Cardiovascular risk

++ (?)

+ (?)

Endometrial carcinoma risk



PCOS, polycystic ovary syndrome; IGT, impaired glucose tolerance. (Adapted from ref. 8.)

PCOS, polycystic ovary syndrome; IGT, impaired glucose tolerance. (Adapted from ref. 8.)

[weight in kg divided by height in m2]), women with PCOS have a higher percentage of body fat and a larger waist-to-hip ratio (7) than their matched controls.

Clinical manifestations of PCOS are different in obese women with PCOS compared with lean women with PCOS (8) (Table 1). Obese women with PCOS report more menstrual irregularities and more oligo-/amenorrhea than their lean counterparts. Obesity is also associated with an increased prevalence of infertility in PCOS and in the general population as well. The risk of miscarriage is also increased in obese women, whether or not they have PCOS. Moreover, obesity increases the risk of complications during pregnancy, such as gestational diabetes and pre-eclampsia (9). Finally, obese women with PCOS tend to have higher hirsutism and acne scores than lean PCOS women.

Biochemical features of PCOS also differ between obese and non-obese women affected by the syndrome. Insulin resistance is associated with PCOS, but obesity seems to amplify the degree of insulin resistance and hyperinsulinemia (10) (Fig. 1). Obese women with PCOS also have a higher risk of developing glucose intolerance or diabetes than lean women with PCOS (11). Dyslipidemia is more frequent in PCOS and is often characterized by high triglycerides, high cholesterol, and low high-density lipoprotein levels (3). Obese women with PCOS usually have higher levels of triglycerides than normal-weight women with PCOS (8).

Growth hormone (GH) and insulin-like growth factors (IGF)-1 and -2 are involved in ovarian function and possibly in the pathogenesis of PCOS. GH pulse amplitude and GH mean levels over 24 hours are reduced in obese PCOS compared with lean women with PCOS (6). IGF-1 and IGF-2 levels in women with PCOS do not seem to differ from those in women without PCOS, but the bioavailability of these hormones seems to be influenced by body weight in women with PCOS. Lean women with PCOS appear to have a higher IGF bioavailability (by IGF binding protein [IGFBP]-1 suppression and greater GH stimulation of IGF), whereas obese women with PCOS seem to have lower IGF bioavailability (because of higher insulin levels and reduced GH stimulation) (6).

Sex hormone levels are also influenced by the degree of overweight. women with PCOS have lower sex hormone-binding globulin (SHBG) levels, with more pronounced SHBG reduction in obese women with PCOS, especially if they present with abdominal obesity (6). Lower SHBG levels increase the bioavailability of sex hormones, and therefore increase hyperandrogenemia in obese PCOS women. Both total and free testosterone levels are increased in obese PCOS compared to normal-weight women with PCOS because of the combination of increased androgen production and lower SHBG (6).

Fig. 1. Insulin sensitivity in lean and obese women with polycystic ovary syndrome (PCOS) or subjects with type 2 diabetes mellitus (NIDDM). Women with PCOS are less insulin sensitive than normal women, and obese women with PCOS are the most insulin resistant of all groups. (Adapted from ref. 10).

Estrogens are elevated in obese women in the general population. In obese women with PCOS, the abundance of adipose tissue induces high peripheral aromatization of androgens into estrogens. This effect, combined with low SHBG levels, increases the free estradiol fraction, which probably accounts for the increased risk of endometrial cancer observed in women with PCOS. Moreover, for reasons not completely understood, obese women with PCOS have lower luteinizing hormone (LH) levels than lean women with PCOS, and very obese women with PCOS often have LH levels comparable with those of women without PCOS (6).

2.1.2. Factors Predisposing to Obesity

Knowing that obesity is highly prevalent in PCOS and influences the clinical and biochemical manifestations of the syndrome, it would be important to summarize environmental factors predisposing to it. Obesity has increased dramatically in the past few decades. Human beings have been evolving for thousands of years in a world where the individuals who were capable of surviving despite long winters, wars, famines, and other starvation situations had a genetic advantage. In the past 50 years or so, our environment has changed to a world of high food availability—often densely caloric—and low energy expenditure. The consequent obesity epidemic has led the scientific community to try to identify the factors predisposing to obesity (12).

