Hormone Therapy for Women: What Are the Known Risks? |Hormone Therapy| Sara Gottfried, M.D.

Discussion about hormone therapy for perimenopause and menopause is complex. Questions about safety, timing, and duration typically require a nuanced conversation about a person’s history, diet, lifestyle, risk factors, genetics, blood tests, and cognitive function. This article illustrates why the discussions are complicated and how I approach them. I practice Precision Medicine, which means that every patient receives a highly customized protocol based on their particular goals, risk factors, inputs, and outputs. 

By way of background, I recorded a video (below) explaining the role of estrogen in the body and its many protective functions. Then I tackle the very confusing data from the large observational study, the Nurses’ Health Study as well as from the two largest randomized trials on hormone replacement therapy, HERS trial and the Women’s Health Initiative. I point out the flaws in the design of these trials and why we can’t generalize the data to all women or women younger than the average age in the HRT trials. Then I continue to discuss my very specific criteria for prescribing bioidentical hormones and what types I currently prescribe. 

You can watch the full video here.

What Does Estrogen Do for Women before Perimenopause and Menopause?

When I say “estrogen,” I am primarily referring to estradiol, the most abundant form of estrogen throughout a woman’s reproductive years. 

What does estradiol do? 

When I say “estrogen,” I am primarily referring to estradiol, the most abundant form of estrogen throughout a woman’s reproductive years. 

What does estradiol do?

  • It is cardioprotective by increasing nitric oxide bioavailability so it helps with the dilation of blood vessels.1
  • It improves lipids (your cholesterol).2
  • It is anti-inflammatory, anti-platelet, and antioxidant in function.
  • It regulates microRNA. MicroRNAs play an important role in gene expression. I predict we will be hearing more about this over the next 10 years.4
  • It is anti-apoptotic.Apoptosis is programmed cell death. Estrogen helps to regulate and modulate cell death so it occurs when it should and not prematurely—and not out of control as occurs with cancer.
  • It prevents arterial stiffening, which is one of those factors we see in women after they go through menopause.6
  • It is anti-atherogenic.7

Hormone Therapy: A Historical Perspective

Now we have outlined the roles of estrogen in the body, we can move on to the discussion about hormone therapy. 

It is part of our shameful past that the medical establishment pushed hormone therapy on women for perimenopausal symptoms before there were randomized trials to show that it was safe. Instead we had observational data and preclinical data (animal studies), and they led us astray.Those data showed that hormone therapy in women was cardioprotective. 

Think back to the list of estrogen benefits above. One of the decisions behind giving hormone therapy to menopausal women was to replace the decline in estrogen levels in an attempt to continue the cardioprotective benefits. Almost 30 years ago, the findings of the first large-scale observational study, the Nurses’ Health Study, appeared to validate this approach because it showed that women who took hormone therapy had about half the risk of cardiovascular disease (CVD) as those who did not.

However, it turns out the women who took hormone replacement therapy (HRT) were different than the women who didn’t take HRT. They tended to be healthier, saw their doctors more, they were wealthier, got more screening tests, essentially, two different populations that made the findings less relevant.

The HERS Study

The first randomized trial that I want to discuss is the Heart and Estrogen/progestin Replacement Study (HERS) study.  Published in 1998, HERS was a randomized trial to reduce the risk of recurrent coronary heart disease (CHD) events in postmenopausal women with heart disease.

It should be noted here that at the time the #1 prescription in the U.S. was Premarin, conjugated equine estrogen, which is derived from pregnant mare urine. This was administered together with a form of synthetic progestin called medroxyprogesterone acetate. You have to give progestin or natural progesterone orally or vaginally to counter the effects of estrogen at the level of the endometrium—in other words, to reduce the risk of endometrial precancer and cancer. Medroxyprogesterone acetate is not bioidentical. The combination of conjugated equine estrogen and medroxyprogesterone were very commonly prescribed. 

HERS looked at a group of women who had active cardiovascular disease (CVD) with an average age of 67. When given conjugated equine estrogen and medroxyprogesterone acetate, what happened? They showed an increased risk of CVD. 

