Hello, I share your concern about the unblinded nature of the study with respect to anginal symptoms which are notoriously difficult to assess and heavily affected by a patient's knowledge of "blockages".
Regarding the MI endpoint discussed in Dr. Mandrola's commentary, the FAME2 trial was conducted at 28 different sites - none of which were using high-sensitivity troponin. The trial allowed for the use of CK-MB and its not clear how many used this biomarker for the assessment of MI in followup (although it was the recommended biomarker for peri-procedural MI). Furthermore, the study results you discuss were 1 year data. Examining the 3 year and 5 year followup data from FAME2, the following rates of spontaneous MI are noted:
PCI vs Medical therapy:
3yrs: 5.8% vs 6.8% (NS)
5yrs: 8.1% vs. 12.0% 0.66 (95% CI 0.43–1.00).
The relatively delayed separation of MI curves is very biologically plausible given what we know about plaque progression in well treated patients, and consistent with sub-study data from ISCHEMIA.
Furthermore, hs-Tn assays increase MI detection by about 5% (3.3% to 8.1%) which is 2.45x fold increase. We cannot apply this directly to FAME2, but certainly there would have been more MIs detected had high sensitivity assays been used. For those of caring with patients with suspected ACS on a daily basis, "true unstable angina" is increasingly rare.
Lastly, its worth reminding us all that only one sham/blinded controlled trial has been done in Afib: SHAM-PVI. A relatively small study of 126 patients which showed a 60% reduction in afib in the ablation group compared to 35% in the sham group at 6 months. And reported a modest improvement in symptom burden with ablation. These curves are likely to overlap rather than diverge over time given what we know about the biology of Afib.
Lots of work to be done in both stable PCI and AFib with respect to identifying the appropriate patients, study design including endpoints and followup duration. But there is reason to believe physiology-guided PCI reduces spontaneous MI.
First, I thoroughly enjoy “this week in cardiology,” your analyses and commentary as well as your X posts!
As an iCards guy, I completely agree that the trial was poorly done. But, I do think we need metrics that guide us in treatment decisions. In my very small experience, there have been some instances where I have medically managed a lesion, only to repeat a cath a few years later and find the lesion to be much worse and more difficult to fix than it would have been at the time of the first angio, or to have the patient present with ACS in a more unstable situation. I do think that achieving adequate longer term follow up is a problem with many cardiology trials. As a specialty, let’s continue to aggressively pursue features and metrics that help us to angioplasty/stent when most appropriate.
Despite the 2018 commentary, the EARLY TAVR trial (an unblinded study comparing observation with TAVR in severe aortic stenosis) used death, stroke, or unplanned hospitalization for cardiovascular causes (including TAVR) as its primary endpoint. As you might expect, the primary outcome was positive but was largely driven by the latter component. The Kaplan–Meier curves immediately reveal the issue, with very early separation followed by largely parallel curves thereafter.
When the results were released, I discussed this potential bias with cardiologist colleagues, but their response was simply that they hoped early TAVR would rapidly be incorporated into the guidelines.
Thanks to the ORBITA studies, we now know that it is possible to include sham controls even in procedural trials, and that the results can sometimes surprise us. There is little justification for not including such controls.
I don’t think this is an issue with FFR as a procedure (or iFR). It’s a way to “measure” the physiologic significance of an anatomical coronary lesion. In fact, I believe more recent studies have generally confirmed that you actually do LESS PCI when you base decisions of physiology rather than “eyeball” estimations of anatomic pathology.
The problem, as you note, is the lack of sham control/procedural blinding on an endpoint susceptible to such bias. This issue has reared its ugly head often (Trilluminate comes to mind but there are many others).
This study is now also quite old. We have (thankfully) Orbita 1 and 2 in the interim. There will be no excuse for studies in this new era to NOT have sham control.
