New protocol improves diabetic acidosis management in ED

ORLANDO – A new protocol for managing mild diabetic ketoacidosis in the emergency department led to a substantially improved discharge rate at a large teaching hospital.

Prior to development of the Emergency Valuable Approach and Diabetes Education (EVADE) protocol, almost all patients who presented with diabetic ketoacidosis (DKA) were admitted to the hospital, and most of those were admitted to the intensive care unit. However, in the most recent month for which outcomes data for the EVADE protocol are available, 24 of 106 patients (23%) managed according to the protocol were discharged from the ED and 82 were admitted, Marianne Chojnicki, a registered nurse and certified diabetes educator at the Joslin Diabetes Center, Boston, reported at the annual meeting of the American Association of Diabetes Educators.

The average length of stay in the ED for those who were discharged was 20 hours. The average length of stay in the hospital for 23 patients with a non-ICU admission was 136 hours, and the average length of stay in the ICU for 59 patients with an ICU admission was 33 hours, followed by 77 hours in the hospital after ICU discharge.

The findings have important implications for the care of patients presenting with DKA, which is documented in up to 9% of all hospital discharge summaries among patients with diabetes, and which is associated with mortality of up to 2%, Ms. Chojnicki noted.

"That’s a significant number of lives that are lost each year from a preventable event," she said.

In an effort to improve the discharge rate to at least 10%, Joslin Diabetes Center physicians, nurses, and diabetes educators worked together with Beth Israel Deaconess Medical Center in Boston to develop the EVADE protocol, which stemmed from a clinical approach used to manage DKA, and which is initiated in the ED based on clinical triggers.

They thought that a significant number of admissions were avoidable in a well-managed, coordinated system. The goal was to provide optimal cost-effective management for DKA, Ms. Chojnicki said.

From the outset, diabetes education was recognized as a key component of this protocol, she noted.

The education is individualized and covers multiple topics, such as identification of the circumstances leading up to DKA, a review of insulin action and technique of administration, and a review of ways to prevent DKA recurrence.

Evaluating the EVADE protocol

To evaluate the effectiveness of the protocol, a system was put in place to identify all patients who presented to the ED with blood glucose higher than 300 mg/dL and bicarbonate of 20 mEq/L or less. The patients were flagged for enrollment to be managed according to the EVADE protocol if their status was determined to be caused by DKA, and if they consented.

This process began in September 2013, and the first several months were spent refining management and defining quality indicators.

Month 8 results suggest that the protocol is effective for increasing discharge rates and reducing ICU admissions.

"Our initial goal was to discharge 10%, so we are above goal," Ms. Chojnicki said.

Overall, patients with mild to moderate DKA are going home from the ED, she said.

Typically, those with moderate DKA are admitted if they have a secondary diagnosis, and those patients are going to non-ICU beds.

"Only the sickest of the sick – basically those who are acidotic – are the patients going to the ICU," she said.

Preliminary data from an ongoing review of the results by a health economist show that despite the fact that the sickest patients are going to the ICU, the number of ICU days is decreasing in the wake of the EVADE protocol implementation.

This seems counterintuitive, but the latest data show that before EVADE, the average time spent in the ICU was 47.8 hours, and after EVADE the average time spent in the ICU was 31.6 hours.

"We shaved off 16.2 hours in one of the highest-cost areas of the hospital," Ms. Chojnicki said.

Notably, only one patient who was treated according to the EVADE protocol returned to the ED within 30 days, and that was a patient with issues related to alcohol abuse, she noted.

The findings suggest that mild DKA can be optimally managed in the setting of identification of the triggering factor, stabilization of glucose and electrolyte imbalances, patient education, and close follow-up, and that such management allows for more patients to be discharged from the ED, she said.

Ms. Chojnicki reported having no disclosures.

The EVADE protocol, step by step

The EVADE protocol for eligible patients begins with initiation of intravenous fluids but withholding insulin until initial labs are available. Once lab results confirm diabetic ketoacidosis, the staff initiates IV insulin.

A team including an attending endocrinologist, two nurse practitioners, and a certified diabetes educator are alerted when a patient is enrolled into the protocol.

Patients with all forms of DKA, including mild DKA without an anion gap, are managed on insulin drip with an endpoint goal of reaching a bicarbonate level of 21 mEq/L.

