Science1 distinct publisher3 min readPublished
Two August analyses of the same cohort found calcium scores only change predictions at borderline and intermediate PREVENT risk. CTs ordered for other reasons did the job as well.
The Scientist · Science desk
Compiled by The ScientistSomething wrong?How this is made
Put numbers on the window and it is small. PREVENT sorts ten-year risk into low below 3%, borderline 3% to 5%, intermediate 5% to 10%, and high at 10% or above [8]. If calcium only moves the decision in the middle two bands, the whole zone of usefulness runs from 3% to just under 10%, and the borderline slice where one author calls the scan a tiebreaker is two percentage points wide [15]. Outside that zone, the scan is a charge without a consequence: running both tools in people referred for cardiovascular screening improves prediction on the order of 1% or 2% [12], at $100 to $400 the patient generally pays alone [7].
Michael Shapiro of Wake Forest University School of Medicine, whose paper appeared in the Journal of the American College of Cardiology on Aug. 18, gave the ceiling plainly: "Adding CAC to PREVENT for everybody only modestly improves statistical prediction" [3]. His next sentence is the one that undercuts the scan's second life as a personalized wellness measure [19]: a score of zero "should not be used as a veto against statin therapy" [4]. Scores run from 0 past 1,000 [9], statins are cheap and long established [14], and neither tool supplies a numeric trigger in either direction, only a discussion the guidelines now say should begin as early as age 30 [13].
That leaves the July Circulation paper doing the practical work. If CTs ordered for unrelated reasons predict coronary heart disease, cardiovascular disease and death over a decade about as well as the purpose-built scan [6], then for any patient with a prior chest or abdominal CT the calcium information has already been bought, and the marginal imaging cost of using it is zero [16]. The guidelines have already opened that door, extending incidental calcium findings to borderline risk for men 40 and up and women 45 and up [11].
The caution sits in the sample. Both August papers drew on the same source: just over 6,000 people aged 45 to 79 in the Multi-Ethnic Study of Atherosclerosis, followed for ten years [1]. Two papers agreeing on one cohort is consistency rather than independent replication [17], and calling them trials, as STAT does, is generous for a reanalysis of observational follow-up [20]. In the JAMA paper the headline result was null and the reclassification surfaced only when the borderline and intermediate groups were looked at separately [5], which is the weakest place for a finding to live and the place a coverage policy would have to rest on.
A payer reading both papers has a defensible narrow position available: fund the dedicated scan inside the 3% to 10% band, and pay for calcium to be read and reported on the scans that were taken anyway.
Ranked by verification strength, evidence, and original report placement.
The two studies published in August both examined data following just over 6,000 people aged 45 to 79 for 10 years in the Multi-Ethnic Study of Atherosclerosis, analyzing whether CAC scores improved prediction of who would later have a heart attack or stroke.
Both August studies found that CAC scores did make a difference in predicting later heart attack or stroke, but only when PREVENT risk was classified as borderline or intermediate.
"Adding CAC to PREVENT for everybody only modestly improves statistical prediction," said Michael Shapiro, co-author of the study published Aug. 18 in the Journal of the American College of Cardiology and a preventive cardiologist at Wake Forest University School of Medicine.
Shapiro: "The main message is that not everyone needs a CAC scan and CAC zero should not be used as a veto against statin therapy... CAC was most helpful as a tiebreaker when a patient was at borderline risk and the statin decision was unclear."
The JAMA study published Wednesday found that overall, CAC did not change the predictions made by PREVENT; in a secondary analysis of people at intermediate or borderline risk, substantial reclassification emerged, according to co-author Nilay Shah, a preventive cardiologist at Northwestern University Feinberg School of Medicine.
CAC scans are not typically covered by insurance and can cost the patient $100 to $400, while PREVENT risk estimates rely on standard information gathered in a doctor's visit.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Peer-reviewed findings, single-outlet secondhand reporting
Three named peer-reviewed publications (JACC, JAMA, Circulation), a specified cohort with sample size, age range and follow-up period, and on-record quotes from two study co-authors plus one uninvolved cardiologist. Weakening it: only one publisher reports the cluster, no effect statistics are given, the '1% or 2%' improvement figure is unattributed to a specific study or metric, and the article calls observational cohort analyses 'trials.'
Guideline-level adoption, no utilization data
There is concrete institutional adoption to point at: PREVENT underpins current ACC-AHA guidelines and those guidelines endorse acting on incidental calcium findings for a broader population than in 2018. What is missing is any measured uptake — no scan volumes, no payer coverage counts, no clinic implementation data — and the article's claim of 'growing popularity' for these tests carries no supporting figures.
Substance hedged, study strength slightly oversold
The substantive framing is deflationary rather than promotional — the headline job for CAC is narrowed to a two-to-seven-point risk window and the article says most people do not need the tests. The overstatement is in evidentiary presentation: two analyses of one observational cohort are called 'trials' and their concordance is presented as convergence, and an unattributed '1% or 2%' figure stands in for effect size. Net mild positive.
Author, guideline, and imaging-revenue interests visible but undisclosed
Two of the three quoted cardiologists are co-authors of the studies they are interpreting, so they have a direct stake in how the findings are characterized; the guidelines that supply the framing come from professional societies whose members read and act on them; and the underlying test is a largely self-pay $100-$400 imaging service whose demand contracts if the narrow-use conclusion holds. The article supplies no funding or conflict-of-interest disclosures, which is why this scores moderate rather than low.
Named expert attribution, no cross-outlet corroboration
Findings are traceable to identified journals, dated publications, and named clinicians including one independent commentator, which supports moderate confidence. But the cluster has a single publisher, no primary-document access is evidenced, several stated numbers lack attribution, and the design of the underlying work is described inconsistently, so the assessment cannot rise above mid-range.
science
US life expectancy hit 79 in 2024, and the cushion over 2019 is about 73 days1 distinct publisher
science
One donor cornea, 400 implants: the graft shortage starts looking like a factory problem1 distinct publisher
science
A four-year visa clock for postdocs now rests on one Massachusetts docket1 distinct publisher
science
SNAP soda bans cut purchases 12%. The exclusion list decides what that is worth1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 26, 2026