Measurement of body composition in children

INTRODUCTION — The measurement of body composition may include direct or indirect measurements of body fat, lean body mass, and bone mass, and sometimes of the distribution of fat between the visceral or subcutaneous compartments. The choice of method depends on which of these compartments are of interest, whether the measurement is for clinical purposes or research, and what degree of precision is required.

The main methods used to estimate body composition are discussed here. Measurements of growth in children and disorders of under- or overnutrition are discussed separately. (See "Measurement of growth in children" and "Definition, epidemiology, and etiology of obesity in children and adolescents" and "Poor weight gain in children younger than two years in resource-abundant settings: Etiology and evaluation".)

THEORETICAL MODELS OF BODY COMPOSITION — Theoretical models of body composition divide the body into two, three, or multiple compartments:

●In the two-compartment model, the body is divided into the fat and fat-free mass. Bioelectrical impedance and deuterium dilution are methods used to assess body composition based on the two-compartment model. The two-compartment model is useful in clinical practice because of the ease with which body fat and fat-free mass can be measured and the simplicity with which their changes during health and disease can be assessed. However, the two-compartment model is subject to error because the methods used to measure body fat and fat-free mass are based upon the assumption that the chemical composition of these tissue stores remains constant across a broad range of ages and disease states.

●In the three-compartment model, the body is divided into fat, fat-free mass, and bone. Dual-energy x-ray absorptiometry (DXA) is a method to assess body composition based on the three-compartment model. The body composition of children has been measured using this method [1-3]. Fat-free mass and body fat increase with age throughout childhood, with significant divergence by sex beginning in puberty, when fat mass normally increases in females and lean mass increases in males [1-4]. Several resources report differences among racial/ethnic groups, but this factor predicts a minimal component of the overall variance within any population. Nonetheless, these estimates provide general reference values for healthy children and useful comparative indices to assess nutritional deficits in children who are ill [5].

●Multicompartment models reduce the errors inherent in the two-compartment model by dividing the body into components based on its atomic, molecular, cellular, or tissue composition (figure 1) [6]. Although these models describe body composition more precisely, the techniques required to measure these components are limited to research settings.

ESTIMATES OF ADIPOSITY — Numerous measures are used for clinical assessment of obesity in children and adolescents and include body mass index (BMI), weight-for-height, skinfold thickness, and waist circumference or waist-to-hip ratio (WHR). BMI is the standard measure of obesity for children older than two years; waist circumference and waist-hip ratio are used primarily for research purposes. (See "Definition, epidemiology, and etiology of obesity in children and adolescents".)

Body mass index — BMI is the most practical way to evaluate the degree of excess weight and risk for metabolic syndrome [7]. It is calculated from the weight and square of the height and can be determined using a calculator for boys (calculator 1) and for girls (calculator 2). In children, BMI varies with age and sex, so that BMI percentiles are used for defining overweight (BMI ≥85^th percentile) or obesity (BMI ≥95^th percentile) using standard growth charts (figure 2A-B). For children with severe obesity, extended growth charts are available: (figure 3A-B) or CDC extended BMI growth charts. (See "Definition, epidemiology, and etiology of obesity in children and adolescents" and "Measurement of growth in children", section on 'Body mass index'.)

BMI is an imprecise measure of adiposity because it does not distinguish between fat mass and fat-free mass [8]. A meta-analysis and systematic review reported that BMI in children had high specificity (approximately 0.95) but poor sensitivity (approximately 0.69) in detecting obesity, meaning that over one-quarter of children with BMIs below the BMI-based obesity threshold had excess adiposity, as measured by more direct methods [9].

The same BMI percentile does not represent the same percentage body fat at different ages or stages of pubertal maturation, for males and females, or among children with different muscle mass. As examples:

●In children <6 years, BMI may overestimate adiposity. One study suggests that higher cutoffs should be used in this age group for defining overweight [10].

●BMI tends to overestimate adiposity in children who are short or who have relatively high muscle mass [9,11]. Conversely, it tends to underestimate adiposity in children who are very tall or who have reduced muscle mass (eg, due to low levels of physical activity). Thus, children of different stature but the same BMI do not necessarily have the same body composition [12].

