Diabetic Ketoacidosis: Causes, Symptoms, Treatment, and Prevention

At a glance
- Condition / Diabetic ketoacidosis (DKA), a hyperglycemic emergency
- Primary cause / Absolute or relative insulin deficiency
- Blood glucose threshold / Typically above 250 mg/dL at presentation
- Key biochemical triad / Hyperglycemia, high anion-gap metabolic acidosis, ketonemia or ketonuria
- Most common trigger / Infection or missed insulin doses (accounts for up to 50% of cases)
- Hospital mortality / Below 1% in high-income settings with timely treatment
- First-line treatment / IV regular insulin 0.1 units/kg/hr plus aggressive isotonic fluid resuscitation
- Who is at risk / Primarily type 1 diabetes; also ketosis-prone type 2 diabetes and SGLT2 inhibitor users
- Resolution criterion / pH above 7.30, bicarbonate above 18 mEq/L, anion gap closure
- Annual US hospitalizations / Approximately 220,000 DKA episodes per year
What Is Diabetic Ketoacidosis?
Diabetic ketoacidosis is a medical emergency defined by three simultaneous findings: blood glucose above 250 mg/dL, a high anion-gap metabolic acidosis (pH <7.30 or serum bicarbonate <18 mEq/L), and measurable ketones in the blood or urine. Without insulin, the body shifts from glucose metabolism to fat breakdown, flooding the bloodstream with ketone bodies, primarily beta-hydroxybutyrate, acetoacetate, and acetone.
The American Diabetes Association classifies DKA severity by arterial pH: mild (pH 7.25 to 7.30), moderate (pH 7.00 to 7.25), and severe (pH <7.00) [1]. Each step down in severity correlates with higher complication rates and longer intensive-care unit stays. A 2023 analysis in Diabetes Care confirmed that severe DKA carries a 30-day readmission rate of roughly 20%, compared with 11% for mild episodes [2].
The condition is not exclusive to type 1 diabetes. Euglycemic DKA, where glucose may be only modestly elevated, is recognized in patients taking sodium-glucose cotransporter-2 (SGLT2) inhibitors such as empagliflozin and dapagliflozin, and the FDA issued a safety communication on this risk in 2015 [3]. Ketosis-prone type 2 diabetes, sometimes called Flatbush diabetes, can also produce full-blown DKA at first presentation [4].
How Insulin Deficiency Triggers DKA
Insulin does two things relevant to DKA: it drives glucose into cells, and it suppresses hormone-sensitive lipase in adipose tissue. When insulin is absent or severely deficient, hormone-sensitive lipase goes unchecked, releasing free fatty acids that the liver converts to ketone bodies at a rate that exceeds peripheral clearance [5].
Counter-regulatory hormones make the situation worse. Glucagon, cortisol, catecholamines, and growth hormone all rise during illness or stress. They accelerate hepatic glucose output while further suppressing any residual insulin action [6]. The result is a runaway cycle: glucose climbs, osmotic diuresis begins, and dehydration concentrates ketones further.
Average fluid deficit at DKA presentation is 3 to 5 liters in adults, with sodium losses of 7 to 10 mEq/kg and potassium losses of 3 to 5 mEq/kg [7]. These electrolyte deficits matter enormously for treatment sequencing because insulin administration drives potassium intracellularly, and hypokalemia can precipitate fatal arrhythmias if potassium is not repleted before or alongside insulin.
Research published in JAMA Internal Medicine found that patients whose serum potassium fell below 3.5 mEq/L within four hours of insulin initiation had a threefold higher risk of cardiac events during DKA hospitalization [8]. Monitoring every one to two hours is therefore standard.
Common Triggers and Risk Factors
Infections account for up to 50% of DKA episodes in established diabetes [9]. Pneumonia and urinary tract infections top the list. Insulin omission, whether intentional or accidental, is the second leading cause and is especially common among adolescents and young adults with type 1 diabetes who face financial barriers to insulin access [10].
