Background: It almost seems that when it comes to the use of the antifibrinolytic agent tranexamic acid (TXA) in trauma, one argument has just been completed and another one comes up right behind it.
Let’s take a step back. Most agree that the evidence clearly supports the role of the early in-hospital administration of TXA in major trauma (in conjunction with balanced blood product transfusion practices). Given the benefit of in-hospital use, and the evidence supporting most benefit with earlier use, it seemed to make intuitive sense to bring this out into the prehospital setting closer to the point-of-injury – many agencies have done just that. However, this particular area of use did not have any associated high-quality evidence.
Led by a team out of the University of Pittsburgh (the same group that brought us the PAMPer trial in 2018), Guyette et al just released the Study of Tranexamic Acid During Air Medical and Ground Prehospital Transport (STAAMP) trial, published in JAMA Surgery.
This Post is Co-Published with Our Friends Over at St. Emlyn’s Blog
Paper: Guyette FX et al. Tranexamic acid during prehospital transport in patients at risk for hemorrhage after injury: A double-blind, placebo-controlled, randomized clinical trial. JAMA Surg 2020. PMID: 33016996
Clinical Question: Does the prehospital use of TXA to patients at risk of hemorrhage from trauma result in a lower 30-day mortality compared to placebo?
What They Did
- Pragmatic, phase 3, prospective, multicenter, double-blind, placebo-controlled randomized clinical trial
- Patients transported to 4 participating level one trauma centers were screened for inclusion based on broad inclusion criteria over a 4-year period
- Computer randomization occurred using a 1:1:1:1 ratio random allocation sequence with a block size of 12
- Each treatment kit contained a prehospital and corresponding in-hospital phase assignment
- Each treatment kit comprised a 10mL solution of either TXA or sterile water, which was diluted into a 100mL of 0.9% normal saline and infused over a specified time period
- Prehospital and in-hospital teams were blinded to the content of the drug kit
- Prehospital treatment arms received either a 1g bolus of TXA (given over 10 minutes) or placebo
- Patients randomized to the prehospital placebo group were allocated to receive further placebo in hospital
- Patients randomized to the prehospital TXA group were randomly assigned on hospital arrival to receive either
- No further TXA (abbreviated group) – 1g TXA total
- 1g TXA infused over 8 hours (standard group) – 2g TXA total
- Further bolus of 1g TXA followed by another 1g infused over 8 hours (repeat bolus group) – 3g TXA total
- In-hospital placebo arm groups received further placebo
- All other pre- and in-hospital care occurred according to usual protocols
- Note at the time of the study, the use of prehospital TXA was not a standard protocol
- Study outcomes were primarily centered around 30-day mortality (see below)
- In addition, prespecified subgroup analyses for 30-day mortality was conducted for the following:
- Patients not requiring blood transfusion
- Those with significant traumatic brain injury (head abbreviated injury scale score >2)
- Patients enrolled at scene versus a referring hospital
- Those requiring early (within 24 hours) operative intervention
- History of vitamin K antagonist medication use
- History of antiplatelet medication use
- Those requiring massive transfusion (≥10 units of blood within first 24 hours)
- Note that the trial was approved under an exception from informed consent (EFIC)
- Community consultation occurred prior to enrollment
- Study subjects had the option to verbally opt out or wear an opt-out bracelet
Inclusion
- Injured patients transferred from scene or from a referring hospital to the trauma center
- within 2 hours of injury and at least one of the following:
- one episode of hypotension (systolic blood pressure ≤90mmHg
- one episode of tachycardia (heart rate ≥110 beats per minute)
- age ≥18 years and ≤90 years
- within 2 hours of injury and at least one of the following:
Exclusion
- Age <18 years or >90 years
- Lack of intravenous or intraosseous access
- Isolated fall from standing
- Documented cervical cord injury
- Prisoner
- Pregnancy
- Traumatic arrest >5 minutes
- Penetrating brain injury
- Isolated drowning
- Isolated hanging
- Objecting to study either
- Voiced at scene
- Wearing a STAAMP study opt-out bracelet
Outcomes
- Primary outcome was 30-day mortality
- Secondary outcomes included:
- 24-hour and in-hospital mortality
- Blood component resuscitation volumes at 6- and 24-hours from admission
- Incidence of multiorgan failure, ARDS, nosocomial infection, early (within 24 hours) seizures, venous thromboembolism, coagulopathy and hyperfibrinolysis (as measured by INR and TEG)
- Volume of crystalloid resuscitation in first 24 hours
Results
- A power analysis was conducted – an enrollment of 994 subjects using a 16% mortality estimate to provide 90% power to detect a 7% difference in 30-day mortality between treatment arms
- Carried out from May 1, 2015 to October 31, 2019
- 6559 patients were screened
- 927 patients eligible for enrollment
- 24 patients found ineligible or withdrew consent
- Total 903 randomized – 447 in TXA arm, 456 in placebo arm
- Patient characteristics
- Mean age 42 years
- Majority male (74.8% placebo; 73.2% TXA)
- Majority blunt mechanism injury (85.3% placebo; 83% TXA)
- Majority transported directly from scene (86.6% placebo; 85.2% TXA)
