Do continuous glucose monitors make healthy people healthier and longer-lived?
Claim attributed to Consumer metabolic-health companies (Levels, ZOE, Nutrisense, January AI, Signos) and influencers such as Jessie Inchauspe ("Glucose Goddess"), via a sensor-plus-subscription model. , Sold as a sensor plus recurring app subscription. Marketing reframes normal post-meal glucose rises as harmful "spikes" to be minimised. Several scientific advisors hold direct commercial stakes (Michael Snyder co-founded January AI; ZOE's PREDICT investigators are consultants to or employees of ZOE).
CGMs measure glucose accurately and reveal that responses to food differ between people. That a healthy person can act on those numbers to flatten "spikes" and live a longer, metabolically healthier life has never been tested for a single hard outcome, and the most rigorous review finds no glycemic benefit in normoglycemic users.
A CGM tells a healthy person, accurately, that their breakfast moved their glucose; it does not tell them that smoothing that curve buys them a single extra day of life.
What it’s supposed to target
- Continuous glucose sensing
- Post-meal glucose spikes
- Glycemic variability
- Behavioral feedback
A continuous glucose monitor is a measurement device, not a treatment, so its “mechanism” is really a behavioral theory: by showing your post-meal glucose spikes in real time, it should let you adjust food and habits to flatten glycemic variability, on the assumption that smaller spikes mean better metabolic health and slower aging in healthy people.
Each step is plausible but unproven outside diabetes. Big glucose swings clearly matter in diabetes, but in metabolically healthy people spikes are a normal response, the “optimal” glucose curve is undefined, and there is little evidence that chasing flatter lines with a CGM improves any hard outcome. Real data, with the leap to “this makes healthy people healthier” resting on assumption, not trials.
Mechanism is theory, not proof. A plausible pathway explains why something might work, not whether it does. The verdict rests on the evidence below, not the elegance of the theory.
What would have to be true
Post-meal glucose excursions within the normal range must be harmful to otherwise healthy people. UNSUPPORTED: no validated 'harmful spike' threshold exists, and normative data treat such rises as ordinary physiology.
Seeing CGM readings must reliably change a healthy person's eating in a beneficial direction. WEAK: behaviour change appears only inside structured programmes, not from the data alone.
Flatter glucose curves must translate into better metabolic markers in non-diabetics. NOT SHOWN: the 2026 meta-analysis finds no appreciable glycemic benefit in normoglycemic users.
Those marker changes must then produce fewer hard events and longer life. UNTESTED: no RCT has measured diabetes incidence, cardiovascular events, or longevity in healthy CGM users.
What the evidence actually shows
The device is real; the benefit in healthy people is not measured
The most direct evidence is a 2026 systematic review and meta-analysis (23 studies, 1,074 non-diabetic participants across 11 countries, including 7 RCTs). It states plainly that CGM improved glycemic control in prediabetes, whereas 'no appreciable glycemic benefit was observed in healthy normoglycemic populations.' It also found no significant body-mass effect (SMD -0.25, P=0.19) and concluded CGM is 'unlikely to achieve weight management goals independently without accompanying behavioral interventions.' The favourable cohort work cited in marketing, ZOE's PREDICT study (n=1,002), demonstrates only that glucose responses to identical meals vary between people (population CV ~68%). Variability is not benefit: showing that curves differ is not evidence that flattening yours improves any outcome.
A 'spike' within range is ordinary physiology
A multicenter normative study of 153 healthy non-diabetic people (ages 7 to 80) found they spent a median ~96% of the day between 70 and 140 mg/dL, with time above 140 averaging only ~2.1% (roughly 30 minutes a day) and readings above 180 mg/dL exceptionally rare. An independent 2022 mini-review in Sensors adds that the 'effects of diet, exercise, and stress on glucose regulation in metabolically healthy individuals are still largely unknown,' and flags an 'inevitable time delay' and accuracy limits that make readings hard to interpret, notably during exercise. In a healthy body, a post-meal rise is the system working, not a wound to be cauterised; there is no validated definition of a 'harmful spike' within the normal range.
Studies, graded, and who paid
PREDICT (n=1,002) confirms responses to identical meals vary widely; the device works.
No validated threshold exists; normative data show healthy adults sit in range ~96% of the day.
2026 meta-analysis: no appreciable glycemic benefit in normoglycemic people; weight loss only with a behaviour programme.
Zero human outcome trials. Untested, not disproven.
