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 Melanie Madorsky, and I'm the cardiac ICU clinical pharmacist at Memorial Hermann Texas Medical Center in Houston, Texas, and I will be moderating today's webcast. A recording of this webcast will be available to registered attendees. Log into mysecm.org and navigate to the My Learning tab to access the 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. SECM 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 SECM. SECM does not recommend or endorse any specific test, position, 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 Q&A. Our first presenter today is Brad Hedick, TGY2 Critical Care Resident at Aurora St. Luke's Medical Center in Milwaukee, Wisconsin. Our second presenter is Sophia Brown, TGY2 Critical Care Resident at Northwestern Memorial Hospital in Chicago, Illinois. And our third presenter is Catherine Moore, TGY2 Critical Care Resident at Methodist University Hospital in Memphis, Tennessee. And now, I'll turn things over to our first presenters. So, as mentioned, my name is Brad Hedick and I'm a resident with Aurora St. Luke's Medical Center. Today's study that I'll be presenting is the American Journal of Emergency Medicine, Efficacy of Ketamine for Initial Control of Acute Agitation in the Emergency Department, a randomized study. As anyone staffing in the emergency department knows, agitation can be a common presentation often characterized by behavioral issues including destructiveness, disorganization, or dysphoria. Although non-pharmacologic management is typically preferred, medications may be necessary when verbal de-escalation fails with the goal of calming the patient enough to further assess their altered state. Multiple organizations have released consensus statements recommending the use of benzodiazepines or first or second generation antipsychotics for the management of acute agitation. Benzodiazepines are preferred for undifferentiated agitation or when related to alcohol withdrawal, while antipsychotics are preferred for underlying psychiatric etiology. And a combination of the two classes may be used for more rapid sedation. In 2017, ACEP released an updated clinical policy statement suggesting that ketamine may also be a reasonable treatment for severely agitated patients in the emergency department, although this was a level C recommendation, meaning it was based off of expert consensus in the absence of adequate published literature. The previous trials have all been retrospective reviews or prospective observational studies, but they have suggested safety and efficacy of ketamine for acute agitation. In one retrospective case series with 27 patients, intramuscular ketamine given at doses of 40 to 400 milligrams was not associated with any significant adverse events when treating acute agitation in the ED, although patients did experience a transient increase in blood pressure or heart rate. Riddle and colleagues also evaluated ketamine as a first-line agent for agitation in the emergency department. They performed a single center prospective observational study with 98 patients using IV or IM ketamine at doses ranging from 0.3 mg per kg to 4 mg per kg. The study determined that ketamine was faster at controlling agitation compared to other agents, which included midazolam, lorazepam, haloperidol, or a combination. And when we think about the PK of all of these agents, these results actually do make a lot of sense. This table is included as a reference to highlight these differences. Ketamine would be expected to work quicker as it has a faster onset of action of about 1 minute for IV and 5 to 15 minutes for the IM route compared to benzodiazepines, which may take up to 30 minutes, and antipsychotics, which range from 15 minutes to 60 minutes. Now before we move on to the specifics of the trial, I would like to gauge current practice of our audience members. So what is your preferred agent for management of acute agitation in the emergency department? All right, and it looks like majority are saying a combination of benzodiazepine and antipsychotic followed by benzodiazepine on its own and antipsychotics at the end. No one is preferring ketamine at this time. And I would say this distribution is fairly similar to what I've seen at our site. Typically we are favoring the combination benzodiazepine, antipsychotic, usually lorazepam and haloperidol for that severe agitation. And I also have not seen the use of ketamine for acute agitation in the ED. Now considering the retrospective and observational design of the previous trials, higher quality studies are clearly needed and the study today is the first randomized study identified. Here the authors sought to answer the question, can ketamine be used effectively and safely to treat acutely agitated patients in the emergency department. This was a prospective randomized study that was active January through October of 2018 and the authors specifically evaluated ketamine compared to the combination of haloperidol and lorazepam for acute agitation. All interventions that were included were considered part of the routine management for any participating ED physician. Patients were included if they were adults admitted to the ED with an active diagnosis of combative agitation or enrichment agitation and sedation score equal to 4. Patients were excluded if they had a known diagnosis of pregnancy, schizophrenia or uncontrolled hypertension among others. And all of these criteria were based on published guidelines for the use of ketamine for dissociative sedation in the emergency