The Glucose Gap: What CGM Data Can and Can't Tell You

Overview

The claim that around 88% of American adults are metabolically unhealthy is real, widely cited, and almost always attributed to the wrong thing.

It comes from a study published in 2018 analysing National Health and Nutrition Examination Survey data from 2009 to 2016, covering 8,721 adults. It found that only 12.2% met optimal targets across five cardiovascular risk factors without medication.

It is not a finding from continuous glucose monitors, and it does not describe glucose variability. It measures blood glucose, triglycerides, HDL cholesterol, blood pressure and waist circumference — a broad cardiometabolic assessment using standard clinical measures.

The evidence that wearing a glucose monitor helps people without diabetes is considerably weaker than the statistic implies. Randomised trials show small effects, reviewers have flagged a lack of evidence for benefit in non-diabetic users, and much of the enthusiasm originates from companies selling the devices.

Where the 88% Actually Comes From

The source is a paper titled 'Prevalence of Optimal Metabolic Health in American Adults', published in Metabolic Syndrome and Related Disorders in November 2018, with Joana Araujo as first author.

It analysed data from 8,721 Americans surveyed between 2009 and 2016 through NHANES, one of the most rigorous population health datasets available. The finding was that 12.2% of American adults were metabolically healthy — equivalent to roughly 27.3 million people.

That is where the 88% comes from: the inverse of 12.2%. It is a robust, well-conducted study, and it describes a population surveyed up to a decade before the current interest in glucose monitoring. Citing it as evidence produced by CGMs misattributes both the method and the era.

The Five Criteria

Understanding what was measured matters, because 'metabolically healthy' sounds vague and the definition used here is precise.

Five indicators were assessed: blood glucose, triglycerides, high-density lipoprotein cholesterol, blood pressure and waist circumference. To count as metabolically healthy, an adult had to be at optimal levels on all five — and to achieve that without medication.

That last condition is doing significant work. Someone whose blood pressure is well controlled on medication is not counted as metabolically healthy under this definition, despite being clinically well managed. The study measures how many people are at optimal levels unaided, which is a stricter and different question from how many are at health risk.

Why 12.2% Was Almost 19.9%

The most instructive detail in the study is rarely quoted alongside the headline, and it should be.

Using the earlier ATP III guideline thresholds rather than the updated cut-points, the same dataset produced a figure of 19.9% metabolically healthy. Switching definitions moved the result from roughly one in five to roughly one in eight — without a single new measurement.

That sensitivity means the widely repeated 88% could equally have been 80% depending on which clinical thresholds an author selected. Both numbers describe the same people. Any statistic this responsive to definitional choices should be quoted with the definition attached, and it essentially never is.

The Finding That Deserves More Attention

One result from the study is genuinely striking and rarely makes it into popular coverage: fewer than 1% of adults with obesity met all five optimal criteria.

The researchers also noted that metabolic health prevalence was alarmingly low even among people at normal weight — which is the more useful public health message, because it undercuts the assumption that body weight is a sufficient proxy for metabolic status.

Those two findings together are the substantive content of this research. A person of normal weight cannot assume they are metabolically healthy, and a person with obesity is very unlikely to be. Neither conclusion requires wearing a sensor to act on.

What CGM Trials Actually Found

Continuous glucose monitors are genuinely transformative in diabetes care. The question here is narrower: what happens when people without diabetes wear one?

A 2024 review of randomised trials found that CGM-based feedback reduced haemoglobin A1C by 0.3% and improved time in range by 7%. The reviewers characterised these as small, favourable effects and concluded that more research is needed to demonstrate benefit.

A 0.3 percentage point change in HbA1c is a real effect and a modest one. It is considerably smaller than what routine interventions like sustained dietary change or increased physical activity produce, and it comes at the cost of continuous sensor use.

The Evidence Gap for Healthy Users

Systematic reviews examining CGM use in non-diabetic individuals have identified real potential alongside significant limitations, and the limitations are the part usually omitted.

The promise is in personalisation — using real-time feedback to guide lifestyle changes, optimise activity timing, improve engagement with dietary modification, and identify at-risk metabolic phenotypes. Those are plausible mechanisms and reviewers acknowledge them.

What is missing is outcome evidence. Direct evidence on changes in traditional cardiovascular risk factors was limited, and evidence of impact on hard cardiovascular endpoints remains limited. A narrative review by researchers at UCL and Birmingham Children's Hospital went further, finding a lack of evidence for effective use in people not living with diabetes — including little published evidence on how accurate these devices even are in that population.

