This study aimed to evaluate patterns of local and distant disease recurrence in patients having primary chemotherapy and compared patterns of relapse in patients with a complete pathological response with those who had residual breast disease. This is an observational study using a sequential series of patients treated with primary chemotherapy. They were followed up for a minimum of 5 years. All data was collected prospectively. Three hundred forty-one consecutive patients with breast cancer were treated with up to eight cycles of doxorubicin-based chemotherapy. Clinical and pathological response rates were evaluated and patients were followed up for disease recurrence (local and distant) and overall survival. Fifty-two patients (16.5%) had a complete pathological response to chemotherapy. Distant disease recurrence occurred in nine patients (17.3%) but no local recurrence was observed. In patients not having a complete pathological response, 86 patients (32.6%) subsequently developed metastases. Local recurrence of disease occurred in 12 (4.5%). There was a statistically significant difference in overall survival between patients whose tumours had a complete pathological response compared with patients who had residual disease in the breast following chemotherapy (88% versus 70% at 5 years, p = 0.036). Following primary chemotherapy, about 84% of patients had residual disease in the breast. Surgery is necessary to ensure complete removal of residual tumour and excellent rates of local control are achievable. A complete pathological response is associated with fewer local and distant recurrences as well as improved survival although there are no differences in time to development of metastatic relapse.