Home page - Search Previous issues
01 Are age-related bioimpedance analysis effects also present in patients with spinal cord injury? Relevance to clinical prediction of skeletal muscle mass
Christopher Nuñez, Dympna Gallagher, Ann Spungen, William Bauman and Steven B. Heymsfield
02 Comparison of whole and regional body composition measured by Hologic QDR-2000 and Lunar DPX-L dual-energy X-ray absorptiometry
Lynn M. Bairos, Bess Dawson-Hughes and Ronenn Roubenoff
03 Validity of predicted percentage body fat from skinfolds in Singapore Chinese, Malays and Indians
Paul Deurenberg and Mabel Deurenberg-Yap
04 Comparison of estimated percentage body fat from foot-to-hand, foot-to-foot and hand-to-hand bioimpedance analysis with densitometry in young females
Paul Deurenberg and Frans JM Schouten
05 Validation of dual-energy X-ray absorptiometry in the assessment of change in fat compartments, compared to measurement by magnetic resonance imaging, in HIV-infected adults, SJ Lan, ES Engelson, D Agin2 , P. Homel, J Wang, SB Heymsfield and DP Kotler
06 The adjustment of measures of energy expenditure for body weight and body composition
PSW Davies1 and TJ Cole2
International Journal of Body Composition Research 2003, Vol. 1 No. 1: 11-16
Christopher Nuñez1, Dympna Gallagher1, Ann Spungen2, William Bauman2 and Steven B. Heymsfield1
1Department of Medicine, Obesity Research Center, St Luke’s-Roosevelt Hospital, Columbia University, College of Physicians and Surgeons, New York, NY; 2The Spinal Cord Damage Research Center, Veterans Administration Medical Center, Bronx, NY, Mount Sinai School of Medicine, New York, NY, USA.
Bioimpedance analysis (BIA) is a practical means of potentially evaluating skeletal muscle mass in clinical populations. However, measured extremity resistance at 50 kHz appears not solely determined by muscle size and length. Rather, an ‘age’ predictor term appears in most BIA regression models and the basis of this senescencerelated effect remains uncertain. The aim of this study was to establish if ‘premature’ skeletal muscle atrophy, as induced by spinal cord injury (SCI), alters measured leg resistance at 50 kHz in a pattern similar to that observed with aging. As skeletal muscle atrophy beyond that for age is present in many clinically-relevant populations in whom BIA can be applied, the finding of an age-related BIA pattern in SCI patients would have important implications for underlying BIA theory and prediction formula development. Conventional foot-to-foot 50 kHz BIA was applied to healthy adults and subjects with SCI with an estimate of leg skeletal muscle provided by dualenergy X-ray absorptiometry-measured leg lean soft tissue (LST). Leg resistance (R) and height2/R were set as dependent variables and LST, gender, age, and group (control or SCI) set as potential independent predictor variables in multiple regression models. The hypothesis tested, based upon earlier research, was that after controlling for LST, stature, and age, subjects with SCI would have a lower leg resistance than control subjects. A total of 91 control subjects and 11 subjects with SCI completed the study protocol. Leg R and H2/R were both well correlated with LST (r=0.40 and 0.77, P<0.001). After controlling for LST, age entered as a significant predictor variable in both the leg R and H2/R prediction models. Additionally, group (ie control or SCI) was a significant predictor variable after controlling for other predictor variables in both models with composite r values of 0.63 and 0.80, respectively (both P<0.001). Resistance, according to these models, is lower in subjects with SCI than in controls after adjusting for the other predictor variables. It is concluded that aging and SCI produce similar effects on BIA resistance at 50 kHz, after controlling for lean soft tissue and other covariates. Establishing the mechanisms of these electrical changes is important because BIA is often applied in patients with SCI and other conditions associated with skeletal muscle atrophy. Until these mechanisms are understood and corrected, population-specific BIA formulas will be required in skeletal muscle mass prediction by BIA.
International Journal of Body Composition Research 2003, Vol. 1 No. 1: 17-22
Lynn M. Bairos, Bess Dawson-Hughes and Ronenn Roubenoff
US Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA 02111, USA.
