Diabetic Ketoacidosis

Diabetic ketoacidosis (DKA) is a potentially life-threatening complication of diabetes mellitus in which insufficient insulin activity leads to excessive production of ketone bodies and metabolic acidosis. Blood glucose is usually elevated, although ketoacidosis can also occur without marked hyperglycemia. DKA is most commonly associated with type 1 diabetes, but it also occurs in type 2 diabetes and may be the first recognized manifestation of diabetes. (ncbi.nlm.nih.gov)

The metabolic disturbance combines increased fat breakdown and ketone formation with urinary losses of water and electrolytes. Its manifestations include thirst, frequent urination, dehydration, vomiting, abdominal pain, deep breathing, and, in severe cases, impaired consciousness. Hospital treatment addresses fluid depletion, insulin deficiency, electrolyte imbalance, and the precipitating illness. DKA is important both as an acute medical emergency and as a marker of difficulties in diabetes recognition, treatment access, or continuing care. Its clinical assessment distinguishes it from other causes of ketoacidosis and from the related hyperosmolar hyperglycemic state. (ncbi.nlm.nih.gov)

Metabolic Mechanisms

Absolute or relative insulin deficiency removes restraints on hepatic fuel production and adipose-tissue fat breakdown. Absolute deficiency means that little or no insulin is available; relative deficiency means that available insulin is insufficient for metabolic requirements, particularly when insulin resistance increases. Concurrent elevations of glucagon, catecholamines such as epinephrine, cortisol, and growth hormone amplify this imbalance. Glucagon promotes hepatic glucose production and ketogenesis, catecholamines stimulate glycogen breakdown and lipolysis, and cortisol and growth hormone oppose insulin action. The resulting combination increases fuel release while reducing insulin-dependent glucose utilization. These mechanisms can operate in both type 1 and type 2 diabetes. (ncbi.nlm.nih.gov)

In adipose tissue, diminished insulin action and increased counterregulatory signaling accelerate lipolysis, releasing free fatty acids and glycerol from stored triglycerides. Fatty acids reach the liver and undergo β-oxidation within mitochondria, generating acetyl-CoA. Hepatic ketogenesis converts acetyl-CoA through acetoacetyl-CoA and hydroxymethylglutaryl-CoA into acetoacetate. Acetoacetate can be reduced to β-hydroxybutyrate or undergo spontaneous decarboxylation to acetone and carbon dioxide. Acetoacetate and β-hydroxybutyrate are the principal acid-associated ketone bodies in ketoacidosis; despite its conventional grouping with ketones, β-hydroxybutyrate is chemically a hydroxy acid. Acetone is a neutral, volatile compound rather than an acid-producing component, and its exhalation accounts for the characteristic breath odor. (ncbi.nlm.nih.gov)

Hyperglycemia develops through increased glucose production and reduced peripheral disposal. Hepatic gluconeogenesis uses substrates including glycerol, lactate, and glucogenic amino acids, while glycogenolysis mobilizes glucose from glycogen stores. Reduced insulin-mediated uptake, especially in skeletal muscle and adipose tissue, further raises circulating glucose. Once the filtered glucose load exceeds renal tubular reabsorptive capacity, glucose remains in the urine and retains water within the tubular fluid, producing osmotic diuresis. This causes losses of water and electrolytes, including sodium and potassium. Progressive volume depletion reduces renal perfusion and glomerular filtration, limiting glucose and ketone clearance and thereby reinforcing the metabolic disturbance. (ncbi.nlm.nih.gov)

When ketone-associated acid production exceeds tissue utilization and renal elimination, bicarbonate buffering consumes circulating bicarbonate and blood pH falls. The buffering reaction converts hydrogen ions and bicarbonate into carbon dioxide and water. Falling bicarbonate and pH stimulate ventilation; increased carbon dioxide excretion partially compensates for the metabolic acidosis, sometimes producing deep Kussmaul respiration. This respiratory response does not remove the underlying metabolic acid load. Poor tissue perfusion can additionally increase lactate production, while reduced kidney function limits acid excretion, compounding the disturbance. (diabetesjournals.org)

Potassium concentration in blood does not reliably represent total-body potassium stores. The 2024 adult consensus report describes typical depletion of approximately $3–6\ \mathrm{mmol/kg}$, attributable to osmotic urinary losses, vomiting, and volume-depletion-associated hyperaldosteronism. Nevertheless, initial serum potassium is often normal or elevated: insulin deficiency reduces cellular potassium uptake, and the accompanying metabolic and osmotic disturbances favor redistribution into extracellular fluid. Thus, substantial depletion can coexist with an apparently adequate blood concentration. Restoration of insulin action moves potassium back into cells, while improved renal perfusion increases urinary excretion, exposing the underlying deficit and allowing serum potassium to fall rapidly. (pmc.ncbi.nlm.nih.gov)

Epidemiology and Population Burden

Diabetic ketoacidosis (DKA) occurs in children and adults, both at the first recognition of diabetes mellitus and after diabetes has become established. These settings require different epidemiological measures: DKA at diagnosis is usually expressed as a percentage of newly diagnosed patients, whereas subsequent DKA is measured as episodes per person-year or hospital admissions per population with diabetes. DKA is more frequent in type 1 diabetes, but neither childhood nor type 1 diabetes is an exclusive setting. In the population-based US SEARCH for Diabetes in Youth study, covering diagnoses before age 20 during 2002–2004, DKA was present at diagnosis in 29.4% of participants with type 1 diabetes and 9.7% with type 2 diabetes. These percentages describe risk within each diabetes category, not the proportions of all DKA cases attributable to either category. (cdc.gov)

The burden at childhood diabetes diagnosis has remained substantial. Among 7,612 newly diagnosed participants younger than 20 in SEARCH during 2010–2016, the proportion with DKA within one month of type 1 diabetes diagnosis increased from 35.3% in 2010 to 40.6% in 2016. Across the study period, the proportion was highest among children aged 0–4 years, at 46.7%, compared with 29.7% among those aged 15–19. These are proportions of incident diabetes cases presenting with DKA, rather than annual DKA incidence rates among all children or among everyone living with diabetes. (pmc.ncbi.nlm.nih.gov)

