster of risk factors for cardiovascular diseases. Defining the metabolic syndrome is a difficult task and over the past few years many definitions have been proposed, focusing on the association between obesity, abdominal fat distribution, metabolic and vascular parameters. One of the most commonly used definitions is the one proposed by the National Cholesterol Education Program Expert Panel/Adult Treatment Panel III in the United States (NCEP/ATPIII).55 Accordingly, three or more of the following should be used to define the presence of the metabolic syndrome in women: (i) central obesity (waist circumference > 88 cm); (ii) increased triglycerides (>1.7 mmol/l), (iii) reduced HDL cholesterol (<1.3 mmol/l); (iv) increased blood pressure (≥135/85 mmHg or medication); (v) fasting plasma glucose ≥ 6.1 mmol/l. The prevalence of the metabolic syndrome is very high in the general population, with a great variability according to different environmental factors, ethnicity and geographical areas, but available data suggest that this prevalence may be significantly higher in women with PCOS,56 ranging from 35 to 50%. As expected, the majority of these women are obese, and most of them are characterised by the presence of the abdominal phenotype. As reported above, this may also occur in apparently normal weight women with PCOS.
Menses abnormalities and infertility
PCOS is one of the most common causes of anovulation and endocrine infertility in women. Several studies have clearly demonstrated that menstrual abnormalities are more frequent in obese than normal weight women with PCOS.5 Moreover, there is evidence that a reduced incidence of pregnancy and blunted responsiveness to pharmacological treatments to induce ovulation may be more common in obese women with PCOS.57 In a prospective study carried out in 158 anovulatory women, the dose of clomiphene required to achieve ovulation was positively correlated with body weight.58 Both insulin resistance and hyperinsulinaemia, which parallel the increase in body fat, may be responsible for the alteration of both spontaneous and induced ovulation observed in obese women with PCOS.
Administration of insulin-sensitising agents, such as metformin and troglitazone, was in fact associated with improved menstrual cyclicity in women with PCOS.59 Moreover, a double-blind, placebo-controlled collaborative study, performed in a large cohort of women with PCOS,60 demonstrated that short-term metformin treatment increased both spontaneous and low-dose-clomiphene-induced (50 mg daily for 5 days) ovulation rates. It has also been reported that, compared with normal weight women, obese women with PCOS may have lower ovulatory responses to pulsatile GnRH analogue administration.61 Accordingly, the pregnancy rate after a low-dose human menopausal gonadotrophin or pure FSH administration may be significantly lower in obese women than in normal weight women with PCOS.62 Finally, in studies performed in women with PCOS conceiving after in vitro fertilisation or intracytoplasmatic sperm injection, it was observed that those with obesity had higher gonadotrophin requirements during stimulation, fewer oocytes, a higher abortion rate and a lower live birth rate than their nonobese counterpart.63 In conclusion, a decreased efficiency of the different treatments for anovulation and infertility may be expected in obese women with PCOS. The presence of hyperinsulinaemia is probably the major factor responsible for this undesirable condition. These findings additionally support the negative impact of obesity on fertility in women with PCOS.
