Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Clinical features, evaluation, and diagnosis

INTRODUCTION — 

Diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state (HHS, also known as hyperosmotic hyperglycemic nonketotic state [HHNK]) are two of the most serious acute complications of diabetes. DKA is characterized by ketoacidosis and typically hyperglycemia, while HHS usually has more severe hyperglycemia but no ketoacidosis (table 1).

The precipitating factors, clinical features, evaluation, and diagnosis of DKA and HHS in adults will be reviewed here. The epidemiology, pathogenesis, and treatment of these disorders are discussed separately. DKA in children is also reviewed separately.

●(See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis".)

●(See "Diabetic ketoacidosis in adults: Treatment".)

●(See "Hyperosmolar hyperglycemic state in adults: Treatment".)

●(See "Diabetic ketoacidosis in children: Clinical features and diagnosis".)

●(See "Diabetic ketoacidosis in children: Treatment and complications".)

PRECIPITATING FACTORS — 

A precipitating event can usually be identified in patients with diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic state (HHS) (table 2) [1-3]. The most common events are infection (often pneumonia or urinary tract infection) and omission or inadequate use of insulin therapy. Compromised water intake due to underlying medical conditions, particularly in older patients, can promote the development of severe dehydration and HHS [2,4,5].

Other conditions and factors associated with DKA and HHS include:

●Acute major illnesses such as myocardial infarction, cerebrovascular accident, sepsis, or pancreatitis.

●New-onset type 1 diabetes, in which DKA is a common presentation.

●In established type 1 diabetes, omission of insulin in the setting of gastroenteritis when the patient mistakenly stops insulin because of reduced oral intake.

●Drugs that affect carbohydrate metabolism, including glucocorticoids, higher-dose thiazide diuretics, sympathomimetic agents (eg, dobutamine and terbutaline) [6], and second-generation "atypical" antipsychotic agents [7].

●Sodium-glucose cotransporter 2 (SGLT2) inhibitors, mostly used in type 2 diabetes but also used off-label in type 1 diabetes [8].

●Cocaine use, which has been associated with recurrent DKA [9,10].

●Psychological problems associated with eating disorders and purposeful insulin omission, particularly in young patients with type 1 diabetes [11]. Factors that may lead to insulin omission in younger patients include fear of weight gain, fear of hypoglycemia, rebellion from authority, and the stress of chronic disease.

●Poor compliance with the insulin regimen.

●Malfunction of continuous subcutaneous insulin infusion (CSII) devices, which was initially reported in the early 1980s [12]. Pump malfunction is now uncommon, but system failure due to blockage or leakage in the syringe or the infusion set or connectors, causing an interruption of infusion flow is common, and can lead to DKA [13]. The frequency of DKA with pump therapy in children, however, appears to be no different from that with multiple daily injections of insulin [14]. Fast-acting aspart and lispro-aabc are both approved with some, but not all, insulin pumps in the United States (U-200 lispro-aabc is not approved with any pump). There are no data to know if these faster insulins result in more frequent or severe DKA should an interruption of insulin flow occur. (See "Continuous subcutaneous insulin infusion (insulin pump)", section on 'Pump failure'.)

CLINICAL PRESENTATION — 

Diabetic ketoacidosis (DKA) usually evolves rapidly, over a 24-hour period. In contrast, symptoms of hyperosmolar hyperglycemic state (HHS) develop more insidiously with polyuria, polydipsia, and weight loss, often persisting for several days before hospital admission.

The earliest symptoms of marked hyperglycemia are polyuria, polydipsia, and weight loss. As the degree or duration of hyperglycemia progresses, neurologic symptoms, including lethargy, focal signs, and obtundation, can develop. This can progress to coma in later stages. Neurologic symptoms are most common in HHS, while hyperventilation and abdominal pain are primarily limited to patients with DKA.

Neurologic symptoms — Neurologic deterioration primarily occurs in patients with an effective plasma osmolality (Posm) above 320 to 330 mosmol/kg [15-17] (see 'Plasma osmolality' below). Mental obtundation and coma are more frequent in HHS than DKA because of the usually greater degree of hyperosmolality in HHS (table 1) [18]. In addition, some patients with HHS have focal neurologic signs (hemiparesis or hemianopsia) and/or seizures [18-22]. Mental obtundation may occur in patients with DKA, who have lesser degrees of hyperosmolality, when severe acidosis exists [23]. However, stupor or coma in diabetic patients with an effective Posm lower than 320 mosmol/kg demands immediate consideration of other causes of the mental status change.