Hospital-discharge data capture a different component of population burden. Stephen R. Benoit and colleagues’ US analysis of the National Inpatient Sample and National Health Interview Survey found that the age-adjusted DKA hospitalization rate declined slightly during 2000–2009, then increased by 54.9%, from 19.5 to 30.2 admissions per 1,000 people with diagnosed diabetes during 2009–2014. The estimated admission count rose from 141,704 to 188,950 over those five years. The analysis included all ages; in 2014, rates were highest below age 45, at 44.3 per 1,000 people with diabetes, compared with 1.6 at ages 65–74 and 1.4 at ages 75 or older. These figures count hospitalizations, not unique patients: repeated admissions by one person contribute repeatedly to the numerator. The denominator did not distinguish diabetes type, so the results cannot establish separate hospitalization rates for type 1 and type 2 diabetes. (cdc.gov)

International differences are evident even within comparatively well-resourced health systems. A retrospective collaboration covering 59,000 children aged six months to 14.9 years with newly diagnosed type 1 diabetes in 13 countries during 2006–2016 found an overall age- and sex-adjusted DKA prevalence of 29.9%. Prevalence was lowest in Sweden and Denmark and highest in Luxembourg and Italy; it increased significantly in Australia, Germany, and the participating US population, but decreased in Italy. Some datasets represented whole countries, whereas others covered selected regions or clinical networks. Consequently, these comparisons document geographic variation but do not isolate the causal effects of financing arrangements or access policies; differences in case ascertainment, population composition, diabetes awareness, and diagnostic delay also matter. (pmc.ncbi.nlm.nih.gov)

Socioeconomic disadvantage is associated with DKA both at diagnosis and in established diabetes. SEARCH found associations at diagnosis with lower family income, less favorable insurance coverage, and lower parental education after adjustment for demographic and diabetes-related variables. Insulin affordability provides a more specific documented pathway in established disease. In Lori Randall and colleagues’ study of 164 adults admitted to Grady Hospital in Atlanta during July 2007–August 2010, 73 presented with first-time DKA and 91 with recurrent DKA; insulin discontinuation was identified as the precipitant in 68%. Among participants who stopped insulin, 27% reported lacking money to purchase it. This single-hospital interview study documents reported financial barriers preceding treatment interruption, but does not estimate their national prevalence or demonstrate that insulin prices caused the national increase in DKA admissions. (pubmed.ncbi.nlm.nih.gov)

Precipitating Illnesses and Treatment Interruptions

Infection can precipitate diabetic ketoacidosis (DKA) by increasing insulin requirements beyond the amount available from endogenous secretion or administered treatment. Pneumonia and urinary-tract infection are prominent examples. The 2024 international adult consensus report identified infection in 14–58% of DKA cases across the studies reviewed; this range describes heterogeneous clinical populations rather than a single population-wide rate. Infection raises counterregulatory hormones, including glucagon, catecholamines, cortisol, and growth hormone, while inflammatory cytokines contribute to insulin resistance. These changes increase hepatic glucose production and reduce the effectiveness of existing insulin availability in suppressing fat breakdown and ketone production. Thus, DKA can develop during infection even without complete interruption of insulin administration. (pmc.ncbi.nlm.nih.gov)

Insufficient insulin delivery is another major precipitant. It includes missed administration, an inadequate prescribed regimen, and interruptions caused by difficulty obtaining insulin or supplies. These circumstances are not interchangeable: an episode attributed to “insulin omission” does not by itself establish whether the underlying cause was financial, logistical, psychological, or related to illness. A previously adequate insulin regimen may also become insufficient during acute physiological stress. In people with marked insulin deficiency, interruption removes the restraint on hepatic ketone production; reduced food intake does not eliminate this metabolic requirement. DKA may additionally be the first recognized manifestation of previously undiagnosed type 1 diabetes or type 2 diabetes, rather than the consequence of an interruption in established treatment. (pmc.ncbi.nlm.nih.gov)

Insulin-pump and infusion-set failures provide a distinct mechanism of interrupted delivery. Blockage, leakage, or displacement of the infusion cannula can prevent insulin from reaching subcutaneous tissue despite a programmed infusion. Conventional pump therapy uses rapid-acting insulin and maintains a smaller subcutaneous insulin depot than injection regimens containing long-acting insulin. Consequently, interruption of the continuous infusion can lead relatively quickly to insulin deficiency, hyperglycemia, and ketogenesis. Device-related interruption is therefore mechanistically different from an intentional change in treatment, and identification of pump use alone does not establish the cause of an episode. (diabetesjournals.org)

Established noninfectious precipitants include myocardial infarction, stroke, pancreatitis, surgery, and trauma. These conditions impose counterregulatory stress and can coincide with fasting, vomiting, dehydration, or disruption of usual treatment. The metabolic consequences are particularly important when pancreatic insulin reserve is limited. An acute illness and inadequate insulin delivery can therefore contribute simultaneously rather than functioning as mutually exclusive explanations. (ncbi.nlm.nih.gov)

Medication-associated DKA includes presentations linked to sodium–glucose cotransporter 2 (SGLT2) inhibitors. These drugs inhibit renal glucose reabsorption, increasing urinary glucose loss and lowering blood glucose independently of insulin. This glucose-lowering effect can coexist with insufficient insulin action and continued ketogenesis; reduced insulin availability relative to glucagon provides a proposed mechanism for enhanced fat breakdown and hepatic ketone production. Reported precipitating circumstances include insulin-dose reduction, acute febrile illness, reduced caloric intake, and surgery, although not every reported episode has an identifiable additional trigger. Blood glucose may consequently be much lower than expected in conventional DKA. Glucocorticoids, by contrast, primarily raise glucose through increased gluconeogenesis and insulin resistance; severe hyperglycemia can progress to DKA when insulin availability cannot meet the increased requirement. Their mechanism does not depend on drug-induced urinary glucose loss. (diabetesjournals.org)

