Butte, Nancy; Heinz, Carolyn; Hopkinson, Judy; Wong, William; Shypailo, Roman; Ellis, Kenneth
Accurate assessment of body composition in infants and children is fundamental to understanding the growth process and factors that alter it, such as undernutrition, overnutrition, or chronic disease. Because of theoretical problems underlying existing methods or practical difficulties in implementing techniques in infants and young children, development and validation of methods applicable to pediatric populations are needed.
Sophisticated in vivo body composition methods have been thoroughly evaluated and compared in adults and children (1-5), but not in infants and toddlers. Assumptions underlying two-compartment models that partition the body into fat-free mass (FFM) and fat mass (FM) are influenced by age and maturation. Because of divergent principles underlying existing technology, there is a high likelihood that body composition methods applicable to young children systematically differ, as in adults. In the absence of a gold standard, validation and interchangeability are often determined by comparing methods. Interpretation of findings across studies using different methods requires knowledge of the extent that body composition methods differ. This is especially important in understanding normal growth and development and nutritional, environmental, and genetic factors that affect these processes.
The objective of this study was to compare methods using total body water (TBW), total body potassium (TBK), total body electrical conductivity (TOBEC), and dual-energy x-ray absorptiometry (DXA) for the estimation of body fat mass in infants and toddlers. Gender- and age-related differences between methods also were explored.
MATERIALS AND METHODS
Study Design and Subjects
Repeated body composition measurements were performed on 76 healthy term infants at 0.5, 3, 6, 9, 12, 18, and 24 months of age at the Children's Nutrition Research Center (CNRC). Seventy-two children completed the 24-month study. By study design, the infants were either exclusively breast fed (n = 40) or formula fed (n = 36) from birth to 4 months of age; thereafter, feeding preference was at the discretion of the parents. This study was approved by the Baylor Affiliates Review Boards for Human Subject Research, and informed, written consent was obtained from each child's mother.
Infants were admitted to the CNRC Metabolic Research Unit from approximately 10:00 A.M. to 5:00 P.M. for the series of anthropometric and body composition measurements. First, anthropometric measurements were performed at least 1 hour after feeding, followed by the TOBEC measurement. The infant was dressed, and cotton balls were placed into the diaper to collect the baseline urine sample before the 2H2O dose, which was administered orally by syringe at least 30 minutes after eating to avoid regurgitation of the dose. The 15-minute whole body counting of 40K and the DXA measurement were usually performed on the younger infants while they slept. By 18 to 24 months, measurements could be obtained in infants in the awake state, while the children were entertained with a video inside the counter or near the DXA equipment.
Infants were weighed naked on an electronic integrating scale (Sartorius, Göttingen, Germany). Crown-to-heel length was measured on a recumbent infant board (Holtain Ltd., Crymych, UK).
Total Body Water
Total body water was determined by dilution of an orally administered dose of 99.8 atom percent deuterium oxide (50 or 100 mg of 2H2O/kg body weight). Urine samples were collected before the dose was administered, and 3 to 5 hours after or daily for 10 days. (The higher dose and longer sampling period were used as part of the doubly-labeled water method for estimation of total energy expenditure, to be reported elsewhere.) Before analysis, hydrogen gas was generated from undistilled urine samples by zinc reduction in quartz vessels (6,7). Abundance of 2H in the urine samples was measured by gas-isotope ratio mass spectrometry (Delta-E, Finnigan MAT, San Jose, CA, U.S.A.). Deuterium dilution space was calculated from the average of two postdose urine samples by the plateau method at 0.5 months of age, and from 10 daily urine samples by extrapolation at all other ages. Deuterium dilution space was converted to TBW by dividing by 1.04, and TBW was converted to FFM by dividing by 0.8055, 0.800, 0.796, 0.793, 0.790, 0.785, and 0.781 for boys, and 0.8055, 0.799, 0.794, 0.790, 0.788, 0.784, and 0.782 for girls at 0.5, 3, 6, 9, 12, 18, and 24 months of age, respectively (8).
Total Body Electrical Conductivity
Total body electrical conductivity was used to measure FFM and FM (model HP-2, EM-SCAN, Inc., Springfield, IL, U.S.A.) (9). The child was undressed and swaddled in a sheet to ensure that the arms and legs were fully extended and remained parallel to the main axis of the body and to restrict movement. The child was then placed supine on the TOBEC instrument sled, which was inserted into the TOBEC tube, and the maximum (peak) TOBEC number was recorded. Fat mass of the child was calculated as follows: FM (kg) = WT - [0.0265 √ (TOBEC#xLc) - 0.0313] where TOBEC# is the mean of 5 to 10 repeat measurements corrected for drift in the phantom reading, and Lc is the conductive length in centimeters of the subject calculated as the crown-to-heel length minus the head diameter. The accuracy of the instrument was within 0.3% of the reference value for the phantom, and the precision was less than 1%, which equates to an error of 0.02 kg FM at 0.5 months and 0.04 kg FM at 24 months of age (9).
