The application of evidence-based medicine is at its easiest after large trials and rigorous analyses have been popularized and widely disseminated. The evidence-based clinician must juggle online resources, academic manuscripts, and trade journals, but ultimately can remain current through the cornucopia of open-access educational resources available to the modern emergency medicine physician.
A true test of bedside Bayesianism, however, comes when we are presented with clinical conundrums not yet thoroughly vetted and extensively analyzed by contemporary emergency medicine educators. It is when no clear answer exists that we are forced to faithfully apply the best available knowledge to answer critical questions in real time. As methodical skepticism in emergency medicine continues to permeate and propel our evidence-based culture, it is crucial that we not abdicate our underlying responsibility to do what we think is best for the patient in front of us.
Often, it seems, we treat the absence of compelling evidence as an indication of an intervention's futility or ineffectiveness, which may be a disservice to the patient in need. In reality, a synthesis of the available literature with our clinical experience, individual patient factors, and the medical degrees and extensive body of knowledge we all have fought so hard to obtain is more than sufficient where top-level evidence may be lacking.
Consider the patient dying of a gastrointestinal bleed. It's no rarity—tens of thousands of these patients pass through our doors annually. Emergency medicine residents cut their procedural teeth in resuscitation bays full of GI bleeders and almost every EP has a story about that time he put a Blakemore in. We often lament how little we have to offer these patients while awaiting reticent gastroenterologists.
Octreotide and ceftriaxone are quickly delivered. Proton pump inhibitors are lambasted for their lack of efficacy yet dutifully ordered despite compelling evidence of no benefit and even a signal of harm. We evidence-based clinicians know the literature, and we cling to transfusion trial talismans as we seek to stem the tide of dark blood that seems to flow as quickly as we can replace it. (N Engl J Med 2013;368:1362; http://bit.ly/2NaBTA8; JAMA 2015;313:471; http://bit.ly/2CRG535.)
Curiously, one medication has remained inexplicably absent from our routine resuscitation of the critically ill GI bleeder. Tranexamic acid, or TXA, recently championed in trauma and publicized for every bleeding source from epistaxis (EMN 2018;40:1; http://bit.ly/2Qq5Iux) to abnormal uterine bleeding (EMN 2018;40:5; http://bit.ly/2N7MAmW), often sits unused in the pharmacy as the blood bank is emptied and our patients deteriorate. Despite strong physiologic justification for its use and a level of clinical evidence that should sate even the strictest of EBM evangelists, TXA has yet to find permanent purchase in the GI bleed armamentarium.
We have suspected since 1979 that TXA might benefit critically ill patients with GI bleeds when a small trial of 149 patients demonstrated decreased transfusion requirements and hinted at a mortality benefit. (Scand J Gastroenterol 1979;14:839.) Since then, multiple randomized controlled trials have consistently demonstrated not only improvement in hemodynamic markers and transfusion requirements but also sustained evidence of mortality benefit.
A meta-analysis of seven randomized controlled trials of 1,385 patients in 2008 showed a relative mortality improvement of 39 percent. (Aliment Pharmacol Ther 2008;27:752; http://bit.ly/2xgvIAL.) This was confirmed with a 2014 Cochrane Review, where eight RCTs and more than 1,700 patients found a relative risk for mortality of 0.60, with no difference in venous thromboembolic events. (Cochrane Database Syst Rev 2014 Nov 21; http://bit.ly/2NbRxvc.)
Tranexamic acid reduces clot breakdown by inhibiting the action of plasmin, an active agent in fibrinolysis. The CRASH-2 and MATTERS trials convinced the trauma and emergency medicine communities of TXA's value in reducing death due to bleeding and all-cause mortality in major trauma, and multiple systematic reviews and meta-analyses since then have consistently demonstrated statistically significant reductions in death when TXA is applied in patients with GI bleeds. As is the case in modern evidentiary science, though, medical societies and guideline organizations have (appropriately) called for additional data before making a recommendation to routinely include TXA in the resuscitation of the GI bleeder.
A 2014 trial sequential analysis—a statistical technique that combines information already available from meta-analyses with weighted interim analyses and helps to quantify how much more information is needed before a treatment effect can be confidently confirmed—indicated that 5,500 more patients would need to be included in clinical trials (more than double the number already analyzed) before the mortality benefit of TXA in GI bleed could be confirmed. (BMJ 2014;348:g1421.)
Fortunately, an ongoing RCT, the HALT-IT trial, has recruited nearly 10,000 patients on the way to their goal of 12,000, and should provide definitive confirmation of the benefit of TXA as a life-saving medication. The frustrating part, however, is that the trial isn't set to conclude until next year, and the timeline of publication and knowledge dissemination likely stretches far beyond that.
The question, then, is what we should do in the interim.
Tranexamic acid has shown itself to be safe and effective in a constellation of emergency conditions, and not one but eight randomized trials have reliably established that it saves lives in patients with gastrointestinal bleeding. Overblown fears of precipitating increased rates of venous thromboembolic events have not panned out in large studies across a variety of applications, and the biochemical basis of the drug and its expected action correspond with demonstrated success in clinical trials.
The much-anticipated HALT-IT trial will settle the debate, but in the meantime we are left with consistent and compelling evidence that tranexamic acid should be routinely incorporated into the resuscitation of the critically ill GI bleeder. Our patients deserve nuance and rational integration of the literature where definitive studies are lacking. Our patients deserve TXA.