DEADLY DOGMA

WHAT WE THINK WE KNOW ABOUT WOMEN’S HORMONES AND HEART DISEASE CAN HURT US ALL

By Jerome L. Sullivan, MD, PhD

"It’s not what we don’t know about nutrition that hurts us, it’s what we know for sure that turns out to be dead wrong." – Victor Herbert

Known-for-sure-but-dead-wrong things hurt us by shutting down inquiry that otherwise might protect from us from harm. Dr Herbert’s insight about nutrition applies more broadly to heart disease. What was known for sure about women and heart disease has recently been shown to be dead wrong. Now we know that hormone replacement therapy (HRT) does not protect against heart disease. The harm from decades of bad advice to women may turn out to be trifling in comparison with the injury caused by decades of delay in understanding basic facts about women and heart disease.

The massive Women’s Health Initiative [1] trial has found a slight increase in cardiovascular disease in women who take HRT, not the dramatic decrease expected. The results have been described as "surprising" and "stunning," but there has been virtually no discussion of the most important meaning of the trial. No one has asked why the findings are surprising and stunning. An entirely unexpected result should be an occasion for going back to fundamental questions and rethinking the problem. The most stunning thing about the Women’s Health Initiative data has been absence of curiosity about fundamental issues.

The trial results surprise because of their perceived conflict with a point of view that has been beyond question: the notion that premenopausal women are directly protected from heart disease by their hormones. Young women do have an extraordinarily low heart attack rate, but the reasons for their protection are not known. There is a deeply entrenched belief shared by scientists, physicians and the public that they are protected by the direct actions of female hormones.

Any doubts about the power of women’s hormones were dispelled by many observational studies of postmenopausal HRT and by data that made protection plausible, e.g. that HRT reliably improves lipid profiles. But the observational work was not randomized and could not tell us whether HRT makes women healthy or healthy women preferentially take HRT. And, plausibility does not prove cause. Despite the beneficial impact of HRT on lipids in the randomized trials, no benefit of treatment was seen.

Data that superficially appeared to support protective effects of HRT squelched any questioning of the core belief in the power of women’s hormones. They also helped everyone forget earlier findings on heart disease in menstruating women that were inconsistent with the prevailing belief system. The most important, but by no means the only, was the Framingham Study. In Framingham, it was discovered that simple removal of the uterus with preservation of functioning ovaries, termed a "simple hysterectomy," in a young woman was enough to cancel out all the heart benefit of premenopausal status. In a young woman, removal of just the uterus leaves her intact hormone-producing ovaries in place. If the hormonal explanation of her heart protection was right, she should have continued to be fully protected from heart attacks. But, the Framingham scientists were surprised that protection was lost, to the same degree as if the young woman had experienced natural menopause. At the time, in 1978, they found this remarkable:

The Framingham scientists later found that postmenopausal estrogen use increased heart disease risk. But as the observational studies began to be interpreted as confirming the "obviously true" conclusion that hormones are protective, the Framingham findings were reanalyzed with the opposite conclusion drawn, i.e. that estrogens did not endanger the heart. What was the motive for this published reanalysis? Could it have been a refusal to accept the data when the data contradicted what everyone knew to be true?

A consequence of the pervasive and stubbornly held notion of protective hormones was that any alternative explanation of how young women are protected became superfluous. In the 1970s, I began to reflect on the puzzle identified so clearly by the Framingham scientists. If functioning ovaries are not enough to maintain the low heart risk of young women, then it seemed there must be some other mechanism, a mechanism of great potential importance considering the virtual absence of coronary disease in menstruating women. The novel alternative explanation that I eventually proposed was published in 1981 [3-5] and has come to be called "the iron hypothesis." It was widely ignored for a decade. After some supportive findings were published in 1992, the hypothesis came under energetic, if scientifically flawed, attack.

