Skip to main content


The BCERF program on the Cancer Risks of Environmental Chemicals in the Home and Workplace closed on March 31, 2010. No further updates will be made to this web site. Please go Cornell University’s eCommons web site to access BCERF’s archived research and educational materials (

Vol. 05 Issue 4, Late Fall 2000

Research Commentary: Decision Analysis of Tamoxifen for the Prevention of Invasive Breast Cancer
The Ribbon 

Grann VR, Sundararajan V, Jacobson JS, Whang W, Heitjan DF, Antman KH, Neugut AI. (Herbert Irving Comprehensive Cancer Center, Department of Medicine, College of Physicians and Surgeons, Columbia University, New York) Cancer Journal of Scientific American 6(3):169-78, 2000.

Seema A. Khan MD, Associate Professor of Surgery, Northwestern Memorial Hospital

Women who are at increased risk for developing breast cancer and have completed child bearing now have the option of taking tamoxifen for five years in order to decrease their breast cancer risk. However, for many women this is a difficult decision to make, since tamoxifen does have some adverse effects, and the risk/benefit analysis is not straightforward. In an attempt to clarify some of these issues, researchers from Columbia University have developed a decision analysis model to aid physicians who are counseling women at increased risk of breast cancer regarding preventive therapy with tamoxifen.

The researchers have used a mathematical model to predict overall benefit in terms of quality adjusted survival for three different age groups of women: 35-49, 50-59, and over 60 years. They have made allowances for the side effects of tamoxifen, and the cost of tamoxifen therapy, as well as the cost of treatment of any complications. Quality of life estimates for these calculations were derived from a recent study where women were asked how much added life-time they were willing to trade for time spent in three states: taking chemopreventive medication, diagnosis of and treatment for invasive breast cancer, and time living with metastatic breast cancer.

The model predicted that the greatest preventive benefit of tamoxifen would be seen in women who start treatment early in life (the 35-49 year age group). The actual time gained in this age group was surprisingly small (69 days), but this number needs to be interpreted with the understanding that in models such as these, the benefit of a particular treatment is actually being averaged across a large number of people who are being subjected to the treatment. When we consider individuals, the benefit will be large for a small minority (those who would have developed breast cancer, but did not because of preventive therapy), and zero for the rest (women who would never have developed breast cancer, with or without tamoxifen, and those who develop breast cancer despite tamoxifen).

In fact, most women will not benefit from preventive tamoxifen, because most women would not have developed breast cancer in their lifetimes even without tamoxifen. For example, the threshold for recommending preventive treatment with tamoxifen is a breast cancer risk of about 2% over five years. An 80-year-old woman belonging to this risk group, might have a lifetime chance of developing breast cancer of about 5%, and 95 out of 100 women in this age group, with this risk level, will die of other causes. On the other hand, a 40-year-old tamoxifen eligible woman with a five-year breast cancer risk of 2%, will have a lifetime risk of 25%. It is easy to see from these figures that a woman who is at high enough risk to be eligible for tamoxifen at a young age actually has a much higher lifetime risk for breast cancer than an older woman, even though their short term risk might be similar. In risk benefit calculations, the benefit of the treatment is usually found to be proportional to the risk of disease, and the model used in this study again validates this general principle, showing us that women at higher lifetime risk derive greater benefit.

The second part of the calculation of the risks and benefits of tamoxifen has to do with the adverse effects that might be experienced by women taking tamoxifen. Since the serious side effects of tamoxifen (uterine cancer, stroke, and clots in the deep veins of the legs which can travel to the lungs) become more common with age, older women who take tamoxifen have a higher chance of suffering serious side effects from tamoxifen, but have a smaller benefit as we have seen above. Thus when the higher risks of therapy are balanced against the smaller benefits in older women, the overall gain is small. For young women on the other hand, the benefits are larger, the risks are smaller (women in the tamoxifen arm of the BCPT did not suffer the adverse events associated with tamoxifen use at significantly higher frequency than women in the placebo arm of the trial) and the overall balance is in favor of using tamoxifen for breast cancer prevention.

The authors then looked at the cost of using tamoxifen for breast cancer prevention, and the costs of treating serious side effects, and estimated the cost per life year saved by tamoxifen use. Again, because the benefit is larger in younger women, and the likelihood of side effects is smaller, the cost per life year saved was smaller for younger women.

The results of the model were varied using different estimates of the duration of the beneficial effect of tamoxifen in terms of breast cancer protection. The available data from several different analyses suggest that this protective benefit outlasts the actual duration of tamoxifen use by at least 5 to 10 years, and perhaps longer. The authors redid the calculations assuming a 5, 10, and 15 year duration of tamoxifen benefit after stopping therapy. They found, naturally, that increment in longevity was greater, and the cost per year of life saved smaller, in all age groups as one assumes increasing duration of benefit. Again, the actual numbers they derived are used only for purposes of illustration and do not have any real meaning if they are applied to individuals. Assuming a 15 year duration of benefit after stopping tamoxifen, the quality adjusted survival (i.e. accounting both for the increased lifespan of women benefiting from tamoxifen and the effect of adverse side effects on this gain) is illustrated as follows:

Assuming a 15 year duration of benefit after stopping tamoxifen, the quality adjusted survival (i.e. accounting both for the increased lifespan of women benefiting from tamoxifen and the effect of adverse side effects on this gain) is 105 days for a woman starting tamoxifen at age 35, 66 days if starting at age 50, and 45 days if starting at age 60. The mean cost per quality adjusted life year saved is about $19,000, 41,000 and 67,000 respectively for each of these groups. These costs are comparable to other life-extending interventions, such as mammography, but the cost benefit ratio is substantially less favorable for older women.

The advice that women who are considering tamoxifen can take away from this cost benefit analysis Ñ which echoes the conclusions drawn from other analyses Ñ is that women over the age of 60 or 65 have in general a lower expectation of benefit, at higher cost, from the use of tamoxifen for the prevention of breast cancer. In this age group it may still be advisable for women to take tamoxifen if they are at substantially increased risk of developing breast cancer, but the present threshold of a five year risk of 2% may not be sufficient to warrant the quality of life and financial cost of tamoxifen.

Back to the top