- Email: [email protected]
Is cortisol involved in upper-body obesity?
MedicaI Hypotheses (1989) 30,95-100 0 Lcmman Grouo UK Ltd 1989
Is Cortisol Involved in Upper-Body Obesity? M. P. RIVERA and F. SVEC Section of Endocrinology, Department Orleans, LA 707 72-2865, USA
of Medicine,
LSU Medical
School,
1542 Tulane Avenue,
New
Abstract - Obesity is a major health problem that can be defined as an excess of body fat, associated with hypertension, diabetes and coronary heart disease. Several groups have evaluated the clinical significance of variations in fat cell distribution on these complications. A frequently used index of fat cell distribution is the waist to hips ratio (W/H). A high W/H ratio is said to reflect upper body fat cell distribution while a low waist to hips ratio reflects a lower body type fat cell distribution. Studies have shown that those whose W/H ratio indicate upper body fat cell distribution had a higher prevalence of diabetes and hypertension than those with the lower type. Over the years cortisol has attracted considerable interest as a possible factor in the development and maintenance of obesity. The clinical findings associated with upper body type of obesity are in many ways similar to those of the hypercortisol state. Our hypothesis is that upper body obesity forms a unique subgroup of the obese population and their regional fat distribution is associated with mild cortisol excess. In humans, studies have reported that some obese subjects hypersecrete cortisol and have an increase in the cortisol production rate. Although recent studies would tend to discount any influence of cortisol in human obesity, several factors should be taken into consideration. It is difficult to measure cortisol economy in obese subjects because among other things the measurements are less than precise; and cortisol secretion changes during the day and in response to outside stimuli. Further, obesity is a heterogeneous disorder and not all obese subjects may have the same disorder. Studies showing no difference in cortisol economy have frequently excluded diabetic and hypertensive subjects, most of which exhibit the upper body type of fat cell distribution. By exclusion of these subjects previous studies may have missed a most interesting subgroup and selected toward a study population which does not have upper body obesity. As yet there are no studies which compare the cortisol economy of obese patients who have upper body obesity to those who have lower body obesity. This study should be done not only over a 24 hr. period, but in response to specific stimuli (like meals) since a hormonal abnormality may not necessarily be present at all times.
9.5
96
MEDICALHYPOTHESES
Introduction Obesity is a major health problem in developed countries. It is clearly associated with a higher rate of morbidity and mortality. A survey conducted by the National Center for Health Statistics found a strong association between the incidence of hypertension, hypercholesteremia, diabetes mellitus with body mass index (BMI).(l) The Framingham Study also reported a direct association between the degree of obesity and the occurance of coronary heart disease independent of other risk factors.(2) Third, there is evidence from the American Cancer Society’s studies that there is a higher mortality in obese females who have cancer of the endometrium. breast, cervix. ovary, and biiiary system as compared to their nonobese counterparts.(l, 3) In obese males a similar association was found between cancer of the colon, rectum, and prostate. Lastly, mortality rates in general were found to increase directly with the degree of obesity in the general population.(l, 3) Although one may expect that the degree of obesity needed to cause these increased risks would be large, in actuality an increased risk appears when a subject reaches 20% above ideal body weight. Thus, even at this seemingly minimal amount of obesity there is enough excess risk to justify clinical intervention. Obesity is most easily defined as an excess in body fat. However, the amount of excess fat, is not always easy to determine. Consequently, a series of indirect measurements have evolved for judging the excess of body fat. For example, tables exist for ideal height/weight values, body mass indices, and proportions of fat per total body weight. Although these tables and indices may be adequate retlections of the amount of excess body fat, they still do not take into account differences in the distribution of the fat deposits. Evidence is now being presented that differences in fat distribution may be more important predictors of coronary heart disease, hypertension, and diabetes than the degree of obesity itself. Regional
fat distribution
Adipose location. different vascular enzymes
tissue varies according to its anatomic It has been pointed out that fat cells are in the degree of innervation, volume of bed, quality of surface, concentration of and other intracellular material
depending on their location in the body.(d, 5) Recently, several groups have tried to evaluate the clinical significance of variations in fat cell distribution. At this point the best way of measuring and reporting body fat variation is not yet known. Some have used calipers to determine regional skin fold thickness. Another approach involves the determination of the waist girth and comparing it to the hip girth. A high ratio is said to reflect an upper-body type of fat distribution (androgenoid) while a low waist to hip ratio is said to reflect a lower type body obesity (gynecoid proportions). The potential importance of this particular variation in fat distribution was recently pointed out in a study done by Hartz, Rupley and Rimm. Using the information obtained from a national survey of 32 856 women who were members of a weight reduction organization known as TOPS (Take Off Pounds Sensibly), they found that those whose waist/hips ratios indicated they had upper-body obesity had a higher prevalence of diabetes and hypertension than those with a lower-body distribution.(4) In some cases the effect of fat distribution was quite striking; the incidence of diabetes in a group of moderately obese women increased from 3% in those patients who had lower-body proportions, to nearly a 10% prevalence in those who had an equal degree of adiposity but had upper-body proportions. This trend in general, held true regardless of the degree of obesity. Even those who were only mildly obese suffered an increase incidence of obesity related problems, if their excess adiposity was in the upper part of their bodies.(4) Other studies confirm the importance of regional fat cell distribution. For example, similar findings were obtained by Blair et ai. who used skinfold thickness as their measure of regional fat distribution.(6) Likewise, Kalkhoff et al. evaluated 110 obese women to determine whether there was a relationship between fat cell distribution. blood pressure, glucose tolerance, and plasma insulin levels. They concluded that in healthy obese women there was a continuum that related increasing fat accumulation in the upper part of the body and a progressively higher blood pressure, reduced carbohydrate tolerance. and increased plasma insulin levels(7) In studies we have done in black females we have found that the prevalence of diabetes and hypertension increases significantly with upper
IS CORTISOL
INVOLVED
IN UPPER-BODY
OBESITY?
