diabetes Logo
Diabetic Ketoacidosis

It is a mistake to feel that diabetic ketoacidosis does not occur in Type 2 diabetics.

DKA is a metabolic disorder consisting of three major abnormalities: elevated blood glucose level, high ketone bodies, and a metabolic acidosis with an elevated anion gap. Dehydration and hyperosmolarity are usually present. All patients may not meet all the above criteria.

Precipitating Factors of DKA:

When considering the precipitating factors for the development of DKA it is important to remember that DKA develops due to either an absolute or a relative absence of insulin. In Type 2 diabetes, it is usually a relative lack of insulin unlike that seen in Type 1 diabetics who usually manifest an absolute absence of insulin.

Relative lack of Insulin:

Acute Illness

* Infection or other inflammatory process

* Myocardial Infarction

* Stroke

* Trauma

Endocrine Disorders Drugs:


Calcium channel blockers


Beta-blocking agents




Clinical Features

polyuria, polydipsia, weight loss

anorexia, nausea, vomiting

abdominal pain

confusion, coma (5-10%)

P/E: dehydration, (sunken eyeball, reduced tissue turgor, dry tongue)


hyperventilation (Kussmaul's breathing)

smell of ketone

evidence of precipitation illness ( MI , Infection )

Click to view Pathogenesis of DKA and HONK


Diagnosis requires the demonstration of hyperglycemia, hyperketonemia, and metabolic acidosis. However, a presumptive bedside diagnosis is justified if the patient's urine or blood is strongly positive for glucose and ketones. A thorough search for a treatable infection must be made.

Laboratory Abnormalities:

In general, the laboratory diagnosis of DKA is based on an elevated blood glucose (usually above 250mg/dl), a low serum bicarbonate level (usually below 15 mEq/L), and elevated anion gap, and demonstrable ketonemia. Individually, all of these values may vary considerably, but taken together they help make the diagnosis of DKA. In addition to the above there are several calculations that are important in the evaluation and therapy of the patient with DKA.

Serum Osmolality: Mental status changes can occur in DKA and may be the result of DKA, or some underlying process that may have caused the patient to develop DKA. It has been well documented that mental status changes in DKA correlate with the effective serum osmolality. The effective serum osmolality is calculated as follows:

Serum Osmolality = 2(Na+) + glu/18 + BUN/2.8 Calculated total osmolalities of greater than 340 mOsm/kg H2O are associated with stupor and coma. Calculated values below this level would not explain a patient with coma and an additional cause such as meningitis, or stroke should be considered.

Corrected Serum Sodium Levels: Despite volume depletion, serum sodium may be low, normal, or elevated. This variation has several causes. When trying to determine the degree of dehydration in a patient it is best to use corrected serum sodium level. This can be calculated using the following formula:

Corrected Na+ = [Na+] + 1.6 x [glu in mg/dl] - 100 Often, the initial serum sodium appears low, but when the above calculation in performed, the final value is elevated. This indicates a marked intracellular dehydration.

Anion Gap: The ketoacids produced during DKA are buffered by the serum bicarbonate and then excreted in the urine. This causes a loss of bicarbonate which is a measured anion. As the bicarbonate is lost the anion gap increases.

The three ketone bodies are beta- hydroxybutyrate, acetoacetate, and acetone. Only acetoacetate and acetone are measured in the nitroprusside reaction, but the formation of these ketone bodies favors the development of beta-hydroxybutyrate. Thus, the test for ketone bodies may be only weakly positive even when large amounts of total ketones are present. Acetone does not contribute to the anion gap, but it is measured in the nitroprusside reaction and is a precursor for the regeneration of bicarbonate.

It is not uncommon for the patient to be improving clinically, but to have the nitroprusside test become more strongly positive since acetone is being produced. At this point, the anion gap should be narrowing, even as the nitroprusside test is getting stronger. Additional Laboratory Evaluation: When the patient arrives in the Emergency Department some initial labs should be sent. Many of these common tests will give the data needed to do the above important calculations. A tube should be sent for exact glucose determination, but a bedside test can me used to determine gross blood sugar levels. To determine the degree of acidosis and bicarbonate loss, an ABG should be sent early in the evaluation of a patient considered to have DKA.

The complete blood count often shows an elevation of the white blood cells. This may be, in part, due to hemoconcentration secondary to dehydration. Thus, WBCs of 20,000 occur commonly. Those patients with WBC,s greater than 30,000 who have a bandemia on peripheral smear should be assumed to have an infectious process.

Additional evaluation should take into consideration the best tests to help determine the potential cause of the patient's decompensation into DKA. Urinalysis, chest radiograph, and electrocardiogram should be done on most patients.

Click to view table of Differential Diagnosis


Click to view Algorithm

The treatment goals of the patient with DKA are as follows:

(1) improve hypovolemia and tissue perfusion,

(2) decrease the serum glucose,

(3) reverse ketonemia and acidemia at a steady rate,

(4) correct electrolyte losses and imbalances,

(5) find and treat the underlying cause of the patient's DKA.

Hydration Therapy: Patients with DKA are invariably dehydrated and foremost in the treatment of DKA is restoration of the intravascular volume. Estimates of fluid deficits in the decompensated diabetic is 4 to 10 liters (usually 5-6 liters). Enough fluid should be given to approximate this amount.

