Hi, my name is Bridget Crum. I am a clinical pharmacist in the Medical Intensive Care Unit at Duke University Hospital. I'm also the residency program director for RPGY2 Critical Care Pharmacy Residency and an assistant professor of clinical education for UNC Eshelman School of Pharmacy. Thank you for joining me today as we discuss Glamming Up a GI Bleed. I have no relevant financial disclosures to report. The objectives today are simple. GI bleeding is critical care bread and butter. And today's our opportunity to take a step back and review why we use the therapeutics we always have and to assess if they still make sense. We will do so by reviewing the landmark literature and the management of both ulcer and variceal upper GI bleeding. Briefly, we'll review proton pump inhibitor or PPI dosing in ulcer bleeds and then move on to the medication utilization and variceal bleeds. Lastly, we'll assess new literature to incorporate into practice. We'll start first with upper GI bleeding caused by ulcerations in the upper GI tract. PPI used in ulcer upper GI bleed is well established. In this disease state, PPIs promote clot formation and stability following endoscopic hemostasis. Increasing gastric pH theoretically prevents pepsin-induced clot lysis, but whether that occurs at a pH of greater than 7 or a pH of greater than 4 is currently unclear. So our target or optimal gastric pH is unknown. In 2021, the American College of Gastroenterology or ACG released a set of guidelines that aggregates the currently available data evaluating PPIs for ulcer upper GI bleed quite nicely. I presented that data for you in this table with columns moving left to right being the number of randomized controlled trials, the comparators and dosing strategies used, rates of re-bleed, mortality, and need for surgical intervention. The first row details the comparison of high-dose PPIs to placebo. High-dose PPI therapy is defined as 80 milligrams per day or more. High-dose PPIs are associated with a lower risk of re-bleed, mortality, and surgery. There has no evidence of treatment effect with continuous infusion versus intermittent dosing of PPIs, but the most rigorously designed trials have compared PPI continuous infusions to placebo. It's important to note that the majority of these trials are done in Asia, and we'll talk about why that's important in a little bit. Compared to histamine 2 receptor antagonists or H2RAs, PPIs lower the risk of re-bleeding. And lastly, in the last row, 12 randomized controlled trials have compared continuous infusion to intermittent dosing of PPIs and have shown no significant differences in rates of re-bleeding, mortality, and surgery related to dose, frequency, or route of PPIs. It's interesting to me to note that four of those compared continuous infusion to simply daily dosing of PPI therapy. In the acute setting, most patients receive either a continuous infusion or twice daily PPI when an upper GI bleed is due to a suspected ulcer. The upper GI team at my institution typically recommends a standard duration of PPI therapy, but I've often wondered the appropriateness of that duration. This question was addressed by Chang and colleagues who evaluated every 12-hour dosing versus daily PPI therapy from days 4 to 14 following endoscopy. They enrolled 293 patients with an ulcer upper GI bleed and a Rockwell score greater than or equal to 6. A Rockwell score is a score from 0 to 11 that includes age, comorbidities, shock, bleeding source, stigmata of bleeding, like a clot or a spurting vessel. Patients who received esomeprazole at a frequency of every 12 hours had lower rates of re-bleeding compared to daily esomeprazole. Based on this, it seems reasonable to continue every 12-hour dosing of a PPI for 14 days at least. Short-term PPI use is fairly low risk and inexpensive, but this does assume that patients have good outpatient follow-up and de-escalation of a dosing strategy at follow-up visits. There are a couple of additional considerations that one should keep in mind when evaluating both historical and modern trials of PPIs and ulcer upper GI bleed. The first is ethnicity of the patient population studied. Higher total doses may be desirable in Western populations because PPIs are reported to have lesser pharmacodynamic and clinical effect in Western compared to Asian populations. There's a couple of different reasons for this. The first, Asians have lower parietal cell mass. The second, H. pylori infection is generally higher in Asian populations compared to European and North American populations. PPIs produce greater suppression of gastric acid secretion in the presence of H. pylori infection, and this is