type of fat cell distribution. The waist to hip ratio used as an index of fat cell distribution in our study was a better indicator of risk than weight. Thus, although obesity can be described in general as an excess in body fat, its medical significance depends not only on its amount but also on its distribution. Consequently, as much experimental effort should be devoted towards determining those factors which control the locution of fat cell development as is devoted to elucidating those factors which contribute to the degree of obesity The role of hormones in obesity ’ The pathophysiology of the genesis and maintenance of obesity is not clear. Its understanding will probably require the dissection of genetic, environmental, psychic, hormonal. and metabolic influences. Already a great deal of research has been directed towards the evaluation of whether hormonal abnormalities can be factors in the development of obesity. There is no final agreement as yet to the exact role of hormones in the development and maintenance of obesity and in many cases the data is conflicting. For example, serum Ti has been reported to be or decreased in different normal, elevated, studies of the obese population.(S) Leaving aside variations in technique of hormone analysis, these divergent results, suggest that there may be great differences between study populations. From what was described in the previous section, one might suggest that variations in regional fat distribution might be one factor that could alter results if it is not taken into account and controlled. Those with upper-body fat distribution appear to be at the greatest risk of the morbid consequences of obesity - hypertension and diabetes. Hartz, et al., proposed in their work that this fat distribution and associated morbidity was caused by the abnormal production of a steroid.(4) Which steroid, however, was not known and they suggested a sex steroid. This hypothesis, however, has major flaws because there is no clear reason androgens should be so detrimental to metabolism in women. Another possibility is that the steroid hormone causing the morbid complications of upper-body obesity is cortisol. a steroid already known to be involved in carbohydrate metabolism and the genesis of hypertension. Over the years cortisol has attracted consider-
97 able interest as a possible factor in the development and maintenance of obesity. This is probably because the recognized states of hypercortisolism are so clearly characterized as having glucose intolerance, arterial hypertension, and an obesity which is of a central distribution. The clinical findings associated with upper-body obesity are in many ways similar to those found in the hypercortisol state. In this paper it is hypothesized that upperbody obesity subjects form a unique subgroup of the obese population and their regional fat distribution is associated with mild cortisol excess. The basis of this hypothesis is not only the clinical similarity of the two states. but also a series of observations derived from both basic science and clinical studies. We will try to establish first, that cortisol is involved in obesity and second, demonstrate that it is involved in upperbody obesity. To begin with. studies using the Zucker rat, a model of obesity. have recently reported that the rat’s obesity is associated with hypercortisolemia and a diminished pituitary responsiveness to CRF.(9) The hypercortisolism in this well studied model was thus only recently found. This suggests that at least one component of this genetically determined obesity is a metabolic defect of central nervous system regulation of adrenal function leading to hypercortisolism. As one might predict, adrenalectomy reversed this type obesity.(9, 10) Similar results were found using OB/OB mice. These are mutants of the CS7BL/6J mouse and show hyperphagia, obesity and diabetes, with high circulating insulin levels. Dubug et al. showed that adrenalectomy is an effective tool for ameliorating the severity of many aspects of this syndrome in the OB/OB mice. Importantly, glucocorticoid receptors have been found in both white and brown adipose tissue indicating that both are glucocorticoid target organs(12) and, it has been shown that in vitro differentiation of human adipocyte precursor cells into mature fat cells is triggered by a combination of cortisol and insulin.( 13) This response to cortisol is dose dependent suggesting that glucocorticoids play a role in the development of human hyperplasic obesity by stimulating the formation of adipocytes from precursor cells when insulin is present. The potential significance of this synergistic relationship between glucocorticoids and insulin will be returned to later.
