Arterial plaque.

Before the scan you were arguing about risk — a probability calculated from your age, your numbers and a population table. After the scan you are not arguing about risk. You have the disease. Imaging didn't predict atherosclerosis; it found it.

That distinction is not semantic, and it is the reason your doctor changed tone. Nearly every controversy you've read about cholesterol drugs — whether they're overprescribed, whether the benefit is worth it, whether the absolute risk reduction is trivial — is an argument about treating people with no known disease. You are no longer in that group.

The benefit of lowering ApoB is roughly proportional to how much disease you already have. In someone with clean arteries, a 30% relative risk reduction applies to a tiny absolute risk, and the numbers-needed-to-treat look unimpressive — which is the honest core of the statin scepticism. Apply the same relative reduction to someone with documented plaque and the absolute benefit multiplies, because the baseline risk is many times higher.

So if you were sceptical about lipid-lowering drugs before the scan, that scepticism was defensible. Carrying it unchanged after the scan is applying the right argument to the wrong person.

Finding it early is the win. Plaque is not a one-way street: the imaging trials of the last twenty years have repeatedly shown that if you push ApoB low enough for long enough, plaque volume shrinks and — more importantly — the dangerous soft plaque converts to stable, calcified, boring plaque. People who find this at 45 with a scan are in a far better position than people who find it at 55 in an ambulance.

‘Widow maker’ isn't a type of plaque — it's a location. It refers to a severe blockage in the proximal left anterior descending artery (LAD), the vessel that supplies the front wall of the left ventricle. Occlude it near its origin and you lose a large share of the pumping muscle at once, which is why that particular lesion has a reputation the others don't.

Two people can have identical total plaque volume and very different risk, because one has it spread through small distal branches and the other has it concentrated proximally in the LAD. This is exactly what a plain calcium score cannot tell you and a proper CT angiogram can — which is the case for the imaging in the next section.

Having plaque in the LAD is not the same as having a widow-maker lesion. The frightening scenario is a severe, flow-limiting stenosis in that specific proximal segment. Non-obstructive plaque in the same artery is a signal to treat aggressively, not an emergency — and it is precisely the situation modern therapy handles well.

Most heart attacks are not caused by the tightest narrowing gradually closing. They're caused by a plaque with a thin fibrous cap and a soft, lipid-rich core rupturing or eroding. The contents hit the bloodstream, a clot forms in seconds, and an artery that was perhaps only 40% narrowed the day before is now 100% closed.

A majority of infarcts arise from lesions that were not severely obstructive beforehand. That's why a stress test can be clean weeks before an event: stress tests look for flow limitation, and the dangerous plaque often isn't limiting flow yet. It's also why ‘my arteries are only mildly narrowed’ is reassuring about symptoms and much less reassuring about risk.

If rupture is the mechanism, the objective isn't just less plaque — it's calmer plaque. Thick cap, small lipid core, less inflammation. That's achievable, it's measurable, and it explains why the treatments below are judged on plaque composition and not only on volume.

Lipid-rich, inflamed, thin-capped. This is the plaque that ruptures. On CT it appears as low-attenuation plaque (LAP), and its volume is one of the strongest imaging predictors of future events. High-risk features that travel with it include spotty calcification, positive remodelling (the vessel bulges outward rather than narrowing, hiding the disease from older tests) and the ‘napkin-ring sign’.

Counter-intuitively, calcium is the scar. Dense calcification is what plaque becomes when it stabilises — it's a healed lesion, mechanically stiff and much less likely to rupture. This is why aggressive lipid-lowering can paradoxically increase your calcium score while reducing your risk: soft plaque is converting to stable plaque. A rising calcium score on treatment is not automatically bad news, which is one of the most misunderstood facts in this whole area.

Falling total plaque volume is good. Falling low-attenuation volume, and a rising proportion of calcified plaque, is better — that's the disease being defused rather than merely shrunk.

A non-contrast CT, roughly 10 minutes, about 1 mSv of radiation, typically $100–300, no needle. It counts calcified plaque and returns an Agatston score. Scores: 0 (none detected), 1–99 (mild), 100–399 (moderate), 400+ (extensive).

