Assessment of body composition in infants and children

RESEARCH TECHNIQUES APPLICABLE TO PEDIATRICS

Nutrition Vol. 14, No. 10, 1998

Assessment of Body Composition in Infants and Children JOHN J. REILLY, PHD From the Department of Human Nutrition, University of Glasgow, Yorkhill Hospitals, Glasgow, Scotland, UK ABSTRACT

The aims of this review are to consider 1) the applications of pediatric body composition methodology; 2) criteria for determining choice of method; and 3) some future developments. The major applications are: development and validation of new methods; assessment of growth or nutritional status; public health applications, such as monitoring the increasing prevalence of pediatric obesity; interpretation of data on energy expenditure; and testing the functional significance of variation in body composition. An appreciation of the underlying theoretical models (two-component and multicomponent models) is essential to an understanding of the methodology, as is an appreciation of the fact that infants and children are not “chemically mature.” The two-component model generates methods that, though limited by variation in the composition of fat-free mass, have accuracy that is acceptable so long as the method in question is chosen with care. Criteria for determining choice of method are provided. Multicomponent models have a more rigorous theoretical basis, but require access to techniques that are not universally available and not always practical for pediatric use. Bedside methods, notably bioelectrical impedance and skinfold thickness, can provide acceptable accuracy, but the precision of all methods limits their ability to measure changes in body composition. Nutrition 1998;14:821– 825. ©Elsevier Science Inc. 1998 Key words: fat-free mass, body fat percentage, nutritional assessment, total body water, bioelectrical impedance, skinfolds, total body electrical conductivity, dual energy x-ray absorptiometry

INTRODUCTION

There is increasing awareness of, and interest in, measurement of body composition (body fatness) in pediatrics. This is the result of growing recognition of the importance of body composition, and the development of methods for measuring body composition that are suitable for use in infants and children. The aims of this review are to outline the applications of body composition data in pediatrics, to consider the major issues that the critical user of pediatric body composition methodology must be aware of, and to briefly describe future developments in pediatric body composition methodology. A detailed description of methodology is beyond the scope of this review, and the literature contains a number of excellent reviews with particular reference to pediatrics,1,2 reviews of the basis of body composition methodology,3,4 and reviews that concentrate on practical aspects.5,6 APPLICATIONS OF PEDIATRIC BODY COMPOSITION METHODOLOGY

There are five principal applications of body composition methodology in pediatrics. Some of these are widely known, others much less well known but potentially important.

Interpretation of Data on Energy Expenditure Body composition (fatness and fat-free mass [FFM]), and in particular the size and composition of the FFM, has a profound influence on energy expenditure. Consideration of differences in energy expenditure between groups (between obese versus nonobese children for example) is common7 and must take account of any differences in body composition between such groups. This is particularly important when the groups differ markedly in body size or composition, and so measurement of body composition underpins the measurement of energy balance. A detailed discussion of how to normalize energy expenditure data for body composition is beyond the scope of this review, but it should be noted that simply dividing energy expenditure by FFM, still a common practice, does not normalize energy expenditure for FFM.8 Public Health Applications Application of body composition methodology to public health is not common, but is of considerable importance with the increasing prevalence of obesity and the need to measure the efficacy of interventions intended to ameliorate the problem.9 Measuring obesity is best achieved by measurement of whole body fatness, rather than proxy measures of fatness, since inferences from limited anthropometric data (weight adjusted for height or

Correspondence to: John J. Reilly, PhD, Department of Human Nutrition, University of Glasgow, Yorkhill Hospitals, Glasgow G3 8SJ, Scotland, UK. E-mail: [email protected]

Nutrition 14:821– 825, 1998 ©Elsevier Science Inc. 1998 Printed in the USA. All rights reserved.