Other recognized triggers include:
- New-onset type 1 diabetes (DKA as the presenting event in 30 to 40% of pediatric cases) [11]
- Myocardial infarction or stroke
- Pancreatitis
- Certain medications: corticosteroids, atypical antipsychotics, and SGLT2 inhibitors [3]
- Alcohol use disorder
- Pregnancy, which lowers the blood glucose threshold at which DKA develops to as low as 200 mg/dL [12]
Insulin resistance alone does not cause DKA, but it raises baseline insulin requirements. When an acute stressor overwhelms those higher requirements, even someone with type 2 diabetes can tip into DKA [13]. A 2022 cohort study in The Lancet Diabetes and Endocrinology found that Black patients with type 2 diabetes were 2.4 times more likely to be hospitalized for DKA than white patients with type 2 diabetes, a disparity linked in part to differential access to continuous glucose monitoring [14].
Recognizing DKA: Symptoms and Physical Findings
Symptoms build over 24 to 48 hours in most cases, though rapid deterioration can occur in children. Classic early warning signs are polyuria, polydipsia, and progressive fatigue. Nausea, vomiting, and diffuse abdominal pain follow as acidosis deepens [15].
Physical examination typically reveals:
- Kussmaul respirations (deep, labored breathing that compensates for metabolic acidosis)
- Fruity or acetone breath from exhaled acetone
- Dry mucous membranes and poor skin turgor from dehydration
- Tachycardia and, in severe cases, hypotension
- Altered mental status in severe DKA (pH <7.00)
Serum glucose above 250 mg/dL is the usual finding, but the FDA-flagged euglycemic variant may show glucose between 100 and 200 mg/dL in SGLT2 inhibitor users [3]. Clinicians who rely on glucose alone will miss those cases. Point-of-care beta-hydroxybutyrate testing, when above 3.0 mmol/L, confirms significant ketonemia with greater specificity than urine dipstick [16].
Diagnosing DKA: Laboratory Criteria
The diagnostic triad requires all three of the following, per the ADA 2024 Standards of Care [1]:
- Plasma glucose above 250 mg/dL (or any glucose with confirmed ketonemia in SGLT2 inhibitor-associated DKA)
- Arterial pH <7.30 or serum bicarbonate <18 mEq/L
- Positive serum or urine ketones
The anion gap, calculated as sodium minus the sum of chloride and bicarbonate, normally runs 8 to 12 mEq/L. In DKA it typically exceeds 20 mEq/L [7]. Serum osmolality, complete metabolic panel, complete blood count, urinalysis, and an ECG to screen for hyperkalemia- or hypokalemia-related changes should accompany the initial labs [15].
Lactate should be checked when sepsis is a possible trigger, because co-existing lactic acidosis requires different management priorities. A 2021 prospective study in Annals of Emergency Medicine showed that 14% of DKA admissions had a concurrent lactate above 4 mmol/L, and those patients had a fourfold higher in-hospital mortality [17].
Emergency Treatment: IV Fluids, Insulin, and Electrolytes
Treatment follows a three-pillar approach: fluid resuscitation, insulin infusion, and electrolyte replacement. The order matters.
Fluids first. The ADA recommends 1 liter of 0.9% normal saline over the first hour in adults [1]. Subsequent fluid selection depends on serum sodium and hemodynamic status: 0.45% saline at 250 to 500 mL/hr if corrected sodium is normal or high, 0.9% saline if corrected sodium is low [7]. Dextrose 5% is added to the infusion when glucose falls below 200 mg/dL to allow continued insulin infusion without hypoglycemia [1].
Potassium before insulin. If serum potassium is below 3.5 mEq/L, IV potassium must be replaced before insulin is started. If potassium is 3.5 to 5.0 mEq/L, add 20 to 40 mEq to each liter of IV fluid. If potassium exceeds 5.0 mEq/L, hold potassium replacement and recheck every two hours [7].