- Median transport time – 39 minutes in both groups (IQR 30-50 minutes)
- Median injury severity score – 12 (IQR 5-22)
- Tachycardia present in 71%
- Initial prehospital hypotension present in 22%
- Blood transfusion required within 24 hours of enrollment – 34%
- Operative procedure required within 24 hours of enrollment – 45%
- Patient interventions
- 98% of patients in both groups received their assigned prehospital infusion
- 92-95% of patients received their assigned in-hospital bolus and/or infusions
- Primary outcome (30-day mortality)
- Available in 99% of patients – 9 patients missing data
- TXA arm – 8%
- Placebo arm – 10%
- No statistically significant difference (8.1% vs 9.9%, absolute difference -1.8; 95%CI -5.6% to 1.9%; p=0.17)
- Secondary outcomes
- 24-hour mortality and in-hospital mortality
- No group differences
- 24 hours: (difference 0.15; 95%CI −2.3 to 2.6; adjusted p= 0.98)
- In-hospital: (difference 1.1; 95%CI −2.7 to 4.9; p= 0.94)
- 6- and 24-hour blood transfusion requirement
- No group differences
- 6-hour total blood component transfusion 0; IQR 0-2; adjusted p= 0.97 for both groups)
- Venous thromboembolism incidence
- No group differences
- PE incidence: −1.3; 95%CI −3.3 to 0.5; adjusted p=0.78
- DVT incidence: −1.2; 95%CI −3.3 to 0.5; adjusted p=0.83
- Seizure incidence
- No group differences
- 4; 95%CI −1.0 to 1.9; p=0.94
- Multiorgan failure
- No group differences
- 2; 95% CI −2.4 to 4.7; adjusted p =0.94
- Subgroup analyses
- Mortality based on dosing regimen
- Placebo – 10.0%
- TXA abbreviated administration group – 9.3%
- TXA standard administration group – 7.8%
- TXA repeat bolus administration group – 7.3%
- After adjusting for site, repeat bolus administration group had significantly lower 30-day mortality (7.3% vs 10.0%; difference −2.7%; 95%CI,−5.0% to −0.4%; p=0.04)
- Mortality based on time to treatment and presence of shock
- Post-hoc analysis
- Lower 30-day mortality when administered within 1 hour of injury (4.6% vs 7.6%; difference −3.0%; 95%CI −5.7% to −0.3%; p<0.002)
- Lower 30-day mortality when administered to those with severe shock (SBP≤70mmHg) (18.5% vs 35.5%; difference −17%; 95%CI −25.8% to −8.1%; p<0.003)
- Mortality based on dosing regimen
- 24-hour mortality and in-hospital mortality
Strengths
- This trial emphasizes the possibility of carrying out randomized clinical trials in the prehospital setting
- Inclusion and exclusion criteria were simple, allowing ease of enrollment
- Prehospital and in-hospital teams were appropriately blinded to the treatment administered
- An intention-to-treat analysis was used
- A broad mix of injury patterns were included
- 4 different trauma centers were utilized
- Groups were well-balanced in terms of demographics, prehospital characteristics, and injury severity
- Multiple secondary parameters were addressed including the amount of blood product administered and the incidence of VTE
- Subgroups for analysis were pre-defined
Limitations
- The trial stopped enrolling early – only 93% of planned enrollment was achieved due to slower than expected enrollment and financial limitations
- The trial was ultimately underpowered to determine the primary outcome – it assumed an enrollment of 994 subjects using a 16% mortality estimate to provide 90% power to detect a 7% difference in 30-day mortality between treatment arms. 903 patients were ultimately enrolled
- Prior trials looking at TXA have generally not found mortality differences greater than 3.5%, and so it is unclear why a 7% difference was chosen for the power calculation
- By extension, the subgroups are also underpowered. Hence the benefits in some subgroups are more associations than true statistically significant differences
- The enrolled patients had an overall low injury severity – the median ISS was only 12, only 22% had an initial prehospital systolic blood pressure of <90mmHg, only 34% required blood in the first 24 hours, and overall mortality was <10%
- Using a single HR≥110 as an inclusion criterion could be problematic as this might not be solely ascribed to hemorrhage
- Although a multicenter trial, the trial could not account for disparities in care protocols in the different prehospital and in-hospital settings included
- Given the disparities of capabilities amongst prehospital systems in the US, the generalizability of the results cannot be guaranteed
- Although broad injury mechanisms were included, the effect on specific injury mechanisms were not accounted for
- Some clinical data was missing, which again can affect results
Discussion
- This study is a valiant attempt to conduct a high-quality prehospital trial to provide further evidence to support an established in-hospital treatment modality
- The use of relatively broad and simple inclusion criteria allowed for ease of enrollment by those at the point of care
- This is especially important given the multiple competing priorities in the prehospital acute resuscitation care phase
- Unfortunately, despite this, recruitment overall did not reach full enrollment
- Inclusion of a broad range of injury mechanisms can be both beneficial and problematic
- It is important to account for different mechanisms which can result in variable patient response
- The majority of enrolled patients (83-85%) included had blunt injury. It is difficult to see whether blunt versus penetrating mechanism was the variable that allowed most benefit from TXA, or whether the mere fact the patient suffered a traumatic injury is the only qualifying factor needed.
