Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Epidemiology and pathogenesis

INTRODUCTION — Diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state (HHS, also called hyperosmotic hyperglycemic nonketotic state [HHNK]) are two of the most serious acute complications of diabetes. They each represent an extreme in the hyperglycemic spectrum.

The epidemiology and the factors responsible for the metabolic abnormalities of DKA and HHS in adults will be discussed here. The clinical features, evaluation, diagnosis, and treatment of these disorders are discussed separately. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Clinical features, evaluation, and diagnosis" and "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Treatment".)

EPIDEMIOLOGY — Diabetic ketoacidosis (DKA) is characteristically associated with type 1 diabetes. It also occurs in type 2 diabetes under conditions of extreme stress such as serious infection, trauma, cardiovascular or other emergencies, and, less often, as a presenting manifestation of type 2 diabetes, a disorder called ketosis-prone diabetes mellitus. (See "Syndromes of ketosis-prone diabetes mellitus".)

COVID-19 infection has also been associated with greater risks of DKA in both type 1 and type 2 diabetes. (See "COVID-19: Issues related to diabetes mellitus in adults", section on 'Clinical presentations'.)

DKA is more common in young (<65 years) patients, whereas hyperosmolar hyperglycemic state (HHS) most commonly develops in individuals older than 65 years [1,2]. According to the Diabetes Surveillance System of the Centers for Disease Control and Prevention (CDC), overall, age-adjusted DKA hospitalization rates decreased slightly from 2000 to 2009, then reversed direction, steadily increasing from 2009 to 2014 at an average annual rate of 6.3 percent. In-hospital case-fatality rates declined consistently during the study period from 1.1 to 0.4 percent (figure 1) [3].

Population-based data are not available for HHS. The rate of hospital admissions for HHS is lower than the rate for DKA and accounts for less than 1 percent of all primary diabetic admissions [1,4-6]. The mortality rate for patients with HHS is between 10 and 20 percent, which is approximately 10 times higher than that for DKA [7]. The mortality rate for hyperglycemic crisis declined between 1980 and 2009 [8]. Mortality in hyperglycemic crisis is primarily due to the underlying precipitating illness and only rarely to the metabolic complications of hyperglycemia or ketoacidosis [1,9]. The prognosis of hyperglycemic crisis is substantially worse at the extremes of age and in the presence of coma and hypotension [9-12].

PATHOGENESIS

Normal response to hyperglycemia — The hormonal regulation of glucose homeostasis is discussed in detail elsewhere. (See "Pancreatic beta cell function" and "Insulin action" and "Physiologic response to hypoglycemia in healthy individuals and patients with diabetes mellitus".)

Summarized briefly, the extracellular concentration of glucose is primarily regulated by two hormones: insulin and glucagon. As the serum glucose concentration rises after a glucose meal, glucose enters the pancreatic beta cells, initiating a sequence of events leading to insulin release.

Insulin restores normoglycemia by diminishing hepatic glucose production via reductions in both glycogenolysis and gluconeogenesis and by increasing glucose uptake by skeletal muscle and adipose tissue. Insulin-induced inhibition of glucagon secretion contributes to the decline in hepatic glucose production; this effect is mediated by direct inhibition of glucagon secretion and of the glucagon gene in the pancreatic alpha cells [13-15].

Spectrum of metabolic abnormalities — Two hormonal abnormalities are largely responsible for the development of diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state (HHS) in patients with uncontrolled diabetes [13,14,16]:

●Insulin deficiency and/or resistance.

●Glucagon excess, which may result from removal of the normal suppressive effect of insulin.

Although glucagon excess contributes to the development of DKA, it is not essential. As an example, patients with complete pancreatectomies and who have no pancreatic glucagon will develop DKA if insulin is withheld; however, it takes longer for DKA to develop compared with patients with type 1 diabetes [17].

In addition to these primary factors, increased secretion of catecholamines, cortisol, and growth hormone, which oppose the actions of insulin, also contribute to the increases in glucose and ketoacid production (algorithm 1) [1].

