Introduction Trastuzumab is a highly effective therapy for the treatment of HER-2/neu positive breast cancer. To maximize benefit and minimize unnecessary toxicity, patient selection is essential. Currently HER-2/neu analysis is routinely performed only for primary invasive breast cancers, and trastuzumab therapy is recommended based on primary analysis only. Methods Using immunohistochemistry, we performed a retrospective study comparing HER-2/neu expression in original primary to subsequent metastatic breast cancers. Results Tumors from 382 patients with metastatic breast cancer were studied. In 254 cases (66%) both primary and metastatic lesions were concordant. In 90 cases the primary lesion was HER-2/neu positive with the metastatic lesion negative; whereas, in 37 cases the primary lesion was HER-2/neu negative and the metastatic lesion positive. Primary HER-2/neu immunostaining was associated with a negative predictive value of 35.7%. Although all four groups were similar at diagnosis, survival differences were noted with the best survival experienced by patients with initial primary lesions HER-2/neu negative and subsequent metastatic lesions positive. Patients with hormone receptor and HER-2/neu positive primary lesions who received tamoxifen were more likely to have HER-2/neu positive metastasis. Conclusions The significant discordance between HER-2/neu expression in primary and metastatic tumors suggests that determination of HER-2/neu status in metastatic disease should be attempted.