Diabetic ketoacidosis occurs exclusively in the diabetic population. It is characterized by hyperglycemia and ketonemia. A relative deficiency of insulin and a concurrent excess of stress hormones are responsible for the metabolic derangement. Therapy includes the replacement of fluids and insulin using low doses administered with various techniques.

PATHOPHYSIOLOGY

The major metabolic abnormalities that occur during diabetic ketoacidosis are hyperglycemia and ketonemia. The metabolic derangements can be explained by relative insulin insufficiency and counterregulatory hormone excess. Insulin is the prime anabolic hormone and is responsible for the metabolism and storage of carbohydrates, fats, and proteins. The counterregulatory hormones are glucagon, catecholamines, cortisol, and growth hormone.

Insulin

Ingested glucose is the primary stimulant of insulin release from the $ cells of the pancreas. Insulin acts on the liver to facilitate the uptake of glucose and its conversion to glycogen. Insulin inhibits glycogen breakdown (glycogenolysis) and suppresses gluconeogenesis. The net effect of these actions is to promote the storage of glucose in the form of glycogen.
Insulin's effect on lipid metabolism is to increase lipogenesis in the liver and adipose cells and simultaneously to prevent lipolysis. Insulin promotes the production of triglycerides from free fatty acids and facilitates the storage of fat. The breakdown of triglycerides to free fatty acids and glycerol is inhibited by insulin. The overall result is the conversion of glucose to stored energy as triglycerides.
Insulin's action in protein metabolism is to stimulate the uptake of amino acids into muscle cells and to mediate the incorporation of amino acids into muscle protein. It prevents the release of amino acids from muscle protein and from hepatic protein sources.
Deficiency in the insulin-secretory mechanism of the $ cells of the pancreas is the predominant lesion in diabetes mellitus. This defect results in insulin lack that may be partial or total.
Absolute insulin lack is rare but may be found in insulin-dependent diabetes (IDDM) patients. In the typical non-insulin dependent diabetes (NIDDM) patient, secretory failure involves primarily the initial rapid-release phase of insulin secretion. Minimal insulin inadequacy causes a decrease in the storage of body fuels, and $-cell failure is evident only by the abnormal response to a glucose load abnormal glucose tolerance test. With more severe failure of insulin secretion, not only is fuel storage impaired, but fuel stores are mobilized during fasting, resulting in hyperglycemia. The increase in the blood glucose level is due to increased glycogenolysis and may elicit an increase in insulin secretion if $-cell reserve is present. The glucose metabolism and concentration may then return to normal.
When there is an absolute or relative failure in insulin secretion, hyperglycemia does not produce increased insulin activity. Loss of the normal physiologic effects of insulin results in catabolism, and hyperglycemia and ketonemia occur.

Counterregulatory Hormones

During insulin insufficiency, glucose transport into the cells is inhibited. The physiologic response to cellular starvation and other stresses is to increase the hormones glucagon, catecholamines, cortisol, and growth hormone. These hormones are grouped as counterregulatory hormones because of their anti-insulin effects. The relative roles of each and their mechanisms of action in diabetic ketoacidosis have not been completely elucidated. However, glucagon in excess has been implicated as the main hormone contributing to hyperglycemia and ketonemia. Excess counterregulatory hormone secretion, in conjunction with relative insulin deficiency, is essential to the development of diabetic ketoacidosis. This is shown by the failure of diabetic animals to develop ketoacidosis in the absence of counterregulatory hormones, the elevation of at least one hormone in every case of diabetic ketoacidosis, a delay or reduction of ketoacidosis with pharmacologic blockade of individual stress hormones, and an increase in ketogenic activity when each of the counterregulatory hormones is infused in high physiologic concentrations. No correlation has been reported between the plasma level of insulin during diabetic ketoacidosis and the severity of the ketoacidosis. Finally, the association between antecedent stress and diabetic ketoacidosis has long been recognized. Secretion of the counterregulatory hormones characterizes all major forms of stress.
The counterregulatory hormones are catabolic and, in general, reverse the physiologic processes promoted by insulin. They affect carbohydrate metabolism by increasing glycogenolysis and gluconeogenesis, thereby raising the blood glucose level. Lipolysis is stimulated by glucagon and catecholamines, and this results in increased free fatty acids for conversion to ketones. Protein breakdown is accelerated and provides amino acids for gluconeogenesis. The net effect of relative insulin insufficiency and excess counterregulatory hormones is hyperglycemia and ketonemia .
Hyperglycemia occurs earlier than ketonemia during diabetic ketoacidosis. Glucose is underutilized because of insulin lack and overproduced because of enhanced glycogenolysis and gluconeogenesis. Gluconeogenesis is faciliated by increased levels of glucogenic precursors such as glycerol and amino acids resulting from unopposed lipolysis and proteolysis.
Ketonemia occurs because of increased lipolysis and ketogenesis. Insulin deficiency and excess stress hormones lead to the breakdown of triglycerides and the release of large amounts of fatty acids into the circulation. These fatty acids are assimilated in the liver, where they are converted to ketone bodies. The peripheral utilization of ketones is decreased during insulin insufficiency, and they accumulate in the usual 3:1 ratio ($-hydroxybutyrate to acetoacetate).
Factors known to precipitate diabetic ketoacidosis include omission of daily insulin injections and a variety of stressful events, such as infections, stroke, myocardial infarction, trauma, pregnancy, hyperthyroidism, and pancreatitis. However, in some patients, there is no clear precipitating cause.

