#185 - Allan Sniderman, M.D.: Cardiovascular disease and why we should change the way we assess risk

Hey everyone. Welcome to the Drive Podcast. I'm your host, Peter Attia. This podcast, my website and my weekly newsletter all focus on the goal of translating the science of longevity into something accessible for everyone. Our goal is to provide the best content in health and wellness, full stop, and we've assembled a great team of analysts to make this happen. If you enjoy this podcast, we've created a membership program that brings you far more in depth content. If you want to take your knowledge of this space to the next level, at the end of this episode I'll explain what those benefits are. Or if you want to learn more now, head over to peterattiamd.com subscribe Now without further delay. Here's today's episode. My guest this week is Dr. Alan Snyderman. Alan is a Senior Scientist at the Research Institute of McGill University Health center and the Edwards professor of Cardiology and professor of Medicine at McGill University University. He's the Director of the Mike Rosenblum Laboratory for Cardiovascular Research at Royal Victoria Hospital in Montreal, and he was elected a Fellow of the Royal Society of Canada in 2009. He's also been an enormous mentor of mine for the past 10 years, certainly one of the three or four people I would count that has nearly single handedly provided me with the education that I try to used today to help understand cardiovascular disease. A number of you may recognize the name Alan. I've certainly included him in some of the things I've written about, and also he's been mentioned a number of times on previous podcasts featuring no less than Tom Dayspring and Ron Kraus. His memberships are probably too numerous to mention, but a few of them include the Royal College of Physicians and Surgeons in Canada, the American College of Cardiology, the American College of Physicians, the Canadian Cardiovascular Society, the American Society for Clinical Investigation, the American Federation for Clinical Research, and a number of others. Okay, this is a bit of a complicated topic. We go very deep on cardiovascular medicine and we actually start this podcast not in a way that I intended to, because in the first 10 or 15 minutes Alan lays out some of the most clear, complex and you might say, well, how can those be used the same? But what I mean is clear thinking of complex concepts that you'll ever hear anybody talk about with respect to APOB and risk management. It becomes clear to me as I get engulfed in that discussion that I need to back this way up so that everyone listening to this who doesn't already find themselves steeped in that literature can orient themselves. So my first comment here is don't be dissuaded by how complicated the first 10 or 15 minutes of this podcast is. Instead, just try to sit tight and we'll walk you through the journey of what APOB is and why it is a superior metric for predicting risk of atherosclerosis relative to the far more commonly used metric LDL cholesterol and the metric that is better than LDL cholesterol but still inferior to apob, which is non HDL cholesterol. We also go into great detail about the role of triglycerides, HDL cholesterol, total cholesterol, etc. In the understanding of and prediction of cardiovascular disease. We get into some of the interesting exceptions, I.e. disease states where not knowing APOB poses an enormous blind spot. And we talk about the challenges that face people like Alan as they try to disseminate the most leading edge and cutting edge science on this topic in an environment that is really wed to guidelines and consensus based recommendations that don't necessarily incorporate all of the best evidence. We also explained some concepts like Mendelian randomization and how that's been used to further bolster the case for APOB as the superior metric for risk prediction. And we talk about where coronary artery calcium scoring comes in. This is obviously an important tool used by many physicians and it's important to understand how it's useful and what its blind spots are. So without further delay, please enjoy my conversation with Dr. Alan Snyder. Hey Alan, thanks so much for making time to sit down today. This is one of those discussions I've been meaning to have for over a decade, and I suspect much of what we speak about today will be reminiscent of things we've spoken about, usually in person over a meal over the past decade. So welcome to the show and sorry that we can't be doing this in person.

Thank you so much.

So I think listeners of the show will be pretty familiar with your name because certainly I've brought it up before, as have other people on the show, most notably probably Tom Dayspring. I've also probably referenced on at least one or two occasions the textbook you gave me. God, probably about eight years ago. Do you remember what book you gave me?

I'm trying to remember which I think it was the book I did with Jacqueline de gra.

That is one of them. But the one I'm referring to is actually, I think it's Herbert Starry the Pathology.

Oh, right, right, right, right, right, right, right. Oh, now I'm with you. Yeah, yeah, yeah.

You just immediately blew my mind. I Then went out and bought as many of them as I could find on Amazon. And they were prohibitively expensive. Tell me and tell the listeners, why did you give me that book?

I'm trying to recall. Atherosclerosis, it's a disease in the tissue, and almost everything that lipid people talk about is in plasma. And if we don't understand the natural history of the disease, how can we construct a strategy to prevent it? And although much of my work has been on apob, the more important part, I think, has been on understanding how the natural history of atherosclerosis should direct our prevention strategy. What that leads to is that every major guideline in the world bases their selection of subjects for statin prevention on the 10 year risk of disease. And that was a huge step forward, forward in 1980 and 1990. But it totally, or not totally, but it very fundamentally makes prevention of premature disease almost impossible. When you plug in the numbers to calculate somebody's risk for any of the risk algorithms, The American College of Cardiology, 2019. Aha, multi society. You plug in numbers that belong to that particular patient and what comes out is what you think is the risk for that particular patient. It actually isn't. But what drives that calculation is the age and the sex of that patient. Things like cholesterol, blood pressure, they contribute minimally to the actual calculation of 10 year risk. So what that means is if you're 35, well, there is even a risk calculator for you. But if you get to 40, almost everybody's risk is low at age 40, and it is until you get to about 55, 60. That risk gets you over the threshold for the American prevention guideline treatment. So prevention really starts at 55 to 60. But half, almost half of all infarcts and strokes occur before the age of 60. So how can that be? What Starry and his colleagues established was for the first three decades or so of life, the disease begins, gets a foothold in the artery. But it's only in the fourth decade that you start to develop the lesions that can actually precipitate a clinical event. But risk is low, and yet the event rate is high. How can that possibly be? Well, the answer is stunningly obvious, which we've published. There are a ton more people under 60 than over 60. So the rate of events is low, but the absolute number of events is high. That's problem number one. Problem number two is, say you get to 60 and you didn't have an event. Well, the disease was developing and extending during your 30s. 40s and 50s. So by the time we start to try and prevent an event, the disease is well advanced in the arteries. That, to me, are the two fatal flaws in the ten year risk approach. And we published a paper pointing this out in JAMA Cardiology a few years ago. Borg Nordiskaard and his colleagues have done exactly the same thing with the European guidelines. You can't beat these numbers. So rather than what Stere taught me, and it took some years before we could develop the methodology, of course risk is a good concept, of course it is. But we should be selecting people also based on causes. I can measure your APOB pretty precisely. I could measure your non HDL cholesterol a little less precisely, but pretty well. And I know it's yours. When I calculate the risk. If I said, okay, Peter, you're my patient, you're a healthy guy, I calculate Your risk is 4.1%. Now, what does that number mean? Is that your risk? Nope. It means that out of 100 people at 4.1%, 4.1 of them will have an infarct. But we know that within that category there's a tremendous variance in real risk. Not everybody's at 4.1. Some are higher, some are lower, some are dead on. So if I had two risk algorithms, the philosopher A.J. ayer, the English the logical positive, he was actually darn good on probability. There's a real challenge predicting singular events. I'm either going to have an infarct in the next year or. Or I'm not. It's not really a probability, so I either am or I'm not. If one algorithm said I had a 10% risk and another one said I had a 15% or 20%, whether I have an infarct or not, both of them were right because they said there was sort of a chance you could, and there was a far larger chance you wouldn't. When we say people should be treated with a risk above 7.5%, that means 92.5% of the time nothing will happen. Well, that's not a great incentive, I think, for helping people understand what's truly going to happen. So the way we can deal with this and what we've done is develop what's called a causal benefit model. We measure it non HCL or APOB, and we can project the risk over 20 or 30 years. If you're 30 years old, the period of time you should care about is up to age, at least to 60. And so if you were in a group, let's say, and let's say I make you 35 again. And I say your chances of having an infarct or a stroke before you're 65 are 30%. Now, that's a number you can deal with. That's a number that has meaning. And we could also calculate how much the risk can be reduced by starting at age 35, or how much you lose by starting at age 45, or how much more you lose by starting at age 55. When I gave you that book, I was starting my own journey on trying to construct an alternative to the present risk model in which, with the help of my colleague here at McGill, George Thanosoulis and Michael Pensina from Duke and Carol Pansina from Harvard, we've done something.

About it, Alan, that also brought home another message that had been somewhat left in the garage of my brain. As you may recall, I trained in general surgery. So the kids in med school who are going to go into surgery were not the sharpest tools in the shed like the kids that go into internal medicine. But I still remembered a couple of things from my pathology class and one of the things I remembered from pathology A. So it's the first of the three major classes you take in pathology was something that the professor said, which is, he said no doctor has more experience with what it is to have heart attacks than pathologists because 50% of the people who have a heart attack die on their first heart attack. So he said, I'm seeing 50% of the people who have a heart attack and their first presentation is death. So I kind of remembered that. And it's a very sobering fact to think that half the time. And again, I don't think that's true today, But I think 25 years ago that was the case. The numbers are probably a bit better today. It might be a third of the first events are fatal, but nevertheless it was sobering. So you have this sort of weird factoid that's again off in the recesses of my brain somewhere and then you hand me this textbook. And it actually made sense with what he said because in addition to going through in great detail the pathological staging of atherosclerosis, it was littered with autopsy sections of coronary arteries of people who had died for other reasons, and notably they were quite young. So here's a 26 year old male victim of a gunshot wound. Here's a 27 year old female who died in a motor vehicle accident. Here's a so and so and so and so. And when you look at their coronary Arteries, you realize they already have atherosclerosis. They already have oxidized apob bearing particles engulfed by macrophages and thickened intima. And while they may not have calcification in their arteries yet, or the types of plaque that would rupture within the ensuing weeks or days or months, they nevertheless had atherosclerosis. And they were in their 20s and in their 30s. So all of a sudden what this professor said 20 some odd years earlier made sense, which is this was now an explanation. This was a bridge to explain what otherwise seemed hard to understand. That's the thing I took away from it in the instant you handed it to me as we were literally looking at it in the restaurant.