The deficiency in insulin (either absolute deficiency, or a relative deficiency caused by excess counterregulatory hormones) is more severe in DKA compared with HHS. Since suppression of lipolysis and ketogenesis is more sensitive to insulin than the inhibition of gluconeogenesis, the residual insulin secretion and its systemic activity in HHS is sufficient to minimize the development of ketoacidosis but not adequate to control hyperglycemia [1]. The increase in glucagon levels and activity is also less in HHS generating a smaller decrease in the insulin/glucagon ratio, which generates a lesser stimulus to ketogenesis [7].

In patients with absolute or relative insulin deficiency, DKA and HHS are usually precipitated by stresses that act in part by increasing the secretion of glucagon, catecholamines, and cortisol (table 1). (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Clinical features, evaluation, and diagnosis", section on 'Precipitating factors'.)

DKA and HHS are two extremes in the spectrum of hyperglycemic crisis, and patients can present anywhere along the continuum of diabetic metabolic derangement (table 2). (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Clinical features, evaluation, and diagnosis", section on 'Diagnostic criteria'.)

Hyperglycemia — The serum glucose concentration in HHS frequently exceeds 1000 mg/dL (56 mmol/L), but in DKA, it is generally below 800 mg/dL (44 mmol/L) and often in the 350 to 450 mg/dL (19.4 to 27.8 mmol/L) range [18]. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Clinical features, evaluation, and diagnosis", section on 'Laboratory findings'.)

At least two factors contribute to less severe hyperglycemia in DKA:

●Patients with DKA often present earlier in the course of their acute disease with symptoms of ketoacidosis (such as shortness of breath, abdominal pain, and nausea and vomiting), rather than late with symptoms due to hyperosmolality.

●Patients with DKA tend to be younger and to have a higher glomerular filtration rate. During the first several years of diabetes, patients may have a supranormal glomerular filtration rate, with levels as great as 50 percent above predicted. As a result, patients with DKA generally have a much greater capacity to excrete glucose than the usually older patients with HHS. This severe degree of glucosuria limits the severity of the hyperglycemia. (See "Diabetic kidney disease: Pathogenesis and epidemiology".)

Although the glucosuria associated with DKA (and HHS) initially minimizes the rise in serum glucose, the osmotic diuresis caused by glucosuria usually leads to volume depletion and a reduction in glomerular filtration rate that limits further glucose excretion [18-21]. This effect is more pronounced in HHS, which results in a higher serum glucose than seen in DKA.

Hormonal alterations in DKA and HHS generate hyperglycemia by their impact on three fundamental processes in glucose metabolism (algorithm 1) [1,22,23]:

●Impaired glucose utilization in peripheral tissues

●Increased hepatic and renal gluconeogenesis

●Increased glycogenolysis

Insulin deficiency and/or resistance in diabetic patients impair peripheral glucose utilization in skeletal muscle. However, decreased glucose utilization alone will produce only postprandial hyperglycemia; glycogenolysis and increased gluconeogenesis are also required for development of the often severe fasting hyperglycemia which occurs in DKA and HHS.

Insulin deficiency and/or resistance promote and accelerate hepatic gluconeogenesis for several reasons [13-15].

Insulin deficiency and/or resistance:

●Increase the delivery of gluconeogenic precursors (glycerol from fat and alanine and other amino acids from muscle) to the liver

●Activate multiple enzymes in the gluconeogenic metabolic pathway

●Increase glucagon levels by removing the inhibitory effect of insulin on both glucagon synthesis and secretion

Oxidation of fatty acids, which are delivered in large amounts to the liver due to lipolysis of fat stores, provides the metabolic energy required to drive gluconeogenesis [16,24,25].

The importance of glucagon in the development of hyperglycemia and ketoacidosis in uncontrolled diabetes has been demonstrated by the following observations:

●After discontinuing insulin in a patient with type 1 diabetes, the rate of rise in serum glucose can be markedly attenuated if glucagon release is prevented by infusing somatostatin [13].

