This page has been automatically translated. If there is any discrepancy, the English version is authoritative.

CuringHeartDisease

Always searching for a cure
Peter Megdal, PhD
Join Our Cause

How Low Should Your LDL Cholesterol Be After a Heart Attack?

By: Peter Megdal PhD

How to Use This Article

Choose your format
Easy Read
Deep Dive
YouTube
SPOTIFY
SLIDE DECK
  • Based on peer-reviewed medical research
  • Same evidence cardiologists use
  • Written in clear language for the public
  • Look for underlined terms and related links

Medical disclaimer: This article is for education only and is not medical advice. Always consult your clinician for personal guidance.

Clinical Takeaways

  • Intensive LDL cholesterol lowering stabilizes vulnerable coronary plaques by thickening protective fibrous caps and reducing lipid content with modest volume changes.
  • Substantially lowering LDL cholesterol reduces recurrent heart attacks and ischemic strokes, although cardiovascular event risk never drops completely to zero.
  • Current guidelines recommend achieving an LDL cholesterol level below 55 mg/dL for patients at very high risk after a cardiovascular event.

Easy Read

The Big Idea in One Minute

  • Very low LDL can improve plaque features linked to lower rupture risk. In the HUYGENS study, people who had just had a non-ST-elevation heart attack (NSTEMI) and drove their LDL cholesterol down to about 28 mg/dL had an increase in minimum fibrous-cap thickness approximately twice as large as people on a statin alone. That compares the size of the increases, not the caps themselves, and it does not prove a corresponding drop in heart-attack risk [1].
  • Separately, lowering LDL prevents heart attacks and strokes. HUYGENS measured plaque, not heart attacks. The evidence that lower LDL prevents events comes from different, much larger randomized trials of tens of thousands of people, which showed fewer repeat heart attacks, strokes, and heart procedures [2, 3, 4].
  • Both the amount and the duration of LDL lowering matter. Genetic studies suggest that lifelong lower LDL cholesterol is associated with substantially greater cumulative protection against coronary heart disease [5, 6].
  • Lower LDL has gone with fewer events, but risk never reaches zero. The studies reviewed have not identified a clear lower limit for benefit, but the evidence becomes sparse at extremely low levels. Heart attacks and strokes still happen even at very low LDL.

Figure 1. Illustrative LDL cholesterol levels reported in three different PCSK9-inhibitor trials, each in the group taking the medicine. Only the on-treatment values are shown, because the trials reported their starting values in different ways that cannot be placed side by side. These are not head-to-head comparisons and not the extra effect over placebo; part of each group’s LDL lowering came from the statin. The values are summarized differently: HUYGENS and PACMAN-AMI report means, FOURIER a median [1, 7, 2].

Why One Heart Attack Often Leads to Another

Many people having a heart attack undergo an X-ray of the heart arteries and, when appropriate, have the blocked artery opened with a balloon and a small mesh tube called a stent; treatment depends on the type of heart attack and the arteries involved. A stent fixes one spot. But the arteries of the heart usually contain many other patches of buildup, called plaque. Many of these look small on an X-ray of the arteries, yet they are where a large share of future heart attacks begin.

In the PROSPECT study of 697 people followed for three years after a heart attack or unstable angina, about 1 in 5 had a major heart event. That count combined heart-related death, cardiac arrest, heart attack, and hospital stays for worsening chest pain, so it is not the same as 1 in 5 having another heart attack, and some later events arose from previously untreated plaques that had looked only mildly narrowed at the start [8]. So the real goal after a heart attack is not just to fix the blocked spot. It is to make all the other plaques less dangerous.

What a Dangerous Plaque Looks Like

A plaque sits inside the wall of the artery. At its center is a soft, fatty pool of cholesterol. Over the top is a protective layer of tough tissue called the fibrous cap.

Figure 2. How a heart attack starts (schematic, not to scale). A plaque with a thick cap and a small fatty center is fairly stable (panel 1). A plaque with a thin cap over a large fatty center is fragile (panel 2). If the cap tears, blood touches the fatty center and a clot forms that can block the artery (panel 3). In erosion (panel 4), the fibrous cap stays intact; the artery’s inner lining (endothelium) is lost and a clot forms on the bare surface. No drawing can predict whether a particular plaque will rupture.

The most dangerous plaques have a thin cap stretched over a big, soft, fatty center. When a thin cap tears, blood touches the fatty center and a clot can form. If the clot blocks blood flow long enough, it can cause a heart attack. Plaque erosion and other mechanisms can also cause one. A thicker cap may be harder to tear, though thickness alone is no guarantee.

How Cholesterol Gets Into the Artery Wall

Cholesterol cannot travel through the blood on its own. It rides inside tiny particles, like passengers in cars. The particles slip into the artery wall, get stuck there, and slowly build up plaque. Your LDL cholesterol number measures how much cholesterol the particles carry. Another test, ApoB, is a useful proxy for the concentration of these particles, because each major cholesterol-carrying particle carries one ApoB molecule; the test measures ApoB concentration, so it estimates particle number rather than literally counting particles.

Figure 3. Conceptual illustration. Cholesterol rides in particles, and the particles are what get into the artery wall. Two people can have the same LDL cholesterol but different numbers of particles. ApoB estimates the number of these particles because each typically carries one ApoB molecule; think of it as counting the cars, while LDL cholesterol weighs their cargo. Real particles do not all enter the wall equally, and remnants from the intestine carry a shorter form called ApoB-48; the drawing simplifies this.

Why this matters to you. If you have high triglycerides, diabetes, or excess weight, your LDL cholesterol can look fine while your particle count is still high. Ask your doctor whether an ApoB test makes sense for you.

The HUYGENS Study: Looking Inside the Arteries

A high-resolution imaging catheter that uses light, called OCT, can measure the fibrous cap in living patients. Doctors thread it into the heart’s arteries. It is sharp enough to measure a cap thinner than a human hair.

The HUYGENS study used this camera. It enrolled 161 people who had just had a non-ST-elevation heart attack (NSTEMI) and who had fragile-looking plaques in another artery. Everyone took a statin. Half also took evolocumab, an injection that lowers LDL cholesterol much further, and half took a placebo shot. After about a year, the doctors imaged the same artery segments again. The results describe those measured segments, not every plaque in the heart [9, 1].

Statin plus placebo Statin plus evolocumab
LDL cholesterol after treatment About 87 mg/dL About 28 mg/dL
Change in the protective cap Thickened about 22 micrometers Thickened about 43 micrometers
Fatty-plaque signal on OCT (lipid arc) Shrank Shrank more
Overall plaque size (ultrasound, in a subset; exploratory) Barely changed Shrank slightly

Figure 4. Average increases in the thinnest part of the fibrous cap were 21.5 micrometers with a statin alone versus 42.7 with evolocumab added (P=0.015). These bars show the change, not the final cap thickness. They are model estimates for all 161 treated patients, of whom 119 had paired OCT images. A micrometer is one-thousandth of a millimeter. These imaging changes have not been validated as a prediction of fewer heart attacks for any individual patient [1].

A second study, PACMAN-AMI, started a similar medicine within 24 hours of a heart attack and used three different imaging tools, each measuring something different: ultrasound measured plaque size, a light-based chemical scan measured the fat signal, and OCT measured the cap. Together they showed slightly smaller plaques, a lower fat signal, and thicker caps [7].

The key lesson. Very low LDL cholesterol changed what the plaque is made of much more than how big it is. The plaque did not disappear, but its imaging features shifted toward those linked to lower rupture risk. Whether these changes are what prevents heart attacks has not been proven.

Does It Actually Prevent Heart Attacks?

Better-looking plaque is encouraging, but what people care about is fewer heart attacks, strokes, and deaths. For that, we look at much larger studies that followed tens of thousands of people.

The answer is yes. In the FOURIER study of more than 27,000 people with heart disease, adding evolocumab to a statin lowered the chance of the study’s primary composite of cardiovascular death, heart attack, stroke, hospitalization for unstable angina, or an artery-opening procedure by about 15% over about two years [2]. In absolute terms, that was about 1.5 fewer people out of every 100. Heart attacks fell by about a quarter, and strokes by about a fifth. Similar studies of other cholesterol-lowering medicines found the same pattern [3, 4, 10].

Figure 5. What “15% lower risk” means in real people. Out of 1,000 people on a statin alone, about 113 had a major event over about two years; with evolocumab added, about 98 did, or about 15 fewer per 1,000. The 15% is a relative reduction in hazard; the 15 per 1,000 is the absolute difference, a separate number. Both are approximate, based on the trial’s estimated event rates at a median of 2.2 years, and the “major event” count includes hospital stays for chest pain and artery-opening procedures, not just heart attacks [2].

The effect looks even bigger when you count every event, not just the first one. Many people who have one heart attack or stroke go on to have another. When FOURIER counted all of them, a prespecified total-events analysis estimated approximately 52 fewer primary-endpoint events per 1,000 patients over three years, including repeat events [11].

Figure 6. Estimated events prevented per 1,000 people over three years, counting first events only versus counting repeat events too. These are estimates from a prespecified FOURIER total-events analysis, not raw patient counts, and the two counting methods are not directly comparable; the larger number does not mean the relative risk reduction doubled [11].

A landmark analysis pooling individual patient data from 26 statin trials and about 170,000 people found that every 38.7 mg/dL (1 mmol/L) drop in LDL cholesterol lowered major heart and stroke events by about 22%, and it lowered deaths too [12]. How much LDL falls, and for how long, are major drivers of the benefit, though the medicines differ in how much evidence supports them, when they work, and their side effects [13].

Time Is the Secret Ingredient

The longer LDL cholesterol stays low, the more benefit builds up.

In FOURIER, the benefit was larger after the first year than during it [2]. After the study ended, survivors who chose to continue were all given evolocumab for about five more years. The people who had started it two years earlier still had fewer heart attacks, strokes, and heart-related deaths, even though both groups were now on the same medicine [14]. This later comparison was not blinded and included only people who chose to continue, so it is supporting evidence rather than proof; it is compatible with, but does not prove, the idea that an earlier start leaves a lasting advantage.

Figure 7. Illustrative LDL cholesterol paths over a lifetime; not measured data. The ages, starting levels, and curves are hypothetical. Cumulative exposure is the area under each curve: treatment after a heart attack lowers future exposure but cannot erase decades of past exposure, while people born with low LDL cholesterol accumulate far less. The sharp drop at age 55 is conceptual, not a measured clinical path, and the area under a curve cannot by itself predict anyone’s risk.

What People Born With Low LDL Teach Us

Some people inherit genes that give them low LDL cholesterol from birth. Because these genes are handed out by chance, they act like a natural experiment, and the results are striking.

In one long-term study, Black participants who inherited a gene variant that lowered their LDL cholesterol by about 28% had an 88% lower risk of heart disease over 15 years [5]. That estimate came from one group of people with few heart events, so its uncertainty is wide. Across many such genes, each 38.7 mg/dL (1 mmol/L) of lifelong lower LDL was linked to roughly half the risk of heart disease, roughly three times the benefit per unit seen when people start a statin later in life; that is a rough comparison between very different kinds of studies, not proof that the two are equivalent [6].

Figure 8. What an 88% lower risk looks like. These dots illustrate an 88% reduction applied to an easy-to-picture starting point of 10 in 100; they are not the study’s actual counts. This particular inherited variant, in one group of people, was linked to markedly lower coronary heart disease risk; genetic and drug-trial estimates are not directly comparable [5].

These gene studies are observations, not treatment trials, so their numbers can’t be directly compared with what a medicine started in adulthood can do. There are also limits. One common variant did not lower strokes caused by blockages [15]. In another large study, PCSK9 variants lowered deaths from heart and blood-vessel disease, but did not change deaths from all causes [16]. Low LDL cholesterol is powerful protection against heart disease, but it is not protection against everything.

How Low Should LDL Go?

