Cholesterol and Statins

An Essay on the Most Successful Unfalsifiable Claim in Medicine

Dr. Malcolm Kendrick knew a man — a competitive cyclist, nearly at international level. Extremely thin. Resting pulse of 50 beats per minute. Blood pressure 120/70 — textbook normal. Total cholesterol 3.0 mmol/l, which is very low. Vegetarian. Non-smoker. He was 36 years old when he had a heart attack.²

By every metric the cholesterol hypothesis uses to predict heart disease, this man should have been invulnerable. He had none of the risk factors. His cholesterol was a number most cardiologists would celebrate. He did everything the guidelines told him to do, and his body did everything the guidelines said it should. The hypothesis predicted that people like him don’t get heart disease. He got heart disease.

The system’s response: genetic susceptibility. Not a new study. Not a re-examination of whether the risk factors are actually predictive. A label. The man had a heart attack without risk factors, so he must be genetically susceptible. How do we know he’s genetically susceptible? Because he had a heart attack without risk factors. The reasoning is circular, and it performs a specific function: it prevents the case from counting as evidence against the hypothesis.

In my previous essay, Unfalsifiable, I examined the structure of scientific claims that have been made immune to contradiction — claims where no result, not even a directly falsifying one, is allowed to count against them.¹ The cases were drawn from genetics and virology. That essay laid out the architecture. This one furnishes the building’s largest room.

Unfalsifiable

The cholesterol hypothesis — the claim that cholesterol in the blood causes heart disease and that lowering it with drugs prevents death — is the most commercially successful unfalsifiable claim in the history of medicine. It has survived every contradiction it has encountered, not by answering the contradiction, but by shapeshifting around it. Kendrick, who has spent decades studying this hypothesis and its mutations, compared it to a monster from a 1950s horror film: every time you think you’ve killed it, it gets back up and carries on.²

Kendrick is a follower of Karl Popper, and says so explicitly.² He understands that the strength of a scientific claim lies not in how many studies support it, but in whether it has specified what evidence would count against it — and survived. One black swan refutes the claim that all swans are white, regardless of how many white swans you’ve counted. The cholesterol hypothesis has encountered black swan after black swan. It has responded by reclassifying each one as a different kind of bird.

What follows is a catalogue of those reclassifications — the specific moves by which the cholesterol hypothesis has made itself immune to falsification. Kendrick’s frustration with the cyclist case captures the pattern: “If you are going to suggest that people are genetically susceptible to heart disease, then you also have to attempt to explain the mechanism. Otherwise, you might as well believe in magic. ‘Abracadabra, genetics — heart attack.’”²

Support Independent Investigative Journalism and Research

This work remains free because paid subscribers make it possible. If you find value here, consider joining them.

What paid subscribers get: Access to all books I’ve written plus 1-2 new ones per month. e.g.

The DMSO Book
The Kitchen Remedies Guide
Chlorine Dioxide
The PSA Trap
Breast Cancer
Drilling for Profit
What Your Vet Can’t Tell You

Plus: Access to the Deep Dive Audio Library — 180+ in-depth discussions (30-50 min each) exploring the books behind these essays. New discussions regularly added. That’s 100+ hours of content for less than the price of a single audiobook.

[Upgrade to Paid – $5/month or $50/year]

unbekoming@outlook.com

The Hypothesis That Won’t Hold Still

The cholesterol hypothesis started as something clean and testable. Kendrick traces its original form: if you eat too much cholesterol, the level of cholesterol in the blood rises, and cholesterol is then deposited on artery walls, causing them to thicken and narrow.²

Problem one arrived almost immediately. Cholesterol in the diet does not raise blood cholesterol levels. Ancel Keys himself proved this.² The body regulates its own cholesterol production — eat more, and it produces less; eat less, and it produces more.³ The dietary hypothesis was dead on arrival, killed by its own champion.

