Scope

Glucose is critical to function of the central nervous system because it is the primary fuel for these tissues. Plasma levels of glucose are strictly regulated—even after a large oral glucose load or exercise the serum level should not deviate from a range of 70 to 150 mg/dL.

Epidemiology

More than 23 million individuals in the United States have diabetes mellitus, and this number continues to increase at an accelerated rate, partially because of the worsening obesity epidemic in this country. In almost 6 million of these individuals, however, the diabetes is undiagnosed. In addition to the cost of life and morbidity associated with this disease, the financial expense is enormous. In 2007 the estimated direct and indirect cost of treating diabetes mellitus in the United States was $116 billion and $58 billion, respectively.¹

Structure and Function

Insulin is a 51–amino acid protein produced by the beta cells of the islet of Langerhans in the endocrine pancreas. After the initial protein, preproinsulin, is translated on the rough endoplasmic reticulum, it is cleaved serially first to proinsulin and then to insulin and C peptide. Insulin and its C peptide are stored in a 1 : 1 ratio in secretory granules and released primarily in response to glucose and, to a lesser extent, amino acids. Release can be further potentiated or inhibited by a number of gastrointestinal and systemic hormones.

On release, insulin binds to its membrane-spanning receptor. Binding induces a conformational change in the structure of the receptor so that it becomes enzymatically active; it is now a functional tyrosine kinase that can initiate anabolic pathways.

A newer classification system of diabetes mellitus reflects the pathophysiology of the disease and long-term treatment options. The new system identifies four types of diabetes mellitus: type 1, type 2, gestational diabetes, and “other” (Table 162.1).

Type 1 diabetes mellitus (note the Arabic numbering) has replaced older terms such as type I, insulin-dependent, and juvenile-onset diabetes mellitus. These older terms became confusing for a multitude of reasons. For example, a small subset of “type II” diabetics fail oral hypoglycemic treatment and must be treated with an insulin regimen; are these patients “insulin dependent”? With increasing worldwide childhood obesity, more and more childhood diabetics are being seen who are not “insulin dependent,” and yet their disease cannot be classified as “adult onset.”

Type 1 diabetes mellitus can best be defined as an absolute deficiency of insulin. The mechanism is complex and occurs in approximately 5% of all patients with diabetes mellitus. The process usually begins years before symptoms appear when the patient is exposed to an antigen (e.g., viral infection) that is similar in structure to a protein found in islet beta cells. The immune system begins to produce a humoral and cell-mediated assault on these antigens that leads to progressive destruction of the cells. Destruction of beta cells results in a decrease in insulin levels, and eventually a critical point is reached at which insulin requirements are no longer met and hyperglycemia ensues.

The other categories include patients with relative insulin deficiency; it is important to note that the majority of hyperglycemic patients in the remaining groups do produce insulin, and thus it is easier to conceptualize their insulin deficiency as a balance between insulin and other counterregulatory hormones (e.g., epinephrine, glucagon, cortisol, growth hormone).

Type 2 diabetes mellitus has replaced older terms such as type II, non–insulin-dependent, and adult-onset diabetes mellitus. The process leading to this type of diabetes also begins years before the onset of overt clinical symptoms. For a multitude of reasons, most commonly obesity, peripheral tissues become increasingly resistant to the effects of insulin. Such resistance leads to increased production of insulin by beta cells, which allows years of relative glucose control. Eventually, the relative insulin resistance can no longer be met by increasing beta cell production, and the patient begins to experience hyperglycemia. Additionally, the beta cells may begin to “burn out” and ultimately produce progressively less and less insulin. The onset of clinical symptoms may be insidious in an otherwise healthy patient or abrupt when significant illness produces a spike in the counterregulatory hormones (e.g., epinephrine, glucagons, cortisol, growth hormone) that tends to increase plasma glucose levels. This abrupt onset occurs because a type 2 diabetic is not able to increase insulin production to counteract this rise, as would a nondiabetic.

