Background The steroid-receptor coactivator amplified in breast cancer one (AIB1) is implicated to be a prognostic factor, although the results are not unanimous. Recently its effect was suggested to be modified by paired box 2 gene product (PAX2). Patients and methods Using immunohistochemistry (IHC) AIB1 and PAX2 were investigated in two cohorts of early breast cancer, including systemically untreated premenopausal lymph-node-negative women and pre- and postmenopausal women receiving tamoxifen. Results AIB1 scores were available for 490 patients and PAX2 scores were available for 463 patients. High AIB1 was a negative prognostic factor for distant disease-free survival (DDFS, P = 0.02) and overall survival (OS, P < 0.001) in systemically untreated women, while no prognostic effect was seen in the tamoxifen-treated cohort, indicating AIB1 to be a predictor of tamoxifen response. In systemically untreated patients, PAX2 was not a prognostic factor, nor did it modify the effect of AIB1. However, in ER-positive patients receiving tamoxifen, PAX2 appeared to be a positive prognostic factor in premenopausal patients, while a negative factor in postmenopausal. The interaction between the menopausal status and PAX2 was significant (P = 0.01). Conclusions In an independent cohort of low-risk premenopausal patients, we validate AIB1 as a negative prognostic factor, indicating AIB1 to be an interesting target for new anti-cancer therapies. The effect of PAX2 warrants further studies.