98 Besides these basic science studies, there are clinical studies which give impetus to the concept that cortisol is involved in human obesity. Sinkin, Arendt, and Pattee reported that some obese subjects hypersecrete cortisol.(l4, 15) Approximately one-third of their obese patients secreted excess 17 ketogenic steroids (metabohtes of cortisol) as compared to non-obese controls. Mlynaryk, Gillies, Murphy, and Pattee similarly showed that the cortisol production rate was increased in the obese using a C-14 isotope dilution method.( 16) Although these studies show that cortisol production was elevated in the obese, this is not widely accepted as an indication of pathology because critics point out it is difficult to compare obese to nonobese. Subsequent research in this area quickly moved to finding the best manner in which values could be compared between obese and non-obese subjects. Many suggested that the increased production of steroids in the obese simply reflect an increase in the demands of the larger body without any pathophysiological implication. Realizing this, Migeon, Green and Eckert studied the excretion of 17-hydroxycorticosteroids (metabolites of cortisol) and cortisol production rates in 160 obese patients and 180 normal subjects.( 17) Their comparisons were done in various ways to try to normalize the data. First, it was shown that the cortisol production rate increased with increasing body weight. When cortisol production rate was expressed in terms of surface area per 24 hours, there was only a small range of values for the non-obese subjects, however, the results did not form such a tight band for the obese, even when corrected for surface area. Further, the researchers showed that the mean and standard deviations were not the. same between the two groups suggesting a state of hypercortisolism in some of the obese patients.(l7) There was a great deal of overlap between the two groups and only a fraction of the obese patients were more than two standard deviations above the mean of the non-obese group. This scatter may indicate that obesity is a hetrogeneous disorder and that not all obese patients should be grouped together. Importantly at the time of that study, no evaluation of regional fat distribution was made. Other groups have evaluated cortisol economy in the obese. They have used different parameters by which the obese and non-obese groups were compared. Strain et al., for example, deter-
MEDICAL HYPOTHESES
mined the cortisol production rate in 74 healthy women volunteers.(l8) They used the 24 hour urine creatinine excretion as a measure of lean body mass and expressed the cortisol production rate per gram of urinary creatinine. They found that when expressed this way cortisol secretion was invariate in both sexes and over a wide range of body weights. Thus, they concluded that there was no significant abnormality in cortisol production in the obese. Since that report most studies have tended to confirm these findings. A study done by Streenten et al., evaluated 28 (twenty-eight) obese healthy subjects and compared them to forty (40) lean subjects.(l9) They found no difference in excretion of 17hydroxycorticosteroids and cortisol production rate when normalized by urinary creatinine. and no difference in urinary free cortisol and their response to dexamethasone in both groups. Although these most recent studies would tend to discount any influence of cortisol in human obesity the subject is not as yet answered completely. To begin with, the techniques used to measure cortisol economy are less than precise. Seventeen hydroxysteroids and 17 ketogenicsteroids are only an approximate reflection of cortisol production and subject to variables besides cortisol secretion rate. Additionally, although the C-14 isotope dilution studies used by many investigators seems more accurate, the most recent evidence by an NIH group shows that these values, too, may be subject to error and, as usually determined, are nearly twice as high as the actual values. Whether the results of this technique vary with body size and proportion is not yet known and thus as yet the overall secretion of cortisol in the obese is not known with certainty, especially among those with different body proportions. Next, the role of study group bias has to be considered. It is noteworthy that Migeon, et a1.(17) evaluated a large group of patients, many of which were diabetic and hypertensive. On the other hand, Strain, et al., (18) and Streenten (19) studied a population which excluded these two groups. As the prevalence of diabetes and hypertension is higher in groups with upper-body obesity, it is possible that by excluding those with these two disorders study populations were skewed towards those with lower-body proportions. Thus some research groups might have missed the most interesting subgroup of the obese; those with upper-body obesity. Many researchers discount the possibility that
IS CORTISOL
INVOLVED
IN UPPER-BODY
OBESITY?
99
Thus, it is possible that the abnormality of obesity is associated with hypercortisolism: They would concede that only those with classic cortisol production may be present for only a Cushing’s syndrome have a glucocorticoid imbal- short time in the day and its demonstration obscured by measurements which reflect 24 hour ance with a clear adrenal pituitary abnormality. metabolism. However, there are at least three other states According to our hypothesis some of the states recently described which are associated with hypercortisolism besides classic Cushing’s. For of obesity are associated with hypercortisolism. example, in the last several years it has been This raises the question as to why this does not documented that pregnancy is associated with a exert a negative feedback on the hypothalomus and pituitary gland to inhibit the hyperproducmild degree of hypercortisolism.(20) Pregnancy tion of cortisol. One possibility to account for may be associated with hypertension, glucose this is that these states of obesity are refractory intolerance, and increased fat stores. Second, to cortisol at the level of the hypothalamus many chronic alcoholics have hypercortisolism and this, too. is associated with abnormalities in and/or pituitary gland. There is information from animal studies that this indeed may be the case. blood pressure and carbohydrate metabolism. This syndrome is induced by alcohol intake and Webb, et al., analyzed the binding of tritiated triamcinolone acetonide to cytosolic receptors resolved with abstinence. Finally, depressive from brains and liver homogenates of misty states are associated with hypercortisolism.(21) diabetic mice (Mdb/Mdb) and reported that the In each of these cases the presence of hypercortisolism has been documented only recently; it diabetic mice had a decreased number of recephad been overlooked by many. Importantly, in tors for glucocorticoids compared to appropriately Further, the steroid these three states the hypercortisolism is tran- matched controls.(23) sient although some of the stigmata may last receptor complexes from the diabetic mice showed a decreased ability to form the activated, longer than the hormonal abnormality. nuclear binding form. Thus, these animals Just as an abnormality of cortisol secretion may not be present through the entire life of a showed a decreased sensitivity in these tissues to the actions of glucocorticoids. pregnant, depressed, or alcoholic patient, so too, one may also hypothesize that the production of cortisol may not have to be abnormal throughout Summary an entire day to lead to abnormalities in metabolism. Thus the time at which observations are Many hormonal abnormalies have been demonmade might be important. Cortisol is secreted in strated in obesity. Which if any, is involved in the genesis and maintenance of the obese state a circadian rhythm. Many studies of cortisol is not yet known. Recent evidence indicates that economy involve assays of morning fasting blood levels. a time at which cortisol is thought to be obesity is not a simple monolithic disorder but instead is made up of multiple subtypes. Within highest. However, it is now becoming evident that cortisol may peak at other times in the day these subtypes it is quite possible that certain besides this morning surge. J. H. Liu et al. have groups reflect the result of discrete hormonal abnormalities. It is also becoming evident that recently reported that cortisol increases dramatically after food intake. especially lunch.(22) It the obese should not be grouped together simply by the degree above ideal body weight; discrimmay be that in some genetically susceptible individuals there is a higher postpandrial peak . ination by the location of fat cells is also than normal. This abnormality over a prolonged important. period could cause enough of a change to be Although early studies suggested that cortisol metabolically significant. It is important to note may be hypersecreted in obesity, more recent that the postpyrandrial surge of cortisol after studies disagree with this finding. However, as lunch is different from the morning cortisol surge noted, these studies were frequently designed to in that insulin levels are high in the former and exclude those with diabetes and hypertension low in the latter. As noted before, the simul- (which may be integral findings in some forms taneous presence of cortisol and insulin can of obesity) and may be skewed toward study trigger the formation of adipocytes from populations which do not have upper-body precursor cells. The combination of an elevated obesity. As yet there are no studies which insulin and cortisol may mediate the developcompare the cortisol economy of obese patients ment of hyperplastic obesity.(l3) who have upper-body obesity to those who have
100 lower-body obesity. These studies need to be done with the inclusion of those who have obesity-associated hypertension and diabetes. Studies may need to be done analyzing not only samples collected over a 24 hour period, but also need to involve detailed analyses of cortisol economy over shorter time periods with the stimulus of meals. Indeed, as the hormonal abnormality may not necessarily be present throughout the subject’s entire life, it may be that longitudinal studies are needed at various times through periods of weight gain and weight loss.