CLINICAL PRESENTATION

Most of the clinical manifestations of diabetic ketoacidosis are related to the biochemical derangements. Hyperglycemia causes an increased osmotic load, and intracellular water is lost because cellular membranes are not freely permeable to glucose. Osmotic diuresis produces total body fluid depletion. Dehydration, hypotension, and reflex tachycardia are consequences. Osmotic diuresis also causes loss of sodium, chloride, potassium, phosphorus, calcium, and magnesium. The serum sodium level may be further decreased by a dilution effect of hyperglycemia. The dilutional effect is a serum sodium decrease of about 5 mEq/L for every 180 mg/dL increase of blood glucose. Electrolyte loss may be worsened by repeated bouts of vomiting.
Dissociation of hydrogen ions from circulating ketone bodies is responsible for the development of acidosis and the fall in the serum bicarbonate level. Some of the ketone bodies are oxidized to acetone, a neutral, soluble, volatile substance that causes the characteristic fruity odor on the breath of a patient with ketoacidosis. Hepatomegaly due to accumulation of fat within the liver may occur and should resolve with reversal of ketogenesis.
Acidosis produces other clinical consequences. Compensatory hyperventilation is commonly seen. Exchange of H+ ions for K+ across the intracellular membrane is partially responsible for the elevated serum potassium level seen in patients in diabetic ketoacidosis. Peripheral vasodilatation and vascular collapse can result from acidosis.
There is no apparent correlation between the state of consciousness of the patient and the degree of ketonemia, hyperglycemia, electrolyte imbalance, or acidosis. The most direct correlation is with serum osmolality. Some degree of mental confusion or coma is more likely with serum osmolality levels above 340 mOsm/kg. In addition, if the serum osmolality is <340 mOsm/kg in a patient with diabetic ketoacidosis, some other cause of coma should be sought.
Nausea, vomiting, and abdominal pain are common presenting complaints. The cause of these disturbances is not clear. Gastric dilatation, paralytic ileus, and abdominal tenderness may be present. Although pancreatitis may develop as a result of ketoacidosis, the diagnosis is difficult since the serum amylase level and the amylase clearance can be elevated in both conditions. Of course, the reverse can occur, so that acute inflammatory or hemorrhagic pancreatitis can result in ketoacidosis. In general, abdominal signs and symptoms should disappear with the resolution of ketoacidosis, and carefully repeated evaluation is necessary to rule out a serious intraabdominal disorder.
Inappropriate normothermia can occur, so that infection must be presumed even in the absence of fever.
Diabetic ketoacidosis may develop rapidly or over a few days. If the patient is able to maintain an adequate fluid and salt intake, a state of compensated ketosis may develop. During the early stages of ketoacidosis or when nausea and vomiting occur, the patient may decrease or omit insulin, thus hastening full-blown diabetic ketoacidosis. The typical patient has nausea, vomiting, abdominal pain, weight loss, dehydration, hypotension, tachycardia, hyperventilation or Kussmaul's respirations, and the odor of acetone on the breath.

