Sodium-glucose cotransporter-2 (SGLT2) inhibitors are a class of antidiabetic medications that block sodium and glucose reabsorption at the SGLT2 protein located in the proximal convoluted tubule of the kidneys, resulting in increased urinary glucose excretion and reduced plasma glucose concentrations. The SGLT2 protein is responsible for around 90% of glucose reabsorption, while the SGLT1 protein, located at the distal part of the proximal convoluted tubule, reabsorbs the remainder.1 The increased urinary glucose excretion from SGLT2 inhibitors exerts a diuretic effect on the body, which has been theorized to explain the cardiovascular benefits seen with their use.2 There are currently 6 SGLT2 inhibitor medications approved by the US Food and Drug Administration (FDA) for the treatment of type 2 diabetes mellitus, the reduction of cardiovascular disease risk and chronic kidney disease progression, and the treatment of heart failure. Information regarding dosing and indications for each agent including sotagliflozin, the only nonselective SGLT1 and SGLT2 inhibitor on the market, is listed in Table 1.3–8 Common warnings and significant adverse reactions associated with SGLT2 inhibitor use include genitourinary fungal and bacterial infections, necrotizing fasciitis and lower limb amputations, and ketoacidosis.Historically, SGLT2 inhibitor medications were used primarily as adjunctive agents for chronic glycemic control in patients with type 2 diabetes mellitus. As with most oral glucose–lowering agents, SGLT2 inhibitors are often avoided in patients with critical illness given their impaired elimination in patients with kidney dysfunction, resulting in increased risk for adverse effects such as hypotension and hypovolemia. Therefore, short-acting insulin is commonly used in place of oral agents within critical care settings for acute glycemic control. Resuming or initiating SGLT2 inhibitors during hospitalization can be considered depending on resolution or confirmed absence of end-organ dysfunction, hemodynamic instability, and infection.With the publication of numerous clinical trials studying the use of SGLT2 inhibitors in various disease states over the past 10 years, indications for SGLT2 inhibitors have expanded. Sodium-glucose cotransporter-2 inhibitors have shown consistent morbidity and mortality benefits in those who have or are at risk of cardiovascular disease, chronic kidney disease, and most recently, congestive heart failure. The 2022 Guideline for the Management of Heart Failure by the American Heart Association, American College of Cardiology, and Heart Failure Society of America now strongly recommends the initiation of SGLT2 inhibitors as part of guideline-directed medical therapy in patients with heart failure with either reduced ejection fraction or preserved ejection fraction, regardless of diabetes history; this recommendation is based on the role of SGLT2 inhibitors in reducing cardiovascular death and hospitalizations.9,10 With their incorporation into guideline-directed medical therapy, use of SGLT2 inhibitors has dramatically increased over the past few years. In the United States between 2013 and 2021, SGLT2 inhibitor use increased by over 9% and 18% in patients with heart failure with preserved ejection fraction and those with reduced ejection fraction, respectively.11 As SGLT2 inhibitor use has increased, so has the incidence of adverse effects such as diabetic ketoacidosis.Diabetic ketoacidosis is a serious complication that occurs more commonly in patients with type 1 diabetes mellitus than in those with type 2 diabetes mellitus. It is characterized by absolute or relative insulin deficiency in combination with increased counterregulatory hormones such as catecholamines, glucagon, and cortisol. This initial process results in lipolysis, hepatic fatty acid oxidation to ketone bodies, hyperglycemia due to increased glycogenolysis and gluconeogenesis, and impaired glucose utilization by tissues (Figure 1). While the pathophysiology and mechanism behind diabetic ketoacidosis is well understood, euglycemic diabetic ketoacidosis associated with SGLT2 inhibitor use has not been characterized as thoroughly because of the gradual increase of its incidence. It is theorized that inhibited reabsorption of sodium and glucose at the proximal convoluted tubule may lead to hypovolemia because of the natriuretic and diuretic effects of SGLT2 inhibitors. Glucagon production is also upregulated with SGLT2 inhibitor use and therefore shifts the body into a state of ketogenesis, resulting in euglycemic diabetic ketoacidosis.Diagnosis of diabetic ketoacidosis involves 3 criteria: (1) presence of hyperglycemia (plasma glucose ≥ 200 mg/dL) or prior history of diabetes irrespective of the glucose level; presence of ketosis (β-hydroxybutyrate concentration ≥ 3.0 mmol/L or urine ketone strip 2+ or greater); and presence of metabolic acidosis (pH < 7.3 and/or bicarbonate CO2 concentration < 18 mmol/L).13 These criteria, specifically the previous blood glucose threshold of < 250 mg/dL, have recently been modified and updated in the 2024 American Diabetes