Why is this not surprising? Firstly, there are increased cardiac events in general in women who are so many years away from menopause. The average age of menopause is 51. The average age in the HERS trials was 67. Additionally, the women had established cardiovascular disease. In the first year of treatment with the devils, as I call them, namely conjugated equine estrogen and the medroxyprogesterone acetate, cardiac events increased. What else did the treatment increase? The rate of myocardial infarction (heart attack), the rate of angina (chest pain caused when your heart doesn’t get enough oxygen), and it also increased blood clots, also known as venous thromboembolism

While the results were not very surprising, it was the first randomized trial that showed that in women who already have heart disease, it’s dangerous and provocative to give them estrogen together with progestin.

The Women’s Health Initiative Hormone Trial

Many women are familiar with the Women’s Health Initiative (WHI) trial because it was stopped three years early based on the findings that hormone replacement therapy increased the risk of heart disease, stroke, and blood clots. The WHI trial was one of the largest randomized trials that took a group of women with an mean age of 63 (overall aged 50 to 79). They were supposedly free of cardiovascular disease but I will dispute this below. 

There were two arms to the WHI trial. One arm was a group of women who had a hysterectomy. They gave them conjugated equine estrogen only. The second arm was made up of women with a uterus, who were given conjugated equine estrogen together with medroxyprogesterone acetate. There was also a placebo group.

Similar to the HERS trial, they found the women had increased cardiovascular events. Now we have two randomized trials showing the same result: taking conjugated equine estrogen together with synthetic medroxyprogesterone acetate increased the risk of heart disease.

Design Flaws of the HERS and WHI Trials

When the initial results came out from the WHI in 2002, they did not stratify by age and did not talk about how many years these women were from menopause. The women in the HERS trial and in the WHI are very different from the women I treat with hormone therapy.

In both trials, they were trying to get women who were not symptomatic of menopausal symptoms. The women in the trials weren’t having hot flashes or difficulties sleeping, or decreased libido. Why? Because they knew if you then randomized women to receive estrogen versus placebo, that they were going to know which treatment arm they were in by the symptoms they were suffering with. So they tried to make sure they were collecting patients who were further out from menopause and weren’t having much in the way of symptoms. Overall, fewer than 10% of women in the WHI were symptomatic. 

That’s the first problem I have with these trials: why would you treat women who were asymptomatic with hormone therapy? If I have a patient who is not symptomatic of menopause, I am not going to give her hormone replacement therapy.

What else do we know about the trials? We know that 70% of the women who were in the WHI were overweight or obese with a BMI of 50 or higher. That alone confers some increased risk in terms of insulin signaling and metabolic health. 50% were current or past cigarette smokers so they already have some endothelial damage just by being past or current smokers. More than 35% had high blood pressure that needed treatment.

If we look at how we treat hypertension now compared to 2000 when the WIH was running, we have a much more intensive approach, where we start to treat women and men for hypertension at a lower systolic and diastolic blood pressure rate. When the WHI was performed, many of these women already had vascular dysfunction as evidenced not only by the smoking status or being overweight or obese but by their blood pressure.

Only 10% of the women in the WHI trial were aged 50 to 54. I would say the age where I focus on hormone therapy tends to be women who are 45 to 55. So very few of the women I am treating are represented in the WHI in which 70% were between the ages of 60 to 79. In this latter age range, this is where you see atherosclerotic plaques and vascular dysfunction, so it makes sense that there were increased cardiovascular events in the women in the trial.

Is HRT Risky for Women?

What do these two large randomized trials tell us then about the dangers of hormone replacement therapy for women? We can conclude that it is probably not safe to give women HRT who are more than 10 to 20 years from menopause. However, we cannot take the data and generalize it to younger women or to all women. We can only generalize it to the older women that were studied in these two trials. 

When I prescribe hormone therapy, the age where I focus tends to be symptomatic women who are 45 to 55, or within 5 to 10 years of their final menstrual period (menopause). However, before we proceed with treatment, we will have a conversation about the limited and flawed data we have from the WHI and other smaller trials. 

Another important issue is that we can’t generalize about the hormone therapy used in the trials. The HRT used was conjugated equine therapy and medroxyprogesterone acetate. We can’t take the data showing the increased risk of cardiovascular disease, stroke, and blood clots and  apply that to all forms of menopausal hormone therapy. 

