If you spend any time reading about cardiovascular health online, you'll quickly realize it feels less like an objective medical discussion and more like an ideological war. On one side, mainstream lipidology asserts unequivocally that apolipoprotein B-containing particles, predominantly LDL, are the direct causal agents of atherosclerosis. So, lower the particles, they argue, and you stop the disease. On the other side, though, an increasingly vocal group of researchers, clinicians, and metabolic health advocates argue that this hyper-focus on particles misses the real picture. They point to people with sky-high LDL who reach old age with a coronary calcium score of zero, or individuals with rock-bottom cholesterol who suffer fatal heart attacks. They argue that heart disease is fundamentally an acquired immunometabolic disorder driven by endothelial dysfunction, chronic inflammation, and insulin resistance. So, what is it? Is heart disease caused by the particles floating in your blood or by the state of your artery wall? In this video, we aren't playing internet tribalism or picking sides.
We're going to steel man both arguments, the best data, the best mechanisms, and the unanswered questions for each. And so, this video is looking to answer one question: how does the process of atherosclerosis actually start. That's what we're looking at today. But before we can do that, we need to know a little bit what the artery actually looks like. So let's look at the anatomy. Most people probably picture an artery like a kitchen drain where the grease slowly coats the inside of the pipe until it clogs, but that picture is wrong in an important way.
An artery is a living organ with three layers. The outer adventitia carries nerves and tiny vessels called the vasovasorum that feed the outer wall. The middle layer, the media, is smooth muscle and elastic sheets that constrict and relax. And the innermost layer is the intima. A single sheet of endothelial cells facing the blood with a layer of connective tissue underneath. In human coronary arteries, that layer under the endothelium usually isn't an empty gap. Many adults have what pathologists call diffuse intimal thickening, a cushion of smooth muscle cells and extracellular matrix, including molecules called proteoglycans, that exist before any fatty lesion appears. And that's where atherosclerosis grows, inside the wall, beneath the lining, not stuck on the surface like sludge. The lumen becomes involved later, either by having that lumen narrowed by subendothelial growth or through an erosion or rupture that triggers a clot to form. But all this raises an obvious question how does anything get from the blood into that layer? The first thing a particle meets is a coating on the endothelium called the glycocalyx. Under a standard microscope, the inner surface of an artery looks smooth. However, with specialized electron microscopy, you see a dense mesh of sugar-protein chains projecting into the bloodstream, the endothelial glycocalyx. And the layer does two jobs. First, it regulates physical access.
In fact, a 2025 study by Kang and colleagues using high-pressure freezing electron microscopy showed that the dense inner glycocalyx layer has a partition coefficient of zero for LDL. So meaning its physical mesh work can completely exclude passive LDL leak under baseline conditions. Understand passive is the key word there.
Second, it senses blood flow. The sheer stress of blood moving across the surface, and then it signals the endothelium to make nitric oxide, supporting normal vessel tone and dampening inflammatory signals. But here's the nuance both sides need to hear. The glycocalyx can be modified, but what hasn't been shown is that an intact glycocalyx makes a human artery immune to LDL entry, or how much glycocalyx alteration contributes to human coronary disease. They're still an open question and not a settled fact for either camp. Now, this raises an objection that sounds really convincing on the surface. If the inner glycocalyx has a passive partition coefficient of zero, how can LDL possibly get through? Well, let's talk about that.
The important distinction is between passive leak and active transport. Endothelial cells move LDL through the cell at baseline via receptor-mediated transcytosis, rather than requiring a tear between cells or passive diffusion. Receptors including SRB1 and ALK1 help in LDL transport, and as Kang et al. demonstrated, while LDL temporarily binds the sugar chains in the outer glycocalyx, the glycocalyx itself does not retain LDL. Instead, surface receptors internalize those bound particles and shuttle them straight across into the subendothelial space.
Experiments in mice show that removing endothelial sRB1 reduces LDL delivery into the wall and reduces atherosclerosis. But that does not settle every question about the glycocalyx, right? Demonstrating transport through an intact endothelial cell in high-resolution animal models or tissue culture is not the same as proving how free LDL crosses a completely preserved glycocalyx in a living human coronary artery. And the glycocalyx regulates access to the cell surface and its interaction with LDL remains an active area of research for sure. However, after reading this, the defensible conclusion is that a visible wound doesn't seem to be necessary for LDL to enter, right? So, the bigger claim that a healthy glycocalyx makes an artery completely immune to LDL entry and retention, that simply has not been demonstrated in living, breathing humans. And that brings us to one of the most hotly debated points in the entire debate.
