Pediatric body composition analysis with dual-energy X-ray absorptiometry

Abstract

Pediatric applications of body composition analysis (BCA) have become of increased interest to pediatricians and other specialists. With the increasing prevalence of morbid obesity and with an increased awareness of anorexia nervosa, pediatric specialists are utilizing BCA data to help identify, treat, and prevent these conditions. Dual-energy X-ray absorptiometry (DXA) can be used to determine the fat mass (FM) and lean tissue mass (LTM), as well as bone mineral content (BMC). Among the readily available BCA techniques, DXA is the most widely used and it has the additional benefit of precisely quantifying regional FM and LTM. This review evaluates the strengths and limitations of DXA as a pediatric BCA method and considers the utilization of DXA to identify trends and variations in FM and LTM measurements in obese and anorexic children.

Appendix

BCA techniques have been grouped as either predictive or compartment models [107]. Predictive models make a direct measurement, such as subcutaneous skinfold anthropometry or the body’s bioelectrical impedance to a small electrical current, and then predict body composition using reference data, algorithms and underlying assumptions. Compartment models measure one or more components of body composition, such as total body water (a surrogate for fat-free mass) and then predict the other body components using separate reference data, algorithms and underlying assumptions.

SFT measurements are easy to perform on children but lack precision and have been found to be a poor predictor of FM by many investigators [107, 108]. This is especially true in obese individuals and in the setting of subcutaneous edema. Accuracy can be improved with better mathematical prediction models. The sheer number of such models (more than 100 published) indicates the complexity of the relationships between body fat and SFT measurements and suggests the need for appropriately matching patient parameters to those of the control group upon which the modeling was based.

The flow of an electric current through the body is proportional to its water content. With the assumption that the body is a cylinder and using appropriate calibration models, total body water can be predicted by BIA. Additional assumptions regarding the water content of lean and fat tissues allow BIA to estimate LTM and FM [107, 108]. The technique is easy to perform and requires little operator training or patient cooperation. However, it has significant limitations that include poor accuracy, imprecision caused by variations in tissue hydration and numerous calibration models based on normal data that might not be appropriate for an individual patient [109].

Compartment models tend to be more complicated and invasive than predictive techniques because up to four body compartments (fat, lean, bone mineral and water) are determined separately. Total body mineral content can be measured with DXA. With the assumption that there is no water in fat and that there is a constant level of hydration and density of lean tissues, techniques that measure total body water can be used to estimate LTM and FM given a patient’s weight. DXA is a two-compartment technique initially developed to use differential absorption of X-rays to distinguish bone mineral from overlying soft tissues. Using the above-mentioned assumptions allows the DXA absorption data to estimate fat and fat-free tissue mass. Three- and four-compartment models require a measurement of total body water [107, 110]. Following the injection of a tracer agent of known quantity and biodistribution, the measured concentration of the agent in a body fluid, such as urine or saliva, can be used to determine total body water. Given the patient's weight and measurement of the patient’s body volume, body density can be determined using assumptions about tissue hydration, body fat and lean tissue. Body volume can be measured using displacement techniques such as underwater weighing or air plethysmography. An underlying assumption of compartment models is that the composition of lean tissue is constant for a given patient age and gender. Tissue composition is fairly constant in healthy individuals but varies with age and particularly with pubertal status and gender. This assumption might not hold for patients with altered hydration status or more severe metabolic derangements.

Alternative imaging techniques for BCA include CT and MRI. CT can provide accurate assessment of subcutaneous and visceral fat within the area imaged but is expensive, requires fairly high radiation doses and does not provide total body assessment. MRI provides highly detailed images of fat and lean tissue, from which fat volume can be derived. To be converted to mass values, assumptions regarding levels of hydration and density of lean and fat tissues are made. Because of this, MRI-derived BCA data are not directly comparable to data derived from the above-mentioned techniques [108]. The technique is relatively expensive and many children require sedation.

All of these techniques have advantages and disadvantages [108]. SFT is easy to perform but is imprecise and particularly inaccurate in obese individuals and provides no direct information regarding LTM. BIA lacks sufficient accuracy for individual patients, though it may be suitable to population studies and its results are particularly sensitive to hydration status and changes in body weight. Dilution and displacement techniques can yield accurate results and, when combined with DXA determination of total body mineral, yield a four-compartment technique that is considered the gold standard against which other techniques are measured. They are not suitable for field work, however. Traditional methods for determination of body volume are not feasible for most children. DXA is highly reproducible, easy to perform and uses minimum radiation. Its accuracy can be improved with further refinement of the algorithms used for its body composition modeling. It is not able to differentiate muscle from other lean tissues, such as liver, spleen and other organ tissue, nor can it distinguish adipose tissue from bone marrow fat or fat within solid viscera. Regional FM analysis with DXA does not give a reliable assessment of visceral fat. However, it is widely available and has the largest body of research and clinical data associated with it. It is highly reproducible and accurate and can provide complementary body composition data for patients who require DXA for assessment of BMD and bone mineral content.