Diabetic ketoacidosis

  • Diabetic ketoacidosis (DKA) results either from absolute insulin deficiency or from relative insulin deficiency in the setting of high levels of counterregulatory hormones stimulated by infection or other illness.

  • DKA is characterized by hyperglycemia, ketosis, and acidosis.

  • Treatment of pediatric DKA involves intravenous insulin administration, intravenous fluid administration to correct dehydration, and replacement of electrolyte deficits.

  • Cerebral injury is the most frequent serious complication of DKA in children and is the most frequent cause of morbidity and mortality resulting from DKA.

Etiology, definition, and presentation

Diabetic ketoacidosis (DKA) occurs when serum insulin concentrations are very low in relation to concentrations of glucagon and other counterregulatory hormones (epinephrine, norepinephrine, cortisol, and growth hormone). This situation occurs most commonly in new-onset type 1 diabetes mellitus (T1DM) and in patients with known diabetes during infections or other intercurrent illnesses or with insulin omission (discussed later). In the setting of low insulin concentrations in relation to counterregulatory hormone concentrations, the normal physiologic mechanisms responsible for maintaining adequate energy substrate during fasting and physiologic stress are exaggerated, resulting in hyperglycemia, ketosis, and acidosis. A diagnosis of DKA can be made when the serum glucose concentration is greater than 200 mg/dL (>11 mmol/L) and venous pH is less than 7.30 (or the serum bicarbonate concentration is less than 15 mmol/L) in the presence of ketosis (“moderate” or “large” ketonuria or serum β-hydroxybutyrate >3 mmol/L).

In a child with new onset of T1DM, declining insulin production results from autoimmune destruction of pancreatic β-cells. The concentration of insulin is decreased relative to glucagon, causing excess hepatic glucose production and decreased peripheral glucose uptake in muscle and adipose tissue. , When the serum glucose concentration rises above approximately 180 to 200 mg/dL, the renal threshold for glucose reabsorption is exceeded. This causes glycosuria, which leads to osmotic diuresis and compensatory polydipsia. Low insulin concentrations also stimulate the release of free fatty acids (FFAs) from adipose tissue to fuel ketogenesis. This, in combination with activation of the hepatic β-oxidative enzyme sequence resulting from relative excess of glucagon in relation to insulin, results in markedly increased hepatic ketone production.

Progressive dehydration and increasing acidosis eventually stimulate release of the counterregulatory (“stress”) hormones—cortisol, catecholamines, and growth hormone—which accelerate hepatic glucose output and ketone production. , Infection or other illness or injury can likewise contribute to this process by stimulating release of counterregulatory hormones. Elevated cortisol concentrations augment FFA release from adipose tissue and decrease peripheral glucose uptake. Increased epinephrine concentrations directly increase glycogenolysis and stimulate the release of gluconeogenic precursors from muscle. , Both epinephrine and norepinephrine also stimulate lipolysis and β-oxidation of FFAs. , Catecholamines may also directly inhibit insulin secretion, thereby accelerating DKA in patients with endogenous insulin capacity, such as those with a new diagnosis of T1DM or with type 2 diabetes. , Growth hormone also decreases peripheral glucose uptake and enhances ketone production by increasing FFA release. Elevated concentrations of counterregulatory hormones thus facilitate increased acidosis, hyperglycemia, and dehydration. This, in turn, stimulates further counterregulatory hormone release, creating a vicious cycle resulting in rapid worsening of DKA.

During DKA, intestinal ileus results from potassium depletion, acidosis, and diminished splanchnic perfusion, causing abdominal pain and vomiting, thereby limiting fluid intake. Progressive dehydration eventually leads to diminished tissue perfusion sufficient to cause accumulation of lactic acid, which further contributes to metabolic acidosis. In addition, poor perfusion may result in diminished renal function, limiting the capacity for clearance of glucose and ketones. Ongoing osmotic diuresis and ketonuria in the setting of acidosis also results in urinary losses of electrolytes (potassium, sodium, chloride, calcium, phosphate, and magnesium).