Does something have to damage the artery wall before LDL can get in? So the experimental evidence seems to say probably not. A cool study by Armstrong and colleagues showed that LDL crosses a structurally intact endothelium in cultured cells. However, the particles that enter don't necessarily stay. They can kind of move in and out and they can stay in the tissue or not. And the problem starts when particles bind to the matrix and stay a while. So, how much accumulates depends on several factors, though, right? How many particles are entering, how strongly the matrix holds them, how long they stay, and how effectively they're removed.
That's the framework for everything that follows. Entry is routine physiology. It's going to happen. Retention, though, this is where disease begins. And notice this framework doesn't pick a side, right? Camp A emphasizes how many particles enter? Camp B emphasizes how sticky the wall is, but both are variables in the same equation. But before we hear each side's best take, let's see what the pathology shows directly, whatever camp you're in.
So, number one. Plaque develops inside the intima underneath an endothelial lining that can remain largely intact, not as a scab on a wound necessarily. Two. Early human coronary lesions show extracellular lipid deposits alongside matrix proteoglycans, meaning, hey, we're seeing that fat in that matrix. And then followed later by macrophages moving in. And so that's how we're seeing those on cross-section. We're finding different stages. That's what they see. And three, a mature plaque is a mix of retained lipid, immune cells, smooth muscle cells, remodel matrix, dead cell debris, and often calcium.
And now, to be clear, these findings are across specimens at different stages of the disease. And it's not like we're watching it unfold in one artery in real time. And so the argument isn't really what a plaque contains, it's about what sets this process in motion and what matters most for keeping it going. Does an excess of particles overwhelm an otherwise normal artery wall? Or does an inflamed, metabolically stressed wall make retention far more likely at ordinary particle levels? Let's hear the best case for each. So, first, let's steel man Camp A, the ApoB hypothesis. This is the consensus model currently in cardiology, and its evidence base is large.
So the premise is that every atherogenic particle, like LDL, IDL, VLDL, and Lp(a), carries one ApoB molecule. I think there's no debate there. So the ApoB level reflects how many potentially atherogenic particles are in circulation, not just how much cholesterol they carry. Current U.S. guidelines note that when LDL, cholesterol, and ApoB disagree, ApoB tends to predict risk more accurately. So, Camp A points to three independent lines of human evidence. And I've already talked about this in previous videos, about the case for it, but we'll review here real quickly. First, large population studies showing risk rises steadily with higher ApoB containing lipoprotein levels.
The second piece of evidence is genetics, and people refer to oftentimes a landmark 2006 study where people carrying a PCSK9 variant that lowered their LDL cholesterol by about 40 milligrams per deciliter from birth had an 88% lower incidence of coronary heart disease. And a milder variant lowering LDL by only 20 milligrams per deciliter was associated with about 50% less disease. Across genetic studies, lifelong lower exposure produces roughly two to three times more risk reduction per unit of LDL than drug trials starting later in life.
And third are the randomized trials across statins, ezetimibe, and PCSK9 inhibitors. The benefit tracks how much and for how long the atherogenic blood proteins were lowered. That's the general idea. And one caveat, though, on genetics, Mendelian randomization is often called nature's randomized controlled trial, but that's just an analogy, right? It rests on assumptions and isn't a clinical trial. Even so, to camp A, three different methods pointing to the same way is what causation looks like. ApoB isn't a bystander, it's the cause and the dose and duration of exposure matter.
And the crown jewel of Camp A's mechanistic argument for this, for what starts it, is a landmark study published in Nature in 2002 by Jan Bourne's laboratory. The researchers wanted to answer a basic question. Is LDL actively involved in starting plaque, or does it just get caught up after something damages the wall? To find out, they engineer mice to make human LDL. In one group, the LDL was normal. In the other group, they change the cluster of positively charged amino acids on ApoB that normally grabs the negatively charged proteoglycans in the artery wall. That left the particle with only a weak grip. So essentially, they kind of eliminated how strongly it's going to grip on. Then they put both groups on the same cholesterol-rich diet, which pushed their blood cholesterol levels to pretty high levels, pretty extreme.
