Journal of Applied Physiology Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Appl Physiol 87: 1163-1171, 1999;
8750-7587/99 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wang, W.
Right arrow Articles by Heymsfield, S. B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wang, W.
Right arrow Articles by Heymsfield, S. B.
Vol. 87, Issue 3, 1163-1171, September 1999

Regional skeletal muscle measurement: evaluation of new dual-energy X-ray absorptiometry model

Wei Wang, Zimian Wang, Myles S. Faith, Donald Kotler, Rick Shih, and Steven B. Heymsfield

Obesity Research Center, St. Luke's-Roosevelt Hospital, Columbia University, College of Physicians and Surgeons, New York, New York 10025


    ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

Although there is growing interest in studying muscle distribution, regional skeletal muscle (SM) mass measurement methods remain limited. The aim of the present study was to develop a new dual-energy X-ray absorptiometry (DEXA) model for estimating regional adipose tissue-free skeletal muscle mass (AT-free SM). Relationships were derived from Reference Man data between tissue-system- level components (i.e., AT-free SM, AT, skeleton, and skin) and molecular-level components including fat-free soft tissue, fat, and bone mineral. The proposed DEXA-SM model was evaluated by multiscan computerized axial tomography (CT). Twenty-seven male subjects [age, 36 ± 12 (SD) yr; body mass, 73.2 ± 12.4 kg; 20 were healthy, and 7 had acquired immunodeficiency syndrome] completed DEXA and CT studies. Identical landmarks for DEXA and CT measurements were selected in three regions, including calves, thighs, and forearms. There was a strong correlation for AT-free SM estimates between the new DEXA and CT methods (e.g., sum of three regions, r = 0.86, P < 0.001). Regional AT-free SM measured in the 27 subjects by DEXA and CT, respectively, were 3.44 ± 0.60 and 3.47 ± 0.55 kg (difference 0.9%, P > 0.05) for calves, 10.49 ± 1.77 and 10.05 ± 1.79 kg (difference 4.4%, P < 0.05) for thighs, 1.36 ± 0.49 and 1.20 ± 0.41 kg (difference 13.3%, P < 0.01) for forearms, and 15.29 ± 2.33 and 14.72 ± 2.33 kg (difference 3.9%, P < 0.05) for the sum all three regions. Although the suggested DEXA-SM model needs minor refinements, this is a promising in vivo approach for measurement of regional SM, because DEXA is widely available, relatively inexpensive, and radiation exposure is low.

computerized axial tomography; body composition; adipose tissue; bone mineral content; connective tissue


    INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

SKELETAL MUSCLE (SM) is the largest nonadipose tissue component at the tissue-system level of body composition in humans, and SM plays an important role in physical activity and many biochemical processes (19). There is increasing multidisciplinary interest in studying muscle distribution and regional SM mass (14). For example, investigators from several different disciplines are interested in monitoring appendicular SM mass changes in relation to growth, development, aging, and weight gain or loss (3, 7, 10; D. Gallagher, E. Ruts, M. Visser, S. Heshka, R. N. Baumgartner, J. Wang, R. N. Pierson, Jr., F. X. Pi-Sunyer, and S. B. Heymsfield, unpublished observations.).

Despite their important role in physiological and pathological processes, practical regional SM measurement methods are not well developed. At present, the most accurate in vivo methods of measuring regional SM mass are multiscan computerized axial tomography (CT) and magnetic resonance imaging (MRI) (14). Although these two methods represent a technological advance and are used as reference standards, their application in routine clinical practice and body composition research is limited because of expense and lack of access to instruments. The CT method also exposes subjects to radiation; therefore it cannot be used in evaluating healthy children and premenopausal women.

To assess regional SM mass, two alternatives to CT and MRI are available that are based on anthropometric and bioimpedance techniques (2, 14). If it is assumed that arm and leg muscle cross sections are circular, the use of anthropometric measurements, typically combined with circumferences and skinfold thicknesses, yields assessment of muscle cross-sectional areas (14). The bioimpedance method relies on the conduction of an applied electrical current to estimate conductor volume. Because muscle is the dominant limb conductor, bioimpedance methods can be used for noninvasive assessment of regional SM (14). However, these two methods may not be accurate in individual subject evaluations, and the derived prediction equations may also be population specific. Additional factors (e.g., abnormal fluid distribution) may affect regional SM estimation by conventional single-frequency bioimpedance analysis. It is also questionable whether anthropometric and bioimpedance estimations are either sufficiently accurate or sensitive to be used to monitor, over short time periods, small changes in regional SM mass associated with weight loss or gain in an individual subject.

