Published: December 15, 2023
In today’s blog post, we are discussing blood lipids, such as LDL-cholesterol, HDL-cholesterol, and triglycerides, and how they affect our risk of cardiovascular diseases, such as heart disease or stroke. For some reason, this has become a major hot-button topic, and one thing I have heard a lot from many of you is that you are totally confused about this issue because there are so many voices, some of which are claiming that LDL-cholesterol is THE cause of cardiovascular disease and we should all be taking statin medications, while others claim that it’s all a ploy by big pharma and that the impact of LDL-cholesterol on cardiovascular disease is negligible.
My goal with this blog post is to bring clarity into this landscape not by just adding another voice, but by taking you by the hand and reviewing all of the different pieces of evidence, and all of the various claims on both sides of the argument. My goal is that by the end of the blog post, you will have much more clarity on the science of this important topic of how blood lipids are related to cardiovascular disease, how important lipids such as LDL-cholesterol are relative to other risk factors, which lab tests to have done, and how to interpret them.
Quick disclaimer: I am not in any way involved with the pharmaceutical or supplement industry, and have no financial interests related to the content of this blog post. I am just sharing the data as I see and interpret them. As always, I have prepared a blog post that discusses some of the issues covered in this blog post in more detail, and that also provides all of the scientific references. The link is, as always, in the description box below the blog post.
The Basics of Atherosclerotic Cardiovascular Disease
Let’s first clarify that there are numerous types of cardiovascular diseases. The type we are interested in in this blog post is what we call atherosclerotic cardiovascular disease, or ASCVD. ASCVD is the gradual accumulation of plaque in the artery wall that, over time, grows and constricts the inside of the artery such that blood can no longer flow freely. We call this process shown in the figure below atherosclerosis or arteriosclerosis. In an advanced stage, the plaque can become what we call unstable and rupture, in which case the artery would be entirely closed shut at this point. If this is a critical artery that provides nutrients and oxygen to a heart muscle, then it’ll result in a heart attack. If it’s an artery that supplies blood to the brain, then this will result in a stroke.
The first thing I’d like you to take away from this blog post is that atherosclerosis is clearly a multifactorial disease, with numerous independent risk factors that each promote its progression. These include age, gender, and race, as well as smoking, hypertension, diabetes, and kidney disease. The LDL-cholesterol concentration is also thought of as one of these established risk factors, which seems plausible because the plaque in the artery wall actually consists largely of cholesterol. There are some other blood lipid measures that should be considered, such as HDL-cholesterol, triglycerides, and a LDL-like particle called Lp(a).
So, again, ASCVD is a multifactorial disease, and the progression of atherosclerosis is highest in people who are exposed to several or all of these risk factors. However, each of these is still an independent risk factor, meaning that even if just one is elevated independent of all of the others, it will raise ASCVD risk. The one independent risk factor we are looking at in this blog post is the blood lipid profile, with a specific focus on the fasting LDL-cholesterol concentration.
The Basics of Blood Lipids and Lipoproteins
This section is going to be a little technical, but I hope you’ll struggle through this with me because understanding the different lipoprotein classes in blood will be necessary to understand the finer nuances later in the blog post.
Fats such as triglycerides and cholesterol are not soluble in blood, so they need to be packaged for transport. The transporters of fats in our blood are called lipoproteins. In the fasting state, there are a variety of lipoproteins in our blood. The two you have probably heard about are low-density lipoprotein, or LDL, and high-density lipoprotein, or HDL. There are two others I’d like to mention at this point: very-low-density lipoprotein, or VLDL, and intermediate-density lipoprotein, or IDL.
All of these lipoproteins contain lipids, such as triglycerides and cholesterol, and also proteins. The different lipoproteins differ from one another in the types of lipids and proteins they carry, and in the relative proportions of lipids to proteins. And the relative proportion of lipids and proteins affects their density. That is because lipids have a density of less than 1 g/mL, while proteins have a density of more than 1.3 g/mL. Therefore, lipoproteins that consist mostly of lipids have a density close to 1, whereas lipoproteins that have relatively more protein have a density of around 1.2.
The lipoprotein with the lowest density is VLDL, then IDL, LDL, and HDL is the most dense one. These lipoproteins do not just differ by density, but also by size, with VLDL being the largest and HDL being the smallest. As you can see in the figure above, each lipoprotein class actually spans a range in terms of density and size, so it’s not really accurate to imagine, say, all LDL particles as exactly the same. Instead, the reality is that these are spread out over the entire range in terms of density and size within each of these classes. This will become relevant later.
VLDL particles are secreted by the liver. They do contain cholesterol as well, but they are particularly rich in triglycerides. The VLDL particles then shed triglycerides, mostly in fat tissue, and in that process, they shrink, and now a greater and greater proportion of the lipoprotein particle consists of protein, which, remember, is denser than lipids. This means the density of the particle increases. Click on this animated graphic for an illustration of this process.
So, again, as shown in the figure below, VLDL particles secreted by the liver gradually become IDL and then LDL. The particle loses triglycerides and shrinks, the density increases. And during that process, the relative proportion of protein and cholesterol within the particle increase. The liver can also directly release IDL and LDL particles into the blood. So VLDL, IDL, and LDL are all derived from the liver, and they are related to one another in a way. In fact, we could make the case that they are one and the same particle, just at different stages in their life cycle.
By contrast, HDL particles are different. These are very small and very dense lipoprotein particles that are very rich in protein. While the key proteins in HDL can be made by the liver, but also by the intestines, they pick up most of their lipid cargo, such as cholesterol, as they circulate in the bloodstream. The full lipoprotein particle actually develops while the HDL particle is circulating in the blood.
If you get a lab test of LDL-cholesterol, what this measures is the cholesterol content that is transported in all of your LDL particles. Similarly, HDL-cholesterol is the cholesterol that is being transported in all of your HDL particles. The cholesterol in both cases is the same molecule; it’s just the transport vehicle that differs. And, as we’ll discuss later, which transport vehicle carries the cholesterol in the blood does matter for ASCVD risk.
There is also one additional major difference between these lipoprotein classes, and that is in the type of the major protein they carry (see figure above). This is important because that is what affects the function of these lipoproteins, their removal from blood, and also how they affect ASCVD risk. The key lipoprotein in VLDL, IDL, and LDL is called apo B, and the key lipoprotein in HDL is called apo A. We’ll get back to why this is important later.
OK, that should suffice, and I hope the basics of what lipoproteins are is now a bit more clear. With that, let’s get to the relationship between these lipoproteins and ASCVD.
Do LDL Particles Cause ASCVD?
Do LDL particles CAUSE ASCVD? That, at least, is the claim made by this scientific paper here. And this is not just any paper: it is a position statement, and its authors are some of the most well-known lipid researchers of our time. So why do they come to this unusually strong conclusion? LDL particles CAUSE ASCVD. Scientists usually do not use strong causal language like that unless they have very strong evidence. Let’s go through the evidence they present and think this through together.
The first piece of evidence they present is that people who are born with genetically very high LDL-cholesterol concentrations tend to have a massively increased risk of ASCVD. There are actually reports in the literature of small children with massively elevated LDL-cholesterol of 400, 500, 600 mg/dL or higher who have a heart attack before entering puberty. These are rare cases, but they show that even though ASCVD is a multifactorial disease, if this one risk factor, LDL-cholesterol, is totally out of control, it by itself can cause severe ASCVD and its clinical manifestations at a very young age, and without any evidence of other contributing risk factors such as smoking, hypertension, or diabetes. I actually had an uncle with a genetic defect that gave him very high LDL cholesterol levels. He had his first heart attack at the age of 29, and almost everyone in his family passed away from ASCVD before their 60th birthday.
On the other end of the spectrum, there are also people who have genetically very low LDL-cholesterol levels. Most of these have loss-of-function mutations in a protein called PCSK9. No matter what they eat, some of them have LDL-cholesterol concentrations of 20, 30, or 40 mg/dL all the way through high age. What do we think is their risk of having a heart attack? Yes, indeed, it’s extremely low.
Now, most of these well-documented cases of genetically very high LDL-cholesterol concentrations have LDL-cholesterol of 190 mg/dL or higher, in many cases much higher. And those cases of genetically very low LDL-cholesterol concentrations usually have LDL-cholesterol of less than 70 mg/dL. What about that range between 70 and 190 mg/dL in the middle that most people without major genetic mutations are in? Same here: very consistently across a wide range of studies, even within that range, higher LDL-cholesterol concentrations are consistently associated with greater ASCVD risk.
OK, so LDL-cholesterol is associated with the risk of ASCVD. Association doesn’t equate to causation though, so let’s look next at clinical trials in which LDL-cholesterol was experimentally lowered. We have numerous published trials, including more than 2 Million participants overall, in which the goal was to lower LDL-cholesterol. These include trials of several classes of lipid-lowering medications, including statins, ezetemibe, bempedoic acid, and PCSK9 inhibitors. However, we also have data from interventions in which LDL-cholesterol was mechanically removed from the blood, without any medication, in a procedure called LDL apheresis, which is similar to dialysis in that the blood runs through a machine, and LDL particles are removed from the blood. Irrespective of how LDL-cholesterol was lowered, these trials showed a consistent reduction in the risk of ASCVD, and also a reduction in plaque volume. And the more strongly LDL-cholesterol was lowered by an intervention, the more ASCVD risk was usually reduced.
One thing that is particularly intriguing here is that genetic variants that mimic the effect of some of these lipid-lowering drugs seem to have the same impact on LDL-cholesterol and ASCVD risk as the drug. For example, statin drugs inhibit an enzyme called HMG-CoA-reductase, which is the key enzyme in cholesterol synthesis. Someone who takes a statin has an HMG-CoA reductase enzyme that is less active, and as a result, has lower LDL-cholesterol levels and a lower risk of ASCVD. Now, the interesting part. What if we found people in the general population with different genetics such that they have either very highly active or an inhibited version of HMG-CoA-reductase? Well, those with a genetically inhibited, less active version of HMG-CoA-reductase tend to have lower LDL-cholesterol levels, just like people on statin drugs, and as a result, they also have a lower risk of ASCVD. The exact same thing is the case for PCSK9, the enzyme targeted by the PCSK9-inhibitor class of drugs. In other words, lowering the function of these enzymes genetically has a similar effect on both LDL-cholesterol levels and ASCVD risk as inhibiting them pharmacologically.
In addition to this evidence from large-scale observational, genetic, and intervention studies, we now have solid evidence that LDL particles and other apo-B containing lipoproteins play a crucial role in the etiology of atherosclerosis. In essence, it is the retention of these apo-B containing lipoprotein particles in the space underneath the arterial endothelial layer that initiates the formation of the first atherosclerotic leasons. Consistent with the evidence from observational, genetic, and intervention studies, higher concentrations of LDL particles and other apo B-containing lipoproteins seems to accelerate this process. Also in line with evidence of other ASCVD risk factors, there is also no doubt that other exposures can greatly increase this process, specifically diabetes, hypertension, and smoking.
In conclusion, I agree with the authors of the position paper that the available data strongly suggest that elevated LDL-cholesterol concentrations are a major independent risk factor for ASCVD, and are probably causal. And I would suggest you consider this multitude of evidence, and if you still doubt a role of LDL-cholesterol in ASCVD, I invite you to read some of the original literature.
In science, we rarely ever have such consistent evidence across the board. We should always be open to new data, and we may yet learn something new about the relationship between LDL-cholesterol and ASCVD risk, as we’ll discuss later, but at this time, the available evidence is strong and very consistent.
Is LDL-cholesterol the Best Blood Lipid Measure to Assess ASCVD Risk?
So this begs the question of whether LDL-cholesterol is the best blood lipid measure to assess ASCVD risk.
The short answer is No. What we have learned in recent years is that not only LDL particles are atherogenic, but also VLDL and IDL particles. And we also now better understand that the best predictive blood lipid measure for ASCVD risk is not the amount of cholesterol these lipoproteins carry, but the number of these atherogenic particles in our bloodstream.
