The purpose of this study was to determine, from a review of published data, whether in stage 4 neuroblastoma in children over 1 year of age, the dose or scheduling of induction chemotherapy influenced the response rate in distant metastases. Publications relating to induction chemotherapy since the introduction of cisplatin/epipodophyllotoxin combinations were identified using Medline, Current Contents and personal reference lists. Thirteen publications were identified which described 17 regimens involving 948 children. The doses and the scheduling of the various regimens were compared with a standard regimen OPEC (vincristine, cisplatin, teniposide, cyclophosphamide). These were correlated with the reported response rates in the bone marrow. Due to a lack of standardisation in the nature of restaging investigations, timing of restaging and definitions of response it was difficult to compare all studies. The complete response rate at distant metastases ranged from less than 40% to over 90%. For individual drugs; the comparative doses given in each course ranged up to 4.2 g/m(2) for cyclophosphamide, 280 mg/m(2) for cisplatin, 600 mg/m(2) for etoposide and 4.5 mg/m(2) for vincristine. There was no evidence of any positive correlation between response rate in the marrow and either the dose of any individual drug or the schedule used. In contrast to a previous study which included a number of older studies where disease assessment was even more variable, this analysis has failed to show any justification for the routine use of very intensive induction regimens in this disease. Such an approach should only be taken in the context of randomised trials in which timing and methods of reassessment can be standardised. Until such studies demonstrate superiority either in terms of response rate or progression-free survival lower morbidity regimens should remain the standard therapy.