So my dad, who died 15 years after his quad bypass at 81, would have most likely lived nearly the same timeframe with his stable angina just by taking statins? I questioned the surgery at the time but in 05’ that was the consensus. He had stable angina symptoms for years on exertion so not sure why we went through that awful bypass that changed his personality for the worse.
I had the privilege of serving as chief medical resident at Evanston Hospital under Dr. Thomas Killip, 1976-77. One of the teaching points he made about the Coronary Artery Surgery Study, was that the only patients randomized were the patients that their physicians thought should be randomized, and that it was therefore not surprising that for most of the patients randomized, there was not a significant difference in outcomes. He would say that he took the results as a sign that physician judgement was generally pretty accurate.
As humans with imperfect information, what we believe today as evidence-based, may and often DOES CHANGE with dynamic societal variables, and via ongoing skepticism, study and research. Consider Group Think and Manufactured Consent in the “science of medicine” - for LICENSED MDs who practice in U.S. profit-driven hell. Be conscious of physical externalities which bullshit (academically speaking) AI systems and political lies convey upon us, precipitating declines in our overall wellbeing.
I read your Substack article and then the 2011 NEJM STITCH article with great interest. Apparently 17% of the participants assigned to medical therapy arm in the STITCH trial underwent CABG during follow-up (median time to CABG was within 5 months of randomization in this arm). This is a worrying leakage and the STITCH authors have acknowledged the same in DISCUSSION. Due to such a significant leakage of participants to the other arm, the ITT analyses won't make any sense to me whatever the purists think of ITT. Therefore, I wanted to see if the authors reported a per-protocol analyses and they did (thanks!). Here it is "We also performed a per-protocol comparison of the 537 patients randomly assigned to medical therapy who did not cross over to CABG during the first year of follow-up and the 555 patients assigned to CABG who actually underwent CABG; the hazard ratio for the primary outcome with CABG was 0.76 (95% CI, 0.62 to 0.92; P=0.005)".
Given the fact that there was significant leakage and also because the study was powered to detect 25% reduction (observed HR was about 0.86 as per ITT analysis) in mortality, I do not think that the primary outcome results based on ITT analysis is in anyway conclusive. Moreover, the per-protocol analysis showed significant benefit with CABG. Secondary outcome results also does not support the primary outcome results. So, the STITCH trial result is INCONCLUSIVE and if it was up to me, I would still favour CABG over medical therapy for eligible ischemic heart disease patients with decreased EF. Cardiac surgeons can continue remaining happy.
Remember, though, that per-protocol analyses are biased. We randomise people into the arms of the trial, but we don’t randomise who escapes the medical management group by getting a CABG.
I haven’t studied the paper, but on first principles it seems likely such people are those who are sicker or more symptomatic. When your per-protocol analysis excludes such people, then you are left with participants in your analysis who tend to be less sick or symptomatic, and therefore likely to have better outcomes for reasons unrelated to their CABG.
Perhaps the best estimate of truth lies somewhere between the intention-to-treat and the per-protocol results.
I still take into account the 10 year long term follow up results. The curves started separating at about 2.5 years, and continued to separate albeit slowly until the end of the initial study. It is consistent with the finding of an “eventual” significant difference. And it fits with the paradigm that CABG benefits accrue over time.
However, this highlights once again the benefits of long term secondary prevention medical therapy, even in patients who have had CABG.
And it should be noted the benefit does not extend to PCI in this cohort, based on the Revived study.
I am still curious about the viability substudy, as surgeons still seem to ask for it despite that result.
How many negative studies does it take to reach the (so far) obvious conclusion that there is no therapy, medical or surgical, that makes a clinically significant difference in improving mortality figures for patients with coronary artery disease? There have been clear advances in the treatment of acute disease and that should be the role of the interventional cardiologist---treating acute MI and unstable angina. The curves showing tiny insignificant differences in this comparative study look remarkedly similar to those for the statin studies and practically every other proposed form of medical "treatment". Many will be aghast and say "What, would you do nothing?" My response would be yes for asymptomatic patients regardless of the meaningless risk profiles.
In addition to deaths, I'd like to know if there were significant differences regarding quality of life between the two groups?
Were there any significant differences between the groups on how far they both able to walk without getting breathless. Did one group require more oxygen support, more "rescue" hospitalizations due to pulmonary edema, ascites ... ?
It's one thing to be alive and able to get out for a mile walk without stopping and another to be alive but housebound while tethered to an O2 generator with low quality of life.
Thanks for posting that link. Dr. Mandrola never mentioned harms, and I was wondering about that. While the 5% ARR sounded pretty good at first (better than statins for primary prevention), the absolute harms outweigh the benefits, and it's always important to look at that side of the equation.
