CGM for Diabetic Cats: What the Libre 3 Really Reads

A FreeStyle Libre 3 sensor glued to a diabetic cat usually stops reporting on day 11, not day 14. That median comes from a 2026 study of 20 client-owned diabetic cats, and it is the single most useful number for planning both your budget and your monitoring schedule. The same study found 99.7% of 288 sensor readings in the clinically safe zones of the Parkes error grid, yet the sensor still averaged 19.4 mg/dL below the blood value. Both facts have to be held at once when you are deciding tonight's insulin dose.

What the Sensor Actually Measures

A CGM does not read blood. A short filament sits a few millimetres into the subcutaneous fat and reads glucose in interstitial fluid, then a transmitter pushes a value to your phone every 60 seconds — 1,440 readings a day, against roughly 6 from a 12-hour in-clinic blood glucose curve. The Libre 3 label specifies 14 days of wear, a 60-minute warm-up, no fingerstick calibration and no scanning. Because it samples tissue fluid rather than blood, the number trails the blood value by several minutes, which matters most when glucose is falling fast after an insulin injection.

Accuracy in Diabetic Cats: Three Numbers

The 2026 Journal of Veterinary Internal Medicine study compared sensor readings with an AlphaTrak 2 meter in 20 diabetic cats. Correlation was strong (Spearman r = 0.95), mean absolute relative difference (MARD) was 13.6%, and mean bias was −19.4 ± 46.0 mg/dL. For context, Abbott quotes a MARD of 7.9% for the same sensor in people, so expect roughly double the error in a cat. Clinical accuracy, the part that decides whether you act, held up: 287 of 288 pairs (99.7%) fell in Parkes zones A and B.

One caveat in that dataset: 279 of 288 pairs (97%) came from the euglycemic 70–180 mg/dL and hyperglycemic ranges. The lows, which are the dangerous end, were barely sampled. A 2025 clamp study in 7 healthy cats filled that gap with 474 paired measurements: r = 0.86 when glucose was stable, and 99–100% in zones A and B even during experimental hypoglycemia.

The Error Changes Direction at Low Readings

This is the detail that decides whether a CGM is safe in your house. Sensor error is not a fixed offset; it flips sign as glucose falls.

Blood glucose range (mg/dL)	Mean sensor minus blood (mg/dL)
91–120	−11.7 ± 11.2
66–90	−5.5 ± 9.1
56–65	−1.5 ± 6.0
46–55	+6.3 ± 5.7
Below 45	+15.7 ± 5.6

In the 60–120 mg/dL band the sensor under-reads, so a displayed 70 is probably closer to 80. Below 55 mg/dL it over-reads, so a displayed 50 may really be 35. Treat every reading under 60 mg/dL, and every cat that looks wobbly, as a reason to confirm on a veterinary blood meter before you touch the insulin pen.

How Long a Sensor Lasts on a Cat

Eleven days is the median in the 2026 series, with a range of 4 to 14 days. An earlier validation of the first-generation Libre in 8 outpatient diabetic cats was harsher: 12 of 15 sensor placements (80%) ended in failure or displacement, with a median working life of 7 days (range 2–13), even though correlation with a laboratory analyzer was excellent at 0.96. A sensor that peels off early is normal cat wear, not a mistake you made.

Budget 2 to 3 sensors per month: 30 ÷ 11 = 2.7.
Write the placement date on the calendar; day 10 is when to check adhesion.
If a sensor falls off after 5 or more days, keep the data. The 2026 AAHA guidelines note that a few days of CGM still beats an 8-hour hospital curve.

Placement: Fifteen Minutes Decide the Next Eleven Days

Everything that goes wrong with feline CGM starts at the skin. The published protocol is unglamorous and worth following exactly.

