As a layperson, I wonder what the study would have shown had progesterone been used instead of progestin. But why would such a study even be attempted? There is no money to be made by marketing a compound that can't be patented...
You left out the elephant in the Women's Health Initiative study--all subjects received conjugated equine estrogen + progestin (IIRC the study was sponsored by Wyeth (clearly in hopes that their HRT product, Prempro, would get a boost from the study results, rather than being so dramatically shot down!)). To then generalize (as you have throughout your post) from equine conjugate to all HRT (and then writing as if all HRT is the same formula) is at best imprecise and at worst, sloppy. So I think your strong conclusion, despite all the impressive statistical analysis, misses a crucial point.
Even back in 2000, I thought equine conjugate was a bad idea because potential for immune system interactions. There are other HRT formulas: human recombinant estrogen, formulas using estriol, with or without progestin. Unfortunately, many of the alternative HRTs are compounded, with no large RCTs to guide clinical judgment.
Also, cardio-protection is not the only, nor from the user's standpoint, the most important, reason to choose HRT.
The 19-year re-analysis of the WHI found it had many many issues. For one, they forgot to control for women historically who had taken ER (estrogen therapy) in the placebo arm. Dr. Leon Speroff began early on showing the many issues with the WHI. Then Naftolin, the Yale statistician. Then my book, Safe Hormones, and then the 19-year re-analysis by 12 prestigious institutions, the very original authors of the first July 2002 published WHI findings, retracted their original study!!!! Their new reanlysis, 19 years later, hind site is better than foresight, found clearly that estrogen replacement (with Premarin in this case - ER., was statistically linked to a 23% reduced risk of getting breast cancer and a 44% reduced risk of fatality if you got this disease if you had taken estrogen. The new NIH 7 million medicare study is a retrospective observational study showing decreased all cause premature mortality if on ER, and a significant reduction of all five cancers (colorectal, lung, breast, ovarian and uterine) and only oral had increased risk of ischemic stroke and dementia while the other modes, especially vaginal, were CV and CA protective. None of the large European studies showed issues with HRT and in fact so much benefit that HRT is given for free in many countries to help protect their medical system as the hormones protect their citizens so much, such as Finland, Iceland, Wales, Scotland, italy, on and on and even Britain got in on this a year ago with a sliding fee schedule but free for those making much less money. Why? Socialized medicine and tracking their registries, like CV, CA and death registries, show such consistent benefit. It all depends on which science you are looking at. I have been on HRT for about 26 years a few years after I had breast cancer, almost 30 years ago now. And... going strong.
The following link is the textbook level of evidence based medicine. There is no way to put this politely. Almost nobody here is even at the textbook level of expertise in philosophy of medicine.
I appreciate your point but I'm afraid the example you chose to cite is a poor one. You were more correct on your first assertion. We've got 18 yr follow-up studies on HRT/MHT demonstrating that 20 yr study was poorly designed and newer studies found that women taking any type of HRT have a lower risk of heart disease, osteoporosis, diabetes, dementia, death, and also colon cancer, and other cancers as well. Women taking any type of HRT are less likely to die from breast cancer. It's assertions like these, based on old science, that contribute to millions of women not getting adequate treatment for menopause and unnecessary suffering. https://www.nhmenopausesociety.org/research/hormone-therapy-why-we-should-no-longer-be-afraid-of-the-breast-cancer-risk-says-new-literature-review/
In spite of the flaws in the WHI study, as identified in previous comments, the conclusion of the article is correct - RCTs are the only way to prove medical efficacy over a large population. Could John have found a more valid RCT to support his thesis?
RCTs can be so flawed that their results are misleading and dangerous, as in the reported WHI study.
Studies which focus on mortality end points can leave out the real-world trade-offs between risk of death vs quality of life. We face such trade-offs every time we decide to travel or even to drive to the grocery store and, unfortunately, statistics will never give us a clear way to strike a balance.
