obesity reviews

Do stress reactions cause abdominal obesity and comorbidities?

P. Bjrntorp

Department of Heart and Lung Diseases, Sahlgrens Hospital, University of Gteborg, Sweden Received 31 August 2000; revised 30 October 2000; accepted 6 November 2000 Address reprint requests to: Department of Heart and Lung Diseases, Sahlgrens Hospital, University of Gteborg, Sweden E-mail: Per.Bjorntorp@hjl.gu.se

Summary
Stress embraces the reaction to a multitude of poorly dened factors that disturb homeostasis or allostasis. In this overview, the activation of the hypothalamic-pituitary-adrenal (HPA) axis and the sympathetic nervous system have been utilized as objective measurements of stress reactions. Although longterm activation of the sympathetic nervous system is followed by primary hypertension, consequences of similar activation of the HPA axis have not been clearly dened. The focus of this overview is to examine whether or not repeated activation of these two stress centres may be involved in the pathogenesis of abdominal obesity and its comorbidities. In population studies adrenal hormones show strong statistical associations to centralization of body fat as well as to obesity. There is considerable evidence from clinical to cellular and molecular studies that elevated cortisol, particularly when combined with secondary inhibition of sex steroids and growth hormone secretions, is causing accumulation of fat in visceral adipose tissues as well as metabolic abnormalities (The Metabolic Syndrome). Hypertension is probably due to a parallel activation of the central sympathetic nervous system. Depression and the small baby syndrome as well as stress exposure in men and non-human primates are followed with time by similar central and peripheral abnormalities. Glucocorticoid exposure is also followed by increased food intake and leptin resistant obesity, perhaps disrupting the balance between leptin and neuropeptide Y to the advantage of the latter. The consequence might be stress-eating, which, however, is a poorly dened entity. Factors activating the stress centres in humans include psychosocial and socioeconomic handicaps, depressive and anxiety traits, alcohol and smoking, with some differences in prole between personalities and genders. Polymorphisms have been dened in several genes associated with the cascade of events along the stress axes. Based on this evidence it is suggested that environmental, perinatal and genetic factors induce neuroendocrine perturbations followed by abdominal obesity with its associated comorbidities. Keywords: adrenals, central, cortisol, leptin, neuropeptide Y, Obesity, stress.
obesity reviews (2001) 2, 7386

Introduction
Stress is a common phenomenon in the industrialized world, and provides a background not only to psychological reactions and insufciency but also, when protracted, probably somatic disease. The condition burned out now often occurs in the current discussion of diseases. This

expression denotes a psychological situation of having used up all endogenous energy, ending up in a psychological insufciency that makes further efforts at work as well as a normal psychosocial life severely disturbed, and requires a long period of recovery. This condition is apparently becoming increasingly prevalent, and is supposed to be due to the complex, competitive environment of current society.
73

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

74 Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

obesity reviews

This environment contains efforts of continuously increased production and elevated personal income to improve material conditions. Increased working hours, due to the factors mentioned, has a major responsibility in such a development. Obviously, factors such as personality characteristics, competence, education, and social status are heavily involved in the risk of ending up in a situation of stress and psychological insufciency. It seems likely that this will be followed by abuse of stimulants such as smoking, alcohol and drugs, which probably amplify the problems. In vulnerable personalities this may lead to symptoms of psychiatric conditions, particularly depression and anxiety. There is an interesting theory that when the relationship between demands, capacity and decision latitude at work is unfavourable, the risk for psychological incapacitation is elevated (1). An obstacle in this eld of research seems to be that a holistic view on the problem, combining both psychological and somatic techniques, is still rather rare. This might have a historical background, where certain schools of psychology are not interested in somatic consequences of psychological events. In addition, researchers in somatic disciplines often nd it easier to examine problems at systemic, tissue, cellular and molecular levels than within the complex reactions of the brain, and are often doubtful of the precision and relevance of methods used in psychological research. Stress researchers are frequently employing animal models, where information is only partially pertinent to the more complex human situation. The primitive consequences, where the stress centres in the lower part of the brain are activated, are, however, most likely similar in mammalian species, because they are essential for survival. Examples of animal stress research can be seen in the experiments performed by Henry and co-workers (2). In mouse societies two principal types of stress reactions were observed. One went through activation mainly of the sympathetic nervous system with elevation of heart rate and blood pressure. After a period of strain successful animals ended up in a superior position, with advantages for feeding, reproduction, etc. In the period of strain the hormonalautonomic consequences were increased activity of the sympathetic nervous system, and in a steady state after successful competition, increased testosterone levels in males. This is the so-called ght-ight reaction, with a successful outcome. Another type of stress reaction observed by Henry and co-workers (2) was a defeat reaction. These animals were unable to cope with a competitive situation and developed a reaction characterized by submissive behaviour and helplessness, and were socially handicapped. Their hormonal reaction was dominated by elevated secretion of cortisol and decreased secretions of sex steroid hormones.

Similar experiments have been performed in non-human primates, with essentially the same results. In these studies additional somatic consequences were registered. The animals with a defeat type of reaction accumulated excess depot fat in visceral adipose tissue and developed insulin resistance, dyslipidaemia, hypertension, impaired glucose tolerance and early signs of coronary atherosclerosis (3). Although these fully standardized animal experiments are very informative as far as primitive survival reactions are concerned, they are not necessarily applicable to the human situation because of the sophistication of the human brain. The types of challenges in human life are more complex and varied. In addition, the behavioural reactions in humans are different due to inhibitory mechanisms that are learned, and necessary, in human societies. The socalled poker face might serve as an example. An additional difculty is apparently that the well-dened reaction patterns in animals, described above, seem to be difcult to distinguish in humans. Usually a mixture of these reactions, or a shift from one reaction type to another seems to occur. In order to avoid these difculties a different approach to the problem has been chosen in this overview. First, signs of activation of stress centres in terms of increased activity of the sympathetic nervous system and/or the hypothalamic-pituitary-adrenal (HPA) axis have been measured. Thereafter factors activating these centres have been dened as stressors. The relationship of stress, dened in this way, to somatic disease has then been examined, with abdominal obesity in focus. Although animals respond to stress by several types of reactions, those of major importance in humans are activation of the central sympathetic nervous system and the HPA axis. The central sympathetic nervous system has its regulating centre in the brain stem and sends signals to a multitude of peripheral branches, of which those to central and peripheral haemodynamic regulation are examples. The HPA axis is regulated from centres in the hypothalamic area which via corticotropin releasing hormone stimulate adrenocorticotropin (ACTH) production and thereby cortisol secretion from the adrenals. The activity of the HPA axis is controlled by a negative feedback mechanism where central glucocorticoid receptors (GR) in among other locations, the hippocampus area, are involved. These two stress axes are closely connected at several levels (Fig. 1). The major goal for these activities is to attempt to maintain homeostasis or allostasis. Allostasis is the allowance of some physiologically necessary adaptations, such as a rise of blood pressure when changing from the supine to erect position, or an elevation of insulin secretion after a meal. When these reactions are disturbed and exaggerated an allostatic load is at hand (4). When the input of noxious signals is prolonged, the HPA axis reactivity changes. First, there is a normal response

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

obesity reviews

Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

75

dopaminergic and serotoninergic neurones (6). The details of such involvement are complex and not fully claried.

