The Peter Attia Drive - #253 ‒ Hormone replacement therapy and the Women's Health Initiative： re-examining the results, the link to breast cancer, and weighing the risk vs reward of HRT ｜ JoAnn Manson, M.D.

00:00:00.000 Hey everyone, welcome to the drive podcast. I'm your host, Peter Atiyah. This podcast,

00:00:15.500 my website and my weekly newsletter all focus on the goal of translating the science of longevity

00:00:19.840 into something accessible for everyone. Our goal is to provide the best content in health and

00:00:24.780 wellness full stop. And we've assembled a great team of analysts to make this happen.

00:00:28.920 If you enjoy this podcast, we've created a membership program that brings you far more

00:00:33.320 in-depth content. If you want to take your knowledge of the space to the next level at

00:00:37.340 the end of this episode, I'll explain what those benefits are. Or if you want to learn more now,

00:00:41.740 head over to peteratiyahmd.com forward slash subscribe. Now, without further delay,

00:00:47.780 here's today's episode. My guest this week is Dr. Joanne Manson. Joanne is a professor of medicine 0.97

00:00:55.880 and the Michael and Lee Bell professor of women's health at Harvard Medical School,

00:01:01.260 professor in the department of epidemiology, along with chief of the division of prevention medicine

00:01:06.700 at Brigham and Women's Hospital. She is also a physician, epidemiologist, endocrinologist,

00:01:12.060 and the principal investigator or co-PI of several research studies, including the Women's Health

00:01:17.700 Initiative, which we of course discuss here in length, along with other studies such as the

00:01:21.920 cardiovascular components of the nurse's health study, the vitamin D and omega-3 trials known as

00:01:27.400 VITAL. Her primary research interests include randomized clinical prevention trials of nutrition

00:01:33.220 and lifestyle factors related to heart disease, diabetes, cancer, and the role of endogenous and

00:01:38.240 exogenous estrogens as determinants of chronic disease in women. Joanne has received numerous honors,

00:01:44.800 including the American Heart Association's Population Research Prize, the American Heart Association's

00:01:50.340 Distinguished Science Award, the Research Achievement Award, election to the Institute of Medical of the

00:01:56.100 National Academies and the National Academy of Medicine, membership in the Association of American

00:02:01.480 Physicians, fellowship in the AAAS, and so many other awards that I could spend the rest of the

00:02:07.000 introduction of this podcast going over them. She has published more than 1,200 peer-reviewed articles

00:02:13.020 in the medical literature and is the author or editor of several books and textbooks. She serves as the

00:02:18.460 editor-in-chief of Contemporary Clinical Trials and is the past president of the North American

00:02:23.560 Menopause Society. She is one of the most highly cited researchers in the history of published research

00:02:29.920 and one of the physicians who is featured in the National Library of Medicine's exhibition,

00:02:36.140 History of American Women Physicians. In this episode with Joanne, we spend the entire conversation

00:02:41.800 focusing on one of her main projects, the study that she was involved in called the Women's Health

00:02:48.000 Initiative, of which she was one of the principal investigators. In this discussion, we speak

00:02:53.100 about the reasons for the study, the questions being examined, the study design, inclusion-exclusion

00:02:58.940 criteria. We then go into the nuances of the study, including why it was prematurely stopped and how it was

00:03:06.780 interpreted. Of course, the most important part of this discussion is what the implications are for

00:03:13.260 someone today listening to this. Hormone replacement therapy is potentially one of the most controversial

00:03:19.880 bits of medicine today. And I would argue, and I make this point to Joanne, that the misinterpretation

00:03:27.540 of the Women's Health Initiative some 20 years ago may be one of the greatest missteps of medicine and,

00:03:34.220 by extension, the medical press in the past several decades. I make no bones about my bias here in this

00:03:41.240 podcast, which is that I think the fears of hormone replacement therapy are completely overblown and are

00:03:49.000 generally being propagated by people who are not familiar with the literature, which is why I wanted

00:03:53.920 to sit down with Joanne today. I could think of no better person to sit down with and go through the

00:03:59.280 details of this study than the person who is more familiar with them than anyone else. So, without further

00:04:05.460 delay, please enjoy my conversation with Dr. Joanne Manson. Joanne, it is great to finally be sitting

00:04:17.400 down with you. This is a topic that is arguably as important as any topic that we'll cover in this

00:04:22.500 podcast, and there's probably no better person to speak with about HRT than you. So, maybe just by way

00:04:30.240 of background, well, maybe I'll introduce you with one interesting statistic. Do you know how to

00:04:36.240 actually calculate the H-index? Yes. Okay. So, tell folks what the H-index is, how it's calculated.

00:04:44.300 I'll embarrass you by telling people how high you rank on that. Okay. This is very embarrassing,

00:04:49.980 Peter, but let me start by saying it's great to have a chance to talk with you, and I'm so glad that

00:04:55.240 you're interested in this subject and providing more information to your audience on this subject.

00:05:01.680 The H-index is calculated from the number of publications you have that are highly cited.

00:05:11.320 If, for example, you have an H-index of 100, that would mean that you have at least, we have 100

00:05:18.720 publications that have 100 or more citations each. An H-index of 200 would be 200 publications that each

00:05:30.560 have at least 200 citations are referenced in other publications. You know, we throw those numbers out

00:05:38.540 like 100 and 200. Those are epic H-indexes. I mean, a person with an H-index of 100 has done 10 people's,

00:05:47.960 you know, lifetimes work in their lifetime. Your H-index, Joanne, last I checked, is 305. Is that 0.72

00:05:54.840 correct? It may have crept up even higher. Yeah. Well, I would say that you are generally in the

00:06:04.160 top three H-index rankings in the history of biomedical science. And I know this because I've

00:06:12.040 checked this. I've never interviewed anybody with a higher H-index than you, and I've certainly

00:06:16.180 interviewed a lot of people with a very high H-index. So with that as a little bit of background,

00:06:21.520 let's go into a slightly more... It means I have wonderful colleagues and collaborations going on

00:06:26.900 throughout the world. Yeah, it certainly does. But let's talk more specifically about this

00:06:30.600 discussion. So there is a study that many people have heard me talk about. It's called the Women's

00:06:35.100 Health Initiative. It's a study that you were one of the principal investigators on. And it's also a

00:06:39.880 study that has produced results that I think we would look back today, 20 years later, and say

00:06:45.560 maybe weren't interpreted in the best way. And the implications of that are obviously significant

00:06:50.500 from a public health perspective. But let's go back in time to the beginning of the planning of

00:06:56.700 this study, which I assume would have been in the early 90s, is about when you and your peers decided

00:07:01.680 we needed a randomized experiment to test what was being found in the Nurses' Health Study and other

00:07:07.660 epidemiologic studies. Exactly. It was the very early 1990s.

00:07:13.000 Talk a little bit about what was observed through the observational studies, the epidemiologic studies,

00:07:18.460 and how that shaped the design of the WHI. Yes. So back in the 1980s and 1990s, there were several

00:07:27.560 observational studies. These are not randomized clinical trials, but large observational studies looking at

00:07:35.320 women who chose to be on hormone therapy or whom doctors were prescribing hormone therapy for.

00:07:40.600 And they did tend to have lower rates of heart disease in those studies compared to women not 0.79

00:07:48.240 using hormone therapy. They also seemed to have more favorable outcomes, such as less cognitive decline,

00:07:56.660 lower all-cause mortality rates. They were generally doing better. But we often say that observational

00:08:04.380 studies of this nature cannot prove a cause and effect relationship, but they can generate hypotheses

00:08:13.560 to be tested in randomized clinical trials. But before the randomized clinical trials were even launched

00:08:21.400 in the early 1990s, there was already an increasing practice in clinical medicine to prescribe hormone

00:08:30.880 therapy for the express purpose of trying to prevent heart disease, cognitive decline, and other chronic

00:08:40.660 diseases. So this was a trend that was occurring not only in recently menopausal women, not only when

00:08:49.860 women had hot flashes and night sweats and were in early menopause, but many clinicians were starting to

00:08:57.820 prescribe these hormones for women who were well over a decade, 10, 20, 30 years after the onset of menopause.

