The Peter Attia Drive - #03 - Ron Krauss, M.D.： a deep dive into heart disease

00:00:00.000 Hey everyone, welcome to the Peter Atiyah Drive. I'm your host, Peter Atiyah.

00:00:10.140 The Drive is a result of my hunger for optimizing performance, health, longevity, critical thinking,

00:00:15.600 along with a few other obsessions I've gathered along the way. I've spent the last several years

00:00:19.860 working with some of the most successful, top-performing individuals in the world,

00:00:23.440 and this podcast is my attempt to synthesize what I've learned along the way to help you

00:00:27.840 live a higher quality, more fulfilling life. If you enjoy this podcast, you can find more

00:00:32.100 information on today's episode and other topics at peteratiyahmd.com.

00:00:41.340 In this podcast, I'm speaking with Dr. Ron Krause. First and foremost, Ron is a very close friend

00:00:47.420 and an amazing mentor. I was introduced to Ron probably five years ago and have worked with him

00:00:54.000 closely in a number of capacities. He's always served as one of the three or four lipidologists

00:01:01.500 that when I get stumped on a really difficult clinical case, he's the person that I'm reaching

00:01:05.980 out to, along with a couple of these other folks that hopefully we'll also have on the podcast at

00:01:09.800 some point. He is certainly recognized globally for his research in lipidology. He wears a ton of

00:01:16.420 hats. He's a clinician, obviously a lipidologist. His interest in nutrition, genetics, drug research

00:01:21.860 is quite profound. He obtained, I believe, both his college degrees and medical degrees from Harvard.

00:01:28.500 He's board certified in internal medicine, endocrinology, and metabolism. And currently

00:01:31.740 he's a senior scientist and the director of atherosclerosis research at the Children's

00:01:36.120 Hospital in Oakland. So in this episode, we talk about a lot of stuff, but obviously we're really

00:01:40.320 focusing on atherosclerosis and cardiovascular disease. I was really excited to have this discussion

00:01:46.260 with him because a lot of these topics I've certainly covered in writing. As some of you may know,

00:01:51.580 I have that sort of nine part straight dope on cholesterol series on the blog that I probably

00:01:57.840 wrote about four or five years ago. Obviously some of that's a little bit outdated. And I also was

00:02:03.260 quite deliberate when I wrote that not to be prescriptive, meaning I don't really get into

00:02:08.500 this is the drug treatment you would do for this. And I largely avoided a lot of that stuff. Whereas in

00:02:14.160 this podcast, Ron and I get a little bit into this. This podcast was pretty technical at times. So my hope

00:02:20.340 is that both the curious patient will get a lot out of that and hopefully the physician will get a lot

00:02:26.340 out of this or the person that is also kind of on the front lines of having to make decisions about

00:02:31.720 how to treat dyslipidemia and reduce the risk of atherosclerotic disease. I think a couple of

00:02:37.180 really interesting things in this, I actually for the first time learned about Ron's motivation for

00:02:42.440 this, both his family history and the five legendary articles that he read in the New England Journal

00:02:48.320 of Medicine that largely shaped his career. This is the stuff we talk about with Friedrichsen,

00:02:53.000 Levy and Lees, who I've written about in the cholesterol series, because these guys are sort of

00:02:56.700 the fathers of this space. We certainly get into one of the age old debates about LDL particle size

00:03:03.600 versus particle number. Lots of controversy here. And I don't represent that we've necessarily

00:03:08.540 resolved it, but I think that's the beauty of talking with really smart, sophisticated,

00:03:12.660 nuanced people is they have the humility to say, we don't know the answer. Sometimes

00:03:16.320 we really dive deep into the whole statin discussion for reasons that aren't entirely

00:03:21.380 clear to me. This has become an increasingly controversial area and somehow it's turned

00:03:27.060 into a binary discussion. Statins are good or statins are bad and very few things in life tend

00:03:33.320 to be that binary. So I'm sort of surprised that it's turned into that. And I don't remember if I

00:03:38.640 even say this on the podcast, but certainly people have probably heard me say this before. And I say it to

00:03:42.660 patients all the time. Statins are tools. And the most important thing when you have a tool is

00:03:46.900 knowing how to use it and knowing when to use it. So if you have a Phillips screwdriver,

00:03:50.520 it's really important to know that it's very good at putting Phillips screws into things. It's not good

00:03:56.520 at putting nails into things. It's not good at cleaning windows. So I hope we can shed some light

00:04:03.240 on that. We talk a little bit about the really interesting and recent stuff around chronic inflammation

00:04:08.100 inflammation and the role that that plays in atherosclerosis, even independent of cholesterol

00:04:12.120 levels. And perhaps for me personally, one of the most interesting things we discussed was another

00:04:18.180 very controversial topic, which is around niacin and niospan, which is a branded version of that,

00:04:23.340 which those of you who follow this world will know that niacin was basically kicked to the curb a couple

00:04:28.520 of years ago. And I think that Ron's insights into that are incredibly interesting and actually have even

00:04:34.860 made me re-question or re-evaluate, I guess, my willingness to ever consider using it again.

00:04:41.420 Ron does a great job explaining the HDL paradox, meaning why is it that all the pharmacologic efforts

00:04:47.240 to raise HDL seem to also raise heart disease or at best make it no better. And finally, we end with

00:04:54.880 a discussion of PCSK9 inhibitors, which I suspect will have a completely dedicated podcast to this topic

00:05:00.760 at some point. And we touch on LP little a though, later on in the release of this podcast, we're

00:05:06.380 going to have a dedicated discussion on LP little a. Okay. So with all that said, you can find a ton

00:05:11.240 more information, including a lot of links to the papers that Ron has mentioned, more information

00:05:15.140 about Ron in the show notes, which are at peteratiamd.com forward slash podcast. So without further

00:05:21.460 ado, here is my conversation with Dr. Ron Krauss. Well, I'm here with Dr. Ron Krauss today, and this

00:05:30.920 is a really exciting topic for me. People know I've written about this a lot. I talk about this a lot

00:05:36.060 clinically, but the genesis of this discussion today is that about a month ago, I called Ron to have a

00:05:43.820 discussion with him about one of my patients in particular. It was a patient who had a pretty elevated

00:05:49.480 calcium score, if I recall, pretty significant LAD calcifications. That's the artery in the left

00:05:55.700 side of the heart. But he was very hesitant to do any treatment and he wanted a second opinion. So I

00:06:01.780 thought we should involve Ron. And Ron, I remember I was standing in my kitchen, we spoke for probably

00:06:05.880 half an hour. And at the end of the conversation, I said, you know, Ron, I wish we recorded that

00:06:09.780 conversation because this is exactly the kind of stuff that I think a lot of physicians and patients

00:06:15.620 would benefit from. And so I said, you know, why don't we do this again more formally? And that

00:06:20.960 brings us here today. Let's start with a big question, but an important one for where we're

00:06:27.200 going. And that's basically the pathophysiology of atherosclerosis. A very recent review article

00:06:32.160 I read described it quite eloquently as a smoldering inflammatory condition fueled by lipids. What does that

00:06:40.540 mean? Well, first of all, thank you, Peter, for asking me to talk with you today and address this

00:06:46.500 topic, which, as you know, I have a deep and longstanding interest in. Hopefully I can address

00:06:51.060 the issues that you're also interested in and we'll have a good conversation. Well, atherosclerosis,

00:06:57.820 of course, is the underlying process that leads ultimately to vascular disease, particularly

00:07:04.760 clinical events, heart attack and stroke. It starts in childhood. It's well known that there is early on

00:07:11.540 the buildup of cholesterol in the artery wall that forms what's called fatty streaks. And that's a

00:07:18.080 process that's actually a fairly normal condition, even in young people. And if it doesn't progress any

00:07:24.460 further than that, it's really not hazardous. It's a way that arterial tissue can put cholesterol in

00:07:32.120 its cells. And some of that cholesterol is actually used for various purposes. So that's not necessarily

00:07:37.260 a pathologic process, but it can progress. And when it progresses, there is a combination of factors

00:07:44.660 that conspire to make that fatty streak into a much more toxic process. And it is fueled by lipids,

00:07:52.520 the same lipids that lead to the fatty streak. But under conditions that many of us live under,

00:07:58.440 there are changes in the lipoproteins that are taken up by the artery. And we'll talk about

00:08:04.800 those in more detail in a minute. In particular, susceptibility to oxidation and the change in the

00:08:12.280 properties that allows them to stick more tightly to the artery wall. And when that happens, particularly

00:08:18.760 the oxidative changes, it does trigger inflammation. It's a very early part of this next phase of the

00:08:25.440 disease process. And inflammation is defined in one way as the accumulation of cells in the artery

00:08:32.540 wall that deliver various inflammatory molecules, things that ordinarily, if one bruises oneself or

00:08:40.380 has some sort of an injury, those inflammatory processes cause redness and accumulation, in some

00:08:46.320 cases, of clotting factors. When that happens in the artery, that can convert this fatty streak into

00:08:52.440 something that is much more malignant. And then there is a process that kind of feeds on itself.

00:08:57.400 And it does involve platelets and clotting factors in an important way. It involves a number of

00:09:03.340 inflammatory molecules. And if there's continuing input of these bad atherogenic, if you will,

00:09:10.600 lipoprotein particles, that can actually change the nature of the plaque. And inflammation comes into play

00:09:17.340 in a very serious way when that results in a breakdown of the surface of the plaque, which

00:09:24.100 ordinarily protects it from any kind of serious consequences. So even fatty streak, which can

00:09:30.700 develop into a plaque. And a plaque in this case is essentially a larger fatty streak. The plaque has

00:09:35.580 the cholesterol, but it also has all these other cells. It's a much more complex phenomenon. So that

00:09:40.740 plaque is ordinarily, under relatively benign conditions, encapsulated by a fibrous layer. But

00:09:48.580 inflammation and the release of various molecules can cause that fibrous cap, that protective cap, to

00:09:54.820 weaken and ultimately potentially rupture. And when that rupture occurs, that is the beginning of the end in

00:10:00.020 terms of the process that we're talking about here. Most cases of heart attack and stroke involve this type of

00:10:06.920 acute rupture and ultimately formation of a clot that blocks the arterial flow. So starting from

00:10:13.560 a relatively benign process, this can develop into something that's much more serious.

00:10:18.040 Now, when I was in medical school, I remember in first year pathology lecture, the pathologist said,

00:10:24.200 let's see a show of hands. What is the most common first presentation of heart disease? And,

00:10:29.880 you know, everybody puts up their hands and says, chest pain, left shoulder pain, shortness of breath.

00:10:35.320 And he said, no, no, no, no, no. It's sudden death. That was a little over 20 years ago. Is that still

00:10:40.680 true today? The estimates I've heard, and I think this is arguable because these are really rough

00:10:45.880 rule of the thumb calculations, somewhere in the range of 30%, possibly upwards of 30%, which is still

00:10:51.400 a huge number. It's staggering. It means that one third of people's first brush with the knowledge

00:10:57.080 that they have atherosclerosis is death. And I have patients, and I'm sure you do as well,

00:11:01.800 who have died and come back. And so there is this process where there's an acute event that causes

00:11:09.160 an irreversible change. But for some people, fortunately, we can bring them back. But together,

00:11:13.240 that represents really the basis for calling this disease the silent killer. Because, as you were

00:11:19.480 saying, we don't, in those patients, have premonitory symptoms. Sometimes in retrospect,

00:11:25.240 they are there. And I think that's the important reason for educating the public,

00:11:30.040 as organizations like the American Heart Association does, as to the first signs of heart

00:11:35.000 disease. Because it may be, and it probably is true, that a significant component of that 30%.

00:11:39.800 Right. Upon further querying, there was some exercise intolerance.

00:11:44.440 Didn't recognize it. Yeah. And it's very hard sometimes. One of the things that we'll be talking

00:11:48.280 about is ways of assessing risk. And those are still imperfect. I mean, we can't, with 100%

00:11:53.800 certainty, use any kind of risk predictor to know if somebody's destined to have a heart attack with

00:11:58.760 certainty. Yeah. You said something at the outset, which is, this is a disease that begins in infancy.

00:12:03.000 And I, one of the, I have very few textbooks and or papers that are, I refer to so frequently that

00:12:09.080 they actually sit on my desk in my office so that every time I'm with a patient, I can pull them out.

00:12:13.640 But one of them was a book that was given to me by one of my mentors. I consider you a great mentor,

00:12:18.920 Tom Dayspring, a great mentor, Alan Snyderman, a great mentor. And Alan gave me this textbook of

00:12:24.120 pathology. I believe it's Starry is the author. And while I believe the data represented,

00:12:30.920 there are somewhat dated because it was largely based on the Vietnam cohort and,

00:12:34.440 and Korean. Yes. Where obviously smoking would have been a higher prevalence than today. Right.

00:12:39.320 The fact remains that when you look at autopsies of young people who died of unrelated reasons,

00:12:46.200 homicides, accidents, et cetera, and you look at the histologic sections of their coronary arteries,

00:12:52.840 it's amazing how many of them have lesions that are type three or beyond type three, meaning

00:12:57.960 obviously a type of pathological region where you go beyond fatty street.

00:13:02.040 Yes, indeed. That's right. So a subset of these youths will have more advanced lesions. And the

00:13:07.720 studies that have been done have linked all of the usual risk factors, smoking, certainly diabetes,

00:13:14.280 hypertension, and dyslipidemia, and a lipid disorder. All of those have been associated with the more

00:13:21.400 advanced, uh, lesions in, in those individuals. So as you're pointing out, even, even a more

00:13:26.920 significant plaque development can occur in childhood. I think the thing that's hard for

00:13:31.800 people to understand, uh, and I think it's true of most chronic diseases, but I don't think any disease

00:13:37.560 in any disease, it is as clear as it is with atherosclerosis, which is the compounding nature of

00:13:41.960 the disease. You know, another great example of one of those questions that the professor asks that

00:13:47.000 gets everybody stumped, which is what's the greatest risk for heart disease. You know,

00:13:50.920 is it smoking? Nope. Is it high blood pressure? Nope. Is it dyslipidemia? Nope. It's age.

00:13:55.720 That's right. It's age. I mean, and why is it age? Because it's exposure, it's time,

00:14:00.920 it's area under the curve. Yeah, that's exactly true. You know, age,

00:14:03.720 regrettably is a risk factor that cuts across, uh, many of the diseases, chronic diseases that we

00:14:09.960 have to deal with cancer, for example. Uh, and yeah, that's right. It's a cumulative process

00:14:14.840 that can progress at various rates depending, uh, the condition. So people, as you know,

00:14:21.080 when we talk about who have genetically elevated, severely elevated cholesterol levels will have

00:14:26.920 that process accelerate and have the disease show up clinically early on. And sometimes with these

00:14:31.240 severe genetic disorders, um, in the teens, whereas others, most of the population, fortunately, who do

00:14:37.960 have risk factors show a gradual increase in the manifestation of disease as a result of those risk

00:14:44.440 factors as a function of their age. You know, at last check, and I can't remember if it was JAMA or

00:14:49.480 another journal, but it was about a year ago and they looked at some actuarial data for people out

00:14:54.360 through being past centenarians and the only disease once you normalized for a few things,

00:15:00.840 the only disease that increased monotonically by decade in risk was atherosclerosis from childhood.