The impact of body fat distribution on the phenotype of PCOS
It is well documented that women with PCOS have a high prevalence of abdominal distribution of body fat, even if they are normal weight.64,65 The impact of abdominal obesity on PCOS may be greater than expected, since this phenotype is associated with more pronounced hyperandrogenism and insulin resistance than the peripheral body fat phenotype. As reported above, increased visceral fat development may occur much earlier than general fat excess in the natural history of PCOS, leading in turn to the development of insulin resistance and associated hyperinsulinaemia and, finally, to a hyperandrogenic state. We have repeatedly demonstrated that the androgen profile and basal insulin levels, as well as the insulin response to a glucose load, are significantly higher in the subgroup of women with PCOS having abdominal body fat distribution than in the group with the peripheral type, regardless of BMI.5 This has been confirmed in studies using dual-energy X-ray absorptiometry to define different obesity phenotypes.65 Moreover, Holte et al.66 found a significant association between abdominal fat mass and insulin resistance determined by the euglycaemic–hyperinsulinaemic clamp technique. They also found a highly significant correlation between plasma free fatty acid (FFA) concentrations and insulin resistance, which supports the concept that an increase of FFA flux from the highly lipolytic abdominal fat to the liver and muscles may represent the most important link between abdominal obesity and the insulin-resistant state. Moreover, this subgroup of women with PCOS may have a more unfavourable lipid profile, namely higher triglyceride and very-low-density lipoprotein levels and lower HDL cholesterol concentrations.5 In addition, women with PCOS having the abdominal phenotype present with a higher prevalence of menstrual abnormalities and acanthosis nigricans (a cutaneous marker of insulin resistance) and a tendency towards more severe hirsutism.5
As discussed above, abdominal obesity is associated with profound alterations of both production and metabolic clearance rates of major androgens and reduced SHBG blood levels. In abdominally obese women with PCOS, androgens could, in turn, play a role in regulating tissue metabolism. In fact, at the level of visceral depots, testosterone stimulates lipolysis and, therefore, increases FFA efflux.15 In addition, at the level of the muscle, testosterone modifies the histological structure by increasing type II, less insulin-sensitive fibres. These androgen-dependent mechanisms may have a further important impact on the insulin-resistant state.
In summary, in women with PCOS, abdominal obesity per se may play a key role in determining both altered androgen metabolism and insulin resistance. This may be an important consideration when phenotyping PCOS and in devising therapeutic strategies to reduce both hyperinsulinaemia (and improve insulin resistance) and hyperandrogenism.
The therapeutic efficacy of weight loss indirectly supports the pathogenetic role of obesity in PCOS
The importance of obesity in the pathogenesis of PCOS and the implications that metabolic alterations frequently associated with the obesity state have on long-term health have induced many investigators to evaluate strategies to control weight disorders in PCOS. Obese women with PCOS often report extreme difficulty in losing weight and maintaining weight loss. Nevertheless, when studied, the resting metabolic rate and postprandial thermogenesis do not differ in obese women with PCOS and weight-matched control subjects, suggesting that the same need for caloric restriction relative to energy expenditure is necessary for weight loss in both groups.67 Moreover, no differences in hormonal responses to physical exercises were found between women with PCOS and weight-matched control women.68
The best therapeutic strategy for favouring weight loss in obese women with PCOS has not yet been investigated. However, lifestyle interventions, particularly with hypocaloric diet with or without associated increased physical activity, have proved their efficacy.69 Physical exercise could have an important impact on insulin resistance. In the context of overall glucose homeostasis, a single instance of exercise can markedly increase rates of whole body glucose disposal70 and increase the sensitivity of skeletal muscle glucose uptake to insulin.71 Moreover, regular physical activity is required in order to have a lasting effect on insulin responsiveness.72 A recent study focusing on the effect of a 48-week period lifestyle intervention, including dietary advice and a standardised physical activity program, with or without metformin treatment, showed a significant positive effect on ovulatory performance which was related to the amount of weight loss, rather than the effect of metformin.73 The majority of studies have, however, investigated the impact of hypocaloric diet or modification of the composition of the diet on hormonal and metabolic abnormalities of obese women with PCOS and on the main clinical features, including menses abnormalities, chronic anovulation and infertility. Although there are not many reports in the literature on the effect of weight loss in obese women, all these studies, nonetheless, clearly demonstrate that weight loss improves both endocrine and metabolic abnormalities and that ovulation and fertility may be significantly restored.69
Dietary-induced weight loss improves insulin resistance and hyperinsulinaemia.27 Insulin sensitivity before and after weight loss has also been studied by the use of several methodologies, including the euglycaemic–hyperinsulinaemic clamp technique.69 Both fasting and glucose-stimulated insulin levels tend to significantly decrease even after a weight decrease by approximately 5–10% of baseline values.69 Using the euglycaemic–hyperinsulinaemic clamp technique, one study found that glucose disposal rate returned to baseline values after a few months of hypocaloric diet.66
In a more recent study performed to evaluate the efficacy of hypocaloric diet alone or associated with insulin sensitisers (metformin) or antiandrogens (flutamide) for a period of 7 months in different groups of obese women with PCOS, we found that although all groups improved fasting and glucose-stimulated insulin and insulin resistance, these effects could be attributed to dietary weight loss rather than to specific drugs, which suggests a primary role of weight loss in improving alterations of the insulin–glucose system.74 The effects of a long-term (12 months) low-calorie diet in a group of twenty obese women with PCOS on insulin levels and insulin sensitivity are reported in Figure 2.
