In this study the long-term prognosis was analysed of all 462 consecutive female breast cancer patients who were diagnosed and carefully staged between 1970 and 1980 in a 600-bed community hospital in Eindhoven, south east Netherlands. Follow-up of recurrence and causes of death was obtained until 1 January 1993. Observed survival rates at 5, 10 and 20 years were 66%, 45% and 32%, respectively, and the corresponding breast cancer-specific survival rates were 71%, 54% and 44%. The yearly risk for a recurrence of breast cancer after treatment steadily decreased from 10% the first year to 1% after 10 years. In a multivariate survival analysis both tumour size and nodal status appeared to be equally important prognostic factors in the first 5 years after diagnosis. After 5 years only tumour size had independent prognostic value, which was not significant any more after 10 years. In patients with a tumour size ⩽ 2 cm and without lymph node involvement at diagnosis, the risk for a recurrence was found to be negligible after 10 years. Those patients may be considered cured, although a search for early diagnosis of a second primary breast cancer in this group is still advisable.