The short answer
Diabetic ketoacidosis (DKA) is a decompensated form of diabetes mellitus. Insulin is lacking and counterregulatory hormones are elevated, so fat is broken down into ketoacids that accumulate in the blood and cause acidemia. Diagnosis rests on hyperglycemia, ketonemia or ketonuria, metabolic acidosis and compatible clinical signs, and blood gas analysis is needed to confirm the acidosis. The sources describe treatment as fluid therapy, insulin, electrolyte supplementation, and a search for the concurrent disease that most affected dogs have. A 2019 review reports that roughly 70% of dogs and cats with DKA survive to discharge, and no published canine DKA guideline is among the sources for this page.
Key takeaways
- A 2019 Veterinary Ireland Journal review reports that about 70% of dogs with DKA have a concurrent disease process, which makes finding a trigger part of the workup [1].
- That review states that dogs with ketones above 3.8 mmol/L are more likely to have DKA, and a Canadian Veterinary Journal review cites a study in which dogs with blood ketones of 3.5 mmol/L or more were at higher risk [1][2].
- Standard urine reagent strips use a nitroprusside reaction that detects acetoacetate but not beta-hydroxybutyrate (BHB), the predominant ketone in DKA, according to both reviews [1][2].
- The 2019 review advises that potassium supplementation is likely indicated regardless of the starting potassium value, because insulin drives potassium into cells [1].
- In a 24-case nonrandomized study, median time to resolution of ketosis was 12 hours with intramuscular insulin lispro versus 23 hours with an intravenous infusion of regular insulin [3].
- At least 7% of dogs have recurring DKA episodes, and the median hospital stay for dogs is reported as six days [1].
Doses on this pageEvery dose below is quoted from the cited source. Verify against the current label/formulary and the patient's full clinical picture.
What DKA is and which dogs develop it
The Merck Veterinary Manual describes DKA as decompensated diabetes. In this state, cells turn to free fatty acids for energy because insulin deficiency blocks their access to glucose, and glucagon and other counterregulatory hormones drive breakdown of those fatty acids into ketoacids [4]. The Canadian Veterinary Journal review explains that low intracellular glucose diverts oxaloacetate into gluconeogenesis, pushing excess acetyl-CoA toward ketone formation. It adds that a raised glucagon-to-insulin ratio is needed to amplify ketogenesis [2]. Ketones dissociate into hydrogen ions, and once production outpaces the body's buffering and tissue use, acidemia follows [1][2].
The CVJ review describes DKA as a disease of middle-aged dogs and cats [2]. According to the 2019 review by Poppy Gant of the Royal Veterinary College, it develops most often in undiagnosed or newly diagnosed diabetics not receiving enough insulin, though any diabetic patient is at risk, particularly with comorbidities [1]. In the Bologna insulin study, concurrent disorders found in 13 dogs included pancreatitis, Cushing's syndrome, urinary tract infection and chronic kidney disease [3].
Clinical presentation
Signs are nonspecific. Gant lists lethargy, anorexia, vomiting, dehydration and eventual collapse, with polyuria and polydipsia from the osmotic effect of glucose and ketones [1]. The CVJ review notes that presentation ranges from ketotic patients still eating and maintaining hydration to the more common dehydrated, ketoacidotic patient [2]. Among the 24 Bologna cases, owners reported polyuria/polydipsia in 22, anorexia in 18, lethargy in 17 and vomiting in 13 [3].
Gant's review states that abnormal mentation on arrival has been associated with a poor outcome. It also notes that rapid breathing may be compensation for acidosis but should raise the possibility of aspiration pneumonia in a vomiting patient [1].
Diagnosis and testing pitfalls
The CVJ review lists hyperglycemia, ketonemia or ketonuria, metabolic acidosis, glucosuria and associated clinical signs as the findings that establish DKA [2]. For study enrollment, the Bologna group required blood glucose above 250 mg/dL, blood BHB above 2.5 mmol/L, and venous pH below 7.3 and/or bicarbonate below 15 mEq/L [3]. Gant states plainly that "only practices with the ability to run blood gas analysis will be able to truly diagnose DKA" [1].
Ketone testing
Point-of-care ketone meters measure BHB, and Gant reports they correlate closely with laboratory methods [1]. Where blood gas is unavailable, her review gives ketone values above 3.8 mmol/L in dogs as making DKA more likely. It adds that values under about 1 to 2 mmol/L suggest ketosis is unlikely to be the primary cause of illness [1]. The CVJ review cites a canine study in which dogs below 2.8 mmol/L were unlikely to develop DKA [2].
Urine strips carry several known problems, per the CVJ review [2]:
- The nitroprusside reaction misses BHB.
- N-acetylcysteine, captopril and penicillamine can produce false positives.
- False negatives occur when blood acetoacetate has not yet crossed the renal threshold or when strips have been exposed to air.
Both reviews report that placing plasma on the strip improves sensitivity, with Gant calling heparinised plasma a reasonable rule-out tool [1][2]. The CVJ authors also wrote that because BHB rises as acidemia worsens, strips may underestimate ketosis in the sickest patients and overestimate it as DKA resolves [2].