Clip a 5 × 5 cm patch over the dorsal neck or lateral thorax, then scrub with dilute chlorhexidine and wipe with alcohol; let it dry for at least a minute.
Add 3 to 4 drops of tissue glue to the adhesive side, press the applicator firmly, and hold it in place for 30 seconds after deployment.
Cover the sensor with a stockinette, a snug medical shirt or a transparent film dressing. The AAHA FAQ lists these as the main defence against a cat removing its own hardware.
No swimming or baths while it is on, and no MRI, CT or radiography — those are explicit exclusions.
At removal, soak a cotton swab in acetone and roll it under the adhesive pad instead of pulling.

Reading the Data the Way the Guidelines Do

The 2026 AAHA diabetes guidelines frame a CGM as a decision tool, not a live scoreboard. Place it when there is a medical question: 5 to 14 days after starting insulin or after any dose change, when clinical signs return, or when hypoglycemia is suspected. On a good day the nadir should sit between 80 and 150 mg/dL, readings should mostly stay between 80 and 300 mg/dL, and insulin effect should last close to 12 hours. Judge the shape of the curve, not the 2 a.m. spike that scared you. The urgent low alarm at 55 mg/dL cannot be silenced; the low alarm defaults to 70 mg/dL and can be adjusted in the app.

When a CGM Is the Wrong Tool

The same guidelines say not to place a sensor just to watch a well-controlled cat, because it invites obsessive checking and unsupervised dose changes. Validation also excluded cats with a packed cell volume under 27%, so accuracy in anemic cats is simply unknown. Skin reactions are common but mild: 4 of 20 cats had erythema where the adhesive sat. Call your vet the same day for a reading under 60 mg/dL with clinical signs, a sensor that dies within 24 hours of placement, a cat that stops eating, or repeated readings above 350 mg/dL. Practical background on insulin handling and remission sits in our guides to insulin storage and curves and diabetes in cats aged 8 to 12.

What It Costs

Prices below are per sensor at retail or cash price, before any clinic placement fee. Most clinics charge a separate technician appointment to shave, glue and cover the sensor.

Item	US (USD)	UK (GBP)	China (CNY)
One sensor	$69–99	£60–70	¥400–500
Two-sensor pack, cash	$75–85	£120–140	¥800–1,000
One month at 11 days per sensor	$110–260	£160–190	¥1,100–1,350
Veterinary blood meter, one-off	$50–75	£40–70	¥200–400

The Chinese figure reflects the Libre 2 generation sold locally as 瞬感, which is the model most owners there can actually buy; the Libre 3 is not the mainstream retail product in that market. Cross-check the annual total against our diabetes supplies cost breakdown, and if your cat is heading toward remission, read what a 4-week remission curve looks like before you buy a third sensor.

This article is for education and does not replace a veterinary diagnosis. Sensor placement, insulin dose changes and the interpretation of any glucose reading belong to the veterinarian treating your cat; never adjust a dose on the strength of one number.

Frequently asked

Can I just buy a Libre 3 and put it on my cat myself?

The sensor itself needs a prescription, and the feline technique is different from the human one: a clipped 5 × 5 cm site on the neck or flank, tissue glue, and a cover to stop the cat removing it. The 2026 AAHA guidelines also say a CGM should be placed when it answers a medical question, not for a cat that is already well controlled. Self-placed sensors are a common reason owners change insulin without telling anyone.

The sensor says 55 mg/dL but my cat is acting normally. Do I treat?

Below 55 mg/dL the Libre 3 over-reads in cats, by roughly 16 mg/dL on average in the 2025 clamp study, so a displayed 55 may correspond to a blood glucose in the 40s. Confirm on a veterinary blood meter before doing anything. If the meter reads above 60 and the cat is alert and eating, record it, keep the sensor on, and call your vet before the next injection rather than skipping or doubling a dose on your own.

How many sensors will I need in a month?

Plan on 2 to 3. Median sensor life in diabetic cats was 11 days (range 4–14) in the 2026 study, and an earlier first-generation study saw 80% of placements fail or displace with a 7-day median. At 11 days, 30 ÷ 11 = 2.7 sensors per month, so a 3-sensor budget covers a bad month without a scramble.