Mind & Body · Wednesday, 9 September 2026

01 Briefing what happened

Cholesterol cannot dissolve in blood, so it travels in packets - and what damages an artery is how many packets there are, not how much they carry

Every bit of cholesterol in you rides inside a protein-wrapped packet, because fat does not mix with water. The standard blood test weighs the cholesterol inside those packets, and what lodges in an artery wall is the packet itself. In March a rewritten US guideline added the tests that count them. On 4 September the drug most people expected to be the next statin failed.

1 in 5

people worldwide carrying a high lipoprotein(a) level, the packet type set by genes

it stays largely stable through life [16], and ordinary cholesterol drugs barely shift it [17]

39.8% to 54.3%

of US adults aged 30 to 79 who qualify for a statin, under the 2018 rules and then the 2026 ones

that is 11.2 million more people, counted across a survey population of 77.8 million [32]

1 in 311

people born with familial hypercholesterolemia, an inherited condition that sets cholesterol very high from birth

found and treated early it cuts the chance of coronary artery disease by about 80% [39]

4 of 66

side effects printed on a statin label that the drug actually causes

the other 62 turned up just as often in people swallowing a dummy pill, across trials of more than 120,000 [24]

The lead story — what happened

Cholesterol is a fat and blood is mostly water, so none of it dissolves. Every bit travels wrapped in a coat of protein, in packets called lipoproteins. [2][3]
Most of the cholesterol in you was made by you. About 80% is built inside the body, mostly by the liver, and about 20% comes from what you eat. [3]
Every packet that can damage an artery carries exactly one copy of a protein called apolipoprotein B, shortened to apoB. Count the apoB and you have counted the packets. [1][4]
The standard cholesterol test does not count packets. It reports the weight of the cholesterol inside them, and the LDL figure on most forms is worked out with a formula rather than measured directly. [1]
The two answers disagree more often than people assume. Among 411,125 Korean adults, apoB varied widely at every LDL level, and the gap was widest in people whose blood sugar and blood pressure were poorly controlled rather than in people who were simply heavy. [5]
When they disagree, the count is the one that follows the risk. Among 375,544 UK adults with no heart disease, those whose apoB sat unexpectedly high had 11% more heart attacks and strokes, and those whose apoB sat unexpectedly low had 13% fewer. [6] A 20-year study of 3,042 adults in Athens and an imaging study of 121 statin-treated patients in Romania point the same way. [7][8]
A packet does not block a pipe from the inside like scale in a kettle. It crosses the artery's inner lining, gets trapped in the wall, is chemically altered there, and the immune cells that arrive to clear it become part of the swelling. [3][69]
On 13 March 2026 the two US bodies whose cholesterol advice most doctors follow rewrote it for the first time since 2018. The American College of Cardiology and the American Heart Association added three tests: apoB, a once-in-a-lifetime lipoprotein(a) reading, and a scan for calcium in the heart's arteries. [10][11][12][37]
Lipoprotein(a), said aloud as L-P-little-A, is an LDL packet with an extra protein stuck to it. About one person in five carries a high level, it is set almost entirely by genes, and eating well does not lower it. [10][14][1]
The rewrite moves the starting line from age 40 to age 30, with treatment considered once LDL reaches 160 mg/dL in someone who has no heart disease yet. [13]
That widens the group who qualify for a statin from 39.8% to 54.3% of US adults aged 30 to 79, about 11.2 million more people. [32]
Then the biggest bet on the idea failed. Novartis said on 4 September 2026 that pelacarsen, its lipoprotein(a) drug, did not work, and released no details. [23]

Who is involved

The liver

the organ that makes most of your cholesterol; about 80% of the cholesterol in you is built there rather than eaten [3], and the main way any of it leaves the body is the liver pushing it into bile [64]
The American College of Cardiology and the American Heart Association

the two US bodies whose cholesterol guideline most doctors follow; they rewrote it on 13 March 2026, the first change in eight years [12][11]
Roger Blumenthal

a cardiologist at Johns Hopkins University in the US who chaired the committee that wrote the new guideline [12][11]
Novartis

the Swiss drug company whose lipoprotein(a) drug pelacarsen was widely expected to be the next statin, and failed on 4 September [23]

How it unfolded

1960s US dietary advice starts telling people to limit the cholesterol in their food [41]
2015 that advice is dropped; the guidelines stop mentioning any limit on cholesterol in food [41]
Feb 2026 a review of blinded trials in over 120,000 people finds only 4 of 66 listed statin side effects are caused by the drug [24][25]
13 Mar 2026 the US guideline adds apoB, lipoprotein(a) and a calcium scan, and moves the starting age down to 30 [11][13]
28 Aug 2026 one injection cuts lipoprotein(a) by 95 to 97% and holds it down for 48 weeks in a first-in-human trial [22]
4 Sep 2026 pelacarsen, the leading lipoprotein(a) drug, fails its outcomes trial [23]

Where this points

Watch for the reason pelacarsen failed, because whether a few years of lowering can undo fifty years of build-up decides what every other lipoprotein(a) drug is worth.