00:15:05.720 Yeah. Even cancer actually, you know, I think by the ninth decade, it started to come down.

00:15:11.240 I see. Yeah. Yeah. Yeah. You know what? Yeah. That's right. No, it's definitely different. I mean,

00:15:14.360 the latency period for cancer also, uh, is a factor there as well. So there's this sort of

00:15:19.400 latency period where nothing happens. And then all of a sudden in the older age, it pops up.

00:15:23.480 I'm sure a lot of people listening to this are going to say, okay, well, God, I'm really confused by

00:15:27.640 half the terms you guys just use. So let's unpack some of them beginning with, we use the term,

00:15:32.520 sometimes lipid cholesterol, lipoprotein. We, we throw those terms around interchangeably,

00:15:36.920 but I think it's probably important to give the average person a sense of an understanding. So

00:15:42.200 what is LDL-C versus LDL-P versus ApoB and things like that? Sure. The underlying concept

00:15:50.520 that we are going to address is, is cholesterol because that's really the compound. It's the,

00:15:55.960 it's the molecule that winds up causing plaques. So cholesterol is indeed an important component of

00:16:01.960 the, of the plaque. And it gets into the plaque by uptake of cholesterol from lipoprotein particles.

00:16:08.840 And so lipoproteins are complex spherical macromolecules, big guys, which come in varying

00:16:15.800 sizes and are composed of cholesterol along with other lipids, uh, such as triglyceride. And most

00:16:23.000 importantly, perhaps for distinguishing the various types of lipoproteins is their protein content. So

00:16:29.000 there's a, a variety of different proteins that form, uh, the package that actually a capsule

00:16:34.760 around the lipid. And so that's, let me interrupt you for one second. Is it just to point at

00:16:39.080 clarification? The reason we even need these lipoproteins is that cholesterol is hydrophilic,

00:16:46.040 oh, pardon me, hydrophobic. It repels water. And so therefore to move cholesterol through the

00:16:50.760 bloodstream, you have to package it in something that is hydrophilic or dissolves in water. Is that

00:16:56.120 correct? Right. Yeah. And for the, the technophiles here, um, it's, it's the cholesterol

00:17:00.360 ester. So there's a, there's two forms of cholesterol and it's the, the fatty form of

00:17:04.760 cholesterol is cholesterol ester. The other form is more waxy. So the fatty form is transported

00:17:10.040 right from one tissue to another. And that is the purpose of lipoproteins, not just cholesterol,

00:17:13.560 of course, but triglycerides, as I mentioned, even perhaps more importantly for many functions,

00:17:18.280 energy metabolism and other compounds, such as phospholipids, as well as passengers on the,

00:17:24.200 on the truck, uh, uh, certain vitamins, et cetera. So these are packages that, uh,

00:17:28.600 serve an important biologic function. They're not here to cause heart attacks. Uh, we divide

00:17:33.160 them into various categories, but the common parlance, the sort of the most typical way that

00:17:38.040 we think about cholesterol as a pathologic factor is when it's on LDL. So that's called LDL cholesterol.

00:17:44.280 And that measures the amount of cholesterol on an LDL particle. And LDL is low-density lipoprotein.

00:17:49.960 So that's, and so this is a form of lipoprotein that, uh, is characterized by size and it's also

00:17:57.960 characterized by its density, which is related to its capacity to float because there's, when there's

00:18:02.840 fat in anything, it causes the thing to float. And that's to varying degrees defines different

00:18:07.960 classes of lipoproteins. And it's the LDL that is the most strongly connected to cardiovascular disease

00:18:14.440 risk and the cholesterol in LDL, which is measured commonly clinically as LDL cholesterol is what has

00:18:22.680 been most widely associated with cardiovascular risk and forms the basis for many of our recommendations

00:18:29.640 for lowering risk. But it's important to recognize that this is a tag on a much more complex substance,

00:18:36.440 a particle. And this we'll talk about, I think very shortly, particle is what causes the plaque to

00:18:43.080 develop and it brings the cholesterol with it. So the LDL particle, the low-density lipoprotein itself,

00:18:51.960 the spherical molecule or macromolecule, which carries around cholesterol, esterol,

00:18:56.520 phospholipid, triglyceride, it has a signature on it, doesn't it? Something called ApoB100.

00:19:02.440 That's right. The key protein that holds this particle together, that allows it to form a sphere

00:19:09.320 and to encapsulate the lipid cargo is called ApoB. There are two major forms of ApoB. The one

00:19:17.320 that's found in LDL particles is called ApoB100. And oftentimes that is used as a surrogate for measuring

00:19:27.080 LDL particle concentration, as we'll talk about. And this is a big clinical distinction.

00:19:32.040 I think it is safe to say, at least in my, you know, relatively modest sampling of physicians,

00:19:39.080 most physicians, let alone most patients, are not really clear on the distinction between

00:19:43.880 the number when they say LDL is 100. They don't necessarily realize what they're saying is,

00:19:48.840 the LDL cholesterol is 100 milligrams per deciliter, meaning if you took all of the LDL

00:19:54.440 particles in the body, smashed them apart, gathered the cholesterol ester, the mass per unit volume is 100.

00:20:00.760 And free cholesterol.

00:20:01.640 And that's, yes. And that's very different from saying, how many of these particles do we have?

00:20:07.160 That's right. And that's important. And so that observation, it really forms a very significant

00:20:13.720 component of my history in this field, because I entered the field as a young fellow knowing about

00:20:21.000 LDL cholesterol. I was interested in diet effects and drug effects and heart disease. But I learned

00:20:25.720 about lipoproteins actually from a group of investigators in Berkeley, California, who were

00:20:31.720 part of a team that initially identified lipoprotein particles. And over the course of the next 10 or so

00:20:37.880 years, I dug into that knowledge and discovered that there are subtypes of the various forms of LDL,

00:20:46.920 as well as other lipoproteins, which we can talk about perhaps in a few minutes. But focusing on

00:20:52.680 the LDL, there can be variation in the amount of cholesterol that is carried on an LDL particle.

00:20:59.480 But there's only one ApoB per LDL. So ApoB represents a pretty good signature for an LDL particle. It is

00:21:08.440 found on some other particles, but it's primarily on LDL. But the amount of cholesterol attached to that

00:21:14.440 ApoB as part of this particle can vary, as can other lipid components. And that results in variation

00:21:21.400 in both the size, as I mentioned, and the density, so that some forms of LDL have less cholesterol,

00:21:28.360 and some have more cholesterol. The ones that have less cholesterol are smaller, generally,

00:21:33.240 and the ones that have more cholesterol are larger, but they all have one ApoB.

00:21:37.880 So there can be an important clinical consequence of focusing on LDL cholesterol to the exclusion of

00:21:45.320 ApoB, because it's the particle that is really the agent of damage in the artery. And measuring LDL

00:21:54.200 cholesterol can under-represent the number of LDL particles compared with the measurement of ApoB,

00:22:02.600 which is a much better measure of the number of particles. And when individuals have smaller

00:22:07.240 particles because of this variation in lipid content, they are actually at higher risk of

00:22:12.520 heart disease because those particles have properties that render them more pathologic,

00:22:17.480 more toxic. So there's a double whammy. If you have small particles, first of all,

00:22:24.040 clinical measurement of LDL cholesterol may under-represent the number of particles. And

00:22:28.280 furthermore, those particles themselves are considered by many, although there's still not

00:22:34.120 total consensus on this point, to have greater pathologic properties.

00:22:37.960 So let's use a specific example. So if a patient has a blood cholesterol level,

00:22:45.800 and let's just assume it's LDL measured directly, not even calculated. And the LDL cholesterol is 100

00:22:52.440 milligrams per deciliter. At the Framingham population, that would place them at about the

00:22:57.720 20th percentile. But let's say that patient has an LDL particle number of 1,400 or 1,300 nanomole per

00:23:07.240 liter. And of course, in the units, that tells you it's a number per unit volume. That places them

00:23:12.120 at the 50th percentile. Now, at least to me, the literature is very clear on, in the case of

00:23:20.840 discordance, which of the two is driving risk. I think both the MESA population, the multi-ethnic study

00:23:28.040 of atherosclerosis and the Framingham and Framingham offspring study make it very clear that risk is

00:23:33.800 tracking with the number of particles, not the cholesterol concentration. Do you agree with that?

00:23:39.640 Or am I missing something?

00:23:40.840 No, the data are out there. And just to take a step back, LDL cholesterol, that clinical measurement,

00:23:47.880 has worked reasonably well for a significant subset of the population as a marker for LDL particles.

00:23:55.080 Because most individuals have particles somewhere in the middle of the LDL size and density range,

00:24:01.400 and the cholesterol content in those particles is fairly proportional to the number of LDL particles.

00:24:06.280 But where things break down is the increasing proportion of the population who have different

00:24:12.680 LDL particle distributions. You need to consider that higher LDL particles with normal LDL cholesterol,

00:24:19.240 which I will answer your question, yes, is associated with an increase in risk that's

00:24:23.800 not reflected by LDL cholesterol. So that is the discordance we're talking about on the high end.

00:24:27.960 And conversely, individuals who have high LDL cholesterol, but normal levels of ApoB will

00:24:36.600 tend to have disproportionately less heart disease risk than would be predicted from the LDL cholesterol.

00:24:41.800 So that's the discordance, which I think is pretty well demonstrated in a significant

00:24:46.920 subset of the population at both ends. But underlying that, and again, there's some

00:24:51.720 debate on this issue, it certainly reflects the numbers of particles, and that is the bottom line.

00:24:57.320 But it also reflects the types of those particles. So there's really two features at both ends of that

00:25:02.120 distribution. Smaller cholesterol particles associated with discordance at the low cholesterol

00:25:08.040 to ApoB protein B ratio, and conversely, larger LDL particles associated with less risk at the other end.

00:25:14.600 And there may be systemic factors here at play because, you know, there's some pretty,

00:25:19.240 I think, pretty widely accepted data now. I think 10 years ago, this was a little more obscure. But

00:25:25.080 one of the greatest drivers of the discordance in the wrong direction, meaning the LDL particle is

00:25:30.040 disproportionately higher than the LDL cholesterol, is metabolic syndrome. In fact, there's a very beautiful

00:25:36.520 graph that I've written about at some point in my blog that talks about how, based on anywhere from

00:25:42.280 zero to five of the characteristics that an individual has of metabolic syndrome, the proportion of

00:25:48.200 discordance goes up. So it also could be that as discordance rises, risk rises because of the other

00:25:55.320 factors such as hyperinsulinemia, which itself may contribute to intimal damage, inflammation, and other

00:26:01.720 things. Is that possible?

00:26:02.920 It is absolutely possible and likely to be true. So this brings up a pattern, a lipoprotein

00:26:08.600 pattern that I will take some credit for having to find in my own way about 25 years ago, 27 years

00:26:15.160 ago now, called the atherogenic dyslipidemia or atherogenic lipoprotein phenotype, which is a

00:26:21.080 constellation of lipid changes that includes higher triglyceride, lower HDL cholesterol. So that's the

00:26:28.600 cholesterol in the protective form of lipoproteins being deficient and a predominance of smaller LDL

00:26:35.640 particles. And so that triad, that lipid triad, has defined atherogenic dyslipidemia, and it folds

00:26:42.680 right into the metabolic syndrome.

00:26:44.040 Right. Two of those three make up two of your five criteria for metabolic syndrome, which for the

00:26:49.800 listener would be low HDL cholesterol, high triglyceride, high fasting glucose, high blood

00:26:55.320 pressure, and girth, uh, basically obesity. Exactly. Drunkal obesity. Yeah. Yeah. So I remember

00:27:00.040 being part of the discussions where that metabolic syndrome was defined.

00:27:04.680 Syndrome X.

00:27:05.720 This is very interesting. Yeah. Well, that started with insulin resistance as the centerpiece.

00:27:09.560 Uh, you can put various molecules and various, uh, processes, uh, toward the center. They all

00:27:14.440 contribute. And as you point out, uh, hyperinsulinemia associated with insulin resistance is likely

00:27:19.560 another marker of another process, uh, related to glucose metabolism and its consequences.

00:27:25.400 And then the blood pressure connection is, is an intriguing one, but that's also part of it.

00:27:30.040 But a lot of that is driven by increased girth. I mean, it's of the five conditions, the one that I

00:27:36.040 think is the most prevalent underlying factor that leads to the development of metabolic syndrome,

00:27:41.720 uh, is, is increased, uh, abdominal fat, which is associated with increased.

00:27:46.440 Specifically visceral fat, uh, visceral fat around the internal organs. And you can have

00:27:50.520 metabolic syndrome without that, but the vast majority of people that certainly Caucasians.

00:27:56.120 Yeah. Yeah. Yeah. And, and non-Caucasians, uh, populations such as the, such as East Asians,

00:28:01.080 who don't have, uh, increased waste still, still can have increased fat internally as part of the

00:28:06.120 syndrome, but it's probably acting on an underlying genetic predisposition, which is very common.

00:28:11.400 And so, and there's many other factors that come into play as part of the syndrome,

00:28:15.800 but I would say the dyslipidemia is probably both clinically and pathologically the one that I

00:28:21.240 think has the most substance in terms of a direct causal connection to cardiovascular disease.

00:28:26.760 Yeah. And, you know, I'll tell you, it's a very interesting historical footnote, uh,

00:28:31.240 Gary Taub's a mutual friend of ours. And I don't know if this actually was in any of his books,

00:28:35.400 or I might've read this in one of the outtakes, but it was an interesting footnote,

00:28:39.880 which was basically at the time of the Framingham study, which I'm talking about the very,

00:28:44.360 very first Framingham study, which was really a two-part study, of course. One of the things

00:28:48.920 that came out of that study was that low HDL cholesterol and high triglycerides was four

00:28:54.920 times more predictive of atherosclerosis than elevated LDL cholesterol. Now that rings true

00:29:02.200 with what we just said about metabolic syndrome. Nowhere in the five criteria of metabolic syndrome

00:29:06.360 was high LDL-C. It's low HDL-C and high trig. But it's interesting that the LDL cholesterol story

00:29:14.280 really took off. And at least Gary argued, I believe, if I'm remembering the argument, that

00:29:18.760 part of that had to do with the fact that the, God, I'm blanking on the name of the trial,

00:29:23.640 the first trial, there was the LRCCP, but then there was the one before it, uh, it had a funny name,

00:29:30.200 like improve it or something, but it wasn't. It was like, uh, it was a trial of fibrate, the fibrate

00:29:35.480 trial. The fibrate trial was LRCCP. No, that was, that was close to I mean. So I think that's when

00:29:39.800 you're thinking, I think you're thinking of this LRC. I was, I was, and so part of my training,

00:29:45.320 yeah, that's where I got my training in lipids before I came out to Berkeley, actually. That's,

00:29:49.560 that's how I got my first dose. And, uh, that study was going on at that time. And there was a lot of

00:29:54.680 nail biting. Well, and, and, and the argument here was, look, we kind of spent, we lost a decade

00:29:59.720 and a half between say 1980 and 1995 when we missed the role of insulin resistance, because we really

00:30:08.200 went down this LDL cholesterol rabbit hole and didn't necessarily see the bigger picture. And

00:30:13.720 one of the things I hope we have time to talk about today, because it's actually something that I

00:30:18.280 spend more time scratching my head about than anything else is you look at a drug like niacin,

00:30:23.240 which lowers APO-B LDL-C and raises HDL-C. So in theory, it's doing everything in the right

00:30:30.120 direction. And yet when it comes to outcomes, it's a very confusing picture. So maybe later on this

00:30:36.120 afternoon, we can get to that. Cause I think that's, there's going to be some rich info in

00:30:40.040 there. If you put that on the table, I'll be happy to pick it up when you're ready because I'm,

00:30:43.720 I'm ready to launch into that anytime. I cannot wait. We're, we're definitely going to do that.

