Based on preclinical data for the antitumour effect of zoledronate in osteosarcoma, we assessed whether zoledronate combined with chemotherapy and surgery improved event-free survival in children and adults with osteosarcoma.

In this randomised, multicentre, open-label, phase 3 trial (OS2006), patients aged between 5 years and 50 years with newly diagnosed high-grade osteosarcoma were randomly assigned to receive standard chemotherapy with or without ten zoledronate intravenous infusions (four preoperative and six postoperative). Adults older than 25 years received 4 mg zoledronate per infusion, patients aged 18–25 years received 0·05 mg/kg for the first two infusions and 4 mg for the remaining eight infusions, and younger patients received 0·05 mg/kg per infusion. Chemotherapy comprised high-dose methotrexate based chemotherapy in patients younger than 18 years, and doxorubicin, ifosfamide, and cisplatin in adults older than 25 years; patients aged 18–25 years were treated with either regime at the discretion of the treating centre. Balanced randomisation between the two groups was done centrally with online randomisation software, based on a minimisation algorithm taking into account centre, age, combined with chemotherapy regimen, and risk group (resectable primary and no metastasis vs other). Patients and investigators were not masked to treatment assignment, but the endpoint adjudication committee members who reviewed suspected early progressions were masked to group allocation. The primary endpoint was event-free survival, estimated from the randomisation to the time of first failure (local or distant relapse, progression, death) or to the last follow-up visit for the patients in first complete remission, analysed on a modified intention-to-treat population, which excluded patients found not to have a malignant tumour after central review. Three interim analyses were planned. This trial is registered with ClinicalTrials.gov, number NCT00470223.

Between April 23, 2007, and March 11, 2014, 318 patients, median age 15·5 years (range 5·8–50·9), were enrolled from 40 French centres; of whom 158 were assigned to the control group (chemotherapy alone) and 160 to the zoledronate group, including 55 (17%) patients with definite metastases. The trial was stopped for futility after the second interim analysis. With a median follow-up of 3·9 years (IQR 2·7–5·1), 125 events occurred (55 in the control group and 70 in the with zoledronate group). Event-free survival at 3 years for all 315 randomly assigned patients was 60·3% (95% CI 64·5–65·9); 3-year event-free survival was 63·4% (55·2–70·9) for the control group and 57·1% (48·8–65·0) for the zoledronate group. The risk of failure was not reduced and was even marginally higher in the zoledronate group than in the control group (hazard ratio [HR] 1·36 [95% CI 0·95–1·96]; p=0·094). No major increase in severe toxic effects of grade 3 or higher associated with zoledronate, barring expected hypocalcaemia (45 [29%] of 153 participants in the zoledronate group vs ten [6%] of 155 participants in the control group; p<0·0001) and hypophosphataemia (61 [40%] of 151 in the zoledronate group vs 26 [17%] of 156 in the control group; p<0·0001). No significant difference in orthopaedic complications was noted between the two groups (27 in the control group and 29 in the zoledronate group). Two treatment-related deaths were reported (one from cardiomyopathy in the control group and one from multiorgan failure in the zoledronate group before the first zoledronate infusion).

From the results observed in this study, we do not recommend zoledronate in osteosarcoma patients. Further biological studies are required to understand the discordance between the results of OS2006 trial and preclinical data.

French National Cancer Institute (INCa), Novartis, Chugai, Ligue Nationale contre le Cancer, Fédération Enfants et Santé, Société Française des Cancers et Leucémies de l’Enfant.