Management of diabetic ketoacidosis

Management of diabetic ketoacidosis

Martin S. Lipsky

Diabetic ketoacidosis is a common acute complication of insulin-dependent, or type 1, diabetes mellitus (IDDM). Before the discovery of insulin in the 1920s, patients rarely survived diabetic ketoacidosis. This complication is still potentially lethal, with an average mortality rate between 5 and 10 percent. Therefore, the ability to manage diabetic ketoacidosis is important for any physician who cares for patients with diabetes. This article discusses the diagnosis, evaluation, pathophysiology and treatment of diabetic ketoacidosis in adults.

Diagnosis and Evaluation

The annual incidence of diabetic ketoacidosis in the United States is 14 cases per 100,000 population.[3] Although most episodes of diabetic ketoacidosis occur in patients with previously diagnosed diabetes, about 20 to 30 percent represent new cases of IDDM.[1] Although the risk of diabetic ketoacidosis is greatest for patients with IDDM, the condition may also occur in patients with non-insulin-dependent diabetes (NIDDM) under stressful conditions, such as during a myocardial infarction. During such situations, insulin may be required to control ketosis.


The signs and symptoms of diabetic ketoacidosis vary, ranging from mild nausea to frank coma. Common symptoms are thirst, polyuria, nausea and weakness that have progressed over several days. It can also present fulminantly, with symptoms progressing to coma over the course of several hours. Because of the variable symptoms, diabetic ketoacidosis should be considered in any ill diabetic patient, particularly if the patient presents with nausea and vomiting.

Common clinical findings include tachycardia, tachypnea, dehydration, altered mental status and a fruity breath odor, indicating the presence of ketones. Kussmaul respirations are usually present in patients with severe acidosis, although they may disappear if the acidosis is protracted.

About 30 percent of patients with diabetic ketoacidosis have abdominal pain that is severe enough to mimic an acute abdomen.[4] The serum amylase level is often elevated, further obscuring the diagnosis. Since elevated amylase in diabetic ketoacidosis originates from the parotid gland, pancreatitis may be ruled out by a normal serum lipase level. Another important distinguishing feature is that the abdominal pain in patients with diabetic ketoacidosis improves rapidly with treatment.


Defined as the triad of hyperglycemia, ketosis and acidosis, diabetic ketoacidosis is generally confirmed by a blood glucose level greater than 250 mg per dL (13.9 mmol per L), a pH less than 73, a bicarbonate level less than 15 mEq per L (15 mmol per L) and serum ketones present in a dilution greater than 1:2.[5]

Spot urinalysis for ketones and glucose is a sensitive screen for diabetic ketoacidosis. In the presence of glycosuria and ketonuria, assessment of the blood glucose level, arterial blood gases or venous pH, serum ketones and serum electrolytes is warranted to confirm the diagnosis. A complete blood count, chest radiograph, electrocardiogram and determinations of blood urea nitrogen (BUN) and creatinine are also recommended to detect the precipitating event and to screen for metabolic abnormalities associated with diabetic ketoacidosis.

Other laboratory abnormalities include hypophosphatemia, hypomagnesemia, and hemoconcentration with elevated BUN and creatinine levels. Serum sodium levels may be low, high or normal. When evaluating the serum sodium level, it is helpful to remember that hyperglycemia causes a shift of free water into the extracellular space, diluting the measured sodium concentration. The true sodium level can be roughly estimated by adding 1.5 mEq per L (1.5 mmol per L) for each 100 mg of glucose over 200 mg per dL (11.1 mmol per L).


Numerous factors cause or contribute to the development of diabetic ketoacidosis (Table 1). The most common precipitating factor is infection, which occurs in 25 to 50 percent of patients.[6] Often the infection may appear trivial, such as mild cellulitis or upper respiratory tract infection. Although the presence of fever suggests infection, its absence does not exclude this possibility.[7] White blood cell counts of 20,000 to 30,000 per [mm.sup.3] (20.0 to 30.0 x 109 per L) can occur in patients with diabetic ketoacidosis without infection, but high counts with a left shift or toxic granulations suggest an infection. A chest radiograph is useful for detecting early pneumonia, and urinalysis is necessary to screen for pyuria and ketones. Bacterial cultures should be taken as necessary


Precipitating Causes

of Diabetic Ketoacidosis

Alcohol abuse


Inadequate insulin under stressful conditions



Beta blockers

Calcium channel blockers

Pentamidine (NebuPent, Pentam)



Myocardial ischemia

Missed insulin doses


Renal failure




An electrocardiogram is useful to screen for underlying cardiac ischemia and arrhythmias, and it provides immediate information about serum potassium levels. Peaked T waves suggest elevated potassium, while prolongation of the QT interval with low-amplitude T waves is a sign of hypokalemia.

