Kenneth D. Burman, M.D., M.A.C.P. Chief, Endocrine Section,Washington Hospital Center
Program Director, Endocrine Fellowship,Georgetown University Hospital / Washington Hospital Center
Professor, Department of Medicine,Georgetown University.

Yasser H. Ousman, M.D. Associate Director, Diabetes Team., Staff, Endocrine Section, Washington Hospital Center.

Madhuri R. Devdhar, M.D. Fellow, Endocrinology and Metabolism, Georgetown University Hospital, Washington Hospital Center.

Contents
Contributors
Search

INTRODUCTION

This chapter will discuss the endocrine changes or alterations that can occur in association with obesity. We will discuss the endocrine role(s) of adipose tissue, alterations in the hypothalamic-pituitary-endocrine axis, and the possible changes that can occur in each of these axes in association with obesity.

It is important to emphasize several issues that are relevant. The first question is whether the endocrine changes that can be seen in obesity are secondary to obesity or could, some of them at least, be a contributive factor to the development and/or perpetuation of obesity.

The endocrine axes of the human body are dynamic systems; they frequently show changes in response to stress, disease or other pathological states. For example, changes in thyroid function tests, testosterone, and growth hormone concentrations are commonly seen in response to acute and chronic illnesses, and starvation. These alterations are typically transient and resolve as the subject recovers from the offending insult. These changes are, therefore, thought to be secondary and reflective of a homeostatic mechanism. It is, however, also important to mention that these "adaptive" changes in hormonal dynamics may not necessarily be appropriate and that therapeutic measures aimed at restoring "normal" serum level of perturbed hormones may hasten recovery and improve patient outcome. So far though, such interventions have not shown to be successful.

Obesity is medically defined as an increase in body mass index (BMI). Different phenotypes of obesity may have variable health implications. Abdominal or visceral obesity is considered to be a more hazardous condition than gluteofemoral or gynecoid obesity. Some also distinguish between subcutaneous abdominal obesity where most of the abdominal fat is under the skin and visceral abdominal obesity where the majority of the fat is in the omentum. The latter has even more health implications than the former. Therefore, when discussing complications of and metabolic abnormalities associated with obesity, one has to remember that different phenotypes of obesity exist and may carry different degrees of risk, in particular cardiovascular risk.

Our understanding of the physiology of adipose tissue has improved in the last decade and extensive research has been dedicated to the study of the interactions between the adipose tissue and other body systems, in particular the central nervous system. New hormones and substances with important potential roles in energy balance and food intake have been discovered and our understanding of their action continues to expand. Some of the newly discovered hormones have not had their role fully elucidated in humans, but the future holds promise in not only improving our knowledge of the pathophysiology of obesity but also in developing novel therapeutic approaches to complement our rather limited pharmacological arsenal.

ENDOCRINE FUNCTION OF THE ADIPOSE TISSUE

Adipose tissue has an important role in storing energy in the form of fat for subsequent use by various body organs. Lipid metabolism depends on the integration of a number of signals that include hormones, neuronal pathways and nutrients. The details of how these different signals influence fat metabolism is beyond the scope of this chapter. Besides its role in energy storage, adipose tissue has many other important functions that can be mediated through hormones or substances synthesized and released by adipocytes. These substances are called “Adipocytokines” and they are capable of acting on distant targets in an endocrine fashion or act locally in a paracrine or autocrine fashion. In the following paragraphs, we shall discuss a few of the important adipocytokines in detail.

The hormone leptin is probably the most important adipocytokines because much is known about its actions and its discovery has greatly influenced our understanding of how the adipose tissue "communicates" with other systems in the body, in particular with the central nervous system(1,2). Leptin is clearly a major link between fat and not only the CNS but also other body components (3). The physiologic roles of leptin are numerous; it has an important role as an anti-obesity hormone (4). In fact, a study has shown that obesity is associated with decreasing levels of circulating soluble leptin receptors (SLR), which in turn leads to decreased action of leptin. (5) The mechanism of leptin resistance is not yet fully understood. Decreased transport across the blood-brain barrier and also a decreased ability of leptin to activate hypothalamic signaling in diet-induced obesity may be crucial mediators in the pathogenesis of leptin resistance. (6, 7)This postulated leptin resistance is one major target in the search for a better understanding of obesity and the development of pharmacological tools to treat this spreading disease (8). Leptin also has other important metabolic effects. It has been shown to be implicated in glucose and lipid metabolism (4,9,10), although there are some studies showing that leptin corrected by fat weight did not influence lipid profile and glycemia (11). Leptin has an important role in reproduction. There are observations which suggest the ability of leptin to control directly human ovarian cell functions including proliferation and apoptosis (12).It can also exert actions on the different hypothalamic-pituitary-endocrine axes (4,9,10). Leptin is also formed in the placenta and is widely expressed in the fetal tissues. It stimulates hematogenesis and angiogenesis indicating a possible role in development. Leptin may also be involved in brain development as evidenced by the decreased brain weight and impaired myelination in mice with leptin deficiency (4,9,10). It has also been shown to influence the respiratory drive and hypercapneic response, irrespective of the amount of body fat (13).Leptin increases sympathetic nerve activity in the brown adipose tissue, adrenal gland, kidney, and hindlimb skeletal muscle. It is speculated, therefore, that leptin may have a role in the etiology of hypertension, cardiovascular, and renal changes associated with obesity (14, 15). Leptin has been shown to be an independent predictor of future cardiovascular events despite other risk factors such as C-Reactive Protein (CRP) and lipid status (16-18). Finally, leptin has been implicated in the pathogenesis of breast cancer, gynecological cancer, gastrointestinal cancer, leukemia, prostate cancer and cancer anorexia (19-21). In fact, insulin resistance in obesity has been implicated as the culprit for increased cancer risk and has been postulated to act through molecular mechanisms involving leptin and other adipocytokines (22). Leptin is also known to stimulate colon cancer proliferation via COX-independent pathways (23).