00:09:06.600 So it's very important to understand whether this practice of prescribing menopausal estrogen therapy or

00:09:16.400 estrogen plus progestin therapy was advisable when used for prevention of chronic diseases. This was a very

00:09:27.680 very different question compared to question compared to asking, does hormone therapy reduce hot flashes,

00:09:34.320 night sweats, and should women in their 40s, early 50s, who are just starting to go through menopause and have these

00:09:42.960 symptoms, should they take hormone therapy to treat those symptoms? It was accepted that hormone therapy is

00:09:51.760 effective for treating hot flashes and night sweats. It's actually FDA approved for that purpose. It has an

00:10:01.200 indication for treatment to reduce hot flashes and night sweats. But the question of its use for prevention of

00:10:09.680 heart disease, stroke, cognitive decline, other chronic diseases had never been tested in a randomized

00:10:17.060 clinical trial. And that was the goal of the Women's Health Initiative. Despite the fact that the

00:10:22.060 epidemiology suggested benefits in all of those arenas, people who listen to this podcast are no

00:10:27.500 strangers to the different types of biases that can creep in. And perhaps there's no greater bias that

00:10:33.420 creeps into something like this than the healthy user bias. It could easily be the case that the women who had 0.98

00:10:38.920 access to physicians or who had access to the type of physicians who maybe felt more knowledgeable or

00:10:45.500 provided better care. And part of that could have been the provision of hormones. It could be that

00:10:50.640 they were coming down with fewer cases of chronic diseases, not because of the hormone replacement

00:10:56.100 therapy, but because of other factors that were a part of a healthy lifestyle of which hormones might

00:11:01.380 have been a part of it. So there's no doubt that an RCT is going to be essential to carry out the

00:11:06.960 elucidation of causality here. Let's also talk a little bit about the formulations. I guess if we go back into

00:11:12.720 probably, gosh, the 1960s when the idea first came along to physicians to replace estrogen in a

00:11:20.920 menopausal woman, I think they were just probably using estrogen alone, correct? And they didn't 1.00

00:11:25.080 understand the role of progesterone as opposition to that to prevent endometrial hyperplasia. But where

00:11:31.840 were we in the early 90s? Clearly that had been figured out. What were the formulations that were most

00:11:37.260 common? Well, that's a very important question, Peter. And the most common formulations were

00:11:42.500 conjugated estrogen with and without medroxyprogesterone acetate. And so women who had a

00:11:49.900 hysterectomy could use estrogen alone, but women with an intact uterus needed to take a, what we call a 1.00

00:11:57.400 progestogen, which counteracts the effect of estrogen on increasing the thickness of the uterine lining,

00:12:06.140 the endometrium. And so if women who have an intact uterus take estrogen alone, they have a very high

00:12:13.100 risk of developing endometrial cancer. And early on, they will just have proliferation of the lining of

00:12:20.900 the uterus and increased vaginal bleeding related to taking estrogen without the progestogen. So those

00:12:28.160 were the two formulations that were very commonly used. And also importantly, those were the two

00:12:34.980 formulations that had been extensively studied in the observational studies where the results had

00:12:42.240 looked very promising for a lower risk of heart disease and cognitive decline, all-cause mortality.

00:12:49.740 So it was felt to be important to test the formulations that had contributed so much to the observational

00:12:57.500 study findings. As you say, the women who were taking hormone therapy in the observational studies

00:13:04.580 tended to be a higher socioeconomic status, more highly educated, and more health conscious. And these

00:13:12.980 were all potential confounding factors that may have contributed to their lower risk of chronic diseases.

00:13:21.840 However, it's also important to note that in observational studies, the women who were being

00:13:27.900 prescribed hormone therapy were still largely women in early menopause. They were at least being

00:13:34.120 started in early menopause, even if they continued into mid and later menopause. So that's another

00:13:41.780 important, perhaps biological difference between the women in the observational studies and women in 0.95

00:13:47.880 randomized trials, where in the Women's Health Initiative, the average age was 63 or more than a decade

00:13:55.880 past onset of menopause when the hormone therapy was being started.

00:13:59.860 I know you're not a gynecologist, so you may not know the answer to this, but do we know if the age

00:14:06.060 of menopause is moving over time? I mean, we certainly know that girls are getting their periods 0.65

00:14:12.240 earlier and earlier, even over just two decades. Do we know if menopause is also a moving target?

00:14:18.020 I don't know that that's been studied really rigorously and systematically. The average age of

00:14:25.500 menopause is 51. I believe that stayed relatively constant for quite a while. Although, as you say,

00:14:33.300 the age of puberty and menarche, start of menstrual periods, has become younger over time. So there are

00:14:42.200 those changes, but I don't know that there have been really clear differences in the age at menopause.

00:14:48.860 Do you have a sense of why the conjugated equine estrogen and the MPA, a synthetic progesterone,

00:14:56.240 were the dominant forms of these hormones used in the 80s and in the 90s, which of course then

00:15:02.240 became the precursor for the epidemiology? For example, do we know why there was not just a

00:15:07.720 bioidentical estradiol and progesterone, which of course is, we'll talk about those things later in

00:15:12.880 our discussion because of course those are the most dominant forms used today. But do you have a sense of

00:15:16.860 just historically why that was not the case even at the outset?

00:15:20.180 There are some theories about that, Peter. You know, one theory is that a pharmaceutical company

00:15:26.380 developed the conjugated estrogens. As you know, they were originally, and even many of the forms

00:15:33.580 today, derived from the pregnant mare's urine. And this pharmaceutical company really became the

00:15:40.760 dominant force in terms of hormone therapy. And the synthesis of estradiol, you know, from plants,

00:15:49.040 and it's a more complicated process that really did not get going on a very large scale until more

00:15:56.140 recent decades. But for quite a long time, many, many decades, more than 50 years, there was, you know,

00:16:03.680 the conjugated estrogen available. So the investigators of which you're, how many of you were actually sort

00:16:11.260 of lead PIs? Were there three of you? In the overall Women's Health Initiative, there were initially 16

00:16:17.660 clinical centers and then expanded to 40 clinical centers for most of the duration of the WHI. So

00:16:25.020 there were actually 40 principal investigators throughout the country. Let's go over sort of

00:16:30.420 study design and participant criteria, inclusion and exclusion criteria. So you've already alluded

00:16:34.500 to several of these. And this is a very important distinction. And it's something I think you and I

00:16:38.780 even discussed many, many years ago, not on a podcast, but just over the phone one day, which was

00:16:43.240 women who were having vasomotor symptoms were excluded, correct? No. We're not excluded. This actually is a

00:16:50.060 common misconception that women could not have hot flashes to participate in the WHI. Women who had 1.00

00:16:58.240 very severe hot flashes self-selected out of the study because they wanted, generally, they wanted

00:17:06.000 to be on hormone therapy. At that time, the assumption was that hormone therapy would have

00:17:12.620 very favorable effects. And they were already taking hormone therapy very often for their severe hot

00:17:20.180 flashes. So they did not want to be in the study. But we didn't exclude any woman on the basis of the

00:17:26.740 severity of her symptoms or presence or absence of hot flashes, night sweats. In fact, the majority 0.52

00:17:32.840 of the women in the study, especially those in the earlier younger menopausal ages, did have at least

00:17:41.600 mild or moderate hot flashes.

00:17:44.980 I see. I misunderstood you. I thought the thinking was by minimizing the number of women who had

00:17:50.780 vasomotor symptoms, you would have less dropouts in the placebo group because, of course, women who were 1.00

00:17:56.320 getting a placebo would presumably not be relieved of those symptoms, which are obviously the most

00:18:01.180 responsive to the HRT. You said the average age was 63.

00:18:05.720 Is that correct?

00:18:06.420 What were the exclusion criteria?

00:18:09.060 So women could not have a prior history of breast cancer, endometrial cancer, or any other

00:18:18.320 estrogen-sensitive cancers. And overall, if they had cancer, it had to be, could not be estrogen-sensitive

00:18:25.720 cancer. And it had to be more than 10 years previously. Women could not have a recent heart 1.00

00:18:31.940 attack or stroke or any of those major clinical cardiovascular events, although a very small

00:18:38.620 percentage of the women did have a more remote history of heart attack, stroke, bypass surgery,

00:18:46.160 that type of clinical history. For the most part, the women were healthy in terms of past history of

00:18:54.220 cardiovascular disease, cancer, other conditions. They were certainly allowed to have diabetes,

00:18:59.920 hypertension, high cholesterol, and many of the more common health conditions that women will have

00:19:07.060 and men will have in midlife. 0.99

00:19:10.400 Smoking, not an issue.

00:19:11.840 They were not excluded for smoking, though we had a small percentage, less than 10%.

00:19:16.500 Osteopenia, osteoporosis was not an exclusion?

00:19:19.180 It was not an exclusion. They were not selected on being required to have osteoporosis or osteopenia,

00:19:27.140 but it was a broad range of bone health, similar to what you would expect in the usual population

00:19:33.420 for women age 50 to 79.

00:19:37.660 And what about family history for breast cancer or uterine cancer?