In the results, the controlled mice developed extensive atherosclerosis. The mice with the weakened latch developed dramatically less, even at comparable cholesterol level. And so, to Camp A, this was the smoking gun. Even with sky-high cholesterol, plaque formation slows dramatically when the particle can't anchor to the wall. So, binding isn't a bystander event, it's the early causal step. But there's a critical follow-up from 2007 that both camps want to talk about. The same group followed these mice for up to 40 weeks, and something surprising happened. Plaques started later in the weak-latch mice, but they eventually caught up and reached the same amount of atherosclerosis as the controls.
How? Well, it seems that once the first lesions form, macrophages move into the wall and release a protein called lipoprotein lipase. Think of it as a molecular double-sided tape. One side sticks to the wall's proteoglycans, the other side grabs the LDL particle. So, the particle was actually retained by a completely different mechanism. The researchers showed this directly. In healthy aortas, normal LDL stuck better than the weak-latched LDL. And this gives us an important lesson for both sides, though, right? In these mice, direct ApoB binding appeared to start the whole process.
Once a lesion existed, though, the wall itself changes and it finds new ways to trap particles, and retention speeds up. However, it also showed that even with the genetic modification, some ApoB particles still got retained despite it being way, way less. So, for camp A, retention is still the central step from start to finish. For camp B, this is evidence that the wall is the main driver. Once the disease begins, the artery's own biology actively drives more trapping. However, what neither camp can claim from the study is that changes in the wall build plaque without particles, right? So, particles need to be there. The mice had extremely high cholesterol the entire time. However, the fairest reading is that established lesions can retain particles through additional mechanisms.
So weakening one binding route delays disease without necessarily preventing it. Another popular claim is that small, dense LDL causes plaque, while large, buoyant LDL is harmless. And particle size definitely affects biology, but that conclusion goes a little bit further than the data that I've seen. Both small and large LDL particles are small enough to enter the artery wall. That happens. In fact, except for chylomicrons and some of the VLDLs, all the other particles can enter. Just because a particle is quote unquote large, it does not lose the ability to interact with arterial proteoglycans simply because it carries more lipid, right? That's not how it works. The more useful question is whether size predicts disease after particle number is taken into account.
And that also is not as clear as you'd think it would be. In a 2025 UK biobank analysis of about 207,000 people, particle count predicted coronary disease, while particle diameter and subclasses did not add an association after adjustment for count. Lp(a), however, did add risk. So, what they're saying is the count, the actual particle number was the most important, the sizes didn't matter. But in another 2025 study using the same biobank, but a different calculation for small dense LDL, they did find that small dense LDL was an even better predictor for cardiovascular disease than ApoB when the results were discordant.
But we know that small, dense LDLs certainly have properties that favor retention and are a quote more risky LDL particle. That's how it works. But greater relative risk per particle does not necessarily make a large buoyant LDL harmless, right? Based on the current literature, it seems that they're definitely less risky, but we can't say in that they're harmless. Now, this raises another fundamental question. If ApoB particles cause atherosclerosis, why doesn't plaque develop everywhere the blood goes?
Why are veins almost completely immune? And why are certain arteries remarkably resistant? Let's start with veins. In their normal position, veins carry blood at low pressure, and plaque develops almost only in arteries. However, watch what happens when a surgeon takes a saphenous vein from a leg and sews it into a coronary circulation as a bypass graft. Suddenly, that vein faces arterial pressure. Abnormal flow, and the injury of being harvested and handled, right? What follows is fairly predictable. In the first month, the biggest risk is clotting. Over the first year, the inner wall thickens.
After that, the vein can develop accelerated atherosclerosis. In one autopsy comparison, a majority of vein grafts in the heart showed atherosclerosis after about four and a half years. The blood didn't change, right? The environment did. Veins aren't immune to atherosclerosis. Their normal condition is just used to protect them. That same autopsy result points to the other half of the puzzle, the internal thoracic artery. In the same study, a much smaller percentage of internal thoracic grafts developed atherosclerosis. The internal thoracic artery runs at full arterial pressure and carries the same blood as the coronary, yet rarely develops atherosclerosis, even in people with severe coronary disease.
