Why does your body make cholesterol if it kills us?

Exploring what I think is going on with the #1 killer of Americans, statins and LDL

Jun 17, 2026

Hi there, and welcome to The Next - my take on health, wellness, and company building.

I previously started leading bone broth brand Kettle & Fire, learned more about our how our food and healthcare systems incentivize illness, and wanted to make a dent on this problem. I’m now working on Truemed, which allows qualified individuals to use HSA/FSA funds on lifestyle interventions that can treat, reverse, or prevent chronic disease. Previously, I worked in tech and had no experience in CPG, DTC, or any other 3-letter industries.

If you missed past episodes, I recommend checking out The Great American Poisoning, my take on the chronic disease crisis. Otherwise, let’s dive in!

*******

This month, I’m excited to co-write a guest post with Justin Eaton of Toku. As I age gracefully into my late 30s, my friends and I are starting to take heart disease more seriously, look at cholesterol numbers, and discuss statins and PCSK9 inhibitors.

So much of the conversation around heart disease today is confusing and full of questions. What is certainly true, though, is that today, heart disease is the #1 killer of Americans, while at the same time statins are the most prescribed drug in the US. So - what’s going on?

I wanted to put together a (rather long 😬) post sharing my exploration into the topic heart disease and how we treat it.

Worth noting: neither Justin Eaton (the post’s author) nor I are doctors, nor do we play one on the internet. Please talk to your doctor and/or local LLM for feedback on your specific situation. Take it away Justin Eaton!

***

Why does our body make cholesterol if it kills us?

I was 39, and this was the question I was posing to myself. I had just gotten my Function Health bloodwork back, and almost all of my heart markers were out of range. Not crazy out of range, but elevated. I also realized I was about to be the same age that my uncle was when he died from his third heart attack, and fear started settling in.

Most of the mainstream literature, and my understanding before diving in, said LDL cholesterol is the cause of heart disease and we should be lowering it as aggressively as possible. But there was a growing set of voices saying that LDL is critical to our health, and we’ve been blaming the wrong thing for fifty years. Both sides had credentials. Both had peer-reviewed papers. Both had reasonable-sounding arguments, and confident-sounding leaders.

All the while, cardiovascular disease remained the number one cause of death in America, despite statins being one of the most prescribed drug classes in the history of medicine.

That part stuck with me. If we genuinely knew what caused this, we wouldn’t still be losing the war against it. So I kept reading both sides.

On the mainstream side, I went deep on the most cited meta-analyses on cholesterol and atherosclerosis, pharmaceutical trials, and the work of thought leaders like Peter Libby at Harvard and Brian Ference at Cambridge. On the skeptic side, I went deep on thought leaders like Malcomn Kendrick (he’s quite fun to read on a pretty stale topic FYI), Dr. Aseem Malhotra, and the works of Dr. Nick Norwitz and Dave Feldman through the Citizen Science Foundation.

But somewhere in the middle of reading what felt like my 300th paper on CETP inhibitors (yeah, I didn’t know what that was either), I realized something. I didn’t actually care about LDL cholesterol. What I cared about was that I didn’t want to have a heart attack like my uncle. That changed everything about my framing of the question. I then started asking: why do our bodies make a lot of something that kills us?

Here’s what I found. It’s long. I’m sorry about that, but the topic doesn’t compress well and I don’t want to give you the cliff notes version of a question that took me well over two years to even begin to understand.

I’m not a doctor. I’m a guy with a family history of heart disease, some bad labs, and an openness to being wrong. I’ll tell you where I thought the evidence was strong, where I think it’s weaker, and where I genuinely don’t know (and I’m not sure anyone really does). I’ll link the studies where it makes sense, but this is more of a synthesis of a lot of different perspectives than it is a strong take.

If you came here looking for someone to tell you cholesterol doesn’t matter, this isn’t that piece. If you came here looking for someone to tell you to just trust the mainstream, it’s also not that. If you came here looking for an honest read of somewhere in the middle ground, well, I hope I can deliver part of it.

OK so, what actually happens during a heart attack?

Skip to the next section if not interested in the science here.