The adaptation: the claim shifted from dietary cholesterol to saturated fat. It was saturated fat that raised cholesterol levels. Keys built his case on the famous Seven Countries Study, which found a straight-line relationship between saturated-fat consumption, cholesterol levels, and heart disease across Italy, Greece, Yugoslavia, the Netherlands, Finland, the USA, and Japan.² The study had one critical feature that is rarely mentioned: Keys selected his countries. Data were available for many more. A different selection — Finland, Israel, Netherlands, Germany, Switzerland, France, and Sweden — would have demonstrated the exact opposite relationship.² Within Keys’s own data, populations living in close proximity showed wildly different heart-disease rates despite similar cholesterol levels: in Finland, Karelia had five times the heart-disease mortality of Turku, with nearly identical cholesterol.³ On the Greek islands, Corfu had five times the cardiac deaths of nearby Crete, despite lower cholesterol.³

Problem two: you don’t have a cholesterol level. You have lipoprotein levels. Cholesterol travels inside lipoproteins — LDL, HDL, VLDL — and calling LDL “cholesterol” is a linguistic sleight of hand that allowed the hypothesis to survive a category error.²

The adaptation: certain lipoproteins were renamed. LDL became “bad cholesterol.” HDL became “good cholesterol.” The terminology was inaccurate but effective. Nobody noticed the switch.

Then the contradictions to saturated fat began accumulating. Population after population turned up with high saturated-fat intake and low heart disease, or low saturated-fat intake and high heart disease. The hypothesis adapted again: it wasn’t saturated fat per se, it was the ratio of polyunsaturated to saturated fat. Then it was a lack of monounsaturated fats. Then it was the specific type of saturated fat — pork fat versus beef fat, grass-fed versus grain-fed, alpine cheese versus cheddar.²

James Black, lecturing in 1803, could have been describing exactly this process: “A nice adaptation of conditions will make almost any hypothesis agree with the phenomena. This will please the imagination, but does not advance our knowledge.”²

Each adaptation was presented as a refinement. Collectively, they constitute the signature of an unfalsifiable claim: a hypothesis that changes its shape every time a contradicting fact arrives, so that the fact can be absorbed rather than confronted. As Kendrick observed: “Any hypothesis that has to keep changing all the time to survive the relentless assault of contradictory facts is, in reality, a dead hypothesis.”²

The Paradox Factory

When a hypothesis predicts that high saturated-fat intake leads to high heart disease, and a population is found where the opposite is true, the honest scientific response is to question the hypothesis. The actual response, every time, has been to explain away the population.

The most famous case is France. The French eat more saturated fat than any other nation in Europe. They smoke more than the British, take less exercise, have the same cholesterol levels, the same blood pressure, the same rate of obesity. Their rate of heart disease is one-quarter that of the UK.² The second-lowest rate of heart disease in Europe belongs to Switzerland — the country with the second-highest consumption of saturated fat.²

The system’s response was to invent the French Paradox — a name that functions as a containment device. The word “paradox” implies that the observation is the anomaly, not the theory. Red wine, garlic, and lightly cooked vegetables were offered as explanations. These appeared in The Lancet.² None of them had supporting evidence from randomised controlled trials. None. They were conjured to keep the hypothesis alive.²

Kendrick’s 14-country data, drawn from World Health Organization figures, tells the same story at scale. The seven European countries with the lowest saturated-fat consumption all have significantly higher rates of heart disease than the seven countries with the highest saturated-fat consumption.² When he showed these graphs to a colleague, the colleague’s first response was: “This can’t be right, where did you get this rubbish from?” When shown the WHO source, his response became: “Well there obviously must be other factors involved.”²

That response — there must be other factors — is the engine of unfalsifiability. It can never be exhausted. By 1981, a paper in Atherosclerosis had already identified 246 factors implicated in heart disease.² Today, the number would exceed a thousand. With 246 factors available, any contradiction to any hypothesis can be explained away by invoking some combination of other factors. The hypothesis becomes computationally irrefutable — not because it’s correct, but because the number of possible explanations exceeds the human ability to test them.

Karl Popper had a name for this: ad-hoc hypotheses. Devices scientists use to explain away apparent contradictions to much-loved hypotheses.² The cholesterol hypothesis is their natural habitat.

