In the years since America first came to know and love her as Crissy Snow on “Three’s Company,” Suzanne Somers has made millions of dollars selling clothes, exercise equipment, and books, seven of which have become bestsellers. Her latest book, “Ageless,” has drawn the ire of some doctors for touting the benefits of something called “custom-compounded bioidenticals,” an alternative form of hormone replacement therapy (HRT) for menopausal women that requires a doctor’s prescription but is not FDA approved. Dr. JoAnn Manson, author of “Hot Flashes, Hormones, and Your Health” and a leading expert on HRT, was invited on TODAY to discuss menopause treatments. Read an excerpt of her new book:
Chapter 3
The Rise and Fall and (Cautious) Return of Hormone Therapy
Once prescribed primarily to cool the hot flashes and dry the night sweats of menopause — two things that it is very effective at doing — hormone therapy (formerly called hormone replacement therapy, or HRT) has for several decades also been promoted as a strategy to forestall many diseases that we associate with aging and that tend to accelerate after menopause, including heart disease, memory disorders, and osteoporosis. Indeed, more than two of five menopausal women in the United States were taking hormone therapy in 2001.1 In addition, it was becoming increasingly common for doctors to begin prescribing hormone therapy for women in their 60s and 70s and even for women with a diagnosis of angina (chest pain) or previous heart attack. So strong was the belief that estrogen protected the heart and that women of all ages could benefit, many women were being started on hormone therapy 20 to 30 years after the onset of menopause. This widespread use was unwarranted, especially among older women, given the lack of conclusive data on the long-term health consequences of such therapy.
Recent results from the Women’s Health Initiative (WHI), a major national clinical trial of hormone therapy in healthy women who were on average more than a decade past menopause, not only appear to refute the idea that supplemental estrogen keeps the heart healthy in older women but also suggest that, when taken in combination with a progestogen (as it normally is to protect against endometrial cancer), it may actually increase the risk of heart disease. Moreover, the WHI findings indicate that supplemental estrogen offers no protection against chronic disease overall, and that the health risks associated with estrogen-plus-progestin therapy may outweigh the benefits. However, very few women in the WHI were newly menopausal (within five years of menopause), so the study could not conclusively address the benefits and risks of hormone therapy in that group.
When the first of these unexpected results was announced in July 2002, physicians and other healthcare providers across the country were deluged with calls from frantic patients asking, “What went wrong? I thought hormones were supposed to be good for me — not just relieve my symptoms but keep me healthy!” and, even more urgently, “What do I do about hormone therapy now?” And many doctors found themselves unprepared to provide satisfactory answers to these questions.It’s important not to judge these physicians too harshly. Yes, some were unprepared simply because they had not critically examined the earlier scientific data underlying the belief that hormones were “healthy” for women of all ages and risk factor status. But others were unprepared because balancing the benefits and risks of hormone therapy for any particular patient can challenge even the most knowledgeable of healthcare providers, including those with a nuanced understanding of the strengths and weaknesses of the existing research.
In this chapter, I’ll review how medical research works and how medical thinking regarding menopause hormones has evolved over time — or, as billed in the title, the rise and fall and (cautious) return of hormone therapy. The material covered here is a useful prelude to the detailed look at what we know today about the health benefits and risks of supplemental estrogen that is presented in the next chapter.
What Different Kinds of Studies Tell UsA brief explanation of the types of studies that contribute to the advance of medicalknowledge will help you understand how we reached the crisis of confidence about hormone therapy, after years of widespread belief that it was beneficial for most women.
Laboratory studies, also known as basic or experimental research, investigate the effects of a given intervention on animals, body cells from animals, or body cells from humans. Such studies allow for far greater control than studies of human beings, but results from animals or cells in tightly regulated environments may not always apply directly to actual people living in the real world. Nevertheless, laboratory studies often provide important insights into whether and how a purported medical treatment might work in humans.
Studies of people living in the real world are known as epidemiologic studies. The two main types of epidemiologic studies are observational studies and randomized clinical trials. In an observational study, researchers observe study participants and record their characteristics, behaviors, use of medications, and health outcomes but do not otherwise intervene in the participants’ lives. This type of study uncovers possible relationships between various “exposures” and diseases, but it cannot prove a cause-and-effect link. Two common subtypes of observational studies are casecontrol studies and cohort studies.
Case-control studies gather histories from a group of people who have developed a particular disease (the cases) and a similar group of people who are free of that disease (the controls) and compare the two groups to look for factors that might have contributed to the development of disease. The information is obtained by questioning the subjects or their family members about their history or by reviewing medical, employment, or other archival records. Case-control studies are an efficient way to study connections between exposures and chronic diseases, which usually take many years to develop. However, having a disease may color the participants’ recall of their behaviors and use of medications, which can distort results.
Cohort studies assemble a group of people who typically have some characteristic in common, such as occupation, place of residence, or menopausal status. Unlike a case-control study, none of the participants in a cohort study has the disease of interest at the start of the study. The group is then followed over time via periodic checkups or mailed questionnaires, or by monitoring death certificates, to see who develops the disease. Once enough time has elapsed, researchers can examine the information to test a variety of hypotheses concerning the development of disease. By gathering exposure information from participants before the disease has occurred, cohort studies avoid the problems of faulty recall that can sometimes affect case-control studies.
Both cohort and case-control studies have suggested major health benefits of hormone therapy, including reductions in heart disease and hip fractures, but have also suggested major risks, such as an increased chance of breast cancer and stroke. It’s because of these results that most healthcare providers have for many years cautioned women who were already at higher-than-average risk for breast cancer or stroke not to choose hormone therapy.
