Date: October 14th, 2019

Reference: CRASH-3 Trial Collaborators. Effects of tranexamic acid on death, disability, vascular occlusive events and other morbidities in patients with acute traumatic brain injury (CRASH-3): a randomised, placebo-controlled trial. The Lancet October 2019

Guest Skeptic: Dr. Salim Rezaie currently works as a community emergency physician at Greater San Antonio Emergency Physicians (GSEP), where he is the director of clinical education. He is also the creator and founder of REBEL EM and REBEL Cast, a free, critical appraisal blog and podcast that tries to cut down knowledge translation gaps of research to bedside clinical practice.

Case: A 42-year-old man falls off a backyard deck. He arrives at the emergency department via EMS with a Glasgow Coma Scale (GCS) score of 10 and both pupils reactive. He is hemodynamically stable and sent for a STAT head CT. It demonstrates a traumatic intracranial hemorrhage. You wonder if you should give tranexamic acid (TXA) while you wait for neurosurgery to call you back.

Background: TXA is a synthetic derivative of lysine that inhibits fibrinolysis and thus stabilizing clots that are formed. We have covered TXA as a treatment for bleeding a number of times on the SGEM. The evidence for TXA providing a patient-oriented outcome (POO) has been mixed. It seems to work for epistaxis, fails to cause a decrease in all-cause mortality in post-partum hemorrhage, does not demonstrate an improved neurologic outcome in hemorrhagic strokes but does have 1.5% absolute mortality reduction in adult trauma patients.

REBEL EM has also looked at TXA for those conditions plus a few others. It is unclear if it provides a benefit for gastrointestinal bleeds (GIB). Nebulized TXA shows promise for both post-tonsillectomy bleeding and hemoptysis. However, better studies are needed to confirm these observations.

Zehtabchi et al published a SRMA of TXA for traumatic brain injuries (TBI). They found only two high-quality randomized control trials with 510 patients having TBI that met inclusion criteria. The results were no statistical difference in in-hospital mortality or unfavorable neurologic functional status. However, there was a statistical reduction in intracranial hematoma expansion size with TXA compared to placebo.

Clinical Question: Does TXA have mortality benefit in patients with isolated head trauma?

Reference: CRASH-3 Trial Collaborators. Effects of tranexamic acid on death, disability, vascular occlusive events and other morbidities in patients with acute traumatic brain injury (CRASH-3): a randomised, placebo-controlled trial. The Lancet October 2019

  • Population: Adult patients 16 years and older with traumatic brain injuries with GCS score of 12 or lower or any intracranial bleed on CT scan and no extracranial bleeding treated within 3 hours of injury
    • Excluded: Age less than 16 years of age, extracranial bleeding, or greater than 8 hours since injury (limited to greater than 3 hours from September, 2016)
  • Intervention: TXA 1g infused over 10 minutes followed by an infusion of another 1g over 8 hours
  • Comparison: Saline placebo
  • Outcome:
    • Primary Outcome: Head injury-related deaths within 28 days
    • Secondary Outcomes: Early head injury deaths (<24hrs), all-cause and cause specific mortality, disability, vascular occlusive events (myocardial infarctions, stroke, venous thromboembolism), seizures, complications, neurosurgery, days in the intensive care unit (ICU), adverse events within 28 days and subgroup analyses.

Authors’ Conclusions: “Our results show that tranexamic acid is safe in patients with TBI and that treatment within 3 h of injury reduces head injury-related death. Patients should be treated as soon as possible after injury.”

Quality Checklist for Randomized Clinical Trials:

  1. The study population included or focused on those in the emergency department. Yes
  2. The patients were adequately randomized. Yes
  3. The randomization process was concealed. Yes
  4. The patients were analyzed in the groups to which they were randomized. Yes
  5. The study patients were recruited consecutively (i.e. no selection bias). Unsure
  6. The patients in both groups were similar with respect to prognostic factors. Yes
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. Yes
  8. All groups were treated equally except for the intervention. Yes
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes
  10. All patient-important outcomes were considered. Yes
  11. The treatment effect was large enough and precise enough to be clinically significant. Unsure

Key Results: The CRASH-3 investigators randomly allocated 12,737 patients with TBI to receive either TXA or placebo. There were 9,202 (72%) who were enrolled within 3 hours of injury. The mean age was 42 years, 80% were male, 80% had both pupils reactive and about 2/3 had a GCS less than 12.

No statistical difference in head-injury related mortality with TXA compared to placebo

  • Primary Outcome: Death due to head injury
    • 18.5% TXA vs. 19.8% placebo, RR 0.94 (95% CI 0.86 to 1.02)

  • Secondary Outcomes: The two statistically significant results were less head injuries deaths within the first 24 hours and in the subgroup of patients with milder injuries (GCS 9-15). Disability was similar between both groups. There was no evidence of increased vascular events, seizures, complications or adverse events. We could not find the data on neurosurgery or days in the ICU.

