CAC scoring — direct imaging of plaque
Blood markers and the CGM tell you about the conditions for atherosclerosis. The Coronary Artery Calcium scan tells you whether anything is actually there. It is arguably the single most clinically useful cardiovascular imaging test.
Why CAC scoring deserves its own section
Every other tool in this guide — your lipid panel, your HbA1c, your hs-CRP, your composite risk score — is inference. They estimate how likely it is that atherosclerosis is forming inside your arteries, but none of them can see directly. A Coronary Artery Calcium (CAC) scan is different: it is a low-radiation CT scan that physically measures the calcified plaque inside your coronary arteries. It doesn't predict; it observes.
For anyone facing a difficult statin decision based on borderline biomarkers, or trying to understand whether decades of "high cholesterol" have actually caused structural damage, the CAC score frequently rewrites the picture. The clinical literature has a phrase for the most important finding here — "the power of zero".
The Agatston score — four tiers
A CAC score is reported in "Agatston units" (named after the radiologist who developed the metric). It is measured non-invasively from a short chest CT — typically under 10 minutes, ~1 mSv of radiation (less than a routine mammogram).
Zero calcified plaque detected. In asymptomatic adults, a CAC of 0 carries a markedly lower event rate over the next 5–10 years than any score above zero, even with an elevated LDL or a borderline biomarker pattern. Lower is not the same as none — the MESA figures below put a number on what a zero actually buys. In everyday terms: among people a guideline would not have offered a statin, a zero corresponded to roughly 1 event per 1,000 people per year, and any calcium at all to roughly 4 or 5. Both are small numbers; one is several times the other. This is the strongest single argument in medicine for delaying or avoiding statin therapy in a borderline patient — provided the underlying metabolic and inflammatory environment is also clean.
Important caveat: CAC=0 does not exclude non-calcified (soft) plaque. In younger people, women, and those with active disease that hasn't yet calcified, soft plaque can be present without showing on CAC. If clinical suspicion is high despite CAC=0, CCTA (CT angiography) can visualise soft plaque directly.
Some calcified plaque present, but burden is low. CV event risk remains relatively low, but the plaque-forming machinery has officially been switched on. The clinical priority shifts from "should I do anything?" to "freeze further progression." This usually means addressing the metabolic environment aggressively (glycation, inflammation, lipid pattern) and considering lipid lowering depending on age and other risk factors.
Significant plaque burden. Statistical risk of a cardiovascular event is substantially higher than CAC=0. Pharmacological lipid lowering is routinely recommended at this level — typically statins (or ezetimibe/bempedoic acid if intolerant), often alongside low-dose aspirin depending on bleeding risk. The metabolic and inflammatory axes also need active management; lipid lowering alone is not the whole answer.
Extensive calcified plaque. This warrants aggressive medical management to stabilise the plaque and prevent rupture. Beyond lipid lowering, this level often prompts further evaluation — stress testing, CT angiography to assess soft plaque burden, and in some cases referral to cardiology. The good news: calcified plaque is the stable form. The risk pattern at this level is mostly about preventing new soft plaque, not about the existing calcified plaque suddenly rupturing.
What a zero and a non-zero are actually worth — the 2026 MESA analysis
The tiers above are a scale. This is what the scale means in events, from an analysis of 5,698 adults in the MESA cohort published in JACC in July 2026, which is the most directly useful version of this question yet published: it compares CAC not against a risk score alone, but against the actual statin decision a guideline would have produced.
| What the guideline said about statins | CAC = 0 | CAC > 0 |
|---|---|---|
| Not recommended | 1.2 per 1,000 person-years | 4.5 |
| Considered | 2.7 | 5.2 |
| Recommended | 5.8 | 16.6 |
The first row is the interesting one, and it is not the row anyone expects. Among people a guideline said not to treat, those with any detectable calcium had close to four times the event rate of those with none. That is a substantial difference discovered inside the group nobody was going to look at again — which is the strongest available argument for scanning the person whose calculator output looks reassuring, rather than the person already destined for a statin.
In the group this chapter is really written for — an LDL between 70 and 189 mg/dL, no diabetes, and a genuinely uncertain decision — any calcium at all was associated with roughly twice the event rate or higher, across every risk tier. Adding CAC to the risk categories produced a net reclassification improvement of 0.124, with the movement going in the right direction for both events and non-events.