Omission of insulin, either because of noncompliance or because patients mistakenly believe that insulin is not required on “sick days” when they are not eating well, is another important cause of diabetic ketoacidosis. However, despite a meticulous search for an underlying precipitating event, none can be found in 15 to 30 percent of patients with diabetic ketoacidosis.[1]


The initiating metabolic defect in diabetic ketoacidosis is an insufficient or absent level of circulating insulin. Inadequate insulin creates a biologic “alarm state,” which triggers the excess secretion of counter-regulatory hormones, particularly glucagon. The abnormal insulin-to-glucagon ratio, coupled with excess circulating catecholamine, cortisol and growth hormone levels, initiates a host of complex metabolic reactions, leading to ketonemia, hyperglycemia and acidosis.[2]

One metabolic effect is the release of stored triglycerides from adipose tissue. Also, the abnormal insulin-to-glucagon ratio affects the metabolic fate of triglycerides; instead of being metabolized to carbon dioxide or stored, the incoming fats are converted by the liver to ketone bodies, mainly beta-hydroxybutyric acid and, to a lesser extent, acetoacetic acid.[4] Although acetoacetic acid plays no pathologic role, it causes the characteristic fruity breath odor of diabetic ketoacidosis and is clinically useful because it reacts with nitroprusside (Acetest), allowing the detection of ketone bodies.

As diabetic ketoacidosis resolves with treatment, the beta-hydroxybutyric acid level decreases, but the level of acetoacetic acetate rises transiently. Since nitroprusside reacts only with acetoacetate, ketones may paradoxically increase despite clinical improvement, but they may eventually disappear with successful treatment of diabetic ketoacidosis. Therefore, the nitroprusside test is not a reliable indicator to use to follow therapy.[8]

Insulin also facilitates glucose transport into cells. The combination of insulin deficiency and the excess of counterregulatory hormones stimulates glucose production and this coupled with the impaired peripheral utilization of glucose, leads to hyperglycemia.

Hyperglycemia induces a number of metabolic abnormalities. As the glucose level rises above 180 mg per dL (10.0 mmol per L), the renal reabsorption threshold is exceeded and an osmotic diuresis ensues, leading to a loss of water and electrolytes, principally sodium, potassium, phosphorus and magnesium. Volume depletion causes hemoconcentration, thereby decreasing glomerular filtration of glucose, and creates a cycle of progressive hyperglycemia.

Hyperglycemia also increases the serum osmolality. To compensate, water shifts from the cells into the extracellular space, initially preserving an adequate circulating blood volume at the expense of intracellular water. However, this mechanism eventually fails and, if left untreated, diabetic ketoacidosis results in severe volume depletion and acidosis, leading to cardiovascular collapse and death.



Once the diagnosis of diabetic ketoacidosis is established, fluid and electrolyte therapy should begin immediately. Since the estimated fluid deficit in patients with diabetic ketoacidosis is between 3 and 5 L,[2] the importance of adequate fluid replacement cannot be overemphasized. Generally, replacement is achieved by administering 1 L of normal saline over the first hour, followed by 1 L of fluid over the next two hours.

The choice of intravenous fluids is controversial. Most authorities recommend normal saline if the patient has hypotension or tachycardia, or if the corrected serum sodium level is depressed. If clinical evidence indicates adequate perfusion (e.g., normal pulse, blood pressure and good skin turgor) and the corrected sodium level is elevated to a level greater than 150 mEq per L (150 mmol per L), one-half normal saline is preferable. After the initial fluid bolus, the patient’s clinical response and volume status determines the optimal intravenous fluid rate. Generally, rates between 150 to 250 mL per hour are acceptable, although the patient’s age, hemodynamic status, underlying medical condition and the precipitating event need to be considered.

Despite the marked fluid deficit associated with diabetic ketoacidosis, many physicians do not administer adequate fluids, particularly to older patients, for fear of precipitating congestive heart failure or cerebral edema. However, one study[9] found no evidence that rapid initial fluid replacement in elderly patients with diabetic ketoacidosis precipitates congestive heart failure. Another study[10] found no correlation between cerebral edema and an initially rapid rate of rehydration.