0899-9007/98/$19.00 PII S0899-9007(98)00093-8

822 single skinfolds) are difficult to make with confidence.1,2 A detailed discussion of these points is given by Kuczmarski.10 Functional Significance of Body Composition Variation in body composition within a population can have direct functional consequences. This application has received little attention, but an example of the potential functional importance of body composition is the evidence that for some drugs a good deal of the interpatient variability in pharmacokinetics, and hence response to therapy, might be explained by variation in body composition.11 Alterations in body composition, secondary to negative energy balance, can also cause impairment of immune function and muscle function. Methodological Research and Development For most investigators in the field of body composition, the primary focus is methodological. This is particularly true in pediatrics because there is a need for development and validation of new methods of assessment, either because of the underlying theoretical problems with existing methods (discussed in a proceeding section) or the fact that most of the traditional methods were designed for use in adults and are unsuitable for infants and young children. Assessment of Growth and Nutritional Status Classical whole body or local measures of growth and nutritional status, such as weight, length, or weight adjusted for height (as a percentage of ideal weight, body mass index, ponderal index), triceps skinfold, and mid arm circumference, are useful but rather crude and limited. In many fields of pediatric practice and research, more detailed assessments are necessary, and these can be made by measuring body composition (total body fat or FFM) rather than crude proxies for it.12 The pediatric literature contains a number of examples of this point. These include examples of inappropriate referral for obesity as a result of limitations of the classical measures,13,14 and demonstrations that the classical measures are poorly related to fatness.15 In intervention studies, body composition is a more sensitive marker of outcome than the classical measures,16 –19 and in observational studies of the natural history of chronic disease20 body composition measurements provide a more accurate description of changing nutritional status than measurements of the traditional indices. One important development is the suggestion that overnutrition in children can and should be defined on a biological basis (by assessment of body fatness and identification of its relationship to clinical outcomes) rather than by using proxy measurements for fatness.1,21 Two studies have identified cut-offs for body fatness (as a percentage of body weight) that are associated with increased cardiovascular risk factors in children.1,22 Both have suggested cut-offs of around 27–30% fat in girls and 18 –25% fat in boys. Undernutrition in children might also be amenable to biological definition based on body fatness, and a body fatness of , 5% (boys) and , 12% (girls) might define minimal body weight,1 but in general, defining undernutrition will be more complex than defining overnutrition. One final application is in the identification of abnormal changes in body composition. When large losses or gains of body weight occur, both the fat mass (FM) and FFM are affected in a predictable way.23 In some clinical settings, abnormalities in the relative contribution of FM and FFM to weight gain or weight loss can be used as indicators of underlying pathology. For example, relatively large losses of FFM can be indicative of a cachectic process, whereas more normal composition of loss might indicate “simple starvation.”24