department. Patients were then randomized to a ketamine or combined haloperidol lorazepam group using a computer generated random number table and patients received either ketamine 1 milligram per kilogram IV, 4 milligrams per kilogram IM or haloperidol 10 milligrams with lorazepam 2 milligrams through the IV or IM route depending on IV access. Additionally doses of haloperidol and lorazepam could be decreased at the discretion of the provider based on pre-specified criteria which included significant comorbidities, age, weight and any medication or ingestions prior to admission. An estimated sample size of 84 patients was needed to detect a 30% difference with a one-sided alpha of 0.05 and statistical power of 80%. Statistical tests included the t-test, rank-sum test and Fisher exact test to assess for bivariate associations between the variables and treatment groups which all seemed adequate to achieve their objective. In total the goal sample size was met, however of 152 patients who received sedative medications for agitation or aggressive behavior during the study period only 93 patients were actually enrolled and randomized. Other patients were excluded from the analysis primarily due to unknown patient outcomes which left 41 patients to be analyzed in the ketamine group and 45 patients in the haloperidol lorazepam group. Now that said patients did appear to be well matched at baseline and were typically white males in their late 30s with an overall low incidence of comorbidities with the exception of psychiatric conditions. In this case 43% of patients in the ketamine group and 67% of the haloperidol lorazepam group had a history of a psychiatric condition which may have included depression, anxiety, bipolar disorder or even schizophrenia. Additionally approximately 30% of patients in the haloperidol and lorazepam group received a reduced dose of at least one of the medications. The authors assessed a number of efficacy outcomes with the primary endpoint of adequate sedation within five minutes based on a RAF score of less than or equal to zero. Secondary outcomes included adequate sedation within 15 minutes, time to sedation median RAF score at 30 minutes, additional set of sedation medications administered within 30 minutes and for safety the occurrence of adverse events. It is important to note that if an adverse event was not able to be assessed it was not included in the analysis. Similar to previous studies the authors did determine that ketamine was significantly more effective in achieving adequate sedation with the primary endpoint occurring in 22% of patients compared to 0% in the haloperidol and lorazepam group. Likewise significantly more patients were sedated within 15 minutes and patients achieved a deeper sedation in a shorter mean time with the ketamine. And although there was no significant difference for additional sedative medications 30 minutes later it is noteworthy that the initial dose was not adequate in about one-fifth of patients in both groups. Safety outcomes were also similar to other studies. Patients in the ketamine group experienced significantly more hypertension and tachycardia compared to the haloperidol and lorazepam group. Additionally there was a non-significant increase in hypoxia in the ketamine group and the incidence of QTC being greater than 450 milliseconds was similar between the groups but was only measured in about half of the patients. Other safety outcomes were fairly unremarkable and the authors did not have any documentation of emergence reactions in the ketamine group. Although they did acknowledge that it would have been difficult to determine whether patients had returned to their agitated baseline or if they were experiencing a new emergence reaction. So based on these results the authors concluded that ketamine was significantly more effective than haloperidol and lorazepam for initial control of acute agitation and was not associated with any significant adverse effects in patients with combative agitation. As for my take on the study I thought there were a number of strengths including being the first randomized trial to evaluate ketamine for acute agitation. Additionally the study met their goal sample size, utilized medication doses that have previously demonstrated efficacy, and evaluated clinically relevant endpoints especially when considering a potentially violent patient. However there were also some limitations. For example the authors intended to exclude patients with schizophrenia based on a previously published guideline in which schizophrenia was categorized as an absolute contraindication due to the potential for ketamine to exacerbate symptoms. However as we know about half of the patients were later found to have an underlying psychiatric condition but it was unclear if how many may have actually had schizophrenia versus some other psych condition and it was unclear if this could have contributed to the number of patients who required additional sedative medications. Study design may also have contributed to bias. There was only a core group of ED physicians who participated in the study which ultimately led to 59 patients not being randomized despite being treated for acute agitation. This increased the risk for selection bias in addition to the study being unblinded and given the fact that haloperidol and lorazepam doses could be decreased at physician discretion. There was also an observation bias risk as the outcomes relied on nursing documentation for timing of adequate sedation. As it related to the safety outcomes some data was limited