Stress factors in humans

As stated above the factors eliciting a stress-reaction are dened in this study as circumstances followed by an activation of the HPA axis and/or the central sympathetic nervous system. Such factors have been dened in a large number of publications. There is now a likely consensus that primary hypertension is initiated by frequent activation of the sympathetic nervous system (7). A corresponding development along the HPA axis has been less studied, perhaps due to the complexity of adequate measurements. The HPA axis is exquisitely sensitive to stimuli. A recent study illustrates such reactivity particularly well (8). A group of subjects provided saliva samples for cortisol determinations at random occasions during an ordinary working day. They also reported events and moods during the preceding hours before saliva sampling. The results showed that not only perceived stress during the day inuenced cortisol levels, but also the memory of previous and the thought of anticipated, stressful events. Furthermore, the current mood inuenced on the values being happy was followed by low cortisols and feeling miserable the opposite. The diurnal variation of cortisol was 20-fold and proportional to the impact of the registrations of psychological events. This study demonstrates the variety of factors that inuence the reactivity of the HPA axis. Furthermore, the results also clearly indicate the plasticity of the reactions of the HPA axis. As described above, the reactivity of this axis is changing with frequency and impact of challenges. Therefore, the kinetics of the diurnal secretion and responses to impacts are important areas for examination to determine the status of the regulation of the HPA axis. This requires repeated measurements of circulating cortisol. Sampling of urinary cortisol seems less useful for such purposes. A considerable number of studies have shown that various strains during work, particularly when demands are larger than decision latitude, and in situations with monotonous type of work and lack of control, are followed by elevated secretion of cortisol (912]). Such studies indicate different types of stressors in daily life, but are usually not including examination of potential somatic consequences in terms of disease or disease predictors. In contrast, the Whitehall studies have indicated that a socioeconomic gradient is inversely followed by disease symptoms, including centralization of body fat and the Metabolic Syndrome, but have not measured potential neuroendocrine reactions which might be responsible for such a development (13). We have recently examined this problem in populations of men and women, with the ambition to measure the neu-

Figure 1 Central inputs activate the hypothalamic-pituitary-adrenal axis via corticotropin releasing hormone (CRH) with adrenocorticotropin (ACTH) and cortisol secretion as consequences. Central glucocoticoid receptors (GR) control the activity by feedback inhibition. The central sympathetic nervous system (n.s.) is frequently activated in parallel followed, if protracted, by primary hypertension and increased heart rate. The two stress centres are tightly interconnected and both are often activated simultaneously.

with peaks of cortisol secretion, which upon relief from the input is winding down to basal levels. When frequently repeated there is apparently a phase of sensitization of the system, which reacts with exaggerated cortisol secretion after a given input. This is occurring during the most active phase of the HPA axis, which is during the night in rodents, and early in the morning hours in humans (5). When repeated too often and with sufcient strength of the input the rst sign of malfunction is a delayed down-winding period, cortisol secretion stays elevated for a prolonged period of time. Subsequently, the normal diurnal rhythm with high HPA axis activity in the early morning is disrupted, and morning values tend to be lower. This is developing into a low, steady, rigid diurnal cortisol secretion with little reactivity a burned out HPA axis. Along with this development the controlling, central glucocorticoid receptors become less efcient, and their density becomes down-regulated. Further challenges are followed by atrophy of the signalling dendrites from the GRs, with end results of atrophy and lacunae in the hippocampus area. This has been found after long-term, severe hypercortisolaemia such as in Cushings syndrome, in melancholic depression and in war veterans (4). During this development the central sympathetic nervous system is also often activated in parallel because of the tight functional coupling of the both stress centres. Such activity seems to be further exaggerated in the conditions of severe regulatory malfunction of the HPA axis, such as in the burned out condition, perhaps as a compensatory activity (6). Both the central sympathetic nervous system and the HPA axis are regulated by proximal input of both stimulatory and inhibitory signals from central adrenergic,

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

76 Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

obesity reviews

roendocrina consequences of potential stressors as well as associated endocrine, anthropometric, metabolic and haemodynamic factors. Measurements of saliva cortisol were used as indicators of HPA axis activity in these studies, and the blood pressure and heart rate values as indicators of the activity of the sympathetic nervous system. In women it became evident that free testosterone was most likely an index of central regulation of adrenal activity because it correlated strongly with dihydroepiandrosterone sulphate (DHEAS), a steroid excreted almost exclusively by the adrenals, as well as with various cortisol measurements. Free testosterone and DHEAS were, therefore, also used as a measurement of adrenal activity because of the advantage in comparison with cortisol of showing less variation over the day (14,unpublished).

Figure 2 Cortisol binds to glucocorticoid receptors, which have a particularly high density in visceral fat depots. Lipoprotein lipase is activated via gene transcription and enzyme stabilization. Cortisol also inhibits lipid mobilization in the presence of insulin. Both these activities lead to triglyceride accumulation and retention in visceral adipose tissue (16).

HPA axis perturbations and centralization of body fat

The waist/hip circumference ratio (WHR) and abdominal sagittal diameter are estimations of centralization of body fat, where the sagittal diameter shows the strongest correlations with the mass of intra-abdominal, visceral depot fat (15). These masses are highly dependent on the endocrine status. This eld has been reviewed in detail, where references to original publications are found (16). In brief, cortisol as well as sex steroid and growth hormones are involved. Cortisol activates lipoprotein lipase, the gatekeeper of lipid accumulation in adipocytes. Furthermore, cortisol in the presence of insulin inhibits the lipid mobilizing system. Since these events are mediated by the GR, and the density of these receptors is higher in intraabdominal, visceral than other fat depots, the activity of cortisol leading to accumulation of fat will be accentuated in this adipose tissue. This is seen dramatically in Cushings syndrome, and disappears with successful therapy. This is summarized in (Fig. 2). Growth hormone exerts powerful inhibition of the lipoprotein lipase activity and activates the lipid mobilizing system. The sex steroid hormones exert permissive effects on growth hormone action. The density of the androgen receptor is particularly high in visceral adipose tissue, and is upregulated by testosterone. Consequently, a combination of these hormones has, in principle, actions counteracting those of cortisol (Fig. 3). In summary, this means that elevation of cortisol is followed by visceral fat accumulation. Furthermore, low sex steroid and growth hormone secretions will have the same consequence, because of the insufcient counteraction against cortisol effects, and a combination of these abnormalities will have powerful consequences to direct a larger than normal fraction of total body fat to visceral depots (Fig. 4).

High density of androgen receptors Testostero increases androgen receptor density Lipoprotein lipase inhibited Lipid mobilisation stimulated Triglyceride accumulation and retention diminished
Figure 3 Testosterone is acting via the androgen receptor, the density of which is increased by autoregulation. The receptor has a higher density in visceral than other fat depots. By inhibition of lipid accumulation and stimulation of lipid mobilizing mechanisms, visceral depots are diminishing. These actions provide a permissive amplication of growth hormone effects (16).

This interpretation of data from cellular and molecular studies is in agreement with observations in the integrated system in vivo. Increased cortisol secretion in Cushings syndrome as well as after corticosteroid treatment of disease is typically followed by an increase of central, visceral adipose tissue volume. Such conditions are also usually associated with low levels of growth and sex steroid hormone concentrations, because it is well established that activation of the HPA axis is followed by inhibition of the central gonadal and growth hormone axes at several levels (17) (Fig. 4). Hormonal, as well as visceral fat mass, abnor-

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

obesity reviews

Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

77

Table 1 Correlations between hormones and anthropometric variables in men (14,18) p-values BMI Normal HPA axis Total cortisol Lunch cortisol Stress cortisol Abnormal HPA axis Total cortisol Lunch cortisol Stress cortisol WHR W D

0.035 0.005 ns ns <0.001 <0.001

ns ns ns 0.001* 0.003 <0.001

ns <0.001 ns ns <0.001 <0.001

ns <0.001 <0.001 ns <0.001 <0.001

*Negative correlation. BMI, body mass index; WHR, waist/hip circumference ratio; D, sagittal abdominal daimeter.