"The decision [to keep the patient in the ED for further observation and treatment or admit to the ICU] is based on the severity of the DKA and/or existing concurrent medical illness. If admitted to the ICU, the patient remains in the ICU arm of the DKA protocol, which involves frequent monitoring of glucose, electrolytes, venous blood gas, and urine output. Treatment is aimed at correcting the DKA and associated electrolyte disturbances," explained Maria Koen, a nurse practitioner and certified diabetes educator at Joslin Diabetes Center.

Regardless of whether the patient is admitted or remains in the ED, treatment involves the same IV insulin algorithm, which calls for no initial bolus; the starting rate is based on the initial finger stick, Ms. Koen noted.

Other features of the protocol include hourly finger sticks with adjustment of the dose with a goal of a target blood glucose of 100-180 mg/dL. Every 3 hours, someone on the team assesses venous blood gas and electrolytes.

IV fluid is run concurrently with IV insulin. Once the glucose drops to less than 250 mg/dL, fluid is changed to 5% dextrose in normal saline (D5NS). Once glucose increases again to 250 mg/dL, the protocol calls for discontinuation of D5NS with continuation of IV insulin infusion until the acidosis resolves.

If severe hypoglycemia occurs, D50 (50% dextrose) can be given as a bolus to reach the target of 100-180 mg/dL. The insulin drip is then resumed at 50% of the rate prior to the hypoglycemic event.

"Generally, we use normal saline or lactated Ringer’s as a first choice for IV fluid, at a rate of approximately 1 liter an hour for the first 4 hours, and this may be adjusted for patients who are at risk for volume overload, such as elders or patients with chronic kidney disease or congestive heart failure," she said.

After the first 4 hours, the IV solution may be changed to half of the normal saline, or half of the D5, if blood glucose drops below 250 mg/dL, at a rate of 125-250 cc per hour.

IV fluids may also be discontinued at this point if the patient is able to tolerate oral fluids.

The protocol also calls for electrolyte repletion as needed. Potassium may be added to each liter of IV fluid if the serum potassium reaches less than 5 mEq/L and the patient is voiding. The amount added is at the discretion of the attending physician, Ms. Koen said.

"Keep in mind that as the acidosis resolves, serum potassium will decrease. In such events, if severe hypokalemia occurs, it may be necessary to either decrease the insulin rate or temporarily hold the IV insulin drip," she said, noting that oral potassium and IV or oral phosphate can also be given once a patient can tolerate oral intake.

Insulin pumps should be removed to avoid overlap with the IV insulin, and the team should assess pump function and self-management.

Urine output and ability to tolerate oral intake should be assessed throughout the protocol, and patients should eat as soon as they are able to tolerate oral intake. The team should assess clinical status every 4 hours.

The patients’ care can move beyond the strictures of the protocol once they are either admitted to the ICU or stabilized in the ED. Those not admitted to the ICU can be discharged to home once their anion gap closes, their bicarbonate reaches more than 20 mEq/L, they can tolerate oral intake, and they successfully transition to a subcutaneous insulin regimen.

The transition to subcutaneous insulin requires that the basal dose be determined based on the patient’s home long-acting insulin dose or adjusted based on the nurse practitioner’s or doctor’s recommendations. Dosing for patients who are newly diagnosed or insulin naïve is calculated based on body weight in kilograms x 0.2.

When the patient is able to tolerate oral intake, prandial subcutaneous insulin based on the home prandial insulin dose should be initiated. Dosing in those who are insulin naïve should be based on body weight in kilograms x 0.2 and divided by 3 (but only with team recommendation).

"When transitioning from the IV insulin drip to subcutaneous insulin, if only basal insulin is given we recommend an overlap of the IV insulin drip for 2 hours. If both basal and prandial insulin are given at the same time, the recommendation is to overlap the insulin drip for at least 1 hour, and then the drip may be turned off," Ms. Koen said.

Insulin pumps may be restarted under the guidance of the certified diabetes educator. Electrolytes and glucose should be reassessed 2 hours after the insulin drip is stopped to ensure there is no return of the DKA.

Any patients with pH below 7.3, bicarbonate below 21 mEq/L, or an inability to tolerate oral intake will remain in the ED for continued observation and management. The patient is ready for discharge once pH is greater than 7.3, the anion gap closes, glucose reaches 250 mg/dL, and the patient can tolerate oral intake and has transitioned to subcutaneous insulin.

Following discharge, someone on the team should phone the patient, ideally within 24 hours, to ensure compliance with the postdischarge plan and schedule a follow-up clinic visit.

