Menopause Mondays: Does HRT Cause Cancer?
Last week, we interviewed Dr. Josh Trutt of PhysioAge Medical Group about bio-identical hormones in “Menopause Mondays: Understanding Hormone Replacement Therapy”. We learned that there are different kinds of HRT, and aside from beneficial effects on menopause symptoms, estrogen has been shown to reduce your risk of heart disease, osteoporosis, and Alzheimer’s. This week, we’ve got part two of our interview with Dr. Trutt, asking him the BIG question. HRT and cancer: Are they as closely linked as we’ve been told? (Note: this is a long blog that contains EXCELLENT information – if you don’t have time at work, bookmark it to read through carefully later!)
Q: Let’s get right down to brass tacks. Does HRT cause cancer?
A: It depends what kind of HRT—and what kind of cancer—you are referring to. This is a serious topic that requires more than a blog posting to respond to, as there are dozens of studies to discuss and your personal medical history is relevant. For those of you that want as much detail as possible, please go to http://physioage.com/blog/.
But to boil it down as best I can, here is what the data shows:
Regarding Endometrial (Uterine) cancer:
If you use estrogen without progesterone, it definitely does increase your risk of endometrial cancer. However, progesterone stops this process in its tracks. In fact, all types of progestin, whether bio-identical progesterone or chemically altered patented progestins (such as Provera), clearly decrease the risk of endometrial cancer. This is not controversial –in fact, this is exactly why women started being given progestins as part of HRT in the 1980s. (Before that, they were given estrogen alone.) Although the synthetic progestins do a great job in preventing endometrial cancer, many studies have shown them to increase the risk of breast cancer. Therefore, bio-identical progesterone should be used instead. Bio-identical progesterone prevents estrogen from causing endometrial cancer, and at the same time does not increase (and may decrease) the risk of breast cancer.
Regarding breast cancer:
This a very hot topic, filled with convoluted reporting and misunderstood data. Let’s start with the following:
1. If you use bio-identical estrogen with bio-identical progesterone, studies show no increase in your risk of breast cancer, and a decrease in your risk of endometrial cancer.
2. Many synthetic, not bio-identical progestins, such as Provera (medroxyprogesterone), norethisterone, medrogestone, cyproterone and others, do increase your risk of breast cancer slightly. (Your risk of endometrial cancer, however, decreases, even with synthetic progestins). In addition, the synthetic progestins negate the beneficial effect of estrogen in preventing heart disease. For these reasons, many specialists only recommend bio-identical progesterone.
Therefore, if your specialist says “hormone therapy increases the risk of breast cancer,” your first question should be, “What type of hormones are you referring to?”
Q: I see. You’re saying that the risks vary, but bio-identical estrogen with bio-identical progesterone are OK with respect to breast cancer and endometrial cancer. Can women who have had a hysterectomy take estrogen without progesterone?
A: The data from the WHI and other trials do seem to support that that is safe with regard to breast cancer, in women that have been followed for up to about 10 years. However, the preponderance of data suggests that bio-identical progesterone is certainly neutral and may even be protective against breast cancer, and has other benefits with regard to HDL cholesterol and stroke prevention. This HDL benefit is so pronounced that Elizabeth Connor, one of the cardiologists involved in the PEPI trial, said, “If I were treating a woman primarily because she was worried about heart disease, I would probably see if she wanted to take micronized (natural) progesterone. I was quite impressed with the better effect.” In addition, it seems to have a beneficial effect on blood vessels. For example, in women with heart disease, exercise time to myocardial ischemia improved when they were given estrogen, and improved even more with combination estrogen/progesterone therapy (92 seconds [35 to 149 [p = 0.001)).
Natural Progesterone, but not Medroxyprogesterone Acetate (Provera®), Enhances the Beneficial Effect of Estrogen on Exercise-Induced Myocardial Ischemia in Postmenopausal Women. J Am Coll Cardiol 2000 December; 36(7)
So, the decision of whether to include natural progesterone in the regimen of a woman who has had a hysterectomy may depend on multiple factors in her medical history.
Q: So why does HRT get such a bad rap in the media for causing breast cancer?
