Hello, and welcome to today's Journal Club Spotlight on Pharmacy webcast, which is supported by the Society of Critical Care Medicine's CPP section. My name is Gabrielle Gibson, and I'm the Surgical Trauma Burn ICU Clinical Pharmacist at Barnes Jewish Hospital in St. Louis, Missouri, and I will be moderating today's webcast. A recording of this webcast will be available to registered attendees. You can log in to mysccm.org and navigate to the My Learning tab to access this recording. A few housekeeping items before we get started. There will be a Q&A after each of today's speakers. To submit questions throughout the presentation, type into the question box located on your control panel. You will also have the opportunity to participate in several interactive polls. When you see a poll, simply click the bubble next to your choice. SCCM provides the following disclaimer. This presentation is for educational purposes only. The material presented is intended to represent an approach, view, statement, or opinion of the presenter, which may be helpful to others. The views and opinions expressed herein are those of the presenters and do not necessarily reflect the opinions or views of SCCM. SCCM does not recommend or endorse any specific test, physician, product, procedure, opinion, or other information that may be mentioned. And now I'd like to introduce your speakers for today. Each will give a 15-minute presentation followed by a question and answer session. Our first presenter today is Kate Summers, and she is the PGY-2 Critical Care Resident at Wake Forest Baptist Health in Winston-Salem, North Carolina. Our second presenter is Joseph Oropesa, and he is a PGY-2 Critical Care Resident at UC Health Memorial Hospital in Colorado Springs, Colorado. And finally, our third presenter is Lauren Andrews, and she is the PGY-2 Critical Care Resident at the University of Illinois Hospital and Health Sciences System in Chicago, Illinois. And now I will be turning things over to our first presenter. Kate, it's all yours. Hi, thanks for that introduction. I'm excited to talk to you about this article today. So today we will be discussing an article entitled, Factor Xa Inhibitor-Related Intracranial Hemorrhage Results from a Multi-Center Observational Cohort Receiving Prothrombin Complex Concentrates. And this was published earlier this year by a pharmacy-led group of authors. I think this is an incredibly relevant study because there's so much controversy surrounding how to most effectively neutralize the anticoagulant effects of Factor Xa inhibitors in patients taking these medications and presenting with ICH, or intracranial hemorrhage. PCCs have not been evaluated by the FDA for the indication of life-threatening bleeding in patients taking Factor Xa inhibitors, and previous studies with PCCs are relatively small, so less than 100 patients or so, and they often use variable definitions of hemostatic efficacy. So overall, there's a gap in the literature really evaluating these meds, and these are relatively expensive medications, and we're using these for a high-risk patient population. So today's study really seeks to provide some clarity. Guidelines differ slightly in terms of recommendation, and this largely hinges upon the FDA approval of a target-specific reversal, indexinet-alpha. So the 2016 Neurocritical Care Society and SCCM guidelines recommend four PCC or activated PCC, 50 units per kilogram, if the ICH occurred within three to five terminal half-lives. The next two guidelines listed here were published after indexinet-alpha was FDA approved. So the 2018 American Society of Hematology recommends either four-factor PCC or indexinet-alpha, and the 2019 European Stroke Organization guidelines recommend indexinet-alpha or four-factor PCC if a specific reversal agent, such as indexinet-alpha, is not available. But you can see that overall, this level of evidence and recommendation level is pretty low. To make sure everyone's on the same page throughout this presentation, we'll refer to four PCC and activated PCC throughout this presentation. So I wanted to, and I know we're used to hearing these brand names most oftenly. Four-factor PCC contains non-activated factors 2, 7, 9, and 10, as well as protein CNS, and goes by the trade name Kcentra. The other primary PCC is activated PCC, containing activated factor 7 and non-activated 2, 9, and 10, and this goes by the brand name of FIBA. The purpose of this study was to evaluate the safety and efficacy of prothrombin-complex concentrates for apixaban or riboroxaban-associated intracranial hemorrhage. And though this concept has been tested before, this was a novel study in that the authors really sought to make up for shortcomings of previous similar trials. Specifically, this was a large multicenter trial, whereas previous trials were relatively small with a standardized safety and efficacy criteria. This was a retrospective multicenter cohort study. Currently, this study group is completing an ongoing larger study evaluating outcomes between patients who received PCC versus andexanet-alpha. This study represents the subset of patients treated with prothrombin-complex concentrate for factor Xa inhibitor-related ICH. 20 total medical centers were included, all of which were certified stroke centers. All patients had intracranial hemorrhage that was either spontaneous or traumatic in nature, and all patients had to have received apixaban or riboroxaban in a timeframe that warranted PCC administration. Key exclusion criteria were hemorrhages associated with liver disease or patients who did not have follow-up due to withdrawal of life-sustaining measures within 24 hours. Additionally, the hemostatic efficacy analysis excluded additional patients, and these patients were excluded if they did not have at least one follow-up image of the brain within the first 24 hours after PCC administration, or if the primary bleed was epidural or intraventricular, and this was due to the lack of clear criteria to assess for hemostasis in these hemorrhage types. The primary efficacy outcome was the percentage of patients with excellent or good hemostasis, defined overall as less than 