Case in point: my wife has been on HRT for over 20 years, starting around menopause. ObGyn's often don't approve, as a result of the WHI study, and we have to pay for her HRT drugs because Kaiser won't. But the symptoms she has experienced any time she has gone off of HRT have been more than adequate to convince her to resume HRT. Fortunately, she has a trusted non-Kaiser health professional who is supportive of HRT. The balance is between the certainty of symptoms that make her miserable vs the possible (now largely discredited) increased risk of life-threatening disease.
It seems to me that any study like the WHI should at least document the difference in quality-of-life symptoms between the placebo and intervention arms of the study. Did WHI do this?
There's no such thing as "bad cholesterol". In fact, there's no such thing as "high cholesterol". The idea that LDL causes heart disease is nonsense and I can't believe in this day and age that doctors still believe it.
Here's one of literally thousands of articles I could post showing there's no association between heart disease and "bad cholesterol".
This is an excellent review outlining a common foible in clinical intuition leading to observational research which is not yet science. HRT started in the 1950s when a cardiologist in Massachusetts noted that HRT not only relieved postmenopausal symptoms but seemed to lower the incidents of CVD and AMI among his patients. The rest is history. Your story omits the part where Wyeth (then Wyeth-Ayerst, then Pfizer) the manufacturers of Premarin (a proprietary mixture of estrogens actually extracted from horse urine) spent huge amounts on advertising to make certain this product dominated the market and was demanded by many women. It gained huge cultural acceptance. The large controlled randomized study you cite was necessary to break this advertising juggernaut.
You also omitted to note that while P/E HRT significantly increased the risk of breast cancer in women, Premarin and other E only HRT increased the risk of uterine cancer 500% more.
For all the women more concerned about HRT than maybe the overall point of this writing, I'm not seeing what type of HRT was administered. Premarin and estradiol are definitely different drugs. Within a few months of a hysterectomy, premarin caused deep indentations in nails and my hair continued to fall out in large amounts, for months after the surgery. A friend educated me about the horse urine used in the drug, the side affects and about plant based estrogen. The estradiol had no side affects (for 25 years). This is anecdotal and probably doesn't apply to a majority of women, but if you are in need of HRT for whatever reason, the information may be helpful.
Except that the same WHI group published research 15 years later showing that HRT has no impact on all-cause mortality at all. The 2002 study was flawed and caused much misery to women benefitting from but subsequently denied HRT. Scare-mongering like "HRT led to a nearly 50k women being harmed" needs to stop. doi:10.1001/jama.2017.11217
In many ways, the cholesterol hypothesis of heart disease provided the blue print for big pharma during the covid crisis. The heart attack ‘epidemic’ sweeping America during the 50’s was heavily promoted in the MSM but was almost certainly due to better technology for testing and changing disease classifications. Academics who disagreed with the cholesterol hypothesis were scapegoated, ridiculed and denied research funding. An expensive ‘antidote’ to heart disease (ie the various statins that have been manufactured over the years) was developed and given to a willing public.
Meanwhile, the evidence that over consumption of saturated and/or animal fats cause heart disease remain, 70 years later, ambiguous at best. The evidence that statins provide benefit to all but a sub-group of highly at-risk patients also remains stubbornly ambiguous. The safety profile of these drugs is certainly up for debate and not helped by clinical trial organisations (CTSU in Oxford, I’m looking at you!!) withholding accumulated safety data from the public
Wonder where we’ve seen all this happen over the last two years.......?
Another great example that STILL is directing patient care erroneously is the observation that patients who get iodinated contrast may have increases in serum creatinine. Worse- it was conflated to iodinated contrast can cause not just creatinine rises, but renal failure. When a matched control study was done, it was found that the hospitalized patients who did NOT get iodinated contrast had the same amount of rise in serum creatinine that hospitalized patients who got iodinated contrast had. Iodinated contrast had no role in the creatinine rise. Yet, still, patients are refused the benefits of a CT scan with contrast because of this fallacy.