This is a pattern we see repeatedly in cardiac research where there is usually a low incidence of meaningful endpoints and one or more measures with some subjective elements involved is included in the composite in order to squeak out statistical significance. For example, statin studies include nonfatal MI (often diagnosed with minimal changes in blood markers and ECGs) and/or cardiovascular death which is frequently applied to many sudden deaths without pathological confirmation. Hospitalization for chest pain is another that is often used to soften up the composite. Ditto for "revascularization". The generally tiny difference between the treatment and control groups is then said to be statistically significant and, in the case of statins, warrants "preventive treatment" for life. The best advice to anyone trying to make sense of these types of studies is to disregard the hazard ratios, confidence intervals, regression equations, and other bits of statistical legerdemain; just calculate the absolute percentages for each element in a composite end point and use common sense to determine whether anything useful has been shown. Beware of investigators that constantly repeat relative risk ratios without identifying them as such and don't give the absolute numbers for perspective.
Dr. Mandrola excellent analysis once again, and I am in line with your thinking. My only retort would be in the setting of say an intermediate risk type 1 MI NSTEMI that needs routine invasive strategy, you also routinely argue for culprit only PCI. If not for FFR, IVUS, etc how else could we identify culprits when other clinical data is not certain? I think there is still utility for these technologies even if FAME2 is flawed, using your culprit only PCI argument.
If hospital systems had to pay out a DOLLAR for every stenting or angioplasty, they’d be orphan procedures. Rather than being the golden goose for CFOs and cardiology groups. Ditto joint replacement in 90 year olds, and 6-12 months of neoadjuvant chemotherapy in breast cancer, with marginal benefit over shorter course adjuvant therapy. In 78 year olds.
I believe that stenting and plasty are valuable and effective. In some. I also know that my utterly unexpected MI, with two subsequent stents and a total of 3 days’ admission w/o ICU, nor other complications was billed out at $284,000 (not a mistype) by the for-profit where I was treated. Would have been similar at the ‘nonprofit’ facility in the next town. Glad I had it. Appalled at the cost (forget the ‘but they don’t get paid that’ yada, yada: they were paid a LOT).
Put on an antiplatelet agent costing 10x the similarly effective older generic alternative. Because’reasons’. And because the cardiologist practice gets lunch every day, courtesy of XYZ Pharmaceuticals.
Go to any medium-to-large town in America, and look for the most impressive buildings in town. They will be:
State or county main offices
Insurance building
Hospital (cardiac wing, Ortho center, cancer center, women’s health annex…)
All built with your money. None of which are subject to competitive market pricing. All dependent on mandated/subsidized programs.
I most recently spent time in a metro of 125,000 or so. Two hospital systems that each cover a few blocks. Literally right next to each other. Competitive?
You bet. Competitive PRICING? Don’t be silly.
I’m no grizzled survivalist living in a trailer in the woods, hoarding cans of beans and ammunition. But, boy…sometimes I wonder.
Last year I had an unexpected MI while traveling in Ankara, Turkey. I had 2 urgent stents, spent 3 days in the hospital. Care was excellent. Total cost was $3600. Our healthcare system is seriously out of whack and unsustainable.
The neglect of psychological realities (placebo effects, emotional state, pleasing one's physician) has been a massive cause of wasted resources in healthcare.
Very helpful analysis. I am not sure I agree with the premise that "avoiding an urgent revascularization is a positive..." The premise of the trial rests on the idea that an "urgent" percutaneous coronary intervention (PCI) is dramatically different that an "elective" PCI. In the first place, as you show, most of the "urgent" PCIs in the medical therapy arm were driven by subjective symptoms rather than any objective evidence of clinical harm. In the FAME-2 trial, then, the authors are proposing that doing elective PCI in 100% of patients with FFR<0.80 is clinical valuable because it prevents doing "urgent" PCI in 15% of them otherwise (the difference in "urgent" PCI between intervention [10%] and medical therapy [25%] arms. If one ignores the distinction between "elective" and "urgent" PCI, the proposition is absurd.
The fallacy here I believe is classifying a medical intervention as a "therapy" in one arm and an "adverse event" in the other arm. This methodology plainly stacks the deck in favor of the intervention. They made the same error in the EARLY AS trial.
Can you image if EPs did a trial of catheter ablation for asymptomatic AF and randomized half to ablation and half to medical therapy and then considered ablation for symptomatic AF in follow-up to be an endpoint? Of course catheter ablation would "win." I think we would all recognize this thinking as fallacious.
