Growth and Body Composition: Objectives, Methods, and Measurements Workshop 

Introduction to the National Children’s Study

Mary Hediger, Ph.D., NICHD, NIH, DHHS

Dr. Hediger thanked attendees for their time and participation in the workshop. She reiterated the basic premise of the National Children’s Study (Study)––a federally mandated longitudinal study of environmental factors and children’s health throughout the United States. Dr. Hediger emphasized that the Study will examine children, their development, their families, and their environment. She also pointed out that the Study will be hypothesis-driven, and that environment is being broadly defined.

Dr. Hediger presented a brief overview of Study concepts, priority outcomes, sampling strategy, and timeline. She listed five priority outcomes identified for the Study:

• Undesirable outcomes of pregnancy
• Altered neurobehavioral development and developmental disorders
• Injury
• Asthma
• Obesity and altered physical development.

Dr. Hediger explained that the Study will be a national probability sample, with multistage, cluster design to enable measurement of chemical, physical, and social characteristics of communities and individuals. The Study will be regionalized through a central coordinating center. Enrollment will include recruiting pregnant women during their first or second trimester. Dr. Hediger noted that possibly 25 percent of those recruited will be enrolled prior to conceiving.

Dr. Hediger then outlined several key points used by the Nutrition, Growth, and Pubertal Development (NGPD) Working Group to define the conceptual framework for determining measures that should be considered by the Study:

• Appropriateness: as an indicator of a key body compartment or structure; relevance to one or more Study hypotheses; demonstrated longitudinal continuity; can be used in all (or most) levels of size and fatness within the target age group; is safe; involves minimum subject burden
• Feasibility and applicability: employs survey methodology in the home or clinic, rather than requires specialized laboratory testing; can be used for the whole Study cohort or in substudies
• Technical issues (within and between centers): ensures proven precision and validity; whether technicians and advanced training are necessary; employs standardization; incorporates quality control.

Dr. Hediger discussed the advantages and disadvantages, as well as examples, of three levels of measurement that likely will be considered for assessing body growth and composition:

• Level 1––“Field” methods: currently, widely used; inexpensive; safe; used with the entire Study cohort, with all ages; may be used at home visits; includes anthropometric methods
• Level 2––Laboratory methods: more precise than field methods, but more expensive and less safe; involve more subject burden; may be used to test entire cohort, but not everyone at all ages; may be used for substudies
• Level 3––Laboratory methods (2): use extremely precise methods; most expensive; less safe; most burdensome; may be used within a subsample to address specific hypotheses or to validate Level 1 and Level 2 measures.

In closing, Dr. Hediger reviewed the general format for the workshop and specified overall goals for the workshop:

• Assess methods for measuring growth and body composition
• Pay particular attention to critical concordance between prenatal and postnatal measurements
• Determine the appropriateness and utility of measures for Study use
• Develop a realistic and minimal schedule for timing of data collection
• Specify concurrent measurement of biomarkers (or collection of biospecimens) for interpretation, diagnosis, or prediction
• Identify protocols for ensuring standardization of measurement (technicians), instrumentation, and quality control
• Suggest pilot studies for instrument validation, as well as to provide more accurate equations and to identify promising new techniques.

Body Composition Changes in Pregnancy

Sally A. Lederman, Ph.D., Columbia University

Dr. Lederman began by emphasizing that pregnancy is a time of significant bodily change. During the early stages of pregnancy, those changes affect mainly the mother; it is not until close to term that those changes mostly affect the baby. Dr. Lederman pointed out aspects of pregnancy body composition and their relevance to several Study priority outcomes:

• Birth weight, length, head circumference, and length of gestation
• Child’s body composition at birth
• Child’s neurodevelopment
• Susceptibility to chronic diseases that are related to characteristics of fetal growth, including diabetes and obesity
• Duration of breastfeeding and its effects.

Next, Dr. Lederman discussed two standards used to determine human body composition:

• Two-compartment model: breaks out body weight by fat and lean tissue
• Multicompartment model: also breaks out body weight by fat, but further delineates the components of lean tissue––bone mineral, non-bone mineral, water, and protein.

Dr. Lederman pointed out that the components of lean tissue differ markedly from each other, which is somewhat problematic when determining body composition in pregnancy. She explained how total body water (TBW), density based on underwater weighing (UWW), or total body potassium (TBK) has been used to determine lean tissue in the two-compartment model. Dr. Lederman also pointed out the limitations of traditional two-compartment methods for assessing body composition during pregnancy:

• Cost and need for special laboratory equipment: TBW requires use and measurement of non-radioactive tracer; TBK and UWW require large, complex equipment
• Time consuming and subject burden: TBW requires at least 3 hours for equilibration; UWW requires submerging the subject in water; TBK machine can be claustrophobic.

Dr. Lederman summarized the limitations of two-compartment models using TBW:

• High cost of the marker
• Special facilities needed to assay non-radioactive tracers
• Lengthy measurement time required of the subject (at least 3 hours)
• Misestimating lean tissue, and thus body fat, due to variation in hydration of lean tissue among subjects.

Dr. Lederman discussed multicompartment models for assessing body composition during pregnancy. Typically three- and four-compartment models measure weight, TBW, total body density, and bone mineral. These account for most of the variations in lean tissue during pregnancy.

Dr. Lederman said that there is one main issue that confounds these methods of measurement––pregnancy is characterized by differential increases in various compartments of lean tissue. These compartments change differentially, and they change differently in different women.

Dr. Lederman also pointed out that other standards methods such as dual x-ray absorptiometry (DXA) and magnetic resonance imaging (MRI) currently are not considered appropriate for pregnant women. While DXA has merit, it involves radiation exposure and, therefore, is not used with pregnant subjects. Only one pregnancy study using MRI has been reported. MRI is costly, and both methods require costly, complex instruments.