35% increase in hematoma volume. Remember that the efficacy outcome did exclude patients who did not have follow-up imaging or if the bleed type was not amenable to clear hemostasis assessment. The safety outcome included a larger patient population and was the occurrence of a thrombotic event during hospitalization. Patients were followed to hospital discharge or censored at 30 days. As far as statistical analysis, continuous data were described using a mean and standard deviation or median and interquartile range. Categorical variables were reported with frequencies. For the primary efficacy outcome, a 95% confidence interval with binomial tests and Jeffrey's interval was used, and the threshold for significance for categories being compared was set at 0.05. So let's get into some of the results. Overall, just over half of these patients were male and about 60% were over the age of 75. The majority were of Caucasian race. The mean weight was about 80 kilograms with a BMI of 28. Over three-fourths of the patients were receiving anticoagulation for the indication of atrial fibrillation. Just over half of the patients were receiving apixaban with the other remainders receiving rivaroxaban. The majority of bleeds were either intracerebral or subdural. Over half of these bleeds were traumatic in nature, but the initial GCS or Glasgow Coma score was impressively high at 14 in the safety analysis group and 15 in the hemostatic efficacy group. Lastly, the majority of patients, so over three-quarters of these patients, received four-factor PCC with the remainders receiving activated PCC. Because this was a retrospective study, the initial dose could vary between practice sites or providers, and this could also vary based on criteria for who received a second dose of PCC. The initial dose was 43.8 units per kilogram for patients receiving four-factor PCC or 25.7 units per kilogram in patients receiving activated PCC. The median time from presentation to administration was 2.6 hours, and 5.1% of patients received a second dose. So because of the variability in dosing, I think it's interesting, and I would like to see how you're dosing at your hospital. So when PCC is used for ICH patients and patients receiving Apixaban or Rivaroxaban, does your institution use a standardized dose, and if so, what dose? So we have choices 25 units per kilogram, 50 units per kilogram, less than 25 units per kilogram, or number four, this is kind of our catch-all, if you have variable dosing, if you have fixed dosing that's not weight-based, or another dosing schematic, or the last choice if you do not use PCC at all for this indication. I know at Wake Forest Baptist Health, our guidelines recommend a 50 units per kilogram dose, and this seems consistent here with the majority of patients, but I do see some variability here. I know we thought about and potentially considered going to other dosing schematics as we revised our reversal guidelines recently, but ultimately we decided to keep our recommendation at 50 units per kilogram. Now we can move on to the primary safety and efficacy outcomes. Excellent hemostasis or good hemostasis, so overall this is less than 35% expansion of the hematoma, this was obtained in 81.8% of patients in this primary efficacy outcome group. The 230 patients were excluded from the hemostatic efficacy analysis. As far as the safety analysis, this included 663 patients, and 3.8% of these patients had a thrombotic event, and the vast majority of these were within 14 days of the PCC administration. Only one of these events occurred after anticoagulation was restarted. As far as secondary outcomes, I think the overarching trend here is that this data points to the hemostatic efficacy population being a slightly less acute population. Likely many of the patients that were excluded were high risk, and the hemostatic efficacy subgroup had lower mortality, and I would argue overall better disposition. Also within the efficacy population, the authors found that mortality was more common in patients with poor hemostasis, and this was associated with a significant p-value of 0.02. Overall, I think this was a really helpful and well-designed study, and we should be really proud of our pharmacy colleagues who were involved in this study. However, as with any study, it's not without limitations. The observational nature of this study leads to variable treatment patterns between providers and between sites, as this was a multi-center study. The type and dosing of PCC because of this could vary greatly. Also, you'll notice that there was no control group in this study. No control group was included. Also, these are overall well-appearing patients with a high initial GCS of 14 or 15, depending on the group that you're looking at, so this really might not guide our treatment for those really sick and highly acute presenting patients. Also, though I think it's understandable why patients were excluded from the efficacy analysis, we should notice that 230 patients, so over two-thirds of these patients, were excluded from this analysis. Lastly, some key components of documentation are missing. For example, the timing of the last factor Xa inhibitor dosing was not documented, so here we're really relying on judgment that the PCC was administered appropriately in the first place. Time to initial brain imaging was also not documented, so I think this is important because if there's a substantial delay in the initial brain imaging, then that first size might appear larger than you would expect, and then thus the delta of the ICH expansion might seem lower than it actually was, so if you have a variable, if it's larger time to this initial scan, then the follow-up scan really doesn't have as much of a chance to expand as much. And then lastly, the timing of anticoagulation resumption and prophylactic anticoagulation was not documented, so remember that these patients are on anticoagulation for a reason, so we expect this to affect thrombotic events. There are some notable strengths associated with this study. The first is the external validity with this large patient population and multi-center nature, and also the ability to treat based