The thing I'd emphasize is how when you have atherosclerosis, I mean, as a cardiologist, we're so used to looking at angiograms, we say, oh, there's an LED lesion, well, there's an LAD lesion, but the whole damn artery is diseased, and when you destroy the normal architecture of the artery, you can't restore it. So a lot of our statin prevention therapy is to prevent the complications of disease, not to prevent the disease. And statins lower apob particle number. That's how they work. Fewer apob particles in plasma, fewer get into the arterial wall, fewer get trapped. It's not that complicated.

So let's back up for a minute because I think everything we've just talked about for the last, whatever it's been 10 minutes is in many ways the advanced, Advanced seminar on prevention of atherosclerosis. So now let's go back and set the stage for this, because there are going to be a number of people listening to this who perhaps have not heard previous discussions I've had on this subject matter. So let's back way up and start with what is cholesterol and how does it relate to this thing called apob that you've mentioned a number of times already? Most people would have heard of cholesterol, and most people understand that you can measure it when you take somebody's blood. But this apob thing might be new to some people.

Cholesterol is obviously a fat lipid. It's a critical element in cell structure. It's in all cell membranes. The amount that's in the cell membrane is determinant of the function of the membrane. All the cells in the body can synthesize cholesterol. Only the liver can really break it down in any amount and excrete it. So when we Eat. We absorb cholesterol and fatty acids in the form of triglycerides, and they get resynthesized in the intestine into particles you can't transport. Cholesterol and triglyceride. Triglycerides are the fatty acids tacked onto a glycerol backbone. They don't mix with water, they're not soluble. So you have to put them in particles like soap bubbles. And there are a variety of different particles. The ones from the intestine that take the fat that we absorb, the cholesterol and the triglyceride, they're very large particles. There are very few of them. They have a protein called APOB48 on the outside surface, which gives integrity to the particle, structural integrity. There are also a bunch of other proteins. The chylomicrons deposit. They go to skeletal muscle, adipose tissue and the heart, and the fatty acids are taken out. They're liberated from the triglycerides and rapidly taken up by these three tissues. The particle that's left is now much smaller because most of the triglyceride has been taken out of it. And it goes to the liver and drops the cholesterol off. So the cholesterol we eat in the diet goes to the liver, and it tends to reduce the synthesis of cholesterol in the liver. The liver gets inundated with fatty acids and cholesterol from all over the place. Chylomicrons, HDL LDL particles. There's a system to regulate the mass of cholesterol and triglyceride in the liver, and that's the VLDL APOB system. VLDL particles have a molecule of APOB100 which is longer, twice as long as APOB48. And that gives the structural integrity of the particle. Just one other thing I'll get to. And that particle removes triglyceride and cholesterol from the liver to maintain the balance in the liver. The triglyceride, just as in the case of chylomicron, gets dropped off in adipose tissue and skeletal muscle and cardiac muscle. So the VLDL particle gets smaller and more cholesterol rich, and it eventually becomes an LDL particle. And an LDL particle is a cholesterol rich particle with relatively little triglyceride in it. When we measure the cholesterol in the blood, the total cholesterol is the cholesterol in the VLDL particle, the LDL particle, and the HDL particle. The good guys, quote, unquote, the non HDL particle. Cholesterol is the mass of cholesterol in VLDL particles and LDL particles. Now, isn't that enough? Isn't that all we really need to know? It tells you a lot, no question. But I'm going to give you two people. They both have the same LDL cholesterol. Their cholesterol is 125. One of them tends to have larger LDL particles, one of them tends to have smaller LDL particles. In order to carry the same amount of cholesterol, there got to be more little ones than big ones. So their LDL cholesterol is the same. Is there any difference in their atherogenic risk? And the answer is yes, yes, yes, and yes. The one with the increased number of particles has higher atherogenic risk because any cholesterol in the artery only got there within an apob particle. It doesn't just float in. It gets there within an apob particle, either VLD or ldl that gets into the arterial wall and gets stuck there. And that's the cause of atherosclerosis. There are lots of things that contribute to multiplying or diminishing the cause, but that's the cause sticking of an apob particle within the wall. And because cholesterol gets in there within the particles, knowing the number of particles is more important even than knowing the cholesterol level.

Alan, when did that historically become apparent? If we take a step way, way, way back, if we go back into the 1950s, Ancel Keys was potentially one of the first people to utilize the then nascent assay for measuring total plasma cholesterol. So to your point earlier, that number, let's say you measure 200 milligrams per deciliter. That's simply telling you that that's the sum total of cholesterol in all of the lipoproteins. And what Ancel Keys and others observed, and this was, by now we're probably into 1957, 1958 was, hey, if you stratify people at the bottom 5% and people at the top 5%, so that might be people whose total cholesterol is less than maybe 100 milligrams per deciliter and people whose total cholesterol is more than 200 milligrams per deciliter, there's a stark difference in their mortality, or rather in their risk of cardiovascular disease. And became very interesting. It turned out that there wasn't a great way to predict that. So the amount of cholesterol a person ate did not seem to predict that. But nevertheless, other dietary factors Saturated fat intake, for one, seemed to predict that difference. It would be, what, maybe 20, less than 20, 15 years later that Friedrichson, Levy, Lees and others would start to fractionate those lipoproteins and realize that, well, actually, there's different versions, as you alluded to. There are some that have low density, there are some that have high density. And I don't know exactly when it became clear just how nuanced that was, that APOA is on one and APO B is on the other side. But when did it become clear that there was a discordance between the cholesterol concentration in the LDL particle and the number of LDL particles?

I think Bob Lees, in 1971, he had a paper in Science. He was measuring LDL APOB, the number of LDL particles, and he showed it had no clear relation to plasma triglycerides or to the cholesterol. I mean, there was sort of a relation, but it wasn't exact. And Ron Kroos, of course, and his colleagues and I did at least one paper, one or two papers with him at the very beginning were the ones who actually showed, in the John Goffman tradition from Berkeley, showed that there were important differences in size, and these related to differences in the amount of cholesterol mass per particle. So Ron Krauss and the group there and then a whole bunch of other people and deserve, I think, the credit. But we're way back in the late 70s, 1980 was my first paper, clinical paper, showing with Peter Quiddervich, the late Pete Quiddervich from Hopkins. We looked at a bunch of patients with coronary angiography, and we compared people with clean coronaries like clean to people with diseased coronaries like disease. There was a little difference in triglyceride, there was a difference in cholesterol, but there was a marked difference in apob. And that was the first, I think, clinically solid observations, along with an Italian group that had much close to the same observations, slightly ahead of us, for that matter, saying that, look, particles could be more important than cholesterol. And it seems like forever since then, 1980 to now, trying to, and I think largely now succeeding in developing the evidence that you can say incontrovertibly, particles more than cholesterol. Now, that hasn't moved the American guidelines, but on the evidence side, there are a handful of studies that show that non HDL cholesterol may be equal to apob. There are more studies that actually show APOB is better, but we developed a way of looking at it. Called discordance analysis to identify people who had a high non HDL cholesterol, which is the total cholesterol in the APOB particles, but a low apob total number of particles versus low non HDL high apob. So if you're a cholesterol maven, you gotta bet on the one with the high non hdl. If you're an apob aficionado, you bet on the one with the lower non hdl, higher apob. They all show it's apob.

Now, is the argument, Alan, if one argues that in the at least equivalence of, if not superiority of non HDL cholesterol as the superior metric, or at least equivalent metric, is it because you're arguing a different mechanism of action, or does everybody agree on the mechanism of action and they're simply saying measuring cholesterol content is a good enough proxy for counting the number of particles?

Nobody suggests there's a different mechanism. There are some people who argue that VLDL particles are more atherogenic than LDL particles, and I think they've got a long way to go to prove it. What people did argue was there were problems with the APOB assay and that it costs money. And the reality is the APOB assay was standardized back in 1994. The measurement of HDL cholesterol is not standardized. The measurement of LDL cholesterol, not standardized. The measurement of triglycerides, not standardized.

In terms of the cost argument, because I actually had that argument with a physician as recently as 3 months ago who accused me of getting APOB on patients as a way to upcharge them, even though I don't make any money on labs. But I actually called the lab that we use and said, what's your cash price for APOB? You want to know what it was?

Please.

$2.50. It's a real money maker.

Yeah, yeah. You know how much it is? In my hospital, it's $2. This cost argument has been used without documentation as a killer argument. And there were labs that charged way too much. Welcome to America. In general, your charges are higher than ours in my little country, but that's a function of how much somebody's billing, not a function of what the assay costs. And apob, it really ticks me off because if you take India for a moment, or almost anywhere, if a doctor gets a report now, he gets total cholesterol, triglycerides, non HDL, C LDL, C HDLC 5 numbers. Do you think he actually looks at any of those numbers? He's trying to do a good job. He does. But let's say the triglycerides are high. Can he do anything with that? Nope, because everything is based on LDL C. So he's got, in reality four numbers that are doing nothing.

Let's explain that to people, Alan, because you and I know the ins and outs of that very well. But I think most people here don't understand the difference between the calculated and measured ldl. So let's start with that, and then let's talk about how VLDL has been estimated, and let's bring this all back in terms of some other work you've done, which is understanding the role of triglyceride in apob. So let's start with the basic. You go to the doctor, you get a set of labs done, and the LDL number comes back at 140mg per deciliter. Is that actually what it is, or is that an estimation?

