Background: Medication is commonly used to treat youth depression, but whether medication should be added to cognitive behavioural therapy (CBT) as first-line treatment is unclear. We aimed to examine whether combined treatment with CBT and fluoxetine was more effective than CBT and placebo in youth with moderate-to-severe major depressive disorder. Methods: The Youth Depression Alleviation–Combined Treatment (YoDA-C) trial was a randomised, double-blind, placebo-controlled, multicentre clinical trial. Participants were aged 15–25 years with moderate-to-severe MDD and had sought care at one of four clinical centres in metropolitan Melbourne, Australia. Patients were randomly assigned (1:1) to receive CBT for 12 weeks, plus either fluoxetine or placebo. Participants began on one 20 mg capsule of fluoxetine or one placebo pill per day. All participants received CBT, delivered by therapists in weekly 50-minute sessions and attended interviews at baseline, and at weeks 4, 8, and 12, during which they completed assessments with research assistants. Participants saw a psychiatrist or psychiatry trainee to complete medical assessments at the same timepoints. The primary outcome was change in the interviewer-rated Montgomery-Åsberg Depression Rating Scale (MADRS) score at 12 weeks. The trial was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12612001281886). Findings: 153 participants (mean age 19·6 years [SD 2·7]) were enrolled from Feb 20, 2013, to Dec 13, 2016. 77 (50%) patients were allocated to CBT and placebo and 76 (50%) to CBT and fluoxetine. Participants had severe depression at baseline (mean MADRS score 33·6 [SD 5·1] in the CBT and placebo group and 32·2 [5·6] in the CBT and fluoxetine group), with high proportions of participants with anxiety disorder comorbidity (47 [61%] in the CBT and placebo group and 49 [64%] in the CBT and fluoxetine group) and past-month suicidal ideation (55 [71%] in the CBT and placebo group and 59 [78%] in the CBT and fluoxetine group). 59 (77%) participants in the CBT and placebo group and 64 (84%) in the CBT and fluoxetine group completed follow-up at week 12. After 12 weeks of treatment both groups showed a reduction in MADRS scores (−13·7, 95% CI −16·0 to −11·4, in the CBT and placebo group and −15·1, −17·4 to −12·9, in the CBT and fluoxetine group). There was no significant between-group difference in MADRS scores (−1·4, −4·7 to 1·8; p=0·39). There were five suicide attempts in the CBT and placebo group and one suicide attempt in the CBT and fluoxetine group (odds ratio 0·2, 0·0–1·8; p=0·21), and no significant between-group differences for other suicidal behaviours. Interpretation: We did not find evidence that the addition of fluoxetine (rather than placebo) to CBT further reduced depressive symptoms in young people with moderate-to-severe MDD. Exploratory analyses showed that the addition of medication might be helpful for patients with comorbid anxiety symptoms and for older youth. Funding: Australian National Health and Medical Research Council.