“One pill makes you larger, and one pill makes you small…”

Although Jefferson Airplane’s Grace Slick was writing about something else entirely in 1967, we had an unusual coincidence this week when two events were publicized widely regarding breast cancer risk alterations due to pills. One pill raises your risk for breast cancer, while another makes risk smaller.

First, the pill that makes risk larger. Birth control pills (and hormonal-release IUDs). Treated as if this were the first study of its kind, a very small risk for breast cancer was announced in terms that made it sound larger. A “25% risk increase” is a “relative risk” not an absolute risk. I’ve been writing about this “relative risk vs. absolute risk” for exactly 23 years (introduced in the first lay book on risk assessment in 2000). First of all, I would point out that this controversy has already had hundreds of studies published, and there are so many caveats, I won’t even bother to walk through the story. However, some of the studies show no risk at all (depending on the preparation — high progesterone pills or low) or a slight risk. Either way, the risk is so small that it should not alter recommendations to patients beyond awareness. Besides, birth control pills lower the risk of both ovarian and endometrial cancer.

As for the “relative” vs. “absolute,” I won’t bore you with the math. Just know that the media (and the researchers) much prefer to discuss their results in “relative” terms because the number is ALWAYS larger. In this case, the researchers broke ranks and actually took the time to explain how a 25% relative risk increase is a big pill to swallow, and when couched in absolute risk terms, the reality is much much smaller. For practical purposes, it’s not a major issue (unlike postmenopausal hormone replacement risk, which can generate some concerning numbers, albeit far below most other risk factors).

And now, for the pill that makes risk smaller. The big announcement came this week, again presented in relative terms. But in this case, when converting to absolute benefit, it’s a compelling option. The story dealt with the National Health Service in the United Kingdom endorsing the use of anastrozole (Arimidex) to lower the risk of breast cancer “by 50%.” First, let’s convert this to absolutes. If your absolute risk for breast cancer is 20% lifetime, and you cut that in half, there is an absolute chance of 10% for major benefit (the benefit being never developing breast cancer). If your baseline risk is a higher 50% lifetime, then the absolute benefit is a lowering of your personal risk from 50% to 25%. That is, a one-in-four chance that the pill you took for 5 years kept you from getting breast cancer.

Taking a pill to prevent breast cancer is really a remarkable option when you think about it. Here’s how the discovery was made: When tamoxifen was new on the market (1970s), its use was to take the place of hormone-lowering surgeries (oophorectomy, adrenalectomy, even removal of the pituitary), primarily in the treatment of metastatic disease. Then, its indications were expanded to use as an adjunct treatment for breast cancer patients to lower the risk of metastatic disease later on. THEN, someone noticed that when used in the adjunct setting, women who took tamoxifen were not getting breast cancers in the opposite breast as much as would have been calculated. In fact, the chances of getting contralateral cancer later on, were cut in half for those women who completed 5 years of tamoxifen. NEXT STEP — multiple prevention trials internationally were performed in the 1990s confirming that giving tamoxifen to healthy women lowered risk by one-third to one-half.

Most unusual, however, was the fact that the benefit increased over time — that is, AFTER the drug was stopped. One trial failed to reach statistical significance while the women were on tamoxifen, but as the years went on, subsequent to the discontinuation of tamoxifen, statistical significance was reached! Theories abound as to how this is possible, but the point is….if a person decides to partake in risk reduction using a pill, you can calculate the benefit for the next 20 years even though they only take the drug 5 years.

As the design of the “P-1” trial from the NSABP prompted lively discussions, the greatest controversy was the inclusion of premenopausal women (some other trials included postmenopausal women only). Many of us predicted it would not work in this group of younger women. We were wrong. Not only did it work (not quite as well as in postmenopausal women) but it became the risk-reducing drug of choice for younger women as a result of the favorable side effect profile. Older women were faced with an increased risk of blood clots and endometrial cancer when using tamoxifen for prevention, but the premenopausal women did not have to deal with these issues, generating pretty much an ideal drug, considering we’re recommending this for healthy women (recalling “do no harm”).