●The magnitude of this effect is illustrated by studies in patients who have undergone total pancreatectomy who make neither insulin nor glucagon. In one report, four such patients and six with type 1 diabetes were fasted after having been maintained on intravenous insulin for 24 hours [17]. After withdrawal of insulin, there was a sharp increase in serum glucagon in the patients with type 1 diabetes. Compared with the pancreatectomized patients, these patients had significantly greater increases in blood glucose (225 versus 139 mg/dL [12.5 versus 7.7 mmol/L]) and blood ketone concentration (4.1 versus 1.8 mmol/L) at 12 hours. (See 'Ketone production' below.)

Ketone production — Both insulin deficiency and glucagon excess contribute to the genesis of DKA [17,24,26]. As noted above, however, glucagon is contributory, but not essential, for DKA to occur. Insulin deficiency and resistance (eg, due to high catecholamine levels) will cause enhanced lipolysis from peripheral fat stores, largely related to increased activity of hormone sensitive lipase, which releases free fatty acids and glycerol. The fatty acids are transported, mainly bound to albumin, to the splanchnic bed and are taken up by hepatocytes. Within the hepatocyte cytoplasm, they are "activated" by linkage of the fatty acid to coenzyme A (CoA), forming acyl-CoA (ie, fatty acid-CoA). The combination of low insulin and increased glucagon activity in the liver cells creates conditions that accelerate the entry of the acyl-CoA into the mitochondria. This transport is mediated by a pair of carnitine palmityl transferase reactions [16,24,26-30].

Within the mitochondria, beta-oxidation splits the fatty acids into multiple two-carbon units of acetyl-CoA. This molecule can have one of three fates:

●Enter the Krebs cycle to be oxidized to carbon dioxide (CO₂) and water (H₂O), thereby creating adenosine triphosphate (ATP)

●Be indirectly exported into the cytoplasm where the acetyl-CoA is used to synthesize fatty acids

●Enter the ketogenic metabolic path to form acetoacetic acid

When fatty acid delivery to the mitochondria is high, then beta-oxidation of the fatty acid usually occurs in a hormonal milieu characterized by low insulin and high glucagon activity. Under these conditions, entry of acetyl-CoA into the Krebs cycle becomes rate limiting, and acetyl-CoA instead is converted to acetoacetic acid. This true ketoacid is the first "ketone body" which forms. The acetoacetic acid may then be reduced to beta-hydroxybutyric acid, which is also an organic acid, or nonenzymatically decarboxylated to acetone, which is not an acid [31]. Ketones provide an alternate water-soluble source of energy when glucose availability is reduced.

The factors responsible for the general absence of ketogenesis in HHS are incompletely understood. However, one important issue is the differential sensitivity of fat metabolism and glucose metabolism to the effects of insulin. Studies in humans have demonstrated that the concentration of insulin required to suppress lipolysis is only one-tenth that required to promote glucose utilization [32]. Thus, less severe insulin deficiency, as occurs in HHS compared with DKA, might be associated with sufficient insulin activity to minimize lipolysis (and therefore ketoacid formation) but not enough to block gluconeogenesis, promote glucose utilization, and thereby prevent the development of hyperglycemia [33]. Less marked elevation of glucagon levels and therefore a higher insulin/glucagon ratio also minimizes ketogenesis [7].

There are several observations compatible with this general hypothesis. DKA tends to occur in patients with type 1 diabetes, who produce little or no insulin. HHS, in comparison, is found primarily in older patients with type 2 diabetes, who have decreased, but not absent, insulin effect [16,23,34]. However, this distinction is not absolute, since DKA can occur in patients with type 2 diabetes (see "Syndromes of ketosis-prone diabetes mellitus").

Anion gap metabolic acidosis — DKA typically presents as an elevated anion gap metabolic acidosis. This is caused by the production and accumulation of beta-hydroxybutyric and acetoacetic acids. The anion gap is calculated by subtracting the serum concentrations of chloride and bicarbonate from the concentration of sodium [35]:

By convention, the measured (reported) sodium concentration (not the sodium concentration corrected for hyperglycemia) should be used for the calculation of the anion gap.