Adiponectin is another important adipocytokine which has shown to have a protective role in the pathogenesis of obesity-related disorders. Adiponectin has been shown to suppress the action of inflammatory cytokines like tumor necrosis factor alpha (TNF-alpha) in the atherogenic process (24). It exerts other beneficial effects by lowering dyslipidemia and other risk factors of cardiovascular disease, and also by markedly improving insulin resistance (25-28). Adiponectin is one of the most important adipocytokines implicated in metabolic syndrome and also in its therapeutic options (28). In addition to an anti-atherogenic effect, adiponectin also has a variety of anti-tumor effects. Adiponectin has been shown to slow the progression of growth of tumors and low (serum) levels of adiponectin have been seen to be associated with a number of cancers including breast, gastric and prostate cancer (29-31).

Resistin is an adipocytokine which has been studied closely. Resistin has been linked to obesity, type 2 diabetes, inflammation and atherosclerosis but the results of animal and human studies have been at variance. Some studies show that resistin may represent a link between obesity and insulin resistance via pro-inflammatory pathways. Resistin has also been shown to play a regulatory role in some cancers including breast cancer and also in Polycystic ovarian syndrome (29).

Visfatin, also called pre-B-cell colony enhancing factors is a newer adipocytokine discovered in 2004 and is worthy of mention. It has been shown to exert insulin-mimetic effects. It has also been implicated in angiogenesis (24,25). The proposed role of visfatin in obesity-related disorders still largely remains unknown with a large potential for future research (25).

The adipose tissue secretes a number of other substances as well, which can have systemic or local actions. Actually, morbid obesity is associated with low-grade systemic inflammation and immune activation. These include cytokines such as TNF-alpha, which may play a role in fat metabolism and insulin resistance (32-34). Adipocytes also produce IL-6 (34,35), which in turn stimulates C-reactive protein. C-reactive protein, a systemic marker of inflammation, is increased in obesity. IL-6 and its subsequent inflammation has been postulated to play an etiologic role in the increased risk of thromboembolism observed in obese patients. The adipose tissue is capable of producing plasminogen activator inhibitor-1 (PAI-1) (36). Several components of the renin-angiotensin system (renin, angiotensinogen, angiotensin-converting enzyme and angiotensin II receptors) are expressed by the adipose tissue (37,38). Elevations in the levels of PAI-1 and angiotensin II could explain at least partly the increased cardiovascular risk associated with obesity. Other proteins such as complement factors, adipsin, apelin, pten that are expressed by fat tissue may also have a role in the pathophysiology or the progression of coronary artery disease and type 2 diabetes (39-44, 45). Adipokines have been implicated in the progression of established prostate cancer and also in breast cancer as mentioned earlier (46,47). Also, human adipocytes have been identified as a new source of endocannabinoids and related compounds, although their potential implication in obesity deserves further investigations (48). Endocannabinoids by themselves have been known to regulate food intake and obesity through central mechanisms.

Finally, adipocytes can metabolize sex steroids and glucocorticoids (49). Adipocytes convert androgens to estrogens under the influence of the enzyme aromatase. Another hormone present in the adipose tissue, 11-beta hydroxysteroid dehydrogenase-1 (11beta-HSD1), converts cortisone to cortisol and its overactivity in obesity has been suggested to result in increased local concentrations of cortisol in the adipose tissue (50).

The adipose tissue is a target for the action of sex steroids and cortisol through interaction with their nuclear receptors present in adipocytes (50-55).

In conclusion, it is clear that the adipose tissue is an active organ with multiple roles. The manner in which these roles are performed and their contribution to the health or risk of disease of the human body will likely be elucidated as more discoveries continue to shed light on the mechanism of the complex interaction between adipocytes and body tissues.

Obesity and Sex Hormones

Not only is obesity associated with the development of alterations in the sex hormones (l), it has recently become evident that they may conversely play a very interesting role in the development of different phenotypes of obesity (56). An interesting fact is that obesity, especially if massive, is associated with a state of hyperandrogenism in women in stark contrast to a state of hypoandrogenism in men. The most important factors responsible for altered androgen status in male and female obesity or SHBG(Sex hormone binding globulin) and Gonadrotrophins (57).

Testosterone and SHBG:

Most of the circulating testosterone is bound to transport proteins; Sex Hormone Binding Globulin (SHBG), and albumin. SHBG is produced by the liver. Only about 2% of circulating testosterone is unbound or free, and this is thought to represent the vast majority of the bioactive fraction of the hormone. Some believe that at least part of the bound testosterone is readily available for use by the body target cells. Total testosterone, therefore, reflects the bound and unbound hormone and is greatly dependent on the serum concentration of SHBG. Table 1, shows some common conditions that can affect the serum concentration of SHBG:

Table 1.

Obesity and androgens in men:

Testosterone should be measured in the morning when its serum concentration is at its peak and we recommend repeating an abnormal measurement for confirmation. It is an unresolved question whether total testosterone, free testosterone, or both should be routinely measured. In most patients, we prefer to assess both free and total.

Obesity itself is one of several conditions that can result in a low SHBG. As a result total testosterone is frequently low normal or below the normal range in obese men (58). The free testosterone and bioavailable testosterone (total testosterone minus SHBG bound testosterone), however, are generally normal in this situation. Men with massive obesity can also manifest low levels of free testosterone (59). The reduced testosterone levels in obese males appear to be caused by several complementary factors, including reduced gonadotropin secretion (as explained below) and the balanced effects of insulin. Studies have shown that insulin plays a significant role by its effects on SHBG (inhibition) (60) and testosterone itself (stimulation⁾ (61,62).These changes in total and free testosterone concentrations are reversible with weight loss indicating that they are secondary to obesity (63, 183)^.The reduction in free testosterone seen in massive obesity is not accompanied by a reciprocal increase in luteinizing hormone (LH) suggesting a central rather than gonadal mechanism behind this change (63,64).