Consider the pattern. The Inuit eat massive quantities of saturated fat and have almost no heart disease — explained by omega-3 fatty acids from fish. Emigrant Asian Indians in the UK have low saturated-fat intake and very high heart disease — explained by genetic predisposition to diabetes. Israel has one of the highest polyunsaturated-to-saturated fat ratios in the world, exactly what the hypothesis prescribed, and a high prevalence of cardiovascular disease — explained by the ratio of omega-6 to omega-3.² Switzerland’s heart-disease rate fell while saturated-fat consumption increased by 20 per cent — explained by the specific fatty acid profile of alpine grass-fed cheese.²

UK wartime rationing severely restricted saturated fat, eggs, cheese, bacon and milk for twelve years. Fruit and fish were freely available. The rate of heart disease nearly trebled.² The small decline in heart-disease deaths that proponents attributed to rationing actually began in 1939 — two years before rationing was introduced.² By the time rationing ended in the early 1950s, heart-disease rates were far higher than when it started.

Each of these populations should have falsified the hypothesis. Each was instead absorbed into it, wrapped in a new explanatory layer that required no supporting evidence and could never itself be tested. The monster doesn’t just survive the attack. As Kendrick observed, attacking it makes it stronger.²

The Ad-Hoc Hypothesis That Killed People

The paradox factory doesn’t just generate explanations for inconvenient populations. Sometimes it generates medical interventions — treatments prescribed to millions, based on a protective mechanism invented to patch the hypothesis rather than tested against reality.

Women have lower rates of heart disease than men. Under the cholesterol hypothesis, this requires an explanation — something must be protecting women from the effects of their cholesterol. The explanation offered was female sex hormones. Oestrogen protected the heart. Protection disappeared at menopause.

This idea was based on no evidence. Kendrick is emphatic on this point: there has never been a study showing that female sex hormones protect against heart disease in humans.² But the hypothesis existed for a structural reason — it was needed to explain why high cholesterol didn’t kill women at the same rate as men. Without it, the cholesterol hypothesis had a problem it couldn’t solve.

In 1963, a study of women who had undergone hysterectomies — half with ovaries removed, half with ovaries retained — found no difference in heart disease rates between the two groups.² The sex-hormone hypothesis should have died there. In 1977, a study in the BMJ concluded that “the idea that female sex hormones protect against coronary heart disease should probably be abandoned.”² In 1987, the New England Journal of Medicine reported that normal menopause was not associated with any increase in heart disease risk.²

The hypothesis survived all of this. By the 1990s, millions of women were being prescribed hormone replacement therapy to prevent heart disease — a treatment based on a protective mechanism that had been disproved every time it was tested.

Then the HERS trial — randomised, placebo-controlled, the first proper clinical test — reported its results. HRT increased the rate of heart disease.² The American Heart Association now recommends strongly against using HRT for cardiac protection.

Kendrick considers the sex-hormone hypothesis the most perfect example of a pure ad-hoc hypothesis in the history of medicine. It came into existence for one reason: to protect the cholesterol hypothesis from the female data. It was based on nothing. It was disproved repeatedly. It killed women. And when it finally died, another ad-hoc hypothesis — HDL “good cholesterol” — was wheeled into place immediately, with no more evidence behind it than the one it replaced.²

The EUROASPIRE study provides another demonstration. This European survey followed coronary patients and found that raised blood pressure, raised cholesterol, and low HDL were not associated with heart-disease mortality.² The researchers’ explanation: the lack of association “may be related to the extensive use of antihypertensive and lipid-lowering drugs in this cohort.” A sister paper from the same study reported the actual figures: 58 percent of patients had cholesterol above 5.5 mmol/l and over 50 percent had blood pressure above 150/90 — clear evidence that the drugs were either not being used extensively or not working.² Another instant ad-hoc hypothesis, contradicted by its own dataset.

What EUROASPIRE actually showed was that smoking, existing heart disease, and diabetes are true risk factors. Cholesterol level and blood pressure are not. The researchers did not consider this possibility.²

The Lourdes of the Cholesterol Believer

Familial Hypercholesterolaemia — FH — is the hypothesis’s last resort. When all other arguments weaken, defenders point to FH: a genetic condition that produces very high LDL levels and, in its most severe form, can cause heart disease in children. The argument seems irresistible. If people with extremely high cholesterol die young of heart disease, cholesterol must cause heart disease.