Figure 4 The chronically activated hypothalamic-pituitary-adrenal axis inhibits the secretion of growth hormone (GH), testosterone (T) and 17b estradiol (E2), which counteract cortisol effects. The net result will be accumulation of visceral fat (16,17). CRH: Corticotropin releasing hormone. GnRH: Gonadotropin releasing hormone. GHRH: Growth hormone releasing hormone.

malities are disappearing with removal of cortisol excess. In conditions with primary deciency of growth hormone, such as in hypophysectomy with substitution of all other hormones than growth hormone, there is clearly an increase of visceral fat mass, disappearing with growth hormone substitution. In analogy, excess growth hormone secretion, such as in acromegaly, is associated with small intra-abdominal fat depots. Similar examples are found for the sex steroid hormones. With menopause and decreased concentrations of 17b estradiol, visceral fat masses are elevated, and diminish with hormone replacement therapy. With testosterone deciency in men, occurring in some men with ageing, visceral fat masses increase, and are diminished with adequate testosterone replacement therapy. The tendency for elevation of visceral fat masses with age in both genders may, thus, be due primarily to the relatively low secretion of growth and sex steroid hormones, which are following ageing (16). In summary, there is an excellent agreement between in vivo observations and cellular, molecular data indicating that excess cortisol, as well as low growth and sex steroid hormones are followed by elevated visceral fat mass. The evidence indicates that these associations are causally mediated. As mentioned above, it is important to realize that an increased activity of the HPA axis is followed by inhibition of both the pituitary gonadal as well as the growth hormone axes (17). From this it follows that a decrease of sex steroid and growth hormones may in fact be a consequence of stress

factors acting over the HPA axis. Stress may, therefore, be active on the factors accumulating visceral depot of fat both via both elevated cortisol secretion as well as the secondary decrease of sex steroid and growth hormones (Fig. 4). The next question then is whether or not these associations in examples of clinical, endocrine entities are valid in the general population. In the populations mentioned above this has been examined in both men and women. In the men there are strong relationships between the body mass index (BMI), the WHR, the waist circumference and the sagittal diameter on the one hand, and various cortisol measurements on the other as seen in Table 1. It should be noted that these correlations are most tight with cortisol secretion regulated by an abnormally functioning HPA axis, particularly after challenges, such as after lunch, and reported peceived stress (14). Furthermore, low testosterone and/or growth hormone concentrations alone show such associations, growth hormone secretion estimated as the concentration of insulin-like growth factor I (18). In addition statistical path analyses indicate that both measurements of cortisol and, negatively, of testosterone and/or growth hormone are directly followed by the WHR or the sagittal abdominal diameter (18) (Fig. 5). These observations indicate that the WHR and sagittal abdominal diameter are dependent on elevated cortisol and/or low sex steroid and growth hormones, and that these anthropometric measurements of centralization of body fat might serve as reasonable approximations of the long-term endocrine abnormalities associated with stress. It should be observed that the BMI shows equally strong associations as the measurements of centralization of body fat to the cortisol measurements (see Table 1 and Fig. 5). This will be discussed separately in the next section. In women the situation is partly different. As briey mentioned above free testosterone is probably a measurement of adrenal activity in women because of its close association with DHEAS and cortisol measurements. As seen in

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

78 Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

obesity reviews

vated testosterone values in women are only about 1/10 of those in men, and the cortisol antagonizing effects of testosterone might not be efcient at such low concentrations. From this overview one may conclude that central accumulation of fat is a consequence of long-term activation of the HPA axis with subsequent inhibition of the gonadal and growth hormone axes, and therefore, by denition, a consequence of stress. Genetic susceptibility is most likely an important contributor, as will be discussed in a later section.

Figure 5 In statistical path analyses activity of the hypothalamicpituitary-adrenal axis (HPA) is, in the absence of body mass index (BMI), followed by low testosterone (T) and growth hormone (GH) secretion, and then the waist/hip circumference ratio (WHR) and abdominal sagittal diameter as well as insulin as indicated (r = <0.996). When BMI is included in the absence of WHR and sagittal diameter, BMI directly follows HPA axis activity (r = 0.998) (18).

Obesity and glucocorticoids

It is important to realize the difference between central accumulation of depot fat and obesity. Obesity is, by denition, an accumulation of excess body fat, irrespective of localization, and is usually measured as the BMI in epidemiological settings. Central accumulation of depot fat is dened as the fraction of total body fat that is localized to central, visceral depots. This could, therefore, be a phenomenon which is not necessarily associated with obesity, as dened above. As a rule, however, the BMI and WHR or other measurements of centralization of body fat, are closely associated, a condition of abdominal obesity (14,18). In a preceding section evidence was provided to indicate that centralization of body fat most likely is a consequence of the endocrine perturbations characteristic of chronic stress reactions. Realizing the close connection between centralization of body fat and enlargement of total body fat, one may therefore ask whether or not excess accumulation of total body fat, that is obesity, is also associated with neuroendocrine stress reactions. There is accumulating evidence that this may be the case. Clinical experience clearly indicates that patients treated with corticosteroids have an increased, sometimes voracious appetite and frequently become obese, with a large fraction placed centrally. Patients with Cushings syndrome are often obese in addition to their pronounced accumulation of fat in central depots. In melancholic depression, another hypercortisolaemic condition, increased food intake and obesity are known to occur in certain subgroups (21). These observations suggest that cortisol or other glucocorticoids may be involved in increased energy intake, resulting in obesity. As mentioned above, statistical path-analyses including the WHR or sagittal diameter show that these follow tightly after measurements of HPA axis activity. Similar analyses show that when BMI is included in such calculations, but not the WHR and the sagittal abdominal diameter, BMI is also closely following measurements of the HPA axis, suggesting that activity of this axis is involved also in obesity (18) (Fig. 5). This is also seen in Table 1, where the BMI shows equally strong relationships to cortisol during various conditions as the WHR, waist circum-

Table 2 Correlations between androgens and anthropmetric factors in women (unpublished) T BMI WHR Sag. Diam. ns ns 0.135* FreeT 0.488*** 0.416*** 0.469*** SHBG -0.523*** -0.470*** -0.541*** DHEAS 0.140* 0.221*** 0.200***

T, Testosterone; SHBG, sex hormone binding globuline; DHEAS, Dihydroepiandrosterone sulphate; BMI, body mass index; WHR, wiast/hip circumference ratio; Sag. Diam., Abdominal sagittal diameter. r-values. ns, not signicant. *, **, ***. P-values <0.10, <0.05, <0.01, <0.001.