It seems bizarre to use “avoidance of a procedure” as an endpoint after one arm started by getting an extra procedure. I’d be curious about “total procedures” in the 2 groups.
Also agree about your corollary example of Early TAVR. The only benefit of doing it early is “avoiding” doing it later.
Whenever we give someone an initial diabetes diagnosis, we should cut off their feet. A few percent will ultimately need amputations, so let’s just do it right away—and charge their insurance for it— to avoid later procedures, urgent or otherwise.
My read is that 100% of the intervention group had PCI initially, then 10% had "urgent" PCI in follow-up. 25% of the medical therapy group had "urgent" PCI in follow-up. Number of follow-up "elective" PCIs unknown. So that is at least 110 PCIs per 100 patients in the intervention arm and 25 PCIs per 100 patients in the medical therapy arm.
I would say any trial of a medical intervention in which the intervention is a "therapy" in one arm and an "adverse event or endpoint" in the other arm is invalid a priori.
..."one group got fixed; one group got tablets."...and in the end, not one person was healed because whatever is causing the stenosis or blockages will never be addressed. The root causes of these problems are ignored by doctors and that is a crime. Then again, it is because they have no clue.
The gospel according to mainstream cardiology is that excess cholesterol causes stenosis. They point to the fact that the buildup of cholesterol causes the narrowing of the blood vessel, which it does. Mainstream cardiologists tend to push statins on everyone, because "everyone" has cholesterol that is "too high." (Never mind that if "everyone" has a total cholesterol of, say, 200, then 200 is "normal.") Their theory is that lowering cholesterol via statins will reduce stenosis and save lives, which it won't.
The designated endpoint of Big Pharma-sponsored clinical trials for statins is "reducing cholesterol," which statins do (they also reduce CoQ10, which is necessary for proper muscle function -- and the heart is the hardest-working muscle in the body). Statins also cause statin-induced diabetes and statin-induced dementia, but never mind, because cardiologists only look at the cardiovascular system.
The biggest problem is that while statins absolutely DO reduce cholesterol, they DO NOT reduce mortality (the proper endpoint of clinical trials). A famous trial studied people with high cholesterol on statins vs placebo for 5 years, and they found that the statin cohort could expect to live only 3 or 4 days longer than the no-statin cohort. So after you and your insurance company pay Pfizer up to $40,500 for 5 years of Lipitor (retail: $675/month), you can expect to live less than a week longer.
Newer research points to a different cause of stenosis: inflammation. It works like this... Inflammation in the blood vessels cause microclots, which adhere to the vessel wall, especially at junction points. Cholesterol, which is ever-present in the blood, then covers the clot to prevent it from traveling to the lungs (pulmonary embolism) or brain (transient ischemic attack). In essence, cholesterol is "fixing" the problem, not "causing" the problem. Of course, a layer or two of cholesterol on top of the clots causes further narrowing of the blood vessel, which is stenosis.
So the cure is to reduce inflammation, not reduce cholesterol. The "problem" with that is that NSAIDs are very effective at reducing inflammation. But aspirin doesn't cost $675/month, so Big Pharma has no incentive to reduce inflammation to prevent stenosis.
Suggested reading: "The Clot Thickens: The enduring mystery of heart disease" (2021) by Dr Malcom Kendrick https://a.co/d/blMJd6s
Crix is really good at engaging in the comment section and explaining his comments, but if he doesn’t respond in a day or so, I’ll jump in with the older, mainstream (ACC/AHA) theory of what causes stenosis vs. more recent evidence of its cause.
Hello, I share your concern about the unblinded nature of the study with respect to anginal symptoms which are notoriously difficult to assess and heavily affected by a patient's knowledge of "blockages".
Regarding the MI endpoint discussed in Dr. Mandrola's commentary, the FAME2 trial was conducted at 28 different sites - none of which were using high-sensitivity troponin. The trial allowed for the use of CK-MB and its not clear how many used this biomarker for the assessment of MI in followup (although it was the recommended biomarker for peri-procedural MI). Furthermore, the study results you discuss were 1 year data. Examining the 3 year and 5 year followup data from FAME2, the following rates of spontaneous MI are noted:
PCI vs Medical therapy:
3yrs: 5.8% vs 6.8% (NS)
5yrs: 8.1% vs. 12.0% 0.66 (95% CI 0.43–1.00).