Before going any further into the cholesterol debate, I think it’s worth grounding ourselves in what’s actually happening when someone has a heart attack, since that’s what I was actually trying to work backwards from. The good news is this part isn’t controversial. Both camps agree on the mechanics. They just disagree about the upstream cause.

Our arteries are basically tubes with three layers. The innermost is a thin sheet of cells called the endothelium. Atherosclerosis is what happens when the endothelium is damaged and the middle layer accumulates stuff that shouldn’t be there. Lipid material, inflammatory cells, smooth muscle cells, cholesterol crystals, fibrin, and incorporated remnants from previous clots. This accumulation is called a plaque, and it grows slowly over decades.

Most plaques are stable. The dangerous ones are referred to as ‘vulnerable plaques, and have a thin protective cap covering an inflammatory, lipid-rich core (kind of like whitehead pimple, sorry for the visual). When a vulnerable plaque ruptures, the core hits the bloodstream. Your body reacts the way it would to any injury, it activates the clotting cascade so you don’t bleed out. A protein called fibrin begins to form and blood platelets pile on quickly to basically turn your blood into a Band Aid. Within minutes, you have a clot at the site of the rupture and your body has effectively prevented internal bleeding. However, if the clot is big enough to block the entire artery, the tissue downstream of the blockage starts dying. In a coronary artery, that’s what we call a heart attack.

This matters. The decades of plaque buildup from repeated damage to the endothelium set the table for the main event. You don’t die from the plaque buildup, you die from the final clot that forms when the plaque ruptures.

Now here’s where the cholesterol camps diverge, because they don’t agree on how the plaque got there in the first place.

The orthodox ‘cholesterol is causal’ model says plaque starts when LDL particles cross the endothelial lining and get retained in the artery wall. Once retained, they get oxidized. Macrophages show up to clean them up, eat them, and turn into foam cells. The foam cells trigger an inflammatory response that brings in more immune cells. The whole thing grows and then we get the main event. Without LDL retention in the artery wall, no plaque and no heart attack (put very simply).

The thrombogenic model, championed by Malcolm Kendrick, says something different. Kendrick argues LDL particles can’t simply cross a healthy endothelial lining, the actual polarity of the cells makes this impossible (like mixing oil and water). There needs to be damage first. Whatever caused the damage (high blood pressure, smoking, infection, hyperglycemia, autoimmune attack, microplastics, environmental factors, etc…), the body responds the way it always responds to vessel injury, with a clot. Tiny clots form at sites of the damage. Most get broken down. The ones that don’t get covered over by new endothelial cells. The clot becomes part of the artery wall. Over decades, repeated cycles of damage and incomplete repair (think the scar you still have from 20 years ago) build up the layered structure pathologists actually see when they cut open plaques at autopsy.

Regardless of what causes plaque to start, the catastrophic event at the end is the same. A vulnerable plaque ruptures. A clot forms. The artery blocks. The tissue downstream dies. The actual moment that kills you or dramatically changes your life is (for most heart attacks) fibrin and platelets clotting, regardless of which upstream story you believe about how the plaque got there.

That fact ended up shaping how I think about prevention more than anything else I read. I’ll come back to it.

So what is LDL, and why do we make it?

Before getting into whether LDL is killing us, I had to back up to a more basic question. If LDL cholesterol is responsible for the #1 cause of death globally, why does our body work so hard to make so much of it? A body doesn’t usually go to a lot of metabolic effort to manufacture something that kills it.

And our body really, really wants to make this stuff. Roughly 75% of the cholesterol in your blood is produced by your own liver. Only about 25% comes from your diet. If you eat less, your liver makes more. If you eat more, your liver makes less. There’s an elaborate feedback system designed specifically to keep cholesterol levels within a particular range. Evolution doesn’t usually build elaborate feedback systems for things that are just bad for you.

So why does our body care so much? It turns out cholesterol does a lot.

It’s structural. Every cell membrane in your body uses cholesterol to maintain integrity and fluidity. Without it, cells literally start to fall apart.

It’s the precursor for basically every steroid hormone you make: testosterone, estrogen, cortisol, aldosterone, vitamin D, progesterone, DHEA… without cholesterol, you wouldn’t have hormones!

