| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research Communications |
1 From the Department of Life Sciences, The University of Sydney, Australia; the Gastrointestinal Investigation Unit, Department of Nuclear Medicine, University Department of Surgery, Royal North Shore Hospital, Sydney, Australia; and the Cancer Care Centre, St George Hospital, Sydney, Australia.
2 Supported by grants from The University of Sydney Mechanism B Scheme and The Australian Institute of Nuclear Science and Engineering.
3 Address reprint requests to RD Hansen, Department of Life Sciences, Building MO2, The University of Sydney, New South Wales, Australia 2006. E-mail: hansenr{at}med.usyd.edu.au.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objectives: The objectives of this study were to 1) compare SMM values derived from dual-energy X-ray absorptiometry (DXA) with those calculated by a nuclear method from total body potassium (TBK) and total body nitrogen (TBN) measurement (both: KN) in older subjects, and 2) assess the accuracy of FFM measurement by DXA in these subjects.
Design: TBK, TBN, DXA (model XR36; Norland, Fort Atkinson, WI), bioimpedance, and anthropometric measurements were performed on healthy women (n = 50) and men (n = 25) aged 51–84 y.
Results: Mean SMM by KN was not significantly different from SMM by DXA in either sex. SMM by KN predicted SMM by DXA with an SEE of 2.1 kg (r = 0.95, P < 0.0001 for women and men together). In the men, FFM by DXA agreed well with FFM estimated by TBK, skinfold thicknesses, bioimpedance analysis, and a multicompartment model. In women, FFM by DXA was 4–5 kg less than that by the other methods (P < 0.01). Truncal fat was related to intermethod FFM differences (r = 0.58, P < 0.0001).
Conclusions: These data indicate that 1) either the nuclear or the DXA method can be applied to estimate SMM in healthy older subjects, and 2) the Norland DXA instrument significantly underestimates FFM in older women, in part, because of the influence of truncal adiposity.
Key Words: Skeletal muscle mass • dual-energy X-ray absorptiometry • sarcopenia • total body potassium • total body nitrogen • bioimpedance analysis • aging • anthropometry • humans
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To quantify age-associated decreases in SMM and the effects of interventions, accurate estimation of SMM is required. This has proven to be difficult because there is no direct in vivo means of measuring SMM. There are, however, several methods of indirect estimation, including anthropometric fractionation (10); creatinine excretion (11); whole-body counting and neutron activation to quantify total body potassium (TBK) and total body nitrogen (TBN), respectively (12, 13); computed tomography (14); and magnetic resonance imaging (15). These methods are all time-consuming and technically difficult to perform, and those methods often ranked the highest for accuracy (computed tomography and magnetic resonance imaging) involve considerable radiation exposure or expensive instrumentation.
Dual-energy X-ray absorptiometry (DXA) is a relatively new method of body-composition analysis that involves minimal radiation (16). A whole-body DXA scan divides the body into bone, fat, and lean compartments. With appropriate definition of arm and leg regions, DXA provides an estimation of the fat-free soft tissue (FFST) in the limbs. If it is assumed that this limb FFST value closely represents limb SMM, as discussed by Heymsfield et al (17), then total SMM can be calculated for normal individuals on the basis of the proportion of limb SMM to total SMM (0.75) reported in cadaver studies (18). This method of SMM estimation was shown to agree well with computed tomography–determined SMM in 25 young to middle-aged men (13). In addition, DXA-determined limb SMM was found to correlate strongly with TBK in a sample of 148 women and 136 men aged 20–90 y (6). Because DXA can also quantify whole-body bone mineral density and bone mineral content (BMC) together with total FFM in a 5–6-min scan involving negligible radiation exposure (<0.05 mSv), it is emerging as a popular body-composition assessment tool.
DXA has not been thoroughly validated for SMM estimation in older subjects. The primary aim of this study was, therefore, to compare DXA-derived SMM values with those obtained via TBK and TBN measurement in a sample of 51–84-y-old women and men. Because some uncertainties exist regarding the accuracy of DXA determination of soft tissue components in older age groups (19–21), and few reports on the use of Norland (Fort Atkinson, WI) DXA instruments exist in the literature, a secondary aim was to compare the whole-body FFM values obtained with this instrument with those estimated by bioimpedance analysis, TBK, skinfold-thickness measurements, and a multicompartment model.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). On recruitment, they were all, by self-report, weight stable and apparently healthy. Each subject gave informed consent for the study, which was approved by the Royal North Shore Hospital Medical Research Ethics Committee and Radiation Protection Committee.
|
Dual-energy X-ray absorptiometry
Total body and regional BMC, FFM, and FFST were estimated by whole-body DXA scan (model XR36; Norland) with a scan speed of 25 cm/s and analyzed with version 2.5.2 software. The instrument was calibrated daily with the manufacturer's spine and soft tissue phantoms. The fat-lean phantom uses 77 combinations of acrylic and aluminum and is an advanced approach to DXA calibration (25). The software assumes a weighted linear fat distribution model that permits extrapolation of the fat content of bone-containing pixels (26).
