Prediction of asthma by common risk factors: a follow-up study in Cuban schoolchildren.

ABSTRACT

OBJECTIVE: To determine which common risk factors, including environmental factors, are predictors for the development of asthma in Cuban schoolchildren. METHODS: A longitudinal study was conducted in 1042 schoolchildren without asthma at baseline in 2 Cuban municipalities. Asthma status in 2007, diagnosed using the International Study of Asthma and Allergies in Childhood questionnaire, was related to a set of common risk factors assessed in 2003/2004 in a multivariable logistic regression model. Multiple imputation was used for missing values. The final prediction model was obtained by backward selection (P<.15). The model's prognostic accuracy (R2) and discriminative ability (area under the receiver operating characteristic curve [AUC]) were assessed and internal validation by bootstrapping was performed. RESULTS: A family history of atopic diseases (odds ratio [OR], 2.19; 95% CI, 1.19-4.04), allergic sensitization (OR, 1.83; 95% CI, 0.94-3.55), municipality (OR, 0.34; 95% CI, 0.15-0.74), and use of antibiotics in the child's first year of life (OR, 1.66; 95% CI, 0.89-3.11) were predictors for asthma development. The model had an R2 of 8.0% and a moderate discriminative ability (AUC, 0.69; 95% CI, 0.60-0.78). Internal validation hardly influenced the model's performance. CONCLUSIONS: Antibiotics use, genetic predisposition, and allergic sensitization were predictors of asthma in Cuban schoolchildren. Although known as common risk factors they could only partly predict asthma development. Poverty-related factors, such as low income and education, and parasitic infections, did not have an effect. Other or additional environmental predictors need to be identified, as these are potential targets for prevention and control of childhood asthma in affluent as well as nonaffluent countries.