00:19:42.580 Women were allowed to participate. It was really up to them whether they thought,

00:19:49.180 that their family history was so strong that they did not want to take any chance of being randomized

00:19:54.560 to active hormone therapy. Many women did self-select out of the study for that reason,

00:20:01.380 but they were not excluded by these study investigators on the basis of their family history.

00:20:08.560 And was there a limit as to how long they could be out of menopause before enrollment?

00:20:13.740 No, it was on...

00:20:14.780 Sorry, into menopause, yeah.

00:20:15.860 The criteria were based on age. So the women were 50 to 79 with a mean age of 63. Some of them had gone 1.00

00:20:25.720 through menopause in their early to mid 40s. Some of them had even had hysterectomy with ovaries removed

00:20:34.880 in their 40s or much earlier in life. So there was no exclusion on the basis of, for example,

00:20:43.080 being more than 20 or 30 years past menopause. It was on the basis of age 50 to 79.

00:20:50.880 Wow. Amazing diversity of age there, right? I mean, you think about it's basically three

00:20:55.180 decades worth. What about prior hormone use? What fraction of the women had previously been on 0.97

00:21:02.160 the exact same drugs that they were going to be potentially randomized to? And was there a required

00:21:07.260 period of washout? Okay, very good question. About 25% of the women in the estrogen plus progestin trial

00:21:15.760 had prior use of hormone therapy. And close to 50% of the women with hysterectomy and in the estrogen

00:21:25.340 alone trial had some prior use of estrogen therapy. So women were not excluded for having a past history

00:21:35.300 of using hormone therapy. But especially in the estrogen plus progestin trial, the large majority

00:21:42.160 did not have prior use. It was only about 25% who had prior use. And overall, just looking more

00:21:50.100 specifically at the percentage of women in the trial who had hot flashes, night sweats at the time 0.70

00:21:58.240 of enrollment, it was overall about 45%, 45 to 50% who had some symptoms, mostly mild or moderate hot

00:22:11.440 flashes or night sweats. And a little over 50% did not have any hot flashes or night sweats at the start

00:22:19.440 of the trial. How many women are enrolled in this trial? About just under 30,000, am I remembering that

00:22:24.520 correctly? Yes. So 27,000, a little over 27,000 in the two trials combined. In the estrogen plus

00:22:31.640 progestin trial, close to 17,000. In the estrogen alone trial, close to 10,000. And again, we've stated

00:22:40.280 this, but I think it's very important that people again understand why there are two trials. If a woman 1.00

00:22:45.620 has a uterus, she must be receiving progestin along with the estrogen. So that's the E plus P trial that 0.90

00:22:53.380 we'll talk about. If a woman has had a hysterectomy, then estrogen alone is sufficient. There's no risk

00:22:59.700 of endometrial hyperplasia because there is no endometrium. And there's an E alone trial. Each

00:23:05.420 group has its own placebo? Yes. Separate placebo for each group. Great. So roughly, we're talking about

00:23:11.480 four groups. The E only is roughly 5,000 plus 5,000, where they randomized one-to-one. Exactly.

00:23:17.260 And then the E plus P would have been roughly 8,500 in E plus P and then 8,500 placebo for that

00:23:25.360 group. Correct. Okay. Primary outcome, was there a single primary outcome and was it ASCVD? Was that

00:23:31.720 the main outcome? The primary outcome for both trials was coronary heart disease with the primary

00:23:39.740 safety outcome because of concern, even before the study was designed, was breast cancer. So the two

00:23:48.360 really key outcomes of the trial were coronary heart disease and breast cancer.

00:23:54.560 And obviously, this is incidence of breast cancer, not mortality of breast cancer, correct?

00:23:58.740 Incidence. A diagnosis of breast cancer that's confirmed by medical record review.

00:24:03.520 And coronary artery disease would be what we would think of today as MACE. So major adverse

00:24:09.140 cardiac event, MI, stroke, death as a result of anything, or was it more complicated?

00:24:14.900 The primary outcome was actually the coronary events. So it was heart attacks. It was either

00:24:22.600 non-fatal heart attack or fatal coronary disease, which is most commonly a fatal heart attack.

00:24:30.780 What was the study powered to detect on either of these? Obviously, those numbers,

00:24:38.220 the study subjects is something that I'm sure the investigators thought long and hard about.

00:24:42.300 What was the power analysis suggesting a difference that was anticipated?

00:24:46.760 Remember that the prior hypothesis was that there would be benefit for heart disease and that there

00:24:54.060 would be overall many favorable effects of these hormones on chronic disease outcomes.

00:25:00.420 It was powered to detect an important clinical reduction such as a 20% reduction in heart disease.

00:25:11.480 Do you recall at the time what fraction of the women were taking lipid-lowering therapies?

00:25:18.860 How common would that have been in the early to mid-90s? You know, I think one of the interesting

00:25:24.180 aspects of the WHI is that, as you pointed out, hormones were being used as a preventive treatment

00:25:32.080 for ASCVD. You know, I think today very few physicians would really consider that. I think

00:25:38.060 given the complete explosion we've had in both the availability, variety, and efficacy of lipid-lowering

00:25:46.000 agents, a person who's deemed at risk is going to be managed much more critically with respect to

00:25:51.360 blood pressure and lipids. Again, I don't know that I've ever come across that stat. That may

00:25:56.220 not be something that you know, but I'm just curious as to how prevalent that was.

00:25:59.580 Yes, I do know it. It's right in the paper that we published in JAMA in 2013, and we published

00:26:06.700 repeatedly on the subject of the use of many of these medications for chronic health conditions.

00:26:12.780 So 7% of the study population was taking statins at the start of the study. By later in the trial,

00:26:24.880 during the intervention phase, it was over 25%. There was a very large increase in statin use,

00:26:33.520 even during the trial itself, and with longer follow-up. Obviously, these percentages are even

00:26:40.040 higher, getting to 40%, 50%, very high percentage. What was known about oral estrogen at the time?

00:26:48.980 Again, going back to the early to mid-90s, as far as its impact on coagulability, for example,

00:26:56.440 blood viscosity. Today, oral estrogen is not used very often as a result of that, but at the time,

00:27:02.760 what was known?

00:27:03.840 It was understood that oral medications and oral estrogen goes directly through the portal

00:27:12.480 circulation to the liver and has a direct effect on the liver in increasing the synthesis of clotting

00:27:20.040 proteins. That was understood at that time because similar to oral contraceptives, it had been seen

00:27:28.540 in observational studies that both oral contraceptives and postmenopausal hormone therapy are associated

00:27:35.700 with an increased risk of deep vein thrombosis and even pulmonary embolism. That had been suggested,

00:27:43.080 but it was believed to be relatively rare and that the benefits for heart disease and for other

00:27:51.160 chronic diseases would outweigh those risks of thrombosis.

00:27:56.700 What was your personal hypothesis going in? I mean, there's obviously the hypothesis that is

00:28:01.840 driving the study, but do you remember back, you know, gosh, it's probably 30 years now since you were

00:28:08.540 in the planning phase of this. What did you think was happening mechanistically to explain the observational

00:28:14.920 data? Did you believe the observational data? Did you think that they were being confounded heavily?

00:28:20.080 Do you have a recollection of that?

00:28:21.740 I believe that it was likely that there was at least a small amount of confounding because it was

00:28:29.120 clear that women taking hormone therapy tended to have a higher socioeconomic status, tended to have

00:28:35.260 better access to medical care and to have somewhat more favorable lifestyle behaviors, be more health

00:28:42.200 conscious. Many of these factors were considered in the data analyses. It's not like the observation

00:28:50.060 studies were just looking very crudely at associations. There was adjustment for many of these

00:28:56.920 lifestyle factors and socioeconomic status in some of the studies, but a benefit, a risk reduction for

00:29:05.520 heart disease did tend to persist even after those adjustments. My thought was that women in early 0.99

00:29:13.960 premenopause who are transitioning from having their natural premenopausal estrogen exposure,

00:29:22.740 which many studies suggested was protective, a cardioprotective favorable in terms of risk factor

00:29:31.760 status, in terms of dilating the blood vessels to the heart and increasing blood flow to the heart. There

00:29:38.180 were many studies already suggesting that a woman's own natural estrogen during premenopausal years was one of

00:29:45.360 the reasons why women started to have heart disease later, 10 years or more later than men. My thoughts were that

00:29:53.140 it is likely that starting estrogen in early menopause would translate into at least a slightly lower risk of heart

00:30:03.340 heart disease. But I was skeptical from the very start that the magnitude of risk reduction seen in the

00:30:11.780 observational studies, you know, the 40%, 50% lower risk of heart disease would stand up to a randomized clinical trial

00:30:23.400 assessment of this question. So I thought it was likely to be a small reduction, wasn't sure whether that benefit might be

00:30:32.740 offset by other risks that would be identified. But overall, you know, I thought there's likely to be some

00:30:40.020 confounding going on in the observational studies.