Truncal, abdominal, arm, and leg regions were marked on the scan image by one operator before analysis, as described by Heymsfield et al (17). DXA-derived FFM (FFMDXA) was calculated from the analyzed scan output by summing total BMC and FFST. As an index of truncal adiposity, the ratio of truncal fat to lean tissue (truncal fat:lean) was calculated from the output by dividing truncal fat mass by truncal FFST. The precisions of total BMC, FFMDXA, and FFST measurements were 1.4%, 2%, and 2%, respectively.
Anthropometry
Height was measured to the nearest 0.5 cm with a wall-mounted stadiometer and body weight was measured to the nearest 0.1 kg with digital scales. Skinfold thicknesses were measured in duplicate by one researcher with constant-pressure calipers (Holtain Ltd, Crymych, United Kingdom) at the triceps, biceps, subscapular, and suprailiac sites. The skinfold-thickness measurements were used to estimate percentage body fat with the appropriate age- and sex-specific equations of Durnin and Womersley (27). The precision of percentage fat estimation was 1.1%. Skinfold thickness–derived FFM (FFMSKF) was calculated from percentage body fat and body weight measurements.
Bioimpedance analysis
A bioimpedance analysis (BIA) measurement was taken after each subject had rested supine for 5 min, with electrodes in a tetrapolar configuration, by using a swept-frequency instrument (SEAC model SFB2.3 with associated software; UniQuest, Queensland, Australia). The BIA output measures were used to derive FFM (FFMBIA) by applying the equation of Lukaski et al (28), and total body water (TBW) by using the equation of Kushner and Schoeller (29). The latter equation was derived in a group of men and women aged 
20–70 y; the equation predicts deuterium oxide space from a combination of subject resistance, height, and weight. The precisions of FFMBIA and TBW measurements were 1.6% and 1.4%, respectively.
Data reduction and analysis
Skeletal muscle mass estimation
SMM was estimated from TBK and TBN data (ie, SMMKN) by using the equation of Wang et al (13) as follows:
|
| (1) |
Because this equation was developed empirically by relating TBK and TBN to computed tomography–determined SMM in a multiple regression model, it can be regarded as a surrogate measure of computed tomography–determined SMM and should therefore accurately represent SMM.
DXA-derived SMM (SMMDXA, in kg) was calculated from the sum of arm and leg FFST values (in kg), assuming that this sum represents limb SMM and that limb SMM represents 75% of total body SMM, as discussed above:
|
| (2) |
Fat-free mass based on a 4-compartment body-composition model
A widely used 4-compartment body-composition model assumes that the body consists of fat, protein, water, and mineral compartments (30, 31). FFM (FFM4C, in kg) was calculated from this model as follows:
|
| (3) |
As discussed by Baumgartner (30), the assumed nitrogen content of protein (1 g N = 6.25 g protein) and the ratio of osseous to nonosseous mineral (0.84) involved in these calculations are applicable to subjects in this age range.
Bone mineral and water content of fat-free mass
The mineral content of FFM was calculated by expressing BMC as a percentage of FFMSKF. Similarly, hydration of FFM was estimated by expressing TBW as a percentage of FFMSKF. These measures were included to assess their potential contribution to intermethod differences in body-composition mea-surement because variance in these FFM components is known to affect body density and can thereby reduce the accuracy of 2-compartment models (26, 30, 31).
Statistical analysis
Correlation and regression analysis, Student's t tests, and Bland-Altman analyses (32) were used to compare the 2 methods of SMM estimation and to compare FFMDXA with the other methods (TBK, SKF, BIA, and 4C) of FFM determination. All statistical analyses were carried out with SPSS for WINDOWS (release 6.1.4; SPSS Inc, Chicago). The level of significance was set at P < 0.05 for all analyses.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
| (4) |
|
| (5) |
Paired t tests showed no significant differences between the 2 estimates of SMM in data from women and men (Table 2). Bland-Altman analyses (32) did not show any systematic bias, for either sex, in the differences between the 2 methods as SMM values increased.
|
). The 95% CIs for the regression slope given in Figure 1 include the line of identity. A paired t test showed that although mean SMMKN was 0.32 kg higher than mean SMMDXA, this difference was not significant. Further analysis with the Bland-Altman method (32) revealed that there was no systematic bias in the differences between the 2 methods as SMM values increased (Figure 2).
|
|
Mineral and water content of fat-free mass
Mean (±SD) values for BMC as a percentage of FFM were 5.4 ± 0.6% and 6.2 ± 0.8% for men and women, respectively. Mean values for TBW as a percentage of FFM were 73.0 ± 5.2% and 73.8 ± 4.3% in men and women, respectively.