It is indeed complicated. Observational Studies can be biased. A single large RCT might be flawed, too. I think, here is some critique on that RCT: https://m.youtube.com/watch?v=DTCmprPCDqc&t=4210s
A couple items. First, all “studies” begin with a single observation. It’s the genesis, the seed, for an hypothesis then a testable theory. So observational studies are powerful in that they guide innovation. A serious problem with studies looking for statistical power is the systematic introduction of error that can dilute or eliminate real effects that are actually real as observed in the case study. This happens often when the suspected treatment agent is only effective when interacting with a second unidentified factor, or there is an unidentified factor responsible for the observation. Jumping to a parametric study will only find no effect and the discovery is lost. Second, in the HRT study the assumption that E reducing LDLs as the mechanism reducing cardiovascular events is errant leading to the finding of no effect in the parametric study. Many studies have shown that cholesterol is not the relevant factor in CVD, the relevant factor is vascular inflammation. In fact, the relevant variable in the study mentioned is actually vitamin D status and its effects on reducing inflammation hence CVD. So, while E is a minor agonist of vitD, the heavy lifting with respect to CVD was vitD. So here we have that second unidentified, unappreciated, ignored? variable that is responsible for the observed effect. Don’t get me started on why vitD, a free, safe, readily available agent is systematically villified/ignored by the medical community. I think it has something to do with $$$.
John I wish you expand your thoughts with regard to two issues: (1) you are emphasizing relative risks rather than absolute risk increases; these absolute risk increases seem small per woman, and (2) women take HRT for a reason and place some value on the benefits received. Shouldn't those values be formally traded off against the (small) absolute risk increases? As an aside I'd like to see us de-emphasize the multiplication of risk by some population size, and concentrate on what the risk is per-person.
As a layperson, I wonder what the study would have shown had progesterone been used instead of progestin. But why would such a study even be attempted? There is no money to be made by marketing a compound that can't be patented...
This article is so on the money. And so applicable to our situation today.
The last thing I needed was another substack to follow, but this one seems irresistible.
+10 for the Gary Taubes citation.
You left out the elephant in the Women's Health Initiative study--all subjects received conjugated equine estrogen + progestin (IIRC the study was sponsored by Wyeth (clearly in hopes that their HRT product, Prempro, would get a boost from the study results, rather than being so dramatically shot down!)). To then generalize (as you have throughout your post) from equine conjugate to all HRT (and then writing as if all HRT is the same formula) is at best imprecise and at worst, sloppy. So I think your strong conclusion, despite all the impressive statistical analysis, misses a crucial point.
Even back in 2000, I thought equine conjugate was a bad idea because potential for immune system interactions. There are other HRT formulas: human recombinant estrogen, formulas using estriol, with or without progestin. Unfortunately, many of the alternative HRTs are compounded, with no large RCTs to guide clinical judgment.
Also, cardio-protection is not the only, nor from the user's standpoint, the most important, reason to choose HRT.
The 19-year re-analysis of the WHI found it had many many issues. For one, they forgot to control for women historically who had taken ER (estrogen therapy) in the placebo arm. Dr. Leon Speroff began early on showing the many issues with the WHI. Then Naftolin, the Yale statistician. Then my book, Safe Hormones, and then the 19-year re-analysis by 12 prestigious institutions, the very original authors of the first July 2002 published WHI findings, retracted their original study!!!! Their new reanlysis, 19 years later, hind site is better than foresight, found clearly that estrogen replacement (with Premarin in this case - ER., was statistically linked to a 23% reduced risk of getting breast cancer and a 44% reduced risk of fatality if you got this disease if you had taken estrogen. The new NIH 7 million medicare study is a retrospective observational study showing decreased all cause premature mortality if on ER, and a significant reduction of all five cancers (colorectal, lung, breast, ovarian and uterine) and only oral had increased risk of ischemic stroke and dementia while the other modes, especially vaginal, were CV and CA protective. None of the large European studies showed issues with HRT and in fact so much benefit that HRT is given for free in many countries to help protect their medical system as the hormones protect their citizens so much, such as Finland, Iceland, Wales, Scotland, italy, on and on and even Britain got in on this a year ago with a sliding fee schedule but free for those making much less money. Why? Socialized medicine and tracking their registries, like CV, CA and death registries, show such consistent benefit. It all depends on which science you are looking at. I have been on HRT for about 26 years a few years after I had breast cancer, almost 30 years ago now. And... going strong.