00:30:44.280 When you think back to, again, that same period of time, what was your thinking with respect to the relationship

00:30:49.580 between estrogen and breast cancer? You know, obviously the classical teaching, I mean, I was in medical school in the

00:30:54.460 mid-90s and you were taught chapter and verse that estrogen causes breast cancer. But on the surface, some of the

00:31:01.180 assumptions are a little hard to understand. In the same way that there has historically been the assumption

00:31:04.980 that testosterone causes prostate cancer, except for the observation that men with the highest levels

00:31:10.040 of testosterone, i.e. men when they're, you know, younger, have lower rates of prostate cancer than men

00:31:14.640 when they're older. And similarly, I mean, this was demonstrated as long as 15 years ago that the lowest levels

00:31:19.720 of testosterone were associated with the most aggressive forms of prostate cancer. I don't know how similar the

00:31:25.780 data are for estrogen and breast cancer, but given that most women get breast cancer in menopause and

00:31:32.500 not prior to menopause, they're getting breast cancer, even estrogen-sensitive breast cancer at a time when

00:31:37.180 they have their lowest levels of estrogen. So what was the understanding of the pathophysiology of the

00:31:42.380 relationship between estradiol or estriol even and, or maybe it's estrone, it was deemed to be even more

00:31:49.220 problematic. What was the understood relationship for why that relationship existed?

00:31:53.880 There was an expectation that there would be at least a modest increase in risk of breast cancer

00:32:01.980 with giving either estrogen plus progestin or estrogen alone. In fact, the most surprising

00:32:08.600 finding in terms of breast cancer was that no increased risk of breast cancer was seen with estrogen

00:32:15.980 alone, even though the observational studies had suggested that both estrogen plus progestin and

00:32:22.360 estrogen alone would be associated with increased risk of breast cancer. But there were many, many

00:32:29.260 observational studies suggesting that hormone therapies associated with increased risk of breast

00:32:36.120 cancer, but the thinking tended to be these are estrogen receptor positive breast cancers. They tend to be

00:32:44.620 more favorable outcome types of breast cancer. And that breast cancer mortality would not be appreciably

00:32:54.620 increased. So that was the thought going into it. And also in the observational studies, there was

00:33:01.580 always that concern that differences in mammographic screening patterns could be contributing to greater

00:33:09.740 detection of breast cancer among women taking hormone therapy because most doctors would not continue to

00:33:16.540 prescribe the hormone unless the woman was having regular mammography and showed a normal mammogram

00:33:23.340 without concern about a lesion there. So mammography also could have been contributing more frequent

00:33:30.580 mammography in women on hormones versus women not taking hormone therapy could have contributed somewhat to the

00:33:38.140 increased risk in the observational studies, which was why it was important to look at this question

00:33:43.180 in a randomized clinical trial with uniform surveillance for breast cancer with a mammogram being required every

00:33:53.340 year. So there was an increased risk of breast cancer with estrogen plus progestin. Also, there were denser

00:34:02.300 breasts developing over time. So breast density was looked at in a study of mammograms, several hundred

00:34:11.180 women whose mammograms were examined, and there was a change in increased breast density, which is known to be a risk 0.94

00:34:21.100 factor for breast cancer. The increase in breast density was greater with estrogen plus progestin than with estrogen

00:34:28.300 alone. Now, what was really surprising, as I mentioned, although there was this increase, 25-30% increase in risk of breast

00:34:38.780 cancer seen with estrogen plus progestin, there was no increase in breast cancer seen with estrogen alone. And with longer

00:34:47.260 follow-up, there was the emergence of a reduction in breast cancer close to a 20% reduction seen with the conjugated

00:34:56.300 estrogen. And the view was that this may be something specific to conjugated estrogen, which is a relatively

00:35:04.540 weak estrogen, and may have certain properties similar to tamoxifen, where it may be both serving as an estrogen

00:35:14.780 and an anti-estrogen. But we cannot assume, importantly, we cannot assume that this finding with conjugated

00:35:23.420 estrogen will necessarily apply to all formulations of estrogen alone and certainly will not apply to

00:35:33.340 the combination of estrogen plus progestin. So that's interesting, Joanne. I wouldn't come to that as my

00:35:40.220 first kind of Occam's razor conclusion because the same conjugated equine estrogen was used with the MPA

00:35:47.900 that found a clinically irrelevant but statistically significant increase in the incidence of breast

00:35:54.700 cancer. In other words, I wouldn't conclude from the differences in those two arms that it was the

00:36:01.340 conjugated equine estrogen that was unusually beneficial. No, I'm not. I'm not concluding.

00:36:07.180 Wouldn't it be that the MPA was the difference? Yes, it's the MPA. But the only way to look at

00:36:13.660 the role of the estrogen is in the estrogen alone trial because in the estrogen plus progestin trial,

00:36:21.740 you can't disentangle. It was given a combined pill. Every woman in the trial was taking the

00:36:26.380 combination. So I completely agree with you, Peter. The increased risk of breast cancer seen with estrogen

00:36:33.740 plus progestin was mostly attributable, if not entirely attributable, to the progestin,

00:36:42.780 to the medroxyprogesterone acetate. And of course, the question has been raised, would other types of

00:36:49.020 progestogens, such as the bioidentical micronized progesterone, would that also lead to an increased

00:36:57.180 risk of breast cancer? There are some observational studies that suggest less increase in breast cancer

00:37:03.900 with that particular formulation of progestogen. However, we have no large-scale randomized clinical

00:37:13.100 trials that have done head-to-head comparisons or even really tested long-term the effects of other

00:37:21.180 formulations of progestogen on breast cancer risk. So probably the most prudent and cautious approach is

00:37:29.980 to assume that at least with longer duration of treatment with estrogen plus progestin,

00:37:35.820 there will be an increased risk of breast cancer. Let's go back and talk about the study being

00:37:41.100 stopped. So the E plus P arm, so the CEE plus MPA arm was stopped early, correct? A little over five years?

00:37:48.700 It was stopped after 5.6 years, and it was stopped 3.3 years early.

00:37:55.340 And this was stopped presumably on the basis of the finding for the increased incidence of breast

00:38:01.580 cancer, correct? It was stopped on the basis of the increased risk of breast cancer together

00:38:08.220 with no reduction in heart disease, which was the primary endpoint, and an overall unfavorable

00:38:17.180 risk-benefit ratio as shown through the global index, which looked at all of these chronic conditions.

00:38:23.580 Now, one of the things that I didn't realize until somewhat recently was that the dropout rate

00:38:30.300 was a little bit unusual in that in the placebo arm of the E plus P group in the

00:38:36.300 first, second, third year relative to the final year. Do you recall what those numbers were?

00:38:41.260 The dropout rate is actually substantial in all hormone therapy trials. It's a combination

00:38:48.220 of people do dropout of all trials because it requires a lot of effort to take a medication

00:38:56.060 that's not being prescribed for your health. It's just part of a study. You don't know exactly

00:38:59.980 whether it's active or placebo. So all trials have some dropout over time in terms of compliance,

00:39:07.820 adherence with study medications. Also, there are some side effects of hormone therapy. The women who had an

00:39:14.540 an intact uterus and were taking estrogen plus progestin, some of them continued to have some

00:39:20.220 vaginal bleeding, and some of them did not want to continue to have those symptoms. They may have

00:39:25.820 had some breast tenderness or other symptoms and dropped out. And women dropped out of the placebo

00:39:32.540 arm as well. Some of that is just taking a medication day in and day out, taking a study pill. Fatigue will

00:39:39.260 set in in any randomized trial. Our participants were extraordinarily dedicated. I cannot imagine

00:39:47.340 any group being more committed to women's health and to getting answers for women on these very

00:39:53.660 pressing issues in menopausal women's health. But there's going to be some dropout in any trial. 1.00

00:40:00.300 Do you think the answer might have been different if the trial had gone to,

00:40:04.780 it was originally planned to be about nine years? Over eight years. I mean, I guess we'll never know

00:40:11.500 what it would have shown had it gone longer. So at the time that the study was halted,

00:40:16.060 really the big headline of the study was estrogen causes breast cancer. I use this as a great teaching

00:40:23.500 example when I talk about the difference between relative risk and absolute risk. The relative risk

00:40:28.620 difference in the group that was CEE plus MPA relative to placebo or versus placebo was 24, 25%,

00:40:36.540 correct? Yes. I think it was a little higher. On the surface, that sounds incredibly startling,

00:40:41.100 right? So women who are getting CEE plus MPA have a 25% higher risk of breast cancer during this 5.2 year 0.82

00:40:49.020 period. The absolute risk increase was 0.1%. It was a difference of one case per thousand.

