Diabetic ketoacidosis (DKA) is a complication of diabetes mellitus with concurrent and often severe metabolic derangements associated with hyperglycaemia, glucosuria, metabolic acidosis, ketonaemia +/- ketonuria. Patients with ketonaemia/ketosis are usually still bright, eating and maintaining their hydration. Those with ketoacidosis are dehydrated, clinically unwell (e.g., anorexia, vomiting, lethargy) and typically require hospitalisation and intensive management.

DKA is distinguished from uncomplicated diabetes mellitus (DM) by a relative insulin lack and increased counter-regulatory hormones. The latter are thought to occur secondary to intercurrent disease.

Concurrent disease has been documented in approximately 90% of cats with DKA, with the most common being hepatic lipidosis, chronic kidney disease, acute pancreatitis, bacterial or viral infections and neoplasia (Bruskiewicz et al. 1997).

Heinz bodies, neutrophilia with a left shift, increased ALT and azotaemia is common.

Most cats presenting with DKA are newly diagnosed diabetics or recently diagnosed but poorly controlled diabetics.

Diagnosis

Hyperglycaemia, Glucosuria, Metabolic Acidosis Plus Ketones in Plasma and/or Urine

Traditionally DKA has been diagnosed using urinary ketone dipsticks, which detect acetoacetate but not beta-hydroxybutyrate. However as the latter is the principle ketone body in DKA, measuring serum beta-hydroxybutyrate is a more sensitive indicator of DKA. In humans portable meters that measure beta-hydroxybutyrate in whole blood have largely superseded urine dipsticks. These ketone meters have recently proven useful in diagnosing DKA in cats, although they tend to underestimate beta-hydroxybutyrate at higher values (Zeugswetter, Rebuzzi 2012; Weingart et al. 2012). In the absence of a ketone meter, using urine dipsticks to measure plasma ketones is more sensitive than just checking for ketones in urine. A recent study cited a sensitivity/specificity of urine ketone dipsticks as 100/88% for plasma (with a cut off of 4 mmol/L) and 82/95% for urine (with a cut off of 1.5 mmol/L) (Zeugswetter, Pagitz 2009).

Venous blood gas analysis is used to detect metabolic acidosis (venous pH < 7.35, HCO₃ < 15 mmol/L). If blood gas analysis is not available, plasma bicarbonate levels may be estimated from TCO₂ measured on in-clinic analysers.

Therapy

Cats with ketosis (urinary/plasma ketones but without metabolic acidosis) who are bright and eating can be treated with lente insulin or glargine (Trotman 2010; Boysen 2008) or with regular crystalline insulin at 0.1–0.2 U/kg q 8hours (Nelson, 2005) with subsequent doses based on blood glucose levels, until the ketonuria resolves.

Those cats with systemic signs of DKA such as anorexia/vomiting/lethargy need hospitalisation and therapy with regular insulin and intravenous fluids. Monitoring of sick ketoacidotic patients is intensive. Acid-base, electrolyte and blood glucose levels can change rapidly during therapy, and close monitoring of these parameters is mandatory. For example, it is not uncommon for 3–4 changes of fluid type to be required in the first 24 hours of therapy. If in-house monitoring of electrolytes and acid base status is not available, referral should be considered, unless an outside laboratory can supply a result in a couple of hours.

The primary goals for therapy of DKA are:

 Correct dehydration and electrolyte deficits

 Treat with regular insulin to prevent ketogenesis and reduce hyperglycaemia

 Correct acidosis if required

 Identify and remove any precipitating factors (e.g., infection)

 Provide nutritional support

As intercurrent disease is common, the data base for patients with DKA ideally includes: CBC, full serum biochemistry (+/- T4) and electrolytes, blood gases, fPLI, UA (SG, dipstick, sediment and culture) +/- chest X-rays and abdominal ultrasound, +/- FeLV/FIV.

1. Correct Dehydration and Electrolyte Deficits

Intravenous fluids are essential in management of DKA. In addition to rehydration, fluids will help lower plasma glucose by dilution and increasing glomerular filtration rate. An isotonic fluid such as 0.9% saline or Hartmans/lactated ringers solution (LRS) is a suitable initial choice in most cases. Many cats with DKA are hypovolaemic at presentation and require initial stabilisation with IV fluid boli. Parameters such as blood pressure, HR, pulse quality, mucous membrane colour, capillary refill and mentation dictate whether fluid boli are needed prior to starting rehydration rates.

Fluid deficits are calculated with standard formulae based on the estimate of dehydration; % dehydration x body weight (kg) x 1000 = ml of fluid deficit.