As a population screening tool it's excellent value. A score of zero in a middle-aged person without other red flags carries a very low short-term event rate — the so-called ‘power of zero’ — and a high score reliably identifies people who should be treated hard. As a one-off tie-breaker on whether to start treatment, it earns its cost.

It sees only calcium. It is physically incapable of detecting soft, non-calcified plaque — the kind that ruptures. A person can have substantial lipid-rich plaque and a calcium score of zero, and this is most common in exactly the group that most needs to know: people under about 50, and people with a high burden of recent, uncalcified disease. A zero score means ‘no calcified plaque detected’. It does not mean ‘no plaque’.

It also tells you nothing about where the plaque sits or whether any lesion is flow-limiting.

A CCTA uses iodinated contrast so the lumen and the vessel wall are both visible. That single change lets it see all plaque — calcified, mixed and soft — plus the degree of narrowing in each segment and where the disease actually sits. Roughly 3–5 mSv on older protocols, often lower on modern scanners; typically $500–1,500 self-pay.

A conventional CCTA report is read by a human radiologist and usually reports stenosis in broad bands — ‘mild’, ‘moderate’, ‘severe’ — with a qualitative comment on plaque. That's fine for answering ‘is there a blockage causing symptoms?’ but it isn't quantitative enough to track change over time. Two readers can describe the same scan differently, and ‘mild non-obstructive plaque’ a year later tells you almost nothing about whether you're winning.

Needs a reasonably controlled heart rate (often a beta-blocker before the scan), contrast (so kidney function and contrast allergy matter), and a scanner and reader who do this regularly.

Cleerly is an FDA-cleared AI analysis layered on top of a CCTA — not a different scan. You have a normal CT angiogram; the images are sent for automated analysis that segments every coronary artery and quantifies plaque at every segment.

It converts a descriptive report into numbers: total plaque volume, and volume broken out by type — low-attenuation (soft), fibrofatty, fibrous and calcified — plus percent atheroma volume and stenosis grading per segment. Instead of ‘mild plaque in the LAD’, you get a quantity attached to an artery.

Three practical consequences. First, it directly measures the thing that predicts rupture, low-attenuation plaque volume, rather than the thing that's easy to see. Second, because it's automated and quantitative it's reproducible — the same analysis repeated in two years produces a comparable number, so you can tell whether you're regressing, stable or progressing. Third, it removes reader-to-reader variability from that comparison.

It is still a CCTA underneath: same contrast, same radiation, same need for rate control, and the AI can only analyse image quality it's given. It costs more than the scan alone (often $500–1,500 on top, frequently self-pay). And while quantitative plaque burden is strongly associated with outcomes, there is not yet a randomised trial proving that managing patients by Cleerly numbers improves survival compared with treating them well on ApoB alone. It is the best measuring instrument available; that's a different claim from ‘proven to change outcomes’.

If a scan has already found disease and you want a baseline you can genuinely track against, CCTA with AI quantification is the right instrument. Get the quantitative report, not just the narrative one — and keep the actual numbers, because your future self needs them for comparison.

No radiation, no contrast, cheap, repeatable. It measures wall thickness and looks for plaque in the neck arteries as a proxy for systemic atherosclerosis. Presence of carotid plaque is meaningfully predictive; the intima-media thickness number alone is a weaker signal. Useful as a low-cost interval check between CT scans, and it's the imaging used in most of the supplement trials below.

Answers a different question entirely: ‘is a narrowing severe enough to limit blood flow under load?’ It's a symptom and ischaemia test, not a plaque test. A normal stress test does not mean you don't have plaque — and given that most infarcts come from non-obstructive lesions, a clean result should never be read as an all-clear on atherosclerosis.

The catheter-based reference standard, and the tool used in the classic plaque-regression trials. It's for people with symptoms or severe disease being considered for intervention — not for monitoring someone who feels fine.

If you're starting from scratch and want the fullest picture: a CCTA with quantitative AI plaque analysis, plus the blood panel in the next section. If cost is the constraint: a calcium score plus ApoB and Lp(a) will still change your management, and a zero score with a bad ApoB should not stop you treating.