BODY COMPOSITION IN INFANTS AND CHILDREN METHODOLOGY FOR ASSESSMENT OF BODY COMPOSITION

Direct measurement of body composition (by chemical analysis) is not possible in humans, and so various models for indirect measurement of body composition have been developed. The inability to directly determine body composition means that there is no “gold standard” method for human studies, but measurement of body density has historically been regarded as a “reference method” against which other methods can be validated on the grounds that errors in the body density method are acceptable and quantifiable.1 Two-Component Model of Body Composition Most studies of body composition in children and adults have been based on a theoretical model that defines two body components: FM (weight of all extractable lipid using ether as the solvent), and FFM (weight of the rest of the body minus the FM). Methods of body composition arising from the two-component model rest on assumptions of constant composition of FM, and in particular, the FFM. In adults the assumptions that FFM is 72– 73% water, has 2.66 g/kg of potassium in men and 2.55 g/kg in women,3 and has a density of 1.100 g/cc permit measurement of FFM and FM by measurement of, respectively, total body water (TBW), total body potassium (TBK), and densitometry. Relative constancy of chemical composition of the FFM (chemical maturity) is not achieved until adulthood, and throughout infancy, childhood, and adolescence there is a gradual change in FFM composition to adult values. The FFM of children therefore has lower density, lower mineralization, higher water content, and lower potassium content.25–27 One consequence of this is that the constants used in adult body composition methodology should not be used in pediatrics,25–27 but the literature shows that this is not always appreciated by those measuring body composition of children. Use of adult values in children will substantially overestimate fatness of children when densitometry, predicted density (multiple skinfold thickness method), TBW, and TBK are used. Systematic errors of this kind decrease with increasing age, but are serious and should be avoided. The use of age and sex-specific constants for the composition of FFM is a valid alternative. These constants vary systematically with increasing age, and they can be obtained with reference to the original sources25–27 or reviews.2 This approach represents a modified two-component model. Critical appraisal of the underlying physiology of growth and development should lead to the conclusion that the concept of constants is unphysiological, because FFM composition varies within each age and sex category. However, the use of age and sex-specific constants approach removes systematic errors (biases) from pediatric body composition methods and has been shown to produce robust methods that can have acceptable accuracy relative to a reference method1,2,28 so long as adequate care is taken over choice of method and measurement technique. Criteria for choosing a method are discussed in a proceeding section. Multicomponent Models of Body Composition There is a consensus that errors arising from assumptions about the composition of FFM can be reduced by taking an alternative, more empirical, approach that involves measurement of several components of body composition.1,2,4 This is the basis of multicomponent models of body composition. These require measurement of at least two (three-component model) or more (four- or five-component models) constituents of the FFM. Detailed discussion of the advantages of the multicomponent approach in adults and children are in the literature1,2,4 and will not be repeated here, but the potential advantages in pediatrics are considerable.1,2 Because measuring several constituents of the FFM is fundamental to

BODY COMPOSITION IN INFANTS AND CHILDREN the multicomponent approach, there is a serious disadvantage in that it is technically more demanding, usually involving measurement of TBW (by isotope dilution), body density (usually by hydrodensitometry), and mineral content (dual energy x-ray absorptiometry [DEXA]). In many circumstances in pediatrics such measurements are impractical on the grounds of cost or acceptability to the patient. An additional practical problem is that the necessary technology cannot easily be used in the field or the clinic. Measuring body density by traditional methods is impossible in preschool children.28 Measurement of body density by plethysmography29 represents a potentially important advance here. The availability of DEXA and methods for measurement of stable isotopes in body fluids (for measurement of TBW) is now more widespread. Application of multicomponent models in pediatrics will therefore increase, but methods based on the twocomponent model and “bedside” methods will remain popular and will be the only option under certain circumstances. “Bedside” Methods of Body Composition Measurement “Bedside” methods, which are suitable for use “in the field” or clinic, have considerable practical advantages. Estimation of body composition from skinfold thickness and bioelectrical impedance is relatively inexpensive, generally acceptable to the child, and the equipment is portable. Standardized measurement technique and quality control is central to accuracy and precision in both methods.1,30 Both skinfold prediction equations31,32 and equations based on bioelectrical impedance28,33 have been validated against the modified two-component model and multicomponent models, and there is now a large number of “validated” prediction equations, or variants of each method, from which to choose. An important practical concern is that prediction equations tend to lack crossvalidity, i.e., they tend not to provide accurate estimates in populations other than that in which they were derived (Table I). Very different estimates of fatness can be obtained from the same basic measurement of bioelectrical impedance or skinfolds (Table I) when different prediction equations are used, and this is an inherent feature of body composition.41 One practical consequence is that prediction equations should be validated in each population or sample in which they are used in order to establish confidence in the accuracy of the estimates obtained. This might be particularly important in many disease states, where alterations in the composition of FFM can violate the underlying assumptions. The important and related issue of how to validate body composition methods is relevant here. Validation must consist of a comparison of one method with a reference method, and the reference should consist of densitometry (two-component model) or a multicomponent model.1,2 Validity requires that estimates of the new method in individuals are close to those obtained by the reference method. Agreement between the new method and reference should be evaluated using the appropriate statistical approach, i.e., a Bland-Alman analysis42 or analysis of the size of the prediction error.2 Body composition methods tend to be highly correlated with each other, but correlation is not evidence of agreement.42 The widely different estimates shown in Table I were all highly correlated with each other, and with the reference, but agreement between them was poor.28,34 Reports of high degrees of correlation between a new method and a reference method does not therefore constitute adequate evidence to establish a new method. It should also be appreciated that some degree of error in estimation relative to a reference method is inevitable, arising from technical errors and errors in the assumptions made in the method (including error arising from violation of assumptions inherent in the reference method itself). Errors of up to 3– 4 body fat percentage units are inevitable1–3 and can even be regarded as indicating relatively high accuracy.1,2 Criteria for judgement of