The recent introduction of dual-energy X-ray absorptiometry (DEXA) provides another opportunity to measure regional SM in vivo with lower cost and substantially less radiation exposure compared with CT. The DEXA approach first allows separation of the whole body into two compartments: soft tissue and bone mineral (Mo). Soft tissue can be further separated into fat-free soft tissue and fat by use of the ratio (R value) of X-ray attenuation at the system's two main energy peaks (16). Whole body DEXA systems also allow separation of extremity from trunk measurements, and this permits regional SM mass measurement. In our previous investigation (8), a DEXA model was proposed and applied to estimate appendicular SM mass. A three-compartment model for appendages was suggested as
Appendicular mass = appendicular (SM + fat + bone)
or
Appendicular SM = appendicular mass 
− appendicular (fat + bone) (1)
Equation 1 assumes that appendicular adipose tissue (AT) mass is equal to fat mass. This equation also assumes that appendicular skin and bone marrow are negligible in mass relative to the SM component. Under usual circumstances, ash represents 55% of defatted and marrow-free bone mass. An assumption was therefore made that wet bone mass equals bone mineral content (BMC; i.e., bone ash) divided by 0.55 or multiplied by 1.82. Appendicular SM was then equal to arm and leg mass minus the sum of limb fat and wet bone mass
Appendicular SM = appendicular mass 
− appendicular (fat + 1.82 × BMC)
= appendicular fat-free soft tissue 
− 0.82 × appendicular BMC (2)
In Eq. 2, appendicular fat-free soft tissue and BMC can be estimated by DEXA. When this model was applied in healthy adults (8), there were strong correlations (all P < 0.001) between limb SM predicted by the model and anthropometric limb muscle area (r = 0.82-0.92), total body potassium (TBK; r = 0.94), and total body SM based on TBK-predicted fat-free body mass (r = 0.82).

In 1992, Fuller and colleagues (5) derived another DEXA model for appendicular SM mass estimation. The Fuller model differs from the Eq. 2 model in two respects. First, Fuller et al. accounted for the contribution of major limb tissues to fat-free soft tissue, whereas the model defined by Eq. 2 assumes that all fat-free soft tissue in limbs is SM. Second, the Fuller model assumes that bone ash is 28% of skeleton, whereas the Eq. 2 model assumes bone ash is 55% of bone. Skeleton is composed of several components, including bone, red marrow, yellow marrow, and periarticular tissue (18). High correlations were observed by Fuller et al. between DEXA-estimated arm and leg SM mass and other SM indexes, such as TBK (r = 0.90-0.94).

A validation study that uses an accurate method to estimate regional SM mass is still lacking for proposed DEXA models. An important and unresolved concern is whether and to what extent assumptions applied in DEXA models are associated with errors in regional SM estimation.

The aim of the present study was to develop a refined DEXA-SM model for estimating lower and upper leg and forearm SM mass. Specifically, we compared DEXA model-derived AT-free SM in calves, thighs, and forearms with corresponding regional AT-free SM measured by multiscan CT in adult men.


    METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

Experimental Plan

The new and previous (i.e., Eq. 2) DEXA-SM models were compared with corresponding CT-measured regional appendicular AT-free SM. Trunk SM was not evaluated because DEXA cannot differentiate SM in the trunk from organs and other lean tissues. The aim of the present study was to examine DEXA regional SM models, and the model of Fuller et al. (5) is not explored further in this report because it was designed to quantify muscle content of the entire arm and leg, including the hand and foot. The present studies were carried out over the course of 1 day at the Obesity Research Center's Body Composition Unit.

New DEXA-SM Model Development

The following descriptions represent the mathematical transformations and assumptions on which the new DEXA-SM model is based.

Regional molecular-level body composition. On the basis of DEXA measurements, the arm and leg can be divided into three compartments on the molecular level: fat-free soft tissue, fat, and Mo
Regional mass = fat-free soft tissue + fat + Mo 
= fat-free soft tissue + fat + 1.0436 × BMC (3)
The DEXA-measured fat-free soft tissue is the sum of four components that include water, protein, soft tissue mineral, and a small amount of glycogen. DEXA-measured BMC represents ashed Mo (4), and 1 g of Mo yields 0.9582 g of ash, because labile components such as bound water and CO2 are lost during heating (9, 15). The BMC estimated by DEXA is, therefore, converted to bone mineral as Mo = 1.0436 × BMC.

Regional tissue-system-level body composition. On the tissue system level, arm and leg can be divided into SM, AT, skeleton, connective tissue (CNT), and skin. The small amount of residual tissue, including nervous tissue and blood vessels, is negligible in mass relative to the SM compartment. A five-compartment body composition model was applied on the tissue-system level for arms and legs
Regional mass = SM + AT + skeleton + CNT + skin (4)
Skeleton, CNT, and skin compartments in Eq. 4 can be estimated in the following manner. Skeleton is considered an anatomic structure that includes cortical and trabecular bone, marrow, and cartilage. Previous studies (18) assume that the proportion of skeleton as Mo is a constant (Mo = 0.28 × skeleton; see Ref. 18, p. 284).

