Like many people in North America, I have close ties to diabetes and insulin. We likely all have someone in our lives be it family, friends or ourselves who are diabetic. My connection is somewhat different. Growing up in London, Ontario, most of my family went to high school at a school named Banting. Not really bothering to make any sort of connection beyond the name itself, it wasn’t until I was working at Rutgers and read the The Discovery of Insulin (great book) that the connection made sense. (spoiler alert) Charles Banting was one of the founders of insulin, which transformed the way diabetics were cared for (essentially living in a treatment facility, titrated down to a caloric intake of 500-1000 calories/day and just dealing with that) to regaining a quality of life that resembled someone without diabetes. Most fascinating of all, after their Nobel-Prize, the group of Banting and Best sought a pharmaceutical manufacturer that allowed the patent to remain open. Eli Lilly was the only manufacturer that agreed, and intended for money to never be an issue for someone who needed insulin. Also, Lilly, located near Chicago, had access to a virtually unlimited source of beef and pork pancreas to make the insulin. When I hear about the cost of insulin these days, I often wonder back to what Charles Banting would think of his discovery now.
As dramatic and remarkable the story of insulin is, so too are the actions of insulin in the extremes of diabetes. From a complete deficiency/total resistance leading to ketoacidosis/hyperosmolarity to fatal (or near fatal) hypoglycemia, the spectrum of potential clinical effects of insulin make it one of the most potent drugs in use today. I won’t spout off the ISMP medication safety data, but I’m sure we all know how often insulin is the culprit in drug errors.
A deficiency or resistance to insulin is of course the core pathophysiology of diabetes. In the USA, the ever growing obesity epidemic fuels development of type 2 diabetes. While a major public health concern, and focus of many inpatient and outpatient programs, the extremes of diabetes often lead patients to the emergency department. Diabetic ketoacidosis (DKA) and hyperglycemic-hyperosmolar nonketotic syndrome (HHS) are the hyperglycemic emergencies most often encountered in the ED. While the treatment of DKA and HHS is rather straight forward, there are critical nuances to ensure the most effective and safest care.
A 22 year old male is brought to the ED via EMS for abdominal pain. He’s writhing in pain on the stretcher, retching, and is very agitated. His medical history is significant for type 1 diabetes, for which he had run out of his insulin for a week and has had a flu like symptoms for the past 2 days. He’s tachycardic on the monitor (130’s), normal pressure (105/80), but is tachypneic (95% on RA), temp of 99.8F and the point of care glucose reads 350 mg/dL. The team is having a hard time getting IV access since the patient is so dehydrated, there is a discussion about whether to give NS or LR first.
Fluid – Chloride rich/Chloride poor
There are numerous therapies that should be simultaneously started in the ED for patients in hyperglycemic crisis.[1-3] The most accessible and easiest to begin, however, are IV fluids. Very often IV fluids are started on patients in the ED with little thought in regards to the particular formulation, dose/rate, and therapeutic goal. With hyperglycemic emergencies, IV fluids are best treated as a drug. In the most ideal state, IV fluids would always be treated as such, but few cases illustrate the potential risks and harms of various fluids better than DKA/HHS.
The underlying pathophysiology of DKA is an insulin deficiency (or resistance) and glucagon excess, producing an excess glucose concentration along with dehydration and electrolyte disturbances.  With this hyperglycemic state, the kidneys secret glucose into the urine, producing a hyperosmolar diuretic effect (also pulling in several electrolytes: Na, K, Mg, Ca, P). This excessive diuretic effect and volume loss, combined with decreased oral intake due to vomiting (gastroparesis/glucagon), patients with DKA can present with profound total body water deficits. 
Dehydration sets off numerous neurohormonal responses that do nothing to help improve the physiologic problems with DKA. The release of cortisol, epinephrine, glucagon, and somatostatin act to either enhance ketogenesis via lipolysis or through inhibiting glycolysis. Furthermore, dehydration impairs peripheral tissue perfusion which may lead to decreased use of ketones by these tissues as a energy source.