823 TABLE I. INFLUENCE OF CHOICE OF PREDICTION EQUATION ON IMPEDANCE OR SKINFOLD-DERIVED ESTIMATES OF FATNESS IN PREPUBERTAL CHILDREN*

Impedance study: Densitometry (reference method) Impedance predictions Cordain et al.35 Houtkooper et al.33 Schaefer et al.36 Deurenberg et al.37 Skinfold study: Densitometry (reference method) Skinfold predictions Durnin and Rahaman38 Slaughter et al.31 Johnston et al.39 Brook40 Deurenberg et al.32

Mean body fat (%)

SD

15.9

8.8

9.1 17.3 22.1 24.5

8.7 7.4 9.7 6.3

13.2

8.0

10.4 15.5 8.2 13.1 16.2

6.1 5.5 6.5 7.1 4.9

* Impedance data from 81 children mean age 9 y; skinfold data from 57 boys mean age 9 y; all children measured by densitometry, impedance and skinfolds; fatness estimated from several prediction equations derived from the same basic measurement. (From Reilly et al.,28,34 with permission.)

accuracy of methods are available in the literature.2 It is unfortunate that many investigators in the field are engaged in the pursuit of levels of accuracy beyond the limit of the underlying model. Pediatric Body Composition Methodology: A Guide to Choice of Method In pediatric practice and research the opportunity to validate methods or assess cross-validity is rarely available, and so some guidelines for choice of method are provided in Table II. If adequate answers to the questions listed in Table II are not available, body composition methods should be used with extreme caution. Both impedance and skinfolds can meet the criteria listed in Table II so long as care is taken over choice of hardware43,44 or prediction equation used,28,34 and measurement conditions.44 Table II also notes the importance of measurement precision for assessment of changes in body composition. In many circumstances in pediatrics, the change in body composition is of greater interest than absolute body composition. Imprecision is a major limitation of body composition methodology in this respect,45 though investigators attempting to measure changes in body composition do not always appreciate this. In summary, changes in body composition of physiologic significance, changes to energy balance for example, are undetectable in individuals because of the inherent imprecision of existing methodology.45 Some of the more precise methods might be suitable for detecting changes in body composition of groups, but this depends on the magnitude of the change occurring and the precision of the techniques in the laboratory concerned.46,47 An additional consideration is that the changes in body composition being measured can render the underlying assumptions invalid; for example, changes in skeletal muscle mass can significantly alter the potassium content of FFM.48

824

BODY COMPOSITION IN INFANTS AND CHILDREN TABLE II.

CRITERIA FOR CHOICE OF BODY COMPOSITION METHOD Practical issues Is the method acceptable to the child? Is the cost acceptable? Training needs: quality control/standardization of technique. Is the method safe? Validity Is the method valid? What are the underlying assumptions of the method? Are these assumptions valid in children being studied? (Are the assumptions appropriate for measuring body composition of children or adults?) Is there a published validation study? Was validity assessed appropriately in that study? What is the error in the method and does it meet criteria for acceptability?2 Is the method valid for individuals as well as groups? Is the particular variant of the method used valid in the population of children being studied (i.e., does it have cross-validity; see Table I)? Precision What is the precision of the method? Is this adequate for detection of changes in body composition? Adequate for individuals or groups?

