financial interests or relationships to disclose. DKA and HHS have distinctly different pathophysiology. HHS is less common and occurs in only 13% of hyperglycemia-related emergency admissions, but it also has a higher mortality than DKA. It occurs typically in older patients with type 2 diabetes who have relative insulin deficiency in that they have adequate insulin levels to prevent ketogenesis, but not enough to cause glucose utilization since it takes a 10th as much insulin to suppress lipolysis as it does to stimulate glucose utilization. So the patient presents with extreme levels of hyperglycemia, usually greater than five to 600 mg per deciliter, and consequent osmotic diuresis, leading to profound dehydration, usually with a total body deficit of about nine liters. HHS also has a slower onset than DKA, which is important to note when we discuss treatment, particularly because the brain tissue of those who develop HHS, namely the older patient population, are at higher risk of injury due to rapid shifts in sodium, water, and glucose. So you'll see that treatment of HHS requires less aggressive fluid resuscitation replacement over 24 hours and slower glucose lowering than DKA, with a cold glucose of 180 to 270 mg per deciliter in the first 24 hours. Here are the diagnostic criteria for HHS based on the American Diabetes Association guidelines from 2009, the table on the left, and the Joint British Diabetes Societies, or JBDS, guideline on HHS management that was published in early 2023, which is the table on your right. Both definitions include significant hypovolemia, severe hyperglycemia defined as greater than 540 or 600 mg per deciliter, without significant ketonemia, either with trace ketones in urine or a blood beta-hydroxybutyrate, or BHB, less than three millimoles per liter, and without significant acidity. Both definitions have an absolute insulin deficiency. This hormonal imbalance causes hyperglycemia by a reduction in the amount of glucose and decreases hyperglycemia by increasing glycogenolysis, hepatic gluconeogenesis, and decreased peripheral utilization of glucose. Hyperglycemia then promotes glucose urea and osmotic diuresis. So DKA patients usually present with dehydration and hypovolemia, with a general total body deficit of around six liters. Absolute insulin deficiency will also promote gluconeogenesis and ketogenesis in adipose tissue, resulting in lipolysis that mobilizes free fatty acids. Ketone body formation is also responsible for hyperketonemia and subsequently results in acidosis. This figure does an excellent job summarizing the metabolic pathway for ketogenesis. You might be familiar with the Krebs cycle, but to review, during insulin deficiency, glucose uptake into cells is quite limited, so an alternative energy substrate is required. The breakdown of non-esterified fatty acids allows for fatty acid CoA to enter the Krebs cycle, which will generate ATP. But excess fatty acid CoA will lead to the production of ketone bodies, which are acetoacetate and beta-hydroxybutyrate. The main ketone body is beta-hydroxybutyrate, which accounts for about 80% of production. The incidence of Euglycemic DKA, or EDKA, has increased in the past few years with a rise in the usage of sodium glucose transporter type 2 inhibitors, or SGLT2 inhibitors. The pathophysiology of the EDKA involves relative or absolute carbohydrate deficit, a milder degree of insulin deficiency or resistance, and increased glucagon to insulin ratio. Medications such as SGLT2s promote renal wasting of glucose, so we don't see the serum glucose rise to the levels we usually see in DKA and HHS. And these patients present with Euglycemia with a glucose level around and less than the 250 milligrams per deciliter. Because of normal glucose levels, the patients do not have polyuria and polydipsia, so they're not profoundly dehydrated. They might instead present with malaise, anorexia, or tachypnea. DKA consists of a triad of symptoms. The definition that we're almost familiar with is probably from the DKA guidelines that was published in 2009 by the American Diabetes Association. And they define DKA as hyperglycemia with a glucose greater than 250 milligrams per deciliter, elevated ketones in the blood or urine, and acidosis, which is a pH less than 7.3. The ADA does not discuss in detail any treatment for Euglycemic DKA, pregnant women with diabetes, impaired gluconeogenesis, secondary to alcohol abuse, or patients on sodium glucose co-transporter 2s or SGLT2 inhibitors. The Joint British Diabetes Societies for Inpatient Care, or JBDS, updated their guidelines June of 2021. Their definitions are... Understanding the connection between insulin and inflammation may be helpful in the treatment and preventative approach to DKA and HHS. Using different types of inflammatory biomarkers could help us differentiate between the severity of DKA and could give us insight into the likelihood of development of acute clinical complications, such as cerebral or pulmonary edema. The two most frequent precipitating risk factors of DKA are missed insulin