Why is that? Well, its lining makes unusually large amounts of nitric oxide, and its smooth muscle cells are slower to migrate and multiply. Even so, researchers still call it largely unexplained. No single feature accounts for all of it. Then there's location within the artery. Plaque collects at branch points and bends. There, blood flow becomes disturbed and swirls back and forth instead of moving steadily forward. Steady, one-way flow keeps the lining in a protective, anti-inflammatory state. Researchers describe these flow patterns less as the cause of plaque and more as something that kind of primes them for atherosclerosis.
They condition the wall to respond to bodywide factors like potentially high cholesterol. Now, each camp reads this data differently. Camp A says these local differences fit their retention model. A body-wide exposure can still cause local disease because tissues differ on how susceptible they are. Where Camp B says this is exactly their point. The state of the wall decides where the disease shows up. Both are describing the same observation. Now let's give Camp B the same treatment and build the strongest version of the metabolic and inflammatory model.
In this view, ApoB particles are the raw material that plaque is made from, but they aren't the only determinant of how bad a disease gets or how fast it develops. Instead, it is believed that inflammation, insulin resistance, and endothelial dysfunction play a bigger role in determining whether the artery wall becomes a place where particles take hold. A landmark piece of evidence for this view is the Kantos trial, published in 2017. Researchers enrolled more than 10,000 people who had already had a heart attack and who still had elevated HSCRP, a blood marker of inflammation. Participants were randomized to either placebo or Canakinumab, an antibody that blocks a single inflammatory signal called interleukin-1-beta. At the 150 milligram dose, Canakinumab lowered the combined rate of heart attack, stroke, and cardiovascular death by about 15%, and it did so without changing LDL cholesterol or ApoB.
Since then, researchers have tested colchicine, an older anti-inflammatory drug, as well. In the Colcott trial, which enrolled patients shortly after the heart attack, Colchicine reduced the primary cardiovascular endpoint by 23%. In the Lodoco2 trial, which enrolled patients with stable coronary disease, it reduced the end point by 31%. But we also have the CLEAR Synergy trial, which failed to show any improvement with Colchicine versus placebo. And Camp B also points to metabolism very understandably.
People with insulin resistance and diabetes carry higher cardiovascular risk than their standard LDL cholesterol would suggest. There's an interesting twist, though. Insulin resistance itself raises the number of ApoB particles and slows the rate at which they're cleared, even when LDL cholesterol looks normal. That means part of what we call metabolic risk may actually be a hidden particle problem potentially. So it's hard to separate the two, but Camp B would respond that the metabolic effects on the artery wall matter much more on top of that. Put together, Camp B's argument goes like this: If blocking an inflammatory pathway can reduce cardiovascular events without lowering cholesterol, then cholesterol cannot be the whole story. It's just as important to be clear about what these trials don't show. Every one of them enrolled people who had already had coronary disease.
So they tell us about preventing future events, not about what started the first lesion in the first place. And these papers didn't look at insulin resistance per se, so let's talk about insulin resistance. Does it make the artery wall more vulnerable or does it raise the risk by changing the particles in the blood? I think three studies here will help us look at this question. First, let's start with an experiment that changed the lining of the artery. Researchers took mice already prone to atherosclerosis and removed the insulin receptors specifically from their endothelial cells. What this is doing is simulating someone who has issues with insulin signaling, which could represent some aspect of insulin resistance.
And their measured blood lipids, blood pressure, and glucose tolerance did not differ from the control mice, but their arteries developed more than twice as much plaque. And that's strong evidence for the metabolic argument. Changing insulin signaling in some way, shape, or form can alter the vessel lining and can accelerate disease without a corresponding rise in measured blood lipids. And obviously, there's lots of limitations to this, one specifically being that they had really high cholesterol in both groups and they both had a genetic predisposition to atherosclerosis. However, this can be used as evidence that the wall changes can affect the disease. Then comes the human evidence. The IRIS trial enrolled nearly 4,000 people who had insulin resistance but not diabetes after a stroke or transient ischemic attack, or TIA. They received either placebo or Pioglitazone, a drug that improves insulin sensitivity.
