Abstract Aim of the study To determine whether a risk factor adapted chemotherapy would improve the outcome of non-metastatic bone Ewing’s sarcoma. Methods Standard risk tumours (SR, good histological response to chemotherapy or small unresected tumours) received the previous EW88 chemotherapy. Ifosfamide/etoposide (IE) were introduced after 3 courses of cyclophosphamide/doxorubicine when tumour regression was <50% or during consolidation therapy for the intermediate risk tumours (IR, intermediate histological response 5–30% residual cells or large unresected tumours >100 ml). High risk tumours (HR, histological poor response >30% residual cells or clinical poor response <50% for unresectable tumours), received IE prior high dose busulfan/melphalan with stem cell rescue. Results From 1993 to 1999, 214 patients were enrolled. 5y-EFS and OS were 60% (95% confidence interval (CI), 53–66) and 69% (95% CI, 63–75), respectively. 116 (54%), 46 (21%), 48 (22%) patients were considered as SR, IR and HR of relapse, respectively. No advantage to IE was observed in the IR group. As compared to previous study, tumour with poor histological response to induction chemotherapy seemed to benefit from the consolidation strategy including busulfan/melphalan: EFS were 45% (95% CI, 30–60) and 20% (95% CI, 7–43) for EW93 and EW88, respectively. Despite a risk-adapted strategy, histological response to chemotherapy remains the main prognostic factor in resected tumours, while initial tumour volume is the main prognostic factor for unresected tumours. Conclusion These results showing a potential benefit of a consolidation strategy including busulfan/melphalan as compared to conventional chemotherapy needed confirmation by a randomised trial and were one of the bases of the ongoing EuroEwing99.