Classical symptoms of DKA include polyuria, polydipsia, polyphagia, weight loss, abdominal pain, nausea, and vomiting. Abdominal tenderness, absence of bowel sounds, and guarding are frequent and may even mimic an acute abdomen. Tachycardia and signs of hypoperfusion, such as delayed capillary refill time and cool extremities, are also common, as well as dry mucous membranes, absence of tears, and poor skin turgor. Despite substantial volume depletion, however, hypotension is unusual in children with DKA. Instead, studies have shown that hypertension occurs frequently in children with DKA and that children with more severe dehydration and acidosis are more likely to be hypertensive during DKA. , The cause of hypertension in children with DKA is unknown. Kussmaul breathing and tachypnea are the result of metabolic acidosis and respiratory compensatory mechanisms. Fruity breath odor (acetone) may be present. Hypothermia has also been described.

Although hyperglycemia is part of the definition of DKA, in rare cases, the serum glucose concentration may be nearly normal, so-called euglycemic DKA . This was previously reported mainly in pregnant women, but has recently been documented in patients with T1DM taking sodium-glucose transporter 2 (SGLT2) inhibitor medications. Normal glucose concentrations or even hypoglycemia despite ketosis may also occur in children with known diabetes who administer insulin to treat DKA prior to arrival in the emergency department. In general, however, the persistence and severity of hyperglycemia reflect the severity of dehydration. In the absence of preexisting renal disease or unusually high carbohydrate intake just prior to presentation, blood glucose concentrations in excess of 500 to 600 mg/dL imply that dehydration is of sufficient severity to diminish the glomerular filtration rate and thereby diminish the capacity for renal clearance of excess glucose.

Concentrations of ketone bodies (β-hydroxybutyrate [βOHB] and acetoacetate [AcAc]) are elevated in DKA, resulting in acidosis. Hyperchloremic acidosis frequently coexists with increased anion-gap acidosis; the anion gap reflects the combination of these processes. The ratio of βOHB:AcAc (typically 1:1 in the normal state) is increased during DKA and may be as high as 10:1. During treatment, this ratio returns to normal. The nitroprusside reaction used to test urine ketone concentrations detects only AcAc and not βOHB. As a result, urine testing cannot be relied on to determine DKA severity or treatment response. Bedside blood ketone meters provide a rapid means for measuring βOHB and may be useful in place of or in addition to urine testing, particularly in patients with anuria or oliguria who produce insufficient amounts of urine for ketone testing. Blood ketone measurements are also useful for determining the timing of transition from continuous intravenous (IV) to intermittent subcutaneous insulin administration. Urine ketones may be present even when blood ketones have normalized as a result of urine stagnating in the bladder.

Hyperglycemia results in fluid shifts from the extravascular to intravascular space and a decrease in serum sodium concentration. This decrease can be calculated as an approximately 1.6 mEq/L decrease in sodium concentration for every 100 mg/dL increase in serum glucose above 100 mg/dL (Na corrected = Na actual + [glucose – 5.5 mmol/L]). , Hyperlipidemia may also contribute to a decrease in measured serum sodium concentrations as a result of a laboratory artifact. However, with modern laboratory techniques, these erroneous measurements are uncommon. Typically, serum potassium concentrations at presentation are in the high-normal range as a result of the redistribution of potassium ions from the intracellular to extracellular space. Several processes are responsible for intracellular potassium depletion, including direct effects of low insulin concentrations, intracellular protein and phosphate depletion, and buffering of hydrogen ions in the intracellular compartment. Intracellular potassium stores may be profoundly depleted, and the serum potassium concentration typically declines rapidly with insulin treatment. Serum phosphate concentrations similarly decrease during treatment.