on provider or hospital guidelines. Also, I think the outcomes and definitions were consistent with previous trials, so though it's difficult to directly compare this, I think it does provide helpful framework. Lastly, third to lastly, hemostatic efficacy was adjudicated by a steering group, and I think that the subgroup analyses and secondary endpoints were thorough and appropriate. Like I said, it's hard to directly compare this to prior studies, but I have listed some similar prior studies here. So this top row is the study that we are discussing today, published in 2020. All of these, except for Goerner 2018, have the same definition of hemostatic efficacy of having the volume increased by 35% or less, whereas the Goerner source used 33%, so a similar outcome. And overall, I would argue that these percentages are relatively consistent, so this study does a great job of adding to the evidence that we already have with more robust numbers. The bottom row shows the ANEXA-4 trial, which is the elephant in the room, because this trial did not investigate PCCs, but instead it investigated indexinet-alpha, which is a target-specific antidote. Though we can't directly compare these percentages, we do see a similar efficacy and an increase in thrombosis in the ANEXA-4 trial. So moving forward, I'm eager to see this comparative data. This leads us to our next polling question involving indexinet-alpha, so I'm curious to know if your institution has indexinet-alpha on formulary and if it's used preferentially over PCCs for apixaban or rivaroxaban-associated ICH. So the first choice, indexinet-alpha is on formulary and used preferentially over PCCs for this indication. Number two, indexinet-alpha is on formulary, but you can use either indexinet-alpha or PCCs. Number three, indexinet-alpha is not on formulary and only PCCs are used, or the last if you're not sure. Or have another guideline. All right, I was really interested to see what results we would have here. I personally would fall into the number three, so the majority of people here chose indexinet-alpha is not on formulary and only PCC is used. So I know at my institution, it was definitely a big decision, but we have decided not to add indexinet-alpha at this time to formulary, and really we're just waiting on more robust data, specifically comparative data to PCCs. Overall, the takeaway points from this study is that though this study is not perfect and not without limitations, I think it was overall a well-executed study and provides appropriate conclusions. I think PCC is a reasonable treatment for apixaban and rivaroxaban-associated ICH as it provided appropriate hemostasis and relatively low rates of thrombosis. However, the exact dosing and PCC agent was pretty variable, and this alone warrants a lot more further research. Another thing that warrants further research is comparative efficacy to target-specific reversals such as indexinet-alpha. Thanks for your time, and I will take any questions right now. Thank you, Kate. The first question today, as if I was hoping you could actually comment on the primary efficacy outcome that was chosen, so why do you think this particular study and maybe prior studies have utilized hemostatic criteria that allowed up to 35% expansion? Is this something that you agree with? Would you change that outcome? I do like this outcome in that it's similar to previous studies. I think 35% can be a little bit misleading based on the initial size of the bleed. Obviously, if the bleed was really small to start with, as a lot of the bleeds in the study actually weren't too big as the initial size, then that delta, 35%, is not that much of an increase, whereas if you already have a large bleed, then that can kind of inflate things and you have a little bit more leeway and how much bleed you can have to have up to 35%. So I don't think it's a perfect endpoint, but I do agree with the endpoint that was chosen here. But I also appreciate that things like mortality and disposition, length of stay was also reported because overall I think we can agree that the best treatment for this is preventing the bleed in the first place. Great, and then the next question is, can you comment on how the patient population from this particular study compared to that of the Anexa 4 trial? Yeah, I think I tried to dig into that a little bit. Overall, I know I critiqued this study saying that a lot of these patients were not acutely ill, we have pretty high GCSs, and if you look in the supplementary index, the size of the initial bleed is pretty small. And this is actually pretty similar to the Anexa 4 trial. I know that was another critique of the Anexa 4 trial is that they excluded some of the more sick patients. So this didn't necessarily exclude low GCS patients, but they did not make up a large subset of the patient population in this study. So I think it's pretty similar. The Anexa 4 trial did not report exactly how many of their bleeds were traumatic in nature versus spontaneous, and in this study it was half and half. So I think that would be another interesting thing to see. Great. And then lastly, can you comment on what your thoughts are on the lack of reporting of functional outcomes after these patients have a bleed, and if that's something that you would like to see in future studies? Yeah, I would absolutely like to see that. I think the functional outcome that we saw here was just based on the disposition. So whether they're going to kind of assisted living or home, so the authors did seek to provide that. But I think a lot of these patients do have poor prognosis, unfortunately. So like I said, preventing the bleed in the first place is the best choice, but obviously that's not 100% feasible. So any data that we can see for functional outcome would certainly be helpful in upcoming studies. And I think just the last question, but I think it's a good wrap-up question, Kate, but if it was up to you, which reversal medication would you recommend in a patient presenting with a bleed due to a factor Xa inhibitor? That is a wonderful question. I would personally stick with four-factor PCC. As far as the dose, obviously at my institution I