Fat-free mass estimated by dual-energy X-ray absorptiometry compared with other methods
In contrast with the close agreement between SMM estimates, there were relatively large and significant differences between FFMDXA and several other FFM estimates, particularly in the data set from women (Table 2
). To determine whether these intermethod differences were related to variables such as age, adiposity, fat distribution, BMC as a percentage of FFM, or TBW as a percentage of FFM, a series of Bland-Altman analyses (32) were performed on pooled data from men and women by plotting the difference between FFM4C and FFMDXA against these variables. The difference between the FFM methods was not related to age or percentage fat, but was related to BMC as a percentage of FFM (r = 0.43, P = 0.0001) and TBW as a percentage of FFM (r = 0.26, P = 0.02). There was a highly significant relation (r = 0.58, P < 0.0001) to truncal adiposity, as reflected in the ratio of truncal fat to leg fat (Figure 3).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The strength of agreement between the SMMDXA and SMMKN estimates was striking, given that these methods involve totally independent assumptions. Although neither of the methods has been extensively validated against a gold standard SMM measure in older subjects, the close agreement between the regression line and the line of identity in Figure 1 implies that both methods were in fact measuring the same variable: SMM. Furthermore, because the equation used in this study to derive SMMKN was empirically derived from computed tomography scans in subjects with a considerably wide range of SMM (13), similar results would be expected if the DXA method were to be compared directly with computed tomography–derived SMM in healthy older subjects.
This agreement between the DXA and the KN methods implies that reasonably accurate determinations of SMM in healthy elderly people are possible by either method. The speed of data acquisition (5–6 min for the Norland model XR36 whole-body scan) and the lower radiation exposure with DXA compared with a TBN measurement (TBK involves no radiation exposure) make DXA an attractive option. In our laboratory, a combined TBN and TBK assessment takes 60 min to complete. However, the fact that we found SMMKN values to be more highly correlated with age than were SMMDXA values (in women), considered together with the lower SDs found in the SMMKN data (Table 2), could indicate advantages for the nuclear method.
Use of the nuclear techniques can yield valuable clinical information in addition to SMM estimation. Total body protein, assessed via TBN, is an important indicator of nutritional status and has been shown to reflect the severity of several illnesses (23, 24, 34). Comparison of patient values with age- and sex-matched norms from a healthy reference population can therefore be an important prognostic guide and assist in clinical decision-making (24, 34, 35). Moreover, because TBK is more likely than TBN to show short-term changes, TBK is invaluable in the assessment of acute disturbances in nutritional status, severity of disease, and response to treatment (23, 36). Thus, the combination of TBK and TBN contributes much valuable information to a patient's nutritional assessment, particularly when serial changes associated with illness or treatment are under investigation.
Although our data suggest that DXA correctly estimates SMM via appendicular FFST measurement, we found significant discrepancies between FFMDXA and other FFM measures, particularly in women. This supports the mounting evidence that DXA does not accurately estimate total FFM in older subjects. Roche et al (37) observed that FFM from a multicomponent model was 1.1 kg higher (P < 0.01) than FFMDXA measured by a Lunar Corporation (Madison, WI) instrument in 19 white women aged 54–68 y. Multicomponent FFM was only 0.2 kg higher (NS) than FFMDXA in 8 men in the same age group. Nord and Payne (38) compared percentage fat values obtained by DXA (Norland) and hydrodensitometry in 219 adult subjects. DXA-derived values were consistently higher than hydrodensitometry-derived values, particularly in women, implying a considerable underestimation of FFM by DXA. In contrast, analysis of percentage fat and weight data from Snead et al (21) shows that FFMDXA measurement by Hologic (Waltham, MA) equipment was greater than hydrodensitometry-derived FFM by 3.5 and 4.7 kg, respectively, in women and men aged >60 y. These studies indicate that the difference between FFMDXA and other FFM estimates varies considerably depending on the instrument used, with the Lunar and Norland instruments underestimating and the Hologic instrument overestimating FFM relative to the criterion method.
