The discordance in oestrogen receptor (ER), progesterone receptor (PgR) and human epidermal growth factor receptor 2 (HER2) status between primary and recurrent breast cancer is being intensively investigated and a large amount of data have been produced. However, results from different studies are heterogeneous and often conflicting. To highlight this issue, a meta-analysis of published data was performed.

A literature search was performed using Medline, and all the studies published from 1983 to 2011 comparing changes in ER, PgR and/or HER2 status in patients with matched breast primary and recurrent tumours were included. We used random-effects models to estimate pooled discordance proportions.

We selected 48 articles, mostly reporting retrospective studies. Thirty-three, 24 and 31 articles were focused on ER, PgR and HER2 changes, respectively. A total of 4200, 2739 and 2987 tumours were evaluated for ER, PgR and HER2 discordance, respectively. The heterogeneity between study-specific discordance proportions was high for ER (I² = 91%, p < 0.0001), PgR (I² = 79%, p < 0.0001) and HER2 (I² = 77%, p < 0.0001). Pooled discordance proportions were 20% (95% confidence interval (CI): 16–35%) for ER, 33% (95% CI: 29–38%) for PgR and 8% (95% CI: 6–10%) for HER2. Pooled proportions of tumours shifting from positive to negative and from negative to positive were 24% and 14% for ER (p = 0.0183), respectively. The same figures were 46% and 15% for PgR (p < 0.0001), and 13% and 5% for HER2 (p = 0.0004).

Our findings strengthen the concept that changes in receptor expression may occur during the natural history of breast cancer, suggesting clinical implications and a possible impact on treatment choice.