When I was trained as an epidemiologist at UNC Chapel Hill, my mentor Charlie Pool demonstrated how clinical trials can give a wrong answer despite p-value below 0.05. I don’t remember the details. However, I strongly support the school of thought in epidemiology, which claims that there is no gold standard. Every study design is liable for bias.
I am disturbed that you "still favor functional stress testing for the evaluation of patients with chest pain."
Stress testing whether coupled with nuclear imaging or echo imaging has an unacceptably high rate of false-positives in real world scenarios. These tests are often read by the same cardiologist who does the subsequent unncessary invasive angiogram and stenting.
Functional tests ideally let us know about stenosis >70% but are normal (when not false-positives) in the presence of severe diffuse coronary atherosclerosis. Those non-stenotic high-risk plaques and high overall plaque burden will be totally missed by your functional stress test.
Patients will be told everything is fine when they are at high for a myocardial infarction.
Coronary CT angiogram visualizes and can quantify the burden of coronary atherosclerosis and allow us to make intelligent decisions about preventive medical therapy and allow patients to fully understand their long term risk of coronary events.
The original study report and the 10 year follow-up do not provide sufficient information to allow a clear picture of what produced the difference in the outcomes, as they were defined.
A CT scan has no direct therapeutic effect. Any effect is indirect through changes in therapy.
In 2025 we know that LDL cholesterol in the intima of arteries is the cause of atherosclerotic plaque with all its accompanying pathophysiology.
We also know that the lower the LDL, the lower the risk of an atherothrombotic event; ACS. There is debate about how low the maximal LDL should be (1.4 or 1.8) and exactly how quantify it (calculated LDL, non-HDL, ApoB).
Revascularization reduces the likelihood of an atherothrombotic event, at least at the site of the revascularization, for about 3-5 years statistically.
There is no data is either report on lipid levels. The statin utilization is surprisingly low for a population with a mean BMI of 29. It only got up to a maximum of 59 & 53% of patients in the 2 groups.
The revascularization rate was very identical.
There were 34 leas non-fatal MI/ but there were still 90 non-fatal MIs
So in the end, a statistical success fir the outcomes as they are defined but no real isight into why.
I am a strong believer in the use of CT yo diagnose the presence or absence of atherosclerosis.
I am also a strong believer that once found, ApoB should be <0.7 g/L & nonHDL/LDL <2/1.4 mmol/L respectively. I also believe ApoB < 0.5 & LDL <0.5 is even better.
This was an interesting study but a missed opportunity.
“Lp(a), pronounced “LP little a,” is a novel blood fat (lipid) that is drawing more attention as a marker for severe CAD (coronary artery disease). However it was discovered in 1963, but only lately has much attention focused on its importance as a marker for heart disease.
Lp(a) is an interesting lipoprotein. Although its existence is not new, there has been a surge in interest in measuring levels for predicting heart disease. Lp(a) is a dense type of lipid that, when levels are high (>125 mmol/L), is good at predicting heart disease. Not to confuse you, but another lipoprotein, apoB, is part of the Lp(a) particle and can also be measured and, if high, predicts heart disease.”
Cardiology Made Easy is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Trying to explain outliers is difficult and in the current environment where so much of medical research has been politically corrupted I would not worry about it. Spend your time trying to figure out who so many experts on a variety of issues dealing with both COVID and the vac . medicine has been politicized beyond recognition. Credible people must look at returning ethics and credibility to the Soul of helping people once again before politicians took over…but they are still in the shadows. Move on, if your studies were credible then people’s data does not always have to correlate as you would wish and hope.
The answer is that this diagnostic test and all others that have been investigated thus far do not lead to reductions in future MIs. An absolute risk reduction of 1.6% is not significant no matter what statistical prestidigitations are applied to show "statistical significance". The same can also be said for statin drugs and every other "preventive" measure that has been recommended to date. Those who think they can provide a remotely accurate assessment of risk for future MI are simply indulging in wishful thinking. That is not to suggest that these possibilities shouldn't be investigated but please stop attributing significance to these minute differences. Atherosclerosis is a feature of aging that proceeds at variable rates in different individuals for reasons that are not well understood and this chronic course is punctuated by acute episodes (MIs and sudden death) that are equally not well understood. Treatment of symptomatic and acute disease has advanced significantly over the past few decades and we should continue to focus on that.