The third category is gestational diabetes mellitus, which occurs in about 2% to 5% of all pregnancies, most often in the second or third trimester. It is believed to occur in a manner similar to that of type 2 diabetes. Pregnancy induces increased levels of human placental lactogen, estrogen, and cortisol—all hormones that tend to increase plasma glucose levels. Pregnant women are usually able to produce insulin in sufficient quantities to combat this increase in glucose-elevating hormones; however, susceptible women cannot. This condition most often resolves after delivery, but as one would expect, these women are susceptible to the development of type 2 diabetes later in life. Gestational diabetes mellitus can cause fetal complications, mostly as a result of increased fetal plasma glucose levels. In response to this elevated glucose derived via placental blood, the fetal pancreas increases plasma insulin production, which results in increased fetal birth weight.

The fourth category—“other”—is a catchall that contains all other causes of diabetes mellitus, including genetic anomalies causing malfunctioning insulin protein, insulin receptors, and beta cells in general, as well as other immune-mediated causes. Any significant insult to the exocrine pancreas—be it trauma, chronic pancreatitis, or cystic fibrosis—may result in this type of diabetes. Many common drug-induced causes of diabetes mellitus fall into this category (Box 162.1), as well as endocrinopathies such as hyperthyroidism, Cushing syndrome, and pheochromocytoma. Infectious causes include congenital rubella and cytomegalovirus. Less common causes include genetic disorders that may be associated with diabetes mellitus, including Down syndrome, Klinefelter syndrome, Turner syndrome, Prader-Willi syndrome, Huntington chorea, and porphyria.

Clinical Presentation

The signs and symptoms of diabetes mellitus can be quite varied. The classic manifestations are discussed here, but it is important to note that many patients do not have these symptoms. Typically, patients with type 1 diabetes have had years of progressively decreasing insulin secretion as a result of beta cell destruction. However, most often an infection or other stressor acutely increases levels of the counterregulatory hormones and the patient is unable to counter this spike with an insulin surge. Hyperglycemia (generally defined as a fasting glucose level > 126 mg/dL) then ensues, and the patient suddenly seeks care in the emergency department (ED).

As the plasma glucose level increases and eventually surpasses the threshold of the kidneys to reabsorb glucose, the patient begins to spill glucose into urine. This leads to osmotic diuresis and ensuing polyuria and polydipsia.

Diagnostic Testing

All patients with a history of diabetes mellitus should have an early point-of-care glucose assay performed when seen in the ED.² At a minimum, diabetic patients with systemic complaints or complaints common to hyperglycemia require glucose testing at the time of first assessment. It is important to note that if serial tests are to be performed, there is a small but significant difference between capillary and venous blood glucose levels.³ Additionally, any patient with altered mental status or new neurologic concerns should also have glucose levels tested because patients with hypoglycemia or hyperglycemia may exhibit these changes.

The purpose of laboratory testing in a hyperglycemic patient is to differentiate simple hyperglycemia from DKA and less commonly from HHS. It is important to note that no reliable historical or physical examination findings are sensitive or specific enough to confirm or exclude these acute and serious complications of diabetes in hyperglycemic patients.⁴ A bicarbonate level below 15 mmol/L with an elevated anion gap (varies depending on the laboratory, but the upper limit is generally approximately 16 mEq/L) strongly suggests DKA. A more complete laboratory evaluation for hyper-glycemia includes venous pH, β-hydroxybutyrate (BHB), and possibly serum osmolality. Additional laboratory tests may be necessary as dictated by the clinical picture. It has recently been suggested that acetoacetate (ACA), the standard ketone assayed for by serum and urine “ketone” assays, is neither sensitive nor specific for the diagnosis of DKA.^5,6

Treatment

Patients with significant hyperglycemia, diabetic ketoacidosis (DKA), or hyperglycemic hyperosmolar state (HHS) are almost universally dehydrated and intravascularly depleted. Even though this is certainly more so in the latter two conditions, all conditions require hydration. The initial treatment is intravenous (IV) hydration with normal saline (NS). Insulin is not the initial treatment of these diabetic complications.

Pathophysiology

Hyperglycemia ensues when the endocrine pancreas cannot produce enough insulin to decrease plasma levels of glucose appropriately (>126 mg/dL in a fasting patient). In type 1 diabetes, this is due to an absolute deficiency; in the other types of diabetes, it is due to a relative deficiency of insulin in comparison with the levels of counterregulatory hormones. In this way, glucose levels increase in plasma and eventually overwhelm the renal glucose threshold, thereby resulting in spillage of glucose into urine. By osmosis, water is pulled into urine, which leads to increased urination and dehydration. This in turn leads to increased thirst and polydipsia. The degree of hyperglycemia and dehydration varies greatly from patient to patient and even in the same patient on different occasions.