Now, do we have a measure of the number of these particles in our blood? Yes, we do. Remember that VLDL, IDL, and LDL all carry this protein called apo B? In fact, each of these lipoprotein particles carries exactly one apo B molecule. We can therefore measure the blood concentration of apo B, and that measure is directly related to the sum of atherogenic particles in blood. Therefore, the current consensus among preventive cardiologists and lipidologists is that plasma apo B is the best measure of lipid-related ASCVD risk.
If obtaining a measure of apo B is not an option for you, then the second best option is non-HDL-cholesterol. Non-HDL-cholesterol is total cholesterol minus HDL-cholesterol, so this is something you can all calculate yourself from a standard blood lipid panel. Non-HDL-cholesterol is not a direct measure of the particle number of the atherogenic lipoproteins, but a measure of their cholesterol content, which is strongly associated with the particle number. So non-HDL-cholesterol is the second best measure.
So is LDL-cholesterol useless? No, it is not, because it correlates with apo B and non-HDL-cholesterol. However, LDL-cholesterol has one slight disadvantage: it does not measure anything related to VLDL and IDL particles. And, as I have mentioned, these are also atherogenic particles just like LDL.
To summarize this part: Ideally, measure the apo B concentration in fasting blood or, as a close second and third option, measure non-HDL-cholesterol or LDL-cholesterol, respectively.
Now, there is another lipoprotein that is often forgotten: lipoprotein (a), or Lp(a) (“Lp little a”). Lp(a) is another lipoprotein that is similar to LDL, but it has not just one apo B molecule, but also an additional protein called apo(a).
Now, I realize this is confusing, but please hear me out because this is worth knowing about and understanding fully. The main protein on LDL particles is apo B, and the main one on HDL particles is apo A, with a capital A. And now I am telling you that there is another particle that you probably never heard about that is called Lp(a) and that has both apo B and another protein called apo(a) but not apo capital A? Yes, and I am sorry about the complicated naming of all this stuff, but I swear I had nothing to do with that. Once you wrap your head around it, it’s also not all that complicated. So, again, VLDL, IDL, and LDL particles all have one main protein in common, and that is called apo B. HDL particles all have apo A. Capital A. OK. Then there is Lp(a). It is like LDL in many ways, including that it has apo B, but it has this additional protein on its outer surface called apo(a). It’s a bit outside of these lipoproteins that we normally consider and talk about because it is larger than most LDL particles, at a diameter of about 25 nm. But because it has an additional protein compared to LDL particles, and because proteins are denser than lipids, it tends to be a little denser than LDL. Somewhere between LDL particles and HDL particles (see figure below).
OK, so why is this important enough to mention here? Because Lp(a) is considered even more of a risk factor for ASCVD than LDL particles, for a variety of reasons that are beyond the scope of this blog post. Suffice it to say that we really want to know whether we have low or high levels of Lp(a).
Now, if we do measure our apo B concentration, this does include our Lp(a) particles, because they do also contain one apo B protein molecule. Also, if we measure non-HDL-cholesterol, these do measure the cholesterol content in all lipoprotein particles that are not HDL, including Lp(a). And our most common method to measure LDL-cholesterol, which is actually not a lab test but a calculation, actually is an estimate of the cholesterol concentration not just of LDL particles, but of LDL-particles PLUS Lp(a) particles. In case you are interested in why that is, most of the time, LDL-cholesterol is not actually measured, but calculated using the so-called Friedewald formula:
LDL-cholesterol (calculated) = total cholesterol – HDL-cholesterol – (triglycerides / 5).
HDL-cholesterol is usually measured directly by removing the apo A containing lipoproteins from blood, and then measuring the cholesterol concentration in these. Triglycerides are included in this formula because the triglyceride concentration divided by 5 is, on average, a good approximation of the cholesterol content in VLDL and IDL particles. What remains as the calculated LDL-cholesterol concentration is the cholesterol content in lipoprotein particles that are not HDL, VLDL, or IDL, and these are LDL and Lp(a).
Well, why do I bother you with this complicated stuff, then? Because even though Lp(a) is included in apo B, non-HDL-cholesterol, and LDL-cholesterol, it is so much more atherogenic than other atherogenic lipoproteins that knowing its concentration independently improves our ability to predict an ASCVD event. And the interesting part about Lp(a) is that its concentration in plasma is very low in most people, low being defined as less than 30 mg/dL. However, in about 20% of people, Lp(a) can be elevated, mostly for genetic reasons. We consider an Lp(a) concentration of 30 to 50 mg/dL as modestly elevated, and a concentration of 50 mg/dL or higher as substantially elevated. Because Lp(a) is strongly genetically determined, it’s probably sufficient to just measure your Lp(a) once in your lifetime, and use that information to gauge your overall risk and be more aggressive in addressing your other ASCVD risk factors should the Lp(a) concentration exceed 30 mg/dL and particularly 50 mg/dL. By the way, the conversion here is not straightforward, but if your lab provides Lp(a) concentrations in nmol/L, you can divide these by 2.25 to get an estimate of the concentration in mg/dL.
To summarize, one key measure you want to consider is either apo B or non-HDL-cholesterol or LDL-cholesterol. And then get a measurement of Lp(a) once in your life to determine how aggressively you may need to keep an eye on other ASCVD risk factors.
There are two other ASCVD risk factors I’d like to touch on: HDL-cholesterol and triglycerides.
First, HDL cholesterol. This is the amount of cholesterol carried on HDL particles, and in general, higher levels are associated with lower ASCVD risk. However, the relationship between HDL particles and ASCVD is complex, and a full discussion is beyond the scope of this blod post. In short, the current understanding is that HDL does have active anti-atherosclerotic functions, but that these are related less to its cholesterol cargo than the composition and function of the more than 50 proteins that are associated with it. Also, raising HDL-cholesterol has not so far been found to reduced ASCVD risk. Still, suffice to say that HDL-cholesterol is still used in ASCVD risk prediction models, and is also useful to detect the emergence of the insulin resistance syndrome, as discussed in the last blog post.
An alternative measure to HDL-cholesterol is apo A (most commonly, Apo A1), and sometimes people also look at ratios, such as the total cholesterol to HDL-cholesterol ratio, the LDL-cholesterol to HDL-cholesterol ratio, and the apo B to apo A-ratio. In general, this is unnecessarily complicated. Pay attention to apo B, non-HDL-cholesterol, or LDL-cholesterol. Consider your Lp(a). And then consider your HDL-cholesterol or apo A. These are the three key blood lipid-related ASCVD risk factors.
What about triglycerides, then? In general, elevated triglycerides are associated with an increased ASCVD risk. However, once all of the other risk factors are considered, whether or not we have high triglycerides usually doesn’t add all that much, and therefore triglycerides are usually not considered in risk prediction models. The reason for this is that people with elevated fasting triglyceride levels usually also have low HDL-cholesterol, and are more likely to have hypertension and type 2 diabetes. And once these factors are included in risk prediction models, adding triglycerides usually does not improve the ability of the model to predict ASCVD event risk. That’s why I see the use of triglycerides mostly as a measure of metabolic health, as explained in the last blog post about the insulin resistance syndrome.
Now, what we have discussed so far may sound quite convincing, but you may still be confused because you heard conflicting pieces of information.
Addressing Common Claims Suggesting that LDL-Cholesterol is not a Risk Factor for ASCVD
In this section, we’ll address severala specific claims that question the role of LDL-cholesterol in ASCVD. Specifically, we are going to discuss:
- that many people who have heart attacks have normal LDL-cholesterol levels, so LDL-cholesterol cannot be the cause of ASCVD;
- that lowering LDL-cholesterol reduces the ABSOLUTE ASCVD risk only minimally, suggesting that LDL-cholesterol is not all that important;
- that only small, dense LDL is harmful;
- that other risk factors, such as insulin resistance or metabolic health, are more important than LDL-cholesterol and that LDL-cholesterol can safely be ignored;
- that elevated LDL-cholesterol on low-carbohydrate or ketogenic diets is not atherogenic; and
- that cholesterol cannot be harmful because it’s a natural substance made by the body and needed for numerous important functions.
All of these are common talking points made in YouTube videos with Millions of views, often made by people with MD or PhD degrees, so I do think it is important to address these here because they are a source of much confusion.
Claim #1: Many People Who Have Heart Attacks Have Normal LDL-Cholesterol
The first claim that I have heard often is that many people who have heart attacks have normal LDL-cholesterol levels, so LDL-cholesterol cannot be a cause of ASCVD.
It is undoubtedly true that many people have heart attacks even though they never had particularly high LDL-cholesterol concentrations. However, among the people who have heart attacks, there are also many who don’t have diabetes, who don’t smoke, and who don’t have chronic kidney disease, yet no one questions that diabetes, smoking, and kidney disease are major independent risk factors for ASCVD. Again, it is the cumulative exposure to all of the risk factors that matters.
Also, we should consider that even LDL-cholesterol concentrations that we traditionally consider ‘normal’, such as 90-110 mg/dL (2.4-2.9 mmol/L) are still associated with progression of atherosclerosis, and if paired with other risk factors can certainly contribute to a clinical ASCVD event. At the same time, ASCVD events are exceedingly rare in individuals who have very low LDL-cholesterol concentrations of 50 mg/dL (1.3 mmol/L) or lower. So this claim is rooted partly in an incorrect definition of what ‘normal’ LDL-cholesterol levels are, partly in not considering the independent contributions of other ASCVD risk factors.
Claim #2: Lowering LDL-cholesterol reduces the absolute ASCVD risk only minimally
The second claim is that lowering LDL-cholesterol reduces the ABSOLUTE ASCVD risk only minimally, suggesting that LDL-cholesterol is not all that important.
This claim is related to an important statistical concept, namely the difference between relative and absolute risk reduction. Let’s assume we were to conduct a trial in people with elevated LDL-cholesterol levels, and we randomized them to a lipid-lowering medication or placebo. Over 4 years, a typical finding may be that 4% of the people in the placebo group have a cardiovascular event, such as a heart attack or a stroke, while only 3% in the treatment group have a cardiovascular event. In this example, the RELATIVE risk in the treatment group would be reduced by 25% compared to the placebo group. And that’s the figure that is often reported. However, the ABSOLUTE risk is only reduced by 1%, right? From 4% to 3%, and critics often comment that the pharma companies reporting on their trial data inflate their very modest findings by reporting the relative rather than the absolute risk reduction.
Well, what do you think? Is this cheating, or does it make sense to you to think about the risk reduction in relative terms? To clarify, the assertion that the absolute risk reduction is very small is totally correct, clearly there in the data, so there is no need to debate that.
However, I certainly do not agree with the critics here, at all. And I actually think that their argument is, frankly, not well-considered. Why is that? For one, obviously, the absolute risk reduction is going to be small if, even in the placebo group, only 4% of people have a cardiovascular event. However, the bigger point is that most people are not just interested in staying healthy for another 4 years. Instead, most of us are probably interested in minimizing our risk of a heart attack or stroke for another 10, 15, 20, or even 25 years, right? And so, what if we extrapolated these lines here to, say, 20 years? Now, we need to be clear that we cannot know what would happen, we can only guess. But given the difference seen at 4 years, I dare say that our best estimate would be that the absolute difference in risk would be much greater than the 1% observed after 4 years.
In this regard, I’d like to make an important point: the proper way to think about the relationship between LDL-cholesterol or apo B and ASCVD risk is to consider the concentration multiplied by the duration of exposure. So rather than think about our current LDL-cholesterol level as the risk factor, we should think about the area under our lifetime LDL-cholesterol concentration as the risk factor. Maybe like shown in the figure below for a 60-year-old. Time is on the x-axis, and the LDL-cholesterol level throughout these 60 years is on the y-axis. So if someone had high and increasingly higher LDL-cholesterol concentrations all his life, and starts taking a lipid-lowering medication at the age of 60, their risk of a cardiovascular event is not determined only by the now lowered LDL-cholesterol level, but by the sum of all of the LDL-cholesterol levels throughout their lifetime. And that initially makes for a rather small difference in the overall exposure between someone who is treated compared to someone who is not. See, after 4 years, this area I have shaded yellow in the area under the LDL-cholesterol curve would be the only difference between someone who does start lipid-lowering therapy and someone who does not. Pretty small difference in the overall lifetime exposure to LDL-cholesterol. So, to me, the fact that we can detect a difference in cardiovascular event risk within a few years speaks volumes as to the importance of LDL-cholesterol in ASCVD.