Volume resuscitation with IV crystalloids have the potential to improve the volume depletion, replace electrolyte imbalances and dilute glucose in plasma to reduce hyperosmolarity. The two fluids most commonly recommended as first line are sodium chloride 0.9% (NS, normal saline) and Lactated Ringer’s (LR).
The contents of NS compared to LR shed light on the not so nuanced differences. NS contains a 1:1 ratio of sodium to chloride. Each 100 mL of NS contains 900 mg of sodium, and works out to 154 mEq/L each of sodium and chloride with a final osmolarity of 308 mOsmol/L and pH of 5.0. This ratio of sodium to chloride, osmolarity and pH are hardly ‘normal’ physiological ranges. Compared this to LR which contains the following: Na 130 mEq/L, K 4 mEq/L, Cl 109 mEq/L, Ca 1.4 mEq/L, Lactate 28 mOsmol/L with a osmolarity of 272 mOsmol/L and pH of 6.5. The more physiologic ratio of sodium to chloride with LR can reduce the risk of iatrogenically inducing a hyperchloremic acidosis, falsely closing the anion gap, and may increase the risk of acute kidney injury.[5,6]
Plasmalyte, a branded balanced electrolyte solution (BES), is a crystalloid fluid that has been adjusted to be as physiologic as possible. This crystalloid has Na 140 mEq/L, K 5 mEq/L, Cl 98 mEq/L, Mg 3 mEq/L, acetate 27 mEq/L, gluconate 23 mEq/L with a osmolarity of 294 mOsmol/L And pH of 6.5. In studies comparing standard NS therapy for DKA to BES use, improve lab-based outcomes such as anion gap, and bicarbonate, but have not proven superiority in patient-oriented outcomes.[7-9]
Given the theoretical benefits, and lack of high quality data, most guidelines either make no preference as to which fluid to use initially, or simply suggest using clinical judgement.[1-3] While the total body water deficit may be as high as 3 to 5 L in patients with DKA, there is no agreed upon calculation for fluid replacement.  It is reasonable to use a bolus of 20 mL/kg in adult DKA patients with a systolic blood pressure less than 80 mmHg.  Slower rates of infusion can be targeted for patients who are hemodynamically stable, but the same total volume should be administered over roughly 3 hours. Subsequent fluid choices, after the bolus, should be based on patient specific findings with electrolytes and based on their response to insulin (ie, the need for dextrose containing fluids).
While insulin therapy is fundamental to DKA/HHS management, fluid resuscitation alone may lower plasma glucose by up to 18%.  Therefore, early aggressive insulin therapy with fluid resuscitation may cause excessively rapid glucose correction (or over correction) which can lead to potentially harmful fluid shifts (ie, cerebral edema). This is why, while insulin is critical, the impact of IV fluid therapy on glucose must be considered prior to insulin initiation.
IV fluid is fundamental for DKA/HHS management
BES seems to be, at least theoretically, more helpful with DKA/HHS.
Cost is the biggest drawbacks to physiologic branded BES. LR is a great, cheap alternative
The 22 year old male with DKA now has two large bore peripheral IV lines, running one liter of LR each. A repeat point of care glucose reading still reads 325 mg/dL and the physician orders an insulin infusion to begin STAT. Should this patient recieve an insulin bolus?
While the hyperglycemia of DKA and HHS can be lowered substantially with IV fluids alone, the underlying physiologic process cannot be treated without insulin. Whether the patient has no baseline insulin secretion from their pancreas (DKA) or are profoundly resistant (HHS), exogenous insulin is needed to correct the underlying pathophysiology.