What is pushing on the whole day

The bar and the word are our reading of how hard each one is pushing today. The arrow is where it is heading. The evidence is in the stories below.

Tests nobody orders High→

guidelines have advised one lifetime lipoprotein(a) reading for years, and it is still rarely ordered even in people known to be at risk [16]. Nearly 90% of people found to carry inherited high cholesterol would never have been sent for the test that finds it [38]

Fear of statins High↓

blinded trials in over 120,000 people found only 4 of the 66 listed side effects are caused by the drug [24]. A survey of 261 doctors found their own worries about muscle aches and liver enzymes shape what they prescribe [31]

Damage already banked High↑

a 2026 review names cumulative exposure over the years, not today's level, as a key determinant of risk [34]. Arteries start hardening in late childhood and cause no symptoms until a vessel is over 70% narrowed [3]

Drugs arriving before answers Building↑

a single injection cut lipoprotein(a) by 95 to 97% for 48 weeks in a first-in-human trial reported on 28 August [22]. A week later the leading drug aimed at the same target failed its outcomes trial [23]

Diet advice pulled two ways Building↑

a review of 17 randomised trials found cutting saturated fat helped only people already at high risk, and landed in a political fight over the delayed US dietary guidelines [45]. A 2025 trial found saturated fat moved LDL while egg cholesterol barely did [40]

The rest of the day

43 more stories on this beat.

Each with its own sources. None of these is a link to the story above.

02
The lipoprotein(a) drug meant to be the next statin failed

Novartis announced on 4 September 2026 that pelacarsen had failed, and released no further detail. The drug blocks the making of lipoprotein(a), a cholesterol packet whose level is set by genes and which no ordinary treatment lowers. Cardiologists had expected it to protect the people who do everything right and have heart attacks anyway. David Maron, who leads the American Society for Preventive Cardiology, said only: oh no, oh no. [23]

Why it matters — Every study short of a randomised trial said lowering lipoprotein(a) would prevent heart attacks, and the randomised trial said it did not. That leaves one person in five with a number they have just been told to measure and nothing proven to do about it.
03
One injection, and the number stays down a year

A first-in-human trial run by the Cleveland Clinic in the US tested Kylo-11, a drug that silences the gene for lipoprotein(a) rather than mopping up the particle afterwards. The highest doses cut levels by about 95 to 97%, and the reduction was still holding 48 weeks later, from a single injection. The results were presented on 28 August 2026 at the European Society of Cardiology meeting in Munich and published in The Lancet. [22] It is one of a family of gene-silencing drugs that can cut lipoprotein(a) by more than 90%. [15]

Why it matters — It landed a week before pelacarsen failed. The field now has a drug that lowers the number beautifully and no proof that lowering the number helps anyone.
04
Thirty years of women's blood, and where the line falls

Researchers followed 27,748 US women who were healthy in 1993, measured their lipoprotein(a) at the start, and counted heart attacks and strokes until 2023. Levels above 30 mg/dL raised the 30-year risk of major cardiovascular events and of coronary heart disease. Only above 120 mg/dL, the top 1%, did the risk of ischaemic stroke and cardiovascular death rise clearly, at 63% higher risk of cardiovascular death than the lowest group. [18]

Why it matters — It shows the threshold problem in one dataset. Where the line is drawn decides whether one woman in four gets flagged or one in a hundred.
05
Six-year-olds are already carrying it

Slovenia screens children for inherited high cholesterol, and 1,418 of them, median age six, had their lipoprotein(a) measured too. A quarter of those with a confirmed gene fault had raised lipoprotein(a), and so did about a third of those with high cholesterol but no fault found. In the confirmed cases, about a third of the lipoprotein(a) reading was feeding into the child's LDL figure, because the two packets look alike to the test. [19]

Why it matters — A six-year-old with both risks stacked is already accumulating damage, and the reading their doctor sees is quietly counting one problem as the other.
06
Some packets do more harm, one for one

A 2026 review compared the three families of packets that carry apoB: ordinary LDL, triglyceride-rich packets, and lipoprotein(a). Genetic evidence suggests the last two are several times more damaging per particle than LDL is. But LDL is far more abundant than either, so it still supplies most of the total harm. The authors conclude that lowering LDL stays the main lever even though it is not the nastiest particle. [20]