My last point is that by lowering LDL-cholesterol, we are addressing only one risk factor out of many in this multi-factorial disease. In the short term, we also have similar, fairly small benefits of treating hypertension or diabetes, or of smoking cessation. I feel strongly that if we address each of the risk factors and minimize our exposure even a bit, the sum of several small effects over a short period of time would be expected to amount to a huge relative and ABSOLUTE benefit over a longer period of time, and that’s what most of us are probably interested in. The figure below illustrates this concept schematically. Note that the graphs are not based on actual data; however, the short-term effects are directionally consistent with published trials.
Taken together, given that lifetime exposure matters, and that there are numerous risk factors that each contribute to ASCVD, we simply cannot expect that addressing a single risk factor by itself will have a high absolute benefit in the short term. However, if we stack several seemingly small short-term benefits for each of several ASCVD risk factors, our risk will be dramatically reduced over the longer term.
Claim #3: Only small dense LDL particles are atherogenic
The third claim is that only small dense LDL particles are atherogenic, and that large, buoyant, or fluffy, LDL particles are not harmful. Now, first of all, what is that all about?
Remember the beginning of the blog post when we discussed the basics of lipoproteins? I mentioned that VLDL, IDL, and LDL span a wide range in terms of both density and size. And that means that there are LDL particles that are smaller and denser than others. We call these small, dense LDL particles, while we call those on the other end large, buoyant, or “fluffy” LDL particles. A common claim is that the risk of ASCVD only increases if we have many of these small dense LDL particles, and that the large and fluffy LDL particles are not harmful.
The first question to address is whether it is true that small dense LDL particles are more atherogenic than large, fluffy ones. Yes, there are quite a few pieces of evidence to suggest that these small dense LDL particles have certain properties that make them more atherogenic. For example, they seem to be enriched in certain proteins other than apo B that may make them more atherogenic. They are more susceptible to oxidative damage and to glycation; glycation being when sugar molecules attach to the proteins within the particle. They may more easily enter the sub-endothelial space than larger apo B-carrying particles. And they also stay in circulation for longer than the large, fluffy LDL particles because their binding affinity to the LDL receptor is lower than that for large, buoyant LDL particles. Also remember that small-dense LDL are more dense because they have more protein and less cholesterol than the large fluffy ones. Because the main protein in LDL particles is apo B, this means that the apo B concentration in the blood will be higher at the same LDL-cholesterol level if someone has more small dense LDL particles.
So having more small dense LDL particles than large fluffy ones is probably associated with a greater risk of ASCVD. However, does this mean that large, fluffy LDL particles are not harmful?
First, no one has only small dense, or only large buoyant LDL particles. To understand this point, let’s first clarify how we can figure out what kind of LDL particles we have. There are a number of different methods to determine this, and one method we have used in my lab is to measure the cholesterol content of each of these LDL types separately. So we would not just separate VLDL from IDL from LDL from HDL, but also separate again within each of these classes. So for LDL, we would isolate 12 different types of LDL, and then measure the cholesterol content in each (see the figure below).
Let’s take a look at two participants from one of our own clinical trials. These two have exactly the same overall LDL-cholesterol level, but if we measure the cholesterol content in each of the different types of LDL separately, we see that the blue person has most of the cholesterol within LDL particles that are rather large and fluffy, or of this middling type, whereas they have very little cholesterol in small dense LDL particles. On the other hand, the person in red has most of their cholesterol in small, dense LDL particles. If we quantify the cholesterol content just within the small, dense LDL particles, that value is 60 mg/dL in the blue person, but 88 mg/dL in the red person. Again, that is even though both have exactly the same LDL-cholesterol concentration. Without knowing anything else about these individuals, we would assume that the person in red with more small, dense LDL has a higher risk for ASCVD. But the point to remember is that even in relatively extreme cases like these, both individuals have a mixture of small, dense, and large fluffy LDL particles.
Another important point can be made by contrasting the red and blue people with the green person in the figure. This green person has a lot less cholesterol in their LDL particles overall. In other words, their LDL-cholesterol level is lower, actually 65% lower. Their cholesterol content in small, dense LDL is 26 mg/dL, much lower than that of the red and blue person. So what we are learning here is that if someone has high LDL-cholesterol levels, even if most of those are not small and dense, such as the person in blue, they could still have a relatively high absolute concentration of small dense LDL particles compared to someone with an overall low LDL-cholesterol level. So, my second point is that one of the biggest factors that determines the amount of small, dense LDL particles is the total LDL-cholesterol concentration.
Third, let’s also be clear that the idea that large, fluffy LDL particles are harmless is not consistent with the literature. There is wide consensus that all apo B particles are atherogenic to some degree; it’s just that small dense particles may be a bit more atherogenic. In statistical models predicting who develops ASCVD, total LDL-cholesterol remains strongly associated with ASCVD risk in most studies, even if we add a measure of small dense LDL to the model. In other words, considering the number of small, dense LDL particles, or the cholesterol content in small, dense LDL particles, may improve our ability to predict an ASCVD event, but that is in addition to the predictive value of LDL-cholesterol, not in place of it. This is even more so the case for apo B, as total apo B tends to correlate with the number of small, dense LDL particles.
Fourth, a shift towards more small dense LDL is mostly seen in people who are insulin resistant and potentially glucose intolerant, who have excess visceral and liver fat, who have elevated triglycerides and low HDL-cholesterol, who suffer from low-grade chronic inflammation, and hypertension. In other words, having small, dense LDL is part of the insulin resistance syndrome that we talked about in the last blog post. To illustrate this, let’s go back to our example in the figure above. The person with more small dense LDL in red does have fatty liver disease and fasting triglycerides of 247 mg/dL. The person in blue does not have fatty liver disease, and fasting triglycerides of 87 mg/dL. These differences are not a coincidence, because higher triglycerides can increase the formation of small, dense LDL particles.
Now, why is that relevant? Consider that when we assess the overall risk of someone having a heart attack or a stroke, we consider all of the major independent risk factors for ASCVD: hypertension, diabetes, low HDL-cholesterol, and sometimes triglycerides. Once we consider all of these risk factors, knowing whether our LDL are small and dense or large and fluffy doesn’t improve our ability to predict ASCVD enough to warrant the additional considerable cost to assess the LDL size distribution. That is because many of these additional risk factors are strongly associated with having more small dense LDL. In a way, we are already considering important components of the insulin resistance syndrome in our risk prediction model, so adding another element of the syndrome, namely small, dense LDL, provides little additional information.
To conclude this part, yes, whether we have small dense LDL particles matters, but this does not mean that large, fluffy LDL particles are irrelevant, and if we assess ASCVD risk considering all of the key risk factors, knowing the size and density of our LDL particles adds fairly little to our ability to predict ASCVD event risk.
If you’d still like to get a sense of whether your LDL-particles are mostly small and dense, you can use the insulin resistance syndrome poster I discussed in the last blog post. If your insulin resistance and triglycerides are in the orange or red portions of the poster, you can assume that you have some or a lot of small, dense LDL particles. You can also get this measured directly: just google NMR LipoProfile for a test that is commercially available in the United States. As outlined here, I do not think this is necessary, as it contributes little to our overall ASCVD risk assessment.
Claim #4: Other Risk Factors Are More Important than LDL-Cholesterol
Let’s move on to another claim, and that is that other risk factors, such as insulin resistance or metabolic health, are much more important than LDL-cholesterol. And therefore, the claim is, that we can safely ignore LDL-cholesterol if we are metabolically healthy.
Well, I don’t agree that LDL-cholesterol can be safely ignored, but I agree that taken together, risk factors clustered around insulin resistance may well be more important. Because, as discussed in detail in the last blog post, the insulin resistance syndrome is strongly associated with elevated fasting triglycerides and low HDL-cholesterol, hypertension, low-grade chronic inflammation, and type 2 diabetes. Plus, the insulin resistance syndrome also often modestly raises LDL cholesterol, and also leads to a shift towards more small, dense LDL particles. So, yes, I agree that the insulin resistance syndrome probably constitutes the most severe cluster of several major risk factors for ASCVD. Let’s illustrate this using an example:
Remember Joe Average from our last blog post? At 65 years of age, he is obese and has type 2 diabetes, hypertension, elevated triglycerides and LDL-cholesterol, and low-HDL-cholesterol. This is a common clustering of risk factors that are all, in one way or another, linked to insulin resistance or its main causes, which include low-grade chronic inflammation and excessive visceral and ectopic fat. His risk of having a heart attack over the next 10 years is 39.4%.
Let’s assume he is worried about his heart disease risk and his diabetes, and starts a strict ketogenic diet. Over the next year, he loses most of his excess weight, and his diabetes goes into remission. His blood pressure also drops, and his fasting triglycerides and HDL-cholesterol normalize, but his LDL-cholesterol increases more. This is not an uncommon scenario in someone who adheres to a ketogenic diet for an extended period of time. However, outcomes with any diet vary greatly, so some people may have a less beneficial response to a ketogenic diet, and also not everyone will experience a major increase in LDL-cholesterol levels. However, these changes in Joe are not untypical.
And what happens to his risk of having a heart attack: it drops substantially in spite of the increase in LDL-cholesterol, to 14.5% over 10 years. So yes, this part is correct: we need to consider all risk factors together, and it is well possible to lower your risk substantially even if LDL-cholesterol increases. However, this does not mean that LDL-cholesterol should be ignored. Even with this improved metabolic profile, the reversed diabetes, lower blood pressure, and the higher HDL-cholesterol level, having LDL-cholesterol levels of 240 mg/dL is not ideal. Let’s say his LDL-cholesterol was around 70 mg/dL instead, his risk would drop to 8.7% over 10 years.
Another, related claim is that LDL-cholesterol is only an important ASCVD risk factor in people who are insulin resistant. In this context, it is worthwhile to consider specific studies conducted to assess the relationship between LDL-cholesterol and atherosclerosis specifically in metabolically healthy individuals with an overall very low risk factor profile. For example, the PESA study conducted in Spain showed a linear relationship between LDL-cholesterol and subclinical atherosclerosis even among young, asymptomatic individuals without any evidence of metabolic dysfunction. Similarly, the Cooper Center Longitudinal Study also found that LDL-cholesterol levels are strongly associated with 10-year risk of ASCVD in a group of men and women of overall very low ASCVD risk.
At the same time, it should be noted that even these studies in low-risk cohorts have not characterized important metabolic factors, such as insulin sensitivity, well, usually just focussing on clinical factors such as diabetes or hypertension. It therefore remains somewhat unclear whether the relationship between LDL-cholesterol and ASCVD is modified by perfect insulin sensitivity. In the absence of perfect data, however, it would seem prudent to use the currently available suboptimal data as our best estimate, and that would lead us to conclude that the overall ASCVD risk will be low in insulin sensitive participants, because they will usually not be exposed to some of the common risk factors (diabetes, hypertension, elevated triglycerides, low HDL-cholesterol, low-grade chronic inflammation), but that the relationship between LDL-cholesterol and ASCVD risk will not be fundamentally modified in this population.
I am hoping this makes sense: LDL-cholesterol or apo B should not be seen as the be-all-end-all, but neither should metabolic health. If we value our health, we want to consider all of the risk factors together, and LDL-cholesterol or, better, apo B, clearly is one of the key risk factors.
Claim #5: Elevated LDL-Cholesterol on Low-Carbohydrate or Ketogenic Diets is Not Atherogenic
I know many of you are on low-carb or ketogenic diets, so I wanted to take some time to address this. The claim, in essence, is that someone like Joe Average on a ketogenic diet, as in our prior example, can just ignore his elevated LDL-cholesterol levels of 240 mg/dL because he would NOT suffer any negative consequences from them.