am used to using the 50 units per kilogram, and then we have a max dose. So I think that's what I'm used to, and I know that a lot of the data does point to that. I know more data is coming out using even lower doses, which I think are definitely appropriate. So I personally would use four-factor PCC for the reversal, but I'm eager to see more data moving forward. Great, thank you for your presentation, Kate. I'd like to now introduce our second presenter, and that's going to be Joseph Oropesa from the University of Colorado. The article that I'm going to be talking about today is titled, Early Initiation of Oral Antihypertensive Reduces Intensive Care Unit Stay and Hospital Costs for Patients with Hypertensive Intracerebral Hemorrhage. A little bit of background before we jump into the article. So back in 2017, Hong and colleagues evaluated patients who presented to the ED with an intracerebral hemorrhage, or ICH. They compared those with a resistant hypertension with those who were more responsive to initial antihypertensive therapy. What they found was that those who had resistant hypertension demonstrated increased need for ventilator support, hematoma evacuation, hypertonic saline therapy, and need for nicotinine infusion. Those who had resistant hypertension also demonstrated longer ICU and hospital length of stay. After a multivariate regression analysis, the authors identified that systolic pressures greater than 140 millimeters of mercury and nicotinine infusion were independently associated with longer ICU stays. The second study that preceded this study was published in 2016 by Schott and colleagues, and they also evaluated patients who presented with an ICH and compared those whose blood pressure was managed solely by a nicotinine infusion for blood pressure control versus those who were transitioned to early enteral antihypertensives. And what they found was that those who had early enteral antihypertensives demonstrated a decreased duration of IV nicotinine with a mean difference of 88 hours and significantly reduced estimated total cost per patient with a median difference of about just over $18,000. Where the literature is a little bit hazy still is that it is around the impacts of early enteral antihypertensives on ICU and hospital costs and in patients with hypertensive ICH. And so the authors hypothesized that early enteral antihypertensive initiation in patients with hypertensive ICH reduces the use of intravenous nicotinine, ICU length of stay, and subsequent ICU costs. So they performed a retrospective cohort study and included all patients presenting with hypertensive ICH between January 1st, 2013 and December 31st, 2017. They excluded anybody who presented to the ED with an ICH secondary to an arteriovenous malformation, ruptured aneurysm, tumor, trauma, cerebral amyloid angiopathy, and coagulopathy. They also excluded anybody with a DNR or DNI order within 72 hours of admission and an initial systolic pressure less than 180. They really focused on those with severe hypertension and anybody who had a duration of myocardic infusion less than two hours to really account for any potential confounding. During their initial search, they identified just over 600 patients who were reviewed for inclusion. Of those, about 440 individuals were excluded, primarily for systolic pressures less than 180, closely followed by those who presented with a non-hypertensive ICH. Of note, the skull systolic pressure for anyone that was going to be in the study was less than 140 millimeters of mercury. And they separated those who received the patients into those who received enteral antihypertensive within 24 hours of arrival to the ED and those who had received enteral antihypertensive after 24 hours. Of note, this is the blood pressure algorithm that they used for the management of the enteral antihypertensive for the first five days. This was a protocol that was validated by the same authors in a previously published study. I'm not going to go through every box here for the sake of time, but I did want to point out that the enterally available antihypertensives included amlodipine, licinopril, spironolactone, labetalol, clonidine, and hydralazine. The authors performed a univariate analysis to detect the effect of early initiation of floral antihypertensive. They evaluated the duration and cost of myocardic infusion, renal function, ventilatory support, length of stay in the ICU and hospital, numbers of oral antihypertensive at discharge, and the cost of hospitalization. They also performed a multivariate logistic regression models to determine which variables were independently associated with early initiation of oral antihypertensive after adjusting for initial systolic blood pressure, NIH score, GCS, and ICH scores. A two-tier value of P less than 0.05 was considered to be statistically significant threshold for this study. At baseline, patients were fairly similar in terms of age, gender, history of hypertension, and initial presenting systolic pressure to the ED. Of note, those who were in the control group had significantly higher NIH, GCS, and ICH scores. Also, a significantly greater proportion of the control group were had some form of dysphagia or required a feeding tube placement at some point within the first 24 hours. This is not as surprising given that the initial severity scores were higher for that group as well. In terms of durations of myocardipine, those who had received ventral antihypertensives within the first 24 hours of ED arrival had significantly lower or shorter durations of myocardipine. Just to orient you guys here, on the first P value that you're going to see, this is the total analysis, and the second P value is representative of the P value after adjusting for the systolic pressure, NIH score, GCS, and ICH scores. In terms of cost, not surprisingly, those who had an early enteral antihypertensive had less than 50% of that cost as a median cost than those in the control group. And this was statistically significant both before and after adjustment. It is important to note that those in the study group did have a significantly higher serum creatinine at admission. However, rates