BACKGROUND: Ipilimumab is a fully human monoclonal antibody that binds cytotoxic T-lymphocyte antigen 4 to enhance antitumour immunity. Our aim was to assess the use of ipilimumab after radiotherapy in patients with metastatic castration-resistant prostate cancer that progressed after docetaxel chemotherapy. METHODS: We did a multicentre, randomised, double-blind, phase 3 trial in which men with at least one bone metastasis from castration-resistant prostate cancer that had progressed after docetaxel treatment were randomly assigned in a 1:1 ratio to receive bone-directed radiotherapy (8 Gy in one fraction) followed by either ipilimumab 10 mg/kg or placebo every 3 weeks for up to four doses. Non-progressing patients could continue to receive ipilimumab at 10 mg/kg or placebo as maintenance therapy every 3 months until disease progression, unacceptable toxic effect, or death. Patients were randomly assigned to either treatment group via a minimisation algorithm, and stratified by Eastern Cooperative Oncology Group performance status, alkaline phosphatase concentration, haemoglobin concentration, and investigator site. Patients and investigators were masked to treatment allocation. The primary endpoint was overall survival, assessed in the intention-to-treat population. This trial is registered with ClinicalTrials.gov, number NCT00861614. FINDINGS: From May 26, 2009, to Feb 15, 2012, 799 patients were randomly assigned (399 to ipilimumab and 400 to placebo), all of whom were included in the intention-to-treat analysis. Median overall survival was 11.2 months (95% CI 9.5-12.7) with ipilimumab and 10.0 months (8.3-11.0) with placebo (hazard ratio [HR] 0.85, 0.72-1.00; p=0.053). However, the assessment of the proportional hazards assumption showed that it was violated (p=0.0031). A piecewise hazard model showed that the HR changed over time: the HR for 0-5 months was 1.46 (95% CI 1.10-1.95), for 5-12 months was 0.65 (0.50-0.85), and beyond 12 months was 0.60 (0.43-0.86). The most common grade 3-4 adverse events were immune-related, occurring in 101 (26%) patients in the ipilimumab group and 11 (3%) of patients in the placebo group. The most frequent grade 3-4 adverse events included diarrhoea (64 [16%] of 393 patients in the ipilimumab group vs seven [2%] of 396 in the placebo group), fatigue (40 [11%] vs 35 [9%]), anaemia (40 [10%] vs 43 [11%]), and colitis (18 [5%] vs 0). Four (1%) deaths occurred because of toxic effects of the study drug, all in the ipilimumab group. INTERPRETATION: Although there was no significant difference between the ipilimumab group and the placebo group in terms of overall survival in the primary analysis, there were signs of activity with the drug that warrant further investigation. FUNDING: Bristol-Myers Squibb.