LDL cholesterol What we know How sure are we?
Below 70 mg/dL Treatments that brought LDL below this level reduced heart attacks and strokes in randomized trials Very sure the treatments work; trials tested drugs, not exact LDL targets
Below 55 mg/dL Add-on medicines that brought LDL below this level gave further benefit in randomized trials Very sure the treatments work; trials tested drugs, not exact LDL targets
Around 30 to 40 mg/dL The medicine groups in large randomized trials reached this level and did better Quite sure
Below 20 mg/dL People who got this low did better on average; no major safety signal in the outcomes checked Fairly sure, but these people were not assigned to this level
Below 10 mg/dL Same direction, but few people reach it Less sure

Thousands of people in the big studies reached LDL cholesterol below 20 mg/dL, and lower achieved LDL cholesterol was associated with lower event rates. No major safety signal emerged in the outcomes examined over as long as about eight years, though rare or lifetime effects remain uncertain [17, 18]. But those people were not randomly assigned to such low levels, and no study has directly compared aiming for 20 with aiming for 40.

In plain terms. Lower LDL has generally gone with fewer events, even at very low levels. The studies reviewed have not identified a clear lower limit for benefit, but the evidence becomes sparse at extreme levels, and the best evidence covers levels down to about 30 to 40 mg/dL. No study assigned people to specific very low targets.

Why Risk Never Reaches Zero

Figure 9. In selected FOURIER extension participants originally assigned to evolocumab (median LDL cholesterol about 30 mg/dL), heart-related death, heart attack, or stroke occurred at 2.05 events per 100 person-years during the extension, shown as 2 marked dots among 100 person-years of follow-up; this is an event rate, not the chance that any one person has an event in a given year. This is not the lifetime risk at that LDL level, and extension participants were survivors who chose to continue, not a random sample [14].

Even at very low LDL cholesterol, some heart attacks and strokes still happen, because:

  • Plaque already built doesn’t vanish. Decades of buildup shrink only slightly.
  • A few years can’t undo forty. Lowering LDL at age 60 can’t erase the exposure of earlier decades.
  • Other things damage arteries. Lipoprotein(a), triglycerides, inflammation, smoking, high blood pressure, and diabetes all still matter [19, 20].
  • Not every heart attack comes from a torn cap. Some come from worn artery linings or clots from elsewhere.

Is Very Low LDL Safe?

The body makes its own cholesterol for the brain and hormones, and the evidence so far is reassuring. In studies lasting up to about eight years, very low LDL cholesterol was not linked to a clear increase in memory problems, bleeding strokes, cancer, or other serious side effects in the outcomes examined [21, 17, 18]. Side effects differ by medicine. Statins modestly raise the chance of developing diabetes in people already close to it, and can cause muscle aches, though true muscle damage is rare [22]. The injected PCSK9 medicines did not show a clear diabetes signal in randomized trials, but can cause minor reactions where the shot is given [2]. These findings apply only to the outcomes measured, in selected volunteers, over the years they were followed; very rare or very long-term effects cannot be ruled out, and lifelong safety at extremely low LDL is unknown.

What This Means for You

If you have had a heart attack… Why
Start or continue a strong statin, usually before leaving the hospital It is the foundation of treatment
Know your LDL cholesterol and your goal For people with heart or artery disease, U.S. guidelines aim for below 55 mg/dL if they meet the “very high risk” criteria and below 70 mg/dL otherwise, with a goal below 55 mg/dL considered reasonable for some people who are not in the very-high-risk group; a recent heart attack alone does not automatically put you in the very-high-risk group [23]
If you are not at goal, ask about adding ezetimibe or a PCSK9 inhibitor These add-on medicines lower LDL further and reduce events
Ask for a one-time Lp(a) test, and ask whether ApoB makes sense for you These can reveal risk your LDL number misses
Don’t read a rising calcium score as failure on its own Calcium can rise while plaque becomes more stable, but discuss any new result with your doctor; repeat calcium scans are not a proven way to judge treatment
Stay on treatment The benefit builds over years and fades if you stop
Treat the whole picture No smoking, blood pressure and blood sugar control, cardiac rehab, and your other heart medicines

Questions to ask your cardiologist. What is my LDL cholesterol now, and what should my goal be? Am I in the “very high risk” group? Would adding another medicine help me reach my goal? Should I have my Lp(a) or ApoB checked? How will we track whether my treatment is working?

What We Still Don’t Know

  • Whether starting a second cholesterol medicine in the hospital prevents more heart attacks than starting it a few weeks later. One recent study, AMUNDSEN, tested two things: how many people reached their LDL goal (82% did with the early injection versus 40% with usual care) and, as its main clinical end point, whether early treatment reduced deaths or unplanned heart-related hospital stays at one year. That clinical end point showed no significant difference (adjusted odds ratio 0.94; 95% confidence interval 0.73 to 1.19). The trial was large enough to detect a drop of roughly one-third in these events and did not find one, which argues against a large early benefit; a smaller benefit, or one that appears only after more than a year, has not been ruled out. Only about 4% of the usual-care group received this type of medicine during the year [24].
  • Whether aiming for 20 mg/dL is truly better than aiming for 40 mg/dL.
  • Exactly how much the plaque changes explain the drop in heart attacks.
  • How safe extremely low LDL is over many decades.

Words to Know

Word What it means
LDL cholesterol (LDL-C) The “bad cholesterol” blood test: the amount of cholesterol carried inside LDL particles
ApoB A blood test that is a useful proxy for the concentration of cholesterol-carrying particles, since each typically carries one ApoB
Plaque A buildup of cholesterol, cells, and scar tissue inside the artery wall
Fibrous cap The protective layer covering a plaque’s soft, fatty center
Statin The standard first medicine for lowering LDL cholesterol
PCSK9 inhibitor An injected medicine (such as evolocumab or alirocumab) that lowers LDL cholesterol much further
Ezetimibe A pill that lowers LDL cholesterol by blocking cholesterol absorption
Lp(a) An inherited type of ApoB-containing particle that raises heart risk; usually tested once
OCT A tiny light-based camera that can image the inside of an artery in fine detail
Placebo A dummy treatment used for comparison in studies

Deep Dive

The Plaque Nobody Treated

Picture a patient a few days after a heart attack. The blocked artery has been opened, a stent holds it in place, and the chest pain is gone. It is natural to think the problem has been fixed. It has not. The stent treated one spot in one artery. Scattered through the rest of the coronary tree are other plaques, most of them too small to look worrying on an angiogram, and those are where many of the next heart attacks will come from.

That is not speculation. The PROSPECT study followed 697 patients for three years after an acute coronary syndrome. Over three years, the cumulative rate of the study’s composite of major adverse cardiovascular events was 20.4%, about one in five; this was a composite of cardiac death, cardiac arrest, heart attack, or rehospitalization for unstable or progressive angina, not heart attacks alone. Events traced to the original culprit sites occurred in 12.9% of patients, and events traced to previously untreated plaques in 11.6%; some patients had both, so the two do not add up to the total, and not every event came from a previously untreated plaque. Most of the untreated plaques that later caused trouble had looked only mildly narrowed on the original angiogram [8].

So the real question after a heart attack is not only “Is the stent working?” It is “Can we change the rest of the disease before it causes the next event?” This article follows that question through the evidence: what happens inside plaque when LDL cholesterol is driven very low, whether that translates into fewer heart attacks, how low is low enough, and why risk never reaches zero.

The short answer. In randomized imaging studies after acute coronary events, intensive LDL-C lowering produced substantial improvements in selected markers of plaque vulnerability, including cap thickness and lipid content, with comparatively modest average reductions in plaque volume. Separate randomized outcome trials demonstrate fewer cardiovascular events with LDL-C lowering, but do not establish that the observed plaque-imaging changes mediate that benefit. Coronary disease does not disappear.

See Figure 1 in Part 1.

The Evidence at a Glance

Question What the evidence shows How sure are we?
Does very low LDL-C change plaque? Yes. Caps get thicker, lipid signals shrink, and plaque volume falls slightly. Shown in randomized imaging trials for specific imaging measures; not a validated substitute for events
Does lowering LDL-C prevent heart attacks and strokes? Yes. Large trials show fewer heart attacks, ischemic strokes, and procedures. Proven in large randomized trials
Does it reduce deaths? Yes for statins in pooled trials; for add-on drugs, only in nominal or open-label analyses. Proven for statins; suggestive for add-ons
Is lower still better below 20 or even 10 mg/dL? Lower achieved levels were associated with lower risk, compatible with benefit but subject to confounding and imprecision; no consistent major safety signal was detected in the populations and follow-up studied. Suggestive (not randomized to those levels)
Does time matter? Yes. Benefit grows over the years, and people with lifelong low LDL-C have far less coronary disease. Strongly supported
Can risk reach zero? No. Events still occur even at extremely low LDL-C. Consistent across all data
Do the plaque changes cause the fall in events? Plausible, but not proven. Inference

Inside a Heart Attack

To understand what HUYGENS measured, it helps to know what a dangerous plaque looks like.

See Figure 2 in Part 1.

A plaque is a buildup inside the artery wall, not a deposit stuck on the inside of a pipe. It has several parts:

  • Lipid-rich or necrotic core: a pool of cholesterol, cholesterol crystals, and debris from dead cells. It is soft, mechanically weak, and strongly clot-promoting if blood reaches it.
  • Fibrous cap: a collagen layer, built by smooth muscle cells, that walls the core off from the bloodstream. In autopsy studies of sudden coronary death, ruptured caps were typically thinner than about 65 micrometers (μm), roughly the width of a fine human hair [25].
  • Macrophages: immune cells that swallow cholesterol and become “foam cells.” They release enzymes that can digest the collagen in the cap.
  • Calcium: mineral deposits. Small, spotty calcium in an active plaque may concentrate stress, while large, dense sheets often mark older, more fibrous disease. Extensive calcium remains a marker of plaque burden and risk.

Most heart attacks begin when a thin cap tears (rupture) and blood meets the clot-promoting core. A smaller share begin with erosion, in which the artery’s inner lining wears away over a firmer, less fatty plaque [26]. That difference will matter later, because thickening a cap mainly protects against rupture.

How Cholesterol Gets Into the Wall

LDL cholesterol does not float into the artery wall as loose fat. It travels inside particles, and it is the particles that cross the artery lining, get trapped, and are swallowed by macrophages. Each of these particles, whether LDL, a triglyceride-rich remnant, or Lp(a), carries one molecule of a protein called apolipoprotein B (ApoB): ApoB-100 on particles made by the liver, and the shorter ApoB-48 on chylomicron remnants from the intestine.

See Figure 3 in Part 1.

In plain terms: ApoB is the traffic count; LDL-C is the cholesterol cargo. The standard lab value, LDL-C, measures how much cholesterol the LDL particles carry. The ApoB test measures ApoB mass concentration, which serves as a proxy for particle number because each major atherogenic particle carries one ApoB-100 molecule; it does not literally count particles. Usually the two move together. When they disagree, as can happen with high triglycerides, diabetes, obesity, or small LDL particles, ApoB may better reflect how many particles can enter the artery wall.

For readers who use international units, these are the approximate conversions at the LDL-C levels discussed in this article (rounded; for example, 40 mg/dL is about 1.03 mmol/L, so the rounded values are not exactly equivalent thresholds):

LDL-C (mg/dL) 70 55 40 30 20 10
LDL-C (mmol/L) 1.8 1.4 1.0 0.78 0.52 0.26

Looking Inside a Living Artery

For most of the history of cardiology, nobody could see a fibrous cap in a living person. Angiograms show only the silhouette of the open channel. That changed with catheters that image the artery from the inside.

Figure 10. Approximate resolution of coronary imaging methods (typical ranges; vary by system). Only OCT resolves detail finer than a vulnerable fibrous cap, so cap-thickness results come from OCT, while plaque volume comes from IVUS, which sees through the full wall. Near-infrared spectroscopy (NIRS, not shown) detects the chemical signature of lipid rather than structure. None of these is a microscope: tissue labels are inferences.