FUTURE DEVELOPMENTS IN PEDIATRIC BODY COMPOSITION

Measuring the body composition of infants and young children is problematic because of limitations in the underlying two-com-

ponent model26 and its associated “constants” for the composition of FFM, and because of the practical difficulties associated with measurement. In the absence of a reference method, TBW is often used as a reference in infants and young children, although variation in hydration of FFM must be relatively high.49 Validity of impedance in young children seems adequate50 and the method is practical even in infancy, but the general caveats with bedside methods noted previously apply. Skinfold thickness may be less valid in infancy;12,15 this may be an inherent limitation or may reflect flaws in the existing prediction equations.51 The development of total body electrical conductivity (TOBEC) is promising,15 to the extent that it has been suggested as a reference method against which other methods can be validated, and standards for body composition of infants based on TOBEC have been proposed.52 TOBEC is not a bedside method, the hardware is relatively expensive, and care must be taken in comparing measurements with standards because estimates of body composition are so heavily method-dependent.28,34,53 There is considerable potential for DEXA,54,55 particularly if it is used as part of a multicomponent model rather than alone. The use of DEXA as a “gold standard” is unjustified.47,56,57 Measurement of skeletal muscle mass, as distinct from FFM, has applications in clinical nutrition and in exercise physiology, but the methodology for measuring muscle mass is limited.46 Measurement of the body cell mass or total body nitrogen, as distinct from FFM, though technically difficult, has advantages under certain circumstances.13,58 Because body fat distribution, independent of body fat percentage, is an important risk factor for disease, there are circumstances where its measurement is indicated, but again the methodology for doing so in children is not established.59,60

REFERENCES 1. Lohman TG. Advances in body composition assessment. Current issues in exercise. Monograph No. 3, 1993. Champaign, Illinois: Human Kinetics Publ, 1993 2. Houtkooper LB. Assessment of body composition in youths and relationship to sport. Int J Sport Nutr 1996;6:146 3. Siri I. Body composition from fluid spaces and density: analysis of methods. In: Brozek J, Herschel A, eds. Techniques for measuring body composition. Washington, DC: National Academy of Science, 1961:223 4. Elia M. Body composition analysis: an evaluation of two component models, multicomponent models, and bedside techniques. Clin Nutr 1992;11:114 5. Jebb SA, Elia M. Techniques for the measurement of body composition: a practical guide. Int J Obes 1993;17:611 6. Sutcliffe JF. A review of in vivo experimental methods to determine the composition of the human body. Phys Med Biol 1996;41:791 7. Reilly JJ, Blacklock CJ, Dale E, Donaldson M, Gibson BES. Resting metabolic rate and obesity in childhood leukaemia. Int J Obes 1996; 20:1130 8. Ravussin E, Bogardus C. Relationship of genetics, age, and physical fitness to daily energy expenditure and fuel utilisation. Am J Clin Nutr 1989;49(suppl 5):968 9. Gortmaker SL, Must A, Sobol AM, Peterson K, Colditz G, Dietz WH. Television viewing as a cause of increasing obesity among children in the USA, 1986 –1990. Arch Pediatr Adolesc Med 1996;150:356 10. Kuczmarski RJ. Trends in body composition for infants and children in the US. Crit Rev Food Sci Nutr 1993;33:375 11. Reilly JJ, Workman P. Is body composition an important variable in the pharmacokinetics of anticancer drugs? Cancer Chemother Pharmacol 1994;34:3 12. Davies PSW. Body composition assessment. Arch Dis Child 1994; 69:337 13. Lazarus R, Baur L, Webb K, Blyth F. Body mass index in screening for adiposity in children and adolescents. Am J Clin Nutr 1996;63:500

