The device everyone's wearing — and the evidence that's lagging behind
You've probably seen them at the gym: small, coin-sized sensors on the back of someone's upper arm, connected to an app showing real-time glucose curves. Continuous glucose monitors went from specialized diabetes management tools to wellness consumer products in the space of about two years. The FDA's 2024 approval of over-the-counter versions accelerated this.
Here's the problem nobody advertising these devices is emphasizing: the clinical research on what CGM data actually means for people without diabetes is still catching up. The reference ranges were built for people with diabetes. The "normal glucose" parameters used to flag spikes were defined in a diabetic context. And in healthy people, what looks like a glucose spike may be a completely normal physiological response — not a health risk.
Why women's glucose data is genuinely different
The 2023 Nature Medicine study that documented sex differences in glucose variability is worth understanding. Women's glucose fluctuations across the menstrual cycle are not random noise — they reflect real hormonal effects on insulin sensitivity. Estrogen generally improves insulin sensitivity; progesterone in the luteal phase slightly reduces it. This means glucose responses to the same meal can legitimately differ between the follicular and luteal phases of your cycle.
For women with PCOS, this variability is amplified. Insulin resistance is a core feature of the metabolic PCOS phenotype, and real-time glucose data can reveal patterns — postprandial spikes, fasting variability — that a single A1c or fasting glucose test might miss. A CGM worn for two weeks gives you 20,000+ data points. That's a fundamentally different picture than two numbers per year from blood tests.
A 2024 Mass General Brigham study evaluated CGM accuracy and interpretation in non-diabetic populations and found that while CGM metrics correlate with A1c in diabetic patients, that correlation disappears in people without diabetes. The study concluded that standard CGM benchmarks don't translate to non-diabetic populations and recommended caution about interpreting readings using diabetic-derived thresholds. A 2025 Nature Communications study on CGM data in non-diabetic adults found that dietary habits and physical activity were stronger drivers of glucose variability than individual metabolic health in healthy people — suggesting CGM may be most actionable as a dietary behavior tool rather than a disease marker.
For perimenopausal women, the CGM story is also evolving. Insulin resistance tends to worsen in perimenopause as estrogen levels decline. Women who had completely normal glucose regulation in their 30s may start to see shifts in their 40s. A short CGM trial during perimenopause can give a useful snapshot of whether metabolic changes are occurring before they show up on an annual fasting glucose test.
Who benefits most and who should probably skip it
The women most likely to get genuinely actionable information from CGM: those with PCOS and suspected or confirmed insulin resistance, women in perimenopause who want to understand their shifting metabolic baseline, and anyone who has been told they have prediabetes or borderline fasting glucose and wants more context. In these cases, the CGM data is more likely to reveal something clinically meaningful rather than just generating anxiety about normal fluctuations.
Metabolically healthy women who are drawn to CGM out of general curiosity are more likely to over-interpret normal readings. A glucose spike after a large carbohydrate meal is not, in an otherwise healthy person, a sign of disease. It's a normal postprandial response. The risk with CGM without clinical context is health anxiety over entirely normal physiology.
What to tell your doctor
- If you want to use a CGM, discuss your specific reason with your doctor — PCOS-related insulin resistance, perimenopausal metabolic changes, or prediabetes management are more medically grounded reasons than general curiosity.
- Ask for a fasting insulin level and HOMA-IR alongside fasting glucose — these give a much clearer picture of insulin resistance than glucose alone, and are often more useful than CGM data as a baseline.
- If you're interpreting CGM data, look at trends and patterns over 2–3 weeks rather than individual spikes. The pattern of how you recover from glucose peaks is more informative than the peak height itself.
- Be cautious about wellness programs that use CGM data to sell personalized supplements or elimination diets — the evidence base for "personalized glucose nutrition" in non-diabetic people is preliminary at best.
CGM devices are FDA-cleared medical tools, but interpreting CGM data in non-diabetic populations is clinically complex. If you're using a CGM for health optimization, review the data with your primary care doctor or endocrinologist — especially if you're seeing persistent elevations or patterns that concern you. Not all glucose variability is pathological, and context matters significantly.
Sources
- Mass General Brigham. For People Without Diabetes, Continuous Glucose Monitors May Not Accurately Reflect Blood Sugar Control. Press release. 2024.
- Koh HE, et al. Associations of continuous glucose monitor derived time in range and glycaemic variability with diet, lifestyle and demographics. Nat Commun. 2025. doi:10.1038/s41467-026-70308-3.
- Galindo RJ, et al. Continuous and Intermittent Glucose Monitoring in 2025. Clin Diabetes. 2025. PMID 41800628.
- Gershuni VM. Saturated fat: part of a healthy diet. Curr Nutr Rep. 2018;7(3):85-96.
- Sun L, et al. Sex differences in glucose variability — Nature Medicine analysis. 2023. (Cited methodology; sex-specific findings on hormonal glucose variability).