The Science

Measuring insulin at
picomolar sensitivity

Insulin circulates at 25–600 pM — one million times more dilute than glucose. That is why insulin measurement has never left the laboratory. Until now.

The Challenge

Why insulin is hard to measure

It's a protein, not a small molecule. It can't be sensed continuously. No wearable sensor exists, and none is in the FDA pipeline. Specialty immunoassay labs are slow, expensive, and disconnected from primary care workflows.

Speed
Lab immunoassay
Days — send-out turnaround
Metsulin strip
Minutes — point-of-care result
Setting
Lab immunoassay
Specialty lab only
Metsulin strip
At-home fingerstick
Analytes
Lab immunoassay
Single analyte per assay
Metsulin strip
Insulin + glucose, multiplexed
Bench Validation

The science works.

0.96 pM
Limit of Detection

Insulin circulates at 25–600 pM. We detect far below that range.

98–99%
Spike Recovery

Accuracy across low, medium, and high concentrations.

R² = 0.999
Linearity vs ELISA

Correlated against the gold-standard immunoassay.

94%
Signal Stability (9 weeks)

Bench-validated shelf-life performance.

0.01 s response time·10× reusable cycles

Bench-validated on plasma. Next milestone: whole-blood lateral-flow integration.

Sensor Architecture

How the sensor works

1
Step 1

Whole Blood Sample

~20 µL from a fingerstick. Passive capillary transport into the lateral-flow strip.

2
Step 2

Multiplexed Electrochemistry

NiOx electrocatalytic insulin detection on a closed bipolar electrode architecture, co-located with glucose sensing.

3
Step 3

CDSS Output

Insulin Sensitivity Score + Pancreatic Function Score fed into the 5-phenotype classification engine.

Clinical Decision Support

Five metabolic phenotypes. One test.

The same fasting glucose can reflect five different disease mechanisms. The CDSS classifies each patient and surfaces the appropriate clinical action.

P1

Beta-Cell Dysfunction

P2

Incretin Impairment

P3

Hepatic Insulin Resistance

P4

Muscular Insulin Resistance

P5

Clearance Defect