It’s also critical for brain function. Your brain is roughly 25% cholesterol by weight. It’s used in myelin (the fatty insulation around your neurons) and in synapse function and neurotransmitter production. Fun fact: your brain doesn’t even trust the rest of the body to supply enough cholesterol, it makes its own, locally!

You can see why “let’s drive LDL as low as possible” gets complicated quickly. We’re talking about a molecule the body makes deliberately, regulates carefully, and uses for at least a half dozen essential functions.

There’s actually some real evidence that having cholesterol too low can have negative consequences. Multiple studies in elderly populations have found that very low total cholesterol is associated with higher all-cause mortality, not lower. Some studies have linked low cholesterol to depression and increased suicide risk, with a meta-analysis of 32 studies covering more than 7,000 patients finding significantly lower total and LDL cholesterol in patients with major depressive disorder who attempted suicide compared with those who did not. A meta-analysis of placebo-controlled randomized trials confirmed statins lower testosterone in both men and women, and the effect has been replicated in larger systematic reviews since. The cognitive effects that prompted the FDA’s 2012 statin label warning have been somewhat inconsistent in randomized trials but very real for the subset of people who experience them.

I want to be careful here, because the skeptic side often overstates the above findings. None of it means high LDL is good. None of it disproves that lowering LDL reduces cardiovascular events in the right populations. What it means is that we’re playing with a regulated, multipurpose molecule that the body explicitly chooses to make. Given that, the “lower is always better” framing deserves scrutiny.

Is LDL causing heart disease? Are we even measuring the right thing?

S we know what plaque is, and we know how it kills us. The next question is whether the LDL itself is the thing driving plaque formation, or whether it’s just along for the ride. The mainstream says causal. The skeptics say it’s complicated. And the answer, as best I can tell, is that they’ve both been arguing about a metric that is too broadly defined to settle the question.

The strongest mainstream argument is from a study doing something called Mendelian randomization. The idea is instead of giving someone a drug to lower LDL and following them for five years, you find people genetically programmed to have lower lifetime LDL and compare their cardiovascular outcomes. Brian Ference’s 2012 paper in JACC found that people with naturally lower lifetime LDL have substantially lower cardiovascular event rates. The European Atherosclerosis Society Consensus Panel synthesized this evidence in a 2017 statement declaring LDL definitively causal in cardiovascular disease. These are large scale studies, and shouldn’t be disregarded, even if some of the conflicts of interest in that 2017 statement raised my skeptical eyebrows.

The strongest skeptic argument, in my opinion, is something called the CETP inhibitor graveyard. (Yes, that’s what it’s called and how badly it went.) CETP is the enzyme that swaps fats around between particle types. Inhibit it, you can lower LDL and raise HDL substantially. Pharma thought they had something special. They ran four major trials.

Torcetrapib lowered LDL 20% and raised HDL 60%. The trial stopped early because it increased mortality by 58%. The dalcetrapib trial didn’t help and was stopped. Evacetrapib dropped LDL 37%, raised HDL 130%, and produced exactly zero reduction in cardiovascular events. Anacetrapib eventually showed a modest benefit, but it was attributable to its effect on non-HDL cholesterol rather than the HDL increase. The only one of four to produce any benefit at all.

Four drugs, three lowered LDL successfully, only one helped and another actually killed people. If lower LDL automatically meant fewer cardiovascular events, this couldn’t have happened. Whatever the drugs that do work are doing, it isn’t simply lowering the LDL number.

So both camps have something to hold onto here. Mendelian randomization says lower lifetime LDL is associated with fewer events. CETP graveyard says you can lower LDL without producing any benefit at all. So how can both of these be true? Turns out there actually is a pretty easy way to make both sides correct.

When the mainstream and the skeptics both talk about LDL, they’re almost always talking about LDL-C. The number on your standard lab panel.

LDL-C is, frankly, kind of an archaic metric. It just measures the total volume of cholesterol cargo riding inside your LDL particles. Not the number of particles. Not their size. Not their oxidation status. Not whether you have a separate genetic risk factor like Lp(a) hiding in the mix. We’ve been using it for fifty years simply because, for a long time, it was the only measurement we could affordably run at scale. Once an entire field calibrates its trial data and treatment guidelines to a particular metric, that metric becomes the metric, even when better ones come along.