Leukocytosis is frequent in children with DKA, likely resulting from elevated concentrations of catecholamines and proinflammatory cytokines. In children, new onset of T1DM or insulin omission is a far more common cause of DKA than infection. Therefore, an elevated or left-shifted white blood cell count need not prompt a search for an infectious process in the absence of fever or other symptoms or signs of infection. However, in the presence of fever, careful history, physical examination, and laboratory evaluation to assess for infection are prudent ( eTable 85.1 ).

eTABLE 85.1

Laboratory Monitoring Schedule Recommendation

Laboratory Analysis DKA Confirmed Hour 1 Hour 2 Hour 3 Hour 4 Hour 5 Hour 6 Hour 7 Hour 8 Hour 9 Hour 10 Hour 11 Hour 12
Glucose a X X X X X X X X X X X X X
Sodium b X X X X X X X X
Potassium X X X X
Chloride X X X X
Bicarbonate X X X X
Creatinine X
Magnesium c X X X
Calcium c X X X
Phosphorus c X X X
Blood gas (capillary, venous, or arterial) X X
βOHB b X X X X X X X

If the patient has fever or other symptoms suggesting infection, blood cultures, chest radiograph, and urinalysis should be considered. Bacterial infections, however, are infrequent causes of DKA in children.

βOHB , β-Hydroxybutyrate; BUN, blood urea nitrogen; DKA , diabetic ketoacidosis.

a Repeat glucose every hour while on insulin drip; send first glucose to the lab as well for serum glucose.

b Repeat sodium and βOHB every 2 hours while on insulin drip. When βOHB measurement is not available or is impractical, calculation of the anion gap can also be used to track resolution of DKA.

c May repeat these labs every 4 hours until normalized, especially while on insulin drip.


Frequency of diabetic ketoacidosis at diagnosis

The frequency of DKA at diagnosis varies widely by geographic region, with an overall estimated frequency of approximately 20% to 67%. In the population-based US study SEARCH for Diabetes in Youth, data were collected from self-reported health questionnaires and medical record review. In this study, 36.4% of children and adolescents presented with DKA at the onset of diabetes. , As mentioned, frequency of DKA varies depending on regions: in a collaborative cohort in Germany and Austria, 26.3% and 21.1%, respectively ; in a Swedish data report, 16.9%; and in a Finnish register report, 18.7%. Younger age (<5 years) and female sex were associated with higher likelihood of presenting with DKA. , A delay in diagnosis is associated with a higher likelihood of DKA; two factors contributing to this result are patient age and provider experience. Regions with a higher prevalence of T1DM generally have a lower frequency of DKA, attributed to heightened awareness in providers and thus earlier detection. Interestingly, however, three of The Environmental Determinant of Diabetes in the Young (TEDDY) study participant families who were counseled on diabetes symptoms nonetheless presented with DKA. The nonspecific nature of individual symptoms, such as polyuria, tachypnea, and altered mental status, may cause such symptoms to be misconstrued as urinary tract infection, pneumonia, or meningitis, respectively. Mallare et al. reported a frequency of DKA of 33% in children and adolescents at the initial visit and almost double (59%) in those for whom the diagnosis of diabetes was missed at the initial visit. The diagnosis of diabetes was more likely to be missed in very young children (34% of children ≤5 years compared with 8.5% in those older than 10 years), particularly when these very young children are evaluated by family practitioners rather than pediatricians. Usher-Smith et al. found protective diagnostic factors of a first-degree relative with T1DM, higher parental education, and higher background incidence of T1DM in a systemic review of 46 studies involving 24,000 children from 31 countries.

Frequency of diabetic ketoacidosis in children and adolescents after diagnosis

More data are becoming available describing the incidence of DKA in children and adolescents with established diabetes. Reported frequencies range from 1 to 10 per 100 patient-years. In the Diabetes Control and Complications Trial, the incidence of DKA in adolescents treated with intensive management regimens was 2.8 per 100 patient-years, significantly lower than the incidence in those treated conventionally (4.7 per 100 patient-years). Although this is an older study, it was a time- and resource-intensive study representing a more idealized situation than often encountered in the overall pediatric diabetes population. In a more recent study from the Barbara Davis Center for Childhood Diabetes in Denver, the overall incidence of DKA was 8 per 100 person-years. In that study, factors associated with higher incidence included older age, higher HbA1C (relative risk [RR] of 1.68 per 1% increase in HbA1C in younger children and 1.43 in older children), higher reported insulin dose, Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-4) psychiatric diagnoses, and “underinsurance” reflecting lower socioeconomic status. A multinational study based on registries and audits showed DKA frequency at 5% in Austria and Germany, 6.4% in England and Wales, and 7.1% in the United States. In some studies, DKA is documented to occur with a two- to fivefold increased risk in children and adolescents on continuous subcutaneous insulin infusion therapy (CSII) compared with subcutaneous injections, particularly in the first year of initiation of CSII. , However, lower rates of DKA are achievable for those on CSII with adequate training and resources.