There are, however, other methods that can estimate fat and or lean tissue in pregnant women:

• Ultrasound: currently only used in measurements of the fetus and infant
• Skinfold measurement: although an age-old method, its value is still uncertain; measurement reliability can be poor; it requires careful training and carefully calibrated calipers; changes are not necessarily fat changes––may be that skinfold, rather than “fatfold” is being measured; pregnancy measurements may reflect changes in fluid as well as fat, especially around delivery.
• Bioelectric impedance analysis (BIA): a newer methodology, but only a few studies have validated or used BIA during pregnancy; although body’s impedance does change with body water changes, the relation of impedance to body water may change during pregnancy; theoretical models currently being developed may improve interpretation of BIA data.

Dr. Lederman concluded her remarks by summarizing several key points for the Study:

• Methods that accurately measure total body composition of individuals are costly.
• None of these methods provides data on regional composition.
• Most methods are labor intensive, and all require technical training, including skinfold (SKF) testing.
• Only BIA and SKF measures use simple, portable equipment and are reasonably convenient and quick.
• Less-expensive methods (such as SKF and BIA) are not well-validated. An individual’s fat/water values also may be very inaccurate.
• Developments that may occur in BIA, MRI, or other scanning methods will require validation, preferably against multicompartment models.
• Improved theoretical approaches (for example, Cole modeling in BIA) may prove useful.
• Good studies will be expensive.

Dr. Lederman emphasized that focus should be placed on:

• Defining the research question to be answered
• Selecting the best method to answer the question
• Determining the minimum sample size and timing of measurements required.

Fetal Growth and Prenatal Body Composition Changes

Patrick M. Catalano, M.D., Case Western Reserve University

Dr. Catalano began by pointing out the importance of considering placental growth as well as fetal growth, noting that by the 28th week of pregnancy, 80 percent of the placenta has developed. He emphasized that a healthy placenta is key to fetal growth and development.

Dr. Catalano reviewed studies using ultrasound to estimate fetal body composition. He stressed the wide range in confidence intervals (CIs) shown in the various studies, cautioning that there is substantial variability in precision of measurement using ultrasound.

Dr. Catalano discussed a study that compared growth rates in various body compartments of large-for-gestational-age (LGA) infants with non-LGA infants. He pointed out that the measurement of abdominal circumference is the one measurement using ultrasound that most closely correlates with fetal weight. These measurements are very time consuming.

Dr. Catalano listed commonly used methods to assess neonatal body composition, including:

• Total body electrical conductivity (TOBEC)
• Anthropometrics
• TBW H ₂ ¹⁸O.

He then discussed several studies that have examined the effect of various factors (such as sex of children, smoking mother, or diabetic mother) on fetal body composition. In summarizing these findings, Dr. Catalano noted that this research has underscored the importance of estimating body composition when studying the effect of defects in the uterine environment on fetal growth and development.

He also presented data using stepwise regression to determine factors correlated with fetal growth and body composition:

•   Maternal factors
  -         Age
  -         Height
  -         Pregravid weight
  -         Weight gain
  -         Pregravid body mass index (BMI)
  -         Gravidity
  -         Parity
  -         Ethnicity
  -         Education
  -         Smoking status
  -         Group (gestational diabetes mellitus/normal glucose tolerance)
• Paternal
  -         Height
  -         Weight
• Fetal
  -         Gestational age
  -         Gender.

Dr. Catalano noted that not surprisingly, gestational age correlates best with body weight at time of birth. He then summarized findings to date of an ongoing 8–10 year follow-up study of 89 infants from the original study:

For those factors within the parental data category, there was no significant difference among tertiles based on:

• Group
• Maternal height
• Weight gain
• Parity
• Glucose screen
• Paternal height
• Paternal weight.

Within the neonatal birth data category, there was no significant difference among the tertiles in:

• Gestational age at delivery
• Gender
• Z-score
• Birth weight
• Fat mass
• Percent of body fat
• Lean body mass.

When reviewing follow-up data, there was no significant difference among the tertiles in:

• Age at follow-up
• Fasting glucose
• Cholesterol
• LDL
• Triglycerides
• Weight/thigh ratio.

Infant Growth and Development (Birth–3 Years)

Nancy F. Butte, Ph.D., Baylor College of Medicine, and Kenneth J. Ellis, Ph.D., Baylor College of Medicine

Dr. Butte first described common anthropometric measurements used to determine body composition in infants––weight, height, BMI, body circumferences, and skinfold thicknesses. She then discussed research using National Center for Health Statistics (NCHS)/CDC growth reference data to:

• Monitor infant’s growth status in relation to reference median
• Weight-for-age (used with young infants)
• Length-for-age (used with young infants)
• Weight-for-length (used with young infants)
• BMI (2-year-olds and older).

Dr. Butte cited recent work conducted at the Children’s Nutrition Research Center (CNRC) of the U.S. Department of Agriculture’s Agricultural Research Service as examples of how these measurements can be interpreted, as well as how they relate to body composition and to each other:

• Height-for-age: assesses stature; related to weight, but not to body composition in terms of fat mass
• Weight-for-age: assesses total body mass; moderately correlated with sum of skinfolds and fat mass, and somewhat less correlated with percent fat mass
• Weight-for-length: assesses body fatness; moderately correlated with sum of skinfolds, fat mass, and percent fat mass.

Dr. Butte pointed out the characteristics of an ideal weight/length ⁿ index as an indicator of adiposity. It should be:

• A good indicator of weight and fat mass
• Uncorrelated with length
• Derived by estimating regression coefficient of log weight on log length at each age.

Given those criteria, Dr. Butte said that BMI may be a fairly sound anthropometric measure.

Dr. Butte presented summary data from the CNRC Growth and Nutrition Study to illustrate use of anthropometric measurements in infants. In this study, the growth and body composition of 40 breast-fed (BF) and 36 formula-fed (FF) infants were monitored from birth to 24 months of age. The study found no significant differences between the BF and FF infants in weight, length, BMI, head circumference, body circumference, and skinfold thickness. However, Dr. Butte pointed out that because babies are born with a low ratio of body fat, it may be that skinfold thickness measures may not be sensitive enough to detect differences.

Dr. Butte discussed changes in body composition in the first 2 years of life, pointing out that:

• Growth involves not only quantitative changes in body size, but also qualitative changes in body composition.
• Chemical composition of fat-free mass (FFM) changes throughout infancy.