Every atherogenic particle — LDL, VLDL, IDL, remnants, Lp(a) — carries exactly one apolipoprotein B molecule. So ApoB is a direct count of the particles that can enter and lodge in an artery wall. LDL-C, by contrast, measures the cholesterol cargo inside those particles, which varies. Two people with identical LDL-C can have very different particle counts, and the one with more particles has more disease. This is why ApoB is the better target, and why ‘my cholesterol is okay’ can coexist with a scan full of plaque.

Rough bands, in mg/dL: population average is around 100–110, which is not a healthy target — it's the average of a population riddled with atherosclerosis. Guidelines put roughly <80 for high risk and <65 for very high risk (established disease). The imaging trials in which plaque actually regressed generally drove particle numbers well below that — equivalent to an LDL-C in the 50s or lower.

With documented plaque, the defensible aim discussed with a cardiologist is ApoB under about 60–65 mg/dL, and lower if the disease is extensive. There is no established threshold below which lowering stops helping or starts harming; the trials that went lowest did best.

Down, and then stays down. Plaque responds to cumulative exposure over years, not to a good quarter. A single ApoB reading is a snapshot; the area under the curve over a decade is what your arteries actually experience. Check about 6–8 weeks after any dose change, then every 6–12 months once stable.

An LDL-like particle with an extra protein attached. It is roughly 80–90% genetically determined, barely moves with diet or exercise, and is both atherogenic and prothrombotic. Around one in five people carry an elevated level, and most have never been tested. It is a common explanation for exactly your partner's situation: real plaque with unremarkable standard cholesterol.

Elevated is usually taken as >50 mg/dL or >125 nmol/L (nmol/L is the better unit — insist on it). You measure it once: it doesn't meaningfully change, so it's a risk-stratifier and not a progress marker. If it's high, the response today is not to chase the number but to drive everything else — ApoB, blood pressure, inflammation — harder to compensate.

This is the most active area in the field. Three drugs are in phase 3 outcome trials: pelacarsen (Novartis, Lp(a)HORIZON — 8,323 patients with established cardiovascular disease and Lp(a) ≥70 mg/dL, primary follow-up completed mid-2026, filings expected thereafter), olpasiran (Amgen, OCEAN(a)-OUTCOMES — 7,297 patients, completing late 2026) and lepodisiran (Lilly, phase 3 data expected 2029). Phase 2 reductions were dramatic — 80% for pelacarsen, over 90% for olpasiran and lepodisiran. What nobody knows yet is whether lowering Lp(a) lowers events; that is precisely what these trials will answer, with the first outcome data due imminently. If you have high Lp(a) and known plaque, this is the space to watch.

High-sensitivity C-reactive protein is a general marker of inflammation. Atherosclerosis is an inflammatory disease, and residual inflammatory risk is a real phenomenon: people who get their ApoB to target but keep a raised hs-CRP still have elevated event rates. It's the second axis after particles.

Conventional bands are <1 mg/L low, 1–3 average, >3 high. With established plaque, aim <1. Interpret it carefully: any infection, injury, recent hard training session or flare will spike it, so never act on a single value — repeat it a few weeks later before concluding anything.

Visceral fat loss, exercise, sleep, smoking cessation, treating gum disease, and statins (which lower hs-CRP independently of LDL). Colchicine has randomised evidence for reducing events in coronary disease via inflammation and is prescribed in some clinics for exactly this — a conversation for your cardiologist.

Fasting insulin, HOMA-IR, triglyceride-to-HDL ratio and HbA1c. Insulin resistance drives small dense LDL, raises triglyceride-rich remnants, damages endothelium and raises blood pressure — it's an engine behind plaque even when LDL-C looks fine. A triglyceride/HDL ratio above about 2 (mg/dL) is a useful rough flag. This is frequently the missing piece in the ‘good cholesterol but real plaque’ picture.

Not a blood test, but it belongs in the same list — mechanical stress on the artery wall is an independent driver of plaque growth and rupture. Home cuff, upper arm, validated device. With established disease, most cardiologists target well under 130/80.