Over a median of 4.8 years, stroke or heart attack occurred in 11.8% in the placebo group and 9% of the treatment group. So a treatment that improves insulin sensitivity reduced important vascular events even before diabetes developed. But Pioglitazone also affects lipid metabolism and inflammation, so the trial cannot isolate which effect produced the benefit. And this was a prevention after vascular event trial, not a test of what started plaque in the first place.
Now, here comes a common counterpoint that people talk about, and it's particle-based in nature. And they say, hey, in human studies that directly measure insulin sensitivity, insulin-resistant participants without diabetes still will have more LDL particles, even though their LDL cholesterol appears similar to that of insulin-sensitive participants. And that means that a normal looking LDL cholesterol result can actually hide a higher particle burden. We know that as discordance, right?
It's a big thing. And so they'll say risk beyond LDL isn't automatically risk beyond ApoB, meaning that ApoB tells more of the real story. And these patients who are insulin-sensitive, is it actually something with their metabolism or is it the particle number causing the issue? And now, this does not show that particles explain all the risk of insulin resistance. There's a very clear risk there, but it shows why the two explanations are difficult to pull apart and why both sides of the argument have legitimate points. And Camp B also has one more argument, and it isn't about how many particles get into the artery wall, it's about how tightly the wall holds onto them once they're inside.
Just beneath the artery lining is a mesh called the extracellular matrix made by the artery's own cells. Think of it as almost like a net. So, strands in that net carry a negative charge. And the LDL particles carry a positive charge on one location of them, and so theoretically the two can naturally stick together. However, the net isn't fixed in lab experiments. Some metabolic signals, like insulin resistance, made it stickier, though not everyone tested did. So, in theory, a stressed artery wall could catch more particles even when the amount in the bloodstream stays the same.
And this raises the real question, does a stressed wall make the net stickier? Or does the net only become sticky after something goes wrong? And two findings can potentially help answer that. First, when researchers examined human coronary arteries on autopsy, the earliest stage they could see of the disease was that extracellular matrix with lipids collecting in it. Immune cells came later, which suggests they show up in response to the trapped lipids. The immune cells don't appear to be what starts it. Second, in one experiment LDL was injected into non-diabetic and diabetic mice, and both groups show that they did hold on to some plaque in specifically plaque-prone spots of the arteries. However, the diabetic mice held on to what eight times more, indicating just what a real risk factor's diabetes can be.
And so here's the fair conclusion that I could come up with. A stress wall seems to turn up the volume on retention, but it doesn't appear to be the on/off switch, meaning it's the only thing that matters. And a healthy artery may catch fewer particles, but it appears it can still catch them even before signs of damage or injury. Now let's tackle an argument you may hear from side B about LDL, and it's like this. Native LDL cannot form foam cells because ordinary LDL receptors shut down when the cell has enough cholesterol. And this argument isn't made up. It's based on a landmark. Nobel Prize winning science by Michael Brown and Joseph Goldstein, and they prove that when a macrophage in a petri dish takes in native LDL through the classical LDL receptor.
Rising cholesterol inside the cell triggers a biological thermostat that shuts the receptor off to prevent overload. Their science in the dish was 100% correct, but extrapolating that single petri dish experiment to an intact human artery is where the argument starts to break down. And the classical LDL receptor is only one way cholesterol enters an immune cell. Think of it as a front door with this smart lock. Once the cell has enough cholesterol, the lock is engaged. The problem is the macrophage has other ways in. First is fluid phase macropinocytosis. Say that three times fast.
Macrophages continuously drink pockets of surrounding fluid when ApoB particles are stalled and concentrated in the narrow space. Beneath the endothelium, macrophages gulp down that fluid and the native LDL suspend in it, completely bypassing the front door. Second is particle aggregation. Once trapped in the matrix, LDL particles can clump together, not necessarily oxidize, they kind of hang out. Macrophages recognize those clumps and engulf them through phagocytic pathways that don't have an on-off switch. So, oxidation is not required. Physical clumping is enough. And Brown and Goldstein prove that the front door has a lock. They never proved that an immune cell in a trapped, concentrated environment cannot drown in cholesterol through back doors and windows, right? But to be fair, the other findings were also experimental in nature, and we do not know how this plays out in a real.