Other tests
Gant notes that glucometers not designed for animals can read low, and gives a canine anion gap reference of 12-24 mEq/L [1]. Her review recommends an extended database of blood count, chemistry with electrolytes, urinalysis with culture, and thoracic and abdominal imaging. It advises against testing for hyperadrenocorticism at emergency presentation because of the high likelihood of a false positive [1]. Because diabetic patients may not mount a visible urine sediment response, she states that urine culture is always recommended [1].
Classification
None of the cited sources gives a formal staging system for canine DKA. The CVJ review separates ketosis without acidosis from ketoacidosis [2]. Merck distinguishes hyperosmolar hyperglycemic state from DKA, defining it as serum glucose above 600 mg/dL and osmolality above 320 mOsm/kg, and notes that mixed forms occur [4].
Treatment approaches described in the sources
Merck summarizes the goals as rehydration with IV fluids, insulin to reduce hyperglycemia and ketosis, potassium and phosphorus support, and treatment of underlying disease such as pancreatitis or infection [4].
Fluids
For patients in shock, Gant describes boluses such as 10 ml/kg over 15 minutes while reassessing perfusion. She states that no improvement after 20-30ml/kg should prompt a search for other causes of shock [1]. She notes 0.9% saline has been the traditional choice, while human research suggests buffered crystalloids are not linked to more cerebral edema [1]. Her review also warns that sodium shifts greater than 0.5 mEq/h risk complications from fluid moving between compartments [1].
Insulin
Gant writes that "there is no conclusive evidence that any particular protocol is superior," and that ketosis resolves only with insulin [1]. She cites one canine study that found no increase in complications when insulin began within six hours of presentation [1].
Merck reports regular insulin is most commonly used [4]. Its example regular insulin infusion for dogs is "up to 2.2 U/kg, IV CRI, every 24 hours, titrated to effect," with glucose checked every 1-2 hours and a caution that glucose "should not decrease by > 50-100 mg/dL/h" [4]. Merck also describes an intermittent regimen:
- 0.2 U/kg IM initially, then 0.1 U/kg IM every 60 minutes.
- 0.1-0.3 U/kg SC every 4-6 hours once glucose is below 250 mg/dL.
- 2.5%-5% dextrose added to fluids as glucose falls [4].
Merck states that NPH, lente or detemir can be started in dogs once they are rehydrated and eating and drinking on their own [4].
The Bologna study compared intramuscular insulin lispro with a regular insulin infusion [3]. Lispro was given at 0.25 U/kg initially and 0.125 U/kg IM every 3 h once glucose reached 250 mg/dL [3]. The study included 24 cases in 23 dogs. Ketosis resolved faster with lispro (12 h vs 23 h), while differences in acidemia, ketoacidosis and hospital stay were not significant [3]. A single lispro bolus achieved the target glucose drop in only 6 of 11 dogs [3]. Among the limitations, the authors cite a small sample without a power calculation, no randomization, and the time gap between the two treatment periods [3].
Electrolytes, bicarbonate and supportive care
Gant reports that nearly all patients are total-body potassium depleted, masked by dehydration and acidosis [1]. In the Bologna study, transient hypokalemia developed despite supplementation in 9 of 11 lispro dogs and 10 of 13 regular insulin dogs [3]. Hypophosphatemia developed in 75% and 61.4% of measured cases, respectively [3]. Gant notes dogs are somewhat more resistant than cats to hemolysis from low phosphorus. She also describes empirical magnesium (0.4mmol/kg/day) when hypokalemia fails to respond [1]. She writes that bicarbonate is "rarely necessary," citing risks including cerebral edema and worsening hypokalemia [1].
The same review suggests that broad-spectrum antimicrobials such as amoxicillin clavulanate are likely appropriate pending culture where sepsis is documented on cytology [1]. It also advises considering assisted feeding after more than three days of anorexia in dogs [1].
Monitoring in hospital
Gant describes electrolyte checks every four to six hours initially, close glucose monitoring during insulin infusions, and weighing every four to six hours to track hydration [1]. She cites a study finding subcutaneous glucose monitors clinically accurate in DKA, with better accuracy in well-hydrated patients [1]. The Bologna protocol measured glucose hourly for 24 hours, BHB every 4 hours until 1.0 mmol/L or lower, and venous blood gas every 8 hours for the first day [3]. The CVJ review notes that no veterinary studies have set a BHB value for ending intensive treatment. It proposes relying on falling BHB alongside clinical improvement [2].
Prognosis and client communication
Gant reports that about 70% of dogs and cats survive to discharge, with a median stay of six days in dogs, and that at least 7% of dogs have repeat episodes [1]. She states owners should be told about likely cost before treatment begins. She adds that patients with severe acidemia, electrolyte problems or comorbidities may warrant referral after stabilization [1].
For discharge, her review advises against trying to find the ideal insulin dose in hospital and suggests starting at the low end of the range [1]. The CVJ authors propose training willing owners to check blood ketones at home and seek care if BHB exceeds 3.5 mmol/L or is climbing [2]. Merck lists lente as the insulin of choice for dogs and recommends spaying intact diabetic females [4].
The protocol comparison data cited here come from a small, nonrandomized study. Its authors note that small case numbers are a common limitation in research on the topic [3].