References B R. Foster W R. Hirsch T. Van Irallie T B. I. Burton Health implications of obesity. International Journal of Obesity 9: 155-169, 1985. P M. 2. Garrison R J. Feinleih M. Castelli W P, McNamara Cigarette smoking as a confounder of the relationship between relative weight and long term mortality: The framington heart study. J. Am. Med. Assoc. 249: 2199-2203, 1983. L. Variations in mortality by weight 3. Lew E A. Garfinkle among 750 000 men and women. J. Chron. Dis. 32: 563-76. 1979. of 4. Hartz A J. Rupley D C, Rimm A A. The association girth measurements with disease in 32 856 women. Am. J. of Epidemio 119: 71-80. 1984. E. Wertheimer E. Comparative physiology of 5. Shafrir adipose tissue in different sites and in different species. p 427-29 in Handbook of physiology Section 5: Adipose Tissue. (Renold A E, and Cahill G F. eds.) American Physiological Society, Washington, D.C. 1965. 6. Blair D, Habicht J P. Sims E A, Sylwester D, Abrahan S. Evidence for an increased risk of hypertension with centrally located body fat and the effect of race and sex on this risk. Am. J. Epidemio 119: 526-40, 1984. R K. Hartz A H, Rupley D, Kissebah A H. 7. Kalkhoff Kelber S. Relationship of body fat distribution to blood carbohydrate tolerance. and plasma lipids in pressure, healthv obese women. J. Lab Clin. Med. 102: 621-27. 1983. ’ 8. Jung R. Endocrinological aspects of obesity. Clinics in Endocrinology and Metabolism 13: 597-611, 1984. 9. Cunningham J J, Calles-Escandon J. Garrido F. Carr D B. Bode H H. Hypercorticostenuria and diminished
MEDICAL
HYPOTHESES
pituitarv responsiveness to CRF in obese Zucker Rats. Endocrinology 118: 98-101, 1986. Freedman M R. Horwitz B A. Stern J S. Effect of adrenalectomy and glucocorticoid replacement on development of obesity. Am. J. Physio. 249: R595-R6U7. 1986. 11. Dubuc P U. Wilden N J. Adrenalectomy reduces but does not reverse obesity in OB/OB mice. Int. J. Obes. 10: 91-Y8. 1986. 12. Feldman D. Evidence that brown adipose tissue is a glucocorticoid target organ. ‘Endocrinology 103: 2091-2097. 1978. 13. Hauner H, Schmid P. Pfeiffer E R. Glucocorticoids and insulin promote the differentiation of human adipocyte precursor cells into fat cells. J. Clin. Endo and Metabol. 64: 832-835, 1987. 14. Simkin B. Urinary 17-Ketosteroid and l7-Ketogenic steroid excretion in patients. New England J. of Med. 264: 974-77. 1961. 1.5.Szenas P. Pattee C J. Studies of adrenocortical function in obesity. J. Clin Endocrinol Metab 19: 344-350. 1959. 16. Mlynaryk P. Gillies R R. Murphy B, Pattee C J. Cortisol production rates in obesity. Journal of Clin Endocrinol & Metab 22: 587-591, 1962. 17. Migeon C J. Green 0 C, Eckert J P. Study of adrenocortical function in obesity. Metabolism 12: 718-738, 1963. IX. Strain G W. Zumoff B, Strain J J, Levin J. Fukushima D K. Cortisol production in obesity. Metabolism 29: 980-85. 1980. D H. Stevenson C T. Dalakos T G. Dennick 19. Streeten L G. Fellerman H. The diagnosis of hypercortisolism: Biochemical criteria differentiating patients from lean and obese normal subjects and from females on oral contraceptives. J Clin Endocrin 29: 1191-1211. 1969. M D. Rueckert S 0. Ehrlich 20. Nolten W E. Lindheimer E N. Diurnal patterns and regulation of cortisol secretion in pregnancy. J Clin Endocrin Metabol 51: 466-472. 1980. D D. Yen 21. Mortola J F, Liu J H. Gillin J C. Rasnaussen S S. Pulsatile rhythms of adrenocorticotropin (ACTH) and cortisol in women with endogenous depression: evidence for increased ACTH pulse frequency. J Clin Endocrin Metabol 65: 962-68. 1967. D D. Characterization 22. Liu J H, Kazer R R. Rusnussen of the twenty-four hour secretion patterns of andrenocorticotropin and cortisol in normal women and patients with Cushing’s Disease. J Clin Endrocrin Metabol 64: 1027-1035. 1987. 23. Webb M L, Hvnn J J, Schmidt T J, Margules D L. and and Litwack G. Decreased glucocorticoidbindin receptor activation in genetically diabetic (MDB f MDB) mice. J Steroid Biochem 25: 645-57. 1986.