Living human being. But if one mechanism is brought up as an argument, I thought it was fair to bring up another mechanism to counter it. And so when I was digging through this research, spending hours learning about this dense vascular biology, I had this moment where I felt exactly like a Mugatu in Zoolander, right? Where he looks around the room and says, I feel like I'm taking crazy pills because when you step back and look at the actual data without an ideological axe to grind, you realize both sides are screaming at each other without listening to a single word the other is saying. It seems like they're arguing over two halves of the exact same biological process. So here's an analogy that kind of helped make this stick for me. So think of atherosclerosis as a seed and soil process. So ApoB particles, they are the seeds.
So it is the non-redundant, necessary raw material. Every particle carries that positive amino acid latch capable of binding to the artery wall, and it carries a cholesterol payload that feeds the lesion. Without the seed, literally nothing can grow, no matter how fertile the ground is. The artery wall and your metabolic health are the soil. High blood pressure, smoking, insulin resistance, and systemic inflammation determine whether that soil is barren rock or fertile ground. And you can only have a few seeds, but if your soil is intensely fertile, plaque can take root. Conversely, you can have a ton of seeds, but if your soil is relatively infertile and resistant, plaque development will be significantly slower. And it sounds so simple when I put it that way, right? So definitive.
But here's the thing that really struck with me after reading all this literature. Anyone who speaks with absolute dogmatic certainty about this topic is being way too brazen. The vast, vast majority of the science we have is mechanistic, it comes from isolated cells in petri dishes. Genetically engineered mice, arterial cross-sections, or autopsies of diseased patients. And if anything's interventional, maybe it's looking at the plaque buildup, but it's definitely not looking at the actual coronary arteries and how the disease starts. We do not have a magic microscope that lets us watch individual ApoB particles interact with the glycocalyx inside living, beating, human coronary arteries in real time. We're dealing with probabilities. Tissue models and population statistics. And when people on the internet talk as if they have this 100% figure out, you have to ask yourself where that confidence is actually coming from.
I always like to point out that I'm not an expert. I'm certainly not an expert in vascular biology. I know, I'm just one beggar to another beggar telling me where to find food and sharing what I found in my research. But the truth is, both camps have legitimate points that the other side stubbornly refuses to acknowledge. Let's start with traditional cardiology camp. In daily clinical practice, there's a tendency to downplay the importance of metabolic health, glycocalyx integrity, and endothelial dysfunction. Now, to be completely fair, those factors aren't ignored. They are written in almost every single major guideline, but practically speaking, the clinical focus often defaults to almost entirely lipid-lowering pharmacology. If you have rampant insulin resistance, severe hypertension, and systemic inflammation, simply lowering the particle count doesn't mean you'll magically fix your problem.
Protecting the vessel wall itself matters enormously. However, the confidence for which the countercultural narrative also speaks is just as puzzling to me. There's this idea floating around the low-carbon metabolic health space that if your insulin is low, your triglycerides are down, your HDL is high, you're essentially bulletproof, right? That you will have an intact glycocalyx and a healthy endothelium, and that will completely prevent ApoB from getting trapped. And I have to be honest with you, based on everything that I looked at here, that does not seem to be the case. Obviously, the data is not robust and it's not amazing, but what I looked at here, it seems like ApoB particles are going to enter your subendothelial space, whether you like it or not. Transcytosis is a continuous 24-7 physiologic reality in every living human being.
It's not a question of if particles enter or encounter the matrix, it's a question of how many and how fast. But this is also where the metabolic and low-carb community actually raises a fascinating point that deserves real attention. Can pristine metabolic health, an intact glycocalyx, and low inflammation significantly slow down the retention and aggregation of particles? I think the answer is almost certainly yes. But the real question is, and the question our entire medical community needs to answer is: by how much? Does excellent metabolic health slow down the process by two years, 20 years, a lifetime? We genuinely do not know the answer to that question yet. And the concern from the traditional lipidologist is that atherosclerosis is an area under the curve disease, meaning if you flood your arteries with high numbers of ApoB particles for 30, 40, or 50 years, relying on metabolic health to make you bulletproof is a massive, unproven gamble. But in the end, heart disease isn't an ideological war.
It's a biological collision between a particle, an artery wall, and time. How important each component is depends on your interpretation of literature. But if anyone tries to sell you a one-size-fits-all blanket answer that claims that one side's 100% right, the other side's 100% wrong, you should probably start wondering if you're getting the whole story.
