Background: Giant cell tumour of bone (GCTB) is a very rare, aggressive, and progressive osteolytic tumour for which no standard medicinal treatment or chemotherapy exists. We report interim safety and efficacy results from a phase 2 study of denosumab in patients with GCTB. Methods: We did an international, open-label, parallel-group, phase 2 trial of patients with histologically confirmed GCTB and radiographically measurable active disease. Eligible patients were adults or skeletally mature adolescents with radiographic evidence of at least one mature long bone who were at least 12 years old and weighed at least 45 kg. We divided patients into three cohorts-those with surgically unsalvageable GCTB (cohort 1), those with salvageable GCTB whose surgery was associated with severe morbidity (cohort 2), and those who transferred from a previous study of denosumab for GCTB (cohort 3). Patients in cohorts 1 and 2 received 120 mg of subcutaneous denosumab every 4 weeks with loading doses on days 8 and 15 of the first cycle; those in cohort 3 continued the regimen from the previous study. Investigator-determined disease status and clinical benefit were assessed every 4 weeks. Our primary endpoint was the safety profile of denosumab in terms of adverse events and laboratory abnormalities. Prespecified secondary endpoints were time to disease progression in cohort 1 and the proportion of patients without any surgery at 6 months in cohort 2. Safety analyses included all patients who received at least one dose of denosumab. Efficacy analyses included all eligible patients who received at least one dose of denosumab. This study is registered with ClinicalTrials.gov, identifier NCT00680992. Findings: 282 patients, including ten adolescents, were included between Sept 9, 2008, and March 25, 2011. Of the 281 patients analysable for safety, three (1%) had osteonecrosis of the jaw and 15 (5%) hypocalcaemia. The most common grade 3-4 adverse events were hypophosphataemia, which occurred in nine (3%) patients, and anaemia, back pain, and pain in extremities, each of which occurred in three patients (1%). Serious adverse events were reported in 25 (9%) patients. No treatment-related deaths were reported. On the basis of investigators' assessment of disease status, 163 of 169 (96%) analysable patients in cohort 1 had no disease progression after median follow-up of 13 months (IQR 5·8-21·0). In cohort 2, 74 of 100 (74%) analysable patients had no surgery and 16 of 26 (62%) patients who had surgery underwent a less morbid procedure than planned. Median follow-up in cohort 2 was 9·2 months (IQR 4·2-12·9). Interpretation: Adverse events were consistent with the known safety profile of denosumab. Denosumab was associated with tumour responses and reduced the need for morbid surgery in patients with GCTB. Denosumab represents a new treatment option for patients with GCTB. Funding: Amgen.