That's an estimation. It's almost always a calculation. And there are at least eight different methods to calculate LDL cholesterol. So if there are eight different methods, they don't all give the same answer, or you wouldn't have eight different methods. LDL cholesterol can also be measured directly. That assay has never been validated in disease patients, and no one has ever published a paper showing that it's more accurate in terms of disease identification than calculated LDL cholesterol. And yet people have paid good money for that lab test. There's no question that the number of LDL particles is a more accurate index of risk than the LDL cholesterol. The VLDL cholesterol is a cholesterol that's in the vl. Very Low density lipoprotein particles, the particles that come out of the liver. That cholesterol is atherogenic. There's a lot of triglyceride in that particle. So the people who measure triglycerides say, well, the triglycerides are high. That must be the problem. And there's no question that people with high triglycerides are at increased risk of heart disease. But the people with the high triglycerides who are at increased risk of heart disease have a higher number of LDL particles and VLDL particles. It's the particle. And when you're measuring the triglyceride, you're just measuring a blob of liquid in a bunch of particles, and you need to know the number of them. So it's an important number in the sense of if you're a lipoprotein guy trying to figure things out. If it's extremely high, it increases the risk of pancreatitis. But I haven't seen any solid evidence that triglyceride itself is pro atherogenic. What's atherogenic is the cholesterol inside the VLDL particles. It's the number of those particles that get into the wall. Now there's a complicating reality because in general, all I need to know is the APOB. But there is a disorder called remnant type 3 dyslipoproteinemia. And that's a very specific, highly atherogenic condition that manifests with high triglycerides, high cholesterol, but get this, you know, low apob. So when I measure my lipids in apob, I can recognize that. But if you don't measure the apob, and this applies to most of the people who are listening to this podcast, if they go to see their doctors, that condition can't be diagnosed.

Can you explain that condition? Walk us through what's happening Pathophysiologically, how does that person have high cholesterol, low apob, high triglyceride?

I'll try. The normal metabolism is the VLDL particles get broken down sequentially as the triglyceride is removed and they get converted to LDL particles, some of them are removed by the liver along the way. In type three, that process breaks down. And for reasons that are not well understood, under age of 30, 35 or 40, people develop high triglycerides and high cholesterol. Because the vldl particles aren't being broken down to LDL particles, that process stalls. Those particles circulate a long time in the blood, and while they're circulating, cholesterol gets deposited into them. So they become very cholesterol rich, like really, really cholesterol rich. And those people have a very high risk of coronary disease. Peripheral vascular disease. It's a commoner syndrome than familial hypercholesterolemia that gets day and night pressed day and night. This one is easily treatable almost all the time. FH needs to be treated much more challenging. But type 3 cannot be diagnosed in most patients in the United States because APOB is not measured. The technology that we used to use to diagnose it, it's all gone. Nobody uses it anymore. It's old fashioned. But it can be diagnosed based on the triglycerides, total cholesterol, and apob. There's a formula that we devised so we can recognize those people and say, hey, you baby, 38 years old, but you got a big problem here. And with treatment, we can take your big problem away.

And so the phenotype of that patient is that they have relatively few particles, but they have so much cholesterol because the VLDLs are so large and so cholesterol full.

Yeah, that's right.

So what is it, given the relative lack of particles that makes that such a dangerous condition?

I'm not sure we know, really. Compared to normal, there are 40 or 50 times as many of these particles. But I'm not sure I understand why it's so dangerous. In terms of particle number, I do know that it means you can't just use the APOB when you're trying to make a diagnosis. When I follow a patient, I really just look at the apob. For normal patients that I'm treating with statins, I only have to get one number right.

It seems that those particles are more atherogenic, sort of in the way that an LP is more atherogenic. So if you did a thought experiment and you said you take three people who all have the same APOB concentration, but they could have three very different predicted risks. If one of them has a very, very high LP concentration, another one has a normal phenotype, and another one has a type 3 as we've just described, the first and the third have a much higher risk, suggesting that on a particle per particle basis, they have more.

Atherogenic particles or particles that contribute to clinical events. Yeah, I agree. For my practice, I measure a lipid panel, I measure an LP in everybody and I measure apob. I measure the LP once when LP is high, but APOB is normal. LP may not add that much to risk, but when you got two of them, it's a double whammy. And that I use as another piece of information in trying to frame for the patient the potential futures that they face.

And so, Alan, what is the treatment for the patients that are type 3s? These patients with many cholesterol rich vldls, despite a normal APOB, statins and or.

Fibrates, they usually respond very well.

The fibrates in that patients remind me their triglycerides are normal or elevated.

Triglycerides are elevated. They're VLDL particles, so the triglycerides are elevated. We made up a algorithm that's, I think it's the APOB app. APOB app where you could plug in the total cholesterol, triglyceride and APOB and you get the diagnosis of any of the atherogenic apop. Dyslipoproteinemias.

Yeah, that was a fantastic app. I can still use it on my desktop, but for some reason it stopped working on the phone. Am I the only one to have that issue?

It stopped working on Google and app, but it's on the web. It's directly available from the web.

What's the URL?

Www.APObapp.app Right.

App, I think.

No, I think it's just APOB app.

Okay. I saved it in my browser because I remember it was a counterintuitive place. I just actually relied on it. I looked something up a month ago on it. So it's a great little. And it walks you through all the diagnostic steps.

Part of the argument against apob. People say it makes things too complicated. If I explain to a patient that they've got a lot of bad particles, cholesterol particles, they get it. When I review the results of how well somebody's doing on statins, if their triglycerides were high to begin with, they're unlikely to normalize. The HDL cholesterol's unlikely to normalize. Their APOB is good. They're good. That's my target of therapy because it's the total number of atherogenic particles. And Nordiskaard had a lovely paper recently in. Was it Jammuk Cardiology on a discordance analysis of non HDL cholesterol and apob, showing that APOB was a more accurate index on statin treatment than on HDL cholesterol. It's a lovey paper.

Yeah. I mean, frankly, I find it much easier to explain to patients what APOB is than to explain what non HDL cholesterol is.

I do, too. I mean, a non number is hard to explain. And it's interesting to me that with all the emphasis that so many of the lipid guidelines have put on non HDL cholesterol, they all still say LDL cholesterol. And the American guidelines clearly state APOB and non HDL are better than ldl. But the world has remained. It's a phenomenon that I don't really understand how resistant the lipid world has been to change. But I think it's important to understand because we'll understand things like Afghanistan and the financial crisis in 2008 and a whole series of bad decisions by good people thinking as hard as they could. But when everybody in the room has the same opinion going in, it's a bad way to solve problems.

Yeah. I mean, are you optimistic? I mean, Is this just a question of time? I mean, in 10 years, will kids in med school be learning about APOB instead of LDL?

I'm pessimistic. Europe. The 2019 guidelines were very pro APOB. The evidence from Mendelian randomization, like the newer technologies, Mendelian randomization, they've just been slam dunk for apob.

Let's explain that to folks, because I want to talk about the causality of this, and this might be the perfect way to actually explain the causality of APOB in the context of this tool. So can you explain to folks what a Mendelian randomization is? People see this all the time in studies, but I don't think it's entirely clear for the average person what it means.

I'll try, okay. It's not my expertise, but I'll try. The conventional ways of taking things apart. With prospective observational studies like Framingham, there's a limited amount of the certainty of your conclusions because of confounding you can't deal with. You take measurements at age 20 and you follow someone for the next 30 years. Well, a lot of things change in the next 30 years that you don't have a handle on. Your inferences are probable, but not causal. What Mendelian randomization allows you to do is to come a lot closer to causality because, for example, you can identify groups of genes that are associated where changes in the gene are associated with a little lower cholesterol or a little higher cholesterol. And when you lump together a bunch of those different genes that that can have different makeups because you can change the makeup of a gene pretty easily. You can see fairly substantial differences in cholesterol. So what you've got is information on somebody that's fixed at birth. And you see, is that associated with a difference in outcome? You've gotten rid of a lot of stuff in the middle. And what a number of Mendelian randomizations have shown is that APOB includes all the information in triglycerides, LDL cholesterol, and even HDL cholesterol. It sums them, which in the sense of VLDL and ldl makes perfect sense. So there are caveats in Mendelian randomization. You can't just push a button and say, give me the answer. But George Davy Smith, really arguably one of the founders of Mendelian randomization are not arguably. He was. He's the author of a number of the Mendelian randomization saying APOB incorporates and therefore beats triglycerides and LDL cholesterol So that's a huge level of information that isn't even mentioned in almost any of the guidelines.

Yeah. So let's make sure people understand everything you just said, because you said a lot of things in there. When you prospectively follow a cohort, the way the Framingham cohort was followed, or the Framingham offspring, or the MESA cohort, or any of these cohorts have been followed, you can take a bunch of people and you could measure their APOB or their LDL C or whatever metric it is that you are trying to determine if it, in fact has a causal relationship to the disease of interest. You can follow them over decades. And you would demonstrate, as has been demonstrated, that the people with higher B, higher LDL C, higher non HDL C, and lower HDL C all have a higher risk of developing atherosclerosis over time. But it's hard to say that that's causal just based on that information, because over the ensuing 20 years that you follow them, they are free to make other choices that may impact those variables of interest and other variables. So the Mendelian randomization attempts to get around that by saying at the time of. I was going to say birth, but really at the time of conception, we all get randomized to a set of genes. We get assigned a set of genes. I guess they're not perfectly random because they come from our parents, but for the purpose of not changing, they are indeed a random assignment that is fixed. If we can identify which genes map to which phenotype, and we can figure out the genes that map to the phenotype of our interest, namely driving up or down a variable of interest, such as apob, then we don't really have to worry about the confounders that occur in between, because the genes can't change. Just to put a bow on that, basically. Now, when you see a difference in outcome, it's much more likely to be causally related to the phenotype of interest because the gene has not changed, that underlies it. Now, what are some of the ways that we can get tripped up with Mendelian randomization? I mean, there's some pretty big ones.

Yeah. Before we get there, HDL cholesterol was the rage. Okay. The total rage, because the epidemiological evidence couldn't be clear.

In fact, it was four times more clear. My recollection was that Framingham demonstrated low HDL C was four times more predictive of cardiac events than high LDL C. Am I remembering that correctly?