| # | Study | Type | Size | Funding / COI | Key limitations |
|---|---|---|---|---|---|
| 1 | Liao 2026, systematic review + meta-analysis (Eur J Med Res) | Systematic review and meta-analysis | 23 studies; 1,074 non-diabetic participants; 11 countries (7 RCTs) | Independent Chinese government research programmes (National Key R&D; Guangdong Provincial S&T; TZI-ZJU). Authors declared no competing interests. | Very recent; full risk-of-bias and heterogeneity appraisal not independently re-run. Pooled mean-glucose improvement is driven by prediabetic subgroups, not healthy users. |
| 1002 | Berry 2020, PREDICT 1 (Nat Med) | Large observational metabolic-phenotyping cohort | ~1,002 UK twins and unrelated adults; 100 US validation | Industry-funded Supported by ZOE (Zoe Global Ltd) plus Wellcome/MRC/BHF; multiple authors are consultants to or employees of ZOE, which sells products built on PREDICT. | Establishes between-person variability only; does not test whether flattening curves improves any health outcome. |
| 57 | Hall 2018, 'glucotypes' (PLoS Biol) | Small exploratory observational cohort | 57 participants (38 normoglycemic, 14 prediabetic, 5 diabetic); 30 in standardized-meal subset | Independent Declared 'no competing interests'; funded by NIH/NSF/Stanford. But senior author Michael Snyder co-founded January AI, a commercial CGM company, an undisclosed conflict (per source 6). | Small, hypothesis-generating; authors call for long-term outcome studies. Documents patterns, not benefits of acting on them. |
| 153 | Shah 2019, normative CGM profiles (J Clin Endocrinol Metab) | Multicenter prospective observational (reference ranges) | 153 healthy non-diabetic participants, ages 7 to 80 | Independent Leona M. and Harry B. Helmsley Charitable Trust (non-commercial). Lead author discloses Dexcom/Sanofi advisory and consultancy ties; funding source itself is independent. | Descriptive normative data, not an intervention; establishes what 'normal' looks like, not whether changing it helps. |
| 8915088 | Holzer 2022, CGM in healthy adults (Sensors) | Narrative mini-review / perspective | No primary sample | Mixed 'No external funding.' One co-author sits on the Abbott Advisory Board; others declare none. | Narrative review, not systematic; corroborates that outcome evidence in healthy people is limited and CGM has lag/accuracy constraints. |
Independent, non-commercial evidence (the government-funded meta-analysis and charity-funded normative data) is null or cautionary for healthy people, while the upbeat narrative leans on cohort studies tied to sellers.
Across sources the same gap recurs: plenty of mechanism-and-variability data, no hard-outcome trial in healthy non-diabetics.
Unproven ≠ disproven
This is unproven, not disproven: no trial has measured diabetes incidence, cardiovascular events, or lifespan in healthy CGM users, so the longevity claim rests on absence of data plus normal-physiology reasoning.
A genuine signal could still emerge for specific subgroups (e.g. prediabetes), where the meta-analysis already shows glycemic benefit.
Where claim and evidence diverge
The marketed chain runs 'CGM shows data, you flatten spikes, you get healthier and live longer.' Only the first link is established. No study connects acting on the data to any hard endpoint in healthy people.
The money trail
The product is a sensor plus recurring subscription sold direct-to-consumer as 'wellness', which sidesteps the evidentiary bar for medical claims and rewards broad use over evidence-gated use.
The most favourable foundational work is tied to sellers: ZOE consultants/employees authored PREDICT, and the glucotypes senior author co-founded January AI, a stake the 2018 paper did not disclose.
The honest read
A CGM will give a healthy person accurate, genuinely interesting data and may nudge short-term habits inside a structured programme. It has not been shown to improve metabolic health or extend life by 'optimising' an already-normal glucose curve, and the strongest pro-CGM evidence comes from the companies selling it.
What would change this verdict
A randomised controlled trial in healthy non-diabetics showing that CGM-guided eating cuts hard endpoints (diabetes incidence or cardiovascular events) versus standard advice.
A validated, prospectively tested threshold defining a 'harmful' post-meal glucose excursion within the normal range, linked to worse outcomes in healthy people.
Sources
- Liao X, Li Y, Tang S, et al. Continuous glucose monitoring in non-diabetic populations: a systematic review of observational and interventional studies with meta-analysis. Eur J Med Res. 2026.
- Berry SE, Valdes AM, Drew DA, et al. Human postprandial responses to food and potential for precision nutrition (PREDICT 1). Nat Med. 2020;26:964-973.
- Hall H, Perelman D, Breschi A, et al. Glucotypes reveal new patterns of glucose dysregulation. PLoS Biol. 2018;16(7):e2005143.
- Shah VN, DuBose SN, Li Z, et al. Continuous Glucose Monitoring Profiles in Healthy Nondiabetic Participants: A Multicenter Prospective Study. J Clin Endocrinol Metab. 2019;104(10):4356-4364.
- Holzer R, Bloch W, Brinkmann C. Continuous Glucose Monitoring in Healthy Adults: Possible Applications in Health Care, Wellness, and Sports. Sensors (Basel). 2022;22(5):2030.
- Michael P. Snyder (biography), co-founder of biotechnology companies including January AI. Wikipedia, accessed 2026.
People also ask
- Do continuous glucose monitors help healthy people without diabetes?
- A 2026 meta-analysis found no appreciable glycemic benefit in normoglycemic people. CGMs accurately reveal that glucose responses to food differ between people, but acting on that data has not been shown to improve metabolic health in healthy users.
- Are glucose spikes harmful for healthy people?
- There is no validated threshold defining a harmful spike in healthy people. Normative data show healthy adults sit in range about 96% of the day, so post-meal rises within range are not established as harmful.
- Can a CGM help a healthy person lose weight?
- Weight loss appeared only when a CGM was paired with a structured behaviour programme, not from the device alone. A monitor gives accurate, interesting data and may nudge short-term habits, but smoothing a normal glucose curve has no proven payoff.
- Does CGM-guided eating help you live longer?
- There are zero human outcome trials, so this is untested rather than disproven. No study connects acting on CGM data to any hard endpoint like diabetes incidence, cardiovascular events or lifespan in healthy, non-diabetic people.
Part of our guide: Longevity devices and therapies, fact-checked
Caveat is journalism, not medical advice. We check public claims against published evidence; we don’t diagnose, treat, or tell you what to take.