Initially, one to two liters of normal saline is given within the first hour followed by 1 L/hour for the next several hours. This initial management should be guided by the patient's general condition and response, with more or less fluid as indicated. After the first 3-4 hours, as the clinical condition of the patient improves, with stable blood pressure and good urine output, fluids should be changed to 1/2 normal saline at 250-500cc an hour for 3-4 hours. Ongoing reassessment is critical. When dehydration does not appear severe, rehydration rates one-half as fast as the above regimens have been studied with good results and less electrolyte disturbance. This may be considered in those patients who appear only minimally dehydrated.

Insulin: Insulin has several actions in managing DKA. These include decreasing glucagon release from the pancreas and limiting glucagon's effect on the liver. This decreases gluconeogenesis and ketogenesis in the liver. Additionally, the insulin allows glucose uptake and utilization by peripheral tissues.

Current recommendations for insulin therapy include an initial intravenous insulin bolus of 0.1 to 0.4 U/kg body weight followed by a continuous intravenous infusion of 0.1 U/kg/hour. This usually amounts to 5-10 U/hour in the typical adult. The goal of treatment should be to lower the serum glucose of the patient by 75-100 mg/dl/hour. The rate can be doubled every hour if this rate is not achieved. Ongoing severe difficulty in controlling the glucose levels may indicate the presence of a severe underlying infection. The ketosis and acidemia in DKA take longer to resolve than the elevation of glucose. For this reason, the insulin therapy must be continued even when the blood glucose levels have improved to near normal levels. When the glucose levels begin to approach 250 mg/dl, insulin infusions are continued, but the fluid composition is changed to include 5-10% dextrose in water to avoid hypoglycemia. Potassium: Regardless of the serum potassium level at the initiation of therapy, during treatment of DKA there is usually a rapid decline in the potassium concentration in the patient with normal kidney function.

General recommendations for potassium replacement are as follows. If the patient does not have marked elevation of potassium, is not in renal failure, the ECG does not show evidence of hyperkalemia beyond peaked T-waves, potassium therapy is initiated once good urine output has been established. Potassium is usually added to the intravenous fluids and should not exceed 40 mEq per liter of intravenous fluids. Some authors recommend spitting the potassium replacement as KCL and KPO4. The potassium level should be checked every one to two hours initially since this is when the greatest shift occurs. After the patient has stabilized the potassium can be checked every 6 to 8 hours.

Bicarbonate Therapy: The use of bicarbonate in the treatment of DKA is highly controversial. Current recommendations for bicarbonate therapy are as follows. Use of bicarbonate is considered unnecessary when the blood pH is greater than 7.1. For those patients with pH between 6.9 and 7.1 there are no clear guidelines. If the patient is elderly or very debilitated there may be some benefit to the bicarbonate in this range. If it is given it should be given with the intravenous fluids and not as IV push. For those patients with pH below 6.9 bicarbonate should be added to the intravenous fluids. One ampule of bicarbonate has 44 mEq of sodium bicarbonate. Attempts should be made to create an isotonic fluid with the bicarbonate being added to either one-half normal saline or D5W.

Phosphate: Phosphate is normally an intracellular substance that is dragged out of the cell during DKA. Similarly to potassium, at presentation the serum level may be normal, high, or low while the total body supply is depleted. Despite this depletion, replacement of phosphate has not been shown to affect patient outcome and routine replacement is not recommended.


In most instances, it may be necessary to start treatment with a broad spectrum antibiotic without waiting for specific proof of the presence of an infection and a culture and sensitivity test.

Frequency of monitoring in DKA
  • In severely ill patients, electrolytes including potassium, pH and serum creatinine should be monitored hourly for the first 4 hours then at 4 hourly intervals, over the next 12 hours.

  • Vital signs (eg; pulse, temperature, respiration, blood pressure and mental status) should be monitored with similar frequency.

  • Once the glucose is < 15 mmol/L, the pH is near normal and the patient is eating and drinking well, the frequency of laboratory blood tests can be reduced further. At this stage, start checking the urine for ketones to ascertain whether these are clearing.

Click to view Flow Sheet for Monitoring Treatment


Brain Edema: Clinical brain edema occurs in less than one percent of the pediatric population and even less frequently in adults. When it does occur the mortality rate is high It is probably prudent to prevent overvigorous correction of severe hyperosmolarity and hypernatremia.

When this complication does develop it typically has a rapid onset of severe headache and depression of the mental status. CT scan will show characteristic changes. Treatment must be started rapidly with intravenous mannitol and intubation as indicated. Adult Respiratory Distress Syndrome: This complication usually occurs during therapy with fluids, insulin, and electrolyte replacement.

Hyperchloremic Acidosis: This complication can be recognized by a low bicarbonate level, low to normal pH, normal anion gap, and an increased serum chloride level. The cause of this condition is multifactorial. It may be minimized by switching to hypotonic fluids during therapy and by using smaller amounts of chloride during therapy (KPhos rather than KCl).

Hypokalemia: As the patient is being treated for DKA, the volume expansion, and insulin therapy can rapidly lower potassium. To avoid sudden decompensation due to severe hypokalemia, it is prudent to recheck a serum potassium, following each liter of fluid. If large doses of insulin are required to control the patient,s blood glucose, the potassium level will need to be checked more frequently.

Hypoglycemia: As discussed previously, during DKA therapy, the serum glucose typically normalizes before the ketotic state has been corrected. To reverse this state it is necessary to continue insulin therapy after the glucose levels have improved. Without close monitoring, this can result in life-threatening hypoglycemia. To help avoid this, glucose measurements should be done frequently, and as the glucose level nears 250 mg/dl, the insulin infusion rate should be slowed, and glucose infusion with D5W should be started.

Back to Consensus