Endocrine responsiveness of primary breast cancers with borderline estrogen receptor expression (ER [1%-9%]) remains unclear. We aimed at investigating differences in endocrine responsiveness, prognosis, and clinicopathological characteristics between the ER (1%-9%) cohort and the ER cohort or ER (≥10%) cohort. Eligible literature published from inception to November 20, 2016 was retrieved from the PubMed database on the basis of Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Data on survival outcomes were extracted and pooled odds ratios (ORs), 95% confidence intervals (CIs), and 2-tailed P values are reported. P values of the χ test for comparison of clinicopathological characteristics among included patients in the ER (1%-9%) cohort and the other 2 cohorts were calculated respectively. The analysis included 6 studies with 16,606 patients. Significant differences were detected between the ER (1%-9%) cohort and the other 2 cohorts on the basis of clinicopathological characteristics respectively. When taking all of the patients into analysis without consideration of treatment modality, the ER (1%-9%) cohort presented better prognosis than the ER group in terms of 5-year disease-free survival (OR, 1.47; P = .046) and 5-year overall survival (OR, 1.23; P = .046). However, patients with ER (1%-9%) breast cancer who received endocrine therapy seemed to have a prognosis similar to those without any endocrine therapy (P = .684) and those with ER carcinoma who received endocrine therapy (P = .145). Patients with ER (≥10%) tumors had better endocrine responsiveness compared with their ER (1%-9%) counterparts (OR, 0.52; P = .034, ER [1%-9%] vs. ER [≥10%]). Our results indicate that primary breast cancer patients with ER (1%-9%) expression gained no significant survival benefit from endocrine therapy, but manifested overall better prognosis than those with ER cancer.