Association (ADA) consensus report to account for the increased use of SGLT2 inhibitors and their associated risk of euglycemic diabetic ketoacidosis, defined as a blood glucose level less than 200 mg/dL in addition to the other typical signs of diabetic ketoacidosis.12,14 Euglycemic diabetic ketoacidosis currently accounts for approximately 10% of all diabetic ketoacidosis cases. Current guidelines report that among patients taking an SGLT2 inhibitor who present with diabetic ketoacidosis, up to one-third had a glucose level below 200 mg/dL, and 71% had a level below 250 mg/dL.15Diabetic ketoacidosis can be further classified into mild, moderate, and severe categories based on the magnitude of metabolic acidosis, ketone levels, and mental status, as shown in Table 2. Although patients may not fit all criteria to be defined as mild, moderate, or severe, these categories may be useful in guiding level of care for patients and treatment modalities. For example, a clinician may choose to manage a patient with mild diabetic ketoacidosis outside of the intensive care unit (ICU) with subcutaneous (SQ) insulin rather than an intravenous (IV) insulin infusion. Ultimately, however, these decisions may need to be individualized on the basis of the patient’s clinical presentation and the provider’s clinical judgment.12The updated consensus report has also moved away from relying on the anion gap to assess diabetic ketoacidosis severity, acknowledging that many patients exhibit complex acid-base disturbances. These imbalances may arise from volume depletion and electrolyte loss, vomiting-induced alkalosis, and hyperventilation-related alkalosis as seen in Kussmaul breathing.12Diabetic ketoacidosis can occur at any age but is most frequently observed in young adults aged 18 to 44 years with type 1 diabetes mellitus. Notably, up to 21% of adults may present with diabetic ketoacidosis as their initial manifestation of type 2 diabetes mellitus.12 Several risk factors are associated with the development of diabetic ketoacidosis, including new-onset type 1 diabetes mellitus, prolonged fasting or carbohydrate restriction, alcohol or substance use, and the use of certain medications such as corticosteroids, immune checkpoint inhibitors, and SGLT2 inhibitors. The most common precipitants of diabetic ketoacidosis are infection and omission or inadequate administration of insulin.12 Euglycemic diabetic ketoacidosis appears to occur more frequently in individuals with type 1 diabetes mellitus who are using an SGLT2 inhibitor, although reported incidence varies across studies. Research indicates that patients with type 1 diabetes mellitus using an SGLT2 inhibitor have a 5- to 17-fold increased risk of diabetic ketoacidosis compared with those not taking these medications. In patients with type 2 diabetes mellitus on SGLT2 inhibitor therapy, the risk of diabetic ketoacidosis is roughly doubled.12 In addition to the standard risk factors for diabetic ketoacidosis, specific risks associated with SGLT2 inhibitor use include reduced oral intake/prolonged fasting, hepatic dysfunction, and underlying autoimmune conditions.13 There are many overlapping risk factors between diabetic ketoacidosis and euglycemic diabetic ketoacidosis, such as infection, pancreatitis, and chronic liver disease. Patients undergoing surgical procedures also are at high risk of SGLT2 inhibitor–induced euglycemic diabetic ketoacidosis, and discontinuation of therapy prior to scheduled surgical procedures, ideally 3 to 4 days, is recommended.15The management of diabetic ketoacidosis centers on fluid and electrolyte replacement, insulin therapy, and identification and treatment of the underlying precipitating factors. The 2009 ADA report does not address SGLT2 inhibitor–associated diabetic ketoacidosis, because the report predates the first FDA-approved drug from this class.14 The 2024 ADA report acknowledges the occurrence of euglycemic diabetic ketoacidosis and suggests adding 5% or 10% dextrose to crystalloid fluids at the initiation of insulin therapy, given that blood glucose levels are typically below 200 mg/dL and these patients are at a high risk of hypoglycemia at baseline.12 Aside from this adjustment, treatment targets remain consistent with standard diabetic ketoacidosis management and are detailed further below and in Figure 2.Patients with mild diabetic ketoacidosis who are hemodynamically stable, cognitively intact, and able to receive close nursing supervision may be managed in the emergency department or an inpatient step-down unit. Management in these settings includes frequent blood glucose monitoring, SQ insulin administration, and fluid replacement. Ideally, blood glucose levels should be checked every 1 to 2 hours, while laboratory assessments—including electrolytes, pH, and β-hydroxybutyrate—should be performed every 4 hours to monitor for diabetic ketoacidosis resolution. This approach may help reduce ICU admissions, which are associated with higher health care costs. In contrast, patients with severe diabetic ketoacidosis or a critical illness as the underlying cause should be managed