Dual-energy X-ray absorptiometry (DXA) scanners have become the standard instruments for measuring whole and regional body composition. In multicenter clinical trials Lunar and Hologic instruments are often intermixed. However, little is known about the differences between these instruments and the potential errors that such a practice may cause. Therefore, we assessed the degree of agreement between Hologic QDR-2000 (fanbeam) and Lunar DPX-L (pencil-beam) absorptiometers. Seventy-nine healthy volunteers, 41 women and 38 men, underwent two DXA scans on the same day to compare lean, fat, and bone mineral density in the whole body, appendages, and trunk. There were significant differences between the scanners for women and men in whole-body lean, fat, bone, and percentage body fat; leg lean, fat, and bone, and arm lean and fat (all P≤0.0001). Women also showed significant differences in estimates of trunk fat (P=0.006) and arm bone (P≤0.0001). These differences increased with increasing body mass, but not with age. We concluded that Lunar and Hologic densitometers give different estimates of regional and whole-body soft-tissue composition, which are biased by increasing mass. Multicenter trials that combine Hologic and Lunar data should address systematic differences between these instruments.
International Journal of Body Composition Research 2003, Vol. 1 No. 1: 23-29
Paul Deurenberg1 and Mabel Deurenberg-Yap2
1Nutrition Consultant in Singapore; 2Director, Research and Health Information Management Division, Health
Promotion Board, Singapore, adjunct Associate Professor, National University of Singapore.
Body composition was measured using a chemical four-compartment model in 291 Singaporean Chinese, 2 Malays and Indians of both sexes, and body fat percentage (%BF) obtained via this model was used as a reference. In addition biceps, triceps, subscapular and suprailiac skinfolds were measured following the Durnin and Womesley protocol and %BF was predicted using age and sex specific prediction formulas from the sum of biceps and triceps and from the sum of all four skinfolds. In Singapore females, especially Malays and Indians, predicted mean %BF from two skinfolds was an underestimation. Mean %BF predicted from four skinfolds was also underestimated in Malays and Indian females, but not in Chinese females. The differences in validity from predictions based on two or four skinfolds could be explained by differences in subcutaneous fat pattern, with the Singaporean females having a more truncal fat pattern than the Scottish population in which the formulas had been developed. In males, predicted mean %BF from two skinfolds was underestimated only in Indians. Mean %BF from four skinfolds did not differ from the reference value in Chinese, Malay and Indian males. The bias of predicted %BF was positively correlated with level of body fatness and negatively with age in both gender groups, resulting in considerable underestimations of %BF in fatter and younger subjects. Differences in validity of predicted %BF across the ethnic groups could be explained by differences in body fatness and age across the groups. It is concluded that the Durnin and Womersley equations are not valid in Singaporeans because of a different body fat distribution (in females) and because of a different age-related increase in body fatness (in males and females), compared to the population in which the formulas were developed.
International Journal of Body Composition Research 2003, Vol. 1 No. 1: 31-35
Paul Deurenberg1 and Frans JM Schouten2
1Nutrition Consultant, Singapore; 2Department of Nutrition and Epidemiology, Wageningen University, The Netherlands
Body composition was measured in 59 normal weight, healthy females, aged 20 to 24 years. Their BMI ranged from 17.4 to 29.4 kg/m2 and their %BF as measured by densitometry (underwater weighing) from 16.6 to 35.6 per cent. Body fat was also estimated from bioelectrical impedance using the traditional foot-to-hand approach (total body impedance), foot-to-foot approach (Tanita) and hand-to-hand approach (Omron). From total body impedance %BF was predicted using a published prediction equation including weight and age as additional predictors. For the foot-to-foot and hand-to-hand impedance the formulas incorporated in the instruments were also used. These formulas use weight and age as additional predictors. The correlation coefficients of estimated %BF with measured %BF ranged from 0.70 (foot-to-hand) to 0.75 (hand-to-hand), all values P<0.01. The correlation of estimated %BF from foot-to-foot with BMI was higher (r=0.93, P<0.01) than with actual %BF (r=0.71, P<0.01), which suggests that the formula used overemphasizes weight in the prediction formula. The mean bias (measured minus predicted) of %BF was –0.3 ± 3.4, –0.9 ± 3.6 and 1.0 ± 3.2 percentage points body fat for foot-to-hand, foot-to-foot and hand-to-hand respectively. All biases were positively correlated with the level of %BF, resulting in underestimations of %BF in the fatter subjects. The number of subjects with a bias greater than 5 percentage points body fat was 8, 12 and 8 for foot-to-hand, foot-to-foot and hand-to-hand impedance respectively. The current data do not show a clearly better validity of predicted %BF for the total body impedance approach compared to the foot-to-foot or hand-to-hand approach.
