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
Source: View publication →
In patients with acute traumatic brain injury, tranexamic acid administered within 3 hours of injury is safe and reduces head injury-related mortality specifically in patients with mild-to-moderate injuries.
Key Findings
Study Design
Study Limitations
Clinical Significance
The CRASH-3 trial established that early administration of tranexamic acid is a safe and feasible intervention for acute traumatic brain injury, carrying no increased risk of thromboembolic complications. While it may not reverse outcomes in patients with devastating neurologic devastation (GCS 3), its use offers a tangible mortality benefit to those with mild-to-moderate head trauma. Because TXA is globally accessible, inexpensive, and easy to administer, these findings support its routine inclusion in early pre-hospital and trauma-bay treatment protocols for head-injured patients presenting within 3 hours.
Historical Context
Following the landmark 2010 CRASH-2 trial—which demonstrated that TXA reduced mortality in trauma patients with significant extracranial hemorrhage—there remained hesitation regarding its use in isolated traumatic brain injury. Clinicians feared that promoting hemostasis in the brain might inadvertently trigger cerebral thrombosis or exacerbate ischemic injury. Several smaller studies offered mixed or underpowered results. As the largest trial of its kind, CRASH-3 was designed to definitively answer whether the antifibrinolytic benefits of TXA translated to the intracranial compartment. Its findings dissolved long-standing safety concerns regarding TXA in TBI and shifted international trauma paradigms.
Guided Discussion
High-yield insights from every perspective
How does tranexamic acid's mechanism of action directly target the pathophysiology of secondary brain injury following acute traumatic brain injury?
Key Response
Tranexamic acid is a synthetic analog of the amino acid lysine. It reversibly binds to plasminogen, preventing its activation into plasmin. Because plasmin degrades fibrin clots, TXA stabilizes established clots. In acute TBI, early administration of TXA prevents the early expansion of intracranial hemorrhage, which is a primary driver of elevated intracranial pressure, mass effect, ischemia, and subsequent secondary brain injury.
Based on the CRASH-3 trial findings, which specific subset of TBI patients experiences the greatest mortality benefit from TXA, and how does the timing of administration dictate management in the ED?
Key Response
The trial demonstrated that TXA reduces head injury-related death primarily in patients with mild-to-moderate TBI (GCS 9-15) and those with bilaterally reactive pupils. Patients with severe TBI (GCS 3-8 or unreactive pupils) did not show a mortality benefit, likely because their brain injury was already unsalvageable at baseline. Crucially, the benefit is highly time-dependent; TXA must be given within 3 hours of injury, with earlier administration yielding greater mortality reduction. Administration after 3 hours provides no benefit and was excluded from the primary efficacy endpoint.
The CRASH-3 trial excluded patients with severe extracranial bleeding. How does this exclusion alter the neurocritical care and trauma management paradigm for polytrauma patients who present with both severe TBI and massive torso hemorrhage?
Key Response
CRASH-2 established TXA's survival benefit in severe extracranial hemorrhage, while CRASH-3 focused on isolated TBI. Although CRASH-3 showed no isolated mortality benefit for severe TBI (GCS 3-8), a polytrauma patient with severe TBI and massive systemic bleeding still absolutely requires TXA under CRASH-2 protocols. The fellow must synthesize these trials to understand that severe TBI does not contraindicate TXA in polytrauma, and systemic resuscitation priorities to prevent hemorrhagic shock (which exacerbates secondary brain injury via hypotension) take precedence.
Given the CRASH-3 data demonstrating an excellent safety profile with no increase in vascular occlusive events or seizures, should TXA be universally protocolized for prehospital administration by EMS in all suspected TBIs before a definitive GCS can be assessed?
Key Response
This addresses the pragmatism of protocolizing TXA. Because the therapeutic window is extremely narrow ('time is brain') and the safety profile in the trial was robust, shifting administration to the prehospital setting maximizes the early treatment benefit. Attendings must weigh the risk of overtriage (administering it to patients with concussions but no intracranial hemorrhage) against the logistical reality that accurate pre-hospital GCS scoring is difficult, and delaying administration until a definitive CT scan or trauma bay assessment is completed costs salvageable brain tissue and lives.
Scholarly Review
Critical appraisal through the lens of expert reviewers and guideline development
The CRASH-3 trial protocol was amended prior to unblinding to restrict the primary efficacy analysis to patients treated within 3 hours of injury, based on external evidence from the CRASH-2 and WOMAN trials. How does this mid-trial adaptive change impact the statistical power and the interpretability of the intention-to-treat analysis?
Key Response
While statistically justified because it was performed prior to unblinding (preserving Type I error integrity), this amendment reduced the effective sample size for the primary efficacy outcome. It raises deep methodological discussions regarding the handling of patients already enrolled after the 3-hour mark (who were analyzed for safety but not efficacy) and challenges researchers to evaluate how trialists can integrate contemporaneous external validity without compromising the internal validity and originally powered effect sizes of their ongoing studies.
How does the reliance on 'head injury-related death' as the primary endpoint, rather than 'all-cause mortality' or a validated 'functional neurological outcome' metric, introduce potential adjudication bias or limit the clinical impact of the trial's conclusions?
Key Response
A critical reviewer would flag that cause-specific mortality is subjective to adjudicate, especially in complex trauma patients where multi-organ failure, ARDS, or sepsis frequently contribute to death. While head injury-related death decreased in the mild-to-moderate group, the lack of a definitive signal in all-cause mortality or a robust improvement in long-term functional independence (e.g., Glasgow Outcome Scale-Extended) suggests the trial may detect a disease-oriented outcome rather than a definitively patient-oriented functional improvement.
In updating the Brain Trauma Foundation guidelines, what level of recommendation and strength of evidence should be assigned to the administration of TXA for mild-to-moderate TBI, and how should guidelines reconcile the mortality benefit with the lack of significant improvement in disability outcomes among survivors?
Key Response
Guidelines must weigh mortality benefits against morbidity and resource utilization. CRASH-3 showed a relative risk reduction in mortality for GCS 9-15 without increasing the incidence of severe disability, though it did not drastically improve the rate of good functional recovery. The committee must determine if a Level B recommendation for prehospital or early ED TXA in mild/moderate TBI is warranted based solely on survival and the drug's low cost and high safety profile, while specifically noting the evidence gap regarding long-term functional independence measures in current literature.
Clinical Landscape
Noteworthy Related Trials
CRASH-2 Trial
Tested
Tranexamic acid (1g loading dose, 1g infusion over 8 hours)
Population
Adult trauma patients with or at risk of significant bleeding
Comparator
Placebo
Endpoint
All-cause mortality at 4 weeks
RESCUEicp Trial
Tested
Decompressive craniectomy
Population
Patients with traumatic brain injury and refractory intracranial hypertension
Comparator
Standard medical management
Endpoint
Extended Glasgow Outcome Scale (GOS-E) at 6 months
WOMAN Trial
Tested
Tranexamic acid (1g intravenous injection)
Population
Women with postpartum haemorrhage after vaginal or caesarean delivery
Comparator
Placebo
Endpoint
Death from postpartum haemorrhage or hysterectomy within 42 days
Tailored to your role
Want this tailored to you?
Add your specialty or training stage to get role-specific takeaways and more questions.
Personalize this analysis