Photo: The Thermogram Center; www.thermogramcenter.com.

One of the more newsworthy health care stories of late was the widely reported National Cancer Institute data showing a major 6.7% decrease in breast cancer rates among US women in 2003.

Much of the media coverage, especially in “alternative” medicine circles, took a “See, we told ya so,” stance, attributing the downturn in breast cancer to American women’s widespread rejection of hormone replacement following the Women’s Health Initiative (WHI) reports in 2002. Some health reporters viewed the new NCI figures as proof-positive that HRT causes breast cancer, and that HRT cessation will prevent it.

That breast cancer dropped by nearly 7% is certainly good news, and while it is tempting to attribute it to changes in HRT, I’m not sure it is quite so simple.

The data from 2004 showed a leveling off, with no real additional decrease after the sharp downturn in 2003. That decrease, starting in mid-2003, was only in women who were 50 years of age or older, and was more evident for estrogen-receptor-positive cancers (14.7%; 95% CI, 11.6 to 17.4). The decreases were similar for localized disease and more advanced disease, and were more evident in primary cancers but not in contralateral second primary or later breast cancers (Ravdin P, et al. Breast Cancer Res Treat. 2006; 100: Suppl S2).

Changes in incidence between 2001 and 2004, but not including this rapidly changing period from mid-2002 to mid-2003, also shows an annual decrease in among women aged 50 or older. But during that time period, there was an increase of 1.3% in breast cancer in women under age 50, a decrease of 11.8% for women between 50 and 69, and a decrease of 11.1% for women 70 and older (Ravdin P, et al. N Engl J Med. 2007; 356(16): 1670–1674).

The compelling question is, why this sharp drop, followed by a stabilization at a lower rate in women 50 and older?

There are several plausible explanations: 1) A decline in hormone use since the first WHI report, clearly the preferred explanation in recent media reports; 2) A decrease in the rate of screening mammograms in general; 3) A possible decrease in annual breast exams in women over 50 who chose to stop HRT; 4) An error in the NCI Surveillance, Epidemiology, and End Results (SEER) data; 5) A general decrease in overall cancer rates of cancers; or 6) Some other positive influence on breast health and breast cancer risk reduction.

Reporting flaws in the SEER data is the least likely explanation, since there are no reports of any significant change in incidence of other types of cancer besides breast cancer during this period. In addition, all nine SEER registries showed the same trend; a uniform error across nine of the most accurately kept disease registries in the world is theoretically possible, but highly unlikely.

The big downturn could be related to a major decrease in screening mammography. For 2003, there was indeed a 3.2% decrease in screening mammograms among women aged 50–65 years, compared with the year 2000. There could also be a change in the frequency and pattern of screening mammograms in women who formerly used HRT, and now do not.

It is true that women who receive HRT are likely to also receive annual mammograms. Women who discontinue HRT, who are 50 and older, may also be disinclined to get annual breast exams, especially now that they are being told they do not need an annual Pap smear. So, once women discontinue HRT, might they also discontinue their annual mammograms? Might they be going to the doctor less frequently? This is a real possibility, and if so, it would mean, basically, that there is no actual downturn in incidence, but that many breast cancers that might have been discovered went undetected because the women weren’t going for their annual exams.

Although I do think that, in fact, women who stop HRT may initially delay annual screening mammograms and doctor visits, there are no published data showing a decrease in screening specifically in women who stop HRT. We really do not know what impact HRT cessation has on breast cancer screening and detection.

The observed decrease in incidence began in mid-2002, and occurred shortly after publication of the first WHI report in July, 2002, which demonstrated a slight increase in the risk of breast cancer among women who had used HRT for 4 years or more (Rossouw J, et al. JAMA. 2002; 228: 321–333).

By the end of 2002, use of conventional HRT declined by approximately 38% in the U.S. and there were 20 million fewer prescriptions written in 2003 than in 2002 (Buist D, et al. Obstet Gynecol. 2004; 104: 1042–1050; Hersh A, et al. JAMA. 2004; 291: 47–53).