In obese patients, the serum concentration of pituitary gonadotropins, LH and FSH, is usually normal, but may be reduced especially in massive obesity.

The response of gonadotropins to gonadotropin releasing hormone, GnRH, or to clomiphene was reported to be normal in previous studies indicating a normal hypothalamic-pituitary responsiveness to stimuli (65,66).

Serum levels of testosterone rise in response to human chorionic gonadotropin (HCG), administration which suggests normal gonadal function in obese individuals. In a study by Isidori et al, however, the response of testosterone to LH/HCG stimulation in obese men was reduced in comparison to the response in nonobese men (67). Testosterone response to LH/HCG stimulation correlated negatively with serum leptin levels. The authors suggested that elevated levels of leptin may contribute to the development of reduced androgens in male obesity. Another study showed that obese men with high BMI values may show a reduction of LH secretion, probably due to impaired secretion of GnRH at hypothalamic level, producing a reduced LH secretory mass per secretory burst without any change in burst number (68).

Some studies found increased serum levels of estradiol and estrone in obese men (66,67,69). One hypothesis suggests this may result from increased peripheral conversion of androgens to estrogens through the action of the enzyme aromatase present in the adipose tissue (70). Overall, these data appear to support the concept that hypotestostronemia in obese men may be partly due to feedback suppression of the pituitary by the elevated estrogen, which may be associated with an increased responsiveness of pituitary gonadotropin secretion to stimulation by GnRH (71). However, the concept that increased estrogen availability in human obesity may play a role in determining abnormalities of androgens contradicts the emerging role of estrogens in the regulation of adipose tissue (72,73). Further studies are warranted in this regard.

Fewer data are available on the effect of obesity on sperm count and fertility in men. Studies have not shown impairment of fertility or sperm count in relation to obesity. However, these parameters have been described as being reduced in subjects with massive obesity (74). This is explained by the fact that increased estrogen levels in very obese men partially suppress the pituitary and cause mild hypogonadotrophic hypogonadism (74).

Several randomized studies assessed the impact of testosterone administration in elderly men with low normal testosterone values (levels usually in the range of 250 to 400 ng/dl) on several metabolic and anthropometric measures. Testosterone given intramuscularly or topically results in mild but significant increase in lean body mass and reductions in fat mass and may have a favorable effect on glucose homeostasis (75). The impact of such an intervention on mortality and cardiovascular event rates is not known. This type of therapy is not approved by the FDA for the treatment of obesity. Careful attention must be given to the causes and significance of a low serum total and free testosterone in an obese patient and treatment options must be carefully considered. Rarely would exogenous testosterone therapy be indicated and this would only be when a definable etiology can be specifically identified.

In conclusion, obesity can result in low androgen levels due to reduced concentrations of SHBG as well as changes in gonadotrophins.The total testosterone levels are low due to reduced levels of SHBG. The free testosterone is generally normal except in some men with massive obesity in whom it may be reduced. These changes are considered secondary to obesity and do not by themselves indicate hypogonadism nor the need for testosterone supplementation. When evaluating men for erectile dysfunction, the same observations apply whether or not obesity exists; hypogonadism is not a common cause for erectile dysfunction, which often has a vasculogenic etiology. Therefore, one should not necessarily conclude that an obese man with erectile dysfunction and a low total but normal free and bioavailable testosterone has hypogonadism as a cause for his erectile difficulties. Testosterone administration in these circumstances rarely improves erectile function. Interpretation of serum testosterone levels must always be placed in the corresponding clinical context and the decision to utilize testosterone therapy has to be made carefully.

Obesity and Androgens in Women:

It is well known that an increase in body weight and fat tissue is associated with several abnormalities of sex steroid balance in both premenopausal and postmenopausal women. In fact, it has been shown that women with central obesity, especially have higher circulating androgen levels (1). This has been seen to be due to lower circulating levels of SHBG which is at least partially dependant on higher circulating insulin in

abdominally obese women and on the inhibiting capacity of insulin on SHBG liver synthesis (76-78).

Androgens in obesity in the context of HPA(Hypothalamic-Pituitary Adrenal )

axis:

It has been shown that that hyperactivity of the HPA axis may lead to an inhibition of the HPG (hypothalamic pituitary gonadal) axis. In contrast, an increase of the HPA axis activity appears to favor increased androgenization in obese women, although the intricate mechanisms are yet to be studied.

Relationship between leptin and sex hormones:

Leptin has been shown to participate in the regulation of HPG axis at all levels and thereby in the regulation of sex hormones. Leptin appears to facilitate GnRH secretion indirectly by modulating several interneuron secretory neuropeptides (79), regulates gonadotrophin secretion by the increased expression of leptin receptors on the hypothalamus (80), directly stimulate LH and, to a lesser extent, FSH, by mechanisms probably involving nitric oxide activation in the pituitary gonadotropes (81). Leptin receptors are also widely expressed in the human ovaries (82) and testes (83)indicating a regulatory role. Another important hormone in obesity is Ghrelin. Studies have shown that androgens have a negative role on ghrelin. On the other hand, how ghrelin would affect sex hormones in the setting of obesity is still under investigation.

Androgen receptor and obesity

There is an interaction between androgen receptor polymorphisms, like the number of CAG repeats, and obesity and also different obesity phenotypes. Whether these polymorphisms play a role in the development of different obesity phenotypes remains largely unknown (84,85).

Androgens and metabolic syndrome:

The current weight of evidence suggests that low endogeneous testosterone is may be responsible as an etiologic factor in the development of obesity, insulin resistance and cardiovascular risk factors in men whereas in women androgen excess and altered insulin action have a complimentary effect in the development of metabolic syndrome (86,87).