LABORATORY

Laboratory abnormalities that are always present during diabetic ketoacidosis include elevated levels of blood glucose, $-hydroxybutyrate, and acetoacetate. Similarly, glucosuria and ketonuria are consistent findings. A decreased pH, low serum bicarbonate level, and decreased PCO2 are present because of metabolic acidosis with respiratory compensation.
The serum sodium level is variable. More water is lost than solutes, and even if the serum sodium level is low, the patient is hypertonic. Initially, the serum potassium level is usually elevated or normal, but it falls as the acidosis is corrected and potassium shifts intracellularly. The serum chloride level may be low if excessive vomiting has occurred. An increased anion-gap acidosis is present because of ketonemia.
The diagnosis of diabetic ketoacidosis should be suspected based upon the clinical presentation previously described. Confirmatory laboratory findings include a blood glucose level greater than 300 mg/dL, a bicarbonate level less than 15 mEq/L, a serum acetone level greater than 2:1 dilution, and a pH less than 7.3. Venous blood should be drawn for a complete blood cell count and determinations of serum glucose, electrolytes, blood urea nitrogen (BUN), creatinine, phosphorus, calcium, magnesium, and acetone. Arterial blood gases are essential. Urinalysis and a chest roentgenogram to search for infection, and an ECG to identify acute myocardial infarction and hyperkalemia, are necessary.

Differential Diagnosis

The differential diagnosis of metabolic coma in a diabetic patient includes hypoglycemia, nonketotic hyperosmolar coma, alcoholic ketoacidosis, and lactic acidosis. A rapid differentiation can be made in the emergency department . The result of the analysis of blood gases should be available in a few minutes, and acidosis, if present, will be confirmed. While the physician is awaiting other laboratory results, a drop of blood can be tested for blood glucose and serum ketones.
Reagent strips that measure blood glucose levels can be interpreted visually, or can be read in a reflectance meter. Visual interpretation identifies a range but not a precise number, and this method reliably distinguishes between hyperglycemic and hypoglycemic levels. Semiquantitative estimation of serum ketones can be made by testing a drop of blood with a nitroprusside-impregnated tablet. The nitroprusside reaction measures acetoacetate but not $-hydroxybutyrate. This test can be misleading if most of the serum ketones are in the form of $-hydroxybutyrate. Additionally, lactic acidosis may occur simultaneously with diabetic ketoacidosis. Measurement of serum lactate levels may be indicated to determine the contribution of lactic acid to the metabolic acidosis. Finally, in mixed acid-base disturbance, the pH may not accurately reflect the degree of acidosis. The anion gap can assist in identifying unmeasured acids.

TREATMENT

Once the diagnosis of diabetic ketoacidosis has been established, therapy must be started immediately. Specific therapeutic goals include rehydration, correction of electrolyte and acid-base imbalance, reversal of the metabolic consequences of insulin insufficiency, treatment of precipitating causes, and avoidance of complications.
A variety of therapeutic approaches are advocated. Regardless of the approach used, frequent monitoring of the effects is essential. The levels of blood glucose, the anion gap, potassium, and carbon dioxide should be determined every 1 to 2 h until recovery is well-established. A flow sheet to record vital signs, level of consciousness, intake and output, therapeutic measures, and blood chemistry determinations is recommended. Complete clearing of hyperglycemia and ketonemia usually requires 8 to 16 h.

Fluid Administration

Rapid fluid administration is the most important initial step in the treatment of diabetic ketoacidosis. The average patient in diabetic ketoacidosis has a water deficit of 5 to 10 L and a sodium deficit of 450 to 500 mEq. Normal saline is the most frequently recommended fluid for initial rehydration even though the extracellular fluid of the patient is hypertonic. The normal saline does not provide free water to correct intracellular dehydration, but it does prevent a too-rapid fall in extracellular osmolality and excessive transfer of water into the central nervous system (CNS). Most authors favor alternating the administration of normal saline with the administration of half-normal saline.
The first liter of fluid should be administered rapidly, usually over 1/2 to 1 h. During the first 3 to 4 h, 3 to 5 L of fluid may be required. The blood glucose level and ketone body concentration fall after fluid administration and before implementation of any other therapeutic modality. With rehydration, tissue perfusion is restored, improving the effectiveness of insulin, and renal blood flow increases, allowing the excretion of ketone bodies.
The fluid should be changed to a hypotonic solution after the initial replacement of intravascular volume or if the serum sodium level is 155 mEq/L. Central venous pressure or pulmonary artery wedge pressure should be monitored during fluid replacement in the elderly patients or those with heart disease.