Clinical Presentation

Diabetic ketoacidosis (DKA) commonly develops over hours to a few days, with increasing thirst and urine output followed by more pronounced systemic illness. Polyuria denotes excessive urination, while polydipsia denotes excessive thirst; both reflect the fluid losses accompanying glucose-driven osmotic diuresis. Unintentional weight loss, weakness, fatigue, and malaise may precede presentation, particularly when DKA is the first manifestation of diabetes mellitus. As fluid depletion and acidosis progress, these relatively nonspecific symptoms may give way to vomiting, conspicuous respiratory changes, and impaired consciousness. The sequence is not fixed, and not every patient exhibits every feature. (ncbi.nlm.nih.gov)

Nausea, vomiting, and abdominal pain are prominent manifestations. Pain can be sufficiently severe to resemble an acute abdomen, creating overlap with gastrointestinal illness or a surgical emergency. In a prospective Atlanta hospital study published in 2002, Guillermo Umpierrez and Amado X. Freire examined 189 DKA episodes collected during a 13-month period beginning in October 1995; abdominal pain occurred in 86 episodes, or 46%. Pain was associated with the severity of metabolic acidosis rather than the degree of hyperglycemia or dehydration. Nevertheless, a concurrent disorder can account for abdominal symptoms: surgical cases in that study included appendicitis and cholecystitis. (pubmed.ncbi.nlm.nih.gov)

The distinction between pain caused by metabolic decompensation and pain associated with a precipitating illness is therefore clinically important. DKA-associated discomfort commonly subsides as dehydration and acidosis resolve. The 2024 adult consensus report identifies persistent abdominal pain after metabolic recovery, or pain occurring without severe acidosis, as circumstances requiring further evaluation for another cause. Vomiting also compounds fluid losses, so gastrointestinal symptoms may intensify the circulatory manifestations of the episode. (doi.org)

Kussmaul respiration is conspicuously deep, often rapid breathing produced by respiratory compensation for metabolic acidosis: increased ventilation lowers carbon dioxide and partially offsets the fall in blood pH. A fruity or solvent-like breath odor results from exhaled acetone, a volatile product of acetoacetate breakdown. The odor and breathing pattern are distinct observations, and neither is invariably present; their absence does not exclude DKA. An observational series of 71 patients presenting to a tertiary hospital in Sukkur, Pakistan, during August 2019–February 2020 documented Kussmaul breathing in only a minority of presentations. (pmc.ncbi.nlm.nih.gov)

Examination commonly reveals signs of dehydration and reduced circulating volume, including diminished skin turgor, tachycardia, and hypotension. Reduced peripheral perfusion accompanies more substantial volume depletion. However, preserved blood pressure does not necessarily indicate preserved fluid balance, because hyperglycemia-related osmotic shifts can temporarily support intravascular volume. Many patients remain alert, whereas severe presentations may involve confusion, drowsiness, stupor, or coma. Altered consciousness is consequently a marker of more serious illness, not a prerequisite for the clinical presentation of DKA. (diabetesjournals.org)

Diagnostic Criteria and Laboratory Interpretation

The 2024 adult consensus report requires all three diagnostic components: diabetes or hyperglycemia, ketosis, and metabolic acidosis. The first is satisfied by plasma glucose ≥200 mg/dL (11.1 mmol/L) or a prior history of diabetes mellitus, irrespective of presenting glucose. The second requires blood β-hydroxybutyrate ≥3.0 mmol/L or urine ketones ≥2+ on a semiquantitative strip. The third requires venous pH <7.3 and/or bicarbonate <18 mmol/L. Thus, neither hyperglycemia alone nor a positive ketone test establishes DKA; conversely, glucose below 200 mg/dL does not exclude it in someone with established diabetes. (diabetesjournals.org)

Direct blood measurement of β-hydroxybutyrate, the predominant circulating ketoacid in DKA, provides a quantitative assessment of ketonemia. Laboratory assays use serum or plasma, while handheld point-of-care instruments use capillary blood. Nitroprusside-based urine strips primarily detect acetoacetate and do not detect β-hydroxybutyrate; some nitroprusside formulations also detect acetone. These tests therefore measure a different fraction of the ketone pool rather than an interchangeable substitute for blood β-hydroxybutyrate. The 2023 laboratory guideline by David B. Sacks and colleagues recommends specific blood β-hydroxybutyrate measurement for diagnosis and permits its use during monitoring. (diabetesjournals.org)

The distinction becomes particularly important when interpreting serial results. Early in DKA, predominance of β-hydroxybutyrate can produce less impressive nitroprusside positivity than the degree of ketoacidosis would suggest. During recovery, β-hydroxybutyrate is converted back to acetoacetate, so urine ketones may remain positive or appear to increase despite biochemical improvement. Persistently positive urine ketones therefore do not necessarily indicate continuing ketoacid production. Analytical interference also matters: sulfhydryl-containing drugs can produce false-positive nitroprusside results, while deteriorated test strips and inappropriate specimen storage can cause false-negative readings. (diabetesjournals.org)

The initial laboratory assessment combines glucose and ketones with venous blood gases, electrolytes, and renal-function measurements. Venous pH establishes acidemia; bicarbonate characterizes the metabolic component. Sodium and chloride support interpretation of osmotic and acid–base disturbances, while potassium identifies an immediately important electrolyte abnormality. A normal or elevated serum potassium does not establish normal total-body potassium stores, because extracellular redistribution can mask substantial losses. Creatinine and urea help characterize impaired kidney function associated with volume depletion or pre-existing renal disease. Measured or calculated serum osmolality assesses the additional hyperosmolar burden and helps identify overlap with hyperosmolar hyperglycemic state; it is not one of the three required DKA components. (diabetesjournals.org)

The anion gap is an adjunct, not an indispensable diagnostic criterion. Using measured serum concentrations in mmol/L and omitting potassium, it is calculated as