The CCTA, along with a Coronary Ca++ score - a roughly 20 min testing involving a peripheral IV dye infusion with CT scanning for both - provides an added visual dimension which frequently influences decisions regarding how vigorously to treat risk factors, and it's exceptionally helpful in separating pulmonary from coronary causes of exertional dyspnea.
This test has its uses. But “first test for anyone who complains of CP” is certainly not it, for the multitude of reasons you’ve listed.
From a trial interpretation standpoint, the change in their primary endpoint from initial study to their long-term follow-up papers, as well as the implausibility of effect size, are the aspects that are curious for me.
I agree that pts with normal or non-occlusive CAD at CCTA are less likely to be coded as "non-STEMI" when they are hospitalized with pneumonia or COPD and are incidentally found to have a minor Troponin leak.
However, let's assume that these incidental Troponin leaks do indeed occur less frequently in the group who had CCTA, possibly due to a combination of lifestyle changes that decrease general inflammation and higher use of aggressive statins, ASA and beta-blockers. I suspect many of the patients in Scot-heart with the endpoint of "non-fatal MI" were indeed found to have Tn elevations when being evaluated for non-cardiac complaints. Are these minor incidental Tn elevations clinically important?
In my hospital, a patient with pneumonia and a hsTn of 20 (>15 considered abnormal)could be coded as "non-STEMI". Does this number/designation have clinical or prognostic value?
I would argue about preventive drugs and NNT argument. It is not absoulte difference in patients taking statin and aspirin that matters, but who were those patients. CCTA is able to detect patients with severe atherosclerotic burden, who get the most benefit from statins and aspirin, while patients in stess-test group who were taking preventive drugs possibly were not really highest risk patients and would not have prescribed statins/aspirin were their CCTA data available. So may be it was result not only of the number of patients takins statins/aspirin, but also of precision in selecting those to whom prescribe it. The same can be applied to life-style modification.
How many stopped smoking, lost weight, or started exercising. Going from a NNT of 2 to 3 only required a 1% increase in RRR. The scare tactic in this trial seemed to work. I see it as a good thinig and people should scrutinize the totality of the changes that came after the testing.
Isn't the simplest explanation that a p value is simply a p value; so long as p>0, it may be that chance alone explains the result. And, for this reason, replicability is essential for this -- and any -- result.
Not saying that it did, but here's how it could decrease future MI by that much.
Let's say I have 2 fifty year old males in my primary care office. Both at intermediate risk of heart disease causing their stable angina. The one gets a stress test. It doesn't show anything that would benefit from heart catheterization (note I'm not sending patients for PCI in stable chest pain anyway). But the message he *unintentionally* gets is that he doesn't need a stent so he's fine. The next guy gets the stress test. Same thing. But then gets the CCTA. I note the plaque buildup in his arteries. He asks his friends if they ever had a CCTA. In 2014 it was way less common so no one has ever heard of it. He starts researching about it. He starts following all sorts of people and groups with a high calcium scores and plaque. And starts implementing all sorts of lifestyle changes we aren't able to measure. And so ends up switching out the daily cereal he has been eating for maybe sweet potatoes, carrots, meat, and some eggs. He starts cooking at home instead of buying takeout or going to a restaurant. He starts biking to work. And it's easy to venture a guess as to how that may be possible.
The part that is more interesting to me is the following: How many people suffered or died in the following 10 years? It's exactly the same. Did the higher statin use in the CT group predispose to more infections? Did they get thyroid disease from the substantial iodine infusion with the CT? Did they die of bleeding disease from aspirin? Did they get cancer? Did they end up on more PCSK9 inhibitors? Did they follow up with doctors more often, get more covid "vaccines" and die because of that? Or did he switch to a vegan diet to address his "high cholesterol" and ended up dying of a hip fracture? So all in all, I think we tend to focus on things we measure. It's hard to measure impacts on "whole person health" accurately.