due to lack of availability specifically the QTC which was only able to be collected in about half of the patients in each group. To me this was a little concerning when considering the large number of patients who ended up having an underlying psychiatric condition at baseline as these patients may have been on other QT prolonging medications. And lastly in the in both groups more than 90% of the administered doses were via the IM route. While this is likely more realistic due to a potential lack of line access or uncooperative patients in the emergency department this limits the generalizability to IV dosing. And now having reviewed the results of the study do you see yourself recommending ketamine for acutely agitated or combative patients in the emergency department? All right, now it looks like there is a pretty even split, maybe slightly favoring the ketamine group. As I mentioned, I personally have not witnessed ketamine being used for acute agitation. Our providers tend to use the more conservative doses of the haloperidol or lorazepam, and then redosing them more frequently as needed. But that said, I do feel that I would be comfortable recommending the use of ketamine for acute agitation. Overall, I felt that this trial did strengthen the case for its use. And some other key takeaways that kind of support my reasoning was that IM ketamine appeared to be more appropriate than the other agents when considering those severely agitated or combative patients. Just thinking about someone who may pose a safety risk to either themselves or to our staff, the shorter time to onset and ability to achieve adequate sedation with ketamine would be ideal to gain better control of that situation. However, when recommending ketamine, I think it's also important to acknowledge that it may require subsequent doses or other agents to maintain that sedation. As we saw in the study, about one-fifth of the patients did require additional sedative medications within 30 minutes. So to me, this just suggests that ketamine may provide more time for that initial assessment and ability to address the underlying etiology, but that first dose may not be the final solution in all cases. And for those of you who may not have been sold on ketamine just yet, know that there are other studies in process, one of which is the RACT trial, which is a randomized study comparing the time to adequate sedation for intramuscular ketamine compared to the combination midazolam and haloperidol for this similar patient population. And the results are expected at the end of this year. Now that concludes the content that I have for you. So at this time, I can field any questions that you may have. Nice job on your presentation, Brad. Our first question is, benzodiazepines are preferred for agitation related to alcohol withdrawal. What impact do you think not excluding these patients had on the study? Well, I think, so since benzodiazepines are generally preferred, it's still the standard of care. I think the ketamine is, this study is just giving us an alternative to our current options. I think that we have to do a real-world study in order to assess this, to just strengthen the data and our current results. Okay, and another question. What is the average change in QCC with Haldol, and is this clinically relevant? I don't know the exact number of the QCC with Haldol. I think it is still fairly low. So it may not be to a clinically relevant level. However, in the study, we did not have any baseline QCCs. We just had that it was over 450. And so if a patient, let's say, already had a QCC greater than 500 and was still administered additional doses, I think that would still pose a safety risk and increase that risk for TRSAT. Okay, so that concludes our Q&A session. Thank you, Brad. I'd now like to introduce our second presenter, Sophia Brown. Hi, everyone. My name is Sophia Brown, and I'm a PGY-2 critical care pharmacy resident at Northwestern Memorial Hospital. And today I'll be presenting my journal club on the comparison of a hydromorphone versus fentanyl-based sedation regimen in extracorporeal membrane oxygenation, a propensity-matched analysis. I have no actual or potential conflicts of interest to disclose related to this presentation. Extracorporeal membrane oxygenation, or ECMO, is a form of life support that can support the heart and or lungs in patients with reversible respiratory or cardiac failure. Veno-arterial is a modality which supports both the heart and lungs, taking blood from a vein and returning it to an artery to circulate through the body. Veno-venous is a modality which supports the lungs only. Both cannulas are placed into veins, and the patient's heart circulates the blood. The number of ECMO runs has greatly increased over recent years, from about 3,400 runs in 2010 to over 12,000 in 2019. The Extracorporeal Life Support Organization, or ELSO, guidelines recommend thorough sedation during cannulation and during the first 12 to 24 hours after, but advise that sedation should be minimal thereafter based on the patient's level of anxiety or discomfort. Most of the literature surrounding drug properties in ECMO are from ex vivo studies which aimed to determine the impact of drug sequestration and protein binding on drug recovery from ECMO circuits. Lipophilic protein-bound drugs were found to have significant losses in drug recovery. Specifically, the mean drug loss of fentanyl in the ECMO circuits was found to be 97%. The table below shows the lipophilicity, as indicated by the log P value, and protein binding of fentanyl compared