Ciammaichella M. M., Galanti A., Rossi C.
Dirigenti Medici
U.O.C. Medicina Interna I per l’Urgenza
(Direttore: Dott. G. Cerqua)
A.C.O. S. Giovanni - Addolorata - Roma, Italia
 

NONKETOTIC HYPEROSMOLAR COMA

KEYWORDS
nonktetotic hyperosmolar coma

 

BIBLIOGRAPHY
TABLE: Drugs and Procedures that May Cause Hyperosmolar Coma
TABLE:
Neurologic Manifestations of Nonketotic Hyperosmolar Coma




Nonketotic hyperosmolar coma is characterized by severe hyperglycemia, hyperosmolality, and dehydration, but no ketoacidosis. This metabolic derangement occurs primarily in diabetics but may occur in nondiabetics under certain circumstances. Many names have been used to identify this entity, but the term nonketotic hyperosmolar coma is used in this chapter.
This syndrome shares many features with diabetic ketoacidosis, including hyperglycemia and hyperosmolality, but the lack of ketoacidosis is its main distinguishing feature. Nonketotic hyperosmolar coma is much less common than diabetic ketoacidosis. Nonketotic hyperosmolar coma and diabetic ketoacidosis are part of a continuum, and when present in pure form, they represent the opposite ends of a spectrum with regard to lipid mobilization. In general, a patient with nonketotic hyperosmolar coma has a blood glucose concentration greater than 800 mg/dL, usually 1000 mg/dL or more, a serum osmolality greater than 350 mOsm/kg, and a negative test for serum ketones. By comparison, the average blood glucose level of a patient in diabetic ketoacidosis is usually less than 600 mg/dL, the serum osmolality is rarely above 350 mOsm/kg, and the test for serum ketones is strongly positive.
Nonketotic hyperosmolar coma occurs most commonly as the initial manifestation of non-insulin dependent diabetes (NIDDM). Most known diabetics who develop this syndrome have mild, NIDDM controlled by diet or oral hypoglycemic agents. A small minority of insulin-dependent patients on parenteral therapy develop nonketotic hyperosmolar coma. Both extremes nonketotic hyperosmolar coma and diabetic ketoacidosis have been reported to occur in the same patient.


