In a trial conducted in southeast Scotland between April 1974 and December 1979, 332 women with invasive breast cancer of Stage I and II with histological evidence of node involvement, or who had operable or inoperable Stage III disease, were randomized, after primary local therapy (mastectomy, node biopsy and radiotherapy for all except the inoperable disease patients who underwent radiotherapy alone) to receive 12 4-weekly intravenous injections of 5-fluorouracil (5-FU), 700 mg/m2 or no systemic therapy. After a median follow-up of 15 years from randomization, no difference is shown between the two groups in terms of distant relapse (hazard ratio (HR) = 1.02; 95% CI 0.78-1.32), event free survival (HR = 1.23; 95% CI 0.97-1.56), or total survival (HR = 1.19; 95% CI 0.93-1.52). Locoregional relapse is significantly reduced by 5-FU administration (HR = 1.88; 95% CI 1.20-2.96). The results are similar for the trial as a whole or when mastectomy patients are considered alone. Toxicity was minimal with marrow suppression in only 19 of 147 patients receiving more than one injection; only five patients discontinued therapy due to nausea and vomiting. However, retrosternal pain occurred in 16 patients, nine of whom had their treatment curtailed as a result. Seventy-seven per cent of patients have died, the majority from breast cancer. Only 1.2% of deaths are considered attributable to cardiac causes of 5-FU is not associated with excess cardiac deaths in these patients.(ABSTRACT TRUNCATED AT 250 WORDS)