Mainstream cardiology has been trying to move past LDL-C for years. Here’s what they’ve been trying to move toward:

ApoB is a protein that sits on the surface of every LDL particle. Measure ApoB and you’re literally counting the number of LDL particles in someone’s blood, not the volume of cholesterol inside of them like LDL-C. The European Society of Cardiology recognized ApoB as a more accurate risk marker than LDL-C in 2019. Canada allowed ApoB to replace LDL-C as a primary target in 2021. The American College of Cardiology followed in 2022. Simply put, two people could have the same LDL-C number but a wildly different ApoB number.

Particle size is another metric to look at. Pattern A (large, buoyant LDL) is associated with lower cardiovascular risk than Pattern B (small, dense LDL), independent of total LDL-C. Small dense particles enter the artery wall more easily, oxidize more readily, and stay in your bloodstream longer because it’s harder for your blood to clear them. Just looking at the LDL-C number on your panel doesn’t tell you the size of these particles. This is powerful to combine with ApoB where we are trying to zero in on the count of small LDL particles in the total volume of LDL-C.

Oxidized LDL (OxLDL) is another blood test that measures the weight of oxidized LDL which is the form that actually drives plaque formation. Native LDL is largely benign. The atherogenic process really starts when LDL gets oxidized, which is a process heavily influenced by metabolic health. Again, we’re taking a much narrower look here than the typical LDL-C. OxLDL would be captured as a subset of your LDL-C metric, but if you don’t measure it, you have no idea how much of your LDL is oxidized.

Lipoprotein(a), or Lp(a), is a separate genetic risk factor that affects roughly 20% of people. It’s an LDL-like particle with an extra protein attached that makes it both more atherogenic and more pro-thrombotic (causes more clotting) than regular LDL. This is largely determined by genes and doesn’t really respond to statins. Most doctors don’t test for it because there hasn’t been a good treatment. However, the total weight of our Lp(a) is also reflected in our LDL-C numbers.

Now here’s another area where the cholesterol argument gets interesting with the above framing. The LDL skeptics often point to a 2009 paper in the American Heart Journal that looked at lipid panels of 136,000 patients hospitalized with coronary artery disease. Roughly half had LDL-C below 100 mg/dL when admitted (i.e. what we consider ‘healthy’ LDL levels). The skeptics use this to argue LDL clearly isn’t the whole story. The mainstream used the same data to argue we should be lowering LDL targets even further.

But there’s a third reading here. What if half those people had LDL-C below 100 but their ApoB was high, or their particles were mostly small and dense, or their LDL was heavily oxidized, or they had elevated Lp(a)? We don’t know, because the standard panel didn’t measure any of that. The LDL-C number was just too crude to tell us what was actually happening in those particles.

If you revisit the Mendelian randomization data with this perspective, those genetic findings start to make a lot more sense. Seeing fewer events in folks naturally predisposed to low LDL is a legitimate observation, but the underlying driver could potentially be explained by particle quality and count rather than the volume of cholesterol. Reducing a lifetime of exposure essentially means fewer opportunities across a large population for particles to get trapped, oxidized, and trigger an inflammatory cascade. In this light, the mainstream view that lower levels correlate with better outcomes across a population doesn’t actually contradict the skeptic’s point that LDL-C is a crude, often deceptive tool for assessing an individual’s risk profile in isolation. It turns out both camps could likely be right at the same time.

This was the moment when the cholesterol debate stopped feeling like a real disagreement and started feeling like an artifact of historical data limitations. We’ve been arguing about whether the number on the LDL-C test is causal, when the more honest answer is that it likely depends on the status of those particles that number is summarizing, and the LDL-C number simply doesn’t tell us.

This is roughly where I’ve ended up: that the composition of your LDL is likely what really matters, and the upstream metabolic health impacting LDL is far more important than the number itself. The total LDL-C is a useful starting point, but an increasingly outdated proxy. As ApoB, particle size, oxidation status, and Lp(a) testing become standard, I suspect a lot of what currently looks like a disagreement between the camps will resolve into “you were both partially right, the metric was lying to both of you.”