Abdominal pain in DKA — Patients with diabetic ketoacidosis (DKA) may present with nausea, vomiting, and abdominal pain; although more common in children, these symptoms can be seen in adults [24]. Abdominal pain is unusual in HHS. In a review of 189 consecutive episodes of DKA and 11 episodes of HHS, abdominal pain was reported in 46 percent of patients with DKA compared with none of the patients with HHS [25]. Abdominal pain was associated with the severity of the metabolic acidosis (occurring in 86 percent of those with a serum bicarbonate ≤5 mEq/L but only 13 percent of those with a serum bicarbonate ≥15 mEq/L) but did not correlate with the severity of hyperglycemia or dehydration.

Possible causes of abdominal pain include delayed gastric emptying and ileus induced by metabolic acidosis and associated electrolyte abnormalities [15]. Other causes for abdominal pain, such as pancreatitis, should be sought when they occur in the absence of severe metabolic acidosis and when they persist after the resolution of ketoacidosis.

Physical examination — Signs of volume depletion are common in both DKA and HHS and include decreased skin turgor, dry axillae and oral mucosa, low jugular venous pressure, tachycardia, and, if severe, hypotension. Neurologic findings, noted above, also may be seen, particularly in patients with HHS. (See 'Neurologic symptoms' above and "Etiology, clinical manifestations, and diagnosis of volume depletion in adults".)

Patients with DKA may have a fruity odor (due to exhaled acetone; this is similar to the scent of nail polish remover) and deep respirations reflecting the compensatory hyperventilation (called Kussmaul respirations).

DIAGNOSTIC EVALUATION — 

Both diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state (HHS) are medical emergencies that require prompt recognition and management (table 3).

Initial evaluation — The initial evaluation of patients with hyperglycemic crises should include assessment of cardiorespiratory status, volume status, and mental status. The initial history and rapid but careful physical examination should focus on:

●Airway, breathing, and circulation (ABC) status

●Mental status

●Possible precipitating events (eg, source of infection, myocardial infarction)

●Volume status

The initial laboratory evaluation of a patient with suspected DKA or HHS should include determination of:

●Serum glucose

●Serum electrolytes (with calculation of the anion gap), blood urea nitrogen (BUN), and plasma creatinine

●Complete blood count (CBC) with differential

●Urinalysis and urine ketones by dipstick

●Plasma osmolality (Posm)

●Serum or blood beta-hydroxybutyrate (by enzymatic assay or point-of-care ketone meter [if available], respectively)

●Arterial or venous blood gas if the serum bicarbonate is substantially reduced; arterial blood gas if hypoxia is suspected

●Electrocardiogram

Additional testing, such as cultures of urine, sputum, and blood, serum lipase and amylase, and chest radiograph should be performed on a case-by-case basis. Infection (most commonly pneumonia and urinary tract infection) is a common precipitating event. Thus, cultures should be obtained if there are suggestive clinical findings. Recognize that infection may exist in the absence of fever in these patients [26-28]. Testing for coronavirus 2019 (COVID-19) should be performed if clinically suspected, as infection is associated with intensive care unit admission for DKA [29]. (See "COVID-19: Issues related to diabetes mellitus in adults", section on 'Clinical presentations'.)

Measurement of glycated hemoglobin (A1C) may be useful in determining whether the acute episode is the culmination of an evolutionary process in previously undiagnosed or poorly managed diabetes or a truly acute episode in an otherwise well-managed patient. However, people with rapidly progressive or "fulminant" type 1 diabetes may have lower A1C levels at diagnosis [30].

Laboratory findings — Hyperglycemia and hyperosmolality are the two primary laboratory findings in patients with DKA or HHS; patients with DKA also have a high anion gap metabolic acidosis (table 1).

A variety of other laboratory tests may be affected. The impact of hyperglycemia, insulin deficiency, osmotic diuresis, and fluid intake in each individual patient leads to variable laboratory findings, depending upon the relative importance of these factors.

Serum glucose

●Classic presentation – The serum glucose concentration may exceed 1000 mg/dL (56 mmol/L) in HHS [16,31], but it is generally less than 800 mg/dL (44 mmol/L) and often in the 350 to 500 mg/dL (19.4 to 27.8 mmol/L) range in DKA [31,32]. The mechanism underlying the hyperglycemia in DKA and HHS is reviewed in detail separately. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis", section on 'Hyperglycemia'.)

