Predictive Models for Calculating Body Weight and Assessing the Likelihood of Overweight and Obesity in Young Children

Abstract

Purpose. To develop mathematical models for calculating body weight (BW) and assessing the likelihood of overweight and obesity in young children based on anamnestic parameters, clinical characteristics and dynamic anthropometric indicators. Materials and methods. A prospective examination of 305 children from birth to two years of age was carried out: 202 children were analyzed for the development of mathematical models and 103 – for the subsequent control of the correct operation of the models. The main group (Gr1) consisted of 101 children with large BW at birth, among which group 1A (Gr1A) – 54 patients with obesity or excess BW, group 1B (Gr1B) – 47 children with normal BW at the age of two. Among normal birth children of the comparison group (Gr2), group 2A (Gr2A) – 15 two-year-old children with excess BW, group 2B (Gr2B) – 86 children with normal BW. The analysis of anamnestic data, clinical and laboratory parameters, direct and derived anthropometric parameters of children at birth, at the age of 1, 3, 6, 12 and 24 months of life was carried out. Results. A prognostic model has been developed with high level of evidence (p<0.001) and sufficient accuracy to determine the BW of young children, taking into account the category of physical development and the value of the coefficient of harmony at birth, dynamic anthropometric characteristics at 1 and 12 months of life, the period of administration the first complementary foods and the genetic growth of the child. The multivariate risk model for overweight and obesity in two-year-old children born with large weights includes: Z-score of BW at birth over 1.82, at 12 months over 1.92, introduction of first complementary foods before 5 months of life. The developed model and classification scheme are characterized by good predictive accuracy (80.9% (95% CI 72.4–89.5%), p<0.001). The results of external validation testify to the high power of the developed prognostic model and classification scheme for the probability of overweight and obesity in two-year-old children born large for gestational age. Conclusion. A comprehensive assessment of the totality of ante- and dynamic postnatal predictors provide s high evidence of the model for calculating BW, accuracy and reproducibility of predicting overweight and obesity in young children.