A: Good question. For one thing, specialists virtually never take the time to notice which studies were done on bio-identicals and which were done on synthetics. The news media certainly does not make that distinction. In addition, many of us have it in our heads that hormones “should” cause cancer, and therefore interpret the data with significant bias. (In fact one of the principal investigators in the WHI trial has stated that he felt his duty was “to get women off of hormones.” It is certainly concerning when such an important public health issue cannot be examined objectively.)
The theory that estrogen causes breast cancer stems from three observations:
- Women produce much more estrogen than men
- Breast cancer is 100 times more prevalent in women
- The earlier a woman starts getting her period and the later she reaches menopause (i.e., the more years she is exposed to estrogen), the greater her risk of breast cancer.
Because of these straightforward observations, researchers wondered if perhaps estrogen was the cause of breast cancer in women.
The problem is, this theory has been contradicted in various ways:
- Women taking estrogen after menopause should have a higher risk of breast cancer! After all, they are exposed to it for many more years. But they do not. Even the WHI trial found that women who took estrogen alone, without Provera, had no increased cancer risk; in fact they had a slight decreased risk. It was the Provera that caused the increase in breast cancer.
- If estrogen causes breast cancer, then women who don’t take HRT should have a progressively LOWER risk of cancer after menopause, since they are not making any estrogen! But in fact we find the opposite: A woman’s risk of breast cancer increases as she grows older, despite the loss of estrogen.
- Estrogens do not directly make breast cells cancerous. However, estrogen can induce cell proliferation. So, the WHI modified their original hypothesis to say: since carcinogens are all around us, the higher the rate of breast cell proliferation (which is caused by estrogen), the more likely a mutation might occur during one of the cell divisions, and thus lead to cancer. The problem with this theory is that the endometrium (the lining of the uterus) is much more sensitive to cell proliferation caused by estrogen than the breast; in fact estrogen’s job in the uterus is to thicken the endometrial lining each month! In fact the uterus is so sensitive to this that women who take estrogen alone, without progestins, DO have a greatly increased risk of uterine cancer (progestins/progesterone prevents this)– and yet they DON’T have an increased risk of breast cancer—not even in the WHI trial. If early onset of menses or late menopause caused breast cancer from increased estrogen exposure, we would certainly ALSO see increased endometrial cancer in these women. But we don’t.
- Lastly, birth control pills, which used to contain far more estrogen than HRT does, should therefore increase the risk of breast cancer. Although there is some controversy on this, most published studies do not demonstrate increased risk. (I would point out that birth control pills all use synthetic progestins, and they are the likely cause of any small increase in breast cancer.)
- There is more to discuss on this topic, including how Tamoxifen works; suffice it to say that Tamoxifen clearly has mechanisms of preventing breast cancer besides blocking estrogen.
When all of this is taken together, it does not appear that estrogen is a direct cause of breast cancer. Therefore, it is not surprising that the studies on Premarin (without progestins) in menopause showed no increased risk of cancer! The basic science, the smaller controlled trials, and the epidemiology are all aligned.
It’s important to note that we still have more to learn: there is evidence to suggest that what matters most is the way a woman metabolizes estrogen: some metabolites are more carcinogenic than others, and not all women metabolize estrogen the same way. There is also evidence to suggest that the route of administration (oral vs transdermal) may matter. And we haven’t even mentioned the role of testosterone in protecting against breast cancer, which is worthy of a separate post (perhaps we will do a “Part 3”). Just as it is inappropriate to make healthcare decisions based on soundbites from Katie Couric (“hormones cause cancer!”), so too a single blog posting cannot provide all of the relevant information. I have detailed conversations with my patients about the nuances of these studies, and you should discuss them with your specialist as well.
(Adapted from oncologist Avrum Bluming’s 2009 paper in The Cancer Journal entitled “Hormone Replacement Therapy: Real Concerns and False Alarms.”)
Q: Interesting! But what about the Women’s Health Initiative?
A: It’s time to strip away some of the mystique from the Women’s Health Initiative trial. It’s gotten an order of magnitude more attention than it deserves… but to explain why, we will have to take a short trip into the sometimes uncomfortable world of statistics.