Not sure. It's that Multiple. Yeah, but it's multiples. And it turns out, as we know now, at least from the CTP inhibitors, that, that you can't manipulate HDL and change outcomes. And that's one of the elements of demonstrating an overall causal relationship. And the Mendelian randomizations show HDL is not causal, whereas they show APOB is and cholesterol is too.

By the way, those are two very important studies, Alan. I mean, and both of those have been in the last 10 years.

Yeah, it's a incredible technical advance in being able to examine questions and look at numbers of people that are. Would be unimaginable in conventional studies. The Mendelian random, they're always talking hundreds of thousands of people because they've got these huge data banks with genes and those numbers get you around the confounding of things. You have huge numbers. But it's like any methodology, no method is perfect. This one can mislead you too, particularly when you've got a sequence of associated variables. For example, people showed using Mr. That triglycerides were causal or associated with increased risk, but when you took into account the non HDL cholesterol or the apob, it disappears. So when you've got a linked metabolic chain, you've got to be careful that you've gone to the end of it. You've got the real actor, not act one leading to that. You've got the real Persona dramatis.

Which is why it's surprising that HDL didn't, at least at the first order, demonstrate causality. Because there's no doubt that phenotypically the high triglyceride low HDL phenotype is so associated with metabolic syndrome that it makes up two of the five criteria.

That's an incomplete description. That's like you describing yourself as six feet tall. I wish. And not giving your weight and letting me guess your bmi. You cannot characterize any phenotype without the apob. It really drives me around the bend when people speak saying, I got somebody because I got their triglycerides and their hdl. Well, I say, okay, what's their apob? How can you pretend you've evaluated the system when you haven't counted the number of atherogenic particles? Because they could be normal, they could be high, or you could have a type 3. They don't know and it's not a phenotype. There is no phenotype without putting an APOB in there. They're lipoprotein particles. They're disorders of lipoprotein particle Metabolism, of course, the triglycerides and cholesterol are important, but my analogy, I didn't do a good analogy there, but it's so fundamental that it drives me to distraction as to why you wouldn't want to know. A core element of knowledge, but it doesn't seem to bother many of my friends.

You walked through the pathophysiology of how the apob bearing particle wreaks havoc in the artery wall many, many years before we see clinical events. And you also mentioned that there are other factors that can amplify or exacerbate that. I can't remember exactly how you said it, but that was the gist of it. Well, two of those things that are widely accepted to exacerbate risk are smoking and hypertension. In fact, smoking and hypertension probably carry a greater risk for atherosclerosis than apob. Or is that not the case?

It all depends the way you think about it. Because if you just say, what's the risk? Somebody with hypertension faces, they have high risk, no question. But then you say, what is hypertension? The last 30 or 40 years, there have been almost an infinite number of basic science studies on hypertension. And when you were in medical school, and even before that, when I was in medical school, we talked about pathophysiology of hypertension. And what strikes me is we don't talk about the pathophysiology of hypertension anymore, but the basic science goes on in rats, just healthier than ever. And there isn't anything I know of that's come out of that basic science that's been clinically useful in the last 30 years. The drugs we use, we use them because they work. So what is hypertension? It's a higher blood pressure than we should have. And where is the disease that produces that higher blood pressure? Is it resistance? We don't have a clue. Okay, we don't have a clue. And it strikes me it's the same thing as much of the debate in lipids about apob or the drunk looking for the key under the light, because this is where the light is, not where he lost it. Everybody who's anybody has the same viewpoint. My bet is it's in the proximal aorta. My bet is that it isn't that complicated. We lose elastance in the proximal aorta, and that's systolic hypertension, thank you very much. What could accelerate that process?

What's the mainstream view? That this is renal?

When I read hypertension, I get lost because I get page after page after page of peripheral arteriolar tone and very complex metabolic studies and very sophisticated animal models. There's some renal left. It's a miasma for me, an absolute miasma.

I hadn't heard about the proximal aorta, so say a bit more about that.

Well, this is me. The proximal aorta is elastic. And if you look at a flow curve, a hydrostatic pressure curve, when we're young, it's rounded because as the left ventricle ejects blood rapidly into the aorta, the aorta expands, so it absorbs some of that energy. You know that Windcassel that they mentioned in school? That's not that big a deal. But the energy is partially captured, partially regained. But the wall isn't battered. The wall can give way. Me personally, just in the middle of my brain, imagine that if those elastic fibers start to go, then the wall's stiff. So now when the left ventricle ejects blood, the pressure goes up more rapidly and it falls more rapidly and diastole. And that's why you get systolic hypertension with normal diastolic pressures. So my bet would be if I was not the age I am, I would be looking at factors like cardiac output, again, which used to be way back when, or factors that alter the behavior of the proximal aorta as much as something that's, to me, pathophysiologically, much more likely to be involved. So once I got hypertension, okay, then I've got a driving force to push particles into the wall.

And so you think it's the actual increase in the pressure of the plasma.

And the response of the wall? I think there are responses to the wall. The wall thickens up, it gets harder for particles to go through.

Does it also damage the endothelium? Do you think that plays a role?

That's right. I don't understand endothelial dysfunction. It's more a language thing to me than it is a reality. I know the endothelium is critically important. It functions abnormally, and that's endothelial dysfunction. How that fits into the overall thing, I don't know. My bet is APOB particles are part of the process of inducing endothelial dysfunction. But I don't know that clearly, experimentally.

So going back then, to the question at the top, does it make sense to even compare hypertension to apob? They both seem to play a causal role. Is one more causal than the other? Or is that a silly question because they're not Binary and static.

I think that's not the right question. I think our blood pressure goes up as we age. I mean, hypertension involves so much of the population. It's not clear to me what the word disease means. The prevalence as we age is so high that to me it becoming almost a aging process because we're lasting a lot longer than we were probably designed to go. So you have this repetitive injury to the proximal aorta. It gets a little progressively less able to deal with it. So by the time we're 50, what, percent, 60% have higher blood pressure. I mean, the figures are staggering.

Is it really that high?

I'm not sure. Don't quote me on that. But it's high, high, high.

But doesn't APOB also rise with age?

It does rise with age, but not that much. When we look at people at age 35, we can pretty accurately categorize the group they belong to at age 35, not that they won't change someone. So if you're high at age 35, you got about a 95% chance of staying high. 5% will go out of the high zone. They won't go low, low. So if you're high at age 35, I wouldn't bet anything's going to move you down. That's why I think it's such a good signal for when we should start thinking about treating people. And if you're low, some people go from low towards high, but the majority don't and we keep following them. But if you're high, no, we've published a fair amount of this. If you're high, it's not 100%, but it's about 90% that you're going to be high.

Is there a gender difference? At least clinically? I seem to see women, as they go through menopause, experience dyslipidemia that men wouldn't experience over that same decade or even five year transition.

I think there are changes and APOB goes up with menopause. I'd like there to be more data. I think part of the reason it's held APOB back is that people didn't measure it. So they were sort of. Well, what I measured has to be important because I can't answer your question. Hopefully more data will be coming. But I agree with you. People can change at the menopause. So I'm not saying we don't keep looking at people, but when you have somebody at age 35 to 40 who's high, the odds are high that they're Going to stay high.

Are we doing a better job treating hypertension than dyslipidemia?

I have no idea. The incidence of coronary disease is going up in the last five years and that's despite statin therapy and that's the obesity, diabetes. So I think we've been too quick to congratulate ourselves at how well we're doing. There are many reasons that treatment is not succeeding as well as it should. And I think the complexity of the lipid phenotype, of the lipid model is part of the answer. It's easy for me. I get the APOB where I want it to go.

Yeah. I mean, an explanation for your observation would be if in the last five to 10 years the incidence of atherosclerosis or major adverse cardiac events is rising despite the advances we have, you would argue, or could argue that if we're measuring LDL C and that's our proxy for treatment. But as dyslipidemia is growing in the metabolic context, meaning if you have more medicine and more insulin resistance and more type 2 diabetes, we know that those phenotypes are associated with greater discordance between APOB and LDL C, suggesting that you'd have a greater and greater portion of the population that is being undiagnosed or being under diagnosed because you're treating their LDL C and you believe that it's lower than their risk actually is because their APOB is higher. I know you know what I just said. I hope the listener understands what I just said.

Yeah, what you just said was important. It's another example, an unfortunate, sad example, that trying to quantify lipoproteins based just on lipids is not adequate. You're not capturing all the information that you should.

So let's talk about smoking for a second. Do you have a sense? I know it's again, not the thing that you study day and night, but what is it about smoking that drives risk of atherosclerosis so much?

Truth is, I don't know. It certainly does and it drives a treatment decision. I think smokers are wonderful human beings who deserve to be treated as human beings. But I don't think that if you choose to continue smoking, that should prioritize you for statin treatment. Whereas based on my concern about people who have high APOB levels, they should be treated because of their longer term risk. Smoking gets you up the category to have your life saved. So bad behavior gets you closer to having your life saved.

Is that a Canadian thing?

No, it's an American Thing, it's in your risk calculator. If you're a smoker, your risk goes up.

Oh, I see what you're saying. Not based on a coverage issue, but simply based on a change in prioritization.

That's right, because everything is risk. Bad behavior increases risk so you get more medical attention. Now, I think bad behavior is our responsibility to help people deal with. But I don't think it should put you to the head of the line for preventive therapy.

I mean, I would argue there is no line. Anybody who wants preventative therapy should be getting it. Are we resource limited on that front?

We're not giving patients the information that they deserve. We published a paper in circulation, I think it's 2019, which I referred to before, looking at what are the costs of delay of intervention starting at age 35, 45 and 55. And if you're non HDL is low, yes, you'll get some gains starting at 35, but it's not a lot. It's actually quite small. Your gain is in the people with the high non hdl. We don't have to be giving pills to everybody at the age of 35. If we use our physiological and epidemiological knowledge. There's about 20% of the population is that evident high risk and they should consider it. And part of the information they need to know is how much do I gain now? How much is gained now versus how much is lost by waiting. And that's why these methodologies of calculating benefit are so important.