Ms. Koen reported having no disclosures.

ORLANDO – A new protocol for managing mild diabetic ketoacidosis in the emergency department led to a substantially improved discharge rate at a large teaching hospital.

Prior to development of the Emergency Valuable Approach and Diabetes Education (EVADE) protocol, almost all patients who presented with diabetic ketoacidosis (DKA) were admitted to the hospital, and most of those were admitted to the intensive care unit. However, in the most recent month for which outcomes data for the EVADE protocol are available, 24 of 106 patients (23%) managed according to the protocol were discharged from the ED and 82 were admitted, Marianne Chojnicki, a registered nurse and certified diabetes educator at the Joslin Diabetes Center, Boston, reported at the annual meeting of the American Association of Diabetes Educators.

The average length of stay in the ED for those who were discharged was 20 hours. The average length of stay in the hospital for 23 patients with a non-ICU admission was 136 hours, and the average length of stay in the ICU for 59 patients with an ICU admission was 33 hours, followed by 77 hours in the hospital after ICU discharge.

The findings have important implications for the care of patients presenting with DKA, which is documented in up to 9% of all hospital discharge summaries among patients with diabetes, and which is associated with mortality of up to 2%, Ms. Chojnicki noted.

"That’s a significant number of lives that are lost each year from a preventable event," she said.

In an effort to improve the discharge rate to at least 10%, Joslin Diabetes Center physicians, nurses, and diabetes educators worked together with Beth Israel Deaconess Medical Center in Boston to develop the EVADE protocol, which stemmed from a clinical approach used to manage DKA, and which is initiated in the ED based on clinical triggers.

They thought that a significant number of admissions were avoidable in a well-managed, coordinated system. The goal was to provide optimal cost-effective management for DKA, Ms. Chojnicki said.

From the outset, diabetes education was recognized as a key component of this protocol, she noted.

The education is individualized and covers multiple topics, such as identification of the circumstances leading up to DKA, a review of insulin action and technique of administration, and a review of ways to prevent DKA recurrence.

Evaluating the EVADE protocol

To evaluate the effectiveness of the protocol, a system was put in place to identify all patients who presented to the ED with blood glucose higher than 300 mg/dL and bicarbonate of 20 mEq/L or less. The patients were flagged for enrollment to be managed according to the EVADE protocol if their status was determined to be caused by DKA, and if they consented.

This process began in September 2013, and the first several months were spent refining management and defining quality indicators.

Month 8 results suggest that the protocol is effective for increasing discharge rates and reducing ICU admissions.

"Our initial goal was to discharge 10%, so we are above goal," Ms. Chojnicki said.

Overall, patients with mild to moderate DKA are going home from the ED, she said.

Typically, those with moderate DKA are admitted if they have a secondary diagnosis, and those patients are going to non-ICU beds.

"Only the sickest of the sick – basically those who are acidotic – are the patients going to the ICU," she said.

Preliminary data from an ongoing review of the results by a health economist show that despite the fact that the sickest patients are going to the ICU, the number of ICU days is decreasing in the wake of the EVADE protocol implementation.

This seems counterintuitive, but the latest data show that before EVADE, the average time spent in the ICU was 47.8 hours, and after EVADE the average time spent in the ICU was 31.6 hours.

"We shaved off 16.2 hours in one of the highest-cost areas of the hospital," Ms. Chojnicki said.

Notably, only one patient who was treated according to the EVADE protocol returned to the ED within 30 days, and that was a patient with issues related to alcohol abuse, she noted.

The findings suggest that mild DKA can be optimally managed in the setting of identification of the triggering factor, stabilization of glucose and electrolyte imbalances, patient education, and close follow-up, and that such management allows for more patients to be discharged from the ED, she said.

Ms. Chojnicki reported having no disclosures.

The EVADE protocol, step by step

The EVADE protocol for eligible patients begins with initiation of intravenous fluids but withholding insulin until initial labs are available. Once lab results confirm diabetic ketoacidosis, the staff initiates IV insulin.

A team including an attending endocrinologist, two nurse practitioners, and a certified diabetes educator are alerted when a patient is enrolled into the protocol.

Patients with all forms of DKA, including mild DKA without an anion gap, are managed on insulin drip with an endpoint goal of reaching a bicarbonate level of 21 mEq/L.