A value >12 mmol/L supports high-anion-gap metabolic acidosis but does not identify its cause. Approximately one-third of adults presenting with DKA have mixed acid–base disturbances; vomiting-related metabolic alkalosis can partially offset ketoacidosis. Hyperchloremic acidosis during recovery can also dissociate bicarbonate from ketone clearance. Where ketone testing is unavailable, the gap retains diagnostic utility, but cannot directly demonstrate or quantify ketonemia. (diabetesjournals.org)

Severity, Atypical Forms, and Differential Diagnosis

The 2024 adult consensus classification grades DKA by ketone concentration, metabolic acidosis, and mental status—not by blood glucose alone. Its categories are shown below; not every variable must fall within the same category, and classification requires clinical judgment. A lower glucose concentration does not establish that an episode is mild. (pmc.ncbi.nlm.nih.gov)

The classification lists pH or bicarbonate as alternative measures of acidosis severity, rather than requiring both thresholds simultaneously. (pmc.ncbi.nlm.nih.gov)

Euglycemic DKA is defined in this framework as DKA with plasma glucose <200 mg/dL (11.1 mmol/L), while the ketosis and metabolic-acidosis criteria remain satisfied. Associated circumstances include reduced food intake, pregnancy, prior insulin administration, and exposure to sodium–glucose cotransporter 2 inhibitors (SGLT2 inhibitors). Thus, “euglycemic” describes the glucose concentration, not the severity of the acid–base disturbance. (profketandhatariya.com) SGLT2 inhibition increases urinary glucose loss by reducing renal glucose reabsorption, allowing substantial ketogenesis to occur without marked hyperglycemia; reduced carbohydrate availability can compound this presentation. (diabetesjournals.org)

Hyperosmolar hyperglycemic state (HHS) differs from DKA chiefly in the predominance of hyperglycemia and hyperosmolality without significant ketoacidosis. The adult criteria specify glucose ≥600 mg/dL (33.3 mmol/L), effective serum osmolality >300 mOsm/kg or total serum osmolality >320 mOsm/kg, β-hydroxybutyrate <3.0 mmol/L, and pH ≥7.3 with bicarbonate ≥15 mmol/L. Mild ketonemia therefore does not automatically exclude HHS. (diabetesjournals.org) DKA and HHS nevertheless form an overlapping spectrum: mixed DKA–HHS combines substantial hyperosmolality with ketonemia and metabolic acidosis. HHS usually evolves over several days to weeks, whereas DKA commonly develops over hours to one or two days; neither the tempo nor the glucose concentration alone reliably separates them. (ncbi.nlm.nih.gov)

Starvation ketosis occurs in prolonged fasting or malnutrition, generally with normal or low glucose and comparatively mild acidosis. Alcoholic ketoacidosis can produce marked ketonemia and high-anion-gap metabolic acidosis despite normal or low glucose; its interpretation depends on the alcohol and nutritional history rather than on glucose alone. (ncbi.nlm.nih.gov) A 2023 Japanese hospital case series illustrates the biochemical overlap: 68 of 83 patients evaluated for suspected alcoholic ketoacidosis had β-hydroxybutyrate concentrations above 3.0 mmol/L, the ketone threshold also used in DKA. Vomiting and nausea each occurred in 71%. Ketone concentration therefore establishes the presence of ketosis but does not, by itself, identify its diabetic origin. (pubmed.ncbi.nlm.nih.gov)

Lactic acidosis is distinguished by lactate accumulation rather than ketones as the principal measured acid burden; relevant contexts include sepsis and impaired tissue perfusion associated with dehydration. (ncbi.nlm.nih.gov) Methanol and ethylene glycol ingestion also cause high-anion-gap metabolic acidosis. An increased osmolal gap can support investigation for toxic alcohol exposure, but a normal gap cannot exclude poisoning, especially after the parent alcohol has been metabolized. (pubmed.ncbi.nlm.nih.gov) Salicylate poisoning commonly produces respiratory alkalosis together with metabolic acidosis, sometimes leaving pH near normal. Exposure history, substance-specific blood measurements, lactate, and ketone testing distinguish these biochemical patterns more reliably than an elevated anion gap alone. (acmt.net)

Hospital Treatment and Biochemical Resolution

Hospital assessment establishes circulatory stability, consciousness, fluid losses, biochemical severity, and the precipitating illness. Adults with uncomplicated mild or moderate DKA may receive treatment in an emergency department or step-down unit where close nursing observation is available; severe DKA, altered consciousness, or critical illness such as sepsis generally requires intensive care. Reassessment includes blood pressure, pulse, fluid balance, and mental status. The 2024 adult consensus specifies bedside glucose measurements every 1–2 hours and blood sampling every 4 hours for electrolytes, phosphate, creatinine, β-hydroxybutyrate, and venous pH until resolution. (pmc.ncbi.nlm.nih.gov)

Intravenous fluids restore circulating volume and tissue perfusion, while insulin corrects hyperglycemia and suppresses ketone production. Isotonic saline and balanced crystalloids, such as lactated Ringer’s solution, are used for volume replacement. Saline’s chloride load can produce a separate metabolic acidosis during recovery. In a 2020 subgroup analysis by Wesley H. Self and colleagues involving 172 adults from two cluster-randomized trials at a United States academic center, median time to DKA resolution was 13.0 hours with balanced crystalloids and 16.9 hours with saline; the study used the resolution criteria then in effect. (doi.org)

Continuous intravenous short-acting insulin is administered through validated clinician-directed protocols, with adjustment according to biochemical response. Selected uncomplicated mild or moderate cases can instead receive subcutaneous rapid-acting insulin at closely supervised intervals, provided fluid replacement and frequent monitoring remain available; this approach is not intended for severe or complicated DKA. (doi.org) Glucose commonly falls below 250 mg/dL (13.9 mmol/L) before ketoacidosis resolves. Added dextrose prevents hypoglycemia while allowing insulin to continue clearing ketones. Potassium replacement addresses depleted body stores and the fall in serum potassium as insulin shifts potassium into cells. Marked hypokalemia requires correction before insulin proceeds. (pmc.ncbi.nlm.nih.gov)

