Millions of women in the United States have taken hormone replacement therapy, a treatment once heralded as a likely remedy for the health issues women face at menopause. A common form of the therapy — a cocktail of horse estrogens and synthetic progesterone — was approved by the Food and Drug Administration (FDA) to treat osteoporosis and the symptoms of menopause. The cocktail was also thought to keep heart disease at bay.

The evidence in favor of hormone therapy came from a series of observational studies. Randomized, controlled clinical trials of the drugs had not been conducted, but many doctors were convinced that the cocktail worked.

All that changed in 2002 when scientists conducted a randomized, double-blind, controlled trial as part of a large epidemiological study called the Women’s Health Initiative (WHI). The trial showed that combination hormone therapy using conjugated horse estrogen and synthetic progestin protected women from hip fractures but raised the risk of invasive breast cancer, heart attack, stroke, and blood clots.

In the trial, more than sixteen thousand postmenopausal women were randomly assigned to receive either the hormone cocktail or a placebo. The women were in the trial for an average of five and a half years before the health risks of the hormone combination became apparent and the trial was stopped. The FDA asked manufacturers to place new warning labels on the drugs and advised consumers to use the lowest effective dose or talk to their doctors about alternative therapies.

But Gerardo Heiss and other WHI researchers wanted to know more. “If a woman chooses to take the preparation,” Heiss asks, “what happens when she stops?”

From July 2002 to March 2005, Heiss and his colleagues followed up 15,730 women who had participated in the WHI trial. The women received annual breast exams and mammograms, and biopsies when needed. Heiss and his colleagues showed that some effects of combination hormone therapy were still present after two and a half years.

Women who had been taking the cocktail were 27 percent more likely than those who had been taking the placebo to develop invasive breast cancer during the follow-up study. But the absolute risk of breast cancer in the hormone replacement therapy (HRT) group was lower in 2005 than it had been when the women stopped taking the drugs in 2002.

Heiss says the randomized assignment of participants to the two groups in the original trial canceled out the effects of differences in age, ethnicity, hypertension, smoking status, use of cholesterol-lowering drugs, and previous history of breast cancer and heart disease.

The researchers also examined the number of heart attacks, strokes, and blood clots during the follow-up study and found no significant difference between the HRT and placebo groups, indicating that the initial surge in the risk of cardiovascular events had ebbed after the HRT trial was stopped. The number of deaths due to all causes was not significantly different between the two groups. Heiss and his coworkers also found that the beneficial effects of the cocktail — decreased risk of hip fractures and colorectal cancer — disappeared once the therapy was stopped.

Heiss says that women who appeared to benefit from the combination hormone therapy in studies prior to the WHI trial were better educated, had better access to health care, saw their physicians more often, and underwent cancer screening more regularly than the women in the WHI trial. “As a group, they showed a considerably more favorable health profile and survival than women who were not taking the preparation,” he says. Also, the observational nature of those early studies made it difficult to account for women who had stopped HRT due to side effects.

The lingering risk of breast cancer points to the need for continued screening, including breast exams and mammograms, for postmenopausal women who have stopped taking combination hormone therapy, Heiss says.

One limitation of the follow-up trial is that it took place over three years; the trial did not provide data about diseases that might develop over a longer period. Also, the women participating in the follow-up had been told at the end of the WHI trial whether they had been taking HRT or a placebo. “Women possibly could have changed their behavior, once they realized the risks associated with taking the estrogen-plus-progestin combination. But we checked for their behavior, adjusted for it, and found it to be completely unrelated to the findings,” Heiss says.

The findings apply to combination hormone therapy that uses a specific mix of natural and synthetic hormones. Some doctors prescribe bioidentical hormones, which have the same chemical structures as naturally occurring human hormones, but the drugs’ long-term health effects have not been studied in large clinical trials or observational studies. Pharmaceutical companies are loath to conduct expensive, large-scale studies of bioidentical hormones because these compounds, having been invented by nature, cannot be patented.