,o.
Is Cortisol Involved in Upper-Body Obesity? M. P. RIVERA and F. SVEC Section of Endocrinology, Department Orleans, LA 707 72-2865, USA
of Medicine,
LSU Medical
School,
1542 Tulane Avenue,
New
Abstract - Obesity is a major health problem that can be defined as an excess of body fat, associated with hypertension, diabetes and coronary heart disease. Several groups have evaluated the clinical significance of variations in fat cell distribution on these complications. A frequently used index of fat cell distribution is the waist to hips ratio (W/H). A high W/H ratio is said to reflect upper body fat cell distribution while a low waist to hips ratio reflects a lower body type fat cell distribution. Studies have shown that those whose W/H ratio indicate upper body fat cell distribution had a higher prevalence of diabetes and hypertension than those with the lower type. Over the years cortisol has attracted considerable interest as a possible factor in the development and maintenance of obesity. The clinical findings associated with upper body type of obesity are in many ways similar to those of the hypercortisol state. Our hypothesis is that upper body obesity forms a unique subgroup of the obese population and their regional fat distribution is associated with mild cortisol excess. In humans, studies have reported that some obese subjects hypersecrete cortisol and have an increase in the cortisol production rate. Although recent studies would tend to discount any influence of cortisol in human obesity, several factors should be taken into consideration. It is difficult to measure cortisol economy in obese subjects because among other things the measurements are less than precise; and cortisol secretion changes during the day and in response to outside stimuli. Further, obesity is a heterogeneous disorder and not all obese subjects may have the same disorder. Studies showing no difference in cortisol economy have frequently excluded diabetic and hypertensive subjects, most of which exhibit the upper body type of fat cell distribution. By exclusion of these subjects previous studies may have missed a most interesting subgroup and selected toward a study population which does not have upper body obesity. As yet there are no studies which compare the cortisol economy of obese patients who have upper body obesity to those who have lower body obesity. This study should be done not only over a 24 hr. period, but in response to specific stimuli (like meals) since a hormonal abnormality may not necessarily be present at all times.
9.5
96
MEDICALHYPOTHESES
Introduction Obesity is a major health problem in developed countries. It is clearly associated with a higher rate of morbidity and mortality. A survey conducted by the National Center for Health Statistics found a strong association between the incidence of hypertension, hypercholesteremia, diabetes mellitus with body mass index (BMI).(l) The Framingham Study also reported a direct association between the degree of obesity and the occurance of coronary heart disease independent of other risk factors.(2) Third, there is evidence from the American Cancer Society’s studies that there is a higher mortality in obese females who have cancer of the endometrium. breast, cervix. ovary, and biiiary system as compared to their nonobese counterparts.(l, 3) In obese males a similar association was found between cancer of the colon, rectum, and prostate. Lastly, mortality rates in general were found to increase directly with the degree of obesity in the general population.(l, 3) Although one may expect that the degree of obesity needed to cause these increased risks would be large, in actuality an increased risk appears when a subject reaches 20% above ideal body weight. Thus, even at this seemingly minimal amount of obesity there is enough excess risk to justify clinical intervention. Obesity is most easily defined as an excess in body fat. However, the amount of excess fat, is not always easy to determine. Consequently, a series of indirect measurements have evolved for judging the excess of body fat. For example, tables exist for ideal height/weight values, body mass indices, and proportions of fat per total body weight. Although these tables and indices may be adequate retlections of the amount of excess body fat, they still do not take into account differences in the distribution of the fat deposits. Evidence is now being presented that differences in fat distribution may be more important predictors of coronary heart disease, hypertension, and diabetes than the degree of obesity itself. Regional
fat distribution
Adipose location. different vascular enzymes
tissue varies according to its anatomic It has been pointed out that fat cells are in the degree of innervation, volume of bed, quality of surface, concentration of and other intracellular material
depending on their location in the body.(d, 5) Recently, several groups have tried to evaluate the clinical significance of variations in fat cell distribution. At this point the best way of measuring and reporting body fat variation is not yet known. Some have used calipers to determine regional skin fold thickness. Another approach involves the determination of the waist girth and comparing it to the hip girth. A high ratio is said to reflect an upper-body type of fat distribution (androgenoid) while a low waist to hip ratio is said to reflect a lower type body obesity (gynecoid proportions). The potential importance of this particular variation in fat distribution was recently pointed out in a study done by Hartz, Rupley and Rimm. Using the information obtained from a national survey of 32 856 women who were members of a weight reduction organization known as TOPS (Take Off Pounds Sensibly), they found that those whose waist/hips ratios indicated they had upper-body obesity had a higher prevalence of diabetes and hypertension than those with a lower-body distribution.(4) In some cases the effect of fat distribution was quite striking; the incidence of diabetes in a group of moderately obese women increased from 3% in those patients who had lower-body proportions, to nearly a 10% prevalence in those who had an equal degree of adiposity but had upper-body proportions. This trend in general, held true regardless of the degree of obesity. Even those who were only mildly obese suffered an increase incidence of obesity related problems, if their excess adiposity was in the upper part of their bodies.(4) Other studies confirm the importance of regional fat cell distribution. For example, similar findings were obtained by Blair et ai. who used skinfold thickness as their measure of regional fat distribution.(6) Likewise, Kalkhoff et al. evaluated 110 obese women to determine whether there was a relationship between fat cell distribution. blood pressure, glucose tolerance, and plasma insulin levels. They concluded that in healthy obese women there was a continuum that related increasing fat accumulation in the upper part of the body and a progressively higher blood pressure, reduced carbohydrate tolerance. and increased plasma insulin levels(7) In studies we have done in black females we have found that the prevalence of diabetes and hypertension increases significantly with upper
IS CORTISOL
INVOLVED
IN UPPER-BODY
OBESITY?
