Hormones and Mood
Estrogen Replacement Therapy: Risk Discussion
UPDATE 7/2005: There is a presentation of this issue by our national Centers for Disease Control, which has been reviewed by a hospital-based team to make sure it is practical information as well as accurate. That basically replaces this entire page. I've left what I wrote earlier in case someone bookmarked this page for one of the references, but instead of reading this page, you should read the CDC's page (it loads slowly: nice colors and formatting...)
My earlier page on this topic, now out of date:
As you probably know, there was a recent study called the Women's Health Initiative (WHI) that showed more cardiovascular risk than expected.WHI Especially after this study, there is an intense debate going on about estrogen and progesterone treatment risks. I am not an expert in this area. However, some of my patients are already on estrogen, or want to consider it because they can see how directly their mood or anxiety symptoms relate to hormone.
So, I'm going to attempt to summarize an extremely complicated area of research. There are many differing opinions among experts on this topic. My patients understand they are more than welcome to do their own research on this. This page will be revised as I learn more, and you are welcome to send me information you think would help with that.
First, you should understand research so far has not focused on mental health benefits of hormone therapy, such as treating to benefit mood, anxiety, or cognition (clarity of thought, memory). Instead, the focus has been on heart/stroke protection, which was thought to be one of the benefits; and on breast cancer risk, which has been recognized as a probable risk of treatment for years. None of the major studies has specifically focused on helping women decide if estrogen treatment (with or without progesterone) has enough mental health benefit to merit taking some risk otherwise.
The biggest studies are the WHI, the HERS project HERS and a major "meta-analysis" including data from both these studies and several previous high-quality studies.Nelson et al So far, a small increased risk of breast cancer appears to be fairly well confirmed. From what I've seen, this is not usually debated (compared to the cardiovascular risk issue, which is intensely debated).
How big a breast cancer risk increase? Let's look at the WHI results overall, then we'll focus on the breast risk. For the 16,000 women the WHI studied, the authors give their results "per 10,000 patient-years" (the increase per year if 10,000 women received estrogen and progesterone; versus a group of women not taking those hormones). They found, per 10,000 patient-years of exposure to Premarin/Provera (horse estrogen/synthetic progesterone):
Their conclusion: this regimen should not be initiated or continued for prevention of coronary heart disease (see how narrow their focus).
Now, back to breast cancer. Hang on, this one paragraph is going to get pretty thick. It will ease up after that... In the HERS study, there was no increased risk. In the big review by Nelson and colleages there was no increase in breast cancer seen in fourteen of 18 observational studies and 7 of 8 meta-analyses.Nelson et al However, when they pooled all the data from all those sources, they did find that using estrogen increased risk, and the risk tended to increase with time. They report this as "relative risk" -- the likelihood of getting breast cancer relative to someone not taking hormones. The relative risk they observed ranged from 1.23-1.35, meaning that there is a 23-35% risk increase (an RR of 1 means no difference in risk observed). Similarly, the WHI reported an RR of 1.26.
So, it looks like the breast cancer risk is for real. However, a mood/menopause expert, Dr. Marjorie Shuer, points out that breast cancer is a very small risk in the United States compared to cardiovascular disease and stroke. From the data she presented at a recent conference,Shuer the likelihood of an 80 year old dying from cardiovascular disease is roughly 10 times greater than breast cancer, and stroke is about 3 times greater a risk. She then adds data on osteoporosis:
You can see she's making a case for use of estrogen. She then points out how dramatically alcohol increases estrogen levels (at least 10 times higher in one study) and wonders how this factor alone might have affected breast cancer risk in the major research studies, by artificially increasing estrogen levels. Similarly, she notes that getting blood clots with estrogen is related to dose -- and that with Premarin, you can't measure the level (for more on that particular point, here are details). Her point: we still really don't know how to interpret the WHI, in the big picture of things. For example, she is adamant that Premarin (used in the WHI) is not a good representative of the new estrogen formulas available, which can now deliver estradiol, the main estrogen product of human ovaries, by pill or patch.
Here's another way of looking at these "relative risk" figures. If you're thinking about estrogen but hesitating because of breast cancer risk, are you looking at your alcohol consumption as well? Alcohol is an established risk factor for breast cancerSellers . Similarly, there are excellent data showing that exercise, even just plain old walking, lowers cardiovascular risk. If you hesitate to consider estrogen replacement treatment because of cardiovascular risk data, are you walking every day too?
May I interject some exercise data -- from the WHI no less -- to give you some comparisons in terms of risk evaluation? Women who regularly exercised, including "just" walking, at the highest exertion levels in the WHI, had a relative risk (RR) for cardiovascular disease of 0.47. In other words, they were only 47% as likely to have a heart attack or similar heart problem as sedentary women (if there were no benefit of exercise, the relative risk would have been 1.0, meaning the exercisers were 100% as likely to have a heart attack as the sedentary folks -- see how that works?). Now you have to admit, if you were really making your health decisions on the basis of "relative risk" of cardiovascular disease, you'd be walking 5 times a week.
So, when you try to boil all these studies down and make a decision about risk and benefit of estrogen, are you really thinking it through or are you reacting to numbers? When I try to think it through, I find that there is no clear answer. For women who are having profound mental health symptoms that are so severe they must contemplate a treatment of some kind, from what I've learned here I believe the "relative risks" of estrogen treatment in the perimenopause are close to but slightly higher than those of antidepressants, for example. That overall risk picture might shift if a woman had a high osteoporosis risk, or a strong family history of breast cancer.
Remember, we don't have direct data on benefit from hormones to compare to these known risks. There is a growing research on Alzheimer's Disease, for example, which seems to suggest that estrogen replacement may have strong protective value against AD.Zec When we have more experience treating women for mental health symptoms with estrogen, we'll be able to characterize the benefit more clearly. For now, I was very impressed by the enthusiasm Dr. Shuer expressed based on her experienceShuer and will continue to research this area. And for now, when possible I am working with a patient's OB-GYN to select doses and monitoring plans.
A note from a paper by Dr. Shuer, parked here until I have a better place to put it (bottom line: don't measure estradiol when you're taking Premarin or birth control pills; it ain't the same):
Comparing endogenous estrogen with conjugated equine estrogen is not scientifically valid, since the human radioimmunoassay for 17[beta]-estradiol does not measure levels of estrone sulfate, equilin sulfate, 17[alpha]-dihydroequilin sulfate, equilin, equilenin, 17[beta]-dihydroequilin, or 17[beta]-dihydroequilenin, which are all metabolic products of conjugated equine estrogen.
The same holds true for ethinyl estradiol, which is the synthetic estrogen found in most oral contraceptives. Taking oral contraceptives usually results in ovarian suppression; thus, there is no production of 17[beta]-estradiol, the most prevalent endogenously produced estrogen. Thus, one has to question the measurement of 17[beta]-estradiol in the women who were taking oral contraceptives. None of your readers would measure an imipramine level in a patient taking fluoxetine; however, measuring 17[beta]-estradiol in patients taking oral contraceptives is its psychiatric equivalent and is therefore inaccurate. Shuer (b)