The utility of body mass index as an indicator for lipid abnormalities in non-fasting children

Background: Many studies have reported the association between body mass index (BMI) and fasting lipid profiles in children. However, little information exists about the screening of dyslipidemia in the non-fasted state. This study assessed whether BMI can predict non-fasting lipid abnormalities in children. Methods: Using gender-separated analysis, 3895 boys and 3866 girls (aged 11-12 years) were investigated. Total cholesterol (TC), triglyceride (TG) and HDL-cholesterol (HDL-C) were measured, and non-HDL-C (=TC-[HDL-C]) was calculated. A BMI z-score was employed as the weight status. Gender-specific 95th percentiles of TC, TG and non-HDL-C were defined as "elevated", with the 5th percentiles of HDL-C defined as "reduced". Results: TG and non-HDL-C were positively, and HDL-C was negatively correlated with the BMI z-score in both genders. Both obese (2 < BMI z-score) and overweight (1 < BMI z-score <= 2) were associated with the risks of elevated TG, non-HDL-C and reduced HDL-C. In both genders, a receiver operating characteristic curve demonstrated that the utility of predicting the above lipid abnormalities was moderate; the areas under the curve ranged from 0.60 to 0.70. The optimal cut-off for the BMI z-score for predicting elevated TG, non-HDL-C and reduced HDL-C were 0.52, 0.55 and 0.51 in boys and 0.34, 0.38 and 0.35 in girls, respectively. Conclusions: The BMI could be an indicator of abnormalities of non-fasting TG, non-HDL-C and HDL-C in both genders.