PATHOPHYSIOLOGY

Any explanation of the pathogenesis of nonketotic hyperosmolar coma must explain why extreme hyperglycemia develops and why ketoacidosis does not. Neither of these questions has been answered with certainty. Simply put, extreme hyperglycemia develops because ketoacidosis does not. The failure of ketoacidosis to occur allows the underlying process to continue unrecognized and much higher levels of glucose to result. In addition, glucose enters the extracellular space more rapidly than it is excreted, leading to profound hyperglycemia.
When a patient with NIDDM is subjected to stress, the ? cells of the pancreas respond to the increased glucose concentration by increasing the secretion of insulin. Continued diabetogenic stress eventually exhausts the insulinogenic reserve of the ? cells, and plasma insulin levels fall. Because of the increased insulinogenic capacity of the patient with mild diabetes, higher levels of blood glucose occur before this reserve is depleted. If the patient is receiving insulin therapy, supplemental insulin allows additional time for ?-cell recovery and further prolongs the time required for exhaustion of the insulin reserve. In addition, elevated levels of glucagon may promote gluconeogenesis in the liver, resulting in massive hyperglycemia.
The reason ketoacidosis does not occur is not well understood, and explanations conflict. Some have found low levels of free fatty acids (FFAs), with normal insulin, glucocorticoid, and growth hormone levels. They believe that inhibition of lipolysis occurs because of relatively higher circulating insulin levels or lower lipolytic hormone levels than are present with diabetic ketoacidosis. When lipolysis is inhibited, the precursors required for ketone body formation are not released and ketoacidosis does not develop. It is known that the quantity of insulin required to inhibit lipolysis in adipose tissue is less than the quantity required to promote the utilization of glucose by peripheral tissues. Ketoacidosis may not develop because there is enough circulating insulin to inhibit lipolysis but an insufficient amount to protect against the development of hyperglycemia.
Others have reported similar high FFA levels and low circulating insulin levels in both nonketotic hyperosmolar coma and diabetic ketoacidosis. Additionally, glucagon and glucocorticoids (cortisol) have been found to be increased to the same extent in both conditions. These authors conclude that in nonketotic hyperosmolar coma the FFAs are mobilized to the same extent as in diabetic ketoacidosis, but that the intrahepatic oxidation of the incoming FFAs is directed along nonketogenic metabolic pathways, such as triglyceride synthesis, because of relatively low but adequate insulin action in the liver. Prehepatic and posthepatic insulin levels have been measured. In diabetic ketoacidosis, both pre- and posthepatic insulin levels are low, but in nonketotic hyperosmolar coma, prehepatic insulin levels twice the posthepatic ones have been found. Subscribers to this theory conclude that the liver is selectively bathed in insulin while the periphery is in a diabetic state. They conclude that the available insulin exerts its antiketogenic effect at the hepatic and not the adipocyte level.
Regardless of the pathogenesis of nonketotic hyperosmolar coma, the effect of hyperglycemia in producing osmotic diuresis and fluid and electrolyte imbalance is understood. When relative insulin insufficiency develops in the diabetic patient, osmotically active glucose is located in the extracellular fluid compartment. During insulin insufficiency the cell membrane is not freely permeable to glucose, and water is drawn from the intracellular compartment into the extracellular compartment in an attempt to achieve equal osmolality. The presence of large amounts of glucose in the extracellular compartment tends to preserve that compartment at the expense of cellular volume. Relative expansion of the extracellular fluid volume may protect against hypotension until late in the course of nonketotic hyperosmolar coma.
In addition to the internal shifts in body fluids, an osmotic diuresis also occurs. Normally, antidiuretic hormone (ADH) from the posterior pituitary gland acts to maintain water balance. With severe hyperglycemia, glucosuria produces an increased volume and rate of urine flow through the kidneys. Despite maximum levels of ADH, water can no longer be maximally reabsorbed, and an increased volume of urine results. Total body water is decreased and serum osmolality is increased. Fluid losses during nonketotic hyperosmolar coma range from 8 to 12 L.
Sodium balance is also upset by osmotic diuresis. Sodium reabsorption normally occurs in the distal tubules, mediated by the renin-aldosterone system. The concentration gradient against which sodium must be actively transported into the distal tubules is increased as water reabsorption diminishes. Thus a large proportion of the filtered sodium remains unabsorbed and passes into the urine. Nevertheless, the water loss during osmotic diuresis is greater than the sodium loss, and the patient becomes hypertonic relative to sodium. Prolonged diuresis results in hypovolemia and hypertonic dehydration.
Total body potassium depletion is also a consequence of osmotic diuresis. The distal tubules are under maximal stimulation by aldosterone, and some sodium is reabsorbed in exchange for potassium. Because of the longer duration of osmotic diuresis with nonketotic hyperosmolar coma, potassium depletion is greater than that which occurs with diabetic ketoacidosis and may reach 400 to 1000 mEq. Potassium depletion may not become evident until the patient is rehydrated. Other solutes such as magnesium and phosphate are also lost during osmotic diuresis.


CLINICAL PRESENTATION

Nonketotic hyperosmolar coma is most common in the middle-aged or elderly and occurs most commonly in diabetics, although the majority are undiagnosed at the time of presentation. Patients who depend upon others to meet their needs, such as infants, nursing home patients, and the mentally retarded, are particularly vulnerable to its insidious onset. Inaccessibility to water coupled with an inability to communicate masks the early signs and symptoms.

Precipitating Factors

Minor upper respiratory infection or gastroenteritis is capable of precipitating diabetic ketoacidosis, but an illness of greater magnitude is usually required before nonketotic hyperosmolar coma results. Infection is a common precipitating cause, especially gram-negative pneumonias. Other precipitating illnesses include myocardial infarction, cerebrovascular accidents, gastrointestinal (GI) hemorrhage, acute pyelonephritis, acute pancreatitis, uremia, subdural hematomas, and peripheral vascular occlusion. Chronic renal and cardiovascular disease is common.
Various drugs have been linked to the onset of nonketotic hyperosmolar coma (Table -1). Most are dehydrating agents or have side effects of impairing insulin release from the pancreas or of interfering with the peripheral action of insulin. Thiazide diuretics and diazoxide possess both characteristics and are well-recognized causes of nonketotic hyperosmolar coma. Other drugs causally related to this syndrome include corticosteroids, phenytoin, mannitol, cimetidine, propranolol, calcium channel blockers, and immunosuppressive agents.
Clinical situations that can result in severe dehydration or an excessive glucose load or both may produce this syndrome in nondiabetic patients. These include extensive burns, heatstroke, hypothermia, peritoneal or hemodialysis with a hypertonic glucose solution, and hyperalimentation; these causes are most commonly seen in hospitalized patients.
An occasional patient may have a history of ingesting enormous quantities of sugar-containing fluids. This patient is usually alert and has a lesser degree of dehydration than the usual patient with nonketotic hyperosmolar coma.
The prodromal period during the development of nonketotic hyperosmolar coma is longer than that for diabetic ketoacidosis. Metabolic changes occur over many days to several weeks. Symptoms of polyuria, polydipsia, and increasing lethargy are almost always present but may not be appreciated. Failure to develop ketoacidosis and its clinical manifestations may allow the underlying process to go unrecognized until stupor or coma develops. Decreased responsiveness is the main reason patients receive medical attention.

