Résumé : Background: Rhabdomyosarcoma is an aggressive tumour that can develop in almost any part of the body. Doxorubicin is an effective drug against rhabdomyosarcoma, but its role in combination with an established multidrug regimen remains controversial. Therefore, we aimed to evaluate the possible benefit of early dose intensification with doxorubicin in patients with non-metastatic rhabdomyosarcoma. Methods: We did a multicentre, open-label, randomised controlled, phase 3 trial involving 108 hospitals from 14 countries. We included patients older than 6 months but younger than 21 years with a pathologically proven diagnosis of rhabdomyosarcoma. We assigned each patient to a specific subgroup according to the EpSSG stratification system. Those with embryonal rhabdomyosarcoma incompletely resected and localised at unfavourable sites with or without nodal involvement, or those with alveolar rhabdomyosarcoma without nodal involvement were considered at high risk of relapse. These high-risk patients were randomly assigned (1:1) to receive either nine cycles of IVA (ifosfamide 3 g/m 2 given as a 3-h intravenous infusion on days 1 and 2, vincristine 1·5 mg/m 2 weekly during the first 7 weeks then only on day 1 of each cycle [given as a single intravenous injection], and dactinomycin 1·5 mg/m 2 on day 1 given as a single intravenous injection) or four cycles of IVA with doxorubicin 30 mg/m 2 given as a 4-h intravenous infusion on days 1 and 2 followed by five cycles of IVA. The interval between cycles was 3 weeks. Randomisation was done using a web-based system and was stratified (block sizes of four) by enrolling country and risk subgroup. Neither investigators nor patients were masked to treatment allocation. The primary endpoint was 3-year event-free survival assessed by the investigator at each centre in the intention-to-treat population. Patients who received at least one dose of study treatment were considered in the safety analysis. In agreement with the independent data monitoring committee, the study was closed to patient entry on Dec 16, 2013, after futility analysis. This trial is registered with EudraCT, number 2005-000217-35, and is currently in follow-up. Findings: Between Oct 1, 2005, and Dec 16, 2013, 484 patients were randomly assigned to receive each chemotherapy regimen (242 in the IVA group and 242 in the IVA plus doxorubicin group). Median follow-up was 63·9 months (IQR 44·6–78·9). The 3-year event-free survival was 67·5% (95% CI 61·2–73·1) in the IVA plus doxorubicin group and 63·3% (56·8–69·0) in the IVA group (hazard ratio 0·87, 95% CI 0·65–1·16; p=0·33). Grade 3–4 leucopenia (232 [93%] of 249 patients in the IVA plus doxorubicin group vs 194 [85%] of 227 in the IVA group; p=0·0061), anaemia (195 [78%] vs 111 [49%]; p<0·0001), thrombocytopenia (168 [67%] vs 59 [26%]; p<0.0001), and gastrointestinal adverse events (78 [31%] vs 19 [8%]; p<0·0001) were significantly more common in the IVA plus doxorubicin group than in the IVA group. Grade 3–5 infections (198 [79%] vs 128 [56%]; p<0·0001) were also significantly more common in the IVA plus doxorubicin group than in the IVA group, in which one patient had grade 5 infection. Two treatment-related deaths were reported (one patient developed septic shock and one affected by Goldenhar syndrome developed intractable seizures) in the IVA plus doxorubicin group, both occurring after the first cycle of treatment, and none were reported in the IVA group. Interpretations: The addition of dose-intensified doxorubicin to standard IVA chemotherapy did not show a significant improvement in the outcome of patients with high-risk non-metastatic rhabdomyosarcoma. Therefore, the IVA chemotherapy regimen should remain the standard of care for patients with localised rhabdomyosarcoma in Europe. Funding: Fondazione Città della Speranza, Italy, and the Association Léon Berard Enfant Cancéreux, France.