"The decision [to keep the patient in the ED for further observation and treatment or admit to the ICU] is based on the severity of the DKA and/or existing concurrent medical illness. If admitted to the ICU, the patient remains in the ICU arm of the DKA protocol, which involves frequent monitoring of glucose, electrolytes, venous blood gas, and urine output. Treatment is aimed at correcting the DKA and associated electrolyte disturbances," explained Maria Koen, a nurse practitioner and certified diabetes educator at Joslin Diabetes Center.

Regardless of whether the patient is admitted or remains in the ED, treatment involves the same IV insulin algorithm, which calls for no initial bolus; the starting rate is based on the initial finger stick, Ms. Koen noted.

Other features of the protocol include hourly finger sticks with adjustment of the dose with a goal of a target blood glucose of 100-180 mg/dL. Every 3 hours, someone on the team assesses venous blood gas and electrolytes.

IV fluid is run concurrently with IV insulin. Once the glucose drops to less than 250 mg/dL, fluid is changed to 5% dextrose in normal saline (D5NS). Once glucose increases again to 250 mg/dL, the protocol calls for discontinuation of D5NS with continuation of IV insulin infusion until the acidosis resolves.

If severe hypoglycemia occurs, D50 (50% dextrose) can be given as a bolus to reach the target of 100-180 mg/dL. The insulin drip is then resumed at 50% of the rate prior to the hypoglycemic event.

"Generally, we use normal saline or lactated Ringer’s as a first choice for IV fluid, at a rate of approximately 1 liter an hour for the first 4 hours, and this may be adjusted for patients who are at risk for volume overload, such as elders or patients with chronic kidney disease or congestive heart failure," she said.

After the first 4 hours, the IV solution may be changed to half of the normal saline, or half of the D5, if blood glucose drops below 250 mg/dL, at a rate of 125-250 cc per hour.

IV fluids may also be discontinued at this point if the patient is able to tolerate oral fluids.

The protocol also calls for electrolyte repletion as needed. Potassium may be added to each liter of IV fluid if the serum potassium reaches less than 5 mEq/L and the patient is voiding. The amount added is at the discretion of the attending physician, Ms. Koen said.

"Keep in mind that as the acidosis resolves, serum potassium will decrease. In such events, if severe hypokalemia occurs, it may be necessary to either decrease the insulin rate or temporarily hold the IV insulin drip," she said, noting that oral potassium and IV or oral phosphate can also be given once a patient can tolerate oral intake.

Insulin pumps should be removed to avoid overlap with the IV insulin, and the team should assess pump function and self-management.

Urine output and ability to tolerate oral intake should be assessed throughout the protocol, and patients should eat as soon as they are able to tolerate oral intake. The team should assess clinical status every 4 hours.

The patients’ care can move beyond the strictures of the protocol once they are either admitted to the ICU or stabilized in the ED. Those not admitted to the ICU can be discharged to home once their anion gap closes, their bicarbonate reaches more than 20 mEq/L, they can tolerate oral intake, and they successfully transition to a subcutaneous insulin regimen.

The transition to subcutaneous insulin requires that the basal dose be determined based on the patient’s home long-acting insulin dose or adjusted based on the nurse practitioner’s or doctor’s recommendations. Dosing for patients who are newly diagnosed or insulin naïve is calculated based on body weight in kilograms x 0.2.

When the patient is able to tolerate oral intake, prandial subcutaneous insulin based on the home prandial insulin dose should be initiated. Dosing in those who are insulin naïve should be based on body weight in kilograms x 0.2 and divided by 3 (but only with team recommendation).

"When transitioning from the IV insulin drip to subcutaneous insulin, if only basal insulin is given we recommend an overlap of the IV insulin drip for 2 hours. If both basal and prandial insulin are given at the same time, the recommendation is to overlap the insulin drip for at least 1 hour, and then the drip may be turned off," Ms. Koen said.

Insulin pumps may be restarted under the guidance of the certified diabetes educator. Electrolytes and glucose should be reassessed 2 hours after the insulin drip is stopped to ensure there is no return of the DKA.

Any patients with pH below 7.3, bicarbonate below 21 mEq/L, or an inability to tolerate oral intake will remain in the ED for continued observation and management. The patient is ready for discharge once pH is greater than 7.3, the anion gap closes, glucose reaches 250 mg/dL, and the patient can tolerate oral intake and has transitioned to subcutaneous insulin.

Following discharge, someone on the team should phone the patient, ideally within 24 hours, to ensure compliance with the postdischarge plan and schedule a follow-up clinic visit.

Ms. Koen reported having no disclosures.

