76 patients with high risk neuroblastoma were treated with one (41 patients) or two consecutive courses (35 patients) of megatherapy. Autologous bone marrow transplantation was scheduled after each megatherapy. Univariate analysis confirmed two prognostic factors in this heterogeneous study population: no bone lesions before megatherapy and age at diagnosis of less than 2 years with 5-year progression-free survival rates of 51% (P < 0.0007) and 53% (P < 0.025), respectively. Both factors were shown to be of independent prognostic significance using the Cox proportional hazard model. Identification of prognostic factors should help to define the interest and limits of megatherapy. We consider that elective megatherapy followed by innovative treatments appears justified in patients with persisting bone disease. In contrast, megatherapy has to be re-evaluated for patients showing a more favourable response pattern and/or young age, ideally in a randomised, prospective trial.