Moderate — a large, well-conducted cohort producing real numbers, and four reasons not to treat it as more than it is.
- It is observational. It shows that CAC sorts people by risk, not that acting on CAC improves outcomes. No trial has yet randomised people to be scanned or not and measured what happened to them.
- The scores were not blinded. Clinicians and participants knew the results, so treatment decisions were influenced by the very thing being evaluated. That bias runs in the direction of making CAC look useful.
- The risk equation is American. These figures are built on PREVENT and the 2026 ACC/AHA thresholds. UK practice uses QRISK, and NICE does not recommend calcium scoring for routine risk assessment — so the reclassification figure does not transfer directly, and in the UK a CAC scan remains something you seek out rather than something you are offered. The direction of the finding travels; the numbers do not.
- MESA was a relatively healthy cohort, which the authors note will have pushed absolute event rates down. And a net reclassification improvement of 0.124 is a measurable gain rather than a dramatic one, on a metric that is itself contested.
What it changes in practice, if anything. Not the treatment of a high score, which was already clear. It sharpens who is worth scanning: the value is highest exactly where the decision is closest, and it is meaningful even in people a calculator has already dismissed. As the authors put it, calcium scoring is most useful for placing a person’s absolute risk in context and setting the intensity of prevention — not for making the decision on its own.
Not all coronary calcium is the same, and the scan cannot tell you which you have
A calcium score is a single number for something that comes in at least two forms. Imaging that sees plaque structure directly — optical coherence tomography, intravascular ultrasound — separates them:
| What it is | What it means | |
|---|---|---|
| Spotty micro-calcification |
Small scattered flecks within young, inflamed, lipid-rich plaque | The worrying pattern. Associated with more inflammation and a higher chance of rupture — which is the event that causes most heart attacks |
| Dense macro-calcification |
Larger, more solid deposits in older plaque | The reassuring one. Associated with less inflammation, and generally read as plaque that has stabilised rather than plaque that is dangerous |
A CT calcium score counts total calcium and cannot distinguish between them. So two people with identical scores can carry quite different risk, and the number alone does not say which. That is a real limit of the test, and it sits alongside the more familiar one — that the scan is blind to soft, non-calcified plaque, which is where a good deal of near-term risk lives.
This is also the most plausible explanation for something that confuses people, and it should be labelled as an explanation rather than a finding. A calcium score sometimes rises after someone improves their diet, activity and metabolic health, or after starting a statin. The proposed reason is that as inflammation settles, soft plaque calcifies and stabilises — so the number goes up while the plaque gets safer. That is an inference from the wider calcification literature, not something any single trial has demonstrated, and it should not be used to wave away every rising score. The practical position is narrower and holds either way: a rising score on its own is not proof of getting worse, and a falling score is not a target worth chasing. What the score is genuinely good at is the first one — establishing whether there is plaque at all, and roughly how much.
🔬 The CAC paradox — why the score isn't a treatment target
A common misconception is that the goal of treatment is to drive a high CAC score back down to zero. This is not how it works, and chasing this goal can be actively counterproductive.
Calcification is your body's defence mechanism, not the disease itself. Plaque begins life as soft, fatty, and unstable — and unstable plaque is what ruptures, throws clots, and causes heart attacks. As the body attempts to heal damaged vessel walls, it deposits calcium into the soft plaque to harden and stabilise it. Hardened, calcified plaque is far less dangerous than the soft plaque it replaced.
The clinical goal is therefore not to lower the CAC score, but to halt new plaque formation and stabilise what's already there. Ironically, when you successfully treat coronary disease with high-intensity statins, your CAC score may temporarily rise — because the soft plaque is being converted into safer, dense, calcified plaque. This "calcium paradox" is well-documented in the literature: statin-treated patients show greater Agatston score progression than untreated patients, yet have lower rates of heart attacks. Plaque volume goes down even as calcified density goes up.
Practical implication: if you're already on a statin and have a CAC scan, the score should not be interpreted the same way as a treatment-naïve score. Serial CAC scans cannot be used to measure statin efficacy — for that you'd need CT angiography measuring plaque volume directly. The 2020 National Lipid Association statement is clear on this.
If your CAC score is above zero — the action plan
The strategy is the same logical sequence regardless of the absolute score: reduce ongoing damage, stabilise what exists, and reduce the supply of new plaque-forming particles. The intensity scales with the score.