Weight loss also induces significant benefits on hyperandrogenism. Harlass et al.75 first reported a decrease in total and free testosterone in a small group of obese anovulatory women after modest weight loss. In subsequent larger noncontrolled studies,9,69 a significant reduction of total and free testosterone levels after dietary-induced weight loss was confirmed. Additional controlled studies have obtained similar findings,69 although other studies did not report any effect.76 Other studies have shown a significant increase of SHBG concentrations, which is consistent with a reduction of the bioavailable free androgens.69 Moreover, there are reports on the beneficial effects of dietary-induced weight loss over a long period of time on hirsutism and total testosterone concentrations (Figure 2). Jakubowicz and Nestler77 performed a leuprolide test before and after 6-month weight loss following a hypocaloric diet and found that the 17a-hydroxyprogesterone response returned to baseline values, suggesting a decrease in the activity of P450c17α, a key enzyme involved in ovarian androgen production.
Whether diet composition may have a preferential role on androgen and metabolism has been investigated in two studies.78,79 They were performed in small groups of obese women with PCOS, randomly prescribed a high protein (30%), low carbohydrate (40%) versus a low protein (15%), high carbohydrate (55%) hypocaloric diet for a short time (1–3 months, respectively). In both the studies, the final results were consistent with a significant effect of weight loss on total testosterone and fasting or glucose-stimulated insulin levels, without any significant effect of diet composition, suggesting that hypocaloric restriction, rather than diet composition, is responsible for the beneficial effect of weight loss on hormones and metabolism.
Several factors may be responsible for these effects. The most evident is represented by the reduction of insulin levels associated with improved insulin-resistant state, which implies a reduction of both the stimulatory effect of insulin excess on ovarian steroidogenesis and the inhibiting effect of hepatic SHBG synthesis. In addition, leptin concentrations decrease after weight loss, and this may represent another mechanism favouring the reduction of ovarian steroid secretion in obese women with PCOS, excess leptin being involved in the negative regulation of endocrine ovarian function, as reported above.
Chronic anovulation is a common feature of women with PCOS and the restoration of normal menstrual cycles and of ovulatory function represents the primary goal to be achieved for many women with PCOS. Evidence exists that dietary-induced weight loss may improve both menses abnormalities and spontaneous ovulation in the majority of affected women (reviewed by Pasquali et al.9). On the other hand, it should be noted that available data on the consequences of weight loss on menses and ovulatory abnormalities among obese women with PCOS have often been obtained in uncontrolled open studies,80–82 in studies including a control group which failed to complete the study program83 (Table 2) or with a lack of weight loss during therapy.84 Mechanisms responsible for the beneficial effect of weight loss on menses alterations and fertility depend on the coexistent reduction of both hyperinsulinaemia and hyperandrogenism.
Table 2. Efficacy of lifestyle intervention in anovulatory infertile women. Reproduced with permission from Clark AM, et al. Hum Reprod 1995;10:2705–1283
Completed, n= 67 (Mean ± SD or %)
Drop-out, n= 20 (Mean ± SD or %)
The original cohort included 87 infertile obese women, most of whom had PCOS, and treatment consisted of a long-term lifestyle intervention program, including physical activity and hypocaloric diet. Those who completed were compared with those who dropped out.
Pregnancies (cumulative: spontaneous or assisted reproductive technologies)
This is further confirmed by studies using insulin sensitisers, such as metformin and tiazolidinediones, in women with PCOS with insulin resistance, demonstrating that by improving insulin resistance and associated hyperinsulinaemia, androgen levels may decline, menses abnormalities may improve and ovulation may be restored in most women, often regardless of weight loss and even after just a few weeks of pharmacological treatment.69,84–87,60 However, two recent metanalyses on the effects of metformin on ovulation88,89 clearly demonstrated that where metformin was used as a sole agent, ovulation was achieved in 46% of recipients compared with 24% in the placebo group; where metformin and clomiphene were used in combination, 76% of recipients ovulated versus 42% of those receiving placebo. The final conclusion that equal or better ovulation rates than those achieved by metformin have been reported by using lifestyle intervention to achieve weight loss is relevant for the purposes of this review.