The severity of the metabolic acidosis and the increase of the anion gap are dependent upon a number of factors:

●The rate and duration of ketoacid production

●The rate of metabolism of the ketoacids

●The rate of loss of ketoacid anions in the urine

●The volume of distribution of the ketoacid anions

●The rate of renal net acid excretion

The rate of ketoacid anion excretion depends upon the patient's volume status, renal function, and the degree to which glomerular filtration is maintained. Patients with relatively preserved extracellular fluid (ECF) volume and higher renal function can excrete large quantities of ketoacids (as much as 30 percent of the ketoacid load), and thereby minimize the elevation in anion gap [36]. The magnitude of this effect was illustrated in a study of patients with DKA; ketone production averaged 51 mEq/hour, while net acid excretion with the ketoacid anions averaged 15 mEq/hour or 30 percent of the ketoacid load [37]. The conversion of acetoacetic acid to acetone can neutralize another 15 to 25 percent of the acid load [37,38]. Rarely, patients excrete ketoacids so efficiently that they present with only a small, or no, elevation in serum anion gap (they have a pure non-gap, or hyperchloremic, metabolic acidosis) [36,39-41].

Renal excretion of ketoacid anions increases when patients are treated with intravenous isotonic fluids to correct the hypovolemia. Excretion of ketoacid anions reduces the anion gap, but to the extent that they are excreted as sodium and potassium salts, the severity of the systemic acidemia is unchanged [36,39,42,43]. Sodium or potassium ketoanion salts represent both "decomposed" bicarbonate and also "potential bicarbonate." When the ketoacids are generated, the protons mainly combine with bicarbonate to form CO₂. Thus, in essence, bicarbonate anions in the serum are replaced with ketoacid anions ("decomposed bicarbonate"). If the ketoacids are metabolized, then a bicarbonate anion is regenerated ("potential bicarbonate"). Therefore, the urinary loss of ketoacid anions as sodium or potassium salts represents the loss of "potential bicarbonate." Furthermore, excretion of beta-hydroxybutyrate and acetoacetate as sodium and potassium salts will reduce the anion gap and convert the anion gap acidosis to a hyperchloremic, or non-gap, acidosis. Thus, almost all patients with DKA who have relatively intact renal function will develop some degree of hyperchloremic (normal anion gap) metabolic acidosis when they are treated with isotonic saline and insulin, due to the urinary loss of potential bicarbonate. These principles are discussed in detail elsewhere. (See "The delta anion gap/delta HCO3 ratio in patients with a high anion gap metabolic acidosis", section on 'The delta AG/delta HCO3 in ketoacidosis'.)

Studies have demonstrated that a portion of the anion gap acidosis in most patients with DKA is due to yet another organic acid: D-lactic acid. A small, but clinically significant, fraction of ketoacids, acetone, and dihydroxyacetone phosphate (a product of glycolysis) can each be converted to D-lactic acid [44]. D-lactic acid can account for as much as 8 to 10 mEq/L of the anion gap elevation and bicarbonate reduction in patients with severe DKA [44]. (See "The delta anion gap/delta HCO3 ratio in patients with a high anion gap metabolic acidosis", section on 'D-lactic acidosis and toluene inhalation'.)

Plasma osmolality and sodium — Plasma osmolality is always elevated in patients with HHS but less so with DKA (table 2). The increase in plasma osmolality created by hyperglycemia pulls water out of the cells, expands the ECF, and thereby reduces the plasma sodium (Na) concentration. If a patient with normal serum electrolytes (Na = 140 mEq/L) rapidly developed a glucose concentration of 1000 mg/100 mL and no urine was made, then that patient's serum sodium concentration would fall to value between 119 and 126 mEq/L and the osmolality would increase to a level between 294 and 308 mosmol/L. However, the osmolality usually increases to a greater degree because a large volume of relatively electrolyte-deficient urine is excreted during the evolution of the hyperglycemic state. The loss of this electrolyte-free water further raises the osmolality [18]. In patients with ketoacidosis, high plasma acetone levels also contribute to the elevated osmolality.