CNT includes mainly tendons, fascia, and periarticular tissue. In Reference Man, separable CNT weighs 1.6 kg (including 0.5 kg of tendons and fascia, 0.6 kg of periarticular tissue, and 0.5 kg of other CNTs; see Ref. 18, p. 276). Assuming that separable CNT distributes homogeneously within the whole fat-free soft tissue compartment (54 kg in Reference Man), the proportion of fat-free soft tissue that is CNT is equal to 1.6/54 or 0.0296. Accordingly, CNT = 0.0296 × fat-free soft tissue.

Skin includes the epidermis and dermis. The hypodermis, which underlies skin, is included in the AT component. The 70-kg Reference Man contains 2.6 kg skin; this indicates that neglect of the skin mass may cause a large overestimation of SM mass. In the present study, skin mass in the region of interest was estimated as the product of skin surface area (S), skin thickness, and assumed constant skin density (0.0011 kg/cm3) (see Ref. 18, p. 284)
Skin mass = 0.0011 × <IT>S</IT> × skin thickness (5)
Calf, thigh, and forearm can be simplified into a truncated cone, so that skin surface area S can be calculated as 0.5 × pi  × (R1 + R2) × L, where L is truncated cone length (in cm) and R1 and R2 are the top and bottom scan truncated cone diameters (in cm). Skin thickness (epidermis + dermis) varies in different regions. The mean skin thickness of legs and arms is 0.26 and 0.17 cm, respectively (see Ref. 18, p. 48). Equation 5 can thus be converted into
Skin mass = 0.0011 × 0.5 × &pgr; × (R<SUB>1</SUB> + R<SUB>2</SUB>) × <IT>L</IT>
 × skin thickness = 0.00173 × (R<SUB>1</SUB> + R<SUB>2</SUB>) × <IT>L</IT>
 × skin thickness (6)

New DEXA-SM model. In the present study, Reference Man data (18) were applied to link the molecular level components with tissue-system-level components as outlined in Table 1. According to Reference Man data, the tissue- system-level components have a known chemical composition that is assumed to be constant. On the basis of the quantitative relationships shown in the table, four simultaneous equations were derived that relate components at the two body composition levels
Fat-free soft tissue = 1.000 × AT-free SM + 0.200 × AT + 0.530 × skeleton + 0.900 
× skin + 0.987 × CNT (7)
Fat = 0.800 × AT + 0.190 × skeleton + 0.100 
× skin + 0.013 × CNT (8)
Mo = 0.280 × skeleton (9)
CNT = 0.0296 × fat-free soft tissue (10)
By resolving the four simultaneous equations, a model for regional SM prediction was derived
Regional AT-free SM = 0.971 × fat-free soft tissue 
− 0.250 × fat − 1.724 × BMC − 0.875 × skin = 0.971 × fat-free soft tissue − 0.250 × fat − 1.724 × BMC − 0.00151 × (R<SUB>1</SUB> + R<SUB>2</SUB>) × <IT>L</IT> × skin thickness (11)
where fat-free soft tissue, fat, BMC, diameters, and length can be estimated by regional DEXA measurements. Skin thickness is assumed to be 0.26 cm for legs and 0.17 cm for arms (see Ref. 18, p. 48).

                              
View this table:
[in this window]
[in a new window]
  		Table 1.   Fractions of tissue-system level components as fat-free soft tissue, fat, and bone mineral

Subjects

The new and previous regional DEXA-SM models (Eqs. 11 and 2) were evaluated in healthy male subjects and in men with acquired immunodeficiency syndrome (AIDS). The AIDS patients had varying degrees of body mass loss since the onset of their illness, although all were clinically stable at the time of study. Each volunteer signed an informed consent, and the study was approved by the Institutional Review Board at St. Luke's-Roosevelt Hospital. Some subjects in this study participated in another body composition project (20) that was unrelated to the present report.

Multiscan CT Measurements

The protocol required multiscan cross-sectional CT images (four for calf and thigh and three for forearm). The anatomic locations of the CT cross-sectional images used for calf, thigh, and forearm measurement are described in Table 2 (see Refs. 11, 12, 17). A Somatom DRH scanner (Siemens, Erlangen, Germany) was used in the study. Subjects were placed on the scanner platform with their arms extended above their heads to minimize beam-hardening artifacts. Each CT image was completed at 125 kVp, with a scanning time of 4 s at 170 mA. Slice thickness was set at 4 mm.

                              
View this table:
[in this window]
[in a new window]
  		Table 2.   Anatomic locations of calf, thigh, and forearm CT scans

An observer who was familiar with cross-sectional CT anatomy encircled the muscle area in each scan with a digital light pen. On the basis of early determinations by Kvist et al. (11, 12), Sjöström (17), and our laboratory (20), attenuations ranged from -190 to +120 Hounsfield units for anatomic SM and from -190 to -30 Hounsfield units for AT. The attenuation range from -29 to +120 Hounsfield units was thus applied to measure the cross-sectional area of AT-free SM. The distances between cross-sectional images were obtained from the CT frontal scanogram to the nearest millimeter. Regional AT-free SM (in kg) was calculated as
Regional AT-free SM = 0.00104 × &Sgr; [<IT>A</IT><SUB>i</SUB> × (<IT>B</IT><SUB>i</SUB> + <IT>B</IT><SUB>i+1</SUB>)/2] (12)
where 0.00104 is assumed constant density (kg/cm3) of AT-free SM (18), Ai is distance (cm) between adjacent scans, and Bi and Bi+1 are the cross-sectional areas (cm2) of AT-free SM in adjacent scans.