Bicarbonate

Sodium bicarbonate is given to correct the negative effects of acidosis. At a pH of 7, peripheral vasodilatation, decreased cardiac output, and hypotension can occur. Respiratory and CNS depression can occur with severe acidosis (pH less than 6.8).
The hazards of excessive alkali replacement can outweigh the potential benefits. These include paradoxical spinal fluid acidosis, hypokalemia, impaired oxyhemoglobin dissociation, rebound alkalosis, and sodium overload.
It has been established that cerebrospinal fluid (CSF) acidosis is deleterious to brain function. Systemic acidosis per se does not cause mental aberration as long as the CSF is protected against large pH changes. When sodium bicarbonate is administered in large doses, the carbon dioxide loss is diminished, and extracellular fluid and levels of carbon dioxide and bicarbonate increase. Carbon dioxide diffuses freely across the blood-brain barrier, but bicarbonate diffuses into the CSF much more slowly. The difference in the rates of movement into the spinal fluid results in an increase in CSF carbonic acid, a fall in CSF pH, and paradoxical spinal fluid acidosis.
Alkali administration causes a shift of potassium intracellularly. In a patient who already has total body potassium depletion, severe hypokalemia could result. During acidosis, the oxyhemoglobin dissociation curve shifts to the right, facilitating the off-loading of oxygen at the tissue level. This beneficial effect of acidosis could be lost with sudden restoration of the pH toward normal. Final complications of excessive sodium bicarbonate administration include overcompensated rebound alkalosis and sodium overload.
Current recommendations are to administer sodium bicarbonate in modest amounts, i.e., 44 to 100 mEq, when the pH is less than 6.9 for adults. Some studies have shown that use of bicarbonate therapy in patients with diabetic ketoacidosis has provided no beneficial effects in terms of clinical recovery or biochemical variables, even with a pH as low as 6.9. Hydrogen ion production ceases when ketogenesis stops; excessive hydrogen ions are eliminated through the urine and through the respiratory tract, and ketone body metabolism results in the endogenous production of alkali.

Potassium

The deficiency of total body potassium is created by insulin deficit, acidosis, diuresis, and frequent vomiting. The potassium deficit is about 3 to 5 mEq/kg. The initial serum potassium level is usually normal or high because of a deficit of body fluid, diminished renal function, and intracellular exchange of potassium for hydrogen ions during acidosis. Initial hypokalemia indicates severe total body potassium depletion, and massive amounts of potassium for replacement are required during the next 24 h.
The goals of potassium replacement are to maintain a normal extracellular potassium concentration during the acute phases of therapy and to replace the intracellular deficit over a period of days or weeks. With initiation of therapy for diabetic ketoacidosis, the serum potassium concentration falls. This is due to dilution of extracellular fluid, correction of acidosis, increased urinary loss of potassium, and the action of insulin in promoting reentry of potassium into the cells. If these changes occur too rapidly, precipitous hypokalemia may result in fatal cardiac arrhythmias, respiratory paralysis, and paralytic ileus. These complications are avoidable if the pathophysiology is understood and the effects of therapy are frequently monitored.
The ability of insulin to drive potassium into the cells is directly proportional to the insulin concentration. Low-dose insulin therapy provides greater stabilization of the extracellular potassium concentration during the early stages of therapy.
Early potassium replacement is now a standard modality of care. Some authors recommend that small doses of potassium (20 mEq) be added to the intravenous fluid given initially. Others favor administering potassium within the first 2 to 3 h, when insulin therapy is initiated, or after volume expansion has been accomplished. If oliguria is present, renal function must be evaluated and potassium replacement must be decreased. Potassium determinations every 1 to 2 h and continuous ECG monitoring for changes reflecting the potassium concentration should be employed. From 100 to 200 mEq of potassium during the first 12 to 24 h is usually required. Occasionally, as much as 500 mEq of potassium may be necessary.