Parental weight is positively related to child's body weight, especially if both parents are obese (12). However, being born from a diabetic mother increases the risk of obesity in teenage years, even after adjustment for the mother's BMI (13). Indeed, a higher birthweight has led to a higher BMI in childhood and in adulthood in many longitudinal studies (14). Finally, many studies have also shown that BMI at any age usually tracks into the subsequent years (15-17).

Normally, BMI increases rapidly after birth, then decreases over a few years to reach a minimum (between 3 and 7 years), and thereafter increases slowly until adulthood (but seems to continue to increase in adulthood according to more recent findings in adult cohorts). Age at adiposity rebound is defined as the age where the minimum BMI is reached, before the second rise of BMI (Fig. 2). An early adiposity rebound has been associated with a higher BMI later in life (18). Others have argued that this effect was simply related to BMI tracking, because a higher BMI in young age is associated with early adiposity rebound and by consequence to a higher BMI in the following years (19).

Age (years)

Age (years)

Fig. 2. Body mass index (BMI) growth charts for British girls, 1990. The age at adiposity rebound is illustrated by a dot on each BMI centile line. (Adapted from ref. 19.)

Another factor predisposing to obesity might be low socioeconomic status (20). Similarly, individuals with higher stress levels and lower level of life satisfaction have shown greater weight gain in longitudinal studies, especially in women (20). Former smokers usually have higher BMI than non-smokers or current smokers (21).

Instinctively, dietary intake and physical activity should be the two main environmental determinants of overweight and obesity. Unfortunately, both are very difficult to measure with precision, and studies have not always been able to prove the contribution of one or the other. In cross-sectional studies, frequent consumption of vegetables is usually associated with lower risk of obesity (21). Conversely, high dietary fat intake has been associated with greater weight gain in some longitudinal studies (15). However, a systematic review of the literature has underlined the inconsistency of these studies, the difficulty of estimating food intake from questionnaires, and the problem of underreporting in food diaries (12).

Physical activity level offers the same challenges concerning estimation by questionnaires, but shows a little more consistency. Many longitudinal studies have shown that higher level of physical activity is associated with less weight gain (16,17). Conversely, inactivity was associated with weight gain in other cohort studies. Finally, sedentary lifestyle, often measured as time spent watching television, has been related to higher risk of developing obesity and diabetes in both men (22) and women (23).

Overall, overweight and obesity are prevalent in women with PCOS and might play an important role in its pathophysiology. Furthermore, when women with PCOS are obese, they usually present with more severe clinical and biochemical features of the syndrome. Accordingly, prevention of obesity is critical in order to prevent PCOS and/or its consequences, and factors that predispose to obesity may start very early in life and are numerous throughout the life cycle.

2.2. Birthweight and PCOS

As mentioned previously, birthweight may be related to body weight later in life. Some authors have tried to determine if a direct link exists between birthweight and PCOS. In a population of women aged 18-25 years , Michelmore et al. (7) showed that an increased birthweight was predictive of polycystic ovaries on ultrasound, but not related to symptoms of PCOS. On the other hand, in the same cohort, low birthweight was related to insulin resistance (which is consistent with many other cohort studies). Using birth records from 1952 to 1953, Cresswell et al. (24) traced 235 women in their 40s. Polycystic ovaries were present on ultrasound in 21% of them, who were found to have a slightly but nonsignificant higher birthweight. A longer gestation, however, was significantly related to polycystic ovaries on ultrasound, especially in lean women. Furthermore, women born with heavier weights and from mothers with greater BMIs were more likely to have polycystic ovaries on ultrasound and were more likely to have a greater BMI themselves.

Sadrzadeh et al. (25) assessed the population of an infertility clinic and compared women with PCOS with women with diminished ovarian reserve capacities or tubal obstruction (as the reference group). They gathered information retrospectively using questionnaires. In this study, birthweight was not related to PCOS, but higher actual BMI (as we already know) and delayed menarche were. In a prospective cohort followed up from birth and re-assessed at 31 years old, Laitinen et al. (26) did not find a relationship between PCOS symptoms (hirsutism or oligomenorrhea) and birthweight, prematurity, or growth retardation. However, they did find that obesity, especially of central distribution, was related to PCOS symptoms.