00:40:55.820 Exactly. The women in the placebo group were getting breast cancer at an incidence of four 0.93

00:41:01.740 cases per thousand women. The women in the CEE plus MPA group were getting it at an incidence of

00:41:07.340 five cases per thousand. The absolute difference of that being one case per thousand. It was about

00:41:13.420 the opposite that was seen in the CEE versus placebo alone, although I don't think it reached

00:41:19.500 statistical significance at 5.2 years, did it? No, it didn't.

00:41:22.780 It took longer follow-up to see that reduction. Yes.

00:41:27.420 To see that there was indeed a reduction. In many ways, it seems that that was the headline

00:41:33.500 that dominated the world. I don't have any recollection of it at the time, Joanne, because

00:41:37.240 I was in my surgical residency when the study was first published, and I must admit, I wasn't paying

00:41:42.180 attention to this literature. I'm embarrassed to admit, I don't even recall this paper coming out.

00:41:46.920 What is your recollection of that time? How much it sort of captured the imagination and fear of the

00:41:51.980 world? Well, the medical community was shell-shocked. There was a sea change in clinical practice,

00:41:59.900 a seismic shift in clinical practice. But let's think about what the shift was. Two major things

00:42:07.880 changed. There was a dramatic reduction in use of hormone therapy, 70% to 80% reduction in use of

00:42:16.640 hormone therapy. So women who were being prescribed hormone therapy in the past for prevention of 0.98

00:42:24.060 chronic diseases, prevention of heart disease, stroke, cognitive decline, those women were no longer

00:42:31.520 being prescribed hormone therapy. And that was a positive thing. That was a positive change in

00:42:38.040 clinical practice because the WHI showed that when hormone therapy is used for prevention of

00:42:47.040 chronic disease purposes in women who have average age of 63 and women in mid to later menopause on

00:42:54.400 average, the risks outweighed the benefits. So it was a favorable change. But the unfavorable change in

00:43:00.040 clinical practice was that the results were extrapolated to women in their 40s and 50s who were taking

00:43:08.560 hormone therapy for treatment of bothersome, even distressing, hot flashes, night sweats, and were in

00:43:15.840 generally good health. And they were being denied hormone therapy to relieve these symptoms. And that was an

00:43:24.880 inappropriate extrapolation of the findings. That was a negative outcome. Women never should have been 1.00

00:43:34.000 denied hormone therapy for the treatment of bothersome, distressing, hot flashes, night sweats to improve

00:43:42.240 their quality of life, especially generally healthy women in early menopause who have such low absolute 1.00

00:43:48.900 rates of adverse events. Now, as you say, in the WHI, even in the overall cohort with an average age of 63,

00:43:58.900 most of the adverse events were relatively rare, such as one extra case of breast cancer or heart attack

00:44:09.020 or blood clot per 1,000 women per year. I mean, it's still important and it can add up with longer term use.

00:44:17.620 However, it's a low absolute risk. And I think that the results were perhaps blown out of proportion,

00:44:29.000 especially in terms of the use of hormone therapy for a clear indication, an FDA-approved indication of

00:44:36.660 treatment of hot flashes and night sweats among women in early menopause who had even lower absolute

00:44:42.960 risks than one extra case per 1,000 women per year. And that type of extrapolation really should not have

00:44:52.900 happened. So there were some positive outcomes that hormone therapy was being used less for inappropriate

00:45:02.300 purposes, such as prevention of heart disease, stroke, cognitive decline, but also this unfavorable

00:45:10.140 change that it was being used less commonly for treatment of bothersome hot flashes and night sweats

00:45:18.440 in early menopause.

00:45:20.280 I guess what I find most troubling is that when I speak with most physicians today, unless they're really,

00:45:26.760 really steeped in this world, and most aren't, they can't reiterate what you just said. The only thing

00:45:33.260 that they seem to understand, and by extension, the only thing that their patients seem to be led to

00:45:38.100 believe, you know, really the enduring legacy of the WHI from an insight perspective is that hormones,

00:45:47.200 in particular estrogen, cause breast cancer, and that HRT is synonymous with breast cancer.

00:45:54.720 And there are really two enormous inaccuracies in that that are so inaccurate that they're almost a

00:46:02.840 parody, right? And when we've discussed them both, but I think it's always worth bringing it back

00:46:06.100 for the listeners so that they aren't lost in the details. The first is that nothing about this

00:46:11.720 study suggested that estrogen is causing breast cancer. If anything, this study suggested MPA

00:46:18.540 is causing breast cancer based on the fact the group that was only receiving estrogen had a

00:46:24.400 reduction in the incidence of breast cancer, while the group that received estrogen plus MPA is the

00:46:29.960 group that saw this small, potentially statistically significant, but potentially clinically

00:46:34.860 insignificant increase. The second thing that seems to get lost from this, and again, it's sad that

00:46:40.640 most doctors and patients who are contemplating HRT don't recognize this, is that the study only found

00:46:45.980 an increase in the incidence of breast cancer to the tune of one case per thousand, but no difference

00:46:51.660 in mortality. One has to wonder what the cost of that was. Let's talk about some of the other outcomes

00:46:59.380 here. First of all, would you agree with that, by the way, that that would be perhaps a more

00:47:02.940 charitable interpretation of the WHI? Two caveats. One, I think we cannot assume what was found with

00:47:09.900 conjugated estrogen will apply to all other formulations of estrogen, and I think there is

00:47:15.640 very strong evidence from several lines of biology that estrogen plays a role in breast cancer, including

00:47:24.620 the fact that when women have their ovaries removed, it lowers their risk of recurrence, lowers the risk of

00:47:31.540 development of breast cancer, lowers the risk of recurrence, and I think there is quite a bit of

00:47:37.380 evidence that estrogen does play a role. The higher estrogen, estrone, would be specifically the type of

00:47:45.100 estrogen that, for example, is associated with increased adiposity, and postmenopausal women is linked to a higher

00:47:53.640 risk of breast cancer. So I think there are several lines of evidence that estrogen is a factor in breast

00:48:01.880 cancer. And this was actually also known, even before the WHI, that many of the observational studies

00:48:09.480 had linked hormone therapy to an increased risk of breast cancer. But I agree with you that it seems to be,

00:48:18.480 when you're talking about conjugated estrogen, it's the combination of estrogen plus the medroxyprogesterone

00:48:25.820 acetate. It may be specific to that particular progestin. We don't know. We have very limited

00:48:31.900 research in terms of randomized clinical trials, large-scale studies of other formulations of

00:48:38.600 progestin. In terms of breast cancer mortality, the results in the WHI are very close to a statistically

00:48:48.800 significant increase in breast cancer mortality with estrogen plus progestin. It doesn't quite make

00:48:56.740 statistical significance, but it is getting toward a significant increase in risk. On the other hand,

00:49:04.500 the conjugated estrogen alone was associated with a statistically significant reduction in breast

00:49:12.520 cancer mortality. So really diametrically opposite effects of the combination with the medroxyprogesterone

00:49:21.080 acetate of the estrogen with the medroxyprogesterone acetate versus the conjugated estrogen alone. I think

00:49:28.280 that the absolute risks are still low. And that is a very important point to emphasize that we are

00:49:36.920 talking about one extra case per thousand women per year. For a woman who starts out with a high 1.00

00:49:45.320 baseline risk because of a strong family history or other risk factors, this would certainly be something

00:49:50.660 she would want to avoid. But for a woman in early menopause at usual risk of breast cancer, 0.71

00:49:56.220 who would derive the benefits of symptom relief and she's suffering from disrupted sleep, very

00:50:03.060 bothersome hot flashes and night sweats that interfere with her day-to-day activities, this would be a risk 0.97

00:50:10.440 that she would probably be willing to accept with an understanding that all medications have some risks and

00:50:18.260 that she could be monitored closely with mammography and breast exams. And, you know, usually there would not be

00:50:26.220 a fatal form of cancer diagnosed. Overall, we did not see any significant increase in total cancer,

00:50:36.300 total invasive cancers with either estrogen plus progestin or estrogen alone. And interestingly,

00:50:42.960 with the combination estrogen plus progestin, we saw a significant reduction in the endometrial cancer,

00:50:51.220 the uterine cancer developing over time and colorectal cancer also seem to be at least borderline reduced.