Insulins numerous secondary messenger actions begin by its binding to the alpha subunit of the insulin receptor on muscle, liver, renal, adipose and other tissues.[11,12] This receptor activity initiates a conformational change in the beta-subunits of the insulin receptor to be capable of phosphorylation actions. Tyrosine residues on these beta-subunits are phosphorylated and the subsequent activity of tyrosine kinases then activates phosphatidylinositol-3 and MAP-protein kinase pathways. These pathways set off further intracellular actions that result in reversal of the catabolic action of insulin deficiency of particular benefit in DKA/HHS. In the liver, insulin inhibits glycogenolysis, conversion of fatty acids to keto acids, and conversion of amino acids to glucose as well as improving glycogen synthesis (GLUT-2 receptors). In muscle tissue, insulin promotes glycogen synthesis through increased glucose transportation (through GLUT-4 receptors in muscle, cardiac and adipose tissue), and in adipose tissue insulin promotes the esterification of fatty acids. Other effects of insulin extend to promoting protein synthesis, regulating gene transcription and cell proliferation. Thus, many of the underlying metabolic derangements leading to fatty acid beta-oxidation are resolved and reversed by insulin and cannot be done without insulin.
Insulin available today is a recombinant form of human insulin. While porcine and bovine preparations are still available internationally, the human recombinant forms are the mainstay in the USA. Among the various types of insulin products available, the most rationale classification is based on their onset of action and duration of action: rapid acting, short acting, and long acting insulins.
While these recombinant insulins are administered with the goal of controlling glucose, they fail to match physiologic insulin responses. Since the administration of exogenous insulin is first absorbed to the peripheral circulation, not the portal circulation where endogenous insulin distributes to, it influences hepatic glucose metabolism differently than normal physiologic insulin.[11,12] Similarly, peripherally administered insulin does not act as rapidly in response to prandial glucose changes in both rapidly increasing and decreasing concentrations. The available recombinant insulin preparations, and drug therapy strategies using them in combinations, aims to match this process as closely as possible.
Fortunately in the case of DKA/HHS, we aren’t necessarily concerned with matching physiologic insulin. In this extreme of diabetes, we aim to administer insulin to reach near-maximal glucose uptake, inhibition of gluconeogenesis and lipolysis. In order to do so, insulin is administered via continuous IV infusion. To determine the appropriate insulin preparation for IV administration, let’s review the types of insulin based on their subclass.
Rapid acting insulin
The pharmacokinetics of insulin products are depended on specific substitutions on the underlying amino acid chain of insulin.[11,12] Insulin aspart is created by substituting the B-28 aspartic acid for proline (Novolog). Insulin lispro (Humalog) is produce with the reversal of proline-lysine on positions B-28, B-29 and insulin glulisine (Apidra) has two substitutions of lysine at B-3 and glutamic acid at B-29. These substitutions allow these insulins to be administered subcutaneously and rapidly absorbed. Their rapid absorption is a result of the insulin dissociating to monomers almost immediately after injection.
Short acting insulin
Insulin regular, as the name would suggest is simply the recombinant version of the endogenous human insulin amino acid chain.[11,12] The absorption of insulin regular following subcutaneous injection is slower than that of rapid acting insulins since hexamers are formed at neutral pH and dissociates slowly at physiologic pH. For many outpatients managing diabetes, this makes insulin regular administration less convenient since for a basal/bolus regimen it must be given 30-45 min prior to a meal compared to 15 minutes before a meal with rapid acting and potentially after a meal with insulin glulisine. Furthermore, the duration is up to 2 hours longer. This may have benefits, but also can lead to hypoglycemia if the dose is high.
Insulin regular is the most commonly administered insulin preparation via IV. While it is possible to administer aspart or lispro IV, regular insulin has maintained its role for this route. There is no evidence comparing the theoretically easier titratability of the rapid acting products given IV, however, they can be used if certain institutions aim to reduce the number of different insulin preparations in a medication safety initiative.
Intermediate and long acting insulin
Intermediate and long acting insulin preparations are not ideal for management of DKA/HHS. Since their pharmacokinetics have been optimized to have longer onsets (to prevent a ‘peak’ effect hypoglycemia) and longer durations of activity, it makes rapid titration for DKA/HHS not feasible.[11,12] There may be an emerging role for long acting insulin preparations in patients with mild or borderline DKA, however, this role is not practice ready at this point.