Table 2, free testosterone as well as DHEAS and, negatively, sex hormone binding globulin are strongly associated with the measurements of total BMI and centralization of body fat (WHR and sagittal diameter) in women (unpublished). Administration of androgens to normal or transsexual women is followed by increased visceral fat mass (19,20). This evidence suggests that adrenal hormones, including androgens are involved in the regulation of central fat masses also in women, and consequently that the WHR and sagittal abdominal diameter are probably reasonable measurements of long-term stress activation of the HPA axis in women. The associations between the hormonal measurements and BMI will be discussed in the next section. One may wonder why low testosterone in men and high testosterone in women is associated with elevated central body fat mass. The explanation might be that both are consequences of increased activity of the HPA axis, where the effects of cortisol outweigh those of testosterone. Even ele-

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

obesity reviews

Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

79

ference or sagittal abdominal diameter. Taken together, these observations open up the possibility that both obesity and centralization of body fat may have the same origin, namely often repeated or chronic activation of the HPA axis. There is interesting experimental evidence to suggest that glucocorticoids might be involved in the pathogenesis of obesity, also elucidating the mechanisms involved. Recent experiments using the rat show that adrenalectomy is followed by marked sensitivity to leptin. With graded replacement of glucocorticoids leptin sensitivity is diminished in parallel, and with administration of larger doses of glucocorticoids overeating is induced, resulting in obesity in spite of elevated leptin levels (22). These experiments demonstrate that corticosteroids induce leptin secretion and a condition of leptin resistant obesity, induced by glucocorticoid excess. Human obesity is also characterized by elevated leptin levels (23), a leptin resistant type of obesity. The localization of this resistance is apparently not at the level of the leptin receptor to a signicant degree (24), and its origin is currently not known. There is further evidence from experimental studies in humans. Elevated levels of leptin follow administration of glucocorticoids with variations in duration and apparently different effects in obesity and between the genders (2527). In one study food intake was measured by a technique where food was available freely in liquid form from a dispenser. This method allows full control of energy intake, but may not be considered to mirror natural conditions. It is also known that this method is followed by overconsumption of energy. Nevertheless, the group taking glucocorticoids was, in spite of elevated leptin levels, consuming more energy than the control group (27). We have recently repeated this study in 15 moderately overweight women who received 25 mg prednisolone during 7 days. Leptin levels rose in all women, who also showed increased food intake measured with an eating monitor. The effect was apparently not diminished over the time of measurements. Effects on total energy intake, measured by the less precise technique of diet history, could not be discovered, although there was a tendency to increased carbohydrate intake (unpublished). Taken together this evidence suggests that glucocorticoids are able to increase food intake in spite of elevated leptin levels. This opens up the interesting possibility that cortisol may be involved, not only in the centralization mechanisms of body fat, but also in the accumulation of total body fat, that is obesity, provided that these effects are remaining with a sufciently long duration. If this turns out to be the case cortisol may induce leptin resistant obesity, the common form of obesity in humans. This problem is of particular interest against the background that obesity, in general, is associated with overproduction of cortisol, although this is apparently more pronounced in

central obesity (28). This is an area of considerable interest for further research. Recent experiments suggest the possibility that the other regulatory axis of food intake regulation, the neuropeptide Y (NPY) system, might also be involved in the effects of glucocorticoids. Stress-induced cortisol-secretion seems to be followed by elevated NPY secretion (29). Increased circulating levels have been reported after strenuous physical stress (30) as well as after mental stress (31). However, in the rat restraint stress diminishes NPY in the arcuate nucleus (32) and NPY mRNA in the amygdala (33), but after repeated stress and high salt intake plasma levels of NPY rise (34). Although these results indicate relationships between stress-related cortisol and peripheral NPY levels in humans, the problem needs further evaluation. In summary, the evidence available suggests that glucocorticoids blunt the efciency of the inhibiting branch of food intake control, the leptin system. There is also a possibility that glucocorticoids can stimulate the food intake branch, the NPY system, and therefore disrupt the balance of food intake regulation to the advantage of NPY effects. If conrmed, this could, in the long-run, be expected to be followed by overeating and obesity. This hypothesis is summarized schematically in Fig. 6. With this background one may wonder if cortisol elevation with stress might induce the phenomenon of stress-eating, and thereby contribute to the generation of obesity. The basic concept for stress-eating to occur is that increased food intake would be a response to emotional

Figure 6 Neuroendocrine background to abdominal obesity, a hypothesis. With cortisol excess (right panel) the balanced system (left panel) the secretion of neuropeptide Y (NPY) is stimulated and leptin effects blunted (leptin resistance). This is followed by increased food intake and elevated total body fat mass (outer circle) with too large a fraction in visceral depots (inner circle), directed there by excess cortisol and low gonadal and growth hormones. The result will be abdominal obesity, as a consequence of stress-eating.

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

80 Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

obesity reviews

stress, and that eating would subsequently diminish the distress, presumably via central opioids. Animal data with acute physical stressors have shown varying results (35), while stressful housing conditions have been shown to be followed by increased food intake (36). Also in human studies conicting results have been reported. Preoperative stress showed no effect (37), while results of another study suggest overeating before the stress of school examinations. This seemed to be more pronounced for fat intake and in girls (38). There are considerable individual variations. A reasonably stable predictor of stress-induced eating seems to be a restrained attitude towards eating. This area has recently been subjected to a comprehensive review (35). Another critical recent review concludes that the evidence now available is awed by various errors, making conclusions difcult (39). Stress-eating is complex to study for several reasons. First, the denition of stress is highly variable and individually different as described in a section above. Second, the presumed end result, increased food intake in relation to stress, is not easy to measure accurately, and studies have often been relying on self-reports. In this review stress has been dened as a measurable neuroendocrineautonomic response. Studies of stress-eating utilizing such measurements might be clarifying. This area is of particular interest when considering the background of the current obesity epidemic in industrialized parts of the world. This has started in parallel with a presumed decrease of physical activity, but also in parallel with the development of a very competitive society, where overeating due to stress might be involved.

decrease below a certain steady state, obese or lean, results in secondary neuroendocrine and autonomic adaptations which confound interpretations. Another way to intervention is to attempt to normalize the neuroendocrine and endocrine abnormalities by, for example, substitution with sex steroids or growth hormone to levels normal for age. Such studies have shown clear global improvements or even normalization of abnormalities in abdominal obesity (40). The crucial intervention is, of course, that on cortisol secretion. This is difcult to perform because there are, unfortunately, currently no safe means for this purpose. Cushings syndrome may, however, be taken as a model. When the hypercortisolaemia is removed, the condition is cured. Furthermore, and the other way around, when other primates than humans are exposed to standardized submissive stress, an identical picture as that seen in abdominally obese humans is develops (3). The consistent statistical association between abdominal obesity and stress-inducing environments in humans (see next section) is another argument in the same direction, because it seems unlikely that such environments would follow moderate abdominal obesity. Further arguments will be discussed below in relation to depression and the small baby syndrome.

Associations between environmental factors and neuroendocrine and autonomic stress reactions
In population studies of men and women several associations have been found between stress reactions measured as neuroendocrine and autonomic consequences, as described above, and a number of environmental factors. In our earlier reports, where information was obtained from the study of men born in 1913 and from the study of women in Gothenburg, associations were sought mainly between measurements of the WHR and environmental factors. The rational for using the WHR as a surrogate measurement of long-term stress was reviewed in a preceding section. The study of men comprised approximately 1000 subjects, and the study of women about 1400 subjects, probands in cohorts of about 200 in different age strata from both before and after menopause. In men, associations were found between the WHR and poor education, low social class, and physical types of poorly paid work. In addition divorce and living alone showed positive correlations. This was also found for alcohol consumption and smoking (41). Similar ndings were made in women with some interesting differences. Women are apparently less sensitive to having experienced divorce, and to live alone. Furthermore, particularly consumption of strong liquor showed correlations to the WHR. In women there was also information

Are neuroendocrine-endocrine perturbations secondary to abdominal obesity?