Eliminate the sandpaper — glycation & inflammation
To stop new plaque forming, you must stop damaging the arterial wall. This is the upstream lever — and it's the one most underused in conventional cardiology.
- Crush glycation — keep HbA1c low and avoid repeated sharp glucose spikes. AGEs are what physically damage the endothelium, creating the lesions that LDL particles then enter. Use the CGM section to learn your specific spike-triggering foods.
- Crush systemic inflammation — target hs-CRP < 1.0 mg/L. Eliminate ultra-processed foods, manage stress, prioritise sleep, address dental health and gut issues. Without inflammation, even high apoB struggles to drive rapid progression.
Manage the cargo — lipid optimisation
If vessel walls have historical damage (proven by CAC > 0), the cholesterol particles in circulation need to be as harmless as possible — fewer in number, and the larger, more buoyant variety.
- Optimise particle size — keep your TG/HDL ratio very low (target < 0.87 mmol/L). This ensures LDL particles are large, buoyant, and less likely to penetrate compromised vessel walls. See Section 3 — your Integrated Risk card already shows this ratio.
- Reduce particle count — at CAC 100+, pharmacological lipid lowering (statins, ezetimibe, bempedoic acid, PCSK9 inhibitors) is routinely recommended. The mechanism is simple: in a damaged vessel, lower the absolute number of apoB particles in circulation, and the statistical probability of any one particle getting trapped drops correspondingly.
- Low-dose aspirin — for CAC 100+ (and especially > 400), low-dose aspirin is often discussed with patients to reduce platelet aggregation around stable plaque. Bleeding risk needs to be weighed against benefit — this is a GP / cardiology conversation, not a self-prescription.
Targeted nutritional synergies
Two supplements have specific, mechanism-driven evidence in established coronary disease — beyond the general anti-inflammatory diet pattern.
Vitamin K2 (MK-7) — K2 activates Matrix Gla Protein, which inhibits calcification of soft tissue. That is a real role, distinct from K1's part in clotting. Whether it is aligned with what you want is exactly the question this page cannot yet answer — see the calcification-type section below, and note that the investigators themselves raise the possibility that slowing calcification could turn out to be unhelpful, since calcified plaque is the stabler kind. Hard endpoints are needed to settle it in either direction.
An earlier version of this page said there was no trial showing K2 slows calcification progression on repeat scanning. That is no longer true, and the correction matters more than the original claim did. VitaK-CAC — published in JAMA Cardiology in June 2026 — randomised 180 people (90 per group) with symptomatic coronary disease and a CAC score between 50 and 400 to 360 µg of MK-7 daily or placebo for two years. 150 completed, 75 per arm.
CAC, Agatston units — medians Placebo MK-7 Baseline 145 135 Two years 214 184 Difference in progression About 29% less by Agatston score, and about 42% less by calcium mass. p = 0.02, holding after adjustment for covariates What that does and does not establish
- Both groups still calcified. Nobody's score fell. MK-7 slowed the rise; it did not reverse anything.
- And a secondary finding that resolves the tension on this page, in its own terms. The rise in CAC score tracked the number of non-calcified plaques that became partially calcified during the trial (p = 0.04). So MK-7 appears to slow new calcification forming in developing plaque, rather than acting on calcium already laid down. That is a different thing from unpicking a stable deposit — which matters, because stable deposits are the ones you would not want to disturb. On stenosis, the comparison was null rather than the progression absent: stenosis severity progressed in roughly a third to two-fifths of both arms — about 33% on MK-7 and 41% on placebo — with no significant difference between the groups. Slowing the calcium score did not slow the narrowing.
- The endpoint is a surrogate, and the trial says so. It was not designed or powered for heart attacks, strokes or deaths, and whether slower calcification means fewer events is unanswered. On this site that distinction is the whole argument — see the ezetimibe chapter, where a drug was written off on a surrogate, and the inflammation agents, where one succeeded on a biomarker and failed on events.
- It is a small trial with a long, interrupted life. 180 randomised and 150 completing, with data collected from 2012 to 2022 with gaps, at two Dutch hospitals. The authors' own word for the effect size is “modest”.
- The population was specific: symptomatic coronary disease, 78% already on a statin, and no high-risk plaque at baseline. It does not transfer cleanly to a well person with an incidental score.