Physical Findings

There are no specific physical findings. Virtually all are significantly dehydrated. Fever, hypotension, and tachycardia may be present. Shock is especially common if gram-negative pneumonia is present. Respirations are variable. Kussmaul's breathing is not a feature of uncomplicated nonketotic hyperosmolar coma, but hyperventilation may be present if the patient is acidotic for other reasons. Shallow respirations with hyperpnea and tachypnea are usual. The smell of acetone on the breath is absent.
The most prominent physical findings are neurologic (Table -2). Almost all patients exhibit some disturbance in mentation, ranging from inappropriate response to confusion, drowsiness, stupor, or coma. The higher the osmolality, the greater the obtundation. The average osmolality for a comatose patient with nonketotic hyperosmolar coma is 380 mOsm/kg. Depression of the sensorium does not correlate with the glucose concentration or with the pH of the plasma or cerebrospinal fluid.
The most common focal signs are hemisensory deficits or hemiparesis or both. Approximately 15 percent of the patients exhibit seizure activity, usually of the focal motor type (85 percent). Grand mal seizures can occur. Tremors, fasciculations, and a variety of other neurologic abnormalities including aphasia, hyperreflexia, flaccidity, depressed deep tendon reflexes, positive plantar response, and nuchal rigidity may be seen. In one series by Arieff and Carroll, 12 of 33 patients with nonketotic hyperosmolar coma were initially diagnosed as probably acute stroke. This diagnosis was not confirmed in any of the patients.
Considering the age of the patient population and the frequency of neurologic findings, it is not surprising that the misdiagnosis of stroke or organic brain syndrome is common. Nonketotic hyperosmolar coma must be suspected in every elderly, dehydrated patient with glucosuria or hyperglycemia, especially if they are mild diabetics and receiving diuretic drugs or glucocorticoids.

Laboratory

Confirmation of the diagnosis is with laboratory findings. The essential tests are blood glucose levels, calculated and measured serum osmolality, and serum ketone levels. A reasonable approximation of the blood glucose and serum ketone levels can be made promptly at the bedside by use of the nitroprusside test and glucose reagent strips. Additional tests should include a complete blood cell count and levels of electrolytes, blood urea nitrogen (BUN), creatinine, and arterial blood gases.
Serum electrolytes display a variable pattern. Serum sodium values usually range from 120 to 160 mEq/L, but because water is lost in excess of sodium through osmotic diuresis, the patient is almost always hypertonic. There is a sodium decrement of 1.6 mEq/L for every 100 mg/dL increase in blood glucose. Total body potassium depletion is usually severe. Potassium loss in nonketotic hyperosmolar coma is greater than that with diabetic ketoacidosis because of the longer duration of osmotic diuresis, GI loss, and prior diuretic use.
The BUN level is almost always elevated because of extracellular volume depletion and underlying renal disease. The initial BUN level is elevated out of proportion to the creatinine level. Ratios of BUN to creatinine may be 30:1. Prerenal azotemia resolves with volume replacement, but most patients have underlying chronic renal impairment.
Metabolic acidosis due to the accumulation of ketone bodies is not a feature of nonketotic hyperosmolar coma, but metabolic acidosis due to other causes can occur. In most series, 30 to 40 percent of the patients have a mild metabolic acidosis attributed to accumulation of lactic acid or due to uremia. However, in a significant number of these cases, no cause of the acidosis can be identified. Striking elevations of creatine phosphokinase (CPK) in patients with nonketotic hyperosmolar coma have been reported and are attributed to the complication of rhabdomyolysis. This condition as a complication of severe hyperosmolality has been reported.
Because of the frequency of underlying chronic disease and precipitating illnesses, a search for a precipitating cause must be made. Urinalysis, chest roentgenogram, ECG, and cultures of the blood, urine, and sputum should be performed. Because of the frequency of fever and neurologic signs, including nuchal rigidity, a CT scan and lumbar puncture may be necessary. Typical findings in the spinal fluid include a normal opening pressure, a markedly elevated glucose level (usually 50 percent of the serum value), a normal or slightly elevated protein level, and an osmolality identical to that of the serum.


