John D. and
Catherine T. MacArthur
Research Network on Socioeconomic Status and Health
Body Composition Summary prepared by Mary Fran Sowers in collaboration with the Allostatic Load working group. Last revised 1997. Table of Contents d. Cortisol e. Female Sex Steroid Hormones f. References Introduction and Rationale It is increasingly well appreciated that body composition and body composition topology are related to disease conditions. This has been shown with heart disease, diabetes, gall bladder disease, certain cancers, osteoporosis, and arthritis (see Sowers, 1996). In field based studies, measures of body composition have historically been limited to simple measures such as weight adjusted for various height power indices as measures of adiposity. An example of this is the body mass index (Quetelet index) which is weight divided by the square of height. While weight adjusted for height is highly correlated with the amount of fat mass (correlations of 0.80 to 0.90), this is a notoriously poor approach for the measurement of muscle mass, protein status or lean tissues. The importance of the lean compartment has now been shown to be a better predictor of bone mass and its change. It is in the lean compartment that metabolism of glucose takes place and where insulin resistance is manifest. 
This section will describe methodologies for three approaches to body composition. The first method, underwater densitometry, refers to the classic approach to determining body composition based on principles promulgated by Archimedes. While not feasible for field studies, it is the approach, which until the last 5 years, served as the gold standard for validating other methods (Siri, 1956 and 1961; Brozek, 1963). The second methodology, dual x-ray densitometry (DXA), is rapidly replacing underwater weighing as the gold standard for body composition and is logistically more viable than is underwater weighing (Cullum et al., 1989; Mazess et al., 1989; Kelly et al.,1988). The third methodology, electrical impedance, is highly viable in field settings and has the capacity for describing the distribution of the water compartment in various diseases. This information can then be used to estimate amounts of fat and lean (Lukaski and Bolonchuk, 1988). Table 1. Ranking of Methods for Assessment of Skeletal Muscle Mass in Human Beings
Ranking system: ascending scale, 1 = least and 5 =
greatest The term densitometry refers to general procedures of estimating body composition from body density, under the assumption that the density of any material is the function of the proportion and densities of its components. Until recently, body composition was characterized by underwater weighing which generated knowledge of two compartments, the fat and the fat-free masses. The fat free mass is a heterogeneous compartment which could be further subdivided according to its primary constituents: water (73.8%), protein (19.4%), and mineral (7.8%). The underlying assumptions of this kind of body composition were as follows:
Since the 1940’s, numerous techniques to estimate body composition have been developed, with the primary gold standard for estimation being the underwater weighing technique methods (Siri, 1956 and 1961; Brozek, 1963). This approach has a number of technical and logistical limitations.
The technique, which is grounded in the determination of fat mass and then calculates the fat-free mass by extension, is estimated to have an error of approximately 4% when relatively homogenous populations are studied (Akers and Buskirk, 1969). While the requirements for calculation are not necessarily complex, accurate data require highly standardized implementation and facilities which are not widely available. Furthermore, a major limitation is that the technique requires a highly motivated participant who is willing to don bathing attire, complete effective spirometry, is willing and capable of exhaling prior to submersion in a water tank and can be confident in that tank for a period of 30 seconds. With the development of dual photon and dual x-ray densitometry (DXA) in the late 1980’s and early 1990’s, another approach was available (Cullum et al., 1989; Mazess et al., 1989;1992). The estimated fat content in bone-free lean tissue is derived by the constant attenuation of pure fat (Rf = 1.18-1.21) and the attenuation of bone-free lean tissue (Rl = 1.399). Given the constancy of these two values from subject to subject, the ratio of the attenuation at the lower energy relative to the higher energy in soft tissue for the low and high energy x-rays (40 keV and 70 keV) is a function of the proportion of the R values for fat and lean in each pixel. The relative ease of access to dual x-ray densitometry has revolutionized the capacity to undertake body composition studies. The implementation is relatively simple and takes minimal cooperation from the participant. Nonetheless, there are still issues which must be addressed. a. Dual x-ray densitometry cannot be undertake in the morbidly obese for two reasons. First, the table is only mechanically stable to weights between 260-290 pounds (depending upon the manufacturer), but deforms with loads beyond those weights, jeopardizing the entire system. Secondly, an assumption underlying the use of DXA assumes that measurements are not affected by the anteroposterior thickness of the body. Consistently studies have shown that thickness greater than 25 cm has an effect, violates the assumption, and typically overestimates the fat mass (Laskey et al., 1992). b. The sensitivity to change in hydration status has the potential to affect the bone-free lean tissue; however, studies indicate that this is a relatively minor source of error (Going et al., 1993). c. The estimation of body composition is a function of the 40-50% of the pixels which do not contain bone. Thus, measurement of regions of the body including thorax and arm which may have relatively fewer pixels without bone are more prone to measurement error.
