Where a calcium score changes the PREVENT decision

Design

Prospective observational cohort (MESA), 10-year follow-up

Population

6,098 US adults aged 45–79 without cardiovascular disease

Primary outcome

Fatal and non-fatal atherosclerotic events (366 events)

Effect

ΔC statistic 0.02 (0.01–0.03); NRI 0.095; borderline-risk event rate 1.9% (CAC 0) to 14.3% (CAC ≥300)

A JAMA analysis (6 October) of 6,098 adults aged 45 to 79 from the Multi-Ethnic Study of Atherosclerosis (MESA) asked what a coronary artery calcium (CAC) score adds to the PREVENT equations, which the 2026 US dyslipidaemia guideline uses to decide who should be scanned.

Across everyone, adding CAC improved discrimination only slightly: the C statistic rose by 0.02 from 0.73. But inside the borderline band (3% to under 5% ten-year risk), ten-year event rates ranged from 1.9% with a score of zero to 3.9% for 1–99, 7.4% for 100–299 and 14.3% for 300 or more.

That is where a scan earns its cost: it splits one risk category into groups whose outcomes differ several-fold, and turns an uncertain statin conversation into a clear one. The study is observational and from the US; MESA did not enrol South Asians as a separate group, and PREVENT has not been validated in India.

• Use CAC selectively, in patients whose PREVENT ten-year risk is 3% to under 10%
• A score of 300 or more at borderline risk carried a 14% ten-year event rate in this cohort
• A zero score at borderline risk carried about 2% — a reasonable basis for discussion about deferring
• Do not scan the low- or high-risk groups to decide treatment; it adds little there
• Recognise that PREVENT has not been calibrated for Indian patients

Why it matters

It gives the new guideline's selective-scanning advice a concrete evidence base.

Don't overread it

Observational data from a US cohort — it predicts events but does not show that scan-guided treatment improves outcomes.

The statistics, in plain English

A C statistic gain of 0.02 is small: the model gets slightly better at ranking who will have an event. The net reclassification improvement of 0.095 means about 10% more people moved to a more accurate risk category than to a less accurate one.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.