DEXA Measurements

Subjects were scanned by using a whole body DEXA system (model DP-4, Lunar Radiation, Madison, WI) with a cerium-filtered X-ray source and peak energies of 40 and 70 keV. The DEXA system software first divides pixels into Mo and soft tissue compartments. Soft tissue is then further separated by system software into fat-free soft tissue and fat (16). Whole body fat was evaluated in subjects and is reported as a percentage of body mass.

After completion of the whole body scan, regions of interest (i.e., calves, thighs, and forearms) were measured by manual DEXA analysis software. The landmarks that define calves, thighs, and forearms were identical to those used in multiscan CT measurements. Specifically, knee joint and ankle, and symphysis caudal edge and knee joint were used as the upper- and lower-limit points of calf and thigh, respectively. Elbow and wrist were selected as the upper- and lower-limit points of forearm (Table 2). The subjects' arms were extended above their heads during CT and below their heads during DEXA tests. This presents a problem in measurement of identical landmarks of upper arm in both the DEXA and CT methods. The upper arm, therefore, was not measured in the present study.

For skin mass estimation (Eq. 6), the region length (L) and diameter (R1 and R2) were obtained from the DEXA frontal scanogram by manual-analysis software.

Statistical Analysis

Simple linear regression analysis was used to describe the relationship between regional AT-free SM quantified by CT and DEXA-SM models. Mean differences between regional AT-free SM measured by CT and DEXA were tested for statistical significance by paired Student's t-test; P < 0.05 was considered statistically significant. The differences in regional AT-free SM estimates between CT and DEXA-SM model were then related to the mean of the two estimates as described by Bland and Altman (1).

The precision of the new DEXA-SM model was estimated on three healthy subjects. Regional DEXA manual analysis was repeated 10 times by a single trained observer, and the average between-measurement coefficient of variation (CV) was calculated.

Group results are presented as means ± SD. Data were analyzed by using SAS version 5 (SAS Institute, Cary, NC).


    RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

Body Composition Analysis

The subjects included 20 healthy male subjects and 7 male patients with AIDS (Table 3). The subjects' characteristics were (mean ± SD) as follows: age, 36 ± 12 yr; body mass, 73.2 ± 12.4 kg; body mass index, 23.2 ± 2.9 kg/m2; and fat, 13.9 ± 7.1%.

                              
View this table:
[in this window]
[in a new window]
  		Table 3.   Subject characteristics and baseline results for body composition

Molecular-level body composition analyses (including fat-free soft tissue, fat, and BMC for calves, thighs, forearms, and total) are presented in Table 4. Regional skin masses estimated by DEXA were 0.67 ± 0.07 kg for calves, 1.11 ± 0.13 kg for thighs, 0.19 ± 0.02 kg for forearms, and 1.96 ± 0.20 kg for the three regions.

                              
View this table:
[in this window]
[in a new window]
  		Table 4.   Regional body components (in kg) measured by DEXA

Regional AT-free SM weight, as measured by CT, was 3.47 ± 0.55 kg for calves, 10.05 ± 1.79 kg for thighs, and 1.20 ± 0.41 kg for forearms (Table 5). The sum of AT-free SM as measured by CT for all three regions was 14.72 ± 2.33 kg.

                              
View this table:
[in this window]
[in a new window]
  		Table 5.   Regional adipose tissue-free skeletal muscle mass (in kg) estimated by CT and DEXA-SM models

DEXA-SM Prediction

New model. The between-measurement CV for the new DEXA-SM model was 1.7% for calves, 2.0% for thighs, and 2.4% for forearms.

The results of regional SM estimates by the new DEXA model and corresponding CT SM estimates are presented in Table 5. There was no significant difference between DEXA and CT SM estimates for calf, and the mean DEXA value was 0.9% lower than that estimated by CT. There were, however, significant differences between new DEXA model SM estimates and CT estimates for thighs (P < 0.05, +4.4%), forearms (P < 0.001, +13.3%), and the sum of three regional muscle groups (P < 0.05, +3.9%).

Simple linear regression analysis demonstrated high correlations between regional SM estimates by the new DEXA-SM model and by CT for calves (r = 0.73, P < 0.001), thighs (r = 0.89, P < 0.001), forearms (r = 0.74, P < 0.001) (Table 6), and the three-region SM sum (r = 0.86, P < 0.001; Table 6 and Fig. 1).