Kendrick calls FH the “Lourdes of the cholesterol believer” — the pilgrimage site for those whose faith is wavering.² It functions as an unfalsifiable claim through what he describes as a one-way evidence valve: only evidence that confirms the hypothesis is allowed through.

The mechanism is simple. People with FH who develop early heart disease come to medical attention. They are identified, studied, and added to the evidence base. People with FH who do not develop heart disease — and there are many — are never identified in the first place. They live normal lives, undiagnosed, invisible to the research literature.²

When researchers have actually looked for these invisible cases, the results have been striking. Dutch researchers traced FH family members back to 1800 and found that mortality was normal throughout the nineteenth and early twentieth centuries — in fact, lower than the general population during periods when infectious disease was prevalent. Even at its peak between 1935 and 1964, mortality was less than twice that of the general population.⁴ A South African study found a father with the FH gene who reached 84 with no heart disease, while his son — carrying the identical gene — required bypass surgery before 50.² Same gene, same cholesterol level, opposite outcome.

An Oxford study followed over 500 FH patients for several years. Between ages 20 and 59, about three percent died of coronary heart disease. Between ages 60 and 74, less than two percent died — fewer than in the general population.⁴ In the age group where heart disease is most common, FH patients fared better than everyone else.

Ravnskov summarises: 40 percent of people with familial hypercholesterolaemia lived a normal life span.⁴ The researchers concluded that environmental factors were more important than cholesterol levels in determining longevity. Hussey, reviewing the same literature, notes that FH individuals who develop heart disease exhibit risk factors associated with insulin resistance — elevated triglycerides, blood glucose, obesity, hypertension — not high LDL per se.⁵

The strongest argument against FH as proof that cholesterol causes heart disease is the one Kendrick makes plainly: most people who die of heart disease do not have raised LDL levels, and most people with raised LDL levels do not die of heart disease. How can a raised LDL level be the cause of heart disease when it is not present? How can it be present and not cause heart disease?²

The Trial That Cannot Fail

If the hypothesis itself is protected from falsification by ad-hoc explanations and one-way evidence valves, the clinical trials that supposedly test it are protected by a different mechanism: they are designed so that failure cannot register.

Ravnskov’s analysis of the TNT (Treatment to New Targets) statin trial is detailed and devastating.⁴ The trial started with 18,469 patients with evident coronary heart disease. Of these, 15,464 were deemed eligible — the first cut removed over 3,000 people for unstated reasons. The remaining patients were then given a test dose of atorvastatin, the drug being studied. This led to the exclusion of a further 5,462 patients — those who experienced adverse effects, died, had a vascular event during the test phase, or showed “lack of compliance.”

From the original 18,469 patients, only 10,001 — 54 percent — made it into the trial. Everyone the drug might have failed on, everyone who showed intolerance, everyone who got sicker, was removed before the experiment officially began. The trial population represented an unusually strong and healthy selection of heart disease patients. The drug was then tested on a group pre-screened for its ability to tolerate the drug.⁴

The results of the TNT trial showed almost identical total mortality in both groups: 5.6 percent in the low-dose group, 5.7 percent in the high-dose group. Fewer people died of cardiovascular disease in the high-dose group, but more died of other causes — causes the researchers did not specify, and about which they ignored requests for detail.⁴

The trial also reported only “non-procedure-related” heart attacks — excluding those that occurred during hospital operations or diagnostic investigations. If those excluded events had been more common in the low-dose group, the researchers would have reported them. Their omission tells you where they fell.⁴

In the IDEAL trial, which used less aggressive patient exclusion, almost 90 percent of participants had side effects and almost half of these were recorded as serious. The researchers dismissed this: the frequency was “as expected,” they said, because elderly coronary patients naturally experience common colds and minor injuries over five years. But common colds and minor injuries had never been reported in any previous trial, and cannot be classified as serious.⁴

Ravnskov’s conclusion is plain: “Taken together with the unwillingness to record obvious signs of organ dysfunction as side effects, the figures for statin side effects are obviously completely unreliable.”⁴