type of fat cell distribution. The waist to hip ratio used as an index of fat cell distribution in our study was a better indicator of risk than weight. Thus, although obesity can be described in general as an excess in body fat, its medical significance depends not only on its amount but also on its distribution. Consequently, as much experimental effort should be devoted towards determining those factors which control the locution of fat cell development as is devoted to elucidating those factors which contribute to the degree of obesity The role of hormones in obesity ’ The pathophysiology of the genesis and maintenance of obesity is not clear. Its understanding will probably require the dissection of genetic, environmental, psychic, hormonal. and metabolic influences. Already a great deal of research has been directed towards the evaluation of whether hormonal abnormalities can be factors in the development of obesity. There is no final agreement as yet to the exact role of hormones in the development and maintenance of obesity and in many cases the data is conflicting. For example, serum Ti has been reported to be or decreased in different normal, elevated, studies of the obese population.(S) Leaving aside variations in technique of hormone analysis, these divergent results, suggest that there may be great differences between study populations. From what was described in the previous section, one might suggest that variations in regional fat distribution might be one factor that could alter results if it is not taken into account and controlled. Those with upper-body fat distribution appear to be at the greatest risk of the morbid consequences of obesity - hypertension and diabetes. Hartz, et al., proposed in their work that this fat distribution and associated morbidity was caused by the abnormal production of a steroid.(4) Which steroid, however, was not known and they suggested a sex steroid. This hypothesis, however, has major flaws because there is no clear reason androgens should be so detrimental to metabolism in women. Another possibility is that the steroid hormone causing the morbid complications of upper-body obesity is cortisol. a steroid already known to be involved in carbohydrate metabolism and the genesis of hypertension. Over the years cortisol has attracted consider-
97 able interest as a possible factor in the development and maintenance of obesity. This is probably because the recognized states of hypercortisolism are so clearly characterized as having glucose intolerance, arterial hypertension, and an obesity which is of a central distribution. The clinical findings associated with upper-body obesity are in many ways similar to those found in the hypercortisol state. In this paper it is hypothesized that upperbody obesity subjects form a unique subgroup of the obese population and their regional fat distribution is associated with mild cortisol excess. The basis of this hypothesis is not only the clinical similarity of the two states. but also a series of observations derived from both basic science and clinical studies. We will try to establish first, that cortisol is involved in obesity and second, demonstrate that it is involved in upperbody obesity. To begin with. studies using the Zucker rat, a model of obesity. have recently reported that the rat’s obesity is associated with hypercortisolemia and a diminished pituitary responsiveness to CRF.(9) The hypercortisolism in this well studied model was thus only recently found. This suggests that at least one component of this genetically determined obesity is a metabolic defect of central nervous system regulation of adrenal function leading to hypercortisolism. As one might predict, adrenalectomy reversed this type obesity.(9, 10) Similar results were found using OB/OB mice. These are mutants of the CS7BL/6J mouse and show hyperphagia, obesity and diabetes, with high circulating insulin levels. Dubug et al. showed that adrenalectomy is an effective tool for ameliorating the severity of many aspects of this syndrome in the OB/OB mice. Importantly, glucocorticoid receptors have been found in both white and brown adipose tissue indicating that both are glucocorticoid target organs(12) and, it has been shown that in vitro differentiation of human adipocyte precursor cells into mature fat cells is triggered by a combination of cortisol and insulin.( 13) This response to cortisol is dose dependent suggesting that glucocorticoids play a role in the development of human hyperplasic obesity by stimulating the formation of adipocytes from precursor cells when insulin is present. The potential significance of this synergistic relationship between glucocorticoids and insulin will be returned to later.
98 Besides these basic science studies, there are clinical studies which give impetus to the concept that cortisol is involved in human obesity. Sinkin, Arendt, and Pattee reported that some obese subjects hypersecrete cortisol.(l4, 15) Approximately one-third of their obese patients secreted excess 17 ketogenic steroids (metabohtes of cortisol) as compared to non-obese controls. Mlynaryk, Gillies, Murphy, and Pattee similarly showed that the cortisol production rate was increased in the obese using a C-14 isotope dilution method.( 16) Although these studies show that cortisol production was elevated in the obese, this is not widely accepted as an indication of pathology because critics point out it is difficult to compare obese to nonobese. Subsequent research in this area quickly moved to finding the best manner in which values could be compared between obese and non-obese subjects. Many suggested that the increased production of steroids in the obese simply reflect an increase in the demands of the larger body without any pathophysiological implication. Realizing this, Migeon, Green and Eckert studied the excretion of 17-hydroxycorticosteroids (metabolites of cortisol) and cortisol production rates in 160 obese patients and 180 normal subjects.( 17) Their comparisons were done in various ways to try to normalize the data. First, it was shown that the cortisol production rate increased with increasing body weight. When cortisol production rate was expressed in terms of surface area per 24 hours, there was only a small range of values for the non-obese subjects, however, the results did not form such a tight band for the obese, even when corrected for surface area. Further, the researchers showed that the mean and standard deviations were not the. same between the two groups suggesting a state of hypercortisolism in some of the obese patients.(l7) There was a great deal of overlap between the two groups and only a fraction of the obese patients were more than two standard deviations above the mean of the non-obese group. This scatter may indicate that obesity is a hetrogeneous disorder and that not all obese patients should be grouped together. Importantly at the time of that study, no evaluation of regional fat distribution was made. Other groups have evaluated cortisol economy in the obese. They have used different parameters by which the obese and non-obese groups were compared. Strain et al., for example, deter-
MEDICAL HYPOTHESES