                              
View this table:
[in this window]
[in a new window]
  		Table 6.   Regression equations between CT-measured SM and DEXA model-predicted SM



View larger version (36K):
[in this window]
[in a new window]
  		Fig. 1.   Regional adipose tissue-free skeletal muscle (SM) predicted by new dual-energy X-ray absorptiometry (DEXA)-SM model (left) and previous DEXA-SM model (right) vs. computerized axial tomography (CT)-measured regional adipose tissue-free SM. Three regions are shown: calf (A and D), thigh (B and E), and forearm (C and F). Lines of identity are shown, and corresponding regression equations are summarized in Table 6.

Bland-Altman analysis indicated that the difference between SM measured by DEXA and SM measured by CT (SMDEXA and SMCT, respectively) was not significantly associated with the mean of the two estimates for any of the three regions (Table 7) or for total regional SM (Table 7 and Fig. 2).

                              
View this table:
[in this window]
[in a new window]
  		Table 7.   Bland-Altman analyses for DEXA-SM models



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 2.   Regional adipose tissue-free SM predicted by new DEXA-SM model (A) and previous DEXA-SM model (C) vs. regional adipose tissue-free SM measured by CT. Sum of SM in 3 body sites (calf, thigh, and forearm) is presented. Lines of identity are shown in A and C, and related regression equations are summarized in Table 6. B and D: Bland-Altman analyses for corresponding linear relationships plotted above as SM difference (Delta ) between DEXA-model and CT (kg) and mean SM for DEXA model and CT. Regression equations for these analyses are presented in Table 7. Dashed lines, 95% confidence limits.

Previous model. Regional SM estimates predicted by the previous DEXA model are presented in Table 5. Regional SM estimates by the previous DEXA model were significantly higher than those measured by CT [36.0% for calves (P < 0.001), 27.2% for thighs (P < 0.001), 45.0% forearms (P < 0.001), and 30.7% for total (P < 0.001)].

Simple linear regression analysis demonstrated slightly stronger correlations and smaller standard errors of the estimate (SEEs) between SM estimates by the previous DEXA model and by CT (Table 6 and Figs. 1 and 2); e.g., r = 0.94 and SEE = 0.94 kg vs. r = 0.86 and SEE = 1.20 kg, for the three-region SM sum vs. the new model, respectively.

Bland-Altman analysis indicated that the differences between SMDEXA and SMCT were significantly associated with the mean of the two estimates for forearm SM but not for other regions, including the three-region SM sum (Table 7 and Fig. 2).


    DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

Despite clinical and research relevance, because of methodological limitations, regional SM mass remains underinvestigated in the study of human body composition. Although anthropometry and bioimpedance methods are widely applied in clinical and field studies, these methods are not yet sufficiently developed for estimating regional SM in individual subjects or in small study samples. The present study examined the possibility of estimating regional SM mass by using the widely available DEXA technique.

New vs. Older Models

The previously reported DEXA-SM model (8) was applied in the present study to estimate regional SM mass. As expected, on the basis of theoretical considerations, the earlier model overestimated SM mass in all three appendicular regions. The main reason for the overestimation of SM by the earlier model is the use of simplifying assumptions that did not consider the magnitude of nonmuscle compartments included within regional DEXA fat-free soft tissue estimates. Specifically, the previous model assumed that appendicular fat-free soft tissue estimated by DEXA is equal to appendicular SM mass. In other words, the previous model did not consider the contributions of skin, CNT, and AT to the fat-free soft tissue. When skin, CNT, and the fat-free portion of AT are ignored for model simplification purposes, the three compartments that are not SM are incorporated into the SM compartment. This assumption thus causes overestimation of SM mass. In the present study, for example, the mean skin masses estimated by DEXA were 0.67 kg in calves, 1.11 kg in thighs, and 0.19 kg in forearms (Table 4). Skin accounts for 10% of DEXA fat-free soft tissue in the three regions. Neglect of the skin mass thus causes a corresponding overestimation of regional SM mass. Another example is Reference Man, with 2.6 kg of skin, 1.6 kg of separable CNT, and 15 kg of AT. The 15-kg AT component contains 20% or 3.0 kg of fat-free soft tissue (18). For Reference Man, with 28 kg of SM, neglect of the three compartments thus causes a 7.2-kg or 26% total estimation error. This may partially explain why the previous DEXA-SM model overestimated regional SM compared with the CT method.

In the present study, a refined model was proposed that overcomes some of the limited assumptions of the previous DEXA model. There are two specific improvements in the new model compared with the previous model. First, our new model considers five major tissue-level compartments in the appendages (i.e., AT-free SM, AT, skeleton, CNT, and skin). Only a small residual mass, including nervous tissue and blood vessels, is not considered in the model. Second, we considered the quantitative interrelationships between five tissue-level compartments and three molecular-level compartments (Table 1). On the basis of these quantitative relationships, four simultaneous equations (Eqs. 7-10 ) were established, and the new regional DEXA-SM model (Eq. 11) was derived. The results showed only small mean differences between regional SM estimates made by the new DEXA-SM model and corresponding CT estimates (Table 5). Strong correlations for regional SM estimates were observed between the new DEXA-SM model and CT. Bland-Altman analyses (1) demonstrated that new DEXA-model and CT-regional SM differences were not significantly associated with the mean of the two SM estimates. These observations support the validity of the new DEXA-SM model.