16% sounds quite a lot to me, especially if you were one of the dead patients who might not have died if they'd been on the other arm of the trial! Isn't there a danger of confusing statistical significance with real importance?
The 16% is a relative risk reduction---a figure that is often quoted in order to make the difference between two options appear much more significant than it really is. And confusing "statistical significance" with importance or practical significance is an error made by many in the analysis of medical or other scientific papers.
That's not how it works. If there's not statistical significance, you're saying there's a greater than 5% chance that the 16% reduction you are observing is random noise and if you run this trial again you may get a completely different number.
Not quite! The significance level of 5%, means that *if* it's all random noise, then the probability of getting results as extreme as the actual results is 5%. The only rigorous way to reverse this and get the probability of a definite hypothesis (like it's all random noise) is to use Bayes' theorem which raises a host of further problems. An alternative is to extend the idea of confidence intervals (as explained at https://doi.org/10.1177/2059799119826518) which leads to a tentative conclusion of a 97.5% probability that the CABG option is better. However, it’s important to remember any problems with the data, and that this is an average and in particular cases there may be reasons to favour one option rather than another.
In the eighties, when people were talking about the pros and cons of Canadian style health care and universal insurance, I heard that a study had shown that more people died in the US getting their CABG pronto, than did Canadians on the wait list for one.
CAD is stable if 1) you know u have it 2) you stop smoking , eating etc 3) get decent medical treatment . Lots of deaths in US of people who most of us would think have CAD who never went to doc or got treated . It is accepted as " another Mcd cheeseburger guy had a heart attack and died " . If the same number of people died of an infectious disease, no one would leave the house.
I am not a cardiologist, but I seem to remember being told that beta blockers don't improve mortality. They improve blood pressure (and reduce heart rate, of course). People don't live longer, they just die with "normal" blood pressure. Does this then imply that neither beta blockers nor surgical intervention improves mortality? Yes, I realize that it was not just beta blockers that were given to people in the medical arm of this study.
I wonder about the effective of much of what we offer as medical providers. It would be nice if there was a non-medical, non-surgical arm of the trial where people were coached on exercise appropriate to their health and coached on nutrition. That being said, if I was the patient, I would probably opt for an intervention of some sort out of fear. Which just goes to show the power that doctors have over patients in poor health.
I am noticing a trend in research where we use studies that only last a couple of years to extrapolate benefit over a lifetime. I understand the need to get new medications out there but where are the studies that follow patients for 15-20 years plus? A 60yo male should still have 20+ years of life expectancy. Will he get there with severe CAD on medical therapy alone? My gut response would be no way.
There is STITCHES (STITCH Extended Study), 9.8 years which appears to my eye to show a reduction in all cause mortality and CV-related mortality especially in younger patients. I am looking forward to reading more about it from Dr Mandrola.
Good observation. Trials happen because they are funded. Funding is limited. Long-term fu is costly. Most trials last 1-5 years. But you are right; in an ideal world, we would have longer trials.
By the time a patient is followed-up for 10-20 years, another new intervention will probably hatch out making the intervention of the original trial obsolete and no longer much relevant.
My very competent cardiologist keeps saying at my annual post CABG visit (now I am 81 years old, 9 years out from surgery, and still solo skydiving too), “I feel sure that the statin you take (20 mg atorvastatin qd) has stimulated plaque regression in you”. What’s all this buzz about the pleotropic effects of statins?
Pleiotropic effects of a drug are actions other than those for which the agent was specifically developed. These effects may be related or unrelated to the primary mechanism of action of the drug, and they are usually unanticipated. Pleiotropic effects may be undesirable (such as side effects or toxicity), neutral, or, as is especially the case with HMG-CoA reductase inhibitors (statins), beneficial. Pleiotropic effects of statins include improvement of endothelial dysfunction, increased nitric oxide bioavailability, antioxidant properties, inhibition of inflammatory responses, and stabilization of atherosclerotic plaques. These and several other emergent properties could act in concert with the potent low-density lipoprotein cholesterol-lowering effects of statins to exert early as well as lasting cardiovascular protective effects. Understanding the pleiotropic effects of statins is important to optimize their use in treatment and prevention of cardiovascular disease.
Beneficial Cardiovascular Pleiotropic Effects of Statin
So my dad, who died 15 years after his quad bypass at 81, would have most likely lived nearly the same timeframe with his stable angina just by taking statins? I questioned the surgery at the time but in 05’ that was the consensus. He had stable angina symptoms for years on exertion so not sure why we went through that awful bypass that changed his personality for the worse.