00:30:47.000 So, okay. Let's talk a little bit about a paper that you were an author on this year. It was the

00:30:51.720 European atherosclerosis society consensus statement. Now you and I were joking about

00:30:56.200 this a while ago that you almost couldn't believe this paper needed to be written,

00:31:00.520 but sometimes there's a benefit in writing it. And what was the conclusion of that paper

00:31:04.120 or more to the point consensus statement? It was more than just a paper. I mean,

00:31:07.000 it was really a tour de force. So this, this paper assembled a multiple lines of evidence

00:31:12.360 addressing the question, does LDL cause heart disease? Is LDL a causal factor for heart disease?

00:31:19.560 And just to be clear, the counter argument is sure people with high LDL are more likely to get heart

00:31:25.960 disease. That can't be disputed. The epidemiology is clear. The counter argument is, but LDL is not

00:31:31.480 a causal role. That's right. And it's associated with that efforts to lower LDL cholesterol are not

00:31:38.280 fully justified as a means of attacking the cause. I don't want to be responsible for having stated that

00:31:44.840 incorrectly because I still can't quite believe anybody would hold that opinion. But that was my

00:31:51.080 understanding that led to the group coming together to counteract that perception that lowering LDL was

00:31:57.960 not beneficial. But there are, there are many people, I mean, not that I spend terrible amounts

00:32:01.880 of time on Twitter, but it's a pretty commonly held view, at least in the vocal minority that love to

00:32:08.440 write about this and talk about this, that, Hey, LDL cholesterol is a myth. Like heart disease has nothing to do

00:32:13.400 with this. And the problem is, and it did come out in the paper to some extent, but I'll tell you,

00:32:17.640 there is a second component to that effort that is still being written. It was planned and will be

00:32:23.880 a two part series. The first part is assembling all the evidence from epidemiology, clinical trials,

00:32:29.400 genetics, et cetera, that speak to the causality. And the second one was really relating all of this

00:32:36.120 information to the role of LDL in the pathophysiology of atherosclerosis. And that paper

00:32:40.920 is a work in progress, but it collectively, those two papers assemble just about all the evidence

00:32:47.880 one needs to support the use of LDL cholesterol.

00:32:51.400 And when will that second paper be out?

00:32:53.400 I can't tell you.

00:32:54.040 We don't even know.

00:32:54.600 It has taken longer than we thought.

00:32:56.520 Well, we'll certainly link to the first one in the show notes because that was published in

00:33:00.840 early 2018.

00:33:01.720 Yeah. Yeah. And we expected the next year or so. But one thing I do want to say, because there's

00:33:06.200 a caveat and part of my life as a researcher, as well as a clinician, is recognizing the complexity

00:33:13.640 of what we're dealing with. In discussions such as this, it's important to keep the concepts

00:33:19.320 straightforward and understandable to the best degree possible. But the flip side of that is the

00:33:25.240 risk of oversimplifying a complex situation. So, when I just said that the evidence is that lowering

00:33:31.880 LDL cholesterol is beneficial, that's not always true. And so, one can point, if one is so inclined,

00:33:40.520 to the evidence that under certain conditions in certain populations, with certain approaches,

00:33:45.720 lowering LDL cholesterol does not result in reduced heart disease risk. And to the extent that you

00:33:51.240 consider that to be a fatal flaw in the argument, that can be, I think, very misleading because it's not.

00:33:58.040 The fact is that LDL is causal, but there are other circumstances that modify that causality to

00:34:04.280 the extent that some forms of LDL under certain conditions, and this may not be uncommon,

00:34:09.400 can be elevated without pathologic consequences. And so, lowering LDL in those cases may not give

00:34:15.160 benefit in that. We know that there is heterogeneity in the clinical response when one looks at

00:34:20.120 cardiovascular protection with LDL-lowering treatment. So, I have to absolutely extend the simple

00:34:26.200 notion of LDL causality to saying that one has to look very carefully at the arguments

00:34:32.280 against LDL causality because they latch on to pieces of information that are really misleading.

00:34:39.480 Just because lowering LDL cholesterol is not always beneficial doesn't mean that LDL is not

00:34:44.200 pathological. And the second component of that is the focus on LDL cholesterol. That goes back to our

00:34:49.960 initial discussion here today as a marker for a causal mechanism, but it's the particles that are

00:34:56.360 causal. And LDL cholesterol, as we just talked about, does not always mirror the number of LDL particles.

00:35:02.280 Now, I don't think we should necessarily take the time to go through the paper in incredible detail,

00:35:07.240 but it did touch on eight criteria for causality. Plausibility, strength, biological gradient,

00:35:14.440 the temporal sequence, the specificity, consistency, coherence, and then the relative risk reduction or

00:35:22.360 risk reduction with an intervention. Among those, I found the Mendelian randomization to also be very

00:35:28.760 compelling. So, you know, when I talk about this with people, I generally talk about the natural

00:35:35.640 experiments such as the people with PCSK9 mutations, both hypo function or, you know, gain of function,

00:35:41.640 loss of function, PCSK9, the FH patients, the Mendelian randomization, and the intervention.

00:35:47.800 If you were going to bring up three points from the paper that you think probably are most relevant,

00:35:54.200 what would they be? Well, you've just touched on probably the number one strongest argument. And

00:36:00.440 it's really where we, those of us who have been in the field for decades, started with being impressed

00:36:07.800 with the role of genetic elevations of LDL. Very, very strong evidence. I would put that

00:36:14.280 probably right at the top. And you talked about this condition, familial hypercholesterolemia,

00:36:18.680 when there's two doses of an abnormal gene, the LDL levels can skyrocket. I referred to that

00:36:24.200 a little while ago as the condition that can lead to heart disease early in childhood. It's unequivocal.

00:36:30.680 In fact, the reason I got a little bit, I was a little bit taken aback by the need to do this

00:36:36.120 more extensive review, which I think, by the way, was quite a good exercise, both for those of us who

00:36:42.040 did it and people hopefully who read it. But all you have to do is look at an eight-year-old child

00:36:46.520 with cholesterol levels that are eight or nine times normal, who's a candidate for liver and heart

00:36:51.720 transplant to know that that's it, that's causal. But the genetics support it beyond that.

00:36:57.000 Right. Now, in those cases, the genetic defect is one in the LDL receptor, correct?

00:37:01.720 That's right.

00:37:02.440 So, closing the loop on how this works, right? The body makes cholesterol. The body, you know,

00:37:08.440 so every cell in the body makes cholesterol. Then cholesterol gets recirculated, ends up mostly back

00:37:13.400 in the liver. It gets secreted. Some of it in bile gets reabsorbed, and this process continues. But

00:37:19.400 it's this LDL clearance, mostly via LDL receptors in the liver, that seems to be where a lot of these

00:37:25.640 genetic things go awry.

00:37:27.240 That's right. Yeah, the liver really is the factory as well as the disposal plant, if you will. Most of

00:37:33.560 the cholesterol that winds up in the blood is released in terms of lipoproteins that are

00:37:38.120 synthesized by the liver. And then they come back to the liver ultimately after they've done their

00:37:45.480 thing, so to speak, delivered their cargo or interacted with cells in various ways and come

00:37:50.440 back to the liver. And a large portion of that return is mediated by these receptors that latch

00:37:56.120 on to ApoB. It's ApoB that is kind of the key that binds to the lock that snaps up the LDL in the

00:38:03.960 liver and degrades it and excretes it into bile. And that's one of the ways we dispose of cholesterol.

00:38:09.720 There are other mechanisms involving HDL, but the receptors are a key determinant and do represent

00:38:17.160 a mechanism by which most of the drugs that we use to lower cholesterol act to increase

00:38:23.800 LDL receptor-mediated disposal of LDL particles. And I'm just going to connect that concept to

00:38:29.640 something you brought up earlier, and that is the duration of exposure to, in this case,

00:38:36.360 high levels of LDL. We talked about it as a function of age. The longer the written number

00:38:41.640 of years, the longer the exposure. But there's also a dynamic aspect to LDL metabolism that we

00:38:47.720 just touched on. That is, particles are produced actually precursors of LDL, which are called VLDL,

00:38:53.240 which carry mostly triglyceride, and apoprotein B as well. Those particles are being actively secreted.

00:39:01.400 They interact with peripheral tissues and receptors and other transporters that handle

00:39:07.080 various lipids in various ways and enzymes, etc. There's a lot of processing that goes on.

00:39:12.360 And then what's left comes back to the liver through LDL receptors. Now, if that process happens

00:39:18.360 briskly, if there is a nice, fast turnover, if you will, of those particles, you can just see that there's

00:39:25.640 less time for the arteries to be exposed to any of those pathologic forms of lipoproteins.

00:39:32.120 They can be scooped up. But what underlies, at least to me, a common concept, a common underlying

00:39:39.880 factor that connects various lipid traits to heart disease risk is the extent to which they influence

00:39:46.600 the circulation time. That is, the length of time that a particle is circulating in the blood,

00:39:51.080 so that if particles are not being cleared by LDL receptors efficiently, they will circulate longer

00:39:57.960 and have more opportunities for mischief. What is the typical half-life of a VLDL particle,

00:40:04.440 so very low-density lipoprotein, IDL, intermediate density, and low density?

00:40:08.840 I'm probably going to get this wrong, and this is where I don't want to have to go back and look at

00:40:13.880 the textbook. Well, directionally, though.

00:40:16.040 Yeah. So, particularly for a larger VLDL triglyceride, the half-life is half an hour,

00:40:21.000 an hour, two hours. It's pretty rapid because that particle is rapidly subject to enzymatic

00:40:26.120 digestion. Well, let me just jump to the LDL. So, the LDL that are formed from those precursors,

00:40:32.520 you know, more like 12 to 24 hours or longer. So, it turns out that smaller particles have a longer

00:40:39.560 resonance time because they are less avidly removed by LDL receptors. So, there's a range

00:40:44.200 of circulation times for LDL, in some cases, you know, days actually, and then IDL are somewhere

00:40:50.280 in the middle. The intermediate density lipoproteins are IDL is what you're referring to. That's the

00:40:54.440 step between VLDL and LDL that is involved in their metabolism. So, it's these longer exposure

00:41:02.040 times, and these intermediate particles can include and do include very pathologic forms as well. And so,

00:41:10.120 there are disorders, and metabolic syndrome, coming back to that, is one of them, where the clearance

00:41:16.120 of those intermediate particles, which also comprise what we call remnants, partial breakdown products

00:41:22.200 of VLDL on the way to forming LDL, those intermediate particles can have much longer exposure times,

00:41:29.000 and they can be subject to various pathologic effects involving oxidation, the acquisition of

00:41:34.680 partial digestion products of various lipids that cause them to be more toxic. And those particles

00:41:40.600 can be damaging even with a shorter residence time because they're so toxic. So, that gets into what I

00:41:48.440 was referring to earlier as a somewhat greater complexity of, beyond thinking just about LDL cholesterol,

00:41:54.120 thinking about LDL particles, thinking about the types of LDL particles in terms of the pathophysiology,

00:41:58.840 and then also thinking about the role of these remnant lipoproteins. They all participate in this

00:42:04.200 potential risk. And it comes down, in my view, in the end, to the length of time that a particle with certain

00:42:10.520 pathologic effects is circulating in the blood.

00:42:12.920 And this VLDL cholesterol, VLDL remnant problem is one that is unfortunately very often missed, even by relatively

00:42:20.000 astute clinicians. You know, your textbook case is these type 3s, these hypertriglyceridemias who have

00:42:26.920 normal APOB, normal LDL-C, you think they're relatively low risk, you sort of miss the fact

00:42:33.320 that their VLDL cholesterol is 75 milligrams per deciliter, and they have these just devastating

00:42:39.960 atherosclerotic plaques.

00:42:41.600 Yeah. And that was, again, one of my learning experiences. When I was at the NIH, I was at the

00:42:46.680 NIH at a time working with Drs. Fredrickson and Levy, when the various-

00:42:52.880 Just for the listener, it's important for them, I think, to understand the luminaries,

00:42:56.800 that you just described, right? Fredrickson, Levy, and Lees. I mean, let's put these guys

00:43:00.800 in context, right? This is, you had people that figured out that there was this thing

00:43:05.440 called cholesterol, but it was really those three that did the pioneering work in the 1970s

00:43:10.960 that laid the groundwork for fractionating, figuring out all of the different subparticles.

00:43:16.800 Late 60s.

00:43:17.200 Late 60s, early 70s, yeah.

00:43:19.200 Late 60s. No, it was late 60s. I'll just give you a slight anecdote since you paused

00:43:24.560 me on this. I've always been interested in heart disease because it runs in my family.

00:43:30.080 As a medical student, I read a series of five articles that was published in the New England

00:43:34.860 Journal of Medicine, and I think you're referring to the three authors, Fredrickson, Levy, and Lees,

00:43:39.320 that completely transformed me. It was a epiphany because those five articles describe lipid disorders

00:43:47.440 in terms of genetic, subgenetic types, different lipoprotein profiles that had differing metabolic

00:43:55.000 effects, different consequences, and were influenced differently by various diets. And I thought this

00:43:59.780 was absolutely the most important lead I can imagine. So I made it my business when it came

00:44:05.720 to deciding what I was going to do after my medical training is to come back, is to come

00:44:10.100 to NIH and work with those guys. And I was fortunate enough to be able to do that. And it was, it was

00:44:14.880 really in the era where what was called the Fredrickson typing system identified these various

00:44:21.060 forms. And the one you just referred to, the genetic forms of lipids that we don't often always

00:44:27.200 consider, it was called type three. And that was characterized by abnormalities in receptor

00:44:34.040 mediated clearance of these remnant particles through a mutation in apoprotein E or variant

00:44:39.660 of apoprotein. And that also was fascinating. And I happened to be in California when the apoEs

00:44:44.080 were discovered. So I sort of feel like one of these characters that just, you know, shows up at

00:44:48.880 the appropriate time. It was called the orange-rich peptide initially. The history is just, I wrote a,

00:44:55.340 we wrote a review actually of the early history of lipoprotein research, which I would commend

00:44:59.200 to the audience here. Let's put that on the list here. It was in a journal of lipid research in

00:45:02.720 2016. In fact, I just got a fan letter for that article because anybody that's seriously interested

00:45:08.480 in this field should probably understand the origins. In fact, I got a letter after that paper

00:45:13.560 was published. I got an email from Joe Goldstein, who is the other, one of the other icons in the

00:45:18.280 field who said, he said, everybody that goes into lipid research should read this paper. So I'm going

00:45:23.180 to, it's not a book. I don't make any money off of it.