Insulin

Familiarity with a particular insulin replacement regimen, continuous monitoring, and attention to detail are the most important factors in successfully treating a patient in diabetic ketoacidosis. The amount and route of insulin administration are of secondary importance.
Large doses of insulin are not required to reverse the metabolic derangements of diabetic ketoacidosis and hypoglycemia, osmotic disequilibrium, and hypokalemia are more frequent with large-dose insulin therapy.
Low-dose insulin techniques for treatment of diabetic ketoacidosis are simple, safe, and effective. Techniques for continuous intravenous infusion and intramuscular, subcutaneous, and intravenous bolus therapy have been developed. Blood insulin concentrations of 20 to 200 µ units/mL inhibit gluconeogenesis and lipogenesis, stimulate the uptake of potassium by peripheral tissues, and achieve maximum rates of fall of blood glucose concentrations. A continuous insulin infusion of 1 units/h raises the plasma insulin concentration by 20 µ units/mL. Similarly, 5 units/h produces a therapeutic level of 100 µ units/mL. This level is generally sufficient to achieve normal metabolic homeostasis. The half-life of insulin given intravenously is 4 to 5 min, with an effective biological half-life at the tissue level of approximately 20 to 30 min.
Hypoglycemia using low-dose insulin techniques is almost nonexistent as long as monitoring is done carefully. With low-dose insulin therapies and proper potassium replacement, the occurrence of hypokalemia is less than 5 percent. The more gradual, even insulin effect achieved by low-dose therapy avoids rapid osmotic fluid shifts and the development of cerebral edema.
All low-dose insulin techniques are effective in reversing the metabolic consequences of insulin insufficiency . In continuous intravenous infusion of low doses of insulin, 5 to 10 units of regular insulin are administered per hour. The effect of insulin begins almost immediately after the initiation of the infusion, and a priming intravenous bolus is not required. Continuous insulin administration ensures that a steady blood concentration is maintained in an effective range, and this technique allows flexibility in adjusting the insulin dose. When the infusion is stopped, the insulin already in the blood is quickly degraded, providing greater control of the amount of insulin given in comparison with the intramuscular or subcutaneous routes.
Serious complications with continuous intravenous low-dose insulin infusion are minimal. The main disadvantage is that it requires an infusion pump and frequent monitoring to ensure that insulin is being administered in the desired amount. A separate intravenous site for the insulin infusion is desirable but not required.
The technique of low-dose intramuscular or subcutaneous insulin therapy is better suited to a hospital environment in which constant nursing supervision is not always possible. The main disadvantage to this approach, more marked with the subcutaneous route, is that insulin absorption may be erratic in a hypotensive, peripherally vasoconstricted patient. Erratic absorption may result in a delay in achieving adequate insulin levels. Further, delayed absorption can produce deposits of insulin that may later be absorbed, causing hypoglycemia. These problems can largely be eliminated by ensuring adequate hydration of the patient and by using small enough doses of insulin to preclude accumulation of large insulin deposits.
The onset of action of insulin given intramuscularly is delayed in comparison with that of insulin given intravenously. The most current protocols recommend an initial dose of 20 units of insulin intramuscularly, intravenously, or divided between these routes, followed by 5 to 10 units/h intramuscularly. The half-life of insulin given intramuscularly is 2 h. Hourly injections produce a continuous, effective blood concentration of insulin.
There are common problems with insulin therapy regardless of the technique used. The incidence of nonresponders to low-dose therapies is 1 to 2 percent. Infection is the main reason for failure to respond to low-dose insulin therapy. If the patient fails to respond to low-dose insulin therapy in the first hour, most protocols recommend doubling the infusion rate or administering an intravenous bolus of insulin. The insulin dose is increased in a similar fashion each hour until a satisfactory response is achieved.
Glucose should be added to the intravenous fluid when the blood glucose concentration falls to 250 mg/dL. Insulin therapy should not be stopped just because the blood glucose level declines but should be continued until the ketonemia and acidosis have cleared. Intravenous insulin should not be abruptly discontinued. An overlap period in which subcutaneous insulin is given should precede discontinuation of the insulin infusion.

Phosphate Replacement

The role of phosphate replacement during the treatment of diabetic ketoacidosis remains controversial. Hyperphosphatemia is the initial finding in most cases of diabetic ketoacidosis. However, one author estimates that up to 90 percent of patients have acute hypophosphatemia within 6 to 12 h after initiation of therapy for diabetic ketoacidosis. The decrease is primarily due to a sudden shift of phosphate from the extracellular to the intracellular compartment following insulin administration and accelerated glucose storage. Phosphate is found in all body tissues, and this sudden shift deprives them of this essential constituent. Hypophosphatemia is usually most severe 24 to 48 h after the start of insulin therapy.
Phosphate plays an integral part in the conversion of energy from adenosine triphosphate (ATP) and in the delivery of oxygen at the tissue level through 2,3-diphosphoglyceric acid (2,3-DPG). In addition, many important enzymes, cofactors, and biochemical intermediates depend upon phosphate. Acute phosphate deficiency has been associated with a variety of clinical disorders including neuromuscular paralysis leading to respiratory failure and possibly myocardial dysfunction.
Acute hypophosphatemia can be corrected by intravenous or oral administration of phosphorus. Several oral forms are available but may cause diarrhea and be erratically absorbed. A commercial intravenous preparation (KH2PO4 plus K2HPO4) containing K+ at a concentration of 4 mEq/mL and phosphorus at a concentration of 96 mg/mL can be used. Five milliliters of this commercial potassium phosphate preparation added to 1 L of intravenous fluid provides approximately 20 mEq of K+ and 480 mg of PO42+.
Hypophosphatemia is not associated with untoward consequences until a serum concentration of less than 1.0 mg/dL is reached. Phosphorus supplementation is not indicated and should not be given as long as the level remains above this concentration. Some authors have recommended the use of potassium phosphate salts instead of potassium chloride as a means of potassium replacement during therapy for diabetic ketoacidosis. Early routine use of potassium phosphate solutions to replace potassium should be discouraged. The need for phosphorus replacement, if at all, occurs several hours after therapy for diabetic ketoacidosis has begun, and potassium is usually required much sooner.
Several undesirable side effects from phosphate administration have been reported. These include hyperphosphatemia, hypocalcemia, hypomagnesemia, metastatic soft tissue calcifications, and hypernatremia and dehydration from osmotic diuresis. The serum phosphate level should be monitored during treatment of diabetic ketoacidosis, but the case for routine phosphate replacement has not been made.