in the ICU.12Although patients with diabetic ketoacidosis may exhibit varying degrees of hypovolemia, fluid therapy serves a key role beyond initial resuscitation and restoration of intravascular volume. Adequate fluid resuscitation helps to increase tissue and renal perfusion, enhancing renal clearance of glucose and ketones and correcting electrolyte abnormalities. Reducing counterregulatory hormones with fluid therapy can also improve insulin sensitivity and further contribute to a decline in glucose levels. Patients with moderate to severe hypovolemia may require more aggressive fluid resuscitation while those with cardiac compromise (eg, heart or kidney failure) may benefit from more cautious repletion with small fluid boluses and frequent hemodynamic monitoring.12Guidelines recommend using either 0.9% normal saline or balanced crystalloids, such as Ringer lactate, for initial resuscitation due to the fact that safety and efficacy data have not shown consistent benefit for one agent over the other. However, some studies favored Ringer lactate, as its use was linked to faster resolution of diabetic ketoacidosis and a shorter hospital length of stay.12 An important consideration when using 0.9% normal saline is the potential for patients to develop hyperchloremic normal anion gap metabolic acidosis, which may occur with the use of large volumes. An initial infusion rate of 500 to 1000 mL/h during the first 2 to 4 hours is recommended with further fluid replacement dependent on the patient’s hemodynamic status and intake-output balance. Dextrose should be added once the blood glucose level falls below 250 mg/dL to prevent hypoglycemia and enable continued insulin administration, since hyperglycemia typically improves earlier than ketoacidosis. As stated previously, dextrose-containing fluids should be administered immediately in the management of euglycemic diabetic ketoacidosis, because the blood glucose level tends to be less than 200 mg/dL upon diagnosis, putting patients at increased risk for hypoglycemia.12Due to the complexity of diabetic ketoacidosis management, many institutions have sought to develop standardized protocols to streamline care, enhance patient safety, and improve outcomes. Some centers have implemented fully nursing-driven protocols covering all aspects of diabetic ketoacidosis treatment, while others have adopted hybrid approaches tailored to their available resources and staffing. One increasingly popular strategy is the 2-bag method, which uses 2 IV fluid bags—containing varying concentrations of dextrose and electrolytes—to facilitate continuous insulin and fluid administration. This approach enables rapid adjustment of dextrose infusion rates while maintaining stable insulin delivery, which is beneficial in later stages of diabetic ketoacidosis management, as hyperglycemia resolves more quickly than acidosis. Several studies have demonstrated that implementing this method is associated with a shorter time to anion gap closure and faster resolution of acidosis.16–19Patients may also need electrolyte repletion and monitoring due to imbalances in potassium and phosphate levels. Although patients may have depleted total body potassium due to volume losses, they may present with a normal or high serum potassium level due to acidosis and extracellular shift of potassium with insulin deficiency. As acidosis resolves with insulin and fluid therapy, patients may experience hypokalemia. Therefore, potassium repletion, either through maintenance fluids or as separate supplementation, is recommended to maintain the potassium level within the range of 4 to 5 mmol/L. Detailed dosing guidelines are provided in Figure 2. Patients may develop a low phosphate level due to extracellular shifts and urinary losses; however, phosphate repletion is not routinely recommended unless there is evidence of deficiency, such as a muscle weakness with serum phosphate level below 1.0 mmol/L. Bicarbonate repletion should only be considered in cases of severe acidosis (pH < 7.0) due to potential vascular side effects. In other situations, bicarbonate repletion has not been shown to improve clinical outcomes and may increase the risk of hypokalemia and cerebral edema.12 Patients may also present with hyponatremia due to fluid shift from intracellular to extracellular space in the presence of hyperglycemia, although this effect may be less pronounced in patients with euglycemic diabetic ketoacidosis.20Insulin therapy is the mainstay of treatment for diabetic ketoacidosis. Intravenous regular, short-acting insulin is typically preferred due to its rapid onset, short half-life, and ease of titration; however, it does require dedicated IV catheter access. Although short-acting insulin infusions are generally preferred for more severe cases of diabetic ketoacidosis that require ICU-level care, administration protocols may differ between institutions. The infusion is typically initiated at a rate of 0.1 U/kg/h as a continuous infusion, although nurse-driven protocols with variable dosing may also be used. Once dextrose is introduced for a blood glucose level below 