Validation of a bioelectrical impedance analysis equation to predict appendicular skeletal muscle mass (ASMM)

Clinical Nutrition (2003) 22(6): 537–543 r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0261-5614(03)00048-7

ORIGINAL ARTICLE

Validation of a bioelectrical impedance analysis equation to predict appendicular skeletal muscle mass (ASMM) U. G. KYLE,1 L. GENTON,1 D. HANS,2 C. PICHARD1 1 Clinical Nutrition, Geneva University Hospital, Geneva, Switzerland, 2 Nuclear Medicine, Geneva University Hospital, Geneva, Switzerland (*Correspondence to: CP, Clinical Nutrition ,Geneva University Hospital, Geneva1211, Switzerland)

Abstract" Rationale: Appendicular skeletal muscle mass (ASMM) is useful in the evaluation of nutritional status because it re£ects the body muscle protein mass. The purpose of this study was to validate, against dual-energy X-ray absorptiometry (DEXA), a BIA equation to predict ASMM to be used in volunteers and patients. Method: Healthy men (n = 246 men, BMI 25.372.9 kg/m2) and women (n =198, 24.173.6 kg/m2), and heart, lung and liver transplant patients (213 men, BMI of 24.674.4 kg/m2; 113 women, BMI 23.075.2 kg/m2) were measured by BIA (XitronTechnologies) and DEXA (Hologic QDR 4500). A BIA equation to predict ASMM (kg) that included height2/resistance, weight, gender, age and reactance, was developed by means of multiple regressions. Results: ASMM (kg) Volunteers Patients

Men DEXA 25.873.6 22.172.8

Women BIA 25.773.4 22.673.5*

DEXA 17.372.5 15.272.8

BIA 17.272.4 15.273.0

Mean7SD, paired t-test between BIA and DXA * Po0.01 Mean di¡erence (Bland-Altman) for volunteers was 0.171.1kg, r =0.95, SEE1.12 kg and for patients
0.471.5 kg, r =0.91, SEE 1.5 kg.Best ¢tted multiple regression equation was
4.211 + (0.267  height2 / resistance) + (0.095  weight) +(1.909  sex (men = 1, women = 0)) + (
0.012  age) + (0.058  reactance). Conclusions: BIA permits the prediction of ASMM in healthy volunteers and patients between 22 and 94 years of age. A slightly larger, though clinically not significant, error was noted in patients. r 2003 Elsevier Ltd. All rights reserved.

impediment to determining ASMM is the lack of suitable, easy and non-invasive methods for estimating ASMM. Earlier studies support the validity of DXA estimates of ASMM (4, 5). However, DXA is not a ‘portable’ method and measurement cost and technician skill limit its use in field studies. Bioelectrical impedance analysis (BIA) has been used to determine the fat-free mass (6). Recent studies indicate good correlation between limb electrical resistance, measured at 50 kHz, and ASMM by DXA (7, 8). This observation suggests that limb skeletal muscle can be estimated from BIA-measured resistance. Janssen et al. (9) recently developed a BIA equation, validated against magnetic resonance imaging (MRI) that provides valid estimates of total skeletal mass in healthy adults varying in age and adiposity. Lee et al. (10) developed anthropometric prediction models to estimate total body skeletal muscle mass, using skinfold and limb circumferences. Baumgartner et al. (11) estimated ASMM from anthropometric parameters, including hip circumference and grip strength in elderly

Key words: appendicular skeletal muscle mass; validation; bioelectrical impedance analysis; dual X-ray absorptiometry

Introduction Aging is associated with a gradual loss of skeletal muscle mass or sarcopenia. Wasting diseases and many health related conditions also result in decreases in skeletal muscle mass (1). Appendicular (or limb) skeletal muscle mass (ASMM) accounts for475% of total skeletal muscle (2) and is the primary portion of skeletal muscle involved in ambulation and physical activity. In the elderly, skeletal muscle mass loss may be masked by weight stability resulting from a corresponding increase in total body fat mass (1). Crosssectional data suggest that the loss of ASMM is greater with aging than the loss on non-skeletal muscle mass (3). Thus, the evaluation of ASMM can contribute important information to the assessment of nutritional status because it reflects the body protein mass. A major 537

