Tamoxifen for the prevention of breast cancer

The complexity of the effects of tamoxifen has recently been summarised: “Tamoxifen is an antioestrogen with complex pharmacology encompassing variable species-, tissue-, cell-, gene-, age- and duration of administration-specific effects, from oestrogen-like agonist actions to complete blockade of oestrogen action. This complexity is consistent with the various, and sometimes paradoxical, effects that have been associated with tamoxifen administration in animals and humans.”¹ The report concluded that there was sufficient evidence that tamoxifen increased the risk of endometrial cancer and conclusive evidence that it reduced the risk of contralateral breast cancer.

Because of this complexity of effects, a consensus has existed until recently that using tamoxifen for preventing breast cancer was questionable even in clinical trials, and that only trials in women at high risk of breast cancer could be justified.² This month, however, investigators from the National Cancer Institute-National Surgical Adjuvant Breast and Bowel Project decided to release the initial results of the Breast Cancer Prevention Trial, based on their assessment that a reduction in breast cancer had been shown in the group receiving tamoxifen (p 1187).³ Should we now conclude that the evidence is sufficient to warrant the widespread use of tamoxifen for preventing breast cancer in high risk women, or do the problems that precluded the preventive use of tamoxifen outside clinical trials⁴ remain unanswered?

The most fundamental question is whether the reduction in incidence will translate into a comparable reduction in mortality. In fact, it would have been surprising not to see any preventive effect of tamoxifen, and the reduction in incidence observed in the trial (45%) is of the size expected from the reduction in the incidence of contralateral cancers in trials of tamoxifen used as adjuvant therapy in women operated on for breast cancer.⁵ There are, however, two reasons to suspect that the prognosis of cancers diagnosed in women taking tamoxifen might be worse than that in the general population. Firstly, the biology of these cancers is likely to be different, owing to the selection of tamoxifen resistant cancers. Secondly, adjuvant tamoxifen is associated with a 25% reduction in mortality, but we don’t know its efficacy in women who were already taking tamoxifen when their cancer was diagnosed. If adjuvant tamoxifen is less effective in these women, or if only well differentiated cancers are prevented by tamoxifen, the reduction in mortality from breast cancer might be less (possibly much less) than the reduction in incidence, with important bearings on the balance between risks and possible benefits of preventive tamoxifen.

A second problem is the duration of tamoxifen treatment. In the B-14 adjuvant trial of the National Surgical Adjuvant Breast and Bowel Project almost twice as many distant recurrences and deaths were observed in patients who were randomly assigned to take tamoxifen for 10 years as in those who stopped after 5 years.⁶ Obviously, these findings relate to the growth of micrometastatic disease and cannot be applied directly to the preventive effect of tamoxifen, but they cast doubts on the long term effects of tamoxifen on the incidence of breast cancer.

The long term toxicity of tamoxifen is also a matter of concern: available data after 5, 10, and15 years of follow up confirm the increase in the incidence of endometrial cancer and of thromboembolic complications and provide some suggestion of ocular toxicity, but these effects are not common and might be more than balanced by the reduced risk of coronary heart disease and osteoporosis. However, with few exceptions, these data are derived from studies using short term treatments (2-5 years), and little is known about the effects of continuing tamoxifen beyond 10 years. Another issue that we know little about is the interaction of tamoxifen with other treatments, including hormone replacement therapy, given concurrently or in sequence.

The point is that if we plan to put healthy women in their 60s, 50s, or even earlier (40% of the women in the Breast Cancer Prevention Trial were aged 35-49) on tamoxifen, we should try to figure out what to do 10 years later, when, hopefully, most of these women will still be healthy but still at risk of breast cancer (at higher risk, indeed, because of age). So far, women in the Breast Cancer Prevention Trial have been followed on average for four years, and, owing to the future likely contamination of the control group, we cannot expect this trial to contribute significantly to resolving this problem.

As a consequence, we need to continue the other two trials (one in Italy and one in Britain) that are currently under way. Their participants should be clearly informed that the uncertainty about the balance between the possible risks and benefits of taking tamoxifen for preventing breast cancer is not substantially modified by the early results of the Breast Cancer Prevention Trial. For the same reason, prescription of tamoxifen for preventing breast cancer outside clinical trials should continue to be discouraged.

A final problem is the definition of the level of risk of breast cancer that could justify, in the future, the start of a preventive regimen associated with costs, side effects, and long term risks. The eligibility criteria used in the Breast Cancer Prevention Trial are very inclusive, since all women aged 60 or older, and younger women with an equivalent (or greater) risk were eligible to participate. On the other hand, we don’t know if tamoxifen is equally effective in preventing breast cancer in all risk groups. For instance, some evidence suggests that some women at very high risk, such as the carriers of a germline BRCA mutation in high risk families, are predisposed to develop hormone independent tumors, which are less likely to be affected by the preventive action of tamoxifen.⁷ This issue might be addressed in the future, when some of the participants in the trials will be tested for germline mutations and by appropriate subgroup analyses, but at present the use of tamoxifen for prevention is not justified even in these women.

News p 1187