The Anxiety Problem

A specific concern raised in the literature is that continuous data may be actively unhelpful for some healthy users.

CGM output can be difficult to interpret without clinical context and may lead to unnecessary anxiety or confusion in otherwise healthy people. Glucose fluctuates normally throughout the day in people without diabetes — rising after meals, varying with exercise, sleep and stress. Much of what a monitor displays as a dramatic curve is ordinary physiology.

Without a reference for what normal variation looks like, a user can reasonably interpret a routine post-meal rise as evidence of harm and begin restricting foods on that basis. The risk is not that the sensor is wrong; it is that a healthy person is being handed an unfamiliar data stream with no framework for reading it.

Who Is Selling the Story

The commercial context is worth stating plainly, because it shapes which findings circulate.

Much of the popular material on glucose variability in healthy people originates from companies whose business is selling glucose monitoring subscriptions or personalised nutrition programmes built on them. That does not make their claims false, and it does mean the enthusiasm reaching consumers is coming substantially from parties with a direct financial interest.

The regulatory framing reinforces the point. Dexcom's Stelo was the first CGM from a major medical device manufacturer cleared for over-the-counter use, and Abbott's Lingo received FDA approval in June 2024 explicitly targeted at wellness users rather than people managing diabetes. Stelo lacks high and low glucose alerts and is unsuitable for insulin users. These are positioned as general wellness products, paid for out of pocket — a category with a lower evidentiary bar than medical devices.

Where CGMs Genuinely Help

None of this argues against continuous glucose monitoring as a technology. It is one of the more important advances in chronic disease management of recent decades.

For people with type 1 diabetes it is close to essential, replacing intermittent finger-prick testing with continuous data and alerts that prevent dangerous hypoglycaemia. For many people with type 2 diabetes it meaningfully improves control. In those populations the evidence is strong and the benefit is clear.

For someone without diabetes, the honest summary is that a CGM will show you real data about your own physiology, that the trials show small effects on glucose measures, that outcome evidence is largely absent, and that interpretation is genuinely difficult. It may be interesting and worth the money to a curious person. It is not established as a health intervention, and the 12.2% figure that usually introduces it was measured by other means entirely.

Conclusion

The statistic is real and misattributed. Only 12.2% of American adults met optimal metabolic health criteria in a 2018 study of NHANES data covering 8,721 people — a finding based on blood glucose, triglycerides, HDL cholesterol, blood pressure and waist circumference, not on glucose monitors.

It is also more fragile than it appears. Applying earlier clinical thresholds to the identical dataset produced 19.9% rather than 12.2%, meaning the familiar 88% could as easily have been 80%. The findings that survive definitional choice are the important ones: fewer than 1% of adults with obesity met the criteria, and metabolic health was low even among people at normal weight.

The case for wearing a monitor without diabetes is much thinner than the case that metabolic health is a widespread problem. Randomised trials show a 0.3% HbA1c improvement and 7% better time in range, reviewers report limited evidence of benefit and unclear accuracy in this group, and the data can produce anxiety in people with no framework for reading it.

This article summarises published research for general information and is not medical advice. Anyone concerned about their metabolic health should seek assessment from a qualified clinician rather than relying on a consumer device.

Data Source and Attribution

NHANES study (PubMed)CGM systematic reviewThe Conversation (CGM evidence)

Metabolic health prevalence figures, criteria, sample size and the ATP III comparison come from 'Prevalence of Optimal Metabolic Health in American Adults: National Health and Nutrition Examination Survey 2009-2016', published in Metabolic Syndrome and Related Disorders in 2018. Continuous glucose monitoring effect sizes come from a 2024 review of randomised controlled trials, and evidence limitations in non-diabetic populations come from published systematic and narrative reviews. Regulatory status of over-the-counter devices reflects FDA clearances announced in 2024.

FactsFigs reviews, cleans, and cross-checks every source dataset before shaping it into a data story. Each visualization is created and designed in FactsFigs Design Studio — an internal tool developed and owned by FactsFigs — and is the original work of a FactsFigs author, not an AI-generated copy of any existing graphic. Individual assets within a visual may or may not be produced with AI tools, but the design of the visual itself is solely FactsFigs' own.

This content is for information only and is not medical advice, diagnosis or treatment guidance. Figures reflect published research at the time of writing.

2026-07-20