Thank you for sharing the study! Overall it doesn't really change my bias towards doing or not doing the CT scans as a diagnostic but it changes the conversation a bit.
"I note the plaque buildup in his arteries. He asks his friends if they ever had a CCTA. In 2014 it was way less common so no one has ever heard of it. He starts researching about it. He starts following all sorts of people and groups with a high calcium scores and plaque. And starts implementing all sorts of lifestyle changes we aren't able to measure. And so ends up switching out the daily cereal he has been eating for maybe sweet potatoes, carrots, meat, and some eggs. He starts cooking at home instead of buying takeout or going to a restaurant. He starts biking to work. And it's easy to venture a guess as to how that may be possible'
We wish such easy lifestyle changes had that much of an effect. Using statins in people who have already had MIs probably doesn't even have the risk reduction seen in this study.
So you don't like using a tool that was developed using funding from a pharmaceutical company that channels people into using medications they make and they come with a wide variety of side effects but half of patients will have some difficulty stopping and 1/4 will have a terrible time stopping the medication especially once they've been on it for a while. Kudos to you!
I don't routinely use the form, may use it if I'm trying to gauge therapy. I am more cautious with dx and tx, but overall my goal is to have an informed consent discussion which includes medications along with lifestyle things such as time outside, exercise, limiting screen use, proper stress management, sleep, nutrition, spirituality. Granted, I have the luxury of time in a direct primary care practice. And almost always patients choose the lifestyle route as the primary treatment modality.
Above my pay grade but this related diagnostic method was just recently studied. My eye doctor has had this ability for years and she has also used it as a basis to send eye patients to hospital in an ambulance. My scans have always been excellent.
Am I correct in assuming that an exercise or chemical "stress test," with MPI, gives the same information to the cardiologist as an "FFR" during a CCTA?
When I was trained as an epidemiologist at UNC Chapel Hill, my mentor Charlie Pool demonstrated how clinical trials can give a wrong answer despite p-value below 0.05. I don’t remember the details. However, I strongly support the school of thought in epidemiology, which claims that there is no gold standard. Every study design is liable for bias.
I am disturbed that you "still favor functional stress testing for the evaluation of patients with chest pain."
Stress testing whether coupled with nuclear imaging or echo imaging has an unacceptably high rate of false-positives in real world scenarios. These tests are often read by the same cardiologist who does the subsequent unncessary invasive angiogram and stenting.
Functional tests ideally let us know about stenosis >70% but are normal (when not false-positives) in the presence of severe diffuse coronary atherosclerosis. Those non-stenotic high-risk plaques and high overall plaque burden will be totally missed by your functional stress test.
Patients will be told everything is fine when they are at high for a myocardial infarction.
Coronary CT angiogram visualizes and can quantify the burden of coronary atherosclerosis and allow us to make intelligent decisions about preventive medical therapy and allow patients to fully understand their long term risk of coronary events.
The original study report and the 10 year follow-up do not provide sufficient information to allow a clear picture of what produced the difference in the outcomes, as they were defined.
A CT scan has no direct therapeutic effect. Any effect is indirect through changes in therapy.
In 2025 we know that LDL cholesterol in the intima of arteries is the cause of atherosclerotic plaque with all its accompanying pathophysiology.
We also know that the lower the LDL, the lower the risk of an atherothrombotic event; ACS. There is debate about how low the maximal LDL should be (1.4 or 1.8) and exactly how quantify it (calculated LDL, non-HDL, ApoB).
Revascularization reduces the likelihood of an atherothrombotic event, at least at the site of the revascularization, for about 3-5 years statistically.
There is no data is either report on lipid levels. The statin utilization is surprisingly low for a population with a mean BMI of 29. It only got up to a maximum of 59 & 53% of patients in the 2 groups.
The revascularization rate was very identical.
There were 34 leas non-fatal MI/ but there were still 90 non-fatal MIs
So in the end, a statistical success fir the outcomes as they are defined but no real isight into why.
I am a strong believer in the use of CT yo diagnose the presence or absence of atherosclerosis.
I am also a strong believer that once found, ApoB should be <0.7 g/L & nonHDL/LDL <2/1.4 mmol/L respectively. I also believe ApoB < 0.5 & LDL <0.5 is even better.