There are similarities between the model proposed by Fuller et al. (5) and our new model because both are based on assumptions from Reference Man data. However, Fuller's model was designed for analysis of the entire arm and leg. It is doubtful that either DEXA or CT have sufficient accuracy to estimate very small muscles, particularly those surrounded by the bones of the hand and feet. Although we did not evaluate the model of Fuller et al. in our regional SM study, these investigators proposed a novel solution for estimation of the limb skin mass component. Specifically, Fuller's model estimates total limb skin mass on the basis of an assumption that each arm is 9% and each leg is 18% of weight- and height-derived whole body surface area. In the present study, we estimated regional skin area, including calf, thigh, and forearm, from the DEXA frontal scanogram. Adaptations of the Fuller body-surface area approach may be simpler than the DEXA-analysis method, and it would be useful to explore this possibility in future studies.

Multiscan CT as the Criterion

In the present study, multiscan CT was applied as the criterion for measurement of regional SM. Multiscan CT is based on the assumption that there exists a linear change between adjacent cross-sectional SM areas. When the number of cross-sectional images is small (i.e., the distance between adjacent cross-sectional images is large), this assumption may cause error in measurement of regional SM mass. In the present study, the number of cross-sectional CT images for each region was four for calves, four for thighs, and three for forearms (Table 2). A question thus arises concerning whether or not additional images might improve the accuracy of CT regional muscle estimates. Recently, Lönn et al. (13) measured lower leg SM by using MRI for both 7 scans and 70 contiguous scans. The authors reported that 7 scans and 70 scans gave similar muscle plus skin volumes (4.53 ± 1.20 vs. 4.33 ± 1.18 liters, respectively). There was no significant difference between 7 scans and 70 scans in estimates of muscle plus skin (P > 0.05; SE = 3.3%). These results support the present CT protocol, which was designed to measure regional SM mass.

Model and Study Limitations

Error sources. Methods such as the one developed in the present study have two sources of error: model and measurement. The total observed SM estimate error is a composite of at least these two main error sources.

In the present study, our emphasis was on reducing model errors, although our new model also included more predictor terms and, by necessity, more measurement error. Although our regional SM estimates with the use of the new model showed improved agreement (i.e., by Bland-Altman analyses) with CT-SM values compared with those by the earlier DEXA-SM model, we assume that the additional new model terms introduced more measurement error. Accordingly, our regional SM estimates by the new model had lower r2 values and higher SEEs than their previous counterparts.

There are also several possible sources of model error that should be considered. A series of assumptions were made during the development of the new DEXA-SM model. For example, we assumed that the three regions (e.g., calf, thigh, and forearm) can be described mathematically as a truncated cone. The separable CNT was assumed to be homogeneously distributed within the whole fat-free soft tissue compartment as a constant proportion (i.e., 2.96%). Skin thickness was assumed constant at 0.26 cm in the legs and at 0.17 cm in the arms. These assumptions, although necessary for development of the new DEXA model, may cause model error in estimating regional SM mass.

The quantitative interrelationships between three molecular-level compartments and five tissue-system-level compartments were used to develop several constants, as defined in Table 1. For example, the proportions of AT as fat and fat-free tissue were assumed constant at 80 and 20%, respectively (18). Another assumption is that the proportions of skeleton as Mo and fat are constant at 28 and 19%, respectively (18). These proportions actually vary within and between subjects and may cause model errors.

The balance between model and measurement errors is an important consideration for all multicomponent models. The new proposed regional SM model would appear to be a reasonable alternative to the older model by providing overall estimates that agree in magnitude with those by CT and also are highly correlated with regional SM estimates by CT. The possibility does exist, however, to develop simple region and population-specific SM prediction formulas in the future that are devoid of any underlying conceptual model and that might also have higher r2 values and lower SEEs than those of the new model presented in this report.

Male study cohort. The quantitative interrelationships between compartments, from which the new DEXA-SM model was derived, is based on the data of Reference Man, defined as a 20- to 30-yr-old Caucasian man with a body mass of 70 kg and 19% body fat (18). It is unknown whether and to what extent these assumed constant relationships apply to other populations, such as women, elderly subjects, and obese subjects. In addition, some physiological and pathological factors may also influence the quantitative interrelationships between the three molecular-level compartments and five tissue-system-level compartments. Thus further studies are needed to evaluate the model's applicability in other populations, such as women, athletes, the elderly, and obese patients, and in other regions such as the upper arm. These studies can be supplemented with chemical and anatomic data derived from cadaver and in vitro tissue studies that provide additional or improved model coefficient information.