538

SKELETAL MUSCLE MASS AND BIOELECTRICAL IMPEDANCE

subjects. Pietrobelli et al. (8) found BIA to be valid for estimating arm and leg skeletal muscle mass. Estimation of regional muscle mass by segmental BIA measurements, compared to MRI (12–14), requires further validation before it can be applied in clinical settings. Currently there are no BIA-determined prediction equations to estimate ASMM. The purpose of this study was to validate, against dual-energy X-ray absorptiometry (DXA), a BIA equation to predict ASMM to be used in volunteers and patients. Subjects and methods Volunteers Four hundred and forty-four healthy ambulatory Caucasians (246 men and 198 women) aged 22 –94 years (Table 1) were included in this study. Subjects were non-randomly recruited through advertisement in local newspapers and invitations to participate in the study sent to members of elderly leisure clubs. Although subjects were non-randomly selected, statistical analysis revealed no difference in height, weight and body mass index (BMI) between subjects in this study and agematched healthy men and women (n = 3170) in Geneva. Exclusion criteria were active medical treatment or hospitalization within 3 months of measurement or physical handicap that might interfere with body composition measurement (amputation, paralysis, etc.). Three women, aged over 65 years who had BMIs ofo18.0 kg/m2, but no history of recent weight loss or illness were considered healthy and therefore not excluded. Each subject was first measured by BIA, then by DXA. All subjects signed an informed consent statement and the study protocol was approved by the Geneva University Hospital Ethics Committee.

Patients Patients (213 men and 113 women) aged 18–70 years (Table 1) who were seen as part of the pre-transplant evaluation or during post-transplant followup examination were also included in this study. Patients with clinically detectable ascites or other fluid abnormalities require therapeutic correction were excluded (15). Bioelectrical impedance analysis Body height was measured to the nearest 0.5 cm and body weight was measured to the nearest 0.1 kg on a balance beam scale. Height and weight of both groups were normally distributed. Briefly, an electrical current of 50 kHz and 0.8 mA was produced by a generator, Xitron 4000B (Xitron Technologies, Inc, San Diego, CA, USA), and applied to the skin using adhesive electrodes (3 M Red Dott, 3 M Health Care, Borken, Germany) with subject in supine position. The skin was cleaned with 70% alcohol. Standard whole body procedure as previously described was used (16) The resistance and reactance were measured by the BIA generator and used to mathematically derive ASMM (16,17) using the formula V ¼ rx height2/resistance in which conductive volume (V) is assumed to represent ASMM, r is the specific resistivity of the conductor, height (ht) is taken as the length of the conductor, and body resistance (R) is measured with four surface electrodes placed on the right wrist and ankle. Reactance (the related BIA factor of phase angle) was included as an additional potential measure of soft tissue composition (8). Short- and long-term reproducibility of resistance measurements indicate coefficients of variation of 1.8–2.9% (18,19). In our data, reproducibility was r ¼ 0:999 for measurements taken in the same subject within one week (n ¼ 29) and 0.977 for repeat measurements up to 1 month (n ¼ 40) or 2.5% variance.

Table 1 Anthropometric and bioelectrical impedance (BIA) characteristics of volunteers and patients Men

Women

Volunteers

20-94 years

o55 years

455 years

20-94 years

o55 years

455 years

N Height Weight Body mass index Resistance Reactance Height2/resistance Fat-free mass Patients N Height Weight Body mass index Resistance Reactance Height2/resistance Fat-free mass