Let me start by stating clearly that I know of no evidence, or even a potential mechanism or hypothesis that would suggest that the relationship between LDL-cholesterol and ASCVD should be different on a low-carb diet in someone like Joe here. As I have explained above, this 66-year old Joe will have a greatly reduced risk of ASCVD compared to the 65-year old Joe because of the normalization of his body weight, the reversal of his diabetes and hypertension, and his lower triglyceride and higher HDL-cholesterol levels. No one should be denying that. However, the elevated LDL-cholesterol should still be seen as a risk factor, and should ideally be lowered to reduce his ASCVD risk further. For those geeks among you, in science speak: I see no evidence or plausible mechanism that being on a low-carb diet could be an effect modifier in the relationship between LDL-cholesterol and ASCVD risk.
Now, that said, there is one particular phenotype that manifests in some people on a ketogenic diet that is worth discussing some more. This phenotype has been described by Dave Feldman, Nick Norwitz, and their colleagues, and is called the lean mass hyperresponder, or LMHR, phenotype. LMHRs are lean people who have normal LDL-cholesterol levels on a mixed diet, but develop very high LDL-cholesterol concentrations exceeding 200 mg/dL on a ketogenic diet, intriguingly combined with also very high HDL-cholesterol concentrations of 80 mg/dL, or higher, and low triglycerides of 70 mg/dL, or lower.
Considering these criteria, Joe would not fit LMHR criteria because his triglycerides are too high and his HDL-cholesterol too low. However, if Joe on the ketogenic diet had lost a little more weight to a BMI of 22, his triglycerides would have dropped to 50 mg/dL and his HDL-cholesterol had risen to 90 mg/dL, he would be a typical lean mass hyperresponder. And just by entering these slightly changed numbers into a heart disease risk calculator, we can see right away that these changes would reduce his heart disease risk further, to 12% over 10 years. So yes, the more favorable blood lipid profile in someone with the LMHR phenotype would have beneficial effects on the overall ASCVD risk.
However, Dave and Nick have proposed the hypothesis that the ASCVD risk associated with the massively elevated LDL-cholesterol concentrations in lean mass hyperresponders may be even lower than traditional risk estimation models suggest. In other words, they propose that elevated LDL-cholesterol levels may be less atherogenic in people with this particular phenotype.
Why would they think this? After all, this would be the first circumstance that I can think of that such massively elevated LDL-cholesterol levels would NOT be associated with a similarly increased risk of ASCVD. Well, Dave and Nick have pointed out, correctly I think, that the lipid and lipoprotein metabolism in these LMHRs is very unique, and that we do not have data on the relationship between LDL-cholesterol and ASCVD risk that were generated in a population even remotely resembling this phenotype. That is because their LDL-cholesterol is elevated not because of a genetic defect in the handling of lipoproteins, as in people with genetically elevated LDL, but because the lipoproteins are shuttling more fats around the body, because fats are the primary source of fuel on a ketogenic diet. So they argue that this is mostly a metabolic adaptation to changed fuel use in the body on a ketogenic diet. And that the visible sign of that adaptation is this very unusual combination of high LDL-cholesterol levels with unusually high HDL-cholesterol levels and low triglycerides. I think personally that this is a fair assessment, because we do indeed not see that combination ever outside of lean, metabolically healthy people on a ketogenic diet, and certainly not usually in people with genetically elevated cholesterol levels. Now, whether the cause of the elevated LDL-cholesterol matters for the associated ASCVD risk is an entirely different question, of course.
I would also add that to be a lean mass hyperresponder, you probably have to have dietary and lifestyle habits that are very different from that of the general population: I haven’t seen much data on this, but I would guess that LMHRs exercise a lot, and consume very little junk food or sugar-sweetened or alcoholic beverages. Otherwise they would not be in a position to be lean mass hyperresponders on a ketogenic diet because they would not actually be lean or eating a ketogenic diet. So, yes, for these reasons, I do agree that Dave and Nick have identified people with a very unique phenotype. And I think their hypothesis that the elevated LDL-cholesterol levels of LMHRs on keto may not cause a progression of atherosclerosis, or at least not as much as we would otherwise think, is interesting and certainly worthy of study.
To their credit, Dave and Nick have initiated a small, crowd-funded pilot study with 100 such lean mass hyperresponders who had been on a ketogenic diet for an average of 4.7 years. The overall goal of the study is to assess whether they develop atherosclerotic plaque over time. This will be assessed by repeated CT angiograms to quantify coronary artery calcium deposition and the amount of coronary artery plaque. The main, longitudinal results of the study are still pending (see figure below).
However, the team has completed the baseline data collection, including the measurement of coronary artery calcium and coronary artery plaque. They then matched 80 of these LMHRs to participants from another, existing cohort study, the Miami Heart Study. The goal here was to match participants in terms of the major ASCVD risk factors, with the exception of the serum lipid profile. And as you can see in the table below, these groups were well matched for age, gender, race, and all of the major ASCVD risk factors, except for BMI, total-, LDL-, and HDL-cholesterol, and triglycerides. However, all of these are part of the LMHR phenotype, so that difference is by design. In terms of the major ASCVD risk factors, the main difference, by design, was in LDL- and HDL-cholesterol levels, with an average LDL-cholesterol of 272 mg/dL in the LMHRs on keto compared to 123 mg/dL in the matched contros. That’s certainly a sizeable difference, and based on everything we know, atherosclerosis should progress faster in those on keto because of these massively elevated LDL-cholesterol levels. And in fact, because these LMHRs had been exposed to these high LDL-cholesterol levels already for an average of 4.7 years, they already may have more coronary plaque than the controls.
To assess whether this is indeed the case was the objective of a cross-sectional analysis that was conducted at baseline. The results were presented by the principal investigator, Dr. Matthew Budoff, at a conference last week (December 8, 2023). Notably, after 4.7 years on a ketogenic diet, the CT angiography data did not show any differences between these two groups in terms of coronary artery calcium scores or coronary artery plaque scores. The total plaque score was zero in more than 50% of the keto group, again, in spite of having had very substantially elevated LDL-cholesterol levels for an average of 4.7 years. Now, that is an interesting finding that seems to support Dave and Nicks hypothesis that these isolated elevations in LDL-cholesterol in LMHRs on keto are less atherogenic. However, I do think a nuanced discussion is required to draw the right conclusions from these results. (By the way, if you’d like to hear these investigators discuss their cross-sectional findings, here is a link to an interview they have recently given).
First, I saw a lot of comments by lipidologists or cardiologists online who stated that there was no chance that 4.7 years of LDL-cholesterol of, on average, 272 vs. 123 mg/dL could have resulted in differences in coronary calcium or plaque. Many of them felt that these data are therefore totally meaningless.
So let’s address that assertion. I’d say yes, for sure, if these were the first 4.7 years of life, this would be too short. Atherosclerotic plaque can develop in less than 5 years, but that would require much higher LDL-cholesterol levels. However, these were not the first 4.7 years of life of these participants. Instead, in both the LMHR group and the matched controls, 4.7 years before this baseline assessment, the average age in both groups was about 50, and we can assume that they had a similar exposure to LDL-cholesterol up to that point. Maybe as shown in the figure below, on average. Remember, it’s the cumulative exposure to lifetime LDL-cholesterol that matters, and by the time they turned 50, participants in both groups had already accumulated a whole bunch of LDL-cholesterol exposure. Probably something like 5,000 mg/dL years, which has been suggested as a threshold at which ASCVD risk starts to manifest.
How did I arrive at 5,000 mg/dL years? Well, this looks like the average LDL-cholesterol concentration over the first 50 years of their lives was about 100 mg/dL, right? And 100 mg/dL multiplied by 50 years, and we arrive at the average cumulative lifetime exposure for participants in each group at the age of 50: 5,000 mg/dL years. Hope this makes sense?
So that was already the estimated lifetime exposure to LDL-cholesterol when the LMHRs began their ketogenic diet. Then they went on keto, and their LDL-cholesterol shot up to about 270 mg/dL. At least, that’s our best estimate. So now, at the age of 55, which is when this baseline assessment took place, we do have quite a substantial difference in the overall cumulative lifetime exposure to LDL-cholesterol (shaded in yellow in the figure above). So I don’t agree that there was no chance that a difference could be seen. In my opinion, the overall lifetime cumulative exposure to LDL-cholesterol could certainly have lifted some LMHRs into the territory where we see substantial and rapid progression of atherosclerosis, and much more likely so than for the matched controls.
At the same time, I do think it is important to acknowledge that the time on keto COULD have been too short. We also need to be clear that this was a cross-sectional baseline analysis of a small pilot study. These are interesting and promising data, but they are the very first step in a long journey that hopefully will include long-term follow up of these participants for at least another 5, better, 10 years, and ideally also a second, much larger study before we can with good justification and some confidence conclude that these elevated LDL-cholesterol concentrations in LMHRs are indeed less atherogenic. Time is a huge factor here, and my best estimate is still that LMHRs will develop atherosclerotic plaque at a greater rate than the matched controls. I’d be very happy if I was wrong, but that is my best estimate at this time.
We therefore also need to be clear that these data do not in the slightest affect our general understanding of the role of LDL-cholesterol in ASCVD risk, as I have seen several people suggest online. We have a huge body of evidence from animal, observational and genetic studies, and even data from randomized controlled high-quality trials with more than 2 million participants, and the cross-sectional baseline data from a 160-participant pilot study certainly cannot even remotely change our general understanding of the relationship between LDL-cholesterol and ASCVD. So the assertion that these data raise questions about the lipid-heart hypothesis is more than silly.
Where the value of this line of investigation lies is that the lipid and lipoprotein phenotype is indeed very unusual in LMHRs, and LMHRs almost certainly also have an unusually healthy lifestyle. This offers the possibility that we could learn something fundamentally new about lipid metabolism and its relationship to ASCVD, or about ASCVD in general. That would be the case, for example, if we were to find that these massively elevated LDL-cholesterol concentrations do not lead to the expected progression of ASCVD even over the longer term, or at a much slower than expected rate. Maybe there is something special about the composition or metabolism of LDL particles in LMHRs on a ketogenic diet, or maybe the other dietary or lifestyle factors that lead to someone having an LMHR phenotype are somehow very protective. I do think we should be open and curious about these possibilities.
So, overall, I am excited about this line of work and hope that the investigators will be able to longitudinally follow these LMHRs for, ideally, at least another 5, better 10 years. And ideally, start another cohort with a larger sample size. In the meantime, let me reiterate that I think the prudent approach is to assume that the relationship between LDL-cholesterol and ASCVD risk holds in this population, until we have much more solid data. As promising as these presented baseline data are, they are much too limited to be considered solid evidence. In other words, these baseline data do not change my view of the role of LDL-cholesterol in ASCVD, in general or in LMHRs on ketogenic diets. As much as I am intrigued by this hypothesis, I certainly would not bet my life on it being right at this point. If I had such massively elevated LDL-cholesterol on a ketogenic diet, I would modify my diet to lower my LDL-cholesterol levels, or seek lipid-lowering therapy.
And let me close by stating again that the hypotheses about LMHRs should NOT be generalized to others on ketogenic or low-carb diets. If you follow a low-carb or ketogenic diet, and you have elevated LDL-cholesterol, but you are not very lean, or you also have low-HDL-cholesterol, or high triglycerides, or other risk factors, such as smoking or kidney disease, none of what we discussed here applies to you in my opinion. I know of no mechanism through which low-carb or keto diets plausibly could change the impact of LDL-cholesterol on ASCVD risk, outside of this very specific LMHR phenotype. And even for LMHRs, please remember that we are currently working with a hypothesis in the very early stages of clinical testing.
Claim #6: Cholesterol Cannot Be Harmful Because it’s a Natural Substance Made By the Body and Needed for Numerous Important Functions
And with that, let’s tackle the last claim, and that is that cholesterol cannot be harmful because it is a natural substance made by the body and needed for numerous important functions.
Now, I think if you have watched any other blog post on this channel, YOU may be able to address this. Are there no substances that are natural, that the body can make, that have important functions in the body, and that can still increase the risk of chronic disease if their concentration in our blood is too high?
