Resumen:Background: The results of multivariate analyses to identify potential predictors of response to fluoxetine or placebo separately in 96 child and adolescent outpatients with major depressive disorder from a recent controlled trial are presented. Methods: A variety of clinical, demographic and laboratory factors were examined as possible predictors of response to fluoxetine or placebo using logistic regression models. Results: No single variable or combination of variables strongly predicted response to fluoxetine. For the placebo group, a younger age, a shorter duration of depressive episode, and a lower socioeconomic status predicted response with an overall predictive power of 81%. Conclusions: This study is limited by the small sample size and should be considered hypothesis generating rather than confirming.