A third technique is electrical impedance (BIA). All BIA devices consist of essentially an alternating electrical current source (usually about 1/4 volt), cables and electrodes for inducing the current into the body and for sensing the voltage drop due to impedance from body tissues and a system for measuring impedance. This approach operates under the assumption that the electronic conduction in biological tissues is mainly ionic, that is, electrical charges are transferred by ionized salts, bases and acids dissolved in body fluids. Thus, simplistically conceived, the body has highly conductive intracellular and extracellular materials which are separated by insulating layers of materials such as lipids which can be measured. The measures generated by the technique, which are resistance and reactance, can then be used to derive estimates of body water and, by extension, lean tissue and fat mass. (Boulier et al, 1990). The electrical impedance approach has several notable characteristics.
Because there is less experience in the direct measurement of total body water and the derived body composition from intracellular and extracellular fluid, the normative data and interpretative value of BIA has not expanded as rapidly as it has for DXA. Waist/Hip Ratio The ratio between waist circumference and hip circumference if often used as an index of the distribution of adipose tissue, although this ratio is influenced by some other body tissues. The utility of the waist/hip ratio (WHR) as an indicator of relative adipose tissue distribution in children and youth is not established, but it is useful in adults. Percentiles are available for the WHR in a large French sample (8,646 males; 9,747 females) 17 to 60 years of age (Tichet et al., 1993) and a large Danish sample (1,527 males; 1,467 females) 35 to 65 years of age (Heitmann, 1991). WHRs of middle-aged women are influenced by menopausal status. Mean ratios change slightly from premenopause (0.73 + 0.05, n = 550, 43.4 + 3.9 years) to perimenopause (0.74 + 0.06, n = 168, 49.2 + 3.8 years) but increase significantly after menopause (0.78 + 0.06, n = 1133, 58.6 + 4.7 years) (Sonnenschein, Kim, Pasternack, & Toniolo, 1993). An elevated WHR is usually accepted as an indication of proportionally more abdominal adipose tissue, but the accuracy of the WHR in distinguishing abdominal visceral adipose tissue from saturated adipose tissue is not defined. In obese individuals, changes in visceral adipose tissue after weight loss are not related to changes in the WHR (van der Kooy et al, 1993). Among adults, mean WHRs are larger in Mexican American than in White females from San Antonio, Texas, but are slightly larger in White than in Mexican American males. WHRs of young adult Black and White women and men in the U.S. differ only slightly (Kaye et al, 1993; Slattery et al, 1992). Among older American Blacks in men (35 years), WHR is similar to that of Europeans while the mean WHR for Black females is markedly larger than Europeans (Croft et al, 1993). In the San Antonio sample, there is a significant social class difference in the WHR among Mexican American males but not among females. The age-adjusted ratio in males increases from 0.92 in barrio residents (low SES) to 0.94 in residents from a transitional neighborhood to 0.95 in suburban residents (high SES). Corresponding age-adjusted means in Mexican American females are, respectively, 0.84, 0.83, and 0.84. In contrast, the WHR does not show a significant social class difference among White males and females in San Antonio (Malina and Stern, 1992). Though not differentiating between saturated adipose tissue and visceral adipose tissue, estimates of android (upper body segment, trunk) and gynoid (hip and thigh region) distributions of adipose tissue from total body DXA scans indicate the sex difference and changes in relative adipose distribution associated with menopause. The proportion of android adipose tissue is greater in postmenopausal women, while the proportion of gynoid adipose is greater in premenopausal women. In both groups of women, however, the proportion of android fat is less than in men. Thus, with the transition to menopause, the relative adipose distribution of women changes in the direction of that observed in men (Ley, Lees, & Stevenson, 1992). Estimates of abdominal visceral adipose tissue and saturated adipose tissue areas increase with greater body mass index (BMI). Overweight women have proportionally more saturated adipose tissue than normal weight females until about 60 years of age, after which they have proportionally more visceral adipose tissue (Enzi et al, 1986).