dose and presence of infection. So High Sensitivity C-Reactive Protein, or HSCRP, is a sensitive marker of infection in adults. Levels below 0.3 milligrams per deciliter are considered physiological, and levels above 10 milligrams per deciliter are indicative of bacterial and viral infections, as well as severe tissue damage. These levels increase by a thousandfold within 24 to 72 hours as an acute phase reactant to inflammation, infection, or organ trauma. It has a long half-life of 19 hours and may help identify concomitant infections. Interleukin-6, or IL-6, is a pro-inflammatory cytokine, while interleukin-10, or IL-10, is an anti-inflammatory cytokine. It's estimated that levels are increased and may correspond with the severity of DKA. C-peptide is used to assess B-cell function, and it's more accurate than insulin levels, since insulin assays also detect exhaustion as insulin for those patients who are administering insulin. C-peptide is produced in a one-to-one ratio to insulin, and it's the gold standard to measure in DODGE. There have been numerous studies investigating the relationship between these biomarkers as predictors of DKA outcome. They're all small studies and conducted mainly in pediatric patients, but the outcomes are similar. So I'd like to highlight this study by Karavanchi and colleagues, because I thought I had a nice summary of the results. They studied 38 newly diagnosed children who had type 1 diabetes and DKA, and they did this over a period of two years. Blood samples were drawn at presentation and before initial hydration was given, at six to eight hours, again at 24 hours, and again at five days, or 120 hours. 25% of these patients were males. The mean age was 7 1⁄2 years old, and there were 16 patients that presented with mild DKA, 16 patients with moderate DKA, and six patients with severe DKA. Baseline values of CRP and C-peptide are depicted here. And I think what's interesting is that when you look at the initial white blood cell count and CRP, which we generally correlate with infection, they are both elevated at baseline. By 24 hours, these values have normalized, even the authors emphasize, without the use of any antibiotics. This is consistent in what has also been previously published of leukocytosis in hyperglycemic crisis without infection, which probably suggests anti-inflammatory effects of insulin therapy, and that CRP and white blood cell count might be elevated in DKA, even in the absence of infection. Also of note, for IL-6, there was no difference between the trends for mild, moderate, or severe DKA and resolution, but there was a significant difference between baseline levels of IL-10 for mild DKA versus moderate to severe DKA. The IL-10 levels were significantly higher on presentation, and then they fell by the six-hour mark. Some acute complications of DKA, such as cerebral or pulmonary edema. Acutely, though, what biomarkers can we use to help us diagnose and identify resolution of hyperglycemic emergencies? For this, we need to look at tests that tell us the degree of ketone production and acidosis. There are three main tests for ketones. You may be the most familiar with the urine ketone test, or the nitroprusside test. This is where acetoacetate and acetone will react with an iron complex to generate a red-purple color on a dipstick. While this is a non-invasive and quick method to assess for ketosis, it can provide a high false negative rate in DKA because nitroprusside reacts with acetoacetate and acetone, which, as you may recall, consists of only 20% of the ketones provided in the body. It can also react with sulfohydrols for medications. The second test for ketosis is by measuring serum beta-hydroxybutyrate, or BHB. This represents approximately 80% of ketones present in the blood during DKA. So monitoring BHB in plasma gives us a real-time indication of the severity of ketosis and reflects the impact of treatment much more quickly than urinary ketone measurements. In hyperglycemia, when glucose is greater than 250 mg per deciliter, a BHB level of more than 1.5 mmol per liter has a sensitivity ranging from 98 to 100% and a specificity of 85%. With a BHB of 3 mmol per liter, the sensitivity increases to almost 100% and a specificity of 94%. This is a FDA-approved assay for ketosis, but the turnaround time is usually quite long. It can take up to one to four hours, depending on your institution's lab turnaround time. And it can be quite costly per test. And when used in serial measurements, the cost can add up quite quickly. Three beta-hydroxybutyrate is measured using a portable ketone meter. This point-of-care test is ultra-rapid, so results are available within 10 seconds, but it has a limited measurement range where results greater than 4 mmol per liter should be disregarded. And accuracy in patients with peripheral shutdown and shock is unclear. Nonetheless, because the specificity of the BHB test, either at the serum or as a point-of-care, the JBDS guidelines recommend attaining BHB over urine ketones. Keep in mind that two-thirds of physiologically relevant ketone