250 mg/dL, the insulin infusion rate should be reduced to 0.05 U/kg/h and maintained to target a blood glucose level of 150 to 200 mg/dL for the duration of treatment until resolution. For patients presenting with euglycemic diabetic ketoacidosis, patients may be initiated on an insulin infusion at a reduced rate of 0.05 U/kg/h to avoid hypoglycemia.Subcutaneous insulin may be less favored in moderate to severe diabetic ketoacidosis due to the drug’s delayed onset and prolonged duration of action. Although the 2009 ADA report does not recommend using SQ insulin until the patient’s serum glucose level reaches 200 mg/dL,14 the updated 2024 ADA report supports the use of SQ rapid-acting insulin as initial therapy for mild to moderate diabetic ketoacidosis.12 Patients with mild or moderate euglycemic diabetic ketoacidosis may also be treated with SQ rapid-acting insulin (eg, 0.1 U/kg every 2 h), although no specific agent, route, or dosage has been formally recommended for this patient population.The 2009 ADA report14 defines diabetic ketoacidosis resolution as a blood glucose level less than 200 mg/dL in addition to the presence of 2 of the following criteria: (1) a serum CO2 level of 15 mEq/L or more, (2) a venous pH greater than 7.3, or (3) an anion gap of 12 mEq/L or less. However, the 2024 ADA report no longer recommends use of the anion gap as a criterion for resolution; the use of plasma ketone levels is instead suggested.12 Relying on the anion gap can be misleading, particularly in patients with hyperchloremic metabolic acidosis, which may develop after administration of large volumes of 0.9% normal saline. The presence of hyperchloremic metabolic acidosis can lead to a normal anion gap acidosis and potentially delay the transition to SQ insulin if misinterpreted as ongoing ketoacidosis.12,14 As a result, the 2024 ADA consensus statement defines diabetic ketoacidosis resolution as a blood glucose level of less than 200 mg/dL, a serum CO2 of 18 mmol/L or more or a pH of 7.3 or higher, and plasma ketones less than 0.6 mmol/L12 (see Table 3 for key changes between the 2 guidelines). Although the anion gap is no longer recommended as a first-line diagnostic or resolution criterion, institutions may choose to use it in situations where plasma ketone measurements are unavailable or are obtainable but may not be rapidly available to guide transition to SQ insulin in a prompt manner. In such cases, clinicians should exercise judgment, particularly in accounting for the effects of hyperchloremic metabolic acidosis when determining diabetic ketoacidosis resolution. Patients presenting with SGLT2 inhibitor– associated diabetic ketoacidosis should hold the offending agent during diabetic ketoacidosis treatment and resolution. Resuming use of SGLT2 inhibitors after treatment for euglycemic diabetic ketoacidosis is not encouraged if no other risk factors were attributed to the event. However, this decision should be discussed with all relevant stakeholders, physicians, and/or advanced practice providers regarding the risks and benefits of stopping SGLT2 inhibitor therapy given its role in guideline-directed medical therapy for various cardiorenal conditions. Resumption of SGLT2 inhibitor therapy in patients who have experienced diabetic ketoacidosis should be only when necessary and approached with caution; a clear precipitating factor for the diabetic ketoacidosis should be identified and addressed and modifiable risk factors minimized (eg, adherence to diabetes sick day guidelines). If the SGLT2 inhibitor was the sole identifiable trigger or no precipitant can be determined, reinitiation may not be advised.21 Patients should be advised to discontinue SGLT2 inhibitor therapy 3 to 4 days before scheduled surgery and during periods of critical illness or prolonged fasting to reduce the risk of euglycemic diabetic acidosis.15As the use of SGLT2 inhibitor therapy increases in both inpatient and outpatient settings, it is essential for clinicians to recognize and identify risk factors for euglycemic diabetic ketoacidosis while optimizing treatment outcomes. Patients receiving SGLT2 inhibitor therapy should be educated on the signs and symptoms of diabetic ketoacidosis as well as prevention strategies, especially if they have other risk factors such as a ketogenic diet or scheduled surgical procedures. Treatment for SGLT2 inhibitor–induced diabetic ketoacidosis similarly mirrors traditional diabetic ketoacidosis management strategies and often requires ICU admission because of the need for continuous IV insulin infusion. However, given the increased use of short-acting IV insulin for mild to moderate cases of diabetic ketoacidosis, this insulin may be used to treat euglycemic diabetic ketoacidosis as well, thus reducing ICU resource utilization and ultimately patient cost. Institutions that use step-down or intermediate care units and are familiar with more frequent glucose monitoring or insulin administration may be capable of handling cases of euglycemic diabetic ketoacidosis, especially with early identification and diagnosis.
Hu et al. (Fri,) studied this question.