ORLANDO – A new protocol for managing mild diabetic ketoacidosis in the emergency department led to a substantially improved discharge rate at a large teaching hospital.

Prior to development of the Emergency Valuable Approach and Diabetes Education (EVADE) protocol, almost all patients who presented with diabetic ketoacidosis (DKA) were admitted to the hospital, and most of those were admitted to the intensive care unit. However, in the most recent month for which outcomes data for the EVADE protocol are available, 24 of 106 patients (23%) managed according to the protocol were discharged from the ED and 82 were admitted, Marianne Chojnicki, a registered nurse and certified diabetes educator at the Joslin Diabetes Center, Boston, reported at the annual meeting of the American Association of Diabetes Educators.

The average length of stay in the ED for those who were discharged was 20 hours. The average length of stay in the hospital for 23 patients with a non-ICU admission was 136 hours, and the average length of stay in the ICU for 59 patients with an ICU admission was 33 hours, followed by 77 hours in the hospital after ICU discharge.

The findings have important implications for the care of patients presenting with DKA, which is documented in up to 9% of all hospital discharge summaries among patients with diabetes, and which is associated with mortality of up to 2%, Ms. Chojnicki noted.

"That’s a significant number of lives that are lost each year from a preventable event," she said.

In an effort to improve the discharge rate to at least 10%, Joslin Diabetes Center physicians, nurses, and diabetes educators worked together with Beth Israel Deaconess Medical Center in Boston to develop the EVADE protocol, which stemmed from a clinical approach used to manage DKA, and which is initiated in the ED based on clinical triggers.

They thought that a significant number of admissions were avoidable in a well-managed, coordinated system. The goal was to provide optimal cost-effective management for DKA, Ms. Chojnicki said.

From the outset, diabetes education was recognized as a key component of this protocol, she noted.

The education is individualized and covers multiple topics, such as identification of the circumstances leading up to DKA, a review of insulin action and technique of administration, and a review of ways to prevent DKA recurrence.

Evaluating the EVADE protocol

To evaluate the effectiveness of the protocol, a system was put in place to identify all patients who presented to the ED with blood glucose higher than 300 mg/dL and bicarbonate of 20 mEq/L or less. The patients were flagged for enrollment to be managed according to the EVADE protocol if their status was determined to be caused by DKA, and if they consented.

This process began in September 2013, and the first several months were spent refining management and defining quality indicators.

Month 8 results suggest that the protocol is effective for increasing discharge rates and reducing ICU admissions.

"Our initial goal was to discharge 10%, so we are above goal," Ms. Chojnicki said.

Overall, patients with mild to moderate DKA are going home from the ED, she said.

Typically, those with moderate DKA are admitted if they have a secondary diagnosis, and those patients are going to non-ICU beds.

"Only the sickest of the sick – basically those who are acidotic – are the patients going to the ICU," she said.

Preliminary data from an ongoing review of the results by a health economist show that despite the fact that the sickest patients are going to the ICU, the number of ICU days is decreasing in the wake of the EVADE protocol implementation.

This seems counterintuitive, but the latest data show that before EVADE, the average time spent in the ICU was 47.8 hours, and after EVADE the average time spent in the ICU was 31.6 hours.

"We shaved off 16.2 hours in one of the highest-cost areas of the hospital," Ms. Chojnicki said.

Notably, only one patient who was treated according to the EVADE protocol returned to the ED within 30 days, and that was a patient with issues related to alcohol abuse, she noted.

The findings suggest that mild DKA can be optimally managed in the setting of identification of the triggering factor, stabilization of glucose and electrolyte imbalances, patient education, and close follow-up, and that such management allows for more patients to be discharged from the ED, she said.

Ms. Chojnicki reported having no disclosures.

The EVADE protocol, step by step

The EVADE protocol for eligible patients begins with initiation of intravenous fluids but withholding insulin until initial labs are available. Once lab results confirm diabetic ketoacidosis, the staff initiates IV insulin.

A team including an attending endocrinologist, two nurse practitioners, and a certified diabetes educator are alerted when a patient is enrolled into the protocol.

Patients with all forms of DKA, including mild DKA without an anion gap, are managed on insulin drip with an endpoint goal of reaching a bicarbonate level of 21 mEq/L.