The relatively delayed separation of MI curves is very biologically plausible given what we know about plaque progression in well treated patients, and consistent with sub-study data from ISCHEMIA.
Furthermore, hs-Tn assays increase MI detection by about 5% (3.3% to 8.1%) which is 2.45x fold increase. We cannot apply this directly to FAME2, but certainly there would have been more MIs detected had high sensitivity assays been used. For those of caring with patients with suspected ACS on a daily basis, "true unstable angina" is increasingly rare.
Lastly, its worth reminding us all that only one sham/blinded controlled trial has been done in Afib: SHAM-PVI. A relatively small study of 126 patients which showed a 60% reduction in afib in the ablation group compared to 35% in the sham group at 6 months. And reported a modest improvement in symptom burden with ablation. These curves are likely to overlap rather than diverge over time given what we know about the biology of Afib.
Lots of work to be done in both stable PCI and AFib with respect to identifying the appropriate patients, study design including endpoints and followup duration. But there is reason to believe physiology-guided PCI reduces spontaneous MI.
Dr. Mandrola,
First, I thoroughly enjoy “this week in cardiology,” your analyses and commentary as well as your X posts!
As an iCards guy, I completely agree that the trial was poorly done. But, I do think we need metrics that guide us in treatment decisions. In my very small experience, there have been some instances where I have medically managed a lesion, only to repeat a cath a few years later and find the lesion to be much worse and more difficult to fix than it would have been at the time of the first angio, or to have the patient present with ACS in a more unstable situation. I do think that achieving adequate longer term follow up is a problem with many cardiology trials. As a specialty, let’s continue to aggressively pursue features and metrics that help us to angioplasty/stent when most appropriate.
Interesting read, thanks!
Despite the 2018 commentary, the EARLY TAVR trial (an unblinded study comparing observation with TAVR in severe aortic stenosis) used death, stroke, or unplanned hospitalization for cardiovascular causes (including TAVR) as its primary endpoint. As you might expect, the primary outcome was positive but was largely driven by the latter component. The Kaplan–Meier curves immediately reveal the issue, with very early separation followed by largely parallel curves thereafter.
When the results were released, I discussed this potential bias with cardiologist colleagues, but their response was simply that they hoped early TAVR would rapidly be incorporated into the guidelines.
Thanks to the ORBITA studies, we now know that it is possible to include sham controls even in procedural trials, and that the results can sometimes surprise us. There is little justification for not including such controls.
I don’t think this is an issue with FFR as a procedure (or iFR). It’s a way to “measure” the physiologic significance of an anatomical coronary lesion. In fact, I believe more recent studies have generally confirmed that you actually do LESS PCI when you base decisions of physiology rather than “eyeball” estimations of anatomic pathology.
The problem, as you note, is the lack of sham control/procedural blinding on an endpoint susceptible to such bias. This issue has reared its ugly head often (Trilluminate comes to mind but there are many others).
This study is now also quite old. We have (thankfully) Orbita 1 and 2 in the interim. There will be no excuse for studies in this new era to NOT have sham control.
This is a pattern we see repeatedly in cardiac research where there is usually a low incidence of meaningful endpoints and one or more measures with some subjective elements involved is included in the composite in order to squeak out statistical significance. For example, statin studies include nonfatal MI (often diagnosed with minimal changes in blood markers and ECGs) and/or cardiovascular death which is frequently applied to many sudden deaths without pathological confirmation. Hospitalization for chest pain is another that is often used to soften up the composite. Ditto for "revascularization". The generally tiny difference between the treatment and control groups is then said to be statistically significant and, in the case of statins, warrants "preventive treatment" for life. The best advice to anyone trying to make sense of these types of studies is to disregard the hazard ratios, confidence intervals, regression equations, and other bits of statistical legerdemain; just calculate the absolute percentages for each element in a composite end point and use common sense to determine whether anything useful has been shown. Beware of investigators that constantly repeat relative risk ratios without identifying them as such and don't give the absolute numbers for perspective.