00:45:25.380 Well, we're going to make sure that the people who want to get smart on this read it.

00:45:30.440 Anyway, paying back to these remnants, I think particle for particle, the remnants are probably

00:45:34.400 the most pathologic particles of all because of this rampant atherosclerosis when there's

00:45:38.700 elevation of remnants. You know, fortunately, this condition of type three is fairly rare. It's

00:45:44.160 like one in 10,000, but it does illustrate that.

00:45:46.520 I've seen biology.

00:45:47.600 Yeah, I've seen several.

00:45:48.740 You've seen several.

00:45:49.420 Right. So what I was starting to say, I actually, initially, what I started to say when I was at NIH

00:45:53.540 in my training, I saw all these things, you know, all the types were there.

00:45:56.140 You collected them.

00:45:56.880 Yeah. We had patients who had the type three, we had type one, which was a serious elevation

00:46:02.000 of triglyceride. Type two was familial hypercholesterolemia. It goes on and on. So, yes. And it's really

00:46:09.360 striking. I tell my students, I lecture on this to students at Berkeley, and I show them pictures

00:46:15.840 of what are called xanthomas, which are, you know, deposits of cholesterol and lipids and tissues.

00:46:21.300 And you have these very characteristic lesions that used to be rampant in patients before we had

00:46:27.840 adequate recognition and treatment as a manifestation of the underlying pathology that also affects the

00:46:34.040 arteries. The cholesterol that winds up in the arteries can also break into the skin. And these

00:46:38.700 are just very striking illustrations of the role of genetics. So it gets back to your earlier question

00:46:44.140 about causality. There are so many situations where genetics helps in establishing causality.

00:46:50.780 And this certainly is one of them. After that, you asked me what are, you know, one, two, and

00:46:56.360 three, right? You asked me how many, what my top three are. I didn't answer that. And before we

00:47:00.680 leave FH, I think the other nice thing about the PCSK9 mutation, you know, FH is only showing you the

00:47:07.400 change in one direction. But with PCSK9, you see both directions. That's right. I believe the

00:47:13.620 hyper-functioning were the first people identified, correct? Yes. So these people had an enzyme,

00:47:19.980 PCSK9. It hyper-functioned. Therefore, this enzyme's, one of its roles is to degrade the LDL

00:47:27.400 receptor. And so they had fewer LDL receptors. They had more LDL. They looked a lot like FH patients,

00:47:33.540 correct? Yeah. Yeah. And just to be, again, for the technical, who was on the phone? Oh,

00:47:38.420 it's actually not, it's not an enzyme. It kind of behaves like an enzyme, but it actually

00:47:42.300 drives LDL receptors into the garbage disposal machinery in the cell called lysosomes. And it

00:47:48.820 causes the LDL receptors to be broken down. But it's the same end result as you get less LDL

00:47:54.120 receptors, higher LDL cholesterol. And then the mutations in the other direction are the ones

00:47:58.460 that led to the development of PCSK9. The loss of function, antibodies to PCSK9 mimic the loss

00:48:05.020 of function mutations, which lower LDL. And that's one of my, you know, when I get to tell

00:48:10.580 these stories one day, I'll look back and say, that was my aha moment. Cause I think that paper

00:48:14.240 came out in 2006 in the New England Journal of Medicine. It was either 04 or 06, but I remember

00:48:18.340 this well, which was the discovery of those families with the hypo-functioning PCSK9.

00:48:24.500 These people walked around with an LDL cholesterol between 10 and 20 milligrams per deciliter.

00:48:29.440 Two things about them stood out. The first, they never got heart disease. The second,

00:48:34.620 they didn't seem to suffer any other consequences that you might concern yourself with.

00:48:39.400 Right. Yeah. One of those patients, classically, was I think a physical therapist or somebody that

00:48:48.660 was very active and had an LDL of in the teens, which is, you know, one-sixth normal and is doing

00:48:56.180 fine. Yeah. So that's another use of genetics in a way to confirm that lowering of LDL, not necessarily

00:49:04.600 using all the ways that LDL can be lowered, but at least certainly that form of LDL lowering is

00:49:10.700 healthy. And probably most forms of LDL lowering are healthy. We have very little evidence to the

00:49:15.900 contrary. Do you want to say anything about the Mendelian randomization? It's, it's, it's, it's,

00:49:20.580 you know, I think it's not the most intuitive concept to people, but it is actually a very powerful

00:49:24.880 concept. So the principle of Mendelian randomization, first of all, Mendel, Gregor Mendel was a 19th century

00:49:32.240 monk who discovered the principle of inheritance of traits in peas, actually. But the idea is that

00:49:39.640 these genetic variants are randomly distributed in the population. So that assumption underlies

00:49:45.840 this concept of Mendelian randomization, because then you can say that the occurrence of a genetic

00:49:52.920 variant in the population or, or a collection of variants can be either single or multiple variants

00:49:58.400 that are associated with a biomarker, such as LDL cholesterol can be used to test the causality of

00:50:06.820 LDL by looking at another relationship. And that is the association of those genetic variants with the

00:50:14.340 disease process. So for example, and so I'm going to give you the example because it's really a little

00:50:20.760 bit abstract without an example, and we'll talk about PCSK9. So the PCSK9 loss of function mutation

00:50:27.160 causes a lowering of LDL. There is independent evidence that that mutation is associated with

00:50:33.700 reduced cardiovascular disease risk, nothing to do with LDL, just the genetic association with

00:50:38.880 outcomes. That relationship parallels very closely the relationship of that variant to LDL cholesterol.

00:50:46.460 The difference is that the risk associated with a genetic variant, the risk of heart disease

00:50:51.800 is actually less than would be predicted from the LDL cholesterol using standard risk relationships,

00:50:57.580 because this is lifelong exposure. This gets back to the exposure issue. So, so a genetic marker like

00:51:03.180 this in the Mendelian randomization model tells you that lifelong exposure to a genetic variant that

00:51:09.420 either raises or lowers risk has, um, effects that can be attributed to LDL because that the LDL change

00:51:17.460 is, it predicts that, that risk relationship. Yeah. Whereas a lot of the conventional risk models are

00:51:23.280 basically looking at maybe a decade of risk or something like that. And they're always going to

00:51:27.020 fall short both under and overestimating long-term risk. So in the paper that you referred to that,

00:51:31.600 uh, that came out of this European consensus, there was a heavy dose of Mendelian randomization. Uh,

00:51:36.760 one of the, either the main author or one of the key authors did a very good job of, um, uh, of showing

00:51:42.820 how the genetic markers, uh, for, uh, that are connected to LDL receptor levels, both. So PCSK9,

00:51:50.140 uh, others that are associated with higher, uh, receptor activity as well, predict cardiovascular risk

00:51:57.280 much more robustly than do the results of clinical trials. Clinical trials only last four or five,

00:52:03.100 six years. Yep. So that risk reduction, which is parallel, but is displaced because the magnitude

00:52:08.920 of that effect is, is blunted because it's not a lifelong exposure. So it's a very instructive.

00:52:13.940 Yeah. Let's go back to something else you said a few moments ago that I think is, I would say 10

00:52:20.480 years ago. I don't remember when Jim Otpost's analysis came out, but Jim, who's an incredibly

00:52:27.060 thoughtful person in this field, wrote a paper that basically said, once you normalize for the number,

00:52:33.100 of LDL particles, the size doesn't matter. Now, a moment ago, you said that your intuition is that

00:52:40.580 that actually that's not correct. That particle for particle, a small particle is more atherogenic.

00:52:45.860 Is that a fair assessment of your thought? It is. It is definitely an assessment of my thought.

00:52:51.560 And unlike what we've just been talking about, we don't have a way showing incontrovertibly that

00:52:57.580 that's true. So when tries to sort of dance around this question by using statistics,

00:53:03.540 can you use statistics to factor out everything with which the particle size? And let me clarify

00:53:09.540 one other thing before I go on. And that is the reason particle size got on the map was that I wrote

00:53:14.840 a paper, I think in the eighties, actually, probably 30 plus years ago, in which the only test we had was

00:53:21.440 particle size measurement. And we showed the particle size, small particle sizes related to risk,

00:53:25.280 but we also said it was associated with lower HDL and higher triglycerides. So that was the

00:53:29.480 definition of this triad that we just talked about. And we never said that the particle size

00:53:33.900 was independent. We never actually said that. We never said it was related to independently to

00:53:37.880 risk. It was a marker for this whole syndrome. The particle size context that I was referring to

00:53:43.820 early on in our discussion today was not the size of particles, but the numbers of particles

00:53:49.540 of differing sizes. And that's a somewhat different concept. So there are techniques that measure,

00:53:55.560 that give a number for whether most of the LDL in the blood is larger or smaller.

00:53:59.380 And we typically bifurcate this at like something like 20.5 nanometers or something like that.

00:54:03.820 And the other thing I'm going to say, because this is law, I'm going to say it because there's an

00:54:08.140 opportunity to say it. If you measure those particle sizes correctly, the distribution in the

00:54:13.680 population is bimodal. That means there is a discrete subset of the population that has smaller

00:54:21.060 LDL particles. Now that says nothing about their heart disease risk. That says that there's something

00:54:26.100 going on that tips in a quantum way towards the small LDL trait. And that is the marker for the

00:54:34.460 metabolic syndrome. So that's not about heart disease risk. That's a metabolic marker. The heart

00:54:39.340 disease risk depends on the magnitude of that small LDL mode. If you have a lot of LDL particles

00:54:46.900 that are small, that's bad. Now, the argument that Atvos and others have made using statistics that

00:54:53.600 I will come back to in a minute as to why I think this is a flawed approach is that if you knew certain

00:55:00.420 statistical corrections for interrelationships of various particles with each other, there is a

00:55:06.320 significant relationship to risk of larger LDL as well. So that's not untrue. A larger LDL can be

00:55:12.880 pathologic as well.

00:55:13.880 I mean, if I recall, as long as something is less than 70 nanometers, it can enter the subendothelial

00:55:19.080 space. So any small, any large or small LDL can enter the space.

00:55:23.640 So it gets back to residence time.

00:55:24.880 Yes. How long do they stay? How likely do they be retained?

00:55:27.000 So it turns out, so this is going to be a little bit complicated, but I'm going to try,

00:55:31.940 maybe you can help me.

00:55:33.060 No, no. We have an audience that's willing to handle complicated.

00:55:35.780 Help me work through this with you because I know what I want to say and I just want to make

00:55:39.640 sure that I say it clearly. Let's talk about larger LDL. So in this large LDL mode, that signifies

00:55:47.060 sort of the flip side of metabolic syndrome. It generally identifies people who have not only larger LDL,

00:55:54.100 but higher HDL cholesterol and lower triglyceride. So that's a low risk syndrome.

00:56:00.180 And often low insulin.

00:56:01.760 That's right. And so quite apart from the question, are these particles better or worse?

00:56:07.480 They signify a metabolic profile where there's a pretty brisk circulation of those particles and

00:56:15.180 the exposure to the artery wall is very low. Now, if you have an LDL receptor defect, it also

00:56:21.320 causes large LDL. Large LDL accumulates.

00:56:23.420 Right. FH patients have large LDL particles.

00:56:25.880 Right. So why is that bad? It's because of the residence.

00:56:28.880 The residence time is much higher.

00:56:30.380 Right. And so neither I, well, there's some people that have taken some of my own work

00:56:35.720 and taken it to an extreme that I don't feel is justified. And that is to say that large LDL

00:56:41.120 are not atherogenic at all. They certainly can be. There's no question about that. The question

00:56:47.180 is, are they equally atherogenic to smaller particles or any other particles in the LDL

00:56:53.940 spectrum? And I'll just say this, the work that you're referring to did not address that. All it

00:57:00.300 said was that you can show that this large LDL are associated with risk. And then the second thing is

00:57:05.740 that if you adjust total LDL particles for the peak size of LDL, the peak size is not associated

00:57:13.920 with risk. So those are two different statistical manipulations, neither of which disprove the

00:57:19.360 hypothesis that smaller LDL carry more atherogenic risk. And we have an example of something I'm going

00:57:26.240 to tell you again genetically, which may or may not be something that your audience is familiar

00:57:32.280 with. There's another genetic syndrome that involves a variant that affects a region of the

00:57:39.640 genome, which is responsible for synthesizing a protein called SORTILIN, S-O-R-T-I-L-I-N.

00:57:46.880 That genetic variant was discovered probably six or seven or eight years ago now, and was associated

00:57:52.420 with both high risk of cardiovascular disease and high LDL cholesterol. And in fact, the association

00:58:00.820 of that genetic variant with cardiovascular disease was as strong, if not slightly stronger,

00:58:07.180 than the associations of genetic variants and the LDL receptor itself. So this was a new player in the

00:58:14.520 spectrum of causal factors, again, relating a genetic mechanism that raises LDL to an effect on

00:58:22.220 cardiovascular disease risk through a pathway that doesn't involve the LDL receptor. This is not an

00:58:27.540 LDL receptor story. What we published as part of the initial description of this variant and its

00:58:34.040 relationship to cholesterol metabolism is that in two independent populations using two independent

00:58:39.560 methods, it's specifically associated with small and very small LDL, not large LDL at all. So here's a

00:58:45.620 genetic variant that as far as we can tell affects... So it's not affecting clearance.

00:58:50.660 Well, we don't know about clearance. I can't tell you about clearance. All I can say is it's not LDL

00:58:54.920 receptor-mediated clearance. There may be... We don't know. This is...

00:58:59.100 It could be Neiman-Pixy, one like-one clearance for all we know.

00:59:01.800 So if you're going to ask me later on what one of the experiments that I would do if I had all the

00:59:06.600 resources in the world that would relate to this mechanism, because the genetic association,

00:59:11.760 it's not clear exactly what's being affected by the genetic variant. It's not clear how it works.

00:59:17.200 But what's definitely clear is that that variant is associated in terms of lipoprotein changes or

00:59:24.520 anything else that we can measure in the usual risk factor range, exclusively with small and very

00:59:29.720 small. It's even a subtype of small LDL, but it's that collection of smaller particles that is somehow

00:59:36.400 affected by this variant.