of this did not significantly impact the rates of AKIs between the two groups. Those who were in the control group had a significantly higher proportion of patients that required ventilatory support and were in the ICU more than two times the amount of those in the study group, and this was statistically significant. Again, not surprisingly, in terms of hospital length per se, it was also significantly longer for those in the control group, which was statistically significant prior to adjustment, but statistical significance was lost after adjusting for systolic pressures, NIH, GCS, and ICH scores. Those in the study group had a significantly higher proportion of patients with favorable outcomes, which are patients who are in the ICU more than two times the favorable outcomes, which are patients who survived with an MRS score of zero to two, and there was no difference in terms of mortality or systolic pressure at discharge. However, the overall median cost was significantly lower in those in the study group and had received early internal antihypertensives, which was statistically significant both before and after adjusting for the potential confounding for distribution at baseline. So, the authors concluded that early initial or initiation of oral antihypertensives may significantly reduce nicartopine infusion, ICU length per se, and total cost of hospitalization, and this study demonstrates the significant financial impact of early initiation of oral antihypertensives in patients with hypertensive ICH that is safe and independent of baseline clinical characteristics. Some of the notable strengths for this study is that they did utilize meaningful outcomes, which included assessing both the duration and the cost associated with nicartopine, the length of stay both in the ICU and in the hospital, and then also the overall cost of healthcare that these patients experienced. They also looked…their methodology did address the research question and the outcomes matched the hypothesis. I also appreciated that they performed a multiple univariate analysis and step-by-multivariate regression analysis to help account for some of that confounding and to be able to detect some of the associations that they did. But similar to other studies, there are some pretty significant limitations to this study, particularly the unequal distribution of patient populations at baseline, which would actually favor the results that the patient that was observed in this study. It's also important to note that this is just a single-center retrospective design, which can limit some of that external validity. Also, they only included those patients presenting to the ED with severe hypertension, which can limit some of the generalizability of the results in terms of applying these results to all patients with hypertensive ICH. In terms of takeaway, standardized approach to early enteralized hypertensive medications in patients with hypertensive ICH and severe hypertension may reduce the duration of IV antihypertensives, ICU length of stay, and healthcare costs. Based on these results, I think it's important to have a discussion about standardizing approach to blood pressure management in this patient population and consider enteral medications sooner with earlier discussion of titration and optimization. Something that I would like to see is further evaluation using more of a prospective multicenter evaluation to evaluate this blood pressure algorithm, as I think that that would further elucidate its utility more widespread as opposed to just a single-center retrospective perspective. For our first polling question, how early are you initiating enteral antihypertensives in patients with hypertensive ICH at your institution? Is it less than 24 hours from ED arrival, between 24 and 48 hours after ED arrival, later than 48 hours after ED arrival, or are you starting enteralized hypertensives on a case-by-case basis? So, yeah, this is something, in terms of our institution, we do currently approach this more on a case-by-case basis, but I would be interested to see if starting earlier would be associated with potentially shorter ICU stays and provide potential cost savings for the ICU. Our second polling question is, do you utilize a standardized order set or approach to managing blood pressure in patients with hypertension-associated ICH? So, for us, we currently do not have a standardized algorithm that we are utilizing for the management of our patients with severe hypertension-associated ICH, but it's something that I think that discussion and potentially development of a protocol might provide some benefits for our patient population here. So, with that said, I will close the end of this presentation, and I will take any questions. Thank you, Joe. Our first question is, would you recommend using a similar protocol with aggressive oral antihypertensive, so, for example, amlodipine 10 milligrams or lisinopril 40 milligrams, on day one for your hypertensive intracranial hemorrhage patients? That's a great question. As far as I don't know that I would utilize the same protocol, something that I would be interested in seeing is perhaps shorter-acting agents, something like Captopril, since it has a quick onset, usually within 15 minutes, and has a fairly short half-life, usually about 46 hours, in comparison to something like lisinopril. So, if I were to use that, I would like to see maybe some more shorter-acting agents done on day one, just in case the patients are very responsive, and I would hate to potentially cause any downstream infarction by lowering the blood pressure too much. But I would be also interested to see further evaluation of this protocol, just because I think it's a little bit limited based on the current level or amount of evidence to support its use. Great. I think that's a good segue into this next question. So, we know that blood pressure fluctuation during the first 48 hours in hypertensive intracerebral hemorrhage patients can affect these patients' outcomes. Were the authors in this study able to control fluctuations of blood pressure in their patients by utilizing these enteral options that they started? So, in