ESPE Abstracts (2014) 82 P-D-1-3-133

Can We Predict the Risk of Obesity?

Antonio de Arribaa, Javier Pascualb, Mario de Franciscob, Marta Ferrera, Ignacio Martínezb, José Ignacio Labartaa & Ángel Ferrándezc

aPediatric Endocrinology Unit, Miguel Servet Hospital, Zaragoza, Spain; bGodday Solutions SL, Zaragoza, Spain; cAndrea Prader Foundation, Zaragoza, Spain

Background: Recognizing the risk of developing obesity is essential to implement preventive measures to avoid the increasing prevalence of obesity in adulthood.

Objective: To evaluate predictive factors that may be associated with overweight and obesity in early adulthood.

Method: A regression analysis of different variables of body composition in a normal population have been done. The sample consisted of 122 boys and 120 girls followed longitudinally from birth to adulthood. Simple data and their variations in different periods of time have been used (from birth to 3 years measured every 3 months). Variables studied: weight, height (H)/length, sitting height (SH), BMI, waist circumference, waist to hip (WH), skin folds (tricipital (TS) and subscapular (SS)), cephalic perimeter (CP), arm perimeter (AP), thigh circumference (TC), biacromial diameter (BD) and biiliac diameter. Overweight threshold has been established for each time period depending on the values of BMI and waist circumference of the total sample when both parameters were >+1 S.D.

Results: Two different predictive models have been obtained (for boys and girls) that can predict at 3 years of age the risk of of being overweight in early adolescence.


P(overweight)=1/(1+e(25,4662×BMI319,1074× BMI2_753,3415xSH1_510,9374xWH0+2,6447× TS0_75d025+1,6515×BD0_75d075+2,5140×BD2)))


P(overweight)=1/(1+e((27,3543+27,9645×BD0_75d 025:BD3d02523,7834×WH0_5d025:SH1d025+ 4,8652×BD1d1+12,2667×TC2+75d0257,8747× BD0_5d02534,3391×SS2_25d025+48,9172×SS2_25d 025:TC2_75d0250,8233×TC1_5d13,0773× BD0_75d0252,5331×SH1×_25d025)))

The variables with the greatest impact on risk estimation were BMI index at 3 years of age in girls, and the interaction of the variation presented by the subscapular skinfold between 2 and 2.25 years of age associated with the difference between thigh circumference measured at 2.5 and 2.75 years of age in boys. These formulas have a sensitivity of 92.13% for females and 91.86% for males and a specificity of 91.3% for females and 92.59% for males.

Conclusions: Longitudinal studies provide great information about the growth and development of populations, which allow to design predictive models of developing overweight/obesity in adulthood. 2. A formula for each sex has been achieved with high sensitivity and specificity that can predict the risk of being overweight from 3 years old. However, these models are of great complexity that limits its application in the daily practice and therefore easier formulas must be obtained.

Article tools

My recent searches

No recent searches.

My recently viewed abstracts