IV Insulin regular – IV administration Insulin bolus vs no bolus
As previously discussed, DKA and HHS cannot be managed without exogenous insulin. Given the pharmacokinetic, lack of prospective research and years of practical experience, insulin regular is often used as the IV infusion insulin. In determining the initial dose of insulin, it’s important to recall the goals of therapy. In particular, the effect of insulin on plasma glucose. With any administration of insulin, the plasma glucose should decrease. However, this is not the goal for DKA/HHS management. In fact, it’s probably best to consider this an adverse event that must be managed. Insulin administration supports the correction of the metabolic derangement while the causative (infection, myocardial infarction, dehydration, other illness) event is resolved.
Insulin infusion rapidly achieves near-maximal glucose uptake to the liver and skeletal muscle, inhibits gluconeogenesis as well as lipolysis.[11-13] These actions shift metabolism away from fatty acid beta-oxidation back towards glycolysis thereby preventing further production of ketones. In the process of rebalancing metabolism, blood glucose will decline ideally at a rate of 50-75 mg/dL/hr. Once plasma glucose levels of 200 mg/dL for DKA and 300 mg/dL for HHS are achieved, maintenance fluids should include dextrose. Without the understanding of the pathophysiology we’ve been discussing, this would seem counter productive. But since we understand both glucose (d-glucose aka dextrose) and insulin are necessary for aerobic cellular respiration, glucose supplementation at some point in treatment will be necessary to allow for clearance of ketone bodies and correction of acidemia. Furthermore, understanding of the management euglycemic DKA caused by the newer diabetes agents (gliflozins), starvation or pregnancy will help solidify the understanding of the role of insulin in DKA.[Euglycemic DKA post]
Insulin infusions may be preceded by a IV bolus of 0.1 units/kg. While there is evidence to suggest that there is no benefit to insulin boluses in adults with DKA, they may solve a more operational pharmacy problem. That being, the time from order entry, IV compounding and delivery of the insulin drip to the bedside. If this process will take more than 15 minutes, a reality in many pharmacies, then an insulin bolus MAY be reasonable in sick DKA patients (those with very low bicarb, very high anion gaps and likely ICU admission). However, in other cases where patients are not as ill, or pharmacy turnaround times are rapid, it is preferable to start the insulin drip at a rate of 0.14 units/kg/hour.[14,15]
Determining the therapeutic goal for the insulin drip may be dependent on the institutional protocol and geographic location of your practice. The three major guidelines for DKA each recommend a different endpoint.[1-3] For the American guidelines, insulin is to be titrated to a particular glucose level. For DKA, once glucose reaches 200 mg/dL the insulin rate should decrease to 0.02-0.05 units/kg/hour or transition to 0.1 units/kg subcutaneous rapid acting insulin every 2 hours to a goal glucose of 150-200 mg/dL. For HHS the same strategy is recommended by targeting a glucose goal of 200-300 mg/dL. Canadian guidelines on the other hand, recommend adjusting the rate to target a closure of the anion gap and supplementing dextrose as needed. British guidelines suggest a middle of the road approach, using both ketone clearance and correction of glucose as a marker of therapy. Given the variations in recommendations, and low quality of evidence, any of the above strategies may be reasonable.
Insulin is necessary to treat DKA/HHS, but it should not distract from the underlying problem
Hypoglycemia is an adverse event, not a therapeutic goal for insulin
IV insulin can be regular, aspart or lispro.
The insulin drip finally arrived from pharmacy, but the patient received a bolus of 0.1 units/kg IV insulin regular. After 20 mL/kg of LR, the patient was also started on an infusion of LR at a rate of 150 mL/hr. On a repeat VBG, the potassium now reads 2.8 mmol/L.
Alongside the declining glucose level, potassium concentrations will also decrease on top of already total body potassium deficit due to osmotic diuresis and insulin administration. [15-18] In fact, prior to insulin administration, potassium levels should be collected. If the potassium is less than 3.3 mmol/L, insulin should be held until potassium supplementation increases this level to somewhere between 4-5 mmol/L. It is also possible for patients to initially present with a mild hyperkalemia due to osmotic shifts. In these patients, who are still potassium depleted, once insulin (and perhaps bicarb) is started, the potassium will rapidly decline.