Diagnosis and Diagnostic Testing

The purpose of laboratory testing in a hyperglycemic patient is to differentiate simple hyperglycemia from DKA and less commonly from HHS.

It is important to ascertain the probable cause of the hyperglycemia. Although dietary indiscretion and medication noncompliance do play a role, these diagnoses should be considered only after excluding more serious causes. The most concerning causes can be grouped into two classes: infection and infarction.

Complete a thorough evaluation for possible sources of infection in all diabetic patients.⁷ Chest radiography is indicated to search for pneumonia in patients with historical and physical examination findings suggesting pneumonia, patients in whom a thorough history and physical examination cannot be obtained, clinically ill patients, and patients at the extremes of age.

Infarction-related causes of hyperglycemia include acute coronary syndrome (acute myocardial infarction and unstable angina), pulmonary embolism, and cerebrovascular accident. It is important to note that acute coronary syndrome is very likely to be manifested in an atypical manner in diabetic patients (e.g., new-onset congestive heart failure without any history of chest pain or dyspnea without chest pain).⁸ Any hyperglycemic patient with these findings should undergo a complete ED evaluation for acute coronary syndrome. A computed tomography scan of the brain or chest may be required if cerebrovascular accident or pulmonary embolism is a concern.

Treatment

Because hyperglycemia is most often associated with some degree of dehydration, the primary modality of treatment should be rehydration with NS. Early insulin therapy is contraindicated before determining electrolyte levels. After the patient is significantly rehydrated, laboratory studies have excluded additional complications such as DKA, and electrolyte status has been stabilized, subcutaneous insulin can be administered.

IV bolus administration of insulin has no role in the treatment of hyperglycemia. Administration of insulin via a continuous drip is not indicated, except in very special circumstances in which exceedingly tight glucose control is required (e.g., during a progressing cerebral vascular accident) for the treatment of simple hyperglycemia. In fact, very tight glucose control in an ill patient has been suggested to have no effect on patient outcome other than significantly increased rates of hypoglycemia.⁹ The dose of insulin depends on the degree of hyperglycemia after hydration and the patient’s previous exposure to insulin therapy. Patients with known diabetes treated with insulin therapy may be given their usual dose after hydration. Patients new to insulin may be given low-dose subcutaneous insulin with the goal of decreasing glucose to acceptable levels at a rate of 100 mg/dL/hr.

A guideline for subcutaneous regular insulin dosing is presented in Table 162.2. This guideline is appropriate for hyperglycemic patients who have little to no previous experience with subcutaneous insulin. Those managed with insulin regimens may do better with one approximating their typical dosage. In addition, this guideline assumes that the patient has first been rehydrated and remains hyperglycemic.

Table 162.2 Regular Subcutaneous Insulin Dosing Guideline*

^* See text for a discussion of modifications of this guideline. Patients treated with regular insulin regimens should be given their usual dosage if appropriate for their condition.

It is important to note that euglycemia may not be a realistic or even appropriate goal in these patients while they are in the ED; longer-term (over a period of days to weeks) personalization of an insulin or oral hypoglycemic regimen by the patient’s primary care provider or inpatient physician is preferred. The ED goal may be simply to rehydrate the patient and then use subcutaneous insulin to further decrease the patient’s glucose level. Targeting a “normal glucose” level in patients new to insulin therapy is fraught with risks, mostly notably hypoglycemia. Moreover, no target “maximum allowable glucose level” before discharge of the patient has been established.

Treatment and Disposition

The initial medication for patients with new-onset diabetes is most often a low-dose sulfonylurea. A good choice of sulfonylurea is glyburide (1.25 to 2.5 mg orally once a day) or glipizide (2.5 to 5 mg orally once a day). These doses may not allow strict glucose control but are appropriate early therapy and pose little risk for hypoglycemia. When starting these medications, patients should be instructed to take them with an early meal or breakfast and to eat regular meals throughout the day. Metformin (850 mg orally once a day) is an appropriate choice for initiating diabetes therapy when a nonsulfonylurea drug is preferred. This drug, when used alone in initial therapy, poses a very low risk for hypoglycemia and may be a good choice for obese patients.¹⁰

All patients with new-onset diabetes should have follow-up arranged in an expedited manner. Patients with comorbid conditions or acute disease should be admitted for inpatient evaluation and treatment.