What’s the difference in cost and profit between stent vs medicine only?
Do you believe that CT Angio with FFR should be the first line of investigation in an individual with chest pain that suggests potential angina ?
Dr. Mandrola excellent analysis once again, and I am in line with your thinking. My only retort would be in the setting of say an intermediate risk type 1 MI NSTEMI that needs routine invasive strategy, you also routinely argue for culprit only PCI. If not for FFR, IVUS, etc how else could we identify culprits when other clinical data is not certain? I think there is still utility for these technologies even if FAME2 is flawed, using your culprit only PCI argument.
If hospital systems had to pay out a DOLLAR for every stenting or angioplasty, they’d be orphan procedures. Rather than being the golden goose for CFOs and cardiology groups. Ditto joint replacement in 90 year olds, and 6-12 months of neoadjuvant chemotherapy in breast cancer, with marginal benefit over shorter course adjuvant therapy. In 78 year olds.
I believe that stenting and plasty are valuable and effective. In some. I also know that my utterly unexpected MI, with two subsequent stents and a total of 3 days’ admission w/o ICU, nor other complications was billed out at $284,000 (not a mistype) by the for-profit where I was treated. Would have been similar at the ‘nonprofit’ facility in the next town. Glad I had it. Appalled at the cost (forget the ‘but they don’t get paid that’ yada, yada: they were paid a LOT).
Put on an antiplatelet agent costing 10x the similarly effective older generic alternative. Because’reasons’. And because the cardiologist practice gets lunch every day, courtesy of XYZ Pharmaceuticals.
Go to any medium-to-large town in America, and look for the most impressive buildings in town. They will be:
State or county main offices
Insurance building
Hospital (cardiac wing, Ortho center, cancer center, women’s health annex…)
All built with your money. None of which are subject to competitive market pricing. All dependent on mandated/subsidized programs.
I most recently spent time in a metro of 125,000 or so. Two hospital systems that each cover a few blocks. Literally right next to each other. Competitive?
You bet. Competitive PRICING? Don’t be silly.
I’m no grizzled survivalist living in a trailer in the woods, hoarding cans of beans and ammunition. But, boy…sometimes I wonder.
I’m increasingly ashamed of my profession.
Last year I had an unexpected MI while traveling in Ankara, Turkey. I had 2 urgent stents, spent 3 days in the hospital. Care was excellent. Total cost was $3600. Our healthcare system is seriously out of whack and unsustainable.
“Then one gets fixed and feels fixed, and the other group gets meds and feels unfixed.” Love this phrase.
The neglect of psychological realities (placebo effects, emotional state, pleasing one's physician) has been a massive cause of wasted resources in healthcare.
Very helpful analysis. I am not sure I agree with the premise that "avoiding an urgent revascularization is a positive..." The premise of the trial rests on the idea that an "urgent" percutaneous coronary intervention (PCI) is dramatically different that an "elective" PCI. In the first place, as you show, most of the "urgent" PCIs in the medical therapy arm were driven by subjective symptoms rather than any objective evidence of clinical harm. In the FAME-2 trial, then, the authors are proposing that doing elective PCI in 100% of patients with FFR<0.80 is clinical valuable because it prevents doing "urgent" PCI in 15% of them otherwise (the difference in "urgent" PCI between intervention [10%] and medical therapy [25%] arms. If one ignores the distinction between "elective" and "urgent" PCI, the proposition is absurd.
The fallacy here I believe is classifying a medical intervention as a "therapy" in one arm and an "adverse event" in the other arm. This methodology plainly stacks the deck in favor of the intervention. They made the same error in the EARLY AS trial.
Can you image if EPs did a trial of catheter ablation for asymptomatic AF and randomized half to ablation and half to medical therapy and then considered ablation for symptomatic AF in follow-up to be an endpoint? Of course catheter ablation would "win." I think we would all recognize this thinking as fallacious.
Totally agree.