generally underreported in most trials. And then lastly, Asians are more likely to be slow metabolizers of PPIs due to a genetic polymorphism in the CYP P450-2C19 enzyme. PPIs are metabolized by 2C19, so Asians have a greater pharmacodynamic response with more effective and prolonged inhibition of acid secretion at lower doses. Leontitis and colleagues evaluated PPI data by population studied. In this forest plot evaluating rates of rebleeding in European and North American populations compared to Asian populations, you can see that the Asian trials consistently favor PPI use and tend to drive aggregate outcomes in rates of rebleeding. When it comes to mortality, there does appear to be some discordance in outcome in European and North American trials compared to Asian trials. The last consideration I want to touch upon is route of administration. In practice, I frequently hear that we must use IV PPI therapy for 72 hours, even if we use an intermittent dosing strategy. I think this is largely extrapolated from the infusion trials, but it's important to keep in mind that the extent of pH rise is similar with enteral administration compared to IV. The onset of pH rise may be a few hours quicker with IV therapy, but that likely has minimal impact on effect or outcome. So I think it's reasonable to use IV for the first or second doses of an intermittent therapy, but enteral is probably OK once hemostasis has been achieved. The 2021 ACG guidelines recommend high dose PPI therapy for three days following upper GI bleed due to ulcers following endoscopy with hemostasis. High dose is defined as 80 milligrams a day or more of IV or enteral therapy. The strongest evidence that we have is for a PPI infusion compared to placebo. Because of the ethics of doing intermittent PPI versus placebo trials, I think that the PPI infusion will continue to be a part of our guidelines. For convenience, I prefer intermittent dosing over a continuous infusion. Please note the enteral administration, either PO or via tube, is an option even on day two following endoscopic hemostasis. High risk patients or those with a raw call score greater than or equal to six should stay on twice daily PPI for 14 days after endoscopy, and this is a new recommendation based on the Cheng trial that we discussed earlier. Now we'll transition into the therapeutics of variceal upper GI bleed. Antibiotics in variceal upper GI bleeds are used to prevent bacterial translocation. This table summarizes the classical literature in patients with cirrhosis and upper GI bleeding due to varices. Ramola and colleagues were the first to describe the enteral administration of antibiotics, both reduced rates of infection and reduced mortality. Additionally, repeated trials with primarily fluoroquinolone prophylaxis following hemostasis reduces infection. While the studies were likely too small to show a statistically significant reduction in mortality, there's certainly a noticeable theme here. In a nice meta-analysis, Bernard and colleagues aggregated this data, demonstrating that the odds of being free of infection were higher with antibiotic prophylaxis, as was survival. Seftriaxone resistance has increased over time, so seftriaxone was evaluated in a multicenter randomized controlled trial of patients with cirrhosis and primarily variceal upper GI bleeding. Patients were randomized to receive norfloxacin, which is notably no longer available in the United States, or seftriaxone for seven days. Possible or proven bacterial infection was higher in the norfloxacin group, which was primarily driven by possible infection. Proven was not different between groups. Most of this difference was explained by a high rate of infections by quinolone-resistant organisms. In their Cox regression analysis, transfusion requirements, failure to control bleeding, and norfloxacin prophylaxis were associated with higher risk of possible or proven infection. Mortality was similar between groups. The 2017 AASLD guidelines recommend short-term prophylaxis with seftriaxone 1 gram IV daily for patients with cirrhosis and GI hemorrhage. The maximum recommended duration is seven days. Interestingly, the guidelines indicate that antibiotics can be discontinued when hemorrhage has resolved and the patient is hemodynamically stable. In my practice, this is a pretty big opportunity, where RGI providers consistently recommend seven days even in stable patients. In the spirit of antimicrobial stewardship, I recommend discontinuing earlier as the patient is stabilized. Transitioning into the second of our three pharmacologic treatment modalities in variceal bleed is the role of octreotide. Octreotide is a somatostatin