What does the Confidence Interval (CI) tell us?
When a study like the WHI says, “Women who take PremPro are 1.24 times more likely to get breast cancer than women who don’t take PremPro,” we want to know if that Relative Risk of 1.24 is “real,” or just the result of random chance in their study. The Confidence Interval helps answer that by providing a range for the Relative Risk: statistics tells us that if we allow for variation due to random chance, the Relative Risk will still fall somewhere within that Confidence Interval 95% of the time. (Sometimes the 99% CI is used.)
If the range of the Confidence Interval includes the number 1.0, the result is not considered statistically significant.
Let’s say we are studying whether smoking causes skin cancer. We do a trial and compare 10,000 smokers with 10,000 non-smokers and we find that the smokers had a Relative Risk for skin cancer of 0.8. That means, the smokers were only 80% as likely to get skin cancer as non-smokers. We think, “Wow! Smoking protects against skin cancer!” But then we look at the Confidence Interval: it’s 0.7 to 1.35. If the Confidence Interval in our study falls both above and below 1.0 it means: if we repeat this study multiple times, sometimes skin cancer will occur more often in smokers– Relative Risk will be greater than 1.0– but sometimes the smokers will get skin cancer less often than the non-smokers (Relative Risk less than 1.0), simply due to random chance. In that case, skin cancer is not significantly correlated with smoking in our study.
Now that we’ve cleared up what the Confidence Interval is, there is an important point to make: Let’s say the Relative Risk in our study on smoking and skin cancer was 0.8, and the Confidence Interval in was 0.4 to 0.9 – statistically significant!
(For help with this I leaned heavily on two fantastic papers: “Epidemiology Faces Its Limits,” by one of my favorite science writers, Gary Taubes; and a beautiful deconstruction of the WHI trial by oncologist Avrum Bluming, “Hormone replacement therapy: Real concerns and false alarms.” See references below.)
Q: Does that mean that smoking lowers the risk of skin cancer?
A: No. It simply means that the results of our study were not due to random chance. What if I told you that all the smokers in this study were African-Americans (a population with very low rates of skin cancer), and the non-smokers were all Caucasian (whites have the highest rate of skin cancer). You’d say, well of course the smokers got less skin cancer! The results are still statistically significant (if you repeated that study 100 times, the African Americans would always get less skin cancer) but that does not mean that smoking lowers the rate of skin cancer. That study is statistically significant, but suffers from poor methodology. So that is point number two: Even if a study reaches statistical significance, that does not mean the correlation definitely exists! You have to look closely at the study itself.
We are almost ready to look at the WHI results. The last point I want to make is that large epidemiologic studies like the WHI or the Million Women Study are more susceptible to confounding factors and observer bias, than are small double-blind prospective trials. Some examples of this:
- Large epidemiologic trials looking at whether alcohol causes cancer have been confounded by the fact that many people who drink a lot also smoke.
- Large cohort studies often use questionnaires, with no way of validating whether the answers filled in are correct.
- “Recall Bias”: women who are diagnosed with breast cancer may take the time to remember in much greater detail what years they used birth control or HRT, whereas healthy women may not make as much effort to answer with great accuracy. Again, there is often no way to verify their answers.
Because of these issues, in large epidemiologic studies like the WHI, unless the lower limit of the Confidence Interval is at least 3, it generally should not be considered a strong, reliable risk factor.
Even if the results are replicated in multiple trials, a large epidemiologic trial with a RR below 2 should not make the headlines the way the WHI has.
So with that in mind, let’s look at the table below.
(Adapted from Bluming AZ. Tavris C. Hormone Replacement Therapy: Real Concerns and False Alarms. Cancer J. 2009;15:93–104.)
