We report the medium-term (median follow-up=52 months) results of a prospective randomised trial of multimodal therapy (neoadjuvant chemotherapy, Patey mastectomy, postoperative radiotherapy and adjuvant hormone therapy) (n=56) versus initial hormone therapy (n=52) for locally advanced primary breast cancer. Compared with multimodal therapy, initial hormone therapy was associated with reduced number of therapies for disease control (mean=3.6 versus 4.9) and mastectomy rate (31%). Multimodal therapy conferred better initial locoregional control and a longer disease-free interval. Nevertheless, there was no statistically significant differences in the rates of survival, metastasis and uncontrolled locoregional disease, as well as in the time to metastasis between the two therapy groups. Regardless of the therapy groups, oestrogen receptor positivity conferred a lower metastasis rate, better survival and locoregional control. Thus, initial hormone therapy may be a reasonable option for managing locally advanced primary breast cancer, especially for oestrogen receptor-positive tumours.