Do you have a sense, Alan, of what fraction of the population has relatively normal APOB? Relatively normal 12 triglycerides and yet has accelerated atherosclerosis through some combination of yet to be identified polygenic risk factors. So you see atherosclerosis that runs in families and there's not an obvious cause. Right? They don't have FH, they don't have LP. And frankly their APOB is harboring around called the 50th percent of all the population. But they disproportionately get afflicted young. So they're all having first events before 60.

I don't know. I think that's where research needs to be done. And I would look at factors that affect the trapping of the APOB particle within the arterial wall. I think Kevin John Williams from Philadelphia, they've done amazing job, amazing job in putting this all together.

And wouldn't that factor into our decision making then? I mean, would we, if we had two people who were both say 40, just say they were identical in terms of lipids and lipoproteins. They both had the same blood pressure. They were both not smokers, et cetera. But one had that family history, the other did not. Are you treating them different?

It would factor into my decision. If the APOB is actually low, I'd be less inclined to let it influence me.

Let's say they're both at the 50th percentile.

If they're 50 60th percentile, that makes me more antsy. The higher they are, the more antsy I get. A lot of these decisions at the individual level actually aren't that difficult when you're speaking to a particular patient because they have their own objectives and goals. If we give people medications, our medications have risks associated with them. I don't think we fully understand the relationship of statins and diabetes. I don't think we do. So statin therapy is, like, amazing, but it's not a no cost, so. But when you talk to a individual human being, at least at this stage of my career, it's been easier to make clinical decisions individually than to write rules for groups.

Yeah, there's no question about that. Let's now look at yet another predictive tool, which is the Coronary Artery Calcium Screen Score. So maybe tell folks what a CAC is. I suspect a number of people listening to this will have had it, but enough will not have. So it's worth explaining what the test is.

Coronary calcium is an important step forward in cardiovascular imaging, and it's a process where you can accurately and pretty safely determine, using X ray techniques, whether there's calcium bone in the coronary arteries. And calcification is a feature of advanced atherosclerosis. There's very strong evidence that people who have coronary calcification are at higher risk of a heart attack or stroke than people who do not have coronary calcification. That's an important piece of information, but there's several facts you also have to appreciate. First, the frequency of a positive coronary calcium goes up as we age. So does the risk of disease. So by the time a man is 60, all American men are at high risk. According to your current guidelines, women are five to 10 years later. So at the point where the test is most commonly positive, I don't even need it, because that person. We were talking a while back about the natural history and stereo and stuff. That stuff is for real. Our arteries in the majority of us have become substantially transformed in bad ways by the time we're 60. Now, in people who are younger than 60, can this give you extra information? If you're on the cusp of saying, should you be treated or not? And I think the answer is yes. If you or the patient says, I want more information, I'm not convinced. I personally are in a situation where I should be taking medications and you have a positive coronary calcium, I think that can be extremely helpful. It's what's taken as the corollary that if your coronary calcium is negative, you're okay. That's the problem for me, from my knowledge and interpretation of literature and the pathological studies, coronary calcification is a advanced disease, means advanced diseases present. When people have a heart attack, they don't just have one little area of their arteries that are abnormal. That's the area where the plaque broke or the endothelium eroded, but the artery is diseased, and there's a chance of an event, a sonometer down or a sonometer closer. So if somebody has a high apob, the fact that their coronary calcium is negative doesn't mean they don't have a lot of disease and that the disease isn't developing at a rapid rate. It could well be there. So there's an argument saying, well, if your coronary calcium is negative, nothing's going to happen to you in the next five or 10 years, and maybe but the disease is going to develop and we can't make the disease go away. We can modify the effects of the disease, modify the consequences. But when we talk about, I mean, LDL cholesterol and APOB levels are now so low, but you still have an artery that's destroyed, you're going to have a substantial number of events. John Wilkins and Don Lloyd Jones from Northwestern, they have a paper in Jaha and it's a terrible paper to read because it's so complicated. And I wish they hadn't presented it in as complex a form as they did. They're friends of mine, so I can criticize them, but within it are the observations that starting to treat, waiting for coronary calcium is a bad idea. So I'm a conservative physician. It may not be my politics, but I want to protect patients. Give them the option because it's the patient's choice. Of course it is. Give them the option to have the best outcome possible when they appear to be in danger. So I wouldn't use a negative coronary calcium to change my clinical decision when I have a high APOB or another cause of vascular disease present. So I think it's a Good test, but relatively limited utility.

For me, the way I've talked about it, maybe even on the podcast, but certainly the way I talk about it with patients every day it would seem, is I describe it as a two by two matrix. So we think about how this test is helpful in people who are young and people who are old. Now, I usually use 50 as the cutoff. It sounds like you use 60 as the over under, but let's just say it's somewhere in that sixth decade. And then is it zero or is it non? Zero. I agree with what I'm hearing is your assessment, which is in the older patient, the positive score is not very informative. So when I have a 70 year old patient who's calcium score score is 50, it's sort of like, so what? You're normal. That doesn't tell me much. Conversely, when I have an older patient whose score is zero and they're adamant about not getting treatment, it becomes an easier decision to accept because you can say, well gosh, you're pretty fortunate to be 73 years old and despite having an APOB of 140 milligrams per deciliter, your calcium score is zero. There must be some other protective mechanisms in you.

The bounds of our knowledge are really quite limited and it's important that we admit that to ourselves and to our patients as well. I look at it maybe a bit differently. I say, yes, I would say the same thing. And I would say, yeah, but even.

So, to me there's an asymmetry which I want to come to in a moment. On the flip side of things, the young patient who has a positive calcium score, really, that's a four alarm fire.

That's a no brainer.

Regardless of the APOB, if you're under 50 and you have a speck of calcium in your coronary arteries, even if it's a low enough spec that it would predict a 10 year risk of 4%, that's still utterly unacceptable.

If it's positive, it's positive. It's only going to go up.

And more to your point, it's what it says about the milieu of the entire system. So that might be the one area in the middle of your left anterior descending artery where you're at such an advanced stage that you've already laid calcium there. It's sort of like looking at the concrete that's been poured over Chernobyl and trying to infer what's going on in the 10 miles around Chernobyl. It's all bad.

Yeah, that's correct.

Where I find the most challenge, there's the groupthink that says if a person's calcium score is 0, no treatment is needed. And this kind of gets back to your paper from JAMA, I think 2018, maybe 2017, looking at the 30 year risk, the causal model which I want to come back to, you mentioned it very briefly at the outset, but it's so important that I want to now use this as a jumping point to go there, which is you take that 45 year old person who you expect their calcium score to be zero, it is zero, but their APOB is higher than it should be or you would like it to be. That calcium score of 0 hasn't really added much information to my decision making.

No, because your time horizon is different. We use 20, 30 years. If you're 45, you want to get to further than 55, your career, your children, your enjoyment of things. Surely you're not just planning to age 55 and you should be thinking well what am I going to be like at 65 or 75? That's reasonable. And it's also by taking it out to that you can get to numbers that are really meaningful. When somebody's at a 30% chance, like one in three, that's a number most people can understand and it starts to become a truly a meaningful number for an individual. When somebody's at 7.8% risk, that's tough to absorb in any way that means something. When you're in one of these higher risk groups, it doesn't mean you're doomed, but you're in company with a lot of folks who are and we can say that absolutely accurately. Yes, it's limited. It may not be. You make your bet, I might even.

Be more risk afraid. You're saying that if a person's 10 year risk is 5%, how can you get somebody excited about a 5% event in the next decade? If somebody's 50? So somebody's 50 and you say you've got a 5% event risk by 60 and you're saying, well the two things that are wrong with that are one, you shouldn't just be thinking about being 60, you should be thinking about being 80, which I completely agree with. We expect you to live a minimum of 30 more years, if not more. And secondly, you're saying, well, 5% is not that much to get excited about. Well, let me turn the table. I want you to pretend you're 50, Alan. And I say to you, Alan, there is a 5% chance that in the next Decade, you will die on a commercial plane. How would you change your behavior as a result of that?

I don't think I would.

You wouldn't stop flying?

No, because there's a 95% chance I won't.

Okay. I think 5% is a bigger risk than people realize.

I get it. It's bigger than one in a thousand. Okay.

Well, not only that, in the case that I just gave you, the treatment would be don't fly for 10 years, which is a real impediment to your lifestyle, versus the treatment to lower lipids, which is far less of a burden and the risks of it can be combated.

Let me come back to you. What are the confidence intervals on that number of 5%?

Well, that's a good question. Let's say I can tell you it's plus or minus 2%.

How do you tell me that? Where's that number?

Have you ever seen that number at the population level? Let's pretend I'm able to tell you that at the population level.

You've never seen it? I've never seen it.

Well, we can't do it at the individual level, of course.

No, no, no, no, no. I'm saying what are the confidence intervals for that prediction? Have you ever seen them published? And the answer is nope. I'm not sure how much of your audience does confidence intervals and this kind of thing, but as scientists, for any result, we get the result and we get the range of possible results, then we know how accurate the prediction is. If the confidence, let's say it's 5% and the confidence interval is 4.5 to 5.5, well, you're there. If the confidence interval is 0 to 70, which it could be, by the way, maybe it isn't. So I'm saying we got an industry that captured clinical care that doesn't include air.

Well, why is that? That seems, now that you mention it, almost impossible to believe.

Please reassure your audience I'm telling the truth. Because we've become less critical. This process of forming opinion in medical care that's appeared over the last 30 or 40 years has damped down the essential element of science, which is challenge different viewpoints, the contention of ideas, the creation of an experiment to say this is the right way or someone with a different view creates a different experiment. Science is a democratic ick where legitimate, contending legitimate views have equal opportunity, where differing views that are legitimate have a chance to contend. I'm not saying every crazy theory has equal. It doesn't.

So that's exactly what I was Going to come back and ask you, Alan, is who is the arbitrator of what's a legitimate differing hypothesis?