Why it matters — LDL is not the most dangerous particle, and it still causes most of the damage, because there is so much more of it. Treatment has to follow the harm, not the ranking.
07
A test recommended for years and still not ordered

A 2026 review notes that guidelines from the European Atherosclerosis Society and others have recommended at least one lifetime lipoprotein(a) measurement for years. It is still rarely done, even in people already known to be at cardiovascular risk. The authors argue that leaving it out means those people's risk is underestimated, and that drugs reaching late-stage trials are what finally changed the calculation. [16]

Why it matters — The reason to measure was never how dangerous the thing was. It was whether anything could be done about it.
08
Two labs, two answers, same blood

A 2025 review sets out a problem sitting underneath every recommendation to test: lipoprotein(a) assays are not standardised, so different laboratories can return different numbers for the same sample. Results come in mg/dL in some countries and nmol/L in others, and converting between them is not exact. The same review reports that both very high and very low lipoprotein(a) have been linked to worse survival, which nobody can yet explain. [17]

Why it matters — A test being added to routine care worldwide still returns an answer whose scale depends on which machine ran it.
09
Sixty-two of 66 statin warnings do not hold up

Researchers at Oxford University in England, funded by the British Heart Foundation, pooled blinded trials covering more than 120,000 people and checked every side effect printed on statin labels. Of 66 listed effects, only four showed any link to the drug: liver test changes, minor liver abnormalities, urine changes and tissue swelling. Memory loss, depression, disturbed sleep, weight gain and impotence were reported just as often by people swallowing a dummy pill. [24][25]

Why it matters — Millions of people stopped a drug that prevents heart attacks because of a list of warnings the trials do not support, and the authors want the leaflets rewritten.
10
Fear, not the drug, ends most statin courses

A comment published in The Lancet alongside that review makes the point plainly: fear of muscle pain, tiredness, disturbed sleep and fuzzy thinking is a major driver of people stopping statins. The symptoms themselves are real experiences. What the blinded trials show is that they arrive at the same rate whether the pill contains the drug or nothing at all. This is the nocebo effect, the mirror of placebo, in which expecting harm produces it. [26][70]

Why it matters — Naming the mechanism is not calling anybody a liar. It is saying that a real symptom can have a cause other than the thing it followed.
11
150,405 people, and no effect on thinking

One of the most persistent statin fears is memory loss. Researchers pooled 42 randomised trials covering 150,405 people taking statins, ezetimibe or PCSK9 inhibitors, three different ways of lowering cholesterol. The rate of neurocognitive problems was 0.99 times the rate on placebo, meaning no difference at all. Splitting it by drug class changed nothing: statins 0.94, ezetimibe 1.11, PCSK9 inhibitors 1.00, every result crossing the line of no effect. [27]

Why it matters — The brain does need cholesterol, and it makes its own, which is why lowering the amount in the blood does not appear to reach it.
12
The muscle damage that is real

An 83-year-old man was put on a high dose of atorvastatin after a heart attack. He developed progressive weakness in his hips and shoulders, exhaustion and repeated falls, and was at first treated as simply frail. Doctors excluded nerve, hormone and blood-vessel causes, stopped the drug, and he recovered with physiotherapy. The case was published in 2026 as a warning about how statin muscle injury hides in older people. [28]

Why it matters — Mostly not the drug is a statement about a population. It is not a statement about the person in front of you, and in this man it was the drug.
13
Women stop statins more often

A 2025 review of the adherence literature found that women stay on statins less than men do. They are more likely to stop or switch because of side effects, less likely to believe statins are safe or effective, and more often recorded as statin intolerant. The review also reports more muscle symptoms and more new diabetes among women, and greater susceptibility to nocebo effects, with risk rising with age and dose. [29]

Why it matters — Two explanations point in opposite directions, more real harm or more expectation of harm, and the review does not settle which.
14
A trial built only from people who could not take statins

CLEAR Outcomes enrolled 13,970 people at high cardiovascular risk who could not tolerate statins, and gave half of them bempedoic acid and half a dummy pill. Nearly half the participants were women. Muscle pain occurred in 5.4% of women on the drug against 5.9% on the dummy, and in 5.8% of men against 7.6%. More people quit the dummy pill than quit the real one. [30]

Why it matters — It is the cleanest available test of statin intolerance, and in a group chosen for muscle pain the drug produced no more of it than nothing did.
15
The prescriber's own worries