As an incidental point, the Gail risk assessment model was introduced to the public when it was chosen as the means to calculate risk for entry to the NSABP P-1 tamoxifen prevention trial. It was validated as accurate by the P-1 trial, at least when it comes to predicting the number of breast cancers that would occur in a high-risk cohort. (At the individual level, however, the initial version of the Gail model was only slightly better than flipping a coin.)

To clarify the nature of these drugs: tamoxifen and raloxifene are both SERMS (selective estrogen receptor modulators) working as antiestrogens in some body locations (breast, by blocking estrogen receptors) and as estrogen in other locations (bone). And when it comes to the uterus, there are differences — tamoxifen acts like an estrogen in postmenopausal women while raloxifene is neutral. Thus, the name SERMs which implies, “estrogen-like in some locations, but anti-estrogen in others.”

As for the aromatase inhibitors (there are more than just the 3 we all hear about), they are pure antiestrogens, everywhere. Circulating estrogen is diminished across the board, rather than receptor blockade. Thus, the greatest concern is the lowering of bone density, sometimes requiring treatment. There is also a common side effect of aching joints, or even non-specific aching, of uncertain mechanism.

Fast forward to today’s options for “pharmacologic risk-reduction.” (The original term was “chemoprevention” but that word stuck in patients’ throats, scaring off candidates, so the name was changed.) Tamoxifen remains the drug of choice for premenopausal women. Raloxifene (Evista) is the drug of choice for postmenopausal women (not quite as powerful as tamoxifen, but no endometrial cancer, only a slight risk of blood clots, and beneficial for bone density). And if more risk reduction is warranted than Evista can provide, the aromatase inhibitors can be used for prevention. The story this week was anastrozole (Arimidex), but the others work as well, offering an approximate 50% risk reduction if the patient can take the drug 5 years (yes, there’s some protection with shorter treatments).

Is this new info? No. These studies were completed a long time ago and FDA approval followed for the SERMs. I have many patients who completed their tamoxifen, or their raloxifene, or their aromatase inhibitor, and now have protection (a lower level of risk) for at least 20 years, maybe more.

Now here’s the Kicker, and the reason I became heavily involved in this arena. Analysis of participants in the P-1 trial revealed a subset where the benefit was even greater than those women at high-risk due to a positive family history. The most dramatic result was in the women with “atypical hyperplasia” on a prior biopsy where a 90% risk reduction was realized. This launched an entire industry aimed at “searching for atypical hyperplasia” to identify the best candidates for pharmacologic risk reduction (thus, ductal lavage, nipple aspirate fluid, random FNA cytology, ductoscopy, etc). After all, 90% risk reduction is close to the same benefit as preventive mastectomies.

As the data from subsequent trials emerged and matured, the benefit was adjusted to a 75% risk reduction range, which is still remarkable. Nevertheless, given the large number of candidates for pharmacologic risk reduction, few women opt for the 5-year plan. In fact, the “industry” sort of dried up, not because the science was off, but rather, the lack of patient interest in pharmacologic risk reduction, even from those at high risk, even if submitting to any of the measures to find “atypia.” No point in performing ductal lavage if the patient is going to decline pharmacologic risk reduction. Over 500,000 women take Evista for its other FDA-approved indication — bone density improvement — but try suggesting it for breast cancer risk reduction, and patients balk. The NSABP even launched a study to understand why women didn’t buy into the NSABP’s huge and expensive studies (13,000 women in P-1 and 20,000 in P-2). It largely remains a mystery as to why pharmacologic risk reduction has been underutilized, but it’s still available through “shared decision-making” between the patient and her health care provider.

Maybe the reluctance has something to do with next line of lyrics in “White Rabbit”….. And the ones that mother gives you don’t do anything at all.

END