In science, it's called the experiment. That's what's different about science. An experiment is done to test a hypothesis. If the hypothesis is sustained, you can continue to hold the hypothesis. If the hypothesis falls, then you must reframe your understanding. And we do experiments to gain understanding. It's our tool, but it's not as easy as it sounds. Our experiments are. The methods could become complicated. The methods we use to analyze them statistically can become complicated. Are complicated, and the conclusions we draw from them may or may not be correct. Error occurs. When I was a medical student, the major medical meeting of the year was at Atlantic City, and the big professors would contend the elephants, you know, and there were different views and they would argue them out. And it was a contentious, open battle. Evidence based medicine came along, which has lots of pluses in which it says we should use randomized clinical trials as part of our knowledge base. But we developed the belief that it was easy to assess knowledge. And it's not. It's not easy to assess experiments all the time. Some of them are straightforward, most of them actually aren't. Most of them, there's uncertainty involved. And it's important that we acknowledge the uncertainty and say, well, maybe there's another way of looking at this that's even better. But we developed tools like statins. There's very good evidence, can save lives. So this process of consensus came along saying this is too complicated for regular doctors. And it is. They don't have the time to analyze all the evidence. So we'll analyze the evidence for them and write it out in a way that tells them what's the best we can do now. And there's a lot of good in that, but there's potential weakness. And I think where a lot of the weakness is happening, because you get a view that becomes the conventional view and it hangs on longer than it should. Science is about change. If we're still saying the same things we said 30 years ago could be a problem, because we should have learned how to say it better, more accurately. In these consensus conferences, I don't know if your listeners appreciate the recommendations are unanimous. The Supreme Court isn't unanimous. Often there's a majority view. The minority says this, that, and the other thing. And it can turn out that the minority. Over time, we see the wisdom in the minority. Any process that has unanimous recommendations has a weakness. Any process where the decisions become larger than the individuals who Propose them has a weakness. Like the recommendations are the American College of Cardiology, American Heart association, blah blah, and about 30 other groups. There are actually 100 people. They're good people, but there's only whatever number it is, 100 of them. But by cloaking it in the anonymity of the group, they become impervious to criticism, even when there's obvious inadequacies. And that's not science. If you write a paper, I could do an experiment and try and overturn your paper or confirm it. But when you have the guidelines, they're the judgment that's cast in stone till the next group of guidelines. Depending on which people write them, it can influence what you see.

I completely agree with that. I would add even more complexity to it, which is if you go through what the critical steps are in the elucidation of knowledge, the first, presumably would be the formation of a hypothesis. The second might be designing an experiment to test that hypothesis, then conducting that experiment, analyzing the results of that experiment and interpreting it. Again, I'm oversimplifying a little bit, but these are quite discrete steps. And as you pointed out, any one of these steps offers infinite ways to do it wrong and a relatively few ways to do it right. I'll give you an example that's near and dear to both of our hearts. So just yesterday I was having an email debate with a friend. I forget what spawned it. Oh, he had sent me an article about something lipid related and it somehow led to a discussion about the Fourier trial, which for listeners is the trial that looked at one of the two PCSK9 inhibitors. This was Repatha. And it demonstrated that on patients who were on a very high level of statins and had a very low LDL C, I think their average LDL C was in the neighborhood of 70 milligrams per deciliter. Over a five year period they had a reduction in cardiac revascularization, but no change in mortality. His point was, how can these drugs be tolerated? How is it that we live in a society where insurance companies are paying for these drugs or people are using these drugs and doctors are prescribing these drugs where they didn't even demonstrate a reduction in mortality. All they demonstrated was a reduction in revascularization. I can't remember if there was a reduction in events because I sometimes confuse Odyssey and Fourier. To which I said, well, it's really interesting because the time course of that study was so short in a group of patients who were already so heavily statinized that it's my interpretation of that study when it came out, which is probably six years ago, it's actually a miracle it showed anything at all. Because if these patients are walking around with an LDL C that's at the fifth percentile of the population, and then you give them another agent that lowers LDL C to the first percentile of the population, and we're talking about a disease that takes at least four decades to take hold, and you study them for just five years. Would you really expect to see an event difference? Which, by the way, you did see in Odyssey, probably, because using a very similar drug, those patients were started out at a higher level of LDL C. So you saw potentially a greater risk reduction. So here you could have two relatively smart people looking at the same, presumably well done experiment with the same reasonably legitimate statistics, but we have a different lens for what the disease is and therefore draw a completely different conclusion. Is your point that no consensus can ever basically resolve that? And the way that medicine has to progress is that each of us needs to progress on the basis of our own understanding. Or how would you referee the debate between my friend and I?

We can only understand as individuals there is no such thing as a group understanding. We can't get beyond ourselves. Really, it's not possible. And when you talk about thinking, I can't think about every issue that's out there that's important either in medicine or my car or my woodworking. I have to delegate a lot. I'm writing a paper now about familial hypercholesterolemia. And there's a study that's the core study for demonstrating that patients with familial hypercholesterolemia in a human. You can demonstrate a genetic abnormality are at much, much, much, much higher risk than people with similar LDL cholesterols, but without the genetic abnormality. Now, that study has been accepted by everybody as being well done and decisive. And I mean by everybody, everybody.

Just to make sure I understand what you said. You can have two patients who are phenotypically identical, but the one who has the genetic abnormality of fh, which is simply a phenotypic disease, by the way, it's not a genetic. The disease is defined by the phenotype. You're saying they carry a residual risk that's not present in the wild type.

That's correct. Three to five times that risk, 300%.

That's unbelievable.

I thought so too. I read the study and I read it again and I reread it. I couldn't understand it. Then it dawned on me. After a lot of rereadings, I think they made a mistake in their methodology. I think they made a grade school error. Now, maybe I'm wrong because I'm writing this up. I'll subject it to review and criticism. And given that I'm saying every authority in the field is wrong. Chances of coming out on positive and this are pretty slim. But the error in the methodology that I think is present is so simple, it's decisive. Just putting maybe I'm wrong, but my bet is I'm actually right is that thinking is damn difficult and we have to continue to think and discuss with ourselves. HDL cholesterol. Everybody believed it till they didn't. I mean, how many dissenters were there until everybody said it was obvious there was nothing there? And we're human beings, we search for validation. But science isn't about that. Real science. And it turns out that the world is a very slippery thing to get your mind around. I gave the example at the beginning of this talk about risk. If I asked you what risk was, I wouldn't do that to you because you're the boss of this podcast. But risk is the number of events per standard number of people over a defined period of time. We leave out the standard number of people. When we say your risk is 5%, we say, well, that's 1 in 20. Well, it's per standard group of people. That's why I said risk is low. If you're under 60, the number of events per hundred people is low, but the number of hundred people is a lot larger than over 60. That's why the absolute number of events is so high. So a difference of words, tremendous difference in action. Because it means I don't use the ten year risk model. Okay? And even though that has been published and even though that has been reproduced by other investigators, that has not changed the guidelines. And that's wrong. When I look at the guidelines and I have enormous respect for people who serve on them, I can look at the literature on an issue. I can tell you who wrote the guideline. There was a recent guideline from Europe that was negative about apob. One paper cited one paper and that became the guideline. So I'm not talking about throwing out the process. The process of reviewing knowledge in a group is positive. I'm saying that we better watch out for the process. We need to do much more to ensure that the process includes a multiplicity of views so we don't wind up with bad decisions. And if you don't think we can't make bad decisions. Just look at all the bad decisions the politicians and the business guys make. We're no different.

So if you did what you're suggesting, Alan, if you brought in a more heterogeneous group of views and greater diversity of priors, by definition, you could probably never arrive at a unanimous decision.

So what's wrong with that?

Nothing's wrong with that.

My question is, well, is everything right with that?

Yeah, exactly. My question is, what does the guideline look like? So now let's take it back to. The doctor who has to see 40 patients in a day will never listen to this podcast, let alone read the show notes of this podcast, which are probably going to be 150 pages of the backup of everything we've talked about and every study that's been cited, et cetera, et cetera. They literally just want to know what to do in the moment with the person sitting in front of them. And currently they look to the guidelines, which are unanimous, expert run, revised often enough that it gives you the feeling that, hey, they're keeping up with the science. Right? Every five to 10 years, I'll get a new one of these. Now it's going to be this complicated legal document that's like reading the dissents and reading the in favors. And I mean, again, I'm not saying that we shouldn't be doing this, but how does it translate to the field?

How do the guidelines read now?

Oh, they're pretty miserable. No mistake about it.

Not pretty miserable. They're totally miserable. I mean, you talked about legal documents.

But there's a summary and that's what the doctors typically read, right? The doctors will read the one page executive summary that basically says, and that gets hammered home.

And do you know what? I get this rebuttal all the time. We've got to do this because the doctor who sees 40 patients isn't going to do anything unless we dumb this down. I don't think most doctors are dumb. I think they're caring. I think doctors want to do good jobs. I think we need to learn more about how we present information. I don't think we should ever compromise on truth because when you do, you say, I've got to do this in order to get to here. And I don't want to use your political history in the United States negatively, but Afghanistan, I can't think of anybody, right, left or center, who thinks that the process getting in and throughout it represents the best efforts of the minds of America. It doesn't. You guys are really smart, wonderful people. But it's what happens when you boil the options down to two, and you gotta get it down to a one liner. And life isn't a one liner, not real life. And if we're gonna learn to make decisions that are to present information to our patients, we're gonna have to learn how to deal with doubt. That's what it's all about.

Yeah. It's funny, just as you said that I was about to say, the real problem here, Alan, is we are not trained to be comfortable with uncertainty.

I think we are. I think that's what medicine's all about.

