Diabetic Dog Monitoring: Curves, Sensors, and a Dog That Is Not Controlled

The 2026 AAHA guideline for dogs says the dog is the measure: "the presence or absence of clinical signs is often a more accurate marker of DM control than any glycemic data obtained." The numbers still matter, because they are the only way to catch low blood sugar and the only way to find out why a dog is not getting better. This page is what the numbers are, how they are taken, which ones to doubt, and the guideline's fixed order of things to check when the signs will not go away. The disease itself, the insulins and the diet are on the main diabetes page.

Spot checks tell you almost nothing

A single reading is the thing owners most want and the guideline trusts least: "Spot-checking BG is unreliable". BG is blood glucose, the sugar reading, in the milligrams per deciliter US meters use. The reason is that a dog's sugar moves through the day in a curve, from a high before the injection down to a low point and back up as the insulin wears off. One reading cannot say where on that curve it was taken, and the guideline allows isolated values only to lower a dose, never to raise one: "Using 'spot checks' or isolated BG values by themselves is not recommended as a sole reason to increase an insulin dose but can sometimes be used to decrease the dose (if hypoglycemia is verified)."

The glucose curve

A curve, BGC in the guideline's shorthand, is a day of readings: one before the morning meal and injection, then one every two hours for ten to twelve hours, at the clinic or at home. It shows three things: how low the sugar goes (the nadir), how long the insulin lasts, and how far the readings swing. "The ideal nadir is a BG of 80–150 mg/dL", and over the day "The BG readings should ideally range between 80 and 300 mg/dL during the BGC as long as the BG remains <200 mg/dL for most of the dosing interval assessed." The guideline keeps repeating the caveat: "Glucose curves are not perfect, so always interpret them in light of clinical signs and body weight changes."

If you do curves at home, the guideline wants a meter calibrated for animals, and it is blunt about the alternative: "Although glucometers designed for use in people are readily accessible to pet caregivers, the task force does not recommend their use owing to lack of veterinary-specific calibration, which can lead to inaccurate result reporting." Blood is taken from the edge of the ear, the inside of the lip, a paw pad that does not bear weight, or the elbow callus. It also says something few owners are told: "Not all clients are suited to the task of obtaining a home BGC", and repeated pricking of a dog that hates it "can negatively impact the human-animal bond." It is allowed to say no to home curves and have them done at the clinic.

The sensor

The other way to see a day, or two weeks, is a continuous glucose monitor, a CGM, a coin-sized sensor stuck to the skin, the same FreeStyle Libre people use. It reads every 15 minutes for up to 15 days, catches overnight lows a curve would miss, and the guideline says that although it is not calibrated for animals it appears "relatively accurate" for following trends. Two warnings come with it. The sensors fall off: "the sensors do not always function or remain attached to the patient for the full 15 days, but the information gained from even a few days of reporting is usually more helpful than a standard BGC." And the low readings are the ones to doubt. A low number on the sensor in a dog that looks fine should be checked with a proper meter before anyone changes a dose.

The guideline also says something about owners, gently. Having the numbers on your phone all day can turn into changing doses on your own, calling about every reading, and wanting the sensor on the dog permanently. So "The task force recommends clinicians only use a CGM when medically necessary." If your vet does not want a sensor on a dog that is doing well, that is the guideline, not indifference.

Fructosamine and urine strips

Fructosamine is a blood test that reflects the average sugar over the previous week to ten days. The guideline is honest about its limits: "well-controlled diabetics can have elevated fructosamine concentrations. Conversely, pets with uncontrolled diabetes can have normal levels", so "fructosamine trends are often more useful than isolated values." Urine glucose strips, which many owners are still told to use, get a thumbs down: "routine UG measurement is not typically recommended", UG being urine glucose, partly because the strips underestimate, and partly because "a negative UG reading only means that BG was below the renal threshold", which can mean good control or too much insulin. Urine ketone strips are a different matter and are worth keeping for a dog that is unwell.

