Time above 140 mg/dL on CGM tracked cardiovascular risk in people without diabetes

Design

Cross-sectional analysis of a community cohort with blinded CGM

Population

1,356 Framingham participants without diabetes or cardiovascular disease, mean age 59

Primary outcome

PREVENT 10-year cardiovascular risk estimate and individual risk factors

Effect

Per SD more time >140 mg/dL: 2% higher estimated risk; 21–26% higher odds of hypertension and dyslipidaemia

Framingham Heart Study participants without diabetes or cardiovascular disease, 1,356 in all with a mean age of 59, wore a blinded continuous glucose monitor for up to 10 days. The investigators related CGM measures and data-driven glycaemic profiles to 10-year cardiovascular risk estimated with the PREVENT equations and to individual risk factors.

Each standard deviation more time above 140 mg/dL (15.4% of the day) was associated with a 2% higher estimated 10-year risk and 21–26% higher odds of hypertension and dyslipidaemia, after adjusting for fasting glucose. Profiles with more CGM dysglycaemia did not line up neatly with prediabetes defined by fasting glucose and HbA1c, and were associated with 6–7% higher estimated risk.

This is a single cross-sectional snapshot in a largely white American cohort, and the outcome is a risk score rather than events. It adds to the case that HbA1c and fasting glucose miss some dysglycaemia, but it is not a reason to fit sensors to people without diabetes.

• Treat CGM in people without diabetes as a research measure, not a screening test
• Keep using HbA1c and fasting glucose to define prediabetes and diabetes
• Where a patient shows you consumer CGM traces, check blood pressure and lipids rather than interpreting the glucose alone
• Remember the cohort was 93% non-Hispanic white; the pattern may differ in South Asian patients

Why it matters

It suggests two people with the same HbA1c can carry different glucose exposure, which matters as consumer CGM spreads.

Don't overread it

This was cross-sectional and the outcome was an estimated risk score — it cannot show that CGM-detected dysglycaemia causes events.

The statistics, in plain English

A 2% higher estimated risk per standard deviation is a small shift in a modelled number, not a measured rise in heart attacks. Odds ratios of 1.21–1.26 for hypertension describe association at one time point; they say nothing about which came first.

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.