Imaging tool What it is good at What it cannot do well
OCT (light) Measures cap thickness, lipid arc, macrophage-like signals, calcium, clot Cannot see total plaque volume behind lipid
IVUS (ultrasound) Measures plaque volume and how the artery has remodeled Too coarse to measure a cap
NIRS (spectroscopy) Detects the chemical signature of lipid Cannot measure caps or volume on its own
Coronary CT Shows plaque across the whole heart without a catheter Cannot see caps or fine detail

None of these is a microscope. The tissue labels they produce are inferences from how light or sound behaves. An OCT increase in cap thickness, a NIRS decrease in lipid, and an IVUS decrease in plaque volume are complementary findings, not interchangeable ones.

The HUYGENS Experiment

HUYGENS (High-Resolution Assessment of Coronary Plaques in a Global Evolocumab Randomized Study) was built to answer one question: if LDL cholesterol is pushed very low right after a heart attack, do the dangerous-looking plaques elsewhere in the heart change? It was a phase 3, multicenter, randomized, double-blind, placebo-controlled imaging trial [9, 1]. Amgen, which makes evolocumab, funded it, and several investigators reported financial ties to the company [9].

Who Took Part

The investigators deliberately chose people whose remaining plaques already looked vulnerable. Patients had a non-ST-elevation heart attack (NSTEMI) treated with angiography and a procedure on the culprit artery. They needed LDL-C above a threshold that depended on their statin use (at least 130 mg/dL on no statin, 80 mg/dL on a low- or moderate-intensity statin, or 60 mg/dL on a high-intensity statin), and a separate, untreated artery segment at least 40 mm long showing lipid-rich plaque on OCT: a cap no thicker than 120 μm and a lipid arc greater than 90 degrees [9].

Everyone took a maximally tolerated statin. Half were also randomized to evolocumab 420 mg, an injected PCSK9 inhibitor, and half to a placebo injection, monthly through week 48, with repeat OCT and IVUS at about week 50 to 52 [9]. HUYGENS included 161 randomized and treated patients (80 evolocumab, 81 placebo). Of the treated patients, 135 had evaluable imaging at follow-up, and a smaller subset had paired IVUS, The registry results report the OCT endpoints for the full analysis set of all 161 treated patients, of whom 119 had paired OCT measurements of the primary endpoint, so the analysis did not directly reimage all 161; the IVUS results come from a smaller subset [1]. Participants averaged 60.5 years of age, and 28.6% were women.

How Low LDL Went

Mean LDL-C at enrollment was 141.3 mg/dL. At follow-up it was 28.1 mg/dL with evolocumab versus 87.2 mg/dL with placebo [1]. The placebo group was not untreated: many patients entered on no statin or a low dose, so their fall from about 141 to 87 mg/dL largely reflects starting or strengthening a statin. Every HUYGENS comparison is statin versus statin plus evolocumab.

What Happened to the Plaque

See Figure 4 in Part 1.

The protective cap thickened by an average of +42.7 μm with evolocumab versus +21.5 μm with placebo, least-squares means from the trial’s analysis of covariance in all 161 treated patients (the full analysis set, per the trial’s registry results); a difference of 21.2 μm (95% CI 4.7–37.7; P=0.015) [1]. The lipid arc shrank more (−57.5 versus −31.4 degrees), and the macrophage signal fell more. Plaque volume, measured by IVUS in an exploratory subset, changed only modestly: percent atheroma volume change −2.29 versus −0.61 percentage points, an arithmetic difference of −1.68 points (not an adjusted contrast, and not directly comparable with PACMAN-AMI’s model-based estimate).

In plain terms. The average increase in cap thickness was about twice as large with evolocumab. That compares the size of the increases, not the caps themselves, and it does not mean heart-attack risk was halved. The big change was in what the plaque is made of, not in how large it is. These results describe the selected nonculprit segments that were imaged, not every plaque in the coronary tree.

A secondary review by trial investigators also reported fewer evolocumab-treated patients with very thin caps (below 65 μm) after treatment [27]. The full HUYGENS results table is in the Technical Appendix.

Two later HUYGENS analyses looked for patients who responded differently. Patients with high Lp(a) showed a larger difference in cap thickening between evolocumab and placebo [28], and patients who reached lower ApoB levels had more favorable plaque findings than those who did not [29]. Both are small, post hoc subgroup analyses: they were not randomized to those levels, they need replication, and they do not establish an independent treatment target or mechanism. Details are in the Technical Appendix.

What HUYGENS Cannot Tell Us

HUYGENS provides randomized evidence that adding evolocumab to statin therapy after NSTEMI improves several OCT-derived features associated with plaque vulnerability in selected nonculprit coronary lesions. It does not establish that these imaging changes directly prevented later cardiovascular events.

HUYGENS was not designed to show fewer heart attacks, strokes, or deaths. That would take tens of thousands of patients, not 161. It imaged only at the start and at one year, so it cannot say how quickly caps change. It studied people chosen because their plaques already looked vulnerable. And it cannot show that thicker caps caused any later drop in events. For that, we need other kinds of trials.

Not a One-Off: Other Trials Saw the Same Thing

A single 161-patient trial would be easy to dismiss. HUYGENS is persuasive because other trials, using different drugs and different imaging tools, saw the same pattern.

PACMAN-AMI randomized 300 heart attack patients to alirocumab, another PCSK9 antibody, or placebo, both on top of rosuvastatin, starting within 24 hours after urgent PCI for the heart attack, and assessed nonculprit plaques with complementary imaging methods. Among the 300 randomized participants, paired IVUS was available in 265; the NIRS lipid analysis included 263 patients, and the OCT cap-thickness analysis 221 patients (105 alirocumab, 116 placebo) who had a fibroatheroma at both time points. All three measurements came from the same randomized trial. LDL-C fell to 23.6 mg/dL versus 74.4 mg/dL. Three complementary imaging modalities, measured in overlapping patients within the same trial, showed concordant favorable changes: percent atheroma volume fell more (difference −1.21 percentage points), the NIRS lipid signal fell more (difference −41.24), and cap thickness rose more (+62.67 versus +33.19 μm; difference +29.65 μm) [7]. These differences are the trial’s model-based estimates. When light, sound, and chemistry agree within one trial, the finding is more convincing than any single measure, though it is still one trial.

Figure 11. HUYGENS and PACMAN-AMI: change over about one year in each treatment arm. Both trials added a PCSK9 antibody (evolocumab in HUYGENS, alirocumab in PACMAN-AMI) to statin therapy after myocardial infarction. Bars are within-arm mean changes as reported; the between-group differences were +21.2 μm (P=0.015) and +29.65 μm (11.75–47.55) for cap thickness, and −1.68 (arithmetic) and −1.21 (−1.78 to −0.65) percentage points for percent atheroma volume. For PACMAN-AMI, simple subtraction of the arm means gives 29.48 μm, slightly different from the trial’s model-adjusted 29.65 μm because the trial’s mixed-effects model accounts for repeated measurements and for the multiple arteries imaged in each patient; the text uses the adjusted figure. The two panels use different units and should not be compared with each other. HUYGENS percent atheroma volume was an exploratory endpoint in an IVUS subset. Analysis populations differ by measure: HUYGENS cap results are model estimates in the full analysis set of 161 treated patients (119 with paired OCT; 135 with any evaluable follow-up imaging) and its PAV comes from a smaller IVUS subset; PACMAN-AMI’s PAV comes from 265 patients with paired IVUS and its cap thickness from 221 patients with a fibroatheroma at both time points. Sources: [1, 7].

GLAGOV randomized 968 patients (846 with evaluable follow-up IVUS) with stable coronary disease to evolocumab or placebo for 76 weeks, with the main IVUS scan at week 78. LDL-C averaged 36.6 versus 93.0 mg/dL. Percent atheroma volume changed by −0.95 percentage points with evolocumab versus +0.05 with placebo (between-group difference −1.0 percentage point; 95% CI −1.8 to −0.64, reproduced verbatim from the published trial), and 64% versus 47% of patients had measurable regression [30]. These are within-arm changes plus the trial’s reported between-group contrast, not percent reductions in plaque mass. In an exploratory subgroup that reached about 24 mg/dL, 81% regressed. GLAGOV also shows the limits: about a third of treated patients did not regress, and a tissue-analysis substudy found no significant differences in plaque components [31].

Earlier statin and ezetimibe trials told the same story at higher LDL-C levels: lower LDL-C meant less plaque growth or modest regression [32, 33, 34, 35], and thicker caps on OCT [36]. YELLOW III documented favorable within-person changes in plaque-imaging measurements over 26 weeks; because it was uncontrolled, it cannot establish how much of the change was due to evolocumab [37]. A 2025 single-arm evolocumab CT/PET study reported less noncalcified plaque and lower microcalcification activity, while calcified plaque volume rose; its imaging endpoints measure different things and should be read individually [38]. Not every trial was positive: ODYSSEY J-IVUS found no significant difference in its plaque endpoint over 36 weeks [39].

Figure 12. Change in IVUS percent atheroma volume in each arm of five randomized imaging trials, one panel per trial, with each arm’s achieved LDL-C. Compare arms only within a panel: populations, durations (9 to 24 months), and LDL-C measures (GLAGOV is time-weighted) differ, so the panels are not evidence of a common dose–response relationship. HUYGENS values come from an exploratory IVUS subset. SATURN reported median changes and time-weighted LDL-C, whereas the other trials report means. ASTEROID is omitted because it had no comparator arm, and ODYSSEY J-IVUS because absolute achieved LDL-C was not available in the sources reviewed. Sources: [34, 35, 30, 1, 7].

What does a one-point drop in plaque volume mean? Percent atheroma volume (PAV) is the share of the vessel’s area taken up by plaque. A change from 40% to 39% is a one-percentage-point drop. It does not mean 1% of all plaque disappeared, and it does not mean a narrowing opened by the same amount.

Figure 13. Conceptual illustration of the scale of a one-percentage-point change in percent atheroma volume (PAV); the open channel (lumen) is drawn the same size in both panels, and the outer vessel boundary shrinks slightly. PAV is the share of the vessel’s area occupied by plaque, averaged along a segment. A fall from 40% to 39% is the scale of change in the IVUS trials. It is not 1% of all plaque disappearing, and it does not mean the open channel widened by a matching amount, because arteries also remodel.

A side-by-side table of all the imaging trials is in the Technical Appendix.

Different, Not Much Smaller

If plaque volume barely changes, how could a plaque become less likely to rupture? Mechanics offers a plausible answer. Rupture happens when stress concentrates in a thin cap stretched over a soft core. Less lipid and a thicker cap should lower that stress even if the plaque is almost the same size. That offers a plausible explanation for how outcome trials can show meaningful reductions in events while imaging trials show only small changes in volume, though it remains an interpretation, not a demonstrated cause.

Figure 14. Illustrative schematic—not an observed patient or a guaranteed treatment response. Conceptual diagram of plaque stabilization: not drawn to scale and not derived from patient images. Randomized imaging trials show cap thickening, less lipid, fewer macrophage-like signals, and a small decrease in plaque volume. Calcium changes and lesion anatomy are illustrative, not measured transitions in a single plaque. Randomized trials do not show plaque elimination, calcium removal, or routine lumen enlargement.

The chain of reasoning from lower LDL-C to fewer heart attacks has several links. Some are measured directly; others are inferred.

Figure 15. How lower LDL-C relates to plaque imaging and to clinical events. Fewer ApoB particles and fewer events are both established, the second in large outcome trials. Lipid, macrophage-like, and cap findings are parallel imaging surrogates measured together, not a proven sequence. Whether these imaging changes mediate the reduction in events has not been established.

Observational imaging supports the idea that these features matter. In the CLIMA study, untreated plaques that combined four high-risk OCT features carried a several-fold higher one-year risk of events [40]. Linking a feature to risk, however, is not the same as proving that changing the feature lowers risk.