Alternatively, in a recent report from a UK birth cohort, Ong et al. (27) showed that a low birthweight was associated with greater weight gain in subsequent years and a higher androgen level at 8 years in both boys and girls. Similarly, Ibanez et al. (28) found that being small for gestational age increased the risk of anovulation at 15 years (Fig. 3).

Overall, these studies are not convincing for a direct link between birthweight and PCOS. However, they underscore again the importance of a higher actual body weight in the clinical manifestations of PCOS, which might be related to birthweight. They also highlight the relationship between low birthweight and important characteristics of PCOS, particularly the development of insulin resistance.

2.3. Insulin Resistance

2.3.1. Insulin Resistance and PCOS

There is growing evidence that insulin resistance plays a key role in the pathophysiology of PCOS (3). Hyperinsulinemia increases androgen levels by several mechanisms (29), including by increasing androgen production and greater bioavailability by lowering SHBG. It has also been shown that women with PCOS are characterized by both increased peripheral insulin resistance and P-cell dysfunction. For the same level of glycemia, women with PCOS have higher basal plasma insulin levels but lower insulin secretory response to meals compared with weight-matched control women (30). Hyperinsulinemic euglycemic glucose clamp studies have demonstrated a significant decrease in

Fig. 3. Fractions of appropriate for gestational age (AGA) and small for gestational age (SGA) subpopulations distributed by number of ovulations detected over 3 months of study at age 15 years. Girls born SGA have significantly fewer ovulations over 3 months at age 15 years compared with girls born AGA. (Adapted from ref. 28.)

Fig. 3. Fractions of appropriate for gestational age (AGA) and small for gestational age (SGA) subpopulations distributed by number of ovulations detected over 3 months of study at age 15 years. Girls born SGA have significantly fewer ovulations over 3 months at age 15 years compared with girls born AGA. (Adapted from ref. 28.)

insulin-mediated glucose disposal in women with PCOS (10). Hyperinsulinemia results from increased basal insulin secretion and reduced hepatic clearance of insulin in both lean and obese PCOS women, but obesity also brings on increased hepatic glucose production (30).

We will not review here the cause nor the implication of insulin resistance in PCOS because it will be extensively discussed in Chapters 24 and 26. But we will review environmental and prenatal factors that might be implicated in the development of the insulin resistance typical of PCOS.

2.3.2. Early Predictors of Insulin Resistance

Many studies have confirmed that low birthweight is associated with central obesity and insulin resistance after adjustment for BMI. Hofman et al. (31) showed that prematurity or being small for gestational age was related to diminished insulin sensitivity measured at 4-10 years old. In the Nurses Health Study, low birthweight was associated with higher risk of type 2 diabetes, even after adjustment for BMI and all the other confounders (32).

Adiposity rebound has also been linked to development of insulin resistance. In Helsinki, a birth cohort followed up from 1934 showed that an early adiposity rebound (before 5 years vs after 7 years old) was related to an increased risk of type 2 diabetes (33). In India, a 30-year follow-up of a birth cohort also showed that early adiposity rebound was associated with the development of impaired glucose tolerance and diabetes mellitus (34). In this cohort, birthweight was not statistically predictive of the oral glucose tolerance test result, but a low weight in infancy (1 or 2 years old) was associated with abnormal glucose tolerance later in life, especially if followed by a rapid weight gain and higher weight at 12 years of age. However, as noted earlier, the adiposity rebound might also be a manifestation of higher BMI tracking into childhood.

In general, BMI tracks into childhood, and longitudinal studies have shown that a high BMI, even in childhood, was predictive of higher BMI and insulin resistance in adolescence and adulthood. The Minneapolis Children's Blood Pressure Study (35) followed individuals from 7 to 23 years of age and showed that BMI in childhood was predictive of early adulthood BMI, fasting insulin levels, lipid profile, and systolic blood pressure. The Bogalusa Heart Study (36) followed individuals (8-17 years old at baseline) over 11 years to identify the factors related to the clustering of features of the metabolic syndrome into adulthood. The strongest predictor was the initial BMI, which remained statistically significant even after adjustment for initial fasting insulin levels.