bodies are strong acids. So the resolution of ketosis will often be reflected in the normalization of your pH and anion gap closure. For measurement of acid doses, we generally rely on venous pH per guideline recommendations both from ADA and JBDS. This largely stems from the 1998 Brandenburg study where they found that pH on venous blood gas in patients with DKA was only 0.03 lower than pH on arterial blood gas. Since this difference is not clinically significant, there's almost no reason to perform the more painful EVGs. Although if your patient has an art line in place, it's not unreasonable to draw arterial blood gases. And lastly, because many of us were taught to calculate the anion gap and to trend it for a resolution of DKA, but clinical interpretation of the anion gap is important, especially in hypoalbuminemic patients and because it's not reliable when other factors for acid doses are present. For this reason, the 2021 JBDS guidelines say not to use anion gap in either diagnosis or resolution of DKA. Resolution of HHS is depicted on the right in green with the ADA 2009 guideline definition on the left when osmolality and mental status have returned to baseline and the JBDS 2023 guideline definition on the right, which is a bit more specific with osmolality of less than 300 millisomals per kilo, urine output returned to greater than 0.5 milliliters per kilo per hour, mental status returned to baseline and glucose levels less than 270. In the past few years, there have been more data on utilizing subcutaneous insulin therapy in the treatment of DKA, HHS and ADKA. We know that treatment of DKA and HHS include first restoration of circulating volume, that's with administration of fluids, more rapidly in DKA patients and then slower over 24 hours for patients with HHS. Secondly, we focus on the clearance of ketones and suppression of ketogenesis through the administration of insulin. With the 2009 ADA guidelines, even noting that insulin therapy is effective regardless of the route of administration, either through intravenous insulin infusions, subcutaneous injections of rapid acting insulin and or adjunctive subcutaneous long acting insulin. With intravenous insulin infusions, we can either use the traditional or continuous titratable IV insulin infusions versus a fixed rate IV insulin infusion, which is a set insulin rate with titration of dextrose containing fluid to maintain euglycemia. And this is typically seen with two bag methods or an EDKA. Today, we are going to focus on the role of subcutaneous rapid acting insulin, Lispro or Aspart, and on early administration of long acting insulin therapy. While continuous infusion of regular insulin is generally recommended due to well-established efficacy studies, its quick onset of action and shorter half-life, it can lead to rapid decreases in plasma glucose and ketones. It requires frequent titration and monitoring and thus ICU admission. From the pharmacokinetic profile of rapid acting insulins, Lispro and Aspart, and the profile for subcutaneous regular insulin, we can see how the longer half-life and duration of action could potentially act as a safe and effective treatment option where we're still providing enough insulin to increase peripheral tissue utilization glucose, reduce gluconeogenesis and glycogenolysis. Dr. Umpieres and his colleagues were the first to explore the use of subcutaneous insulin for the treatment of mild to moderate DKA. The aim of the study was to compare efficacy and safety subcutaneous insulin Lispro versus low-dose IV insulin infusion of regular insulin in the treatment of patients with uncomplicated DKA. Patients were excluded if they were hypotensive despite one liter of normal saline infusion with the rest of the exclusion criteria listed above on the table. Patients treated with intravenous insulin were admitted to the intensive care unit or ICU. They were given an initial bolus of 0.1 units per kilogram followed by a continuous infusion of 0.1 units per kilo per hour until the glucose was less than 250 mg per deciliter. And then the insulin infusion was decreased to 0.05 units per kilo per hour until resolution of DKA. Patients that were treated with the subcutaneous Lispro were managed on a general medicine floor in a step-down unit. They were given an initial subcutaneous insulin Lispro dose of 0.3 units per kilo that was followed by 0.1 units per kilogram subcutaneous injection every hour until the glucose was less than 250 milligrams per deciliter. At that point, the dose was decreased to 0.05 units per kilogram per hour until the resolution of DK8. The fluids in both groups were changed to a dextrose containing IV fluid once the glucose decreased to less than 250 milligrams per deciliter. Baseline characteristics were similar between the two groups. There were no difference in hospital length of stay and duration of treatment with the resolution of DK8 which was within 10 to 11 hours. The authors also note there were no difference in episodes of hypoglycemia. However, treatment with hourly subcutaneous insulin injections is