●BMI also varies with pubertal maturation stage, and this should be taken into account when BMI measures are used for research [13-15].

Weight-for-height — Weight-for-height measurements are another means to assess adiposity and are the preferred method for clinical assessment of obesity or failure to thrive in children younger than two years old. In older children, overweight, simple obesity, and severe obesity are sometimes defined as weight-for-height measurements greater than 110, 120, and 140 percent of expected, respectively [16]. Expected body weight is determined from the weight percentile that is proportionate to the height percentile for chronologic age using standard growth curves (figure 4A-B and figure 5A-B).

For children older than two years, BMI has replaced weight-for-height estimates as the standard measure for adiposity because weight-for-height measurements do not correlate well with body fat and are affected by height to a greater degree. (See 'Body mass index' above.)

Combined prediction models — Combined prediction models use multiple patient characteristics to predict adiposity; these models can be useful for population research but have little role in routine clinical care. As an example, a prediction model using height, weight, age, sex, and ethnicity was reasonably well correlated with adiposity (as measured by the deuterium dilution method) [17]. The model was subsequently validated in a separate group of children and adolescents from 19 countries (n = 5693) [18]. The investigators developed an Excel spreadsheet to facilitate the calculation [19].

FAT DISTRIBUTION — Adults with abdominal adiposity (also called central, visceral, android, or male-type obesity) are at increased risk for heart disease, diabetes, hypertension, and some forms of cancer. Similar data exist for children [20]. Data from the Bogalusa Heart Study generated age-specific thresholds for body mass index (BMI) and waist circumference in children and adolescents that were predictive of having three or more risk factors for cardiovascular disease (high blood pressure; low high-density lipoprotein cholesterol concentration; or high low-density lipoprotein cholesterol, triglyceride, glucose, or insulin concentration, adjusted for age) (figure 6A-B) [21]. A more recent study suggests that the 95^th BMI percentile is a useful threshold to predict elevated visceral adipose tissue and cardiometabolic risk in children and adults [22].

Abdominal obesity is evaluated clinically by measuring the waist circumference or the ratio of waist circumference to the hip circumference (waist-to-hip ratio [WHR]). Waist circumference is measured just above the uppermost lateral border of the right ilium at the end of a normal expiration. The hip circumference is measured in a horizontal plane at the level of maximum circumference of the hips and buttocks.

Waist-to-hip ratio — Men with a WHR of 0.95 or more and women with a WHR of 0.85 or more are considered to be at increased cardiovascular risk. Similar definitions have not been developed for children, although they, too, are at increased risk of cardiovascular disease if they have abdominal adiposity [23-26]. In a study of 1001 adolescents from Spain, WHR predicted the greatest number of components of metabolic syndrome compared with waist circumference [26].

Waist circumference — Waist circumference percentile curves have been developed for Canadian [27], British [28], Italian [29], Spanish [30], and Australian children [31]. Data from the Third National Health and Nutrition Examination Survey (NHANES III) were used to estimate age-, sex-, and ethnicity-specific waist circumference percentiles for children and adolescents in the United States (table 1) [32].

The waist circumference values at different percentiles describe the existing population of children and adolescents and do not establish a standard of what "should be." Some, but not all, definitions of the metabolic syndrome in children include waist circumference (table 2), and these features do not track consistently to metabolic syndrome in adulthood. Nonetheless, waist circumference combined with BMI may serve as a useful clinical tool to estimate obesity-related comorbidities [25,33]. (See "Overview of the health consequences of obesity in children and adolescents" and "Metabolic syndrome (insulin resistance syndrome or syndrome X)", section on 'Children and adolescents'.)

Waist-to-height ratio — The waist-to-height ratio is another measure of abdominal adiposity that has been associated with cardiovascular risk. A meta-analysis evaluated the diagnostic test accuracy of waist-to-height ratio for identifying cardiovascular risk factors (eg, hyperlipidemia, hypertension, hyperglycemia, central obesity) in children 5 to 19 years of age (n = 85,281) [34]. Diagnostic accuracy of the model was high (sensitivity 0.86, specificity 0.86, area under the summary receiving operator characteristic curve 0.91). Accuracy was somewhat better for adolescents than for children <12 years and for males than for females.

