The Oxford Overview of adjuvant endocrine trials [1] indicates that 5 years of adjuvant tamoxifen reduces recurrence by 41% and deaths by 34% in women with oestrogen receptor (ER)-positive breast cancers. At 5 years, in all patients studied, the recurrence rate was 25.8% in controls but 13.9% on tamoxifen. There was a substantial 'carry over' effect of tamoxifen such that even after 15 years of follow up mortality was about 30% less in tamoxifen-treated patients. The effect of tamoxifen was greater in patients with ER-positive, PR-positive as compared with ER-positive PR-negative tumours. These data indicate a substantial effect of tamoxifen but it is clear that approximately half of patients are resistant to tamoxifen de novo (early relapses) or acquire resistance if we assume that women who relapsed later had an initial response to tamoxifen. The potential reasons for resistance include activated growth factor pathways overriding the inhibitory effects of the drug either via nuclear or membrane ER. Of ER-positive PR-negative tumours, 30% are HER1/2-positive, as compared with about 10% of ER-positive PR-positive tumours, and this difference may account for their lower activity of tamoxifen in PR-negative tumours. Modern aromatase inhibitors (AIs) are more effective in reducing relapse compared with tamoxifen whether AI treatment is initiated after surgery (ATAC and BIG1-98 trials) or after 2–3 years of tamoxifen (ITA, ARNO/ABCSG). At present, it is difficult to distinguish any differences in effectiveness between the three agents used in these trials (anastrozole, letrozole and exemestane) but small differences in toxicity patterns are beginning to emerge. The reason for the greater effectiveness of AIs is not clear. In randomized studies of neoadjuvant endocrine therapy [2] and in the anastrozole adjuvant trials (ATAC and ARNO/ABCSG) [3], the AIs used were particularly more active than tamoxifen in the ER-positive PR-negative subgroup of tumours, but this was not seen in the BIG1-98 and IES trials. Studies on letrozole resistant human mammary tumour cell lines show that growth factor pathways such as MAPK (mitogen-activated protein kinase) are activated, and sensitivity to the AI can be restored by growth factor pathway inhibitors [4]. Also, AI resistance can be reduced by combined treatment with fulvestrant in animal models. These data suggest mechanisms whereby AI resistance may be circumvented in patients and point to new approaches to adjuvant treatment.