Tuesday, June 15, 2010

Breast Cancer Prevention


New data from long term follow-up of women participating in the STAR trial, a study comparing Tamoxifen and Raloxifene (known as SERMS) for prevention of breast cancer, show strong and persisting benefits of reduced breast cancer risk after stopping therapy (see abstract). This is an important addition to our understanding. Based on follow-up of 18,747 women in the STAR trial the updated results highlight the value of extended follow-up of participants in trials to better understand the overall impact of interventions for cancer prevention. We summarized these key questions for cancer prevention in an earlier blog posting and in a recent article.

Moving from the results of the STAR trial to implementing prevention requires that we have an approach to identify women who are at increased risk of breast cancer. While numerous models have been developed to prediction of breast cancer, it is worth remembering that any of these models does not tell us who will get cancer and who will not, but rather we can place women according to their level of risk, much above average, above average, average, etc. This is the approach we use in Your Disease Risk and is based on evidence from communication science about understanding of risk. For decisions regarding use of a drug to prevent cancer, the challenge is balancing the risks of side effects against the known benefits. The higher the risk group taking the preventive agent the stronger the balance of benefits exceeding risks.  This balance, of course, applies throughout medicine whether we talk about drugs, surgery, or other medical interventions.

Take for example the strong genetic predisposition to breast cancer that is present in women who carry the BRCA1 or BRCA2 mutation. Lifetime risk in these women is around 50 to 60% so among those with the mutation we have about a 50/50 chance of knowing who will get breast cancer. On the other hand the combined data from 10 studies now clearly show that for these high risk women bilateral oophorectomy halves their risk of breast cancer (see article).

To bring results from a randomized trial of efficacy to the patient population we must weigh risks of side effects against the expected benefits of therapy for individual patients (Glasziou and Irwig set out an approach for evidence based individualized treatment).  We previously estimated a scenario to weight risks and benefits under the assumption that we wanted to prevent more breast cancers than major medical events that might be caused as side effects from taking drugs, see article. We show that among the 26 million women in the US between ages 50 and 69, 7.8 million have risk above a break even point and that over 22,000 cases of breast cancer can be avoided each year by using a SERM in this group. The new data add a refinement to these estimates as the benefit of taking a SERM persists after stopping and the side effects are reduced after stopping. This tips the ratio of benefits to risks more strongly in favor of taking drug to prevention breast cancer. The overall reduction in risk remains at about 50%, or halving the risk of breast cancer during the years the drug is taken and this benefit continuing after stopping.

What are the side effects? Tamoxifen causes endometrial cancer while Raloxifene does not. Pulmonary embolism and deep-vein thrombosis are now well recognized side effects of Tamoxifen. Raloxifene had significantly lower rates of these events in the STAR trial. Overall 2.47 thromboembolic events were reported per 1000 women per year among those taking Raloxifene. One approach to estimating benefits and risks assumes women have a 4% risk of breast cancer over the next 5 years. See related report.  The benefits and risks are summarized below for 1000 women over 7 years

Raloxifene
Tamoxifen
Invasive breast cancer prevented
15
20
Noninvasive breast cancer prevented
16
20
Endometrial cancer caused
0
2.25
Thromboembolic events caused
2.47
3.3
Benefit to risk ratio
31 vs. 2.47 or 13:1
40 vs. 5.55 or 7:1

If we rerun these numbers using a lower cut point for the 5-year risk of breast cancer, the benefit to risk ratio will come closer to 1.

Of course the importance placed on side effects from SERMS may differ among women. In our estimate, we assumed a deep vein thrombosis (a side effect of SERMS) was equal to a prevented breast cancer. Other options may apply in individual decision making. Tools to help women and their health care providers balance risks and benefits may soon help lay out these considerations.

No comments:

Post a Comment