Why is my CGM reading different from a finger prick?

Two different fluids, two different timelines

When people first wear a continuous glucose monitor (CGM) such as a Dexcom or FreeStyle Libre, they frequently test with a traditional capillary fingerstick meter and find numbers that disagree by 20, 30, or even 50 mg/dL (1.1 to 2.8 mmol/L).

This rarely means either device is broken. They are measuring glucose in two physically distinct biological compartments operating on a time delay.

1. Interstitial fluid lag (5 to 15 minutes)

A finger prick samples capillary blood from microvessels in your fingertip. Glucose absorbed from your gut enters the bloodstream first.

A CGM sensor wire sits in the subcutaneous tissue layer beneath the skin and measures interstitial fluid — the fluid surrounding your cells. Glucose must diffuse out of the capillaries and filter through the vascular endothelium before reaching interstitial tissue.

• When glucose is stable (flat trend arrow): Capillary blood and interstitial fluid have reached equilibrium. The CGM and fingerstick will usually match within 5% to 10%.
• When glucose is rising rapidly (e.g., 30 minutes after pasta): Blood glucose spikes first. The fingerstick might register 165 mg/dL while the CGM still reads 130 mg/dL with an upward arrow.
• When glucose is falling rapidly (e.g., post-exercise or after bolus insulin): Blood glucose drops first. The fingerstick may read 85 mg/dL while the CGM still displays 115 mg/dL with a downward arrow.

In fast-changing states, the fingerstick tells you what your blood sugar is now, while the CGM reflects where it was 5 to 15 minutes ago.

2. Permissible clinical tolerance (MARD)

No consumer glucose meter provides clinical laboratory perfection:

• Traditional fingerstick meters adhere to ISO 15197 standards: 95% of readings must fall within $\pm 15%$ of laboratory reference values when blood glucose is above 100 mg/dL.
• Modern CGMs achieve a Mean Absolute Relative Difference (MARD) of roughly 8% to 9%, which means individual readings regularly vary by $\pm 10%$ to $15%$ from a lab draw.

If your true blood glucose is 120 mg/dL:

• A completely accurate fingerstick meter can legally read anywhere between 102 and 138 mg/dL.
• A completely accurate CGM can read between 105 and 135 mg/dL.
• Together, they can diverge by 25 mg/dL while both remain within regulatory safety limits.

3. Compression lows overnight

If you wake up to an alarming low reading (such as 55 mg/dL) that promptly shoots back up to 95 mg/dL upon standing, check how you were sleeping.

Lying directly on top of the sensor compresses the local capillaries and temporarily blocks blood perfusion around the filament. The local cells continue consuming glucose, draining the interstitial fluid immediately around the sensor wire and triggering an artificial "compression low."

Which reading should you trust?

1. When symptoms disagree with the CGM: If your sensor reads 120 mg/dL but you feel shaky, sweaty, lightheaded, or confused, trust your symptoms and confirm immediately with a finger prick.
2. Before dosing corrective insulin: When glucose is fluctuating rapidly or directional arrows point straight up or down, base critical meal or correction boluses on a fingerstick.
3. Use the CGM for trends and rate of change: The greatest value of a CGM is not an isolated single point in time, but the velocity curve showing whether you are stabilizing or continuing to climb.