It seems bizarre to use “avoidance of a procedure” as an endpoint after one arm started by getting an extra procedure. I’d be curious about “total procedures” in the 2 groups.
Also agree about your corollary example of Early TAVR. The only benefit of doing it early is “avoiding” doing it later.
Whenever we give someone an initial diabetes diagnosis, we should cut off their feet. A few percent will ultimately need amputations, so let’s just do it right away—and charge their insurance for it— to avoid later procedures, urgent or otherwise.
Same bat-quit crazy logic!
My read is that 100% of the intervention group had PCI initially, then 10% had "urgent" PCI in follow-up. 25% of the medical therapy group had "urgent" PCI in follow-up. Number of follow-up "elective" PCIs unknown. So that is at least 110 PCIs per 100 patients in the intervention arm and 25 PCIs per 100 patients in the medical therapy arm.
I would say any trial of a medical intervention in which the intervention is a "therapy" in one arm and an "adverse event or endpoint" in the other arm is invalid a priori.
So, how the heck do you decide to jump in with the PCI?
..."one group got fixed; one group got tablets."...and in the end, not one person was healed because whatever is causing the stenosis or blockages will never be addressed. The root causes of these problems are ignored by doctors and that is a crime. Then again, it is because they have no clue.
Could you send the studies u mention
Sure. Here's the trial I mentioned that studied 5 years' statin use to produce 3-4 days longer life:
https://bmjopen.bmj.com/content/5/9/e007118
Go to Page 2, under RESULTS, see the last sentence of the 2nd paragraph:
"The median postponement of death for primary and secondary prevention trials were 3.2 and 4.1 days, respectively."
I’m not sure what you mean. Could you be more specific?
The gospel according to mainstream cardiology is that excess cholesterol causes stenosis. They point to the fact that the buildup of cholesterol causes the narrowing of the blood vessel, which it does. Mainstream cardiologists tend to push statins on everyone, because "everyone" has cholesterol that is "too high." (Never mind that if "everyone" has a total cholesterol of, say, 200, then 200 is "normal.") Their theory is that lowering cholesterol via statins will reduce stenosis and save lives, which it won't.
The designated endpoint of Big Pharma-sponsored clinical trials for statins is "reducing cholesterol," which statins do (they also reduce CoQ10, which is necessary for proper muscle function -- and the heart is the hardest-working muscle in the body). Statins also cause statin-induced diabetes and statin-induced dementia, but never mind, because cardiologists only look at the cardiovascular system.
The biggest problem is that while statins absolutely DO reduce cholesterol, they DO NOT reduce mortality (the proper endpoint of clinical trials). A famous trial studied people with high cholesterol on statins vs placebo for 5 years, and they found that the statin cohort could expect to live only 3 or 4 days longer than the no-statin cohort. So after you and your insurance company pay Pfizer up to $40,500 for 5 years of Lipitor (retail: $675/month), you can expect to live less than a week longer.
Newer research points to a different cause of stenosis: inflammation. It works like this... Inflammation in the blood vessels cause microclots, which adhere to the vessel wall, especially at junction points. Cholesterol, which is ever-present in the blood, then covers the clot to prevent it from traveling to the lungs (pulmonary embolism) or brain (transient ischemic attack). In essence, cholesterol is "fixing" the problem, not "causing" the problem. Of course, a layer or two of cholesterol on top of the clots causes further narrowing of the blood vessel, which is stenosis.
So the cure is to reduce inflammation, not reduce cholesterol. The "problem" with that is that NSAIDs are very effective at reducing inflammation. But aspirin doesn't cost $675/month, so Big Pharma has no incentive to reduce inflammation to prevent stenosis.
Suggested reading: "The Clot Thickens: The enduring mystery of heart disease" (2021) by Dr Malcom Kendrick https://a.co/d/blMJd6s
"Why Are Statins So Dangerous?" (2025) by A Midwestern Doctor https://www.midwesterndoctor.com/p/why-are-statins-so-dangerous
Crix is really good at engaging in the comment section and explaining his comments, but if he doesn’t respond in a day or so, I’ll jump in with the older, mainstream (ACC/AHA) theory of what causes stenosis vs. more recent evidence of its cause.