●Normoglycemic DKA – Normoglycemic DKA, in which the serum glucose is normal or near normal, has been described, particularly in patients with poor oral intake, treatment with insulin prior to arrival in the emergency department, in pregnant women, and with sodium-glucose cotransporter 2 (SGLT2) inhibitors [8,33-36].

When SGLT2 inhibitors block the SGLT2, the resulting glucosuria can minimize or prevent the development of hyperglycemia, despite very low insulin levels/activity and development of ketoacidosis. SGLT2 inhibitors and their adverse effects are reviewed in detail elsewhere. (See "Sodium-glucose cotransporter 2 inhibitors for the treatment of hyperglycemia in type 2 diabetes mellitus", section on 'Diabetic ketoacidosis'.)

Patients with normoglycemic diabetic ketoacidosis generally require both insulin and glucose to reverse the ketoacidosis. (See "Diabetic ketoacidosis in adults: Treatment", section on 'Fluid replacement' and "Diabetic ketoacidosis in adults: Treatment", section on 'Insulin'.)

Serum ketones — Three ketone bodies are produced and accumulate in DKA: acetoacetic acid, which is the only one that is a true ketoacid; beta-hydroxybutyric acid (a hydroxyacid formed by the reduction of acetoacetic acid); and acetone, which is derived from the decarboxylation of acetoacetic acid. Acetone is a true ketone, not an acid. Testing for serum ketones is generally performed whenever ketoacidosis is suspected or the urine ketone test is positive. Urine ketone bodies are detected with nitroprusside tests, while serum ketones can be detected with either a nitroprusside test or by direct assay of beta-hydroxybutyrate levels. Direct assay of beta-hydroxybutyrate levels is preferred, particularly for monitoring response to therapy. (See "Diabetic ketoacidosis in adults: Treatment", section on 'Monitoring'.)

Nitroprusside testing — This chemical develops a purple color in the presence of acetoacetic acid (and to a much lesser degree, acetone). Urine dipstick testing with nitroprusside tablets or reagent sticks is widely utilized, and results are available within minutes.

Although semiquantitative serum ketone testing with nitroprusside was the major methodology used for detecting elevated blood ketoacid and acetone levels for many years, it has now been replaced in most hospitals by enzymatic measurement of serum beta-hydroxybutyrate levels.

Enzymatic measurement of serum beta-hydroxybutyrate — The serum nitroprusside test for ketone bodies has been largely replaced by enzymatic assays for beta-hydroxybutyrate [37]. Several beta-hydroxybutyrate assay instruments are commercially available [38-41]. Enzymatic assays for serum beta-hydroxybutyrate eliminate the problems associated with nitroprusside testing (eg, false positives and false negatives). One limitation of most of these assays, however, is that beta-hydroxybutyrate is not quantitated above a level of 6 mmol/L. An ideal assay for "ketoacids" would measure the concentrations of both acetoacetate and beta-hydroxybutyrate.

Point-of-care bedside analyzers to measure capillary blood beta-hydroxybutyrate are increasingly available [42,43].

Anion gap metabolic acidosis — The serum anion gap is calculated as follows:

Serum anion gap = Serum sodium - (serum chloride + bicarbonate)

By convention, this calculation uses the actual measured plasma sodium concentration, not the sodium concentration corrected for the simultaneous glucose concentration. When quick beta-hydroxybutyrate assays are available, the anion gap is less important for diagnosing and monitoring DKA, but it still has diagnostic utility. For example, a large anion gap may be the critical diagnostic clue when DKA coexists with metabolic alkalosis (eg, due to vomiting) [44].

The serum bicarbonate concentration in DKA is usually moderately to markedly reduced. In contrast, the serum bicarbonate concentration is normal or only mildly reduced in HHS (table 1). Patients with DKA usually present with a serum anion gap greater than 20 mEq/L (reference range approximately 3 to 10 mEq/L). In patients with DKA, the elevated anion gap metabolic acidosis is caused by the accumulation of beta-hydroxybutyric and acetoacetic acids. However, the increase in anion gap is variable and determined by several factors: the rate and duration of ketoacid production, the rate of metabolism of the ketoacids and their loss in the urine, the bicarbonate distribution space, and the volume of distribution of the ketoacid anions. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis", section on 'Anion gap metabolic acidosis'.)