Overall, insulin resistance is strongly related to weight changes. A low birthweight is probably predictive of higher risk of impaired glucose tolerance and diabetes mellitus, especially if followed by a rapid weight gain, early adiposity rebound, and higher BMI into childhood and adolescence. However, weight gain in adulthood, overweight, and central obesity are still independently and strongly associated with insulin resistance and usually have a greater influence on the actual risk of impaired glucose metabolism. Insulin resistance alone does not cause PCOS in every woman. This might be explained by genetic predisposition(s) or perhaps by in utero imprinting or the timing of development and maturation of those women.

2.4. Precocious Pubarche and PCOS

Puberty is a critical period for the future appearance of PCOS because it involves sexual maturation and the initiation of ovulation and ovarian steroidogenesis. Moreover, it has been shown that the timing of menarche is influenced by body weight and adiposity and that insulin resistance typically increases during puberty.

Precocious pubarche (PP; also called precocious adrenarche) has been associated with higher risk of developing PCOS symptoms, but the link with PP, which is defined by appearance of pubic hair before 8 years of age in girls, has been more extensively investigated, especially by Ibanez and colleagues. PP is usually followed by normal puberty, menarche, and growth; but appears to predispose to functional ovarian hyperandrogenism (37). Early pubarche has been associated with increased risk of polycystic ovaries on ultrasound, hyperandrogenism, hyperinsulinism, and altered lipid profile (38) during and after pubertal development. Girls with PP also have a higher risk of developing anovulation a few years after menarche (39). Finally, Ibanez et al. (40) have found that girls 6-18 years old who had presented with PP displayed higher waist circumference, waist-to-hip ratio, total fat mass, and percentage fat mass than girls matched for age, pubertal stage, and BMI. These girls also presented with higher abdominal fat, which was positively correlated to the level of insulinemia and androgenemia.

Birthweight also influences pubertal evolution. Girls with PP had a lower birthweight than matched control girls, and those with hyperandrogenism had an even lower birthweight. If they presented with hyperinsulinism on top of that, their mean birthweight was lower again (41) (Fig. 4). Prospective follow-up of girls with PP until after menarche showed that the subgroup with lower birthweight had decreased insulin sensitivity, abnormal ovarian function, and higher triglycerides and low-density lipoprotein levels compared to girls with PP and normal birthweight (42).

In summary, these studies have shown that girls having a low birthweight followed by early pubarche are at increased risk of developing PCOS. Those who also presented with hyperinsulinemia were at highest risk to develop PCOS after puberty, and prepurbertal intervention with an insulin-sensitizing drug in these individuals prevented the progression towards PCOS (43). Therefore, birthweight and/or PP might be significant modifiers of the relationship between insulin resistance and PCOS.

2.5. Dietary Intake and Other Lifestyle Factors

There are few published studies of the influence of diet on PCOS. Wild et al. (44) investigated the difference in lipid content of the diet between women with PCOS and a control group of eumenorrheic women of similar age, as well as other lifestyle habits. Women with PCOS were heavier than controls, had a diet characterized by more saturated fat and less fiber, and were more sedentary. They did not find any difference in alcohol consumption and smoking habits, which is consistent with other observations. On the other hand, Wright et al. (45) did not find any difference in the composition of the diet or physical activity level of women with PCOS compared with age- and race-matched control women. Intriguingly, lean women with PCOS had a lower total caloric intake than lean control

Fig. 4. Precocious pubarche, hyperinsulinism, and ovarian hyperandrogenism in girls—relation to reduced fetal growth. (Adapted from ref. 41.)

women. The authors suspected that women with PCOS had to restrict themselves to maintain normal body weight.

Carmina et al. (46) looked at the differences between women with PCOS from the United States and Italy. American women with PCOS had a higher BMI and worse insulin resistance and lipid profile than the Italian women with PCOS. The only difference in diet found was the higher consumption of saturated fat in the American women. Therefore, diet composition might influence the manifestations of PCOS rather than causing PCOS.