International Journal of Body Composition Research 2003, Vol. 1 No. 1: 37-43
SJ Lan1, ES Engelson2, D Agin2 , P. Homel3, J Wang4, SB Heymsfield4 and DP Kotler2
1Department of Nutrition and Health Sciences, Taipei Medical University, Taiwan; 2Gastrointestinal Division, St Luke’s-Roosevelt Hospital Center, New York, NY, USA; 3Office of Research Support, Beth Israel Medical Center, New York, NY, USA; 4Weight Control Unit, Body Composition Unit, Department of Medicine, St Luke’s-Roosevelt Hospital Center, Columbia University, College of Physicians and Surgeons, New York, NY, USA.
Fat redistribution has been described in HIV-infected people. Magnetic resonance imaging (MRI) provides a criterion measure of total and regional adipose tissue. Dual-energy X-ray absorptiometry (DXA) also can quantify total and regional fat contents. We evaluated the ability of DXA to estimate changes in regional body fat, as determined by MRI scanning, by analyzing data from two longitudinal trials: 25 HIV-infected subjects given growth hormone for fat redistribution with excess visceral adipose tissue (VAT), and 30 malnourished, HIVinfected women treated with exercise training and/or protein supplements. Changes in subcutaneous adipose tissue (SAT) and VAT measured by MRI and regional fat measured by DXA were compared. The limits of agreement between values from the criterion and surrogate (predicted) measures were analyzed by the methods of Bland and Altman. The results showed that the change in limb fat measured by DXA reflected change in SAT by MRI (R2=0.70, P<0.001) while change in trunk fat measured by DXA reflected change in VAT by MRI (R2=0.72, P<0.001). The estimates of the slope and intercept values of the prediction equations were confirmed by bootstrap resampling. Bland-Altman analyses showed no systematic errors over the range of changes in SAT or VAT. The limits of agreements for the prediction of change were +3.2 l for SAT and +1.5 l for VAT, respectively. The precision of the limits of agreements were (–0.4; +0.4) and (–0.3; +0.1) l for SAT and VAT, respectively. The standard errors of the estimate were 1.8±1.9 l for SAT and 1.0±1.2 l for VAT. We conclude that DXA measurements of regional body fat reflect changes in fat compartments as measured by MRI. The prediction models are appropriate for epidemiological studies, but should be applied with caution in clinical situations.
International Journal of Body Composition Research 2003, Vol. 1 No. 1: 45-50
PSW Davies1 and TJ Cole2
1Children’s Nutrition Research Centre, Department of Paediatrics and Child Health, University of Queensland, Royal Children’s Hospital, Brisbane, QLD, Australia; 2Centre for Paediatric Epidemiology and Biostatistics, 30 Guilford Street, London, UK.
The appropriate adjustment of data relating to energy expenditure for body weight and body composition is critical for the correct interpretation of those data. This paper discusses the historical development of methods that have been used to try and adjust energy metabolism data for body weight and body composition. We have also shown that simply expressing energy expenditure relative to body weight, or fat-free mass as per kilogram, is not the best adjustment in many cases and that a power function of body weight or body composition should be used. Details of how such a power function can be calculated and interpreted are given. Finally, we have described how potential differences in energy expenditure between groups can be appropriately examined, and attempted to explain why simple expression of energy expenditure relative to body composition does not adjust appropriately.