Total Body Potassium
Total body potassium was estimated from the 40K naturally present in the child's body using the CNRC whole body counter (10). 40K emits high-energy gamma rays at the constant rate of 200.4 photons/min per gram potassium. Photons are detected by 12 photon-sensitive NaI(T1) detectors arranged in two arrays above and below the child's body in the low-background whole-body counter. For the 15-minute counting, the younger children were swaddled and placed in a plastic bassinet. The older children were secured on a hammock with straps. A custom set of four phantoms is used for routine quality control of the instrument. The in vivo precision for TBK measurements of infants and toddlers is less than 2.5%, which equates to an error of 0.04 kg FM at 0.5 months and 0.27 kg FM at 24 months of age (10).
To convert TBK to FFM in infants 12 months of age or less, equations were used based on data published by Fomon et al. (8): Boys: TBK/FFM (g/kg) = -0.8557×10-4 × age(mo)2 + 0.9928 × 10-2 × age(mo) + 1.92; Girls: TBK/FFM (g/kg) = -0.1002×10-3 × age(mo)2 + 0.011 × age(mo) + 1.92
For 18-month-old children, the conversion factors of 2.27 and 2.30 g/kg were used for boys and girls, respectively. For 24-month-old children, the conversion factors of 2.32 and 2.33 g/kg were used for boys and girls, respectively.
Dual-Energy X-ray Absorptiometry
A DXA scan was used to estimate bone mineral content, fat, and lean mass at 0.5, 12, and 24 months only. The whole body was scanned in the single-beam mode with an aluminum table pad designed to improve instrument linearity and reduce radiation dose (QDR-2000, Infant Whole Body, analysis version 5.56-5.71P; Hologic, Inc., Waltham, MA, U.S.A.). For the DXA scan the child, lightly dressed and diapered, was placed supine on the pad with an interposing paper sheet. No sedation was used. A step phantom was placed beside the child for calibration. The precision of the DXA was determined for bone mineral (1.2-4.1%), fat (3.0-4.2%), and lean (1.0-1.5%) compartments with repeated measurements on pigs, weighing 4.6 to 15.7 kg (11). In terms of FM, an average precision of 3.6% equates to an error of 0.02 kg FM at 0.5 months and 0.11 kg at 24 months of age.
Values are reported as means ± standard deviation (SD). Commercially available software (Minitab, ver. 12; Minitab Inc., University Park, PA, and BMDP ver. 7; BMDP Statistics Software, Inc., Los Angeles, CA, U.S.A.) was used for data description and statistical analysis. The FM values estimated by four body composition methods at seven time points (0.5, 3, 6, 9, 12, 18, and 24 months postpartum), were compared using repeated measures analysis of variance. (ANOVA; BMDP5V, 2V). A significant interaction between method and time was encountered; therefore, the data were subdivided by time points and reanalyzed using repeated measures ANOVA, followed by Bonferroni multiple comparisons at 5%. The intraclass correlation, which expresses the between-subject variance as a percentage of the total variance (12), provides a measure of relative agreement between methods. The value ranges from 0 to 1; the greater the value the greater the agreement. Mean differences ± SD for all pairwise comparisons are presented, to calculate Bland-Altman limits of agreement (mean ± 2 SD) (13).
Mean ± SD birth weight and length of the 76 infants were 3.42 ± 0.44 kg and 50.56 ± 2.24 cm, respectively. Mean gestational age was 39.1 ± 1.3 weeks. Weight, length, TBW, and TBK at each age are displayed in Table 1. Although all parameters tended to be higher in boys than in girls, TBK and TBW attained statistical significance (p ≤ 0.05).
Fat mass values estimated from TBW, TBK, TOBEC, and DXA measurements, are summarized in Table 2. The rank order from highest to lowest estimate of FM changed with age. At 0.5 months of age TBW produced the lowest estimate of FM, and DXA produced the highest. At 24 months, TBK provided the lowest estimate, and TBW provided the highest. Although all methods were positively correlated with one another (p = 0.001-0.05), the degree of concordance varied with age and among methods (Table 3).