There is usually a total body sodium and potassium deficit, but serum levels may be low, normal or increased. Metabolic acidosis, hypoinsulinaemia and hyperglycaemia drive potassium extracellularly, and help mask the potassium deficit. Commencing IV fluid therapy and insulin plus correcting metabolic acidosis will drive potassium back into cells and enhance urinary potassium losses. This means that profound hypokalaemia can develop during therapy for DKA and is one of the reasons why insulin should ideally be withheld until serum potassium exceeds 3.5 mol/L. Replacement of potassium should be based on measurement of serum potassium (see Table 1) using IV potassium chloride.

Phosphate shifts from the intra and extracellular compartments in parallel with potassium, hence hypophosphataemia commonly develops during the first 12–24 hours of therapy for DKA. Severe hypophosphataemia can cause life threatening haemolysis in cats. To minimise the risk of severe hypophosphataemia, some authors recommend replacement of 1/3–1/2 of the total potassium deficit with KPO₄. Certainly if phosphorous levels drop below 0.5 mmol/L or if haemolysis occurs, KPO4 (4.4 mmol/ml K and 3 mmol/ml PO₄) is added to 0.9% saline and delivered at 0.03–0.12 mmol/kg/h. KPO₄ should not be added to calcium containing fluids such as LRS, as CaPO4 may precipitate. Phosphorous supplementation is not recommended in cats with renal disease, hypercalcaemia, hyperphosphatemia or oliguria.

Table 1. Potassium supplementation for iv fluid therapy: mmol = mEq

NB: Max infusion rate is 0.5 mmol/kg/hr of potassium - if this rate does not increase serum K⁺, check Mg levels as hypomagnesaemia has been associated with refractory hypokalaemia.

Reduced plasma ionised magnesium (Mg) has also been reported in cats with DKA and has been associated with refractory hypokalaemia. If the latter occurs, serum Mg should be measured. Clinical signs of hypomagnesaemia do not usually develop until total Mg drops below 0.4 mmol/L or ionised Mg (iMg) below 0.2 mmol/L. Hypomagnesaemia can be corrected with MgSO₄ constant rate infusion (CRI) at 0.5–1 mmol/kg/24 hours. Mg levels should be rechecked at least once daily if supplementation is commenced. Mg toxicity has been reported in cats including weakness, flaccid muscle tone, vomiting, respiratory depression, mental dullness and hypotension.

The typical DKA cat is 6–10% dehydrated; this fluid deficit should be replaced over a 24 hour period unless the patient is in shock, has heart failure or anuric/oliguric renal failure. One author suggests that 20% of the calculated fluid deficit should be replaced in the first hour, 30% in the next 5 hours and 50% in the next 18 hours (Boysen 2008). In addition maintenance fluids at 1.5–2 times maintenance (3–4 mls/kg/hour) should also be provided. Fluid should be delivered via an infusion pump or syringe driver to ensure accurate delivery rates.

After initiating treatment with 0.9% saline or Hartmans, the fluid type, rate and supplementation with K/PO₄/Mg should be adjusted based on serial assessments of hydration, electrolytes and serum phosphate. For example serum Na and K should be measured every ~ 6 hours until they are stable in the normal range.

2. Start Insulin Therapy

Insulin is usually started within 2–4 hours of initiating IV fluid therapy. A delay of up to four hours may be indicated if there is marked hypokalaemia at presentation, allowing time for IV fluids to correct the hypokalaemia before starting insulin.

The goal of insulin therapy is to slowly reduce serum glucose by no more than about 3–4 mmol/hour. More rapid reduction of blood glucose can cause large shifts in osmolality with potentially serious consequences. Regular crystalline insulin (Actrapid) is recommended as it is rapid in onset and can be given intravenously in a low dose CRI or intermittent IM injections. IM administration should be avoided until the patient is rehydrated as insulin absorption can be unpredictable.

A separate IV line from your fluids is used to deliver a CRI of IV regular crystalline insulin at 1.1 U/kg/day (2.2 U/kg/day has been recommended in dogs and a recent study shows that it appears to be safe in cats (Claus et al. 2010) but my preference is to err on the side of caution with the 1.1 U/kg/day dose). An accurately calibrated infusion pump or syringe driver is needed; Add 1.1 U/kg of regular crystalline insulin to 250 ml 0.9% NaCl or Hartmans, run out 50 ml as insulin binds to plastic and glass surfaces. Guidelines for infusion rates based on blood glucose are outlined in Table 2, blood glucose should be checked 1–2 hourly and the infusion adjusted accordingly (NB rates and concentrations may need to be changed from those in Table 2 based on an individual's response to therapy). The CRI of insulin can be stopped and longer acting insulin (e.g., lente or glargine q 12 hours) started once the cat is eating and hydrated. Typical starting doses of lente insulin or glargine are 0.25–0.5 U/kg BID.