Imaging — Magnetic resonance imaging (MRI) or computed tomography (CT) can be used to measure visceral adipose tissue. The technique usually quantifies adipose tissue in a single-slice cross-section at the level of the L4/L5 lumbar disc. The subcutaneous fat (outside the abdominal musculature) may be measured in the same image. These measures of visceral adiposity correlate with insulin resistance, triglycerides, hepatic steatosis, and other components of the metabolic syndrome [35,36]. The technique is used for research in obesity and metabolic disease and does not contribute to clinical care.

MEASURES OF BODY COMPOSITION — Measurement of body composition, particularly body fat, can be an important part of nutritional assessment because fat is the major source of stored energy in the body and a good indicator of nutritional state. In addition, lean body mass indicates the water and protein content, and bone density indicates the calcium and mineral stores [37]. The methods used to measure body composition differ in their ease of determination, cost, accuracy, use of radiation, and utility for estimating body fat (table 3) or other body compartments relevant to nutritional status (eg, bone density). The most commonly used techniques are described below. The results from different methods are not interchangeable [38].

Anthropometrics — The most widely used anthropometric measurements to assess body composition are the triceps skinfold (TSF) and the mid-upper arm circumference (MUAC). The TSF reflects body fat and the mid-upper arm cross-sectional area reflects lean body mass. Because the precision and reproducibility of these measures is low, they add little to the clinical evaluation for most patients [39]. For the same reason, most research studies seeking a measure of body composition will use one of the more precise measures discussed below when possible. These anthropometric measurements have not been included in the National Health and Nutrition Examination Survey (NHANES) since 1994.

Skinfold thickness — Measurements of multiple skinfold thicknesses are used widely because the technique is noninvasive, inexpensive, and easy to perform. However, measurements of skinfold thickness often are less accurate than are measurements of height or weight, particularly in obese subjects [40]. One study compared the agreement between eight widely used skinfold thickness equations and the value obtained on the basis of measurements of body density, body water, and bone mineral content [41]. The use of skinfold measurement calculations over- or underestimated individual fat mass by approximately 10 percent. There is poor correlation between the various standards that are used [42,43]. A study and a systematic review of the literature substantiate these concerns [38,39]. In contrast, in a study carried out in healthy Mexican schoolchildren 4 to 18 years of age (n = 288), skinfold thickness compared well with the four-compartment model for fat mass in the group as a whole, but the correlation varied with sex and age [38]. Measurement may be particularly compromised in conditions such as edema, obesity, and muscular or lipid dystrophies [38].

Although skinfold thickness measurements have a limited role in routine clinical care, they may be useful in the long-term monitoring of nutritional therapy in individual children who are malnourished. As an example, for disorders in which muscle mass is often diminished (eg, cerebral palsy or Rett syndrome), the suprailiac skinfold thickness measurement alone may provide a useful measure of truncal adiposity/nutritional status [44].

The TSF is measured as follows:

●The child's arm should hang freely

●The upper arm skinfold (skin and fat minus the underlying muscle) should be pulled out 1 cm above midpoint

●The calipers should be applied 1 cm in depth at the measured midpoint

●The calipers should be released, and the reading should be obtained to the nearest 1 mm as soon as the needle is steady

●The average of duplicate measurements should be recorded [45]

Mid-upper arm measurements

●Mid-upper arm circumference (MUAC) – MUAC is the circumference of the upper arm, measured at the midpoint. A MUAC of <115 mm is used as a threshold for diagnosing malnutrition in children 6 to 59 months of age.

The MUAC is closely correlated with muscle mass in children, although it does not distinguish between muscle and subcutaneous fat. It is a clinically practical tool for evaluating wasting in a malnourished population, especially in resource-limited settings in which time, equipment, and trained personnel are limited. (See "Malnutrition in children in resource-limited settings: Clinical assessment", section on 'Mid-upper arm circumference'.)

●Mid-arm muscle circumference (MAMC) – The MAMC is calculated from the MUAC and triceps skinfold (TSF) as follows:

Because MAMC incorporates TSF as a measure of subcutaneous fat, it may be a more accurate measure of muscle mass than MUAC, especially in children who are not malnourished. MAMC measurements can be used to monitor changes in body composition in children who are critically ill or when other measures cannot be obtained [46,47].

