RAG CHECK: Has New Research Really Shown That High Cholesterol Is Not Linked to Heart Disease?

There is a kernel of genuine and genuinely interesting science behind this claim. Recent research has found groups of people with extraordinarily high LDL cholesterol who nevertheless have surprisingly little detectable coronary atherosclerosis, and within some of these carefully selected groups, researchers have failed to find a statistically significant relationship between LDL concentration and measured coronary plaque. Other studies, particularly among older adults, have reported complicated or even U-shaped relationships between cholesterol and mortality.

Those findings deserve attention. They challenge some overly simplistic ideas about cholesterol and cardiovascular risk.

What they have not demonstrated is that LDL cholesterol has no relationship with atherosclerotic cardiovascular disease.

That distinction matters enormously.

🟢 What the New Research Really Found

One of the most important studies behind the recent controversy is the KETO-CTA study, published in JACC: Advances in 2024.

Researchers studied a highly unusual group of people following carbohydrate-restricted ketogenic diets. Many were lean, physically healthy and metabolically quite different from the population normally associated with high cholesterol.

Their LDL cholesterol had risen dramatically after adopting carbohydrate restriction. These individuals are sometimes described as Lean Mass Hyper-Responders (LMHRs).

The study included people with LDL-C of at least 190 mg/dL. In the matched analysis, average LDL-C was approximately 272 mg/dL in the ketogenic group versus roughly 123 mg/dL among controls.

Those are extraordinarily different cholesterol levels.

Yet when researchers examined their coronary arteries using coronary calcium and CT angiography, the high-LDL ketogenic group did not have significantly greater coronary plaque burden than the matched control group.

Even more provocatively, within the ketogenic cohort, the investigators did not find a statistically significant relationship between LDL-C concentration and the amount of coronary plaque detected.

It is therefore reasonable to say:

Among this particular population of lean, metabolically healthy people who developed very high LDL while following carbohydrate-restricted diets, higher LDL was not significantly associated with greater measured coronary plaque burden during the period studied.

That is quite different from saying:

“Scientists have discovered that LDL doesn't cause heart disease.”

The first statement describes the study. The second extrapolates far beyond it.

🔴 “The Study Proves High LDL Is Harmless”

The study cannot establish this.

The participants were highly selected. They weren't representative of the general population. Their extreme LDL elevations had generally existed for several years rather than for their entire lives. Coronary atherosclerosis normally develops over decades.

Most importantly, the study did not randomly assign thousands of people to very high or low LDL concentrations and follow them for 30 years to determine who suffered heart attacks and strokes.

It examined arterial plaque in a relatively small and unusual population.

That makes the research scientifically interesting but limits the conclusions that can legitimately be drawn from it.

Cholesterol Is Not Simply “Grease in the Pipes”

Older explanations of cholesterol sometimes gave people an unfortunate mental picture: eat cholesterol, cholesterol enters the bloodstream, cholesterol sticks to arteries, artery becomes clogged.

Human biology is considerably more complicated.

Cholesterol is essential to life. It is used in cell membranes, steroid hormones, bile acids and numerous other biological processes.

Because cholesterol is not water-soluble, the body transports it through the bloodstream inside structures called lipoproteins.

Consequently, when doctors measure “LDL cholesterol,” they aren't simply counting lumps of cholesterol floating through the bloodstream. They are estimating the amount of cholesterol being transported inside LDL particles.

That distinction has become increasingly important.

ApoB: Counting the Trucks Instead of the Cargo

Imagine cholesterol as cargo being transported around the bloodstream in delivery trucks.

LDL-C measures approximately how much cholesterol cargo is being carried.

Another measurement called Apolipoprotein B, or ApoB, gives information about the number of potentially atherogenic particles circulating in the bloodstream.

Each LDL, IDL, VLDL-remnant and Lp(a) particle carries one ApoB molecule.

So two people can theoretically have similar LDL-C concentrations while having different numbers of atherogenic particles.

This is one reason modern cardiovascular medicine increasingly pays attention to ApoB, non-HDL cholesterol and Lp(a), rather than treating total cholesterol as the entire story.

The 2026 ACC/AHA dyslipidaemia guideline reflects this development, giving ApoB and Lp(a) greater roles in cardiovascular risk assessment.

That isn't evidence that LDL has become irrelevant. It means cardiovascular risk assessment has become more sophisticated.

🟢 Does LDL Actually Cause Atherosclerosis?

If the only evidence were “people with high LDL have more heart attacks,” we would have a classic correlation-versus-causation problem.

But observational correlation isn't the only evidence we have.

The evidence comes from epidemiology, genetics, pathology, familial hypercholesterolaemia and randomized clinical trials.

The Genetic Evidence

Human genetics provides something resembling a gigantic natural experiment.

