Anatomy of a Rupture: How Plaque Forms, and Why It Fails

The mental model most people carry for a heart attack is plumbing: a pipe that narrows, year by year, until one day nothing gets through. It is an intuitive picture, and it is wrong in a way that matters — because it points attention at the wrong number, and it makes the actual event look like bad luck rather than the end of a long, describable process.

What is really there is a composite structure under pressure. An atherosclerotic plaque has a soft, greasy core that carries no mechanical load at all, and over it a cap of collagen a fraction of a millimetre thick that carries all of it. Blood pressure pushes outward against that cap roughly a hundred thousand times a day. For decades it holds. Then, in some people, on some morning, it does not.

The model below lets you build that structure from scratch — one ApoB particle at a time — and then load it until it fails. There are two numbers on it worth watching: the stress in the cap, and the strength of the cap. Everything else in the disease is a way of moving one of those two lines towards the other.

How to read the model

The artery is shown in longitudinal section, cut open lengthways, with blood flowing left to right and the plaque growing from the lower wall. The circular inset at the top right is the same lesion in cross-section — the view an angiogram or an intravascular ultrasound would give you. The panel splits into what the model tells you and what you tell the model.

• •Failure analysis (read-only). The safety factor is the headline: cap strength divided by peak cap stress. Above about 3 the structure is over-built, around 1.5 it is marginal, and at 1.0 it fails. Below it are three meters — the stress against the strength line, the cap thickness against the 65 micrometre threshold, and the stenosis, which is deliberately faded, because it is the number that misleads.
• •Drive the lesion. Two tabs. "Grow it from particles" runs an agent simulation: set an ApoB level, press Run, and watch particles enter the wall, be retained, oxidise, recruit monocytes, become foam cells, die, and leave a necrotic core behind. A simulated age counter runs alongside it. "Set it directly" skips the biology and hands you the four mechanical variables: cap thickness, core size, inflammation and calcification.
• •Load and events. Systolic pressure sets the load. The event buttons compress real timescales: twelve years untreated, eighteen months of a high-intensity statin, two years of training, a single pressure surge, or the alternative failure mode — erosion.

The phenotype label in the top left corner names what you have built, using the pathology classification: adaptive intima, intimal thickening, fibroatheroma, thin-cap fibroatheroma, or fibrocalcific. Rupture is not scripted — when stress exceeds strength, the cap tears, whatever combination of settings got you there.

Part one: how the plaque gets built

Press Run with the ApoB slider around 130 mg/dL and let it play. The sequence the particles go through is the sequence in the pathology literature, and each step is a place where the process can be interrupted.

• 1.Entry. ApoB-containing lipoproteins cross the endothelium into the artery wall. This is not damage — it happens continuously in everyone, and the flux is proportional to how many particles are in the blood. Sites where flow is slow and disturbed, like the inside of a branch point, are more permeable, which is why plaques appear in the same handful of locations in almost everybody.
• 2.Retention. Most particles leave again. Some bind to the proteoglycan chains of the wall matrix and stay. This is the rate-limiting step of the whole disease — the response-to-retention model — and it is why residence time matters as much as concentration.
• 3.Modification. A trapped particle oxidises and aggregates. It is no longer recognised by the normal LDL receptor and has become something the immune system reads as damage.
• 4.Uptake and recruitment. The activated endothelium displays adhesion molecules; monocytes stick, migrate in, and mature into macrophages that take up the modified particles through scavenger receptors. Here is the design flaw: the ordinary LDL receptor switches off when a cell has enough cholesterol, but scavenger receptors do not. The macrophage keeps eating until it is a bloated foam cell.
• 5.Death. Foam cells accumulate cholesterol crystals, hit endoplasmic reticulum stress, and die. Early on this is harmless — neighbouring macrophages clear the corpses, a process called efferocytosis, and the wall stays tidy.
• 6.Failed clearance. In an advanced lesion that cleanup breaks down. The receptors that recognise dying cells are cleaved, dying cells display "do not eat me" signals, and the corpses undergo secondary necrosis instead, spilling their contents. That spill is the necrotic core: extracellular lipid, cholesterol crystals, proteases and tissue factor. Mechanically it is dead weight. Chemically it is loaded.