Three cautions keep this in proportion:

  • A thicker cap is not a guarantee. The 65 μm cutoff comes from autopsy studies [25]. Thicker caps can still rupture, OCT measurement of such thin structures varies, and cap thickness varies along its length.
  • Erosion is different. Heart attacks from erosion are driven more by injury to the artery lining and inflammation than by cap thinning [26]. Cap thickening mostly addresses rupture, which is one reason events still occur at very low LDL-C.
  • Calcium may go up, not down. High-intensity statins increased IVUS calcium measures while reducing plaque volume [41], and statin users in the PARADIGM CT cohort had slower growth of noncalcified plaque but faster growth of calcified plaque [42]. A rising calcium score on treatment does not, by itself, mean treatment has failed, and serial calcium scores are not a validated way to judge treatment. Claims that LDL lowering “dissolves calcium” are not supported.

Nor does regression necessarily reopen the artery. A reduction in percent atheroma volume does not necessarily translate into a comparable increase in lumen diameter. Coronary arteries can remodel outward or inward, and plaque regression may occur without a clinically meaningful change in the narrowing seen on an angiogram. The imaging trials discussed here do not establish routine reversal of fixed coronary narrowing.

From Better-Looking Plaque to Fewer Heart Attacks

Imaging trials show what happens to plaque. Outcome trials, which enroll thousands of patients and count actual heart attacks, strokes, and deaths, show whether it matters. This is the strongest evidence in the story.

Figure 16. How the evidence in this article is weighted. Randomized outcome trials establish that lowering LDL-C prevents the events they tested; randomized imaging trials show that it changes imaging measures of plaque; the remaining tiers support, but cannot by themselves prove, claims about very low targets, long-term legacy, and lifelong benefit. Each claim in the text is labeled by the kind of evidence behind it.

The Numbers That Matter to Patients

Relative reductions sound large; absolute reductions show what a person gains. In FOURIER, the primary composite (cardiovascular death, heart attack, stroke, hospitalization for unstable angina, or coronary revascularization) occurred in 9.8% with evolocumab versus 11.3% with placebo over a median 2.2 years: a 15% relative hazard reduction and an absolute difference of about 1.5 percentage points, or about 15 fewer people per 1,000 [2]. That composite is not the chance of a heart attack alone, and the original FOURIER trial did not establish a reduction in death.

See Figure 5 in Part 1.

Trial What was tested Follow-up Event rate, treatment vs control Difference per 1,000 (people; FOURIER-OLE: events per 1,000 person-years)
MIRACL [43] High-dose statin vs placebo, started in hospital 16 weeks 14.8% vs 17.4% About 26
PROVE IT–TIMI 22 [10] Intensive vs moderate statin 2 years 22.4% vs 26.3% About 39
IMPROVE-IT [4] Adding ezetimibe to a statin 7 years 32.7% vs 34.7% About 20
ODYSSEY OUTCOMES [3] Adding alirocumab after ACS Median 2.8 years 9.5% vs 11.1% About 16
FOURIER [2] Adding evolocumab, stable disease Median 2.2 years 9.8% vs 11.3% About 15
FOURIER-OLE [14] (open-label extension; not a randomized trial endpoint) Earlier vs later evolocumab (open label) Median 5 more years CV death, MI, or stroke 2.05 vs 2.58 per 100 person-years About 5.3 fewer events per 1,000 person-years (rate difference, not people)

Each trial counted its own combination of events in its own population over its own follow-up, so the rows are not a ranking. FOURIER-OLE compares original randomized assignment in selected survivors, all on open-label evolocumab. A detailed version of this table, with relative effects, LDL-C levels, and mortality, is in the Technical Appendix.

Figure 17. Absolute benefit in the major outcome trials, shown as small multiples: each panel gives one trial’s control and treated primary-endpoint event rates on its own y-axis scale, with the difference expressed per 1,000 patients treated. Each trial used a different composite endpoint, population, comparator, and follow-up period, so the panels cannot be compared with one another or read as a ranking; they show the scale of benefit within each trial. *FOURIER-OLE shows cardiovascular death, MI, or stroke during the open-label extension, comparing patients originally assigned evolocumab with those originally assigned placebo (2.58 vs 2.05 events per 100 person-years); because this is a rate, not a percentage of patients, its panel has its own y-axis label; it is an exploratory comparison in selected survivors, not a new randomized trial. Sources: [43, 10, 4, 3, 2, 14].

Figure 18. Selected outcome estimates from the trials discussed: primary endpoints for the randomized trials, the open-label extension comparison for FOURIER-OLE, and the main clinical end point for AMUNDSEN. Each trial used a different composite endpoint, population, and follow-up, and reported a different effect measure (risk ratio, hazard ratio, or adjusted odds ratio), so each effect measure is shown in its own panel with its own axis, for orientation, not pooled or ranked. Absolute differences are crude differences in each trial’s reported event rates over its own follow-up (for FOURIER-OLE, per 100 person-years). FOURIER-OLE is not a newly randomized outcome comparison: it compares selected surviving participants by original assignment during an open-label extension in which all received evolocumab. AMUNDSEN (gray) is shown for its main clinical end point, which was tested second in the trial’s prespecified hierarchy, after the LDL-C target endpoint; the trial was powered for it. PROVE IT–TIMI 22’s interval is derived from its reported 16% (5%–26%) relative reduction. MIRACL’s relative risk is shown as the trial reported it, estimated from a Cox proportional hazards model. Sources: [43, 10, 4, 3, 2, 14, 24].

Fewer Heart Attacks and Strokes; Death Is Harder to Move

The benefit is concentrated in nonfatal events. In FOURIER, heart attacks fell from 4.6% to 3.4%, strokes from 1.9% to 1.5%, and coronary procedures from 7.0% to 5.5% [2]. ODYSSEY OUTCOMES and IMPROVE-IT showed similar reductions in heart attack and ischemic stroke [3, 4]. These are secondary analyses that support the main findings.

Among the add-on trials (IMPROVE-IT, ODYSSEY OUTCOMES, and FOURIER), none showed a reduction in death that was statistically significant under its prespecified rules. ODYSSEY OUTCOMES’ lower all-cause death rate was nominal only. This does not prove that there is no mortality benefit; it means one should not be claimed as proven for these drugs individually.

Counting Every Event

Trials usually count only each patient’s first event. But a patient who has one heart attack often has another, or a stroke, or a stent. When FOURIER counted every event, the estimated number of events prevented more than doubled, although the relative risk reduction did not: about 22 first events but 52 total events prevented per 1,000 patients treated for three years, counting first and repeat events rather than additional people helped (total events: rate ratio 0.82, 95% CI 0.75–0.90; first events: hazard ratio 0.85) [11].

See Figure 6 in Part 1.

The Big Picture

The Cholesterol Treatment Trialists pooled individual data from 26 statin trials and about 170,000 people. Each 38.7 mg/dL (1 mmol/L) reduction in LDL-C lowered major vascular events by about 22%, and all-cause death fell by about 10% per mmol/L, with no increase in cancer or other non-cardiovascular deaths [12]. A meta-regression of randomized trials found broadly similar proportional risk reductions per unit of LDL-C lowering across statins, ezetimibe, and PCSK9 inhibitors, which act through LDL-receptor pathways [13]. That supports a common LDL-mediated mechanism, though it does not mean every drug has identical clinical effects or safety profiles.

In plain terms. The absolute LDL-C reduction and how long it lasts are major determinants of benefit, but therapies differ in evidence, timing, off-target effects, and tolerability.

Do the plaque changes explain this benefit? Plausibly, in part. The same drugs at similar LDL-C levels improve plaque in imaging trials and reduce events in outcome trials. But the imaging and outcome trials enrolled different people, and no trial has shown that the imaging changes account for the event reduction. The plaque findings make the benefit biologically coherent; the outcome trials establish the benefit for the endpoints they tested.

Time Is the Hidden Ingredient

Perhaps the most important lesson in the newer data is that benefit builds over time.

The Lag Effect

In FOURIER, the relative reduction in cardiovascular death, heart attack, or stroke was about 16% in the first year and about 25% after the first year [2]. Earlier statin trials showed the same pattern: benefit grows the longer treatment continues. That is consistent with, though it does not prove, plaque gradually becoming more stable.

The Legacy Effect

FOURIER did not demonstrate a significant cardiovascular mortality benefit during its median 2.2-year randomized phase. Afterward, 6,635 surviving participants who chose to continue entered an open-label extension, FOURIER-OLE, in which everyone received evolocumab; median extension follow-up was 5.0 years, and maximum evolocumab exposure across the parent trial and extension reached about 8.4 years; median LDL-C during the extension was about 30 mg/dL [14].

Even though both groups were now on the same drug at similar LDL-C, those originally assigned evolocumab, about two years earlier, had fewer subsequent events. Cardiovascular death, heart attack, or stroke occurred at 2.05 versus 2.58 events per 100 person-years (HR 0.80; 95% CI 0.68–0.93), and cardiovascular death at 0.68 versus 0.90 per 100 person-years (HR 0.77; 95% CI 0.60–0.99), as reported in the results table of the primary FOURIER-OLE publication, where events were counted during extension follow-up by original randomized assignment [14]. This is an exploratory comparison among selected survivors in open-label follow-up, not a new blinded randomized trial, so it cannot establish a mortality benefit by itself. It is compatible with a legacy effect from earlier, longer LDL lowering, but selection into the extension and open-label treatment mean it cannot show that early treatment caused the difference, or what carried it forward.

Figure 19. FOURIER-OLE: event rates per 100 person-years during the open-label extension, when all 6,635 participants received evolocumab and median LDL-C was about 30 mg/dL. Patients first randomized to evolocumab had 2.2 more years of low LDL-C exposure. Even in that group, cardiovascular death, MI, or stroke occurred at about 2 per 100 person-years. Extension participants were selected survivors and treatment was open label. Source: [14]. These are event rates in a selected-survivor, open-label extension; they cannot prove a randomized mortality benefit.

How Fast Things Change

  • LDL-C falls within days with a PCSK9 antibody and within two to four weeks with a statin. In AMUNDSEN, median LDL-C was 16 mg/dL six weeks after evolocumab was started before stenting, versus 56 mg/dL with standard care [24].
  • Plaque changes in the randomized imaging trials were measured after roughly 9 to 24 months. YELLOW III documented within-person changes at 26 weeks, but it was uncontrolled, so it cannot establish the size of the drug’s effect or how early it began [37]. A trial that imaged only at the start and at one year cannot show that caps began thickening within days.
  • Clinical events: MIRACL showed fewer recurrent ischemic events within 16 weeks [43]; FOURIER’s curves separated within about six months [2]; FOURIER-OLE’s exploratory cardiovascular-death difference emerged only during the open-label extension [14].

Figure 20. Timeline separating measured LDL-C changes, measured plaque changes, and observed clinical benefit. Each cell shows when a study made a measurement, not when a biological effect began; the figure cannot show when stabilization started, and no unmeasured time points are extrapolated. The time of a study measurement is not necessarily the time a biological benefit began; clinical benefit can begin before imaging is repeated, and no fixed biological lag is implied. A blank cell means not measured, not no effect.

Exposure Over a Lifetime

Why should time matter so much? Because atherosclerosis reflects the total, accumulated exposure of the artery wall to ApoB particles over a lifetime. Genetic studies indicate that cumulative exposure drives risk [44]. A person treated after a heart attack lowers their future exposure but carries decades of past exposure and the plaque it built.

See Figure 7 in Part 1.

How much a patient gains depends on three things together: baseline risk, how much LDL-C falls, and how long it stays down. The same LDL reduction prevents far more events in a high-risk patient, and larger sustained reductions prevent more.

Figure 21. Evidence summary, not a patient-specific prediction. Lines apply the Cholesterol Treatment Trialists’ estimate of a 22% lower risk of major vascular events per 38.7 mg/dL (1 mmol/L) of LDL-C lowering over about five years [12] to three illustrative baseline five-year risks. The same LDL-C reduction prevents more events in higher-risk patients, and larger sustained reductions prevent more. Longer duration adds further benefit beyond what a five-year estimate shows (see the lifetime-exposure figure). Reductions above about 80 mg/dL extend beyond much of the trial data. Assumptions: the CTT rate ratio (a relative risk from a meta-analysis, not a trial hazard ratio) is applied uniformly at every baseline risk and scales log-linearly with LDL-C reduction. This is not a patient risk calculator.