analog and splanchnic vasoconstrictor. It inhibits vasodilator release, this is hormones like glucagon, for example, which leads to reductions in portal pressure. Octreotide's impact on portal pressure, though, even with a continuous infusion, is incredibly short. It's only about five minutes, and that could be for a couple of different reasons. It could be tachyphylaxis or maybe even rapid desensitization. In a study by Escorcel and colleagues, mean arterial pressure, hepatic venous pressure gradient, and azygous blood flow were evaluated at different octreotide dosing strategies. In this first graphic, patients received either placebo, which is the light dotted line, a 50-mic bolus, which is the heavier dots, or a 500-mic bolus, which is the line of octreotide. You can see that the effect of bolus administration is pretty transient. In the second evaluation, patients received a 50-mic bolus followed by either placebo infusion, which is the light dots, a 50-mic per hour infusion, which is the heavier dots, or a 250-mic per hour infusion, which is the line. And again, the effect of on hepatic venous pressure gradient and azygous blood flow were transient with returns to the level of placebo within an hour. In this last graphic, you can see the effect of bolus dosing compared to a continuous infusion. Again, the boluses have marked but transient reduction in portal venous flow. Infusions have overall pretty minimal effect on blood flow. It's likely that octreotide's mechanism in the setting of variceal bleed is largely due to reductions in intestinal blood flow. Variceal bleeding tends to increase intestinal blood flow. By inhibiting glucagon release, octreotide blunts this increase in blood flow, which may influence risk of re-bleeding. Octreotide has been studied in multiple randomized controlled trials following sclerotherapy or bending in variceal upper GI bleed. These trials noticeably use different dosing strategies and durations, and they all show variable effects on rates of re-bleeding, need for transfusion, and mortality. In aggregate, octreotide plus local intervention likely helps with hemostasis and resource utilization, but it probably does not impact mortality overall. In general, octreotide is considered fairly low risk, is pretty inexpensive, and it has a potential benefit, though one that hasn't been consistently demonstrated across all trials. Most likely, its benefit is in reduction of rates of re-bleeding. In combination with endoscopy, octreotide is reasonable at a dose of 50 mics times one, and then 50 mics an hour for two to five days. Since most re-bleeding occurs within the first 48 hours of endoscopy, I typically recommend 48 to 72 hours as long as the patient remains stable. It's always important to monitor for adverse drug reactions, most commonly bradycardia and QTC prolongation, which are rare but can occur. And lastly, we can't forget about PPIs in variceal upper GI bleed. The 2007 ACG guidelines make no recommendation for PPI use and discuss one study favoring use, which we'll touch upon in a minute. The 2017 AASLD guidelines make no mention of a role for gastric acid suppression in the setting of variceal upper GI bleed, but it's become tradition. Over 95% of patients receive a PPI on presentation, and 45% are continued on long-term therapy after variceal upper GI bleed. If you are as interested in transitions of care as I am, this number must make you gasp just a little bit. The literature supporting PPI use for secondary prophylaxis after variceal ligation is less than robust. Shaheen and colleagues in a single-center, double-blind, randomized controlled trial of 42 patients demonstrated that giving a PPI daily reduced the ulcer size by 50% but didn't reduce the number of ulcerations. It's a great hypothesis-generating study. Follow-up and larger studies have shown no clinical benefit of PPIs in transfusion requirements for bleeding or outcomes. Long-term prophylaxis, meaning a PPI given for eight weeks with a 13-month follow-up, has also demonstrated no effect on the incidence of a new bleeding event. As we are all aware, long-term PPI use in general is associated with rare but notable adverse drug effects, but the risk of PPIs in patients with cirrhosis are potentiated. Rates of C. diff in cirrhosis may be higher than in non-cirrhotics, and this is probably due to SVP prophylaxis, PPI use, and overall inpatient antibiotic exposure, but note the confidence interval for this is pretty wide. The odds of SVP are higher for patients with cirrhosis on PPIs compared to those who are not, with the risk increasing with dose and duration of use. The degree of hepatic encephalopathy is also greater