Yes, there are, of course, countless numbers of such substances. Glucose is a good example that’s similar to cholesterol: glucose is certainly a natural substance. If blood glucose drops too much, we die, so for sure, glucose is essential for our survival, and therefore, we can safely assume it has important functions. If we don’t eat enough glucose, our body will therefore produce it endogenously. But if we have too much glucose in our blood, as in diabetes, our risk of chronic disease increases. Or how about insulin? It’s natural, we can make it, but too little or too much of it will kill us. We need just the right amount of glucose and insulin in our blood for good long-term health. Too little and too much will be a problem. Why do we accept this for glucose, insulin, and countless other substances, but for cholesterol, we keep making this argument that it’s important for our cell membranes and an important substrate for hormone or bile acid synthesis, so too much cholesterol in the blood cannot be bad? It’s puzzling, and I hope you can see it’s a silly argument.
Still, let’s address the question of whether low LDL-cholesterol levels are associated with negative health consequences. The available data do suggest that while certain lipid-lowering medications can have adverse effects, these are not per se related to the LDL-cholesterol levels being too low, and in fact, lowering LDL-cholesterol to 11-45 mg/dL in extreme lipid-lowering szenarios was not found to be associated with negative health consequences. This is consistent with individuals who have similarly very low LDL-cholesterol levels for genetic reasons, who also do not seem to suffer any negative health consequences.
Summary and Conclusions
In conclusion: the available evidence clearly, consistently, and strongly supports the idea that apo B containing lipoproteins play a causal role in the development of ASCVD. To say this explicitly, the evidence is overwhelming, and I find it highly concerning, unfortunate, and potentially very harmful that some influencers just brush the incredibly strong and consistent data aside with usually rather superficial or poorly considered arguments.
I therefore strongly recommend regularly measuring the fasting plasma apo B concentration, and also to use apo B to guide nutritional or drug therapy. Non-HDL-cholesterol, as assessed by subtracting HDL-cholesterol from total cholesterol, or LDL-cholesterol are good alternative measures. For both apo B and LDL-cholesterol, levels below 80 mg/dL are optimal in the context of primary prvention. For non-HDL cholesterol, I would aim for a concentration of 100 mg/dL or lower. Disclaimer: these are not official guidelines, but my synthesis of how I think of these concentrations for the primary prevention of ASCVD. Please note that commonly lower targets are set in secondary prevention, i.e., in individuals who have suffered a heart attack or stroke.
I also recommend getting a measurement of Lp(a) at least once in your lifetime to improve ASCVD risk prediction and to guide therapy. If Lp(a) is elevated above 30 mg/dL and particularly above 50 mg/dL, then other ASCVD risk factors should be more tightly controlled.
Among the other lipid-related risk factors, it is also good to regularly assess HDL-cholesterol and triglycerides, both as ASCVD risk factors but also as convenient measures to monitor the emergence of the insulin resistance syndrome, as explained in the last blog post, which I have linked below. For HDL-cholesterol, an optimal concentration is 60 mg/dL or higher in men and 70 mg/dL or higher in women.
With all of this discussion of lipid-related risk factors, let me emphasize again that ASCVD is a multi-factorial disease, and that it is critical to use all of the known independent risk factors in the assessment of risk, and to guide treatment. These include age, gender, race, diabetes, hypertension, smoking, chronic inflammatory conditions, chronic kidney disease, HDL-cholesterol, Lp(a), and apo B or non-HDL-cholesterol or LDL-cholesterol.
In your evaluation of ASCVD risk, particularly consider strongly that what matters most for each risk factor is the exposure multiplied with the duration of the exposure. This line of research strongly suggests that earlier intervention and – even better – prevention are golden to minimize overall lifetime exposure to all of these risk factors.
And lastly, let me emphasize that this blog post is not a plug for drugs. In fact, one of the primary objectives of this website is to help you live a long and healthy life without needing drugs. My objective with this blog post was simply to provide unbiased guidance on how to use blood lipids to assess ASCVD risk. Whether or not drugs may be helpful to lower an elevated risk certainly needs to be evaluated on a patient-by-patient basis, with an experienced physician, hopefully after other methods of lowering the risk have been considered, and after a careful discussion of drug side effects. So clearly, medication can be a useful option, but in my opinion, it would be much better to develop solid diet, exercise, and lifestyle habits, ideally early in life, to keep all of those risk factors we can influence in the optimal or close to optimal range for as long as possible. That would minimize the risk of ASCVD without the need for any meds. And if any of these risk factors are already elevated, we should, in my opinion, still prioritize addressing the root cause, which usually includes a poor diet, chronic overeating, a sedentary lifestyle, poor sleep, and chronic stress. Whether or not medications are indicated on top of that needs to be decided between you and your doctor.
OK, that was a very long blog post, thank you for hanging in there with me. I very much hope that this blog post addressed all of your questions and removed any confusion you may have had. If you still have any questions, please feel free to post them below. I also hope that it wasn’t too technical. Feel free to give me some feedback on that below.
References
- Feingold. Introduction to Lipids and Lipoproteins. Endotext 2021.
- Ginsberg. Lipoprotein physiology. Endocrinology Metabolism Clinics of North America 1998; 27: 503-19.
- Ference et al.; Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel. European Heart Journal 2017; 38: 2459-72.
- Wiegman et al.; Familial hypercholesterolemia in children and adolescents: gaining decades of life by optimizing detection and treatment. European Heart Journal 2015; 36: 2425-37.
- Horton et al.; PCSK9: a convertase that coordinates LDL catabolism. Journal of Lipid Research 2009; 50 Suppl.: S172-7.
- Masson et al.; Role of non-statin lipid-lowering therapy in coronary atherosclerosis regression: a meta-analysis and meta-regression. Lipids in Health and Disease 2020; 19: 111.
- Ference et al.; Variation in PCSK9 and HMGCR and risk of cardiovascular disease and diabetes. New England Journal of Medicine 2016; 375: 2144-53.
- Libby. The changing landscape of atherosclerosis. Nature 2021; 592: 524-33.
- Sniderman et al.; Apolipoprotein B particles and cardiovascular disease: a narrative review. JAMA Cardiology 2019; 4: 1287-95.
- Cole et al.; Use of apolipoprotein B in the era of precision medicine: time for a paradigm change? Journal of Clinical Medicine 2023; 12: 5737.
- Bilgic and Sniderman. Low-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol and apolipoprotein B for cardiovascular care. Current Opinion in Cardiology 2024; 39: 49-53.
- Kronenberg et al.; Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis: a European Atherosclerosis Society consensus statement. European Heart Journal 2022; 43: 3925-46.
- Schoch et al.; Update of HDL cholesterol in atherosclerotic cardiovascular disease. Clinical Investigations in Arteriosclerosis 2023; 35: 297-314.
- Farnier et al.; Triglycerides and risk of atherosclerotic cardiovascular disease: an update. Archives of Cardiovascular Disease 2021; 114: 132-9.
- Krauss. Small dense low-density lipoprotein particles: clinically relevant? Current Opinion in Lipidology 2022; 33: 160-6.
- Rizvi et al.; Lipoproteins and cardiovascular disease: an update on the clinical significance of atherogenic small, dense LDL and new therapeutical options. Biomedicines 2021; 9: 1579.
- Tsai et al.; New automated assay of small dense low-density lipoprotein cholesterol identifies risk of coronary heart disease. The Multi-Ethnic Study of Atherosclerosis. Arteriosclerosis Thrombosis and Vascular Biology 2014; 34: 196-201.
- Ibanez et al.; Progression of early subclinical atherosclerosis (PESA) study. Journal of the American Colleage of Cardiology 2021; 78: 156-79.
- Abdullah et al.; Long-term association of low-density lipoprotein cholesterol with cardiovascular mortality in individuals at low 10-year risk of atherosclerotic cardiovascular disease. Circulation 2018; 138: 2315-25.
- Norwitz et al.; Elevated LDL cholesterol with a carbohydrate-restricted diet: evidence for a “lean mass hyperresponder” phenotype. Current Developments in Nutrition 2021; 6: nzab144.
- Norwitz et al.; The lipid energy model: reimagining lipoprotein function in the context of carbohydrate-restricted diets. Metabolites 2022; 12: 460.
- Ference et al.; Impact of lipids on cardiovascular health. Journal of the American College of Cardiology 2018; 72: 1141-56.
- Sabatine et al.; Efficay and safety of further lowering of low-density lipoprotein cholesterol in patients starting with very low levels: a meta-analysis. JAMA Cardiology 2018; 3: 823-8.
55 Responses
Perhaps the single best piece of science communication and explanation that I’ve read on this topic over the last 10 years. Great work, and thank you for putting this together! Please keep the articles like this coming.
Thank you, Tyler.
This is a well-written, detailed, yet easy to understand explanation of the current accepted science re: lipids and ASCVD. This obviously took time to pull together and write. The illustrations are also very helpful. Thank you for all your efforts. This should be required reading for all podcasters who tell their audience that an elevated LDL is not a long term risk for major cardiovascular events.
Thank you, Karen. I think the issue is that most of those podcasters would not be convinced otherwise, no matter what evidence they are presented with. And the problem is that it is so easy to make some of these arguments: the issue of low short-term absolute risk reduction in lipid-lowering trials, for example, is correct, and easily verifiable. So it’s easy to make a statement that seems correct and then distort the conclusions a little bit and make it seem like a no-brainer that “we’ve been lied to” or that “it’s all about big pharma and big money” …
Best,
Mario
Fantastic! I read every word.
Thank you!
Thank you for an excellent presentation of the topic. I have two questions that I hope you can shine some light on:
1) You cite FH as primary evidence for causality of LDL-C for ASCVD. If I understand correctly, FH prevents proper recycling of LDL from the blood. For most people, this is not the reason for elevated LDL. Does the analogy in this case still hold?
2) You suggest trying lifestyle interventions to mitigate ASCVD risk. While dietary measures and exercise work very well to improve metabolic health, lower blood pressure, etc, unfortunately they often do not lower and in some cases even increase LDL. What non-pharmacological measures exist to reduce risk arising from LDL-C?
Thank you! Bernhard
Hi Bernhard,
Great questions.
Re 1.) You raise a reasonable concern, and it is likely that some people with FH do have other disturbances in lipid and lipoprotein handling that may contribute to the increased ASCVD risk. However, it is intriguing that there are several different genetic mutations that lead to both elevated LDL-cholesterol and an increased ASCVD risk, and the increase in ASCVD risk is pretty consistently associated with the LDL-cholesterol level. If the impact of these genetic mutations were mostly not related to LDL-cholesterol, but some other less obvious molecular mechanism, then one would expect the relationship between LDL-cholesterol and ASCVD risk to be a bit more random and messy. Still, we do not have 100% certainty on any line of evidence that it really is the LDL-chol, or better, the apo B-carrying lipoproteins, that are the culprit. Because even in the clinical lipid-lowering trials, one could make the case that statins, for example, have other effects on the body than just lowering lipids that could explain partly their impact on ASCVD risk. However, considering the large number of different lines of evidence, and that these are quite consistent across populations and different lines of study, the consensus in the field is that apo B-containing lipoproteins are likely causal. For example, you can find genetic variations in the genes targets by statins and PCSK9 inhibitors, and the genetic variations exactly mimic the LDL-cholesterol and ASCVD risk effects of statins and PCSK9 drugs (REF 7 in the blog post). It’s pretty remarkable how consistent some of this evidence is.
Re 2.) We’ll talk about this in detail in future videos, but in general, fiber content and the fatty acid composition play major roles in regulating apo B and LDL-chol. Less trans-fatty acids and long-chain saturated fatty acids and more unsaturated fatty acids is one dietary change that can have a pretty substantial impact. More fiber is another. And for people who do keto, eating a bit more carbs with every meal can make a huge difference. Lastly, avoiding overeating and fatty liver are key aspects of keeping blood lipids, including TGs and LDL-chol, lower. I talk about that in these blog posts here:
https://nourishedbyscience.com/personal-fat-threshold/
https://nourishedbyscience.com/insulin-resistance-syndrome/
Best,
Mario
Hi Mario, thank you very much for your insightful reply. I am looking forward to future videos discussing alternatives to reducing ASCVD risk than lipid-lowering medication.
