The aim of this analysis was to assess how the clinical response to chemotherapy corresponded to long-term prognosis in patients of less than 35 years of age. A retrospective analysis was made of response and survival data of 609 premenopausal patients who had been treated by four cycles of neoadjuvant chemotherapy followed by surgery and/or radiotherapy. Patients were stratified into three age groups (group 1, < or = 35 years; group 2, 35-40 years; group 3, > or = 41 years). Objective and complete clinical response rates were significantly higher in the youngest patients (below 35 yrs: P = 0.005 and P = 0.001, respectively) in stark contrast to a particularly poor outcome of this subpopulation. Five-year local recurrence rates were 31% in the youngest patients, compared with 26% and 16% in groups 2 and 3, respectively (P = 0.0007). Group 1 patients also had significantly higher 5-year metastatic relapse rates (41% versus 35% and 28%; P = 0.007) and 5-year survival figures were 70%, 82% and 84% for groups 1, 2 and 3 respectively (P = 0.002). Finally, stratification by age and by response revealed that, whilst the outcome of the youngest patients was highly dependent on their response to primary chemotherapy, complete responders showed disease-free survival rates at 5 years that were lower than these of older patients, whatever their response. Despite a seemingly better control of the primary tumour by chemotherapy, the patients in the youngest age group remained at a high risk for local and metastatic relapse. This apparent paradox may be in part attributable to rapid disease progression of micrometastatic tumour subpopulations that are refractory to chemotherapy.