Two different methods to determine steroid receptors were analysed with respect to their ability to estimate prognosis in primary breast cancer patients. The immunohistochemical assay (IHA) was compared with the dextran-coated charcoal (DCC) method of receptor determination. A random sample of 281 patients with invasive ductal carcinoma was drawn from 841 consecutive patients with primary breast carcinoma treated at Odense University Hospital between 1 January 1980 and 31 December 1990. Receptor determination by the DCC method had been carried out previously in 164 patients for the oestrogen receptor and in 132 patients for the progesterone receptor. The former group was reassessed by IHA with the antibody ER1D5, and the latter with the antibody PgR-ICA. The median follow-up time was 8.3 years (range 2.9-12.9 years). A cutoff of zero was used for the DCC method. Immunohistochemical results were quantified by counting in systematically random sampled fields of vision and values above zero were considered to be positive. Overall agreement of positive and negative cases was 86% for the oestrogen receptor and 83% for the progesterone receptor. Although the study included a limited number of patients, receptor positive cases fared better than negative cases in all situations. Investigation of the prognostic power revealed that classification based on IHA allowed better discrimination of patients than classification based on the DCC method. The reason for this difference might be because distinction between benign and malignant tissue is possible using the IHAmethod. Thus, IHAresults appear to be more clinically relevant.