INTRODUCTION: As one of the most promising biomarkers in depression course and treatment, the brain-derived neurotrophic factor (BDNF) has been identified as a possible predictor of antidepressant treatment outcomes in the adult population. However, studies on this subject in the developmental population are lacking. We aimed to verify the changes in serum BDNF, proBDNF, and BDNF/proBDNF ratio levels during antidepressant treatment in adolescents diagnosed with depressive episodes. We also investigated these parameters as predictors of antidepressant treatment outcomes. MATERIAL AND METHODS: Thirty female inpatients, aged 11-17, diagnosed with a first-lifetime depressive episode were assessed at two time-points: before (t0) and after (t1) the period of antidepressant treatment and compared with thirty age-matched healthy girls. The assessment at t0 and t1 involved BDNF and proBDNF serum levels (analyzed with the ELISA method) and standardized depressive symptoms scales. RESULTS: BDNF serum levels decreased significantly with the antidepressant treatment in the studied group (p = 0.0224). This effect was still present in the subgroup of responders (p = 0.0179) but not among non-responders (p = 0.2184). Responders had a significantly higher initial BDNF/proBDNF ratio than non-responders. The BDNF/proBDNF ratio at t0 could predict the remission status at t1 with a sensitivity of 66.67 % and a specificity of 81.25 % (p = 0.0327). CONCLUSIONS: BDNF serum levels tend to decrease with successful antidepressant treatment in adolescent girls treated for the first episode of depression. Pretreatment BDNF/proBDNF ratio should be considered a possible biomarker predictive of antidepressant treatment response in adolescent girls.