Body mass index trend as a new parameter for evaluating children's nutritional status

Objectives: The objective was to identify a body mass index (BMI) trend as a new indicator and predictor of children's nutritional status, replacing absolute BMI, and to demonstrate that a local BMI trend is more appropriate for this purpose than a national (or international) BMI trend.

Study design: An entire school year group of primary school children were subject to a longitudinal 8-year cohort study.

Methods: BMI was measured three times during primary school education-on enrolment to primary school, then in the fifth and eighth grades. The BMI values obtained were used to create gender- and age-based percentile curves for children born in 1998/1999 (Osijek Percentile Curves (OPC) study) in the Osijek-Baranja County, Croatia. Every BMI result obtained was linked to one of the five percentile ranks (PRs) according to threshold percentiles 3, 10, 90 and 97; hence, the PRs were <3, ≥3 to <10, ≥10 to <90, ≥90 to <97 and ≥ 97. Percentiles and PRs were specified for each BMI value in a dual fashion, i.e. according to OPC and the entire Croatian territory (Croatian Percentile Curves (CROPC)-2008) cross-sectional study. After integrating the BMI values at all three time points, two BMI trends appeared-one based on the OPC criteria and one based on the CROPC-2008 criteria. Individual trends were grouped into three descriptive BMI categories: 'healthy', 'at risk' and 'unhealthy'.

Results: For the OPC and CROPC-2008, the 'healthy' trend characterised 2097 (78.72%) and 1975 (74.14%) children, respectively; the 'at risk' trend marked 434 (16.29%) and 458 (17.19%) children, respectively; and the 'unhealthy' trend was found in 133 (4.99%) and 231 (8.67%) children, respectively. The biggest difference between the OPC-based and CROPC-2008-based results was detected in the 'unhealthy' trend. According to CROPC-2008, there were almost twice as many children in that category. For the purpose of study design and comparing results obtained based on OPC and those obtained based on CROPC-2008, the BMI of one boy was used. His BMI values on starting primary school, in the 5th and in the 8th grades were 19.05 kg/m², 26.11 kg/m² and 27.11 kg/m², respectively. Analysis of PRs based on OPC and CROPC-2008 reveals that the same boy was represented by different trends: '<90 <90 <97' and '<90 <97 <97', respectively. In terms of residence, girls from urban areas have higher BMI values than rural girls when starting primary school.

Conclusion: This study suggests that a BMI trend could be a better indicator and predictor of children's nutritional status than absolute BMI. In addition, a cohort study is preferential to a cross-sectional study as a trend in an OPC-based study can be personalised. Also, local BMI reference values defined for each generation of children are more advantageous in this context than national BMI reference values.