- And the commercial thread should be visible. The trial used a branded MK-7 preparation, and the group that developed and monitored it sits within the same university. That does not make the result wrong — it was randomised, blinded and placebo-controlled, which is what counts — but supplement trials funded or supplied by the manufacturer are the norm rather than the exception, and it is worth knowing which ones are.
Where that leaves the grade. Moderate for slowing calcification progression — one properly randomised trial, a real effect, an unrelated group would need to reproduce it. Ungraded, because untested, for anything that actually matters to a person: no trial has shown K2 prevents a heart attack. If you take it, take it on that footing.
The safety point that outranks all of the above. Vitamin K in any form opposes warfarin and the other vitamin K antagonists, and can reduce their effect — which for someone anticoagulated after a clot, a valve or in atrial fibrillation is a serious matter. Do not start, stop or change a K2 supplement while taking warfarin without speaking to whoever manages your INR. This does not apply to the DOACs — apixaban, rivaroxaban, edoxaban, dabigatran — which work by a different mechanism.On dose, since the gap is easy to miss: the trial used 360 µg/day, and most supplements on the shelf contain 100–200 µg. A bottle bought on the strength of this trial may well not contain the amount the trial tested.
- High-dose EPA omega-3 — REDUCE-IT showed prescription-grade purified EPA (icosapent ethyl, 4g/day) stabilised plaque and reduced cardiovascular events on top of statin therapy, though that result is contested because the placebo was mineral oil, which raised LDL-C, apoB and hs-CRP in the control arm; the parallel STRENGTH trial, using corn oil with EPA+DHA, was neutral. See the calculator page for the detail. The effect was independent of the modest TG-lowering, suggesting direct plaque-stabilising action via reduced vascular inflammation.
- Cut refined sugar and ultra-processed food — the dietary drivers of the oxidative and inflammatory load that oxidises LDL and turns it from passing-through to plaque-forming. This is the diet section's central message, and CAC above zero makes it non-negotiable. Moderate
On seed oils specifically, because this page previously said something the rest of the site does not. An earlier version of this list named “industrial seed oils” alongside refined sugar. That was wrong, and it contradicted the oils chapter, which sets out the evidence: replacing saturated fat with seed oils lowers cardiovascular risk in trials, and the case against them rests on mechanism and on omega-6 ratios rather than on outcomes. Extra-virgin olive oil remains the better default for other reasons. The claim has been removed rather than softened, because two pages of the same site should not disagree. Contested is the fair grade for the anti-seed-oil position, and it belongs to the chapter that examines it properly.
The bottom line on CAC
How long a zero lasts — the "warranty period". A score of zero is not permanent, and the useful question is how long it can be relied on before repeating. For most people at low to intermediate risk the interval commonly quoted is around five years, shortening in the presence of diabetes, a strong family history, smoking or a high Lp(a) — the same factors that make plaque more likely to appear in the interval. The practical consequence is that a zero buys a defined period of reassurance rather than an indefinite one, and that repeating it too early wastes a scan while repeating it too late defeats the point of having done it.
A CAC score is an early-warning system, not a death sentence. A score of zero in a borderline biomarker pattern is one of the most reassuring findings in cardiology — it earns you the "power of zero" and often legitimately lets you delay statin therapy while you keep working on the metabolic environment. A high score isn't a verdict either: by combining a low-glycation, low-inflammation metabolic profile with strategic lipid management, you can effectively "freeze" the plaque in place, harden it so it cannot rupture, and dramatically reduce your actual cardiovascular risk.
Most importantly: the CAC scan gives you data that no blood test can. If you've been agonising over whether to start a statin based on borderline numbers, a CAC scan often makes the decision for you — in either direction.
One important caveat for younger patients and women: CAC sees only calcified plaque. In patients under ~50, in women, and in anyone with elevated Lp(a) or strong family history of premature cardiovascular disease, soft (non-calcified) plaque may be present even with a CAC of zero. Lester et al. (2009) found 47% of younger adults with CAC=0 still had subclinical carotid atherosclerosis on ultrasound. A carotid ultrasound (measuring CIMT and identifying plaque) is the complementary test that catches this — see Section 15's Phase 2 imaging discussion for the full clinical rationale. In these populations, the two tests together provide much stronger reassurance than CAC alone.
UK access: Most private health screening services offer CAC at £200–400 (Bupa, Vitality, Nuffield, One Welbeck, the Princess Grace Hospital). NHS access is limited and clinical-indication only at present, though it is becoming more available through preventive cardiology clinics.