In 1977 the National Surgical Adjuvant Breast and Bowel Project initiated a prospectively randomized clinical trial for women with primary operable breast cancer and positive axillary nodes. In this study 1891 patients were randomized to receive L-phenylalanine mustard and 5-fluorouracil (PF) either with or without tamoxifen (T). In this interim report findings are presented concerning disease-free survival (DFS) and survival as related to age and to estrogen receptor (ER) and/or progesterone receptor (PR) content of the tumor. The median follow-up time is 3 yr. Patients 50 yr of age or older with either 1-3 or more than 3 positive axillary nodes had a markedly longer disease-free survival on PFT than did those receiving PF adjuvant therapy (p less than 0.001). The effectiveness of PFT was related to the levels of tumor receptors. Patients 50 yr old or more with both tumor ER and PR levels of 10 fmole or more ("high") displayed the greatest benefit in disease-free survival from PFT (p = 0.004). Analyses by age indicated that it is more appropriate to divide patients of 50 yr or older into two age groups, 50-59 and 60-70 yr old. In the former the survival results were poorer on PFT when tumor PR was low, whereas, regardless of receptor levels, those 60-70 yr old experienced an advantage on PFT. In women under 50 yr of age, there was no difference in disease-free survival (p = 0.64), but survival results favored the PF over the PFT treated (p = 0.06). Patients under 50 yr with tumor ER and PR levels under 10 fmole ("low") had a poorer survival when given PFT (p = 0.003). Those whose tumors demonstrated a high ER and a low PR also had a shorter survival on PFT (p = 0.01). The observation of no benefit in younger patients when both receptor levels were high, but a benefit in older patients with receptor-poor tumors, indicates that, at least according to the conditions of this study, the difference between the two age groups cannot be explained by the association of age with receptor content. Multivariate analyses considered the effects of the number of positive nodes, age, ER, and PR. They support the conclusion that, while nodes and ER exert strong prognostic influences in both PF- and PFT-treated patients, the PR content of tumors is a stronger predictor of the effectiveness of PFT therapy than is ER content.(ABSTRACT TRUNCATED AT 400 WORDS)