When Heart Biomarkers Mislead: LDL Cholesterol, HDL and CRP Under Fire

What happens when the results from well-controlled clinical trials produce disappointing results? Could heart biomarkers mislead us? For decades, cardiologists have focused enormous attention on two laboratory measurements: LDL cholesterol and high-sensitivity C-reactive protein (hsCRP). Elevated LDL cholesterol (LDL-C) is considered a major risk factor for atherosclerosis. Elevated CRP is considered an important marker of inflammation and cardiovascular risk.

The conventional wisdom has been straightforward for many years: Lower “bad” LDL cholesterol and reduce inflammation and you should reduce heart attacks, strokes and cardiovascular deaths. If you could raise “good” HDL cholesterol, outcomes should be even better.

What Happens When The Story Does Not Have a Happy Ending?

Usually, that story works reasonably well. It has been repeated countless times by healthcare professionals to millions of patients for decades.

But then along come clinical trials that refuse to cooperate. Some extraordinarily expensive, carefully conducted randomized controlled trials have produced results that should make anyone who follows heart disease research at least a little uncomfortable. Drugs have dramatically improved the laboratory numbers doctors care about, yet patients did not necessarily fare better.

That does not mean LDL cholesterol is irrelevant. It does not mean inflammation has nothing to do with atherosclerosis. It does raise a more interesting question: Are we sometimes confusing a biomarker with the disease itself?

When Heart Biomarkers Mislead: The Latest Shock from ZEUS

The newest challenge comes from a clinical trial called ZEUS. Novo Nordisk developed an experimental monoclonal antibody called ziltivekimab. It targets interleukin-6 (IL-6), an important signaling molecule in inflammation.

ZEUS enrolled more than 6,300 people with atherosclerotic cardiovascular disease, chronic kidney disease and evidence of inflammation. Participants had hsCRP levels of at least 2 mg/L. This was not merely an attempt to nudge a laboratory number in the right direction. The drug successfully engaged its intended biological target. IL-6 activity was inhibited and hsCRP declined as expected. In other words, very high-risk individuals had their marker of inflammation reduced.

Then came the number that mattered most. The hazard ratio for major adverse cardiovascular events was 0.99.

In other words, ziltivekimab was essentially indistinguishable from placebo for preventing cardiovascular death, nonfatal heart attack or nonfatal stroke.

Novo Nordisk Announced the Results of ZEUS on July 31, 2026

Let me quote the drug company directly:

“While ziltivekimab demonstrated target engagement and inhibition of the IL-6 pathway, as reflected by expected reductions in free IL-6 and high-sensitivity C-reactive protein (hsCRP) respectively, this did not translate into major adverse cardiovascular events (MACE) risk reduction versus placebo in people with atherosclerotic cardiovascular disease (ASCVD), chronic kidney disease (CKD) and inflammation.”

That result stunned many cardiologists. Why?

Because ziltivekimab appeared to do what scientists asked it to do biologically. It suppressed the IL-6 pathway and substantially reduced hsCRP. But changing those inflammatory signals did not translate into fewer cardiovascular events in this population.

There is an important qualification. At this writing, ZEUS has been reported as top-line results rather than a complete peer-reviewed publication. Novo Nordisk says full results will be presented at a scientific meeting, and two other ziltivekimab cardiovascular outcome studies, HERMES and ARTEMIS, are continuing.

So it would be premature to declare the inflammation hypothesis dead. Nevertheless, ZEUS asks an uncomfortable question:

If you successfully suppress an inflammatory pathway and dramatically lower a biomarker associated with cardiovascular risk, why don’t cardiovascular events fall too?

CRP Predicts Risk. But Is CRP the Culprit?

A biomarker can be an excellent predictor of disease without itself being the thing that must be changed to prevent the disease. Smoke is an excellent marker of fire. Removing the smoke without extinguishing the flames would not save the house.

For years, hsCRP has been associated with cardiovascular risk. A particularly influential study was the JUPITER trial, published in 2008.

JUPITER enrolled apparently healthy people who did not have conventionally high LDL cholesterol but did have elevated hsCRP. Rosuvastatin (Crestor) lowered LDL cholesterol by about 50% and hsCRP by about 37%, and cardiovascular events were significantly reduced.

JUPITER strengthened interest in inflammation, but it could not tell scientists whether the benefit resulted primarily from lowering LDL cholesterol, reducing inflammation or some combination of effects. That is why trials directed specifically at inflammation have been so important.