T2DM has been occurring with increasing frequency in older children and adolescents. Certain racial/ethnic groups in the United States are disproportionately affected, including Native Americans, Hispanics, and African Americans. DKA can be the clinical presentation of T2DM in youth, estimated at 5% to 10%. Youth with T2DM may also present with hyperglycemic hyperosmolar syndrome (HHS), also referred to as hyperosmotic hyperglycemic nonketotic coma (described more fully later).

Morbidity and mortality associated with diabetic ketoacidosis

Mortality in children presenting with DKA is approximately 0.25% to 0.30%. Most of the mortality in DKA occurs in children with cerebral edema, accounting for 57% to 87% of deaths. Neurologic sequelae of DKA are described later. Other causes of morbidity and mortality include sepsis and secondary infection, electrolyte abnormalities (e.g., hypokalemia), arrhythmias, rhabdomyolysis, cerebral infarction, thrombosis, pneumomediastinum, subcutaneous emphysema, and pulmonary edema.

Management guidelines ( fig. 85.1 )


Restoration of adequate peripheral perfusion and hemodynamic stability with bolus administration of IV fluids (0.9% saline or other isotonic fluids) should begin as soon as possible. Typical patients require an initial fluid bolus of 10 to 20 mL/kg. Repeated boluses may be necessary if ongoing hemodynamic instability is present. Studies have shown that clinical assessments of dehydration severity in children with DKA tend to be inaccurate. The average degree of dehydration for most patients is approximately 7% to 9% of body weight. This figure should be used as a basis for determining the total volume of fluids to be replaced. The estimated fluid deficit, along with maintenance fluid requirements, should be replaced over 24 to 48 hours using 0.45% to 0.90% saline, generally initially with 0.90% saline, then transitioning to 0.45% saline after several hours. Replacement of ongoing urinary fluid losses is usually unnecessary because osmotic diuresis typically resolves rapidly after beginning DKA treatment. However, in circumstances of persistently high urine output, or profuse vomiting or diarrhea, replacement of ongoing losses may be considered.

• Fig. 85.1

Diabetic ketoacidosis treatment algorithm in children and adolescents. *Lower dose of 0.025–0.05 U/kg/h may be considered in certain circumstances, such as young age. ABCDs, Airway, breathing, circulation, differential diagnosis; βOHB , β-hydroxybutyrate; D10, 10% dextrose; Dx, diagnosis; Endo, Endocrinology; GCS, Glasgow Coma Scale; Kphos/K-acetate, potassium phosphate/potassium acetate; NS, normal saline; SC, subcutaneous.

Whether variations in IV fluid administration protocols might contribute to risk of cerebral injuries resulting from DKA has been a topic of debate. A recent large prospective multicenter study, the Pediatric Emergency Care Applied Research Network (PECARN) Fluid Therapies Under Investigation in DKA (FLUID) Trial, assessed neurologic and neurocognitive outcomes of children with DKA treated with fluid regimens that varied in rate of administration and sodium content. , This study found no differences in acute or postrecovery outcomes of children treated with 0.45% versus 0.90% NaCl content fluids nor in children treated with more rapid versus slower fluid infusions. The study findings suggest that a range of fluid protocols can be safely used to treat children with DKA and that fluid infusions should not be restricted because of concerns about causing cerebral injury (CI). IV fluid administration should be adjusted according to each patient’s hemodynamic state and fluid balance.