Lp-PLA2 and myeloperoxidase (vascular-specific inflammation), homocysteine, ferritin and thyroid function. These refine the picture; none of them displaces ApoB, blood pressure and inflammation as the things you actually manage.

Every 6–8 weeks after any change: ApoB and a lipid panel, plus liver enzymes if starting a statin. Every 6–12 months when stable: ApoB, hs-CRP, HbA1c, fasting insulin, blood pressure log. Once, ever: Lp(a).

There's no consensus interval, and more imaging is not automatically better. A common practical approach is a repeat quantitative CCTA at 18–24 months after a major treatment change — long enough for plaque composition to shift measurably, short enough to correct course. Repeating it annually mostly buys radiation and anxiety. Carotid ultrasound is a reasonable no-radiation interval check in between.

ApoB at or below target and holding. hs-CRP under 1. Blood pressure under 130/80. Then at rescan: total plaque volume stable or lower, low-attenuation volume lower, and a higher proportion of what remains calcified. That last combination — even with total volume flat — is a win, because the plaque has been defused.

The strongest version of the argument goes: in people without known disease, absolute risk reduction over five years is small, numbers-needed-to-treat run into the dozens or hundreds, all-cause mortality benefit in low-risk primary prevention is unimpressive, industry funded most of the trials, and side effects are under-reported. That case has real merit for a healthy 45-year-old with an average ApoB and a calcium score of zero.

Every element of that argument is about baseline risk. Relative risk reductions from lowering ApoB are remarkably consistent — roughly a fifth fewer major vascular events per 1 mmol/L (~39 mg/dL) of LDL reduction, holding across primary and secondary prevention. What differs is the size of the number you're taking a fifth of. Documented plaque moves you into a group where that fifth is worth a great deal.

Statins also do something specific to your problem: the imaging trials showed they shrink plaque and thicken fibrous caps. That's not a lipid abstraction — it's the lesion becoming less likely to rupture.

Muscle aches are the common complaint, and blinded n-of-1 trials show a large share are nocebo — symptoms recur on placebo as often as on drug. Real intolerance exists and deserves respect, but it's usually manageable by switching agent, lowering the dose, or dosing intermittently. Rhabdomyolysis is rare. New-onset diabetes risk is real and small; the cardiovascular benefit outweighs it in people with established disease. Liver enzyme elevation is usually transient.

If statins genuinely don't suit you, the honest conclusion isn't ‘skip lipid-lowering’ — it's ‘use one of the alternatives below’.

PCSK9 is a protein that degrades your LDL receptors. Block it with a monoclonal antibody (evolocumab, alirocumab — a fortnightly or monthly self-injection) and the liver keeps far more receptors, clearing LDL particles aggressively. On top of a statin they typically drop LDL-C a further 50–60%, taking people to levels statins alone essentially cannot reach.

Two large outcome trials — FOURIER (evolocumab) and ODYSSEY OUTCOMES (alirocumab) — showed significant reductions in cardiovascular events in people with established disease. Directly relevant to plaque: GLAGOV used intravascular ultrasound and showed adding evolocumab to a statin produced measurable plaque regression, with regression in around two-thirds of patients versus about half on statin alone, at a mean achieved LDL in the 30s.

The honest caveats: absolute risk reductions over the trial periods were meaningful but not enormous, FOURIER did not demonstrate a mortality benefit over its follow-up, and the trials tested PCSK9 added to statins rather than instead of them. Safety has been notably clean, including at very low LDL levels — one of the more reassuring findings in modern cardiology. Cost and insurance approval are the practical barriers.

A small interfering RNA against PCSK9, given as an injection twice a year after loading doses. Similar LDL reduction to the antibodies with unbeatable adherence. Its cardiovascular outcome trials are still reading out, so the event-reduction evidence is younger than for the antibodies.

Blocks intestinal cholesterol absorption. Modest alone (roughly 15–20% LDL reduction), cheap, generic, and very well tolerated. IMPROVE-IT showed adding it to a statin reduced events further — important because it confirmed the benefit tracks the LDL reduction rather than anything statin-specific. Often the most sensible second agent before escalating to injections.