I had the privilege of serving as chief medical resident at Evanston Hospital under Dr. Thomas Killip, 1976-77. One of the teaching points he made about the Coronary Artery Surgery Study, was that the only patients randomized were the patients that their physicians thought should be randomized, and that it was therefore not surprising that for most of the patients randomized, there was not a significant difference in outcomes. He would say that he took the results as a sign that physician judgement was generally pretty accurate.
I live in Florida. I need a cardiologist who thinks like this. My current one is full-on conventional medicine. Any recommendations?
Definitely get a 2nd opinion for a conservative medical approach, if possible. I personally would only have CABG in an MI situation only.
As humans with imperfect information, what we believe today as evidence-based, may and often DOES CHANGE with dynamic societal variables, and via ongoing skepticism, study and research. Consider Group Think and Manufactured Consent in the “science of medicine” - for LICENSED MDs who practice in U.S. profit-driven hell. Be conscious of physical externalities which bullshit (academically speaking) AI systems and political lies convey upon us, precipitating declines in our overall wellbeing.
I read your Substack article and then the 2011 NEJM STITCH article with great interest. Apparently 17% of the participants assigned to medical therapy arm in the STITCH trial underwent CABG during follow-up (median time to CABG was within 5 months of randomization in this arm). This is a worrying leakage and the STITCH authors have acknowledged the same in DISCUSSION. Due to such a significant leakage of participants to the other arm, the ITT analyses won't make any sense to me whatever the purists think of ITT. Therefore, I wanted to see if the authors reported a per-protocol analyses and they did (thanks!). Here it is "We also performed a per-protocol comparison of the 537 patients randomly assigned to medical therapy who did not cross over to CABG during the first year of follow-up and the 555 patients assigned to CABG who actually underwent CABG; the hazard ratio for the primary outcome with CABG was 0.76 (95% CI, 0.62 to 0.92; P=0.005)".
Given the fact that there was significant leakage and also because the study was powered to detect 25% reduction (observed HR was about 0.86 as per ITT analysis) in mortality, I do not think that the primary outcome results based on ITT analysis is in anyway conclusive. Moreover, the per-protocol analysis showed significant benefit with CABG. Secondary outcome results also does not support the primary outcome results. So, the STITCH trial result is INCONCLUSIVE and if it was up to me, I would still favour CABG over medical therapy for eligible ischemic heart disease patients with decreased EF. Cardiac surgeons can continue remaining happy.
Remember, though, that per-protocol analyses are biased. We randomise people into the arms of the trial, but we don’t randomise who escapes the medical management group by getting a CABG.
I haven’t studied the paper, but on first principles it seems likely such people are those who are sicker or more symptomatic. When your per-protocol analysis excludes such people, then you are left with participants in your analysis who tend to be less sick or symptomatic, and therefore likely to have better outcomes for reasons unrelated to their CABG.
Perhaps the best estimate of truth lies somewhere between the intention-to-treat and the per-protocol results.
I still take into account the 10 year long term follow up results. The curves started separating at about 2.5 years, and continued to separate albeit slowly until the end of the initial study. It is consistent with the finding of an “eventual” significant difference. And it fits with the paradigm that CABG benefits accrue over time.
However, this highlights once again the benefits of long term secondary prevention medical therapy, even in patients who have had CABG.
And it should be noted the benefit does not extend to PCI in this cohort, based on the Revived study.
I am still curious about the viability substudy, as surgeons still seem to ask for it despite that result.
How many negative studies does it take to reach the (so far) obvious conclusion that there is no therapy, medical or surgical, that makes a clinically significant difference in improving mortality figures for patients with coronary artery disease? There have been clear advances in the treatment of acute disease and that should be the role of the interventional cardiologist---treating acute MI and unstable angina. The curves showing tiny insignificant differences in this comparative study look remarkedly similar to those for the statin studies and practically every other proposed form of medical "treatment". Many will be aghast and say "What, would you do nothing?" My response would be yes for asymptomatic patients regardless of the meaningless risk profiles.
In addition to deaths, I'd like to know if there were significant differences regarding quality of life between the two groups?
Were there any significant differences between the groups on how far they both able to walk without getting breathless. Did one group require more oxygen support, more "rescue" hospitalizations due to pulmonary edema, ascites ... ?
It's one thing to be alive and able to get out for a mile walk without stopping and another to be alive but housebound while tethered to an O2 generator with low quality of life.
https://www.thennt.com/cms/nnt/coronary-heart-bypass-surgery-for-prevention-of-death/
Thanks for posting that link. Dr. Mandrola never mentioned harms, and I was wondering about that. While the 5% ARR sounded pretty good at first (better than statins for primary prevention), the absolute harms outweigh the benefits, and it's always important to look at that side of the equation.