Dual-energy x-ray absorptiometry — The above methods inform the two-compartment model of body composition. In contrast, DXA estimates fat-free mass, body fat, and bone mineral density by using the differential absorption of x-ray or photon beams of two levels of intensity. DXA relies on the principle that the intensity of an x-ray or photon beam is altered by the thickness, density, and chemical composition of an object in its path [48]. In children, the scan takes approximately 10 minutes. The average radiation dose, depending on the instrument and body size, is 0.04 to 0.86 mrem, less than the average exposure of a chest radiograph [49]. The precision of DXA (coefficient of variation) is less than 2 percent [50]. DXA has been used to measure body composition in children and is becoming more readily available in the clinical setting [1-3]. The technique is limited because estimates of fat mass become less accurate as the individual's trunk thickness increases and it appears more accurate at a group level, compared with the individual level [51,52]. That said, a systematic review and study support its ability to accurately and reproducibly measure fat-free mass [38,39]. The figures show reference curves for bone mineral density (figure 7) and bone mineral content (figure 8) [3]. (See "Overview of dual-energy x-ray absorptiometry", section on 'Children'.)

Quantitative magnetic resonance — The use of quantitative magnetic resonance (QMR) has been increasingly adopted to assess body composition in human studies. It has been shown to be a valid and precise method for a noninvasive measurement of body composition. The QMR measurement does not use ionizing radiation, and results are independent of fat-free mass hydration. A study in young children demonstrated reasonably good correlation with results of air-displacement plethysmography [53]. However, it appears to underestimate fat mass in individuals with greater degrees of adiposity [38]. The routine use of this technique is limited to research due to its expense.

Air-displacement plethysmography — Air-displacement plethysmography is a method for determining body volume and hence, fat mass. Using the same whole-body measurement principle as underwater weighing (see 'Hydrodensitometry' below), air-displacement plethysmography measures a subject's mass and volume, from which the whole-body density can be determined. The subject is placed inside the device; computerized pressure sensors then determine the amount of air displaced by the subject's body. Body fat and lean muscle mass can then be calculated [54].

Compared with the four-compartment model (fat, fat-free mass, total body water, and bone), the accuracy and precision of air-displacement plethysmography are good in school-aged children (± 2 percent for fat mass), providing that child-specific equations for lung volume are used [55-57]. In full-term infants up to 10 kg, the method has modest accuracy but good reproducibility [56,58-60]. There are fewer data in preterm infants, but evidence suggests similar results [61].

Infrequently used techniques

Hydrodensitometry — Underwater weighing is the oldest method for determining body density; it has been largely replaced by dual-energy x-ray absorptiometry (DXA) (see 'Dual-energy x-ray absorptiometry' above). Body density is estimated from the weight of the subject in air and water, using appropriate correction factors for temperature, air in the respiratory tract, and constants for the densities of fat and fat-free tissue [62]. Hydrodensitometry can be performed only in children who can hold their breath for a period of time and have no contraindication to complete underwater submersion. In addition, the technique is not reliable in younger children because of the change in density of lean body mass during maturation [63]. (See "Normal puberty".)

Isotope dilution — In isotope dilution techniques, the volume of a body compartment is determined by the ratio of the dose of a tracer, administered orally or intravenously, to its concentration in the body compartment after a sufficient equilibration period has elapsed. Total body water, as an example, is estimated by isotope dilution using the stable isotopes of deuterium (2H₂O) or oxygen (H₂¹⁸O) [64]. Once total body water has been determined, the fat-free mass of the body can be calculated, based upon the assumption that the hydration of lean tissue is constant across a broad range of ages and for both genders. Potential errors can occur when hydrogen isotopes are used because they exchange with nonaqueous hydrogen resulting in overestimation of total body water by 5 to 10 percent. Isotope dilution is impractical in the clinical setting because of the expense of the stable isotopes and the analytical equipment.