Works upstream of statins in the same pathway but is activated only in the liver, not muscle — which is why it rarely causes muscle symptoms. CLEAR Outcomes showed event reduction in statin-intolerant patients. It's the obvious option for someone who genuinely can't take statins, and it combines with ezetimibe.

These agents are additive because they attack different steps. A common ladder: high-intensity statin, add ezetimibe, then add a PCSK9 inhibitor if ApoB is still above target — substituting bempedoic acid for the statin where intolerance is real. Most people with established plaque can reach an ApoB in the 50s with two or three of these.

Once routine, now much more selective. In primary prevention the bleeding risk substantially offsets the benefit, which is why blanket recommendations were withdrawn. In established disease the calculus shifts back toward benefit, but this is genuinely individual — it depends on your plaque burden, bleeding risk and other medications, and it is a decision for your cardiologist, not a supplement-aisle choice.

Rupture causes events by triggering clot. Lipid-lowering addresses the plaque; antiplatelet therapy addresses what happens if one ruptures anyway. They're complementary, which is also the mechanistic argument the fibrinolytic supplements below lean on — with far less evidence behind them.

This is the one with genuine outcome and imaging data. REDUCE-IT gave 4 g/day of icosapent ethyl (purified EPA, prescription — Vascepa) to people with high triglycerides on statins and found roughly a 25% relative reduction in major cardiovascular events. EVAPORATE then imaged it: over 18 months, icosapent ethyl produced significant regression of low-attenuation plaque volume versus placebo. That's the exact target — the soft plaque that ruptures.

The comparator was mineral oil, which may have raised LDL and inflammation in the placebo arm and could exaggerate the effect — a genuine and unresolved criticism. And critically, the result is for purified EPA at 4 g/day, not for a standard fish-oil capsule. Mixed EPA/DHA supplements at 1–2 g have repeatedly failed to reduce events (STRENGTH, VITAL, ASCEND). Dose and formulation are the whole story here.

If you want this effect, it's a prescription conversation about icosapent ethyl, not a trip to the supplement aisle. If using over-the-counter fish oil, understand you're taking it for triglycerides and general cardiovascular health, not for the EVAPORATE plaque effect. Watch for a modest increase in atrial fibrillation and bleeding risk at high doses, especially alongside antiplatelets.

A fibrinolytic enzyme from fermented soybeans (nattō). Mechanistically it degrades fibrin and has mild blood-pressure and lipid effects — a plausible story for plaque, since fibrin is part of what stabilises a growing lesion.

A 2022 trial in Frontiers in Cardiovascular Medicine followed 1,062 participants for a year and reported that 10,800 FU/day reduced carotid plaque size by around 36% — a large effect for any intervention, let alone a supplement. The finding was strongly dose-dependent: doses at or below 3,600 FU/day did essentially nothing. An earlier head-to-head (Ren 2017) reported nattokinase performing comparably to simvastatin on carotid plaque.

Nearly all the positive plaque work comes from a single research orbit and has not been independently replicated by other groups. Trial design details — blinding, true placebo control, imaging methodology — are weaker than a cardiology trial would demand. And a well-conducted placebo-controlled trial, NAPS, found no effect on plaque. An effect that large from an oral enzyme should be replicated before it's believed; as of now it hasn't been.

Graded honestly: ●●○○○ — a genuinely intriguing signal at high dose, not established.

The dose in the positive trials is far above what most shop products supply — typical capsules are 2,000–4,000 FU, which is in the range that failed. Check the FU (fibrinolytic units) on the label, not the milligrams. The real caution is bleeding: it is fibrinolytic, so combining it with aspirin, clopidogrel, warfarin, a DOAC, or high-dose fish oil raises bleeding risk, and it should be stopped well before any surgery or dental procedure. Anyone on antiplatelet or anticoagulant therapy — which includes many people with diagnosed plaque — should clear it with their cardiologist first rather than stacking it silently.

A proteolytic enzyme originally isolated from silkworm gut bacteria, marketed for ‘dissolving’ arterial plaque and clearing fibrin. It's frequently paired with nattokinase in high-dose cardiovascular stacks.

