Tamoxifen (T) is a synthetic antiestrogen having both agonist and antiagonist properties. It was originally developed as a contraceptive that was successful in rats but not in women. In fact, ovulation was induced in some. Its estrogen agonist is noted in the bone, liver, endometrium, and vaginal epithelium, whereas its antiagonist effects are mainly on the breast duct endothelium and the central nervous system. T was shown to block estrogen binding to human tumor estrogen receptors (ER) in a rat and this led to testing breast cancer in women.

In the 1970s and 1980s, T was shown to be effective when used alone or in combination with chemotherapy in the treatment of advanced breast cancer. Proven efficacy was noted in preventing reoccurrences and prolonging survival when used as adjunct therapy in stage I and II breast cancer postoperatively. Because of safety, compliance, and minimum side effects it was then evaluated as a preventive agent in women who were at high risk for breast cancer. This led to the National Surgical Adjuvant Breast and Bowel Project (NSABP) P-1 study that began in 1990 and was reported in 1998.

Raloxifene (R), also a selective estrogen-receptor modulator, was found to have estrogen-like activities that decreased low-density lipoprotein cholesterol and improved bone mineral density with a decrease in fractures secondary to osteoporosis. It was also found to have antiestrogen activity and appeared to be effective in treating breast cancer when used in a large dosage. These attributes were similar to T. This led to the multiple outcomes of raloxifen evaluation trial of T compared to a placebo in postmenopausal women with osteoporosis and subsequent Food and Drug Administration approval for the prevention of osteoporosis. Although not a primary outcome, it was noted that R decreased incidence of breast cancer. The raloxifene use for the heart trial evaluated postmenopausal women with coronary heart disease and found that, compared to a placebo, T decreased the incidence of breast cancer and vertebrae fractures but not the risk of cardiovascular events.

Although both T and R are selective estrogen-receptor modulator with similar activity, the results are somewhat different. The decrease in invasive breast cancer is 49%; 69% in ER-positive cancers in the P-1 study but had no effect on ER-negative cancers. Lobular carcinoma in situ or atypical hyperplasia was reduced 56% and 80%, respectively, as well as noninvasive breast cancer by 50% in the T group. In the multiple outcomes of raloxifen evaluation study there was a 76% decrease in invasive breast cancer and 90% decrease in ER-positive cancers but no impact on preinvasive cancers. In contrast the raloxifene use for the heart study noted a 44% reduction in breast cancer and 55% lower instances of ER-positive cancers. There was no reduction in ER-negative or noninvasive cancers.

These results then led to the Study of Tamoxifen and Raloxifene (P-2) prospective randomized trial comparing T with R in the prevention of breast cancer. There were 310 breast cancers in R patients compared to the 247 in the T group (risk ratio [RR], 1.24; 95% confidence interval [CI], 1.05–1.47). The difference in noninvasive breast cancers was 137 in the R patients compared with 111 in the T group (RR, 1.22; 95% CI, 0.95–1.59).

Killacky et al , in 1985, reported on 3 patients who developed endometrial cancer after having received T. This led to a question of whether the 2 were related. Interestingly, these cancers were diagnosed at 2, 12, and 14 months after starting T. In the P-1 study, 15 women developed endometrial cancer in the placebo group compared to 36 in the T group (27 were postmenopausal). In situ cancers were similar in the 2 arms. The RR was 2.53 (95% CI, 1.35–4.97, 2.3/1000 women). All cancers were stage 1 in the T group and no deaths from cancer occurred. In the P-2 trial there were 37 endometrial cancers in the R group and 65 in the T group (RR, 0.55; P = 0.003). This was an annual rate of 1.23 and 2.5/1000 women, respectively. There have been 2 deaths in each group. Hyperplasia with atypia was present in 22 T patients and 4 R women (RR, 0.17; 95% CI, 0.04–0.51).

The current article in this issue of the journal by Runowicz and associates reviewed the gynecological conditions (adverse effects) in the NSABP P-2 study. Self-reported symptoms noted higher incidence of bothersome hot flashes and vaginal discharge in T group compared with the R women. These symptoms in the P-1 and P-2 of T were similar. None or slight vaginal dryness was similar in both groups, however, in the P-1 study 69% reported no or slight vaginal dryness compared with 50% of T patients in the P-2 study. Other benign conditions such as leiomyomas, ovarian cyst, polyps, and endometriosis were statistically more frequent in the T group compared with R group. Twice as many hysterectomies were done in T compared to the R group; indications for hysterectomies were unknown. Only about 1 in 5 could be related to endometrial cancer.

The authors rightly suggested that differing gynecological effects should be considered when counseling women with an intact uterus when options for breast cancer prevention are discussed. Risks and benefits always enter into our discussion for options and probably even more so for preventive measures. Just exactly how do we use these data? Endometrial cancer is, of course, a concern. Although T had a higher annual rate of endometrial cancer, it should be remembered that the R group had an increase over the placebo normal rate (1.23 vs 0.7/1000 women). Another concern is the fact that there were 34 ovarian cancers in the R group compared to 21 T with 15 total deaths compared with 4 total deaths from endometrial cancers. In the T group there were 63 fewer breast cancers compared to those taking R. This compares to 28 more endometrial cancers but 13 fewer ovarian cancers in the women taking T. As far as these 3 cancers are concerned there were 48 more cancers in the R group. In addition, there were 25 fewer noninvasive breast cancers in T compared to the R group.

Several of the benign uterine conditions in the T group compared to the R group may have been identified because there were twice as many hysterectomies. Endometriosis is, of course, a histologic diagnosis. Were the myomas of clinical significance or incidental findings on pathological evaluation? Are these benign situations a major concern compared to the difference in the gynecological and breast cancers in the 2 groups? Most would probably agree the cancer risk is of greater significance than gynecological symptoms, particularly since many of these symptoms can be effectively managed.