Excess cortisol secretion and its consequences are clearly seen in Cushing’s syndrome as a wasting of lean body mass, particularly muscle. Body fat mass and distribution are also changed. Clinical evidence suggests that the changes in adipose tissue are dependent on the prevailing insulin concentrations. These in turn are determined by the insulin resistance, created by the glucocorticoid excess. Therefore, the interactions of insulin become important for the net effect of glucocorticoids on muscle mass as well as on adipose tissue mass and distribution. This is why body composition of both lean and fat mass are needed. Glucocorticoids probably exert direct effects on the insulin sensitivity of both liver and muscle (Baxter & Rosseau, 1979; McMahon, Gerich, & Rizza, 1988). It has been hypothesized that these effects are mediated via free fatty acids (FFA) in studies with rats under stress (Guillaume-Gentil, Assimacopoulos-Jeannet, Jeanrenaud, 1993). The consequences were insulin resistance of hepatic gluconeogenesis as well as of glucose metabolism in muscles. These perturbations were obliterated by the administration of an inhibitor of fatty acid oxidation. These results suggest that the effects of glucocorticoids on the mobilization of FFA from adipose tissue are of primary importance for the insulin resistance of that condition in the integrated system. The important regulatory effects of insulin would, in conditions with cortisol excess in vivo, be expected to be a balance between a slight inhibition of lipolysis and a blunted antilipolytic effect of insulin. There are complex interactions between steroid and peptide hormones in the regulation of adipose tissue metabolism and the subsequent effects on body fat mass. In the condition of excess cortisol secretion of central origin, sex steroid secretion is probably inhibited through interactions by corticotropin releasing factor (CRF) on gonadotropin releasing hormone (Olsen & Ferin, 1987). Therefore the net effects will mainly be those of cortisol and insulin, whereas the effects of GH and sex steroid hormones are less evident, which contributes to the total effect of cortisol excess on body fat due to central inhibitory mechanisms. Glucocorticoids also exert marked effects on lipid mobilization in adipose tissue with interactive effects by other hormones. Fain, Kovacev, and Scow (1965) showed that glucocorticoids, particularly in the presence of GH, act by increasing lipid mobilization in vitro in adipose tissue from rats. It is important to interpret the results in relation to the species studied. In human adipose tissue, the acute short-term effects of cortisol appear to be absent. In long-term (days) experiments under fully controlled conditions, the effects are easier to detect. In the presence of insulin, cortisol has a weak inhibitory effect on catecholamine-induced lipolysis, but cortisol induces a marked "permissive" effect by GH on lipolysis (Cigolini & Smith, 1979; Ottosson, Lönnroth, Björntorp, & Edén, 1995). Mobilization of triglycerides from adipose tissue is dependent not only on lipolysis but also on the capacity of the adipocyte to re-esterify FFA. There is evidence that glucocorticoids may inhibit this process by interactions with glucose transport into adipocytes (Carter-Su & Okamoto, 1987), glucose being obligatory for re-esterification. The net effects of glucocorticoids on lipid mobilization in vivo have been found to result in marked increases of circulating FFA both in man (Divertie, Jensen, & Miles, 1991) and rats (Guillaume-Gentil et al, 1993). Studies in vitro of adipose tissue from patients suffering from Cushing’s syndrome with hyperinsulinemia have shown that the lipolytic sensitivity to catecholamine is unchanged or inhibited (Rebuffé-Scrive, Krotkiewski, Elfverson, & Björntorp, 1988). The elevated FFA concentrations after cortisol excess may be due either to a weak antilipolytic effect of insulin or an inhibited glucose transport, leading to diminished