Cm Kg Kg/m2 O O Cm2/O Kg

246 175.177.7 77.779.9 25.372.9 457747 54.679.8 67.979.1 60.376.7

144 177.976.9 79.379.7 25.072.6 451742 60.077.1 71.078.5 63.075.9

102 171.176.9 75.479.8 25.873.2 467751 47.177.9 63.578.2 56.676.0

198 162.176.2 63.3710.0 24.173.6 564760 60.4710.7 47.376.6 42.975.0

85 165.375.4 62.678.1 22.972.7 555756 66.979.2 49.876.2 44.974.6

113 159.775.8 63.9711.2 25.173.9 571762 55.579.1 45.376.2 41.474.7

Cm Kg Kg/m2 O O Cm2/O Kg

213 172.777.0b 73.3713.3b 24.674.4a 516791b 46.3713.9b 59.5710.9b 55.277.6b

127 174.276.9b 72.9714.9b 24.175.1 522798b 49.1714.1b 60.0710.6b 55.678.1b

86 170.376.5 73.8710.8 25.472.9 507778b 42.2712.6a 58.9711.3a 54.776.9a

113 160.775.2a 59.4714.5a 23.075.2a 6207103b 50.1714.2b 42.978.1b 40.376.4b

80 161.675.2b 59.0712.9a 22.574.5 618780b 51.9714.0b 43.076.6b 40.576.0b

33 158.674.6 60.6718.0 24.076.6 6257146a 45.7713.9b 42.6711.1 39.977.3

Paired t-Test between volunteers and patients

a

P o 0.05,

b

P o 0.001.

CLINICAL NUTRITION

Dual energy X-ray absorptiometry Body composition was determined with DXA, Hologic QDR 4500A (Hologic Inc., Waltham, MA, USA), Enhanced 8.26 Whole-body software version. In this scanning technique, an X-ray generator emits switched pulsed radiation of two energies, 100 and 140 kVp, in a fan-beam mode. As they pass through the body, these two X-ray beams are attenuated due to the absorption and scattering of the photons. The attenuation is measured for every pixel of the body surface by a linear array of 216 detectors. A complex development measurement allows determination of bone mineral and soft tissue densities. Soft tissue can further be partitioned into body fat and fat-free mass, since they have different attenuation characteristics. The scanner was calibrated for bone mineral component with a rotating drum and for fat with an external lucite-aluminum phantom (20). The calculation of ASMM has been previously described in detail (4). With the use of specific anatomic landmarks, the legs and arms are isolated on the skeletal X-ray planogram (anterior view). The arm encompasses all soft tissue extending from the center of the arm socket to the phalange tips and contact with the ribs, pelvis and greater trochanter is avoided. The leg consists of all soft tissue extending from an angled line drawn through the femoral neck to the phalange tips. The system software provides the total mass, bone mineral mass, lean and fat mass extremities for each region. The fat and bone mineral-free portion of the extremities were assumed to represent ASMM along with a small and relatively constant amount of skin and underlying connective tissue (21). Radiation dose per individual was 2.6 mSv. To our knowledge, no information on precision of QDR 4500A for measuring body composition is available, but the precision of DXA QDR 2000, a former model of Hologic, is 1.0% for FFM and 2.0% for FM (22). The FM derived from DXA measurements in previous studies correlates well with the FM determined by hydrodensitometry and total body K40 measurements (23,24).

Statistics Descriptive statistics were calculated for height, weight, body mass index and bioelectrical impedance parameters, including resistance, reactance and height2/ resistance and are expressed as mean 7 standard deviation (x7SD). ASMM value measured by DXA was used as the criterion measurement. Stepwise multiple regressions were used to derive a prediction equation by BIA. Predictor variables, entered into the BIA model in the order of highest correlation coefficient and smallest SEE, were height2/resistance, weight, gender, age and reactance. The prediction equation for BIA was developed by using a double cross-validation technique (25).

539

The subjects were split into two samples of Odd and Even on the basis of age (odd youngest, even second youngest). Thus the two groups were evenly matched for the age of the subjects. An equation was developed for each, with the opposite group being used to crossvalidate each equation. If the equations proved to be similar (evaluated by a comparison of multiple r values and visible inspections of graphs), groups were combined and a single equation was developed using the entire sample. The equation was separately validated ino55 and over 55 year subjects. Simple regressions were calculated to test correlations of ASMM between DXA and BIA. T-test was used to test differences between methods. Bland and Altman analysis was calculated according to methods previously described (26) to assess the agreement between two clinical measurements. The difference between the values is plotted against their mean, the mean being the best available estimate of the true value. This analysis allows for the calculation of bias (estimated by the mean differences), the 95% confidence interval for the bias, and the limits of agreement (two standard deviations of the difference) (26). The technical error was calculated as TE ¼ OSn i ¼ 1 (ASMM1—ASMM2 )2/ 2n (27). It represents the dispersion of the differences from a normal distribution, 95% of the values for a measurement should be within plus or minus twice the technical error values of the corresponding true value. Ideally, body composition methods should have low TE and high correlation coefficients (27). Statistical significance was set at P p 0.05 for all tests.

