More Data Needed on Hormone Use and Breast Cancer Rates
Cancer registries around the country are sending NCI their most recent statistics, and all eyes will be watching to see whether the incidence of breast cancer declined in 2004 for a third straight year, ending a rise that began in the early 1980s.
Epidemiologists noted last year in the Annual Report to the Nation on cancer that incidence rates leveled off in 2002 and 2003, but they wanted another year of data before deciding whether the change was a true reversal of a trend or a random fluctuation.
"We look at trends, and in general we like to see 3 consecutive years," says Dr. Brenda K. Edwards of NCI's Division of Cancer Control and Population Sciences (DCCPS), the senior author of the Report. "2004 will give us what we need."
If the decline persists (an answer is expected in April), Dr. Edwards and many others will be asking why the rates went down. The likely answer is multiple reasons, including screening and prevention programs.
But a growing number of researchers are now considering the declining use of hormone replacement therapy as a factor in breast cancer incidence.
Millions of women stopped taking the drugs to treat menopausal symptoms after a large national study, the Women's Health Initiative (WHI), reported in July 2002 that certain hormones increased the risk of breast cancer and also heart disease.
Many clinicians have wondered whether the declining use of hormones might eventually influence breast cancer rates.
Last month at the San Antonio Breast Cancer Symposium, researchers said the answer was yes. They presented a new analysis of incidence data that showed a dramatic overall decline of 7 percent between 2002 and 2003.
The largest decline - 12 percent - occurred in women who had estrogen receptor-positive breast cancer. This type of cancer may depend on hormones for continued growth.
The only plausible explanation for the results was a decrease in the use of hormone therapy that happened at about the same time, the researchers said.
Dr. Peter Ravdin of the University of Texas M.D. Anderson Cancer Center presented the results and a scientific abstract (a study has not yet been published).
"When we saw that breast cancer rates had gone down, we looked at various risk factors and the role of screening, but hormone therapy stood out," says Dr. Kathy Cronin, a mathematical statistician in NCI's DCCPS and co-author of the abstract.
"Screening might still play a role as well, and of course we're waiting for the 2004 data," Dr. Cronin adds.
The researchers caution that epidemiological data cannot show cause and effect.
Their hypothesis is that the effect of taking away the hormones was to slow the growth of some tumors so that many small tumors went undetected when women went in for their mammograms.
"No one thinks stopping hormone therapy is preventing the initiation of breast cancer, but everyone feels that it is slowing the growth of tumors by taking away some of their fuel," says Dr. Donald Berry of M.D. Anderson, who led the research.
"All we've seen so far is this precipitous drop in incidence," Dr. Berry adds. It is possible that small tumors not detected in 2003 might eventually grow and be detected later.
He believes the future incidence data will offer clues about what happens biologically when hormones are stopped - whether the effect is to slow tumor growth, to stop growth, or potentially even to cause tumors to regress.
If incidence rates remain low for several years, that is pretty good evidence of a substantial slowing of tumor growth and maybe even stopping, says Dr. Berry.
If, on the other hand, rates go back up and other risk factors remain constant, then stopping the hormones is probably only slowing growth.
In the coming months, the researchers will use statistical modeling to assess the relative contributions of hormone therapy, screening, and other factors on sharp declines in incidence rates. "That's what we can do right now," says Dr. Cronin.
The modeling can simulate the effects of stopping hormone therapy for the population, and a resurgence of hormone use could be factored in, should one occur.
Modeling could also address the fact that after hormone therapy is stopped, a woman's breast density changes, and small tumors are easier to detect by mammography.
The effects on mortality, if there are any, would not be seen for many years and would be difficult to assess given the complexity of the disease, says Dr. Berry.
"This is an intriguing hypothesis, and I would guess that the declining use of estrogen plus progestin certainly plays a role," says Dr. Leslie Ford, associate director for clinical research in NCI's Division of Cancer Prevention (DCP).
"But I don't think we can totally rule out other factors such as screening and the preventative effects of tamoxifen and raloxifene," says Dr. Ford. "It would be too neat a story to say that we announced the results of WHI and breast cancer went away. It's just not that simple."
The picture is certainly complicated. Some women, for instance, may have gone off hormones and then come back on later, but at a different dosage.
Furthermore, any analysis would have to focus solely on estrogen-plus-progestin therapy because estrogen-alone therapy is not associated with an increased breast cancer risk.
Prescription rates for both types of hormone therapy fell rapidly after the WHI results appeared. The new analysis suggests that incidence rates were affected almost immediately.
The short time frame is surprising but may be consistent with what epidemiological studies have shown, according to Dr. Christine Berg of NCI's DCP, who has been treating women and studying the effects of hormones on breast cancer for 25 years.
She points to a study, published in The Lancet in 1997, that says the risks of hormone therapy decline within 1 to 4 years after the therapy is stopped. By the fifth year after stopping, the risk is back to where it was before the hormone use began.
Given the preponderance of evidence on the adverse effects of hormones on breast cancer and also vascular disease, women who are still taking hormones should consider stopping them, says Dr. Berg.
"Physicians need to discuss this increasing body of evidence with their patients," she says.
The message for women is to avoid estrogen-plus-progesterone replacement therapy, adds Dr. Jo Anne Zujewski, who oversees breast cancer trials for NCI's Cancer Therapy Evaluation Program.
"We have made this recommendation before because there are serious risks and there are not long-term benefits," says Dr. Zujewski. If hormone therapy is recommended for debilitating menopausal symptoms, the duration of use should be limited, she says.
Several weeks before the San Antonio meeting, researchers in California published their state's incidence statistics for 2004 in the Journal of Clinical Oncology after the data were finalized early.
At least in California, the decline in 2003 rates persisted for another year and could not be explained by variables such as screening.
"Since the national data won't be released until later this spring, we thought it was important to show that in California the drop continued in 2004," says Dr. Christina Clarke of the Greater Bay Area Cancer Registry and Stanford Comprehensive Cancer Center.
Her team looked at data on mammography rates from Kaiser Permanente to see if this might explain the drop, but screening rates were stable "any way you look at it," she says.
"I don't think there's a study design that will allow us to look at this phenomenon and answer the question of causation," says Dr. Clarke. "We're going to continue to watch the trend and think about it and rule out possible factors as we can."
Dr. Berry predicts that the national data for 2004 will show a plateau, and he is more interested in what might be learned from 2005 and 2006 data. But he has confidence in the current hypothesis.
"People say that now we have to wait and see what happens, but that's not necessary in my view," says Dr. Berry. "We have very compelling information that withdrawing the hormones is causing the lion's share of the effect."
"The future information will tell us about the biology of the effect, but the effect itself is very clear," he adds.
Dramatic shifts in cancer trends are rare, but they do occur and can be triggered by a single event, such as a public figure's diagnosis with cancer.
"We've seen this before when Betty Ford got breast cancer and when Ronald Reagan was diagnosed with colon cancer," says Dr. Edwards. "These events have an impact on people's behavior, and in this case it would be the WHI results."
By Edward R. Winstead