The considerations above, both as far as fat distribution and obesity, suggest that the associated neuroendocrineendocrine abnormalities in, particularly, central obesity is the primary factor followed by abdominal obesity and its complications. The increased cortisol secretion in central obesity is now well established (28). There is also the possibility that this is secondary to the obese state. Cortisol turn over has been reported to be elevated, due to a peripheral metabolism to less active metabolites. This would, in turn, increase cortisol secretion secondarily due to a less efcient feedback inhibition of the HPA axis. If this is the correct interpretation then it is difcult to understand why only central obesity would have elevated cortisol and peripheral symptoms of a relative hypercortisolism, and not peripheral obesity at the same BMI. An apparent solution to this problem would be to examine what is left of the central regulatory abnormalities after treatment to normal body weight. Unfortunately the results of such interventions become inconclusive, because weight

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

obesity reviews

Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

81

on psychiatric symptoms or equivalents, including depressive moods, consumption of antidepressant drugs and anxiolytics. In addition, sleeping difculties were frequent (42). Similar ndings have subsequently also been reported in studies from other laboratories (43). These previous population studies were not primarily designed for analyses of the relationships between stress and somatic abnormalities, making the possibilities for detailed analyses fairly limited. The information obtained was, however, of sufcient interest to encourage further studies directly focused on this problem. Therefore, new populations were selected, where data were assembled to allow more detailed examinations of the relationships between environment, stress-reactions, and somatic abnormalities. In these recent studies the previous observations have been replicated with some additions. The psychiatric relationships to the WHR were also found in men (44). In addition, certain personality types were more prone to show such associations, particularly men with dependent, anxiety prone personalities (45). In early studies the association to the WHR was clearly stronger than those to the BMI, in fact the latter associations were in certain instances negative (41,42). In the new studies strong associations are found not only to the WHR, but also to the BMI. An apparent explanation for this difference might be that there is a 30-year difference between the birth year of the men, and between 25 and 40 years in the women. During this period of time there has been a considerable increase in the prevalence of obesity in Sweden. This might have caused a greater impact of obesity in the newly performed studies. In the new studies measurements of the activity of the HPA axis were included, allowing studies of the direct associations between environmental and genetic (see below) factors and the regulation of neuroendocrine and autonomic centres, as well as the consequences for somatic variables. In essence, similar results were obtained with the measurements of HPA-axis activity as with the anthropometric measurements, reemphasizing that measurements of centralization of body fat masses might be considered as useful indicators of long-term activation of neuroendocrine abnormalities (46,47). The direct measurements of detailed HPA-axis activity give, however, more detailed information on the current status of the kinetics of neuroendocrine regulations, and better possibilities to trace the origin of the abnormalities. Physical inactivity may be involved in the pathogenesis of abdominal obesity. This is indicated by answers to questionnaires by both men and women (50,unpublished). Men with socioeconomic handicaps often have physical types of works (41). Physical training is followed by adaptations towards a lower activity of the central sympathetic nervous system with low pulse rate and blood pressure (49). This would presumably be the reverse with physical

inactivity. Intervention studies separating physical inactivity and moderate daily activity with information on neuroendocrine and autonomic adaptations do not seem to have been performed. There is no doubt that physical activity even at moderate levels would decrease the preponderance to develop obesity. Whether physical inactivity would have an additional, specic effect on abdominal obesity is, however, not claried.

Metabolic and haemodynamic consequences of stress reactions

The focus of this review has been the association between stress and obesity, particularly central, abdominal, visceral obesity. There are, however, other statistical companions to stress with abdominal obesity. These are metabolic perturbations, with insulin resistance in a central position, which are characteristic of the Metabolic Syndrome. These perturbations are also likely to be consequences of the endocrine abnormalities following stress reactions, with elevated cortisol secretion as well as associated decreases in growth and sex steroid hormones. These mechanisms are largely known, and this eld has recently been reviewed (50). Frequently associated hypertension has been considered to be a consequence of hyperinsulinaemia, following insulin resistance (51). It is, however, equally likely that the hypertension associated with the Metabolic Syndrome is caused by parallel activation of the both stress centres, the HPA axis and the central sympathetic nervous system. These centres are closely functionally coupled (17), and it is difcult to activate one centre without activating the other (Fig. 1). This problem has also been reviewed in detail recently (50).

Depression
Depression is a classical condition with an activated HPA axis and central sympathetic nervous system. Depression is known to be followed by a substantially increased risk to develop cardiovascular disease and type 2 diabetes mellitus in prospective population studies. Recent data show that depression or depressive symptoms are associated with somatic symptoms characterizing the Metabolic Syndrome, including visceral accumulation of body fat. Treatment with a serotonin reuptake inhibitor seems to be followed by improvement not only of the neuroendocrine but also the metabolic abnormalities in men with depressive symptoms (52). This information suggests the possibility that the risk factor pattern in depression follows the neuroendocrine perturbations of that condition, and therefore the risk of somatic disease is increased. This then supports the suggested pathway for development of the Metabolic Syndrome and central obesity, described above.

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

82 Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

obesity reviews

The small baby syndrome

The small baby syndrome is a condition where subjects, born small for gestational age, develop central obesity and the Metabolic Syndrome at adult age. Recently it has been shown that such individuals display perturbations of the regulation of the HPA axis, much of the same character as described above for subjects in the population studies. The background seems to be preprogramming of the regulation of central neuroendocrine axes, which is possible to induce in experimental studies in rats by agents such as glucocorticoids, lipopolysaccharides, cytokines and immune or stress challenges to the pregnant mother. It, therefore, seems possible that similar neuroendocrineautonomic, anthropometric, metabolic and haemodynamic abnormalities found in the population may in fact be due to perinatal conditions. The quantitative importance of this pathway in relation to environmaental pressures in adult life and genetic abnormalities is not yet known (53).

Prospective studies of stress exposure

The best evidence for the correctness of the interpretations presented above of the pathogenesis of the Metabolic Syndrome and central accumulation of visceral fat comes, however, from controlled experiments in a non-human primate. When Cynomolgus monkeys are subjected to psychosocial stress, some individuals respond with a depressive, defeated reaction. These monkeys show enlarged adrenals, a diminished suppression of cortisol secretion by a low dose dexamethasone, decreased sex steroids, elevated androgens in females, visceral fat accumulation, the Metabolic Syndrome and early signs of coronary atherosclerosis and glucose intolerance (54). This is an end result of controlled, long-term stress exposure and is identical to the results we have reported in humans with psychosocial stress as summarized above, and, thus, offers strong support to the contention that stress may be the rst part of the chain of events leading to somatic disease. Similar observations may in fact be made in men. Men exposed to poor working conditions develop with time a gradually increasing visceral fat mass and the Metabolic Syndrome (13,46). Stress-induced cortisol secretion is increasing and the dexamethasone suppression is becoming blunted (46). These observations are also supportive of the interpretations provided for the connection between environmental stress and somatic risk factors.