Compensatory hyperventilation reduces the partial pressure of carbon dioxide (CO₂) and mitigates the fall in arterial pH. However, severe ketoacidosis can reduce the pH below 7, especially if hyperventilation is compromised.

Plasma osmolality — Plasma osmolality (Posm) is always elevated in patients with HHS but less so with DKA (table 1). The typical total body deficits of water and electrolytes in DKA and HHS are compared in a table (table 4). In patients with HHS, the effective Posm is typically >320 mosmol/kg (reference range approximately 275 to 295 mosmol/kg).

Effective Posm (or tonicity; in mosmol/kg) is the portion of total osmolality that is generated by sodium salts and glucose (and, if present, mannitol or sucrose). Effective osmoles do not penetrate most cell membranes and can cause movement of water across membranes to achieve osmolal equilibrium. Effective Posm does not include "ineffective" osmoles, such as urea, because urea is rapidly permeable across most cell membranes and its accumulation does not induce major water shifts between the intracellular spaces (including the brain) and the extracellular water space [45].

Effective osmolality can be estimated with either of the following equations, depending upon the units for sodium (Na) and glucose:

Effective Posm = [2 x Na (mEq/L)] + [glucose (mg/dL) ÷ 18]

Effective Posm = [2 x Na (mmol/L)] + glucose (mmol/L)

The Na concentration in these equations is the actual measured plasma Na concentration and not the corrected Na concentration. The Na is multiplied by two to account for the osmotic contribution of sodium's accompanying anions (primarily chloride and bicarbonate). Eighteen is a factor to convert glucose units from mg/dL into mmol/L.

If the Posm is measured, using a freezing point reduction osmometer, the result is the total osmolality. Because effective osmolality excludes urea osmoles (BUN), it can be estimated as:

Effective Posm = Measured Posm - [BUN (mg/dL) ÷ 2.8]

Effective Posm = Measured Posm - BUN (mmol/L)

2.8 is a factor to convert urea concentration from units of mg/dL into mmol/L.

Serum sodium — Most patients with DKA and HHS are mildly hyponatremic [46]. However, patients with HHS who have a marked osmotic diuresis may present with a normal or even elevated serum Na concentration, despite a markedly elevated serum glucose concentration that can exceed 1000 mg/dL (56 mmol/L) [47]. These patients have a markedly elevated effective Posm and often have neurologic symptoms that can include seizures and coma (see 'Neurologic symptoms' above). Inadequate water intake contributes to the hyperosmolality and is a particular problem in hot weather and in older individuals who may have an impaired thirst mechanism [48].

The measured serum Na concentration in uncontrolled diabetes mellitus is variable because of the interaction of multiple factors, some that lower Na concentration and others that raise it. The increase in Posm created by hyperglycemia pulls water out of the cells, and this reduces the plasma Na concentration. Theoretical calculations suggested that, in the absence of urine losses, the serum Na concentration should fall by approximately 1.6 mEq/L for each 100 mg/100 mL (5.5 mmol/L) increase in glucose concentration, but subsequent studies in experimental participants and patients found higher ratios. We use the following ratio: The Na concentration will fall by approximately 2 mEq/L for each 100 mg/100 mL (5.5 mmol/L) increase in glucose concentration. The "corrected" Na concentration can then be approximated by adding 2 mEq/L to the plasma Na concentration for each 100 mg/100 mL (5.5 mmol/L) increase above normal in glucose concentration (calculator 1). (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis", section on 'Plasma osmolality and sodium' and "Diabetic ketoacidosis in adults: Treatment", section on 'Fluid replacement'.)

Pseudohyponatremia/pseudohyperchloremia — Some patients with uncontrolled diabetes develop marked hyperlipidemia and have lactescent serum. This can create electrolyte artifacts when certain analyzers are utilized. Hyperlipidemia will displace water and thereby reduce the plasma water phase fraction below its normal 93 percent. If any dilution step in the analysis requires an exact volume of plasma (or serum), then a reduced amount of water (and electrolyte) will be analyzed. This can result in "pseudohyponatremia." This artifact occurs if the analysis is performed by flame photometric analysis or with an indirect potentiometric, or ion selective electrode, analyzer. Direct potentiometry does not require a dilution step and is not susceptible to this artifact [49-51]. (See "Diagnostic evaluation of adults with hyponatremia".)

However, hyperlipidemia can have an opposite artifactual effect on the chloride concentration when certain chloride analyzers are employed, generating marked pseudohyperchloremia [52]. (See "Serum anion gap in conditions other than metabolic acidosis", section on 'Negative serum anion gap'.)