The measured serum sodium concentration in uncontrolled diabetes mellitus is affected by the interaction of multiple factors, some that lower and others that raise it:

●Hyperglycemia raises the ECF osmolality (and tonicity) and shifts water from the intracellular fluid (ICF) space to the ECF space. Expansion of the ECF dilutes serum sodium and reduces its concentration.

●Glucosuria generates an osmotic diuresis which causes the excretion of sodium and potassium salts, and water. The sum of the concentrations of sodium and potassium salts in the urine is generally much lower than the sodium concentration in blood. This indicates that electrolyte-free water is being excreted in the urine. Loss of electrolyte-free water will raise the serum sodium concentration and plasma osmolality.

●The variable intake of water and the loss of free water in vomitus or nasogastric suction will also impact the serum sodium concentration and plasma osmolality.

Physiologic calculations suggest that, in the absence of urine losses, the serum sodium concentration should fall by approximately 1.6 mEq/L for each 100 mg/100 mL (5.5 mmol/L) increase in glucose concentration [45]. However, when hyperglycemia was induced in six healthy subjects by the administration of somatostatin (to block endogenous insulin secretion) and the infusion of hypertonic dextrose solution, the relationship between the fall in sodium concentration and the increase in glucose concentration was not linear and the reduction in sodium was greater than had been previously reported [46]. These authors suggested that the sodium should fall by approximately 2.4 mEq/L for each 100 mg/100 mL (5.5 mmol/L) increase in glucose concentration. A number of prior and subsequent studies reported a fall in sodium between 1.6 and 2.4 mEq/L for each 100 mg/100 mL increase in glucose [47]. It must be emphasized that these ratios are only very approximate estimates of potential water deficits (or less commonly, excess). However, they do provide a rough estimate of how much the serum sodium concentration will rise as the hyperglycemia is corrected. Because there is so much variation in the literature and so many variables affect this relationship, we recommend the use of a simple ratio of a 2 mEq/L decline in sodium for each 100 mg/100 mL (5.5 mmol/L) increase in glucose concentration above the normal range.

Potassium — Patients presenting with DKA or HHS have a potassium deficit that averages 300 to 600 mEq (table 3) [48-50]. A number of factors contribute to this deficit, particularly increased urinary losses due both to the glucose osmotic diuresis and the excretion of potassium ketoacid anion salts (the ketoacid anions are filtered primarily as sodium salts, but some of the sodium is reabsorbed in the distal renal tubule in exchange for potassium as a result of volume contraction-related secondary hyperaldosteronism). Gastrointestinal losses and the loss of potassium from the cells due to glycogenolysis and proteolysis also may play a contributory role.

Despite these large total body potassium deficits, the serum potassium concentration is usually normal or, in one-third of patients, elevated on admission [18,48,51]. This is mainly due to hyperosmolality and insulin deficiency [22,48,52]. (See "Causes and evaluation of hyperkalemia in adults".)

●The rise in plasma osmolality causes osmotic water movement out of the cells. Potassium also moves into the ECF as a result of at least two mechanisms [45]:

•The contraction of the ICF space increases intracellular potassium concentration and favors passive potassium exit through potassium channels in the cell membrane

•The frictional forces between solvent (water) and solute result in potassium being carried across the cell membrane with water (this process is called solvent drag)

●Insulin normally promotes potassium uptake by the cells. Therefore, insulin deficiency contributes to elevated serum potassium levels.

Acidemia per se plays a small role in the shift of potassium from the ICF to the ECF in patients with DKA. Although a transcellular exchange of potassium with hydrogen ions does increase serum potassium in some forms of "inorganic" metabolic acidosis [53], it is less important in organic metabolic acidoses such as ketoacidosis or lactic acidosis [48,53-55]. The important role of hyperosmolality and insulin deficiency is illustrated by the observation that hyperkalemia also often occurs in HHS, despite the absence of acidosis [18]. (See "Potassium balance in acid-base disorders", section on 'Metabolic acidosis'.)

