A Drug That Needs Further Testing

The FDA has approved the administration of tamoxifen to prevent breast cancer in women who have never had the disease but who are at high risk of developing it. The decision is based upon a single study, prematurely terminated last March.

At that time, sponsors of the study, called the Breast Cancer Prevention Trial, issued a joint statement announcing that any additional information that could be gained from continuing the study did not outweigh the benefits of making the treatment available to the participants in the placebo group and other women at high risk of developing breast cancer. In short, they felt, tamoxifen was so effective in preventing breast cancer that the researchers could not in good conscience keep this information from the study participants or the public.

Because this conclusion will significantly affect the lives of women over the country, we should all be aware of what this trial really found and what is still unresolved. I believe that the FDA’s approval was premature, and I am worried about the forthcoming advertising blitz, announced by the manufacturer of tamoxifen, to persuade healthy women that this drug is safe and appropriate.

The Breast Cancer Prevention Trial began in 1992, to test whether tamoxifen, a drug shown to help prevent the recurrence of breast cancer in some patients, would also prevent it from developing in the first place.

Accordingly, 13,175 women identified as being at high risk of developing breast cancer were randomly assigned to receive tamoxifen or a placebo; they were followed for an average of four years. On Sept. 16, the findings from this research were published in a medical journal, where the work could be scrutinized for the appropriateness of its methods and for the validity of its conclusions.

Because the study was terminated early, there are relatively few cases for comparison: 175 of 6,599 (2.7%) women in the placebo group versus 89 of 6,576 (1.4%) women in the tamoxifen group developed invasive breast cancer. So we are talking here about only a 1.3% difference in occurrence. This difference has been reported, however, as a 48% decrease in risk of development. That sounds impressive, but it is a statistical gimmick to heighten the appearance of a meaningful difference; 1.4% of the women in the tamoxifen group who got cancer is indeed nearly half of the 2.7% in the control group who got it, but the small absolute number of cases (89 versus 175) makes this comparison preposterous, especially when we consider that 212 patients dropped out of the study and were lost to follow-up.

Second, there is no blood test to confirm whether a woman is actually taking tamoxifen. Because the side effects of the drug--including hot flashes and vaginal irritation--cause many women to stop taking it, we cannot be certain what percentage of the women in the tamoxifen-treated group were complying with the regimen.

In a parallel study from Italy, about 26% of the participants dropped out of the study, most within the first year. In a related study from England, 42% of participating women discontinued the tablets they were taking. In the American study, 23.7% of the women were reported to have stopped taking their tamoxifen. That comes to 1,559 patients included in the analysis who cut short the period of time they took tamoxifen. Yet once assigned to take tamoxifen, study subjects were analyzed as if they had faithfully taken the drug.

Third, one of the goals of the study was to determine the effect of tamoxifen on the death rate from breast cancer. Obviously, this will never be accomplished now that the study has been cut short. Only nine deaths in the entire study population were attributed to breast cancer. We do not know the long-term effect on the rate of breast cancer deaths resulting from giving tamoxifen to apparently healthy women--and this concern is separate from the known risk of increased uterine cancer and blood clots in women who take it.

Finally, most important, this single study’s findings have not been replicated and confirmed. On the contrary, they have been disputed by the English and Italian studies of many thousands of women, which found no protective effect of tamoxifen. That is why other countries have been far more cautious than the United States about endorsing tamoxifen as a breast cancer preventive.

There are other questions that should be answered. If tamoxifen really has a protective effect, how long should the medication be taken as a preventive? At what age should it be started? How does hormone replacement therapy fit into the picture?

The American study may be promising, but it is premature to accept its conclusions as the basis of valid clinical practice, let alone as the basis of national policy and guidelines.