2.5.1. Eating Disorders and PCOS

Some authors have noticed a link between binge eating or bulimia nervosa (BN) and PCOS. In a twin cohort (47) the presence of polycystic ovaries on ultrasound was associated with a higher score on the bulimia investigation test-Edinburg (BITE) questionnaire. Using the same questionnaire, McCluskey et al. (48) compared women with PCOS with women suffering from organic endocrin-opathy. They found that women with PCOS had higher scores on the BITE questionnaire and that about 6% of them had scores suggestive of BN compared with only 1% in the control group. McCluskey also found that among patients suffering from BN, 56% had menstrual irregularities and 76% had polycystic ovaries on ultrasound (49). Moreover, when patients with BN who had polycystic ovaries on ultrasound were successfully treated, they normalized their ovarian morphology (50).

In contrast to those findings, Michelmore et al. used the Eating Disorders Examination questionnaire in a group of young women (18-25 years old) who volunteered for a study on women's health and did not find any association between binge eating or overeating and polycystic ovaries on ultrasound or PCOS (defined as polycystic ovaries on ultrasound plus one feature of the syndrome) (51).

In summary, binge eating might be related to the development of polycystic ovaries or even PCOS, but more data are needed to confirm such an association. Also, the mechanisms by which periods of binge eating could induce PCOS are not known. It is possible that such eating behavior causes a physical or psychological stress that induces insulin resistance independent of body weight, which in turn contributes to PCOS.

2.6. Stress and PCOS

Psychological stress has been associated with PCOS in a few studies. Lobo et al. (52) showed that women with PCOS had higher scores on the Life Events Inventory questionnaire and had higher numbers of major life events. They also demonstrated that women with PCOS had higher urinary levels of norepinephrine metabolites, which is a reflection of norepinephrine turnover and, according to the authors, a surrogate of psychological stress. Trent et al. (53) investigated the quality of life of adolescent girls with PCOS using the Child Health Questionnaire-Child Self-Report Form. They found that teenagers with PCOS scored lower on general health perception, physical functioning, general behavior, and limitations in family activities compared to healthy age-matched adolescents.

Based on these limited number of studies, it is reasonable to conclude that women with PCOS usually have a higher level of stress. However, because all the evidence is cross-sectional, it is impossible to assume any causal relationship, that is, is it the stress that increases the risk of PCOS or simply that women with PCOS suffer more stress because of their condition. In fact, McCook et al. (54) used the Health-Related Quality of Life Questionnaire adapted for women with PCOS to evaluate the impact of overweight, hyperandrogenism, menstrual problems, and infertility on the quality of life of women with PCOS. They found that weight was the major concern for those women and that this level of distress was positively correlated to BMI. The other specific categories of distress (infertility, body hair) were also correlated with each respective clinical manifestation.


Twin studies have shown that genetic factors alone cannot adequately explain the development of PCOS in women. Many environmental factors probably modulate the clinical expression of genetic predisposition. Obesity is the most important of these factors and contributes directly to insulin resistance and hyperinsulinemia, two key factors in the pathogenesis of PCOS. We have reviewed the most important environmental and neonatal predictors of obesity and insulin resistance, which include low birthweight. Low birthweight seems also to modify the relation between insulin resistance and PCOS in that girls with premature pubarche and hyperinsulinism are at higher risk of developing PCOS if they were born with a lower weight. Finally, diet composition, eating disorders, and psychological stress have been associated with PCOS, but available studies are conflicting and not conclusive for causality.


Many environmental factors have been associated with the development of PCOS. Some of them might be causal and others might significantly modify the relationship between causal factors and the clinical development of PCOS. However, most of the studies reviewed in this chapter were cross-sectional and only defined associations. In order to critically assess the causal or modifying effect of environmental factors in PCOS, more long-term and large prospective studies are needed. These studies should focus on the relationship between postpurbertal development of PCOS and birthweight, prepubertal progression of weight, characteristics of puberty, stress or stress management, eating behaviors, and evolution of insulin resistance. Finally, long-term intervention studies might help elucidate the possible causal effect of diet composition in the phenotypic expression of PCOS.

Breaking Bulimia

Breaking Bulimia

We have all been there: turning to the refrigerator if feeling lonely or bored or indulging in seconds or thirds if strained. But if you suffer from bulimia, the from time to time urge to overeat is more like an obsession.

Get My Free Ebook

Post a comment