Background There is great heterogeneity of clinical presentation and outcome in paediatric depression. Alms To identify which clinical and environmental risk factors at baseline and during treatment predicted major depression at 28-week follow-up in a sample of adolescents with depression. Method One hundred and ninety-two British adolescents with unipolar major depression were enrolled in a randomised controlled trial (the Adolescent Depression Antidepressants and Psychotherapy Trial, ADAPT). Participants were treated for 28 weeks with routine psychosocial care and selective serotonin reuptake inhibitors (SSRls), with half also receiving cognitive-behavioural therapy (CBT). Full clinical and demographic assessment was carried out at baseline and 28 weeks. Results Depression at 28 weeks was predicted by the additive effects of severity, obsessive-compulsive disorder and suicidal ideation at entry together with presence of at least one disappointing life event over the follow-up period. Conclusions Clinicians should assess for severity, suicidality and comorbid obsessive-compulsive disorder at presentation and should monitor closely for subsequent life events during treatment.