Bioelectrical impedance analysis — Bioelectrical impedance analysis (BIA) relies upon an electrical current to measure the fat and fat-free mass of the body [65,66]. The lean tissues of the body, because of their dissolved electrolytes, are the major conductors of electrical current, whereas body fat and bone are relatively poor conductors. Thus, BIA primarily measures total body water from which an estimate of fat-free mass is obtained. BIA frequently is used because it is noninvasive, portable, and inexpensive. However, small changes in body water result in large differences in the estimate of fat-free mass. Placement of the electrodes is site-specific and can compound errors of measurement. The placement of the electrodes (hand-to–foot, hand-to-foot, or foot-to-foot), will affect the capacity to predict body fat percentage. In addition, BIA is influenced by sex, age [67], disease state, and level of fatness (because total body water and relative extracellular water are greater in obese individuals) [49]. The data suggest BIA remains problematic in its ability to assess percent body fat, fat mass, or fat-free mass [68] and is less reliable than DXA [39,69]. The same appears to hold true in newborns and preterm infants [61].

Other — The following techniques are occasionally used in research:

●Neutron activation analysis – Neutron activation analysis allows the direct elemental analysis of the body. Virtually all of the major elements of the body can be assayed in vivo: hydrogen, oxygen, carbon, nitrogen, calcium, phosphorus, sodium, chlorine, and potassium.

The method uses the principle that exposure to a given dose of neutrons generates a known amount of radioactivity within a given type of tissue [70]. When an atom captures a neutron, the atom is transformed to another nuclear state of the same chemical element. The new atom can be stable or radioactive, but it will have excess energy that must be released. If the new atom is radioactive, it will decay over time with a known half-life. Thus, when the body is exposed to neutrons, gamma rays are emitted immediately (prompt) and for some time thereafter (delayed); these gamma rays can be measured in a counting chamber. The particular element being examined can be identified by the characteristic energy of the electromagnetic radiation it emits and its rate of decay. The use of this technique is limited by its expense and radiation exposure.

●Total body potassium – Total body potassium counting is another method that determines elemental composition. It measures the content of radioactive potassium (40K) in the body while the child lies between detectors in a well-shielded metal counting chamber [71]. The radioisotope emits a characteristic high-energy gamma ray (1.46 MeV) that, in turn, serves as a marker for total body potassium, body cell mass, and fat-free mass. Although this method is noninvasive and safe, the limited availability of the instrument, particularly one with sufficient sensitivity to measure small infants, renders this method impractical in the clinical setting.

Emerging technologies — Emerging technologies to measure body composition include:

●Three-dimensional optical scanners can be used to create surface renderings of the body in three-dimensional space [72]. Anthropometry measures including circumferences, lengths, volumes, and surface areas can be generated from the scans. The automated measurements have been combined and used to derive and validate body composition equations for children [73].

●A D3-creatine dilution method for measurement of the total body creatine pool size can serve as a noninvasive index of muscle mass [74]. Validation studies are needed for infants and children [72].

●Bioimpedance spectroscopy determines total body water and extracellular water volume [75]. Three body composition compartments then can be derived from the measurements.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Healthy diet in children and adolescents".)

SUMMARY AND RECOMMENDATIONS

●Body composition – The choice of method for measuring body composition depends on whether the measurement is for clinical purposes or research and what degree of precision is required.

•Indirect measures – Body mass index (BMI) is a clinically practical way to estimate adiposity. It is sufficiently correlated with adiposity to provide a clinical estimate of weight-related medical risk, and it is noninvasive and inexpensive. However, it is also imprecise and the same BMI percentile does not represent the same percentage body fat at different ages or stages of pubertal maturation, for males and females, or among children with different muscle mass. (See 'Body mass index' above.)

•Direct measures – Body composition can be measured directly using anthropometrics, dual-energy x-ray absorptiometry (DXA) and air-displacement plethysmography. These measures vary considerably in their precision and expense; they generally are not valuable for the routine clinical evaluation of a patient. For the purposes of research studies requiring a precise measure of body composition, DXA or air-displacement plethysmography typically are used. (See 'Measures of body composition' above.)

●Fat distribution – Fat distribution can be estimated by waist-to-hip ratio (WHR) and waist circumference, measures that are associated with weight-related medical risks in adults. The association is less clear for children and adolescents. The primary utility of these measures is for research rather than for clinical care. (See 'Fat distribution' above.)