Serum potassium — Patients presenting with DKA or HHS have a potassium deficit that averages 300 to 600 mEq (table 4) [46,53,54]. A number of factors contribute to this deficit, particularly increased urinary losses due both to the glucose osmotic diuresis and to the excretion of potassium ketoacid anion salts. Despite these total body potassium deficits, hypokalemia is observed in only approximately 5 percent of cases [55,56]. The serum potassium concentration is usually normal or, in one-third of patients, elevated on admission [31,46,57]. This is due to a shift of potassium from intracellular fluid to extracellular fluid (ECF) caused by hyperosmolality and insulin deficiency [6,45,46]. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis", section on 'Potassium'.)

Insulin therapy shifts potassium into cells and lowers the potassium concentration. This may cause severe hypokalemia, particularly in patients who present with a normal or low serum potassium concentration [54]. Careful monitoring and timely administration of potassium supplementation are essential. (See "Diabetic ketoacidosis in adults: Treatment", section on 'Potassium replacement'.)

Serum phosphate — Patients with uncontrolled hyperglycemia are typically in negative phosphate balance because of decreased phosphate intake, an acidosis-related shift of phosphate into the ECF when metabolic acidosis exists, and phosphaturia caused by osmotic diuresis. Despite phosphate depletion, the serum phosphate concentration at presentation is usually normal or even high because both insulin deficiency and metabolic acidosis cause a shift of phosphate out of the cells and as a result of ECF volume contraction [58,59].

This transcellular shift is reversed and the true state of phosphate balance is unmasked after treatment with insulin and volume expansion. In a review of 69 episodes of DKA, the mean serum phosphate concentration fell from 9.2 mg/dL (3 mmol/L) at presentation to 2.8 mg/dL (0.9 mmol/L) at 12 hours and, in some patients, to levels as low as 1 mg/dL (0.32 mmol/L) [60]. (See "Diabetic ketoacidosis in adults: Treatment", section on 'Phosphate repletion (rarely needed)'.)

Serum creatinine — Most patients with uncontrolled hyperglycemia have acute elevations in the BUN and serum creatinine concentration, which reflect the reduction in glomerular filtration rate induced by hypovolemia. High acetoacetate levels can also artifactually increase serum creatinine levels when certain colorimetric assays are utilized [61]. However, most laboratories now utilize enzymatic assays, which are not affected by this artifact [62].

Serum amylase and lipase — Acute pancreatitis may precipitate or complicate DKA. Serum amylase and lipase are generally used to diagnose acute pancreatitis, but each of these enzymes is often elevated in patients with DKA who do not have any other clinical or radiological evidence of pancreatitis [63-67]. Therefore, the diagnosis of pancreatitis in patients with DKA should be primarily based upon clinical findings and imaging. (See "Clinical manifestations, diagnosis, and natural history of acute pancreatitis", section on 'Imaging'.)

The mechanisms for non-pancreatitis-associated hyperamylasemia and hyperlipasemia in DKA are not well understood, but the following observations have been made:

●In a review of 134 consecutive episodes of DKA in patients without evidence of acute pancreatitis on computed tomography (CT) scan, elevations of serum amylase and lipase (threefold or higher in some patients) were seen in 17 and 24 percent, respectively [63]. Abdominal pain was present in 19 percent of the patients in this series.

●The source of these nonspecific amylase elevations is most often salivary, though some may also be of pancreatic origin [64,65,67]. The source of nonspecific lipase elevations is not known.

●The rise in amylase correlates with pH and Posm, while the rise in lipase correlates only with Posm [63]. Values peak within 24 hours of presentation [67].

●In 100 consecutive cases of DKA, 11 did have acute pancreatitis confirmed by CT scan. The most common causes of pancreatitis in these patients were hypertriglyceridemia and chronic alcohol intake [68]. Two of the 10 evaluable patients (one was comatose) did not have abdominal pain.

Leukocytosis — The majority of patients with hyperglycemic emergencies present with leukocytosis, which is proportional to the degree of ketonemia [6,69]. Leukocytosis unrelated to infection may occur as a result of hypercortisolemia and increased catecholamine secretion [70]. However, a white blood cell count greater than 25,000/microL or more than 10 percent bands increases suspicion for infection and should be evaluated [71].