Insulin should be administered intravenously via continuous infusion. An infusion rate of 0.1 U/kg per hour is typically used. Lower insulin dosages (0.025–0.05 U/kg per hour) are used in some centers. To date, there have been few studies comparing standard insulin dosages with lower dosages. These studies generally have not found substantive differences in outcomes but have involved small sample sizes or have been retrospective and nonrandomized. , A larger prospective study is necessary.

Insulin administration results in resolution of acidosis and hyperglycemia via suppression of ketogenesis and hepatic glucose output (gluconeogenesis) and promotion of peripheral glucose uptake. An initial bolus or loading dose of insulin is not recommended. Maximal suppression of ketogenesis is achieved rapidly with an insulin infusion (0.05–0.10 U/kg per hour). Even in the absence of insulin administration, the serum glucose concentration often decreases substantially with initial rehydration, reflecting improvements in renal perfusion and decreased counterregulatory hormone concentrations. This decline in glucose concentration during the initial period of rehydration should not be interpreted as indicating excessive insulin administration.

Serum glucose concentrations typically normalize before ketosis and acidosis resolve. To continue insulin administration at dosages sufficient to allow resolution of ketosis, dextrose should be added to the IV fluids. Transition to dextrose-containing fluids should occur when the serum glucose concentrations decline below approximately 250 mg/dL. The “two-bag system” for dextrose administration allows a rapid response to changes in serum glucose concentration and is cost-effective. Two bags of IV fluids with identical electrolyte content but different dextrose content (0% and 10%) are administered simultaneously with the relative rates of administration frequently adjusted to increase or decrease the dextrose concentration while maintaining a constant overall rate of administration of fluid and electrolytes.


Serum potassium concentrations often decline rapidly during treatment and potassium replacement is mandatory. Typical patients require potassium administration at concentrations of 30 to 40 mEq (occasionally, up to 80 mEq) per liter of IV fluids. A combination of potassium salts (potassium chloride, potassium phosphate, or potassium acetate) is often used to decrease chloride administration and replace phosphate deficits. Adequacy of renal function should be considered before administration of potassium.

Phosphate replacement in children with DKA may be beneficial. Theoretically, low 2,3-diphosphoglycerate (DPG) levels in red blood cells may occur in association with hypophosphatemia, leading to decreased tissue oxygen delivery. , Clinical relevance, however, has not been established. Although the risk of hypocalcemia during DKA treatment is increased with phosphate replacement, symptomatic hypocalcemia is uncommon when phosphate is administered slowly and in the more modest concentrations recommended in most DKA treatment protocols. , Severe hypophosphatemia during DKA has been shown to be associated with rhabdomyolysis and hemolytic anemia. , Therefore, monitoring of serum phosphate concentrations is necessary and treatment of severe hypophosphatemia is essential.

Hypomagnesemia and hypocalcemia may also occur during DKA treatment but are generally mild, rarely requiring treatment. However, monitoring of serum calcium and magnesium concentrations is recommended.

Correction of acidosis

Routine bicarbonate administration is contraindicated in children with DKA, as acidosis generally corrects rapidly with insulin and fluid administration, and hemodynamic instability resulting from acidosis is rare. Bicarbonate administration is associated with several possible adverse effects, including an increase in the risk of hypokalemia and a theoretic increase in tissue hypoxia resulting from a leftward shift in the hemoglobin-oxygen dissociation curve. , Paradoxical acidosis of the cerebrospinal fluid (CSF) has also been documented with bicarbonate administration, likely resulting from diminished respiratory drive and a rise in the partial pressure of carbon dioxide (P co 2 ), which readily crosses the blood-brain barrier, generating CSF acidosis. , Bicarbonate administration has also been associated with an increased risk of DKA-related CI. In very rare circumstances (severe hemodynamic instability not responding to standard measures or potentially life-threatening hyperkalemia), bicarbonate administration may be considered.