the previous study that evaluated the utilization of this protocol, they did look at functional outcomes and found no difference between the pre- and post-protocol. However, they didn't necessarily report day-by-day blood pressure management. But you're absolutely right. It would be important to make sure that we are not having extreme fluctuations in blood pressures initially on presentation, just to help, again, avoid any potential development of further infarct and damage to the brain. Great. I think that concludes our Q&A session, Joseph. Great job. Thank you. You're welcome. And now I'd like to introduce our final presenter, Lauren Andrews. Lauren, the floor is yours. Good afternoon, everyone, and thank you again for joining us today. My name is Lauren Andrews, and as mentioned previously, I'm currently concluding my critical care PGY-2 at the University of Illinois at Chicago College of Pharmacy. Today, I will be discussing medication safety during hyperglycemic emergencies. As we all know, DKA and HHS are life-threatening complications of diabetes, with an incidence of greater than 37,000 cases reported annually. These hyperglycemic crises have reported mortality rates of less than 1% and 5% to 16% respectively. Additionally, hypoglycemia is associated with a two- to three-fold increased mortality rate. Precipitating factors that may result in a decreased amount of insulin production or utilization include the five I's, infection, which is the most common, inadequate insulin therapy, initial-onset diabetes, infancy, and myocardial infarction. Other factors include pancreatitis, stroke, and medications. Hyperglycemic crises are treated with aggressive rehydration, insulin therapy, and electrolyte replacement in an effort to correct acidosis and alter mental status in addition to hyperglycemia. The 2009 American Diabetes Associated Guideline, which was published over 10 years ago, is one of the most commonly referred to documents from the management of DKA and HHS. Highlighted in red, we see recommendations for insulin infusion initiation and a range for adjustment following the attainment of disease-dependent serum glucose goals. However, many institutions have taken it upon themselves to devise their own algorithms for more concise stepwise insulin infusion titration, including my own. As evidenced by the NICE sugar trial published by New England Journal of Medicine in 2009, we are familiar with the concept that intensive glycemic control focusing on rapid correction and euglycemic targets has been associated with negative outcomes such as hypoglycemia and mortality in the critically ill population. Although NICE sugar focused on the intensity of glucose control and non-DKA hyperglycemia, the principles employed within their studies still hold true. The Kaplan-Meier curve depicted here provides a visual representation of the results of their primary outcome, 90-day mortality. The probability of survival between the two groups was 11% greater in the conventional control group than in the intensive control group at the 90-day mark. With the adoption of less intense glucose management strategies in the ICU population following NICE sugar, it is reasonable to believe that patients in acute hyperglycemic crises may also benefit from the strategy, since similar concerns would exist in those receiving aggressive insulin treatment. In an effort to improve patient outcomes and promote medication safety when addressing hyperglycemic emergencies in the ICU, a pharmacist-led group of investigators conducted a pre-post study within their large academic medical center over a four-year period entitled Moderate Intensity Insulin Therapy is Associated with Reduced Length of Stay in Critically Ill Patients with Diabetic Ketoacidosis in Hyperosmolar Hyperglycemic State. This study was published in the Journal of Critical Care Medicine in May of 2019 and included adult ICU patients receiving continuous insulin infusions for the treatment of DKA-NIHHS. It was funded by a grant from the NIH National Center for Advancing Translational Sciences. The purpose of this study was to evaluate the impact of moderate intensity insulin therapy on outcomes in patients experiencing hyperglycemic emergency with the hypothesis that this approach would decrease hospital length of stay and prevalence of hypoglycemia. Thus, hospital length of stay was defined as the primary efficacy endpoint with the prevalence of hypoglycemic events or blood glucose less than 70 as the primary safety endpoint. Secondary endpoints included those included in the boxes there, ICU length of stay, average blood glucose, glycemic variability or the standard deviation of blood glucose concentrations collected, severe hypoglycemic events with a serum blood glucose of less than 40, time to anion gap closure, and probability of prolonged length of stay. Of the 205 adult ICU patients screened for inclusion, four of those were excluded due to pregnancy. Other exclusion criteria included the presence of insulin-specific antibodies and deviation from the study protocol. The population was evenly split between the pre-intervention or high-intensity therapy group and post-intervention moderate intensity group. Based on the sample size calculation of 170 patients, the study met power to detect a 25% difference in the primary efficacy endpoint of hospital length of stay. Appropriate statistical tests were utilized to analyze the collected data. Here we can see that the pre-intervention high-intensity group was managed in accordance with the ADA consensus statement with a focus on rapidly normalizing blood glucose levels. The post-intervention or moderate intensity group was designed with a focus on mitigating Compared to the previously defined standard of practice, the elements of this protocol included liberalized target blood glucoses of 200 to 300, mixed per deciliter, avoidance of an initial IV bolus dose of insulin, conservative starting infusion rates of 0.05 to 0.1 units per kick per hour, as well as gradual insulin titrations of no greater than 25%, and a target decline in insulin titrations of no greater than 25%, and a target decline in blood glucose of 26 