Genetic factors
An update of the genetics of human obesity has recently appeared (55). There are, clearly, genetic factors involved

in the neuroendocrine and autonomic aberrations described above. For example, homozygotic twins show closely similar diurnal secretion of cortisol (56). Association studies suggest several possibilities for polymorphisms of candidate genes involved in the syndrome described in this study. The control of the activity of the HPA axis provided by central glucocorticoid receptors shows abnormalities in the men examined with abnormal HPA-axis regulation (14,46) (women, to be examined). This is associated with a polymorphism of the glucocorticoid receptor gene. A longer allele (4.5 kb) than that considered normal (2.3 kb), obtained after digestion with the restriction enzyme Bcl I, is associated with a poor control of stimulated cortisol secretion (57) as well as with abdominal obesity, insulin resistance and hypertension (5760). Homozygotes (4.5/4.5 kb) are found in about 14% of the men. Heterozygotes (4.5/2.3 kb) also show associations, although less pronounced. The localization of this polymorphism is in the rst introne, and it may, therefore, not be excluded that it has a functional signicance for initiation of transcription. Another polymorphism is found in the 5-anking (promoter) domain of the glucocorticoid receptor gene locus and is associated with total cortisol secretion (61). Both these polymorphisms suggest that there might be variants of functional importance in the 5end or promoter region of the glucocorticoid receptor gene. We are, therefore, currently attempting to sequence these domains. A microsatellite in the rst coding exon of the gene is apparently normal (unpublished). We could not conrm in our material of men that a polymorphism in this exon is associated with obesity (62), which was recently reported (63). In the women studied potential genes associated with hyperandrogenicity have been examined. Focus has been set on microsatellite length, which apparently shows large variations in genes regulating steroid hormone metabolism and signals to target cells. This is a novel aspect of genetic associations to malfunction and disease. Short stretches of microsatellites in, for example, the transactivating domain of the androgen receptor gene result in a receptor with strong signalling power and vice versa. These quantitative relationships seem particularly interesting in the pathogenesis of certain human diseases, such as diabetes type 2 or hypertension, where a gradual, quantitative deterioration of regulatory mechanisms occurs. The microsatellite stretches are apparently instabile between generations, making these types of polymorphisms interesting also for evolutionary changes (64). Testosterone is transformed into 17b estradiol by the enzyme aromatase. The aromatase gene has a microsattelite in the 5th intron where short length variants have previously been shown to be associated with disease in women (65). A short microsatellite with tetranucleotide (TTTA) repeats was found to be associated with elevated free

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

obesity reviews

Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

83

testosterone and low 17b estradiol levels (unpublished), which is the expected result of a poorly functional enzyme. The activity of the enzyme has, however, not yet been determined. Furthermore, a short microsatellite with <19 trinucleotide (CAG) repeats was found in the transactivating domain of the androgen receptor gene suggesting strong androgen signals to target cells in women with hyperandrogenicity (unpublished). Associations between polymorphisms in candidate genes and abnormalities in the regulation of the central sympathetic nervous system have also been examined. Blood pressure elevation is directly associated with polymorphisms in the dopamine receptor 2 as well as in the b-adrenergic receptor 2 genes (unpublished). Several polymorphisms in the leptin receptor gene show negative relationships to blood pressure even in subjects with obesity and elevated leptin levels. Among hypertensive probands, who have elevated BMI and leptin levels, only one out of 64 subjects were found to have a leptin receptor gene with polymorphisms. The blood pressure differences are not only statistically, but also clinically signicant, amounting to approximately 25/15 mmHg in systolic/diastolic blood pressures in men with or without polymorphisms (66). This is of interest because there is recent evidence that leptin activates the central sympathetic nervous system, followed by elevated blood pressure (67). Polymorphisms of the leptin receptor gene are apparently protecting from the hypertension frequently following obesity, and might act via a blockage of the leptin signals to the central sympathetic nervous system. These ndings might provide an explanation to the fact that all obese subjects are not hypertensive. Taken together these results of molecular genetic studies with the candidate gene approach have been possible by the initial results of the detailed phylogenetic expression of disease symtoms, with focus on abdominal obesity and its associated complications. A new direction of such studies is the focus on central, regulatory genes. Most of the results are associations to abnormalities and require further studies to evaluate potentially functional importance. The glucocorticoid receptor gene results seem particularly promising and are pursued further. Moreover, the differences in microsattelite alleles may have direct functional importance as suggested by detailed experimental manipulations in model systems, including experimental animals (unpublished). Taken together these early results suggest that genes regulating neuroendocrine and autonomic systems are of interest for further studies.

Summary and conclusions

The focus of this review has been to summarize the evidence indicating a potentially causal relationship between stress and, particularly, abdominal obesity. The endocrine

stress reactions are characterized by elevations of activities of the HPA axis and the central sympathetic nervous system. The approach in this study has been to start out with the physiological consequences of stress in terms of the neuroendocrine-autonomic reactions, and to examine how these are related to environmental factors, which then might be characterized as stress factors. This is an apparently easier approach to the problem than starting out from supposed stress factors, where the perception of stress is highly varying among individuals, and may or may not be associated with biological responses. Using this method one can, on a population basis, identify a number of psychosocial and socioeconomic handicaps which are followed by either increased cortisol secretion or other perturbations of the control of the HPA axis with subsequent inhibition of the gonadal and growth hormone axes, as well as activation of the central sympathetic nervous system. These reactions are closely statistically associated with visceral fat accumulation. There is considerable evidence from clinical and intervention studies as well as research at the cellular and molecular levels that these neuroendocrine perturbations have the consequences of accumulation of depot fat in visceral depots because of the high density of steroid hormone receptors which mediate activation of fat accumulating mechanisms. It is, therefore, tentatively concluded that a biological neuroendocrine-endocrine stress-response is followed by visceral fat accumulation, and consequently, that elevated visceral fat mass is a consequence of a long-term stress reaction. This is individually different and most likely based on genetic susceptibility. Centralization of body fat and obesity, dened as elevated fat mass irrespective of localization, are usually tightly statistically associated. Clinical experience from Cushings syndrome and corticosteroid therapy suggests that glucocorticoids increase food intake and cause obesity. In statistical path-analyses perturbations of HPA axis regulation are directly followed by obesity measurements, suggesting some association between the HPA axis and obesity. Experimental interventions in humans with administration of glucocorticoids show elevated leptin levels as well as increased food intake. Elevated leptin levels, suggesting a leptin resistant obesity characterize human obesity. Furthermore, survey of literature suggests that human obesity is a condition with elevated cortisol secretion. Results from studies in experimental animals suggest that glucocorticoids may cause leptin resistant obesity. There is emerging evidence that these hormones may also activate the NPY system, resulting in both inhibition of satiety mechanisms and activation of food intake mechanisms. Taken together this evidence opens up the possibility that not only central obesity, but obesity in general might be a consequence of elevated activity of the HPA axis. Such increased activity is the result of perceived stress.

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

84 Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

obesity reviews

The neuroendocrine-autonomic stress reaction is probably followed by metabolic abnormalities, the Metabolic Syndrome, via the endocrine perturbations following longterm, repeated activation of the HPA axis, as well as elevation of blood pressure via a parallel activation of the central sympathetic nervous system. Depression may serve as a model for the condition described, because here the neuroendocrine-autonomic abnormalities are similar as well as apparently the somatic risk factor pattern and increased risk to develop somatic disease. Furthermore, the small baby syndrome, characterized by central obesity and the Metabolic Syndrome, has recently been found to be associated with similar central abnormalities, which in experimental studies are possible to induce by manipulations during pregnancy. The strongest support for the correctness of the interpretations offered for a chain of submissive stress-neuroendocrine-autonomic perturbations-anthropometric, metabolic and haemodynamic abnormalities comes, however, from controlled experiments in non-human primates, where this has been directly shown. A similar situation can be identied in men subjected to stressful working conditions. The evidence summarized in this study, thus, suggests that abdominal obesity is a condition with a neuroendocrine background affecting the HPA axis as well as the central sympathetic nervous system with subsequent inhibition of the gonadal and growth hormone axes. These might be primary events where subsequently the leptin and NPY systems become unbalanced, favouring the latter. The background to such a development is likely of central origin via factors that are commonly dened as stress. In order for abdominal obesity to develop a positive energy balance must be established. The neuroendocrine cascade described may facilitate an increased energy intake in susceptible subjects. A decreased physical activity will also be followed by a positive energy balance and might be another consequence of current environments. It has been suggested that there is a closer control of energy balance at higher levels of energy output, and less tight regulation at lower levels of physical activity. One might speculate that the lack of stress releaving function of physical activity might add to the impact of the syndrome described above.