Bicarbonate and phosphate are not routine components of treatment. A 1986 randomized study by Lawrence R. Morris, Mary Beth Murphy, and Abbas E. Kitabchi included 21 adults with initial arterial pH of 6.9–7.14 and found no significant improvement in metabolic recovery with bicarbonate. (pubmed.ncbi.nlm.nih.gov) The 2024 consensus restricts consideration of bicarbonate to severe acidemia, specifically pH below 7.0. Phosphate replacement is reserved for clinically important deficiency, particularly low phosphate accompanied by muscle weakness or respiratory or cardiac compromise. (pmc.ncbi.nlm.nih.gov)

Biochemical resolution requires plasma β-hydroxybutyrate below 0.6 mmol/L together with either venous pH at least 7.3 or bicarbonate at least 18 mmol/L; glucose should ideally also be below 200 mg/dL (11.1 mmol/L). Urine ketones can remain positive as β-hydroxybutyrate converts to acetoacetate, while chloride-related acidosis makes the anion gap an unreliable recovery endpoint. (pmc.ncbi.nlm.nih.gov) Transition to maintenance subcutaneous insulin therefore includes treatment overlap rather than abrupt cessation of intravenous insulin. The American Diabetes Association’s 2026 hospital standards specify basal insulin administration 2–4 hours before stopping the infusion, preventing a gap in insulin action and recurrent ketoacidosis. Subsequent basal, mealtime, and correction insulin requirements are individualized using prior treatment, nutritional intake, renal function, and hypoglycemia risk. (doi.org)

Children, Pregnancy, and Other Special Clinical Settings

The 2022 International Society for Pediatric and Adolescent Diabetes (ISPAD) guideline defines pediatric DKA by blood glucose above 11 mmol/L (approximately 200 mg/dL), venous $pH<7.3$ or bicarbonate below 18 mmol/L, and β-hydroxybutyrate at least 3 mmol/L or moderate-to-large ketonuria. Pediatric fluid assessment incorporates body weight, age-related maintenance requirements, circulation, and estimated dehydration, rather than applying adult infusion schedules unchanged. Because physical examination can misestimate fluid deficits, ISPAD uses approximate deficits of 5%, 7%, and 10% of body weight for mild, moderate, and severe DKA, respectively, with deficit replacement generally planned over 24–48 hours alongside maintenance fluids. Previously administered fluids are included in these calculations. (ispad.org)

Pediatric monitoring includes hourly vital signs, fluid intake and output, capillary glucose, and neurological assessment using the Glasgow Coma Scale or an equivalent instrument. Electrolytes, blood gases, and available blood ketone measurements are reassessed every 2–4 hours, or more frequently when indicated. New or worsening headache, declining consciousness, or unexpected slowing of heart rate prompts assessment for cerebral injury. Younger children, particularly those under five years, require heightened neurological observation; deterioration may begin before treatment, not only during fluid replacement. (ispad.org)

During pregnancy, placental and other insulin-antagonizing hormones increase insulin resistance, particularly later in gestation. Short periods of reduced food intake can produce comparatively rapid lipolysis and ketogenesis. Pregnancy-associated respiratory alkalosis also increases renal bicarbonate loss, reducing buffering capacity. Meanwhile, fetoplacental glucose consumption, increased renal glucose filtration, and expanded plasma volume can limit the degree of hyperglycemia despite substantial ketoacidosis. Consequently, DKA may develop at glucose concentrations below 200 mg/dL (11.1 mmol/L). (academic.oup.com) Maternal metabolic instability and the risk of fetal loss require coordinated medical and obstetric care. In a Mayo Clinic retrospective cohort covering January 2004–December 2021, 71 episodes occurred in 58 pregnancies among 51 women. No maternal deaths occurred, but 10 pregnancies (17.2%) ended in fetal demise—six miscarriages and four stillbirths. These are outcomes in an affected hospital cohort, not a population-wide risk estimate. (pmc.ncbi.nlm.nih.gov)

In older adults, frailty, cardiovascular disease, infection, and medication effects can complicate presentation and recovery; isolated hyperosmolar hyperglycemic state and mixed presentations are relatively more frequent than isolated DKA. Fluid replacement is individualized to comorbidity and repeatedly reassessed. In heart failure and advanced chronic kidney disease, reduced capacity to tolerate sodium and water loading makes unmodified standard replacement schedules hazardous. The 2024 adult consensus emphasizes smaller fluid increments and frequent hemodynamic assessment in these settings rather than assuming that every patient has a large, readily replaceable volume deficit. (diabetesjournals.org)

Dialysis-dependent kidney failure changes both fluid and potassium interpretation. Little or no urine production prevents the usual glucose-driven osmotic diuresis, so cellular dehydration may coexist with extracellular volume expansion. Hyperkalemia can reflect impaired excretion without the typical urinary depletion of total-body potassium. Hospital assessment therefore incorporates residual urine output, recent dialysis, weight relative to usual dialysis weights, and evidence of congestion. Dialysis can correct hyperkalemia, fluid overload, and acidosis, but does not replace insulin needed to suppress ketogenesis. Reduced renal insulin clearance also increases susceptibility to treatment-associated hypoglycemia. (diabetes.org.uk)

Ketosis-prone diabetes describes heterogeneous presentations in which DKA does not necessarily imply permanent, absolute insulin deficiency. The $A\beta$ classification separates patients by islet autoantibody status ($A+$ or $A-$) and preserved or absent β-cell reserve ($\beta+$ or $\beta-$). Evaluation includes antibodies such as GAD65, IA-2, and ZnT8, together with fasting and stimulated C-peptide after recovery, when acute glucose toxicity is less likely to distort secretion measurements. A prospective cohort assessed reserve 3–4 weeks after correction of ketoacidosis. Some antibody-negative patients with preserved reserve subsequently maintain glycemic control without insulin under supervised follow-up, whereas others remain insulin-dependent or relapse; neither body habitus nor the initial DKA episode alone establishes the long-term diabetes phenotype. (pmc.ncbi.nlm.nih.gov)