00:51:01.020 So I think the clinical message and the message for the public is that hormone therapy has very complex

00:51:09.820 effects. It has a complex matrix of benefits and risks that vary according to a woman's age,

00:51:21.220 her time since menopause, her underlying health status, and decision-making about hormone therapy really has to be

00:51:30.300 individualized, personalized, and women themselves play such an important role in the shared decision-making. 0.98

00:51:40.120 Because many women will say, I do not want to take hormones no matter what, and even if a woman's at low risk,

00:51:46.280 and the doctor or clinician may think that they really should consider hormone therapy, that's the woman's decision, 1.00

00:51:54.640 and you respect it. However, other women, let's say, will say that they want to take hormone therapy, 1.00

00:52:03.300 even though they know that there may be X, Y, or Z increased risk, because their symptoms are so bothersome,

00:52:13.740 they're so distressing, they're being disrupted, their sleep is being disrupted,

00:52:17.880 they are not being able to have work productivity, their day-to-day activities are affected, their quality of life is

00:52:25.860 really impaired. And it is very important that women be able to help make that decision together with their clinician

00:52:35.400 that they share in that decision-making.

00:52:37.680 I don't disagree with any of that. I guess the only issue I would potentially highlight is that in the scenario

00:52:43.920 you described where the physician thinks that it's the right thing to do, the patient is experiencing

00:52:50.060 vasomotor symptoms, the patient already has osteopenia, and their bone mineral density is only going to

00:52:57.220 decline with time, and they're not at unusually high risk for breast cancer. That scenario where the patient

00:53:03.300 is in, you know, almost a fear mode of saying, oh my god, no way I want, you know, HRT, it's going to cause

00:53:08.740 breast cancer. I guess what I'm arguing is that's a very ill-informed decision. So yes, it's their

00:53:13.500 decision, and yes, they should be able to make whatever decision they want, but that doesn't mean

00:53:17.180 it's a good decision, and that doesn't mean it's a decision based on good data. Because in reality,

00:53:21.540 it's not based on data. That's my fear, is that this is no longer about the data, because I think the data,

00:53:27.180 what you and I are discussing today is the data. And the data really don't make a very strong case

00:53:32.180 for avoiding HRT outside of select circumstances, and yet I worry that the last 20 years of women 1.00

00:53:39.500 entering menopause have been put into two categories, right? Either there are women who really want HRT, 0.91

00:53:45.520 but they can't find physicians who will give it, or there may be a physician who understands the data

00:53:50.760 at the level you and I are discussing it, but the patients themselves have been so frightened off it

00:53:54.820 by misinterpretations of the data. I think the conversation has to take place. And, you know,

00:54:01.840 it is important for the clinicians to be informed themselves about benefits and risks and be able

00:54:08.580 to discuss the benefits and risks in a very knowledgeable way with the patient. But my view

00:54:15.320 is that if a patient feels strongly that she doesn't want to take hormone therapy, maybe, you know, her mother

00:54:20.940 develop breast cancer while she was on the hormones, whether or not it was directly due to the hormones.

00:54:27.540 So she is going to have a lot of fear and anxiety surrounding use of hormone therapy, and that will 0.98

00:54:33.640 affect her overall well-being. That's a factor in the benefit-risk equation. So I would generally say,

00:54:42.020 don't push someone by just showing that the absolute risks in terms of number of cases,

00:54:48.340 cause versus prevented, might be favorable for them because their emotional well-being is a very

00:54:55.340 important part of the equation. However, I do agree that we need to emphasize absolute risks,

00:55:02.540 that absolute risks are low, and that there are better candidates and there are worse candidates.

00:55:09.180 It's amazing how the pendulum has swung. So in the 1980s, 1990s, the perception was that hormone therapy

00:55:18.140 was good for all women. Women were being routinely started on hormone therapy. Then after the WHI in 1.00

00:55:24.880 the early 2000s, the pendulum was in the opposite direction, that hormone therapy is bad for all

00:55:32.320 women. And now I think it is coming, the pendulum is coming to rest in a more appropriate place,

00:55:37.640 that hormone therapy is good for some, but not all women. And the best candidates are women in early 1.00

00:55:44.380 menopause who have moderate to severe or bothersome hot flashes and night sweats and are in generally

00:55:51.180 good health. Those are women who will derive quality of life benefits and have very minimal 1.00

00:55:57.660 absolute risk from hormone therapy. One of the things I've thought a lot about is how many women, 1.00

00:56:03.900 if you go back to 2002 and the data are published, but the media has a different take on it. The media

00:56:11.400 has the take on it, which is kind of how we're discussing it now. Not very exciting. So maybe

00:56:17.220 that's why that's not the take the media had, but I play the what if game, which is how do things shake

00:56:22.360 out? So instead of the pendulum going so far to the other side of women can't and shouldn't under 1.00

00:56:29.400 any circumstance have HRT because they're going to get breast cancer, it turns into, no, there will be 1.00

00:56:34.880 additional cases of breast cancer. I've done sort of a back of the envelope calculation based on

00:56:38.920 the assumption that somewhere between four and five million women probably missed out on HRT in the

00:56:44.200 last 23 years, 22, 23 years as a result of the study, which saved, mean it reduced about 4,500

00:56:52.520 cases of breast cancer. There's benefit to that. Didn't really reduce mortality from breast cancer, 0.92

00:56:58.720 but here's what's interesting. Even though it's not a primary outcome, HRT reduced the incidence of hip

00:57:05.020 fracture by about one and a half percent in absolute terms. That's remarkable.

00:57:09.560 Yes, there was the benefit for hip fracture and there were benefits for-

00:57:15.860 In other cancers as well.

00:57:16.980 Yes, there were benefits for endometrial cancer with estrogen plus progestin. But here's the problem

00:57:22.120 with the whole argument for using it for bone health. Women in their forties and fifties have a low 1.00

00:57:29.640 risk of osteoporotic hip fracture. They may have a hip fracture from a traumatic incident, but it's not

00:57:37.040 likely to be related to osteoporosis. So it's when they get into their late 60s, 70s, 80s, that they're

00:57:44.580 more likely to have the osteoporotic fractures. And if we were to treat women from early menopause into

00:57:53.400 their 70s, 80s, that would be very long duration of hormone therapy use, which would lead, especially

00:57:59.380 combination hormone therapy, would lead to an increased risk of breast cancer. And once women 1.00

00:58:03.920 go off of hormone therapy, bone loss is very rapid. So all of the benefits to the bones in terms of

00:58:12.740 preserving bone mineral density, that dissipates very quickly within a matter of a few years after

00:58:20.400 stopping hormone therapy. So that if a woman is taking hormones in her forties and fifties,

00:58:27.660 and then she stops, let's say at age 60, then by the time she gets to the age where her risk of hip

00:58:34.920 fracture is very substantial in her seventies and eighties, she's really not going to have a persistent

00:58:40.780 sustained benefit from the hormone therapy. And we looked within the group at high risk of

00:58:47.500 osteoporotic fracture, whether they had a favorable, overall favorable outcome in terms of the global

00:58:54.360 index combination of all of these outcomes with hormone therapy. And overall, there really was no

00:59:00.700 group of women, irrespective of their risk of osteoporotic fracture, who had a clearly beneficial

00:59:09.140 ratio from benefit risk ratio from estrogen combined with progestin when used for chronic disease

00:59:17.440 prevention. Again, this does not mean that it isn't a very effective treatment for hot flashes and

00:59:24.000 night sweats, and that women in early menopause would have a favorable benefit risk profile taking

00:59:31.600 into account the quality of life benefits. We haven't talked about age differences. I think that

00:59:38.940 overall, it's so important for women to understand that the absolute risks of these hormones are much

00:59:46.600 lower in early menopause than in later menopause. And in many ways, timing is everything when it comes to

00:59:55.000 hormone therapy, because not only are the absolute risks, the risk of having adverse events on hormone

01:00:02.360 therapy lower in the younger women, women in early menopause, but also we saw some signals that especially

01:00:09.860 with estrogen alone, the women in their 50s, the youngest women in the study were 50 to 59, those women

01:00:18.340 tended to do quite well in terms of heart disease and heart attack rate compared to the women on 1.00

01:00:25.060 placebo. There was a signal there for benefit, also a signal there for favorable outcomes in terms of all

01:00:32.020 cause mortality. So overall, there were favorable signals with estrogen alone in the younger women,

01:00:39.700 and also lower risks of adverse events, suggesting that certainly for estrogen alone in a woman who's

01:00:48.060 had a hysterectomy, if she is symptomatic with hot flashes, night sweats, the benefits of treatment are

01:00:54.900 likely to outweigh the risk. And even for combination estrogen plus progestin, despite what may be a small

01:01:02.400 increase in risk of breast cancer, the benefits are likely to outweigh the risk, even combination