Results A total of 444 healthy volunteers between ages 22 and 94 years and 326 patients were included. Table 1 shows their anthropometric and bioelectrical impedance analysis characteristics. The weight was significantly higher in volunteers than patients. The resistance and height2/ resistance were significantly higher and the reactance significantly lower in patients than volunteers. The reactance was significantly lower in volunteers and patients455 year than in those o 55 years. The prediction equations developed in the odd (n = 222) and even numbered subjects (n = 222) are shown in Table 2. The BIA-predicted ASMM by the odd or even equations was not different from the DXA-measured ASMM. The odd equation resulted in a predicted ASMM of 22.175.3 kg in the even-numbered subjects (r = 0.977, SEE = 1.14 kg). The even equation resulted in a predicted ASMM of 21.875.1 kg in the odd numbered subjects (r = 0.976, SEE = 1.11 kg). Thus, the correlation coefficients (r values) and SEE were similar between the even and odd samples and the crossvalidation showed similar results. The regression lines (Odd = 0.95370.01 and Even = 0.95370.01) were

540

SKELETAL MUSCLE MASS AND BIOELECTRICAL IMPEDANCE

Table 2 Contribution and order of entry of independent variables to the BIA model for ASMM Cumulative Prediction variables BIA Height2/resistance + Weight + Gender + Age + Reactance Bmi Height Height, weight Height, weight, age

2

Individual 2

r

SEE

P value

r

SEE

P value

0.917 0.924 0.933 0.948 0.953

1.53 1.46 1.38 1.21 1.15

0.0001 0.0001 0.0001 0.0001 0.0001

0.917 0.668 0.644 0.164 0.038 0.123 0.688 0.816 0.830

1.53 3.05 3.15 4.84 5.19 4.95 2.96 2.27 2.18

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

n = 444 subjects, r2 value of the validity coefficient, SEE=Standard error of the estimate. Table 3 Prediction equation for appendicular skeletal muscle mass (ASMM), using all volunteers
4.211 + (0.267*height2 / resistance) + (0.095*weight) + (1.909*sex (men = 1, women = 0)) + (
0.012*age ) + (0.058*reactance) DXA-measured ASMM 22.075.3 kg BIA-predicted ASMM 21.975.2 kg, r = 0.976, SEE = 1.12 kg, TE 0.65 kg ASMM =

n = 444 subjects, r = validity coefficient, SEE = standard error of the estimate, TE = technical error (see methods)

Table 4 Comparison of appendicular skeletal muscle mass by DXA and BIA as estimated by equation determined in volunteers Volunteers

Patients

Age (years)

DXA(kg)

BIA(kg)

t-test

DXA(kg)

BIA(kg)