The architecture extends beyond individual trials. Every statin trial is sponsored by the company whose drug is being tested. The companies design the studies, control the data, perform the analyses, and decide whether to publish. Researchers sign contracts preventing publication without company permission. Dr. Marcia Angell, former editor of the New England Journal of Medicine, described the arrangement precisely: the drug company decides what the data show, what the conclusions are, and whether the results will even see daylight.²

Positive results are more financially rewarding than negative ones. Researchers who produce them are more often invited as speakers, more often chosen for future projects. The EXCEL trial — the first statin trial — showed higher mortality in the treatment group after just one year: 0.5 percent in the lovastatin groups versus 0.2 percent in the placebo group. Had the trial continued and the trend held, the difference would have become statistically significant. No final results were ever published.⁴

The blinding of statin trials is itself questionable. Patients knew their cholesterol levels at baseline. A simple cholesterol test during the trial — easily obtained outside the study — would reveal whether they were taking the drug or the placebo. Ravnskov notes that it is not unreasonable to assume a substantial proportion of patients and doctors knew which group they belonged to, and that such knowledge could unintentionally influence the results.⁴

Then there is the matter of how results are presented. The 4S trial, the most positive statin trial ever conducted, showed a relative risk reduction of 29 percent in total mortality. In absolute terms, that was a 3.3 percent reduction over five years. The chances of being alive at the end of the trial on placebo were 88.5 percent. On a statin: 91.8 percent.² Relative risk — the large, impressive number — is what appears in press releases. Absolute risk — the small, honest number — is what appears in the fine print, if it appears at all.

The system is designed so that statins cannot fail. Not because they work, but because the trials are structured to exclude failure, suppress adverse outcomes, and amplify marginal benefits through statistical presentation.

The Surrogate Shuffle

When a drug trial cannot show that people actually live longer, the endpoint gets changed to something measurable that isn’t survival. This is called a surrogate outcome.

Ravnskov describes the mechanism. Instead of counting deaths, researchers measure whether coronary arteries widen by a fraction of a millimetre on an angiogram, or whether a blood marker moves in the desired direction. A widening of the vessel is called “regression” and is attributed to disappearing plaques. The drug is then approved on the basis of this laboratory change, even though no one has demonstrated that the laboratory change translates to fewer deaths.⁴

When the National Heart, Lung and Blood Institute ran an angiographic trial with cholestyramine, the results were disappointing: coronary arteries widened a fraction of a millimetre in four treated patients — but also widened in four untreated patients. To extract a positive finding, the researchers used a one-tailed t-test, a statistical method that assumes the result can only go in one direction. Their justification: “the weight of laboratory and epidemiological evidence suggested that reduction of blood cholesterol would retard coronary artery disease.” The result could only go one direction because they decided in advance it could only go one direction.⁴

By the time the one-tailed test was deployed, seven controlled cholesterol-lowering trials had already resulted in an increase in coronary mortality in the treatment groups.⁴

The surrogate shuffle is the clinical-trial equivalent of the paradox factory. When the real endpoint — death — doesn’t cooperate, you measure something else and call it equivalent. When the surrogate doesn’t cooperate either, you adjust the statistics. The drug never fails. The endpoint just moves.

The Magic Word

When all other defences fail — when the paradoxes pile up, the trials disappoint, the surrogates don’t move — there is one final word that ends every argument: multifactorial.

Heart disease is multifactorial, we’re told. Which means that any single contradiction can be explained by the interaction of other factors. Kendrick, in The Clot Thickens, calculated the possible interactions from the 20 risk factors in the UK’s Qrisk3 calculator: 2.4 quintillion.⁶ With that many possible interactions, any contradictory finding can be argued away. The word “multifactorial” doesn’t explain anything. It makes explanation unnecessary.

The Qrisk3 calculator asks about age, sex, smoking, diabetes, cholesterol ratio, blood pressure, BMI, chronic kidney disease, rheumatoid arthritis, lupus, migraines, severe mental illness, antipsychotic medication, steroid use, atrial fibrillation, erectile dysfunction, family history, ethnicity, and postcode.⁶ Every one of these factors is associated with cardiovascular disease. The problem is explaining how such different things — a postcode, a psychiatric medication, a clotting disorder — all cause the same disease through the same mechanism. Under the cholesterol hypothesis, they can’t. Hence: multifactorial. The word is not an explanation. It is the absence of one, dressed in a lab coat.