A survey called STATRIP asked 261 doctors, almost all family doctors and internal medicine physicians, what they believe about statins. They reported clear concerns about interactions with other medicines, muscle aches, raised liver enzymes and stomach problems. The survey was designed to explain why so many patients never reach their cholesterol targets, and it concluded that a substantial part of the reason sits with the treating doctor rather than the patient. [31]

Why it matters — The gap between what trials show and what gets prescribed runs through a person who read the same leaflet the patient did.
16
Eleven million more people become eligible

Researchers applied both the 2018 and the 2026 US cholesterol guidelines to a nationally representative health survey of adults aged 30 to 79. Statin eligibility rose from 39.8% to 54.3%, a relative increase of 36.2%, or 11.2 million additional adults out of a weighted population of 77.8 million. That happened even though the new risk calculator, called PREVENT, generally estimates lower absolute risk than the one it replaced. [32] A companion review sets out the rest of the direction: calcium scans, combination drugs, and targets as low as 30 mg/dL in the highest-risk patients. [36]

Why it matters — A calculator predicting less danger produced more prescriptions, because the guideline changed what counts as a reason to treat, not only the arithmetic.
17
Europe rewrote its version first

The 2025 focused update of the European guidelines, written jointly by the European Society of Cardiology and the European Atherosclerosis Society, adopted risk calculators called SCORE2 that count non-fatal events as well as deaths and run to age 89. The authors say this corrects a long-standing underestimate of risk in women and in younger people. The update also drops the old step-by-step approach after a heart attack, in favour of a strong statin plus ezetimibe immediately. [33]

Why it matters — Two continents rewrote the same advice within a year and both moved the same way: earlier, and at more people.
18
The word for what is really being measured is years

A 2026 paper on when doctors should escalate treatment states the shift plainly: beyond hitting a target today, cumulative exposure over time has emerged as a key determinant of risk. It argues for early, intensive and sustained lowering rather than a cautious ladder of steps. It also reports the size of the gap between advice and practice: fewer than one in three people who have already had a cardiovascular event reach the recommended level. [34]

Why it matters — The guidelines have started counting years rather than readings. The practice underneath them has not caught up.
19
Taiwan names the half-tolerant

The 2026 consensus statement from the Taiwan Society of Lipids and Atherosclerosis introduces a category it calls suboptimally tolerable statins. It covers people who cannot hold the recommended dose because of side effects, real or perceived, even when they do not meet the formal definition of statin intolerance. It then sets out a tiered plan, starting with ezetimibe and bempedoic acid in the highest-risk patients and escalating to PCSK9-blocking drugs if targets are still missed. [35]

Why it matters — Most people who quit a statin are not intolerant by the textbook definition, so the textbook definition was leaving them uncounted and untreated.
20
Most inherited high cholesterol is never found

Familial hypercholesterolemia passes silently through families for generations and is highly treatable. A Mayo Clinic study published in Circulation: Genomic and Precision Medicine ran DNA testing across a research population and compared who it found with who current screening rules would have flagged. Nearly 90% of the people carrying the condition would never have been sent for genetic testing, and none of them knew they had it. About one in five had already developed coronary artery disease. [38]

Why it matters — The screening rules use cholesterol level and family history, and this is exactly what they miss: the families where nobody knew either.
21
What the inherited version costs when missed

About one person in 311 has familial hypercholesterolemia. Their LDL runs above 190 mg/dL as adults and above 160 mg/dL as children, and eating well and exercising are usually not enough to bring it down. Found early and treated, the chance of coronary artery disease falls by around 80%. For a child diagnosed with it, statin treatment may need to begin at age eight to ten. [39]

Why it matters — Treatment starting at eight is the clearest statement anyone makes that this harm is counted in years rather than in readings.
22
The egg was never the main problem

US dietary advice told people to limit the cholesterol in their food from the 1960s onwards, and in 2015 the guidelines dropped every mention of it. A trial published in the American Journal of Clinical Nutrition in July 2025 separated the two things an egg breakfast usually brings: the cholesterol in the yolk, and the saturated fat sitting next to it on the plate. The saturated fat moved LDL and the egg cholesterol barely did. [40][41]

Why it matters — For fifty years the advice named the ingredient that was easiest to point at, and the one doing the work was on the same plate.
23
The counter-finding, in stroke survivors

Not every study agrees. Researchers followed 1,367 US stroke survivors recruited between 1999 and 2018 and tracked deaths through to 2019. Each extra 100 mg of dietary cholesterol per 1,000 calories a day was linked to a 16% higher risk of dying from any cause and 15% higher from cardiovascular causes. Eating more than one egg a day was linked to a 40% higher risk of death from any cause. [42]

Why it matters — This is an observational study in people who have already had a stroke, which is a different question from what an egg does to a healthy person's LDL.
24
What eggs do to the numbers, honestly