She concluded that age- and sex-specific constants are required for use of two-compartment body composition methods based on TBW, TBK, and densitometry.

Dr. Butte then discussed various standard methods for determining body composition in infants and toddlers, noting that the complexities in using these methods with pregnant women also apply to this age group. She pointed out that the CNRC study found that there were unsystematic, but significant, differences among these measurement methods:

• TBW, TBK, TOBEC, and DXA estimations of fat mass (FM) in infants are not interchangeable.
• Wide limits of agreement imply that individual estimates vary significantly.
• Rank order of the methods and magnitude of method differences were a function of age.
• Discrepancies may be due to underlying assumptions regarding composition of FFM (which changes with age).

Dr. Butte emphasized that these data indicate that it is important to choose one measurement method and use it throughout the Study. If for some reason that method changes, it will be extremely important to document that change in detail.

Dr. Butte ended her presentation by briefly discussing the multicomponent model as the “gold standard” for determining body composition. Although comprehensive––measuring FFM, protein, TBW, bone mineral content (BMC), non-bone mineral, and glycogen––this method is still based on certain assumptions and caveats and is tedious and burdensome to carry out.

In discussing the various models used to measure infant body composition, Dr. Ellis cautioned that it is important to understand that while some instruments may yield a whole range of calculations, they are still only measuring one thing. He then briefly listed what these methods are actually measuring.

• One-compartment (1-C) model: simplest model; measures body weight
• Two-compartment (2-C) model: basic model; distinguishes fat from non-fat tissues; Wt=Fat+FFM
• Four-compartment (4-C) model: currently most widely used; breaks out fat, water, protein, and minerals; Wt=Fat+[water+protein+minerals].

Dr. Ellis presented results from a review of studies that used these various methods to measure body composition in infants. He further cautioned, however, that it was likely that each study used its own defined measurement and that different machines and equipment were used in each study. Therefore, before comparing results across studies, it is critical to know what equipment and software were used.

Dr. Ellis discussed a recent study that measured activity levels in 3 to 5-year-olds over a 12-month period. BMC, lean tissue mass, and fat were calculated at baseline and then after 1 year. The researchers described two limitations of assessing body composition in this age group:

• Retention. It was difficult to keep children in this age group in the study.
• Reliability of measurement. It was difficult to obtain reliable measurements due to movement artifacts (that is, it is extremely difficult for a child of this age to remain quiet long enough to obtain a reliable measurement).

Dr. Ellis presented breakouts of data collected by the CNRC. He concluded by discussing the evolution of density-based models, noting that density of FFM is not a constant. Rather, it is a function of age and gender.

Childhood Growth, the Midgrowth Spurt, and the Adiposity Rebound (Ages 4–9 Years)

Alan D. Rogol, M.D., Ph.D., University of Virginia

Dr. Rogol reviewed the main factors that control human growth:

• General health and nutrition
• Intrauterine environment
• Genetics.

He also briefly discussed the regulatory role played by hormones. Dr. Rogol described growth as a sensitive indicator of health, nutrition, and genetic potential. He noted that growth has incremental, hypertrophic, and reparative properties.

Dr. Rogol referred to the infant, childhood, puberty (ICP) model of linear growth as an example of one of the various methods currently being used. He explained that the ICP is a mathematical model of the human growth curve that attempts to relate growth to the underlying dynamics of hormonal and other control.

Next, Dr. Rogol summarized findings from various studies that have identified several key characteristics of the mid-childhood growth spurt:

• It tends to occur around age 7 in boys and 6.7 in girls.
• It may possibly be cyclical.
• It may possibly be seasonal.
• If there is a hormonal basis, it is largely unknown.

If the mid-childhood growth spurt is cyclical and/or seasonal, Dr. Rogol cautioned that it cannot be represented in regular population-based growth charts.

Dr. Rogol defined adiposity rebound (AR) as the point of maximal leanness or minimal BMI, expressed as the ratio of weight to height squared. Children have a rapid increase in BMI during the first year. After a child reaches 9 to 12 months of age, BMI declines and reaches a minimum, usually around age 5 or 6. At that point, BMI begins to gradually increase through adolescence and most of adult life. Dr. Rogol also noted that:

• The time of AR may be a critical period in childhood for the development of obesity; however, this hypothesis is still being tested.

• A younger age at AR is associated with a higher BMI in adolescence and in early adulthood.

Dr. Rogol listed several methods of measurement and suggested timeframes for conducting those measures within the Study:

• Height: twice each year

• Weight

• BMI

• Upper-to-lower ratio

• Abdominal circumference

• Skinfolds: three measurements, once each year

• DXA body composition

• Peripheral quantitative computed tomography (QCT)

• Hormonal measurements, specifically, DHEA-S (consider using saliva) and IGF-I.

He closed by noting that DXA, peripheral QCT, and hormonal measurements would likely not be realistic for the entire Study cohort. Therefore, he suggested that these measures could be taken using a small, but representative, cohort.

Puberty, Growth, and Body Composition Development in Adolescence (Ages 8–18 Years)

John H. Himes, Ph.D., M.P.H., University of Minnesota, Minneapolis

Dr. Himes noted that his discussion would focus on reviewing general patterns and issues that could help frame the Study, given the extended timeframe for studying subjects in this age group. He began by presenting working definitions of key terms:

• Growth is an increase in size, number, or mass.

• Maturation is achievement of adult status, morphology, or function.

• Body composition is the absolute and relative contributions of elements, molecules, or tissue masses to the whole body.

Dr. Himes underscored that puberty is a process, a series of interactions related to endocrinal and hormonal changes. Puberty is a time of rapid growth and maturation. Because there is so much change occurring during puberty, the Study will need to be very selective about what will be measured, as well as how and when. Dr. Himes pointed out that the Study will examine morphological manifestations of puberty––skeletal growth and body composition.

Dr. Himes next discussed several seminal studies, including the Child Research Council longitudinal study of Denver children that compared different patterns of growth and maturity among boys and girls. The CRC compared median combined radiographic thicknesses of subcutaneous fat from the forearm, thigh, calf, deltoid, and hip in girls and boys. The researchers found that boys tend to lose fat whereas girls gain fat during adolescence.