Intensive monitoring is essential for children with DKA, and most should be treated in a pediatric intensive care unit (ICU) or other unit with similar capacities. , Blood glucose concentrations are typically measured hourly and electrolyte concentrations every 2 to 4 hours. Determination of serum pH (every 2–4 hours) is helpful, particularly because serum bicarbonate concentrations may not increase during the first several hours. Free-flowing venous blood gas samples generally are sufficient and arterial samples are rarely needed. Failure of acidosis to improve during treatment should prompt evaluation of the adequacy of insulin infusion; fluid balance; and a search for other causes, such as hyperchloremia, renal failure, sepsis, or even appendicitis. Monitoring of the βOHB level during DKA is prudent, as its clearance (to <1 mmol/L) is a good indicator of DKA resolution. ,

All fluid intake and output should be accurately recorded. Consideration of fluids and other management that may have occurred prior to admission to a pediatric ICU is important. Vital signs and mental status should be monitored hourly. One study showed a high frequency of prolonged QT interval corrected for heart rate (QTc) in children with DKA, and arrhythmias have been described in rare cases. Therefore, cardiac monitoring is recommended.

Diabetic ketoacidosis–associated complications

CI has been recognized as a complication of T1DM in children since 1936. It is essentially a clinical diagnosis, based on the deterioration of mental state during treatment of DKA. Signs and symptoms that should prompt consideration of CI include severe headache, recurrence of vomiting, irritability, lethargy, or other mental status changes. Some patients progress to coma, respiratory arrest, and cerebral herniation. Most episodes of CI occur several hours after the initiation of DKA treatment. However, 5% to 20% of cases occur at the time of presentation, prior to the initiation of therapy. CI remains the leading cause of death and morbidity in children with T1DM. The frequency of CI associated with DKA remains unchanged despite clinical efforts to the contrary.

The reported mortality from CI varies and is in part dependent on the criteria used to define CI. Rates as high as 50% to 90% have been reported, but more recent studies , report lower rates of 21% to 24%. Overall, the incidence of CI is approximately 0.5% to 0.9% within DKA presentations. In other words, approximately 1 in 600 children with DKA die as a result of CI. Morbidity is significant; in particular, debilitating neurologic sequelae occur in 21% to 26% of children with DKA-related CI. , Although frank CI is uncommon, substantial data suggest that subclinical or asymptomatic CI occurs in many children with DKA, perhaps even in the majority. Recent data suggest that subtle brain injury may also be associated with DKA, even in the absence of clinically apparent CI.

The pathophysiology of CI remains enigmatic; several causative theories have been proposed. Initial theories proposed that rapid osmotic declines during DKA treatment caused brain cell swelling, with increased intracranial pressure leading to CI. Accumulating evidence, however, contradicts this theory, most notably the lack of increase in risk of CI in children in the PECARN FLUID trial who were treated with more rapid infusion of hypotonic fluids. Alternatively, cerebral hypoperfusion (caused by hypocapnia and volume depletion) prior to DKA treatment and the effects of reperfusion during DKA therapy have been proposed to be involved in CI. Direct effects of inflammatory cytokines and other substances that affect blood-brain barrier function have also been proposed to play a role, as well as increased activity of brain ion transporters and activation of brain microglia.

Epidemiologic studies of risk factors for CI show that children with higher initial blood urea nitrogen concentrations, lower initial P co 2 concentrations, and greater acidosis at the time of presentation of DKA seem to be at greatest risk. , , , A lesser rise in measured serum sodium concentration during DKA treatment has also been associated with increased risk of CI, as has treatment with bicarbonate. Early administration of insulin (within the first hour) has also been associated with increased CI risk in one study.