This was a large multinational study looking at a very important question. The CRASH-3 Trial Collaborators need to be commended for successfully completing this study.  All studies will have some limitations that need to be considered when interpreting the results. Here are five that we identified and wanted to discuss:

1. Selection Bias: We are unsure if there was any selection bias in CRASH-3. Patients were eligible if the recruiting clinician was uncertain as to the appropriateness of TXA. No denominator was provided for how many people were screened. They state that “almost all patients with TBI who met inclusion criteria were recruited” but do not provide the actual number. Because of the subjective nature of the inclusion and exclusion (based on recruiting clinicians’ uncertainty) this could have introduced selection bias into the study.

2. Wide-Confidence Intervals: The confidence intervals were wide for point estimate of the primary outcome (relative risk head injury related death). It ranged from a large effect size (0.86) but also crossed the line of no difference (1.02). That does not mean we can conclude TXA does not work but rather it did not demonstrate a statistical benefit. The point estimate did favour TXA over placebo (RR 0.94). The width of the 95% confidence interval and the upper limit crossing the 1.0 decreases the certainty of any conclusion that can be made.

3. All-Cause vs. Head Injury Mortality: The authors considered many patient-oriented outcomes. While their primary outcome was head-injury related mortality, a more important POO would be overall mortality. The patient and their family usually do not care what they died from but whether or not they did die. We saw this in the WOMAN trial where the overall mortality was not decreased but the paper highlighted the statistically significant decrease in post-partum hemorrhage related deaths.

4. External Validity: The study was done in 29 countries with the majority being middle to low income countries. Canada had one centre and the USA had no participating centres. They did do an analysis based on income that was not pre-specified. This did not show a statistical difference. However, the question remains on whether TXA would have a clinically important impact in the USA healthcare setting.

5. Subgroup Analysis: While the subgroup analyses were pre-specified, they are underpowered to draw any strong conclusions. We should be cautious not to over-interpret these results.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We agree that TXA does appear to be safe but think the claim about decreased head-injury related death is misleading. This is because it was only for a subgroup of patients. If TXA is given, then giving it as soon as possible seems reasonable.

SGEM Bottom Line: We cannot recommend the routine use of TXA for patients with isolated traumatic brain injuries at this time.

Case Resolution: You decide not to give TXA and decide to leave it up to the neurosurgeon.

Clinical Application: Clinicians will have different thresholds of evidence that will convince them to change practice. The effect size of TXA in patients with TBI was small, the confidence intervals were wide and the upper end of the 95% confidence interval crossed the line of no statistical difference. The evidence was even less convincing when looking at the more patient-oriented outcome of all-cause mortality. While the subgroup analyses showing efficacy in certain cohorts was interesting, the groups were underpowered to provide a clear answer on how to apply this information.

We have a number of concerns how this trial will be interpreted and applied. One aspect is the “spin” on the article. People may highlight the “significant” relative reduction of death in a subgroup rather than focusing on the lack of statistical difference of an absolute reduction in overall mortality. This was seen in the WOMAN trial and we need to be cautious not to do this with CRASH-3 by over hyping the results.

Dr. Salim Rezaie

It is also well recognized that the efficacy of treatment decreases when applied outside the strict environment of a research trial. Applying this protocol in the community setting would likely dilute and negate any possible benefit of TXA for patients with isolated TBI.

We are also concerned that a quality metric will be created to monitor adherence to providing TXA to these adult isolated TBI patients within 3 hours. This could be tied to physician or hospital pay despite the weak underlying evidence. We have seen this happen with tPA for acute ischemic stroke and the 30cc/kg of IV crystalloid within the first hour in septic patients.

Let us not forget about cognitive load and attention. Adding an additional treatment which may or may not provide a patient-oriented benefit is just one more thing to consider when managing a trauma patient. While we are paying more attention to that individual, we will be decreasing our attention to the many other undifferentiated patients in the department. Those other patients might have time dependent emergencies whose treatment could be impacted negatively.

The low cost of TXA may also be used as an argument for using it in these patients. However, this is irrelevant if it does not provide a patient-oriented outcome. There are also millions of people who suffer from TBI every year worldwide. A small number multiplied by millions of patients will add up to a lot of money.

What Do I Tell My Patient?  You tell the family that their loved one has a serious brain injury. You have called the neurosurgeons and they are coming shortly to talk about the best treatments.

Keener Kontest: Last weeks’ winner was E.S. Rosenberg an EMT from Israel. This is a back-to-back win for Eliyahu. He knew Asciano Sobrero is the person that invented nitroglycerin. We caught a few people on this question thinking it was Alfred Nobel; He was responsible for commercialized nitroglycerin.

Listen to the SGEM podcast to hear this weeks’ question. If you think you know the answer, send an email to with “keener” in the subject line. The first correct answer will receive a cool skeptical prize.

Other FOAMed:

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.