Bolus Insulin for DKA in Adults

Washington University Emergency Medicine Journal Club
-May 20, 2021


You are working an EM-2 shift one steady afternoon when you encounter Miss X, a 25-year-old woman with a history of type-1 diabetes. She presents with nausea, vomiting, and vague abdominal discomfort for the last couple of days. She also endorses fevers and chills, dysuria, and left flank pain, and is worried she might have a “kidney infection.” She reports compliance with her long-acting insulin, but has had poor PO intake the last couple of day and has not been using her Novolog. Her initial blood sugar in triage was 430 and her ketones were 4.8. 

You enter the room to find a pleasant young woman who appears uncomfortable but is not in any distress. She reports ongoing nausea and some mild epigastric discomfort, as well as suprapubic pain and left flank pain. She is tachycardic, febrile to 38.1, and has a stable BP. She has no epigastric tenderness on your exam but does have suprapubic and left CVA tenderness. 

Fairly certain that your patient has DKA and pyelonephritis, you order labs and begin IV hydration with a liter of LR. Her UA comes back with nitrites, leukocyte esterase, bacteria, and WBCs and you order IV ceftriaxone. Her BMP comes back with a blood sugar of 443, a CO2 of 10, and an anion gap of 22. You go into the DKA order set and being clicking boxes, but when you come to the box for the insulin bolus you aren’t sure what to do. You are aware, from your time in the Children’s ED, that current guidelines recommend against the use of bolus insulin for DKA in pediatric patients given the increased risk of cerebral edema in this patient population, but you aren’t sure how this applies to adults. You wonder if there is any downside to giving bolus insulin to adults, but also wonder if there is any utility. You end up ordering the bolus and admit the patient to the ICU, but after your shift you decide to search the literature and see if there is any evidence to support or refute the use of bolus insulin for DKA in adults… 

PICO Question

Population: Adult patients presenting to the ED with diabetic ketoacidosis
requiring a continuous infusion of IV insulin

Intervention: Continuous IV insulin without a bolus of IV insulin

Comparison: Standard continuous IV insulin infusion with an initial IV insulin bolus

Outcome: Episodes of hypoglycemia, time to normalization of anion gap, duration of IV insulin therapy, need for ICU admission, ICU and hospital length of stay

Search Strategy:

PubMed was searched using the terms “(DKA of (diabetic ketoacidosis)) AND (bolus OR load)” resulting in 154 citations ( Three relevant articles were identified among these. A fourth article was identified by using Google Scholar using the terms “bolus insulin DKA adults.”

Article 1: Goyal N, Miller JB, Sankey SS, Mossallam U. Utility of initial bolus insulin in the treatment of diabetic ketoacidosis. J Emerg Med. 2010 May;38(4):422-7. doi: 10.1016/j.jemermed.2007.11.033. Epub 2008 Jun 2. PMID: 18514472. Answer Key

Article 2: Sarra J, Fatma H, Rym H, Hana H, Hela M, et al. Utility of Bolus of Insulin in Diabetic Ketoacidosis Management in Emergency Department. J Anest & Inten Care Med. 2019; 8(5): 555749. DOI: 10.19080/JAICM.2019.08.555749. Answer Key.

Article 3: Brown HD, Tran RH, Patka JH. Effect of Bolus Insulin Administration Followed by a Continuous Insulin Infusion on Diabetic Ketoacidosis Management. Pharmacy (Basel). 2018;6(4):129. Published 2018 Dec 7. doi:10.3390/pharmacy6040129. Answer Key

Article 4: Kitabchi AE, Murphy MB, Spencer J, Matteri R, Karas J. Is a priming dose of insulin necessary in a low-dose insulin protocol for the treatment of diabetic ketoacidosis? Diabetes Care. 2008 Nov;31(11):2081-5. doi: 10.2337/dc08-0509. Epub 2008 Aug 11. PMID: 18694978; PMCID: PMC2571050. Answer Key.

