Background: Aromatase inhibitors (AI) are superior to tamoxifen (TAM) in postmenopausal estrogen receptor positive (ER+) breast cancer patients. TAM is standard of care for premenopausal ER+ patients. The SOFT/TEXT trials showed, however, that ovarian function suppression combined with the AI exemestane significantly improves disease-free survival compared to TAM in premenopausal patients. Evidence on perimenopausal women at breast cancer diagnosis is scarce as they are usually excluded from studies. We therefore aim to provide evidence for the optimal treatment of ER+ breast cancer patients 45-50 years of age, using data from the population-based Netherlands Cancer Registry (NCR). Methods: All pre-menopausal Dutch women, without prior history of malignancy, diagnosed between 2004-2007 at age 45-50 years with an ER+, non-metastatic, primary invasive breast cancer for which they received chemotherapy and endocrine treatment, were identified through the NCR. Cox proportional hazards regression was used to calculate hazard ratios (HR) according to the time-dependent percentage of endocrine treatment duration during which an AI was used, categorized into < 25%, 25%-75% or > 75%. Analyses were adjusted for chemotherapy, trastuzumab use, lymph node status, grade, progesterone receptor, HER2 status, and clinical T stage. Results: A total of 346 events were observed in 2387 patients. Recurrence-free survival (RFS) was significantly improved for patients who received an AI > 75% of treatment duration (adjusted HR 0.59; 95% Confidence Interval 0.42-0.83; P = 0.002) compared to patients who received AI < 25% of treatment duration. No difference was observed between the 25-75% and < 25% groups. There was a significant trend with increasing percentage of the use of AI (p = 0.001). Conclusions: An AI for > 75% of treatment duration significantly improves RFS and should be considered in ER+ breast cancer patients aged 45-50 years at diagnosis.