This was an interesting study but a missed opportunity.
So what is your perspective on Lp(a) testing?
“Lp(a), pronounced “LP little a,” is a novel blood fat (lipid) that is drawing more attention as a marker for severe CAD (coronary artery disease). However it was discovered in 1963, but only lately has much attention focused on its importance as a marker for heart disease.
Lp(a) is an interesting lipoprotein. Although its existence is not new, there has been a surge in interest in measuring levels for predicting heart disease. Lp(a) is a dense type of lipid that, when levels are high (>125 mmol/L), is good at predicting heart disease. Not to confuse you, but another lipoprotein, apoB, is part of the Lp(a) particle and can also be measured and, if high, predicts heart disease.”
Cardiology Made Easy is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Trying to explain outliers is difficult and in the current environment where so much of medical research has been politically corrupted I would not worry about it. Spend your time trying to figure out who so many experts on a variety of issues dealing with both COVID and the vac . medicine has been politicized beyond recognition. Credible people must look at returning ethics and credibility to the Soul of helping people once again before politicians took over…but they are still in the shadows. Move on, if your studies were credible then people’s data does not always have to correlate as you would wish and hope.
The answer is that this diagnostic test and all others that have been investigated thus far do not lead to reductions in future MIs. An absolute risk reduction of 1.6% is not significant no matter what statistical prestidigitations are applied to show "statistical significance". The same can also be said for statin drugs and every other "preventive" measure that has been recommended to date. Those who think they can provide a remotely accurate assessment of risk for future MI are simply indulging in wishful thinking. That is not to suggest that these possibilities shouldn't be investigated but please stop attributing significance to these minute differences. Atherosclerosis is a feature of aging that proceeds at variable rates in different individuals for reasons that are not well understood and this chronic course is punctuated by acute episodes (MIs and sudden death) that are equally not well understood. Treatment of symptomatic and acute disease has advanced significantly over the past few decades and we should continue to focus on that.
The CCTA, along with a Coronary Ca++ score - a roughly 20 min testing involving a peripheral IV dye infusion with CT scanning for both - provides an added visual dimension which frequently influences decisions regarding how vigorously to treat risk factors, and it's exceptionally helpful in separating pulmonary from coronary causes of exertional dyspnea.
Thanks for the detailed summary.
This test has its uses. But “first test for anyone who complains of CP” is certainly not it, for the multitude of reasons you’ve listed.
From a trial interpretation standpoint, the change in their primary endpoint from initial study to their long-term follow-up papers, as well as the implausibility of effect size, are the aspects that are curious for me.
I agree that pts with normal or non-occlusive CAD at CCTA are less likely to be coded as "non-STEMI" when they are hospitalized with pneumonia or COPD and are incidentally found to have a minor Troponin leak.
However, let's assume that these incidental Troponin leaks do indeed occur less frequently in the group who had CCTA, possibly due to a combination of lifestyle changes that decrease general inflammation and higher use of aggressive statins, ASA and beta-blockers. I suspect many of the patients in Scot-heart with the endpoint of "non-fatal MI" were indeed found to have Tn elevations when being evaluated for non-cardiac complaints. Are these minor incidental Tn elevations clinically important?
In my hospital, a patient with pneumonia and a hsTn of 20 (>15 considered abnormal)could be coded as "non-STEMI". Does this number/designation have clinical or prognostic value?
Very good critique. Something is off.
I would argue about preventive drugs and NNT argument. It is not absoulte difference in patients taking statin and aspirin that matters, but who were those patients. CCTA is able to detect patients with severe atherosclerotic burden, who get the most benefit from statins and aspirin, while patients in stess-test group who were taking preventive drugs possibly were not really highest risk patients and would not have prescribed statins/aspirin were their CCTA data available. So may be it was result not only of the number of patients takins statins/aspirin, but also of precision in selecting those to whom prescribe it. The same can be applied to life-style modification.
How many stopped smoking, lost weight, or started exercising. Going from a NNT of 2 to 3 only required a 1% increase in RRR. The scare tactic in this trial seemed to work. I see it as a good thinig and people should scrutinize the totality of the changes that came after the testing.