Which raises the natural next question: if our body regulates cholesterol so carefully, how does it get out of range in the first place?

Why poor metabolic health is the real question to answer

This is the part where I think most of the typical “LDL is bad” arguments tends to ignore or disregard the real upstream metabolic issue.

Insulin resistance, chronic inflammation, hyperglycemia, hypertriglyceridemia, low HDL, small dense LDL, glycated proteins… these are not separate problems. These are one problem showing up in different lab tests, and are all symptoms of metabolic dysfunction.

Here’s what’s happening when your metabolic health is in a bad place (warning, this may get a bit technical. If too much, skip forward to the next section):

Your body’s job is to keep blood sugar in a tight range. You eat carbs, blood sugar goes up, pancreas releases insulin, cells take up sugar.

When you eat too many carbs too often, or carry around too much extra body fat, your cells start ignoring insulin. Your pancreas pumps out more, trying to get the message through. Now your blood is full of both sugar and insulin, and neither is doing its job. This is insulin resistance, and it’s the upstream of upstreams for a huge number of diseases.

Here’s what happens to your cholesterol when this is going on. The mechanisms are well described in atherosclerotic lipid research like this, if you’re interested in a deep dive, but to cover it simplify:

Your liver, faced with an influx of excess sugar and surplus energy, begins converting this energy into fat, packaging it into an overabundance of triglyceride-rich particles known as VLDL (very low-density lipoproteins). Upon entering the bloodstream, a critical enzyme called CETP begins exchanging the contents of these VLDL particles with your other lipoproteins with the object of getting this energy stored away as quickly as possible. This process primarily involves shifting triglycerides into your LDL and HDL particles while pulling cholesterol out of them. This triglyceride-rich LDL is then processed by another enzyme, hepatic lipase. When hepatic lipase strips the triglycerides away to be stored as fat, the structure of the LDL collapses, resulting in an abundance of smaller, denser, and potentially more atherogenic LDL particles. This same process also breaks down HDL particles, resulting in lower total HDL. So, this single upstream metabolic problem of insulin resistance leads to this classic higher risk profile of high triglycerides, low HDL, and an overrepresentation of small, dense LDL particles, appearing as three distinct issues on a standard lab report but sharing one root cause.

There’s a second issue going on, too. When blood sugar runs high for long enough, sugar molecules randomly attach to proteins in your blood, including the surface of your LDL particles. This is glycation: the same process that turns bread crispy brown when you toast it. Glycated LDL is more readily oxidized, more inflammatory, and harder to clear. So high blood sugar gives you both small dense LDL particles and damaged, oxidized ones.

But taking a step back into why this matters is that metabolic dysfunction isn’t only an upstream cause of heart disease. It’s an upstream cause of basically everything killing people in the modern world. Type 2 diabetes is metabolic dysfunction by definition. Obesity is closely tied to it. Some researchers call Alzheimer’s “Type 3 diabetes” because the brain pathology shows insulin resistance signatures. There are connections to certain cancers, fatty liver disease, and depression.

If you wanted to pick one upstream variable to fix and watch the largest number of downstream conditions improve, metabolic health would be hard to beat.

So, where do statins fit in?

One of the most heated debates around the cholesterol argument always seems to land on statins. And since it was what I was about to be prescribed, I wanted to take a fair look at both sides.

There’s reasonably strong evidence that statins reduce cardiovascular events, at least in certain scenarios. The core of the argument from the pro-statin side rests on the concept of Relative Risk Reduction (RRR), which expresses the percentage by which a treatment reduces the risk of an event compared to the control group. For instance, the Cholesterol Treatment Trialists’ Collaboration out of Oxford pooled data and found about a 22% relative risk reduction in major vascular events for every 1 mmol/L drop in LDL-C, a compelling and consistent effect across populations.

Where the debate shifts is when the statin skeptics reframe this argument by focusing on Absolute Risk Reduction (ARR), which measures the raw difference in event rates between treated and untreated populations rather than using percentages to tell the story.