Dr. Himes summarized several issues of maturational status relative to somatic growth and body composition during adolescence:

• Timing of peak adolescent growth velocity varies among individuals and among body dimensions and composition components.

• Maturation status is associated with variation in growth and body composition variables even within small chronological age groups.

• Amount of maturation-related variation varies but may be substantial.

• Maturation-related variation is not random. That is, components respond in a predictable way to changes in maturity.

He also outlined the implications of the effects of maturation-related variation in growth and body composition during adolescence:

• For individuals, there is an increase in misclassification and bias when using reference data based only on chronological age.

• For groups, there is an increase in the standard deviation (SD); there is a decrease in precision of estimates; likewise, there is a decrease in detection of differences, effects, and change.

Dr. Himes presented data from several studies, including NHANES III, which assessed stages of sexual maturity independent of chronological age against certain variables, such as stature, FFM, and variation in percent body fat. He also reviewed findings from a study of white 1 to 13-year-old girls in Minnesota, pointing out that girls with later onset of puberty tended to have less mean total body fat than girls with earlier onset of puberty.

Dr. Himes concluded by summarizing several implications of puberty on assessing growth and body composition in the Study. In particular, the Study will need to:

• Conduct more frequent measurements than in childhood to accommodate the rapid growth and change that occur during puberty.

• Measure a wide range of dimensions and components of body composition.

• Measure maturational status on multiple occasions.

• Develop valid and reliable noninvasive measures of maturation.

• Use maturation variables in analyses.

Dr. Himes again emphasized that if attempting to detect differences among adolescents, what is being measured, as well as when it is being measured, will be significant in terms of the magnitude of those changes being detected.

Lean Body Mass, Skeletal Muscle, and Body Water

Jack Wang, M.S., Columbia University, and Ira Bernstein, M.D., University of Vermont

Dr. Bernstein focused his presentation on reviewing some of the specific considerations related to fetal measures of body composition that could possibly be included in the Study. He pointed out that fetal measures are unique and somewhat specific. He discussed the basic framework for current techniques used to assess fetal body composition:

• Standard obstetrical ultrasound equipment

• Well-defined anatomic planes

• Limited to subcutaneous fat depots (80 percent of fat in neonate)

• Anatomic locations
  –         Abdominal wall
  –         Proximal upper and lower extremity.

Dr. Bernstein noted that currently data are restricted to 2-D imaging. He also explained that the discussion centered on two measurement sites––the abdominal wall and the proximal upper and lower extremities—because these measures have been validated based on comparisons among newborns in multiple clinical studies.

He next summarized relevant validation studies that examined abdominal wall thickness (to measure neonatal fat) and proximal extremities, that is, the humerus/femur (to measure neonatal fat and lean body mass).

Dr. Bernstein offered two considerations for the Study:

• Two-dimensional (2-D) ultrasound imaging of the fetus can provide validated indices of body composition. Anatomic locations include abdominal wall (length), proximal humerus, and proximal femur (area).

• 3-D imaging has not been critically examined for measures of fetal body composition and no validated measures exist. Therefore, significant pilot data would be required to establish reliable measurement techniques.

He concluded by noting preferred timeframes for observations and measurements:

• Observations should occur over the second half of pregnancy (beginning at or beyond 20 weeks gestation).

• Serial scans should occur every 4 to 6 weeks until delivery, based on discretely defined gestational age windows (for example, at weeks 20–24, 25–28, 29–32, and 33–36).

Mr. Wang began by explaining that his presentation would focus on discussing currently available techniques that could be used to measure FFM, TBW, skeletal muscle mass (SMM), and body dimensions during the five phases of the Study. Each method had been scored based on several criteria:

• Risk
• Participant burden
• Reliability
• Accuracy
• Feasibility
• Cost.

First, Mr. Wang described current methods for measuring FFM, as well as specific variables measured by each technique:

• Anthropometry: weight, length, circumference, width, skinfold thickness

• UWW: body volume (BV)

• Tracer dilution: TBW, ECW

• 40K counting: TBK

• Creatinine: 24-hour or 28-hour urinary excretion

• 3-methylhistidine: 48-hour urinary excretion

• CT: total/regional adipose tissue and skeletal muscle volume, organ sizes

• BIA (SF, MF): bioelectrical impedance

• Ultrasound: subcutaneous fat thickness (prenatal growth and development)

• IR: subcutaneous thickness

• IVNA: body elements

• DXA: total/regional body fat, FFM, bone mineral

• MRI: total/regional adipose tissue and skeletal muscle volume, organ sizes

• GNRA: body elements

• ADP (PP, BP): BV

• TBSS: total/regional BV, circumference, length, width.

Mr. Wang then discussed the issue of reliability of skinfold measurements to determine body circumference. He reported on a study of 26 clinical nurses who were trained to measure body circumference at five body locations and skinfold at five different body locations. After observing the trainer’s demonstration and practicing on at least 10 subjects, 73 percent of the nurses achieved skill levels for body circumferences less than 2 percent different from the trainer’s reading. These same nurses achieved skill levels for skinfold measurements that differed by less than 20 percent from the trainer’s reading.

Mr. Wang also reviewed findings from a number of studies that examined various other measurement techniques:

• Ultrasound assessment of body composition in obese adults––chiefly overcoming the limitations of the skinfold caliper
• UWW as a technique to measure percent of body fat
• 3-D photonic scanning to assess body volume in the head, torso, left arm, right arm, left leg, and right leg

• BIA to predict TBW and FFM in healthy and HIV-infected children and teens

• BIA to estimate SMM.

Mr. Wang next discussed the scale used to rate the various methods used to measure FFM, TBW, SMM, and body dimensions. He explained that each method was assigned a score based on the six previously defined criteria. The highest score was 12; the lowest was 0.

He concluded by summarizing the following issues for consideration by the Study when finalizing how body composition will be measured:

• Minimize participant’s burden to help ensure long-term retention.
• Minimize participant’s risk from participation.
• Establish study sites convenient to participants.
• Use a multiple-measurement approach.
• Standardize protocol and instrumentation.
• Centralize data production and management.
• Establish a long-term quality assurance protocol.
• Conduct pilot studies.