Repeated ANOVA indicated that the differences among methods depended on time. There was no statistical evidence that the differences among methods depended on gender or feeding mode. Because of the significant interaction between method and time, methodologic differences were examined separately at each age. Significant differences among methods were encountered at each time interval: 0.5, 3, and 6 months (p = 0.001), 9 months (p = 0.05), 12 months (p = 0.001), 18 months (p = 0.004), and 24 months (p = 0.001).
Mean differences ± 1 SD for all pairwise comparisons are presented in Table 4. At 0.5 months, results from all methods differed significantly from one another (p < 0.05). At 24 months, TBK differed from TBW, TOBEC, and DXA, and TOBEC differed from DXA (p < 0.05). The mean differences relative to the mean FM were greatest at 0.5 months of age (0.23-0.79), and decreased thereafter (0.00-0.23). The limits of agreement (mean difference ± 2 SD) indicate that substantial differences were seen among methods for individual estimates of FM.
These findings indicate that methods using TBW, TBK, TOBEC, and DXA for estimation of FM in infants and toddlers were not interchangeable. The wide limits of agreement also imply that the methods were not interchangeable for group or individual estimations. The rank order of the methods and the magnitude of the method differences were a function of age, which makes it difficult to correct for systematic biases. Method differences were not affected by gender or infant feeding mode.
The FM was estimated indirectly from TBW, which was measured by isotope dilution and converted to FFM. Errors in the implementation of the technique include analytical precision, isotope equilibration within the body, corrections for exchange of label with nonaqueous hydrogen or oxygen, and estimation of the hydration of FFM. The precision of the mass spectrometric determinations of 2H is between 1% and 2% (14). Theoretical and experimental evidence indicates that oxygen exchange is slightly more than 1% and that deuterium exchange is slightly less than 4% of TBW (14). Total body water may be calculated by the plateau or back-extrapolation approach. Because of the rapidly changing state of the infant at 0.5 months, we chose to use the plateau approach; thereafter, we used back extrapolation.
In theory, the plateau approach overestimates TBW, because only urinary loss of label is accounted for in the equilibration phase, and the back-extrapolation approach underestimates TBW because mixing of the label is not instantaneous. In practice, the two approaches yield similar results in adults (14) and infants (15); measurement errors probably override these theoretical considerations. Conversion of TBW to FFM is particularly problematic in the rapidly chemically maturing infant. We applied hydration factors published by Fomon et al. (8) that were derived from 2H dilution measurements. The TBW of children from birth to 10 years of age were obtained by interpolation-extrapolation from TBW data available on 44 boys and 42 girls approximately 6 months of age, and 11 boys approximately 9 years of age (8). A value of 1.3% was used to adjust for the exchange of deuterium with the exchangeable hydrogen of organic molecules. Had Fomon et al. used a value of 4.0% to adjust for deuterium exchange, the hydration constants would be 0.5% lower, which would result in a 5% to 6% difference in FM.
Measurement of TOBEC has been shown to be an accurate and precise method for the estimation of FFM and FM (9,16,17). It has the advantage that it is rapid, safe, noninvasive, and independent of the hydration of FFM. The instrument was calibrated against carcass analysis of minipigs weighing 1.0 to 10 kg. The calibration equation (9) is recommended for infants weighing between 2.8 and 10.0 kg. Although there is no indication of nonlinearity to preclude extrapolation of the equation beyond 10 kg, there are no validation data; therefore, the authors formally recommended an upper limit of 10 kg. In our study, some toddlers were physically too large for the instrument, and their weights exceeded the calibration range. De Bruin et al. (17) compared TOBEC and TBW estimates of FFM and FM in 50 infants aged 1 to 12 months. Total body electricity conductivity resulted in slightly, but significantly higher FFM values in boys (1.5%) and girls (4%). In our study, mean differences in FM derived from TOBEC and TBW were not statistically significant in infants more than 3 months of age.
Errors involved in the estimation of FM from TBK most likely stem from the conversion of TBK to FFM, not from the actual counting of 40K. The high level of precision of the CNRC whole-body counter for small children was achieved by improved NaI(T1) detectors, reducing the number of detectors in the counting array, bringing the detectors closer to the body, and increasing the counting time (10). We used conversion factors published by Fomon et al., (8) based on interpolations of TBK data of 26 boys and 14 girls approximately 6 months of age from the laboratory of Romahn and Burmeister (18) and of 64 boys 8 to 10 years of age from the laboratory of Forbes (19). At 6 months of age, the mean values of TBK were 40.4 and 40.3 mEq/kg body weight compared with our values of 37.9 and 36.0 mEq/kg body weight for boys and girls, respectively. In terms of length, the mean values of TBK were 4.54 mEq/cm in the Romahn and Burmeister database and 4.27 mEq/cm in ours. Mean weight and length of the infants studied by Romahn and Burmeister were similar to our infants at 6 months (7.74 kg, 68.6 cm). It is unlikely that our lower values are caused by higher FM, because the ratio of TBK to length was also lower. Romahn and Burmeister used liquid scintillation detectors, whereas we used NaI(T1) detectors. Liquid scintillation detectors do not completely resolve the photopeaks of 137 Cs (0.66 megaelectron volts [MeV]) and 40K (1.46 MeV) (20). At the time of the study conducted by Romahn and Burmeister, 137Cs from nuclear fallout could have contributed to higher TBK values in their infants. The conversion factors adopted by Fomon et al. therefore would underestimate FFM and overestimate FM.