From this lens, where statins clearly seem to work best is secondary prevention, meaning people who’ve already had a heart attack, stroke, or established cardiovascular disease. The number needed to treat (NNT) to prevent one major non-fatal cardiovascular event here is roughly 1 in 39 over 5 years, and 83 to prevent one death. This means if we treated 39 people who already had a cardiovascular event with statins for 5 years, one major cardiovascular event would be prevented attributable to the statin. For those with established disease, the math leans towards taking one.

Where the math gets less convincing is in primary prevention, meaning healthy people with ‘elevated cholesterol’ who have not yet had an event. The absolute risk reduction is much smaller. According to NNT.com, you’d need to treat 104 people for 5 years to prevent one non-fatal cardiovascular event, with no statistically significant mortality benefit at all in this group. The skeptics emphasize this with a separate meta-analysis that calculated the median postponement of death within trial duration to be only 3.2 days in primary prevention. The mainstream counter is that 5 years isn’t the right window, and the real benefit comes from decades of cumulative LDL exposure reduction. That’s a defensible position consistent with the Mendelian randomization data, but it’s also an extrapolation beyond what we’ve actually measured in trials.

Either way you slice it, the 22% relative reduction is statistically right, but the absolute benefit is small. And the 103 other people you treated to prevent that one non-fatal heart attack are still on the drug, exposed to its systemic effects, both good and bad, beyond lowering LDL.

That last part is where it gets interesting, because statins do a lot more than just lower LDL.

Statins block an enzyme called HMG-CoA reductase, which kicks off something called the mevalonate pathway. The mevalonate pathway doesn’t just make cholesterol. It also makes CoQ10, the molecule your mitochondria literally cannot function without. It makes dolichols, heme A, and several other compounds your cells use for everyday operations. So when you take a statin, you’re not just lowering LDL. You’re also throttling an entire upstream metabolic pathway that produces a bunch of things your cells need.

This also matters because of where statins act. HMG-CoA reductase exists in basically every cell in our bodies, and statins act on it everywhere. When you block a critical metabolic pathway in all of our cells, it should be no surprise there are likely side effects.

Diabetes risk is the most important one. Multiple meta-analyses of randomized trials have found about a 9 to 13% relative increase in new-onset diabetes in statin users, with higher dosing pushing that to 18%. This is somewhat ironic because type 2 diabetes is itself a major cardiovascular risk factor. The math still favors statins in higher-risk patients. But it changes the calculus for people whose underlying issue is metabolic dysfunction in the first place.

Muscle pain is the next most discussed, and this is where the mevalonate pathway issue shows up directly. CoQ10 is essential for mitochondrial energy production, especially in muscle, and plasma CoQ10 levels measurably drop 25 to 50% in statin users. So when people report muscle pain on statins, there’s a biochemical explanation that isn’t just in their head.

That said, the statin trial data tells a confusing story. Real-world surveys find 10 to 20% (at least) of statin users report muscle symptoms. Randomized trials, however, find this to be 1 to 5%, much closer to placebo. Part of that gap could be a genuine nocebo effect (people expect side effects and, in turn, perceive they’re experiencing them). Part of it is likely structural to the trial itself.

Most major statin trials use a “run-in period” where everyone takes the drug for several weeks before being officially randomized. People who can’t tolerate it drop out before the trial starts, so the randomized population is made up of people who already tolerate statins well. This is a known methodological feature, not a conspiracy theory, but it likely accounts for some of the gap between trial-reported and real-world side effect rates.

It’s also worth noting that nearly every major statin trial was funded by the pharmaceutical company that makes the drug, and the patient-level data behind the most influential meta-analyses is held by the CTT Collaboration at Oxford, who don’t release it for independent reanalysis. The CTT itself maintains it doesn’t take direct pharma money for the meta-analyses. But it operates inside a research unit that has received hundreds of millions in pharma funding over the years, and the writing committees on its papers include many investigators who ran the original industry-funded trials. None of this means the data is wrong. But, it does mean the data has been examined less independently than you might assume, given how often it’s cited.

So my view is that statins aren’t poison, but they don’t seem to be a panacea. The case is strongest in secondary prevention. The case is weaker in primary prevention, especially in people whose elevated LDL might be downstream of metabolic dysfunction the statin could make slightly worse.

