Background: Treatment decisions in recurrent breast cancer are usually based on the estrogen (ER), progesterone (PgR) and HER2 receptor status of the primary tumour. Retrospective studies suggest that discordance between receptor expression of primary and recurrent breast cancer exists. Methods: A pooled analysis of individual patient data from two large prospective studies comprising biopsy of recurrent lesions obtained from consenting patients was undertaken. Tissue was analyzed for ER, PgR by immunohistochemistry and HER2 by FISH. Receptor status of recurrent disease was compared with that of the primary tumour. Recruiting clinicians assessed whether or not receptor discordance affected subsequent systemic treatment. Results: Two hundred and eighty-nine patients underwent biopsy. Recurrent biopsy specimens were obtained from locoregional recurrence in 48.1% and from distant metastases in 51.9%. Distant sites included skin/soft tissue (25.0%), bone/bone marrow (19.2%) and liver (15.8%). Benign disease or second primary cancer was observed in 7.6% of biopsies. Discordance in ER, PgR or HER2 between confirmed primary and recurrent breast cancer was 12.6%, 31.2% and 5.5%, respectively (all p <0.001). Biopsy results altered management in 14.2% of patients undergoing biopsy (95% confidence intervals 10.4-18.8%, p = 0.0001). The duration between primary and recurrent disease, the site of recurrence and the receptor profile of the primary tumour did not affect discordance rates. Conclusions: There is substantial discordance in receptor status between primary and recurrent breast cancer. The number needed to biopsy in order to alter treatment was 7.1. Patients with recurrent breast cancer should have tissue confirmation of receptor status of recurrent disease.