Saturday, December 16, 2006

Jane Zones, Board Member

“Cancer is a Preventable Disease,” According to the NCI

Dr. Leslie Ford, who is the National Cancer Institute’s (NCI) Associate Director for Clinical Research, strongly endorsed a broad educational effort to induce women to take drugs to “prevent” breast cancer. As at NCI’s press conference on the STAR Trial last April, she portrayed raloxifene (Evista) as substantially safer than tamoxifen, even though the published data that came out two months later showed virtually no statistical differences between the two drugs.

Most fascinating was her handling of uterine cancer, a dangerous side effect of these drugs. Ford showed that raloxifene caused fewer uterine cancers than tamoxifen, and asserted that the difference was of “borderline statistical significance.” After using that term a couple of times, she acknowledged that the difference between the two drugs did not reach statistical significance. Then she stated that considerably more women in the tamoxifen group underwent hysterectomy during the trial due to symptoms related to tamoxifen, and claimed that, but for this fact, the advantage of raloxifene over tamoxifen for uterine cancer would have been statistically significant.

During her talk, Ford referred to death as a “devastating side effect,” but dismissed the latest findings of the Breast Cancer Prevention Trial, which showed more deaths in the tamoxifen group than the placebo arm after seven years (not a statistically significant difference, but interesting nonetheless), as too few to make any difference. (A later speaker, reporting long-term follow up of tamoxifen users versus placebo in the IBIS-I trial, showed similar differences—65 deaths in the tamoxifen group, compared to 55 deaths in the placebo group.)

Ford said that alterations in the study design (women on placebo were invited to go on tamoxifen after an average of five years when the study was unblinded early) confounded the data so that you could not use overall survival as an outcome measure. Of course, the decision to unblind the trial was a decision made by the trial designers. In Britain, the IBIS-1 trial continues as unblinded 10 years out, permitting evaluation of overall survival.

Ford claimed that 55 women need to be treated with tamoxifen to prevent one case of breast cancer1 and went on to make a case that this is similar to taking statins or beta blockers, which require 40 to 45 users to prevent a single cardiac event. However, this analogy does not consider the known hazards of tamoxifen and raloxifene, which are considerably greater than known risks of cardiovascular prevention drugs, which have been studied for much longer periods of time.

Ford appeared indignant that a previous speaker had attributed the 2003 decline in breast cancer incidence to widespread discontinuation of hormone therapy in postmenopausal women, rather than to use of chemoprevention drugs .

It was frightening and disheartening to see the top management of the NCI so wedded to this approach, which is not only unpopular with healthy women, but has been shown to cause considerable harm, and is not being studied for long enough periods to show overall survival benefit.