Background on the controversyWhen I went to medical school and Public Health school in the 1970s, the belief among large numbers of public health professionals was that most post-menopausal women should receive hormone replacement therapy (HRT) for the remainder of their life. There were several reasons for this belief. Not only did HRT relieve troublesome and sometimes life-altering post-menopausal symptoms, but it also substantially reduced the risk of osteoporosis with subsequent hip fracture (often a life altering terminal event in older women) and apparently lowered the risk of coronary artery disease, the number one killer of women (as well as men). On the negative side was a slight and inconsistently demonstrated increased risk of breast cancer and a large increase in endometrial cancer. Since the increased risk of breast cancer was small at most, and not well demonstrated, and since endometrial cancer is rare with or without ERT, the overall benefits were considered by many public health professionals to outweigh the risks for most post-menopausal women, though the issue was controversial. Here is a 1999
review of the subject, which summarizes the epidemiological evidence. And here is
an article that showed an overall 21% reduction in mortality in women who took ERT, compared to women who did not take it – though the difference was not quite statistically significant at the 95% confidence level.
Then, in 2002 findings were released from the Women’s Health Initiative (WHI), a large randomized clinical trial on the use of ERT in post-menopausal women, sponsored by the National Institutes of Health (NIH). The most surprising finding of this study was that not only did it now appear that HRT did not
lower the risk of coronary artery disease, but HRT may actually
increase that risk. Consequently, the study was stopped, and the recommendations for the use of HRT in post-menopausal women were radically altered.
Considerations on the use of epidemiological research vs. controlled clinical trialsOne reason for the vast turnaround in the recommendations for HRT was the fact that the new WHI study was based on a randomized clinical trial, in contrast to the older evidence which was based on epidemiological research. Randomized clinic trials pose a significant advantage over epidemiological research because, by virtue of the fact that they randomize the treatment, it is generally assured that pre-existing risks will be fairly equal in the treatment group as compared with the control group. With pre-existing risks equalized, the scene is set for a more valid assessment of the treatment. Epidemiological studies, on the other hand, are
observational, meaning that treatment is determined by the needs and wishes of the individual subjects, rather than the experimenter. Consequently, there is always the possibility that the two groups may have different pre-existing risks, which may influence their relative outcomes (though efforts are made to statistically control for differences in pre-existing risk).
Though these advantages of randomized clinical trials over epidemiological research are well recognized, there are also advantages that epidemiological studies have over clinical trials, and those advantages are less well recognized. One of the biggest advantages of epidemiological studies is that they mimic real life circumstances, as opposed to the often artificial environment of a randomized clinical trial.
With respect to hormone replacement research, probably the most artificial aspect of the Women’s Health Initiative trial was that there was no distinction made between women who needed the HRT to relieve their symptoms and those who had no symptoms. If a woman was randomly assigned to the HRT arm of the study she was put on a continuing dose of HRT, whereas if she was randomly assigned to the control arm of the study she was provided with a placebo, which meant that it was highly unlikely that she would take HRT even if symptoms required it. Symptomatic need for HRT was not part of the study design.
To give you an oversimplified parallel to this situation, suppose that a randomized clinical trial was used to ascertain whether or not water was good for a person’s health, and the treatment group was told to drink an extra gallon of water every night before they went to bed. Suppose then that the study found no health benefit from the water and concluded that water was not important to one’s health. That example of course is extreme and somewhat ridiculous, but I believe that it makes a point. And that point can be imagined by the realization that if the water treatment was limited to people who were thirsty and dehydrated it would be found that water makes a great deal of difference to a person’s health.
Specific overall study results from the Women’s Health Initiative (WHI)Here is a very good
review article from the Mayo Clinic about the WHI findings, and here is a
scientific presentation on the same study. The treatment group was divided into women who took a combination of estrogen and progestin (Prempro) and women who took estrogen only (Premarin). The results for the group that took combination therapy (estrogen plus progestin) were worse than for the group that took estrogen alone, so I’ll summarize the results for the group that took estrogen alone.
In the estrogen only group, compared with the placebo group, the
downside was a slight increased risk of stroke and a slight increased risk of abnormal mammogram (but
without an increased risk of breast cancer). The
advantages for the estrogen only group (over the placebo group), other than the relief of symptoms, were a decrease in the risk of osteoporosis related hip fractures and a decrease in the incidence of colorectal cancer. There was no difference between the estrogen only group and the placebo group with regard to breast cancer, heart disease, or overall death.
Some important caveats about the overall findings of the WHI studyThe average age of women in the WHI study was 63 years, whereas women typically begin HRT much earlier than that. Furthermore, as I alluded to above, the presence or absence of symptoms had no influence on whether or not women were assigned HRT.
When women in the age group 50-59 were analyzed, those who took estrogen only were found to have
less heart attacks and coronary artery disease related deaths than those who took the placebo instead. This very important advantage of HRT (estrogen only) may be considered as an added advantage to the other benefits noted above, and would seem to shift the overall balance significantly in favor of HRT for that age group.
We don’t know why the 50-59 age group experienced a reduced risk for coronary artery disease and related death when they took estrogen, as compared to the placebo controls. But I think that it would be logical to assume that the reason may be related to the fact that many women in that group had post-menopausal symptoms that were relieved by estrogen – which has always been the primary reason for post-menopausal women take HRT in the first place.
RecommendationsGiven the above facts, it is not surprising that the Mayo Clinic article goes on to recommend HRT (estrogen only) for women who have post-menopausal symptoms that are relieved by HRT and for women who have or are at high risk for osteoporosis. Thus, there are two major advantages to taking HRT: to relieve symptoms which often seriously interfere with one’s quality of life, and to reduce the risk of bone fracture, which is often a terminal event in older women (By causing people to become immobile, hip fractures associated with osteoporosis often lead to a large number of serious health conditions, which often lead to a downward spiral.)
And the article goes on to couch its recommendations in cautious language, involving advice that women talk to their doctor about what to do about this, and listing several other ways to reduce the likelihood of bone disease and heart disease. Also, it notes that women with breast cancer or a history of blood clots should avoid HRT.
Lastly, I have to say that there is one short paragraph in the Mayo Clinic article that I feel is overly cautious, to the extent that it is not even consistent with the rest of the article or with the known facts about HRT, and therefore it is also unnecessarily confusing. That is the part that reads, “One of the previously believed benefits of HRT was that it promoted long term health of post-menopausal women, from reducing the risks of heart disease to making bones stronger. But
since that’s no longer the case, here are some alternatives…” As I hope was made clear by the above discussion, we don’t know that “that is no longer the case”. HRT
does promote stronger bones, according to all studies that have looked at that issue. And it
does apparently reduce the risks of heart disease in relatively younger women, and possibly in all women who take it for symptomatic relief. And finally, there is no question that, by virtue of the symptomatic relief that it provides, it improves quality of life for millions of women.