How Low Should LDL Go?

Patients often ask whether 40 is good enough, or whether 20 or 10 is better. The honest answer depends on what kind of evidence you accept, because the lower one goes, the more the evidence shifts from randomized comparisons to observational ones.

LDL-C level What the evidence looks like Bottom line
Below 70 mg/dL Randomized trials (PROVE IT, statin imaging trials) Established for drug regimens that reached these levels (not randomized to the threshold)
Below 55 mg/dL Randomized trials (IMPROVE-IT, ODYSSEY, FOURIER) Established for drug regimens that reached these levels (not randomized to the threshold)
Below 40 mg/dL Randomized trial arms reached it (FOURIER about 30; GLAGOV 36.6) Strongly supported
Below 30 mg/dL Imaging trials (HUYGENS 28, PACMAN-AMI 24); achieved-level analyses Supported for plaque; outcomes observational
Below 20 mg/dL Thousands of patients in FOURIER and FOURIER-OLE reached it, but were not randomized to it Lower risk, no clear safety signal; observational
Below 10 mg/dL Small subsets of FOURIER, FOURIER-OLE, and ODYSSEY Consistent but statistically fragile

No trial cited here randomized post-heart-attack patients between specific LDL-C targets below 55 mg/dL. People in the drug trials were randomized to drugs, not to “20 versus 40.” That is the most important caveat in this area.

Within those limits, the data point one way. In FOURIER, 2,669 participants reached LDL-C below 20 mg/dL (0.5 mmol/L) by week 4; lower achieved LDL-C was associated with steadily lower cardiovascular risk across the range studied, and serious adverse events were not related to achieved LDL-C over the randomized period [17]. In FOURIER-OLE, about 1,604 participants maintained LDL-C below 20 mg/dL; lower levels remained associated with lower risk and no detected increase in serious adverse events over the extension’s follow-up [18]. In ODYSSEY OUTCOMES, patients who transiently reached below 15 mg/dL had fewer subsequent events in an adjusted post hoc comparison (HR 0.72; 95% CI 0.51–0.997), which does not establish that aiming below 15 mg/dL is better [45].

In plain terms. Lower achieved LDL-C has generally been linked to lower event rates across the ranges examined, including very low levels. The studies reviewed have not identified a clear lower threshold; evidence becomes sparse at extreme LDL-C, and achieved-level analyses cannot identify the best target. No randomized trial has shown that aiming for 10 or 20 mg/dL beats aiming for 30 or 40 mg/dL.

How Low Can Risk Go?

There is a different question hiding inside “how low should LDL go.” Suppose LDL-C is pushed nearly to zero. How low can the rate of heart attack, stroke, and cardiovascular death fall? The answer differs sharply between people who already have coronary disease and people who have had low LDL-C since birth.

In People Who Already Have Atherosclerosis

This is not a randomized comparison of LDL-C targets: patients sorted themselves into achieved-LDL groups, so selection and confounding can only partly be adjusted for, and the analysis does not prove there is no biological floor. Within that limit, the peer-reviewed FOURIER-OLE achieved-LDL analysis found risk falling steadily down to the lowest levels studied, below 20 mg/dL [18].

Two points follow. Risk was lowest at the lowest achieved levels studied, an observational association. And the floor is well above zero: among selected participants entering FOURIER-OLE, cardiovascular death, heart attack, or stroke occurred at 2.05 events per 100 person-years of follow-up in those originally assigned evolocumab, the group treated longest; this is a rate by original randomized assignment, not the risk at an LDL-C of 30 mg/dL [14].

These are observational comparisons within trials; patients who reach the lowest levels may differ in ways that statistical adjustment cannot fully remove.

Why the Floor Exists

  • Plaque already built. These patients carried decades of accumulated plaque, and IVUS shows plaque burden falling by only a few percentage points.
  • Short exposure. Two to eight years of very low LDL-C cannot erase forty or more years of higher exposure.
  • Other pathways. Lp(a), triglyceride-rich remnants, and residual inflammation persist. In CANTOS, COLCOT, and LoDoCo2, anti-inflammatory therapy reduced events in patients already on lipid-lowering drugs [19, 20, 46].
  • Other causes. Some strokes come from the heart or tiny vessels, and some heart attacks from erosion, artery tears, or clots carried from elsewhere. Smoking, blood pressure, diabetes, and kidney disease also continue to drive risk.

Nature’s Experiment: People Born With Low LDL

Some people inherit gene variants that lower LDL-C from birth. Because those variants are assigned at conception, largely independent of lifestyle, they act like natural randomized experiments. Their results are dramatic.

See Figure 8 in Part 1.

  • PCSK9 nonsense variants. In the ARIC study, Black participants carrying inactivating PCSK9 variants had about 28% lower LDL-C and an 88% lower risk of coronary heart disease over 15 years. White participants with a milder variant had about 15% lower LDL-C and 47% lower risk [5]. Few carriers had events, so the estimate is imprecise (hazard ratio 0.11; 95% CI 0.02–0.81).
  • APOB truncating variants. Carriers had 43 mg/dL lower LDL-C and 72% lower odds of early coronary disease [47].
  • PCSK9 R46L in large studies. This common variant lowers LDL-C only modestly, yet it was associated with less coronary or ischemic heart disease [48, 15] and lower cardiovascular mortality [16].
  • Across nine variants in six genes (312,321 people), each 38.7 mg/dL of lifelong lower LDL-C was associated with 5% lower coronary risk, about three times the effect per unit of statins started later in life; this is an illustrative comparison across different study designs, not causal equivalence [6].

The genetic data also mark limits. The same R46L variant that lowered coronary risk was not associated with lower ischemic stroke [15], and lifelong lower LDL-C from PCSK9 variants was associated with lower cardiovascular but not all-cause mortality [16]. Across these datasets, reductions in ischemic stroke or all-cause mortality were not consistently demonstrated; estimates differ by variant, population, endpoint, and statistical precision. These are observational genetic associations with population-specific uncertainty. Effect sizes differ across variants, populations, and endpoints, and because they reflect lifelong exposure, their scale cannot be equated with a drug treatment effect started in adulthood; quantitative comparisons with trials are illustrative.

People who inherit two inactivating PCSK9 variants are extremely rare. The best-documented case is a 32-year-old woman with no detectable PCSK9 and LDL-C of 14 mg/dL who was healthy, fertile, and working as an aerobics instructor [49]. She was described at a single point in time; nobody knows her long-term cardiovascular outcome.

Putting the Two Together

Figure 22. Illustrative model, not trial data or a patient risk calculator. Lines show two published log-linear associations with different endpoints, populations, and exposure durations: the Cholesterol Treatment Trialists’ rate ratio of 0.78 (95% CI 0.76–0.80) for major vascular events per 38.7 mg/dL over about five years of statin treatment [12], and the Mendelian randomization estimate of 54.5% (95% CI 48.8%–59.5%) lower coronary heart disease risk per 38.7 mg/dL of lifelong exposure, from 312,321 participants and 9 variants in 6 genes [6]. Each curve is ratio^(LDL-C difference/38.7); shaded bands carry each published 95% confidence interval through the same formula and are not observed trial confidence bands. Extrapolated segments are dashed: in panel B beyond about 40 mg/dL, past the range of most genetic variants, and in panel A beyond about 80 mg/dL, past most trial LDL-C reductions, and the model assumes the log-linear relationship holds at extreme exposures, which has not been shown. The two estimates are shown in separate panels, without individual study or trial points, because their endpoints and populations differ; they must not be read as a head-to-head comparison of statin therapy with lifelong genetic LDL reduction or as a prediction of individual lifetime risk.

The figure compares two published estimates. Applied to a difference of about 75 mg/dL, the statin-trial estimate predicts roughly 38% fewer major vascular events over about five years, while the genetic estimate predicts roughly 78% fewer coronary events over a lifetime. Both are extrapolations, especially the genetic one. They illustrate why duration matters, not what any individual should expect.

The defensible conclusion. Early, sustained, very low exposure to ApoB particles probably drives atherosclerotic coronary events toward a low floor. But no human data show near-zero cardiovascular events, the genetic data suggest stroke and total mortality would remain above zero, and in people who already have coronary disease, events still occur even at extremely low LDL-C.

Is Very Low LDL Safe?

A natural worry is that the body needs cholesterol. It does, but cells make their own, and the evidence so far is reassuring, with honest limits.

  • Memory and thinking. In EBBINGHAUS, 1,204 FOURIER patients underwent formal cognitive testing over a median 19 months, with no difference between evolocumab and placebo and no link to achieved LDL-C [21]. Longer follow-up found no excess of reported cognitive events [18], though that relies on adverse-event reports.
  • Bleeding strokes. High-dose atorvastatin after stroke modestly increased hemorrhagic stroke in SPARCL [50]. FOURIER, ODYSSEY OUTCOMES, and the very-low-LDL analyses showed no significant excess [2, 17, 18], though such strokes are uncommon.
  • Statins modestly raise the risk of new diabetes, about 9% in relative terms, mostly in people already near the threshold [22]. PCSK9 inhibitors and very low LDL-C have not shown a comparable increase [2, 17].
  • The adrenal glands and gonads can make their own cholesterol and also take it up from circulating lipoproteins, and large PCSK9 programs have not shown a clinically important hormonal signal.

Achieved-LDL analyses are observational within trials and subject to selection and residual confounding; any isolated nominal findings are hypothesis-generating, and very long-term safety at extreme levels remains incompletely characterized.

Drug side effects are a separate matter: PCSK9 antibodies cause slightly more injection-site reactions (2.1% vs 1.6% in FOURIER) [2], and statins can cause muscle symptoms, though true muscle injury is rare. Randomized safety data for PCSK9 antibodies cover about two to three years; open-label exposure reached about 8.4 years [14]. In selected FOURIER extension participants, very low achieved LDL-C was not associated with a detected increase in several prespecified safety outcomes over the parent trial and its open-label extension [18]. That is reassuring but limited by participant selection, observational comparisons, and the inability to exclude rare or lifelong effects. Unlimited lifelong safety at extremely low levels has not been, and cannot yet be, established.

What Guidelines Recommend

Guidelines turn many trials into practical goals. The 2026 U.S. multisociety dyslipidemia guideline reintroduced explicit LDL-C goals and divides people with atherosclerotic cardiovascular disease (ASCVD) into two groups [23]. Very-high-risk ASCVD means multiple major ASCVD events, or one major event plus multiple high-risk conditions; a recent heart attack alone does not automatically qualify.

Who What the 2026 U.S. guideline recommends
ASCVD, not very high risk High-intensity statin; LDL-C below 70 mg/dL and non-HDL-C below 100 mg/dL (recommended)
ASCVD, not very high risk Add ezetimibe, a PCSK9 antibody, or bempedoic acid if needed to reach below 70 mg/dL
ASCVD, not very high risk Going further, to below 55 mg/dL (non-HDL-C below 85) (reasonable)
Very-high-risk ASCVD High-intensity statin; LDL-C below 55 mg/dL and non-HDL-C below 85 mg/dL (recommended)
Very-high-risk ASCVD Add ezetimibe and/or a PCSK9 antibody if needed to reach those goals
Very-high-risk ASCVD Bempedoic acid is a further add-on option

“Recommended” and “reasonable” are shown only for the LDL-C and non-HDL-C goals, where they follow the guideline’s own wording for stronger and moderate recommendations. For add-on drugs, the guideline’s recommendation tables give the formal classes, evidence levels, and eligibility criteria, and separate recommendations apply in special situations such as severe hypercholesterolemia [23].