I think it is for things that can't be measured, and I think it's not. I think there's a false degree of confidence that comes from things that can be measured. So you're absolutely right. In the olden days when a surgeon went down to the ER to evaluate somebody for appendicitis, so this was long before the days when every one of those patients. Patients had a CT scan. Right. This is based on the books that I used to read from the great surgeons of the 50s and 60s where this was a purely clinical diagnosis. So, you know, 7% of people in their lifetime are going to have appendicitis. And you know that the pretest probability on this person is a heck of a lot higher than 7% because they're sitting here right in front of you, presenting with these signs or that signs. But you also know that there's an asymmetry in your decision. In other words, there's two wrong decisions that can be made here. Operating on the person without appendicitis and not operating on the person with appendicitis. And you also understand that you have to calibrate your decision making around that uncertainty so that one mistake is more likely than the other. And so I think you're absolutely right. I think most physicians are very good at doing this. That somehow that doesn't translate into the type of uncertainty that I think is necessary to do what you want to be able to do, which is rather than give people a unanimously agreed upon consensus that is so distilled down to simplicity that it borders on being incorrect. You'd like to be able to give them the range of thoughts on a subject, acknowledging that you can't tell them which one is correct.

I think, for example, there are different audiences. There's the practicing doctor, there's a practicing family doctor, there's a practicing internist, there's the academic internist, There are the experts in the field I think at a minimum, you've got to have a range of opinion in the experts in the field. And I don't think we meet that minimum. And I think that when there's doubt, you ought to be able to write. I'm not sure about this. Which we do when we have weak recommendations, but they're also always unanimous. What happens when you put five people in a room? Doctors, experts are no different. The one with the loudest voice can carry the day. The one who claims the greatest expertise in the area can carry the day. And then we wind up with documents that are very difficult to read and read, like 20 years ago, with minor modifications. You can't say anything was wrong. Well, good gracious, we learned it was okay, what's our problem? Who are we here for? Ourselves or our patients? And the truth. And so I think you can write clearly and people can read and make informed choices. In the end of the day, what this is saying is that people are too stupid to make a real choice. And I don't believe that. I believe that when we tell people, when I've treated patients and made recommendations, were they always the right ones? Of course not. They were what I thought was best advice, but I can't claim they were always right. And in knowing I could be wrong, then I calibrate the safest way through for my patient. That's what makes me. If I was good, that's what helped me. If I'm wrong about this, how can I set up something to catch it? How can I hedge my bet here, or I gotta go a little stronger here because I could be missing this? Algorithms help us. Of course they do that. When we treat algorithms just by algorithms, then that isn't what's called clinical medicine or clinical surgery. That's algorithms. And. And we have good health care professionals who are more driven by algorithms and more advanced healthcare professionals who can take the algorithms and work within them and around them. And everybody's doing a good job, but we got to respond. Respect the fact that our understanding is limited and that science requires different views to contend equally and that we need to write the truth.

I mean, do you think that we've seen an acceleration of the forcing function around a uniform voice? I mean, it seems to me that the past 18 months with COVID has really amplified what you're describing. I think there are lots of dissenting views out there for how Covid could be managed, what the potential efficacy is for repurposed drugs. And truthfully, I've struggled to wade through the literature on this stuff. And you'll always find somebody who's made it their mission to understand how this drug or that drug or this intervention or that intervention is the solution to the problem. But there's no denying that such people have been pretty roundly silenced for this. And it begs the question, should we be paying more attention to these views? And when do these views become so fringe and marginal that they're actually harmful?

Because they can be. They can be. The first level of discussion is should be amongst experts where you can call out views that are so divorced from experimental evidence that they're not serious views. Covid's a great example. I heard good people express somewhat different views at different times on different issues. Not the same person, but one person had a view. And you say, gee, that's a good point. But they were all within a channel that made sense to me as a physician, not an infectious disease expert. So I'm not talking about legitimizing any possible view because it's a possible view. I am saying that people were pointing out different aspects of a complex problem. And they might put a little more emphasis here, but the result would be. I understand more clearly that there is a series of choices involved, and it's challenging to go through them. In Canada, the government, we had much longer periods between the first and second vaccination, so more people got their first dose before the second. So I get that. That's a decision in real time. That's a real challenge. You can't be sure that you got it right at the moment you're making it, but you're the responsible person. You gotta deal with it. But it was in the open, so, you know, it was being done. I'm good with that process because I know what happened, and we can assess the outcome. I'm not good with someone just shrieking. What I'm against is saying, because we can't have somebody shrieking, we can't have any debate. That's wrong. That's totally wrong. Because it winds up. Then we make mistakes where we don't have to make mistakes. The responsibilities we have to make recommendations to our patients are so awesome. We need to be humble and say, okay, give me your best shot. I'll give you my best shot. And we're colleagues. We're not enemies. I mean, I disagree with lots of people because I have a scientific viewpoint, but they're not my enemies. And we have ways of discussing this. And if I'm in the room, I can say, hey, you said that. I'll show you line 26 in your paper which contradicts you and he has to respond in front of other people. That's what I believe in is the testing of the argument in a jury of your peers.

Is that culture being watered down in science? Is it the same today as it was 30 years ago?

No, I think that's the weakness. I think that's the crucial weakness. I think inside the room that's what's not occurring as I judge that by the product that comes out.

And why do you think that is? Why has this process become diluted?

People get attached to views because this emphasis on consensus and unanimity. You don't want Snyderman in the room because he's going to argue for apob. You can deal with me. I'm not hard to deal with. You can say, okay, I heard you, but here's what's wrong with you. Your argument, not me personally. Here's what's wrong with your argument.

But has something changed to make it that ideas are more personal? Like what is the factor that has led to or factors that have led to that?

I think people are invested in what.

Came out in a way that they weren't before.

No, I mean, I think people used to become known for their own science. Now, if you're on the guidelines, that's your science. You're very prominent because you're on the guidelines, not that you did the science. You're on the guidelines. And we have so much science that's done like the clinical trials. I mean, I think clinical trials are wonderful, but they're limited tools. I think the composition of the committees needs to be re examined. And I think you ought to be able to criticize what comes out. APOB is my thing. I wrote a critique of the 2018American guidelines. How many references were there on the comparison of APOB and LDL cholesterol and.

Non HDL cholesterol in the guideline? Yeah, five.

There were four. Two from a group opposed and two from me, and mine weren't even correctly cited. That's one issue. They dealt with a whole bunch of issues, but that's not adequate. That wouldn't get you a passing grade in school. And we're not limited now by space in dealing with issues. We can put anything out on the Internet for anybody who wants to read it. And what that represented to me was that I can't tell what happened inside the room was there wasn't a full discussion. We didn't have these processes when I was young. Contending ideas could contend more easily. I think apob. I don't think it's going to happen. Not because there is value, but I think there were a group of people who just weren't interested. And that's sad.

And you think that ultimately there are still that many physicians who are going to defer to the guidelines. Reimbursement's not an issue. Right. APOB bills your insurance company 4 bucks and its cash pay is $2.50. I mean, this is not a cost issue. This is really an awareness issue on the part of physicians. That's really all it comes down to at this point.

Well, the guidelines presented as a cost issue. That's the argument against apob.

Right. But the guidelines must have failed to actually look at the cost then.

That's correct. That's correct. I know that, but I can't say that because the guidelines are the guidelines. That's what I'm objecting to.

Of course you can say it. That's what we're talking about here. I mean, that might be. Your next paper is do a survey of the laboratory cost of APOB and the average Medicare reimbursement rate on it across the United States or something to that effect. Those are valuable exercises, right?

I think they are. And we did a cost analysis on APOB, I think using a charge of $10 and looking at the cost of care, and I think it contributed 0.01% to the cost of care, something like that. Okay. The concept that a $10 test really changes the cost of care.

The whole thing is so idiotic. The last time I looked in the United States, and admittedly, we're spending infinitely more than we should and infinitely more than anybody else. This is 10 year old data. You know, it's probably closer to 10,000 now, but 10 years ago in the United States, our health care spending was $7,000 per person per year. That's 10 years ago. So again, I don't see how it's less than $10,000 per person per year in the United States today. So if the APOB test cost $100, which again is 20 times more than it actually costs, so what? Atherosclerosis, when last I checked, is the number one killer of men and women. It kills women at a rate that is more than 10 times that of breast cancer. I mean, are we missing something here? Is there some part of this that you haven't told me?

That's my frustration, that's my sadness. You're absolutely correct. This is pennywise pound foolish, where it's lives that are being lost and we've calculated the number of lives and heart attacks that could be avoided, lives saved, heart attacks avoided. If we switch to apob, the American College of Clinical Chemistry, the European clinical chemists, they have a series of reports saying APOB can be measured more accurately than LDL cholesterol and non HDL cholesterol. No question. Is that in any of the guidelines? Accuracy of measurement? No. That's what's so hard to deal with when you say you're criticizing the guidelines. Maybe it's just you, who are you? You're a nobody. Which is true. But I'm saying, look, it's a laboratory test. Surely the quality of the measurement is something that should be mentioned. The diagnosis of type 3 can't be done without APOB. So any hypertriglyceridemia, pain, needs an APOB. Not done, not mentioned. Type 3. So I can't beat on you with sadness and stuff. I've been incredibly privileged to have the opportunity to try and understand the world around me. My background is not a would have normally led someone like me to have that chance, nor to work with the quality of people that I've had the privilege of working with. And I've been able to write and record the images of the world that looked real to me and how few human beings ever get that privilege. And I'm just sad that it won't help. It won't help people. I'm sad about that and I'm sad that you go so far in the thinking and then somebody can take it to the next step. It's not like I've done that much. Somebody can see this and say, oh wow, if that's so, let's go there and it's not going to happen.

I don't agree on two levels. The first, I'm going to call you out on this, Alan. For a guy who understands uncertainty and probability and risk, you're using awfully black and white language right now. It's not going to happen. With what certainty can you say that?

So you're right, you're right. And it's self pity and it's not attractive to me either, okay?

