In August 1990, a chemist named Duane Burnett walked into John Clader's office in New Jersey to discuss the results of their failed experiment.
Together, they had spent months building molecules to block a specific enzyme — the one everyone was certain was the key to lowering cholesterol.
Now, cholesterol itself isn't the villain. Almost every cell in your body can make it, and you would die without it. The trouble is the excess that ends up circulating in your blood — because that's what works its way into the walls of your arteries.
But the molecules Burnett and Clader had designed barely touched that enzyme. They were duds.
Buried in their data, though, was a shocking result.
Even though the molecule they had designed didn't work on the specific enzyme they were targeting, the animals' blood cholesterol had dropped. And by a significant amount. What was going on? Had they discovered the key to lowering cholesterol, and thereby heart disease, through a serendipitous moment? Like the discovery of penicillin?
Well, serendipity is a generous word for what they were actually facing. They had a molecule that lowered cholesterol, and no idea how [1].
Table of Contents
- Twelve Years, No Target
- Approved, and Still Unexplained
- They Finally Find the Lock
- The Trial That Ended It
- The Collapse, and Then a Long Silence
- The People Born With the Mutation
- Eighteen Thousand Patients
- What We Had Been Getting Wrong About Cholesterol
- A Smarter Way to Use It
- What Happens If You Aim Lower
- Where Ezetimibe Stands Today
- The Brain: A Maybe, Not a Reason
- Where I Land
- References
Twelve Years, No Target
There was no protein to design against. No structure to model.
But what follows led to a pill that is proven to lower heart disease rates, costs only $6, and yet barely anyone uses it.
Five years after the initial discovery, in the mid-1990s, that problem had landed with a scientist named Margaret Van Heek. She wasn't one of the chemists. She was one of the biologists on the team.
Van Heek's group dosed animals with the mysterious molecule, and then went looking for what came back out. Not the drug. The pieces of it. The fragments the animal's own body had broken it down into.
Some of those fragments were active in their own right.
So they built the next molecule to look like the fragment. Dosed animals with that. Looked again. Built the next one. The animal was doing the chemistry. They were copying it [2].
Van Heek's loop eventually spat out a molecule called SCH 58235. That one became the finished drug.
Here is the mechanism they had stumbled into. Your liver makes cholesterol. A lot of that cholesterol ends up as bile, which is pumped into your gut to help digest fats. Your gut then reabsorbs that cholesterol. Statins work by blocking the liver from producing excess cholesterol. The molecule Van Heek created works by blocking your gut from reabsorbing it. How it did that was still a mystery.
Approved, and Still Unexplained
The drug is called ezetimibe. On the 25th of October 2002, the FDA approved it. The sponsor at approval was Schering-Plough. The molecular target: unknown.
Hang on a minute. A drug went onto pharmacy shelves in the United States, was prescribed to millions of people, and nobody — not the FDA, not the company, not the chemists who had spent twelve years building it — could tell you which protein it worked on.
They knew it blocked cholesterol absorption in the gut. They could measure that. They just couldn't say what it was grabbing hold of to do it.
That is not how drug development is supposed to work. You're meant to find the target, then build the key that fits it. The ezetimibe team had built a key that opened a lock nobody had ever seen.
They Finally Find the Lock
And a much bigger problem was on the horizon.
The target protein, called Niemann-Pick C1-Like 1 (NPC1L1), was eventually discovered in 2004. Ezetimibe had no effect in mice bred without it, which suggested NPC1L1 sat in the pathway responsible for intestinal cholesterol absorption [3].
It was confirmed in 2005 as ezetimibe's target. As that paper put it, ezetimibe is a potent inhibitor of cholesterol absorption that had already been approved for treating high cholesterol — but its molecular target "has been elusive" [4].
What nobody had done yet was ask the only question that actually matters: does it stop people having heart attacks?
The Trial That Ended It
In January 2008, the answer came back, and it was brutal.
The trial was called ENHANCE. Seven hundred and twenty people with an inherited condition that gives them dangerously high cholesterol from birth. Half got a high-dose statin. Half got the same statin plus ezetimibe. Two years of follow-up.
The LDL-cholesterol result was exactly what you'd predict. The combination group's fell much further — to 141, compared with 193 in the statin-only group. About 50 mg/dL further.
But that wasn't what the trial was primarily counting.
ENHANCE measured the thickness of the artery wall in the neck — the carotid. It's a scan. The idea is that thicker artery walls mean more disease building up, so if you're doing something useful, the wall should thicken more slowly. It's a stand-in for heart attacks. A proxy.
And on that proxy, adding ezetimibe did nothing [5].
Remember, this trial didn't look at heart disease rates. What was measured was artery wall thickness. What was not measured was heart attacks.
The Collapse, and Then a Long Silence
But that didn't stop the reaction from being fast and devastating.
At the American College of Cardiology's Scientific Sessions in March 2008, two months after the results, an expert panel was convened to tell cardiologists what to do about the disappointing findings. Steven Nissen of the Cleveland Clinic delivered the panel's line: for clinicians who may have employed this medication before exhausting options with statins, the strongest recommendation the panel could make was to turn back to statins [6].
