CGM: What gets measured gets managed!

Before starting, it’s worth understanding the CGM Black Swan

Why? The first question should always be:

“Do I understand the risk of using CGM readings for insulin dosing — and is that risk the same for all CGMs?”

For some devices, the risk is well characterised and clearly acceptable. For others, it isn’t. That difference matters, and it should shape how much risk we’re willing to take.

Ok, risk understood, now for understanding the measurement challenges of CGM.

TL;DR

Not all CGM systems report the same glucose level on their displays, despite measuring the glucose in the same place!

What!

I get why that feels ridiculous — surely if they all measure in the same place they should report glucose values in the same way? But they don’t.

CGM systems can be grouped into three categories based on what the on-screen number represents — and that on-screen number is the one you act on: treating hypos, giving correction insulin, adjusting food, exercise, or bolus timing.

Those same displayed values also generate your percentage Time in Range (TIR) and Time Below Range (TBR), typically defined as:

• TIR: 3.9–10.0 mmol/L (70–180 mg/dL)
• TBR: <3.9 mmol/L (<70 mg/dL)

They’re the metrics you and your diabetes team use to judge control and risk. So if two devices report systematically different numbers, they can make the same physiology look better or worse on paper, and change decisions!

That’s why knowing which Zone your CGM reads in really matters. It affects day-to-day decisions, how you interpret patterns, what treatment changes you make, and ultimately how well the data reflects your true glucose exposure and long-term health risk.

The Zones

Here we assume you are using a Zone P device.

Zone B: reads below true physiological glucose

These systems typically read lower than venous glucose when above 10.0 mmol/L or 180 mg/dL.

They can trigger low glucose alerts too early, smooth post-meal highs, and make TIR appear higher than actual glucose exposure.

To match the long-term risk profile of a Zone P system achieving 70% TIR, a Zone B device usually requires closer to 75–80% TIR.

But don’t fret: hitting 75–80% TIR on a Zone B system is typically easier than on a Zone P system.

Why?

Because Zone B sensors under-read when abive 10.0 mmol/L (180 mg/dL), so glucose appears to rise more slowly and peak lower than it would on a Zone P CGM system.

This is why 70% TIR on one CGM is not automatically equivalent to 70% TIR on another. It depends entirely on whether the system displays glucose values close to the body’s true exposure or consistently below it.

Where current systems sit

Most established CGM systems fall clearly into Zone P or Zone B. However, a growing number of devices provide no publicly available paired accuracy data, making it impossible to classify them — and this lack of transparency represents an unknown risk on many levels.

CGM systems reading in Zone P

These systems align with physiological glucose exposure and match the CGM systems that provided the data from which current global CGM targets (including 70% TIR) were derived.

CGM systems reading in Zone B

These systems read below venous glucose when above 10.0 mmol/L (180 mg/dL). This results in higher reported TIR percentages that under-report true physiological blood glucose exposure.

CGM systems with unknown measurement zone

These systems cannot be classified because they either publish no performance data, do not report study design in sufficient detail, or include insufficient paired glucose samples. Without this information, it is not possible to state with confidence whether they read in Zone A, P, or B when above 10.0 mmol/L (180 mg/dL)

Zone P vs Zone B CGM Sytems

If a CGM system under-reports glucose above 10.0 mmol/L (180 mg/dL)—particularly during rising glucose—this can materially influence insulin dosing decisions made by users, AID systems, and clinical therapeutic decision-making.

Compared with CGM systems that do not under-report in this range, under-reading devices can distort perceived glucose stability, alter how TIR maps to HbA1c, and ultimately affect interpretation of long-term risk.

Understanding a CGM system’s measurement zone allows users to account for system behaviour when making decisions, enables clinicians to adjust clinical targets appropriately, and supports informed adjustments to AID settings and medication behaviour.

Understanding the physiology: why capillary and venous glucose differ

Capillary blood reflects glucose freshly delivered to tissues, before uptake. Venous blood reflects what remains after the tissues have extracted glucose for energy. This creates a consistent 5–10% difference between them, and after meals the gap can reach 30%.