In order to achieve normokalemia, IV administration of potassium chloride is often necessary. As many traditional pharmacy beliefs note that PO potassium “sticks” better, it may not be feasible to administer via this route given the nausea and vomiting often present with hyperglycemic crisis. To replace potassium, while there is little guidance from the guidelines, it is reasonable to administer runs of 10-20 mEq/100mL, with the empiric assumption each 10 mEq will increase serum potassium by 0.1. So in order to go from 3.3 to 4.0, a total of 70 mEq to be administered over 7 hours (10 mEq/hr) or 3.5 hours (20 mEq/hr) depending on patient tolerance. Once maintenance fluids are started, it is reasonable to add 40 mEq of potassium chloride to each liter of fluid.
Magnesium administration should be a consideration whenever potassium supplementation is necessary. Unacknowledged hypomagnesemia can further potassium losses in the kidneys and worsen the adverse effects of hypokalemia on specific tissues.
In DKA/HHS there is likely a potassium deficit, which can go unrecognized.
Potassium supplementation is necessary before insulin in many cases where hypokalemia is already present
Potassium can’t be bolused, and magnesium should be a concomitant intervention
Along with the VBG findings of hypokalemia, the admitting team is also concerned that the bicarb is still less than 12, despite the pH being 7.25. An order for sodium bicarbonate 150 mEq in 1000mL of D5W is ordered.
The role of sodium bicarbonate in hyperglycemic emergencies are a matter of controversy.  While the evidence for, or against, bicarbonate therapy is limited, it offers some guidance. A review of nine studies (434 patients total) in diabetic ketoacidosis that saw 217 patients receive bicarbonate therapy and 178 patients not receive bicarbonate therapy, did not demonstrate improvements in patient oriented outcomes including cardiac or neurological functions, rate of recovery of hyperglycemia, and ketoacidosis.
While there may be no upside to sodium bicarbonate, there is certainly downside. Sodium bicarbonate therapy increases the already high risk of hypokalemia, may decrease tissue oxygen uptake, induce cerebral edema, and can trap protons in the CNS leading to a local acidosis. The ADA guidelines remain bearish on sodium bicarbonate therapy and suggest there may be a role if pH is less than 6.9. Provided that the patient has the minute ventilation to expire the produced CO2, this may be a reasonable therapy. However, for patients who are tiring from Kussmaul’s respirations or who are intubated without subsequent ventilator setting adjustments, sodium bicarbonate may worsen acidemia.
One small upside to sodium bicarbonate is in its potential role in fluid resuscitation. From the previous discussion regarding chloride rich vs chloride poor solutions, sodium bicarbonate can be substituted for sodium chloride to reduce the risk for iatrogenic hyperchloremic metabolic acidosis, or premature closure of the anion gap. However, sodium acetate, LR or other BES products can achieve the same goal.
Bicarbonate is often a reflex order for DKA/HHS. It’s not that it’s unreasonable, it just has to be a conscious decision of all the physiologic benefits, and potential harms.
Bicarbonate could be a useful chloride sparing strategy
Don’t count on it fixing the blood gas, and it will make hypokalemia worse
- Kitabchi AE, et al. Hyperglycemic Crises in Adult Patients With Diabetes. Diabetes Care Jul 2009, 32 (7) 1335-1343
- Goguen J, Gilbert J. Clinical Practice Guideline: Hyperglycemic emergencies in adults. Can J Diabetes, 2013;37:S72-S76
- Joint British Diabetes Societies Inpatient Care Group. The Management of Diabetic Ketoacidosis in Adults. Available at: https://www.diabetes.org.uk/professionals/position-statements-reports/specialist-care-for-children-and-adults-and-complications/the-management-of-diabetic-ketoacidosis-in-adults
- Cydulka RK, Maloney GE: Diabetes Mellitus and Disorders of Glucose Homeostasis; in Marx JA, Hockberger RS, Walls RM, et al (eds): Rosen’s Emergency Medicine: Concepts and Clinical Practice, ed 8. St. Louis, Mosby, Inc., 2014, (Ch) 126: p 1652-1668.