Once the diagnosis of CI is made, treatment is a matter of urgency and should not be delayed while awaiting imaging studies or further testing. IV 20% mannitol (0.25–1 g/kg infused over 15 minutes) should be initiated as soon as possible. Alternatively, hypertonic (3%) saline (2.5–5 mL/kg over 10–15 minutes) has been used. , One report suggested the possibility of poorer outcomes with 3% saline in comparison with mannitol. Therefore, until additional data are available, mannitol should likely remain the first line of treatment. Ongoing ICU monitoring is essential. Respiratory support by means of endotracheal intubation is likely to be required due to severe alterations in mental status and impaired airway reflexes. Therapeutic hyperventilation in intubated patients, however, has been associated with poorer outcomes. Therefore, decreasing P co 2 below the patient’s own compensation for metabolic acidosis should be avoided in children with DKA except when absolutely necessary to treat impending cerebral herniation. A reasonable approach would be to initially maintain the patient’s current P co 2 level and then gradually allow the P co 2 to increase as acidosis corrects. Central nervous system (CNS) imaging in patients with suspected CI should be used to exclude other etiologies of altered mental status, such as CNS thrombosis or infarction. Imaging may not be necessary if there are improvements after CI therapy.

Hyperglycemic hyperosmolar syndrome

HHS is characterized by extreme elevations in serum glucose (>600 mg/dL) and hyperosmolarity (serum osmolarity >330 mOsm/kg) in the absence of significant ketosis or acidosis (urine ketone concentration <1.5 mmol/L [negative or “trace” on urine dipstick] and serum bicarbonate >15 mEq/L). Although HHS is defined as a condition separate from DKA, 30% of cases occur in combination with substantial ketosis and acidosis, meeting criteria for both HHS and DKA. Until recently, HHS was thought to occur infrequently in pediatrics. An increase in case reports of HHS in children suggests that the frequency may be increasing. , As in adults, HHS in children has a relatively high mortality of 10% to 35%. , The majority of HHS reports in children are in patients who have acanthosis nigricans, are obese, are African American, and have a family history of T2DM. Most cases of HHS are the initial presentation of diabetes, and most of these youths will subsequently have a clinical diagnosis of T2DM.

The occurrence of HHS during DKA poses challenges in terms of recognition and treatment. Generally, dehydration is more profound than the clinical assessment would suggest, reflecting difficulties in clinical evaluation due to obesity and relative preservation of intravascular volume due to hyperosmolarity. Electrolyte losses similarly exceed those of DKA as a result of more prolonged osmotic diuresis. Patients who meet criteria for both DKA and HHS require more prolonged and aggressive fluid and electrolyte replacement therapy than typical children with DKA. Replacement of ongoing urinary losses may be necessary. Frequent reassessment of circulatory status and fluid balance is critical. A high frequency of thrombosis has been described in children with HHS as well as rhabdomyolysis and a malignant hyperthermia-like syndrome. , CI appears to be a rare complication of HHS.

Thrombotic complications

Thrombotic complications are common in children with DKA who have central venous catheters. , Cerebral thrombosis and pulmonary emboli have also been described. Hyperosmolarity may result in direct osmotic disruption of endothelial cells, leading to a release of tissue thromboplastins. Higher levels of vasopressin stimulated by hypertonicity and decreased vascular volume may also contribute to enhanced coagulation. Prophylaxis with low-dose heparin should be considered for children with central venous lines. , ,

Other complications

Acute kidney injury (AKI) is common in children with DKA. One recent study found that 64% of children hospitalized for DKA had AKI. Of children who developed AKI during DKA, 65% had stage 2 or stage 3 AKI, suggesting renal tubular injury rather than simply prerenal azotemia. Greater acidosis and volume depletion are risk factors for AKI. Development of renal failure during DKA is rare—most children return to normal renal function after DKA resolves. However, it is unclear whether the occurrence of AKI during DKA might increase the risk of future diabetic kidney disease.

Rhabdomyolysis , is potentially life-threatening. It is characterized by elevated serum creatine kinase, lactate dehydrogenase, and alanine aminotransferase concentrations due to muscle injury. Rhabdomyolysis may result in renal failure, compartment syndrome, severe hyperkalemia, and other electrolyte disorders leading to arrhythmias. Hyperosmolarity has been thought to be one causative factor; the risk is higher in children who have DKA complicated by features of HHS. Rhabdomyolysis may also result from severe hypophosphatemia during DKA.