Although several protocols for insulin therapy exist, most authorities recommend continuous intravenous infusion of low-dose regular human insulin (Humulin).[11] An intravenous loading dose of 0.1 to 0.2 U per kg of insulin is followed by a continuous infusion of 0.1 U per kg per hour (5 to 10 U per hour). This suppresses ketone body formation and lowers the blood glucose sufficiently without the risk of hypoglycemia and hypokalemia associated with higher-dose regimens.

After administration of an insulin drip for the first hour, the blood glucose level should be determined. A decrease of 50 to 100 mg per dL (2.8 to 5.6 mmol per L) per hour confirms an adequate insulin rate. If the blood glucose level does not change, the insulin rate should then be doubled.[12] Subsequent incremental adjustments in the dosage of insulin are based on the glucose levels. Large insulin requirements (over 20 U per hour) suggest either inadequate volume replacement, infection or insulin resistance.

One important facet of insulin therapy is that the hyperglycemia invariably improves faster than the acidosis. Even if the blood glucose level has dropped, intravenous insulin should be continued until the ketosis has resolved. The addition of 5 percent or 10 percent dextrose to the intravenous fluid when the blood glucose reaches the 200 to 300 mg per dL (11.1 to 16.7 mmol per L) range prevents hypoglycemia as sufficient insulin is continued to correct the acidosis. Gradually, as the acidosis is corrected, the rate of insulin delivery may be lowered to a maintenance level of 1 to 2 U per hour.

When the patient is able to eat and the glucose level is below 250 mg per dL (13.9 mmol per L), serum ketones are negative and the serum bicarbonate level is greater than 21 mEq per L (21 mmol per L), switching from intravenous insulin to subcutaneous injections is appropriate. Patients with diabetes can resume their usual dose of insulin. A reasonable insulin regimen for a patient with newly diagnosed diabetes is 15 to 20 U of long-acting insulin in the morning and 5 to 15 U in the evening, supplemented by regular insulin for elevated blood glucose levels. Because the half-life of intravenous insulin is only five minutes, in order to allow sufficient time for the injected insulin to be absorbed, the insulin drip should be continued for at least one to two hours. In the event that the ketoacidosis has resolved but the patient cannot eat, intravenous dextrose should be continued and’ regular insulin given subcutaneously every four to six hours as needed, based on a sliding scale, until the patient is able to eat.


Because diabetic ketoacidosis causes major potassium deficits, replacement therapy is always necessary, although the timing and rate of therapy may vary.[6] After an adequate urine output is established, and if the serum potassium level is normal or below normal, potassium replacement therapy should be initiated. A rate of 10 to 30 mEq per hour is usually sufficient and can be achieved by adding 20 to 40 mEq of potassium per L of intravenous solution.

Lower delivery rates should be used for patients with a potassium level greater than 4.5 mEq per L (4.5 mmol per L), and higher rates should be used for patients with a potassium level less than 4.0 mEq per L (4.0 mmol per L). Very large doses may be needed for treating patients who have severe acidosis or a low initial potassium level. For patients with a potassium level elevated to a level higher than 5.5 mEq per L (5.5 mmol per L), replacement should begin once the levels fall within the normal range and the patient is producing urine. Monitoring potassium levels is critical, particularly early in treatment when the steepest declines occur.

Patients with diabetic ketoacidosis are also phosphate-deficient.[2] Theoretically, phosphate is necessary for the synthesis of 2,3-diphosphoglycerate, which promotes tissue oxygenation by shifting the oxyhemoglobin curve to the right. Hypophosphatemia is associated with rhabdomyolysis, weakness, diminished cardiac function, hemolysis and respiratory failure.[8]

Despite these theoretic considerations, no definitive clinical evidence is available to suggest that phosphate administration improves outcome. Evidence does show that phosphate therapy can precipitate hypokalemia and hypomagnesemia. Still, most authorities recommend phosphate replacement if levels are less than 1.5 mg per dL (0.5 mmol per L).[13,14] Administration of potassium phosphate, 20 mEq per L, at a rate between 10 and 20 mEq per hour, is the best method of phosphate replacement. Use of this salt replaces both potassium and phosphate simultaneously.


Bicarbonate administration is rarely indicated. Although acidosis affects both the central nervous system and the cardiovascular system, no convincing studies show a significant reduction in mortality or morbidity with the administration of bicarbonate.[1,15] In addition, potential hazards of bicarbonate therapy include hypokalemia, aggravation of tissue hypoxia by shifting the oxyhemoglobin curve to the left, and an increased risk of cerebral edema. Although these risks limit the usefulness of bicarbonate therapy, many authorities still recommend it if the pH is less than 7.0, particularly in the presence of hypotension and arrhythmia.[16] Bicarbonate is administered at a rate of 40 to 80 mEq per hour; once the pH rises above 71, the bicarbonate should be discontinued.