Turn the ApoB slider down below about 60 while the simulation runs and the direction reverses: the entry flux falls below the rate at which lipid leaves, and the core begins to drain. That asymmetry — a core that fills at high particle concentrations and drains at low ones — is the whole rationale for lipid-lowering therapy, expressed as a single slider.

Why the stenosis number lies to you

Watch the faded stenosis meter as the lesion grows. For a long time it barely moves, even as the plaque mass underneath it climbs. That is not a quirk of the model — it is compensatory remodelling, described by Glagov in 1987 and one of the most consequential findings in the field. As plaque accumulates, the artery expands outwards, so the channel through the middle stays roughly the same size while the wall thickens substantially.

The consequence is stark: for the first several decades, the disease progresses and the thing we historically measured — the lumen — does not change. An angiogram is a picture of the hole, not of the wall. Only once outward expansion runs out, typically after plaque occupies something like 40 percent of the vessel's cross-sectional area, does the lumen finally begin to close.

Set the core large, the cap thin and the inflammation high, and look at the two meters together. You can produce a lesion at 30 percent stenosis with a safety factor under 1.5. An angiogram would report that vessel as moderate and unremarkable. The stress meter disagrees, and in that disagreement sits most of the counterintuitive epidemiology of coronary disease — including the finding, replicated many times, that most infarctions arise from lesions that were less than 70 percent narrowed beforehand.

Part two: the mechanics of failure

Once you stop thinking about blockage and start thinking about structure, the physics is almost embarrassingly simple. Stress in the cap scales roughly as pressure times radius divided by cap thickness — the Laplace relation for a thin-walled pressure vessel, with a concentration factor for the geometry. Three levers, and the model gives you all three.

• •Cap thickness is in the denominator. Halve it and you double the stress. Drag the cap slider from 180 down to 90 micrometres and watch the safety factor collapse while nothing else about the lesion changes — same core, same pressure, same narrowing.
• •The core concentrates stress at the shoulders. Because the lipid pool carries no load, the forces route around it and pile up where the cap meets the stiffer wall on either side. That is where caps actually tear in pathology specimens, and it is why a bigger core makes things worse even when the cap thickness is unchanged.
• •Pressure is the load itself. It is also the only one of the three that changes minute to minute. Press ⚡ Pressure surge on a marginal lesion — the settings do not change, the load does, and that is enough.

Cap stresses in the region of 300 kPa are where ruptured caps cluster in the finite-element work on this question, and that number is the reference the model's strength line is built around. It is a population landmark, not a threshold any individual plaque respects.

The other half: strength, and what eats it

Stress alone cannot tell you whether something breaks — you need to know what it can carry. The cap's strength comes from type I collagen, laid down and continuously repaired by smooth muscle cells. Inflammation attacks both sides of that balance: macrophage-derived matrix metalloproteinases digest existing collagen, while inflammatory signalling suppresses the smooth muscle cells that would replace it.

This is worth pausing on, because it is the least intuitive part of the whole picture. Move the inflammation slider and watch what happens: the cap thickness meter does not move at all. The strength marker slides left instead, and the safety factor falls. A cap can be structurally the same thickness and materially far weaker — which is precisely why inflammatory markers add information that a picture of the artery cannot provide.

Calcification: one slider, two opposite meanings

Drag the calcification slider slowly from zero to maximum and watch the stress meter rise, peak, and then fall away. It is the only control in the model that reverses its own effect, and that is deliberate, because calcium in a plaque genuinely does two opposite things depending on how it is arranged.

In the first half of the slider the calcium is spotty: micrometre-scale deposits embedded in a cap that flexes with every heartbeat. Hard inclusions in a soft flexing membrane are classic stress risers — the local stress at the interface can be several times the background — and modelling work has shown microcalcifications sitting inside the cap can amplify stress enough to matter. In the second half those deposits fuse into a dense confluent plate. Now the plaque cannot flex at all. The load is carried by a rigid shield, the cap stops fatiguing, and the lesion becomes one of the stable, boring ones.