Epidemiology

The approximate incidence of DKA is 5 to 8 cases per 1000 diabetics per year. Mortality has remained approximately 4% to 5% for the last decade in patients in whom DKA is diagnosed and treatment begun early in the disease course. Some estimates of mortality in patients in whom diagnosis and treatment are delayed run as high as 14%.¹¹ Other estimates suggest that in as many as 20% of patients DKA is misdiagnosed; these patients therefore have a significantly increased risk for death.

Pathophysiology

A hyperglycemic patient has a deficiency in insulin; type 1 diabetics have an absolute deficiency and type 2 diabetics have a relative deficiency. When this deficiency is significant, patients are unable to uptake glucose from blood into cells and must rely on the metabolism of fat for energy. Fatty acids from blood are metabolized in the liver to three ketone bodies: ACA, BHB, and acetone. Acetone is a minor component and may be removed from the body via the lungs; this is often manifested as a fruity odor on the breath. The other ketone bodies are in equilibrium with their ratio determined by the redox state and relative levels of nicotinamide adenine dinucleotide (NAD) to NADH (the reduced form of NAD); consequently, the majority of ketones generated during DKA are in the form of BHB. The increasing level of BHB causes an acidosis and leads to significant electrolyte shifts, including shift of potassium out of cells into blood. Because of the continued hyperglycemia, the patient experiences hyperglycemic osmotic diuresis, which leads to significant renal potassium loss and total body potassium depletion. The acidosis also causes a decrease in serum bicarbonate levels and eventually overwhelms this buffering system and results in decreased serum pH.

These changes lead to the classic findings of DKA, including dehydration, Kussmaul breathing (deep, sighing type of respiration caused by acidosis-induced stimulation of the central respiratory center, which may be seen in other forms of acidosis), abdominal discomfort, vomiting, and often altered mental status.

Clinical Presentation

The causes of DKA are many, and they are similar to those of hyperglycemia (Box 162.3). The onset may be due to any significant stressor (classically an infection or infarction) or may be due to a deficiency of insulin, usually because the insulin dosage is insufficient, oral therapy is ineffective, or the patient is not compliant with therapy. It is important to note that diabetic patients may have acute conditions—for example, patients with diabetes and pneumonia—and may therefore need increased insulin administration temporarily or DKA will develop; in these cases the cause of the ketoacidosis is both infection and inadequate insulin administration.

Differential Diagnosis, Diagnostic Testing, and Testing Pitfalls

The differential diagnosis of DKA is summarized in Box 162.4. In any patient who may have DKA, the laboratory evaluation shown in Box 162.5 should be performed. No single standard laboratory diagnosis has been established for DKA; however, any diagnosis should include the factors noted in Box 162.6. It should be stressed that the diagnosis of DKA is based mainly on clinical findings; although laboratory evaluation is important, common complicating factors often make laboratory diagnosis difficult. Each of the components of the laboratory diagnosis is fraught with limitations and qualifications.

Hyperglycemia is commonly considered the cornerstone of the work-up; however, DKA in the presence of “euglycemia” is not an uncommon finding. In fact, approximately 30% of patients with DKA have a glucose level lower than 300 mg/dL,¹² and some studies suggest that the longer patients are in a state of DKA, the more likely they are to be euglycemic.¹³ The reasons for this are manyfold. Patients with DKA often have gastrointestinal disturbances that cause vomiting and therefore limited oral intake, with the result that liver glycogen stores will be depleted and gluconeogenesis, which may be inhibited during acidosis,¹⁴ will be the sole source of glucose production. It has also been suggested that patients with poorly controlled diabetes do not store glucose as liver glycogen as readily or efficiently as do nondiabetic individuals or patients whose disease is well controlled. Liver disease of any cause will also limit or prevent glycogen storage and gluconeogenesis.