01:01:09.240 therapy when used for the purpose of treating bothersome, disturbing, hot flashes, night sweats, disrupted

01:01:17.580 sleep, and impaired quality of life. And women should not shy away from the use of hormone therapy and should 0.99

01:01:24.160 discuss the option, weigh the benefits and risks carefully with their healthcare provider, with

01:01:29.540 their clinician, and see if it's the right decision for them. And if they have trouble finding a

01:01:35.080 clinician, they can go to the North American Menopause Society has a website, menopause.org,

01:01:41.960 and they can find the tab for find a healthcare professional, a certified healthcare professional

01:01:48.500 with training in menopause. And they can put in their zip code and find the clinicians in

01:01:54.100 their area who have this expert training. I'm an endocrinologist, and I know that many

01:02:00.480 endocrinologists have the training to talk about hormone therapy and discuss hormone therapy with

01:02:07.940 patients. But many clinicians have had limited training in, you know, use of hormone therapy,

01:02:15.300 and it may be helpful for women to seek out a clinician who has had some additional training

01:02:23.060 in menopause management and hormone therapy pros and cons. And for women who are not good candidates,

01:02:30.200 the good news is that there are non-hormonal options as well. They're not quite as effective

01:02:37.460 as the menopausal hormone therapy for treating hot flashes and night sweats. But some of the antidepressant

01:02:45.020 medications, SSRIs, SNRIs, some of the gabapentin, pentanoid medications, these medications have been found

01:02:54.500 to be quite effective, 40-50% reduction in hot flashes and night sweats with these medications. And there's a new

01:03:04.640 medication that may be approved by the FDA fairly soon that works entirely in the brain in terms of making women 0.99

01:03:12.880 less sensitive to the changes in temperature and has a very beneficial effect in terms of preventing hot

01:03:21.980 flashes and night sweats.

01:03:23.660 I guess I still kind of come back to something that the math doesn't quite add up, right, which is

01:03:28.540 if a woman stays on estradiol for the duration of her life, say from age 50 to 80, we accept that there's 0.97

01:03:37.120 going to be an increase in the incidence of breast cancer, though it doesn't seem to translate to a

01:03:41.960 difference in lifespan. But on the flip side, we are reducing her risk of fracture during a very

01:03:48.540 dangerous window. So the incidence of fracture saved is about one and a half percent, and the mortality

01:03:54.520 of that once she reaches 65, the one-year mortality, depending on the series, is 15 to 30 percent. So even

01:04:01.960 if we just want to do this on the basis of mortality, apples to apples, it doesn't even appear close,

01:04:07.960 does it?

01:04:08.360 Well, let's look at the actual data, because we published in JAMA 2017 the mortality results,

01:04:16.960 the all-cause mortality results by age and time since menopause. And we did see that the younger

01:04:24.240 women, the women who were in their 50s taking either estrogen alone or estrogen plus progestin,

01:04:30.460 had signals for about 30 percent lower mortality, though it was not statistically significant in either

01:04:37.440 trial. Yet the women who were older in their 70s, 70 to 79 randomized to estrogen alone, had a hazard

01:04:49.880 rate 22 percent higher risk of all-cause mortality. It was not quite statistically significant. So it was

01:04:58.380 right at the border of being statistically significant. And with estrogen plus progestin,

01:05:05.540 it was actually quite neutral, almost completely null. So estrogen alone seemed to be just a tad worse

01:05:14.340 in the women, 70 to 79, though it was a tad better among the women, 50 to 59. So again, timing is

01:05:22.140 everything when it comes to the all-cause mortality results. There are favorable signals for the hormone

01:05:30.300 therapy. The women earlier in menopause, age 50 to 59 in the study, quite neutral results in age 60 to

01:05:39.240 69. But in age 70 to 79, there's a bit of a signal that estrogen alone may be linked to a small increase

01:05:49.500 in risk of mortality. That is a little different than what I was asking. I don't think the study can

01:05:54.180 answer the question I was asking, which is really more of conjecture that if a woman is started at

01:06:00.680 the appropriate time, which I think we all agree is during the transition from perimenopause to

01:06:05.480 menopause. And I'm really asking this question through the lens primarily of hip fracture, which I think

01:06:09.740 is just such an underappreciated cause of mortality in both men and women over the age of 65. But estrogen

01:06:16.900 is hands down the most important hormone as it pertains to signal transduction from the strain gauge

01:06:23.760 within the muscle to the osteoblast and osteoclasts. And so what I'm really asking is, it's more of a

01:06:29.080 thought experiment. If we put women on HRT at 50, and they stay on HRT for 30 years, by my calculation, 0.99

01:06:37.220 it's a 15x reduction in mortality. Because even if you accept a slightly higher incidence of breast

01:06:44.560 cancer, it's being more than dwarfed by the reduction in the hip fractures that they will be sustaining

01:06:50.200 15 years after beginning. Again, we don't know the answer, and we'll never know the answer, because I don't

01:06:55.400 think anyone's going to do that study. But the magnitude of that difference is so great. And that's where I just

01:07:00.920 think the discussion is very confusing for people, because they only see one thing, which is breast cancer. And

01:07:06.940 they don't see it through the nuanced lens that we're talking about it, which is why I'm glad we're talking

01:07:10.820 about it. I think this is the discussion people need to understand.

01:07:13.060 I agree that there may be an excessive focus on one isolated outcome, such as breast cancer,

01:07:21.520 when it's really the overall health of the woman. And of course, all-cause mortality is an integration 0.85

01:07:27.500 of all of these life-threatening health conditions. And it was quite neutral with hormone therapy.

01:07:34.280 Overall, the results were what we call null for all-cause mortality. There was no increase in risk

01:07:41.180 or decrease in risk in the overall study population. And in the younger women, there was a signal for some

01:07:48.240 reduction in risk. But a randomized trial has never been done that would look at 30 years of hormone

01:07:58.000 therapy treatment, starting in early menopause and continuing into mid to later menopause. And the

01:08:07.060 reality is we don't know whether the benefits would outweigh the risks. I know that some women will make 1.00

01:08:14.340 the choice to continue to take hormone therapy into middle to later menopause because they started in

01:08:23.020 early menopause. They did well on the hormones. Their quality of life is very good. When they try to stop or

01:08:30.120 reduce the dose, they feel hot flashes coming back and they end up taking these hormones well into

01:08:35.940 their 60s, 70s, even longer. And in the observational studies, this looks like it has a relatively

01:08:44.820 favorable benefit-risk ratio. However, keep in mind, this is a very select group of women who are

01:08:51.760 choosing to stay on hormone therapy long-term because they're doing extremely well with it and

01:08:58.640 they've tolerated it well and they haven't developed any of these interim events such as a heart attack,

01:09:03.360 a stroke, a blood clot in the legs or lungs, or breast cancer or other estrogen-sensitive cancers.

01:09:09.860 That's right. So we're selecting for healthier women. 1.00

01:09:12.220 It's a highly selective group of women. And I think that overall, we just don't know what a 1.00

01:09:18.600 randomized trial would show with long-term hormone therapy. But I think there would be concern that

01:09:23.460 the risk would outweigh the benefits because keep in mind, the WHI showed an increased risk of stroke

01:09:29.320 with both estrogen alone and estrogen plus progestin. And there was cognitive decline and

01:09:35.720 increased risk of cognitive decline among the women 65 and older, although that was not the case

01:09:40.480 in the younger women. But we just don't know how the pattern of benefits and risks would change

01:09:47.760 with longer duration once you get into 15, 20, 25 years of treatment.

01:09:54.900 Yeah. It's pretty clear to me we'll never get another bite at the apple as far as

01:09:59.140 going for that duration. But do you think we'll ever get another bite at the apple to use

01:10:04.280 kind of better practices? I mean, CEE and MPA have largely fallen out of favor. Very few women. 1.00

01:10:10.420 I've never actually seen a woman take those. I'm sure some do. But most women these days are using 1.00

01:10:14.180 the Vevel dot or one of the topical FDA patches, which is just pure estradiol. As you alluded to

01:10:21.040 earlier, most women these days are taking micronized progesterone, which is a bioidentical progesterone

01:10:25.440 or foregoing oral progesterone altogether and using a progesterone-coded IUD, which seems to offer

01:10:30.960 the same degree of local endometrial protection without any of the systemic effects for some women

01:10:36.060 who can't tolerate that. So in some ways, the entire cluster of insights from the WHI are less relevant

01:10:44.060 today, given that the drugs that were tested aren't the drugs that are in mass adoption.