DXA is a precise, safe method involving minimal radiation exposure on the order of 3µSv (0.3 mrem). However, it appears to overestimate FM in early infancy. The accuracy of DXA for body sizes comparable to infants has been evaluated using piglets, and results differed according to the specific software and instrument (21). Brunton et al. (22) evaluated the Hologic QDR-1000 instrument using infant whole-body software (version 5.56) in small (1.58 kg) and large (5.89 kg) piglets. Fat mass obtained by DXA was overestimated by 119% in the small piglets and 29% in the large piglets compared with carcass analysis. Picaud et al. (23) also found that FM was overestimated by 22% in piglets with a true FM greater than 250 g, using the Hologic QDR-2000 instrument (v. 5.64). Application of a correction equation reduced the difference between carcass and DXA results to 1.5 ± 8.8%. The DXA results were robust in response to a change in the hydration factor of FFM when it varied from 83.3% to 78%; the increase in percent FM was only 0.3%. Rigo et al. (24) also have derived equations to correct for the overestimation of fat mass based on results from eight piglets (1.40-5.15 kg) using the Hologic QDR 2000 instrument (v. 5.64P). The errors of the corrected FM decreased as fat content increased: ±30% for a fat content of 100 g and ±12% for a fat content of 500 g. There is some question whether validation and calibrations based on piglets, which have a low FM percent (<6%), are applicable to human infants who normally have a higher proportion of body weight as fat. Our data confirm a systematic positive bias in DXA-derived estimates of FM relative to estimates obtained using TBW, TBK, and TOBEC measurements at 0.5 months and 12 months of age for infants.
All body composition methods applied in pediatrics are indirect, and none can be considered a gold standard. In the absence of a gold standard, validation is often determined by comparing methods. Our analysis showed that FM determinations based on TBW, TBK, TOBEC, and DXA in infants and toddlers do not fully agree. The rank order of the methods and the magnitude of the method differences varied with age. Comparative studies in adults generally display systematic differences between body composition methods. The rank order from highest to lowest for estimates of percent FM was TBK, DXA, TBW, and TOBEC in two comprehensive adult studies (4,5). In contrast, percent FM was lower by DXA than by TBW in another adult study (4). In part, the discrepancies among methods when applied to infants and toddlers may be caused by the limitations associated with two-compartment models (25). Throughout infancy, childhood, and adolescence, there is a gradual change in the chemical composition of FFM. The use of age- and sex-specific conversion factors for the composition of FFM is a viable alternative, but its validity rests on the quality and completeness of the original data from which the constants are derived. Multicomponent models may further improve the accuracy of body composition estimates in infants and toddlers.
In conclusion, methods using TBW, TBK, TOBEC, and DXA for the estimation of FM in infants and toddlers are not completely interchangeable. When comparing results among studies, investigators should be cognizant of the different body composition methods that are used. We have demonstrated that the choice of method can influence both cross-sectional and longitudinal data. Further development and validation of body composition techniques are still needed for use in infants and toddlers.
Acknowledgment: The authors thank the women who participated in this study and acknowledge the contributions of Marilyn Navarrete for subject recruitment; Sopar Seributra and Sandra Kattner for nursing and dietary support; Maurice Puyau, Firoz Vohra, Judy Joo Posada, JoAnn Pratt, Nitesh Mehta, Zahira Colon, Kiyoko Usuki, Shide Zhang, and Deborah Roose for technical assistance; Anne Adolph for data management; Leslie Loddeke for editorial review; and Idelle Tapper for secretarial assistance.