Lipids — Patients with DKA or HHS may present with marked hyperlipidemia and lactescent serum. In a study of 13 patients with DKA, the mean plasma triglyceride and cholesterol levels on admission were 574 mg/dL (6.5 mmol/L) and 212 mg/dL (5.5 mmol/L), respectively [72]. Triglycerides fell below 150 mg/dL (1.7 mmol/L) in 24 hours with insulin therapy. Elevated triglycerides may generate pseudohyponatremia and/or pseudohyperchloremia. (See 'Serum sodium' above.)

Lipolysis in DKA, and to a lesser extent in HHS, is due to insulin deficiency, combined with elevated levels of lipolytic hormones (catecholamines, growth hormone, corticotropin [ACTH], and glucagon). Lipolysis releases glycerol and free fatty acids into the circulation. High levels of serum fatty acids cause insulin resistance at both the peripheral and the hepatic level, and they serve as the substrate for ketoacid generation in hepatocyte mitochondria. Insulin is the most potent anti-lipolytic hormone.

DIAGNOSTIC CRITERIA — 

Diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state (HHS) are distinguished by the absence of ketoacidosis and usually greater degree of hyperglycemia in HHS [26,31,73]. The diagnostic criteria proposed by the American Diabetes Association (ADA) for mild, moderate, and severe DKA and HHS are shown in the table (table 1).

●DKA is characterized by the triad of hyperglycemia, anion gap metabolic acidosis, and ketonemia. Metabolic acidosis is often the major finding. The serum glucose concentration is usually less than 800 mg/dL (44 mmol/L) and commonly between 350 to 500 mg/dL (19.4 to 27.8 mmol/L) [31,32]. However, serum glucose concentrations may exceed 900 mg/dL (50 mmol/L) in patients with DKA who are comatose [74]. In certain settings, such as starvation, pregnancy, treatment with insulin prior to arrival in the emergency department, or use of sodium-glucose cotransporter 2 (SGLT2) inhibitors, the glucose level may be mildly elevated or even normal.

●In HHS, ketoacid accumulation is mild or absent, the serum glucose concentration may exceed 1000 mg/dL (56 mmol/L), the plasma osmolality (Posm) may reach 380 mosmol/kg, and neurologic abnormalities are frequently present (including coma in 25 to 50 percent of cases) [16,26,31]. Most patients with HHS have an admission pH >7.30, a serum bicarbonate >20 mEq/L, a serum glucose ≥600 mg/dL (33.3 mmol/L), and test negative for ketones in serum and urine, although mild ketonemia may be present.

Factors that contribute to the lesser degree of hyperglycemia in DKA, compared with HHS, are discussed elsewhere. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis", section on 'Hyperglycemia'.)

Significant overlap between DKA and HHS has been reported in more than one-third of patients [4,15,17,75].

DIFFERENTIAL DIAGNOSIS — 

The differential diagnosis of hyperglycemic crises includes other causes of ketosis, acidosis, hyperosmolality, and/or coma (table 5).

Alcoholic ketoacidosis and starvation ketosis — Alcoholic ketoacidosis (AKA) and starvation ketosis are other causes of ketoacidosis. Low-carbohydrate diets can also precipitate ketoacidosis [76]. Metabolic acidosis can be relatively severe in patients with AKA. However, when ketoacidosis develops as a result of starvation, it is usually relatively mild. Ketoacid levels with prolonged fasting rarely exceed 8 to 10 mEq/L, and the serum bicarbonate concentration is typically greater than 17 mEq/L [77]. More severe ketoacidosis may develop with prolonged fasting in children and pregnant women [78,79]. (See "Fasting ketosis and alcoholic ketoacidosis".)

Ketoacidosis without hyperglycemia in a patient with chronic alcoholism is virtually diagnostic of AKA [80]. Modest elevations in serum glucose in patients with alcoholic ketoacidosis may reflect underlying unrecognized diabetes or a catecholamine-mediated stress response [81]. Measurement of A1C can confirm chronic hyperglycemia.

Anion gap acidosis — Diabetic ketoacidosis (DKA) must be distinguished from other causes of high anion gap metabolic acidosis including lactic acidosis (which can rarely be associated with, or generated by, metformin, particularly in patients with impaired kidney function); aspirin or acetaminophen toxicity and poisoning with methanol, ethylene glycol, and propylene glycol; D-lactic acidosis; and advanced chronic kidney disease (table 6). Although none of these disorders cause ketoacidosis (table 5), several different types of acidosis may coexist, especially lactic acid and ketoacidosis. (See "Approach to the adult with metabolic acidosis".)