for patients on PPIs. Medical admissions may be longer, and there may be even a signal of higher odds of both short- and long-term mortality in patients with cirrhosis who are receiving PPI therapy. My bottom line here is that PPI use in cirrhosis is common, and it should probably be avoided. It likely gets initiated in the setting of variceal bleed, so as a critical care provider, short-term use does appear reasonable until the bleeding source is identified and endoscopically managed. If a PPI is recommended at all by our GI consult service, I recommend that it's prescribed daily and only for a maximum of 10 days, and really discourage long-term use. I see this as a significant opportunity for improving transitions of care. Our final objective for today is to discuss new literature in upper GI bleeding, so I'd like to touch upon the use of tranexamic acid, or TXA. TXA is an antifibrinolytic that reversibly binds to plasminogen to prevent activation of plasmin. It also blocks plasmin binding to fibrinogen to stabilize clot formation. TXA has been studied in several large trials, including CRASH-2, WIMN, and CRASH-3. In CRASH-2, a large double-blind, randomized controlled trial, TXA reduced 28-day all-cause mortality and death due to bleeding. In postpartum hemorrhage, the WIMN trial demonstrated that TXA reduced death due to bleeding. And in TBI, TXA does not appear to have a specific role. There is one published Cochrane review evaluating TXA in upper GI bleeding. I pulled out the TXA versus placebo data specifically, so there does appear to be a signal for improving outcomes. Notably, in all studied patient populations, TXA does not appear to affect transfusion requirements. In that spirit, the HALT-IT trial was developed and published in Lancet in 2020. HALT-IT specifically evaluates the role of TXA in upper or lower GI bleeding using the same uncertainty principle employed in the CRASH-2 and 3 and WIMN trials. TXA was administered at a dose of 1 gram over 10 minutes, followed by 125 milligrams an hour for 24 hours. Included patients primarily had an upper GI bleed with just under half with suspected varices. Fewer than 10% were receiving anticoagulants. The primary endpoint was death due to bleeding, which was not different between groups. TXA did not affect rates of death due to bleeding in any of the evaluated subgroups, which you can see in the forest plot on the right. There were no differences in any other secondary endpoint, including death due to bleeding at 24 hours and 28 days, re-bleeding, need for surgical or radiologic intervention, transfusion requirements, length of stay, or functional status. Patients who received TXA had a higher risk of seizure, albeit the incidence was rare, and also DVT. The authors concluded that TXA should not be used outside of a clinical trial setting for GI bleed. HALT-IT was a methodologically sound evaluation of TXA in this patient population, using a robust dose and meeting all standards of care for all bleed types, so this is a great study. In summary of all that we have just discussed, for upper GI bleeding due to ulcers, PPIs and reduced re-bleeding after endoscopic hemostasis, I recommend dosing twice daily for convenience, and this can be given IV or enterally for two weeks in patients with a raw call score greater than or equal to six. Alternatively, a continuous infusion can be used for up to 72 hours, followed by twice daily dosing for up to 14 days. In variceal bleeds, antibiotics reduce rates of infection and probably save lives. I recommend Ceftriaxone 1 gram IV daily for a maximum of seven days, but don't forget that you can discontinue earlier if hemorrhage is resolved and the patient is hemodynamically stable. Optreotide likely decreases re-bleeding in the setting of a variceal bleed. Dosed at 50 mics times one, and then 50 mics an hour for two to five days. PPIs may decrease lesion size in the short term after a variceal bleed, and I say this with some reservation based on the limited data that's out there, but have no role for long-term use in variceal upper GI bleeding. If prescribed, the frequency should be daily for a maximum of 10 days. And then lastly, tyrenexamic acid currently does not have a role in the setting of GI bleed. Thank you for listening in today. I hope I've been successful in my attempt to make upper GI bleed exciting again for the critical care clinician, and that you can take away one or two pearls to incorporate into your clinical practice and challenge the status quo. Please do not hesitate to reach out to me with any questions.

upper gastrointestinal bleeding

proton pump inhibitors

gastric pH target

variceal bleeding

octreotide