Tamoxifen for the primary prevention of breast cancer: a review and critique of the concept and trial.

This paper has reviewed the rationale for and design of the NSABP Breast Cancer Prevention Trial and has provided a brief critique of the philosophy of chemoprevention of breast cancer and of certain practical aspects of the trial design. If the assumptions and estimates from the trial protocol are correct, the net benefit of the trial will be moderately large and positive (50 to 77 events prevented). However, over 500 women per year will be treated with tamoxifen unnecessarily. Recalculation of the net-benefit table using another set of reasonable assumptions regarding risks and the trial assumptions regarding benefits shows a negative effect to a small positive effect overall (-17 to 8 events). If the probability of adverse ocular events is included in the net-benefit equations, it appears that more harm than good will result from the intervention. In the face of the uncertainty regarding the net benefit of the trial, ranges of these risks and benefits should be provided to both potential and enrolled participants. The lack of significant benefit to participants seen with the recalculations may raise the question of whether the trial should continue as designed. One option would be to limit trial participation to postmenopausal women only, since 1) breast cancer is more common in postmenopausal women, 2) tamoxifen is more effective in postmenopausal women, 3) cardiovascular disease is more common in postmenopausal women, and 4) reductions in cholesterol levels and preservation of bone mass have only been documented in postmenopausal women (11, 27). Even in this case, however, the fundamental philosophical question of whether large numbers of healthy women should be "treated" with a toxic drug for the primary prevention of a rare event remains.