The 2025 U.S. acute coronary syndrome guideline recommends high-intensity statin therapy and timely addition of nonstatin therapy for those above recommended levels [51]. European guidance after a heart attack recommends at least a 50% reduction and LDL-C below 55 mg/dL [52], and allows a goal below 40 mg/dL after a second vascular event within two years [53]. None of these specific targets was directly tested against another in a randomized target trial; they are expert judgments built on the evidence above.

Should a Second Drug Start in the Hospital?

AMUNDSEN randomized 2,161 heart attack patients and demonstrated that a strategy of giving the first evolocumab injection before PCI substantially improved attainment of the combined LDL-C target at 12 months (82% versus 40%). However, the difference in the trial’s main clinical end point (tested second in the prespecified order, after the LDL-C target), death or unplanned cardiovascular hospitalization, was not statistically significant at one year (14.6% versus 15.4%; adjusted OR 0.94, 95% CI 0.73–1.19). The trial therefore supports faster, more reliable target attainment, but does not demonstrate an additional short-term clinical-event benefit from this initiation strategy [24]. Standard care permitted a PCSK9 inhibitor when indicated, but only 3.8% of the standard-care group received one. The trial was powered (80%) to detect an odds ratio of 0.68 for this end point, so it argues against a large early clinical effect of starting a PCSK9 inhibitor in the catheterization laboratory; it does not exclude a smaller benefit or one that emerges with longer treatment. A prespecified per-protocol analysis suggested fewer events with evolocumab (10.2% versus 14.3%), but the investigators caution that this is likely affected by selection bias. A one-year result does not refute the longer outcome trials. EVOLVE-MI, testing evolocumab within 10 days of a heart attack, had not reported by the search cutoff. For now, early combination therapy is proven to reach lower LDL-C sooner and, in PACMAN-AMI, to improve one-year plaque changes; it is not yet proven to prevent more events than adding the same drug weeks later.

Where ApoB Fits

The guideline gives ApoB a role in assessing residual risk. ApoB is most useful when it and LDL-C disagree, as in people with high triglycerides, diabetes, or obesity, where LDL-C can look “at goal” while particle number remains high. No randomized trial reviewed here has compared ApoB treatment targets with LDL-C targets, so ApoB is best used as an added measure of remaining particle burden rather than as a proven separate target.

The Bottom Line

If someone has already had a heart attack, how much can intensive lipid lowering change the disease and lower the chance of another event?

The disease: Within about a year, lowering LDL-C to roughly 25 to 30 mg/dL produces about twice the cap thickening seen with a statin alone, reduces imaging markers of lipid in untreated plaques, and adds roughly a 1.2 to 1.7 percentage-point reduction in percent atheroma volume compared with control groups. The plaque does not disappear, calcium does not dissolve, and the artery is not restored to normal.

The risk: In randomized trials after an acute coronary syndrome, relative reductions in each trial’s main composite endpoint were roughly 15%, with absolute benefits that differ by trial and follow-up. Across the pooled statin trials, each 38.7 mg/dL of LDL-C lowering reduced major vascular events by about 22%. Heart attacks and ischemic strokes are clearly reduced; a reduction in death is proven for statins in pooled analyses, but for add-on therapies only in nominal or open-label analyses.

How low can risk go: The studies reviewed have not identified a clear lower threshold, evidence becomes sparse at extreme LDL-C, and achieved-level analyses cannot define the optimal target; but events still occur even at extremely low LDL-C. People with lifelong low LDL-C have far less coronary disease, which points to duration of exposure as a major lever, but reductions in ischemic stroke and total mortality were not consistently shown. No evidence shows that cardiovascular events can be driven to zero.

Well established Suggestive Still unresolved
Lowering LDL-C after a heart attack reduces repeat heart attack, ischemic stroke, and procedures, in proportion to the LDL reduction Benefit and safety at sustained LDL-C below 20 and even below 10 mg/dL Whether starting a PCSK9 inhibitor in the hospital prevents more events than starting it weeks later
The benefit continues below 70 and 55 mg/dL in randomized comparisons A lasting legacy benefit from earlier treatment (open-label, selected survivors) Whether any target below 55 mg/dL beats another in a randomized trial
Intensive LDL lowering improves plaque composition and modestly reduces plaque volume over about a year Far greater protection from lifelong low LDL-C The true floor of risk with lifelong near-zero LDL-C
For statins, ezetimibe, and PCSK9 antibodies, benefits generally outweigh recognized harms in the populations studied A cardiovascular-mortality benefit from adding PCSK9 inhibition Lifelong safety at extremely low LDL-C
That plaque changes explain part of the benefit; that ApoB and Lp(a) modify the plaque response How to prevent heart attacks caused by erosion; whether imaging-guided therapy helps

Practical Points to Discuss With Your Cardiologist

  • Start or continue high-intensity statin therapy during the hospital stay after an acute coronary syndrome, unless it is contraindicated or not tolerated.
  • Know your numbers and your risk category. The 2026 U.S. guideline recommends an LDL-C goal below 55 mg/dL for people with ASCVD classified as very high risk and below 70 mg/dL for those who are not, with a goal below 55 mg/dL reasonable in selected patients who are not at very high risk; European ACS guidance generally recommends at least a 50% reduction and LDL-C below 55 mg/dL after an acute coronary syndrome. Ask your clinician which category you are in and what your goal is. If you are not at goal on a statin, ask about adding ezetimibe or a PCSK9 inhibitor.
  • Ask for a one-time Lp(a) test, and consider ApoB, especially if your triglycerides are high or you have diabetes.
  • Do not read a rising calcium score as failure on its own. Discuss what it means in your case.
  • Treat the whole picture. Stop smoking, control blood pressure and blood sugar, attend cardiac rehabilitation, and take prescribed antiplatelet therapy.
  • Stay on therapy. LDL-C rises when treatment stops, and the benefit depends on how long it stays low.

Medical disclaimer: This article is for educational purposes and summarizes published research. It is not medical advice and does not replace individualized care from your physician. Do not start, stop, or change any medication without talking to your doctor.

Technical Appendix

About This Review

This narrative review selectively summarizes landmark peer-reviewed studies available through October 2026 and is not a systematic review or meta-analysis. Where full text was not available, I used the peer-reviewed abstract only when it contained enough detail to support the specific statement. Press releases, registry pages, and news coverage were not used as evidence. Numbers not supported by the sources consulted were excluded. Every value plotted in a data figure was checked against the published abstract or primary report, as listed in the companion figure register; supplementary tables were not used. The bibliographic details of all 54 references (authors, title, journal, year, volume, and pages) were checked against the Crossref registry.

A1. HUYGENS Results

HUYGENS measure (change over about one year) Evolocumab + statin Placebo + statin Difference* P value
Achieved LDL-C, mg/dL 28.1 87.2 about −59 <0.001
Minimum fibrous cap thickness, μm (primary endpoint) +42.7 +21.5 +21.2 0.015
Maximum lipid arc, degrees (secondary) −57.5 −31.4 −26.1 0.04
Macrophage index, mm (exploratory) −3.17 −1.45 −1.72 0.04
Percent atheroma volume, percentage points (IVUS subset; exploratory) −2.29 −0.61 −1.68 0.009
Microchannels and calcium (exploratory) No significant difference NS

*Values and P values are as reported in the published HUYGENS abstract; 135 treated patients had evaluable imaging at follow-up. The cap-thickness values are least-squares means from the trial’s analysis of covariance in the full analysis set of 161 treated patients, 119 with paired OCT (registry results: difference 21.2 μm, 95% CI 4.7–37.7); IVUS values come from a smaller subset. Other differences are simple arithmetic on the reported group means; the P values come from the trial’s own models. Source: [1].

The protocol prespecified the absolute change in minimum cap thickness as the primary endpoint; secondary endpoints were the percent change in minimum cap thickness, the change in average minimum cap thickness, and the change in maximum lipid arc; IVUS plaque burden, macrophages, and microchannels were exploratory. The trial planned about 150 patients so that 125 would complete paired imaging, with 90% power to detect a 51 μm benefit; the observed adjusted benefit, 21.2 μm, was less than half that assumption yet still statistically significant [9]. Imaging trials can only measure patients who survive, stay in the study, and undergo a second catheterization with usable images, so missing follow-up imaging can introduce selection bias.

A2. HUYGENS Subgroup Analyses

These small, post hoc findings are hypothesis-generating and do not establish separate treatment thresholds or independently causal mechanisms. They were based on subsets and not adjusted for multiple comparisons.

  • Lipoprotein(a). In HUYGENS, the difference in cap thickening between evolocumab and placebo appeared larger in patients with high Lp(a) than in those with lower Lp(a), partly because the placebo-plus-statin group with high Lp(a) showed a smaller response [28]. Evolocumab lowered Lp(a) only modestly, and the analysis does not show that lowering Lp(a) caused the benefit. In PACMAN-AMI, baseline Lp(a) did not modify the plaque-volume or cap-thickness response to alirocumab, although patients with the highest Lp(a) had a smaller reduction in the plaque lipid signal [54].
  • Patients who reached lower ApoB levels showed more favorable OCT findings than those who did not [29]. Achieved ApoB was not randomized, and the full adjustment model could not be examined, so this does not establish an optimal ApoB target or an independent causal effect.
  • Achieved LDL-C. Pooling both arms, lower achieved LDL-C and larger LDL-C reductions were associated with greater cap thickening [1]; also non-randomized.

A3. Plaque-Imaging Trials

Trial Population Treatment vs comparator Achieved LDL-C, mg/dL (mmol/L) Imaging, duration Main plaque finding Key limitations
HUYGENS (2022) [1] 161 post-NSTEMI Evolocumab vs placebo, both + statin 28.1 vs 87.2 (0.73 vs 2.25) OCT; IVUS subset; 50 wk Cap +42.7 vs +21.5 μm; lipid arc −57.5° vs −31.4°; PAV −2.29 vs −0.61 135 with evaluable follow-up imaging; denominators vary by endpoint; selected plaque
PACMAN-AMI (2022) [7] 300 acute MI, started <24 h after PCI Alirocumab vs placebo, both + rosuvastatin 20 23.6 vs 74.4 (0.61 vs 1.92) IVUS + NIRS + OCT; 52 wk PAV diff −1.21; LCBI diff −41.24; cap diff +29.65 μm 265 paired IVUS; 81% men
GLAGOV (2016) [30] 968 stable CAD on statin Evolocumab vs placebo 36.6 vs 93.0 (0.95 vs 2.41) IVUS; 76 wk treatment, week-78 IVUS PAV −0.95 vs +0.05; regression 64% vs 47% Not post-MI; composition substudy neutral
REVERSAL (2004) [32] 654 stable CAD Atorvastatin 80 vs pravastatin 40 79 vs 110 (2.0 vs 2.8) IVUS; 18 mo Progression halted vs continued Moderate LDL levels
ASTEROID (2006) [33] 507 CAD Rosuvastatin 40, single arm 60.8 (1.57) IVUS; 24 mo PAV −0.98 No control; attrition
SATURN (2011) [34] 1,385 CAD randomized; 1,039 with follow-up IVUS Rosuvastatin 40 vs atorvastatin 80 62.6 vs 70.2 (1.62 vs 1.82) IVUS; 104 wk Median PAV −1.22 vs −0.99 (P=0.17) No significant difference between the two statins; both regressed; no placebo arm
PRECISE-IVUS (2015) [35] 246 after PCI, Japan Atorvastatin + ezetimibe vs atorvastatin 63.2 vs 73.3 (1.63 vs 1.90) IVUS; 9–12 mo PAV −1.4 vs −0.3; regression 78% vs 58% Open label
EASY-FIT (2014) [36] 70 unstable angina Atorvastatin 20 vs 5 69 vs 78 (1.78 vs 2.02) OCT; 12 mo Cap +69% vs +17% Small; low-dose comparator
ODYSSEY J-IVUS (2019) [39] 206 recent ACS, Japan Alirocumab vs standard care Not shown IVUS; 36 wk No significant difference in primary plaque endpoint Neutral; short duration
YELLOW III (2026) [37] 110 stable CAD, lipid-rich plaque Evolocumab + statin, single arm Not reported in sources reviewed OCT + NIRS-IVUS; 26 wk Cap 70.9 → 97.7 μm; maxLCBI4mm 306.8 → 213.1; PAV −1.38 points No control group
CCTA/PET study (2025) [38] 47 stable, noncalcified plaque Evolocumab, single arm Not reported in sources reviewed CCTA + PET; 18 mo Noncalcified plaque fell; calcified plaque rose No control group

 

Values are not directly comparable across trials. PACMAN-AMI “diff” values are model-based between-group differences; other entries are within-arm changes as reported. PAV, percent atheroma volume; TAV, total atheroma volume; LCBI, lipid core burden index.