Re 1) – I may be mistaken but from what I know one genetic defect associated with FH is a reduction or inhibition to clear LDL from the blood stream. This differs from from what seems to be the more common causes of high LDL (e.g. overeating or fatty liver and probably insulin resistance).
I have seen arguments that LDL that remains in circulation for a long time can get “damaged”, and that damaged LDL is causing ASCVD while LDL that is cleared efficiently presents a lower risk (also addressed in your presentation, comparing type “A” and “B” lipid patterns). The preliminary results from the LMHR study lend some credibility to this hypothesis: this cohort appears to be subject to a high level of LDL that seems less atherosclerotic.
If the LMHR studies (only one so far with a second in the planning stages) pan out, they may open new path to reducing ASCVD risk that does not require pharmaceuticals or dietary interventions like substituting unsaturated for saturated fat (with its own host of concerns, especially vegetable oils). I agree, though, that a ketonic diet demands a lot of effort and is not for everyone. A pill is certainly a simpler path – if it works.
I hope you do not interpret this as a critique – but I continue to be confused about ASCVD risk factors and mitigation strategies and feel we know less about the issue than most practitioners admit.
Regarding your last sentence: I value a good, respectful debate, and am totally fine if someone comes to a different conclusion.
I also understand your impression that we know less about this issue than most practitioners admit. However, from my perspective, we do understand the process really, really well, and among scientists and clinicians working in this field, there isn’t much debate about most of the points that are constantly raised by certain communities on social media. Yes, sure, there are certain issues where we will benefit from more data. But, I will be honest with you: most of the claims I have heard on social media are not well considered, as I have tried to show in this blog post. A lot of what is going on on social media on this topic reminds me of someone saying “OK, this looks like a duck, and it quacks like a duck, and it waddles like a duck, and it swims like a duck, but under these and those circumstances, it may still be a baby swan”.
From my perspective, there are two worthwhile questions that we don’t have conclusive answers to and where the community is right to be interested in more data:
(1.) whether elevated LDL-cholesterol will lead to atherosclerotic plaque in the absence of any triggering event (endothelial dysfunction, inflammation, hypertension, oxidative stress) that initiates damage to the endothelium, and in the absence of any other risk factor. These questions, to me, have not been sufficiently studied to know for sure. At the same time, it’s clear that, without exception, everyone is exposed to some other risk factor(s) in their live, be it air pollution or a temporary acute inflammation due to an infection, so I think this question is academically interesting, but it doesn’t negate that we benefit from having lower apo B. Even a LMHR, for example, assuming LMHR is harmless, could have some health crisis associated with hypertension and inflammation: what would the extremely high LDL-chol levels do to their endothelium then? And once a plaque has been initiated, it tends to grow if LDL-chol remains high.
(2.) whether the exceptional LMHR phenotype is associated with no, or less, atherogenesis than we would otherwise expect. I am still skeptical, but I think it’s a worthwhile question.
Best,
Mario
Mario, overall presentation is excellent in terms of visual impact. I applaud your effort. However, for me, the debunking of claim #2 is not convincing at all; it is based on assumption, NOT evidence; you assume that a relative risk is good enough if we follow those patients for decades and not just a couple of years; you seem to have forgotten that the JUPITER trial lasted less than 2 years; are you okay with that — good enough just for less than 2 years, let’s keep going and hope all is good?
We have an ethical obligation to stop a randomized controlled trial as soon as a a priori-defined clinical benefit has been statistically demonstrated. That’s what the sample size and duration of the study are based on. No ethics committee in the world would permit that about half of the participants (often high ASCVD risk patients) continue to be given placebo pills once the effectiveness to lower the clinical risk has been shown. Therefore, we will never know what the risk reduction actually is at 10 or 20 years. However, my best guess is actually that the relative benefit grows over time, because of the issue of the cumulative lifetime exposure, the gap of which also widens over time.
I personally think this is one of the most poorly represented and most misunderstood topics in this field. As I say in the video, given that the cumulative lifetime exposure to apo B-particles is what matters, and not the current concentration, the fact that we can even detect effects of lipid-lowering therapy so quickly does speak volumes as to the key role that these particles play in ASCVD.
Best,
Mario
Ethical obligation? Based again on what? Assumption? Hope? This is no different like “trust the science”. If we invoke philosophical arguments, I have a couple for you (there are more, but I will just stick with these two):
– Moniz received Nobel prize for lobotomy in 1949. Shall we say that at the time it was unethical to deprive patients of lobotomies, right?
– Insulin shock therapy for psychiatric diseases; it lasted decades. Unethical not to do it, right?
Hope and assumptions violate the falsification theory (Karl Popper) which says that for a theory to be considered scientific, it must be able to be tested and conceivably proven false. Stopping the trial prematurely for ‘ethical purposes’ is highly unethical because it undermines the possibility to be proved false.
Sincerely,
Andrei
I think you misunderstood: if you run a trial, you must define a priori (before you enroll the first participant) what your hypothesis is, what you would consider a clinically relevant differential change between the treatment group(s) and placebo, and what your sample size is and with what rationale. Once a lipid-lowering trial, for example, has shown that the primary endpoint of ASCVD reduction has been reached, even if, in the short-term, the absolute benefit is small, the investigators are ethically obliged to stop the trial. Because continuing to withhold the effective treatment from the placebo group would not be ethical. No ethics committee in the US or Germany, for example, would allow the investigators to continue to withhold effective treatment from people randomized to placebo once the clinical benefit has been proven.
Standards clearly have changed, a lot, in biomedical research in the last 50 years. Many of the studies that were routinely done 50+ years ago would no longer be feasible today.
All I wanted to do is explain why we are not running a trial of a statin or a PCSK9 inhibitor vs. placebo for, say, 20 years (which it sounded like you were sceptical off, no?). And why the small absolute benefits seen in short-term lipid-lowering trials are not to be underestimated, because if several such short-term benefits are stacked by addressing all key risk factors, the cumulative effect over a longer period of time will be very meaningful IMO.
Best,
Mario
I’d like to add an additional consideration: Statins have side-effects that usually take longer than the trials to manifest. Isn’t then one possible outcome that while the statin reduces ASCVD, it could simultaneously cause another problem (e.g. muscular or dementia). All-cause mortality might be a better criterion, but it probably would have to be evaluated over longer periods.
Unfortunately data about statin side-effects is scant and interpretation controversial. Understandably (though perhaps not ethically), pharmaceuticals are not enthusiastic funding long term studies to evaluate concerns that could reduce their business.
The field certainly suffers from biases due to conflicts of interest. My cardiologist suggested I look at the Fourier Trial (Evolocumab and Clinical Outcomes in Patients with Cardiovascular Disease, New Engl J Med, 2017) as “proof” of the benefit of lipid-lowering medication. The reduction of LDL-C achieved is indeed impressive (mean 30mg/dL). My literature search also brought up the following publication: “Restoring mortality data in the FOURIER cardiovascular outcomes trial of evolocumab in patients with cardiovascular disease: a reanalysis based on regulatory data” (BMJ Open, 2021) which posits that outcomes were not properly reported. The authors conclude: “After readjudication, deaths of cardiac origin were numerically higher in the evolocumab group than in the placebo group in the FOURIER trial, suggesting possible cardiac harm.”
I am not in a position of evaluating the claims, but am surprised that (1) the concern appears in a different publication and (2) there is virtually no reaction by the authors of the original publication (aside from an objection in an answer to a journalist). In my scientific field such issues are handled differently: objections are raised, investigated by all parties and independent reviewers, and conclusions published in the same venue.
The discussion here is valuable since it is free from bias.
Yeah, I share that concern about side effects of long-term medication usage. That’s why I feel so strongly about prevention, nutrition, and lifestyle.
Thank you also for sharing that Fourier Trial issue. I had not been aware of that. Reading up on this a little bit, it’s maybe still too early to really know what the correct outcome is, but this type of controversy certainly doesn’t make any of us more eager to take these drugs … You are certainly correct to point out that there simply is too much money at stake here …
Cheers
Mario
Hello,
I have just discovered your site and this post, and wish to add my appreciation for this extremely thorough and accessible presentation on the topic. I am very much that “average Joe” that you reference … a 64 year old male who has been on an intentional change of path in terms of diet and lifestyle for the last couple of years. Most of the changes in my biomarkers are very similar to what you list here (e.g., A1c down from 6.4% to 5.3%, etc., weight from 215 to 185, etc.)
One question I have is about the correspondance between apo(b) and LDLc. My understanding is that one sometimes sees a discordance between the two, with relatively low LDLc but elevated apo(b). This obviously speaks to the importance of checking apo(b). In my case, though, I have what seems to be a more unusual discordance … quite low apo(b), 66 mg/dL (down from 74 a year ago), with somewhat elevated LDLc, 123 mg/dL in the most recent bloodwork (total cholesterol 198, and TG of 97). I am not currently taking statins or any other pharmacological intervention, but am weighing whether that might yet be warranted.
Am I correct that I can, in essence, just focus on the apo(b) number and ignore the LDLc number?
Thanks again,
Mo
Hi Mo,
I hope you understand that I cannot provide medical advice here. I can only repeat what two leading lipidologist say in the paper I have linked below:
– Apo B is a more accurate marker of ASCVD risk than LDL-chol, because apo B is directly proportional to the measure that matters, the number of apo B containing atherogenic lipoproteins, whereas LDL-cholesterol is only correlated with the number of atherogenic lipoproteins
– The lab measurement of apo B is standardized and direct, whereas the most common measurement of LDL-chol is a calculation based on total cholesterol, HDL-cholesterol, and triglycerides. Therefore, the lab measurement of apo B is more accurate than that of LDL-chol.
https://pubmed.ncbi.nlm.nih.gov/37934698/
Best wishes,
Mario
Hi Mario, Very valuable and informative lecture and notes. I am probably still absorbing it and will continue to check back as I think it through. At 70, I can see now have things have changed. I have generally been in the high ranges of LDL (121) but also high HDL (87) and low triglycerides (56). This has not changed but now I am considered to be at higher risk because of the high LDL while previously doctors shrugged it off because I my HDL was high, triglycerides were low, I was fit and BP was normal or low. I have a much better understanding after watching your video. Thank you very much, Cathy
Hi Cathy,
In cases where your doctor tells you that you are ‘higher’ risk or borderline, it may be a good idea to measure both apo B and Lp(a) (fasting blood for both) to help guide treatment decisions.
Best,
Mario
Thanks a lot for the article! It was both interesting and difficult to read:)
1 – After reading, the idea arose – to reduce the consumption of fatty foods. Question -> Did I draw the right conclusion if I am 32 years old, I am thin, there are no chronic diseases, I adhere to the low glycemic index diet.
2 – How much does the fat we eat affect the lipoproteins in our blood?
3 – It is said that vegetable fat is more harmful than animal fat. Is it true?
4 – Do you use Chat Gpt 4 when writing articles? It could be a very good tool to make working with text easier. It can help make the test simpler and easier to read.
Hi Igor,
I will address all of your questions in future posts, and I would have to be very superficial if I were to give you a brief response here. Sorry.
And no, I do not use any AI.
Cheers
Mario
Thank you for your comment. I am just not sure knowing my Lp(a) would change my trajectory as it still seems like the best thing to do is stay fit, eat right, lower stress, sleep well. Even thinking about testing Lp(a) makes me anxious. I have 3 sisters and we are all over 65, Mom lived to 98, Dad to 83. No heart disease in the family. I was mostly mystified when I saw my cholesterol had been much higher in the past and my doc never said anything. Your video really helped explain that. I also have watched this video and I think they make some very good points. While I think they differ on some points, in others they are in agreement with a lot of your videos. Lifestyle changes are best. https://www.youtube.com/watch?v=yr_4RoPhtu4&t=631s
I am curious as to why you suspect that over the long term the LMHRs will have increased risk of CVD. My first question would be: before being shown the results of the LMHR baseline to the Miami Heart baseline, did you expect there to be ZERO increase of plaque in the LMHR group? Based on your article, I’d say you were surprised at that initial finding. So then logically, why is it a vast stretch of the imagination to presume the long term outcome won’t be similar?