Our finding that truncal adiposity was positively related to intermethod FFM differences suggests that the fat distribution model in the DXA system software is a critical factor in determining the accuracy of whole-body soft tissue estimates. The SMM data indicate that the software permits good separation of fat and lean tissue in regions where the boundaries between bone and soft tissue are readily defined, such as the limbs. However, the method appears to have difficulty in separating fat and lean tissue in regions such as the trunk, where bone and soft tissue boundaries are far more irregular and fat content can be more variable (20). This shortcoming in software modeling is highly likely to affect all DXA instruments. Snead et al (21) found that Hologic instrumentation markedly underestimated exogenous fat when lard was placed over the trunk of subjects, but detected the additional fat accurately when it was placed over the thighs.
In addition to fat distribution, 2 other groups of factors could potentially influence the intermethod FFM agreement. First, if the BMC and water content of FFM vary markedly from the assumed normal values used in the original hydrodensitometry analyses, the assumed density of FFM will be incorrect. This will introduce errors when methods such as skinfold-thickness measurements and BIA are used, which are highly influenced by body density analysis (20, 30, 37). The influence of these factors in the current study was reflected in the positive correlations between intermethod FFM differences and BMC:FFM and TBW:FFM. Thus, the errors involved in predicting FFM from skinfold-thickness measurements and BIA are relatively large in comparison with those associated with criterion FFM methods such as hydrodensitometry and isotope dilution (30, 31). Similarly, because the potassium content of FFM has been shown to decline with age (36), the use of an assumed value for TBK:FFM introduces an error into the extrapolation of FFM from TBK (30, 36). Second, technical factors—including instrument make and model, calibration technique, scan speed, and software—are highly likely to affect the accuracy of body-composition measurement by DXA (20, 39). In consideration of these multiple confounding factors, DXA should be used cautiously in the measurement of FFM and fat mass in older subjects and results from different instruments should be interpreted conservatively.
In summary, the results of this study indicated that both DXA and KN measurement can be used to estimate SMM in healthy elderly subjects with reasonable accuracy and precision. DXA has the practical advantage of convenience over the nuclear method. However, when comprehensive body-composition assessment is justified, particularly in the study of disease states in which constancy assumptions are challenged, measurements of TBK and TBN provide more detailed information for nutritional analysis. Although DXA provides FFM values that agree quite well with FFM estimates from other methods in older men, it underestimates FFM in older women, possibly because of inaccuracies in defining truncal fat. Further studies are necessary to confirm whether other DXA instrument models and software can provide accurate assessment of SMM in this age group and in patient groups. Thus, these conclusions apply only to healthy subjects. The use of nuclear techniques is recommended for thorough body-composition investigations of disease states.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
 |
|
 S. M. Roth, M. A. Schrager, M. R. Lee, E. J. Metter, B. F. Hurley, and R. E. Ferrell Interleukin-6 (IL6) Genotype Is Associated With Fat-Free Mass in Men But Not Women J. Gerontol. A Biol. Sci. Med. Sci., December 1, 2003; 58(12): B1085 - 1088. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Salinari, A. Bertuzzi, G. Mingrone, E. Capristo, A. Scarfone, A. V. Greco, and S. B. Heymsfield Bioimpedance analysis: a useful technique for assessing appendicular lean soft tissue mass and distribution J Appl Physiol, April 1, 2003; 94(4): 1552 - 1556. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. D Hansen and B. J Allen Habitual physical activity, anabolic hormones, and potassium content of fat-free mass in postmenopausal women Am. J. Clinical Nutrition, February 1, 2002; 75(2): 314 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Roth, M. A. Schrager, R. E. Ferrell, S. E. Riechman, E. J. Metter, N. A. Lynch, R. S. Lindle, and B. F. Hurley CNTF genotype is associated with muscular strength and quality in humans across the adult age span J Appl Physiol, April 1, 2001; 90(4): 1205 - 1210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Broeder, J. Quindry, K. Brittingham, L. Panton, J. Thomson, S. Appakondu, K. Breuel, R. Byrd, J. Douglas, C. Earnest, et al. The Andro Project: Physiological and Hormonal Influences of Androstenedione Supplementation in Men 35 to 65 Years Old Participating in a High-Intensity Resistance Training Program Arch Intern Med, November 13, 2000; 160(20): 3093 - 3104. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. W. Welch and M. R. Sowers The Interrelationship between Body Topology and Body Composition Varies with Age among Women J. Nutr., September 1, 2000; 130(9): 2371 - 2377. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and men (•) combined. The fine diagonal line is the line of identity; the bold line is the regression line.