TREATMENT

Attention to detail and constant monitoring are necessary. Serial measurements of glucose, electrolyte, and serum osmolality levels are essential. A flow sheet to record therapeutic measures and patient response is recommended. The specific goals of therapy of nonketotic hyperosmolar coma include correction of hypovolemia and dehydration, restoration of electrolyte balance, and reduction of serum glucose and hyperosmolality levels. Reasonable end points that can usually be achieved within 36 h are a blood glucose level of 250 mg/dL, a serum osmolality of 320 mOsm/kg, and a urine output of at least 50 mL/h.

Crystalloid

No agreement exists on the composition of the initial replacement fluid. Some authors advocate the use of isotonic saline (0.9% NaCl), and others recommend the use of half-normal saline (0.45% NaCl). Those who advocate isotonic saline believe that the most immediate threat to life is hypovolemic shock. Even though the patient has lost water in excess of solute and is hypertonic, normal saline is still hypotonic to the patient with nonketotic hyperosmolar coma. Normal saline corrects the extracellular volume deficit, stabilizes the blood pressure, and maintains adequate urinary flow. Once this has been achieved, hypotonic saline can be administered to provide free water for correction of intracellular volume deficits.
Those who recommend half-normal or hypotonic saline as initial fluid therapy argue that any osmotically active solute in the replacement fluid prolongs and enhances the hyperosmotic state. Further, since the patient has lost water in excess of solute, a hypotonic solution is the logical replacement.
All authors agree that if the patient is in hypovolemic shock, isotonic saline should be given until volume has been restored. Most agree that if the patient has significant hypernatremia (155 mEq/L) or hypertension, hypotonic saline should be the initial fluid of choice.
Rarely a patient has hyperglycemia, hyponatremia, and a low or normal osmolality. This indicates a significant excess of water, probably due to the ingestion of enormous quantities. The use of hypotonic saline in this setting can precipitate water intoxication.
There are no controlled studies that compare the advantages of isotonic solutions with those of hypotonic solutions in the initial management of nonketotic hyperosmolar coma. Regardless of the fluid used, there are guidelines to determine the rate and amount of fluid administration.
The average fluid deficit in nonketotic hyperosmolar coma is usually between 20 and 25 percent of total body water (TBW), or 8 to 12 L. In elderly subjects, it is assumed that 50 percent of the body weight is due to TBW. By using the patient's usual weight in kilograms, normal TBW and water deficit (20 to 25 percent of TBW) can be calculated. One-half of the estimated water deficit should be replaced during the first 12 h and the balance during the next 24 h. Ongoing insensible and urinary losses should also be replaced.
Renal cardiac and cerebral function must be monitored. Too rapid correction of glucose and osmolality can result in cerebral edema.

Electrolytes

Electrolyte replacement is an essential part of therapy for nonketotic hyperosmolar coma. In the average patient, for every liter of body water lost, 70 mEq of monovalent ions is concomitantly lost. That translates into 300 to 800 mEq of sodium and potassium that usually needs to be replaced.
The sodium deficit is replenished by the administration of normal saline (154 mEq of sodium per liter) or half-normal saline (77 mEq of sodium per liter). Potassium replacement, as with diabetic ketoacidosis, should be started early in the course of treatment. Potassium supplement should be started within 2 h of the institution of fluid and insulin therapy or as soon as adequate renal function has been confirmed. Most authors recommend the infusion of KCl at a rate of 10 to 20 mEq/h during the acute phase of therapy (24 to 36 h). Potassium should be added to the initial intravenous fluid if the patient presents with hypokalemia. Magnesium levels should be obtained and replacement given if the level is low. Caution is necessary in the presence of renal dysfunction.