The idea is that young women are protected from heart disease so long as their monthly menstrual bleeding keeps the level of iron near or even a bit below the minimum amount needed to prevent anemia. Blood is rich in iron and the low iron status of menstruating women is a well known feature of monthly bleeding. Their formidable protection against heart attack is lost when they stop regular iron loss from menstrual bleeding either by natural menopause or by the simple hysterectomy that so interested the Framingham scientists. If the uterus is missing, there is no way to lose menstrual blood even though the ovaries continue to have hormonal cycles. Young men start to build up an excess of iron, termed "storage iron," in their late teens. Male iron stores continue to rise with age until late middle age, in close parallel with their rising heart attack rate. In women, iron storage and the onset of heart disease are both delayed until after menstrual bleeding stops. The iron hypothesis suggests that the condition of having no storage iron in fact gives young women strong protection against heart attacks. It also implies that prevention of stored iron accumulation after menopause or in men would protect against heart attack. With appropriate monitoring of iron and hemoglobin levels, this can be safely accomplished by regular, frequent blood donation or other methods.

Unfortunately, to date, there has been no properly designed trial to test the iron hypothesis. It has been found that iron is an essential cofactor in the dangerous free radical reactions and inflammation implicated in coronary plaque formation and in the destruction of heart muscle that happens after a coronary blockage. Many studies since 1981 have shown that use of an anti-iron drug, deferoxamine, decreases heart muscle destruction and helps to prevent heart rhythm problems caused by vessel blockage. Also, there has been some observational research with mixed results. None of this work comes close to being a definitive test of the iron idea. A major problem with research to date has been the use of study designs similar to those that falsely proved protection by female hormones. Getting the proper study started has not been helped by the climate of opposition to the iron hypothesis. A properly designed, randomized, prospective trial to test the hypothesis should have been done two decades ago.

A look at iron was apparently not part of the Women’s Health Initiative. This is also unfortunate because HRT can be associated with bleeding from the uterus. The bleeding is not usually the same amount as normal menstrual bleeding, so it would be expected that the average woman on HRT would still quickly lose her iron depleted state compared to younger menstruating women. A few of the Women’s Health Initiative women may have bled enough to remain iron depleted for as long as they continued the therapy. What was the heart attack rate among those few women who stayed iron depleted? Were they relatively protected from heart attack compared with the women in the study whose iron levels soared? This study may be possible after the fact if the Women’s Health Initiative scientists kept frozen serum samples from all their subjects. The measurement of the iron storage protein, ferritin, in serum can give an indication of iron status. The Women’s Health Initiative should look back at the impact of HRT on iron stores and heart attack rates in their subjects. This would not be a gold standard randomized study but might yield useful information for future work.

Not all dogmas are deadly. A hundred years ago the theory of continental drift was believed to be not only wrong but ridiculous. Now it is an accepted idea, however no one died because of its initial enthusiastic rejection. In medical science, fixed beliefs are dangerous to human life and must be much more aggressively questioned. Potentially murderous dogmas today include belief that women’s hormones protect against heart attack and the ingrained notion that it is normal and beneficial to have iron in storage in the body. The new findings are not just an end to the use of HRT. They show us the next step and cry out for new studies and new thinking. If young women are not directly protected by hormones, then what is it that protects them so completely from heart attack at an age when their husbands begin dropping dead on golf courses? Iron depletion may be the main protective factor. If so, we will have a highly feasible preventive treatment for heart disease, i.e. iron reduction therapy. But, whatever the protective factor turns out to be, medical scientists have a duty to forcefully seek the reasons in a spirit of open-minded inquiry. As the Framingham scientists thought so long ago, heart protection in young women may be a pivotal clue to understanding and controlling our leading cause of death.

Additional Reading

[1] Writing Group for the Women’s Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women, Principal results from the Women’s Health Initiative randomized controlled trial. JAMA 2002;288:321-333.

[2] Gordon T, Kannel WB, Hjortland MC, McNamara PM. Menopause and coronary disease: The Framingham Study. Ann Intern Med 1978;89:157-161.

[3] Sullivan JL. Iron and the sex difference in heart disease risk. Lancet 1981;1:1293-1294.

[4] Sullivan JL. Stored iron and myocardial perfusion deficits. Am Heart J 2002;143:193-195.

[5] Sullivan JL. Are menstruating women protected from heart disease because of, or in spite of, estrogen? Relevance to the iron hypothesis. Am Heart J (Editorial in press).

Jerome Sullivan is in the Department of Pathology at the University of Florida.

Copyright 2002 by Jerome L. Sullivan, MD, PhD