There is essentially no human cardiovascular outcome or imaging evidence. What exists is small, mostly older work on inflammation, swelling and post-operative or ENT indications, plus in-vitro fibrinolytic activity. The plaque claim is an extrapolation from mechanism, not a finding. Nor is there good evidence that an orally administered enzyme survives digestion intact in quantities that act systemically on coronary plaque.

Graded honestly: ●○○○○ — plausible-sounding, unproven for this purpose.

Same bleeding consideration as nattokinase, and the two are usually taken together, which compounds it. If your partner is on high doses of both alongside any antiplatelet, that combination deserves an explicit conversation with his cardiologist — not because it's known to be dangerous, but because nobody has studied it and the mechanism points at bleeding.

Blends like Neprinol AFD put serrapeptase, nattokinase, lipase and several proteases in one capsule, and are widely recommended as a straight upgrade on plain nattokinase — more fibrinolytic enzymes, therefore more effect. It's the product people most often arrive with after reading that nattokinase shrinks plaque.

There is no published controlled trial of the blend. The manufacturer's own clinical-data page reports a 12-week reduction in blood viscosity with no participant count, no design, no journal and no citation; everything else it cites is work on the individual ingredients. Combining two things that each have their own evidence doesn't inherit either one's — and here one of them, serrapeptase, is graded a single dot on this very page.

Graded honestly: ●○○○○ — a reasonable mechanism, sold well, never tested as sold.

Worth doing the arithmetic, because plaque is the one nattokinase indication where dose is the whole story: 10,800 FU/day moved carotid plaque in the 2022 trial and 3,600 FU/day did nothing at all.

Neprinol's headline is 15,000 FU of nattokinase per capsule. NSK-SD — the research-grade nattokinase those trials used — is standardised at 20,000 FU per gram, which makes 15,000 FU about 750 mg of it. The entire proprietary blend is 500 mg, and it also has to hold serrapeptase, lipase, three proteases, amla, papain, bromelain, rutin and CoQ10. That doesn't fit.

So either the nattokinase is a far more concentrated grade than the one that was studied, or — as the printed panel suggests, by grouping serrapeptase, nattokinase and lipase under a single ‘15,000 FU’ line — the figure is the combined activity of three enzymes rather than nattokinase alone. Both readings land in the same place: you can't tell which side of the 10,800 FU threshold you're on. If the nattokinase effect is what you're after, a standalone NSK-SD product states its FU and lets you aim at it. That isn't a knock on enzyme blends in general — it's the specific reason this one can't be pointed at plaque.

Several fibrinolytic enzymes at once, and the manufacturer's own label says not to take it without your physician's consent if you're on anticoagulants. The dosing tiers run to nine capsules a day and the bleeding consideration scales with them. Stop 7–14 days before any surgery or dental extraction. Alongside aspirin, a DOAC or clopidogrel, that's a conversation with the prescriber.

This one is more interesting than its reputation. Multiple small randomised, double-blind, placebo-controlled trials using cardiac CT have found aged garlic extract slows coronary calcium progression, and a randomised trial in 55 people with metabolic syndrome using CT angiography found a significantly greater reduction in percent low-attenuation plaque versus placebo over about a year at 2,400 mg/day.

The trials are small, several come from overlapping research groups, and some used aged garlic combined with B vitamins and L-arginine rather than alone. It's imaging evidence, not event reduction. But it is one of the very few supplements with randomised CT-imaging data pointed at the right target.

Graded honestly: ●●○○○ — small but real imaging signal.

Aged garlic extract specifically (the Kyolic-style preparation used in the trials), around 2,400 mg/day. Mild blood-thinning effect, so the same bleeding conversation applies. Cheap, well tolerated, and a more defensible addition than most of what's marketed for arteries.