I’m not anti-statin. I’m anti-pretending statins solve a problem they don’t really solve.

It’s also worth noting that there are newer drugs like PCSK9 inhibitors, ezetimibe, and bempedoic acid that are much more targeted in how they reduce LDL, rather than the broad, sweeping approach that statins take by inhibiting cholesterol production in every cell in our body. Regardless of your stance on LDL, being more targeted in how we reduce it seems to be a much more logical approach, at least in my opinion.

And fortunately, it’s becoming easier and easier to get more targeted! Rather than just taking a LDL measurement and deciding if you should be on statins or not, it’s increasingly possible for one to get a plaque score (via a scan like Cleerly or something) and understand your actual plaque buildup. If you can do this and confirm you have no plaque, in my mind there’s absolutely no reason to get on statins.

Summarizing where I’ve ended up

Heart disease is complicated (revolutionary statement, I know). We’ve been treating it like a single factor disease for over fifty years because that factor was easier to measure and easier to drug.
LDL does a lot of essential things in our body. The “lower is always better” frame ignores that we’re playing with a regulated, multipurpose molecule that the body explicitly chooses to make. Lowering it seems to have real potential upsides for cardiovascular events in the right populations, and real downsides we’re still figuring out.
LDL-C is the wrong metric to fight about. The composition of your LDL almost certainly matters more than the total amount. Small dense particles are more dangerous than large buoyant ones. Oxidized and glycated particles are worse than native ones. Lp(a) is a separate genetic factor most people don’t even know they have but gets lobbed in with LDL-C. As ApoB, Lp(a), particle size, and oxidation status testing become standard and more studied, I suspect a lot of the cholesterol debate will resolve into both sides being partially correct. That being said, this view is contested. Mainstream researchers point to studies showing LDL-C correlates with subclinical atherosclerosis even in people without other risk factors, suggesting LDL matters regardless of context. They may be right. But until those studies measure particle composition directly rather than just LDL-C, we won’t know whether the correlation is about total particles or about the same composition story showing up under a different name.
Metabolic dysfunction is the upstream master variable. It produces the atherogenic LDL composition. It produces the inflammation that makes plaque vulnerable. It produces the prothrombotic state that turns ruptured plaques into heart attacks. It also produces type 2 diabetes, contributes to obesity, and is implicated in a long list of other modern diseases. If you’re going to spend energy fixing one thing in your cardiovascular health, fixing metabolic health probably gets you more benefit than any cholesterol targeted intervention.
The catastrophic event is a clot. Plaque sets the stage. But, the thing that actually kills you is fibrin and platelets occluding an artery in the minutes after a vulnerable plaque ruptures.
Statins may modestly reduce cardiac events, with caveats. They’re most defensible in secondary prevention. They’re least defensible in metabolically dysfunctional patients where the drug may worsen the upstream cause. The more targeted therapies that are currently used when statins don’t work are likely a better answer.

The most confident thing I can say is that the loudest voices in this conversation, on both sides, are usually more confident than the evidence supports.

What I decided to do

Given everything I just walked through, here’s where I started thinking about what to actually do for myself.

The piece I kept coming back to is the catastrophic event. The clot. Both camps in the cholesterol debate agree on this part. The disagreement is upstream. Whatever you believe about LDL, you don’t disagree about the most common type of heart attack itself.

It’s ultimately how I found nattokinase, an enzyme with powerful fibrinolytic activity derived from a fermented soy dish called natto. Fibrinolytic activity means that it breaks down fibrin, and there were demonstrable, double blind placebo controlled human studies showing that it could actually help break down blood clots. Which was the one thing that we seemingly could all agree on being at the core of many cardiovascular events.

One of the most compelling studies showed 1,062 participants over 12 months, taking 10,800 fibrinolytic units per day resulted in a 36% reduction in carotid plaque size measured by ultrasound, plus reductions in carotid intima-media thickness (i.e. it was actually reversing atherosclerosis). There were also notable lipid improvements including 16.5% reduction in triglycerides, 17.9% reduction in LDL-C and a 17% improvement in HDL-C in that same study. Not to mention other studies showing improvements to inflammatory markers, clotting markers, and blood pressure, so candidly, for me it felt like a no-brainer at the intersection of all these conflicting and confusing positions on LDL cholesterol, with no reported downsides.