to hydromorphone, showing fentanyl's significantly higher lipophilicity and protein binding drug properties, which is potentially problematic given this is our general first-line agent in most non-ECMO ICU populations. In a recent international survey sent to SCCM members in January of 2018, it was noted that 77% of responders still use fentanyl as the preferred opioid in ECMO patients. Propofol was the most common medication used for deep sedation, and 73% of responders reported avoiding benzodiazepines for delirium prevention, with the most common treatment of delirium being atypical antipsychotics. I have two polling questions here to get a sense of current practices amongst the members listening today. The first is, what is the most common in-algo sedation regimen used in ECMO patients at your practice site currently? A, a hydromorphone-based regimen with or without propofol, B, a fentanyl-based regimen with or without propofol, C, hydromorphone with or without a benzodiazepine, D, fentanyl with or without a benzodiazepine, or E, other. And we have our results here. It seems that most of our respondents went with a fentanyl-based regimen with or without propofol, and I would say this does correlate to our institution's current practice. We also generally use fentanyl with or without propofol for sedation in these patients. Secondly, what is the most common first-line agent used for rescue delirium and agitation in ECMO patients at your practice site? A, haloperidol, B, olanzapine, C, quetiapine, or D, benzodiazepines. And our results indicate that most of our responders do use cutiapine for delirium and agitation, what that looks like second being haloperidol, and I would say that this also correlates roughly with our distribution here at this institution. We utilize cutiapine the most. Sometimes olanzapine will also be used as well, which it looks like very few, no respondents answered to this. This brings us to the hypothesis that this study was attempting to answer. Will hydromorphone, a more hydrophilic and lower protein-bound medication, result in more days alive delirium-free and coma-free with less narcotic and sedative exposure than fentanyl? This was a single center retrospective observational study done at the University of Pittsburgh Medical Center Presbyterian Hospital from January of 2016 to December of 2018. Adult patients 18 years and over were included if they had been receiving ECMO for over 48 hours, continuous infusion fentanyl or hydromorphone for at least 6 hours, and received fentanyl or hydromorphone for 75% of the time they spent on continuous infusion sedation. Patients were excluded if they had been cannulated for over 24 hours prior to transfer or were receiving continuous infusion paralytics. UPMC changed their analgesic practice from a fentanyl-based and algal sedation regimen to a hydromorphone-based regimen for all ECMO patients in 2018. So the primary outcome was compared between fentanyl sedation from January 2016 through September of 2018, before this practice change, to hydromorphone sedation from September of 2018 to December of 2018, after this practice change. The primary endpoint was the number of days alive, delirium-free, and coma-free between the hydromorphone and fentanyl groups at days 7 and 14. This was calculated by dividing total ECMO days alive, delirium-free, coma-free by total ECMO days evaluable. ECMO days were considered evaluable days when at least a single SAS and ICDS-C score were reported. Delirium was defined as an ICDS-C score greater than or equal to 4, with a coma defined as a SAS score less than 3. The secondary endpoint was narcotic and sedative exposure, which was reported as mean narcotic, benzodiazepine, and antipsychotic use in terms of fentanyl, midazolam, and chlorpromazine equivalents, respectively. Patients were followed starting 12 hours after cannulation for all outcomes to account for increased periprocedural sedation and paralytic requirements. Continuous variables were reported using median and interquartile range, or sample size and percentage, and compared with the Mann-Whitney U test, and categorical data were analyzed using the Fisher EXACT test. Propensity score matching was done in this study to account for the covariates that may influence the primary outcome. Groups were assessed both before and after matching to account for these variables, which included type of ECMO, open chest, SOFA score, age, hepatic failure, and weight. The two groups were matched one-to-one and no cofactors were statistically significant. The caliper was set at 0.1. A best-worst worst-best-case sensitivity analysis was also conducted to address the effect of the days that were non-evaluable due to data not being able to be collected from the electronic medical record. A p-value of less than 0.05 was considered statistically significant. There were 272 patients screened, and after exclusion, 88 were included in the fentanyl group and 60 in the hydromorphone group. After matching, there were 54 patients included in each group. This table depicts the characteristics which the propensity matching was based on. Of these, SOFA score and body weight were statistically significant before matching. No cofactors were significantly different between groups after matching. A total of 1,114 patient days were evaluated before matching and 548 days after matching. The hydromorphone group had a higher proportion of delirium-free coma-free days at day 7 and day 14 before matching, although these were not statistically significant. After matching, the hydromorphone group had