16% sounds quite a lot to me, especially if you were one of the dead patients who might not have died if they'd been on the other arm of the trial! Isn't there a danger of confusing statistical significance with real importance?
The 16% is a relative risk reduction---a figure that is often quoted in order to make the difference between two options appear much more significant than it really is. And confusing "statistical significance" with importance or practical significance is an error made by many in the analysis of medical or other scientific papers.
That's not how it works. If there's not statistical significance, you're saying there's a greater than 5% chance that the 16% reduction you are observing is random noise and if you run this trial again you may get a completely different number.
Not quite! The significance level of 5%, means that *if* it's all random noise, then the probability of getting results as extreme as the actual results is 5%. The only rigorous way to reverse this and get the probability of a definite hypothesis (like it's all random noise) is to use Bayes' theorem which raises a host of further problems. An alternative is to extend the idea of confidence intervals (as explained at https://doi.org/10.1177/2059799119826518) which leads to a tentative conclusion of a 97.5% probability that the CABG option is better. However, it’s important to remember any problems with the data, and that this is an average and in particular cases there may be reasons to favour one option rather than another.
I had the same thought.
In the eighties, when people were talking about the pros and cons of Canadian style health care and universal insurance, I heard that a study had shown that more people died in the US getting their CABG pronto, than did Canadians on the wait list for one.
Interesting natural experiment. I wouldn’t be surprised.
CAD is stable if 1) you know u have it 2) you stop smoking , eating etc 3) get decent medical treatment . Lots of deaths in US of people who most of us would think have CAD who never went to doc or got treated . It is accepted as " another Mcd cheeseburger guy had a heart attack and died " . If the same number of people died of an infectious disease, no one would leave the house.
The real driver tht keeps CABG so popular is the money it makes for institutions and cardiac surgeons!
I am not a cardiologist, but I seem to remember being told that beta blockers don't improve mortality. They improve blood pressure (and reduce heart rate, of course). People don't live longer, they just die with "normal" blood pressure. Does this then imply that neither beta blockers nor surgical intervention improves mortality? Yes, I realize that it was not just beta blockers that were given to people in the medical arm of this study.
I wonder about the effective of much of what we offer as medical providers. It would be nice if there was a non-medical, non-surgical arm of the trial where people were coached on exercise appropriate to their health and coached on nutrition. That being said, if I was the patient, I would probably opt for an intervention of some sort out of fear. Which just goes to show the power that doctors have over patients in poor health.
I am noticing a trend in research where we use studies that only last a couple of years to extrapolate benefit over a lifetime. I understand the need to get new medications out there but where are the studies that follow patients for 15-20 years plus? A 60yo male should still have 20+ years of life expectancy. Will he get there with severe CAD on medical therapy alone? My gut response would be no way.
There is STITCHES (STITCH Extended Study), 9.8 years which appears to my eye to show a reduction in all cause mortality and CV-related mortality especially in younger patients. I am looking forward to reading more about it from Dr Mandrola.
https://www.ahajournals.org/doi/full/10.1161/CIRCULATIONAHA.116.024800
Good observation. Trials happen because they are funded. Funding is limited. Long-term fu is costly. Most trials last 1-5 years. But you are right; in an ideal world, we would have longer trials.
By the time a patient is followed-up for 10-20 years, another new intervention will probably hatch out making the intervention of the original trial obsolete and no longer much relevant.
My very competent cardiologist keeps saying at my annual post CABG visit (now I am 81 years old, 9 years out from surgery, and still solo skydiving too), “I feel sure that the statin you take (20 mg atorvastatin qd) has stimulated plaque regression in you”. What’s all this buzz about the pleotropic effects of statins?
Pleiotropic effects of a drug are actions other than those for which the agent was specifically developed. These effects may be related or unrelated to the primary mechanism of action of the drug, and they are usually unanticipated. Pleiotropic effects may be undesirable (such as side effects or toxicity), neutral, or, as is especially the case with HMG-CoA reductase inhibitors (statins), beneficial. Pleiotropic effects of statins include improvement of endothelial dysfunction, increased nitric oxide bioavailability, antioxidant properties, inhibition of inflammatory responses, and stabilization of atherosclerotic plaques. These and several other emergent properties could act in concert with the potent low-density lipoprotein cholesterol-lowering effects of statins to exert early as well as lasting cardiovascular protective effects. Understanding the pleiotropic effects of statins is important to optimize their use in treatment and prevention of cardiovascular disease.
Beneficial Cardiovascular Pleiotropic Effects of Statin
Circulation. 2004;109[suppl III]:III-39-III-43
May I kindly ask how long you have been taking statins? Thank you.
9 years