t-test

Men 20-94 o 55 4 55

25.873.6 27.473.0 23.573.0

25.773.3 27.272.7 23.772.9

0.58 0.07 0.17

22.173.8 22.673.9 21.373.5

22.673.5 23.073.5 22.173.5

0.001 0.001 0.001

Women 20-94 o 55 4 55

17.372.5 18.372.5 16.572.3

17.272.4 18.572.0 16.372.3

0.61 0.16 0.06

15.272.8 15.272.5 15.273.5

15.273.0 15.372.4 14.874.1

0.96 0.36 0.13

Paired t-test: comparison between DXA and BIA, significance level Po 0.05

virtually identical, with deviation from the line of identities being similar for both samples. Thus a single equation using all 444 subjects was developed for BIA prediction of ASMM, shown in Table 2. The order of entry of predictor variable was height2/ resistance, weight, gender, age and reactance, which each contributed significantly to the BIA Model (Table 3). Height2/resistance accounted for 91% of the variability (SEE 1.5 kg) of the equation, whereas weight alone only accounted for 67% of the variability (SEE 3.0 kg) and height alone accounted of 69% of the variability (SEE 5.0 kg). Inclusion of height, weight and age, without BIA parameters, accounted for 83% of the variability with a SEE of 2.2 kg. BMI was a poor predictor of ASMM (r2 = 0.12, SEE 4.95 kg). The prediction equation developed from all subjects is shown in Table 2. Thus, the inclusion of BIA parameters clearly improved the prediction power and decreased the SEE, compared to anthropometric parameters only. Table 4 shows the mean ASMM and t-test of the prediction equation in healthy volunteers and patients agedo55 and 455 years. Non-significant mean differ-

ence in ASMM between DXA and BIA equations in volunteers ranged from
0.1 to +0.2 kg. Thus the combined equation is valid to predict FFM in healthy volunteers aged o55 and 455 years. Mean difference in ASMM between DXA and BIA equations in patients ranged from
0.4 to +0.6 kg and t-tests were significant for male patients. Figure 1 shows the correlation (top) and mean difference, according to Bland-Altman (bottom), using the combined equation in healthy volunteers (left) and patients (right). Discussion The aim of the study was to develop and cross-validate a prediction equation for estimating ASMM from BIA measurements. Our findings indicate that DXA-measured ASMM was strongly correlated to the BIAderived resistance, normalized for height, (ht2/R) and that the BIA method is a valid method for estimating ASMM in healthy volunteers and patients. The error (SEE) for predicting ASMM was 1.1 kg (5%) in volunteers and 1.5 kg (7.6%) in patients.

CLINICAL NUTRITION

Y = 1.01 + 0.95 * X, r2 = 0.95 SEE 1.12 kg, TE 0.65 kg

ASMM (bia) (kg)

40

Y = 1.05 + 0.97 * X, r2 = 0.91, SEE 1.50 kg, TE 1.12 kg 40

35

35

30

30

25

25

20

20

15

15 10

10 10

Difference ASMM dxa-bia (kg)

541

15

20 25 30 35 ASMM (dxa) (kg)

10

40

6

6

4

4 + 2 SD 2

2 0

0

-2

- 2 SD -2

-4

-4

-6 10

15

20

25

30

35

40

ASMM (dxa + bia)/2(kg)

-6 10

15

20 25 30 ASMM (dxa) (kg)

35

40

+ 2 SD

-2 SD

15 20 25 30 35 ASMM (dxa+bia)/2 (kg)

40

Fig. 1 Correlations (top) and differences (bottom) of appendicular skeletal muscle mass (ASMM) in volunteers (left side) and patients (right side) estimated by dual-energy X-ray absorptiometry (ASMMDXA) and bioelectrical impedance. The difference of ASMM (calculated as ASMMDXA
ASMMBIA) per Bland-Altman) is plotted against the mean of the measurements of ASMM by DXA and BIA. SEE = standard error of the estimate, TE = technical error (see methods). d = men, m = women.

Resistance index (height2/R) had an r2 value of 0.92 and thus accounted for 92% of the variability. Additional prediction variables further improved the prediction equation. In agreement with others (8,9), gender and age were significant independent predictors. Furthermore, age was reported to be a significant predictor, even at frequencies 4 50 kHz (8). Our study further indicates that reactance is a significant independent predictor of ASMM and FFM (28) and was responsible for a small improvement in the prediction equation. Reactance was lower in older than younger subjects and lower in patients than volunteers and would explain about 1.0 kg of the difference in ASMM between patients 455 years and volunteerso55 years. Reactance reflects the cell membrane capacitance, tissue interfaces and non-ionic tissues (29), and has been suggested to be a function of intracellular water and thus of body cell mass (30). Thus the findings of lower reactance, lower body cell mass and lower ASMM found in older and ill subjects are consistent. Our data contradicts Pietrobelli et al. (8) who found that phase angle (the related BIA factor of reactance) failed to improve the predicted ASMM. In a review of a number of studies, Houtkouper et al. (6) found that the r2 for equations developed for estimating fat-free mass ranged from 0.73 to 0.98 and that SEE ranged from 1.7 to 4.1 kg. Janssen et al. (9) reported an r2 value of 0.86 and a SEE of 2.7 kg for total skeletal muscle mass. In our study, the r2 value for ASMM was higher (0.95 and 0.91 in volunteers and