A4. Major Outcome Trials in Detail

Trial Population, timing, follow-up Treatment vs comparator LDL-C, mg/dL (baseline; achieved) Primary endpoint Result Mortality
MIRACL (2001) [43] 3,086 unstable angina or non-Q-wave MI; 24–96 h; 16 wk Atorvastatin 80 vs placebo ~124; 72 vs 135 at 16 wk Death, MI, resuscitated arrest, or recurrent ischemia with rehospitalization 14.8% vs 17.4%; RR 0.84 (0.70–1.00) No significant reduction in death or MI; benefit driven mainly by fewer rehospitalizations for symptomatic ischemia
PROVE IT–TIMI 22 (2004) [10] 4,162 ACS; within 10 days; mean 24 mo Atorvastatin 80 vs pravastatin 40 106; 62 vs 95 Death, MI, UA rehospitalization, revascularization, or stroke 22.4% vs 26.3%; 16% reduction (5%–26%) Death 2.2% vs 3.2%; P=0.07
IMPROVE-IT (2015) [4] 18,144 ACS; within 10 days; median ~6 y Simvastatin + ezetimibe vs simvastatin 93.8; 53.7 vs 69.5 CV death, major coronary event, or nonfatal stroke 32.7% vs 34.7% at 7 y; HR 0.936 (0.89–0.99) No difference
ODYSSEY OUTCOMES (2018) [3] 18,924 ACS 1–12 mo earlier; median 2.8 y Alirocumab vs placebo + statin ~92 (mean); ~40 vs ~93 at 4 mo CHD death, MI, ischemic stroke, or UA hospitalization 9.5% vs 11.1%; HR 0.85 (0.78–0.93) All-cause death 3.5% vs 4.1%, nominal only
FOURIER (2017) [2] 27,564 stable ASCVD; median 2.2 y Evolocumab vs placebo + statin 92 (median); 30 with evolocumab at 48 wk, 56 mg/dL lower than placebo CV death, MI, stroke, UA hospitalization, or revascularization 9.8% vs 11.3%; HR 0.85 (0.79–0.92) CV death HR 1.05 (0.88–1.25)
FOURIER-OLE (2022) [14] 6,635 FOURIER completers; median 5 more years Original evolocumab vs original placebo, all on open-label drug Median ~30 at extension week 12, on open-label evolocumab Same as FOURIER 3.36 vs 3.96 per 100 person-years; HR 0.85 (0.75–0.96) CV death 0.68 vs 0.90 per 100 person-years, HR 0.77 (0.60–0.99); exploratory
AMUNDSEN (2026) [24] 2,161 STEMI/NSTEMI (1,087 vs 1,074); open label; 12 mo Evolocumab before PCI + standard care vs standard care (PCSK9 inhibitor permitted; used in 3.8%) Median 16 vs 56 at 6 wk LDL-C <55 and ≥50% reduction; main clinical end point tested second in a prespecified hierarchy (80% power for OR 0.68) 792/970 (82%) vs 370/934 (40%); aOR 5.54 (4.50–6.82). Clinical composite 159/1,087 (14.6%) vs 165/1,074 (15.4%); aOR 0.94 (0.73–1.19) Not significant

Event rates are Kaplan–Meier estimates for PROVE IT–TIMI 22 and IMPROVE-IT and observed proportions for the others. Endpoints, populations, and follow-up differ, so these figures should not be compared across trials. AMUNDSEN was checked against the full primary JAMA article. CV, cardiovascular; CHD, coronary heart disease; UA, unstable angina.

Endpoint (treatment vs control) IMPROVE-IT (7 y) [4] ODYSSEY OUTCOMES (median 2.8 y) [3] FOURIER (median 2.2 y) [2]
Myocardial infarction 13.1% vs 14.8% 6.6% vs 7.6% 3.4% vs 4.6% (HR 0.73)
Stroke (trial-specific definition) Ischemic stroke 3.4% vs 4.1% Ischemic stroke 1.2% vs 1.6% Any stroke 1.5% vs 1.9% (HR 0.79)
Coronary revascularization 21.8% vs 23.4% (not significant) Ischemia-driven, lower with alirocumab 5.5% vs 7.0% (HR 0.78)
Cardiovascular death No significant difference Not significant (HR 0.88) No difference (HR 1.05)
All-cause death No significant difference 3.5% vs 4.1% (HR 0.85), nominal only No difference

 

These components are secondary or nominal analyses, not multiplicity-controlled results.

A5. Limitations of the Evidence

  • Small imaging trials. HUYGENS randomized 161 patients, 135 of whom had evaluable follow-up imaging; PACMAN-AMI randomized 300, with 265 having paired IVUS. Both selected particular segments, had missing follow-up imaging, and enrolled mostly men. HUYGENS selected patients with already vulnerable-looking plaque.
  • Imaging is not histology. OCT can mistake fibrous tissue for lipid, and macrophage signals are indirect.
  • Industry funding. Amgen funded HUYGENS, GLAGOV, and FOURIER; Sanofi and Regeneron supported ODYSSEY OUTCOMES, and Sanofi, Regeneron, and Infraredx funded PACMAN-AMI; Merck funded IMPROVE-IT; AstraZeneca funded SATURN. The major outcome trials used blinded treatment and independent event adjudication, which limits but does not eliminate the concern.
  • Different populations. The outcome trials differ in timing, background therapy, and endpoint definitions.
  • Achieved-LDL and open-label analyses are vulnerable to selection and confounding.
  • Genetic studies estimate lifelong effects of specific variants; their effect sizes cannot be transferred directly to drugs started in adulthood.
  • Abstract-level sources. The following were reviewed at the abstract or published-summary level rather than in full text: the HUYGENS primary report (supplemented by its registry results) and ApoB analysis, Cohen 2006 (PCSK9 variants), Peloso 2019 (APOB variants), Benn 2010 and Benn 2019 (PCSK9 R46L), and Ference 2012 (Mendelian randomization). These are also marked in the reference list.
  • Sources of figure data. Each data figure plots values reported in the study cited in its caption; conceptual and model figures are labeled as illustrative.
  • Review method. This narrative review is based on selected peer-reviewed studies and guidelines rather than a registered systematic search. Some findings were extracted from peer-reviewed abstracts or published summaries. Numbers not supported by the sources consulted were excluded. Figure values were checked against the published abstracts or primary reports listed in the companion figure register, and each statement in the text is supported by its cited source at the level indicated above (abstract or full text).