The Framingham Confession

The Framingham Heart Study, launched in 1948, was supposed to establish the relationship between cholesterol and heart disease beyond doubt. Thirty years of data produced a finding that should have detonated the hypothesis.

For men above age 47, there was no difference in mortality between those with high cholesterol and those with low cholesterol. The study that was designed to prove cholesterol kills people found that, past the age when heart attacks actually begin, cholesterol level made no difference.³

More than 95 percent of all heart attacks occur in people older than 48.³ Cholesterol was predictive only in the age group where almost no one dies of heart disease, and non-predictive in the age group where nearly everyone does.

The study also found that for each 1 mg/dl drop in cholesterol, there was an 11 percent increase in coronary and total mortality.³ People whose cholesterol fell over the 30-year period died at higher rates than those whose cholesterol rose.

The THINCS group (The International Network of Cholesterol Sceptics) — of which Kendrick is a member — published a systematic review in BMJ Open examining LDL levels and mortality in people over 60. The finding: “High LDL-C is inversely associated with mortality in most people over 60 years.”⁶ People with high LDL lived as long or longer than those with low LDL. The paper was the most widely read article in BMJ Open for six consecutive months.⁶

In a French nursing home study published in The Lancet, the lowest mortality was at a cholesterol level of 7.0 mmol/l, and the highest mortality was at 4.0 mmol/l — a 5.2-fold difference.² At Yale, researchers following 997 elderly people in the Bronx found that about twice as many with low cholesterol had heart attacks compared to those with the highest cholesterol levels.³

The response to these findings has been silence, or absorption. The explanation offered: low cholesterol is a marker of underlying disease, so it’s not the low cholesterol that kills — it’s the disease behind it. But the Framingham data showed that coronary mortality, not just total mortality, increased as cholesterol fell.³ The ad-hoc explanation didn’t match the data it was invented to explain. It didn’t matter. The hypothesis moved on.

Kendrick describes the cumulative effect. The THINCS group published a second paper in 2018 stating that “the idea that high cholesterol levels in the blood are the main cause of CVD is impossible because people with low levels become just as atherosclerotic as people with high levels and their risk of suffering from CVD is the same or higher.” It became the second most downloaded paper from Taylor Francis, the world’s second-largest medical publisher.⁶

The effect on statin prescribing: nothing. Statin prescribing has continued to boom worldwide. It has never been higher.⁶

The Monster Under the Bed, Revisited

In Unfalsifiable, I used a children’s analogy: a friend who claims there’s a monster under his bed, and every time you test for the monster, the story changes to explain why your test didn’t work. The monster is invisible. Then it floats. Then it isn’t hungry. The monster is set up so that no test can ever catch it — not because the monster is clever, but because the story keeps changing.¹

The cholesterol monster has a name now. It was dietary cholesterol, then saturated fat, then the wrong ratio of fats, then the wrong type of saturated fat. It was total cholesterol, then LDL, then small dense LDL, then oxidised LDL. Every test that fails to find the monster is explained by a refinement of the story, and the story is always more complex than the last version — which makes it harder to test and easier to defend.

And every year, the monster medicine gets prescribed to more people.

Where the Thinking Goes When the Blob Dies

Kendrick reached a conclusion about the cholesterol hypothesis that is relevant to every unfalsifiable claim in science: “Eventually I came to realise that the transient satisfaction of giving the cholesterol hypothesis a damn good kicking was merely a distraction. It was never going to get us anywhere.”⁶

Facts could not kill it. Paper after paper, study after study, contradiction after contradiction — the hypothesis absorbed them all. Kendrick compared it to the Terminator: “After you have blown it into smithereens, the hypothesis simply re-forms and starts chasing after you again.”⁶

What was needed, he realised, was not more demolition but an alternative hypothesis. Somewhere else for the thinking to go.