Some people inherit genetic variants that keep LDL relatively low throughout their lives. Others inherit variants that increase it. Across different genes and populations, variants producing lower lifelong LDL exposure are generally associated with substantially lower coronary disease risk, while variants producing higher lifelong exposure generally move cardiovascular risk in the opposite direction.

This approach—particularly Mendelian randomization—helps overcome some of the confounding inherent in ordinary observational research.

Familial Hypercholesterolaemia: Nature's Long-Term Experiment

People with familial hypercholesterolaemia (FH) can experience extremely elevated LDL from childhood because of inherited abnormalities affecting LDL metabolism.

Untreated FH dramatically increases the probability of premature coronary artery disease. In its rare and extremely severe homozygous form, LDL can become extraordinarily high from childhood and accelerated atherosclerosis can occur at remarkably young ages.

That observation is difficult to reconcile with the hypothesis that LDL is biologically irrelevant to atherosclerosis.

Interestingly, the ketogenic KETO-CTA researchers excluded people with familial hypercholesterolaemia. That was scientifically reasonable because they were investigating a different phenomenon, but it also means the KETO-CTA results cannot simply be generalized to people with genetically elevated lifelong LDL.

Randomized Trials Provide Another Test

If LDL were merely an innocent marker correlated with cardiovascular disease, deliberately reducing it shouldn't repeatedly reduce cardiovascular events.

Yet that is what randomized trials have found.

Statins reduce LDL primarily by increasing LDL-receptor activity. Ezetimibe lowers cholesterol through a different biological mechanism involving intestinal cholesterol absorption. PCSK9 inhibitors lower LDL through yet another pathway.

Despite those different mechanisms, lowering LDL consistently reduces cardiovascular events in appropriately selected populations.

A landmark Cholesterol Treatment Trialists meta-analysis involving approximately 170,000 participants across 26 randomized trials found that reducing LDL by approximately 1 mmol/L — about 39 mg/dL — reduced major vascular events by roughly one fifth.

The PCSK9 Experiment

PCSK9 inhibitors provide an especially interesting test because they can reduce LDL dramatically without using the statin mechanism.

In the large FOURIER randomized trial, more than 27,000 people with established cardiovascular disease were studied. Evolocumab reduced LDL by approximately 59%, bringing median LDL to around 30 mg/dL. Cardiovascular events decreased.

Alirocumab subsequently produced compatible cardiovascular findings in another major randomized trial.

So we now have multiple drugs manipulating LDL through different biological pathways and repeatedly producing reductions in cardiovascular events.

🟠 But the Cholesterol Skeptics Have a Legitimate Point

The old statement “High cholesterol causes heart disease” is scientifically inadequate.

Total cholesterol alone is a particularly crude measurement. Two people could have exactly the same total cholesterol while possessing dramatically different cardiovascular risk profiles.

Their LDL, ApoB, HDL, triglycerides, Lp(a), blood pressure, glucose metabolism, smoking history, kidney function, genetics and existing arterial plaque might be completely different.

So the modern question isn't simply “What's your cholesterol?” It is increasingly: “What is your lifetime exposure to atherogenic particles, what other cardiovascular risks do you carry, and what evidence is there that atherosclerosis is actually developing?”

The Importance of Time: LDL “Pack-Years”

One of the most important concepts missing from many internet arguments is cumulative exposure.

Consider smoking. Someone smoking twenty cigarettes per day for three years doesn't carry the same accumulated risk as someone smoking twenty cigarettes per day for forty years.

Their current cigarette consumption is identical. Their lifetime exposure isn't.

Something conceptually similar happens with LDL.

Consider someone whose LDL remained approximately 110 mg/dL until age 50 and then increased to 270 after beginning a ketogenic diet. At age 55, that person has experienced approximately five years of extreme LDL exposure.

Compare that with someone with familial hypercholesterolaemia whose LDL has been 270 since childhood. Their cholesterol concentration today might be identical. Their cumulative arterial exposure is profoundly different.

That may help explain why relatively healthy arteries can coexist temporarily with extremely high LDL in some ketogenic hyper-responders. A clean scan today does not necessarily tell us what those arteries will look like after another twenty or thirty years.

A Statistical Trap: Restricted Range

Imagine researchers investigating whether height helps basketball performance. But instead of studying everyone from 5 feet tall to 7 feet tall, they recruit only professional centers measuring between 6'10" and 7'1".

Within that highly selected population they might discover very little correlation between height and basketball performance.

It would be incorrect to conclude: “Height has nothing to do with basketball.”

The researchers selected an unusually narrow portion of the population. This is called range restriction.

Something similar potentially affects studies where nearly everyone already has extremely elevated LDL.

Failure to find a statistically significant LDL-plaque relationship within that selected population does not automatically demonstrate that LDL is unrelated to plaque across the broader population.