The following link is the textbook level of evidence based medicine. There is no way to put this politely. Almost nobody here is even at the textbook level of expertise in philosophy of medicine.
https://plato.stanford.edu/entries/medicine/#RandContTriaEvidBaseMedi
Sorry, now go read.
I appreciate your point but I'm afraid the example you chose to cite is a poor one. You were more correct on your first assertion. We've got 18 yr follow-up studies on HRT/MHT demonstrating that 20 yr study was poorly designed and newer studies found that women taking any type of HRT have a lower risk of heart disease, osteoporosis, diabetes, dementia, death, and also colon cancer, and other cancers as well. Women taking any type of HRT are less likely to die from breast cancer. It's assertions like these, based on old science, that contribute to millions of women not getting adequate treatment for menopause and unnecessary suffering. https://www.nhmenopausesociety.org/research/hormone-therapy-why-we-should-no-longer-be-afraid-of-the-breast-cancer-risk-says-new-literature-review/
A couple of points:
In spite of the flaws in the WHI study, as identified in previous comments, the conclusion of the article is correct - RCTs are the only way to prove medical efficacy over a large population. Could John have found a more valid RCT to support his thesis?
RCTs can be so flawed that their results are misleading and dangerous, as in the reported WHI study.
Studies which focus on mortality end points can leave out the real-world trade-offs between risk of death vs quality of life. We face such trade-offs every time we decide to travel or even to drive to the grocery store and, unfortunately, statistics will never give us a clear way to strike a balance.
Case in point: my wife has been on HRT for over 20 years, starting around menopause. ObGyn's often don't approve, as a result of the WHI study, and we have to pay for her HRT drugs because Kaiser won't. But the symptoms she has experienced any time she has gone off of HRT have been more than adequate to convince her to resume HRT. Fortunately, she has a trusted non-Kaiser health professional who is supportive of HRT. The balance is between the certainty of symptoms that make her miserable vs the possible (now largely discredited) increased risk of life-threatening disease.
It seems to me that any study like the WHI should at least document the difference in quality-of-life symptoms between the placebo and intervention arms of the study. Did WHI do this?
There's no such thing as "bad cholesterol". In fact, there's no such thing as "high cholesterol". The idea that LDL causes heart disease is nonsense and I can't believe in this day and age that doctors still believe it.
Here's one of literally thousands of articles I could post showing there's no association between heart disease and "bad cholesterol".
https://drmalcolmkendrick.org/2017/03/20/cholesterol-lowering-proven-or-not/
This is an excellent review outlining a common foible in clinical intuition leading to observational research which is not yet science. HRT started in the 1950s when a cardiologist in Massachusetts noted that HRT not only relieved postmenopausal symptoms but seemed to lower the incidents of CVD and AMI among his patients. The rest is history. Your story omits the part where Wyeth (then Wyeth-Ayerst, then Pfizer) the manufacturers of Premarin (a proprietary mixture of estrogens actually extracted from horse urine) spent huge amounts on advertising to make certain this product dominated the market and was demanded by many women. It gained huge cultural acceptance. The large controlled randomized study you cite was necessary to break this advertising juggernaut.
You also omitted to note that while P/E HRT significantly increased the risk of breast cancer in women, Premarin and other E only HRT increased the risk of uterine cancer 500% more.
