Skip to main content
ARS Home » Plains Area » Houston, Texas » Children's Nutrition Research Center » Research » Publications at this Location » Publication #105957
Title: PREDICTION MODELS FOR BONE MINERAL CONTENT IN SCHOOL-AGE CHILDREN

Author
item Ellis, Kenneth
item SHYPAILO, ROMAN - BAYLOR COLL OF MEDICINE
item Wong, William
item Abrams, Steven

Submitted to: New York Academy of Sciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/9/1999
Publication Date: 5/1/2000
Citation: N/A

Interpretive Summary: When measuring bone mineral content and density in children, a major limitation of dual-energy X-ray absorptiometry (DXA) is the lack of appropriate reference standards for growth. Hence the usefulness of DXA for assessing bone status in children remains controversial. Therefore, we developed a prediction model for total body bone mineral content in children based on age, gender, ethnicity, body size, and body composition. We performed a cross-sectional study of 1571 healthy Caucasian, African- American and Mexican-American children of both sexes, 5 to 18 years of age. We used multiple regression analysis to derive anthropometric models first and then additional models that include body composition data for lean tissue and fat mass. Both types of models were developed because abnormal bone growth can be accompanied by altered muscle development. These models should be very useful when comparing healthy children from other studies or rethnic groups. In addition, when clinicians need to evaluate the bone status of children of diseases that may involve abnormal bone growth, these references will provide the expected normal values for gender, age, and body size. The clinician should use caution in the interpretation of the body composition prediction models because many illnesses involving abnormal bone mineralization often also have altered lean tissue mass development.

Technical Abstract: In pediatrics, one major limitation of dual-energy x-ray absorptiometry (DXA) for the assessment of bone in children is the lack of adequate reference standards. Our aim, therefore, was to develop an anthropometric- based (ethnicity, gender, age, weight, and height) prediction model for bone mineral content (BMC) in children. We examined 1571 healthy children (446 males, 1125 females) from three ethnic groups (723 European-, 433 African-, and 415 Mexican-American). Whole-body BMC, bone area (BA), and lean tissue mass (LTM) were obtained using a Hologic QDR-2000 instrument. Using step-wise multiple regression analysis, the best anthropometric prediction equation (r2 = 0.936, SEE = 160 g) was: BMC = - 503 + [ (6.98 + 1.75 x Sex) x 10-3] x Ht2.31 + 16.4 x Wt + 37.7 x Age - 77.3 x Sex - 45.8 x Race. If LTM and BA are included, the equation with the lowest SEE was: ln(BMC) = (3.6 - 0.15 x Sex) x ln(BA) + 0.008 x LTM - 0.0015 x %Fat + 0.9 x xSex - 2.01, with r2 = 0.98 and SEE = 0.066. When the prediction models were used to assess BMC status in children with various diseases, their values were often found below the normal range. Additional longitudinal studies are needed to test the prediction accuracy of these prediction models for bone growth in children receiving therapy.