That alternative has existed since the 1850s. Karl von Rokitansky proposed that atherosclerotic plaques are the remnants of blood clots formed on damaged artery walls — not deposits of cholesterol leaking in from the blood, but the end-products of a clotting and repair process.⁷ He called it the encrustation hypothesis. It was taken up and refined by researchers including Duguid, Ross, Smith, and most recently Kendrick himself, who calls it the thrombogenic hypothesis.

The process Kendrick describes is straightforward: the endothelium (the lining of the artery wall) is damaged. A blood clot forms over the damage, the way a scab forms on skin. New endothelial cells grow over the clot, drawing it into the artery wall. Normally, the clot is then broken down and removed. But if the damage is repeated, or the clots are larger and harder to clear, or the repair systems are impaired, the remnants accumulate. Over years, they become plaques.⁷

This model explains what the cholesterol hypothesis cannot: why so many different risk factors — smoking, diabetes, high blood pressure, chronic kidney disease, rheumatoid arthritis, stress, lupus, antipsychotic medication, steroid use — all lead to the same disease. They all damage the endothelium or impair the clotting and repair process. They don’t need to be connected through cholesterol, because cholesterol was never the mechanism. The mechanism is damage and clotting.⁷

The thrombogenic hypothesis is falsifiable. It predicts that anything that damages the endothelium or increases blood clotting should increase cardiovascular disease. It predicts that anything that protects the endothelium or improves clot clearance should reduce it. These predictions can be tested, and every risk factor Kendrick examines — he works through dozens — fits the model.⁷

This is where the 20 factors in the Qrisk3 calculator — the ones that, under the cholesterol hypothesis, required the magic word “multifactorial” — suddenly resolve into a single process. Smoking damages the endothelium. Diabetes damages the endothelium. Rheumatoid arthritis and lupus cause vasculitis — inflammation that strips the endothelial lining. Chronic kidney disease impairs the clotting and repair system. Steroid use raises cortisol, which activates the renin-angiotensin system and damages artery walls. Antipsychotic medications do the same. Even postcode — a proxy for deprivation, stress, and toxic exposure — maps onto a process of repeated endothelial damage and impaired repair.⁷

The cholesterol hypothesis needed 246 ad-hoc explanations and the word “multifactorial” to account for these connections. The thrombogenic hypothesis needs one mechanism.

Unfalsifiable claims don’t just block the evidence against them. They block the alternatives. When the cholesterol hypothesis occupies the entire conceptual space, there’s no room for questions about endothelial damage, clotting factors, or the thrombogenic process. Funding flows toward cholesterol research. Careers are built on cholesterol research. The journals publish cholesterol research. An alternative that explains the data better cannot gain traction because the unfalsifiable claim has colonised the infrastructure.

Kendrick quotes the environment directly: researchers were “intentionally discouraged from pursuing alternative questions.” To challenge the cholesterol hypothesis publicly was to be labelled a denialist. An article in the Annals of Internal Medicine in 2017, by one of the world’s most influential cardiologists, was titled: “Statin Denial: An Internet-Driven Cult With Deadly Consequences.”⁶

The language of heresy. Not the language of science.

Why This Matters

Hundreds of millions of people worldwide take statins.² The best statin trial ever conducted — the 4S trial, a secondary prevention study — showed a 3.3 percent absolute reduction in total mortality over five years, which translates to an average increase in lifespan of approximately two months over thirty years of treatment.² In primary prevention — healthy people without existing heart disease — no trial has shown that statins extend life by a single day.²

Meanwhile, statins block the production of coenzyme Q10, a substance essential to energy production in every cell, and found in particularly high concentrations in heart muscle. Q10 and cholesterol share a biosynthetic pathway — block one, and you block the other. Merck knew this when they launched lovastatin in 1987. They applied for a patent combining their statin with Q10 supplementation.² They never marketed the combination. The biochemist Karl Folkers, who first described Q10’s molecular structure, reported in 1990 that lovastatin lowered Q10 concentrations and that heart function declined as a result.⁴ If statins impair the energy production of the very organ they’re supposed to protect, the implications extend to every patient on long-term therapy.

Heart failure is not reported as a side effect in statin trials, partly because patients with heart failure are routinely excluded before the trials begin, and partly because heart failure in elderly patients is attributed to their underlying condition rather than their medication.⁴ The question of whether the sharp rise in heart failure deaths over recent decades — during the same period that statin use expanded massively — has any connection to the drug is one that nobody is funding.