For all the women more concerned about HRT than maybe the overall point of this writing, I'm not seeing what type of HRT was administered. Premarin and estradiol are definitely different drugs. Within a few months of a hysterectomy, premarin caused deep indentations in nails and my hair continued to fall out in large amounts, for months after the surgery. A friend educated me about the horse urine used in the drug, the side affects and about plant based estrogen. The estradiol had no side affects (for 25 years). This is anecdotal and probably doesn't apply to a majority of women, but if you are in need of HRT for whatever reason, the information may be helpful.
Except that the same WHI group published research 15 years later showing that HRT has no impact on all-cause mortality at all. The 2002 study was flawed and caused much misery to women benefitting from but subsequently denied HRT. Scare-mongering like "HRT led to a nearly 50k women being harmed" needs to stop. doi:10.1001/jama.2017.11217
In many ways, the cholesterol hypothesis of heart disease provided the blue print for big pharma during the covid crisis. The heart attack ‘epidemic’ sweeping America during the 50’s was heavily promoted in the MSM but was almost certainly due to better technology for testing and changing disease classifications. Academics who disagreed with the cholesterol hypothesis were scapegoated, ridiculed and denied research funding. An expensive ‘antidote’ to heart disease (ie the various statins that have been manufactured over the years) was developed and given to a willing public.
Meanwhile, the evidence that over consumption of saturated and/or animal fats cause heart disease remain, 70 years later, ambiguous at best. The evidence that statins provide benefit to all but a sub-group of highly at-risk patients also remains stubbornly ambiguous. The safety profile of these drugs is certainly up for debate and not helped by clinical trial organisations (CTSU in Oxford, I’m looking at you!!) withholding accumulated safety data from the public
Wonder where we’ve seen all this happen over the last two years.......?
Another great example that STILL is directing patient care erroneously is the observation that patients who get iodinated contrast may have increases in serum creatinine. Worse- it was conflated to iodinated contrast can cause not just creatinine rises, but renal failure. When a matched control study was done, it was found that the hospitalized patients who did NOT get iodinated contrast had the same amount of rise in serum creatinine that hospitalized patients who got iodinated contrast had. Iodinated contrast had no role in the creatinine rise. Yet, still, patients are refused the benefits of a CT scan with contrast because of this fallacy.
It is indeed complicated. Observational Studies can be biased. A single large RCT might be flawed, too. I think, here is some critique on that RCT: https://m.youtube.com/watch?v=DTCmprPCDqc&t=4210s
@1h22min
A couple items. First, all “studies” begin with a single observation. It’s the genesis, the seed, for an hypothesis then a testable theory. So observational studies are powerful in that they guide innovation. A serious problem with studies looking for statistical power is the systematic introduction of error that can dilute or eliminate real effects that are actually real as observed in the case study. This happens often when the suspected treatment agent is only effective when interacting with a second unidentified factor, or there is an unidentified factor responsible for the observation. Jumping to a parametric study will only find no effect and the discovery is lost. Second, in the HRT study the assumption that E reducing LDLs as the mechanism reducing cardiovascular events is errant leading to the finding of no effect in the parametric study. Many studies have shown that cholesterol is not the relevant factor in CVD, the relevant factor is vascular inflammation. In fact, the relevant variable in the study mentioned is actually vitamin D status and its effects on reducing inflammation hence CVD. So, while E is a minor agonist of vitD, the heavy lifting with respect to CVD was vitD. So here we have that second unidentified, unappreciated, ignored? variable that is responsible for the observed effect. Don’t get me started on why vitD, a free, safe, readily available agent is systematically villified/ignored by the medical community. I think it has something to do with $$$.
John I wish you expand your thoughts with regard to two issues: (1) you are emphasizing relative risks rather than absolute risk increases; these absolute risk increases seem small per woman, and (2) women take HRT for a reason and place some value on the benefits received. Shouldn't those values be formally traded off against the (small) absolute risk increases? As an aside I'd like to see us de-emphasize the multiplication of risk by some population size, and concentrate on what the risk is per-person.