Acidosis, or decreased serum bicarbonate, and an elevated anion gap are the basic components of the diagnosis. Patients with DKA often vomit considerably and lose acid through this mechanism. They can also become exceptionally dehydrated, which leads to stimulation of the renin-angiotensin system and promotes the loss of both potassium and hydrogen ion in the distal tubule of the kidneys. Medications such as diuretics and antacids may also cause a metabolic alkalosis.

Although the majority of patients have a low serum bicarbonate level and pH, as well as an elevated anion gap, approximately 10% will have at least one of these factors reported as normal.⁵ It should be noted that venous pH is perfectly acceptable and there is no reason to obtain an arterial sample.¹⁵

The final component of the diagnosis is the presence of ketones. As discussed previously, the majority of ketones are in the form of BHB, but the standard laboratory urine and serum examinations assay for ACA. Recent work has suggested that ACA is neither sensitive nor specific for DKA, whereas BHB appears to be very sensitive and specific for this disease.^5,6,16 Additionally, ACA levels may not be high enough to be detected, especially in the dilute urine of a patient experiencing hyperglycemic osmotic diuresis. However, as the patient is rehydrated and treatment begins, serum BHB is converted to ACA. Thus, in a not-uncommon scenario, a hyperglycemic patient is initially negative for urine ketones but after rehydration and treatment begins to “suddenly spill” them and is falsely labeled as worsening and beginning to be in DKA when in fact the patient already was in DKA but the diagnosis was missed because the urine examination tested for the “wrong” ketone.

Treatment

A treatment plan for DKA is outlined in Box 162.7. Early treatment is similar to that for hyperglycemia. Patients usually exhibit moderate to severe dehydration and should receive NS. The average patient without a history of congestive heart failure or renal failure often requires 5 to 8 L of fluid over the course of the hospitalization. Potassium levels must be checked and hypokalemia corrected concurrent with the administration of insulin because patients may have significant (often hundreds of milliequivalents) total body potassium depletion. Insulin will drive extracellular potassium into cells and, because of the total body depletion, may cause an exaggerated serum hypokalemia that may lead to cardiac arrhythmia.

Other electrolytes such as magnesium and phosphorus are less crucial and may be administered as usual. Correction of even moderate hypophosphatemia has not been shown to be beneficial in these patients. It is therefore recommended that only severe hypophosphatemia (<1 mmol/L) or moderate hypophosphatemia with clinical findings such as respiratory muscle weakness or cardiomyopathy be corrected.¹⁸ If necessary, potassium chloride may be replaced with potassium phosphate for this purpose. Hypomagnesemia may be corrected with IV magnesium sulfate.

Bicarbonate therapy rarely has a place in the treatment of patients with DKA. Although administration of bicarbonate to a patient with metabolic acidosis may seem logical, it is rarely helpful and may cause multiple, significant complications.^19,20 Because bicarbonate cannot cross the blood-brain barrier but carbon dioxide can, administration of bicarbonate may allow increased carbon dioxide to enter cerebrospinal fluid and cause a paradoxic cerebrospinal fluid acidosis. Additionally, bicarbonate administration may worsen the hypokalemia by driving potassium into cells. Studies have clearly shown administration of bicarbonate to a patient with DKA and a pH of at least 6.8 to be of no benefit.¹⁹ No studies specifically support the administration of bicarbonate for any pH in patients with DKA. If bicarbonate administration has any role, it may be only in a patient with shock and impending or present cardiovascular collapse secondary to marked acidosis and dehydration.

Administration of insulin should begin only after the initial hydration and electrolyte correction. IV bolus insulin has no role in treatment. Administration of IV bolus insulin leads to a supraphysiologic serum insulin level that can cause a significant drop in plasma potassium levels and cardiac arrhythmias.²¹ It may also cause hypoglycemia and cerebral and pulmonary edema, and it has been theorized to lead to changes in gene-regulated protein synthesis that may exert its effects for weeks.^22,23 Additionally, because IV regular insulin has a plasma half-life of less than 5 minutes, a continuous, low-dose infusion reaches a steady-state level quickly.²⁴