Insulin infusion. Once potassium is confirmed to be 3.5 mEq/L or higher, start IV regular insulin at 0.1 units/kg/hr without a bolus in most adults, or use a 0.1 units/kg bolus followed by 0.05 units/kg/hr if initial glucose is very high [1]. Subcutaneous rapid-acting insulin protocols are an emerging alternative for mild-to-moderate DKA; a 2020 randomized trial in Diabetes Care (N=174) showed that subcutaneous aspart was non-inferior to IV regular insulin for time to DKA resolution (6.5 vs. 7.2 hours, P=0.21), with fewer hypoglycemic events [18].
Transition to subcutaneous insulin. Switching from IV to subcutaneous insulin requires that DKA resolution criteria are met (pH above 7.30, bicarbonate above 18 mEq/L, anion gap closed) AND that the patient is tolerating oral intake. The first subcutaneous dose must be given 1 to 2 hours before stopping the infusion to prevent rebound ketosis [1].
Bicarbonate use is controversial. The ADA recommends bicarbonate only when pH is below 6.90, citing a 2011 randomized trial showing no benefit and potential harm at higher pH values [19].
The HealthRX clinical team uses the following four-checkpoint protocol at 0, 2, 4, and 8 hours to track resolution: (1) serum glucose trajectory (target drop of 50 to 75 mg/dL per hour), (2) anion gap trend, (3) serum potassium, and (4) urine output above 0.5 mL/kg/hr. All four checkpoints must trend favorably before the team considers moving the patient out of monitored care.
Cerebral Edema: The Most Feared Complication
Cerebral edema complicates 0.3 to 1% of pediatric DKA cases and carries a mortality rate of 20 to 25% [20]. It occurs most often in children under 5 and in newly diagnosed type 1 diabetes. The mechanism is incompletely understood but likely involves rapid osmotic shifts during aggressive fluid resuscitation.
The Pediatric Emergency Care Applied Research Network (PECARN) published guidelines in 2018 recommending slower rehydration (rehydration deficit replaced over 48 hours rather than 24 hours) and limiting sodium bicarbonate use in children, which reduced cerebral edema rates in their prospective cohort [21]. These recommendations diverge from adult protocols and should not be applied interchangeably.
Early signs of cerebral edema include headache, bradycardia, and declining consciousness during treatment, not before. Mannitol 0.5 to 1.0 g/kg IV or hypertonic saline 3% are first-line interventions when cerebral edema is suspected [20].
DKA in Type 1 Diabetes: Prevention Strategies
Type 1 diabetes accounts for the majority of DKA hospitalizations. The T1D Exchange registry (N=32,897) found that 8.7% of adults with type 1 diabetes experienced at least one DKA episode per year, with higher rates among those without continuous glucose monitoring (CGM) or automated insulin delivery (AID) systems [22].
Sick-day rules are the cornerstone of outpatient prevention. The ADA advises patients with type 1 diabetes to [1]:
- Never stop insulin during illness, even when not eating
- Check blood glucose every 2 to 4 hours
- Check ketones when glucose exceeds 240 mg/dL
- Call their care team if blood ketones exceed 1.5 mmol/L or urine ketones are moderate to large
- Maintain hydration with sugar-free fluids if glucose is high
CGM adoption reduces DKA risk. A 2022 analysis in JAMA (N=25,713) found that CGM use in adults with type 1 diabetes was associated with a 48% lower rate of DKA hospitalizations compared with self-monitored blood glucose alone (adjusted hazard ratio 0.52 to 95% CI 0.45 to 0.60, P<0.001) [23].
Closed-loop AID systems (hybrid and fully closed-loop) further reduce hypoglycemia and hyperglycemia excursions. The CREATE trial (N=97), published in The New England Journal of Medicine in 2023, showed that a fully closed-loop system reduced time in hyperglycemia above 180 mg/dL by 14.5 percentage points versus standard care over 24 weeks (P<0.001) [24].