significantly more delirium-free coma-free days at day 7. Additionally, at day 7, when the time period was expanded to 24 hours after cannulation, the hydromorphone group in the matched population maintained an increase in days alive and delirium-free coma-free. There was also a statistically significant difference both before and after matching in the percent of coma-free days but not between the percent of delirium-free days, indicating that the significant difference between the primary outcome was driven by the coma-free days and a reduction in over-sedation. This table depicts the drug utilization outcomes both before and after matching between the hydromorphone and fentanyl groups. Before matching, there was a statistically significant increase in the amount of fentanyl equivalents used in the fentanyl group as compared with the hydromorphone group, nearly four times greater in fentanyl equivalents. There was no significant difference in midazolam equivalents, propofol, or clopromazine equivalents. After matching, similar results were found, a significant reduction in fentanyl equivalents administered in the hydromorphone group and no statistically significant difference in the other requirements. Two subgroup analyses were also conducted. The first found a significant increase in number of days alive delirium-free coma-free at day 7 in V-A ECMO but no difference in V-V ECMO, although both cohorts retained the significant reduction in fentanyl equivalents in the hydromorphone group. The second found a significant increase in number of days alive delirium-free coma-free in the acute respiratory failure ECMO indication but no difference in the non-acute respiratory failure group. The author's conclusions from this trial are that patients receiving hydromorphone-based in-algo sedation while on ECMO support had more days alive in delirium-free coma-free than those receiving fentanyl-based in-algo sedation. Hydromorphone was associated with more coma-free days and about 75% less narcotic exposure than fentanyl-based in-algo sedation. And they report that the results are clinically relevant because ICU delirium is associated with negative outcomes, including increased mortality, length of stay, and mechanical ventilation days. Strengths of this study include its relevance to current practice as there has been a drastic increase in ECMO cannulations in recent years. The endpoints appropriately assess the study questions and the outcomes support the study's initial hypothesis. The study included a patient population with generalizable baseline characteristics and performed a propensity-matched analysis to eliminate the influence of covariates. It also allotted a 12-hour time window between cannulation and data point collection to allow for increased sedation around the time of cannulation. There is a considerable sample size in this study for this specific patient population. Weaknesses of the study include its retrospective observational design, which limited data collection to what was already recorded in the electronic medical record. This led to the fact that not all ECMO days were evaluable, and evaluable days were defined as only requiring one single reported SAS and ICDS-C score per day, which may not be externally valid as most institutions evaluate these on a much more frequent basis, specifically SAS or RAS scores. Delirium and sedation scores used are not consistent with preferred scoring tools. This study used SAS and ICDS-C scoring, whereas current guidelines recommend the RAS sedation score and CAM-ICU to assess delirium. Additionally, SOFA scores were reported to assess ICU mortality instead of SAVE scores, which have been validated to predict survival in VA ECMO patients. The population was all-inclusive, encompassing both VA and VV ECMO types, as well as etiologies including acute respiratory failure and those with non-acute respiratory failure etiologies. Although all patients in this population were cannulated with ECMO, the population is not homogenous and there may be differing sedation requirements based on patient's primary indication for ECMO. The influence of education to nursing and staff regarding sedation assessment due to this new protocol implementation could have also influenced the study outcome. Though with statistically significant results, the study has some key limitations that are important to consider in its interpretation. This propensity match analysis showed an 11.1% absolute risk reduction for ECMO days alive, delirium-free, coma-free in the hydromorphone group at day 7, which correlates to a number needed to treat of 9 patients. However, although a large absolute risk reduction was observed, factors such as the etiology of ECMO as well as education bias that could have surrounded this new protocol implementation should be taken into consideration. Patients did receive fewer ECMO, fewer fentanyl equivalents per day of narcotics in the hydromorphone group, likely attributed to its favorable PK profile. And although the primary outcome was assessed in days alive, delirium-free, coma-free, when stratified, only the percent of coma-free days was statistically significant in both the matched and non-matched populations, which indicates that over sedation occurs less frequently with the hydromorphone group. While this could lead to less delirium, this was not specifically observed in this study, although it was discussed in the study as a clinically relevant outcome. Overall, although there are limitations to its design, the study did account for differences in baseline variables and