Even women who took estrogen and progestin pills for as little as a couple of years had a greater chance of getting cancer. And when they stopped taking them, their odds quickly improved, returning to a normal risk level roughly two years after quitting.
Collectively, these new findings are likely to end any doubt that the risks outweigh the benefits for most women.
It is clear that breast cancer rates plunged in recent years mainly because millions of women quit hormone therapy and fewer newly menopausal women started on it, said the study's leader, Dr. Rowan Chlebowski of Harbor-UCLA Medical Center in Los Angeles.
"It's an excellent message for women: You can still diminish risk (by quitting), even if you've been on hormones for a long time," said Dr. Claudine Isaacs of Georgetown University's Lombardi Comprehensive Cancer Center. "It's not like smoking where you have to wait 10 or 15 years for the risk to come down."
Study results were given Saturday at the San Antonio Breast Cancer Symposium.
They are from the Women's Health Initiative, which tested estrogen and progestin pills that doctors long believed would prevent heart disease, bone loss and many other problems in women after menopause. The main part of the study was stopped in 2002 when researchers saw surprisingly higher risks of heart problems and breast cancer in hormone users.
Since then, experts have debated whether these risks apply to women who start on hormones when they enter menopause, usually in their 50s, and take them for shorter periods of time. Most of the women in the federal study were in their 60s and well past menopause.
So the advice has been to use hormones only if symptoms like hot flashes are severe, and at the lowest dose and shortest time possible. The new study sharpens that message, Chlebowski said.
"It does change the balance" on whether to start on treatment at all, he said.
Even so, most women will not get breast cancer by taking the pills short-term. The increased cancer risk from a couple of years of hormone use translates to a few extra cases of breast cancer a year for every 1,000 women on hormones. This risk accumulates with each year of use, though.
This part was halted when researchers saw a 26 percent higher risk of breast cancer in those on Prempro.
But that was an average over the 5½ years women were on the pills. For the new study, researchers tracked 15,387 of these women through July 2005, and plotted breast cancer cases as they occurred over time.
They saw a clear trend: Risk rose with the start of use, peaked when the study ended and fell as nearly all hormone users stopped taking their pills. At the peak, the breast cancer risk for pill takers was twice that of the others.
Think of it as President George W. Bush's public approval rating, said another study leader, Dr. Peter Ravdin of the University of Texas M.D. Anderson Cancer Center in Houston.
"Bush's popularity may be 50 percent on average, but it might have been descending the whole time he was president," Ravdin said.
In the second part of the federal study, researchers observed just 16,121 women who had already been on hormones for an average of seven years and another group of 25,328 women who had never used them. No results on breast cancer risk in these women have been given until now.
Plotting cases over time, researchers saw in retrospect that hormone users had started out with twice the risk of breast cancer as the others, and it fell as use declined. Among those taking hormones at the start of the study, use dropped to 41 percent in 2003, the year after the main results made news.
In the general population, use of hormone products has dropped 70 percent since the study, said another of its leaders, Dr. JoAnn Manson, preventive medicine chief at Harvard's Brigham and Women's Hospital in Boston.
That corresponds with big drops in breast cancer cases, but some scientists have said this could be due to a fall-off in mammograms, which would mean fewer cancers were being detected, not necessarily that fewer were occurring.
The new study puts that theory to rest. Mammography rates were virtually the same among those taking hormones and those not.
"It is clear that changing mammography patterns cannot explain the dramatic reductions in breast cancer risk," Manson said.
"The data are getting stronger," said Dr. C. Kent Osborne, a breast cancer specialist at Baylor College of Medicine in Houston.
Women who do need the pills should not panic, though the doubling of risk - a 200 percent increase - for long-term users is quite worrisome, cancer specialists say. Although the new study does not calculate risks in terms of actual cases, previous research showed that the average increased risk of 26 percent meant a difference of a few extra cases a year for every 1,000 women on hormone pills, compared with nonusers.
"Hormone therapy remains a good health care choice to relieve moderate to severe menopausal symptoms," says a statement from Wyeth, which made the pills used in the study.
"Most women should be able to discontinue hormones in three to four years," or at least reduce their dose, Manson said.
A future analysis will look at other women in the study who took only estrogen, generally women who have had hysterectomies.