Obesity and ovaries(PCOS)

Polycystic ovarian syndrome(PCOS) is a condition which is defined by the presence of both obesity and hyperandrogenism. Hence, this disorder has been studied extensively in the context of interactions between sex hormones, sexes and obesity. Nearly 50% of women with PCOS are obese (88-91). Hyperandrogenism and anovulation are the hallmarks of PCOS. Women with PCOS also have insulin resistance and are at a high risk for developing impaired glucose tolerance and diabetes mellitus

As in men, obesity in women is associated with decreased levels of Sex Hormone Binding Globulin (SHBG) but here the concentration of free testosterone is increased (76, 92). The increase in free testosterone concentration is usually moderate but some women with PCOS can have much higher testosterone levels which need to be differentiated from other causes of hyperandrogenism in women, such as androgen producing ovarian tumors.

The precise etiology of PCOS is not known but insulin resistance and hyperinsulinemia are thought to play an important role. Hyperinsulinemia decreases the levels of SHBG and increases the ovarian production of androgens (93,94). Therefore, studies have shown that administration of metformin, an insulin sensitizer, to women with PCOS may reduce free testosterone and increases SHBG levels by exerting its action on hyperinsulinemia (95-97).

Anovulation and menstrual irregularities are also major features of PCOS. Again, metformin appears to result in improved fertility in women with the syndrome, implicating hyperinsulinemia as a direct or indirect cause for the menstrual irregularities (98,99). These improvement in fertility and hyperandrogenism can also be achieved with weight loss. The use of metformin therapy in this situation must be evaluated carefully, in consultation with the patient with discussion of pros and cons. This drug is not approved by the FDA for this indication, although it is approved by FDA for the treatment of Diabetes mellitus.

The use of anti-androgens has also been widely implicated in the management of PCOS. Cyproterone acetate is an anti-androgen which exerts its action by increasing the clearance of testosterone, decreasing ovarian androgen secretion and also causes a slight decrease in both HDL and LDL cholesterol levels (100,101). Some studies have shown that other anti-androgens like spirinolactone, flutamide and finasteride not only have an androgen suppressive effect but also have been implicated in improving insulin resistance and reducing hyperinsulinemia both in obese and non-obese women with PCOS (102).

As more and more of the biochemical mechanisms come to light, the treatment of PCOS with a combination of insulin sensitizers and anti-androgens has been emphasized. The most popular anti-androgen being used is flutamide. This compound specifically antagonizes androgen receptors and inhibits androgen biosynthesis (103,104). Anti-androgens and insulin sensitizer combined have been shown to have a synergistic effect in reducing testosterone concentrations, increasing HDL cholesterol and SHBG levels (105).

Obesity in women is also accompanied with enhanced estrogen production as a result of increased aromatization of adrenal androgens into estrogens in the adipose tissue (106, 183). Obesity is a known risk factor for endometrial cancer. Given the lower levels of SHBG seen in obese women, free estradiol levels are elevated and there is increased conversion of estradiol to estriol along with decreased conversion of estradiol to 2-hydroxyestradiol (107).

Gonadotropin levels are normal in obese premenopausal women. In women with PCOS, an increased ratio of serum LH to FSH may be seen (108,109).

Leptin levels are higher in obese women with PCOS than in normal weight women with PCOS. Higher leptin levels in obese women with PCOS are inversely correlated with basal LH levels (110). Leptin may have a modulatory effect on GnRH. Other studies in normal weight women with PCOS found a significant correlation between insulin resistance compared with free testosterone and LH.

In conclusion, obesity is a common feature of PCOS. Hyperinsulinemia is believed to be the main etiological factor behind the development of PCOS. Obesity also leads to hyperestrogenism. Weight loss and/or use of insulin sensitizing agents (mainly metformin) improve insulin sensitivity, reduce insulin levels and improve fertility in women with PCOS. Insulin sensitizers and anti-androgens have an additive effect.

THYROID

More than 99% of T4 and T3 circulate bound to transport proteins. Only a very small, less than 1%, amount of thyroid hormone is unbound or free and represents the biologically active fraction of the hormone. Thyroxin Binding Globulin (TBG) is the major transport protein. The serum concentration of TBG is influenced by several conditions (Table 2) which can result in a significant increase or decrease in total T4 concentration. Therefore, when evaluating thyroid function we measure TSH and free T4. Free T3 can also be measured in selected circumstances, such as hyperthyroidism.

Table 2.

Given the important role of thyroid hormones in the regulation of thermogenesis, it is not surprising that the dynamics of the hypothalamic-pituitary-thyroid axis in obesity have been studied in detail.

In obesity, the serum concentration of T4, T3 and thyroid stimulating hormone, TSH are normal if the subject is in a neutral energy balance state (111-113). Some have reported elevations in serum T3 in obese subjects but this has not been confirmed and is likely related to enhanced carbohydrate intake (114). An important point is whether the subject(s) were in positive energy balance state. Increased caloric intake can, in fact, result in increased T3 and decreased reverse T3 concentrations (115,116). Inversely, caloric restriction in obese subjects results in a low T3 and increased reverse T3 concentrations (116-119). These changes in T3 and rT3 levels normalize upon achievement of a steady state from an energy balance standpoint. Therefore, when evaluating thyroid function in a patient with obesity, it is important to consider this possibility.

Some recent studies have interestingly shown that in morbidly obese individuals, not only is the prevalence of overt and subclinical hypothyroidism high, these subjects have higher levels of T3, FT3, T4, and TSH, probably the result of the reset of their central thyrostat at higher level (120). It has been shown that even slightly elevated serum TSH levels are associated with an increase in the occurrence of obesity (121). There has been shown to be a positive association between TSH, BMI and leptin levels (122). On the contrary, no association between thyroid status within the normal range and BMI was noted in another study (123).