mined the cortisol production rate in 74 healthy women volunteers.(l8) They used the 24 hour urine creatinine excretion as a measure of lean body mass and expressed the cortisol production rate per gram of urinary creatinine. They found that when expressed this way cortisol secretion was invariate in both sexes and over a wide range of body weights. Thus, they concluded that there was no significant abnormality in cortisol production in the obese. Since that report most studies have tended to confirm these findings. A study done by Streenten et al., evaluated 28 (twenty-eight) obese healthy subjects and compared them to forty (40) lean subjects.(l9) They found no difference in excretion of 17hydroxycorticosteroids and cortisol production rate when normalized by urinary creatinine. and no difference in urinary free cortisol and their response to dexamethasone in both groups. Although these most recent studies would tend to discount any influence of cortisol in human obesity the subject is not as yet answered completely. To begin with, the techniques used to measure cortisol economy are less than precise. Seventeen hydroxysteroids and 17 ketogenicsteroids are only an approximate reflection of cortisol production and subject to variables besides cortisol secretion rate. Additionally, although the C-14 isotope dilution studies used by many investigators seems more accurate, the most recent evidence by an NIH group shows that these values, too, may be subject to error and, as usually determined, are nearly twice as high as the actual values. Whether the results of this technique vary with body size and proportion is not yet known and thus as yet the overall secretion of cortisol in the obese is not known with certainty, especially among those with different body proportions. Next, the role of study group bias has to be considered. It is noteworthy that Migeon, et a1.(17) evaluated a large group of patients, many of which were diabetic and hypertensive. On the other hand, Strain, et al., (18) and Streenten (19) studied a population which excluded these two groups. As the prevalence of diabetes and hypertension is higher in groups with upper-body obesity, it is possible that by excluding those with these two disorders study populations were skewed towards those with lower-body proportions. Thus some research groups might have missed the most interesting subgroup of the obese; those with upper-body obesity. Many researchers discount the possibility that
IS CORTISOL
INVOLVED
IN UPPER-BODY
OBESITY?
99
Thus, it is possible that the abnormality of obesity is associated with hypercortisolism: They would concede that only those with classic cortisol production may be present for only a Cushing’s syndrome have a glucocorticoid imbal- short time in the day and its demonstration obscured by measurements which reflect 24 hour ance with a clear adrenal pituitary abnormality. metabolism. However, there are at least three other states According to our hypothesis some of the states recently described which are associated with hypercortisolism besides classic Cushing’s. For of obesity are associated with hypercortisolism. example, in the last several years it has been This raises the question as to why this does not documented that pregnancy is associated with a exert a negative feedback on the hypothalomus and pituitary gland to inhibit the hyperproducmild degree of hypercortisolism.(20) Pregnancy tion of cortisol. One possibility to account for may be associated with hypertension, glucose this is that these states of obesity are refractory intolerance, and increased fat stores. Second, to cortisol at the level of the hypothalamus many chronic alcoholics have hypercortisolism and this, too. is associated with abnormalities in and/or pituitary gland. There is information from animal studies that this indeed may be the case. blood pressure and carbohydrate metabolism. This syndrome is induced by alcohol intake and Webb, et al., analyzed the binding of tritiated triamcinolone acetonide to cytosolic receptors resolved with abstinence. Finally, depressive from brains and liver homogenates of misty states are associated with hypercortisolism.(21) diabetic mice (Mdb/Mdb) and reported that the In each of these cases the presence of hypercortisolism has been documented only recently; it diabetic mice had a decreased number of recephad been overlooked by many. Importantly, in tors for glucocorticoids compared to appropriately Further, the steroid these three states the hypercortisolism is tran- matched controls.(23) sient although some of the stigmata may last receptor complexes from the diabetic mice showed a decreased ability to form the activated, longer than the hormonal abnormality. nuclear binding form. Thus, these animals Just as an abnormality of cortisol secretion may not be present through the entire life of a showed a decreased sensitivity in these tissues to the actions of glucocorticoids. pregnant, depressed, or alcoholic patient, so too, one may also hypothesize that the production of cortisol may not have to be abnormal throughout Summary an entire day to lead to abnormalities in metabolism. Thus the time at which observations are Many hormonal abnormalies have been demonmade might be important. Cortisol is secreted in strated in obesity. Which if any, is involved in the genesis and maintenance of the obese state a circadian rhythm. Many studies of cortisol is not yet known. Recent evidence indicates that economy involve assays of morning fasting blood levels. a time at which cortisol is thought to be obesity is not a simple monolithic disorder but instead is made up of multiple subtypes. Within highest. However, it is now becoming evident that cortisol may peak at other times in the day these subtypes it is quite possible that certain besides this morning surge. J. H. Liu et al. have groups reflect the result of discrete hormonal abnormalities. It is also becoming evident that recently reported that cortisol increases dramatically after food intake. especially lunch.(22) It the obese should not be grouped together simply by the degree above ideal body weight; discrimmay be that in some genetically susceptible individuals there is a higher postpandrial peak . ination by the location of fat cells is also than normal. This abnormality over a prolonged important. period could cause enough of a change to be Although early studies suggested that cortisol metabolically significant. It is important to note may be hypersecreted in obesity, more recent that the postpyrandrial surge of cortisol after studies disagree with this finding. However, as lunch is different from the morning cortisol surge noted, these studies were frequently designed to in that insulin levels are high in the former and exclude those with diabetes and hypertension low in the latter. As noted before, the simul- (which may be integral findings in some forms taneous presence of cortisol and insulin can of obesity) and may be skewed toward study trigger the formation of adipocytes from populations which do not have upper-body precursor cells. The combination of an elevated obesity. As yet there are no studies which insulin and cortisol may mediate the developcompare the cortisol economy of obese patients ment of hyperplastic obesity.(l3) who have upper-body obesity to those who have
100 lower-body obesity. These studies need to be done with the inclusion of those who have obesity-associated hypertension and diabetes. Studies may need to be done analyzing not only samples collected over a 24 hour period, but also need to involve detailed analyses of cortisol economy over shorter time periods with the stimulus of meals. Indeed, as the hormonal abnormality may not necessarily be present throughout the subject’s entire life, it may be that longitudinal studies are needed at various times through periods of weight gain and weight loss.