• Establish a similar database in a normal cohort.

Bone Mineral Content and Density

Steven B. Heymsfield, M.D., Columbia University

Dr. Heymsfield discussed the dynamics of the relation among body weight, skeletal muscle, and bone. Bone mass increases early in life, peaks in adulthood, and then declines. A modern view is that bone actually is generated by mechanical forces––body weight or physical activity. Force exerts tension on skeletal muscle. In turn, skeletal responds by pulling on bone thereby influencing bone mass and bone density. Dr. Heymsfield emphasized that these factors are all interrelated, although the relation between bone and muscle stays more or less stable.

He next briefly described changes in bone properties due to age:

• Increase in cortical thickness/diameter

• Stable mineral and matrix content in mature bone

• Decrease in bone hydration during maturation.

Dr. Heymsfield discussed factors associated with inadequate bone mass accumulation in children:

• Systemic steroid use for longer than 1 month over a year

• Immunosuppressive medication use

• Prior malignancy

• Systemic burns

• Chronic pulmonary diseases

• Chronic renal disease

• Malabsorptive disorders

• Chronic liver disease

• Muscular weakness

• Chronic rheumatic diseases

• Immobility

• Chronic anticonvulsant therapy

• Poorly controlled diabetes mellitus

• Thyroid hormone use

• Anorexia nervosa

• Genetic hypercalciuria

• Inadequate nutrition

• Unhealthy lifestyle.

Dr. Heymsfield also discussed the benefits and weaknesses of methods used to assess bone mineral content and density:

• Quantitative computed tomography (QCT) and peripheral QCT (pQCT): bone composition and architecture, as well as skeletal muscle; expensive; only appendicular bones can be studied; independent of chronological age and body size; false-low bone mineral density (BMD) in small subjects; radiation exposure; reproducibility of BMD +/- 0.3–1.2 percent

• Dual-energy x-ray absorptiometry (DXA): bone composition of multiple bones and body composition; expensive; radiatio exposure; cannot separate trabecular and cortical bone; pediatric influences; reproducibility of BMD +/- 1–1.5 percent

• Quantitative ultrasound (QUS): no radiation; relatively easy to use; inexpensive; unlimited repetition; many pediatric influences including macrostructure effects.

Adiposity and Regional Fat Distribution

Henry S. Kahn, M.D., National Center for Chronic Disease Prevention and Health Promotion, CDC, DHHS

Dr. Kahn began by reiterating the longitudinal implications of examining adipose tissue (AT) for the Study:

• Relation of AT to the etiology of adult chronic disease
  -         Causal and protective roles
  -         Marker roles
• Environmental origins of AT accumulation
  -         Total AT
  -         Regional distribution
  -         Relationships to accumulated lean mass.

Dr. Kahn also emphasized the need to consider conventional field methods, as well as high-techoptions, in terms of feasibility, cost, and relevance to public health.

He reviewed the concepts of lipid overaccumulation and lipotoxicity. Lipid overaccumulation occurs when the flux of lipid fuels exceeds the capacity of peripheral AT to buffer and store energy. This leads to larger waist size and an increase in circulating triglycerides.

Lipotoxicity, a consequence of lipid overaccumulation, is associated with deposition of lipid in ectopic tissues, leading to:

• Insulin resistance in skeletal muscle and the liver

• Decline of pancreatic ß-cell function, which in turn, leads to type 2 diabetes and coronary heart disease (CHD).

Dr. Kahn next discussed the concept of caudal diminution and the suppression of caudal growth. He presented several possible meanings of caudal diminution:

• Marker of growth restraint in specific time windows

• Reduced lower-body AT causes decreased ability to buffer and store lipid fuels

• Reduced lower-body muscle leads to decreased disposal of calories (decreased locomotion)

• Reduced lower-body bone length provides reduced frame for support of functioning AT.

Dr. Kahn followed with a rationale for studying reduced lower-body size (caudal diminution):

• The association with lipid overaccumulation and with raised levels in triglycerides, insulin resistance, type 2 diabetes, and CHD

• Continued uncertainty about which diminished tissue(s) are most important

• Lack of clarity about optimal timeframes for programming upper/lower ratios.

He also pointed out that lower-body measurements are noninvasive and inexpensive.

Dr. Kahn discussed several studies as examples of critical outcomes among an older population for which the predecessor state could have been in the fetus or due to some childhood occurrence. He also reported on a study of persons with type 2 diabetes, which showed an inverse association with hip and thigh circumference.

Dr. Kahn further suggested that measuring head diameter at 18 weeks gestational age and femur length at 20 weeks could be used to track peak growth velocity. These measures might be able to identify environmental factors that could be linked to the failure of the fetus to grow and develop.

Dr. Kahn next discussed fingerprint ridge counts as a measurement technique. He explained that fingerprints are formed by the 19th week of gestation. Thus, if it can be demonstrated that a particular pediatric or adult outcome is associated with fingerprints, then it is logical to conclude that the outcome represents a dynamic that occurred before the 19th week. Dr. Kahn also pointed out that fingerprints are sequentially related to spinal cord segments. Finally, fingerprint measurement is inexpensive, and current technologies are more efficient and less cumbersome than earlier methods.

Dr. Kahn described several measurement options:

• External anthropometry
• Hand morphology: fingerprints and digit lengths
• Imaging: fetal ultrasound; MRI, CT, or liver fat in children and teens
• BIA
• DXA.

He reiterated that external anthropometry is relatively inexpensive, and available for later clinical adoption. This method can be used to measure weight, height, circumferences, sagittal abdominal diameter (SAD), and subcutaneous AT.

Dr. Kahn concluded his remarks by suggesting timeframes and measurement methods for examinations that would include adiposity and regional fat distribution in the Study cohort:

• Mother (external anthropometry): at enrollment, 17 weeks gestational age, and at 37 weeks gestational age

• Fetus (ultrasound): at 17 weeks, 27 weeks, and 37 weeks gestational age

• Newborn (external anthropometry): placental weight and assay for LPL activity, FABP expression

• Child and adolescent (external anthropometry): at ages 3 months and at 1, 3, 5, 8, 13, and 18 years.

