ESPE Abstracts

Background: There is limited information on the ability of demographic or lifestyle factors to predict short term changes in weight status during childhood.

Objective and hypotheses: To study the effect of parental (educational level, BMI status and perception of child’s weight status) and childhood factors (eating habits, sedentary behaviour and physical activity), on 1-year BMI increments by the use of BMI, BMI SDS and BMI SDS conditional gain.

Method: With data from the Bergen Growth Study the relations between demographic and lifestyle factors and one-year BMI increments were explored. Each of the three BMI measures (changes in BMI and BMI SDS, and BMI conditional gain) was analysed separately as a dependent variable with linear regression models. Adjusted regression models were estimated for each BMI measure separately, including all the statistically significant variables from the unadjusted models.

Results: In the unadjusted models, 1-year changes in BMI were correlated to maternal BMI, parental perception, irregular meals and screen time. Changes in BMI SDS were only correlated to irregular meals and screen time. Changes in BMI SDS conditional gain were correlated to maternal BMI, parental perception and irregular meals. In the fully adjusted model raw BMI increments were correlated to parental perception, irregular meals and screen time, BMI SDS increments to irregular meals and BMI SDS conditional gain to parental perception and irregular meals.

Conclusion: Parental perception of child’s weight status, irregular meals and screen time can predict higher 1-year BMI increments. BMI SDS conditional gain adjusts for regression towards the mean, and might therefore be the preferred measure.