The aim of this study was to investigate in a randomized trial the activity of perioperative chemotherapy in patients treated with preoperative chemotherapy for locally advanced breast cancer and to compare it with the preoperative chemotherapy alone. Patients with cT2-3 N0-2 M0 histologically proven breast cancer, with estrogen receptors and progesterone receptors in less than 20% of cells, or with absence of progesterone receptors, received epirubicin 25 mg/m days 1 and 2, cisplatin 60 mg/m day 1, and fluorouracil 200 mg/m daily as continuous infusion. Responding patients were randomized to continue fluorouracil until 2 weeks after surgery (perioperative chemotherapy) or to stop fluorouracil 1 week before surgery. Fifty-eight patients completed six courses of epirubicin, cisplatin and fluorouracil, and were randomized to perioperative chemotherapy (29 patients) or to control (29 patients). The median Ki-67 index remained stable (32-27.5%) in the perioperative chemotherapy arm (P=0.3) and decreased from 55 to 22.5% in the control arm (P=0.01). The rate of pathological complete remission was 41% in both arms (P=1.0). No significant difference in terms of disease-free survival and overall survival was observed between the two arms. Perioperative chemotherapy failed to show an increase in the pathological complete remission rate. A biological effect on Ki-67 expression was demonstrated.