And I'm not saying that to belittle the point of view. What I'm really saying is there's a beautiful story about the guy that's hitting the stone, right? The mason who's hammering away on the stone for 10,000 strikes. And on the 10,000 first strike, when he hits the stone, it finally fractures. Now to the person watching it, it was that 10,000 first strike that fractured the stone. But of course the stone mason knows that it was that strike plus the 10,000 that came before it. And so you just have to accept the non linearity of the advancement of knowledge. It's embarrassing that I would even attempt to try to say that to you because you know that so much better than I do. Of course. But maybe in this one situation, Alan, you're just closer to it than I am and therefore you're just too close to the political environment of it. Because at the end of the day, these consensus statements are largely political. They're far more based on who you are and who you know and how long you've been on the committee than the strength of the evidence. But I wouldn't bet against the truth. I think that in the end the truth generally wins in science. Not always, but it also depends on your time course. It just wouldn't surprise me if in 10 years, 20 years, we're going to look back at this just as we are looking back today at hdlc. I mean, that was not that long ago, Alan, that people thought HDLC was everything. And today, thanks to the CTEP trials, the misters, we now know HDL C is A much more complicated than we ever thought and B, probably not something we should be trying to manipulate in an effort to improve outcomes.

At the end of it all, that's what I'll hope for. You need to pick yourself up and keep fighting. We have a new observation that we think is even different, more startling than anything you've heard and we'll try to quit, is wrong. Part of this is for yourself, just to see. Can you understand what looks like chaos? Is there any pattern to what looks like. There's no pattern. How does blood sugar and lipids and how do they combine to hurt us? So just to have the privilege of looking at those questions and to write down your thoughts in a manuscript and to publish the manuscript, so few people ever have that privilege. I mean, I've just been extraordinarily rewarded for the very little bit that I've been able to achieve. So I have a short fuse and it annoys me. Stupidity annoys me, ignorance annoys me, and pretentiousness annoys me deeply.

Look, I think those are all things worth being annoyed at. Another one of your strengths, Alan, that I've observed over the 10 years that we've known each other, maybe longer, is you have an insatiable curiosity about things that are theoretically or in quotes, theoretically outside of your lane. And we've probably exchanged as many emails on the mechanisms of insulin resistance as we have on everything that's related to lipids. So I think that that's a very important part of your success, which you've been very modest in downplaying is the breadth to which you think about this problem. I think that's why you've been able to spot patterns that are not obvious to others. Before we wrap, there's one thing I want to go back to because we spoke about it quite briefly, but I think it is so important. And when we spoke about it at the outset, we hadn't given the listener the full landscape of the disease. But now that we have, I think they will understand more what you mean when you talk about the causal model of risk, this 30 year causal model versus just a risk calculator. So let's go back to that. You go to a Framingham risk calculator and you plug in a few variables. I'm this old, I do smoke or I don't smoke. My HDL cholesterol is this, my blood pressure is this, what is my 10 year risk? And you hit click on the button and a number comes up. And most of the guidelines say if that number is below something. A typical response would be if the 10 year risk is below 5%, there is no need to treat, carry on as normal. I'm exaggerating a little bit, but that's the gist of it. I practice medicine a very different way because I have a very different goal. I have an objective for how long a person might be able to live disease free. And I also tend to be influenced by people like you who have taught me how long it takes for this disease to take hold. So the analogy I would use is, is imagine there was a calculator that told you how much money you should be saving for retirement, but it could only predict 10 years into the future. It wouldn't be a very useful tool because a 30 year old isn't going to be retired in 10 years. So how has your thinking, which has influenced me greatly, how did you come to the 30 year model that is based on the causal relationship? What are the types of yields that we see there?

Part of it is utilitarian. 30 years was the longest stretch we had reliable data for. And part of it was that the period of uncertainty is greatest between age 30 and I'm saying 60, but 55, whatever, that's where the tenure falls down. And we originally did these models based on 10 year risk. And you could say, you're right, we're going to miss most premature Disease, let's just lower the risk. Instead of 5%, we'll make it 2.5%. And you could do that. The problem is it's not cost effective because you're multiplying the number you need to treat more than you would need if you start off with a co. Now, let's say I categorize you by your APOB or your non HDL cholesterol. Let's say you have a high APOB. So you're in this group of about 25% of the population, because I use 25%, 75% to get you there. So that group has, over 30 years, a 30% event rate. It's a lot. It's not just the APOB that's doing it.

Age is still a big driver.

Age is a driver. But they'll be fatter at the start. Their blood pressure will be a little higher at the start. It's not that the other things are being left out. You're catching them. You're just selecting differently. And I would submit you're selecting more accurately, because I can measure APOB or non HDL cholesterol with much less error than I can measure your risk. The variables that we know Michael Pancini's calculated, we account for about 20% of the variance in risk. That's about it. That's a small number. So from a utilitarian perspective, by stretching out and using the idea of cause and precision of measurement, I group you, then I can present a patient with what we know about the group and they can make their decision. There are people who say, thank you, I'll pass. They come back in six months, a year, we can have the same discussion again. I didn't lose anything. They've lost some, and we can show that. But these aren't decisions that you can't revisit. If someone starts on therapy, they can decide to stop it. Patients are free agents. I mean, I respect that. And it's up to us to present our arguments and continue to present them as we frame them to their, what we think is their best interest and the best we can do. So I like to maintain contact with my patients who are in this sort of thing to make sure that they know they can change their mind. But I can also show them Progress. If their APOB started at 120 and we're down to 60, I know they're taking the medication. I really do. Okay. Otherwise it would be 120. And they can see and appreciate the extent of change that's occurred for me and for the patients I've cared for, that tends to work.

One of the things that I find very helpful to explain this, because it's nonlinear and I think nonlinearity is not innate to us. I don't think evolution needed us to be able to think nonlinearly. I think linear thought was good enough, and that included linearity with time and chance. But this problem doesn't lend itself to that. And therefore I think it's not intuitive to somebody who hasn't filled with numbers a lot or hasn't spent time looking at things that compound. Even if you were just to look at a risk based model that was short term, there's a significant benefit to reducing risk from 4% to 2% over a decade. Because even though that might only mean there's a 2% risk reduction over the next decade, the amount that that amplifies over two, three and four decades is amazingly counterintuitive, just in the way it is if you think about saving for retirement. If you're 40 years old or hopefully younger when you start saving for retirement, if you're 30 years old, when you really start in earnest to saving for retirement and you have one option that's going to generate a return of 7% and another that generates a return of 10% or 9%, there's a 2 percentage point difference. That's not going to yield an enormous difference by the time you're 40, but by the time you're 70, it does. I've done the math on these and you can easily show how it can literally double the nest egg just based on relatively small changes up front. So I think that's also part of the issue here is just understanding what compounding does and how it works in your benefit when you're talking about investing, but understanding how it works against you when you're talking about a disease like atherosclerosis.

We've done some of these calculations and you're right. Starting early pays off later. And when you start later, what you're doing essentially is to try and modify the disease that's already present. When you're starting early, what you're actually doing is stopping disease from developing. And if you can stop a lesion developing, that's 100% success trying to treat, modify through one process a complex set of outcomes at this end, much less likely to succeed. So stopping disease is perfect prevention, treating disease is partial prevention. And it has to be partial, can't be more.

Yeah, I think that's a very important point because although we haven't stated it explicitly, the undertone of everything we've said is not only that apob is causal, but that it's a necessary but not sufficient driver of atherosclerosis. And I think that's important because necessary but not sufficient creates a lot of confusion in medicine, doesn't it? Necessary means you have to have an apob particle traffic a lipid into an artery wall. If that doesn't happen, you don't get atherosclerosis. Sufficient means if that's the only thing that happens. Do you necessarily get disease? No, you don't. Each of us could rattle off tons of patients who somehow make it into their 90s with high APOB and don't have disease. But from a prevention standpoint, it's much easier to go after something that's necessary because you only have to block that versus once the disease has already taken hold. You're advancing something that's multifactorial. So I think that's a very important point you raise, and I don't think it gets made enough. Well, Alan, you were a little reluctant to sit down. You didn't see the value in doing a podcast, but I'm glad we twisted your arm a little bit. It's funny, I didn't realize until a little while ago that you were 80. You always come across as so much younger to me, so that explains to me why you've got lots of miles left in terms of the APOB work and the crusade here. So I'm very optimistic.

It was nice to Dinner would be even nicer, but this was nice to speak with you again.

We'll make that dinner happen again at some point soon. Alan, thanks so much.

You're very welcome.

Thank you for listening to this week's episode of the Drive. If you're interested in diving deeper into any topics we discuss, we've created a membership program that allows us to bring you more in depth exclusive content without relying on paid ads. It's our goal to ensure members get back much more than the price of the subscription now. To that end, membership benefits include a bunch of things. One Totally kick ass comprehensive podcast show notes that detail every topic, paper, person, thing we discuss on each episode. The word on the street is nobody's showing notes rival these monthly AMA episodes or Ask Me Anything episodes. Hearing these episodes completely access to our private podcast feed that allows you to hear everything without having to listen to spiels like this the Qualys, which are a super short podcast that we release every Tuesday through Friday, highlighting the best questions, topics and tactics discussed on previous episodes of the Drive. This is a great way to catch up on previous episodes without having to go back and necessarily listen to everyone. Steep discounts on products that I believe in but for which I'm not getting paid to endorse, and a whole bunch of other benefits that we continue to trickle in as time goes on. If you want to learn more and access these member only benefits, you can head over to peterattiamd.com subscribe. You can find me on Twitter, Instagram and Facebook all with the ID Peter atiamd. You can also leave us a review on Apple Podcasts or whatever podcast player you listen on. This podcast is for general informational purposes only and does not constitute the practice of medicine, nursing or other professional health care services, including the giving of medical advice. No doctor patient relationship is formed. The use of this information and the materials linked to this podcast is at the user's own risk. The content on this podcast is not intended intended to be a substitute for professional medical advice, diagnosis or treatment. Users should not disregard or delay in obtaining medical advice from any medical condition they have, and they should seek the assistance of their healthcare professionals for any such conditions. Finally, I take conflicts of interest very seriously. For all of my disclosures and the companies I invest in or advise, please Visit Peter Attia md.com about where I keep an up to date and active list of such companies.

Sa.