There was a steep decline in prescriptions for Premarin and PremPro, the two most common forms of HRT, starting in 2002, and especially in 2003. There were 62 million prescriptions written in 2000, 61 million in 2001, 47 million in 2002, 27 million in 2003, 21 million in 2004, and 18 million in 2005 (Drug Topics. Drugs by unit in the United States in specific years. www.drugtopics.com/drugtopics/). The period of sharpest decline appeared to start in 2002 and accelerated in 2003.

Reduced HRT use could have had some impact, but it is difficult to attribute the major breast cancer downturn in 2003 to a reduction in HRT that only started several months earlier. Development of breast cancer does not happen that fast, nor would its prevention, on a population basis.

What about other medical or lifestyle factors that might have influenced a downturn in breast cancer rates? Was there something else that emerged in the period from 2000 to 2003? Use of drugs such as tamoxifen, raloxifene, nonsteroidal anti-inflammatories and statins was certainly widespread during this period, and there is some evidence that these medications can reduce overall risk of breast cancer. However, there were no major increases in use of any of these from 2000 to 2004 as compared to previous years.

Wider use of vitamin D, green tea or soy products also deserve some thought. There are data showing potential positive influences of each of these on breast health and even in reducing breast cancer. But again, we do not have any data showing increased utilization of these products in this period of 2002 to 2004.

The bottom line is, a thorough look at the new NCI data as well as other breast cancer trend data make it difficult to draw a definitive conclusion about what caused the observed drop. We’ll simply have to wait for further studies.

Women who were in the PremPro arm of WHI when that study was discontinued are being followed for clinical outcomes. A report on this is expected later this year, and could shed additional light on how discontinuation of HRT affects incidence of breast cancer.

Experts on this topic are hesitating to render any final opinion until that report is published, while also admitting that “the ultimate understanding of the effect of cessation of hormone-replacement therapy will be complex; it will probably depend on more than one mechanism and will be affected in different ways by various forms of postmenopausal hormone-replacement therapy.”

More time and study could answer another interesting question: Will broad cessation of HRT result in the delayed appearance of clinically detectable tumors but not an actual long-term reduction in breast cancer incidence? Dropping HRT may only slightly or temporarily slow the growth of small not-yet-detected tumors. If this is the case, then as use of HRT stabilizes again at a certain utilization rate, the incidence of breast cancer would likely rise again at some point in the future.

It is my humble opinion that since the WHI and other studies show only a slight increase in breast cancer after combined conjugated equine estrogens and progestins for 5 years or more, any decrease in breast cancer seen immediately after WHI probably reflects a delayed growth of already pre-existing breast tumors as a result of the decreased use of HRT.

Dr. Leon Speroff, an expert in the subject of breast cancer and HRT, reminds us of a very important feature of this latest data: the short latent period between when hormone therapy was discontinued, and the reduction in prevalence of breast cancer. He asserts that case-control and cohort studies show an increase in breast cancer risk only in current hormone users, and a rapid reduction after hormone cessation. These reports are consistent with statistics in Switzerland where the peak of breast cancer increased in women ages 60–64. The increase occurred only for Stage I and Stage II disease and for estrogen-receptor positive tumors in women who used hormones (Bouchardy C, et al. BMC Cancer. 2006; 6: 78–85). These hormone effects are consistent with other reports of better outcomes and better-differentiated tumors in hormone users (Gertig D, et al. Breast Cancer Res Treat. 2003; 80: 267–273. Pappo I, et al. Ann Surg Oncol. 2004; 11: 52–58).

Another interesting observation is that no studies have revealed any increase in in-situ breast cancers among hormone users. If hormones were causing new breast tumors, we would expect to see this.

These are complicated and even uncomfortable issues, especially for alternative-minded clinicians. I think our bias is to blame hormone use for any negative effect on breast cancer. On the other hand, we should be alert to bias in favor of hormone use. If Dr. Speroff and the research he cites is correct, then tumors that emerge later may be of higher stage and grade, and consequently poorer outcomes. Is it possible, that by leading to earlier diagnosis of early-stage, well-differentiated tumors, hormone use could actually have a favorable effect on breast health? It’s certainly something to think about.