In this table, the association of breast cancer with Conjugated Equine Estrogen (Premarin) in the Women’s Health Initiative has a CI from 0.59 to 1.01, which crosses 1.0 and is therefore (barely) not a significant association. (Premarin certainly did not INCREASE the risk of cancer, as the CI is almost entirely below 1.0, but we can’t say it significantly decreased the risk either; we say instead that there was no significant association.) Premarin/Progestin (PremPro), French Fries and Grapefruit did barely reach statistical significance in those studies. As you can see, if you are worried about taking HRT, you should also certainly worry about eating grapefruit, using an electric blanket, and whatever you do, don’t become a flight attendant.
At the bottom of the chart are two risk factors that are “real”. As you can see, the link between tobacco smoke and lung cancer is incontrovertible. And when a link is “true”, it only gets stronger with subsequent studies. With cancer and HRT you seem to get a new answer every other year (as researchers strive desperately to find a significant link), whereas with smoking and lung cancer, the link only became stronger with each subsequent study.
At the bottom I put the link between estrogen and endometrial (not breast) cancer. This is a true link—estrogen taken without progesterone will certainly increase the risk of endometrial cancer.
**Progesterone—indeed, even synthetic progestins— unquestionably prevent estrogen from causing endometrial cancer. This is not a controversial issue. This is why progestins started being given with estrogen in the 1980s.
I hope this gives at least an idea of how overblown the news coverage of the WHI trial has been.
Q: Aha — it’s time to step away from the WHI trial. What about women with a family history of breast cancer? Should they avoid HRT?
A: Not at all. Some very good studies have looked at this issue. In 1997 a prospective study was done in Iowa looking at 42,000 women with a family history of breast cancer. Here is what they found:
Among women with a family history of breast cancer, those who currently used HRT and had done so for at least 5 years developed breast cancer at an age-adjusted annual rate that was not statistically significantly higher than the rate in women who had never used HRT.
Among women with a family history, those who used HRT had a significantly lower risk for total mortality than did women who had never used HRT, including total cancer-related mortality. The age-adjusted annual mortality rate for women using HRT for at least 5 years was 46 deaths per 10,000 person-years ; this is roughly half the rate seen in women who had never used HRT (80 deaths per 10,000 person-years).
CONCLUSIONS: These data suggest that HRT use in women with a family history of breast cancer is NOT associated with increased incidence of breast cancer, and IS associated with a significantly reduced total mortality rate (!!).
The role of hormone replacement therapy in the risk for breast cancer and total mortality in women with a family history of breast cancer. Ann Intern Med. 1997 Dec 1;127(11):973-80.
Those results are probably exactly opposite to what you have heard on the news, from your friends and, unfortunately, from your specialist. Here is another study that also looked specifically at HRT in women with a strong family history of breast cancer:
This study looked at 3300 women who had benign breast biopsies at the start of the study, and followed them for quite a long time– median of 17 years. Here again, the relative risk of breast cancer was lower in the women who used estrogen replacement therapy. But that is not the most interesting part: The table below shows the interaction between estrogen usage and a history of breast cancer in a first degree relative (mother, sister, or daughter):
The women with a strong family history of breast cancer were the ones who benefited the most from estrogen! These results are fairly remarkable. Dozens of studies have looked for a possible relationship between HRT and breast cancer; the majority are statistically non-significant, and very few have as clear a statistical significance as this one.
Exogenous estrogens reduced breast cancer risk in both women with and without a family history of breast cancer. However, the protective effect of estrogens was more pronounced among women with a family history of breast cancer.
Influence of exogenous estrogens, proliferative breast disease, and other variables on breast cancer risk. Cancer. 1989 Mar 1;63(5):948-57.
Q: Estrogen is MORE protective in women with a family history of breast cancer? But why would that be? It seems counterintuitive…
A: The relationship between HRT and breast cancer is incompletely understood; for example, the available evidence suggests that BRCA1 mutation carriers would be better off using HRT:
Journal of the National Cancer Institute, 2008
We conducted a matched case–control study of 472 postmenopausal women with a BRCA1 mutation to examine whether or not the use of HT is associated with subsequent risk of breast cancer.
Among postmenopausal women with a BRCA1 mutation, Hormone Therapy use was not associated with increased risk of breast cancer; indeed, in this population, it was associated with a decreased risk.