Hyperemesis of pregnancy — Hyperemesis of pregnancy is characterized by excessive vomiting, weight loss, and ketonuria, and it typically occurs during the first trimester of pregnancy. Serum bicarbonate and pH may be normal or elevated due to concomitant metabolic alkalosis. (See "Nausea and vomiting of pregnancy: Clinical findings and evaluation".)

Metabolic encephalopathy — Toxic metabolic encephalopathy is usually a consequence of systemic illness, and its causes are diverse. In patients with diabetes, both severe hypoglycemia and severe hyperglycemia can result in coma. Measurement of fingerstick (capillary) or plasma glucose may be diagnostic (table 5). (See "Acute toxic-metabolic encephalopathy in adults".)

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Hyperglycemic emergencies".)

INFORMATION FOR PATIENTS — 

UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Diabetic ketoacidosis (The Basics)" and "Patient education: Hyperosmolar hyperglycemic state (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Precipitating factors – A precipitating factor can usually be identified in patients with diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic state (HHS). Common factors include (table 2) (see 'Precipitating factors' above):

•Infection (eg, pneumonia, urinary tract infection, coronavirus 2019 [COVID-19])

•Omission or inadequate use of insulin therapy in people with existing diabetes (including insulin pump malfunction)

•New-onset diabetes (more common with type 1 diabetes)

•Drugs (eg, glucocorticoids, sodium-glucose cotransporter 2 [SGLT2] inhibitors)

●Clinical presentation – Signs and symptoms of DKA are the result of:

•Hyperglycemia (eg, polyuria, polydipsia, weight loss)

•Volume depletion (eg, decreased skin turgor, dry oral mucosa, tachycardia)

•Metabolic acidosis (eg, nausea, vomiting, abdominal pain, deep and rapid [Kussmaul] respirations)

Symptoms of DKA evolve rapidly over 24 hours (table 3). As the degree and duration of metabolic disturbances progress, neurologic symptoms, including lethargy and obtundation, may develop.

In contrast, symptoms of HHS develop more insidiously (eg, days to a week). (See 'Clinical presentation' above.)

●Diagnostic evaluation – The initial evaluation of patients with hyperglycemic crises should include assessment of cardiorespiratory status, volume status, and mental status (table 3).

The initial laboratory evaluation of a patient with suspected DKA or HHS should include determination of serum glucose; electrolytes (with calculation of the anion gap), blood urea nitrogen (BUN), plasma creatinine, and serum phosphorus; complete blood count (CBC) with differential; urinalysis and urine ketones by dipstick; plasma osmolality in selected patients (eg, severe hyperglycemia and dehydration accompanied by central nervous system [CNS] findings); serum ketones; venous pH or arterial blood gas (if the serum bicarbonate is substantially reduced or hypoxia is suspected); and electrocardiogram. (See 'Diagnostic evaluation' above.)

●Laboratory findings – DKA and HHS primarily differ according to the presence of ketoacidosis and the degree of hyperglycemia (table 1). Hyperglycemia and ketoacidosis are the primary laboratory findings in patients with DKA, whereas HHS is often characterized by more severe hyperglycemia in the absence of ketonemia.

•Serum glucose – In patients with new-onset diabetes presenting as DKA, glucose is elevated, often in the 350 to 500 mg/dL (19.4 to 27.8 mmol/L) range. In patients with a prior history of diabetes, glucose may be normal, minimally above normal, or elevated.

Normoglycemic (eg, glucose <200 mg/dL) DKA may occur with poor oral intake, treatment with insulin prior to arrival in the emergency department, in pregnancy, and with use of SGLT2 inhibitors.

•Serum ketones – Three ketone bodies are produced in DKA: one ketoacid (acetoacetic acid), one hydroxyacid (beta-hydroxybutyric acid), and one neutral ketone (acetone). Urine ketone bodies are detected by a dipstick with nitroprusside tests, while serum ketones can be detected by enzymatic assay of beta-hydroxybutyrate levels or with a nitroprusside test. Enzymatic assay of beta-hydroxybutyrate levels is preferred. (See 'Serum ketones' above.)