Every CGM system must choose which physiological signal it aligns with. That choice determines how glucose appears on the screen, how quickly levels seem to rise or fall, and ultimately how AID algorithms and users behave in response to the glucose values and trend arrows.

The identical twin problem: same TIR, different outcomes

Imagine two identical twins with type 1 diabetes. One uses a Zone P CGM; the other uses a Zone B system. Both hit 70% TIR for 20 years. One develops complications earlier and has consistently higher HbA1c. Why?

Set aside the biological variation in HbA1c that we know is real. Yes — people with genuinely similar glucose exposure can still end up with different HbA1c values because of factors like differences in red blood cell lifespan, or varying rates of vitamin C use to protect red blood cells, as the very clever Dr James Hempe has described. Other biological mechanisms likely contribute as well, but they’re beyond the scope of this piece. If you want the full dive on what this variation is, why it happens, and why it genuinely matters for understanding future health risk, read this article on HbA1c and CGM.

But here’s the key move: identical twins share essentially the same biology. So if identical twins show different HbA1c values despite the same apparent CGM-measured exposure, what explanation is left?

The remaining candidate is measurement error. Specifically, a Zone B device is systematically under-reporting true glucose exposure. The CGM makes time-in-range look better than it really is, but HbA1c acts as the audit trail and exposes the mismatch.

In short: when biology is held constant, a persistent HbA1c gap in the face of “identical” CGM data implies the CGMs are reporting a different glucose level despite measuring the same glucose in the fat tissue (interstitial space) — and HbA1c is calling its bluff.

The health cost?
The literature suggests that for each 10 mmol/mol rise in HbA1c (≈1.0%), the risk of both microvascular and macrovascular complications increases by roughly 30–40%. So a 5 mmol/mol (0.5%) rise maps to about a 15–20% increase.

But read that carefully: this is relative risk, not absolute risk. It does not mean your chance of complications is suddenly “30–40%”. It means your risk increases by 30–40% relative to whatever your baseline risk already was.

That distinction matters a lot. If your baseline absolute risk is low, a big-sounding relative increase may translate into a tiny real-world change. For example, a 40% relative increase on a 1% baseline risk raises it to 1.4% — not nothing, but hardly the apocalypse. On the other hand, if your baseline risk is already high, the same relative increase is a serious deal: a 40% relative rise on a 20% baseline risk pushes it to 28%.

This relative-versus-absolute risk confusion is exactly why headlines like “risk goes up 40%” regularly mislead people. If you want a great general explanation of why this matters and how to think about it properly, Peter Attia lays it out clearly in his studying studies guide.

How TIR leads to HbA1c — and why CGM zone affects both

Mounting evidence shows a clear relationship: each 10% change in TIR corresponds to about a 5 mmol/mol (≈0.5%) change in HbA1c. Importantly, Zone P systems were used to generate the data for these relationships.

Newer long-term studies show that a sustained 5% improvement in TIR predicts around a 20% reduction in risks for eye and kidney disease. But again, this is based on Zone P CGM data. If your CGM measures in Zone B, a reported 70% TIR very likely will mean a higher physiological glucose exposure, than someone hitting 70% TIR on a Zone P CGM system.

A 5% improvement in TIR on a Zone B system is likely equivalent to the same benefit for a 5% improvement on a Zone P system.

But in absolute terms, real-world glucose exposure is different, because Zone B systematically under-reads.

That means: to reach the same long-term risk profile, the reported TIR you’d need on Zone B is likely higher — plausibly by about 5–10% — to compensate for the device’s bias.

So again, this is the relative–absolute trap: equal TIR does not mean equal biological reality underneath. HbA1c (or other outcome-anchored measures) is what tells you whether the “improvement” was real or just a nicer-looking graph.

A practical analogy: the degree classification problem

If you study for four years and your tutor marks generously, telling you you’re averaging 70%, you expect a first-class degree. But if their marking runs 10% high, your true score might be closer to 60% — and your future opportunities change dramatically.

CGM works the same way. A generous (Zone B) CGM can make your glucose exposure look “first class” while your true exposure is not.