non-evaluable ECMO days. Based on its PK properties, we can conclude that hydromorphone-based analgo-sedation in patients on ECMO is associated with less over sedation and narcotic exposure than more lipophilic drugs like fentanyl. Based on these results and the information about their pharmacokinetic profiles, it seems as though hydromorphone may have a more favorable PK-PD profile in this population. Given the strengths and limitations of this study, it may not be enough to solidify hydromorphone's place in therapy as a first-line agent, but serves as hypothesis-generating evidence to consider it based on patient-specific factors. Future directions relating to ECMO effects on drug absorption and choice of therapy would include an even larger study population with a prospective study design to increase number of evaluable ECMO days, evaluation of specific patient populations based on ECMO etiology using preferred scoring tools to assess sedation and delirium, and longer follow-up time to assess post-ICU cognitive function. That concludes my presentation for today, and I will now take any questions that anyone may have. Good job, Sophia, on that presentation. So one question is, to your knowledge, have the results of this trial changed the way your practitioners and pharmacists practice at your institution in terms of what agents they choose? Great question. Thank you for that. To my knowledge, the results specifically of this trial have not changed the way we practice specifically. At our institution, the majority of practitioners will still use fentanyl and propofol in these patients, unless for a very specific patient population, for example, patients that may be experiencing tachyphylaxis on high doses of fentanyl infusions, but overall based specifically on this study, no. Okay, thank you. And then in your slide, you have that within the appropriate place and therapy, you would use propofol if deeper sedation is indicated. So do you have any concerns with the use of propofol, considering it is highly lipophilic and also highly protein-bound, considering the findings of fentanyl in this study with those properties? Yes, that's another good point to point out, and I did consider that when making that conclusion on that slide. Fentanyl does have a similar lipophilicity and protein binding to fentanyl, so that is an important thing to consider. I think that in that recommendation, that would only be specifically in patients that really were requiring a deeper sedation. So for example, in maybe some of these patients that were included that were requiring VVECMO due to an ARDS indication that really were requiring a deeper level of sedation that was not being able to be reached. In addition, based on propofol versus midazolam and benzos, I kind of chose the propofol route just due to potential for renal accumulation in some of these critically ill patients that we might see with a benzodiazepine. Okay, thank you for answering those questions. That concludes our Q&A session. Thank you, Sophia Brown. Now, I'd like to introduce our final presenter, Catherine Moore. Okay, thank you for the introduction. Today, I'm going to be talking about the effects of a high-dose 24-hour infusion of tranexamic acid on death and thromboembolic events in patients with acute gastrointestinal bleeding or the HALT-IT trial. So, GI bleeding is a common disease state that leads to significant morbidity and mortality in hospitalized patients. In fact, it has been reported that 30-day mortality can be up to 11% in patients experiencing an acute GI bleed. Patients can have a wide variety of clinical presentations, and there are recommended treatments based on the etiology of the bleed. I've listed a few common etiologies on this figure, including peptic ulcer disease, esophageal varices, and malignancy. In general, upper GI bleeding is more common than lower GI bleeding. When evaluating the severity of GI bleeds, the ROCKALL score can be used to predict mortality. So, there are two variations, one for pre-endoscopic GI bleed and one for complete GI bleed after the etiology is known. It's calculated using the ABCDE method, A for age, B for blood pressure and signs of shock, specifically systolic blood pressure and heart rate. C is for comorbidities such as renal failure, malignancy, and liver disease, among others. And then D is included in the complete ROCKALL score or post-endoscopy and stands for diagnosis or source of bleed. So, the pre-endoscopic ROCKALL, which is used in HALT-IT, ranges from 0 to 7, and a score of 4 equates to approximately a 25% risk of mortality. So, in HALT-IT, investigators study the use of tranexamic acid, or TXA, in this acute GI bleed population. While TXA has been studied in other acute bleeding populations, no current guidelines recommend its use in GI bleeding. So, now diving further into the mechanism of TXA, normally, plasminogen becomes bound to an activator and becomes plasmin in the presence of fibrin. This complex goes and binds to fibrin at a lysine binding site and degrades it. TXA is a competitive reversible inhibitor at the lysine binding site on plasminogen. Once it's bound, TXA prevents plasmin activator complex from binding to fibrin, preventing the breakdown of fibrin to degradation products, and ultimately stabilizing the fibrin matrix of the thrombus. So, when it comes to its use in life-threatening bleeds, TXA has shown mortality benefit in several major trials, including CRASH-2 for major trauma, WOMEN for postpartum hemorrhage, CRASH-3 for TBI, and MATTERS for military-based trauma. In these