14. Dietz WH. Childhood obesity. In: Suskind RM, Lewinter Suskind L, eds. Textbook of paediatric nutrition. New York: Raven Press, 1993: 279 15. De Bruin NC, Velthoven KAM, Stifnen T, Juttmann RE, Degenhart HJ, Visser HA. Body fat and fat-free mass in infants. Am J Clin Nutr 1995;61:1195 16. Tsou Yau KI, Chang MH. Growth and body composition of preterm, small for gestational age infants. Early Hum Dev 1993;33:117 17. Krew MA, Kehl RJ, Thomas A, Catalano PM. Relation of amniotic fluid c-peptide levels to neonatal body composition. Obs Gynaecol 1994;84:96 18. Gilmour CH, Sertipal-Walerius JM, Jones JG, et al. Pulse dexamethasone does not impair growth and body composition of very low birthweight infants. J Am Coll Nutr 1995;14:455 19. Vaisman N, Zadik Z, Duchan R, Voet H, Lotan D, Drukker A. Changes in body composition of children with chronic renal failure during growth hormone treatment. Pediatr Nephrol 1994;8:201 20. Thomson MA, Quirk P, Swanson CE, et al. Nutritional growth retardation is associated with defective lung growth in CF. Nutrition 1995;11:350 21. Hoerr SL, Nelson RA, Lohman TG. Discrepancies among predictors of desirable weight for black and white obese adolescent girls. J Am Diet Assoc 1992;92:450 22. Dwyer T, Blizzard CL. Defining obesity in children by biological endpoint rather than population distribution. Int J Obes 1996;20:472 23. Forbes GB. Body composition: influence of nutrition, physical activity, growth and ageing. In: Shils ME, Young VR, eds. Modern nutrition in health and disease. 8th ed. Baltimore: Williams and Wilkins, 1998:533 24. Jebb SA. Energy metabolism in cancer and human immunodeficiency virus infection. Proc Nutr Soc 1997;56:763 25. Westrate JA, Deurenberg P. Body composition in children. Am J Clin Nutr 1989;50:88 26. Fomon SJ, Haschke F, Ziegler EE, Nelson SE. Body composition of

BODY COMPOSITION IN INFANTS AND CHILDREN

27. 28. 29. 30. 31. 32. 33. 34. 35. 36.

37. 38. 39.

40. 41.

reference children from birth to age 10 years. Am J Clin Nutr 1982; 35:1169 Lohman TG. Applicability of body composition techniques and constants for children. In: Pandolf KB, ed. Exercise and sports science reviews. New York: Macmillan, 1986:325 Reilly JJ, Wilson J, McColl JH, Carmichael M, Durnin JVGA. Ability of bioelectrical impedance to predict FFM in prepubertal children. Pediatr Res 1996;39:176 McRory MA, Gomez TD, Bernauer EM, Mole PA. Evaluation of a new air displacement plethysmograph for measuring human body composition. Med Sci Sports Exerc 1995;27:1686 Kushner RF, Gudivaka R, Schoeller DA. Clinical characteristics influencing bioelectrical impedance analysis measurements. Am J Clin Nutr 1996;64:s432 Slaughter MH, Lohman TG, Boileau RA, et al. Skinfold equations for estimation of body fatness in children and youth. Hum Biol 1988;60: 709 Deurenberg P, Pieters JJL, Hautvast JGAJ. The assessment of body fat percentage by skinfold thickness measurement in childhood and young adolescence. Br J Nutr 1990;63:293 Houtkooper LB, Going SB, Lohman TG, Roche AF, Van Loan M. Bioelectrical impedance estimation of fat free body mass in children and youth: a cross-validation study. J Appl Physiol 1992;72:366 Reilly JJ, Wilson J, Durnin JVGA. Determination of body composition from skinfold thickness: a validation study. Arch Dis Child 1995;73:305 Cordain L, Whicker RG, Johnson JE. Body composition determination in children using bioelectrical impedance. Growth Dev Ageing 1988;52:37 Schaefer F, Georgi M, Zieger A, Scharer K. Usefulness of bioelectrical impedance and skinfold measurements in predicting fat-free mass derived from total body potassium in children. Pediatr Res 1994;35: 617 Deurenberg P, Kasters CS, Smith H. Assessment of body composition by bioelectrical impedance in children and young adults is strongly age-dependent. Eur J Clin Nutr 1990;44:261 Durnin JVGA, Rahaman MM. Assessment of the amount of fat in the human body from measurements of skinfold thickness. Br J Nutr 1967;21:681 Johnston JL, Leong MS, Checkland EG, Zuberbahler PC, Conger PR, Quinney HA. Body fat assessed from body density and skinfold thickness in normal children and children with CF. Am J Clin Nutr 1988;48:1362 Brook CGD. Determination of body composition of children from skinfold measurements. Arch Dis Child 1971;46:182 Reilly JJ, Wilson J, Durnin JVGA. Measuring the body composition of elderly subjects: a comparison of methods. Br J Nutr 1994;72:33

