Objective: To fully explain the dynamic and comprehensive etiology of the trajectory associated with adiposity indices. Methods: This study involved data of 5572 children, aged 6–11 years, as part of the Taiwan Children Health Study (TCHS). The present study introduced four distinct BMI trajectories, identified previously among children: persistently healthy weight; late-onset overweight or obesity; persistent overweight or obesity; and declining BMI class. Logistic regression was used to examine the effect of non-modifiable factors on BMI trajectory classes. Generalized estimating equations were used to examine the effect of dynamically modifiable factors on either BMI trajectory classes or adiposity indices. Results: Compared with class 1 (persistently healthy weight), class 2 exhibited a significantly increased risk of weight gain and fat mass, affected by lower family incomes and poor-quality sleep. Class 3 had a higher risk of persistent obesity and abdominal obesity, affected by higher birthweight and sedentary lifestyle. Class 4 approached a healthy weight due to increased physical activity, which was associated with a decrease in body fat and central obesity. Conclusions: We found crucially non-modifiable and modifiable factors that could describe each high BMI growth pattern, and calculated their modifiable contributions to adiposity indices. Modifiable factors that focus on those crucially dynamic factors are recommended for preventing obese growth trajectories.

Original languageEnglish
Pages (from-to)270-276
Number of pages7
JournalObesity Research and Clinical Practice
Issue number3
Publication statusPublished - May 1 2018


  • Children
  • Genetic determinant
  • Modifiable factors
  • Obesity index
  • Trajectory

ASJC Scopus subject areas

  • Endocrinology, Diabetes and Metabolism
  • Nutrition and Dietetics


Dive into the research topics of 'Comprehensive determinants of growth trajectories and body composition in school children: A longitudinal cohort study'. Together they form a unique fingerprint.

Cite this