However, observational studies are open to the criticism that the apparent benefits seen in these studies may simply reflect the fact that women who choose to use hormone therapy tend to be healthier, have greater access to medical care, and embrace health-promoting habits, such as eating a nutritious diet and exercising regularly, more readily than women who do not choose to use hormones. In short, the argument goes, it’s not that taking estrogen promotes health, it’s that being healthy promotes the taking of estrogen. To address this concern, researchers typically use statistical techniques to factor out the effect of variables, such as age, diet, physical activity, smoking, and so forth, that may vary between hormone users and nonusers. This strategy is effective, though not perfect, because it’s impossible to anticipate, measure, and account for all the factors that could conceivably distort, or confound, the relationship between an exposure and disease. So there’s always some possibility that the results are at least partially due to external factors (so-called “confounders”) rather than the exposure of interest.
Enter the randomized clinical trial, which is less susceptible to this so-called “healthy-user” bias than observational studies. In a clinical trial, whose purpose is often to test a particular medical treatment, researchers actively intervene in participants’ lives by assigning them to the therapy under investigation or to a control group. The most rigorous type of trial — the “gold standard” by which all other studies are usually judged — is the randomized, placebo-controlled, double-blinded clinical trial. In these carefully controlled studies, half of a group of volunteers is assigned at random — via a figurative flip of the coin — to an active treatment (a drug, for example), and the other half is assigned to something that looks like the drug but is in fact an inactive placebo (often called a dummy or sugar pill). Neither the doctors conducting the trial nor the participants knows who is getting the medication and who is getting the placebo — hence the term “double-blinded.” The double-blinding ensures that the doctors do not treat the participants differently, and that the patients themselves do not behave differently, based on knowledge of their treatment status.
After a predetermined amount of time, the number of people in the treatment group who have developed the outcome (“endpoint”) being studied — for example, death, heart attack, hip fracture, or, as we saw in the last chapter, hot flashes and other symptoms — is compared with the number in the placebo group who have developed the same outcome. If the trial is large enough, the randomization process ensures that the people in the treatment group are virtually identical to those in the placebo group in terms of age, lifestyle, general health, menopausal status, and other possibly important factors. Because the only characteristic that differs between the two groups is the treatment under investigation, it is extremely likely that any difference in health outcomes found in the two groups is attributable to that treatment. This is the main advantage of a randomized clinical trial over an observational study.
Understanding the basic distinction between observational studies and clinical trials will take you a long way in interpreting epidemiologic data. But many other issues must be considered when evaluating scientific research. Indeed, a well designed observational study can be better than a poorly designed clinical trial for arriving at a sound scientific conclusion. Although this list is not exhaustive, the elements of a good study, whether an observational study or clinical trial, include the following factors.
- A representative study population: A study of a particular therapy may be of
limited value if the study participants are not typical of the general public who use
that therapy. As an extreme example, the first clinical trial of supplemental estrogen
to reduce heart disease was actually conducted in men, not women. Because
sex hormones differ dramatically between men and women, it’s unlikely that the
results of such a study in one gender will apply to the other. (The trial showed
clearly that estrogen didn’t help men, but the story is more complex in women.) - Use of actual disease outcomes: Because it often takes many years for chronic
diseases to develop, many studies look at intermediate changes, such as changes in
blood levels of cholesterol, changes in bone density, or proliferation of the uterine
lining, as proxies for the actual disease of interest, such as a heart attack, osteoporotic
bone fractures, or endometrial cancer. However, these changes do not
inevitably reflect or lead to changes in the actual disease outcome. Research looking
at concrete disease outcomes such as heart attacks is more convincing and
reliable. - Large number of participants (size matters!): The larger the study, the smaller
the possibility that its findings — whether positive, negative, or neutral — might have
resulted by chance alone. The play of chance is such an important issue that scientists
routinely employ sophisticated numerical methods to assess whether a study
finding is “statistically significant” — that is, unlikely to have occurred simply by
chance. - Consistency of the evidence: Consistency is actually not a characteristic of one
particular study, but rather of the whole body of research on any given topic. Perhaps
the most convincing evidence that an effect may be real is consistent results
from a number of researchers at different times, using different study designs, and
involving different groups of people. An example would be the connection between
cigarette smoking and lung cancer that’s been seen in so many different studies.
But keep in mind that scientific research progresses in fits and starts, and there is
rarely a straight-line path to the right answer or a complete absence of conflicting
findings. Indeed, the progress of medical knowledge has been compared to the
stock market — sometimes it goes up, sometimes down, but in the long run, the
trend is (hopefully) in the right direction.
In the case of hormone therapy, a roller-coaster analogy, at first glance, might seem more appropriate because the ride ends up right back where it started. In a narrow sense, as we will see, this is true: the original reason for taking menopausal hormones — for relief of menopausal symptoms — is, for most women, still the only compelling reason for taking them today. In a larger sense, however, the rollercoaster analogy falls short, because researchers have made great strides, especially in the last decade, in understanding the impact of hormone therapy on many aspects of women’s health beyond the relief of symptoms. Incidentally, keeping these research principles in mind will help you be an educated consumer of the health news reported by the media, which tends to magnify the importance and implications of new studies, especially those that contradict common conceptions, while sometimes glossing over their limitations or failing to provide a context for the research.
Excerpted from “Hot Flashes, Hormones, and Your Health” by Dr. JoAnn Manson. Copyright 2007 by Dr. JoAnn Manson. Permission granted by