References B R. Foster W R. Hirsch T. Van Irallie T B. I. Burton Health implications of obesity. International Journal of Obesity 9: 155-169, 1985. P M. 2. Garrison R J. Feinleih M. Castelli W P, McNamara Cigarette smoking as a confounder of the relationship between relative weight and long term mortality: The framington heart study. J. Am. Med. Assoc. 249: 2199-2203, 1983. L. Variations in mortality by weight 3. Lew E A. Garfinkle among 750 000 men and women. J. Chron. Dis. 32: 563-76. 1979. of 4. Hartz A J. Rupley D C, Rimm A A. The association girth measurements with disease in 32 856 women. Am. J. of Epidemio 119: 71-80. 1984. E. Wertheimer E. Comparative physiology of 5. Shafrir adipose tissue in different sites and in different species. p 427-29 in Handbook of physiology Section 5: Adipose Tissue. (Renold A E, and Cahill G F. eds.) American Physiological Society, Washington, D.C. 1965. 6. Blair D, Habicht J P. Sims E A, Sylwester D, Abrahan S. Evidence for an increased risk of hypertension with centrally located body fat and the effect of race and sex on this risk. Am. J. Epidemio 119: 526-40, 1984. R K. Hartz A H, Rupley D, Kissebah A H. 7. Kalkhoff Kelber S. Relationship of body fat distribution to blood carbohydrate tolerance. and plasma lipids in pressure, healthv obese women. J. Lab Clin. Med. 102: 621-27. 1983. ’ 8. Jung R. Endocrinological aspects of obesity. Clinics in Endocrinology and Metabolism 13: 597-611, 1984. 9. Cunningham J J, Calles-Escandon J. Garrido F. Carr D B. Bode H H. Hypercorticostenuria and diminished
MEDICAL
HYPOTHESES
pituitarv responsiveness to CRF in obese Zucker Rats. Endocrinology 118: 98-101, 1986. Freedman M R. Horwitz B A. Stern J S. Effect of adrenalectomy and glucocorticoid replacement on development of obesity. Am. J. Physio. 249: R595-R6U7. 1986. 11. Dubuc P U. Wilden N J. Adrenalectomy reduces but does not reverse obesity in OB/OB mice. Int. J. Obes. 10: 91-Y8. 1986. 12. Feldman D. Evidence that brown adipose tissue is a glucocorticoid target organ. ‘Endocrinology 103: 2091-2097. 1978. 13. Hauner H, Schmid P. Pfeiffer E R. Glucocorticoids and insulin promote the differentiation of human adipocyte precursor cells into fat cells. J. Clin. Endo and Metabol. 64: 832-835, 1987. 14. Simkin B. Urinary 17-Ketosteroid and l7-Ketogenic steroid excretion in patients. New England J. of Med. 264: 974-77. 1961. 1.5.Szenas P. Pattee C J. Studies of adrenocortical function in obesity. J. Clin Endocrinol Metab 19: 344-350. 1959. 16. Mlynaryk P. Gillies R R. Murphy B, Pattee C J. Cortisol production rates in obesity. Journal of Clin Endocrinol & Metab 22: 587-591, 1962. 17. Migeon C J. Green 0 C, Eckert J P. Study of adrenocortical function in obesity. Metabolism 12: 718-738, 1963. IX. Strain G W. Zumoff B, Strain J J, Levin J. Fukushima D K. Cortisol production in obesity. Metabolism 29: 980-85. 1980. D H. Stevenson C T. Dalakos T G. Dennick 19. Streeten L G. Fellerman H. The diagnosis of hypercortisolism: Biochemical criteria differentiating patients from lean and obese normal subjects and from females on oral contraceptives. J Clin Endocrin 29: 1191-1211. 1969. M D. Rueckert S 0. Ehrlich 20. Nolten W E. Lindheimer E N. Diurnal patterns and regulation of cortisol secretion in pregnancy. J Clin Endocrin Metabol 51: 466-472. 1980. D D. Yen 21. Mortola J F, Liu J H. Gillin J C. Rasnaussen S S. Pulsatile rhythms of adrenocorticotropin (ACTH) and cortisol in women with endogenous depression: evidence for increased ACTH pulse frequency. J Clin Endocrin Metabol 65: 962-68. 1967. D D. Characterization 22. Liu J H, Kazer R R. Rusnussen of the twenty-four hour secretion patterns of andrenocorticotropin and cortisol in normal women and patients with Cushing’s Disease. J Clin Endrocrin Metabol 64: 1027-1035. 1987. 23. Webb M L, Hvnn J J, Schmidt T J, Margules D L. and and Litwack G. Decreased glucocorticoidbindin receptor activation in genetically diabetic (MDB f MDB) mice. J Steroid Biochem 25: 645-57. 1986.
,o.