Complications and Prognosis

Complications may reflect the presenting metabolic crisis, its precipitating illness, or changes occurring during correction. Volume depletion and impaired organ perfusion can already be present at admission, whereas hypoglycemia commonly develops during insulin treatment. The 2024 adult consensus report summarized treatment studies in which glucose below 70 mg/dL occurred in 16–28% of DKA cases and glucose below 40 mg/dL in approximately 2%. Hypokalemia can emerge as insulin moves potassium into cells despite pre-existing total-body depletion; severe potassium disturbances can provoke cardiac arrhythmias. Treatment-associated hypoglycemia and severe hypokalemia are associated with increased inpatient mortality, although these associations do not establish that either disturbance independently caused death. (diabetesjournals.org)

Hyperchloremic metabolic acidosis is distinct from persistent ketoacidosis. Chloride-rich fluids, particularly 0.9% saline, can produce a normal-anion-gap acidosis during recovery, while urinary loss of keto-anions reduces the bicarbonate that could otherwise be regenerated through their metabolism. Consequently, bicarbonate may remain low after ketone production has subsided. This recovery-phase disturbance is usually self-limiting and has few clinical consequences. In Wesley H. Self and colleagues’ 2020 subgroup analysis of randomized fluid trials involving 172 adults, saline recipients had higher chloride and lower bicarbonate concentrations than recipients of balanced crystalloids, illustrating how fluid composition can alter biochemical recovery. (ispad.org)

Acute kidney injury is frequently associated with the initial dehydration and reduced renal perfusion rather than being solely a treatment complication. Brenden E. Hursh and colleagues’ Canadian study, published in 2017, examined 165 children aged 18 years or younger hospitalized with DKA during 2008–2013: 106, or 64.2%, met acute kidney injury criteria, and two required hemodialysis. Severe acidosis and markers of volume depletion were associated with kidney injury; no deaths occurred in that cohort. These findings describe a selected hospital population, not a universal pediatric incidence. (pubmed.ncbi.nlm.nih.gov) Thrombosis is another possible complication: dehydration and acidosis may alter clot structure, but thromboembolic risk is better documented in hyperosmolar presentations than in isolated DKA. (diabetesjournals.org)

Cerebral injury encompasses more than clinically visible cerebral edema. ISPAD’s 2022 guidance places clinically overt DKA-associated cerebral injury in children at approximately 0.5–0.9% of episodes, with reported case-fatality of 21–24%. Neurological deterioration can begin before treatment, demonstrating that fluid administration cannot be its sole cause. Proposed mechanisms include cerebral hypoperfusion, inflammatory activation, and disruption of the blood–brain barrier; their relative contributions remain incompletely established. The 2018 PECARN fluid trial found no significant neurological-outcome differences among its tested infusion rates and saline concentrations, rather than demonstrating that every possible fluid regimen is equivalent. (ispad.org) Osmotic demyelination syndrome is a separate, uncommon neurological complication associated with rapid correction of hyponatremia, particularly in hyperosmolar or mixed presentations; its incidence is uncertain. (diabetesjournals.org)

Case-fatality depends strongly on population and outcome definition. A CDC analysis of U.S. National Inpatient Sample data for 2000–2014 found that age-adjusted in-hospital case-fatality among DKA admissions declined from 1.1% to 0.4%. In 2014, the corresponding age-specific proportions were 0.1% below age 45 and 2.6% at age 75 or older. These were deaths during admissions with DKA, not necessarily deaths attributable exclusively to ketoacidosis. (cdc.gov) Precipitating illnesses and comorbid conditions—including infection and advanced kidney disease—also influence prognosis. (diabetesjournals.org)

Hospital survival does not eliminate subsequent morbidity. A U.S. Nationwide Readmissions Database study of adults with type 1 diabetes hospitalized for DKA during 2010–2018 found that risk-adjusted 30-day all-cause readmission increased from 16.0% to 19.4%; DKA-specific readmission increased from 8.1% to 11.3%. These admission-based outcomes distinguish recurrent ketoacidosis from readmission for other illnesses and should not be interpreted as the proportion of all people with diabetes experiencing recurrence. (pmc.ncbi.nlm.nih.gov)

Recurrence and Prevention Programs

Recurrent episodes require assessment beyond the immediate precipitating illness: the same interruption in insulin delivery, difficulty obtaining supplies, or psychosocial barrier may remain after biochemical recovery. The American Diabetes Association’s 2025 standards recommend reviewing previous hyperglycemic crises at every clinical encounter in people at risk and screening those who have experienced DKA for social determinants of health, including inadequate access to medication and monitoring equipment. Assessment therefore distinguishes an isolated precipitant from continuing barriers that require health-care or community-service coordination rather than education alone. (doi.org)

Structured diabetes education covers recognition of illness-related metabolic deterioration, interpretation of glucose and blood or urine ketone measurements, insulin administration, identification of delivery problems, and communication with clinical services. “Sick-day” education explains why infection or other physiological stress can increase insulin requirements even when food intake falls; programs provide individualized plans rather than assuming that routine management remains adequate during illness. The 2026 ADA standards also emphasize clear arrangements for whom patients and caregivers can contact outside normal clinic hours and where different levels of illness should be assessed. Education is recommended for people with type 1 diabetes, those with type 2 diabetes who have experienced a hyperglycemic crisis, and other high-risk groups. (pmc.ncbi.nlm.nih.gov)

Evidence for structured education includes the United Kingdom’s Dose Adjustment For Normal Eating (DAFNE) program. In a 2014 analysis by Jackie Elliott and colleagues, complete baseline and one-year follow-up data were available for 939 of 1,651 adults with type 1 diabetes. DKA rates fell from 0.07 episodes per patient-year during the year before training to 0.03 during the following year; the estimated relative risk was 0.39, corresponding to a 61% reduction. This was a before-and-after observational analysis, not a randomized comparison of DKA prevention, and only 57% of participants had complete data. The findings support an association between structured training and fewer emergencies without establishing that education alone caused the reduction. (pubmed.ncbi.nlm.nih.gov)

