Tamoxifen or raloxifene in postmenopausal women for prevention of breast cancer: a tale of two choices--counterpoint.

Postmenopausal women 50 years or older with an estimated 5-year Gail model breast cancer risk of >1.66% or prior history of lobular carcinoma in situ now have two choices of prevention drug therapy: approved by the Food and Drug Administration for breast cancer risk reduction tamoxifen and raloxifene. It has been estimated that f2 million women living in the United States would realize a health benefit from 5 years of tamoxifen given as primary breast cancer prevention therapy (1, 2); yet, only 5% to 30% of risk-eligible women choose to take tamoxifen because of potential side effects and incomplete efficacy (3-5). Serious side effects that can occur with tamoxifen, such as deep venous thrombosis, pulmonary embolism, and uterine cancer, occur primarily in women over 50 years. Treatment of elderly, average-risk women with raloxifene in the Multiple Outcomes of Raloxifene Trial was associated with fewer breast cancers without the increase in uterine cancer observed with tamoxifen. However, raloxifene was associated with a 3-fold increase in thromboembolic phenomenon, similar to what would be expected with tamoxifen (6). Subsequently, the National Surgical Adjuvant Breast and Bowel Project Study of Tamoxifen and Raloxifene (STAR) trial was launched in high-risk postmenopausal women with the hope that raloxifene would have similar efficacy but fewer serious side effects than tamoxifen. The initial assessment of the primary end point of the STAR Trial would seemingly indicate that this hope was realized. Compared with tamoxifen, women randomized to 5 years of raloxifene in the STAR trial had a similar incidence of invasive breast cancer and similar quality of life but a lower incidence of serious side effects (7, 8). Women randomized to raloxifene had 30% fewer thromboembolic events, 25% fewer uterine cancers, and 21% fewer cataracts compared with those randomized to tamoxifen. These differences were significant for cataracts and the combination of deep venous thrombosis and pulmonary emboli (7). Marring this otherwise perfect story is the disquieting observation that in the STAR trial the incidence of noninvasive breast cancer was 40% lower for women randomized to tamoxifen than those randomized to raloxifene, and ductal carcinoma in situ comprised 54% of these in situ cancers (7). The clinical relevance is obvious in that >20% of newly diagnosed breast cancers projected for 2007 will be ductal carcinoma in situ (9). Management of ductal carcinoma in situ , which often incorporates 5 years of antihormone therapy following surgery F radiation makes this a life-altering event. Although not quite statistically significant at the time of the analysis (P = 0.052), the reduced incidence of in situ cancers in women randomized to tamoxifen versus raloxifene in STAR is consistent with results in prior placebo-controlled trials. In trials in high-risk women, tamoxifen has been observed to reduce ductal carcinoma in situ by 37% to 49% compared with placebo (10, 17). Raloxifene has not been compared with a placebo in high-risk women, but in average to low-risk postmenopausal women, raloxifene was not associated with either a significant or numerical reduction in ductal carcinoma in situ compared with placebo (see Fig. 1; refs. 12, 13). The observation that raloxifene is similar to tamoxifen in preventing invasive but less efficacious in preventing in situ cancer seems to counter the existing paradigm but that most invasive cancers are derived from associated in situ lesions. What then might be the explanation for the differences between raloxifene and tamoxifen with regard to emergence of in situ cancer? This seeming paradox could be explained if changes in gene expression giving rise to in situ cancer are different than those associated with progression from in situ to invasive cancer, and tamoxifen modulated those genes key to development of breast precancer differently than raloxifene. Although there does not seem to be any gene universally upregulated or down-regulated in in situ or invasive cancer, several investigators have noted differences in patterns of change in gene expression between normal and in situ cancer and between in situ and invasive cancer (14-16). Further, Frasor et al. (11) have reported that there is a set of at least 60 genes up-regulated or downregulated by tamoxifen but not by raloxifene. If, as generally thought, hormone receptor–positive invasive cancer emerges from in situ cancer, the implications are that tamoxifen might be more efficacious than raloxifene in the long run for preventing invasive disease