patients , respectively) and the SEE was smaller (1.1 or 5.0% and 1.5 kg or 7.6% in volunteers and patients, respectively). The BIA method was within 5% error in 68% and 52%, and within 10% in 93% and 83% of volunteers and patients, respectively. Because the model development size is large, the model should be applicable to a large proportion of a white adult population. Validation in patients also permits the estimation of ASMM in patients without altered hydration due to disease or drug treatment. The slightly lower r2 value and higher SEE noted in patients suggests that patients are different from healthy subjects. The predicted ASMM was significantly higher in male patients, with the largest difference (0.871.6 kg) noted in male patients 455 years. We can only speculate on these differences. There were no clear trends and differences were inconsistent. Patients were measured when they were seen as part of the pre-transplant evaluation or during post-liver, lung or heart transplant followup examination. Larger ASMM differences were noted in some patients evaluated prior to liver transplantation, which might suggest the presence of sub-clinical ascites and/or altered hydration. Although there were no clear trends, overestimation of ASMM was noted in some patients at extremes of BMI (low or high) and patients with lower than expected FFM (FFM below the 10th percentile). Pietrobelli et al. (8) suggest that, although at an empirical level, the development of good skeletal muscle mass prediction models is possible, additional BIA

542

SKELETAL MUSCLE MASS AND BIOELECTRICAL IMPEDANCE

studies are probably necessary to improve the understanding of appendicular conduction pathways. Our study confirms that men have higher ASMM than women and younger subjects had higher ASMM than older subjects. The ASMM in woman was twothirds the ASMM of men. Our ASMM results (Table 4) are similar to the New Mexico Elder Health Survey (11) (22.5 and 14.5 kg in men and women, respectively) and to the younger adults in the Rosetta Study (11) (27.3 and 17.7 kg, respectively). We also found that patients had significantly lower ASMM than volunteers. The ASMM in our study was 76% of total skeletal muscle mass reported by Janssen et al. (9) (29.677.2 kg), which is close to the 75% of the total skeletal muscle being ASMM, as reported in reference man (2). Study limitations DXA was the reference method used in our study. One limitation of this method is subject thickness, which may lead to an overestimation of % FFM in large people (30). However, only one volunteer and eight patients exceeded a BMI of 35 kg/m2 in our study and therefore errors due to this limitation should be minimal. Differences in FFM, and therefore ASMM, have been reported between dual-energy X-ray absorptiometry instruments by different manufacturers (Hologic versus Lunar) (31). We have no information on the comparability of regional estimates across instruments from different manufacturers. These differences would affect the final BIA equation. The second limitation is that the measured ASMM includes not only appendicular skeletal muscle but also non-muscle components as skin, neurovascular tissues, and connective tissue and interstitial fat. Some authors describe an increase with age of these last three elements, which may result in an overestimation of ASMM in elderly compared to other reference methods (21,32). The ASMM determined by DXA includes estimated skin, connective tissue and intramuscular fat and thus would be higher than CT determined ASMM. ASMM was 2.1 (8.1%) and 0.9 kg (5.2%) lower in men and women, respectively, when using regression equation to adjust for differences between CT and DXA (33). However, in clinical application, comparison of direct predicted ASMM would be better than using adjusted ASMM that is based on a number of assumptions which may vary with age and gender and has not been validated in women, obese and elderly subjects (33). ASMM as presented in this study permits comparison of patients to healthy subjects, and of longitudinal followup in individual subjects and group of subjects. Furthermore, soft tissue algorithms rely on a fixed hydration of FFM (0.73 ml/g), which may be questionable in the elderly. Nevertheless, Visser et al. (34) validated DXA in the elderly against a four-compartment model. The described limitations apply to all DXA instruments.