studies, the mortality benefits of TXA were most significant when TXA was administered within three hours of the onset of injury or bleeding. Prior to HALT-IT, no major randomized controlled clinical trials had been conducted evaluating the role of TXA compared to placebo in acute GI bleeding. A meta-analysis of eight studies was published in 2014, suggesting that there was a significant decrease in all-cause mortality with the use of TXA in the acute GI bleed population. However, these studies were smaller. Many had comparator groups consisting of acid-reducing medications, and the duration of therapy was anywhere from two to seven days. The mortality benefit found in this meta-analysis was driven by a study by Barer et al., in which 58% of the patients were known to have peptic ulcer disease, and therapy was continued for seven days. So before we dive into the details of HALT-IT, I'll pose the first polling question. Does your institution currently use TXA in the setting of acute GI bleeds? Okay, so about 90% of you answered no, which would be in line with the current guidelines and is also what we practice at my current institution. So now let's dive into HALT-IT. The purpose of HALT-IT trial was to answer the clinical question, does IV TXA reduce five-day death due to bleeding in adult patients with acute GI bleed compared to placebo? HALT-IT was an international, randomized, double-blind, placebo-controlled clinical trial, and it was conducted in 164 hospitals across 15 countries in Europe, Africa, Asia, and Australia. It included patients with significant upper or lower GI bleeds. Significant bleeding was defined as risk of bleeding due to death and included patients with hypotension, tachycardia, signs of shock, or likely to need transfusion, endoscopy, or surgery. In order for patients to be enrolled, the clinician also had to be substantially uncertain as to whether or not to use TXA. So if the physician or patient felt that either TXA or placebo would be harmful, they would not be included. If the patient had a clear indication for or a clear contraindication to TXA, they were excluded from this trial. The patients were randomized to receive either TXA or placebo. The TXA group received a loading dose of one gram over 10 minutes, followed by a three gram administered over 24 hours as a slow infusion. The matching placebo group received a normal saline bolus, followed by an isotonic fluid placebo infusion. The primary outcome of this study is death due to bleeding within five days. Please note the asterisks as we will revisit this on the following slide. Select pertinent secondary outcomes include death due to bleeding within 24 hours, death due to bleeding at 28 days, all-cause mortality at 28 days, incidence of re-bleeding at 24 hours, five days, and 28 days as three separate outcomes, blood product transfusion, thrombotic event, and seizures. So circling back to the primary outcome, it is important to note that this was amended while the trial was underway. So the initial intent of the study and the sample size calculations were based on a primary outcome of all-cause mortality. However, while the trial was ongoing, the investigators found that over half of the deaths were due to non-bleeding causes. So prior to unblinding the data, the primary outcome was amended to death due to bleeding in five days. So based on the amended primary outcome, 12,000 patients would be needed to detect a 1% absolute risk reduction from 4% to 3% with an 85% power and significance level of 0.05. Looking at the patient's baseline characteristics, they were well-matched between the two groups. Of note, about 58% of patients were randomized greater than eight hours after the suspected onset of their bleeds, 89% in both groups had suspected upper GI bleeds, and about 46% in both groups had suspected variceal bleeding. Slightly less than half of the patients in both groups had signs of shock, but only 87% of patients were maintaining a systolic blood pressure greater than 90. Less than 10% of patients in each group were known to be taking anticoagulants. And finally, a quarter of the patients had a Rockwell score of one to two, which would put them at the lowest risk of mortality due to GI bleeding. So for the primary outcome of death due to bleeding within five days, this occurred in 3.7% of patients in the TXA group and 3.8% of patients in the placebo group. And this difference was not statistically significant. There were four pre-specified subgroups of the primary endpoint, which were stratified by time to treatment, suspected location of the bleed, cause of bleeding, and the pre-endoscopy Rockwell score. No statistically significant differences existed between the two groups for any of these subgroups. For the secondary outcomes of death due to bleeding within 24 hours, death due to bleeding within 28 days, pre-bleeding within five days, and units of blood and FFP transfused, there were no significant differences found between the two groups. And the other secondary outcomes not displayed, there were again no significant differences. Looking at the secondary outcomes related to safety, there was no difference in arterial thrombotic events, which included MI and stroke. However, for venous thrombotic events, 0.8% of the patients in the TXA group versus 0.4% of patients in the placebo group did have VTE events, which was considered a significant increase. So in this table, you can see the breakdown of the pulmonary emboli and deep vein thrombosis. In addition to venous clots, there was also a significant increase in the amount of