The TSH response to Thyrotropin Releasing Hormone, TRH, has been found to be increased in patients with obesity although this was not significantly different from control subjects (118,119,124,125). Obese subjects in whom TSH values are slightly elevated but who have normal T4 and T3 concentrations display increased TSH response to TRH in comparison to obese subjects with normal TSH. In this case the exaggerated TSH response to TRH probably reflects a mild degree of thyroid failure or "subclinical hypothyroidism" which is not an uncommon finding particularly in elderly women. Subclinical hypothyroidism, however, is not more common in obese subjects than in subjects of normal weight. Treatment of subclinical hypothyroidism does not result in significant weight loss. This is important as some patients with obesity may attribute their failure to lose weight to "slow metabolism" and not infrequently the thyroid gland is looked at as the culprit, even when the thyroid function tests are perfectly normal.

As mentioned above, T3 levels fall during fasting or caloric restriction. This fall in T3 results from decreased T3 production. Simultaneously reverse T3 levels increase due to decreased reverse T3 clearance. Overfeeding has the opposite effects with resulting increase in T3 and decrease in reverse T3.No effects are seen on T4 in these situations. These changes in T3 levels are not related to the adipose tissue mass since they also occur in normal weight individuals, but are thought to be related to dietary components. Burman KD etal studied the effect of prolonged TRH infusion on TSH secretion as well as TT3 and reverse T3 in 9 euthyroid obese subjects. This study revealed that TSH responsiveness to prolonged TRH infusion is diminished during fasting and does not return to control (fed) values despite 5 days of refeeding a 1500 Kcal diet. Also net T3 increases observed during the TRH infusion are greater in the fed period than in the fasting or refeeding periods. Moreover, 5 days of refeeding a 1500 Kcal diet (40% carbohydrate, 40% fat, 20% protein) did not return the T3 to its original fed value whereas rT3 which was increased in the fasting state was completely restored to control values.(125a)

In conclusion, the function of the hypothalamic-pituitary-thyroid axis in obesity can be essentially normal. Overfeeding and caloric restriction can affect the concentrations of T3 and reverse T3 with return to normal when a steady state of the caloric balance has been reached.

Another important point to be noted here is the relationship between TSH and lipid concentrations. It is a well-known and proven fact that hypothyroidism is associated with hyperlipidemia. A recent study showed that within the range of TSH that is considered clinically normal, increasing level of TSH was associated with less favorable lipid concentrations (126). Thus there is a linear association between subclinical hypothyroidisma and hyperlipidemia as well (127).

Synthetic thyroid hormones as well as various other thyroid hormone preparations have been and are being used as adjunctive measures to induce or facilitate weight loss. It is clear that supraphysiologic doses of thyroid hormones are not only associated with loss of fat, but also lean tissue. Some authors have suggested that small doses of T3 (20 ucg a day) may be helpful when used in conjunction with very low caloric diets (128-130). The argument is that such low doses can facilitate weight loss, ameliorate hyperlipidemia, prevent hypoglycemic episodes while on very low caloric diets. These doses have been shown to improve hormonal profile, but without a major influence on either body weight or basal metabolic rate in small studies. This approach is controversial and most societies have argued against this approach. Moreover, there are no large randomized placebo-controlled trials evaluating the efficacy and side-effects of low dose T3 treatment for obesity. The long term side effects of such therapy relate mainly to cardiac and osseous events and, therefore we do not recommend such an intervention.

The other school of thought is based on the presence of two subtypes of thyroid hormone receptors. TRalpha and TRbeta. Selective TRbeta activation causes a modest increase in the metabolic rate and cholesterol reduction without the side effect of cardiac acceleration which can be seen with other non-selective thyroid preparations. .Studies are being performed with one these selective TRbeta agents(KB-141) and they have evolved to be potentially promising as adjunctive therapy in obesity (131-134).

ADRENAL GLANDS

Cortisol is mainly bound to Cortisol-Binding Globulin (CGB or transcortin) and less to albumin. About 10% of circulating cortisol is free or unbound and this fraction represents the bioactive portion of the hormone. CBG concentration can be increased or decreased in a number of conditions (Table 3).

Table 3.

The dynamics of the hypothalamic-pituitary-adrenal (HPA) axis in obesity have been examined. Patients with Cushing's syndrome display a number of clinical features that resemble those seen in patients with the metabolic syndrome or syndrome X. These features include abdominal obesity, insulin resistance, impaired glucose homeostasis, hypertension, and lipid abnormalities. These similarities led to the hypothesis that a dysregulation of the HPA axis in the form of "functional hypercortisolism" could potentially be a cause for abdominal obesity and its different metabolic consequences.

The serum concentration of cortisol is essentially normal in obesity. Some have reported a decrease in the morning cortisol in obese subjects as compared with normal weight subjects (135-137). The 24 hour urine free cortisol excretion is also normal or sometimes mildly elevated in obesity. If complicated by conditions such as depression and/or alcoholism it may also be slightly elevated (136).

Adipose tissue is involved in the metabolism of cortisol. The enzyme 11 Beta-hydroxysteroid dehydrogenase-1 (11beta-HSD-1) which converts cortisone to cortisol, is expressed in adipose tissue (138). It appears that in obesity more cortisol is derived from cortisone due to the increased activity of this hormone, which could simply be due to increased fat mass (139). Urine studies in obesity also show an increase in the ratio of tetrahydrocortisol to tetrahydrocortisone indicating an relative increase in the pathways leading to cortisol formation (139)(through the action of 11Beta-HSD-1). Therefore, in obesity, 11HSD1 activity in subcutaneous adipose tissue is increased, putatively explaining the link between obesity and other features of the metabolic syndrome. Against this background, inhibition of 11HSD1 has become a major therapeutic target in metabolic syndrome (140,141).

Since serum cortisol is not increased in obesity, it is possible that there is an increase in the local production of cortisol in the fat tissue and this in turn could lead to increased local action of cortisol with the subsequent metabolic consequences known to occur in obesity. This hypothesis is purely speculative at present.

The cortisol response to a variety of stimuli such as corticotropin releasing hormone, CRH, or meal ingestion is increased in obesity (142). An increase in cortisol response to ACTH has also been seen in obese men and obese premenopausal women, but not in obese postmenopausal women, indicating a possible role for estrogens in that response (143).