The BIA methods used may be criticized, but have been optimized for this study, namely: Water and electrolyte abnormalities are known to influence body composition measurements, including BIA measurements. To limit the impact of such interference, BIA measurements were performed before IV fluids for medications and treatment for dehydration were started, and patients with edema, and dehydration were excluded. Whole body impedance measurements were used in this study to estimate ASMM. Although segmental impedance measurements to determine ASMM are an interesting concept, we found that segmental (limb) impedance measurements were less accurate in predicting individual limb ASMM than whole body impedance and were therefore not used in this study. This appears to be due to the non-uniform conduction of impedance through various tissues, which distorts the contribution of different body segments. Conclusions The study validated a BIA equation to predict ASMM. BIA permits the prediction of ASMM in healthy volunteers between 22 and 94 years and patients between 18 and 70 years. A slightly larger, though clinically not significant, error was noted in patients. Acknowledgements We thank the Foundation Nutrition 2000Plus for its financial support.

References 1. Gallagher D, Ruts E, Visser M, et al. Weight stability masks sarcopenia in elderly men and women. Am J Physiol Endocrinol Metab 2000; 279: E366–375 2. Snyder WS, Cook MJ, Nasset ES, Karhausen LR, Howells GP, Tipton IH. Report of the task group on reference man. 23rd International Commission on Radiological Protection. Oxford, UK: Pergamon, 1975 3. Kyle UG, Genton L, Hans D, Karsegard L, Slosman D, Pichard C. Age-related differences in fat-free, skeletal muscle, body cell mass and fat mass between 18 and 94 yrs of age. Eur J Clin Nutr 2001; 55: 663–672 4. Heymsfield SB, Smith R, Aulet M, et al. Appendicular skeletal muscle mass: measurement by dual-photon absorptiometry. Am J Clin Nutr 1990; 52: 214–218 5. Wang ZM, Visser M, Ma R, et al. Skeletal muscle mass: evaluation of neutron activation and dual-energy X-ray ablsorptiometry methods. J Appl Physiol 1996; 80: 824–831 6. Houtkooper LB, Lohman TG, Going SB, Howell WH. Why bioelectrical impedance analysis should be used for estimating adposity. Am J Clin Nutr 1996; 64: 436S–448S 7. Nunez C, Gallagher D, Grammes J, et al. Bioimpedance analysis: potential for measuring lower limb skeletal muscle mass. J Parent Enteral Nutr 1999; 23: 96–103 8. Pietrobelli A, Morini P, Battistini N, Chiumello G, Nunez C, Heymsfield SB. Appendicular skeletal muscle mass: prediction from multiple frequency segmental bioimpedance analysis. Eur J Clin Nutr 1998; 52: 507–511 9. Janssen I, Heymsfield SB, Baumgartner RN, Ross R. Estimation of skeletal muscle mass by bioelectrical impedance analysis. J Appl Physiol 2000; 89: 465–471