00:59:37.680 Now, do those patients have elevated triglycerides or low HDLC?

00:59:40.560 No. No. It's exclusively...

00:59:43.180 So that counters the argument that I would say, gosh, maybe these small particles ultimately are

00:59:48.640 just a marker for an inflammatory metabolic dysregulation.

00:59:53.780 They largely are.

00:59:55.020 But this example...

00:59:56.100 This is a subtype. So this gets into what you might consider a nuance. I'll tell you another

01:00:00.900 anecdote. When I first kind of discovered that people had all these different forms of LDL,

01:00:05.860 this is again in the 80s. I don't know how many... This is 35 years ago now. I was invited to give

01:00:10.400 a couple of talks, actually, at various meetings, one of which was in San Diego, actually, and

01:00:15.760 others where I presented this data. And it was using not the current methodologies, but a very

01:00:21.000 elaborate procedure involving the ultracentrifuge, which separates these particles into various

01:00:26.580 fractions. And I have a picture of what I showed on the wall of my office because it's so emblematic

01:00:32.120 of the existence of these discrete forms of LDL. And I remember talking about this to very

01:00:38.740 intelligent and experienced people in the field. And it was considered esoteric. And

01:00:44.100 for about 15 years, nobody paid any attention to it because it was felt... Nobody else had

01:00:49.440 methods to show what we had been showing in large populations. Fortunately, later on, that

01:00:56.140 was remedied by more widely available methodologies that we were partly responsible for.

01:01:03.040 But the bottom line is that the recognition of these various forms of LDL, we tended to simplify

01:01:10.100 to avoid having people think it was too esoteric. So we talked about large and small as if there

01:01:15.460 are only two forms and these two modes, et cetera. And all of that's true. But within the small

01:01:19.960 LDL mode, within both of those, but the small LDL, there's yet another subtype. And it's this

01:01:26.040 very small LDL. So the garden variety small LDL that is generally measured by techniques that are

01:01:34.620 being used, such as NMR and ion mobility method, largely measures the small LDL that you're talking

01:01:40.560 about that's part of this generalized metabolic syndrome. But this very small guy looks like

01:01:45.360 it's another pathway. It's another pathway. And it's a pathway that has a strong genetic

01:01:50.860 association with risk. And we are, I'll just say this to this audience, we are trying to do some

01:01:57.540 studies, and this gets back to the studies I'd love to do, to test the hypothesis. These particles

01:02:02.400 may be secreted directly. So this might be a mechanism that spits out a pathologic form of LDL

01:02:08.840 without going through the VLDL pathway.

01:02:11.220 Without going through the VLDL pathway. That's a hypothesis that we're now testing.

01:02:13.800 So it sounds like almost like an LP little a type issue.

01:02:16.760 In a way it is. Yeah, that's a good point. It's a particle that we don't know how to lower.

01:02:21.340 Like LP little a. Well, we were assigned to learn how to lower. That's another topic, I guess.

01:02:25.580 We're not going to get to LP little a today. That's another topic. But yes,

01:02:28.340 it's a genetic factor that is associated with risk that we don't yet know what to do with.

01:02:32.760 So let's go back to something that you've kind of touched on a little bit, which is

01:02:37.360 can LDL cholesterol slash LDL particle slash ApoB be too low? And I'm referring specifically to a

01:02:45.820 pharmacologic intervention. So I think we've already established that the people

01:02:50.140 so genetically blessed to have hypofunctioning PCSK9 seem to be completely fine. But if someone

01:02:57.560 came along and said, look, I'm walking around at the 30th percentile of the population,

01:03:01.820 I want to walk around at the first percentile of the population. I'm going to pharmacologically

01:03:07.320 lower it. Right.

01:03:08.300 Is there a downside?

01:03:09.580 Right. So I guess I'm hung up on genetics today because I think the best scenario to consider in

01:03:16.880 evaluating the pros and cons of very low LDL are genetic syndromes associated with very low LDL.

01:03:22.460 We just talked about one of them. So PCSK9 loss of function mutations are an example of what you're

01:03:27.880 asking. Those individuals have lifelong exposure to very low LDL. And as far as we can tell, do fine.

01:03:33.600 There are people who have abnormalities in the ApoB protein that results in impaired production of

01:03:40.980 LDL ultimately. And most people do very well. Not only have less heart disease, they tend to live

01:03:46.720 longer. So this is the genetic evidence for the safety and the benefit really of having very low

01:03:52.180 LDL. That doesn't imply that we can extrapolate those genetic observations to all treatments. Now for

01:04:00.500 the PCSK9 inhibitors, if we assume that the use of the antibodies to lower PCSK9 therapeutically mimic

01:04:09.240 the genetic effect, then one would have the same confidence that this would not be hazardous. We

01:04:15.320 don't yet know that there are not other effects of these antibodies. It may not necessarily fully

01:04:19.140 mimic the genetic effect. But by and large, I think it is a vote of confidence that those treatments

01:04:24.520 that lower that particular treatment with PCSK9 inhibition probably does not only lower risk,

01:04:30.360 but have no significant downsides related to the LDL lowering. But we don't know the clinical trials

01:04:36.480 of any of the drugs that lower LDL have not been long enough to know what the lifelong effects might be.

01:04:43.260 So if you look at the most widely prescribed class of drugs for lipid lowering, it's obviously going to

01:04:48.220 be statins. And statins really do two things. You know, they have a direct effect, which is they

01:04:53.040 inhibit the first committed step of cholesterol synthesis. And so that directly lowers the burden

01:05:00.140 of cholesterol, thereby lowering the burden of lipoprotein. But in many ways, their indirect effect

01:05:05.200 is at least as strong, which is the liver in response to this upregulates the LDL receptor and you get

01:05:11.760 enhanced clearance. Now, the latter, we certainly have a genetic model to look at. Do we, for the

01:05:19.640 former, do we know of people who have deficient cholesterol synthesis? I mean, outside of the

01:05:26.300 extreme, we know that there are certainly inborn errors of metabolism that are uniformly fatal.

01:05:30.780 Smith on the opiates.

01:05:31.400 Yeah, yeah, yeah. But sort of outside of those people, are there people walking around with low

01:05:35.500 cholesterol where the defect is in cholesterol synthesis that would give us confidence that,

01:05:40.780 hey, inhibiting cholesterol synthesis can't be that bad?

01:05:43.580 I'd have to say, I don't know that there is such a genetic variant. There are genetic variants

01:05:48.840 in the rate-limiting enzyme, HMG-query reductase, that is the target of statins that affect LDL levels

01:05:55.060 and heart disease risk in the expected direction. But those variants have a modest effect size. They are

01:06:02.300 not big-time players to knock down LDL to those same rate levels. That's what makes the PCSK9 story so

01:06:09.920 exceptional. There's just nothing quite like it. So the answer is, to my knowledge, no, we don't have

01:06:15.640 that kind of evidence that would apply to very low LDLs that are induced by genetic factors.

01:06:24.080 And that's a scenario that I think I clinically struggle with. And I suspect there's going to be

01:06:28.520 at least one other person listening to this that's going to share that struggle, which is,

01:06:32.120 I do get a little bit nervous when I have a patient whose risk of atherosclerosis is so high.

01:06:38.760 For example, a patient with significant family history and a very elevated LP little a, just as

01:06:43.940 an example. So you've got, you know, and you know, I seem to collect these patients. So you've got these

01:06:49.380 folks and they've got a clinical burden of disease. So they've, you know, they've, their CT angiogram

01:06:54.940 shows soft plaque, their calcium score shows that they've got calcifications, their LP little a is

01:06:59.140 through the roof. And they're tolerating their statins, meaning they don't have the myalgias or

01:07:04.680 CK elevations or any of those things. But to get their LDL where it needs to be in a patient like

01:07:10.500 that, I'm going to put to the fifth or 10th percentile. I have effectively, by all means,

01:07:15.720 that I can measure almost shut off cholesterol synthesis. In those patients, I panic because of

01:07:21.780 a couple of papers that I've seen that look at the opposite end of the spectrum, which is,

01:07:25.940 you know, they look at markers of cholesterol synthesis in patients who are medicated and then

01:07:31.540 the risk of dementia. In particular, there's a paper that looked at desmosrolol levels and it found

01:07:37.220 that if the level was below 0.5, which is, you know, generally very low on the scale we look at,

01:07:43.180 and they use that as a cutoff on the receiver operating characteristic curve, the area under that

01:07:49.740 curve, which again can vary from somewhere between about 0.5 to 1, 0.5 meaning it's a useless test,

01:07:55.640 it's a coin toss, one is a perfect test. You know, they're coming in with AUCs of the ROC at 0.87,

01:08:03.260 0.89. That's quite suggestive of this. And certainly biochemically, there's a plausibility

01:08:09.880 to this, right? We understand that every tissue in the body has the ability to borrow cholesterol

01:08:15.440 from elsewhere. That doesn't appear to be the case in the brain. The lipoproteins don't seem to,

01:08:19.920 you know, be able to traffic across the blood-brain barrier. So I guess that's just one area where

01:08:24.460 I certainly don't know an answer, but I've become, I think, clinically much more quick to move people

01:08:31.020 to PCSK9 inhibitors when I get uncomfortable with the degree of cholesterol synthesis. Do you think

01:08:36.780 I'm paranoid? You've opened up a big topic all by itself, I think, and that is the off-target effects

01:08:44.480 of cholesterol. If we consider the target, that's not even the proper term. It's the off-tissue target

01:08:50.160 because it's the tissue targeting of statins to the liver, inhibiting HMG-covid reductase in the

01:08:56.920 liver. That is the therapeutic goal. And that's right. That's actually all you're really trying

01:09:02.640 to do. That's right. But we can't, we have to hit all of these peripheral tissues as well.

01:09:06.440 Oh, that's right. So there's something called pharmacokinetics. So for the last 16 or 17 years,

01:09:11.300 I have been leading a program in studying statin pharmacogenetics, which is...

01:09:16.160 I just want to explain this again, because I know you and I are sitting here in the discussion,

01:09:19.100 we understand this, but I want to make sure the listener understands the point you just made.

01:09:21.640 It's so important. In an ideal world, a statin would be a dream drug if it only inhibited

01:09:29.180 cholesterol synthesis, meaning HMG-covid reductase activity in the liver, such that the liver would

01:09:35.380 upregulate and you wouldn't impact peripheral tissue metabolism, cholesterol metabolism, for example,

01:09:41.540 in the brain and the muscles, et cetera. Unfortunately, that's not the case. So I'll let you continue.

01:09:45.840 Right. So this has to do with at least my exposure to this set of issues through the world of

01:09:53.680 pharmacology. So I have been hanging out with pharmacologists for the last 15 years through

01:09:58.200 this pharmacogenomics program. It's obvious even without that experience that factors that affect

01:10:04.020 the disposition of statins, like any other drug, are important determinants of clinical outcomes.

01:10:09.460 So the disposition, that term refers to getting it to any of the tissues, but you want to get

01:10:18.080 statins to the liver. For the most part, statins are very efficiently removed by the liver. So

01:10:23.280 fortunately, so that tends to offset some of the concerns that you have. However, there is variation

01:10:28.600 in the genes as well as other factors that affect a statin disposition. The nature of the statin itself,

01:10:36.860 its chemical composition, the presence or absence of certain genetic variants, one in particular that's

01:10:43.060 been well studied, that affects the amount of statin in the blood that prevents it from getting

01:10:47.500 at the liver. Under those conditions, there is a greater likelihood that the statin will wind up

01:10:51.760 somewhere else. And that can be muscle, which is the most common symptom, but it can be all the other

01:10:58.280 tissues. And part of my research experience right now is delving in to all of these other effects that go

01:11:06.160 beyond the desired inhibition of HMD-chlorid reductase in the liver that are actually, quote,

01:11:13.960 on-target effects. So I'm going to come back to the brain in a minute because I've been extremely interested

01:11:18.800 in the issues that you described, and I'll come back to that. But there's a condition that's even more,

01:11:24.900 I think, clearly connected to an unexpected effect of statin, and that is its tendency to

01:11:31.600 increase blood sugar and increase the risk of diabetes. That's been demonstrated now.

01:11:37.420 What's the relative risk?

01:11:38.560 It's about, it's, you know, so if you talk to cardiologists, they'll say the relative risk is

01:11:44.080 small because the benefit is much greater. Well, that risk is about 10% on average. We have published

01:11:49.640 a paper.

01:11:50.040 Over what period of time?

01:11:51.580 Over, you know, duration of the clinical trial.

01:11:53.740 So probably about five years.

01:11:54.260 Six or seven years. Yeah, up to six or seven years. But we have, we've published that that

01:11:58.040 risk can be significantly higher in women than men, perhaps as higher as 30% or plus.

01:12:03.340 And do we think that that effect is due to dysregulated glucose uptake in the muscle?

01:12:09.380 There is evidence, and we're deeply involved with these studies, of direct pathologic effects

01:12:15.600 on muscle energy metabolism. That's more closely connected with, obviously, with muscle symptoms

01:12:22.140 than with diabetes. But insulin resistance is certainly another factor that could be involved,

01:12:29.260 and the muscles could be involved, the liver could be involved. So that's one situation where

01:12:33.780 there may be on-target effects, even in the liver, that might contribute to this. We think it's probably

01:12:38.900 not liver. We think there's probably effects either beta cells on cells that produce insulin

01:12:45.280 and or the tissues that insulin acts on, such as muscle.

01:12:50.280 Probably the muscle being the most important.

01:12:52.000 Right. Yeah, right. So it's, the reason I haven't been more specific than that is we really don't

01:12:56.120 know. There's a number of theories, all of which sort of collectively could be true in different

01:13:01.580 individuals. But the net effect is not trivial. It's...

01:13:05.800 Is it dose-dependent?

01:13:07.140 Yes. Well, there are some evidence that it's just a glycemic effect that's dose-dependent. In fact,

01:13:12.720 this was very limited information on this, actually. But there is some...

01:13:17.220 Are there some statins that seem to be... So, for example, when you look at simvastatin,

01:13:21.580 it seems to have a much higher incidence of myalgias or CK elevation.

01:13:25.000 When we look at the entire suite of statins, do we see some that seem higher risk for diabetes,

01:13:31.380 some that seem lower risk?

01:13:32.380 Yes. And again, this is sort of a collection of observations from various sources. One of the

01:13:38.400 statins that's most recently been introduced is something called pitavastatin.

01:13:43.140 Livolo.

01:13:43.700 Yeah, Livolo. That group in France has shown pretty convincingly recently that this is not associated

01:13:50.440 with diabetes risk. This is kind of my go-to statin before I move to a PCSK9 inhibitor.

01:13:56.640 This is my last line statin. It's not that potent, as you know.

01:14:00.060 That's right.

01:14:00.440 That's the problem. You sort of, you know...

01:14:02.000 That's right.

