Development of a prediction equation for insulin sensitivity from anthropometry and fasting insulin in prepubertal and early pubertal children

Academic Article


  • OBJECTIVE - To test the utility of homeostasis model assessment (HOMA) in predicting insulin sensitivity [X10-4 min-1/(μIU/ml)] in children and to develop and compare two new prediction equations for insulin sensitivity in children using demographic and anthropometric measures in the presence or absence of fasting insulin. RESEARCH DESIGN AND METHODS - We studied 156 white and African-American children with complete data (mean age 9.7 ± 1.8 years, 87.8% Tanner Stage 1 or 2). For development of new equations, two-thirds of the children were randomly assigned to a development group, whereas the remaining children were assigned to a cross-validation group. RESULTS - A modified HOMA equation accurately predicted insulin sensitivity, but its utility is similar to fasting insulin alone. Demographic and anthropometric measures alone did not predict insulin sensitivity accurately, even when precise measures of body composition were included in the prediction model. Ethnicity, calf skinfold, and fasting insulin together explained 73% of the variance in insulin sensitivity and accurately predicted insulin sensitivity. The regression of measured versus predicted insulin sensitivity in the cross-validation group was not significantly different from the line of identity (P > 0.05). Mean difference between measured and predicted insulin sensitivity was also not significant (P > 0.05). Some bias was apparent, particularly in white boys. CONCLUSIONS - Ethnicity, calf skinfold, and fasting insulin can accurately predict insulin sensitivity with greater precision than HOMA or fasting insulin alone (R2 = 0.73). Future studies, however, are needed to examine whether a universal equation is possible. A cross-validated prediction equation may be useful in population-based studies when complex measures of insulin sensitivity are not available.
  • Authors

    Published In

    Digital Object Identifier (doi)

    Pubmed Id

  • 28755081
  • Author List

  • Huang TTK; Johnson MS; Goran MI
  • Start Page

  • 1203
  • End Page

  • 1210
  • Volume

  • 25
  • Issue

  • 7