fatty acid re-esterification in adipose tissue, or a combination of both. Again, as with LPL regulation, sex steroid hormones may interfere with several of these regulatory steps of FFA mobilization. In states of hypercortisolemia, however, the secretion of sex steroid hormones is decreased, minimizing such effects, which, however, probably play important roles in the normal situation, particularly in young adults with intact sex steroid hormone secretion. Steroid hormones usually exert their effects via interaction with specific receptors for subsequent interactions of the hormone-receptor complex at the level of the appropriate genes. A glucocorticoid receptor (GR) has been described in adipose tissue from rats (Feldman & Loose, 1977) and human beings (Rebuffé-Scrive, Lundholm, & Björntorp, 1985b). The density of this receptor seems to vary in different regions of adipose tissue, and there is evidence that the density is higher in intra-abdominal (visceral) than in subcutaneous adipose tissues in studies with ligand binding in cytosol preparations in man (Rebuffé-Scrive et al, 1985b) and in intact cells in rats (Sjögren, Weck, & Björntorp, 1994). Also, the steady state GR mRNA levels have been reported to be higher in visceral than subcutaneous adipose tissue in man (Rebuffé-Scrive et al, 1990; Peeke et al, 1993). The effects of glucocorticoids on adipose tissue appear to be regionally specific as shown in Cushing’s syndrome, where there is marked redistribution of storage fat from the periphery to central depots. There are several possible reasons for this. For example, there is a high density of the GR in this depot (Rebuffé-Scrive et al, 1985b, 1990; Sjögren et al, 1994) or the balance between lipid accumulation and mobilization are disturbed, with more pronounced effects of the former. Other conditions with long-term exposure to excess glucocorticoids are followed by redistribution of adipose tissue, having been described in the treatment of asthma and rheumatoid arthritis with glucocorticoids (Krotkiewski, Blohmé, Lindholm, & Björntorp, 1976b). Marin has suggested that obesity, with disproportional increase of visceral fat depots, is a condition with increased cortisol secretion due to a high sensitivity of the hypothalamic-adrenal axis to different forms of stress (Marin et al, 1992). Female Sex Steroid Hormones
Female sex steroid hormone concentrations also regulate adipose tissue mass, although it is not clear whether these effects are direct or mediated via energy intake and/or expenditure (Wade & Gray, 1979). With menopause, an increase in visceral fat mass is preventable by hormonal replacement therapy (Haarbo, Marskew, Gottfredsen, & Christiansen, 1991; Rebuffé-Scrive, Eldh, Hafström, & Björntorp, 1986). Furthermore, the re is some evidence of enlargement of femoral subcutaneous adipocytes to support the demands of pregnancy and lactation and this enlarge is reported to disappear with menopause (Rebuffé-Scrive et al, 1985a). Progesterone competes with the GR (Rebuffé-Scrive et al, 1985b; Xu, Hoebeke, & Björntorp, 1990b) and may protect from glucocorticoid effects during the late luteal phase of the menstrual cycle. In summary, there are substantial reasons to believe that both measures of body composition and body topology are important is states associated with higher corticalism, growth hormone interactions and sex steroid levels. To the degree that these hormone interactions are contributory to allostatic load, body composition and body topology is contributory to allostatic load. Furthermore, the recent development of technology which provides highly precise and logistically feasible measures of lean tissue and adipose tissue as well as its relative anatomic location suggest that it is possible to generate research in this area relative to allotstatic load.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||
   |
||||||||||||||||||||||||||||||||||||||||||||||||||||||
Contact: Judith Stewart
Revised 25 July 1999