Med J 29, 2006). Little is known on the phenomenology and potential short-term prognosis factors of bipolar episodes in this age population. In particular, very few studies examine this issue on inpatients in the European context of free access to care. Objective To describe the phenomenology of acute manic and mixed episodes in hospitalized adolescents and to analyse potential predictive factors associated with clinical improvement at discharge and length of hospitalization. Methods A total of 80 subjects, aged 12-20 years, consecutively hospitalized for a manic or mixed episode. Sociodemographic and clinical data were extracted by reviewing patients' charts. We used a multivariate analysis to evaluate shortterm outcome predictors. Results The sample was characterized by severe impairment, high rates of psychotic features (N = 50, 62.5%), a long duration of stay (mean 80.4 days), and an overall good improvement (86% very much or much improved). Thirtythree (41.3 %) patients had a history of depressive episodes, 13 (16.3%) had manic or brief psychotic episodes but only 3 (3.7%) had a history of attention deficit/ hyperactivity disorders. More manic episodes than mixed episodes were identified in subjects with mental retardation (MR) and in subjects from migrant and/or low socio-economic families. Overall severity and female gender predicted better improvement in GAF scores. Poor insight and the existence of psychotic features predicted longer duration of stay. Conclusion These results suggest that severe manic and mixed episodes in adolescents with BD-I need prolonged inpatient care to improve and that socio-cultural factors and MR should be examined more closely in youth with BD.j Key words bipolar disorder type I -acute episodeadolescent -prognosissocio-cultural factors BRIEF REPORT