We investigated the efficacy and toxicity of induction chemotherapy with cisplatin and etoposide with either bleomycin or ifosfamide in patients with intermediate-prognosis testicular non-seminoma. A total of 84 eligible patients were randomized to receive four cycles of etoposide, ifosfamide, cisplatin (VIP), or four cycles of bleomycin, etoposide, cisplatin (BEP). Intermediate prognosis was defined as any of the following: lymph node metastases 5-10 cm in diameter, lung metastases more than four in number or > 3 cm, HCG 5000-50,000 IU l(-1), AFP > 1000 IU l(-1). The complete response (CR) rates to VIP and BEP were similar, 74% and 79% respectively (P = 0.62). Including the cases in whom viable cancer was completely resected with post-chemotherapy debulking surgery, the percentages of patients who achieved a no-evidence-of-disease status were 80% on VIP and 82% on BEP (P = 0.99). In addition, there were no differences in relapse rate, disease-free and overall survival after a median follow-up of 7.7 years. The 5-year progression-free survival was 85% (95% CI 74-96%) in the VIP arm and 83% (95% CI 71-96%) in the BEP arm, hazard ratio (VIP/BEP) 0.83 (95% CI 0.30-2.28). The VIP regimen was more toxic with regard to bone marrow function; the frequency of leucocytes below 2000 microl(-1) throughout four cycles was 89% on VIP and 37% on BEP (P < 0.001). Our study does not indicate that ifosfamide is superior to bleomycin in combination with cisplatin and etoposide. The sample size in this study is small as the study was prematurely discontinued when data became available from a competing study that showed no improved effectiveness of VIP compared with BEP. Taken together with these data, bleomycin should not be replaced by conventional-dose ifosfamide.