What we don’t have is a massive, randomly controlled trial on whether daily nattokinase reduces cardiovascular events over decades (as much as I’d love one). And we likely never will, as natural compounds lack the same financial incentive that proprietary pharmaceuticals do.

What nattokinase does for me is sit at the exact intersection of nuance in this conversation that I landed on. I certainly don’t think it is a magic pill. It won’t fix upstream metabolic illness, and like any intervention, it may work for some and not for others. But at least for me, when dosed at 10,800 FU (taking Toku), over 6 months it had a profound impact on my triglycerides (32% reduction), LDL-C (38% reduction), ApoB (20% reduction), and blood pressure (systolic average dropped ~10 points). It was so mindblowing that I actually decided to build a brand around it as it felt more along the lines of something everybody should be taking more than statins did.

A final thought

When I started this, I thought I’d find a clear answer. There isn’t one. I’ve changed my mind on parts of this three or four times while writing it, and I’ll probably change it again next time someone publishes something I haven’t read yet. Over two years in, I realized that’s just the nature of this question. The research is genuinely unsettled in a lot of areas. The orthodoxies on both sides are more confident than they should be.

I also know that we are all individuals. We are not a meta-analysis. We are not a randomized controlled trial. While those are incredibly powerful tools for making broader public health decisions, they should never be used to replace your own experience. Nobody cares or should care more about your own health than you, and what works for one person won’t necessarily work for you. That doesn’t mean the other side is wrong, and it doesn’t mean that you are right. It just means that humans are complicated, incredible systems that are all uniquely programmed and we don’t have the full instruction book yet.

What I keep coming back to is that we’re the first generation that even has access to this conversation. My parents and grandparents didn’t have access to the level of detail that blood tests can now show. The commoditization of blood testing and more people taking charge of their health earlier will allow us to be the first in over a hundred years to reverse the trend of heart disease. That’s my hope.

(see references for this post at the bottom!)

😌 Dope stuff on the internet

Some of my favorite things since the last newsletter (note: I don’t get paid to recommend anything here):

📰 Article - I’ve been very into bioelectricity the last few years, and am increasingly bullish on it as a future Big Idea in biology. This post is a good overview of Dr. Michael Levin’s work - expect more to come from me on this topic in future episodes!
📚 Book rec - I just started reading Consider Phelbus, the first of the fairly well-known sci-fi Culture series. So far, I’d say it’s just okay: really am looking for other sci-fi books along the lines of Three-Body Problem, Red Rising, Nexus, and others that I’ve read and really enjoyed.
⌚ Cool product - I signed up for a future Jhourney meditation retreat, which I’m incredibly excited about. I’m also friends with the founders, and they were kind enough to share a code - justinmares1000 - that gets the user $1,000 off any of our retreats this year. Just a heads-up that it’s single-use, so it’ll only work for the first reader to redeem it!
🎵 Music - This new James Grant set is excellent.
🏀 Random - I have absolutely no idea how true this is, but this Chinese researcher is working on extending the time between periods in order to delay menopause. From the AI slop thread: “Before birth control, women were pregnant or breastfeeding through most of their fertile years, and both pause periods. About 100 cycles in a lifetime. Now, with later marriages, fewer kids, and longer lives, the number sits above 400. Each cycle uses up eggs. Run out, menopause starts.”
🔥Hot take - I spent way too long putting together my thoughts on why I’m skeptical of the “LDL directly causes heart disease” argument that Cremieux recently shared online. If you’re curious about it, I wrote my response here - very open to thoughts and feedback, as all thoughts are in draft mode. More long-form heart disease content from yours truly.
🙋‍♂️ Ask - If you know any talented engineers or marketers, we are hiring at Truemed. Come join one of the fastest growing companies in all of consumer healthcare, and make a dent in the chronic disease crisis 💪

***

I’ll be at Edge Esmeralda for the rest of June. Please do say hi if you end up making the trip to beautiful Healdsburg, CA! Otherwise, see you in 30 👋