A comprehensive 2025 review pulled together trials, cohort studies and country-level comparisons. Its summary is genuinely mixed. Recent meta-analyses find eggs raise cholesterol in the blood but have limited or no effect on atherosclerosis or cardiovascular disease. One extra egg a day for four weeks improved HDL and lowered oxidised LDL. Among Japanese patients sent for coronary imaging, three to four eggs a week tracked with less disease across multiple vessels. [43][44]

Why it matters — The number moves and the disease does not, which is the difficulty of this whole subject compressed into one food.
25
Saturated fat, measured and then politicised

A systematic review of 17 randomised trials, published in Annals of Internal Medicine in December 2025, found that cutting saturated fat lowered serious cardiovascular events, but only among people already at high risk. People at low to moderate risk saw no such benefit. The finding landed while the US health secretary, Robert F. Kennedy Jr., was promoting full-fat dairy, red meat and beef tallow ahead of national dietary guidelines that had been delayed. [45]

Why it matters — Both sides of a political argument found something they liked in the same paper, which is what happens when a result is genuinely conditional and the argument is not.
26
Good cholesterol was never that simple

HDL earned the name good cholesterol from observational studies going back decades showing that people with more of it had less heart disease. A 2025 review argues the quality of the particles matters more than the quantity. Inflammation, autoimmune disease and oxidative stress change HDL's structure, and altered HDL can lose its protective properties and start driving inflammation instead. How well HDL pulls cholesterol out of cells may matter more than how much of it there is. [47][2]

Why it matters — It is why every attempt to help people by raising their HDL number has come to nothing. The number was never the job being done.
27
Very high HDL tracked with more deaths, not fewer

A study of 3,758 adults having their lipids measured at a hospital sorted them into five HDL bands and followed them for a median of 5.9 years. The very high group, 90 mg/dL or more in men and 110 or more in women, had the highest cardiovascular death rate at 11.3 per 1,000 person-years and a 52% higher adjusted risk than the normal group. The authors describe a U-shaped curve. [46]

Why it matters — A separate study links higher HDL to an 18% higher rate of new fractures in men, another reason to stop reading the number as a score. [48]
28
The drug that raised the good number and harmed people

Torcetrapib raised HDL and lowered LDL, exactly as designed. Its maker ran a study in 15,000 people to see whether that prevented cardiovascular disease. The opposite turned out to be true: the drug increased it, and it was never approved. A German health institute now uses it as the standard example of why a treatment that improves cholesterol figures is not automatically good for the person swallowing it. [3]

Why it matters — It is the same shape as pelacarsen, twenty years earlier, which is part of why the cardiologists quoted this week sounded shocked rather than resigned.
29
The better number already on your form

Non-HDL cholesterol is total cholesterol minus HDL, which captures the cholesterol riding on every harmful packet rather than on LDL alone. It needs no extra blood test at all. A 2025 consensus by 11 experts across ten Asia-Pacific countries argues it estimates risk better than LDL and should be a primary treatment target, and notes that most guidelines in that region still treat it as secondary. [49][50]

Why it matters — One of the better answers to the counting problem has been sitting free on every lipid panel for decades.
30
The healthiest cholesterol level keeps moving

Danish researchers pooled three Copenhagen population studies covering more than 114,000 people who were not taking cholesterol drugs, and asked which LDL level carried the lowest death rate. In 1991 to 1994 it was 155 mg/dL. By 2010 to 2018 it had fallen to 135 mg/dL. Median LDL in the population fell over the same period too, from 142 to 124 mg/dL. [51]

Why it matters — The level associated with living longest is not a fixed fact of human biology. It moved with the population and with whatever else was being treated.
31
A pill that halves LDL without a statin

A newly approved tablet lowers LDL cholesterol by more than half. It is aimed at two groups: people who cannot tolerate statins, and people whose LDL is still too high despite the cholesterol drugs they already take. Harvard Health describes it as an option in those two situations rather than as a general replacement for statins. [52]

Why it matters — The non-statin shelf is now deep enough that not tolerating statins no longer has to mean going untreated.
32
Your brain makes its own and cannot borrow yours

The brain holds a large share of the body's cholesterol and needs it for cell membranes, for building synapses, for receptor signalling and for the insulation around nerve fibres. The blood-brain barrier blocks the packets circulating in your blood from getting in. So the brain runs a separate supply: astrocytes, the support cells wrapped around neurons, make cholesterol and hand it over to the neurons that need it. [54][56] HDL-like particles made inside the brain help hold the barrier itself together. [55]