Conclusion

In the present report, we derive and then validate a new DEXA model for estimating regional SM in calves, thighs, and forearms. Our earlier DEXA model failed to consider important contributions to regional appendicular SM mass of nonmuscle components. The initial results demonstrate that regional SM mass, as determined by the new DEXA-SM model, is similar in magnitude and highly correlated with that measured by multiscan CT in men. Further studies are needed to evaluate the model's applicability in other regions (such as the upper arm) and in other groups (such as women, elderly, and obese subjects).

This report thus reaffirms DEXA's regional SM measurement potential and provides a new operational model for estimating regional SM by DEXA. As such, this new model and its conceptual design lead the way for future model verification and refinement.


    ACKNOWLEDGEMENTS

This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-42618.


    FOOTNOTES

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Address for reprint requests and other correspondence: Z. M. Wang, Weight Control Unit, 1090 Amsterdam Ave., 14th Floor, New York, NY 10025.

Received 3 December 1998; accepted in final form 18 May 1999.


    REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

1.   Bland, J. M., and D. G. Altman. Statistical methods for assessing agreement between two methods for clinical measurement. Lancet 8: 307-310, 1986.

2.   Brown, B. H., T. Karatzas, R. Nakielny, and R. G. Clark. Determination of upper arm muscle and fat areas using electrical impedance measurements. Clin. Phys. Physiol. Measurement 9: 47-55, 1988.[Medline]

3.   Clarys, J. P., A. D. Martin, and D. T. Drinkwater. Gross tissue weights in the human body by cadaver dissection. Hum. Biol. 56: 459-473, 1984[Medline].

4.   Friedl, K. E., J. P. DeLuca, L. J. Marchitelli, and J. A. Vogel. Reliability of body-fat estimations from a four-compartment model by using density, body water, and bone mineral measurements. Am. J. Clin. Nutr. 55: 764-770, 1992[Abstract/Free Full Text].

5.   Fuller, N. J., M. A. Laskey, and M. Elia. Assessment of the composition of major body regions by dual-energy X-ray absorptiometry (DEXA), with special reference to limb muscle mass. Clin. Physiol. 12: 253-266, 1992[Medline].

7.   Gallagher, D., M. Visser, R. E. De Meersman, D. Sepúlveda, R. N. Baumgartner, R. N. Pierson, Jr., T. Harris, and S. B. Heymsfield. Appendicular skeletal muscle mass: effects of age, gender, and ethnicity. J. Appl. Physiol. 82: 229-239, 1997.

8.   Heymsfield, S. B., R. Smith, M. Aulet, B. Bensen, S. Lichtman, J. Wang, and R. N. Pierson, Jr. Appendicular skeletal muscle mass: measurement by dual-photon absorptiometry. Am. J. Clin. Nutr. 52: 214-218, 1990[Abstract/Free Full Text].

9.   Heymsfield, S. B., J. Wang, S. Heshka, S. Lichtman, and R. N. Pierson, Jr. Chemical determination of human body density in vivo: relevance to hydrodensitometry. Am. J. Clin. Nutr. 50: 1282-1289, 1989[Abstract/Free Full Text].

10.   Hochachka, P. Muscles as Molecular and Metabolic Machines. Boca Raton, FL: CRC, 1994.

11.   Kvist, H., B. Chowdury, L. Sjöström, U. Tylen, and Å. Cederblad. Adipose tissue volume determination in males by computed tomography and 40K. Int. J. Obes. 12: 249-266, 1988[Medline].

12.   Kvist, H., L. Sjöström, and U. Tylen. Adipose tissue volume determinations in women by computed tomography: technical consideration. Int. J. Obes. 10: 53-67, 1985.

13.   Lönn, L., G. Starck, M. Alpsten, S. Ekholm, and L. Sjöström. Determination of tissue volumes: a comparison between CT and MRI. In: L. Lönn, In: Hormonal Regulation of Body Composition: Applied Studies and Methodological Developments (PhD dissertation). Göteborg, Sweden: Göteborg University, 1998.

14.   Lukaski, H. Estimation of muscle mass. In: Human Body Composition, edited by A. F. Roche, S. B. Heymsfield, and T. G. Lohman. Champaign, IL: Human Kinetics, 1996, p. 109-128.

15.   Mendez, J. A., A. Keys, J. R. Anderson, and F. Grande. Density of fat and bone mineral of mammalian body. Metabolism 9: 472-477, 1960.

16.   Pietrobelli, A., C. Formica, Z. M. Wang, and S. B. Heymsfield. Dual-energy X-ray absorptiometry body composition model: review of physical concepts. Am. J. Physiol. 271 (Endocrinol. Metab. 34): E941-E951, 1996[Abstract/Free Full Text].

17.   Sjöström, L. A computer-tomography based multicomponent body composition technique and anthropometric predictions of lean body mass, total and subcutaneous adipose tissue. Int. J. Obes. 15: 19-30, 1991.

18.   Snyder, W. S., M. J. Cook, E. S. Nasset, L. R. Kahausen, G. P. Howells, and I. H. Tipton. Report of the Task Group on Reference Man. Oxford, UK: Pergamon, 1975.

