Improvements in screening and adjuvant therapy for breast cancer are associated with decreased recurrence, which may have the effect of increasing the proportion of patients presenting with first-line de novo versus recurrent metastatic breast cancer (MBC). Here, we describe and compare patients with de novo versus recurrent human epidermal growth factor 2 (HER2)-positive MBC. registHER was a prospective observational cohort study (late 2003-early 2006) of 1,023 patients with HER2-positive MBC. Baseline characteristics, treatment patterns, and clinical outcomes were examined in patients with newly diagnosed de novo (n = 327) compared with recurrent HER2-positive MBC after prior treatment for early-stage disease (n = 674). Patients with de novo HER2-positive MBC were less likely to have lung metastases, more likely to have lymph node, bone, and/or liver metastases and >4 sites of metastases and more likely to receive combined or concurrent chemotherapy and hormonal therapy with or without trastuzumab than those with recurrent HER2-positive MBC. Median follow-up was 29 months. Median progression-free survival was 12.1 versus 9.3 months [hazard ratio = 0.716 (95 % confidence interval (CI) 0.617-0.831)], and overall survival was 41.7 versus 32.8 months [hazard ratio = 0.766 (95 % CI 0.633-0.928)] for patients with de novo versus recurrent HER2-positive MBC, respectively. Patients with recurrent HER2-positive MBC had similar outcomes regardless of whether they received prior adjuvant therapy, excluding hormonal therapy. Despite presenting with more advanced-stage disease and higher tumor burdens, patients with de novo HER2-positive MBC have more favorable clinical outcomes than those with recurrent HER2-positive MBC. These differences may be due to effects of prior drug exposure and could have implications for designing and interpreting clinical trials.