Why it matters — It explains the neurocognitive trial result. Lowering cholesterol in the blood does not lower it in the brain, because the two supplies were never joined.
33
One lost night, and the insulation suffers

A 2026 study examined what losing sleep does to myelin, the fatty sheath that insulates nerve fibres and lets signals travel fast. It found slowed nerve conduction, poorer synchronisation between the brain's two halves, and worse performance on tasks. Profiling the cells that build myelin showed cellular stress and disrupted cholesterol handling. Pushing more cholesterol towards the sheaths prevented both the slowdown and the behavioural effects. [53]

Why it matters — It is a route from a bad night to slow thinking that does not run through tiredness at all.
34
The artery that almost never gets the disease

Surgeons doing heart bypasses prefer the internal mammary artery, which runs down the inside of the chest wall, because it resists atherosclerosis in a way the coronary arteries do not. A 2025 review asks why, given that both see the same blood and the same cholesterol in the same person. It points to that vessel producing higher levels of certain protective proteins, and to gut bacteria producing inflammatory chemicals that reach some arteries more than others. [60]

Why it matters — The same exposure with a different outcome inside one body is the strongest argument that what is in the blood is only half the story.
35
The risk statins do not touch

A 2026 review on why plaques tear argues that lowering cholesterol has cut heart attacks substantially and left behind what it calls residual inflammatory risk. It traces the mechanism through a signalling network called NF-kappaB, which controls inflammation in the artery wall, and argues that sustained activity in one branch of it pushes plaques towards rupture. The strategy of the field, the authors write, is shifting from lipid-lowering towards anti-inflammatory approaches. [62]

Why it matters — The next argument about heart disease will be about inflammation, and no number on the cholesterol panel measures it.
36
Some plaques grow their own blood vessels

A 2026 review examines why some plaques tear open and others sit quietly for decades. The dangerous ones are full of new, fragile blood vessels grown into them from outside. Those vessels relieve the oxygen shortage inside the plaque and weaken it at the same time. They also feed back into inflammation, cell death and tissue remodelling, so each process makes the others worse. [65]

Why it matters — A heart attack is usually a plaque tearing rather than a plaque closing, and what makes one fragile is not how big it is.
37
Medicines that push cholesterol the wrong way

Several widely used drugs raise cholesterol as a side effect. Beta blockers, prescribed for high blood pressure and various heart conditions, lower HDL. Prednisone and other steroids, given to damp down inflammation, raise cholesterol too. The effects are usually small and the benefits of the drugs usually outweigh them, which is why the advice is to change the dose or the drug only when the shift is large. [61]

Why it matters — A cholesterol reading is a snapshot of a body that is also taking other things, and the form it comes back on does not say so.
38
Cholesterol turns up in the cancer literature too

Cancer cells rebuild their cholesterol handling: making more, taking in more, and letting less out. A 2025 review links this to sustained growth signals, resistance to a form of cell death called ferroptosis, and spread through the body. [57] A 2025 paper on HMG-CoA reductase, the enzyme statins block, reports that cholesterol metabolism is switched on in most tumour growth and that this enzyme drives it. Statins show anti-tumour effects in the laboratory, at concentrations far too high to give a person. [58]

Why it matters — The reason a cheap heart drug keeps reappearing in cancer research is that tumours need the same raw material your cell membranes do.
39
Immune cells run on it as well

A 2026 review in the journal Immunity sets out how every immune cell handles cholesterol, and how the products of that handling act as signals rather than only as building material. That includes the molecule itself, its half-built intermediates, and its oxidised derivatives. Those signals shape how immune cells behave. The same work notes that oxidised derivatives help exhaust the T cells that fight tumours. [59][57]

Why it matters — It is the same molecule the artery story is about, doing an entirely different job two tissues away.
40
The way out runs through bile

The body's main route for getting rid of cholesterol is the liver pushing it into bile, and converting some of it into the bile acids that digest fat. [64] A 2026 mouse study examined a channel protein in liver cell mitochondria called aquaporin-8. Knocking it down reduced the transporter that pumps cholesterol into bile, and cut how much cholesterol left the body that way. [68]

Why it matters — The exit is narrow and still poorly mapped, which is part of why almost every drug works on how much cholesterol comes in rather than on how much leaves.
41
A minority report on the whole idea

A 2026 narrative review challenges what it calls the cholesterol hypothesis. It notes that cardiovascular disease went from rare to the leading cause of death in industrialised countries between 1920 and 1950, alongside changes in the food supply including refined carbohydrates, industrial seed oils and trans fats. It argues that Ancel Keys' focus on saturated fat and cholesterol crowded out other explanations, and that the sugar industry's documented influence on nutrition research shaped which questions got asked. [63]