This work is a publication of the U.S. Department of Agriculture (USDA)/Agricultural Research Service (ARS) Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children's Hospital, Houston, Texas. This project was funded in part with federal funds from the USDA/ARS under Cooperative Agreement 58-6250-6001. The contents of this article do not necessarily reflect the views or policies of the UDSA, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
1. Goran MI, Driscoll P, Johnson R, Nagy TR, Hunter G. Cross-calibration of body-composition techniques against dual-energy x-ray absorptiometry in young children. Am J Clin Nutr
2. Gutin B, Litaker M, Islam S, Manos T, Smith C, Treiber F. Body-composition measurement in 9-11-y-old children by dual-energy x-ray absorptiometry, skinfold-thickness measurements, and bio-impedance analysis. Am J Clin Nutr
3. Ellis K. Measuring body fatness in children and young adults: Comparison of bioelectric impedance analysis, total body electrical conductivity, and dual-energy x-ray absorptiometry. Int J Obes
4. Fuller NJ, Jebb SA, Laskey MA, Coward WA, Elia M. Four-component model for the assessment of body composition in humans: Comparison with alternative methods, and evaluation of the density and hydration of fat-free mass. Clin Sci
5. Pierson RN Jr, Wang J, Heymsfield SB, et al. Measuring body fat: Calibrating the rulers. Intermethod comparisons in 389 normal Caucasian subjects. Am J Physiol
6. Wong WW, Lee LS, Klein PD. Deuterium and oxygen-18 measurements on microliter samples of urine, plasma, saliva, and human milk. Am J Clin Nutr
7. Wong WW, Cochran WJ, Klish WJ, Smith EO, Lee LS, Klein PD. In vivo isotope-fractionation factors and the measurement of deuterium- and oxygen-18-dilution spaces from plasma, urine, saliva, respiratory water vapor, and carbon dioxide. Am J Clin Nutr
8. Fomon SJ, Haschke F, Ziegler EE, Nelson SE. Body composition of reference children from birth to age 10 years. Am J Clin Nutr
9. Fiorotto ML, de Bruin NC, Brans YW, Degenhart HJ, Visser HKA. Total body electrical conductivity measurements: An evaluation of current instrumentation for infants. Pediatr Res
10. Ellis KJ, Shypailo RJ. Whole-body potassium measurements independent of body size. Human Body Composition: In Vivo Methods, Models, and Assessment
. New York: Plenum Press, 1993;371-5.
11. Ellis KJ, Shypailo RJ, Pratt JA, Pond WG. Accuracy of dual energy x-ray absorptiometry for body composition measurements in children. Am J Clin Nutr
12. Snedecor GW, Cochran WG, eds. Statistical methods
. 6th ed. Ames, IA: Iowa State University Press, 1967.
13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
14. Schoeller DA. Hydrometry. In: Roche AF, Heymsfield SB, Lohman TG, eds. Human body composition
. Champaign, IL: Human Kinetics, 1996;25-44.
15. Salazar G, Infante C, Vio F. Deuterium equilibration time in infant's body water. Eur J Clin Nutr
16. de Bruin NC, van Velthoven KAM, Stijnen T, Juttmann RE, Degenhart HJ. Quantitative assessment of infant body fat by anthropometry and total-body electrical conductivity. Am J Clin Nutr
17. de Bruin NC, Westerterp KR, Degenhart HJ, Visser HKA. Measurement of fat-free mass in infants. Pediatr Res
18. Romahn A, Burmeister W. Die Körperzusammensetzung während der ersten zwei Lebensjahre Bestimmungen mit der Kalium-40-Methode. Klin Padiatr
19. Forbes GB. Growth of the lean body mass in man. Growth
20. Forbes GB, ed. Human body composition. Growth, aging, nutrition, and activity
. New York: Springer-Verlag, 1987.
21. Lapillonne AA, Braillon PM, Delmas PD, Salle BL. Dual-energy x-ray absorptiometry in early life. Horm Res
22. Brunton JA, Weiler HA, Atkinson SA. Improvement in the accuracy of dual energy x-ray absorptiometry for whole body and regional analysis of body composition: Validation using piglets and methodologic considerations in infants. Pediatr Res
23. Picaud J-C, Rigo J, Nyamugabo K, Milet J, Senterre J. Evaluation of dual-energy x-ray absorptiometry for body-composition assessment in piglets and term human neonates. Am J Clin Nutr
24. Rigo J, Nyamugabo K, Picaud J-C, Gerard P, Pieltain C, De Curtis M. Reference values of body composition obtained by dual energy x-ray absorptiometry in preterm and term neonates. J Pediatr Gastroenterol Nutr
25. Wang ZM, Heshka S, Pierson RN, Heymsfield SB. Systematic organization of body-composition methodology: An overview with emphasis on component-based methods. Am J Clin Nutr
© 1999 Lippincott Williams & Wilkins, Inc.