difficult at most institutions because of the intensity of treatment. So Dr. Umpieres and his colleagues subsequently followed up with a study investigating the use of subcutaneous insulin aspart in the treatment of mild to moderate DK8 to see if they could increase the time interval between glucose checks. The aim of this study, exclusion criteria and dosing for the regular insulin infusion dosing arm was the same as the previous study. They added two subcutaneous aspart insulin arms. Both received an initial insulin aspart subcutaneous bolus of 0.3 units per kilo per hour. One group received a scheduled dose of aspart 0.1 units per kilo every hour and the other arm received aspart 0.2 units per kilo every two hours. Baseline demographics were similar between the two groups. The authors also found no difference in the length of hospital stay, total amount of insulin needed for resolution of hyperglycemia or ketoacidosis or incidence of hypoglycemia. The authors noted that administration of subcutaneous aspart enabled treatment of DK8 in the wards or in the emergency room and it even reduced their cost of hospitalization by 30% without significant changes in hypoglycemic events. The caveat of all of this was that this is an uncomplicated. This study by Mohammed and colleagues was a retrospective cohort study to evaluate the efficacy and safety of early initiation of insulin glargine in combination with an IV insulin infusion in patients admitted to the ICU with DK8. These patients were managed on a continuous insulin infusion using a commercially available computerized clinical decision support system which would adjust the insulin infusion rate based on individualized patient variables. So baseline hemoglobin A1C, weight, height, initial blood glucose levels and rate of glucose change in consumed carbohydrates. They enrolled 380 patients who are relatively young with a mean age of 45 years and a baseline hemoglobin A1C of around 12. The insulin infusion was initiated approximately three hours after diagnosis of DK8 and insulin glargine was on average started 21 hours after DK8 diagnosis at a starting dose of 25 units. 10% of patients experienced a hypoglycemic episodes considered less than 80 milligrams per deciliter but only 0.5% had a critically low glucose level that was less than 50 milligrams per deciliter. The authors found that for every six hour delay in administrating glargine, it was associated, there was a 26 minute increase in a time to DK8 resolution. This was their primary outcome as well as a 3.2 hour increase in duration of insulin infusion and 6.5 hour increase in ICU length of stay. So now this is looking at all 380 patients. When they looked at the subgroup of patients who had definitively received glargine before the resolution of DK8, this was only in 45 patients. They found that each six hour delay in administrating insulin glargine was associated with a 3.25 hour delay in time to DK8 resolution. The study showed that early initiation of subcutaneous glargine concomitantly with an insulin infusion was safe and was associated with a reduction in time to DK8 resolution, ICU and hospital length of stay and shorter insulin infusion duration. Limitations of this study include the generalizability and that the insulin titration algorithm was conducted by a computerized clinical decision support system which can be cost prohibitive to implement at many institutions. The mean time to starting insulin glargine was about 21 and a half hours, which is likely around the time of DK8 resolution. So not too much earlier than resolution. And only 45 patients actually received the glargine prior to resolution of DK8. Although the subgroup analysis shows that there was significant association between time to insulin glargine and time to DK8 resolution. The authors didn't specify their definition of DK8. So we don't know if these patients had mild, moderate or severe disease. Although looking at baseline characteristics, most patients probably presented with moderate to severe DK8. Baseline characteristics were similar between two groups, except that the early glargine group had more severe degree of ketoacidosis with a BHB of 9.3 versus 6.5. And thus also lower presenting pH of 7.2 versus 7.3. The primary outcome time to DK8 resolution was significantly shorter in the early glargine group compared to the control group. So 10 hours versus 13. With a significantly shorter hospital length of stay in the early glargine group by 10 days. When the authors conducted a subgroup analysis of patients that were hospitalized for more than one week, nine in the early glargine group, 19 in the control group, they found that the predominant cause of prolonged hospitalization was due to complicated infection that required ventilator support, surgical interventions, and ongoing IV antibiotics. There was no difference in rates of rebound hyperglycemia, hypoglycemia, severe hypoglycemia, hypokalemia, and all-cause mortality. And the authors concluded that the early combination of long-acting basal insulin with IV insulin infusion within three