The theory is elegant: K2 activates matrix Gla protein, which inhibits calcium deposition in soft tissue, so it should direct calcium to bone rather than arteries. Observational data are supportive and trials in kidney disease and general populations are mixed, with the largest failing to show reduced arterial calcification. Reasonable to take, particularly alongside vitamin D, but don't count it as plaque treatment. ●●○○○

Soluble fibre (psyllium), plant sterols, berberine, red yeast rice (which contains monacolin K — chemically a statin, with statin-like effects and statin-like side effects, at unpredictable doses) and citrus bergamot. These work through the mechanism that matters. None of them approaches prescription potency, and if you're targeting an ApoB in the 50s you will not get there on supplements.

Vitamin E, high-dose antioxidants, niacin (large trials failed to add benefit on top of statins and added harms), chelation therapy (marginal at best in trials, and burdensome), and the general category of ‘artery cleanse’ formulas. CoQ10 is worth taking if statins cause muscle symptoms — the evidence for that is modest — but it doesn't treat plaque.

The single highest-value action if it applies. Smoking damages endothelium, promotes clotting and accelerates plaque directly; risk falls measurably within a year of stopping. Vaping is not established as safe here — it is less studied, not proven benign.

Mechanical stress drives plaque growth and rupture independently of lipids. With established disease, most cardiologists aim well under 130/80. Measure at home with a validated upper-arm cuff rather than relying on clinic readings.

Reducing saturated fat, increasing soluble fibre, more plants, and losing visceral fat. Realistically, diet moves ApoB perhaps 10–20% — genuinely useful, and nowhere near enough on its own when the target is a halving. Treat it as the foundation the drugs build on, not the alternative to them.

If fasting insulin, HbA1c or the triglyceride/HDL ratio are off, fixing that changes particle quality as well as quantity. This is often where the biggest non-drug win sits in someone whose standard cholesterol looked fine.

Zone 2 aerobic work most days plus resistance training twice a week. Effects on ApoB are modest; effects on blood pressure, insulin sensitivity, inflammation and mortality are large. Cardiorespiratory fitness is among the strongest predictors of survival there is. If you have established coronary disease, discuss intensity limits with your cardiologist before starting hard training.

Short and fragmented sleep raise blood pressure, inflammation and insulin resistance. Untreated sleep apnoea is a genuine cardiovascular risk factor and is common, under-diagnosed, and treatable — if you snore, wake unrefreshed or have blood pressure that won't come down, get tested.

Periodontal disease is consistently associated with cardiovascular events, plausibly through chronic inflammatory load. Treating gum disease is cheap, and it's a lever most people ignore.

GLAGOV (evolocumab added to statin, intravascular ultrasound, plaque regression), ASTEROID and SATURN (high-intensity statin regression), EVAPORATE (icosapent ethyl, CT, low-attenuation plaque regression over 18 months), PARADIGM (serial CCTA and plaque progression).

FOURIER and ODYSSEY OUTCOMES (PCSK9 inhibitors), IMPROVE-IT (ezetimibe), CLEAR Outcomes (bempedoic acid), REDUCE-IT (icosapent ethyl), and the Cholesterol Treatment Trialists' meta-analyses for the per-mmol/L relationship between LDL reduction and event reduction.

Lp(a)HORIZON (pelacarsen), OCEAN(a)-OUTCOMES (olpasiran) and the lepodisiran programme, with phase 2 dose-finding data already published and the first cardiovascular outcome results expected from 2026.

The 2022 Frontiers in Cardiovascular Medicine nattokinase trial and Ren 2017, against the null NAPS trial; randomised aged-garlic-extract CT trials on calcium progression and low-attenuation plaque; and the negative mixed-omega-3 outcome trials (STRENGTH, VITAL, ASCEND) that explain why formulation and dose matter.

Cardiovascular outcome trials beat imaging trials, imaging trials beat biomarker changes, and biomarker changes beat mechanism. Most supplement claims in this area live at the mechanism level while being marketed as if they lived at the top.

A single dramatic result from one research group is a hypothesis, not a finding — which is the situation with high-dose nattokinase. Replication by independent groups is what converts one into the other, and it hasn't happened yet.

For the lipid fundamentals underneath this — how ApoB particles cause plaque in the first place, the full statin discussion, diet detail and the standard panel — see the cholesterol guide. For the blood-pressure half of the problem, see the blood pressure guide.