Isn't the simplest explanation that a p value is simply a p value; so long as p>0, it may be that chance alone explains the result. And, for this reason, replicability is essential for this -- and any -- result.
In the last paragraph of your meta-analysis, the editorialist found your findings "intriguing" but advocated for business as usual.
Not saying that it did, but here's how it could decrease future MI by that much.
Let's say I have 2 fifty year old males in my primary care office. Both at intermediate risk of heart disease causing their stable angina. The one gets a stress test. It doesn't show anything that would benefit from heart catheterization (note I'm not sending patients for PCI in stable chest pain anyway). But the message he *unintentionally* gets is that he doesn't need a stent so he's fine. The next guy gets the stress test. Same thing. But then gets the CCTA. I note the plaque buildup in his arteries. He asks his friends if they ever had a CCTA. In 2014 it was way less common so no one has ever heard of it. He starts researching about it. He starts following all sorts of people and groups with a high calcium scores and plaque. And starts implementing all sorts of lifestyle changes we aren't able to measure. And so ends up switching out the daily cereal he has been eating for maybe sweet potatoes, carrots, meat, and some eggs. He starts cooking at home instead of buying takeout or going to a restaurant. He starts biking to work. And it's easy to venture a guess as to how that may be possible.
The part that is more interesting to me is the following: How many people suffered or died in the following 10 years? It's exactly the same. Did the higher statin use in the CT group predispose to more infections? Did they get thyroid disease from the substantial iodine infusion with the CT? Did they die of bleeding disease from aspirin? Did they get cancer? Did they end up on more PCSK9 inhibitors? Did they follow up with doctors more often, get more covid "vaccines" and die because of that? Or did he switch to a vegan diet to address his "high cholesterol" and ended up dying of a hip fracture? So all in all, I think we tend to focus on things we measure. It's hard to measure impacts on "whole person health" accurately.
Thank you for sharing the study! Overall it doesn't really change my bias towards doing or not doing the CT scans as a diagnostic but it changes the conversation a bit.
"I note the plaque buildup in his arteries. He asks his friends if they ever had a CCTA. In 2014 it was way less common so no one has ever heard of it. He starts researching about it. He starts following all sorts of people and groups with a high calcium scores and plaque. And starts implementing all sorts of lifestyle changes we aren't able to measure. And so ends up switching out the daily cereal he has been eating for maybe sweet potatoes, carrots, meat, and some eggs. He starts cooking at home instead of buying takeout or going to a restaurant. He starts biking to work. And it's easy to venture a guess as to how that may be possible'
We wish such easy lifestyle changes had that much of an effect. Using statins in people who have already had MIs probably doesn't even have the risk reduction seen in this study.
I would say the absolute risk reduction of 1.6% isn't that much.
"tend to focus on things we measure": And conversely, tend to ignore what isn't or can't be measured via a supposedly objective test (e.g., pain).
It's unreal how many patients I've seen over the years who had been told to "go see a psychiatrist" because "all your tests are normal."
So you don't like using a tool that was developed using funding from a pharmaceutical company that channels people into using medications they make and they come with a wide variety of side effects but half of patients will have some difficulty stopping and 1/4 will have a terrible time stopping the medication especially once they've been on it for a while. Kudos to you!
I don't routinely use the form, may use it if I'm trying to gauge therapy. I am more cautious with dx and tx, but overall my goal is to have an informed consent discussion which includes medications along with lifestyle things such as time outside, exercise, limiting screen use, proper stress management, sleep, nutrition, spirituality. Granted, I have the luxury of time in a direct primary care practice. And almost always patients choose the lifestyle route as the primary treatment modality.
Above my pay grade but this related diagnostic method was just recently studied. My eye doctor has had this ability for years and she has also used it as a basis to send eye patients to hospital in an ambulance. My scans have always been excellent.
https://heart.bmj.com/content/early/2025/01/03/heartjnl-2024-324705
Am I correct in assuming that an exercise or chemical "stress test," with MPI, gives the same information to the cardiologist as an "FFR" during a CCTA?
If patients knew the NNT values of these interventions, they might rethink the glorification of statins.