The Inflammation Hypothesis Has Won Some Battles

To be fair, anti-inflammatory treatment has sometimes worked. The landmark CANTOS trial tested canakinumab, an antibody against interleukin-1 beta. Unlike statins, canakinumab lowered hsCRP without lowering LDL cholesterol.

The 150-mg dose produced a statistically significant reduction in recurrent cardiovascular events (New England Journal of Medicine, Sept. 21, 2017). That was an important proof-of-concept finding for the inflammation hypothesis. But here are the data:

They were looking at “events” such as heart attacks, strokes or cardiovascular death. There were “4.5 events per 100 person years in the placebo group.” Those patients who received 150 mg of canakinumab experienced 3.86 events per 100 person years. That was after a median follow-up of 3.7 years. So, not even 1 full event was reduced per 100 person years.

There was also a catch. Canakinumab increased serious and sometimes fatal infections. It never became a routine cardiovascular drug.

Another drug that was tested was colchicine. This compound has been used for thousands of years and was originally extracted from the autumn crocus plant. It was used to calm inflammation in joints and prevent gout attacks. Some clinical trials suggested that the drug’s anti-inflammatory activity might have cardiovascular benefits

The CLEAR Trial Upset the Apple Cart

The large CLEAR trial studied colchicine in more than 7,000 patients who had experienced a heart attack. Colchicine was a big disappointment despite its anti-inflammatory properties (New England Journal of Medicine, Feb. 13, 2024).

The Conclusions:

“Among patients who had myocardial infarction, treatment with colchicine, when started soon after myocardial infarction and continued for a median of 3 years, did not reduce the incidence of the composite primary outcome (death from cardiovascular causes, recurrent myocardial infarction, stroke, or unplanned ischemia-driven coronary revascularization).”

If that seems confusing, it should. The bottom line appears to be that reducing inflammatory biomarkers such as hs-CRP has not led to the breakthrough that many cardiologists were hoping for.

Heart Biomarkers Mislead: Now Let’s Consider “Bad” LDL and “Good” HDL

The cholesterol story provides an even more dramatic illustration of how heart biomarkers may mislead. In 2016, cardiologists gathered at the American College of Cardiology meeting to hear the results of an enormous clinical trial called ACCELERATE. The experimental drug was evacetrapib, an inhibitor of cholesteryl ester transfer protein, better known as CETP.

The numbers were astonishing. Evacetrapib reduced “bad” LDL cholesterol substantially while raising “good” HDL cholesterol spectacularly. In the published ACCELERATE trial, HDL cholesterol increased by roughly 133% compared with placebo, while LDL cholesterol decreased by about 31% from baseline at three months.

If “bad” LDL cholesterol is supposed to go down and “good” HDL cholesterol is supposed to go up, evacetrapib appeared to be doing almost everything a cardiologist could hope for.

Then investigators looked at cardiovascular outcomes.

Nothing!

The primary cardiovascular outcome occurred in 12.9% of patients taking evacetrapib and 12.8% taking placebo (New England Journal of Medicine, May 18, 2017). In other words, the drug did not work.

Here are the authors’ conclusions in their own words:

“Although the cholesteryl ester transfer protein inhibitor [CETP] evacetrapib had favorable effects on established lipid biomarkers, treatment with evacetrapib did not result in a lower rate of cardiovascular events than placebo among patients with high-risk vascular disease.”

Think about that for a moment.

A drug dramatically moved both of the cholesterol numbers physicians had been taught to care about in the desirable direction. The “good” went up and the “bad” went down.

And cardiovascular outcomes moved 0.1% after more than two years. Nada, nothing, bupkis.

When Spectacular Cholesterol Numbers Accomplished Nothing

Results from the ACCELERATE trial were especially unsettling because evacetrapib did not just increase HDL. It also substantially lowered LDL. And we have been told for decades that LDL cholesterol is our enemy. The numbers looked great after evacetrapib. The patients did not benefit.

Another Example of How Heart Biomarkers Mislead

Torcetrapib Was Even Worse than Evacetrapib

Before evacetrapib came torcetrapib, Pfizer’s CETP inhibitor. Torcetrapib raised HDL cholesterol dramatically and lowered LDL cholesterol. Once again, the laboratory reports looked terrific. Heart attacks and other cardiovascular complications should have disappeared dramatically.

The ILLUMINATE trial enrolled more than 15,000 patients at high cardiovascular risk (New England Journal of Medicine, Nov. 22, 2007). This was a huge clinical trial!  It had to be stopped early. Patients taking torcetrapib experienced more cardiovascular events and more deaths than patients receiving atorvastatin alone.