There are many other smaller randomized trials. One that I often cite is the ELITE trial, which supports the use of bioidentical hormone therapy within 5 to 10 years of menopause.8 In this study, N=643 healthy postmenopausal women were stratified by time since menopause (Early <6 years, Late >=10 years) to 1 mg oral E2 plus 45mg vaginal progesterone QDx10 days per month. After a median of 5 years, the women with “Early” postmenopause treated with oral estradiol and vaginal progesterone had the lowest progression of subclinical atherosclerosis (as measured by carotid intimal-medial thickness, or CIMT, a standard measure of a blood vessel in the neck.) They also measured coronary calcium score as a secondary outcome, and there was no difference between either postmenopause group and placebo. The conclusion was that this study confirmed the timing hypothesis comparable to the WHI reanalysis mentioned previously.

I prescribe bioidentical hormone therapy. The way I like to do it is to use transdermal estradiol. I prefer to use FDA approved forms of estradiol so that means a patch or a gel, occasionally a cream. I also like to give oral progesterone because that is what’s been studied the most and what is most safe in terms of CVD. Vaginal progesterone in smaller studies may also protect the endometrium. I do not prescribe transdermal progesterone in women with a uterus. I only prescribe oral or vaginal progesterone in women with a uterus. They are the only ways of giving progesterone that are evidence based when considering the safety of the uterus—that is, prevention of build up of the uterine lining.

My Approach to Hormone Therapy: The Critical Window Hypothesis

We need an historical perspective when it comes to hormone therapy. The HERS trial and WHI provided important information but not data that we can generalize to younger women (in the range of 45-55). The way I approach HRT in women is that you have this window of time called the critical window for both CVD and also for brain health. 

The critical window is probably within five years of menopause, possibly 10 but we still do not have definitive evidence. If I start patients on hormone therapy, it’s a decision based on complex data analysis that they are good candidates for hormone therapy. We look at the all of the factors I menioned, from lifestyle to genomics. We look at risk of cardiovascular disease, breast cancer, and Alzheimer’s disease, including dementia. We look at prior surgery, such as removal of one or both ovaries, or a partial or full hysterectomy. If we decide to start hormone therapy, it’s going to be women who are the best candidates, usually 45 to 55 or 60. Once a woman ages beyond 55, we have to question in greater detail whether the benefits outweight the risks, and what alteratives exist. 

Part of this complex data analysis is examining a woman’s history for cardiovascular disease. I don’t mean the classic five factors that include high cholesterol, diabetes, high blood pressure, obesity and smoking. I am referring to the 400+ emerging risk factors including what happened in pregnancy as this can unmask endothelial dysfunction as seen in preeclampsia or small for gestational age baby or or gestational diabetes. 

How do we integrate all of these concepts and help a woman who is in perimenopause and symptomatic or freshly menopausal and symptomatic? The protocols in my book The Hormone Cure reflect what I offer in my practice. I work with women to find the best possible solutions to balance their hormones—including an upgrade to the way they eat, move, think, sleep, and supplement. If I prescribe bioidentical hormones, it’s a decision made together with my patient based on their symptoms and only after we have carefully reviewed all of the risk factors. That is shared decision making at its best.


I am opening up my practice to new patients towards the end of this year. To get on the waitlist, you can sign up here 

  1. Nevzati E, et al. Acta Neurochir Suppl 2015;120:141-5; Somani YB, et al. Am J Physiol Heart Circ Physiol. 2019 Aug 1;317(2):H395-H404.
  2. Guetta V, et al. Circulation. 1996;93(10):1928–37; Fåhraeus L: The effects of estradiol on blood lipids and lipoproteins in postmenopausal women.Obstet Gynecol. 1988;72(5 Suppl):18S–22S.
  3. Kovats S. Cellular immunology 294, no. 2 (2015): 63-9; Miller VM, et al. Platelets 27, no. 1 (2016):32–42; Song CH, et al. Gut Liver. 2018;12, no. 6 (2018):682693; Ramesh SS, et al. Biol Rev Camb Philos Soc. 2019 Dec;94(6):1897-1917.
  4. Ramesh SS, et al. Biol Rev Camb Philos Soc. 94, no. 6 (2019):1897-1917.
  5. Ramesh SS, et al. Biol Rev Camb Philos Soc. 94, no. 6 (2019):1897-1917.
  6. Westendorp IC, et al. Menopausal status and distensibility of the common carotid artery. Arterioscler Thromb Vasc Biol. 19, no. 3 (1999):713–7. 
  7. Naftolin F et al. F1000 Research vol. 8 F1000 Faculty Rev-1576. 3 Sep. 2019
  8. Hodis, HN, et al. N Engl J Med 2016; 374:1221-1231