The Vitamin 8 in 10 Americans Take—But Shouldn’t

CONNOR'S HEALTH NOTES

Calcium in an Artery Is a Marker Before It Is a Cause

Connor Hill · InsightfulWord · September 2, 2026

Nothing here is a reason to start or stop any supplement or prescribed medication. That decision belongs with the clinician who knows the case. What follows is about how arterial calcium is measured and what the number means, which is a separate question and a well-documented one.

Calcification in a coronary artery is one of the most studied findings in cardiovascular medicine, and its interpretation contains a genuine paradox that is worth understanding before any claim about it can be assessed.

The paradox is this. The presence of calcium indicates that atherosclerosis is present, and more of it indicates more disease and higher risk. But calcification is also the process by which an unstable plaque becomes a stable one, and a treatment known to reduce heart attacks measurably increases the calcium score.

Both of those statements are supported. They are not in conflict once the measurement is understood, because the score that is used to summarize calcium is constructed in a way that weights the density of the deposit, and density is precisely what increases when a plaque stabilizes.

That construction was reasonable when it was devised in 1990 and remains clinically useful, but it means a rising score has two entirely different possible interpretations, and distinguishing them requires knowing what else changed.

The broader point extends past cardiology. A biological finding that appears where disease is present is not automatically the mechanism of the disease, and it may be the body's response to it. Scar tissue appears where injury occurred. Fever appears with infection. Neither is the pathology.

Confusing the marker with the cause supports a familiar and rhetorically powerful move: identify something present in diseased tissue, identify a dietary source of it, and conclude that consuming the source produced the disease. The chain has a missing link at every joint.

What follows is how the score is computed, why density cuts both ways, what the score predicts and what it does not, the comparison that breaks under treatment, and when the scan actually changes a decision.

How the Score Is Actually Computed

The standard summary is the Agatston score, and its arithmetic explains most of the interpretive difficulty.

A computed tomography scan of the heart is acquired without contrast. Within each image slice, regions of calcification are identified as areas exceeding a fixed radiodensity threshold, measured in Hounsfield units — the scale on which water is zero and dense bone is well above a thousand.

Each identified area is then multiplied by a weighting factor determined by the maximum density within it. The factor increases in steps as the peak density rises. The weighted areas are summed across all slices and all coronary vessels, producing a single number.

The consequence of that design is that a large area of low-density calcification and a small area of very dense calcification can produce similar scores, and that increasing the density of an existing deposit raises the score without any new territory being involved.

The categories in common use are conventional rather than biological: zero, one to ninety-nine, one hundred to three hundred and ninety-nine, and four hundred or above, with the boundaries chosen for practical stratification.

Alternative measures exist and are computed from the same scan. Calcium volume ignores density entirely. Calcium mass approximates the actual quantity of mineral. Density can be reported separately as a mean. Each answers a different question, and the fact that three summaries can be derived from one scan is itself an indication that no single one captures everything.

Why Density Cuts Both Ways

The behavior of the score under treatment is where the interpretive problem becomes concrete and clinically consequential.

Lipid-lowering therapy reduces cardiovascular events. That is established across a very large randomized evidence base and is not in dispute.

The same therapy is associated with an increase in measured coronary calcium. Plaques treated this way tend to develop denser, more consolidated calcification, and the density weighting in the score converts that stabilization into a higher number.

So a patient on treatment can return a rising score while their actual risk has fallen. The score went up because the plaque became the kind of plaque that is less likely to rupture.

This has a direct consequence for research. A population of people with high scores contains both those with extensive unstable disease and those with stabilized, densely calcified disease at considerably lower risk. Mixing them weakens the observed association between score and events, which is why the score performs differently in treated and untreated populations.

The general lesson is that a measurement designed as a marker of disease burden can be moved by a process that reduces harm, and that when this happens the direction of the number stops being interpretable on its own.

What the Score Predicts and What It Does Not

The evidence on prediction is extensive, and the boundaries of the claim are as important as the claim.

The score adds predictive information beyond conventional risk factors in asymptomatic adults. That has been shown in large prospective cohorts followed for many years, and it is the basis for the scan's place in prevention guidelines.

Its strongest use is at the low end. A score of zero in an asymptomatic person is associated with a low rate of events over the following several years, and it is that finding, more than any high-score finding, that changes management — by supporting a decision not to begin treatment in someone whose calculated risk was borderline.

What the score does not do is diagnose. It does not identify blockages, does not measure how much a vessel is narrowed, and does not assess whether blood flow is limited. Those require different tests entirely.

It also does not detect the disease that has not calcified yet. Non-calcified plaque exists, is not visible on this scan, and matters — which is why a zero score in a person with symptoms is not reassuring in the way a zero score in an asymptomatic person is.

And it is a risk estimate for a population, not a prediction for an individual. A low-risk group still contains people who have events; the score changes the probability, not the outcome.

The Comparison That Breaks Under Treatment

The reason all of this bears on claims about supplements is that the same structural problem defeats most observational studies in this area.

A study comparing people who take something to people who do not is comparing groups that differ in far more than the thing. People take supplements for reasons — a doctor suggested it, a test came back low, a diagnosis was made, a family history exists. Those reasons are also risk factors.

The technical name is confounding by indication, and it produces effects in either direction depending on why people take the substance. Where it is taken by people already unwell, it looks harmful. Where it is taken by the health-conscious, it looks protective.

Statistical adjustment helps only for factors that were measured, which is why observational studies of the same supplement routinely disagree, and why randomized trials in this area have so often failed to confirm the observational finding.

The same applies to the calcium finding specifically. People with higher measured calcium differ systematically from people with less, and any substance more common among them will appear associated with it.

What would settle a causal claim is a randomized trial with the calcium measurement or clinical events as a pre-specified outcome. Those exist for some supplements and not for others, and where they exist their results are the answer regardless of how many observational studies point elsewhere.

There is a second requirement that is easy to overlook. A trial has to be long enough for the outcome to occur. Arterial calcification develops over decades, and a trial running two or three years with a calcium endpoint is measuring a change in a slow process against the precision of the scanner. Where such trials report no difference, the finding is frequently uninformative rather than reassuring, and the distinction belongs in any summary of them.

When the Scan Changes a Decision

The practical question for anyone encountering this test is narrow, and the guidelines are explicit about it.

The scan is used to refine risk estimation in asymptomatic adults whose calculated risk sits in an intermediate band — where the decision about preventive treatment is genuinely uncertain and additional information could move it either way.

Below that band, the estimated benefit of treatment is small enough that a scan is unlikely to change anything. Above it, treatment is indicated on the calculated risk alone and a scan adds cost without altering the decision.

For someone already on preventive treatment, a repeat scan is generally not informative for the reason described above: the score can rise while the treatment is working, and no threshold exists for interpreting that rise.

There are costs on the other side. The scan involves radiation, though a modest dose by cardiac imaging standards. It produces incidental findings in the lungs and elsewhere that generate follow-up, some of which is unnecessary. And it can produce false reassurance in the presence of non-calcified disease.

There is also a sequencing point. The decision that the scan informs is whether to begin preventive treatment, which means the scan belongs before that decision rather than after it. Ordering it in someone already treated, or in someone whose risk is high enough that treatment is indicated regardless, produces a number that cannot change the plan.

The composite point is that arterial calcium is a well-measured marker whose summary score can move for opposite reasons, that its most valuable result is a zero in an asymptomatic person near a decision threshold, and that a finding of calcium in an artery is a description of what is there rather than an account of how it arrived.