The standard insulin protocol has been to start a drip of regular insulin at 0.1 U/kg. More recently it has been suggested that lower doses may work as well with less risk for hypokalemia, and some authors have even suggested the use of subcutaneous insulin.²⁵

Patients treated by insulin drip are at risk for hypoglycemia and hypokalemia, as mentioned previously, and therefore require regular electrolyte and glucose assays. The author recommends alternating a bedside glucose assay with a basic chemistry panel every hour and charting a flow sheet as shown in Table 162.3. Potassium can then be supplemented as necessary. The patient’s fluid should be changed to D₅ NS (5% dextrose solution in NS) when the glucose level reaches 250 mg/dL. When the anion gap has resolved to 15 mEq/L or less, the patient should be given subcutaneous insulin at the usual dose or at a dose similar to those detailed in Box 162.7. Two hours later, the insulin drip can be discontinued; this approach allows subcutaneous insulin administration and the insulin drip to overlap by 2 hours.

It is not necessary nor should it be expected for the pH to return to normal before discontinuing the insulin drip. Because the treatment regimen by itself will often cause hyperchloremic acidosis, monitoring the anion gap is more helpful than monitoring the pH. In a patient with normally functioning kidneys, the chloride level and pH will then be slowly corrected over a period of hours to days. In a patient who has otherwise recovered, mild acidosis at this point should not prevent discharge.

Cerebral edema has long been the most feared complication of pediatric DKA. It occurs in approximately 1% of all cases but carries a mortality rate quoted to be as high as 50%. It has long been held that the treatment regimen, especially rapid rehydration, has been responsible for this complication. However, no data support this belief, and the current literature strongly suggests that there is no casual relationship between cerebral edema in pediatric patients with DKA and their treatment regimen¹⁷; instead, it is more likely that the degree of illness better correlates with the likelihood of this deadly complication. Treatment is similar to that for other causes of cerebral edema while the standard management of DKA continues.

Disposition

All discharged patients should begin an insulin regimen. Oral hypoglycemic medications are insufficient after treatment of DKA.

Patients being treated with an insulin drip need to be monitored closely for all the reasons discussed previously; they are usually admitted to an intensive care unit or other specialized unit that can closely monitor their glucose, electrolytes, and clinical status.

If patients with DKA are hungry and have no other contraindication, they should be allowed to eat. Eating a normal meal will decrease the likelihood of hypoglycemia and hypokalemia.

Pathophysiology

Like DKA, HHS is initiated by a relative lack of insulin; however, the insulin deficiency is usually significantly less profound than that of DKA. Very low levels of insulin (or a low ratio of insulin to counterregulatory hormones) are necessary to prevent ketoacidosis; as insulin levels increase, further gluconeogenesis and glycogen metabolism are sequentially switched off. Because just small amounts of insulin are needed to prevent ketosis, only limited amounts of ketones are produced during HHS. The patient becomes successively more dehydrated secondary to the glucose-driven osmotic diuresis. The dehydration may eventually lead to impaired renal function and an inability to excrete the continually produced glucose, thereby resulting in severe hyperglycemia.²⁷

Clinical Presentation

Typically, these patients may state that they cannot or do not wish to drink liquids; patients with HHS may thus have an impaired thirst sensation or may be unable to obtain water (e.g., elderly or bedridden, psychiatric, or jailed patients).

Despite many similarities between DKA and HHS, they differ in some very important ways (Table 162.4). As discussed earlier in regard to DKA, patients with HHS may often have atypical findings. The diagnosis of HHS is associated with glucose levels higher than 600 mg/dL, but the average glucose level is approximately 900 mg/dL, and it is not uncommon for the glucose level to be well in excess of 1000 mg/dL.

Table 162.4 Comparison of Classic Laboratory Findings in Patients with Diabetic Ketoacidosis and Hyperglycemic Hyperosmolar State

AMS, Altered mental status; BHB, β-hydroxybutyrate; DKA, diabetic ketoacidosis; HHS, hyperglycemic hyperosmolar state; MS, mental status.

Diagnostic Testing and Testing Pitfalls

Diagnostic criteria for HHS are summarized in Box 162.8. As with DKA, it needs to be stressed that the diagnosis is based mainly on clinical findings; although laboratory evaluation is important, common complicating factors often make laboratory diagnosis difficult.