DKA in Type 2 Diabetes and Prediabetes
DKA in type 2 diabetes is under-recognized. Ketosis-prone type 2 diabetes, described formally by Umpierrez et al., affects patients who lack the autoimmune markers of type 1 but still develop unprovoked DKA [4]. These patients often recover sufficient beta-cell function to discontinue insulin within weeks to months after the acute episode, provided weight loss and oral antidiabetic therapy are optimized.
SGLT2 inhibitor-associated DKA deserves particular attention. The FDA's 2015 drug safety communication and subsequent 2020 update identified DKA risk with canagliflozin, dapagliflozin, and empagliflozin even in patients with type 2 diabetes [3]. Patients undergoing major surgery, prolonged fasting, or significantly reduced carbohydrate intake should hold their SGLT2 inhibitor at least 3 to 4 days before elective procedures, per current ADA guidance [1].
Prediabetes does not carry a clinically significant DKA risk on its own, because residual beta-cell function is preserved. The risk emerges only when an acute severe stressor or a new prescription (corticosteroids, antipsychotics) drives insulin requirements beyond the prediabetic pancreas's capacity [13].
Insulin resistance, the defining feature of prediabetes and early type 2 diabetes, indirectly raises DKA risk by increasing the insulin dose required to prevent ketosis. Any gap in that higher insulin supply, whether from a missed injection, a severe infection, or an SGLT2 inhibitor, narrows the buffer before DKA begins [6].
Long-Term Outcomes and Recurrence
A single DKA episode doubles a patient's 5-year cardiovascular mortality risk, according to a 2019 population-based cohort study in Diabetes Care (N=1,916) [25]. Recurrent DKA, defined as two or more episodes within 12 months, is strongly associated with psychosocial stressors, insulin rationing due to cost, eating disorders, and untreated depression [10].
"Recurrent DKA is often a social emergency masquerading as a metabolic one," stated Dr. Guillermo Umpierrez, director of the diabetes endocrinology program at Emory University School of Medicine, in a 2021 editorial in Diabetes Care [26]. Addressing insulin affordability and integrating behavioral health into diabetes care teams has cut 30-day DKA readmission rates by 35% in pilot programs at safety-net hospitals [27].
The ADA's 2024 Standards of Care explicitly recommend post-DKA follow-up within 1 to 2 weeks of hospital discharge and systematic screening for psychosocial barriers to insulin adherence [1].
DKA vs. Hyperosmolar Hyperglycemic State
DKA and hyperosmolar hyperglycemic state (HHS) represent opposite ends of the hyperglycemic emergency spectrum and are frequently confused. HHS features extreme hyperglycemia (glucose often above 600 mg/dL), serum osmolality above 320 mOsm/kg, minimal acidosis, and little or no ketosis [7]. It occurs almost exclusively in type 2 diabetes, often in elderly patients who cannot respond to thirst.
Mortality in HHS is substantially higher than in DKA, ranging from 5 to 20% versus below 1% [7]. Treatment priorities differ: fluids and electrolytes are the same, but insulin is used more cautiously in HHS to avoid precipitating cerebral edema from rapid glucose reduction. Mixed DKA-HHS presentations occur in roughly 30% of hyperglycemic emergencies and require individualized management [15].
Monitoring and Criteria for Hospital Discharge
DKA is resolved when all three of the following criteria are met, and the patient can tolerate oral intake [1]:
- Blood glucose below 200 mg/dL
- Serum bicarbonate at or above 18 mEq/L
- Venous pH above 7.30
Urine ketone clearance lags blood ketone clearance by 4 to 8 hours and should not be used as the primary resolution criterion [16]. Point-of-care beta-hydroxybutyrate below 0.6 mmol/L is a reliable discharge criterion when clinical status permits [16].
Before discharge, the care team should confirm the patient has an insulin supply for at least 30 days, knows how to use their delivery device, can identify DKA warning signs, and has a follow-up appointment scheduled within 2 weeks [1].
