The published literature comparing surgery, with or without adjuvant endocrine therapy, with endocrine therapy alone in older women with operable breast cancer was systematically reviewed.The design used is Cochrane review. Randomised controlled trials retrieved from the Cochrane Breast Cancer Group Specialised Register on 29 June 2005. Eligible studies recruited women aged 70 years or over with operable breast cancer, fit for surgery under general anaesthia. The studies compared surgery (either mastectomy or wide local excision, with or without endocrine therapy) to endocrine therapy alone. Primary outcomes were overall survival (OS) and progression-free survival (PFS). Double data extraction and quality assessment were undertaken. Seven eligible trials were identified of which six had published time-to-event data. The quality of the allocation concealment was adequate in three studies and unclear in the remainder. In each case the endocrine therapy used was tamoxifen. When surgery alone was compared to endocrine therapy alone, there was no significant difference in OS (hazard ratio (HR) 0.98, 95% confidence interval (CI) 0.74 -1.30, P ¼ 0.9), but a significant difference in PFS (HR 0.55, 95% CI 0.39 -0.77, P ¼ 0.0006). When surgery with adjuvant endocrine therapy was compared to endocrine therapy alone, there was no significant difference in OS (HR 0.86, 95% CI 0.73 -1.00, P ¼ 0.06), but a significant difference in PFS (HR 0.65 (95% CI 0.53 -0.81, P ¼ 0.0001) for surgery plus endocrine therapy vs primary endocrine. The regimens have different side effect profiles with one study suggesting increased psychosocial morbidity at 3 months in the surgical arm, which resolves by 2 years. Primary endocrine therapy with tamoxifen is associated with inferior local disease control but noninferior survival to surgery for breast cancer in older women. Trials are needed to evaluate appropriate selection criteria for its use in terms of patient co-morbidity and quality of life. Trials are needed to evaluate the clinical effectiveness of aromatase inhibitors as primary therapy for this population.