•Anion gap metabolic acidosis (DKA) – In DKA, the serum bicarbonate concentration is usually moderately to markedly reduced (eg, <18 mEq/L [18 mmol/L]), and the venous pH is usually reduced due to metabolic acidosis. However, mixed acid-base disorders are common in DKA (eg, DKA with vomiting may cause a mixed metabolic acidosis and metabolic alkalosis). In such cases, venous pH may not be reduced.

In DKA, the calculated anion gap = Na - (Cl + HCO₃) is often elevated, reflecting increased accumulation of beta-hydroxybutyrate and acetoacetic acids. However, the increase in anion gap is variable and determined by several factors. (See 'Anion gap metabolic acidosis' above and "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis", section on 'Anion gap metabolic acidosis'.)

•Serum sodium – Most patients with DKA and HHS are mildly hyponatremic, but patients who have a marked osmotic diuresis may have a normal or even elevated serum sodium (Na) concentration. Most patients with DKA have mild hyponatremia. The sodium concentration will fall by approximately 2 mEq/L for each 100 mg/100 mL (5.5 mmol/L) increase in glucose concentration. The "corrected" Na concentration can then be approximated by adding 2 mEq/L to the plasma Na concentration for each 100 mg/dL (5.5 mmol/L) increase above normal in glucose concentration (table 4). (See 'Serum sodium' above and "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis", section on 'Plasma osmolality and sodium'.)

•Potassium deficit – Patients with DKA or HHS have a potassium deficit (approximately 3 to 5 mEq/kg) (table 4). Despite a total body potassium deficit, the serum potassium concentration is usually normal or, in one-third of patients, elevated on admission due to extracellular shifts of potassium in response to hyperosmolality and insulin deficiency. Insulin therapy lowers the potassium concentration and may cause severe hypokalemia, particularly in patients with a normal or low serum potassium concentration at presentation. Thus, careful monitoring and timely administration of potassium supplementation are essential. (See 'Serum potassium' above.)

•Other findings – Serum amylase and lipase levels are elevated in 15 to 25 percent of patients with DKA and, in most cases, do not reflect acute pancreatitis. The diagnosis of pancreatitis should be based upon clinical findings and confirmed by imaging. (See 'Serum amylase and lipase' above.)

Leukocytosis is common in patients with DKA. It may be related to underlying volume depletion, high levels of stress hormones (cortisol, catecholamines, vasopressin and growth hormone), and/or infection.

●Diagnostic criteria

•DKA – DKA is diagnosed when the triad of anion gap metabolic acidosis, hyperglycemia, and ketonemia is present.

-Metabolic acidosis is often the major finding (usually venous pH <7.3 and/or serum bicarbonate <18 mEq/L [18 mmol/L]) (table 1).

-In new-onset diabetes, hyperglycemia is defined as blood glucose ≥200 mg/dL (11.1 mmol/L). In adults with a prior history of diabetes, blood glucose may be normal, near normal (eg, <200 mg/dL), or elevated. The serum glucose concentration is usually less than 800 mg/dL (44 mmol/L) and generally between 350 to 500 mg/dL (19.4 to 27.8 mmol/L).

-Ketosis is defined by the presence of ketones in the blood (blood or serum beta-hydroxybutyrate ≥3 mmol/L]) or urine (urine ketones ≥2+ on dipstick). Blood or serum beta-hydroxybutyrate is the most reliable measure of ketosis.

•HHS – In HHS, there is little or no ketoacid accumulation, the serum glucose concentration is ≥600 mg/dL (33.3 mmol/L) and may exceed 1000 mg/dL (56 mmol/L), the effective Posm is >300 mosmol/kg, and neurologic abnormalities are frequently present (including coma in 25 to 50 percent of cases) (table 1). (See 'Diagnostic criteria' above.)

●Differential diagnosis – Other causes of ketoacidosis include (see 'Differential diagnosis' above):

•Starvation ketosis is characterized by a history of poor caloric intake (<500 Kcal/day). Blood glucose is typically normal.

•Alcoholic ketoacidosis is characterized by a history of chronic alcohol use disorder, with binge drinking, vomiting, and chronic undernutrition. Blood glucose is typically normal.

•Hyperemesis of pregnancy is characterized by excessive vomiting, weight loss, and ketonuria, typically during the first trimester of pregnancy. Bicarbonate and pH may be normal or elevated due to concomitant metabolic alkalosis.

ACKNOWLEDGMENT — 

The UpToDate editorial staff acknowledges Abbas Kitabchi, PhD, MD, FACP, MACE (deceased), who contributed to earlier versions of this topic review.