Discharge coordination links hospital treatment to continuing diabetes care. The 2024 adult consensus report specifies adequate supplies of insulin, glucose-monitoring and administration equipment, and contact details for clinical assistance. It identifies follow-up within two to four weeks as a potential means of reducing recurrence, alongside assessment of contributing causes. Access to diabetes specialist care and social services is particularly relevant when affordability or supply problems persist. Mental-health assessment also forms part of prevention: depression, eating disorders, and other psychological conditions are associated with DKA risk, and multidisciplinary approaches may include psychological intervention, peer support, coaching, and family-based therapy. (pmc.ncbi.nlm.nih.gov)

A specific example of family-centered intervention is Deborah Ellis and colleagues’ 2008 randomized controlled trial of multisystemic therapy at Children’s Hospital of Michigan. The study involved 127 adolescents aged 10–17 years with persistently elevated glycated hemoglobin. Intensive home-based psychotherapy was added to standard medical care and addressed barriers across family, school, and health-care settings. Participants assigned to the intervention had significantly fewer DKA admissions over 24 months than controls. This evidence concerns a defined high-risk adolescent population, rather than all people discharged after DKA. (pmc.ncbi.nlm.nih.gov)

Monitoring technologies can support prevention but do not eliminate risk. In a 2023 population-based study by Beate Karges and colleagues, 32,117 young people with type 1 diabetes were identified through 511 centers in Austria, Germany, Luxembourg, and Switzerland, using treatment data from 2014–2021. DKA rates were 3.72 per 100 patient-years among users of continuous glucose monitoring, compared with 7.29 among users of blood-glucose monitoring; the adjusted incidence-rate ratio was 0.51. This observational association does not establish causation, and the nonzero event rate demonstrates residual risk. Device provision remains complementary to ketone-testing education, reliable insulin access, and clinical support. (pubmed.ncbi.nlm.nih.gov)

Development of Clinical Knowledge and Research

The discovery of insulin in 1921 established the therapeutic foundation for reversing diabetic ketoacidosis, but did not immediately produce modern DKA-specific protocols. At the University of Toronto, Frederick Banting and Charles Best investigated pancreatic extracts under John Macleod’s direction; James Collip subsequently improved purification for clinical use. Fourteen-year-old Leonard Thompson received Banting and Best’s preparation on January 11, 1922, and Collip’s purified extract on January 23. Commercial insulin production followed in 1923. These developments made insulin replacement practicable; the later standardization of fluid replacement, electrolyte management, laboratory monitoring, and treatment endpoints belonged to a separate history of acute-care research and consensus development. (nobelprize.org)

Formal American Diabetes Association adult consensus statements appeared in 2001 and 2009. The 2009 statement, authored by Abbas E. Kitabchi, Guillermo E. Umpierrez, John M. Miles, and Joseph N. Fisher, organized diagnosis and treatment around hyperglycemia, ketosis, acidosis, and clinical severity. Although its classification table used glucose above 250 mg/dL, it explicitly recognized lower-glucose, or “euglycemic,” presentations. The 2024 update involved five organizations: the American Diabetes Association, European Association for the Study of Diabetes, Joint British Diabetes Societies for Inpatient Care, American Association of Clinical Endocrinology, and Diabetes Technology Society. It lowered the glucose criterion to 200 mg/dL or accepted a prior history of diabetes, while retaining the requirements for ketosis and metabolic acidosis. Direct blood β-hydroxybutyrate measurement gained priority over urine nitroprusside testing, which does not measure this principal ketoacid and can misrepresent recovery. Recognition of lower-glucose presentations therefore became more explicit rather than originating in 2024. (pmc.ncbi.nlm.nih.gov)

A major experimental contribution was the Pediatric Emergency Care Applied Research Network (PECARN) fluid trial, published in June 2018 by Nathan Kuppermann and colleagues. This randomized controlled trial enrolled 1,255 children at 13 United States emergency departments between February 2011 and September 2016; repeat enrollment during recurrent illness produced 1,389 evaluated DKA episodes. Its $2\times2$ factorial design compared faster and slower rehydration with either 0.45% or 0.9% sodium chloride. The primary outcome was deterioration of neurological status, defined by two consecutive Glasgow Coma Scale scores below 14 during the first 24 hours. Neither infusion rate nor saline concentration produced statistically significant differences in mental-status deterioration, clinically apparent brain injury, or neurocognitive function after recovery. These findings challenged the assumption that slower rehydration necessarily protects the brain, but applied to the regimens and population studied rather than establishing the safety of every possible fluid strategy. (nejm.org)

Continuous ketone monitoring has been investigated as a means of detecting rising ketones between intermittent blood tests. In a 2021 feasibility study, Shridhara Alva and colleagues tested subcutaneous sensors in 12 participants following low-carbohydrate diets for 14 days. The sensors used β-hydroxybutyrate dehydrogenase chemistry to measure ketones in interstitial fluid, with capillary measurements as the reference. Following a single retrospective calibration, 91.4% of paired readings fell within the study’s specified absolute-or-relative accuracy limits. This demonstrated technical feasibility, not prevention of DKA or reliable performance during acute ketoacidosis; the authors specifically identified testing in intended patient populations as necessary. (pubmed.ncbi.nlm.nih.gov)

Research on post-DKA transitions addresses the interval between biochemical recovery and sustained outpatient treatment. The 2024 consensus identified real-time continuous glucose monitoring at discharge, continuous interstitial ketone monitoring in high-risk patients, and transition to ultra-long-acting insulin after crisis resolution as research priorities. These proposed evaluations are distinct from established practices such as overlapping intravenous and subcutaneous insulin during hospital transition; they concern whether particular technologies or formulations improve outcomes beyond existing care. (pmc.ncbi.nlm.nih.gov)