References
1. Karasek RA. Job demands, job decision latitude and mental strain: Implications for job redesign. Administrative Sci Q 1979; 24: 285308. 2. Henry JP, Stephens PM. Stress, health and the psychosocial environment: a sociobiological approach to medicine. New York: Springer, 1977. 3. Shively C, Laber-Laird K, Anton RF. Behavior and physiology of social stress and depression in female Cynomolgus monkeys. Biol Psychiatry 1997; 41: 871872.

4. McEwen BS. Protective and damaging effects of stress mediators. N Engl J Med 1998; 338: 171179. 5. Dallman MF, Akana SF, Scribner KA, Bradburg MJ, Walker C-D, Strack AM, Casio CS. Stress, feedback and facilitation in the hypothalamo-pituitary-adrenal axis. J Neuroendocrinol 1992; 4: 516526. 6. Plotsky PM, Cunningham ET Jr, Widmaier EP. Catecholaminergic modulation of corticotropin-releasing factor and adrenocorticotropin secretion. Endocr Rev 1989; 10: 437458. 7. Folkow B. Physiological aspects of primary hypertension. Physiol Rev 1982; 62: 384504. 8. Smyth J, Ockenfels MC, Porter L, Kirschbaum C, Hellhammer DH, Stone AA. Stressors and mood measured on a momentary basis are associated with salivary cortisol secretion. Psychoneuroendocrinology 1998; 23: 353371. 9. Frankenhaeuser M, Lundberg U, Fredrikson M, Melin B, Tuomisto M, Myrsten A-L, Hedman M, Bergman-Losman B, Wallin L. Stress on and off the job as related to sex and occupational status in white-collar workers. J Organizational Behav 1989; 10: 321346. 10. McKinnon W, Baum A, Morokoff P. Neuroendocrine measures of stress. In: Wagner HL (ed). Social Psychophysiology of Emotion: Theory and Clinical Application. Wiley, Chichester: 1988, pp 4364. 11. Lundberg U, Frankenhaeuser M. Pituitary-adrenal and sympathetic-adrenal correlates of distress and effort. J Psychosom Res 1980; 24: 125130. 12. Pruessner JC, Hellhammer DH, Kirschbaum C. Burnout, perceived stress, and cortisol responses to awakening. Psychosom Med 1999; 61: 197204. 13. Brunner E, Marmot MG, Nanchatal K, Shipley MJ, Stanseld SA, Juneja M. Social inequality in coronary risk: central obesity and the Metabolic Syndrome. Evidence from the Whitehall II study. Diabetologia 1997; 40: 13411349. 14. Rosmond R, Dallman MF, Bjrntorp P. Stress-related cortisol secretion in men: relationships with abdominal obesity and endocrine, metabolic and hemodynamic abnormalities. J Clin Endocr Metab 1998; 83: 18531859. 15. Kvist H, Chowdhury B, Grangrd U, Tyln U, Sjstrm L. Total and visceral adipose tissue Volumes derived from measurements with computed tomography in adult men and women: predictive equations. Am J Clin Nutr 1988; 48: 13511361. 16. Bjrntorp P. The regulation of adipose tissue regulation in humans. Int J Obes 1996; 20: 291302. 17. Chrousos GP, Gold PW. The concept of stress and stress system disorders. Overview of physical and behavioral homeostasis. JAMA 1992; 267: 12441252. 18. Rosmond R, Bjrntorp P. The interactions between hypothalamic-pituitary-adrenal axis activity, testosterone and insulin-like growth factor I and abdominal obesity with metabolism in men. Int J Obes Relat Disord 1998; 22: 11841196. 19. Lovejoy JC, Bray GA, Bourgeois MO, Macchiavelli R, Rood JC, Greeson C, Partington C. Exogenous androgens inuence on body composition and regional body fat distribution in obese menopausal women-A clinical research center study. J Clin Endocr Metab 1996; 81: 189203. 20. Polderman KH, Gooren LJG, Ascheman H. Induction of insulin resistance by androgens and estrogens. J Clin Endocr Metab 1994; 79: 265271. 21. Gold PW, Chrousos GP. The endocrinology of melancholic and atypical depression: Relation to neurocircuitry and somatic consequences. Proc Assoc Am Physicians 1999; 111: 2234. 22. Zakrzewska KE, Cusin I, Sainsbury A, Rohner-Jeanrenaud F, Jeanrenaud B. Glucocorticoids as counterregulatory hormones of

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

obesity reviews

Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

85

leptin. Toward and understanding of leptin resistance. Diabetes 1997; 46: 717719. 23. Considine RV, Sinha MK, Heiman ML, Kriauciunas A, Stephens TW, Nyce MR, Ohannesian JP, Marco CC, McKee LJ, Bauer TL, Caro JF. Serum immunoreactive leptin concentrations in normal-weight and obese humans. N Engl J Med 1996; 334: 292295. 24. Uotani S, Bjrbaek C, Torn J, Flier JS. Functional properties of leptin receptor isoforms. Internalization and degradation of leptin and ligand-induced receptor down-regulation. Diabetes 1999; 48: 279286. 25. Newcomer JW, Selke G, Melson AK, Gross J, Vogler GP, Dagogo-Jack S. Dose-dependent cortisol-induced increases in plasma leptin concentration in healthy humans. Arch Gen Psychiatry 1998; 55: 9951000. 26. Dagogo-Jack S, Selke L, Melson AK, Newcomer JW. Effect of dexamethasone on plasma leptin concentration in lean and obese humans: a randomised placebo-controlled study. J Clin Endocr Metab 1997; 82: 32303233. 27. Tattarini PA, Larson DE, Snitker S, Young JB, Flatt JP, Ravussin E. Effects of glucocorticoids on energy metabolism and food intake in humans. Am J Physiol 1996; 271: E317E325. 28. Bjrntorp P. Obesity and cortisol. Nutrition 2000; volume number. (In press.) 29. Zukowska-Grojec Z, Neuropeptide Y. A novel sympathetic stress hormone and more. Ann N Y Acad Sci 1995; 771: 219233. 30. Morgan CA, 3rd Wang S, Southwick SM, Rasmusson A, Hazlett G, Hauger RL, Charney DS. Plasma neuropeptide-Y concentrations in humans exposed to military survival training. Biol Psychiatry 2000; 47: 902909. 31. Guidi L, Tricerri A, Vangeli M, Frasca D, Riccardo Errani A, DiGiovanni A, Antico L, Menini E, Sciamanna V, Magnavita N, Doria G, Bartoloni C. Neuropeptide Y plasma levels and immunological changes during academic stress. Neuropsychobiology 1999; 40: 188195. 32. Pralong FP, Corder R, Gaillard RC. The effects of chronic glucocorticoid excess, adrenalectomy and stress on neuropeptide Y in individual rat hypothalamic nuclei. Neuropeptides 1993; 25: 223231. 33. Thorsell A, Svensson P, Wiklund L, Sommer W, Ekman R, Heilig M. Suppressed neuropeptide Y (NPY) mRNA in rat amygdala following restraint stress. Regul Pept 1998; 7576: 247254. 34. Corder R, Castagne V, Rivet JM, Mormede P, Gaillard RC. Central and peripheral effects of repeated stress and high NaCl diet on neuropeptide Y. Physiol Behav 1992; 52: 205210. 35. Greenoo CG, Wing RR. Stress-induced eating. Psychol Bull 1994; 115: 444464. 36. Morgan MJ. Effects of postweaning environment on learning in the rat. Anim Behav 1973; 21: 815817. 37. Bellisle F, Louis-Sylvestre J, Linet N, Rocaboy B, Dalle B, Cheneau F, LHinoret D, Guyot L. Anxiety and food intake in men. Psychosom Med 1990; 52: 452457. 38. Michaut C, Kahn JP, Musse N, Burlet C, Nicolas JP, Mejean L. Relationships between a critical life event and eating behaviour in high-school students. Stress Med 1990; 6: 5764. 39. Allison DB, Heshka S. Emotion and eating in obesity? A critical analysis. Int J Eating Disord 1993; 13: 289295. 40. Bjrntorp P. Visceral obesity. A Civilization Syndrome Obes Res 1993; 1: 206222. 41. Larsson B, Seidell J, Svrdsudd K, Welin L, Tibblin G, Wilhelmsen L, Bjrntorp P. Obesity, adipose tissue distribution and health in men- the study of men born, 1913, Appetite, 1989: 3744. 42. Lapidus L, Bengtsson C, Hllstrm T, Bjrntorp P. Obesity, adipose tissue distribution and health in women-results from a

