Breast cancer (BC) is the first female cancer in France, accounting for 49,240 new cases in 2004. Approximately 80% of those tumors have positive hormone receptors (HR). Tamoxifen was used in four chemoprevention randomized trials, as well as another SERM (Selective Estrogen Receptor Modulation), raloxifen. This review analyses the updated results of these trials. All trials have shown that the risk of developing HR positive BC was reduced by tamoxifen or raloxifen, but without impact on HR negative BC and overall survival. Moreover, several unfavorable side effects (thrombo-embolic accidents and uterine cancers) have been observed. A new assessment of BC risk factors seems necessary, including not only family history and some histopathological abnormalities (e.g. atypical hyperplasia), but also new elements such as high bone and breast density and thoracic irradiation at young age (Hodgkin's disease). Indeed, tamoxifen efficacy seems optimal in very "high-risk" women. Therefore, the creation of a new and most comprehensive "risk model" is necessary as well as a tailored SERM use (maybe with other compounds), in order to optimize results and reduce potential side effects.