ACTH secretion in obesity may be altered with increased pulsatile frequency but decreased pulse amplitudes. It is not clear whether this has an influence on cortisol release from the adrenal glands (142).

Dexamethasone is used in different dosages to study the suppressibilty of cortisol secretion in the diagnosis of Cushing's syndrome. Obese individuals may have decreased suppression with small doses of dexamethasone (0.5 or 1mg) but they do typically suppress normally with higher doses (143).

Pasquali et al, summarized the main abnormalities of the hypothalamic-pituitary-adrenal axis seen in abdominal/visceral obesity phenotype as follows (144):

Table 4.

In fact, no single marker has the power to detect subtle alterations of the HPA axis in conditions, such as the abdominal obesity and the metabolic syndrome. At present, evaluation of 24 hour urinary free cortisol, , and salivary-free cortisol appear to be promising for these purposes, whereas dynamic tests should be reserved for specific clinical settings (145).

These subtle alterations in the HPA axis could result from peripheral and/or central mechanisms. We have previously mentioned the action of 11Beta-HSD-1 in the fat tissue and subsequent increase in local cortisol production. This, however, does not lead to increased total cortisol and should not be expected to inhibit the central release of ACTH and or CRH. There could be a central defect in the glucocorticoid receptor sensitivity to cortisol. The glucocorticoid receptor is located in the hippocampus. Alternatively, there could be a decrease in the HPA sensitivity to the inhibitory effect of the sympathetic nervous system which is exerted through alpha-2 adrenergic receptors leading to a state of HPA axis "excitation" (146).

In a recent cross-sectional designed study, Björntorp et al, assessed HPA axis function through measurements of salivary cortisol, which is a reasonable reflection of serum free cortisol, and dexamethasone suppression test. They also examined a number of other parameters such as BMI, Waist to hip ratio, testosterone, IGF-1 and lipids (147). The authors found that there was an association between increased BMI, increased waist to hip ratio, low testosterone, low IGF-1, abnormal glucose metabolism, increased insulin, low HDL, and elevated triglycerides, all of which are features of the metabolic syndrome. An abnormal HPA axis dynamic characterized by low morning cortisol, decreased cortisol response to feeding and inadequate dexamethasone inhibition was also seen. The authors concluded that a primary abnormality of the HPA axis is likely to result in visceral fat accumulation and the resultant endocrine and metabolic changes that may be associated. It would be interesting to reevaluate these subjects following weight loss. The question here again is whether the observed HPA abnormalities in this subgroup of obese men reflect an initial alteration in the HPA axis that might have been brought forward or exaggerated due to some genetic factor(s), or whether they represent bystander effects that happen to accompany visceral obesity as it develops.

A decrease in the mineralocorticoid receptor response to circulating corticosteroids was suggested by Jessop et al as an explanation for the decreased HPA axis to glucocorticoid feedback in obesity (148). Apart from this, a more recent study has shown that mineralocorticoid receptor represents an important proadipogenic transcription factor that may mediate both aldosterone and glucocorticoid effects on adipose tissue development. Mineralocorticoid receptor thus may be of pathophysiological relevance to the development of obesity and the metabolic syndrome (149).

In conclusions, obesity is accompanied by alterations in the HPA axis characterized mainly by an increased sensitivity to stimuli and decreased sensitivity to inhibition. An increase in the local production of cortisol in the adipose tissue, abnormal central control of the HPA axis by the sympathetic nervous system, and abnormal glucocorticoid-receptor response may all contribute to the altered HPA axis described in obesity. From the general internist or the clinical endocrinologist standpoint, however, there is little direct clinical impact of such findings. The main clinical question here is whether the patient with obesity has a subtle form of Cushing's syndrome. The laboratory tests commonly used by practitioners to rule out Cushing's syndrome, such as urine free cortisol and salivary cortisol are not increased in obesity. Occasionally a patient with a pseudo-Cushing's may present with slight elevations of the UFC and abnormal suppression with the 1mg overnight dexamethasone test. For these patients a combined CRH/Dexamethasone diagnostic test can be used. This test is relatively accurate in distinguishing pseudo-Cushing's from true Cushing's syndrome.

GROWTH HORMONE

Growth hormone (GH) is secreted by the pituitary gland. Most of GH promoting effects are mediated by Insulin- like Growth Factor-1 (IGF-1). GH also has effects independent of IGF-1. Serum IGF-1 concentration represents the most accurate reflection of growth hormone biologic activity. The liver is the major but not exclusive source of IGF-1.

About 50% of circulating growth hormone is bound to binding proteins. These include a high affinity Growth Hormone Binding Protein (GHBP) which actually represents the extracellular portion of the GH receptor.

IGFs are mostly bound to IGF- Binding Proteins (IGFBPs). IGF-1 is bound to IGFBP3.

Several conditions may affect the concentration of GH (Table 4). Obesity is typically accompanied by a decrease in GH levels and increase in GHBP levels. This is the opposite picture to starvation in which GH levels are increased and GHBP levels decreased. An inverse relation exists between GH levels and percent fat mass. Another major determinant of GH level is age(150); GH levels fall with increasing age and recombinant human growth hormone, rhGH, has been studied as a supplement to counteract the effects of aging on body composition.

Table 5.

The reduction in GH levels in obesity is believed to result from decreased pituitary release of the hormone. In animal models of genetic obesity there is a decrease in the number of somatotroph cells in the pituitary gland. GH responses to a variety of stimuli such stimuli included GHRH, Arginine, and GHRPs (growth hormone releasing peptides) have also been shown to be reduced in obesity(151,152). Cordido et al, on the other hand, found a normal growth hormone response in obese men after combined administration of GH-releasing hormone and GHRP-6, a growth hormone releasing peptide, suggesting a normal pituitary growth hormone reserve(153). Exercise is a known physiologic stimulus of GH secretion. A recent study by Holt et al found no difference in GH response to exercise between young obese and nonobese men(154). A decreased response was seen between elderly obese and nonobese men. However, on multivariate analysis, only age and physical fitness, as reflected by oxygen consumption (VO2 max), predicted GH response. Obesity did not. These findings are in contrast to those by Kanaley et al who examined GH response to exercise in women with and without obesity(155). They found a significantly higher GH response in the group of nonobese women compared to the group of obese ones. The increased response was attributed to increased GH pulse amplitude. Interestingly, after 16 weeks of aerobic training, the GH response did not improve in the group of obese women in spite of a significant improvement in VO2 max. Weight loss is associated with improved GH response to stimuli, however there is uncertainty on how much weight loss is required to completely normalize GH secretion (156,157).