Table 2. CRI insulin adjustments with changes in blood glucose concentrations

Data adapted from Trotman 2010.

a. Hartmans/lactated ringers or Normosol-R can be substituted for 0.9% NaCl
Add 25 ml of 50% dextrose to 250 ml of fluids to make a 5% solution

An alternative to CRI insulin is the low dose intramuscular regime, but caution should be exercised using the IM regime if the cat is dehydrated as insulin absorption may be unpredictable.

A single injection of 0.2 U/kg regular insulin is given IM, followed by 0.1 U/kg IM every 1–2 hours (dilute insulin 1:10 with sterile saline and use 0.3 ml insulin syringes to titrate the dose accurately). Blood glucose should be measured hourly, and once serum glucose drops below 14mmol/L regular insulin injections should be reduced to 0.1–0.4 IU/kg every 4–8 hours and dextrose should be added to the fluids to make a 5% solution (add 50 ml of 50% dextrose to 500 ml of fluids). The insulin dose then needs to be titrated in each patient to keep it in the 11–16 mmol/L range. Longer acting insulin (e.g., lente or glargine q 12 hours) can be started once the cat is eating and hydrated, typical starting doses are 0.25–0.5 U/kg BID.

Can I Use Glargine to Treat Cats with DKA?

A recent abstract (Marshall et al. 2010) describes successful use of IM +/- SC glargine in 15 cats with DKA. The study was retrospective and data about the exact protocol was brief, but all cats were given IM glargine at 1–2 U/cat, and 12 of the 15 cats also received 1–3 U glargine SC, with SC glargine started later in the remaining 3 cats. Intermittent IM glargine was repeated as needed 2–22 hours later, with SC glargine (1–2 U) every 12 hours. 6/12 cats started on a combination of IM and SC glargine were managed with only SC glargine within 18 hours. Although this preliminary study is encouraging, where possible I would advise using one of the well-established protocols with regular insulin.

3. Correction of Acidosis

In most cases specific therapy for metabolic acidosis is not required as insulin and fluid therapy facilitates metabolism of ketone bodies and their urinary excretion. Serum bicarbonate supplementation has tended to fall out of favour in DKA, although the American Diabetes Association lists it as a treatment option if arterial pH is < 7 an hour after starting fluid therapy. In cats and dogs consideration may be given to bicarbonate supplementation if plasma HCO₃ or total venous CO₂ drops below 12 mmol/L (Feldman 2005); these animals are usually profoundly depressed.

4. Identifying and Treating Concurrent Illness

Concurrent disease has been documented in approximately 90% of cats with DKA. A thorough history and physical examination plus haematology, serum biochemistry and urinalysis are mandatory as identification of intercurrent disease will impact directly on therapy and prognosis. A cystocentesis urine sample should be always be cultured, as urinary tract infections are common in DKA and active urine sediment is not always present. Depending on the cat's physical exam findings other tests may also be indicated. For example moderate to severe pancreatitis may be identified with a combination of abdominal ultrasound and fasting feline specific pancreatic lipase immunoreactivity (fPLi).

5. Nutritional Support

Depending on the severity of DKA and intercurrent disease, enteral nutritional support may be needed. Nasoesophageal feeding tubes can be placed with local anaesthetic in critical patients. If ongoing nutritional support is needed, once the patient is stable, an esophagostomy tube can be placed.

Complications of Therapy and Prognosis

DKA is a complex disorder with the potential for high mortality if management is inadequate or if there is severe underlying disease. Complications usually arise from inadequate patient monitoring, or over-aggressive or under-aggressive therapy. Hypoglycaemia, cerebral oedema, hypophosphataemia and hypokalaemia are all potentially life-threatening complications of DKA. The outcome for an individual patient will depend on the severity of intercurrent disease and your ability to manage the metabolic complications of DKA.

In one study of 42 cats with DKA or ketosis without acidosis, 70% of cats survived to discharge. Median hospitalisation was 5 days, with up to 40% having recurrent episodes of DKA (Bruskiewicz et al. 1997). In another study, 7 of 12 cats with DKA achieved diabetic remission; four of these had received glucocorticoids prior to onset of DKA (Sieber-Ruckstuhl et al. 2008).

References

References are available upon request.