to 50 mixed per deciliter per hour. Of the 201 patients included in the study, approximately 64% were male with a mean age of 45 years old. Average BMI and hemoglobin A1c were 26 and 12 respectively, with 46% of patients diagnosed with type 1 diabetes at baseline. Mean initial SOFA score was approximately 5, indicating greater than 20% ICU mortality risk, and a mean Charleston comorbidity index of almost 4, indicating an estimated 10-year survival rate of approximately 50%. The authors also reported that a quarter to one-third of patients in each study group presented with infection. Overall, pre- and post-intervention study populations exhibited no significant differences in baseline characteristics. Of note, data regarding the severity of hyperglycemic events on admission, such as pertinent laboratory values, and the presence of all terminal status was not included. In regard to the primary efficacy and safety outcomes, the authors found a 23.6% decrease in hospital length of stay, and a 97.1% decrease in the prevalence of hypoglycemia in the post-intervention or moderate-intensity group. This group also experienced a significant decrease in ICU length of stay at 38%. Glycemic variability, or the mean population standard deviation, as determined from the average plus or minus standard deviation of each subject's blood glucose values, was reduced by 28.6% in the moderate-intensity group. This significant reduction in glycemic variability indicates decreased fluctuation within individual patient blood glucose values. Despite having a 25.6% higher mean blood glucose in the moderate-intensity group, which was taken over an undefined time frame, there was no significant difference in time to anion gap closure. In addition, there were no reported episodes of severe hypoglycemia in the moderate-intensity group compared to four in the high-intensity group. In order to illustrate the final secondary outcome regarding probability of prolonged length of stay for ICU and hospital, the authors included a Kaplan-Meier curve for each, labeled here as figures one and two. In figure one, we see that the probability of remaining in the ICU at 48 hours after admission was significantly reduced by about 66% when utilizing the moderate-intensity strategy. When directing our attention over to figure two, we see that the probability of hospitalization on day seven and day 14 were significantly reduced by about 50% and 70% respectively when utilizing the moderate-intensity strategy. In addition, multivariate regression identified SOFA score, intensity of insulin therapy, and glycemic variability as independent predictors for ICU and hospital length of stay. Although this was a single-center retrospective study with a relatively small sample size, the provision of a pragmatic study design and clearly defined treatment protocol for the evaluation of outcomes during the pre- and post-intervention periods allowed for an increased likelihood of reproducible outcomes if repeated by an outside party. Additional strengths included sufficient power to detect a difference for both efficacy and safety outcomes and the implementation of minimal exclusion criteria to increase the generalizability of the results. While the authors listed their final limitation as the effect of uncontrolled variables on observed differences in length of stay, including the underlying cause of hyperglycemic crisis, logistical and administration inefficiencies, and timeliness of transition from IV to subcutaneous insulin, there are a few other limitations to overall clinical application that warrant further discussion. In order to appreciate the acuity of the overall patient population, access to the clinical criteria utilized for the diagnosis and stratification of hyperglycemic crisis severity is imperative. Unfortunately, pertinent laboratory values such as pH, serum bicarbonate, anion gap, beta-hydroxybutyrate, and serum osmolality were unavailable. Moreover, one of the study inclusion criteria was admission to the ICU. However, the authors reported a total of 100 of the 201 included patients that required admission to the ICU. This may be confusing to the reader and again raises the question of the acuity of the patients included. As timely blood glucose checks and insulin titrations by nurses are imperative to any insulin protocol success and to decrease risk of harm secondary to hypoglycemia, it is also important to know the patient-nurse ratio, which was not mentioned in the study. A higher patient-nurse ratio, such as on an intermediate care unit versus an intensive care unit, may dictate how often the nurse is able to check blood glucoses and titrate insulin, and thus moderate intensity titrations may be safer in that context compared to aggressive titrations for rapid blood glucose correction. The impressive hypoglycemia and glycemic variability reduction rates in the study need to be interpreted within the context of these human factors as well, even if they are uncontrolled for. Additionally, it is important to know how the protocol was carried out to assess feasibility of implementation at a different institution. For example, if this was a nurse-driven protocol, how were the algorithms followed? Were they posted on the institution's intranet as a paper protocol for the nurse to do manually? Another strategy could have been to utilize an electronic tool embedded within the EHR that could assist with timing of point-of-care glucose testing and subsequent insulin titrations. All of this additional information would help the reader determine if this is a feasible endeavor for their own institution. Based on these results, the authors conclude that moderate-intensity insulin therapy for the management of DKA and HHS reduced glycemic variability and hypoglycemia, which were associated with improvements in ICU and hospital length of stay through the utilization of liberalized target blood glucose goals, avoidance of an initial IV insulin bolus dose, reduced starting insulin fusion rates, and conservative titration strategies for a gradual decline in glucose. As the reader, I would like to echo the significance of these outcomes while also highlighting the