An accompanying imaging study was hardly reassuring. Despite substantially increasing HDL and lowering LDL, torcetrapib did not significantly slow the progression of coronary atherosclerosis (New England Journal of Medicine, March 29, 2007).

Torcetrapib did have other potentially harmful effects. It raised blood pressure and altered aldosterone and electrolyte levels. Consequently, cardiologists could reasonably argue that something peculiar to torcetrapib counteracted whatever benefit its cholesterol changes might otherwise have produced. Evacetrapib is harder to explain away that way.

Does This Mean Lowering LDL Doesn’t Work?

No, but the causes and treatments of cardiovascular disease are likely more complicated than just saying LDL cholesterol is “bad” and HDL cholesterol is “good.” Assuming that lowering CRP with anti-inflammatory drugs could cure heart disease has also turned out to be far more complicated than many cardiologists would have hoped.

Assuming that impressive changes in LDL-C or HDL-C will automatically predict how much a patient will benefit turns out to be debatable.

Heart Biomarkers Mislead When We Confuse Prediction With Prevention

Here is the larger take-home message from all these studies. Biomarkers are not, in themselves, outcomes.

Suppose people with gray hair have a higher likelihood of heart disease than people without gray hair. Gray hair might be considered a “biomarker” for heart disease. But inventing a pill that turns gray hair black would not necessarily prevent heart attacks. That sounds silly because everyone understands that gray hair does not cause heart disease.

The trouble comes when a biomarker sits much closer to the disease process. LDL cholesterol, inflammatory cytokines, CRP, HDL cholesterol, triglycerides and Lp(a) all occupy different positions along enormously complicated biological pathways that are linked to heart attacks.

A marker may predict risk. A pathway may contribute to disease. A drug may change the biomarker. Those three statements are not interchangeable.

What Happens When a Sacred Cow Meets a Randomized Trial?

Medicine has a peculiar problem when a beautiful theory encounters an ugly result. The temptation is to explain away the trial. Maybe the wrong patients were studied. Perhaps the treatment began too late. Maybe the drug hit the wrong part of the pathway. And here’s a favorite explanation: the follow-up wasn’t long enough. If we just waited longer, the results could have been better.

Sometimes those explanations are correct. Torcetrapib is a good example. Its adverse effects on blood pressure and aldosterone provide a plausible reason that its favorable cholesterol changes failed to translate into better outcomes.

But there is another danger. If every negative drug trial is explained away while every positive drug trial is treated as confirmation, a hypothesis can become almost impossible to verify.

When ACCELERATE produced astonishing cholesterol improvements and no cardiovascular benefit, the appropriate response should have been:

That’s strange. What are we missing?

When CLEAR failed to reproduce the cardiovascular benefit seen in earlier colchicine trials:

That’s strange. Why?

And when ZEUS successfully inhibited IL-6 signaling and lowered hsCRP without reducing cardiovascular events:

That’s strange too. What does this tell us about inflammation?

Don’t Throw Out LDL or Inflammation. Throw Out Certainty.

The evidence does not justify declaring LDL innocent. Nor does ZEUS justify announcing that inflammation has nothing to do with atherosclerosis. But these trials should make us much less comfortable with slogans such as:

• “LDL is bad.”
• “HDL is good.”
• “Inflammation causes heart disease.”
• “Lower is always better.”

A biomarker can accurately identify people at increased risk without guaranteeing that changing that biomarker will change their risk by the amount we expect — or at all. For patients, the health outcome ultimately matters far more than a lab report.

If Heart Biomarkers Mislead, What Really Matters?

When evaluating a new cardiovascular medicine, here are some questions to ask:

• Did fewer people have heart attacks?
• What was the absolute risk reduction? 
• Out of 100 people, how many were helped by the drug vs. the placebo?
• Did fewer people suffer strokes? How many fewer?
• Did fewer people die? How many fewer compared to placebo?

Those are harder questions than asking whether a number went up or down. But they are the questions that randomized clinical trials are designed to answer. When excellent research provides answers drug companies don’t like, doctors shouldn’t ignore them until everyone forgets. Everyone should pay especially close attention. That is the only way patients can make informed decisions about their health. It will also open doors to new discoveries.

Final Words

I know this is a very complicated topic. I apologize if this was way too technical. Drug companies like to forget about inconvenient results and move on to the next big thing. Our job here at The People’s Pharmacy is to shine a light on all the data, even if the results do not fit with the predominant paradigm.

If you found this article of value, please share it. If you could encourage acquaintances to subscribe to our newsletter, we would be ever so grateful. Where else will they get this kind of analysis? And if you can donate to our work, it would go a long way to keeping The People’s Pharmacy moving forward.