When considering the diagnosis, any treatment before collection of samples for laboratory tests must be noted. These patients are severely dehydrated and hyperglycemic; any fluid administration—for example, by the emergency medical service—will significantly decrease their glucose level. In this context it is not uncommon for HHS to be diagnosed in patients with an initial glucose level of just 500 mg/dL. Serum osmolarity may also decrease to a lesser degree with initial fluid resuscitation.

Absence of ketosis is required to differentiate pure HHS from pure DKA. However, many patients lie in a spectrum between these two entities and may form a small amount of BHB. Additionally, because most laboratories actually assay for ACA, as discussed earlier, mild ketoacidosis is common secondary to anorexia and vomiting. Another classic finding of HHS is a “normal pH.” However, it is not uncommon to have a mild acidotic hyperosmolar state secondary to lactic acidosis caused by lack of perfusion to peripheral tissues from the severe dehydration. In fact, approximately one half of patients with HHS are believed to have a mild anion gap metabolic acidosis.²⁸ In addition, the pH may temporarily decrease further during the initial fluid resuscitation as the peripherally produced lactic acid is returned to the liver for processing.

Treatment

A treatment plan for HHS is summarized in Box 162.9. As with simple hyperglycemia and DKA, the primary treatment is fluid resuscitation with NS. The average fluid deficit is 9 L.²⁹ Resuscitation includes administration of boluses of 1 to 2 L of NS initially, followed by administration of NS until there is improvement in vital signs that suggests improved hemodynamics, as indicated by the onset of urine output and an improved clinical hydration state. Fluids then can be changed to NS.

Failure to change the fluid at this point is apt to lead to significant and clinically detrimental hypernatremia. Fluid administration can be guided further by the following:

Electrolytes also require aggressive monitoring and replacement. Total body deficits of potassium, magnesium, and phosphate cannot be gauged accurately from the initial laboratory assays and should be supplemented aggressively once urine output has been established. Note that although the initial measured potassium may be normal or even high, the patient is still severely potassium depleted. Therefore, potassium supplementation must be started early in the course of treatment and certainly before the initiation of any insulin therapy.

Phosphate and magnesium supplementation may be more essential in the management of HHS than in the treatment of DKA. Because HHS characteristically develops over a period of days to weeks, total body stores of these electrolytes are more likely to have been significantly affected by the osmotic diuresis. Although compelling studies are lacking, it is probably of greater urgency to supplement these electrolytes early in the course of treatment.^30,31

Insulin administration should begin only after the patient has been fluid-resuscitated to the point of hemodynamic stability and potassium replacement has been started. Insulin has no role in the initial treatment of HHS—early administration can lead to cardiovascular collapse.

The effective osmolarity of the vasculature in patients with HHS is dependent on the high level of glucose present. This level may decrease slowly with the administration of NS as sodium begins to replace glucose in maintaining proper tonicity. However, if insulin is administered early before appropriate NS resuscitation is attained, this cannot happen. Insulin will drive glucose intracellularly, thereby effectively and quickly decreasing intravascular tonicity, which may lead to acute and catastrophic cardiovascular collapse. Additionally, early insulin administration will drive potassium intracellularly and risk the same hypokalemia-induced arrhythmias discussed earlier in the treatment of DKA.

After early fluid and electrolyte administration and when the patient is hemodynamically stable, insulin therapy can begin. Typically, a drip of regular insulin is started at 0.05 to 1 U/kg/hr; many authorities suggest low doses of 2 to 4 U/hr. This is then titrated to control the rate of decrease in glucose to approximately 50 to 100 mg/dL/hr. As with DKA and hyperglycemia, IV bolus administration of insulin has no role in the treatment of HHS.

Disposition

Patients should be admitted to the intensive care unit because frequent neurologic examinations and regular laboratory studies are required during the first 24 hours of treatment. Electrolyte supplementation will need to be adjusted regularly based on the results of these laboratory tests. As with patients with DKA, if patients being treated for HHS are hungry, their mental status has returned to baseline, and no other contraindications are present, there is no reason that they should not eat. They are often ravenously hungry, and eating will act naturally to decrease the likelihood of hypoglycemia and acute electrolyte deficiencies.