CLINICAL NUTRITION

10. Lee RC, Wang Z, Heo M, Ross R, Janssen I, Heymsfield SB. Total-body skeletal muscle mass: development and cross-validation of anthropometric prediction models. Am J Clin Nutr 2000; 72: 796–803 11. Baumgartner RN, Koehler KM, Gallagher D, et al. Epidemiology of sarcopenia among the elderly in New Mexico. Am J Epidemiol 1998; 147: 755–763 12. Fuller NJ, Hardingham CR, Graves M, et al. Predicting composition of leg sections with anthropometry and bioelectrical impedance analysis, using magnetic resonance imaging as reference. Clin Sci 1999; 96: 647–657 13. Lukaski HC. Assessing regional muscle mass with segmental measurements of bioelectrical impedance in obese women during weight loss. Ann NY Aca Sci 2000; 904: 154–158 14. Elia M, Fuller NJ, Hardingham CR, et al. Modeling leg sections by bioelectrical impedance analysis, dual-energy X-ray absorptiometry, and anthropometry: assessing segmental muscle volume using magnetic resonance imaging as a reference. Ann NY Aca Sci 2000; 904: 298–305 15. Kyle UG, Genton L, Mentha H, Nicod L, Slosman D, Pichard C. Reliable bioelectrical impedance analysis estimate of fat-free mass in liver, lung and heart transplant patients. J Parent Enteral Nutr 2001; 25: 45–51 16. Lukaski HC, Bolonchuk WW, Hall CB, Siders WA. Validation of tetrapolar bioelectrical impedance measurements to assess human body composition. J Appl Physiol 1986; 60: 1327–1332 17. Kushner RF, Schoeller DA. Estimation of total body water by bioelectrical impedance analysis. Am J Clin Nutr 1986; 44: 417–424 18. Lukaski HC, Johnson PE, Bolonchuk WW, Lykken GL. Assessment of fat-free mass using bioelectrical impedance measurements of the human body. Am J Clin Nutr 1985; 41: 810–817 19. Jackson AS, Pollock ML, Graves JE, Mahar MT. Reliability and validity of bioelectrical impedance in determining body composition. J Appl Physiol 1988; 64: 529–534 20. Slosman DO, Casez JP, Pichard C, et al. Assessment of wholebody composition using dual X-ray absorptiometry. Radiology 1992; 185: 593–598 21. Gallagher D, Visser M, De Meersman RE, et al. Appendicular skeletal muscle mass: effects of age, gender and ethnicity. J Appl Physiol 1997; 83: 229–239 Submission date: 4 November 2002 Accepted: 17 March 2003

543

22. Mazess R, Chesnut III CH, McClung M, Genant H. Enhanced precision with dual-energy X-ray absorptiometry. Calcif Tissue Intern 1992; 51: 14–17 23. Fuller NJ, Laskey MA, Elia M. Assessment of the composition of major body regions by dual-energy X-ray absorptiometry (DEXA), with special reference to limb muscle mass. Clin Physiol 1992; 12: 253–266 24. Van Loan MD, Mayclin PL. Body composition assessment: dualenergy X-ray absorptiometry (DEXA) compared to reference methods. Eur J Clin Nutr 1992; 46: 125–130 25. Kerlinger FN, Pedhazur EJ. Multiple Regression in Behavioral Research. New York: Holt, Rinehard & Winston, 1973 26. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; i: 307–310 27. Wang ZM, Deurenberg P, Guo SS, et al. Six-compartment body composition model: inter-method comparisons of total body fat measurement. Int J Obesity 1998; 22: 329–337 28. Kyle UG, Genton L, Karsegard L, Slosman DO, Pichard C. Single prediction equation for bioelectrical impedance analysis in adults aged 20-94 yrs. Nutrition 2001; 17: 248–253 29. Foster KR, Lukasi HC. Whole-body impedance — what does il measure ? Am J Clin Nutr 1996; 64: 388S–396S 30. Gudivaka R, Schoeller DA, Kushner RF, Bolt MJ. Single-and multifrequency models for bioelectrical impedance analysis of body water compartments. J Appl Physiol 1999; 87: 1087–1096 31. Genton L, Hans D, Kyle UG, Pichard C. Dual-energy X-ray absorptiometry and body composition : differences between devices and comparison with reference methods. Nutrition 2002; 18: 66–70 32. Kim J, Wang Z, Heymsfield SB, Baumgartner RN, Gallagher D. Total-body skeletal muscle mass: estimation by a new dual-energy X-ray absorptiometry method. Am J Clin Nutr 2002; 76: 378–383 33. Wang W, Wang Z, Faith MS, Kotler D, Shih R, Heymsfield SB. Regional skeletal muscle measurement: evaluation of new dualenergy X-ray absorptiometry model. J Appl Physiol 1999; 87: 1163–1171 34. Visser M, Fuerst T, Lang T, Salamone L, Harris T. Validity of fan-beam dual-enery X-ray absorptiometry for measuring fat-free mass and leg muscle mass. Health, Aging, and Body Composition Study-Dual-Energy X-ray Absorptiometry and Body Composition Working Group. J Appl Physiol 1999; 87: 1513–1520