Bottom Line:

Diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic nonketotic coma are among the most dangerous acute diabetic complications observed in the ED. DKA is diagnosed in over 200,000 patients each year, with an increasing incidence observed between 2003 and 2014. The mainstay of treatment for DKA currently involves volume repletion with crystalloid fluids and administration of exogenous insulin to correct hyperglycemia and promote clearing or serum ketones, with the ultimate goal of correcting the anion gap acidosis. The most recent guidelines from the American Diabetes Association recommend administration of a bolus of IV insulin (0.1 U/kg) prior to starting a continuous IV insulin infusion, although it should be noted that these guidelines have not been updated since 2009 and the use of insulin bolus in adult patients with DKA has become controversial in the interim. We sought to examine the literature to determine whether there is evidence that a bolus of IV insulin prior to a continuous infusion in adult patients with DKA reduces the time to normalization of the anion gap, reduces the duration of the insulin infusion, affects length of stay, or results in increased episodes of hypoglycemia. 

Two randomized controlled trials and two observational studies were identified in the literature. A trial from 2008 (Kitabchi 2008) randomized a small number of patients to receive either a low-dose insulin bolus (0.07 U/kg) following by low-dose infusion (0.07 U/kg/hr), a low-dose infusion without a bolus, or a high-dose (0.14 U/kg/hr) infusion with no bolus. With only 37 total patients enrolled, this study found no significant difference in time to achieve a glucose ≤ 250 mg/dL, time to achieve a  bicarbonate level ≥ 15 mEq/L, or time to achieve a pH ≥ 7.3. There was also no difference in hospital length of stay or episodes of hypoglycemia. It should be noted that this was a very small study that was underpowered detect potentially clinically meaningful differences in outcomes. Additionally, none of the patients in this study were treated with a standard insulin infusion dose of 0.1 U/kg, making it difficult to generalize the results when such dosing is used. 

A second randomized trial (Sarra 2019), published much more recently, did manage to compare an insulin bolus to placebo in patients receiving standard insulin dosing. In this study of 106 patients, the mean time to resolution of acidosis was similar between the bolus and no bolus groups (difference 1.2 hours in favor of the no bolus group, 95% CI -1.4 to 3.8), and found no difference in time to blood sugar control, total dose of IV insulin given, ED length of stay, or the incidence of hypoglycemia. 

Finally, two observational studies were identified (Goyal 2010Brown 2018). The first of these found significant differences in the incidence of hypoglycemia, the rate of change of serum glucose in the ED (60.1 vs. 56.0 mg/dL/hr, p = 0.54), the rate of change in the anion gap (1.9 vs. 1.9 mEq/L/hr, p = 0.66), the difference in the serum glucose level or anion gap at the time of discharge from the ED, length of ED stay, or length of hospital stay between those who received and those who did not receive an insulin bolus. The second study primarily evaluated time to resolution of DKA—defined as a blood glucose less than 250 mg/dL plus 2 of the following: pH > 7.3, anion gap ≤ 12 mEq/L, and bicarbonate ≥ 15 mEq/L. There was no difference in this outcome between the bolus and no bolus groups (median time 15 vs. 15.9 hours, p = 0.24), and no difference in the total amount of insulin given, the incidence of hypoglycemia, or hospital length of stay. Both of these studies are at high risk of selection bias given their observational nature. 

Unfortunately, this body of evidence is of moderate quality at best and it is difficult to strongly recommend either for or against the use of an insulin bolus for DKA in adults. While none of these studies found a clear benefit, they also not demonstrate any significant harm. It seems reasonable to omit an insulin bolus when a continuous infusion can be initiated in a timely fashion, but in cases in which the drip may be delayed, it may be prudent to give a bolus while waiting. Ultimately, the decision will need to be made on a case-by-case basis, but this limited evidence does not seem to support the routine use of an insulin bolus in these patients.