01:14:02.440 But once in a while, you'll save someone. You'll get a guy who can't tolerate anything,

01:14:06.900 but he tolerates this, and you're off to the races.

01:14:09.220 Same here. Exactly. So, it's probably the lower potency, and it's probably its chemical

01:14:14.040 characteristics. Again, what I'm saying right now is still not established in a conclusive way.

01:14:20.440 But it does suggest that there is differences in the statins, and that being one that has

01:14:24.540 less association with risk. And the one that's probably most commonly associated with risk,

01:14:27.900 and we've seen this in studies that we've not yet even published all of them, is a torvastatin

01:14:32.760 appears to be...

01:14:34.000 Lipitor by its...

01:14:35.280 Yeah, have a higher risk. So, there's interesting...

01:14:38.080 And we don't think that that's just due to the fact that it's so ubiquitous.

01:14:41.500 No.

01:14:42.280 So, would that be a reason? Because I got to tell you, I think when I'm confronting a patient

01:14:47.060 for the first time with a statin, I am generally almost tossing a coin between Crestor and Lipitor

01:14:53.140 as a first-line agent, very quick to flip between them if I see any CK bump or LFT bump. But what

01:15:00.520 you're saying would almost suggest that if it's a person who's not incredibly insulin-sensitive,

01:15:05.380 where none of this probably matters, someone who's borderline, you'd lean towards Crestor over

01:15:10.460 Lipitor?

01:15:10.860 Yeah. This is sort of putting on my clinical judgment hat rather than my scientist hat.

01:15:16.660 No, no. But unfortunately, for many of us, we still have to make decisions.

01:15:21.840 You like to use evidence-based criteria, but this is actually one of the things that I struggle with

01:15:27.120 with evidence-based criteria. We don't have the evidence that allows me to give an answer to that

01:15:31.800 other than saying, yes, I agree that that's what I would do. And I'm also going to back up a little

01:15:35.920 bit and make sure that I have not conveyed the impression that atorvastatin is a diabetogenic

01:15:42.240 drug. It's still a minority of the population, and we think there's genetic factors that contribute

01:15:47.440 to that.

01:15:48.140 And we can measure it. I mean, that's sort of the other thing that I sort of tell patients is that

01:15:51.620 this isn't going to sneak up on us one day we wake up with diabetes.

01:15:54.520 I use plenty of atorvastatin, and I'm at the least bit concerned about it in the patient whose risk

01:15:59.560 merits statin treatment. And because, as the cardiologists are quick to point out,

01:16:04.300 the benefits of statin treatment with any of the statins, cardiovascular risk in patients

01:16:09.100 with diabetes far outweighs the risk of actually developing diabetes. And there's even some

01:16:14.660 evidence that the microvascular complications might be improved. So you can argue that statins

01:16:19.520 are not causing a damaging effect through this mechanism. But it does raise just a little bit

01:16:26.060 of a caution. In fact, it's a caution, I think, that should lead to more widespread monitoring

01:16:31.960 of glucose on statins just so that, like, one can detect those individuals who may have an

01:16:36.760 adverse effect. But it's still a minority of the population, and it's not something that should

01:16:40.780 be considered a hazard of drug use. Far from it.

01:16:43.400 Yeah, I think for me, I am, I mean, I sort of, you know, talk to patients and I say, look, I think

01:16:47.200 there are short-term things that we're generally going to figure out in three months, which is

01:16:50.880 myalgias plus or minus a CK elevation. So either, you know, your muscles are going to get sore with

01:16:55.180 or without an elevation in CK, which is a way that we can measure the breakdown of muscle and changes

01:17:01.020 in your liver function tests. We look for those elevations. I don't know about you, but I see a

01:17:04.780 lot more LFT bumps when combined with Zetia than just statin alone. I find a lot of patients I have

01:17:11.160 that tolerate any dose of a statin, then you add a Zetia. It seems to me like 20% of people just

01:17:16.280 immediately have an LFT bump. I haven't seen it that often, but I have, I have seen it, yes.

01:17:20.800 Um, and then I talk about the long-term stuff, which is actually, I think in many ways, what

01:17:25.300 we should be more concerned with because the short-term stuff, like you, you figure that out

01:17:28.920 in 10 seconds, right? But it's the, and the diabetes doesn't worry me as much for all the

01:17:32.840 reasons you've said, which is it doesn't sneak up on you and you can measure the progress,

01:17:36.640 but for maybe I'm being overly cautious and I've had many an argument with many, a cardiologist

01:17:43.140 that I share patients with who, you know, will have a patient on a maximum dose of a statin.

01:17:48.020 And, you know, I'll give you one example. I've got a patient who came to me on 80 milligrams

01:17:51.820 of Lipitor, still wasn't quite at goal. We added the 10 of Zetia. So then he was at goal

01:17:57.720 and given his burden of disease goal for this guy is about a 700 nanomole per liter of LDLP,

01:18:04.360 but you know, he had no cholesterol synthesis that we could speak of. So put him on a PCSK9 inhibitor

01:18:10.940 that took his LDL down to like 200 nanomole per liter. So I said, great, let's back off

01:18:16.200 the Lipitor and his cardiologist just thought like I would, this was malpractice. And, you know,

01:18:22.860 it took many a discussion to just even get that back to 40. And my goal is to hopefully get them

01:18:28.220 down to maybe 20 of Lipitor so that I can actually see some cholesterol synthesis come back.

01:18:33.940 That's interesting.

01:18:34.460 But again, now we're also a little outside of evidence-based medicine and we're,

01:18:39.080 this is more the art than the science.

01:18:41.020 That's right. And it's a fascinating issue to deal with because we are conducting an experiment

01:18:45.800 in the global population, certainly in the U.S., that has never been done before. And that is

01:18:51.340 prescribing statins to millions of people as a lifelong treatment without knowing what the

01:18:58.260 downstream effects are beyond the clinical trial data that we have, which is limited. And we can,

01:19:04.720 you know, call in genetics, but it doesn't necessarily mimic the effects of using a statin drug

01:19:11.040 or, you know, for 40 years. And so I'm just going to say very briefly that I don't want to

01:19:17.420 open the lid on the discussion that many of the naysayers have used to say that statins should be

01:19:24.580 avoided because they can have long-term effects. But I will say...

01:19:29.080 That's like saying driving should be avoided because driving has some negative effects.

01:19:32.760 Right. But there are things that we don't know that certainly I would like to learn,

01:19:38.340 and hopefully our own research project will contribute to this, as to what may be affecting

01:19:44.800 certain subsets of the population with prolonged use by understanding the mechanisms that might be

01:19:51.300 operating in tissues like the muscle, pancreas, and the brain. And so that by understanding those

01:19:57.400 mechanisms and perhaps developing markers for people that may have increased likelihood of

01:20:03.440 these undesirable effects that we may be able to guide our treatment more effectively, that's really

01:20:07.940 the goal of this pharmacogenetics project.

01:20:10.020 Do we understand the mechanism of the myalgias? I tend to give patients ubiquinol, but honestly,

01:20:16.060 the trials are... And I explained to them, I said, look, I use it as a practice, but I can't point to

01:20:21.820 amazing data. What do we think is going on there?

01:20:24.720 We just talked earlier about the European consensus group that I was on. Well, there was a second one

01:20:30.380 that I was on, and it was...

01:20:32.580 Staten side effects.

01:20:33.560 Staten side effects. And so the first paper in that series came out a year or so ago, and that was

01:20:39.220 on myopathy. The second paper is, I think, just out, and it describes the data as related to diabetes,

01:20:47.340 cognitive function, et cetera. The first paper attempted to address the question you asked,

01:20:51.440 among other things, and that is what's going on here. And all we have is a diagram that has many

01:20:56.660 points of attack, where one of the things that we're actually working on, one of my colleagues in

01:21:01.760 my research program, is mitochondrial targeting of statins. There may be on-target effects. That is,

01:21:09.920 a normal response to statins in some people, or maybe in a significant percentage of the population,

01:21:16.940 may affect mitochondrial function in ways that generally are not clinically important,

01:21:22.060 but which in some people could be magnified and lead to changes in muscle function and muscle number,

01:21:29.280 muscle cell number. That's a hypothesis that I'm speaking to now is the lack of a clear single

01:21:37.120 mechanism that we can point to. That's just one of several possibilities. The effects on a coenzyme Q,

01:21:43.840 ubiquinol, which you mentioned, has certainly been out there. But as you say-

01:21:49.260 Very high placebo effect.

01:21:50.560 Trying to reverse that, that's right, has not been that successful. One of my patients,

01:21:54.440 I said, doctor, actually, I'll say this for whatever it's worth. This is one of the sort of

01:21:58.340 clinical pros. I have a doctor who's really a very good observer who wound up getting a liquid form of

01:22:03.840 coenzyme Q as opposed to a capsule. A capsule didn't work. His muscle symptoms dramatically disappeared

01:22:10.240 or it was improved for the liquid form. So maybe there's issues of absorption or exposure. Who

01:22:15.360 knows?

01:22:15.660 Well, I do think there is actually. I mean, we've just empirically used it. We test CoQ10 levels in

01:22:20.200 the blood and I've noticed for what it's worth that virtually every version doesn't show up.

01:22:28.120 The only one I have found, and just to be clear, I don't get paid by this company at all,

01:22:32.340 but there's a brand made by Gero, which I, Gero, I think is probably the best supplement maker I've

01:22:37.560 seen. And we've had some other stuff tested, but Gero's ubiquinol is so readily absorbed because,

01:22:44.220 you know, the clinical trials call for 600 milligrams. I don't think I have a patient on

01:22:48.440 more than 400. In fact, most people at 200 milligrams of the Gero variant reach systemic

01:22:54.320 levels that are above our on statin target. Again, does that mean anything? I actually have no clue,

01:22:59.460 but for what it's worth, I've noticed that on most other versions and variants of ubiquinol,

01:23:05.300 we just do not measure it in the blood.

01:23:07.260 There you go. Yeah. I certainly don't have that experience. It's very interesting to learn,

01:23:11.940 but, but again, how, how, how this, how that's working is. Yeah. Yeah. Right. But the other thing

01:23:16.080 I'll just mention along those lines, and again, it deals with diving deeper into the biology of the

01:23:21.580 system. There's about 20 intermediates on the way to ubiquinol. It's just as complicated a pathway as

01:23:26.320 cholesterol synthesis. And so we don't know whether there may be other targets in that,

01:23:30.480 in that pathway. It just happens to be the end, the end result. So there's lots to learn about

01:23:35.520 how statins impact biology in ways that could affect health. And it's in part because we have

01:23:41.300 such a large population who is doing this experiment basically. And it's also because

01:23:45.500 of the centrality of mechanisms that statins affect. So it's not just cholesterol synthesis,

01:23:51.440 it's, it's precursors of steroid hormones that are in this pathway. And it's these other pathways

01:23:56.600 that can affect intermediates like ubiquinol and other farnaceals. There's, there's lots of other

01:24:02.320 downstream products that are affected by statins that may have biological effects that tell us

01:24:08.880 something about what the drug is doing in the physiological way, but may also have pathologic

01:24:13.920 consequences in some people. What do you make of the evidence for and against the case that,

01:24:19.540 and this is not a common argument, but it is one that's, that shows up enough, which is,

01:24:24.080 okay, statins do reduce events, but it's not by lowering LDL. That's a, an unintended consequence

01:24:32.040 or an intended consequence that is true, true, and unrelated. But the benefits of statins actually

01:24:37.500 come from the endothelial health and or inflammatory reduction. Right. So this is another deep topic,

01:24:47.820 which is obviously clinically important to people understand what it is they're treating and why.

01:24:54.520 It gets back to a couple of issues. One is the causality of LDL, which we discussed a while ago.

01:25:01.720 LDL is causal. That's established beyond doubt. So the benefits of drug-induced LDL lowering,

01:25:08.640 let's say statins, have to be considered as operating in part through that mechanism. Otherwise,

01:25:13.940 right. Let's at least concede that in part, otherwise you're just ignoring reality. So

01:25:19.240 that's number one. Number two is, is that the, is it the most important factor? Is it the only

01:25:22.680 factor? And the answer is, it's certainly not the only factor. There's no doubt in my mind anyway.

01:25:28.200 I'm not sure how much of this is opinion versus evidence, but well, there's certainly anti-inflammatory

01:25:33.060 effects. And there's a, and there's a lot of work that my good friend, Paul Ricker has been

01:25:37.060 involved with, uh, to help establish Paul Ricker's at NIH. No, he's at, uh, he's at Brigham. Okay.

01:25:43.020 Yeah. Yeah. Uh, early on establishing, uh, the importance of C-reactive protein as a marker

01:25:48.580 for inflammatory, uh, risk, um, not just of heart disease, but. And we should detour on that topic

01:25:54.320 because it's so important, right? There was a recent trial that looked at low dose of methotrexate,

01:25:58.360 which is an, that's. No, it was an IL. It was an interleuking one. Weren't there two?

01:26:02.220 There was the IL-1 IL. But the methotrexate isn't out yet. Oh, okay. Maybe I'm just.

01:26:05.460 It's the other one. Okay. Okay. The other one's out. Yeah. Uh, it used an interleuking one.

01:26:10.560 It was one or six. Yeah, yeah, yeah. Yeah. One. And, uh, and so that, that was a very

01:26:14.460 interesting trial, right? Because that, yeah, that was okay. That's right. So that, that was

01:26:17.460 a trial that said, look, we can make no change to the lipoprotein. We reduce inflammation

01:26:22.580 in a subset of patients. These patients had to have an elevated C-reactive protein, if I recall,

01:26:26.900 and you reduced events. Yeah. And we may see the same thing with this methotrexate study. And I

01:26:31.960 think that speaks to. Or not. Yeah. Yeah. Either way, it's an important finding. Exactly.

01:26:37.360 So, you know, the reason I asked the question among others is. Can you pay the statins, right?

01:26:42.240 Yeah. Okay. Because then you can say, well, what does statins do to this pathway? Well,

01:26:47.440 they do lower inflammatory signaling down the same pathway. And, uh, again, Ritzger was a pioneer

01:26:55.940 in studies such as, uh, the Jupiter trial and earlier ones as well, that reanalyzed the predicted

01:27:03.500 benefit or the associated benefit of the L of LDL lowering, uh, against CRP lowering as a marker for

01:27:09.100 inflammation. So this goes back to using CRP as a, as a marker for inflammation and achieving a lower

01:27:14.860 level of CRP and a higher level of LDL was associated with benefit, achieving a lower level

01:27:20.840 of LDL with a higher CRP was benefit. So they both contributed to risk. Uh, and so targeting both

01:27:25.760 LDL to less than 70, I think it was in CRP less than two milligrams per deciliter, I think, but

01:27:33.580 targeting those two risk markers together gave the greatest benefit that each contributed. And I think

01:27:38.840 that model stands up pretty well, uh, that it is a double whammy. And one of the reasons statins are

01:27:45.420 so effective is probably because of their unique ability to hit inflammation as well as, as LDL

01:27:51.080 metabolism and gets back to PCSK9 doesn't quite. That's an intro. That's exactly where I was going

01:27:56.540 to go, which is I've got a number of patients that are coming to me saying, look, Peter, I, uh, you know,

01:28:02.160 I don't care about the cost of the PCSK9 inhibitor. Get me off this statin. I just want to be on a PCSK9

01:28:07.300 inhibitor to which I say, the only issue I take with that is, and I don't have a problem doing

01:28:13.200 it in the patient who is completely statin intolerant. And I have a couple of these patients

01:28:16.680 where they absolutely need to be on lipid lowering therapy and they absolutely can't even tolerate,

01:28:22.100 you know, live a low. So we, you know, there we just do what we have to do, but the trials don't

01:28:27.680 actually tell us how well PCSK9 inhibitors work in isolation, right? Relative to non-treatment.