Obesity and Biomarkers of Insulin Resistance

Stephen R. Daniels, M.D., Ph.D., Cincinnati Children’s Hospital Medical Center

Dr. Daniels explained that he would focus on the “downstream effects” of obesity, noting that obesity and physical development is a priority outcome for the Study.

Given that the most widely used definition of obesity is based on BMI, Dr. Daniels suggested that BMI may be the most useful way to identify obesity in the clinical setting. He also noted that BMI is simple, inexpensive, and relatively easy to interpret. In discussing the best method to evaluate overweight and obese children in the Study, Dr. Daniels suggested that multiple methods may be needed.

Dr. Daniels emphasized the broad metabolic impact of obesity, noting that obesity affects virtually every organ system. He listed adverse outcomes related to obesity:    

• Metabolic: type 2 diabetes mellitus, metabolic syndrome, dyslipidemia
• Orthopedic: slipped capital femoral epiphysis, Blount’s disease

• Cardiovascular: hypertension, left ventricular hypertrophy, atherosclerosis

• Psychologic: depression, poor quality of life

• Neurologic: pseudotumor cerebri

• Hepatic: nonalcoholic fatty liver disease, nonalcoholic steatohepatitis

• Pulmonary: obstructive sleep apnea, asthma (exacerbation)

• Renal: proteinuria, end-stage renal disease.

Dr. Daniels discussed the relationship of obesity with lipids and lipoproteins, noting that obesity:

• Increases triglycerides

• Decreases HDL-cholesterol

• Increases LDL-cholesterol.

He also summarized studies that examined the relation of obesity and cardiovascular disease (CVD), noting that even small differences in blood pressure in obese persons can substantially increase the risk of CVD. Other studies examined obesity and left ventricular mass (LVM), finding obesity to be an important determinant of LVM.

Dr. Daniels discussed whether obesity contributes to development of atherosclerosis. He cited the Bogalusa study that found that obesity accelerates coronary atherosclerosis. He noted that increased BMI was associated with increased prevalence of fatty structures and fibrous plaque. Furthermore, childhood weight and BMI were found to be significant predictors for development of coronary artery disease. Dr. Daniels pointed out that a number of noninvasive methods are now available to evaluate the development of atherosclerosis, suggesting that some of these methods may be feasible for the Study.

Dr. Daniels described insulin resistance syndrome (IRS), noting that IRS is not a specific clinical disease, entity, or diagnosis. He explained that the primary cause is resistance to insulin action, resulting in ß cell compensation, leading to increased insulin production and secretion. Dr. Daniels discussed the two models currently used to assess insulin resistance––homeostasis assessment model (HOMA) and QUICKI.

Dr. Daniels next discussed the relation of obesity with metabolic syndrome. He summarized findings from studies, including NHANES III and the Bogalusa study, indicating that there was a substantial increase in the prevalence of metabolic syndrome in obese 12 to 19-year-olds.

Dr. Daniels ended by pointing out that:

• That the definition of obesity may be complex.

• Obesity results in numerous adverse effects.

• Insulin resistance is important and probably best assessed in a large epidemiologic study by measuring fasting insulin and glucose.

• There is increasing interest in and concern about the metabolic syndrome.

• It is not clear at present how to best define the metabolic syndrome in young individuals.

Air Displacement Plethysmography

Alessandro Urlando, M.S., Life Measurement, Inc.

Mr. Urlando explained the overall design and operational concepts of the BOD POD ^® Body Composition Tracking System. This technology measures body composition through densitometry. The BOD POD determines volume through application of air displacement plethysomography using a precision load cell scale system.

Mr. Urlando cited published research on use of BOD POD in adults. To date, there are more than 50 published studies on body composition assessment in adults using the BOD POD. The adult populations in these studies ranged in age from 18 to 86 years. BMI ranged from 17 to 40 kg/m2. Specific population groups were included in the studies, such as the disabled, chronically ill, severely obese, and the elderly. Most of these studies evaluated the performance of the BOD POD using UWW, DXA, or multicompartment models (3- or 4-C) as the reference methods.

A summary of 25 studies comparing percent of body fat using the BOD POD versus the reference methods in adults showed that on average the BOD POD and the reference methods agreed within 1 percent body fat for adults. Individual agreement between BOD POD and the reference methods was considered “excellent to ideal.”

Mr. Urlando also pointed out that to date, there have been more than 12 published studies on body composition assessment in children using the BOD POD. The children studied ranged in age from 5 to 19 years. The BMI ranged from 13 to 45 kg/m ². Younger children, aged 5 to 7 years, overweight and obese children, various ethnic groups, and children with cystic fibrosis were also studied. Mr. Urlando reported that percent body fat by DXA was highly correlated with the BOD POD, for both baseline and follow-up measurements.

Mr. Urlando next described the PEA POD ^® Infant Body Composition System, which is based on the same air displacement technology as the BOD POD. He noted that the PEA POD was designed to address all the issues associated with more traditional measurement methods, such as practical limitations, training requirements, limited availability, cost, accuracy, and safety.

Mr. Urlando emphasized that the PEA POD can be used to:

• Assess infant growth

• Optimize nutritional and pharmacological interventions

• Investigate and optimize release criteria in NICU settings

• Establish normative body composition data.

He also listed various benefits of this technology. That is, the PEA POD:

• Is accurate and precise

• Is safe and noninvasive

• Provides immediate feedback

• Requires a short test time

• Accommodates most infant behaviors

• Is comfortable, with a heated testing environment

• Is mobile

• Is user-friendly, with minimal operator training requirements

• Includes menu-driven software with data management capabilities.

Mr. Urlando also summarized data from validation studies of the PEA POD compared with the two-compartment model.

Optical Detection of Subcutaneous Fat Thickness

Kenneth J. Ellis, Ph.D., Baylor College of Medicine

Dr. Ellis discussed the preliminary evaluation of the Lipometer by the Children’s Nutrition Research Center (CNRC). He described the basic components and operation of the Lipometer as a tool to measure body composition, including total body fat and FFM. Dr. Ellis pointed out that the Lipometer is a mobile system with few component parts. He then presented examples of screen shots and printouts showing detailed measurements from 15 body sites.