- Although the trial was underpowered (and therefore not able to show an overall mortality benefit in either group), there are some key points to take away:
- The trial emphasizes the safety of administration of prehospital TXA
- The point estimate for the difference between both groups was similar to what CRASH-2 reported, hence further supporting the results from the latter trial
- Similar to CRASH-2, those administered TXA early seemed to achieve the most benefit
- As such, moving TXA administration to the point-of-injury appears to make more sense, especially in trauma systems with longer prehospital times
- Those patients in most severe shock appeared to benefit the most – 18.5% vs 35.5%
- This may allow tailoring of prehospital protocols to target those patients most severely injured
- It appears that the higher dosing regimen of at least 2g (and possibly 3g) may confer additional benefit
- Should this be administered differently? Can we provide a 2g (or higher?) bolus up-front (as is being discussed and adopted in military practice) and omit the 8-hour infusion
- This trial additionally addresses one of the perceived concerns for the use of TXA, the downrange complication of VTE. This has been a particular sticking point in some communities. The STAAMP trial showed a similar incidence in both treatment arms through prospective randomized data collection, reinforcing the argument that TXA can be safely administered to appropriately selected patients in the appropriate time window without significantly higher risk of VTE.
Author Conclusion: “Prehospital administrations of TXA compared with placebo did not result in a lower rate of 30-day mortality in this population. Prespecified dose response analyses demonstrate that receipt of a repeat bolus regimen (3g of TXA) results in significantly lower 30-day mortality compared with placebo. In patients who receive prehospital TXA within 1 hour of injury and in those with evidence of prehospital severe shock, post hoc subgroup analysis suggests that prehospital TXA is associated with lower 30-day mortality. The administration of prehospital TXA during air or ground transport is safe and can be provided to patients at risk for hemorrhage.”
Clinical Take Home Point: Although this trial was stopped early and ultimately underpowered to show a clear benefit in its overall primary objective, it did reinforce several aspects – prehospital administration of TXA is safe, it can benefit the most seriously injured group of bleeding trauma patients, and the incidence of complications (in particular rate of VTE – a big sticking point especially in the US) was no higher than the placebo group when appropriately administered.
Bringing beneficial treatments to point of injury is crucial to further reduce the burden of potentially preventable death from traumatic hemorrhage. EMS agencies should include TXA in their protocols, and can possibly further tailor these to emphasize early administration and use in those in severe hemorrhagic shock.
Primary Author’s Comments – Francis Guyette MD MS MPH FACEP FAEMS – University of Pittsburgh Department of Emergency Medicine
The trial was designed to be pragmatic in its implementation to improve its generalizability but that should not discount the incredible challenge it is to perform a multicenter, prehospital EFIC study. Among the difficulties we incurred were seasonal and secular trends in referral patterns to the trauma centers and difficulty of maintaining the research infrastructure necessary to continue multiple dosing regimens, obtain granular patient information, blood samples and thromboelastograms across several geographically disparate sites. We felt the challenge and expense would be outweighed by the opportunity to examine mechanistic differences and determine the optimal administration of TXA
With respect to the study design, we kept the inclusion criteria simple and as close to the original CRASH criteria as possible. We honed the population by excluding patients who were less likely to have life-threatening hemorrhage (ie ground level falls), more likely to be harmed (administration fo TXA >2 hrs from injury, and developed a sample size calculation based on a treatment effect of 6.5% consistent with what was observed in the MATTERS study rather than the 1.5% in the CRASH2 trial. We believed our population would be more severely injured than the CRASH population based on our enhanced exclusion criteria. We also felt the robustness of our trauma systems would more closely resemble those in the MATTERS trial rather than the variability seen in CRASH2. While we included three different dosing regimens in the study, these were predefined secondary analyses. Powering the study to discern a mortality outcome based on the dosing strata would have been prohibitively expensive.
References
- Drew B et al. The use of tranexamic acid in Tactical Combat Casualty Care: TCCC proposed change 20-02. J Spec Oper Med 2020. PMID: 32969002
- Guyette FX et al. Tranexamic acid during prehospital transport in patients at risk for hemorrhage after injury: A double-blind, placebo-controlled, randomized clinical trial. JAMA Surg 2020. PMID: 33016996
- Morrison JJ et al. Association of cryoprecipitate and tranexamic acid with improved survival following wartime injury: findings from the MATTERs II Study. JAMA Surg 2013. PMID: 23670117
- Shakur H et al. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant hemorrhage (CRASH-2): a randomized, placebo-controlled trial. Lancet 2010 PMID: 20554319
Post Peer Reviewed By: Salim R. Rezaie, MD (Twitter: @srrezaie)