In spite of the reduced GH levels seen in obesity, IGF-1levels do not appear to be significantly different between obese and non obese subjects (158,159).

Individuals with visceral obesity, even in the absence of total body obesity, may, on the other hand, have lower IGF-1 levels than comparable individuals with no abdominal obesity . The serum concentration of unbound or free IGF-1 is increased in obese subjects. Similarly it has been found that GHBP levels are increased in obesity. This suggests that lower levels of GH are accompanied by increased peripheral sensitivity to GH accounting for the normal IGF-1 levels. Increased peripheral sensitivity to GH in obesity is supported by data from Maccario et al. The authors found that the administration of a low dose of rhGH had an enhanced stimulatory effect on IGF-1 secretion in obese subjects compared to normal weight subjects (160). In another study the same authors showed a normal inhibitory response of the somatotroph to IGF-1 administration suggesting that the feed back between the somatotroph and IGF-1 in normal (161). Hyperinsulinemia, frequently found in the obese could help explain these findings. Insulin could result in increased peripheral sensitivity to GH, reduced IGFBP-1 levels and increased (to within normal range) IGF-1 in spite of decreased GH secretion by the somatotroph. High free IGF-1 levels in this case exert a negative feedback mechanism on GH secretion. This hypothesis supports a peripheral rather than a central mechanism by which GH release is decreased in obesity.

Other possible mechanisms included in the altered GH response in obesity are free fatty acids (FFA) and leptin, both of which are increased in obesity. Lee et al, showed that reduction in free fatty acids concentrations in obese subjects through use of Acipimox leads to increased GH response to GH-releasing hormone (162). In animals, leptin has an inhibitory role on GH secretion from the pituitary gland through its effects on GHRH and neuropeptid Y (NPY) at the hypothalamus level (163).

Finally, it is important to remember that in children, obesity is also accompanied by reduced basal and stimulated GH. Nevertheless, linear growth is normal in obese children who reach a final height comparable to that of nonobese controls.

In conclusions, obesity is accompanied by a reduction in basal and stimulated GH secretion by the pituitary gland. The reduction in GH does not appear to translate into similar reduction in IGF-1. Therefore it is unlikely that obesity represents a condition of GH deficiency that is reflected at the tissue level. The use of recombinant growth hormone in elderly and in subjects with visceral obesity has resulted in a number of mild to moderate, but significant positive anthropometric and metabolic effects such as reduced fat mass, increased lean mass, and improved lipid profile. The role of GH replacement/supplementation other than in those individuals with documented GH deficiency as assessed by specific cut off values of GH levels during GH stimulatory testing (using two separate tests), is not clear at this point. Recombinant growth hormone (rhGH) has been extensively studies in the recent era as a treatment for obesity. Studies in the past have shown that, in obesity, rhGH normalized IGF-I levels, induced loss of body weight, mainly body fat, and improved lipid profile without untoward effects on insulin sensitivity.(163a)A metaanalysis of human studies examining the efficacy and safety of recombinant human GH (rhGH) as therapy for obesity in adults was published recently in Jan 09. This indicated that rhGH therapy leads to decrease in visceral adiposity and increase in lean body mass as well as beneficial changes in lipid profile in obese adults, but without inducing weight loss. Administration of rhGH was associated with increases in fasting plasma glucose and insulinemia over shorted periods of time.(163b)The dose of rhGH used in these studies was supraphysiological. Further studiesof longer duration using closer to physiological doses of rhGH and also determining its effect on cardiovascular morbidity are warranted.

Obesity is strongly associated with insulin resistance and is a risk factor for type 2 diabetes (164). Recently available data from the diabetes prevention program (DPP) suggested that life style modifications including weight loss and exercise, and metformin, lower the risk of developing diabetes in patients with impaired fasting plasma glucose and impaired post-glucose load plasma glucose. In these individuals at high risk for developing diabetes, both life style modification and metformin use reduced the incidence of diabetes by 58% and 31% respectively as compared to no intervention. The follow up was for 2.8 years (165).

Visceral abdominal obesity is associated with increased hepatic glucose production and reduced peripheral glucose uptake. A significant correlation exists between waist to hip ratio and insulin sensitivity. Visceral abdominal fat is inversely correlated with insulin sensitivity (166,167).

Insulin levels are elevated in obese individuals and a high correlation exists between insulin secretion and BMI. The hyperinsulinemia seen in obese subjects is mainly due to increased insulin secretion perhaps related to peripheral insulin resistance. In some patients, decreased insulin clearance can also contribute to the elevated insulin levels. Insulin release in lean and obese subjects who have normal glucose homeostasis, is pulsatile. The pulse amplitude is greater in the obese. The insulin response to mixed meals is also normal in this case. When impaired glucose tolerance develops, the insulin secretory dynamics are altered with progressive decrease in the ability of the pancreas to maintain normal glucose homeostasis, initially in response to glucose load, and ultimately in the basal state (168).

The exact mechanism(s) by which obesity contributes to the development of insulin resistance is the subject of extensive research. In animal models insulin resistance itself and hyperinsulinemia can precede the development of obesity.