Insulin

Traditionally it has been taught that the insulin requirement of a patient with nonketotic hyperosmolar coma is less than that of a patient with diabetic ketoacidosis. The difference in insulin requirement was attributed to decreased insulin resistance in the patient with nonketotic hyperosmolar coma because of the absence of acidosis.
Changing concepts have led to a reappraisal of insulin therapy, and continuous intravenous infusion of low doses of insulin and intermittent intramuscular injections of low doses of insulin are effective therapy.
The usual insulin dose is 0.1?units/kg/h, given by continuous intravenous infusion or by intramuscular injection. If the intramuscular route is chosen, 20 units of regular insulin can initially be administered intramuscularly or by intravenous bolus. Often no additional insulin is required after the initial dose. No insulin should be given after the blood glucose level reaches approximately 300 mg/dL.
The reasons for not using large doses of insulin when treating nonketotic hyperosmolar coma are even more compelling than those stated for diabetic ketoacidosis. In addition to producing a more gradual reduction of the glucose concentration, thus avoiding hypoglycemia, hypokalemia, and cerebral edema, low-dose insulin techniques may help to avoid vascular collapse and renal shutdown in the patient with nonketotic hyperosmolar coma.
Cerebral edema in nonketoic hyperosmolar coma is far less common than in patients with diabetic ketoacidosis. It usually occurs when the metabolic abnormalities have largely been corrected. As the patients’ clinical condition is improving, there is an abrupt decrease in the sensorium, increased lethargy, elevated blood pressure and decreasing heart rate. The pediatric population may be more prone to this complication than adults. One study suggests that rapid lowering of the blood glucose below 300 mg/dl during the first 24 h of insulin therapy may contribute to the genesis of cerebral edema. Treatment of this complication is usually ineffectual and the mortality rate is 75%.
A high glucose concentration in the extracellular fluid compartment protects that compartment against hypovolemia at the expense of intracellular water. If the concentration of glucose is rapidly lowered by the administration of large doses of insulin, insufficient extracellular osmotic solute may result in a net intracellular shift of large volumes of water, producing hypovolemia and vascular collapse. Similarly, osmotic diuresis induced by hyperglycemia acts to protect the kidney against acute tubular necrosis (ATN) in the presence of reduced renal perfusion. If the blood glucose concentration is rapidly reduced, the osmotic diuresis decreases, and ATN may result. Acute tubular necrosis after institution of large-dose insulin therapy was reported in 5 of 30 patients in the series studied by Arieff and Carroll.

Glucose

Glucose should be added to the intravenous solution when the blood glucose level declines to 250 mg/dL. It is at this level that further rapid lowering of the blood glucose concentration may result in cerebral edema. Cerebral edema can be recognized clinically by the sudden onset of hyperpyrexia, hypotension, and deepening of coma in spite of biochemical improvement. Though cerebral edema during treatment of nonhyperosmolar coma is rare, it is usually fatal and can be prevented.

Additional Treatment

The role of phosphate replacement during treatment of nonketotic hyperosmolar coma is controversial. The plasma phosphorus level should be monitored during therapy, but a case for routine phosphate infusion has not been made convincingly.
Patients with nonketotic hyperosmolar coma are at risk for the development of arterial and venous thrombosis. Low-dose prophylactic heparin therapy should be considered.
Seizures are usually due to hyperosmolality and electrolyte abnormalities, but CNS mass lesions and meningitis or encephalitis should be considered. Unless the patient has an underlying seizure disorder or has status epilepticus, specific treatment for single seizures is not necessary. Phenytoin may precipitate nonketotic hyperosmolar coma and should be given with caution.

BIBLIOGRAPHY:


1)Arieff AI: Cerebral edema complicating nonketotic hyperosmolar coma. Mineral Electrolyte Metab 12:383, 1986.
2)Arieff AI, Carroll HJ: Nonketotic hyperosmolar coma with hyperglycemia: Clinical features, pathophysiology, renal function, acid-base balance, plasma-cerebrospinal fluid equilibria and the effects of therapy in 37 cases. Medicine 51:73, 1972.
3)Bendezu R, Wieland RH, Furst BH, et al: Experience with low-dose insulin infusion in diabetic ketoacidosis and diabetic hyperosmolarity. Arch Intern Med 138:60, 1978.
4)Gerich JE, Martin MM, Recant L: Clinical and metabolic characteristics of hyperosmolar nonketotic coma. Diabetes 20:228, 1971.
5)Podolsky S: Hyperosmolar nonketotic coma in the elderly diabetic. Med Clin North Am 62:815, 1978.

TABLE
Drugs and Procedures that May Cause Hyperosmolar Coma

Hydrochlorothiazide and other thia- Cimetidine (Tagamet)
zide diuretics Propranolol (Inderal)
Chlorthalidone (Hygroton, Thalitone) Asparaginase (Elspar)
Furosemide (Lasix) Immunosuppressive agents
Ethacrynic acid (Edecrin) Mannitol
Diazoxide (Hyperstat, Proglycem) Peritoneal dialysis
Calcium channel blockers Hemodialysis
Glucocorticoids Intravenous hyperosmolar
Phenytoin (Dilantin) alimentation
Chloropromazine (Thorazine)


TABLE
Neurologic Manifestations of Nonketotic Hyperosmolar Coma

Diffuse Focal
Seizures Focal seizures
Lethargy Todd's paralysis
Confusion Hemisensory loss
Delirium and hallucinations Hemiparesis
Stupor Babinski's reflex
Coma Aphasia
Hemianopsia
Tonic eye deviation
Nystagmus
Hyperreflexia
Choreoathetosis