Why it matters — It is a minority position. Against it sits a genetic study finding that each standard-deviation rise in lipoprotein(a) causally raises the odds of coronary artery disease by about 24%. [21]
42
The nocebo effect has a physiology

A 2026 chapter reviews what brain imaging shows about placebo and nocebo effects. Both are produced by learning and expectation rather than by imagination. Contextual cues shape them: how a treatment is delivered, what brand it carries, what it costs, and even whether the patient is told outright that it is a dummy. The chapter traces the chemistry of placebo pain relief, which runs partly through the body's own opioid system. [70]

Why it matters — It is why the trial says it was not the drug and your symptoms were real can both be true statements about the same person.
43
Where else the plaque goes

Atherosclerosis is not only a heart disease. When it narrows the arteries inside the skull the condition is called intracranial atherosclerotic disease, and it is one of the main causes of stroke. Less blood reaches the part of the brain that artery feeds, and a complete blockage causes an ischaemic stroke. The narrowing usually causes no symptoms at all until a vessel is more than 70% closed. [71][3]

Why it matters — The same process reaches the arteries inside the skull, and it stays just as quiet there, until it is far along.
44
Not a villain, a building material

Cholesterol is a structural part of every cell membrane you have, and the raw material your body uses to build steroid hormones, vitamin D and the bile acids that digest fat. A 2025 review sets out the four systems that keep the amount in balance: making it from scratch, absorbing it from food, converting it into other things, and clearing it away. Other work shows it steering which kind of cell a stem cell becomes, and shows brain support cells storing fats in droplets that feed nerve signalling. [64][67][66]

Why it matters — The molecule with the worst reputation in medicine is one no cell in you can do without, and nearly the whole argument is about one packet type in one place.

02 Lesson why it matters

Your cholesterol number is today's reading. The harm is that number added up over a lifetime.

Artery damage starts in childhood and builds for decades, so a reading taken at fifty says how fast, never how much is already there.

The twist

Lowering the number helps most in people whose number does not yet look alarming, because by the time a reading is alarming enough to act on, most of the building has already happened.

How it works

Packets push into the artery wall every day, starting in childhood
A few stick, and stuck ones are not taken back out
How much sticks depends on how many packets pass and for how many years
The blood test reports how many are passing this week
Nothing on the form reports how many years they have been passing
So the same number means one thing at 25 and another at 65

The same force, elsewhere today

Where this chain is also running, in today's other stories.

The lipoprotein(a) drug that failed

the same chain, two different totals - people born with low lipoprotein(a) get sixty years of it, and a drug started in middle age gives a few
Statins considered from age 30

the guideline moved the starting line earlier rather than the dose higher, which is the same step run forward: subtract years, not milligrams
Most inherited high cholesterol going undiagnosed

someone born with the level runs the total up decades early, which is why treatment can start at eight and why missing nearly 90% of them costs so much
Six-year-olds in Slovenia already carrying raised lipoprotein(a)

the packets are pushing into the wall at age six, long before anything on a form would look wrong

Where you've seen this

Sunburn

skin damage is the total of every hour in the sun, so burns at eight still count at fifty

Smoking

doctors ask how many years times how many packs, never how many cigarettes today

Radiation work

a badge records a running lifetime total, because the daily reading on its own says nothing

Loud work

hearing damage is loudness times hours, which is why a factory measures a whole shift

The catch

Starting late still helps. Statins lower risk in older people too, and a plaque can be made steadier even when it cannot be shrunk. Later buys less, it does not buy nothing.

And the whole of it

Nobody in this has your total. Your doctor has one reading, the guideline committee has averages from other people's decades, and the drug trial had five years. Everyone is working from the piece of the clock they can see, and so are you.

03 Truth what's really going on

What is really going on

The thing that damages an artery is how many cholesterol packets have pushed into its wall over a whole life, and nobody has ever measured that in anyone. So doctors use the things they can measure today, argue about which of them comes closest, and on 4 September Novartis found out the hard way that the closest one is still not the thing: pelacarsen lowered lipoprotein(a) and did not prevent heart attacks. [1][23]

Why it works on us — Calling one packet good cholesterol and another bad attaches a moral word to a delivery vehicle, and a moral word invites you to raise one and cut the other, which is why raising HDL was tried for years and never worked, most starkly with torcetrapib. [2][47][3]