Hormone therapy and the risk of breast cancer in BRCA1 mutation carriers. J Natl Cancer Inst. 2008 Oct 1;100(19):1361-7
Granted, this was a small study. But the study was statistically significant (unlike much of the WHI results), and there is a plausible mechanism by which HRT might protect against breast cancer risk in BRCA1 mutation carriers: estrogen increases expression of BRCA1. In cells that retain one normal BRCA1 allele, estrogen could serve to increase the level of the wild-type protein, and thereby promote genetic stability.
Here is another BRCA1 study:
The authors identified 462 women with BRCA1/BRCA2 mutations. 155 of the women decided to have their ovaries removed prophylactically, which has been shown to greatly decrease breast cancer risk. However, these women all decided to use HRT afterwards to avoid premature menopause symptoms. The women were followed to see if breast cancer developed. The authors found that after having their ovaries removed, the women were protected from breast cancer– and that using HRT did not change that.
CONCLUSION: Short-term HRT use does not negate the protective effect of (ovary removal) on subsequent breast cancer risk in BRCA1/2 mutation carriers.
Rebbeck et al. Effect of short term HRT on breast cancer risk reduction after bilateral prophylactic oophorectomy in BRCA1 and BRCA2 mutation carriers: the PROSE Study Group. J Clin Oncol. 2005;23:7804-7810
Some have suggested that the reason that BRCA1 mutation carriers are protected, is that their tumors are usually estrogen receptor-negative. That sounds reasonable, except it implies that women on HRT get ER-positive tumors more often than women not on HRT; this turns out not to be true. It also suggests that if an ER-positive cancer develops in a women on HRT, she would do very poorly– after all, ER-positive tumors are treated with estrogen blockers! Yet the truth is that most breast cancers are ER-positive, and despite that, surprisingly: women who get breast cancer while on HRT do much better than women who get breast cancer while NOT on HRT.
For example, in 2009 Sener et al reviewed 1055 women who developed breast cancer; half of them had used HRT. Estrogen Receptor status was positive in 66% of HRT users and 64% of non-users. Yet HRT users had a much lower risk of death than HRT nonusers (hazard ratio .438, 95% confidence limit .263 to .729, P = .002).
(The effects of hormone replacement therapy on postmenopausal breast cancer biology and survival. Am J Surg 2009;197:403–7)
Others propose that women on HRT are much more vigilant due to anxiety about breast cancer, and thus they may be catching their tumors sooner, and therefore having a better outcome. This is not the whole story either, as you will see below: when women develop breast cancer while on HRT, their tumors actually have more favorable histology than tumors in women not on HRT. Some of these details remain to be elucidated; the bottom line for now is that women on HRT appear to get breast cancer less often, and when they do get it they have fewer metastatic cancers and fewer fatal cancers.
Q: Wow! We’ve covered women with family history of breast cancer, the WHI trial and its faults and the differences between HRT and effects on various cancers. What about breast cancer survivors? Should they avoid HRT?
A: Actually, in many cases even breast cancer survivors should use hormone replacement therapy. It will decrease your risk of heart disease, colon cancer, Alzheimer’s disease and osteoporosis, making your life far more enjoyable (and possibly longer). As a cancer survivor, it is natural that your focus will always be on preventing cancer recurrence; believe it or not, the literature shows that in breast cancer survivors who have been declared cancer-free, HRT will decrease your risk of recurrence!
That is probably contrary to what you have heard, and we invite you to review the papers at www.physioage.com/blog with your specialist.
The current data supports waiting until you are cancer-free for five years, and then initiating HRT — even if your cancer was ER positive or you have a BRCA1 mutation. The five-year mark is probably somewhat arbitrary, but that is how the studies to date have been done, and it’s best to stick with the literature.
In conclusion, we can gather that it does not appear that estrogen is a significant cause of breast cancer, women who get breast cancer while on HRT do much better than women who get breast cancer while NOT on HRT, and the protective effect of estrogens is more pronounced in women with a family history. Even breast cancer survivors can use HRT because it will decrease risks of other health diseases that menopausal women often face. A huge thank-you to Dr. Trutt for his passion and expertise, and to all you women: go and get the help you deserve!