References

1. Hyperglycemic Crises - Endotext - NCBI Bookshelfncbi.nlm.nih.gov
2. ketogenesis | Pathway - PubChempubchem.ncbi.nlm.nih.gov
3. Ketoacidosis - PubMedpubmed.ncbi.nlm.nih.gov
4. Biochemistry, Ketogenesis - StatPearls - NCBI Bookshelfncbi.nlm.nih.gov
5. Hyperventilation in severe diabetic ketoacidosispubmed.ncbi.nlm.nih.gov
6. Hyperglycemic Crises in Adults With Diabetes: A Consensus Report - PMCpmc.ncbi.nlm.nih.gov
7. Trends in Diabetic Ketoacidosis Hospitalizations and In-Hospital Mortality — United States, 2000–2014cdc.gov
8. Presence of diabetic ketoacidosis at diagnosis of diabetes mellitus in youth: the Search for Diabetes in Youth Studypubmed.ncbi.nlm.nih.gov
9. Increase in Prevalence of Diabetic Ketoacidosis at Diagnosis Among Youth With Type 1 Diabetes: The SEARCH for Diabetes in Youth Studypmc.ncbi.nlm.nih.gov
10. Increase in Prevalence of Diabetic Ketoacidosis at Diagnosis Among Youth With Type 1 Diabetes: The SEARCH for Diabetes in Youth Studypubmed.ncbi.nlm.nih.gov
11. Temporal trends in diabetic ketoacidosis at diagnosis of paediatric type 1 diabetes between 2006 and 2016: results from 13 countries in three continentspmc.ncbi.nlm.nih.gov
12. Recurrent diabetic ketoacidosis in inner-city minority patients: behavioral, socioeconomic, and psychosocial factorspubmed.ncbi.nlm.nih.gov
13. Hyperglycemic Crises in Adults With Diabetes: A Consensus Reportpmc.ncbi.nlm.nih.gov
14. Hyperglycaemic crises in adults with diabetes: a consensus reportpmc.ncbi.nlm.nih.gov
15. Management of Hyperglycemic Crises: Diabetic ketoacidosis and hyperglycemic hyperosmolar statepmc.ncbi.nlm.nih.gov
16. FDA Supplemental Application Review, 204629Orig1s007accessdata.fda.gov
17. Diabetic Ketoacidosis - Endotext - NCBI Bookshelfncbi.nlm.nih.gov
18. Abdominal pain in patients with hyperglycemic crisespubmed.ncbi.nlm.nih.gov
19. Diabetic Ketoacidosis: Clinical Characteristics and Precipitating Factorspubmed.ncbi.nlm.nih.gov
20. Hyperglycemic Crises — Endotextncbi.nlm.nih.gov
21. Hyperglycemic Crises in Adults With Diabetes: A Consensus Reportprofketandhatariya.com
22. Hyperglycemic Crises — Endotextncbi.nlm.nih.gov
23. Alcoholic ketoacidosis evaluated with a point-of-care capillary beta-hydroxybutyrate measurement devicepubmed.ncbi.nlm.nih.gov
24. Utility of the serum osmol gap in the diagnosis of methanol or ethylene glycol ingestionpubmed.ncbi.nlm.nih.gov
25. Management Priorities in Salicylate Toxicityacmt.net
26. Clinical Effects of Balanced Crystalloids vs Saline in Adults With Diabetic Ketoacidosis: A Subgroup Analysis of Cluster Randomized Clinical Trialspubmed.ncbi.nlm.nih.gov
27. Clinical Effects of Balanced Crystalloids vs Saline in Adults With Diabetic Ketoacidosis: A Subgroup Analysis of Cluster Randomized Clinical Trialspmc.ncbi.nlm.nih.gov
28. Bicarbonate therapy in severe diabetic ketoacidosispubmed.ncbi.nlm.nih.gov
29. ISPAD Clinical Practice Consensus Guidelines 2022: Diabetic ketoacidosis and hyperglycemic hyperosmolar stateispad.org
30. Diabetic Ketoacidosis in Pregnancy: Clinical Risk Factors, Presentation, and Outcomespmc.ncbi.nlm.nih.gov
31. The management of DKA in people with end stage renal failure or on dialysisdiabetes.org.uk
32. Diagnosing Diabetic Ketoacidosisdiabetes.org.uk
33. Clinical characteristics and outcomes of patients with end-stage renal disease hospitalized with diabetes ketoacidosispmc.ncbi.nlm.nih.gov
34. End-Stage Renal Disease Increases Rates of Adverse Glucose Events When Treating Diabetic Ketoacidosis or Hyperosmolar Hyperglycemic Statepmc.ncbi.nlm.nih.gov
35. Presence or absence of a known DKA precipitant defines distinct syndromes of “A-β+” Ketosis-Prone Diabetes (KPD) based on long-term beta cell function, HLA class II alleles, and gender predilectionpmc.ncbi.nlm.nih.gov
36. Factors associated with early relapse to insulin dependence in unprovoked A-β+ Ketosis-Prone Diabetespmc.ncbi.nlm.nih.gov
37. Acute Kidney Injury in Children With Type 1 Diabetes Hospitalized for Diabetic Ketoacidosispubmed.ncbi.nlm.nih.gov
38. Rising Readmission Rates After Diabetic Ketoacidosis Hospitalization Among Adults With Type 1 Diabetes Throughout a Decade in the United Statespmc.ncbi.nlm.nih.gov
39. Substantial reductions in the number of diabetic ketoacidosis and severe hypoglycaemia episodes requiring emergency treatment lead to reduced costs after structured education in adults with Type 1 diabetespubmed.ncbi.nlm.nih.gov
40. Multisystemic Therapy for Adolescents With Poorly Controlled Type 1 Diabetes: Reduced diabetic ketoacidosis admissions and related costs over 24 monthspmc.ncbi.nlm.nih.gov
41. Continuous glucose monitoring versus blood glucose monitoring for risk of severe hypoglycaemia and diabetic ketoacidosis in children, adolescents, and young adults with type 1 diabetes: a population-based studypubmed.ncbi.nlm.nih.gov
42. Celebrating the discovery and development of insulin - NIDDKniddk.nih.gov
43. John Macleod – Nobel Lecture - NobelPrize.orgnobelprize.org
44. Hyperglycemic Crises in Adult Patients With Diabetes - PMCpmc.ncbi.nlm.nih.gov
45. Feasibility of Continuous Ketone Monitoring in Subcutaneous Tissue Using a Ketone Sensor - PubMedpubmed.ncbi.nlm.nih.gov