And I want to be fair here. Given the evidence in front of them, that was a reasonable call.
Doctors did what the panel said. Between January 2008 and December 2009, American prescriptions for ezetimibe fell from 1,082 per hundred thousand people to 572. That's a drop of forty-seven percent in under two years [7].
Half the prescriptions gone, on the strength of a scan of the neck.
This idea is still prevalent today, and it might be why your doctor has never discussed ezetimibe with you.
Because the real question we should be asking is: does ezetimibe lower heart disease rates?
The People Born With the Mutation
Here's how ezetimibe's reputation was dragged out of the mud. It started with human DNA.
Ezetimibe blocks the gut protein that absorbs cholesterol. Some people are born with a broken copy of the gene for that protein. Their gut absorbs less cholesterol than everybody else's from the moment they're born, every single day, never a missed dose.
That's a natural experiment. Nature has already run the trial. If blocking the cholesterol gut transporter is genuinely useful, those people should get less heart disease.
In 2014, a consortium sequenced the gene in over twenty-two thousand people with coronary heart disease and nearly seventy thousand without it, and found the carriers.
Their LDL was about 12 mg/dL lower than everyone else's. A modest difference — the kind you'd barely comment on if you saw it on a blood test.
But their risk of coronary heart disease was cut roughly in half — a relative reduction of 53% [8].
Fifty-three percent is a relative figure, and I'll come back to the difference between a relative risk reduction and an absolute one shortly.
But notice what this result rules out.
There's a common argument that cholesterol isn't really the problem — that what actually drives heart disease is insulin resistance, or the specific type of LDL particle, or general metabolic health, and that LDL is just riding along beside the real cause.
This is one datapoint that speaks to that. These carriers weren't randomly healthier. They didn't eat better or exercise more. The single thing separating them from everyone else is one broken copy of one gene that does one job: it reduces how much cholesterol their gut absorbs. Change only that, and the heart disease falls.
When you zoom out from that single datapoint to the whole picture, a review of over 200 studies — including randomised controlled trials involving more than 2 million people — shows the same consistent pattern. Higher LDL-cholesterol in the blood tracks with a higher rate of heart disease, and the authors concluded the evidence shows clearly that LDL causes it [9].
Eighteen Thousand Patients
Coming back to ezetimibe: one genetic study isn't enough to convince the American College of Cardiology. Which brings us to the trial that should have settled it.
IMPROVE-IT was enormous — 18,144 people, all of whom had already been hospitalised with a heart attack or unstable angina. Everyone got a statin. Half also got ezetimibe. The trial ran for years; the median follow-up was six.
And this time, it counted the right things. Not artery wall thickness. Actual events: cardiovascular death, a non-fatal heart attack, a non-fatal stroke, unstable angina bad enough to put you back in hospital, or needing a procedure to reopen an artery.
Here's the result, absolute first — the actual gap between two groups of real people, before any percentage. Relative percentages make drug effects sound enormous.
At seven years, the group on a statin alone had one of those events 34.7% of the time. The group who also took ezetimibe: 32.7%. That is an absolute difference of two percentage points [10].
On the surface, that doesn't sound like much. But there are three reasons it's a big deal.
First, before IMPROVE-IT, nobody had ever shown that adding a non-statin drug on top of a statin delivers further benefit. That was an open question, and a lot of very smart cardiologists had concluded the answer was probably no. It answered the question ENHANCE never asked. Same drug. Real endpoints.
Second, it lines up with the genetics. We have a genetic study and a large randomised controlled trial — two completely different paths, both pointing the same way.
And third, that 2% absolute difference was only after seven years. Extrapolate it across a lifetime and the difference becomes much larger. Let me explain.
What We Had Been Getting Wrong About Cholesterol
Two years later, in 2017, a European consensus panel published something that reframes every number I've just given you.
Their argument is that we have been thinking about LDL the wrong way. It isn't a number you check when you're sixty and then fix. It's an exposure that accumulates — and what damages your arteries is the total dose over a lifetime. The causal effect of LDL is determined by both the absolute magnitude and the cumulative duration of exposure [9].
Magnitude and duration. How high, and for how long. Blockages building up in the arteries is a multi-decade disease.
And this is where the genetics stops being a curiosity and becomes the whole argument. Because when you compare people born with genetically lower LDL against people who lower their LDL by the same amount with a statin started in middle age, you do not get the same result.
Lifelong exposure to lower LDL was associated with a 54.5% reduction in the risk of coronary heart disease per unit of LDL — roughly a three-fold greater reduction than the same drop achieved with a statin started later in life [11].
Three times the protection, for the same drop in cholesterol, if you start early enough.
Relate that back to the 2% absolute risk reduction seen with ezetimibe in IMPROVE-IT. That was over seven years. Instead, we need to be thinking about the absolute risk reduction over 30, 40, 50 years.