hours of DK8 diagnosis led to faster DK8 resolution and shorter length of stay without increasing high. When the revised British Diabetes Satiety Guidelines came out in 2021, they had already recommended that if a patient was already on a long-acting insulin at home to continue it at the usual dose and time, even at the onset of DK8 presentation. So this was based on some smaller studies on the administration of glargine earlier in the DK8 process. The two larger studies that I presented just now were published after the 2021 revision, and they further solidified their recommendation that early administration of long-acting insulin in conjunction with a standard IV insulin infusion is both safe and effective at decreasing the duration of DK8 length of insulin infusion and hospital length of stay. For patients who may not have been on a long-acting subcutaneous insulin at home, the guidelines are a bit less clear on when to start long-acting insulin, just that it should likely occur within 60 minutes to six hours, which usually is before the full resolution of DK8. So based on the small volume of data available, I would think that potentially subcutaneous administration of long-acting insulin is probably not the most effective initially for an extremely dehydrated patient, mainly from an absorption standpoint. But it seems reasonable to initiate a long-acting insulin prior to resolution of DK8. And at our institution, we now initiate Glargine when the initial blood glucose levels fall below 250 mg per deciliter. Basal insulin should be administered two to four hours before stopping intravenous insulin in order to prevent recurrence of DK8 and rebound hyperglycemia. Managing the transition from intravenous to subcutaneous insulin can be complicated. And what better way to help than with algorithms and protocols? Pharmacist-driven protocols, either at the bedside or created by the pharmacy staff, have consistently been shown to improve patient outcomes and decrease cost of care. Our extensive training, coupled with our in-depth knowledge of insulin preparations and dosing regimens can help ensure a safe and effective transition from intravenous to subcutaneous insulin therapy. DK8 management often requires tailoring the insulin regimen to patient-specific needs. And pharmacists can work closely with the healthcare team to use best practices to adjust insulin dose and regimens according to the patient's glucose levels, response to treatment, and other individual factors. Pharmacists are also vital in assisting with safety outcomes and protocol adherence, helping to ensure that the transition from intravenous to subcutaneous insulin is carried out in accordance with established best practices, reducing the risk of complications. We also help prevent hypoglycemia by carefully calculating the appropriate starting dose of subcutaneous long-acting insulin and providing guidance on titration of therapy. We often facilitate multidisciplinary collaboration and ensure that the transition from intravenous to subcutaneous insulin is coordinated with other healthcare providers, whether that includes nurses, physicians, or dieticians. Effective teamwork is essential for a smooth transition. And lastly, pharmacists contribute immensely to quality improvement initiatives by regularly reviewing and updating the transition protocol based on patient outcomes and emerging best practices. This continuous assessment leads to safer and more efficient transitions off of insulin infusions. When we initially brought up the idea to start long-acting insulin glarging earlier on in the DKA process and prior to resolution of DKA at our institution, we were met with quite a few obstacles. Some of them included resistant to change because why change something that's worked for years? Sometimes the existing culture within the organization might not be conducted in a way that is effective might not be conducive to change or they may promote counterproductive behaviors. Maybe the staff are not aware of new literature or guideline recommendations. And maybe there is some fear in increased amounts of workload during the trial maturation period of a new protocol or procedure. Maybe there is a lack of leadership support because without strong support from top-level management, many quality improvement projects might struggle to gain traction and resources. So effective communication is key, especially getting buy-in with key stakeholders since misunderstandings and lack of clear communication channels can really hinder a project's progress. Perhaps you have inadequate resources, maybe a shortage of funding or time or skilled personnel that can definitely hinder projects implementation and sustainability. If you get overzealous and expand the scope of the project beyond its original boundaries, it can lead to delays and resource overruns. Inadequate project management practices like lack of clear goals, timelines or accountability can also lead to project failure. We're all stretched thin, so balancing other ongoing activities and priorities with quality improvement