Individualised prognosis for risk of developing abdominal obesity in the paediatric population.

ABSTRACT

AIMS: To develop and validate risk scores for predicting abdominal obesity in South American children and adolescents based on extrinsic and intrinsic variables. METHODS: Children (n = 358) and adolescents (n = 369) from seven South American cities from the South American Youth Cardiovascular and Environmental (SAYCARE) Study. The primary outcome was abdominal obesity. Potential predictors were based on sociodemographic, maternal, environmental, and behavioural factors and nutritional status. In multilevel logistic models, associated variables were tested to build the scores, which were internally validated. RESULTS: We identified 120 children and 98 adolescents who were abdominally obese. We found at least five variables associated with the outcome in children with unacceptable predictive capacity. However, in adolescents, we found that biological sex, age, maternal body mass index (BMI), active commuting by bike, soft drink consumption (for risk score A), and weight (for score B) can predict abdominal obesity. Both scores, A and B, showed acceptable performance in the ROC curve [areas under curve 0.70 (95% CI 0.56-0.82) and 0.95 (95% CI 0.89-1.00), respectively]. CONCLUSION: The SAYCARE risk scores present accurate, individualised estimates for identifying adolescents who are at risk of developing abdominal obesity. However, these have not been externally validated.