What the vet does with the low point

The guideline's dose rules hang on the nadir, the lowest reading of the day, and they are worth knowing so that a change makes sense when it comes.

Below 60 mg/dL is too low. If it was a sensor reading in a dog that looked fine, it is checked with a proper meter first. If it is real, the dog gets a small meal and the dose comes down: "Consider reducing the insulin dose by 10–25% at the next scheduled injection time." If there were signs of low sugar, insulin is held until the reading is back above 250 mg/dL and restarted at 25 to 50% less.
Between 60 and 150 mg/dL is the target. "If clinical signs are controlled, no adjustment is needed." If the dog still has signs with a good low point, the guideline looks at how long the insulin is lasting, at day-to-day swings a sensor would show, and at other diseases, rather than at the size of the dose. The answer is a different schedule or a different insulin, not a bigger dose.
Above 150 mg/dL with signs means a small increase, "0.5-1 U/dose in small to medium-sized dogs" or "1-2 U/dose in medium to large breed dogs", then another curve or sensor in 5 to 14 days. Above 150 with no signs and a steady weight means no change and a recheck in one to three months.
A dose that has climbed past 1.5 units per kilogram without the sugar coming below 250 mg/dL is the guideline's signal to stop raising it and look for what is blocking the insulin, which is the list below.

When it is not working

An uncontrolled dog is one that still has the signs: "An uncontrolled diabetic dog has persistent clinical signs (e.g., PU/PD/PP, weight loss)." When that happens the guideline gives vets a fixed order of things to check, and the order is worth knowing because the first three are about the house, not the dog.

The insulin and the injection are checked first: the vet watches how you shake or roll it, how you draw it up, which syringe you use, and where the needle goes, and replaces insulin that is out of date, cloudy in the wrong way, or discolored.
The food comes next. "It is essential for clients to be honest about food additives/toppers and treats offered regularly at home." The vet is not judging you; a handful of cheese a day can be the whole problem.
Low sugar hiding as high sugar is ruled out. Too much insulin can push the sugar so low that the body floods it back up, and a spot check catches only the rebound. The fix is less insulin, not more, which is why a sensor or a curve comes before any increase.
Then the vet looks for something blocking the insulin. The guideline sets a threshold: "Consider insulin resistance when intermediate-acting or long-acting insulin doses approach or exceed 1.5 U/kg/injection and BG remains greater than 250 mg/dL." The first suspects are medicines. It tells vets to ask "about steroid-containing creams, eye or ear drops, and progestins that may be transferred from a family member, such as topical hormone-replacement therapies." A person's hormone gel on the hand that pets the dog is on the list. After medicines come the diseases: dental disease, urinary infection, pancreatitis, Cushing's, an underactive thyroid, kidney disease, and in an intact female, her hormones.
Cushing's is tested for, but not too soon. HC is the guideline's shorthand for it: "Generally, endocrine testing for HC should not be performed before diabetic regulation has been attempted for at least 2–4 wk, because unregulated diabetes can lead to false-positive cortisol results." The urine cortisol screening test is not recommended in diabetic dogs at all, because it is so often positive.
The thyroid is tested later still: "In most cases, delay testing for hypothyroidism for 4–6 wk after DM diagnosis and treatment initiation", because untreated diabetes lowers the thyroid number by itself.
Whatever was found is treated before the insulin is touched again.
Only then is the insulin itself changed. In the guideline's words, where IR is insulin resistance: "If the patient has never been regulated, has no cause of IR identified, and has been given only one insulin type, consider switching to a different insulin formulation."
A specialist is called in if none of that has worked.

If your dog has been on rising doses for weeks without getting better and none of this has happened, ask for the list. It is the guideline's, not ours. The signs of the two emergencies, and what to do in the first minutes, are on the emergencies page.

Sources — 23 references