population study in Gothenburg. Sweden Appetite 1989; 12: 2535. 43. Wing RR, Matthews KA, Muller LH, Meilahn EN, Platinga P. Waist to hip ratio in middle-aged women. Associations with behavioral and psychosocial factors and with changes in cardiovascular risk factors. Arterioscl Thromb 1991; 11: 1250 1257. 44. Rosmond R, Lapidus L, Mrin P, Bjrntorp P. Mental distress, obesity and body fat distribution in middle-aged men. Obes Res 1996; 4: 245252. 45. Rosmond R, Eriksson E, Bjrntorp P. Personality disorders in relation to anthropometric, endocrine and metabolic factors. J Endocrinol Invest 1999; 22: 279288. 46. Rosmond R, Bjrntorp P. Occupational status, cortisol secretory pattern and visceral obesity in middle-aged men. Obes Res 2000; 8: 445445. 47. Rosmond R, Bjrntorp P. Endocrine and metabolic aberrations in men with abdominal obesity in relation to anxiodepressive inrmity. Metabolism 1998; 47: 11871193. 48. Seidell J, Bjrntorp P, Sjstrm L, Sannerstedt R, Krotkiewski M, Kvist H. Regional distribution of muscle and fat mass in mnnew insight into the risk of abdominal obesity using computed tomography. Int J Obes 1989; 13: 289303. 49. Bjrntorp P, Holm G, Rosmond R, Folkow B. Hypertension and the Metabolic Syndrome: Closely Related Central Origin? Blood Press 2000; 9: 7182. 50. Bjrntorp P. Neuroendocrine perturbations as a cause of insulin resistance. Diab Met Res Rev 1999; 15: 427441. 51. Reaven GH. Role of insulin in human disease. Diabetes 1988; 37: 15951607. 52. Bjrntorp P. Epidemiology of the relationship between depression and physical illness. In: Thakore JH (ed). Depression a StressRelated Disorder: Physical Consequences and Potential Treatment Implications. Wrightson Biomed: England, in press. 53. Bjrntorp P. Hypothalamic origin of prevalent human disease. In: Pfaff D, Arnold A, Etgen A, Fahrbach S, Rubin R (eds). Hormones, Brain and Behavior. Academic Press: San Diego, in press. 54. Jayo JM, Shively CA, Kaplan JR, Manuck SB. Effects of exercise and stress on body fat distribution in male Cynomolgus monkeys. Int J Obes 1993; 17: 597604. 55. Chagnon YC, Prusse L, Weisnagel SJ, Rankinen T, Bouchard C. The human obesity gene map: the 1999 Update. Obes Res 2000: 89117. 56. Linkowski P, Van Onderbergen A, Kerkhofs M, Bosson D, Mendlewicz L, Van Cauter E. Twin study of the 24-h cortisol prole: evidence for genetic control of the human circadian clock. Am J Physiol 1993; 264: E 173181. 57. Rosmond R, Chagnon YC, Holm G, Chagnon M, Perusse L, Lindell K, Carlsson B, Bouchard C, Bjrntorp P. A glucocorticoid receptor gene marker is associated with abdominal obesity, leptin, and dysregulation of the hypothalamic-pituitary-adrenal axis. Obes Res 2000; 8: 211218. 58. Weaver JU, Hitman GA, Kopelman PG. An association between a Bcl I restriction fragment length polymorphism of the glucocorticoid receptor locus and hyperinsulinemia in obese women. J Mol Endocrinol 1992; 9: 295300. 59. Buemann B, Vohl MC, Chagnon M, Chagnon YC, Gagnon J, Perusse L, Dionne F, Despres JP, Tremblay A, Nadeau A, Bouchard C. Abdominal visceral fat is associated with a Bcl I restriction fragment length polymorphism at the glucocorticoid receptor gene locus. Obes Res 1997; 5: 186192. 60. Watt GCM, Harrap SB, Foy CJW, Holton DW, Edwards HV, Davidson HR. Abnormalities of glucocorticoid metabolism and the renin-angiotensin system: a four corners approach to identi-

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386

86 Do stress reactions cause abdominal obesity and comorbidities

P. Bjrntorp

obesity reviews

cation of genetic determinants of blood pressure. J Hypertens 1992; 10: 473482. 61. Rosmond R, Chagnon YC, Chagnon M, Perusse L, Bouchard C, Bjrntorp PA. polymorphism on the 5-anking region of the glucocorticoid receptor gene locus is associated with basal cortisol secretion in men. Metabolism 2000; (In print.) 62. Rosmond R, Chagnon M, Bouchard C, Bjrntorp P. A Tsp5091 polymorphism in exon 2 of the glucocorticoid receptor gene in relation to obesity and cortisol secretion. BM J 2000; (In print.) 63. Lin RC, Wang Wy Morris BJ. High penetrance, overweight, and glucocorticoid receptor variant: a case-control study. BMJ 1999; 319: 13371338.

64. Comings DE. Polygenic inheritance and micro/minisattelites. Mol Psychiatry 1998; 3: 2131. 65. Dunning AM, Healey CS, Pharaoah PD, Teare MD, Ponder PA, Easton DF. A systematic review of genetic polymorphisms and breast cancer risk. Cancer Epidemiol Biomarkers Prev 1999; 19: 843854. 66. Rosmond R, Chagnon YC, Holm G, Chagnon M, Prusse L, Lindell K, Carlsson B, Bouchard C, Bjrntorp P. Hypertension in obesity and the leptin receptor gene locus. J Clin Endocr Metab 2000; (In print.) 67. Haynes WG, Morgan DA, Walsh SA, Mark AL, Sivitz WI. Receptor mediated regional sympathetic nerve activation by Leptin. J Clin Invest 1997; 100: 270278.

2001 The International Association for the Study of Obesity. obesity reviews 2, 7386