Ancillary measures help to assure a positive outcome for patients with diabetic ketoacidosis, and careful monitoring is probably the most important of these. Table 2 lists some important parameters to follow that allow the clinician to spot trends and adjust therapy appropriately.


Since infection is a common precipitating cause, antibiotic treatment is indicated for patients with an infection. Many patients with diabetic ketoacidosis have gastric stasis, and in obtunded patients, nasogastric suction is useful to avoid aspiration. Although bladder catheterization is initially helpful in monitoring urine output in severely ill patients or patients who have other complicating factors, the catheter should be removed as soon as possible to avoid urinary tract infection. Hypoxic patients require supplemental oxygen. For patients who have persistent hypotension after the first 2 to 3 L of fluid have been administered, pressor agents may be necessary to maintain perfusion; hemodynamic monitoring may also be needed to direct further fluid therapy

The author thanks Jack Brunner, Jr., M.D., a certified endocrinologist at the Mercy Hospital Endocrine and Diabetic Care Unit in Toledo, Ohio, for reviewing the manuscript.


[1.] Sanson TH, Levine SN. Management of diabetic ketoacidosis. Drugs 1989;38:289-300. [2.] Foster DW, McGarry JD. The metabolic derangements and treatment of diabetic ketoacidosis. N Engl J Med 1983;309:159-69. [3.] Faich GA, Fishbein HA, Ellis SE. The epidemiology of diabetic acidosis: a population-based study. Am J Epidemiol 1983;117:551-8. [4.] Siperstein MD. Diabetic ketoacidosis and hyperosmolar coma. Endocrinol Metab Clin North Am 1992;21:415-32. [5.] Rumbak MJ, Kitabchi AE. Diabetic ketoacidosis: etiology, pathophysiology and treatment. Compr Ther 1991;17:46-9. [6.] Walker M, Marshall SM, Alberti KG. Clinical aspects of diabetic ketoacidosis. Diabetes Metab Rev 1989;5:651-63. [7.] Gale EA, Tattersall RB. Hypothermia: a complication of diabetic ketoacidosis. Br Med J 1978;2:1387-9. [8.] Israel RS. Diabetic ketoacidosis. Emerg Med Clin North Am 1989;7:859-71. [9.] Carroll P, Matz R. Uncontrolled diabetes mellitus in adults: experience in treating diabetic ketoacidosis and hyperosmolar nonketotic coma with low-dose insulin and a uniform treatment regimen. Diabetes Care 1983;6:579-85. [10.] Rosenbloom AL. Intracerebral crises during treatment of diabetic ketoacidosis. Diabetes Care 1990;13:99-33. [11.] Blackett PR, Lera T Jr, Garnica A, Schaefer GB, Domek D, Parker M. Diabetic ketoacidosis at the Children’s Hospital of Oklahoma: a review on presentation and management. J Okla State Med Assoc 1990;83:594-601. [12.] Berger W, Keller U. Treatment of diabetic ketoacidosis and non-ketotic hyperosmolar diabetic coma. Baillieres Clin Endocrinol. Metab 1992;6:1-22. [13.] Fisher JN, Kitabchi AE. A randomized study of phosphate therapy in the treatment of diabetic ketoacidosis. J Clin Endocrinol Metab 1983;57:177-80. [14.] Wilson HK, Keuer SP, Lea AS, Boyd AE III, Eknoyan G. Phosphate therapy in diabetic ketoacidosis. Arch Intern Med 1982;142:517-20. [15.] Riley LJ Jr, Cooper M, Narins RG. Alkali therapy of diabetic ketoacidosis: biochemical, physiologic, and clinical perspectives. Diabetes Metab Rev 1989; 5:627-36. [16.] Morris LR, Murphy MB, Kitabchi AE. Bicarbonate therapy in severe diabetic ketoacidosis. Ann Intern Med 1986;105:83640.

NURTIN S. LIPSKY, M.D. is the director of the RSM Medical Foundation in Toledo, Ohio. He also teaches at the Mercy Family Practice Center and is an associate professor at the Medical College of Ohio in Toledo. Dr. Lipsky graduated from the Medical College of Pennsylvania, in Philadelphia, and completed a residency in family medicine at the University of California, Irvine, Medical Center.

COPYRIGHT 1994 American Academy of Family Physicians

COPYRIGHT 2004 Gale Group