Inflammation — Hyperglycemic crises are proinflammatory states that lead to generation of reactive oxygen species and oxidative stress. Studies have shown elevated proinflammatory cytokines including tumor necrosis factor-alpha and interleukin (IL)-1B, IL-6, and IL-8. Lipid peroxidation markers, as well as plasminogen activator inhibitor-1 and C-reactive protein (CRP), are also increased [56]. Proinflammatory factors returned to near-normal levels within 24 hours of insulin therapy and resolution of hyperglycemia. The proinflammatory state in DKA results in in vivo activation of T-lymphocytes with de novo emergence of growth factor receptors [57].

A variety of eicosanoids, including prostaglandins, are involved in the pathogenesis of diabetes mellitus and its complications [58]. Some are protective and others accelerate organ dysfunction, including pancreatic beta-cell destruction. Prostaglandins accumulate during DKA, increase in the circulation before elevation of epinephrine, and return promptly to normal levels with insulin therapy [59,60].

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 topic (see "Patient education: Hyperosmolar hyperglycemic state (The Basics)")

SUMMARY

●Epidemiology – Diabetic ketoacidosis (DKA) is more common in younger patients with type 1 diabetes, though it can occur in type 2 diabetes. HHS occurs less frequently and is associated with higher mortality, at least in part due to underlying comorbidity. (See 'Epidemiology' above.)

●Metabolic abnormalities – DKA and hyperosmolar hyperglycemic state (HHS) are part of a spectrum, representing the metabolic consequences of insulin deficiency, glucagon excess, and counterregulatory hormonal responses to stressful triggers in patients with diabetes (table 2). (See 'Spectrum of metabolic abnormalities' above.)

•Hyperglycemia – Glucose concentrations are typically lower (usually <800 mg/dL [44 mmol/L]) in DKA compared with HHS (usually >1000 mg/dL [56 mmol/L]). Patients with DKA present earlier because of symptoms and generally can excrete glucose more effectively than older patients with HHS. (See 'Hyperglycemia' above.)

Hyperglycemia results from impaired glucose utilization, increased gluconeogenesis, and increased glycogenolysis (algorithm 1). Gluconeogenesis results from delivery of precursors to the liver from breakdown of fat and muscle and is promoted by insulin deficiency and glucagon excess. Glycogenolysis is stimulated by catecholamines and a high glucagon-to-insulin ratio. Osmotic diuresis further contributes to elevated blood glucose. (See 'Hyperglycemia' above.)

•Ketone production – Ketoacidosis results from lipolysis, with synthesis of ketones from free fatty acids in the liver mitochondria. Insulin levels in HHS are insufficient to allow appropriate glucose utilization but are adequate to prevent lipolysis and subsequent ketogenesis. (See 'Ketone production' above.)

•Metabolic acidosis – The elevated anion gap metabolic acidosis characteristic of DKA is caused by the production and accumulation of beta-hydroxybutyric and acetoacetic acids. The severity of the metabolic acidosis is dependent upon a number of factors, including the rate and duration of ketoacid production, the rate of metabolism of the ketoacids, and the rate of acid excretion in the urine. (See 'Anion gap metabolic acidosis' above.)

•Plasma osmolality and serum sodium – Plasma osmolality is always elevated in patients with HHS but less so with DKA (table 2). The marked hyperosmolality seen in HHS is only in part due to the rise in serum glucose. It is also due to the glucose osmotic diuresis that causes urine electrolyte-free water loss. These factors also influence the measured serum sodium concentration, which is variable because of the interaction of multiple factors, some that lower and others that raise it. (See 'Plasma osmolality and sodium' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Abbas Kitabchi, PhD, MD, FACP, MACE, now deceased, who contributed to an earlier version of this topic review.