Long-term lifestyle intervention may be needed to achieve weight loss and improve fertility. In a study performed by Hoeger et al.,73 a group of 38 overweight or obese women with PCOS were treated with metformin or placebo plus lifestyle intervention (including a 500–1000 kcal deficit per day, weekly behavioural education program and exercise [150 minutes per week]) or metformin alone. In the 23 women completing the trial, they found an overall weight loss of 7–10%; hirsutism did not significantly change, although total and free testosterone decreased by approximately 20% in the combined group treatment, whereas fasting and glucose-stimulated insulin decreased by approximately 25%. Most importantly, all treatments significantly improved ovulation rate, with a greater efficacy in those who lost weight compared with those who did not. Interestingly, however, prediction of three or more ovulations per 24 weeks by a logistic regression analysis indicated that weight loss (OR: 8.97; P= 0.030), rather than metformin treatment (OR: 1.14; P= 0.891), was responsible for this effect. This is in agreement with our own data obtained in a large cohort of women treated for 12 months with a low-calorie diet alone or combined with metformin and/or a pure antiandrogen, flutamide (A. Gambineri et al., unpubl. obs.). These findings further support the beneficial effect of weight loss on ovulation in obese women with PCOS also in the long term and suggest that weight loss should be encouraged in these women despite difficulties in achieving and maintaining adequate compliance on an individual basis. Ultimately, the goal represented by an increased probability of ovulating and, hopefully, becoming pregnant may have a great impact on improving compliance. This has been demonstrated in one study where obese women with PCOS with patent fallopian tubes and chronic anovulation were invited to follow a hypocaloric diet in order to achieve a 5% weight loss in a 6-month period, after which they were forced to achieve further weight loss in the presence of a poor ovulatory response. Overall, after 8–10 months, approximately 80% achieved regular menses, 60% had ovulatory cycles and 40% became pregnant.90
Conclusions and perspectives
The increasing prevalence of obesity among adolescent and young women with PCOS may partly depend on the increasing worldwide epidemic of obesity, although this hypothesis should be supported by long-term prospective epidemiological trials. This may have great relevance in preventive medicine and offer the opportunity to expand our still limited knowledge of the genetic background of the PCOS itself.
PCOS also represents a relatively simple clinical condition but with complex pathophysiological aspects. Further efforts should be made to investigate the potential mechanisms by which obesity affects PCOS, particularly during adolescence and early reproductive age.
Since studies performed in experimental animals suggest that androgen exposure during intrauterine life may affect a women’s susceptibility to develop PCOS later in life, both retrospective and prospective studies in large cohorts or well-selected groups of women should be undertaken. This topic appears to be a promising way of expanding our knowledge on the natural history of PCOS and may open new strategies to prevent it.
Obesity has profound effects on reproductive function in women with PCOS. Although the mechanisms responsible are not clearly elucidated, nonetheless there is consistent evidence that weight loss represents a simple and cost-effective safe way to improve spontaneous ovulatory cycles and pregnancy rates in affected women. Unfortunately, weight loss interventions do not appear to be a common practice among fertility centres and gynaecological clinics, in spite of the clear evidence in the literature. We need well-controlled clinical trials to potentially predict responder women, although some studies found that the majority of those with PCOS and obesity are likely to benefit even from a small weight loss. Treatment of obesity for endocrinological reasons may also improve fertility rates, therefore providing an additional unexpected benefit. Many aspects, however, have still to be defined, particularly regarding the best time to start weight loss according to the age of each woman, the duration of lifestyle intervention programs, the usefulness of additional drugs, such as insulin sensitisers and, finally, whether treatment of obesity by dietary management may help women to avoid premature fetal loss or miscarriages, together with metabolic derangement during pregnancy, such as gestational diabetes or hypertensive disorders.