What a continuous glucose monitor actually measures

A continuous glucose monitor (CGM) is a small biosensor worn on the upper arm or abdomen. A wire-thin filament sits just under the skin in interstitial fluid (the fluid between cells, not blood). The sensor measures glucose in that fluid every 1–5 minutes for 10–14 days.

Interstitial fluid is not blood#

The most important technical point: a CGM does not measure blood glucose. It measures interstitial fluid glucose, which equilibrates with blood glucose with a delay of roughly 5–15 minutes depending on the device and the rate of glucose change. When your blood glucose is rapidly rising or falling, the CGM reading lags. When it's stable, the readings agree closely.

This matters for fast events. The peak of a sugar-driven spike is usually reported later by a CGM than by a fingerstick. The depth of a drop after exercise is similar. For trend analysis over hours and days the lag is unimportant; for capturing the precise moment of a peak it matters.

Accuracy in practice#

Modern devices in the UK (Abbott Freestyle Libre 2/3, Dexcom G7) report Mean Absolute Relative Differences (MARD) versus laboratory plasma glucose of around 7–9% in validation studies. That's the population average. Individual sensors can run higher or lower than capillary blood for the first day or in particular skin sites. Most manufacturers recommend a 24-hour warm-up period before relying on readings.

What it's licensed for, in the UK#

CGMs are regulated as medical devices in the UK; they are widely available on the NHS for people with type 1 diabetes and certain people with type 2 diabetes. They are also sold over the counter for self-paid wellness use (Lingo, Levels, ZOE), framed as educational rather than diagnostic.

A CGM is not a diagnostic device for a healthy person. A pattern that looks alarming to a layperson may be entirely normal; HbA1c, fasting glucose, and an OGTT are what UK guidelines use to assess glucose dysregulation, not CGM curves alone.

What it's good for in healthy users#

The reasonable uses for a non-diabetic, in our reading of the literature:

• Personalising meal choice by comparing your post-meal curves across foods you'd eat anyway.
• Spotting outliers — meals that produce dramatically larger or longer responses than your baseline.
• Linking lifestyle inputs (sleep deprivation, alcohol, illness) to next-day glucose patterns.

What it's not good for in a healthy user: comparing your spikes against published "ideal" ranges. Those ranges were defined in diabetic populations and are not validated as health benchmarks in the general population.

A note on action#

If a CGM repeatedly shows post-meal readings in the 11–14 mmol/L range or a fasting glucose consistently above 7 mmol/L in a non-diabetic, see your GP. That's a fingerstick-and-HbA1c conversation, not a wellness conversation.