patients that experienced seizures in the TXA group. So based on these results, the authors of the study concluded that a high-dose 24-hour infusion of TXA did not reduce death due to GI bleeding within five days, but was associated with an increased risk of VTE events and seizures. So some strengths of this trial are that it has a robust sample size, enrolling 12,000 patients. It is a multicenter randomized placebo-controlled trial of solely IV TXA. It included a full spectrum of GI bleeding presentations. And it has a primary outcome specifically related to the underlying problem. One major limitation, though, is that this primary outcome was amended while the trial was underway. So the authors commented that this outcome was changed due to the observation that most of the mortality was not attributed to bleeding. And the timeframe was also shortened from 28 days to five days to better determine if TXA had an effect on cause-specific mortality. An additional limitation is that the majority of bleeding and death occurred in cirrhotic patients with suspected variceal bleeding, for which we know there are other recommended treatment options. And the receipt of these other treatments was not reported. Finally, this study has a low applicability to patients that were taking anticoagulants at home, as they made up a small portion of the patients in this study. So jumping into a few more discussion points. In previous trials, the benefit of TXA was time-dependent, with the most significant benefits seen within three hours of the onset of bleeding. In HALT-IT, only 10.3% of patients in both groups were randomized within that three-hour time period. Also due to the nature of GI bleeds and then reliance on patients presenting to the hospital when their symptoms are known, it is nearly impossible to determine the exact timing of the onset of bleed and achieve administration of TXA in under three hours. HALT-IT also used a higher dose of TXA and a longer treatment duration than seen in the more contemporary trials, such as CRASH and MATTERS. The authors explained that they chose to extend the treatment to 24 hours due to this high risk of re-bleeding in this patient population. Although Barr et al. found a mortality benefit that was driving the meta-analysis, their trial included more than half of patients primarily presenting with bleeding ulcers, whereas 45% of the study population in HALT-IT had suspected variceal bleeds. It's known that cirrhotic patients may have a mixed fibrinolytic phenotype and may be more prone to bleeding due to lack of clotting factor production. So to be specific, the majority of patients who experienced the primary outcome of death due to bleeding within five days did have suspected variceal bleeds. So the difference in underlying disease could potentially explain the absence of reduction of bleeding deaths found between TXA and placebo. And then finally, TXA has generally been perceived as having a minimal side effect profile. However, evidence from HALT-IT suggests that TXA may not be as benign as previously thought and may even cause harm through increased risk of VTE events and seizures. So since the dose in this trial is significantly higher than seen in other large controlled clinical trials, this could potentially explain the increased incidence of these adverse effects. So to summarize, the HALT-IT trial was a well-designed large multicenter randomized trial that compared TXA to placebo for acute GI bleeds. It showed no five-day mortality benefit when given a 24-hour infusion of TXA, but did find an increased risk of VTE events and seizures. So at this time, I would not recommend TXA be administered for patients presenting with acute GI bleed. So this brings me to my final polling question. Based on the results of this study, would you change your practice at your institution regarding the utilization of TXA in GI bleeds? Okay, so most of you answered no that you don't use it, and so most of you answered no here that you would not change your practice, which is kind of in line with my conclusion as well. So that concludes my presentation, and then I will field any questions that you all may have at this time. Thank you, Katherine. So one question is, did patients in each group have similar baseline risk factors? For example, thrombotic and seizure histories. There's been some data that has arisen that TXA itself may not increase VTE risk. However, it's more so the patient's underlying risk factors. That's a very good question and a good point. The study actually doesn't necessarily report the baseline thrombotic and seizure history. They do break down the major comorbidities in major categories. So cardiovascular, respiratory, liver, renal, but they do not break down that specifically, but that would be a very interesting thing to look at to see. Okay, Katherine. And did they mention onset of VTE and seizures in the trial? The time of onset, they did not necessarily mention when they occurred, no. But that would be, again, another good thing to know to look into. Okay, so that concludes our Q&A session. Thank you, Katherine Moore. And thank you to our presenters today and the audience for attending. Please join us on August 21st, which is the third Friday of the month from 2 to 3 p.m. Eastern Standard Time for the next Journal Club Spotlight on Pharmacy. That concludes our presentation today.

Journal Club Spotlight on Pharmacy

ketamine

acute agitation

emergency department

hydromorphone

fentanyl-based sedation

extracorporeal membrane oxygenation

tranexamic acid

gastrointestinal bleeding