19.   Wang, Z. M., R. N. Pierson, Jr., and S. B. Heymsfield. The five-level model: a new approach to organizing body composition research. Am. J. Clin. Nutr. 56: 19-28, 1992[Abstract/Free Full Text].

20.   Wang, Z. M., M. Visser, R. Ma, R. N. Baumgartner, D. Kotler, D. Gallagher, and S. B. Heymsfield. Skeletal muscle mass: evaluation of neutron activation and dual-energy X-ray absorptiometry methods. J. Appl. Physiol. 80: 824-831, 1996[Abstract/Free Full Text].

J APPL PHYSIOL 87(3):1163-1171
8570-7587/99 $5.00 Copyright © 1999 the American Physiological Society


This article has been cited by other articles:


Home page
 ptjournalHome page
S. Mathur, K. P. Takai, D. L MacIntyre, and D. Reid
Estimation of Thigh Muscle Mass With Magnetic Resonance Imaging in Older Adults and People With Chronic Obstructive Pulmonary Disease
Physical Therapy, February 1, 2008; 88(2): 219 - 230.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
R. D Hansen, D. A Williamson, T. P Finnegan, B. D Lloyd, J. N Grady, T. H Diamond, E. U. Smith, T. M Stavrinos, M. W Thompson, T. H Gwinn, et al.
Estimation of thigh muscle cross-sectional area by dual-energy X-ray absorptiometry in frail elderly patients
Am. J. Clinical Nutrition, October 1, 2007; 86(4): 952 - 958.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Physiol. Heart Circ. Physiol.Home page
S. Volianitis, C. C. Yoshiga, P. Nissen, and N. H. Secher
Effect of fitness on arm vascular and metabolic responses to upper body exercise
Am J Physiol Heart Circ Physiol, May 1, 2004; 286(5): H1736 - H1741.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
C. M. Modlesky, C. S. Bickel, J. M. Slade, R. A. Meyer, K. J. Cureton, and G. A. Dudley
Assessment of skeletal muscle mass in men with spinal cord injury using dual-energy X-ray absorptiometry and magnetic resonance imaging
J Appl Physiol, February 1, 2004; 96(2): 561 - 565.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
K. L. Herbst, B. D. Anawalt, J. K. Amory, A. M. Matsumoto, and W. J. Bremner
The Male Contraceptive Regimen of Testosterone and Levonorgestrel Significantly Increases Lean Mass in Healthy Young Men in 4 Weeks, but Attenuates a Decrease in Fat Mass Induced by Testosterone Alone
J. Clin. Endocrinol. Metab., March 1, 2003; 88(3): 1167 - 1173.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
Z. Wang, S. Zhu, J. Wang, R. N Pierson Jr, and S. B Heymsfield
Whole-body skeletal muscle mass: development and validation of total-body potassium prediction models
Am. J. Clinical Nutrition, January 1, 2003; 77(1): 76 - 82.
[Abstract] [Full Text] [PDF]

Home page
 J. Gerontol. A Biol. Sci. Med. Sci.Home page
M. Iannuzzi-Sucich, K. M. Prestwood, and A. M. Kenny
Prevalence of Sarcopenia and Predictors of Skeletal Muscle Mass in Healthy, Older Men and Women
J. Gerontol. A Biol. Sci. Med. Sci., December 1, 2002; 57(12): M772 - 777.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
J. Kim, Z. Wang, S. B Heymsfield, R. N Baumgartner, and D. Gallagher
Total-body skeletal muscle mass: estimation by a new dual-energy X-ray absorptiometry method
Am. J. Clinical Nutrition, August 1, 2002; 76(2): 378 - 383.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
A. Y. McDermott, A. Shevitz, T. Knox, R. Roubenoff, J. Kehayias, and S. Gorbach
Effect of highly active antiretroviral therapy on fat, lean, and bone mass in HIV-seropositive men and women
Am. J. Clinical Nutrition, November 1, 2001; 74(5): 679 - 686.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
B. M. Prior, C. M. Modlesky, E. M. Evans, M. A. Sloniger, M. J. Saunders, R. D. Lewis, and K. J. Cureton
Muscularity and the density of the fat-free mass in athletes
J Appl Physiol, April 1, 2001; 90(4): 1523 - 1531.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
R. Shih, Z. Wang, M. Heo, W. Wang, and S. B. Heymsfield
Lower limb skeletal muscle mass: development of dual-energy X-ray absorptiometry prediction model
J Appl Physiol, October 1, 2000; 89(4): 1380 - 1386.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
J. A. Levine, L. Abboud, M. Barry, J. E. Reed, P. F. Sheedy, and M. D. Jensen
Measuring leg muscle and fat mass in humans: comparison of CT and dual-energy X-ray absorptiometry
J Appl Physiol, February 1, 2000; 88(2): 452 - 456.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wang, W.
Right arrow Articles by Heymsfield, S. B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wang, W.
Right arrow Articles by Heymsfield, S. B.