01:10:48.700 Do you think it's likely we will see another short-term study, and I call short-term like kind of a five-year

01:10:55.740 study, that tests the same question using the current formulations and potentially does so in a manner

01:11:03.720 that is in line with more current use cases of the drug? Or do you think that that's too much of an undertaking

01:11:09.860 and something that we'll never really get another RCT to address?

01:11:13.400 I think that it is a good trend to move toward the transdermal, estradiol, and the micronized

01:11:22.200 progesterone, these FDA-approved bioidentical formulations of hormone therapy as opposed to the

01:11:31.460 oral conjugated estrogen, medroxyprogesterone acetate. I think that that is a good trend. We have to keep in

01:11:38.940 mind that we really don't know the long-term benefits and risks, but based on other, you know, in terms of

01:11:47.260 clinical outcomes, such as heart attacks, strokes, different forms of cancer, and all-cause mortality. But I think it

01:11:54.820 makes sense based on how these different formulations affect clotting, affect biomarkers for clotting, and

01:12:03.600 cardiometabolic health, blood pressure, all of those parameters. I think that there's reason to use the

01:12:11.700 transdermal and the micronized progesterone preferentially over the older formulations. I think

01:12:18.240 we need more randomized trials of these formulations that are now in more common use, especially looking at

01:12:26.780 breast density, mammographic breast density, to see if there is less suggestion of a future risk in

01:12:35.880 breast cancer based on the change in mammographic breast density, whether or not there is going to be

01:12:42.840 another large-scale trial of the magnitude of the Women's Health Initiative, which would be if you take

01:12:49.260 women in early menopause in order to have similar numbers of health outcomes and similar robust power 0.92

01:12:56.020 to look at health outcomes, it would have to be at least 40, 50,000 women being randomized in their

01:13:04.840 50s and followed for seven, eight years. I think it's really unclear whether such a trial could be

01:13:13.300 mounted, would be extremely expensive. And also, because of the rapid changes in formulations over time,

01:13:21.660 there would be the similar risk that the results might become obsolete by the time the results were

01:13:28.320 available, as you're suggesting is the case with the conjugated estrogen and MPA. But let's keep in mind

01:13:36.840 that even though the WHI was testing the most common formulations at that time in the early 1990s,

01:13:45.820 it did put an end to a practice of prescribing hormone therapy among women in later menopause for

01:13:55.440 the purpose of preventing strokes and heart attacks and cognitive decline, which actually were found in

01:14:02.740 the WHI to be adversely affected by hormone therapy. And women were not getting the benefits for quality of 0.99

01:14:11.220 life that many of them didn't have. Once they were in their late 60s, 70s, they were not having any hot

01:14:18.140 flashes, night sweats, disrupted sleep related to these symptoms. So they weren't getting the benefits

01:14:23.320 and they were getting these adverse outcomes. So it's important to acknowledge that the WHI did put an end to

01:14:32.000 what was an unfavorable practice. The problem was it was over extrapolated. The results were extrapolated

01:14:39.040 to women in early menopause taking hormone therapy for the hot flashes, night sweats. It's really

01:14:44.680 important for women to understand that the WHI results are really not intended to discourage a 1.00

01:14:53.500 woman who's having disrupted sleep, very severe, significant symptoms to discourage her from seeking help

01:15:01.460 and seeking hormone therapy as one of her options for treatment. And in her specific case in early

01:15:08.960 menopause and generally good health, the benefits are likely to outweigh the risk. That's the really

01:15:15.000 key message that women in early menopause should take these symptoms seriously, make sure they find a

01:15:22.520 clinician who will take these symptoms seriously and discuss their options, their treatment options,

01:15:30.100 review with them the pros and cons. And if they are not finding that clinician, they need to go outside

01:15:36.100 their current system and find a clinician who can help them to do that because there are many

01:15:43.120 clinicians out there who are knowledgeable in menopause management, knowledgeable in hormone therapy

01:15:48.480 decision-making, who can help them to make the best choice for themselves.

01:15:54.640 Joanne, what was the reference you gave for women to go and look for finding a provider near them

01:15:59.200 where they enter their zip code? What was the name of that again?

01:16:00.820 Menopause.org website of the North American Menopause Society. And they can go to the Find

01:16:08.980 a Certified Menopause Practitioner tab and it will give them an opportunity to put in their zip code

01:16:16.640 and then it will tell them which clinicians within five miles, 10 miles, have the expertise in menopause

01:16:25.620 management and hormone therapy and other treatments.

01:16:28.260 Well, I could continue talking with you for hours and extract so much more insight. There's so many other

01:16:33.720 topics. We could talk about vitamin D for another few hours.

01:16:36.800 That could be another. That could be another.

01:16:39.180 There's a whole other discussion. Yeah. I'll conclude with two things. I guess one is just I still remain

01:16:43.840 somewhat sad because I think there's a lost generation of women. I think there's 20 years of women who

01:16:48.380 entered menopause who were denied HRT due to the ignorance of their physicians and the irresponsibility

01:16:55.200 of the media. And I look at women like my mother and my mother-in-law who were entering menopause 0.99

01:16:59.820 just as the WHI was coming to its conclusion and who suffered unnecessarily. And I don't know how many

01:17:07.220 millions of women suffered unnecessarily. I think you're right. I think the pendulum is swinging

01:17:11.920 and I'd like to believe that there aren't women that are suffering that way today.

01:17:15.540 The other point I'd make is I really have always respected you and continue to do so because I feel

01:17:20.720 like you're one of the few people who was such an important part of the WHI who has been able to

01:17:26.040 look back at that and acknowledge its limitations. And I think your thinking seems to at least me

01:17:31.760 through your writing to have evolved over time. I don't think that's a property that is necessarily

01:17:36.880 inherent to everyone at your position. So I really applaud you for that. And I think that you're doing

01:17:41.820 more good today by speaking out on the limitations of the WHI than you probably even did by taking part

01:17:47.760 of the WHI. So thank you for that. And thank you for joining me today.

01:17:51.640 Well, thank you, Peter. I've enjoyed talking with you and I hope that these messages are helpful

01:17:56.740 to your audience. Thank you for listening to this week's episode of The Drive. If you're interested in

01:18:02.680 diving deeper into any topics we discuss, we've created a membership program that allows us to

01:18:07.420 bring you more in-depth exclusive content without relying on paid ads. It's our goal to ensure members

01:18:13.040 get back much more than the price of the subscription. To that end, membership benefits

01:18:17.880 include a bunch of things. One, totally kick-ass comprehensive podcast show notes that detail

01:18:23.220 every topic, paper, person, thing we discuss on each episode. The word on the street is nobody's

01:18:28.780 show notes rival these. Monthly AMA episodes or Ask Me Anything episodes, hearing these episodes

01:18:34.480 completely. Access to our private podcast feed that allows you to hear everything without having to

01:18:40.540 listen to spiels like this. The Qualies, which are a super short podcast that we release every

01:18:45.920 Tuesday through Friday, highlighting the best questions, topics, and tactics discussed on

01:18:50.060 previous episodes of The Drive. This is a great way to catch up on previous episodes without having to

01:18:55.540 go back and necessarily listen to everyone. Steep discounts on products that I believe in, but for

01:19:01.300 which I'm not getting paid to endorse. And a whole bunch of other benefits that we continue to trickle in

01:19:06.280 as time goes on. If you want to learn more and access these member-only benefits, you can head

01:19:10.600 over to peteratiamd.com forward slash subscribe. You can find me on Twitter, Instagram, and Facebook,

01:19:17.940 all with the ID peteratiamd. You can also leave us a review on Apple Podcasts or whatever podcast

01:19:24.440 player you listen on. This podcast is for general informational purposes only and does not constitute

01:19:30.000 the practice of medicine, nursing, or other professional healthcare services, including the giving of

01:19:35.500 medical advice. No doctor-patient relationship is formed. The use of this information and the

01:19:41.200 materials linked to this podcast is at the user's own risk. The content on this podcast is not intended

01:19:47.380 to be a substitute for professional medical advice, diagnosis, or treatment. Users should not disregard

01:19:53.860 or delay in obtaining medical advice from any medical condition they have, and they should seek the

01:20:00.020 assistance of their healthcare professionals for any such conditions. Finally, I take conflict

01:20:05.480 of interest very seriously. For all of my disclosures and the companies I invest in or advise,

01:20:11.240 please visit peteratiamd.com forward slash about where I keep an up-to-date and active list of such companies.

01:20:35.480 Thank you.

The Peter Attia Drive - May 08, 2023

#253 ‒ Hormone replacement therapy and the Women's Health Initiative： re-examining the results, the link to breast cancer, and weighing the risk vs reward of HRT ｜ JoAnn Manson, M.D.

Episode Stats

Summary

Transcript