Secondly, as you rightly point out, the LMHR group is probably a pretty healthy group in terms of diet-adherence and exercise. At the end of the day, life is about choice and risk, and I’d contend that the substitution of carbs back into the diet or drugs like statins may be MORE risky to health-span vs the absolute risk of a CVD event. Especially if you are like me and carbs present a rapid influx of cravings that lead to over-indulging in processed foods.
And thirdly, from what I understand, HDL is the least well understood piece of the puzzle when it comes to lipidology. So is it a stretch to assume that elevated HDL in the LMHR is precisely the mechanism which may completely negate the risk of elevated LDL? Is there another phenotype out there who has a higher HDL than the LMHR? And if so, what can we learn from them?
Hi Paul,
All good questions. It’s not a stretch to think that LMHR could be an exception. As I’ve tried to explain in the post, the association between LDL-chol and ASCVD is pretty strong and consistent, and based on everything we know at this point about such massively elevated LDL-cholesterol levels, I would expect greater progression of atherosclerotic plaque than in similarly metabolically healthy people with lower LDL-cholesterol levels.
I was a bit surprised by the baseline cross-sectional data, but not hugely, because (a) the data from the presentation leave a lot of questions open, which I hope the paper will address, and (b) this is a very small cross-sectional analysis that has a lot of limitations and potential issues.
All that said, do I think it is possible that the LMHRs will not experience greater than control progression in their atherosclerotic plaque? Yes, absolutely, and I even have a few hypotheses that may explain why that may be. And yes, these do include differences in the composition of the key lipoproteins, both in terms of lipid and – maybe more importantly – their protein cargo. Because while I was talking in the video about apo B, apo (a) and apo A only, these lipoproteins do carry around numerous other proteins, some of which are through to be involved in atherosclerosis. I also hypothesise that LDL-cholesterol will drop in LMHRs in the postprandial period, and much more so than in people eating mixed diets. Partly as a result of that, I think it’s reasonable to hypothesize that the average apo B concentration throughout the day is not all that much elevated in LMHRs compared to controls.
So, yes, lot’s of potential reasons why LMHRs may differ a lot in their ASCVD risk from what we are currently assuming. However, it’s important to keep in mind that all of these are just hypotheses at this point, and given the extensive data we have from all other populations, our best estimate should remain that atherosclerotic plaque will progress at a faster rate in LMHRs than in controls. Also, IF we found LMHRs to not experience a progression of ASCVD, they would be the first population ever to demonstrate an uncoupling of massively elevated LDL-cholesterol from ASCVD progression. This would be a pretty big deal, for LMHRs, but also for this field.
Best,
Mario
Mario please watch Dr. Paul Mason’s short video “hard science on the real cause of heart disease” and provide a comment. He gives compelling evidence against the lipid hypothesis and cites clotting as a risk factor that you did not seem to include in your list of independent risk factors of CVD.
Hi Paul,
I am aware of the claim that blood clotting rather than the retention of apo B-carrying lipoproteins is causal in atherogenesis. It’s an interesting hypotheses, and there are some associative data supporting the idea. For example, atherosclerotic plaque do contain material from platelets and red blood cells, and the lipid model doesn’t provide a good explanation for how these get there. Still, the data supporting the idea that blood clots are the only cause are basically non-existent. If anything, the current data at best support the HYPOTHESIS that microclots in the circulation could play a role in atherogenesis. In my opinion, they are far from definitive, and I also don’t understand why proponents of the blood clotting hypothesis see it as evidence against a role of blood lipids in ASCVD, as we have substantially more and substantially stronger evidence for that.
I am following this field, and will be interested in further evidence, but I see nothing in the existing scientific literature that even remotely makes me feel that app B-carrying lipoproteins are unimportant.
Best,
Mario
Regarding: “ASCVD is a multifactorial disease”.. I’m interested in two other factors that are known to be strongly associated or possibly cause ASCVD but I rarely saw them mentioned in most blogs and YouTubes in the past. The first is the role of Uric Acid. This has received MUCH more attention in the last year after Dr. Perlmutter’s book came out, Drop Acid.
The second is far more rare than elevated UA. Iron overload, mostly caused by genetic Hemochromatosis. It has been documented in many research studies that it causes systemic inflammation including heart and cardiovascular problems. Luckily once you discover you have the condition it is easily controlled.
I’d like to see your thoughts on the science around Uric Acid and ASCVD. Thanks!
Hi James,
Good questions.
Iron overload is clearly an ASCVD risk factor, but that’s a genetic condition, and there are numerous genetic conditions that are treated as such. In the blog post, I covered those risk factors that are broadly applicable to the general population. Even though, among men, one could certainly debate whether iron intake on the high end of normal may carry some risk.
Uric acid is one of many emerging ASCVD risk factors where, I think, we slowly have enough data to consider them, and also to understand their relationship to other ASCVD risk factors (which is important to be able to determine whether a risk factor is independent of others). I may make a separate video about that in the near future, as it has been requested by others as well.
Cheers
Mario
Hi Mario,
Have there been any studies on the following?
Where “heavy” arterial calcification is present is there any proven benefit of taking statins? It would seem like “closing the stable door after the horse has bolted”. From what I can find on the web calcified plaque is considered “stable” and is less of a risk than hot/soft plaque.
The current mantra appears to be the anyone with a high calcium score is immediately put on a high dose statin .
It would seem more appropriate to firstly make all the lifestyle changes possible to address the risk factors covered your excellent video.
The aim would be to halt any further formation of soft plaque. Statins (starting at low dose) might then be required should the relevant lipid markers remain high even after lifestyle changes.
It appears that nothing much can be done about already calcified plaque (although it is a current hot topic that this may be possible with supplementation of Vitamin K2).
If you have covered any of this in your previous videos could you advise the relevant links.
Can you cover the subject of the various states of arterial plaque and their ramifications in a future video?
A very good question relating to a complex issue. Allow me to give you a brief response here only: in general, if someone has had an ASCVD event, the consensus is to always prescribe a statin to lower apo B as much as possible, and also because some of the other non-lipid-related effects of statins may help prevent another event. It’s a little less clear in people with calcified plaque who have not had an ASCVD event, but I’d still say that most preventive cardiologists would use an “all of the above”-kind of approach to try to lower all of the risk factors as much as possible to prevent the development of additional plaque and a worsening of the existing ones. Overall, I’d say that this is supported by the evidence, even though in general statin use in primary prevention (before any ASCVD event) is a little less solidly supported by data than in secondary prevention.
Cheers
Mario
Many thanks for your reply , hopefully , you will make a video on the subject in the future. I have seen that there are a number of surgical procedures available to physically remove arterial plaque although I am not sure if these are used on coronary arteries. The animations look a bit scary and appear to drill out/abrade the artery deposits. One process appears to “zap” calcium deposits using a wire. I can’t remember if they use an electrical current or some sort of shock wave to disintegrate the calcium particles. All a bit worrying, any comments?
G’Day Dr. Kratz
I became aware of your work following your presentation on Dr Ford Brewer’s webcast and immediately related with your discipline and demeanour. A year ago, a coronary CT scan produced an Agatston Score of 880 (total) with 660 on the LAD. My former GP had me on 10 mg Rosuvistatin for the past 15 years but I was not satisfied with his explanation of what the scan results implied for my coronary vascular health. A bit of online research directed me to Dr. Brewer’s Prevmed website and YouTube channel. I believe Prevmed and Dr. Brewer have much to offer in helping prevent ASCVD and was tempted to adopt his his approach to using low dose statins to reduce endothelial inflammation and progression of soft plaque. I was preparing to confront my new GP and her recommendation to increase my Rosuvastatin to 20 mg, but after considering your review, I choose to accept my new GP’s advice. I consider this a prudent choice as I have not experienced any negative side effects of Rosuvastatin. Please keep up the good work.
Thank you.
Thank you Mario for a channel focused on the facts.
I understand that any response will be general in nature and not medical advice for me specifically.
I’m a 55 year old male with a CAC score of 486. I have none of the standard risk factors for ASCVD beyond family history and being male. No hypertension, no smoking, no drinking, no obesity, no diabetes, no high cholesterol, no kidney disease.
Average LDL-C since 2006 is 76, HDL 41, total 131 and triglycerides 72.
My goal is to determine the root cause (maybe it isn’t possible) of my ASCVD so I have the best chance of limiting its progression. My doctor immediately prescribed 40mg of atorvastatin, which I took for several months before stopping it due to muscle issues. That reduced my LDL to 45, and ApoB to 31. I believe those results suggest that my ApoB was not high even before starting the statin. C-Reactive Protein was 0.31 while on the statin. I think that suggests I didn’t have chronic inflammation.
I was overweight and very likely pre-diabetic prior to 2006 for at least 5 years and maybe more. I lost more than 30lbs beginning in 2006 and have maintained at or below that level since then. BMI has been less than 27 since 2006 and less than 26 most of the time. Glucose and A1C have all been within normal ranges since then. I made drastic changes to my sugar and calorie intake starting in 2006 so it seems that my diet would not have been a major contributor to my ASCVD during the past 17 years.
1. What does research say about root causes of ASCVD for individuals with this profile?
2. Is it likely that my CAC score has been high for a long time and I simply did not know it?
3. What does research say about the efficacy of statins for those with high CAC scores and who don’t have high LDL or high ApoB?
My research suggest that a low dose of a high-potency statin or a low-potency statin such as pravastatin could be beneficial to get the non-lipid lowering effects of statins.
Kindly,
Chris
Hi Chris,
Yours is a bit of a frustrating case, I assume. If this was me, these are the things I would look into some more:
– Measure my Lp(a) (“Lp little a”) concentration. Even though your LDL is not elevated, Lp(a) could be high and could put you into a higher-risk category. Lp(a) is thought to be highly genetic, and could explain the familial history piece. While there are currently no medications approved that can meaningfully lower Lp(a), several are in end-stage clinical testing, so this would be relevant to know.
– Another genetic anomaly that could predispose to ASCVD. Something that immediately comes to mind in a man is hemochromatosis (iron storage disease), or Wilson’s Disease (copper storage disease), or something similar.
– A chronic inflammatory condition, such as an autoimmune disease.
I would also keep a really close eye on my blood pressure (regular measurements at home) and blood sugar levels (maybe wear a CGM for a couple of weeks to see if you experience frequent spiking patterns), and determine my HOMA-IR (see my video/blog post about measuring insulin resistance).
It is possible that the ASCVD developed as a consequence of your genetic risk + being overweight, insulin resistant, and pre-diabetic for a few years, but that alone strikes me as unlikely, as it sounds like none of these parameters was ever really bad? No manifest diabetes, no BMI > 30? Do you know your LDL-chol prior to 2006? But still, it is possible that you’ve had coronary plaque and possible calcified plaque for some time without knowing it. It’s impossible to know for sure at this point.
My own view would be to (a) identify the cause(s), and (b) minimize all factors that could promote progression. That would mean lipid-lowering medication, but I would not just tolerate side effects, but try different ones or different doses. I would agree that even at fairly low LDL-chol/apo B, statins may be beneficial, partly by lowering LDL-chol/apo B more (from 76 to 45 mg/dL is meaningful) and partly through its other effects (plaque stabilization, anti-inflammatory actions).
As I know you understand, this is not medical advice, and I strongly suggest you discuss these points with a physician.
Hope this is helpful. Best of luck!
Best wishes,
Mario
I did have my Lp(a) checked and it is < 10 nmol/L so I'm good there. Unfortunately I don't have cholesterol data prior to 2006.
My highest BMI was around 30, but it wasn't sustained there for more than a year or two. As with cholesterol data, I don't have other medical numbers prior to 2006, but my baseline fasting glucose in 2006 was 130. I was at my highest weight at that time so I doubt it was ever much higher than that.
I'm working with a functional medicine doctor presently hoping to identify the cause, but I'm becoming more skeptical that I'll ever really know. My current cardiologist was not interested in that research so that is what led me to a functional medicine doctor. I know functional med doctors have a bad name in traditional medicine circles. The one I'm seeing is a board certified internal med doc and has practiced internal medicine for 20 years so hopefully I'm in good hands. Her interpretation of my blood tests was that I'm not currently insulin resistant. My HOMA-IR from the point-in-time insulin and glucose blood tests is 1.4.