- Nor’azim Mohd Yunos, Rinaldo Bellomo, Michael Bailey, et al. Association Between a Chloride-Liberal vs Chloride-Restrictive Intravenous Fluid Administration Strategy and Kidney Injury in Critically Ill Adults. JAMA 2012; 308(15): 1566-1572.
- Van Zyl DG, Rheeder P, Delport E. Fluid management in diabetic-acidosis–Ringer’s lactate versus normal saline: a randomized controlled trial. QJM. 2012 Apr;105(4):337-43.
- Mahler SA, Conrad SA, Wang H, Arnold TC. Resuscitation with balanced electrolyte solution prevents hyperchloremic metabolic acidosis in patients with diabetic ketoacidosis. Am J Emerg Med. 2011 Jul;29(6):670-4
- Chua HR, Venkatesh B, Stachowski E, et al.Plasma-Lyte 148 vs 0.9% saline for fluid resuscitation in diabetic ketoacidosis. J Crit Care. 2012 Apr;27(2):138-45
- Oliver WD, Willis GC, Hines MC, Hayes BD. Comparison of Plasma-Lyte A and Sodium Chloride 0.9% for Fluid Resuscitation of Patients With Diabetic Ketoacidosis. Hosp Pharm. 2018 Oct;53(5):326-330.
- Cydulka RK, Maloney GE. Chapter 118 Diabetes mellitus and disorders of glucose homeostasis. In; Rosen’s Emergency Medicine: Concepts and Clinical Practice, 9th Edition. Elsevier
- Powers AC, D’Alessio . Endocrine Pancreas and Pharmacotherapy of Diabetes Mellitus and Hypoglycemia. In: Brunton LL, Hilal-Dandan R, Knollmann BC. eds. Goodman & Gilman’s: The Pharmacological Basis of Therapeutics, 13e New York, NY: McGraw-Hill; . http://accesspharmacy.mhmedical.com/content.aspx?bookid=2189§ionid=172482821. Accessed January 18, 2019.
- Kennedy M, Masharani U. Pancreatic Hormones & Antidiabetic Drugs. In: Katzung BG. eds. Basic & Clinical Pharmacology, 14e New York, NY: McGraw-Hill; . http://accesspharmacy.mhmedical.com/content.aspx?bookid=2249§ionid=175222393. Accessed January 18, 2019.
- Triplitt CL, Repas T, Alvarez C. Diabetes Mellitus. In: DiPiro JT, Talbert RL, Yee GC, Matzke GR, Wells BG, Posey L. eds. Pharmacotherapy: A Pathophysiologic Approach, 10e New York, NY: McGraw-Hill; . http://accesspharmacy.mhmedical.com/content.aspx?bookid=1861§ionid=146065891. Accessed January 18, 2019
- Kitabchi AE et al. Is a priming dose of insulin necessary in a low-dose insulin protocol for the treatment of diabetic ketoacidosis? Diabetes Care. 2008;31(11):2081
- Goyal N, et al. Utility of initial bolus insulin in the treatment of diabetic ketoacidosis. J Emerg Med;4 doi:10.1016/j.jemermed.2007.11.033
- Ennis ED, Stahl EJ, Kreisberg RA. The hyperosmolar hyperglycemic syndrome.
Diabetes Rev 1994;2:115e26.
- Mahoney CP, Vlcek BW, DelAguila M. Risk factors for developing brain herniation
during diabetic ketoacidosis. Pediatr Neurol 1999;2:721e7.
- Murthy, K et al. Profound Hypokalemia in Diabetic Ketoacidosis: A Therapeutic Challenge. Endocrine Practice. 2005; 11:5 p 331
- Huang CL, Kuo E. Mechanism of Hypokalemia in Magnesium Deficiency. JASN Oct 2007, 18 (10) 2649-2652
- Viallon A, Zeni F, Lafond P, Venet C, Tardy B, Page Y, Bertrand JC . Does bicarbonate therapy improve the management of severe diabetic ketoacidosis? Crit Care Med 1999; 27: 2690– 2693
- Glaser N, et al. Risk factors for cerebral edema in children with diabetic ketoacidosis. N Engl J Med 2001; 344: 264– 269