Objective To determine the validity of using an historical maximum spontaneous regression rate (reportedly 0±1.1% in those with lung metastases after nephrectomy) in clinical trials of treatments for patients with metastatic renal cell carcinoma (RCC), as the eligibility criteria for most studies will select patients with better performance status (and thus excluding those who are unlikely to respond) and more modern staging methods would potentially reduce the number of false-positives. Patients and methods A multicentre randomized, placebo-controlled, double-blind trial was recently completed in which 197 patients with metastatic RCC from 17 study centres across Canada were randomized to receive placebo or recombinant interferon c-1b (60 mg/m 2 ) subcutaneously once every 7 days until disease progression. All tumour responses were validated by an independent response committee unaware of the treatment. Results The median (95% con®dence interval) overall response rate (complete, CR, and partial, PR) for those on interferon-c was 4 (1.4±11.5)% and for those on placebo was 6 (2.5±13.2)% (P=0.75). In the six patients who were receiving placebo the CR and PR (three each) was considered to represent spontaneous remission. Of these six patients (aged 44±64 years) ®ve had undergone nephrectomy, one a tumour embolization, four had clear cell carcinoma and one an adenocarcinoma, and all had regression of lung and/ or lymph node metastases. Conclusion The lack of ef®cacy of interferon-c in this trial underlines the importance of continued research to identify alternative therapeutic agents or combinations of agents in phase II studies. However, the threshold response rate for initiating phase III trials should be increased to 18% in the phase II trials, i.e. three times the response rate on placebo.