He reported on an assessment comparing DXA and the Lipometer as methods for monitoring body fatness in school-aged children. This study, which is ongoing, is following 325 children and young adults between the ages of 3 and 26. Dr. Ellis presented preliminary findings, pointing out several benefits of the Lipometer, as well as potential drawbacks, and noted that the weaknesses will likely be resolved and clarified with further study.

• Strengths:
  -         Is easy to use
  -         Alerts the operator if the procedure is not correct
  -         Takes only about 5 minutes to measure all 15 sites
  -         Results are not operator dependent
  -         Has good reproducibility
  -         Correlation with skinfold measurements at each site (r = 0.30–0.85)
  -         Can be used to predict percent fat with reasonable accuracy.
• Weaknesses:
  -         Occasional low reading (especially among African Americans)
  -         Change in skin pigmentation may influence values
  -         Anatomical site must be correct.

Quantitative Ultrasound

David J. Helowicz, Sunlight Medical, Inc.

Mr. Helowicz described several new products that use ultrasound technology to assess skeletal development in children and adults. He pointed out that bone age testing can help identify growth patterns and deficiencies in young children. Mr. Helowicz pointed out several strengths of the BonAge product. It is:

• Ultrasound based

• Radiation-free

• Objective, providing on-the-spot measures of skeletal development in children ages 5 to 18

• Precise

• Operator-independent.

Mr. Helowicz explained that BonAge measures the ossifying cartilage structures of the wrist, which provides a view of skeletal development. BonAge measures the velocity of an ultrasound wave transmitted through the wrist, using a proprietary gender-and ethnically based algorithm to generate a numeric bone age score. These scores are available on site, and do not require interpretation by the physician. A full-color report presents bone age, height, and adult height prediction, compared with age-, gender-, and ethnicity-matched normal curves.

Mr. Helowicz presented a video demonstrating operation of the BonAge assessment tool, as well as overviews of several new prototypes of other products currently under development that utilize ultrasound technology for body composition assessment.

Questions and Open Discussion

At the close of the morning and afternoon sessions, participants asked questions and offered comments relevant to the Study design, hypotheses, and outcomes that would influence suggestions and findings regarding measuring body composition within the Study. These discussion points included:

• Recognition that it will be critical not only to define measurements, but also the timing/staging for when those measurements will be taken

• What measurements fit best within the Study will depend on each specific question or hypothesis

• New technologies are emerging that may alter the entire approach to measuring body composition

• Whether “localized findings” will be excluded due to IRB guidelines

• The need to recognize that what is being measured may be an outward manifestation of an internal dynamic related to growth, and to connect internal dynamics with observable occurrences

• Accommodation of sample collection and storage

• What is currently considered a topic of interest for Study inclusion may cease to be of interest within the next 5–10 years

• Reiteration that the Study hypotheses have been formulated in the broadest terms to capture as much information from as many subjects as possible

• Emphasis on defining types of feasible measurements, populations to be measured, and timing of measurements within a cohort of 100,000

• Measurement continuity, in particular, transitioning from the prenatal to postnatal stage

• Determining measures of predictable risk and understanding that those measures may be different in children than in adults

• Establishment of norms and how to define them within the Study context

• Recognition that “normal” ranges have not been determined for some types of measurement

• Limitations of technologies currently being used to measure body composition, particularly in special populations (for example, children with structural malformations or physical disabilities).

Breakout Sessions

Robert J Kuczmarski, Dr.P.H., R.D., National Institute of Diabetes and Digestive and Kidney Diseases, NIH, DHHS

Dr. Kuczmarski explained that the workshop, so far, had:

• Provided an overview of the Study
• Described the conceptual framework of growth and body composition developed by the Nutrition, Growth, and Pubertal Development Working Group
• Described the three levels of appropriate measurements
• Offered presentations on variety of growth and development topics (the “what” of the workshop)
• Listed the workshop objectives
• Presented findings from other Study workshops.

According to Dr. Kuczmarski, the purpose of the breakout sessions was to focus on the “how” by developing a consensus on the most promising methods for measuring growth and body composition across the lifespan of a large longitudinal study. Workshop participants were assigned to one of the following age/status groups:

• Pregnancy
• Fetal growth and preterm infants
• Infants (birth to age 3 years)
• Childhood (ages 4–9 years)
• Adolescence (ages 8–21 years).

• (There is an overlap in age ranges for childhood and adolescence.)

Dr. Kuczmarski commented that the challenges for the breakout groups increased with increasing age of Study participants. He noted that a particular challenge for the Study is incorporating evolving technologies that will become state-of-the-art methods over the next
15–20 years. Each group considered the following questions:

• Which methods do you consider to be most promising for use in this age range for the Study? Why?

• How often and when (at what ages) should measurements be taken? Optimally? Mandatory minimum?

• Are there any unique opportunities for data collection in this age range? Are there any special concerns? For example, should neonates be measured on a gestation-corrected or chronological age schedule? If measured at gestation-corrected ages, when should the approach change to chronological age?

• What are the current problems or barriers, if any, with using these methods in the Study? Do you believe it is possible to overcome these problems/barriers? How?

• For those methods that are most promising, what research, additional validation, or pilot study would be needed to make this instrument(s) a viable option for measuring growth and/or body composition in the Study?

• Are these instruments or methods sufficiently developed to provide continuity across time?

In their discussions, the groups specifically identified the following information:

• Age (or appropriate range)
• Measurement(s)
• Level: 1, 2, 3
• Rationale
• Substudy rationale for levels 2 and 3
• Technical concerns
• Concurrent measurements
• Pilot/validation needed.