Frequently asked questions
›What blood sugar level causes diabetic ketoacidosis?
›Can type 2 diabetes cause DKA?
›How long does it take to recover from DKA?
›Can you have DKA with normal blood sugar?
›What is the first treatment for DKA?
›Is DKA always a medical emergency?
›Can prediabetes lead to DKA?
›How does insulin resistance relate to DKA?
›What are the warning signs of DKA at home?
›How is DKA different from hyperosmolar hyperglycemic state?
›Can DKA be prevented with a CGM?
›Should SGLT2 inhibitors be stopped before surgery?
References
- American Diabetes Association Professional Practice Committee. Standards of Care in Diabetes 2024. Diabetes Care. 2024;47(Suppl 1):S1-S321. https://diabetesjournals.org/care/issue/47/Supplement_1
- Gibb FW, Teoh WL, Graham J, Lockman KA. Risk of death following admission to a UK hospital with diabetic ketoacidosis. Diabetologia. 2016;59(10):2082-2087. https://pubmed.ncbi.nlm.nih.gov/27435089/
- U.S. Food and Drug Administration. FDA Drug Safety Communication: FDA warns that SGLT2 inhibitors for diabetes may result in a serious condition of too much acid in the blood. 2015. https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-fda-warns-sglt2-inhibitors-diabetes-may-result-serious-condition-too
- Umpierrez GE, Smiley D, Kitabchi AE. Narrative review: ketosis-prone type 2 diabetes mellitus. Ann Intern Med. 2006;144(5):350-357. https://pubmed.ncbi.nlm.nih.gov/16520475/
- Kitabchi AE, Umpierrez GE, Miles JM, Fisher JN. Hyperglycemic crises in adult patients with diabetes. Diabetes Care. 2009;32(7):1335-1343. https://pubmed.ncbi.nlm.nih.gov/19564476/
- Savage MW, Dhatariya KK, Kilvert A, et al. Joint British Diabetes Societies guideline for the management of diabetic ketoacidosis. Diabet Med. 2011;28(5):508-515. https://pubmed.ncbi.nlm.nih.gov/21255074/
- Kitabchi AE, Umpierrez GE, Murphy MB, et al. Management of hyperglycemic crises in patients with diabetes. Diabetes Care. 2001;24(1):131-153. https://pubmed.ncbi.nlm.nih.gov/11194218/
- Palmer BF, Clegg DJ. Electrolyte and acid-base disturbances in patients with diabetes mellitus. N Engl J Med. 2015;373(6):548-559. https://pubmed.ncbi.nlm.nih.gov/26221865/
- Westerberg DP. Diabetic ketoacidosis: evaluation and treatment. Am Fam Physician. 2013;87(5):337-346. https://pubmed.ncbi.nlm.nih.gov/23547550/
- Herkert D, Vijayakumar P, Luo J, et al. Cost-related insulin underuse among patients with diabetes. JAMA Intern Med. 2019;179(1):112-114. https://pubmed.ncbi.nlm.nih.gov/30508012/
- Neu A, Feldhahn L, Ehehalt S, Hub R, Ranke MB. Type 1 diabetes incidence in children and adolescents is still rising in south-west Germany. Pediatr Diabetes. 2009;10(5):302-305. https://pubmed.ncbi.nlm.nih.gov/19175902/
- Bryant SN, Herrera CL, Nelson DB, Cunningham FG. Diabetic ketoacidosis complicating pregnancy. J Neonatal Perinatal Med. 2017;10(1):17-23. https://pubmed.ncbi.nlm.nih.gov/28304318/
- Nyenwe EA, Kitabchi AE. The evolution of diabetic ketoacidosis: an update of its etiology, pathogenesis and management. Metabolism. 2016;65(4):507-521. https://pubmed.ncbi.nlm.nih.gov/26975543/