Acute pancreatitis has been described in case reports of both children and adults with DKA, but it occurs rarely. Far more frequent are benign elevations in serum amylase or lipase, occurring in 24% to 40% of children with DKA. These elevated pancreatic enzyme concentrations typically normalize rapidly with DKA treatment and are not associated with clinical features of pancreatitis.

Although neurologic deterioration in children with DKA is most frequently caused by diffuse CI/cerebral edema, focal cerebral infarctions with and without hemorrhage and cerebral thrombosis have also been described. Other rare complications of DKA in children include pulmonary edema, cardiac arrhythmias, renal failure, intestinal necrosis, and rhinocerebral mucormycosis. ,

Neuropsychologic sequelae

Adverse neurodevelopmental outcomes in children with diabetes have previously been attributed to recurrent hypoglycemia. However, recent data suggest that chronic hyperglycemia and hyperglycemic extremes such as occur with DKA may play a more central role in causing neuropsychologic sequelae. Short-term effects on neurocognitive performance involving complex skills, such as inhibiting an overlearned response and learning of complex novel information, have been described. Long-term deficits in attention, memory, and IQ , have also been found to be associated with DKA, as have alterations in cerebral microstructure.

Healthcare costs associated with diabetic ketoacidosis

Healthcare costs for DKA vary by geographic regions in the United States. Comparing costs is also often complicated by variations in healthcare systems, methods of reporting costs (e.g., hospital costs vs. payer costs), and contractual arrangements. Several studies document that admissions for DKA are an important driver of diabetes-related healthcare costs in the United States. In addition, among patients with known diabetes admitted for DKA, readmissions within 1 year of discharge are common. One strategy to decrease the frequency of DKA is to promote awareness in the general population, in communities, and among providers. A successful campaign to heighten awareness of signs and symptoms of DKA took place in Parma, Italy. In the Parma campaign, simple messages regarding signs and symptoms of diabetes were provided to practitioners and schools, and free access to care was arranged. Compared with neighboring regions, where the frequency of DKA was quite high at 78%, the Parma region observed a very low frequency of 12.5% during the 8 years of the campaign. Of note, the campaign was relatively inexpensive, costing $23,000 for the 8 years of the campaign.

Another more targeted strategy is linked to recognition of diabetes risk. In children enrolled in prevention studies (mainly siblings of probands with T1DM), the frequency of DKA is far less than that of the general population; less than 4% of those participating in the Diabetes and Prevention Trial 1 presented in DKA, and 63.3% were asymptomatic. For children and adolescents with diagnosed diabetes, multidisciplinary and intensive team management approaches have been shown to decrease the frequency of DKA. Unfortunately, obtaining sufficient reimbursement for intensive case management in the United States has been challenging, despite demonstrated savings to the healthcare system. In a relatively small study, the costs for emergency and hospital visits for those not involved in intervention more than exceeded (125%) the costs of intensive case management. This included only hospital charges and did not include additional costs, such as missed days of work and school, patient and family anxiety, and cognitive and long-term health impact of recurrent DKA and poor glycemic control. In a larger study of home-based psychotherapy for adolescents with poorly controlled diabetes, admissions for DKA were reduced by almost half over a 2-year period, resulting in a cost savings of $23,886 to $72,226 (the range reflecting hospital costs and third-payer costs, respectively). These examples emphasize the need for preventive rather than crisis-based approaches to the pediatric diabetes population.

Management of DKA with utilization of βOHB measurement has been shown to reduce overall care costs during hospitalization by reducing length of stay in the ICU, overall time in hospitalization, reduced laboratory measurements, and clinical assessments.

Very young children with diabetes are the most likely to present in DKA and constitute the age group with the most rapid rise in incidence of diabetes. These data suggest that there are important opportunities for prevention strategies in this age group. Major efforts are needed to address healthcare disparities overall in children with diabetes, and prevention of DKA is no exception.

Key references

Only gold members can continue reading. Log In or Register to continue

May 20, 2021 | Posted by in RHEUMATOLOGY | Comments Off on Diabetic ketoacidosis

Full access? Get Clinical Tree

Get Clinical Tree app for offline access