COMPLICATIONS AND MORTALITY

Complications related to the disease state include aspiration of gastric contents by an unconscious patient, vascular stasis and deep vein thrombosis, and disseminated intravascular coagulation (DIC). Rhabdomyolysis during diabetic ketoacidosis has also been recently reported. Protection of the airway and evacuation of gastric contents are indicated in an unconscious patient. Prophylactic heparin therapy may help to prevent thrombotic complications.
Major complications related to the therapy of diabetic ketoacidosis include hypoglycemia, hypokalemia, paradoxical spinal fluid acidosis, and cerebral edema. The goal of therapy of diabetic ketoacidosis is to produce a gradual, even return to normal metabolic balance. Rapids shifts of the levels of water, electrolytes, and other solutes can be avoided by using isotonic saline as the initial intravenous fluid, refraining from excessive bicarbonate administration, replacing potassium early in the course of treatment, and using low-dose insulin techniques. Above all, a basic understanding of the pathophysiology of diabetic ketoacidosis, constant monitoring of the patient, and attention to detail are essential to prevent complications of treatment.
The development of cerebral edema during the treatment of diabetic ketoacidosis, especially in young patients, is a continuing problem. An extensive review found no specific treatment variables that contributed to the development of cerebral edema. Variables included overhydration, rapid osmolar changes, hemodynamic compromise and hypoxia. Young age and new-onset diabetes were the only identified contributing risk factors.
Approximately one-half of the patients who developed cerebral edema had premonitory symptoms of severe headache, incontinence, change in arousal or behavior, pupillary changes, blood pressure changes, seizures, bradycardia, or disturbed temperature regulation. Early recognition of neurologic deterioration and treatment of cerebral edema by hyperventilation and mannitol is the best hope for survival of this disastrous complication.
In general, the greater the presenting serum osmolality, blood urea nitrogen (BUN), and blood glucose concentration, the greater the mortality. There is also increased mortality for patients presenting with a serum bicarbonate level of less than 10 mEq/L.
Of the factors responsibile for precipitating diabetic ketoacidosis, infection and myocardial infarction are the main contributors to high mortality. Half the patients in diabetic ketoacidosis die when myocardial infarction is the precipitating event. Additional factors that reduce the chances of survival include old age, severe hypotension, prolonged and severe coma, and underlying renal and cardiovascular disease.

BIBLIOGRAPHY:

1)Fisher JN, Kitabchi AE: A randomized study of phosphate therapy in the treatment of diabetic ketoacidosis. J Clin Endocrinol Metab 57:177, 1983.
2)Fisher JN, Shahshahani MN, Kitabchi AE: Diabetic ketoacidosis: Low-dose insulin therapy by various routes. N Engl J Med 297:238, 1977.
3)Foster DW, McGarry JD: The metabolic derangements and treatment of diabetic ketoacidosis. N Engl J Med 309:159, 1983.
4)Kitabchi AE: Low-dose insulin therapy in diabetic ketoacidosis: Fact or fiction? Diabetes Metab Rev 5:337, 1989.
5)Page MM, Alberti KGMM, Greenwood R, et al: Treatment of diabetic coma with continuous low-dose infusion of insulin. Br Med J 2:687, 1974.
6)Rosenbloom AL: Intracerebral crises during treatment of diabetic ketoacidosis. Diabetes Care 13:22, 1990.