825 42. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;I:307 43. Smye SW, Sutcliffe J, Pitt E. A comparison of four commercial systems used to measure whole body electrical impedance. Physiol Meas 1993;14:473 44. National Institutes of Health. Bioelectrical Impedance Analysis in Body Composition Measurement: National Institutes of Health Technology Assessment Conference Statement 1994. Dec 12–14. 1994:1 45. Jebb SA, Murgatroyd PR, Coward WA, Goldberg GR, Prentice AM. In vivo measurement of changes in body composition. Am J Clin Nutr 1993;58:455 46. Nelson ME, Fiatorone MA, Layne JE, et al. Analysis of body composition techniques and models for detecting change in soft tissue with strength training. Am J Clin Nutr 1996;63:678 47. Jebb SA. Measurement of soft tissue composition by DEXA. Br J Nutr 1997;77:151 48. Murray LA, Reilly JJ, Choudhry M, Durnin JVGA. A longitudinal study of changes in body composition and basal metabolism in physically active elderly men. Eur J Appl Physiol 1996;72:215 49. Hewitt MJ, Going SB, Williams DP, Lohman TG. Hydration of fat-free body in children and adults: implications for body composition assessment. Am J Physiol 1993;265:E88 50. Goran MI, Kaskoun MC, Carpenter WH, Poehlman ET, Ravussin E, Fontieville AM. Estimating body composition of young children using bioelectrical resistance. Am J Clin Nutr 1992;56:835 51. Davies PSW, Lucas A. The prediction of total body fatness in early infancy. Early Hum Dev 1990;21:193 52. De Bruin NC, Velthoven KAM, De Ridder M, et al. Standards for total body fat and fat-free mass in infants. Arch Dis Child 1996;74:386 53. Ellis KJ. Measuring body fatness in children and young adults. Int J Obes 1996;20:866 54. Goran MI, Driscoll P, Johnson R, Nagy TR, Hunter G. Cross-calibration of body composition techniques against DEXA in young children. Am J Clin Nutr 1996;63:157 55. Picaud JC, Rigo J, Nyamugabo K, Milet J, Senterre J. Evaluation of DEXA for body composition assessment in piglets and term human neonates. Am J Clin Nutr 1996;63:157 56. Laskey MA. DEXA and body composition. Nutrition 1996;12:45 57. Kohrt WM. Body composition by DXA: tried and true? Med Sci Sports Exerc 1995;27:1349 58. Allen JR, Baur LA, Waters DL, et al. Body protein in prepubertal children with PKU. Eur J Clin Nutr 1996;50:178 59. Brambilla P, Manzoni P, Sironi S, et al. Peripheral and abdominal adiposity in childhood obesity. Int J Obes 1994;18:795 60. Goran MI. Energy expenditure, body composition and disease risk in children and adolescents. Proc Nutr Soc 1997;56:195