I will look into using a CGM and keep my BP monitored. I don't have any symptoms of the two genetic anomalies you mentioned. I do have very minimal eczema, which could indicate some type of autoimmune or inflammatory condition. Perhaps there are tests other than C reactive protein to identify chronic inflammation?
Thank you so much for providing such a thoughtful response.
Chris
Fantastic essay that ties together many specifics of blood lipids that had always been a bit unclear to me. Very well written in easy to understand language, thank you.
One aspect that has always bothered me is the notion of an independent risk factor as it seems to be a flawed premise. Metabolic syndrome presents itself as a cluster of symptoms (obesity, low HDL, high triglycerides, high A1c, etc.) as an example, as do so many of these related conditions. Thus a single indicator will always have an extremely weak correlation, if any, as it’s rarely present without the others. The overall pattern is what’s important which is why a ketogenic diet has a dramatic impact on Joe Average’s risk profile, while simply putting him on statins will be marginal at best.
Nick and Dave’s work on the LMHR’s would seem to bear this out. Also, why wouldn’t the exact same mechanism they describe also apply to Joe Average on a ketogenic diet? His risk plummeted from 39.4% to 14.5% on the keto diet using the (questionable IMO) scoring method. Where is the evidence that additionally lowering his LDL (perhaps using statins) would improve his risk? It would only do so if one believes the components are significantly important in isolation which doesn’t seem to be the case.
Much of this thinking is driven by drug development IMO, which uses proxy endpoints (such as lowered LDL) rather than measuring fewer dead people. In theory it makes sense as it’s easier to measure lipids than enroll enough participants over a long enough time to accumulate enough bodies. The problem though, is that proxy endpoints can be easily abused either through incompetence or outright malice. Developing blockbuster drugs hyper-focused on single parameters is missing the forest among the trees at best.
Hi Jeff,
Thank you for your thoughtful comment.
The concept of ‘independent’ risk factor is based largely on whether a risk biomarker is STATISTICALLY independently associated with ASCVD, i.e., after adjustment for all of the other known risk factors. That said, I agree with you that it’s hard to make the case that any biomarker by itself is totally pathophysiologically independent of all other factors.
That said, I think the risk estimation using the existing data is the best we can currently do. If we got new data that there is actual effect modification, such that massively elevated LDL-cholesterol is no longer associated with ASCVD in some people, then we could update the risk prediction models. Currently, no such effect modifier exists, and I think telling people to ignore their sky-high LDL-chol based on what currently amounts to a hypothesis is irresponsible. We need to consider that in all of lipid research, there is not a single instance where we have found a disconnect between LDL-chol and ASCVD (i.e., a true effect modifier).
Cheers
Mario
Hi, Thanks again for your very clear explanations.
In addition to LDL, apoB, and lipoprotein(a) levels, could an ultrasound of the arteries also provide information or confirm the level of risk, especially in a dynamic perspective (e.g., testing every 3-5 years) by quantifying the presence and amount of deposits?
You mentioned risk factors, highlighting meat consumption: What causes meat to raise LDL/apoB levels in the blood? Does this differ depending on the type and quality of meat consumed (for example, lean meats like chicken breast, or meat from extensive farming with grass feeding vs. soy/corn feeding)?
Thanks in advance for your response,
Jean-Jacques
Hi Jean-Jacques,
It’s mostly fatty meat, because that’s very rich in long-chain saturated fatty acids, which raise LDL-cholesterol when compared to unsaturated fatty acids. It’s not a huge effect, and impacts on ASCVD risks are a bit uncertain, as I will discuss in a future video.
Cheers
Mario
Hi Mario, thanks for your article very useful.
I understand that the various factors you mentioned have to be taken into account with regard to the overall risk of ASCVD.
If I’m just concerned by a slightly elevated level of apoB (93mg/dl), what can I put in place in a non-drug way to lower this (type of food, supplements, etc.)?
Little bit of context, 33 years old, good shape, good die, exercises often. (TG:54mg/dl, HDL: 55mg/dl, LDL: 126mg/dl)
Thanks
Hi,
and in which metric is 93mg/dl slightly elevated (recommendation of whom)?
Hi Michael, 93mg/dl for ApoB. From what i have read, the link between number of particles LDL and ASCVD is pretty clear. The more particles we have, over a longer period of time, exposes us to greater risk of ASCVD.
Why i consider its slighty elevated, i rely on the study of Framingham (Contois, John H et al 2009), so i’m in the 45 percentile of the population.
i would like to be in the 20 percentile of the population (78mg/dl ApoB) or less.
What is glaringly missing is there is still no method proposed which even remotely explains HOW ( ie a mechanism ) by which high LDL or ApoB is indeed atherogenic .. yet there are very plausible mechanisms for how atherosclerosis occurs with LDL not being a causative ( eg Dr Kendrick’s clot hypothesis or indeed just the degradation of the glycocalyx by blood sugar /insulin ) .. without any mechanism, high LDL /ApoB present in CVD could well be an indicative bio marker .. the fireman rather than the arsonist at the fire scene .
Hallo Mario,
in the paper:
Physiological Bases for the Superiority of Apolipoprotein B Over Low-Density Lipoprotein Cholesterol and Non–High-Density Lipoprotein Cholesterol as a Marker of Cardiovascular Risk
you can find following statements:
“What physiological process is served by the cho-
lesterol that was exported from the liver within VLDL
particles and returned as VLDL or LDL particles? Is this
just a futile cycle of cholesterol or could it be that cho-
lesterol and CE are essential elements to form VLDL
particles? Alternatively, could it be that at least in cer-
tain circumstances VLDL particles also export excess
cholesterol from the liver? These are questions without
answers at present.”
Is this really true, that in 2024 we do not know why there is cholesterol circulating in Lipoproteins in our bloodstream?
In the same paper it is mentioned, that the peripheral cells do NOT use or uptake this cholesterol from Lipoproteins!
Perplexing.
Greetings Michael
Great question. All cells can make their own cholesterol, so uptake from the circulation is indeed not necessary. However, almost all tissues express the LDL-receptor molecule that binds to apo B and can remove LDL particles from the circulation, so that suggests that cells do to some degree get some of their cholesterol from circulating apo B-carrying lipoproteins. Still, it has been estimated that about 70% or so of all LDL-particles are removed from the blood by the liver, so that part would seem like a futile cycle given that LDL was originally secreted (as VLDL) by the liver cells.
My best guess is that the liver needs to use cholesterol as packaging material to assemble VLDL particles. Exporting TGs in VLDL particles is only one of two ways through which the liver can get rid of triglycerides. I speak about this here:
https://nourishedbyscience.com/insulin-resistance-syndrome/
Cheers
Mario
I have some issues with this video. I will just bullet point them, plus I am too lazy to look up many of the references.
– You seem to have ignored methods of lowering cholesterol and/or LDL that didn’t work out. Niacin lowers LDL and triglycerides and raises HDL. Bile acid sequestrants were tried and failed. CETP inhibitors make the bloodwork look great but also failed to show any benefit.
– I haven’t seen any studies on PCSK9 inhibitors alone that show benefit. There are some where they were given along with statins, and some where they were given alone and lowered the surrogate markers.
– Are there any data on the general population showing any benefit of lowering total cholesterol/LDL/ApoB. There are so many people taking statins that there should be some reduction in CVD but AFAIK it is only increasing.
– You mentioned LP(a) but I don’t recall you saying that it is usually counted along with LDL. Maybe I missed it.
You didn’t mention any possible pleiotropic effect of statins. One of the first studies to show any benefit to statins (see lazy comment above) concluded “We lowered cholestreol and got a [barely signicant] reduction in heart attacks.” No, they gave a statin and got that result.
Elliot,
Respectfully, I have poured over 1,000 of papers in this field over the past 30+ years, and citing 23 key ones here. You counter this by making statements without providing any evidence. Not sure how you expect me to respond. It’s very easy to find a few outliers in the vast sea of literature on this topic. Because of course, there are some exceptions, but do you really think it is enough to fundamentally question what is presented here?
And, more specifically, if you were me, do you think these inconsistent data points are sufficient to tell all of my readers to ignore their serum lipids?
Best,
Mario
Thanks so much, Mario. I did a self-experiment to see if I could reduce my apoB (currently at 103 mg/dL) through dietary change. I cut out eggs, limited saturated fat and other sources of cholesterol, and ate large amounts of legumes for two months. It had no change on my apoB and increased my triglycerides from 79 mg/dL to 101 mg/dL. Given that feasible dietary change seems to have no effect on my apoB and that my Lp(a) is 126 nmol/L, would you recommend talking to my doctor about drug therapy? I am not aware of family history of CVD. My doctor said everything was fine and not to worry. But based on your recommendations I think it might be worth lowering my apoB through drug therapy.
Hi Ty,
I know you understand that I cannot provide medical advice here, so everything I suggest below is some food for thought, to be discussed further with a physician.
Yours is a borderline case. Lp(a) of 126 nmol/L is equivalent to 50 mg/dL, which indicates that your Lp(a)-related risk is elevated. That means that your other risk factors should ideally be more strongly controlled. Apo B of 103 mg/dL, by itself, if someone doesn’t have any other ASCVD risk factors, would probably not be treated by most doctors, but with the Lp(a) being elevated, some cardiologists would probably consider a low-dose lipid-lowering medication, given that dietary change wasn’t successful at lowering apo B. One thing to understand here is that the absolute ASCVD risk is likely low in the near future (such as the next 10 years), so the main reason one would consider lowering apo B at this stage is to lower the risk of an ASCVD event 20 or even 30 years from now, to lower the cumulative lifetime exposure. Right, now is the time to think about what your cumulative lifetime apo B area-under-the curve will be by the time you are 60, 70, or 80. And I think that relatively small near-term benefit of initiating drug therapy would only make sense if the lipid-lowering medication was well tolerated (no muscle-related side-effects, no increase in insulin resistance and glucose intolerance). That’s how I would think about this, and would probably discuss these points with a preventive cardiologist to carefully weigh the benefits vs. the potential risks.
Best wishes,
Mario
In terms of blood tests what of the PLAC test?
Chris,
You can read up on this here:
https://www.aafp.org/pubs/afp/issues/2020/0101/p44.html
It’s probably not covered by insurance to have this measured, I would assume, and I personally would not pay for it.
Cheers
Mario
Hi Mario,
Loved the article. I’m currently not convinced on LDL alone and have been wading through the research. I would love to hear your thoughts on this opinoin paper. It covers more than just statins on LCD:
https://journals.lww.com/co-endocrinology/fulltext/2022/10000/statin_therapy_is_not_warranted_for_a_person_with.14.aspx
by Diamond, David M.a; Bikman, Benjamin T.b; Mason, Paulc
It appears to agree on Lp(a), though has a different perspective on LDL on LCD. My take home is to get Lp(a) tested (as you recommend) and CAC.
Cheers
Matt
Hi Matt,
I agree with the authors that LDL-cholesterol/apo B is only one risk factor among many, and that we need to consider the overall risk by considering all risk factors. In contrast to the authors, however, it is my position that any elevation in LDL-cholesterol/apo B, including from low-carb diets, should be seen as a risk factor, until proven otherwise. As I explain in the video, it is certainly possible to lower the risk of ASCVD on a low-carb diet even if that diet raises LDL-cholesterol and apoptosis B, for example if weight is lost and glucose tolerance, insulin resistance, and blood pressure are improved. I personally would still not be comfortable with very high LDL-cholesterol chronically. Specifically, I have pretty high LDL-cholesterol if I follow a ketogenic diet, and I would not follow that diet over the long term for this reason alone. For me, my LDL-cholesterol comes down about 80 mg/dL if I include just enough carbs in my diet to move out of ketosis.
All of that said, it is possible that future research teaches us something new about the relationship between LDL-cholesterol/apo B and ASCVD, for example in the context of low-carb diets. I follow the LMHR research with a lot of interest, but at this time, we don’t have anything conclusive yet, just a hypothesis and some early stage pilot data. Not enough to ignore these as risk factors at this point.
Cheers
Mario