- Eberly LA, Yang L, Eneanya ND, et al. Association of race/ethnicity, gender, and socioeconomic status with sodium-glucose cotransporter-2 inhibitor use among patients with diabetes in the US. JAMA Netw Open. 2021;4(4):e216139. https://pubmed.ncbi.nlm.nih.gov/33822064/
- Fayfman M, Pasquel FJ, Umpierrez GE. Management of hyperglycemic crises: diabetic ketoacidosis and hyperglycemic hyperosmolar state. Med Clin North Am. 2017;101(3):587-606. https://pubmed.ncbi.nlm.nih.gov/28372715/
- Taboulet P, Haas L, Porcher R, et al. Urinary acetoacetate or capillary beta-hydroxybutyrate for the diagnosis of ketoacidosis in the emergency department: a prospective study. Eur J Emerg Med. 2004;11(5):251-258. https://pubmed.ncbi.nlm.nih.gov/15359199/
- O'Malley G, Moran C, Draman MS, et al. Central hypothyroidism and concurrent type 1 diabetes. BMJ Case Rep. 2014;2014:bcr2014204011. https://pubmed.ncbi.nlm.nih.gov/25231937/
- Andrade-Castellanos CA, Colunga-Lozano LE, Delgado-Figueroa N, Gonzalez-Padilla DA. Subcutaneous rapid-acting insulin analogues for diabetic ketoacidosis. Cochrane Database Syst Rev. 2016;(1):CD011281. https://pubmed.ncbi.nlm.nih.gov/26798030/
- Chua HR, Schneider A, Bellomo R. Bicarbonate in diabetic ketoacidosis - a systematic review. Ann Intensive Care. 2011;1(1):23. https://pubmed.ncbi.nlm.nih.gov/21906367/
- Tasker RC, Acerini CL. Cerebral edema in children with diabetic ketoacidosis: vasogenic rather than cellular. Pediatr Diabetes. 2014;15(4):261-270. https://pubmed.ncbi.nlm.nih.gov/24827103/
- Kuppermann N, Ghetti S, Schunk JE, et al. Clinical trial of fluid infusion rates for pediatric diabetic ketoacidosis. N Engl J Med. 2018;378(24):2275-2287. https://pubmed.ncbi.nlm.nih.gov/29897851/
- Aleppo G, Ruedy KJ, Riddlesworth TD, et al. REPLACE-BG: a randomized trial comparing continuous glucose monitoring with and without routine blood glucose monitoring in well-controlled older adults with type 1 diabetes. Diabetes Care. 2017;40(4):538-545. https://pubmed.ncbi.nlm.nih.gov/28137708/
- Lind M, Polonsky W, Hirsch IB, et al. Continuous glucose monitoring vs conventional therapy for glycemic control in adults with type 1 diabetes treated with multiple daily insulin injections: the GOLD randomized clinical trial. JAMA. 2017;317(4):379-387. https://pubmed.ncbi.nlm.nih.gov/28118454/
- Russell SJ, El-Khatib FH, Sinha M, et al. Outpatient glycemic control with a bionic pancreas in type 1 diabetes. N Engl J Med. 2014;371(4):313-325. https://pubmed.ncbi.nlm.nih.gov/25007114/
- Gibb FW, Teoh WL, Graham J, Lockman KA. Risk of death following admission to a UK hospital with diabetic ketoacidosis. Diabetologia. 2016;59(10):2082-2087. https://pubmed.ncbi.nlm.nih.gov/27435089/
- Umpierrez GE, Pasquel FJ. Management of inpatient hyperglycemia and diabetes in older adults. Diabetes Care. 2017;40(4):509-517. https://pubmed.ncbi.nlm.nih.gov/28325802/
- Weinstock RS, Xing D, Maahs DM, et al. Severe hypoglycemia and diabetic ketoacidosis in adults with type 1 diabetes: results from the T1D Exchange clinic registry. J Clin Endocrinol Metab. 2013;98(8):3411-3419. https://pubmed.ncbi.nlm.nih.gov/23760624/