“No statistically significant association detected” does not mean “scientists proved there is no association.”

The Controversial Longitudinal KETO-CTA Study

The researchers subsequently followed approximately 100 ketogenic hyper-responders and measured changes in coronary plaque.

The resulting longitudinal research produced even more dramatic headlines because investigators reported that neither LDL exposure nor ApoB significantly predicted plaque progression, while existing plaque burden strongly predicted subsequent progression.

That finding, if confirmed over longer periods and in larger populations, could be important.

However, the 2025 JACC: Advances publication titled “Longitudinal Data From the KETO-CTA Study” was subsequently marked RETRACTED.

Retraction does not automatically establish that underlying data were fraudulent. Papers can be retracted for numerous reasons. But a retracted paper should not be treated as settled clinical evidence.

A newer 2026 manuscript/preprint has reported prospective results again. Until that work completes the appropriate peer-review and replication process, its findings should be considered provisional and hypothesis-generating rather than definitive.

Something Often Missing From the Headlines

The ketogenic participants didn't universally remain plaque-free.

The newer longitudinal analysis reported progression of non-calcified coronary plaque in a substantial portion of the cohort, although absolute plaque burden remained relatively low for many participants.

So the result wasn't “people with massive LDL didn't develop atherosclerosis.” It was closer to: “People showed differing amounts of plaque progression, and within this relatively small, unusual cohort, differences in LDL/ApoB did not statistically explain those differences.”

What About Studies Showing Higher Cholesterol Is Associated With Longer Life?

These studies also exist.

Particularly among elderly populations, researchers have sometimes observed U-shaped relationships between LDL and mortality.

Mortality can appear elevated at both very low and very high cholesterol concentrations.

That sounds alarming until another major epidemiological problem is considered: reverse causation.

Cancer, chronic inflammation, frailty, malnutrition, liver disease and severe chronic illnesses can lower cholesterol.

Consequently, an observational study might find low cholesterol ↔ increased mortality without demonstrating low cholesterol → increased mortality.

Instead, the biological pathway might partly be: serious illness → falling cholesterol + increased mortality.

Heart Disease Isn't One Disease

LDL is principally implicated in atherosclerotic cardiovascular disease — ASCVD.

That includes coronary artery disease, many myocardial infarctions, ischemic stroke and peripheral arterial disease.

But cardiovascular disease also includes arrhythmias, congenital heart abnormalities, myocarditis, cardiomyopathies, valvular disorders and other conditions that LDL isn't expected to cause.

So even saying “LDL causes heart disease” is scientifically imprecise.

RAG CHECK — The Claims

Why This Isn't Just “Correlation”

Correlation alone doesn't establish causation. Absolutely correct.

But the LDL hypothesis doesn't rest solely upon correlation.

• Epidemiology: greater lifetime LDL/ApoB exposure tracks with greater ASCVD.
• Pathology: ApoB-containing particles are retained within arterial walls during atherogenesis.
• Genetics: inherited variants altering lifelong LDL exposure alter coronary risk in predictable directions.
• Familial hypercholesterolaemia: extreme lifelong LDL produces dramatically accelerated atherosclerosis.
• Statin randomized trials: lowering LDL reduces cardiovascular events.
• Ezetimibe randomized trials: lowering LDL through another pathway reduces events.
• PCSK9 randomized trials: lowering LDL through yet another pathway reduces events.
• Dose and duration: larger and longer LDL reductions generally produce larger reductions in cardiovascular risk.

Any one of these observations could potentially be challenged. Explaining all of them simultaneously without LDL/ApoB participating causally becomes considerably more difficult.

What the New Research May Actually Be Telling Us

The genuinely interesting message from these controversial studies may eventually prove more important than the headline claiming cholesterol doesn't matter.

The research suggests that people respond very differently to the same LDL concentration.

LDL is not destiny. Neither is ApoB.

Atherosclerosis develops within a complicated biological system involving particle exposure, endothelial function, arterial retention, inflammation, blood pressure, insulin resistance, smoking, genetics, kidney function, thrombosis, age and—perhaps most importantly—time.

The next generation of cardiovascular medicine therefore appears increasingly likely to combine blood biomarkers with direct assessment of disease.

Final RAG Verdict

The responsible scientific position therefore lies between two extremes.

The old slogan “cholesterol is bad and high cholesterol means you're going to have a heart attack” is obsolete and unnecessarily simplistic.

But replacing it with “new research proves high cholesterol doesn't matter” simply substitutes one oversimplification for another.

The emerging picture is considerably more interesting.

LDL and ApoB matter. But the person carrying those particles, their genetics and metabolism, their other cardiovascular risks, how long they have been exposed, and what is actually happening inside their arteries matter enormously too.

That isn't the collapse of the cholesterol hypothesis. It is cardiovascular science becoming more precise.