"The decision [to keep the patient in the ED for further observation and treatment or admit to the ICU] is based on the severity of the DKA and/or existing concurrent medical illness. If admitted to the ICU, the patient remains in the ICU arm of the DKA protocol, which involves frequent monitoring of glucose, electrolytes, venous blood gas, and urine output. Treatment is aimed at correcting the DKA and associated electrolyte disturbances," explained Maria Koen, a nurse practitioner and certified diabetes educator at Joslin Diabetes Center.

Regardless of whether the patient is admitted or remains in the ED, treatment involves the same IV insulin algorithm, which calls for no initial bolus; the starting rate is based on the initial finger stick, Ms. Koen noted.

Other features of the protocol include hourly finger sticks with adjustment of the dose with a goal of a target blood glucose of 100-180 mg/dL. Every 3 hours, someone on the team assesses venous blood gas and electrolytes.

IV fluid is run concurrently with IV insulin. Once the glucose drops to less than 250 mg/dL, fluid is changed to 5% dextrose in normal saline (D5NS). Once glucose increases again to 250 mg/dL, the protocol calls for discontinuation of D5NS with continuation of IV insulin infusion until the acidosis resolves.

If severe hypoglycemia occurs, D50 (50% dextrose) can be given as a bolus to reach the target of 100-180 mg/dL. The insulin drip is then resumed at 50% of the rate prior to the hypoglycemic event.

"Generally, we use normal saline or lactated Ringer’s as a first choice for IV fluid, at a rate of approximately 1 liter an hour for the first 4 hours, and this may be adjusted for patients who are at risk for volume overload, such as elders or patients with chronic kidney disease or congestive heart failure," she said.

After the first 4 hours, the IV solution may be changed to half of the normal saline, or half of the D5, if blood glucose drops below 250 mg/dL, at a rate of 125-250 cc per hour.

IV fluids may also be discontinued at this point if the patient is able to tolerate oral fluids.

The protocol also calls for electrolyte repletion as needed. Potassium may be added to each liter of IV fluid if the serum potassium reaches less than 5 mEq/L and the patient is voiding. The amount added is at the discretion of the attending physician, Ms. Koen said.

"Keep in mind that as the acidosis resolves, serum potassium will decrease. In such events, if severe hypokalemia occurs, it may be necessary to either decrease the insulin rate or temporarily hold the IV insulin drip," she said, noting that oral potassium and IV or oral phosphate can also be given once a patient can tolerate oral intake.

Insulin pumps should be removed to avoid overlap with the IV insulin, and the team should assess pump function and self-management.

Urine output and ability to tolerate oral intake should be assessed throughout the protocol, and patients should eat as soon as they are able to tolerate oral intake. The team should assess clinical status every 4 hours.

The patients’ care can move beyond the strictures of the protocol once they are either admitted to the ICU or stabilized in the ED. Those not admitted to the ICU can be discharged to home once their anion gap closes, their bicarbonate reaches more than 20 mEq/L, they can tolerate oral intake, and they successfully transition to a subcutaneous insulin regimen.

The transition to subcutaneous insulin requires that the basal dose be determined based on the patient’s home long-acting insulin dose or adjusted based on the nurse practitioner’s or doctor’s recommendations. Dosing for patients who are newly diagnosed or insulin naïve is calculated based on body weight in kilograms x 0.2.

When the patient is able to tolerate oral intake, prandial subcutaneous insulin based on the home prandial insulin dose should be initiated. Dosing in those who are insulin naïve should be based on body weight in kilograms x 0.2 and divided by 3 (but only with team recommendation).

"When transitioning from the IV insulin drip to subcutaneous insulin, if only basal insulin is given we recommend an overlap of the IV insulin drip for 2 hours. If both basal and prandial insulin are given at the same time, the recommendation is to overlap the insulin drip for at least 1 hour, and then the drip may be turned off," Ms. Koen said.

Insulin pumps may be restarted under the guidance of the certified diabetes educator. Electrolytes and glucose should be reassessed 2 hours after the insulin drip is stopped to ensure there is no return of the DKA.

Any patients with pH below 7.3, bicarbonate below 21 mEq/L, or an inability to tolerate oral intake will remain in the ED for continued observation and management. The patient is ready for discharge once pH is greater than 7.3, the anion gap closes, glucose reaches 250 mg/dL, and the patient can tolerate oral intake and has transitioned to subcutaneous insulin.

Following discharge, someone on the team should phone the patient, ideally within 24 hours, to ensure compliance with the postdischarge plan and schedule a follow-up clinic visit.

Ms. Koen reported having no disclosures.