A Smarter Way to Use It
By 2022 the question had shifted. Not does this drug work any more, but how should we actually be using it?
The standard approach when someone's cholesterol isn't low enough is to push the statin dose higher. The trouble is that statins hit diminishing returns quickly — each doubling of the dose buys only a few more percent of LDL lowering, while the dose itself keeps climbing.
So the RACING trial tested the alternative. Roughly 3,780 people with established heart disease, in South Korea, randomised to either a high-dose statin on its own, or a moderate-dose statin plus ezetimibe. Three years of follow-up.
The combination held its own — 9.1% of the combination group had a major event, against 9.9% on the high-dose statin alone. And more of the combination group got their cholesterol down under 70 [12].
Two drugs working on two different routes, at sensible doses, doing the job of one drug at a high dose.
What Happens If You Aim Lower
Which leaves one last question, and it got answered this year.
Everything so far tells you that lowering LDL-cholesterol reduces heart disease. It doesn't tell you where to stop. The guidelines have long pointed at a target of 70 for people with established heart disease. Is lower actually better, or is 70 good enough?
Ez-PAVE randomised 3,048 people with heart disease to one of two targets: the standard under-70, or a more aggressive under-55. Three years.
The lower-target group ended up at a median LDL of 56. The standard group, 66. Ten points apart.
Over those three years, the lower-target group had a major cardiovascular event 6.6% of the time. The standard-target group: 9.7%. Three percentage points, absolute — about a third fewer events, from aiming ten points lower. And that's after only three years, let alone fifty [13].
And here's the detail that ties the whole thirty-six-year journey together. The Ez-PAVE trial is named after the drug they used to get there. Two-thirds of that lower-target group were on ezetimibe at three years — a higher proportion than were on a high-intensity statin.
The drug that was written off in 2008 for failing an artery scan is the tool that hit the target in the trial that proved aiming lower works.
Where Ezetimibe Stands Today
So where does that leave it, right now, in 2026?
Ezetimibe is off patent. In the United States you can buy it for around six dollars a month [14].
On side effects, it's about as close to placebo as a real drug gets. For a very small minority of patients, it can cause a bit of tummy upset.
But despite all of this encouraging data, and how cheap it is, ezetimibe is still barely used.
One caveat to all of this, and I'm presenting it so you have the complete picture. Every trial in this story — IMPROVE-IT, RACING, Ez-PAVE — enrolled people who already had heart disease. Every single one. There is no trial asking whether this drug prevents a first heart attack in someone who hasn't had one.
And there probably never will be.
Because the drug is off patent. A trial like that costs hundreds of millions of dollars and takes a decade, and no company on earth has a reason to spend that money proving a six-dollar generic works. The question isn't unanswered because it's scientifically hard. It's unanswered because nobody stands to profit from the answer.
That's a structural problem with how we fund evidence, and it isn't going to be solved by me pointing at it.
The Brain: A Maybe, Not a Reason
There's one more thing worth putting on the table.
You may have seen headlines suggesting cholesterol-lowering drugs may increase dementia. The evidence we have today points in the opposite direction.
A large genetic study of over a million people found that lowering cholesterol through ezetimibe's target, and through the statin target, was linked to substantially less dementia [15].
That matches the real-world observational data we have as well — but beware of over-interpreting it. The risk reduction in that genetic study is huge, around seven times lower, which doesn't sound right.
So it's not yet proven. But the trend is encouraging that cholesterol-lowering therapies are associated with reduced rates of dementia.
Where I Land
So here's where I land, and I want to be clear this is my opinion and it runs ahead of the guidelines.
I personally take a low-dose statin plus ezetimibe, and I have done for a number of years now. I aim for an LDL-cholesterol below 55 mg/dL, which is highly aggressive for primary prevention — as in, to prevent my first heart attack or stroke.
Which raises the obvious follow-up question. If lower is better, how low should you actually go — and what happens at the bottom? That argument has been running almost as long as this one, and it's finally been settled.
References
1. https://www.forbes.com/2002/10/31/cx_mh_1031mrk.html
2. https://pubmed.ncbi.nlm.nih.gov/9336320/
3. https://www.science.org/doi/10.1126/science.1093131
4. https://www.pnas.org/doi/10.1073/pnas.0500269102
5. https://www.nejm.org/doi/full/10.1056/NEJMoa0800742
7. https://pmc.ncbi.nlm.nih.gov/articles/PMC4215424/
8. https://www.nejm.org/doi/full/10.1056/NEJMoa1405386
9. https://academic.oup.com/eurheartj/article/38/32/2459/3745109
10. https://www.nejm.org/doi/full/10.1056/NEJMoa1410489
11. https://www.jacc.org/doi/10.1016/j.jacc.2012.09.017
12. https://pubmed.ncbi.nlm.nih.gov/35863366/
13. https://www.nejm.org/doi/full/10.1056/NEJMoa2600283
14. https://www.costplusdrugs.com/medications/ezetimibe-10mg-tablet/
15. https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/alz.70638
