Primary objective: The study evaluated the accuracy of body mass index (BMI) in detecting hyperinsulinaemia during oral glucose tolerance testing (OGTT) in severely obese children.
Research design: A cross-sectional study was carried out.
Materials and methods: A total of 118 obese children and adolescents (49 females and 69 males) aged 6-19 years were consecutively studied at an outpatient paediatric clinic. Hyperinsulinaemia was defined as a value of log-transformed fasting insulin >/= 80th percentile and OGTT hyperinsulinaemia as a value of the log-transformed area under the curve (AUC) of insulin >/= 80th percentile. The study hypothesis was tested using a logistic regression model with hyperinsulinaemia as the outcome variable and the z-score of BMI corrected for age (z-BMI(age)) as the predictor variable. Receiver-operator characteristic (ROC) curves were used to evaluate accuracy.
Results: The mean (SD) BMI for age of the children was 28.6 (4.0) kg m(-2), corresponding to 2.2 (0.5) standard deviation scores. The odds ratio (OR) of OGTT hyperinsulinaemia was 2.0 (95% CI 1.2-3.3; p = 0.007) for each unit increase of z-BMI(age) and the corresponding ROC-AUC was 0.74 (95% CI 0.61-0.86; p = 0.0001). In comparison, the OR of fasting hyperinsulinaemia was 1.1 (95% CI 0.7-1.7; p = 0.716) for each unit increase of z-BMI(age) and the corresponding ROC-AUC was 0.49 (95% CI 0.35-0.62; p = 0.863).
Conclusion: BMI is reasonably accurate in detecting OGTT hyperinsulinaemia in severely obese children.