Substances originating in the adipose tissue and capable of exerting distant or local actions may explain the link between obesity and insulin resistance. Free fatty acids (FFA) are released in excess by abdominal fat (169,170). Increased concentrations of FFA cause insulin resistance and decreased glucose uptake by impairing insulin mediated glucose transport. Impairment at the level of the insulin receptor and insulin receptor substrate-1 (IRS-1) are likely reasons for the impaired insulin mediated glucose uptake (171).

Tumor necrosis factor-alpha (TNF-a) is another potential link between obesity and insulin resistance (172,173). Although TNF-a has been shown to alter basal and glucose stimulated insulin secretion and to produce insulin resistance in isolated cell lines, the contribution of TNF-a to insulin resistance in humans is uncertain (174,175). Similar conclusions can be drawn in regard to leptin. Leptin in vitro is capable of inhibiting insulin release but its role in the pathogenesis of insulin resistance in humans has not been confirmed (176).

Resistin, a peptide isolated from the mouse, is produced by the adipose tissue and its expression is decreased by PPAR gamma agonists (177). Serum levels of resistin are elevated in obese mice. Immunoneutralization of this peptide improves insulin sensitivity in the mouse. mRNA levels of resistin in freshly isolated human adipocytes are very low but the role of resistin in human obesity and insulin resistance is unclear at this point (178).

Ghrelin, a potent growth hormone secretagogue, is produced mainly by the stomach. Administration of ghrelin results in increased GH release, hyperglycemia, and decrease insulin concentrations (179). Tschöp et al, found a negative correlation of circulating ghrelin levels with insulin, leptin levels, and percent body fat (180). Ghrelin is another peptide which may turn out to have a role in the pathogenesis of insulin resistance accompanying obesity (181).

In conclusion, obesity, in particular, abdominal or visceral obesity, is strongly associated with insulin resistance, and increases the risk for type 2 diabetes. The link between obesity and insulin resistance continues to be explored. Weight loss as well as metformin use improve insulin resistance and lowers the risk of developing type 2 in individuals with impaired glucose tolerance. As our understanding of the interaction between obesity and insulin resistance improves we should be able to develop new therapeutic strategies to treat insulin resistance and prevent or delay the development of type 2 diabetes in subjects at a high risk for the disease.

RECOMMENDATIONS

As discussed in the previous sections, several endocrine alterations can be identified in association with obesity. In most cases, these alterations are reversible with weight loss and, therefore, they appear to be a consequence of obesity. Emphasis has been focused on the hypothalamic-pituitary-adrenal axis and the possibility that some "subclinical" alterations in this axis may be at the origin of obesity. At this stage this hypothesis needs further testing. What is true is that the interaction between the adipose tissue and the body is far more complex than once believed, and the future will certainly provide more decisive data on the precise mechanisms of these interactions and their contribution to the development and/or the maintenance of obesity.

Certain endocrine syndromes are known to result in obesity. From the clinical practitioner's perspective it is important to remember these syndromes and to be suspicious should a patient with obesity display one or more of the clinical features seen in these disorders.

Hypothyroidism is a common clinical problem and can, of course, occur in obese patients and could contribute to the presence of symptoms such as fatigue and inability to concentrate. Hypothyroidism is under diagnosed in the general population and specifically obese patients. Routine screening of patients who present with obesity with a sensitive TSH assay and free T4 is reasonable, although there are no specific guidelines with regards to this. As previously mentioned, subclinical hypothyroidism does not appear to be a contributor to obesity. There is debate on whether or not to treat subclinical hypothyroidism. The presence of anti-thyroid antibodies in these patients increases the risk of progressive thyroid failure. Treatment of subclinical hypothyroidism decreases cholesterol levels. In the presence of mildly elevated TSH and normal Free T4 and T3 concentrations, we generally recommend to repeat these tests after a few weeks and if the TSH elevation persists, to treat with thyroid hormones. The benefits and potential disadvantages of this approach should be discussed with the patient.

Cushing's syndrome is frequently included in the differential diagnosis of obesity and patients with abdominal obesity have many features in common with patients with authentic Cushing's. However, Cushing's syndrome due to excessive endogenous corticosteroids is rare. However, if there is a reasonable suspicion that it may be present, the patient should be screened. Attention should be focused on these symptoms and signs that are more specific to Cushing's such as proximal muscle weakness, purple striae, thin, bruised skin, hypokalemia, and osteopenia.. In these circumstances we recommend a 24 hour urine free cortisol (UFC) measurement. Repeatedly normal UFC over time are highly accurate in helping to rule out Cushing's syndrome. An overnight dexamethasone suppression test (1 mg) may also be helpful although false positive results can occur due to obesity. In patients with obesity and depression or alcoholism, who may have an elevated UFC, we recommend a CRH/Dexamethasone test which has good discrimination power to separate patients with "Pseudocushing's" from patients with true Cushing's. Occasionally, even the most sophisticated testing may not be sufficient to make a definitive diagnosis of Cushing's disease and it may be necessary to closely follow a patient with repeated examination. Referral to a specialist is indicated in these more difficult cases.

Insulinomas are rare neuroendocrine tumors. They typically present with symptoms and signs of hypoglycemia and perhaps weight gain. The diagnosis requires confirmation of the hypoglycemia usually through a fasting test with measurement of glucose, insulin and C-peptide. A referral to an endocrinologist is indicated.

Hypogonadism and growth hormone deficiencies can both lead to abdominal obesity. The latter is usually suspected in the setting of surgery or disease of the hypothalamus-pituitary axis, the former is not very common but should be kept in mind in males with other symptoms or signs suggestive of androgen deficiency. The treatment of these two conditions can result directly and indirectly (by improving conditioning, muscle strength, and stamina) in weight loss, improved metabolic profile, and improved bone density. An estimate of bioactive testosterone can be determined by measuring a morning free or bio-available testosterone (and total testosterone), and repeating the measurement at least once. LH should be measured if the testosterone is clearly low to distinguish between primary versus central hypogonadism. Ruling out GH deficiency requires dynamic testing and is indicated in the appropriate setting.