major takeaway from this study, which is the need for a thorough revision of the 2009 ADA guidelines to better address patient safety in addition to treatment efficacy. In terms of clinical applicability, utilization of this moderate-intensity insulin strategy resulted in a number needed a treat of three in order to prevent one episode of hypoglycemia when compared to the control group. This is particularly important because many patients experiencing hypoglycemic emergency may not experience the adrenergic manifestations of hypoglycemia, such as sweating, nervousness, fatigue, hunger, and tachycardia, potentially leading to worsening morbidity and already critically ill population. If the outcomes of this study were to be further evaluated within a randomized control trial, some additional data points to consider might include an expansion of demographic identifiers to include variables like race in an effort to identify health disparities in the patients included for analysis, prior insulin dependence to address issues of non-adherence versus the need for a regimen revision due to inadequate dosing, and baseline renal dysfunction or replacement therapy in order to determine potential increased risk of hypoglycemia associated with impaired insulin clearance. Another factor that may have an impact on the interpretation of future analysis is the inclusion of a clear method for transitioning from IV to subcutaneous therapy upon DKA HHS resolution, as there are mixed recommendations in the literature for how to safely and effectively accomplish this. In regards to my own institution, UI Health is committed to reviewing our hospital's overall insulin-related hypoglycemia data as part of our Medication Safety Review Committee's Zero Harm Initiative. Based on data collected over a 10-month period, our first endeavor was to revise our current hyperkalemia treatment protocol in order to address the most commonly associated intervention preceding hypoglycemia at our institution. However, we also identified an opportunity to improve our IV insulin infusion protocol. Although we currently have a nurse-driven algorithm for titrating IV insulin infusions in the setting of hyperglycemic emergencies, it lacks clear instruction for transitioning from IV to subcutaneous therapy and thus is currently being updated. To reflect on the information we discussed during the session, our first polling question asks, in what scenarios or scenario have you most frequently witnessed hypoglycemic events in your ICU patients? Is it A, IV insulin infusion titrations, IV to subcutaneous insulin transitions, hyperkalemia management, or another reason? Okay, so a decent number of you said that IV insulin infusion titrations were the leading cause for hypoglycemic events in your ICU patients, which may mean that the conclusions of this study could be valuable and something to use at your own institution. As I mentioned on the last slide, UI Health identified hyperkalemia management as the leading cause for our hypoglycemia cases. However, for those of you that answered A or B, I'm curious to see what your response to this next polling question will be. So here, does your institution have a clearly delineated physician or nurse-driven insulin infusion protocol for patients treated with DKA-HHS? Yes, no, unsure, or people do that? Okay, so a majority of you all use a physician or nurse-driven insulin infusion protocol. So, makes it a lot easier to document this information. And those of you that didn't respond with yes, this might be something that will allow you to revise your own current insulin titration methods. I've included my contact information here at the end of my presentation for reference, and with that, I'm happy to open the floor for questions. Thank you, Lauren. The first question, though the difference was not statistically significant, can you hypothesize why gap closure was actually faster in the moderate intensity group? It's hard to infer because there weren't a lot of other laboratory values that were included in the study, so that we could completely review the acuity of the patients, but I'm curious if, as the gradual titrations were occurring, you know, the concomitant IV infusion therapies of fluid rehydration and electrolyte repletion also assisted in the correction of their anion gap. So, not only the insulin intervention, but also the additional interventions for treating DKA-HHS having an impact on that. Great. And you sort of alluded to this throughout your presentation, and you sort of alluded to this throughout your presentation, that the authors even noted that length of stay as their primary efficacy outcome obviously was wrought with a lot of confounding that they couldn't even account for within the study. Would you have chosen a different primary efficacy outcome if you could redo this study? You know, when you're doing a study like this, the major goal is to identify how to potentially save money at your institution to find value for, you know, utilizing pharmacist time to do these kinds of things. So, I think from an administration standpoint, hospital length of stay is really important. I think for me, I might have focused on ICU length of stay instead if I was looking at something from the administration side, but I think even from a safety standpoint, my biggest focus would be the prevalence of hypoglycemia and actually utilizing that as the primary focus versus trying to validate my outcomes based on length of stay. Great. Thank you. And I would like to thank everyone today for joining us. This will conclude our last Q&A session, and I want to give a specific thanks to our speakers today, so Kate, Joe, and Lauren, and all of you again for attending this late on a Friday as well. I'd like to ask all of you to please join us on July 17, 2020 from 2 to 3 p.m. for the next Journal Club. This will conclude our presentation today, and everyone have a great weekend.

critical care medicine

prothrombin-complex concentrates

Factor Xa inhibitor-related intracranial hemorrhage

hemostasis

thrombosis

retrospective design

early enteral antihypertensives

hypertensive intracranial hemorrhage

ICU length of stay