References

  1. Nicholls SJ, Kataoka Y, Nissen SE, et al. Effect of evolocumab on coronary plaque phenotype and burden in statin-treated patients following myocardial infarction (HUYGENS). JACC Cardiovasc Imaging. 2022;15(7):1308-1321. doi:10.1016/j.jcmg.2022.03.002. PubMed: https://pubmed.ncbi.nlm.nih.gov/35431172/
  2. Sabatine MS, Giugliano RP, Keech AC, et al. Evolocumab and clinical outcomes in patients with cardiovascular disease (FOURIER). N Engl J Med. 2017;376:1713-1722. doi:10.1056/NEJMoa1615664. PubMed: https://pubmed.ncbi.nlm.nih.gov/28304224/
  3. Schwartz GG, Steg PG, Szarek M, et al. Alirocumab and cardiovascular outcomes after acute coronary syndrome (ODYSSEY OUTCOMES). N Engl J Med. 2018;379:2097-2107. doi:10.1056/NEJMoa1801174. PubMed: https://pubmed.ncbi.nlm.nih.gov/30403574/
  4. Cannon CP, Blazing MA, Giugliano RP, et al. Ezetimibe added to statin therapy after acute coronary syndromes (IMPROVE-IT). N Engl J Med. 2015;372:2387-2397. doi:10.1056/NEJMoa1410489. PubMed: https://pubmed.ncbi.nlm.nih.gov/26039521/
  5. Cohen JC, Boerwinkle E, Mosley TH Jr, Hobbs HH. Sequence variations in PCSK9, low LDL, and protection against coronary heart disease. N Engl J Med. 2006;354:1264-1272. doi:10.1056/NEJMoa054013 [abstract reviewed]
  6. Ference BA, Yoo W, Alesh I, et al. Effect of long-term exposure to lower low-density lipoprotein cholesterol beginning early in life on the risk of coronary heart disease: a Mendelian randomization analysis. J Am Coll Cardiol. 2012;60:2631-2639. doi:10.1016/j.jacc.2012.09.017 [abstract summary reviewed]
  7. Räber L, Ueki Y, Otsuka T, et al. Effect of alirocumab added to high-intensity statin therapy on coronary atherosclerosis in patients with acute myocardial infarction: the PACMAN-AMI randomized clinical trial. JAMA. 2022;327:1771-1781. doi:10.1001/jama.2022.5218
  8. Stone GW, Maehara A, Lansky AJ, et al. A prospective natural-history study of coronary atherosclerosis (PROSPECT). N Engl J Med. 2011;364:226-235. doi:10.1056/NEJMoa1002358
  9. Nicholls SJ, Nissen SE, Prati F, et al. Assessing the impact of PCSK9 inhibition on coronary plaque phenotype with optical coherence tomography: rationale and design of the randomized, placebo-controlled HUYGENS study. Cardiovasc Diagn Ther. 2021;11(1):120-129. doi:10.21037/cdt-20-684. PMCID: PMC7944215. PubMed: https://pubmed.ncbi.nlm.nih.gov/33708484/
  10. Cannon CP, Braunwald E, McCabe CH, et al. Intensive versus moderate lipid lowering with statins after acute coronary syndromes (PROVE IT–TIMI 22). N Engl J Med. 2004;350:1495-1504. doi:10.1056/NEJMoa040583. PubMed: https://pubmed.ncbi.nlm.nih.gov/15007110/
  11. Murphy SA, Pedersen TR, Gaciong ZA, et al. Effect of the PCSK9 inhibitor evolocumab on total cardiovascular events in patients with cardiovascular disease: a prespecified analysis from the FOURIER trial. JAMA Cardiol. 2019;4:613-619. doi:10.1001/jamacardio.2019.0886
  12. Cholesterol Treatment Trialists’ (CTT) Collaboration. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376:1670-1681. doi:10.1016/S0140-6736(10)61350-5. PubMed: https://pubmed.ncbi.nlm.nih.gov/21067804/
  13. Silverman MG, Ference BA, Im K, et al. Association between lowering LDL-C and cardiovascular risk reduction among different therapeutic interventions: a systematic review and meta-analysis. JAMA. 2016;316:1289-1297. doi:10.1001/jama.2016.13985
  14. O’Donoghue ML, Giugliano RP, Wiviott SD, et al. Long-term evolocumab in patients with established atherosclerotic cardiovascular disease (FOURIER-OLE). Circulation. 2022;146:1109-1119. doi:10.1161/CIRCULATIONAHA.122.061620
  15. Hopewell JC, Malik R, Valdés-Márquez E, Worrall BB, Collins R; METASTROKE Collaboration of the ISGC. Differential effects of PCSK9 variants on risk of coronary disease and ischaemic stroke. Eur Heart J. 2018;39:354-359. doi:10.1093/eurheartj/ehx373
  16. Benn M, Tybjærg-Hansen A, Nordestgaard BG. Low LDL cholesterol by PCSK9 variation reduces cardiovascular mortality. J Am Coll Cardiol. 2019;73:3102-3114. doi:10.1016/j.jacc.2019.03.517 [abstract reviewed]
  17. Giugliano RP, Pedersen TR, Park JG, et al. Clinical efficacy and safety of achieving very low LDL-cholesterol concentrations with the PCSK9 inhibitor evolocumab: a prespecified secondary analysis of the FOURIER trial. Lancet. 2017;390:1962-1971. doi:10.1016/S0140-6736(17)32290-0
  18. Gaba P, O’Donoghue ML, Park JG, et al. Association between achieved low-density lipoprotein cholesterol levels and long-term cardiovascular and safety outcomes: an analysis of FOURIER-OLE. Circulation. 2023;147:1192-1203. doi:10.1161/CIRCULATIONAHA.122.063399
  19. Ridker PM, Everett BM, Thuren T, et al. Antiinflammatory therapy with canakinumab for atherosclerotic disease (CANTOS). N Engl J Med. 2017;377:1119-1131. doi:10.1056/NEJMoa1707914
  20. Tardif JC, Kouz S, Waters DD, et al. Efficacy and safety of low-dose colchicine after myocardial infarction (COLCOT). N Engl J Med. 2019;381:2497-2505. doi:10.1056/NEJMoa1912388
  21. Giugliano RP, Mach F, Zavitz K, et al. Cognitive function in a randomized trial of evolocumab (EBBINGHAUS). N Engl J Med. 2017;377:633-643. doi:10.1056/NEJMoa1701131. PubMed: https://pubmed.ncbi.nlm.nih.gov/28813214/
  22. Sattar N, Preiss D, Murray HM, et al. Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials. Lancet. 2010;375:735-742. doi:10.1016/S0140-6736(09)61965-6
  23. Blumenthal RS, et al. 2026 ACC/AHA/AACVPR/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management of dyslipidemia. Circulation. 2026;153(17). doi:10.1161/CIR.0000000000001423. Copublished in J Am Coll Cardiol. 2026;87(19):2624-2757. doi:10.1016/j.jacc.2025.11.016. PubMed: https://pubmed.ncbi.nlm.nih.gov/41824552/
  24. Montalescot G, Ferrari E, Souteyrand G, et al. LDL cholesterol lowering with evolocumab before percutaneous coronary intervention for acute myocardial infarction: the AMUNDSEN randomized clinical trial. JAMA. Published online August 29, 2026. doi:10.1001/jama.2026.17302. PubMed: https://pubmed.ncbi.nlm.nih.gov/42666092/
  25. Burke AP, Farb A, Malcom GT, et al. Coronary risk factors and plaque morphology in men with coronary disease who died suddenly. N Engl J Med. 1997;336:1276-1282. doi:10.1056/NEJM199705013361802
  26. Jia H, Abtahian F, Aguirre AD, et al. In vivo diagnosis of plaque erosion and calcified nodule in patients with acute coronary syndrome by intravascular optical coherence tomography. J Am Coll Cardiol. 2013;62:1748-1758. doi:10.1016/j.jacc.2013.05.071
  27. Fujino M, Nicholls SJ. PCSK9 inhibitors and coronary atherosclerosis: insights from recent intravascular imaging clinical trials. Cardiometab Syndr J. 2023;3(2):103-109. doi:10.51789/cmsj.2023.3.e14
  28. Di Giovanni G, et al. Impact of evolocumab on plaque phenotypic changes in patients with acute coronary syndrome and elevated lipoprotein(a) levels: a HUYGENS secondary analysis. Eur J Prev Cardiol. 2026;33(11):2047-2054. doi:10.1093/eurjpc/zwaf211
  29. Fujino M, et al. Achieved levels of apolipoprotein B and plaque composition after acute coronary syndromes: insights from HUYGENS. Atherosclerosis. 2025;403:119145. doi:10.1016/j.atherosclerosis.2025.119145 [abstract reviewed]
  30. Nicholls SJ, Puri R, Anderson T, et al. Effect of evolocumab on progression of coronary disease in statin-treated patients: the GLAGOV randomized clinical trial. JAMA. 2016;316:2373-2384. doi:10.1001/jama.2016.16951. PubMed: https://pubmed.ncbi.nlm.nih.gov/27846344/
  31. Nicholls SJ, et al. Effect of evolocumab on coronary plaque composition. J Am Coll Cardiol. 2018;72:2012-2021. doi:10.1016/j.jacc.2018.06.078
  32. Nissen SE, Tuzcu EM, Schoenhagen P, et al. Effect of intensive compared with moderate lipid-lowering therapy on progression of coronary atherosclerosis: a randomized controlled trial (REVERSAL). JAMA. 2004;291:1071-1080. doi:10.1001/jama.291.9.1071
  33. Nissen SE, Nicholls SJ, Sipahi I, et al. Effect of very high-intensity statin therapy on regression of coronary atherosclerosis: the ASTEROID trial. JAMA. 2006;295:1556-1565. doi:10.1001/jama.295.13.jpc60002
  34. Nicholls SJ, Ballantyne CM, Barter PJ, et al. Effect of two intensive statin regimens on progression of coronary disease (SATURN). N Engl J Med. 2011;365:2078-2087. doi:10.1056/NEJMoa1110874. PubMed: https://pubmed.ncbi.nlm.nih.gov/22085316/
  35. Tsujita K, Sugiyama S, Sumida H, et al. Impact of dual lipid-lowering strategy with ezetimibe and atorvastatin on coronary plaque regression in patients with percutaneous coronary intervention: the multicenter randomized controlled PRECISE-IVUS trial. J Am Coll Cardiol. 2015;66:495-507. doi:10.1016/j.jacc.2015.05.065. PubMed: https://pubmed.ncbi.nlm.nih.gov/26227186/
  36. Komukai K, Kubo T, Kitabata H, et al. Effect of atorvastatin therapy on fibrous cap thickness in coronary atherosclerotic plaque as assessed by optical coherence tomography: the EASY-FIT study. J Am Coll Cardiol. 2014;64:2207-2217. doi:10.1016/j.jacc.2014.08.045
  37. Kini AS, Vengrenyuk Y, Liu J, et al. Intracoronary imaging and transcriptomic characteristics before and after maximum lipid-lowering therapy: YELLOW-III study of evolocumab and maximum statin therapy in chronic coronary disease. JACC Cardiovasc Imaging. 2026;19(4):516-534. doi:10.1016/j.jcmg.2025.12.013. PubMed: https://pubmed.ncbi.nlm.nih.gov/41758107/
  38. Han D, et al. Effects of evolocumab on coronary plaque composition and microcalcification activity by coronary PET and CT angiography. JACC Cardiovasc Imaging. 2025;18:589-599. doi:10.1016/j.jcmg.2025.01.005. PubMed: https://pubmed.ncbi.nlm.nih.gov/40178463/
  39. Ako J, Hibi K, Tsujita K, et al. Effect of alirocumab on coronary atheroma volume in Japanese patients with acute coronary syndrome: the ODYSSEY J-IVUS trial. Circ J. 2019;83(10):2025-2033. doi:10.1253/circj.CJ-19-0412
  40. Prati F, Romagnoli E, Gatto L, et al. Relationship between coronary plaque morphology of the left anterior descending artery and 12 months clinical outcome: the CLIMA study. Eur Heart J. 2020;41:383-391. doi:10.1093/eurheartj/ehz520
  41. Puri R, Nicholls SJ, Shao M, et al. Impact of statins on serial coronary calcification during atheroma progression and regression. J Am Coll Cardiol. 2015;65:1273-1282. doi:10.1016/j.jacc.2015.01.036
  42. Lee SE, Chang HJ, Sung JM, et al. Effects of statins on coronary atherosclerotic plaques: the PARADIGM study. JACC Cardiovasc Imaging. 2018;11:1475-1484. doi:10.1016/j.jcmg.2018.04.015
  43. Schwartz GG, Olsson AG, Ezekowitz MD, et al. Effects of atorvastatin on early recurrent ischemic events in acute coronary syndromes: the MIRACL study. JAMA. 2001;285:1711-1718. doi:10.1001/jama.285.13.1711. PubMed: https://pubmed.ncbi.nlm.nih.gov/11277825/
  44. Ference BA, Ginsberg HN, Graham I, et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel. Eur Heart J. 2017;38:2459-2472. doi:10.1093/eurheartj/ehx144
  45. Schwartz GG, Szarek M, Bittner VA, et al. Transiently achieved very low LDL-cholesterol levels by statin and alirocumab after acute coronary syndrome are associated with cardiovascular risk reduction: the ODYSSEY OUTCOMES trial. Eur Heart J. 2023;44:1408-1417. doi:10.1093/eurheartj/ehad144. PubMed: https://pubmed.ncbi.nlm.nih.gov/36879424/
  46. Nidorf SM, Fiolet ATL, Mosterd A, et al. Colchicine in patients with chronic coronary disease (LoDoCo2). N Engl J Med. 2020;383:1838-1847. doi:10.1056/NEJMoa2021372
  47. Peloso GM, Nomura A, Khera AV, et al. Rare protein-truncating variants in APOB, lower low-density lipoprotein cholesterol, and protection against coronary heart disease. Circ Genom Precis Med. 2019;12:e002376. doi:10.1161/CIRCGEN.118.002376 [abstract reviewed]
  48. Benn M, Nordestgaard BG, Grande P, Schnohr P, Tybjaerg-Hansen A. PCSK9 R46L, low-density lipoprotein cholesterol levels, and risk of ischemic heart disease: 3 independent studies and meta-analyses. J Am Coll Cardiol. 2010;55:2833-2842. doi:10.1016/j.jacc.2010.02.044 [abstract reviewed]
  49. Zhao Z, Tuakli-Wosornu Y, Lagace TA, et al. Molecular characterization of loss-of-function mutations in PCSK9 and identification of a compound heterozygote. Am J Hum Genet. 2006;79(3):514-523. doi:10.1086/507488. PMCID: PMC1559532
  50. Amarenco P, Bogousslavsky J, Callahan A 3rd, et al. High-dose atorvastatin after stroke or transient ischemic attack (SPARCL). N Engl J Med. 2006;355:549-559. doi:10.1056/NEJMoa061894
  51. Rao SV, O’Donoghue ML, Ruel M, et al. 2025 ACC/AHA/ACEP/NAEMSP/SCAI guideline for the management of patients with acute coronary syndromes. J Am Coll Cardiol. 2025;85:2135-2237. doi:10.1016/j.jacc.2024.11.009. PubMed: https://pubmed.ncbi.nlm.nih.gov/40013746/
  52. Byrne RA, Rossello X, Coughlan JJ, et al. 2023 ESC guidelines for the management of acute coronary syndromes. Eur Heart J Acute Cardiovasc Care. 2024;13:55-161. doi:10.1093/ehjacc/zuad107. PubMed: https://pubmed.ncbi.nlm.nih.gov/37740496/
  53. Mach F, Baigent C, Catapano AL, et al. 2019 ESC/EAS guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Atherosclerosis. 2019;290:140-205. doi:10.1016/j.atherosclerosis.2019.08.014. PubMed: https://pubmed.ncbi.nlm.nih.gov/31591002/
  54. Koskinas KC, Häner J, Ueki Y, et al. Association of lipoprotein(a) with changes in coronary atherosclerosis in patients treated with alirocumab. Circ Cardiovasc Imaging. 2024;17:e016683. doi:10.1161/CIRCIMAGING.124.016683. PubMed: https://pubmed.ncbi.nlm.nih.gov/39561225/

Transparency Note: This article was created with assistance from AI tools. The final content has been carefully reviewed and edited by the author, who is responsible for its accuracy. This site is supported by an Advisory Board of researchers. All content is for educational purposes only and does not constitute medical advice.

Editorial and AI Transparency statement

Share This Post

Related Articles

Watch on YouTube

Get notified about new articles!

AI App

Heart Risk Calulator

Educational family-history heart risk calculator with H-score insights, visual family tree input and shareable PDF reports.

Read why this app is so important here.