01:28:34.440 So, and then that, so that's just sort of the evidence-based reason for, hey, it would be ideal

01:28:39.340 if we could at least keep you on some modest amount of a statin. And then secondly, there's,

01:28:44.140 it's not clear the PCSK9 inhibitor attacks that other mechanism.

01:28:47.040 There's a mechanistic aspect to this. And it's not just inflammation down the interleukin CRP pathway.

01:28:52.520 It's also, uh, there's nitric oxide synthesis affects nitric oxide being a vasodilator. So I think

01:28:58.400 there is a good mechanistic case, uh, as well as the clinical evidence case for not abandoning

01:29:04.380 statins in favor of PCSK9 unless of course the patient's statin intolerant.

01:29:08.220 Then where do you think the effect? So, so one of the things I sort of try to divide this as men,

01:29:12.720 women, primary, secondary prevention, right? That's a nice two by two square. I don't think

01:29:18.500 there's any, I mean, again, I'm not talking about the blogosphere or Twitter, but like if you

01:29:22.840 actually look at evidence, is there any dispute in the efficacy of statins in secondary prevention

01:29:28.100 for men or women? No. So no, there's no dispute. So we won't even need to talk about that. Let's

01:29:32.660 talk about primary prevention, the variability in clinical trials. When you look at primary

01:29:37.800 prevention, especially in women, but I think for men to some extent as well, both in the NNT. So

01:29:42.960 that means the number needed to treat. So, and you know, you, the NNT of course being the reciprocal

01:29:47.420 of the absolute risk reduction. So an NNT of a hundred means you have a 1% risk reduction. You need

01:29:52.660 to treat a hundred people to prevent an event. You know, when you look at the summary data on this,

01:29:57.840 it's amazing how all over the map it is. So how would you rate the strength of the evidence?

01:30:05.200 Just, just a quick question, Ron, I'll just give you 30 seconds to answer this. No,

01:30:08.780 the strength of the evidence in primary prevention for statins in either men or women.

01:30:14.260 Okay. Well, let me start with women and let me also specify that you need to think about

01:30:19.560 cardiovascular events and then mortality from cardiovascular disease.

01:30:23.700 Right. So we'll talk about major adverse cardiac events. So myocardial infarction,

01:30:28.020 stroke versus death.

01:30:29.660 So the, there is evidence again, importantly from the Jupiter trial, which, which had a very large

01:30:35.380 enrollment of women, but collectively in other trials as well for the benefit of statins for

01:30:42.400 high risk primary prevention in women. And the high risk caveat is really important here because

01:30:49.020 it really speaks to this very fine gradation between high risk primary prevention and secondary

01:30:55.480 prevention, because if you're at a high enough risk and haven't had a heart attack, you may be

01:31:00.540 five seconds away from the heart attack. So.

01:31:03.800 Yeah. Or you may have had a silent MI that was missed.

01:31:05.860 Already. Right. So that distinction could be a little bit fuzzy, but technically at least the

01:31:11.400 evidence is there for primary prevention of cardiovascular events in women. It's not,

01:31:15.160 it's not really there, at least to my reading. And I may be wrong about this because I haven't

01:31:19.320 gone back and double checked, but I haven't seen evidence for benefit on mortality for primary

01:31:25.460 prevention.

01:31:26.380 In women?

01:31:26.940 In women.

01:31:27.300 Yeah.

01:31:27.800 And in men, it's kind of weakish for mortality, but it's certainly present for events.

01:31:32.780 And is it also possible that, you know, one of the things I try to explain to, if I'm giving a

01:31:38.200 lecture or something to students is you look at a paper and you see no statistical benefit. The

01:31:45.500 first question you must ask yourself was, was this study powered adequately to detect a difference?

01:31:50.940 And it's very often the case that there could have been an effect, but we didn't, you know,

01:31:55.360 it was too small to see with the number of subjects. I feel like in some of the primary

01:31:59.600 prevention trials, duration might be the bigger issue. It's, if you're only looking at seven

01:32:04.740 years, which is probably the outer limits of where these trials look, it might simply be that

01:32:11.320 when you're talking about a 1% absolute risk reduction over seven years, it's not really

01:32:18.140 that interesting. Of course, a 1% risk reduction over seven years, over 30 years is an enormous

01:32:24.240 difference. Alan Snyderman is actually working on a paper now that looks at 30 year risk because

01:32:30.660 of course all the risk calculators are based on 10 year risk, which I think has value, but

01:32:34.740 also has great limitation. And I guess I think that's part of what I struggle with this clinically,

01:32:40.680 which is, you know, we never want to expose a patient to something unnecessarily. And there's

01:32:46.500 no substance you're going to put in your body, whether it's, you know, vitamin C versus a

01:32:50.440 statin that comes with zero risk. So how do you weigh that benefit?

01:32:54.820 Yeah. Well, you know, this really speaks to need for much better markers that could help us predict

01:33:02.400 benefit versus adverse effects so that we can identify not just the high risk candidate for

01:33:10.460 statins based on conventional markers, but individuals in whom we have evidence for mechanisms that would

01:33:16.980 be benefited by statin treatment that would argue strongly for benefit. And on the other side,

01:33:23.060 markers, ideally that could assess risk for adverse effects, not just symptoms, but underlying

01:33:29.140 pathology, for example. And again, without X, I want to come back briefly to your question about

01:33:35.900 cognitive function, because I sort of escaped talking about that because I feel this is...

01:33:41.220 Because you forgot because you're cognitive function.

01:33:43.280 No, as a matter of fact, quite the opposite. I put it in a part of my brain that

01:33:49.180 has made me not forget that there can be effects that are long-term that we simply can't assess

01:33:57.580 easily. And so one of them could be cognitive function. But having said that, there's no evidence

01:34:02.240 and in our consensus paper that just came out on statin intolerance, no evidence really.

01:34:07.380 Well, at the population level, I would say it's the opposite. At the population level,

01:34:10.500 I think we see that the risk, if anything, goes down a little bit.

01:34:13.160 Yeah. And there's a cardiovascular...

01:34:14.620 There's a vascular element.

01:34:15.800 That's exactly right.

01:34:16.700 Right.

01:34:17.020 It's a vascular element that even contributes to classical Alzheimer's disease. But again,

01:34:21.620 the data are really not compelling. They're not compelling.

01:34:24.500 No, this is the challenge. This is personalized medicine means at the end, I don't care about

01:34:28.620 the population. I care about the one person sitting in front of me at this moment when we have

01:34:32.980 to make a decision.

01:34:33.680 Exactly. Exactly. But what's intriguing and what your question opened up, and I have to

01:34:38.880 be careful because I don't want to go down the rabbit hole on this one, to the extent

01:34:43.940 that statins influence mechanisms that interact with cholesterol transport in the brain, an

01:34:50.080 apoprotein E, which we mentioned is a variant that causes type 3 hyperlipidemia. Apoprotein

01:34:56.000 E is a key protein that has a variant, another variant that increases the risk for Alzheimer's

01:35:01.920 disease. Well, we're looking at ways that cholesterol metabolism intersects with that pathway. And

01:35:07.200 inconceivably, that's just one example of a gene that could interact with statins. And if the

01:35:13.440 statins are getting past the blood-brain barrier in some cases, and inflammation might be a factor

01:35:19.240 that could predispose to that, you could start to have imagined effects that could go either

01:35:25.180 way, but could perhaps potentiate dysfunction. That's just a completely off, you know, out of

01:35:30.980 left. No, no, look, I mean, that's exactly, I mean, that's, that's sort of the hand-waving

01:35:35.060 that I use, right? So when I have patients that have one or two copies of an APOE4 gene,

01:35:40.460 I am that much more careful with them with statin use and until proven otherwise.

01:35:45.820 Yeah. Yeah. And we don't know, and we don't know, you know, I mean, I actually, there's a,

01:35:49.800 there's a wonderful group of APOE4 carriers that has formed a sort of a support, very large,

01:35:56.460 that I've interacted with. They came to my lab. There's about 30 members of this group that came

01:36:01.900 and we gave a talk about APOE4 and nutrition, actually. I had to sort of be careful because

01:36:10.000 these people were just anxious for advice. What do we do? I've got this APOE4 variant. What do I do

01:36:16.280 with my diet? And I told them, you know, we've done a lot of studies along these lines, but we don't

01:36:20.620 know what the effect on disease risk is. But there's one other cumulative effect that I'm going to

01:36:25.720 mention without, again, turning the tables on the value of statins. And that is getting back

01:36:31.360 to the muscle effect. Because if, if there is a adverse effect on mitochondrial function, which

01:36:37.280 we can, we can show that and others have shown that. And if there are individuals who, for one

01:36:42.940 reason or another, are at risk for muscle wasting, called sarcopenia, muscle wasting, conceivably that

01:36:51.040 on-target effective statins, the extent that it is affecting muscle tissue, could over decades

01:36:58.360 conceivably impact the rate at which this muscle wasting could occur. And so there may be a time,

01:37:05.040 and this is just, I'm throwing this out at the, probably we're near the end of this discussion,

01:37:09.300 as a kind of just a thought, is it may argue for after a certain age, not worrying so much about

01:37:16.620 high-dose statin treatment unless the patient is at really high risk. And because one may be adding

01:37:22.800 in the elderly individuals an exposure.

01:37:25.680 Yeah, you're creating a new risk in the presence of another risk reduction.

01:37:29.520 Right. And it's the experiment that we probably are never going to be able to do. And so it just

01:37:34.820 has to sort of be cognizant. I've had patients, and again, I actually take care of a number of

01:37:40.140 physicians, actually, in my clinical work, who ask me, you know, I've reached the age of 75 or 80,

01:37:44.820 and you've done well, I've had family history, but I'm doing fine now. And it's a difficult question

01:37:49.300 to answer because we don't have the evidence base beyond age 75. That was really where the

01:37:55.120 cholesterol guidelines that came out a few years ago stopped, because that's where the evidence

01:37:59.480 stopped. We do have benefit in older individuals who are at higher risk, but we don't know what the

01:38:04.620 trade-off is in terms of adverse effects that people don't really need to be taking it any longer.

01:38:09.780 Now, you were involved in one of the ATP guidelines, correct?

01:38:12.520 Yeah. Well, I was on the ATP-4, on the last panel that handed it off to the American Heart

01:38:17.880 Association, American College of Cardiology. I was connected with it when it was part of the NIH

01:38:22.400 guidelines.

01:38:23.280 Now, you took issue with something, didn't you?

01:38:25.620 Well, there were some concerns that I and others had regarding the scope of the effort.

01:38:32.020 Originally, and this is my own personal experience with it, when I was brought on, I was expecting,

01:38:37.000 and we originally talked about having a wide range of topics to discuss, not just LDL lowering,

01:38:42.600 but how we manage triglycerides, the role of HDL, emerging risk factors, LP little a. We had

01:38:49.620 probably 15 or 16 high-priority questions, and I was intending to be involved heavily on the

01:38:57.440 discussion of some of these other risk factors. I thought the LDL story was obviously important,

01:39:03.860 and one had to address, should we update the guidelines? That's what it was really designed

01:39:08.060 to be. And it turned out, in the end, that because of various changes in support of the

01:39:14.300 whole program, the number of questions was reduced down to just three, and they all had to do with

01:39:20.540 LDL lowering and the evidence for it. And it was very heavily dependent on trials that were available

01:39:28.840 to provide evidence-based conclusions. So that means that the effort was really limited to a very

01:39:36.320 strict interpretation of evidence-based guidelines, and that led to an abandoning of the LDL targeting

01:39:43.300 that we had. The LDL levels that had been used to guide management of risk were abandoned because

01:39:51.160 there was no study that actually addressed specifically that question, that is, does reaching a

01:39:57.040 certain target of reduced risk? And I personally got rather discouraged at the limitation that we

01:40:04.000 were under. Eventually, it sort of got a little bit softened, actually, after I left the committee.

01:40:08.880 I think there was a little bit more tolerance of a potential role, but basically, by being that

01:40:15.060 narrow, we not only lost a lot of important questions in lipid management that just weren't

01:40:19.240 addressed. Triglyceride, you know, et cetera, et cetera. Metabolic syndrome wasn't really touched

01:40:25.120 on it all. So it was disappointing to me. And I felt my LPA, right. And so I felt my role on the

01:40:30.900 LDL lowering side was sort of limited. You know, I'd actually been involved with being consultants

01:40:35.600 to companies that made LDL lowering drugs. So I had to actually excuse myself from even the

01:40:39.260 decision process. So it was really a point where I just wasn't able to contribute the way I thought

01:40:43.420 I could. So I decided to step off. I could continue this discussion probably for six more hours,

01:40:48.300 but there is one other thing I want to go back to that we did talk about. And I think it's the

01:40:52.360 niacin question. Right. You haven't forgotten that one. No, I sure have not. So what does

01:40:58.200 niacin do, right? So first of all, I don't think anybody actually really understands the mechanism

01:41:01.620 by which niacin lowers ApoB. I mean, is that generally? Yeah, that's right. I would say that's

01:41:08.040 a fair statement. But niacin is a drug that lowers LDL, particle cholesterol. It also raises HDL

01:41:14.800 cholesterol quite significantly. In fact, I've seen it raise HDL cholesterol more than I've seen

01:41:19.280 anything in those CTEP inhibitors. So it seems like the dream drug. It's doing everything right.