Statin side effects are systematically underreported. In Rhode Island, a survey of practicing doctors found that serious side effects reported to the FDA corresponded to only one percent of the numbers actually seen.⁴ In Ontario, two-thirds of heart disease patients and three-quarters of those prescribed statins solely for high cholesterol had stopped taking them within two years.⁴

Astronaut and physician Duane Graveline experienced two episodes of total amnesia on Lipitor. The first erased six hours of memory. The second erased everything after high school — his medical training, his career at NASA, his marriage, his children. Every doctor he consulted denied any connection to the drug. When he published his experience, hundreds of patients contacted him with similar stories — memory loss, confusion, disorientation — all associated with statin treatment. The FDA’s response, according to Graveline: “The subject is still being reviewed.”⁴

Cholesterol is the second most concentrated substance in the brain, after the adrenal glands. It is essential to the membranes of brain cells, to nerve fibres, and to the synapses that connect them. The rate of cholesterol synthesis is extremely high in the developing brain, which is why children with Smith-Lemli-Opitz syndrome — a genetic error that produces extremely low cholesterol — suffer severe brain malformations.⁴ Statins cause cancer in laboratory animals.² A low cholesterol level is associated with a higher risk of death from cancer.² Several statin trials showed more deaths from violence and suicide in the treatment groups — individually not statistically significant, but all pointing in the same direction.³ The system’s response to each of these signals has been consistent: coincidental, expected, unrelated.

None of this has altered the trajectory. The cholesterol hypothesis is kept alive, as the THINCS researchers wrote, “by reviewers who have used misleading statistics, excluded the results from unsuccessful trials and ignored numerous contradictory observations.”⁶

The test of a theory is not whether it can survive every challenge. It is whether it has specified what would constitute a challenge — and faced it honestly.

In 2004, the US National Cholesterol Education Program issued guidelines so aggressive that millions of additional Americans would require statins to comply. What the Program did not disclose was that most panel members who wrote the recommendations had financial ties to the pharmaceutical companies that stood to gain from the new prescribing targets. The Washington Post reported the hidden conflicts. Critics demanded disclosure. The guidelines remained unchanged.²

By that measure, the cholesterol hypothesis has not been proven strong. It has been made unfalsifiable. And the difference between the two is the difference between a scientific claim and a commercial franchise.

---

References

¹ “Unfalsifiable: An Essay on Theories That Can Never Be Wrong,” Lies Are Unbekoming, March 17, 2026.

² Malcolm Kendrick, The Great Cholesterol Con: The Truth About What Really Causes Heart Disease and How to Avoid It (London: John Blake Publishing, 2008). Kendrick explicitly identifies as a follower of Karl Popper; see also Popper, The Logic of Scientific Discovery (London: Hutchinson, 1959).

³ Uffe Ravnskov, The Cholesterol Myths: Exposing the Fallacy That Saturated Fat and Cholesterol Cause Heart Disease (Washington, DC: NewTrends Publishing, 2000).

⁴ Uffe Ravnskov, The Cholesterol Myths, supplemented by statin trial analyses discussed in Malcolm Kendrick, The Clot Thickens: The Enduring Mystery of Heart Disease (Columbus Publishing, 2021).

⁵ Stephen Hussey, Understanding the Heart: Surprising Insights into the Evolutionary Origins of Heart Disease — and Why It Matters (Chelsea Green Publishing, 2022).

⁶ Malcolm Kendrick, The Clot Thickens: The Enduring Mystery of Heart Disease (Columbus Publishing, 2021). The THINCS papers referenced are: Ravnskov et al., “Lack of an association or an inverse association between low-density-lipoprotein cholesterol and mortality in the elderly: a systematic review,” BMJ Open (2016); and Ravnskov et al., “LDL-C does not cause cardiovascular disease: a comprehensive review of the current literature,” Expert Review of Clinical Pharmacology (2018).

⁷ Malcolm Kendrick, The Clot Thickens (2021), drawing on the work of Karl von Rokitansky, James Duguid, Russell Ross, and Elspeth Smith.