Press 💊 Statin and watch this play out as a package: the core shrinks, inflammation falls, the cap thickens — and the calcium goes up, consolidating into the dense form. This is the resolution of a question that confuses a lot of people looking at repeat calcium scans. A rising calcium score on effective therapy is not the disease winning. It is the lesion turning into scar.

The minutes after the tear

Take a lesion with a safety factor near 1 and press ⚡. The cascade text under the failure panel follows the clock, and every line has a clinical counterpart.

• •t = 0. The cap tears at the upstream shoulder. Tissue factor, the most potent initiator of coagulation in the body and abundant in the necrotic core, meets flowing blood for the first time.
• •Seconds. Platelets adhere to the exposed collagen and core material, activate, change shape, and recruit more platelets.
• •Minutes. The coagulation cascade lays fibrin across the platelet plug. The clot organises and grows into the lumen, and flow past it slows — you can see the blood cells in the model begin to stall.
• •Occlusion. The vessel closes. Whatever the stenosis read a moment earlier — 30 percent, 50 percent, it hardly matters — it is now 100 percent, and the muscle downstream is infarcting.

That last transition is the entire argument of this page in one line. The event was never about the narrowing. It was about a sealed bag of thrombogenic tissue held in by a membrane that had been getting thinner for twenty years.

Press ◌ Erosion to see the alternative. Here the cap never tears; the endothelial lining strips off an intact surface and a clot builds on the exposed collagen. Intracoronary imaging suggests this accounts for around a third of acute coronary syndromes, skews younger, female and towards smokers, and typically produces a smaller mural clot rather than a total occlusion — which is one reason it more often presents without ST elevation. Same artery, same consequence, different failure mode.

What actually moves the safety factor

The model has three intervention buttons, and the honest summary of decades of cardiovascular research is that they act on the two terms of one fraction.

Lower the particle exposure. Plaque volume is driven by ApoB-particle concentration multiplied by years. That multiplication is why treatment started earlier does disproportionately more than the same treatment started late, and why ApoB — a direct count of the particles that get retained — reads the process more directly than the cholesterol they happen to be carrying. Everything on your lipid panel is an approximation of that count.

Take the load off. Blood pressure appears directly in the stress equation, on every beat, in every plaque simultaneously. Press 🏃 Exercise and watch resting pressure fall while the calcium consolidates — the same lump of plaque, moved into a different failure regime.

Protect the collagen. Anything that reduces the inflammatory digestion of the cap — not smoking above all, along with glycaemic control and treating the sources of chronic inflammation — moves the strength line rather than the stress line. Both arrive at the same ratio.

Frequently asked questions

The honest limits of this model

The mechanics are anchored to the vascular literature — the critical cap stress around 300 kPa from the finite-element work of Cheng and colleagues, the 65 micrometre thin-cap definition from Virmani's autopsy series, compensatory remodelling from Glagov, and the observation that most culprit lesions were previously under 70 percent stenosis. But this is a two-dimensional analytical model, not a finite-element simulation. Real cap stress fields depend on three-dimensional geometry, the anisotropic and non-linear behaviour of arterial tissue, and residual stresses in the wall that no slider here represents.

The pathogenesis simulation compresses forty years of cell biology into a few hundred moving dots. The agents are illustrative: the timing constants were chosen so the sequence is legible at human speed, not measured. And rupture in the model is deterministic — stress crosses strength and the cap tears — whereas in life it is probabilistic, which is why many people carry thin-cap fibroatheromas that never fail. Autopsy studies routinely find healed, silent ruptures in people who died of something else entirely.

Use it to understand why coronary disease behaves as strangely as it does: why the silent decades are silent, why the tightest lesion is not the dangerous one, and why lipids and inflammation are separate levers. Do not use it to estimate your own risk.