01:41:25.380 Now it does also create some insulin resistance and it seems to do that at a higher level than

01:41:29.320 even statins do. But do you think that that's the explanation for why the hard outcome data

01:41:34.680 aren't there and why niacin has fallen so far out of favor that, you know, I actually called in

01:41:39.620 niacin for a patient about a year ago. It was, we were kind of doing an experiment. He couldn't

01:41:43.940 tolerate a statin, couldn't afford a PCSK9 inhibitor. You know, we were really reaching for

01:41:49.000 straws and we couldn't even get niacin for him. I mean, we literally couldn't even get his insurance

01:41:52.560 company to pay for niacin. So it is persona non grata. Help me understand that. Okay. So my disclosure

01:41:58.820 is I still use a lot of niacin. So why do I do that in the light of all these trials? So let me just

01:42:06.280 back up then and say that the trials, the high profile trials that led to the dismissal of

01:42:13.300 niacin as a therapeutic option. One was aim high and the other was HPS2. Both involved the following

01:42:22.820 criterion. And that is niacin was tested as an HDL raising agent in the setting of LDL being at a low

01:42:32.960 level as a result of intensive statin plus or minus setia to bring LDL cholesterol levels to a low

01:42:41.360 enough level that it would be considered a non-factor. That is that the niacin effect due to LDL would be

01:42:48.140 minimized. Whereas the HDL... Right. In other words, how do we assess the level of niacin exclusively on

01:42:53.820 its HDL raising properties because we've maximized LDL lower? That's right. So they were designed to test

01:42:59.220 whether the HDL increase with niacin is beneficial. And the answer is from those trials,

01:43:06.120 I'd say unequivocally no. That is the HDL raising effect of niacin is not protective. That's the

01:43:13.180 conclusion to take away. That in my view was beating maybe not a dead horse, but a pretty sick one

01:43:19.780 because I never really was totally convinced that raising HDL itself would be beneficial. Even though

01:43:26.900 I was part of the earliest study that showed that if you genetically treat mice with a gene that

01:43:34.560 raises HDL, you can reduce atherosclerosis. I mean, there is a protective effect, but I was of the

01:43:40.260 opinion that the association of HDL cholesterol across the population with risk fell into the same

01:43:47.160 category as the LDL, small LDL story and the triglyceride story. It's part of a interrelated

01:43:53.000 metabolic syndrome. And I felt that the most important pathologic features of that syndrome

01:43:58.200 were not low HDL, but the high levels of small LDL and triglycerides. And both of those

01:44:04.800 components have been, not just the LDL, but the triglyceride related risk have been pretty clearly

01:44:10.620 shown to be causal. And the HDL was a marker for that. That was actually my guess, just based on my

01:44:18.380 being. And there could be another issue here, right? Which is to raise it. I mean, there are,

01:44:22.580 there are known, you know, HDL lipoproteinemias where very high HDL is actually atherogenic because

01:44:29.100 they're so dysfunctional, they can't actually delipidate, right? Exactly. So the simple-minded

01:44:34.000 idea that raising HDL by any means would be beneficial, just like lowering LDL by any means

01:44:39.100 would be beneficial, was disproven by those studies. Plus all of the CTEP inhibitors.

01:44:44.680 That's right. None of the CTEP inhibitors, right. And so there's something fundamentally

01:44:48.360 pathologic about that effect on HDL that overrides any potential benefit that may be due to other

01:44:58.000 factors. And in the case of the two major niacin trials, you know, there was some lowering of LDL,

01:45:04.860 but as you point out, the HDL increase was fairly substantial. And those particles are basically

01:45:10.340 hung up in plasma. It gets back a little bit to this residence time issue that I was talking about

01:45:16.260 regarding LDL. Well, here's the situation for HDL. Well, these particles are just sort of like the

01:45:21.200 toilet is plugged. Yeah. You prevent them from doing their job. Right. So you're not delivering

01:45:26.100 cholesterol as efficiently back to the liver for excretion. And those particles are hanging around

01:45:30.660 long enough that who knows what properties they may be acquiring that might possibly override some of

01:45:36.840 the benefits of the LDL lowering effect. So does that mean that niacin has no benefit? Well,

01:45:44.420 there was a study that preceded those two studies that was carried out by Greg Brown and actually was

01:45:50.200 one that led to the AIM-HI study being funded by the NIH. And that was called HATS, HDL,

01:45:56.760 atherosclerosis treatment study, HATS. And that was a study- Almost as good as Mr. Fit,

01:46:04.480 which I realize now is the one I was forgetting before the LRCP. Okay. That was a multi-factor risk

01:46:09.480 intervention. Yes. Yeah. Yeah. Yeah. Yeah. Okay. So going back to HATS. Right. So HATS was a statin

01:46:14.820 plus niacin, and there was other combination approach, which included niacin. And that study showed

01:46:20.160 benefit two levels. One was the intended endpoint was quantitative angiographic progression,

01:46:25.800 measuring the narrowing of the coronary vessels, which is now not used- We now realize it's not

01:46:31.000 particularly helpful. Right. It happened to be correlate with endpoints. And in fact, in that

01:46:34.800 study, even though it was initially underpowered to detect the benefit on endpoints, that intervention

01:46:40.540 did reduce risk of events. It was a successful trial. We analyzed data from that study using four

01:46:47.140 different methods to look at lipoprotein particles. And so part of this was to see whether we could learn

01:46:51.260 something more than the standard lipid measurements would tell us. And what it did tell us is that the

01:46:57.460 small and very small LDL lowering achieved by niacin, which it does do, was associated with

01:47:03.940 vascular benefit in that study. Whereas the HDL, to the extent of nothing else was explanatory,

01:47:10.640 and that was independent of all the standard lipids. So we have this in the literature from just one

01:47:15.580 study, and I wish we would have access to other types of data that could support this, but it spoke

01:47:21.300 to a therapeutic effective niacin that would be lost in these larger trials because the statin hammer

01:47:30.640 had lowered the LDL low enough. So that additional benefit is probably just lost-

01:47:34.900 It's too small. It's underpowered.

01:47:36.260 It's lost in the noise. So the failure of those trials speaks to the success of statins and the

01:47:42.040 failure of HDL raising. Those are the two things I take away. But niacin, I think in the patients

01:47:46.900 who have small LDL particles, I still use it, and I like to see-

01:47:51.380 Well, that's what I want to ask you. Talk me through the perfect niacin patient.

01:47:57.000 Well, the perfect patient with niacin is actually twofold. And we weren't going to talk about LPA,

01:48:02.480 but you did bring it up earlier, and I'll mention it again. So it's this genetic factor that is sort of

01:48:07.900 the wild card in atherosclerosis. It's certainly damaging, and we'd like to lower it. So niacin

01:48:13.320 can lower LPA by up to 25% or so. And there is not conclusive, but to me, clinically impressive

01:48:23.560 evidence that if you have a combination of high LPA and small LDL, you're just a time bomb for

01:48:31.100 atherosclerosis. So if you have high levels of, you know, high particle number of the small LDL type

01:48:37.420 plus LP little a elevations, and there's always a strong family history of heart disease, and the

01:48:42.580 patients are going to be at high risk at a young age because niacin lowers LPA and because niacin

01:48:47.800 lowers the small LDL particles.

01:48:49.100 So niacin, you think, specifically targets smaller particles over larger particles?

01:48:54.840 Yeah, preferentially.

01:48:56.400 Sorry, yeah, that would be a better one.

01:48:58.080 It lowers that pathway. It's not clear whether it does that through the VLDL pathway or not.

01:49:02.860 Honestly, we don't know. It gets back to your earlier statement that we don't know the mechanism

01:49:06.440 really as well. We just don't know the mechanism. But in that combination, that really is the

01:49:12.120 patient that, and I've got patients like that who have...

01:49:15.020 And you'd put that patient on that over a PCSK9 inhibitor if they could afford it?

01:49:19.200 Well, there's another angle...

01:49:20.440 Because you're going to get a 30% reduction on LP little a through the PCSK9 inhibitor.

01:49:23.840 That's right. There's another... That's right. That's another angle. Because PCSK9 has a

01:49:27.700 similar new develop LPA-lowering effect. Plus, it has a gangbusters bigger effect on LDL

01:49:32.660 particle, LDL levels. So, no, I wouldn't use it over PCSK9. The argument there is largely

01:49:38.500 financial, honestly.

01:49:39.840 And what insurance will cover. You can buy niacin at the local shop off the shelf,

01:49:45.460 longer-acting niacin in a relatively safe form, for pennies. You're going to pay eight or nine

01:49:51.880 $14,000 for PCSK9. I've had patients like you who have been willing to do that, who have been

01:49:59.660 willing to sort of...

01:50:00.700 Just pay out of pocket.

01:50:01.820 That's right. And that is the same category of patient, and I probably would consider that even

01:50:06.320 more effective. Although, one other thing, a little tweak here, and that is that both statins and

01:50:12.920 PCSK9 inhibitors, because they work, as you said, by upregulating LDL receptors, as in the case of

01:50:20.360 PCSK9, that's the mechanism.

01:50:21.740 Exclusively, yeah.

01:50:22.380 It's right for statins. It's part of the mechanism. That effect primarily lowers medium and larger

01:50:29.420 size LDL particles and has less of a therapeutic effect than smaller particles, and none on the

01:50:35.400 very small LDL. So, there's a gradation of effect which relates to the structure of the LDL

01:50:41.020 cell being more or less capable of interacting with LDL receptors, and the larger particles

01:50:47.320 interact well, and so upregulating LDL receptors.

01:50:50.520 And this explains why we sometimes see a widening of the discordance in the statinized patient,

01:50:56.740 because they're lowering LDL-C more than LDL-P because you're selecting out the larger,

01:51:02.960 more cholesterol dense.

01:51:04.140 That's right. And so, that applies to any mechanism that operates through LDL receptors,

01:51:07.960 which is what most of the drugs do, even azetimibe. So, what does that mean for niacin? Well,

01:51:13.260 niacin lowers the small and very small LDL very nicely. So, there's a complementarity there

01:51:19.240 to statins. And I think that's part of what attracts me to using it. And again, I don't use

01:51:25.180 nowadays because of potency of the statins, the ability to get LDL down to target in such

01:51:30.320 high proportion of patients, the availability of PCSK9. I'd have to say there's fewer patients

01:51:35.000 that I'm starting on niacin, but there's still candidates out there that I think would benefit

01:51:39.440 from this, hitting small LDL and hitting LP in ways that complement the benefits of statins.

01:51:46.360 You know, the only patient I've put on niacin in the last four years is a really interesting case.

01:51:52.300 He's a guy with, he doesn't have FH, but he's clearly got some snips of FH because his LDL is

01:52:00.000 incredibly high, but normal synthesis, normal triglycerides, and normal absorption markers.

01:52:05.740 So, you know, his sterols are normal. This is a clearance problem. Unable to tolerate a single

01:52:12.560 statin, including one milligram of Livolo. So, we went to Repatha and there was no effect.

01:52:20.620 Oh, really?

01:52:21.340 And we switched to Prolulent, no effect.

01:52:25.020 That's really interesting.

01:52:25.880 So, he's missing, there's an epitope that's just...

01:52:28.420 I would like the DNA from that patient because...

01:52:31.560 I would be happy to introduce you to him.

01:52:32.980 I have a candidate genetic variant that might explain that, actually.

01:52:35.600 Okay.

01:52:36.060 It's an LDL receptor. So, you use niacin in that patient?

01:52:39.420 Yeah. I mean, it's like basically going to be what? Niacin and apheresis.

01:52:43.400 Yeah.

01:52:43.640 I mean, it's a really tough case.

01:52:46.080 Yeah. Well, again, I mean, you know, I think it's legitimate to say that there's probably less,

01:52:50.860 fewer candidates for niacin with the availability of this potential for using PCSK9 and certainly

01:52:56.600 with inflammation with statins. But I would say the patients with moderately elevated triglycerides

01:53:00.840 in the 150 to 400...

01:53:02.280 Yeah, yeah. You give them a little phenofibrate?

01:53:04.060 Yes, I do, actually, despite... That's another subset question.

01:53:07.880 That is a subset of patients who may benefit based on subset analysis, subgroup analysis of

01:53:13.600 the clinical trials. But unfortunately, it's just not a very potent agent.

01:53:18.280 But niacin, I will use as a alternative approach to lowering the triglyceride and lowering the

01:53:25.160 small LDL.

01:53:26.320 That's elegant. That's something, I mean, if just on a personal level, that's certainly

01:53:31.060 an amazing and insightful pearl that I've gathered from our discussion today.

01:53:34.860 I appreciate you saying that, Eddie.

01:53:36.360 Well, and with that, Ron, I want to say, you know, first of all, just I consider you a great

01:53:40.780 friend and an unbelievable mentor. I feel so fortunate to have benefited from your knowledge

01:53:46.580 and peers of yours over the last decade. And your generosity is unparalleled in terms of

01:53:54.440 anytime I can pick up the phone and call you and run a tough case by you, you're always there to do

01:53:59.340 so. So I greatly appreciate that. I also think we need to do this again sometime because I literally

01:54:04.820 have twice as many questions as we've got to. I wanted to get into saturated fat, fructose,

01:54:11.360 APOE. There's all these other things that I know you are just an expert on that I know people are

01:54:16.600 going to want to hear about. So I'm going to have to come back to San Francisco and we'll have to

01:54:20.560 continue this discussion, hopefully during the Warriors offseason.

01:54:24.560 Happy to do that, Peter. And thank you for your kind words. And it's been a pleasure talking with

01:54:29.640 you. I think the opportunity to help people understand some of these complex issues is

01:54:34.540 something you've been very good at. And I've been glad to be able to contribute to that.

01:54:38.060 Thank you so much, Ron.

01:54:39.260 Okay.

01:54:41.780 You can find all of this information and more at peteratiamd.com forward slash podcast.

01:54:47.040 There you'll find the show notes, readings, and links related to this episode.

01:54:51.160 You can also find my blog and the nerd safari at peteratiamd.com. What's a nerd safari, you ask?

01:54:57.280 Just click on the link at the top of the site to learn more. Maybe the simplest thing to do is to

01:55:01.540 sign up for my subjectively non-lame once a week email where I'll update you on what I've been up to,

01:55:06.320 the most interesting papers I've read, and all things related to longevity, science, performance,

01:55:11.480 sleep, etc. On social, you can find me on Twitter, Instagram, and Facebook, all with the ID peteratiamd.

01:55:18.340 But usually Twitter is the best way to reach me to share your questions and comments.

01:55:22.220 Now for the obligatory disclaimer. This podcast is for general informational purposes only and does

01:55:26.600 not constitute the practice of medicine, nursing, or other professional healthcare services,

01:55:31.460 including the giving of medical advice. And note, no doctor-patient relationship is formed.

01:55:36.320 The use of this information and the materials linked to the podcast is at the user's own risk.

01:55:41.960 The content of this podcast is not intended to be a substitute for professional medical advice,

01:55:46.460 diagnoses, or treatment. Users should not disregard or delay in obtaining medical advice for any medical

01:55:51.880 condition they have and should seek the assistance of their healthcare professionals for any such

01:55:56.440 conditions. Lastly, and perhaps most importantly, I take conflicts of interest very seriously. For all of

01:56:02.440 my disclosures, the companies I invest in and or advise, please visit peteratiamd.com forward slash about.

The Peter Attia Drive - July 02, 2018

#03 - Ron Krauss, M.D.： a deep dive into heart disease

Episode Stats

Length

Words per Minute

Word Count

Sentence Count

Misogynist Sentences

Hate Speech Sentences

Summary

Transcript