projects is challenging. External factors such as economic shifts, market changes or unexpected events can disrupt quality improvement projects. Inconsistencies in processes and procedures make it hard to implement standardized improvements and navigating regulatory requirements and ensuring compliance is also quite complex. Lastly, maintaining quality improvements over a long-term is challenging. It's easy to revert back to previous practices if the changes aren't consistently monitored and reinforced. So re-education is key, as well as determining and demonstrating the effectiveness and importance of the quality improvement projects. So, with resistance to change, our DKA protocol had not been revised in over a decade. And so when the 2021 British Diabetes Update came out that suggested that we administer subcutaneous glargine earlier, we thought it'd be a good time to readdress our hyperglycemic emergency algorithm and to make it more efficient, especially since we were noticing significant delays in transitioning patients off insulin infusion upon DKA resolution. Sometimes the delays would be up to two days in ICU, which increased length of stay and cost to the patient. We were initially met with quite a bit of resistance since many of our staff felt that our current protocol was good enough and that it would take an immense amount of work for a complete algorithm overhaul. So several pharmacists in our group came together, we gathered some data and we presented our findings on our current time to transition off of insulin infusion into subcutaneous insulin when there was DKA resolution. We presented this to the critical care, endocrine and emergency medicine providers. After that, we gathered a small group of key stakeholders together, which included endocrine and ICU nurses, advanced practice providers or APPs, pharmacists and physicians. We then met as a group to fishbone diagram and identify possible causes of the problem. We sorted the data into useful categories, moving post-its around, which is a helpful visual way to look at cause and effect. And then we met with content experts, endocrine and pharmacy staff to come up with a plan. We were able to start gaining more support once we were able to get the frontline staff on board. So we presented the data and our proposed plan to them. And then it was important to listen and seek their feedback to get their input in the algorithm so that they also felt like they had a stake in the game. We did this by holding small focus groups with the nursing staff and APPs in the ICU. Halfway through revising the algorithm, we decided to formally add early initiation of chlorogene when the glucose decreased to less than 250 milligrams per deciliter, or step two of our current algorithm after initial fluid resuscitation was complete. This increased the scope of our project and led to further delays since we had to rework the project and think about additional safety guards. Our group of pharmacists all had clinical duties, so we had to be diligent with our project management practices to balance our clinical duties with also making this algorithm a priority. We even looked into getting an electronic clinical decision support tool to help improve our insulin infusion titration algorithm. But we did not have sufficient funds for that. We found that what helped the most was to make sure this project was a priority so that we didn't lose momentum. We also split up duties so that it was less time intensive for each of us. But also being careful not to increase the group to too many members and having too many hands in the bucket. External factors play a role in unexpected events, such as a global pandemic, which can disrupt quality improvement projects. So it's important to be flexible in how you work. Maybe you meet virtually instead of in person and find ways to meet in the middle. You can identify your needs You can identify potential external factors that can impact your project and conduct a thorough risk assessment. You should develop contingency plans that outline how your project will adapt in response to these external challenges. You should ensure that your project team is aware of these plans and knows what steps to take when certain factors arise. You also need to stay informed about relevant regulatory and guideline changes or new landmark trials that may impact your protocol. Build flexibility into your project planning and consider using adaptive project management methodologies that will allow you to make gradual adjustments to your project objectives, the scope and your timelines. This kind of flexibility can help your project adapt to changing circumstances. Lastly, you want to make sure that all your hard work initiating this quality improvement project is maintained over the long term. This can be challenging. Implementing new protocols and algorithms require... Thank you for your time. Please don't hesitate to contact me if you have any questions or comments.

diabetic ketoacidosis

DKA

hyperosmolar hyperglycemic state

HHS

pathophysiology

management

insulin therapy