Background Adjuvant endocrine therapy compromises bone health in patients with breast cancer, causing osteopenia, osteoporosis, and fractures. Antiresorptive treatments such as bisphosphonates prevent and counteract these side-effects. In this trial, we aimed to investigate the effects of the anti-RANK ligand antibody denosumab in postmenopausal, aromatase inhibitor-treated patients with early-stage hormone receptor-positive breast cancer. Methods In this prospective, double-blind, placebo-controlled, phase 3 trial, postmenopausal patients with early hormone receptor-positive breast cancer receiving treatment with aromatase inhibitors were randomly assigned in a 1:1 ratio to receive either denosumab 60 mg or placebo administered subcutaneously every 6 months in 58 trial centres in Austria and Sweden. Patients were assigned by an interactive voice response system. The randomisation schedule used a randomly permuted block design with block sizes 2 and 4, stratified by type of hospital regarding Hologic device for DXA scans, previous aromatase inhibitor use, and baseline bone mineral density. Patients, treating physicians, investigators, data managers, and all study personnel were masked to treatment allocation. The primary endpoint was time from randomisation to first clinical fracture, analysed by intention to treat. As an additional sensitivity analysis, we also analysed the primary endpoint on the per-protocol population. Patients were treated until the prespecified number of 247 first clinical fractures was reached. This trial is ongoing (patients are in follow-up) and is registered with the European Clinical Trials Database, number 2005-005275-15, and with ClinicalTrials.gov, number NCT00556374. Findings Between Dec 18, 2006, and July 22, 2013, 3425 eligible patients were enrolled into the trial, of whom 3420 were randomly assigned to receive denosumab 60 mg (n=1711) or placebo (n=1709) subcutaneously every 6 months. Compared with the placebo group, patients in the denosumab group had a significantly delayed time to first clinical fracture (hazard ratio [HR] 0·50 [95% CI 0·39–0·65], p<0·0001). The overall lower number of fractures in the denosumab group (92) than in the placebo group (176) was similar in all patient subgroups, including in patients with a bone mineral density T-score of −1 or higher at baseline (n=1872, HR 0·44 [95% CI 0·31–0·64], p<0·0001) and in those with a bone mineral density T-score of less than −1 already at baseline (n=1548, HR 0·57 [95% CI 0·40–0·82], p=0·002). The patient incidence of adverse events in the safety analysis set (all patients who received at least one dose of study drug) did not differ between the denosumab group (1366 events, 80%) and the placebo group (1334 events, 79%), nor did the numbers of serious adverse events (521 vs 511 [30% in each group]). The main adverse events were arthralgia and other aromatase-inhibitor related symptoms; no additional toxicity from the study drug was reported. Despite proactive adjudication of every potential osteonecrosis of the jaw by an international expert panel, no cases of osteonecrosis of the jaw were reported. 93 patients (3% of the full analysis set) died during the study, of which one death (in the denosumab group) was thought to be related to the study drug. Interpretation Adjuvant denosumab 60 mg twice per year reduces the risk of clinical fractures in postmenopausal women with breast cancer receiving aromatase inhibitors, and can be administered without added toxicity. Since a main side-effect of adjuvant breast cancer treatment can be substantially reduced by the addition of denosumab, this treatment should be considered for clinical practice. Funding Amgen.