Who gains

Makers of apoB and lipoprotein(a) blood tests — the March 2026 US guideline named both, and a lipoprotein(a) reading is now recommended once for every adult, in a country of tens of millions of adults who have never had one. [10][11]
Generic statin manufacturers — the same guideline widens eligibility from 39.8% to 54.3% of US adults aged 30 to 79, about 11.2 million more people, and every major statin type is off patent. [32][11]
Makers of the non-statin drugs — the Taiwan consensus and the European update both push ezetimibe, bempedoic acid and PCSK9-blocking drugs earlier, and a new tablet that halves LDL is aimed squarely at people who cannot take statins. [35][33][52]
Egg producers — US dietary guidance dropped every mention of limiting cholesterol in food in 2015, and the 2025 trial that separated egg cholesterol from saturated fat found the fat did the work. [41][40]
Children found by universal screening — Slovenia screens every child, which is how 1,418 with high cholesterol got a lipoprotein(a) reading at a median age of six; treating the inherited form early cuts the chance of coronary artery disease by about 80%. [19][39]

Who pays

The one person in five with high lipoprotein(a) — they have just been told to get tested, and the leading drug for it failed on 4 September, so a positive result now buys a number and no proven treatment. [14][23]
People who stopped a statin over a symptom the drug did not cause — 62 of the 66 effects on the label showed up just as often on a dummy pill, and women stop more often than men and are more often recorded as intolerant. [24][29]
People with real statin muscle injury — a message that the side effects are mostly not real makes the genuine cases harder to raise, and an 83-year-old was treated as simply frail until the drug was stopped and he recovered. [28][26]
Families carrying inherited high cholesterol who have never been told — screening rules based on cholesterol level and family history missed nearly 90% of carriers in the Mayo Clinic study, and one in five of those found had coronary artery disease already. [38]
People whose apoB is high while their LDL looks normal — the standard test reports the cholesterol weight and misses them, and in 411,125 Korean adults the gap was widest in those with poorly controlled blood sugar and blood pressure. [5][1]

What nobody knows yet

Open questions from across today’s stories — ours included.

01

Why pelacarsen failed.

Novartis announced the failure on 4 September 2026 and released no further detail, so nobody outside the company yet knows whether the drug lowered the wrong thing or whether a few years of lowering was never going to undo fifty. [23]
02

Whether measuring apoB actually changes what happens to people.

Every study so far shows apoB predicts risk better than LDL does, but predicting is not the same as improving, and the strongest support offered for routine use is a computer simulation of cost-effectiveness rather than a trial. [9][6]
03

What counts as a high lipoprotein(a).

The US health agency uses 50 mg/dL, the 30-year women's study found risk rising above 30 mg/dL for heart events but only above 120 mg/dL for stroke and cardiovascular death, and a 2025 review says the assays are not standardised, so two laboratories can disagree about the same blood. [14][18][17]
04

Whether eggs matter at all.

A 2025 trial found saturated fat moved LDL and egg cholesterol barely did, while a cohort of 1,367 stroke survivors found more than one egg a day linked to a 40% higher risk of death from any cause. The two are asking different questions and neither can settle the other. [40][42]
05

What the U-shaped curves mean.

Very high HDL tracked with 52% higher cardiovascular death in one hospital cohort, and the LDL level with the lowest death rate in Copenhagen fell from 155 to 135 mg/dL over three decades. Nobody can yet say whether low or high numbers cause the extra deaths or merely mark people who are already ill. [46][51]
06

Why one artery escapes the disease.

The internal mammary artery resists plaque while the coronary arteries of the same person do not, on the same blood. A 2025 review offers protective proteins and gut-produced inflammatory chemicals as candidates and settles neither. [60]
07

Whether starting statins at 30 pays off.

The 2026 US guideline moved the starting age from 40 to 30 on the strength of lifetime-risk modelling. No trial has followed anyone for the forty years that claim covers, and none is running. [13][32]
08

How much of any one person's statin intolerance is the drug.

Blinded trials show most listed side effects arrive equally on a dummy pill, and a published case describes an 83-year-old whose weakness and falls resolved when atorvastatin was stopped. Population data cannot tell an individual which of those they are. [24][28]

04 Hope carry this

A review of blinded trials covering more than 120,000 people checked all 66 side effects printed on statin labels and found only four the drug actually causes. The other 62, including memory loss and depression, turned up just as often in the people swallowing a dummy pill.

Also true today

Forty-two trials covering 150,405 people taking cholesterol-lowering drugs found no increase at all in memory or thinking problems.
A child found to have inherited high cholesterol and treated for it has about 80% less chance of developing coronary artery disease.
A single injection cut lipoprotein(a) by 95 to 97% in a first-in-human trial, and it was still down 48 weeks later.
In Copenhagen the middle LDL cholesterol level of the general population fell from 142 mg/dL in the early 1990s to 124 mg/dL by 2018.