At the conclusion of discussions, each group reported its findings in a measurement rating form, providing information on:

• Anthropometric dimension or body composition
• Measurement method
• Relative ratings:
  –         Risk (min = 2, med = 1, max = 0)
  –         Participant burden (min = 2, med = 1, max = 0)
  –         Reliability (high = 2, acceptable = 1, low = 0)
  –         Accuracy (high = 2, acceptable = 1, low = 0)
  –         Feasibility (high = 2, med = 1, low = 0)
  –         Cost (low = 2, med = 1, high = 0)
• Time points for measurement:
  –         Specify as year of age, time interval, and age category
  –         Rate the frequency of measurement as realistic/feasible, absolute minimum, etc.
•   Rate longitudinal continuity between ages:
  –         From adolescence to childhood (high = 2, acceptable = 1, unacceptable = 0)
  –         From childhood to infancy (high = 2, acceptable = 1, unacceptable = 0)
  –         From infancy to fetal period (high = 2, acceptable = 1, unacceptable = 0)
  –         From adolescence to adulthood (high = 2, acceptable = 1, unacceptable = 0)
•   Appropriateness of the measure:
  –         Whole Study or substudy only
  –         Field method or restricted to laboratory
•   Associated technical issues (for example, technician training, standardization, quality control issues, etc.—specify issue and level)
  –         None/easily resolved (a )*
  –         Moderately difficult (b)*
  –         Complex (c)*
•   Status of method for the Study
  –         Ready for use (a)*
  –         Requires pilot study (specify) (b)*
• Other considerations, notes, comments.
  *See breakout group reports.

Pregnancy Breakout Group Report

Sally Ann Lederman, Ph.D., Columbia University

Dr. Lederman summarized the pregnancy breakout group’s findings in the following matrixes.

Pregnancy matrix, part 1

NST = nonstretchable tape, R1 = risk, B = participant burden, R2 = reliability, A = accuracy, F = feasibility, C = cost, DXA = dual X-ray absorptiometry, MF BIA = multifrequency bioelectric impedance analysis, D ₂O = deuterium, OGTT = oral glucose tolerance test, HbA _1c = glycosylated hemoglobin, Hct = hematocrit.

Pregnancy matrix, part 2

WS = whole Study, F = field, T = training, S = standardization, QC = quality control, MRI = magnetic resonance imaging, DXA = dual X-ray absorptiometry, MF BIA = multifrequency bioelectric impedance analysis, D ₂O = deuterium, UWW = underwater weight, OGTT = oral glucose tolerance test, HbA _1c = glycosylated hemoglobin, IGF = insulin-like growth factor, Hct = hematocrit.

Fetal Growth Breakout Group Report
Ira M. Bernstein, M.D., University of Vermont

Dr. Bernstein summarized the fetal growth breakout group’s findings in the following matrixes.

Fetal matrix, part 1

*Only option

BPD = bi-parietal diameter, OFD = occipital-frontal diameter, 2-D U/S = two-dimensional ultrasound, R1 = risk, B = participant burden, R2 = reliability, A = accuracy, F = feasibility, C = cost, AC = abdominal circumference, FL = femur length, HL = humerus length, AFA = arm fat area, AMA = arm muscle area, TFA = thigh fat area, TMA = thigh muscle area, 3-D U/S = three-dimensional ultrasound, U = unknown, BIA = bioelectrical impedance analysis, QUS = quantitative ultrasound, GCA = gestation-corrected age.

Fetal matrix, part 2

*Only option

BPD = bi-parietal diameter, OFD = occipital-frontal diameter, 2-D = two-dimensional ultrasound, A = adult to adolescent, B = adolescent to child, C = child to infant, D = infant to fetus, WS = whole Study, F = field, T = training, S = standardization, QC = quality control, AC = abdominal circumference, FL = femur length, HL = humerus length, AFA = arm fat area, AMA = arm muscle area, TFA = thigh fat area, TMA = thigh muscle area, BIA = bioelectrical impedance analysis, QUS = quantitative ultrasound.

Infant to Age 3 Years Breakout Group Report
Adolfo Correa, M.D., Ph.D., M.P.H., National Center on Birth Defects and Developmental Disabilities, CDC, DHHS

Dr. Correa summarized the breakout group’s findings in the following matrixes.

Infant to age 3 matrix, part 1

R1 = risk, B = participant burden, R2 = reliability, A = accuracy, F = feasibility, C = cost, NHANES = National Health and Nutrition Examination Survey, DXA = dual X-ray absorptiometry, D ₂0 = deuterium.

Infant to age 3 matrix, part 2

A = adult to adolescent, B = adolescent to child, C = child to infant, D = infant to fetus, WS = whole Study, F = field, T = training, S = standardization, QC = quality control, BW = body weight, NHANES = National Health and Nutrition Examination Survey, DXA = dual X-ray absorptiometry.

Children, Ages 4–9, Breakout Group Report
Bonny Specker, Ph.D., South Dakota State University

Dr. Specker summarized the breakout group’s findings in the following matrixes.

Childhood matrix, part 1

R1 = risk, B = participant burden, R2 = reliability, A = accuracy, F = feasibility, C = cost, WC = waist circumference, HC = head circumference, DXA = dual X-ray absorptiometry, vBMD = volumetric bone mineral density, pQCT = peripheral quantitative computed tomography, BIA = bioelectric impedance analysis.

Childhood matrix, part 2

A = adult to adolescent, B = adolescent to child, C = child to infant, D = infant to fetus, WS = whole Study, F = field, T = training, S = standardization, WC = waist circumference, AC = abdominal circumference, NHANES = National Health and Nutrition Examination Survey, ATPIII = Adult Treatment Program III, vBMD = volumetric bone mineral density, BMD = bone mineral density, BMC = bone mineral content, BA = bone area, QC = quality control, BP = blood pressure.

Adolescents, Ages 8–18, Breakout Group Report
Babette S. Zemel, Ph.D., University of Pennsylvania School of Medicine

Dr. Zemel summarized the breakout group’s findings in the following matrixes.

Adolescent Matrix, part 1

R1 = risk, B = participant burden, R2 = reliability, A = accuracy, F = feasibility, C = cost, NST = nonstretchable tape, TBD = to be determined, U = unknown, BMC = bone mineral content, LBM = lean body mass, DXA = dual X-ray absorptiometry, pQCT = peripheral quantitative computed tomography, D ₂O = deuterium, 2D:4D = second digit-to-fourth digit.

Adolescent matrix, part 2