(EBM) Crash-2 trial





A. Background

CRASH-2 = Clinical Randomisation of an Antifibrinolytic in Significant Haemorrhage 2

  • [Lancet 2010; 376:23-32]

  • Use of tranexamic acid in trauma patients

⦿ NB:

  • The first CRASH study was on steroids ... in head injury
    • "Corticosteroid Randomisation ... After Significant Head Injury" [Lancet 2004; 364:1321-28]

B. Methods

Large multicentered double-blinded
... randomised control trial

n = 20211 randomised

  • Tranexamic acid = 10096
  • Placebo = 10115

274 hospitals in 40 countries

Regime of tranexamic acid (TXA)

  • Loading dose = 1 g over 10 min
  • Followed by infusion = 1 g over 8 hours

Inclusion

Inclusion criteria is clinical

Adult trauma patient

  • Significant haemorrhage
    • SBP < 90 mmHg and/or HR > 110
    • OR considered to be at risk
  • Within 8 hours of injury

⦿ NB:

Patient is still included when it is uncertain if TXA should be used

Outcome measured

➤ Primary outcome:

  • Death in hospital within 4 weeks
  • Causes of death categorised into:
    • Bleeding
    • Vascular occlusion
      • MI
      • Stroke
      • PE
    • Multiorgan failure
    • Head injury
    • Others

➤ Secondary outcome:

  • Vascular occlusive events
    • MI
    • CVA
    • PE
    • DVT
  • Surgical intervention
  • Blood transfusion
    • and units of blood products transfused

➤ Dependency measured:

  • At discharge
  • Or at Day 28 if in hospital

C. Findings

In the tranexamic acid group...

  • All-cause mortality was reduced

    • 14.5% vs 16%
    • RR = 0.91, p = 0.0035
    • 95% CI = 0.85 to 0.97

  • ↓ Death due to bleeding

    • 4.9% vs 5.7%
    • RR = 0.85, p = 0.0077
    • 95% CI = 0.76 to 0.96

  • ↓ MI

    • Total, i.e. fatal and non-fatal
    • 0.3% vs 0.5%
    • RR = 0.64, p = 0.035

  • ↓ Death due to vascular occlusion

    • 0.3% vs 0.5%
    • RR = 0.69, p = 0.096
    • i.e. Non-significant
    • Includes MI and others

Other findings:

  • Symptom-free survival more likely
    • 14.7% vs 13.3%
    • RR = 1.11, p = 0.0023

➤ No difference in:

  • Surgical intervention
  • Need for transfusion
  • Units of transfusion received

⦿ NB:

  • No difference in transfusion requirement could be due to patients who survived longer have more opportunity to receive blood transfusion

  • Patients tend to be young (Mean 35 y.o.)

    • May not have significant underlying CAD or cerebrovascular disease

  • ISS (injury severity score) is not used

  • Not all recruited patients have...

    • Coagulopathy (not part of inclusion criteria)
    • Significant haemorrhage
      • Hypotension may be due to other causes



CRASH-2 update

  • Exploratory analysis

    • Focusing on death due to bleeding

  • Published in Lancet 2011

Main findings:

TXA has no significant effect on death due to non-bleeding causes

  • RR = 0.94 ... but non-significant (p = 0.13)

➤ Death due to bleeding

  • Significantly reduced with early TXA
    • If ≤ 1 hr injury to Rx time
      • 5.3% vs 7.7%
      • RR = 0.68 (p < 0.0001)
      • 95% CI 0.57 - 0.82
    • If 1 to 3 hours
      • 4.8% vs 6.1%
      • RR = 0.79 (p = 0.03)
      • 95% CI 0.64 - 0.97
  • But significantly INCREASED if TXA is given late
    • When injury to Rx time is 3 hours or greater
      • 4.4% vs 3.1%
      • RR = 1.44 (p = 0.004)
      • 95% CI 1.12 - 1.84

➤ Re increase in mortality when given ≥ 3 hours after injury:

  • Authors found that to be unexpected and have no explanation
  • Patients who arrive at hospital many hours after injury are likely to differ from those who arrive early

Others

Delay in treatment due to written consent requirement

In one of the correspondence to the CRASH-2 update, it was noted that requirement for written consent from relatives delay initiation of Rx by 1.2 hours (95% CI 0.7 - 1.8)

(This is in reference to MRC-CRASH study, not CRASH-2 itself)

Variations in dosage used

  • Studies on tranexamic acid vary in dosage
  • Loading dose 2.5 to 100 mg/kg
  • Infusion 0.25 to 4 mg/kg/h

Study in cardiac surgery

  • Loading dose 10 mg/kg + infusion 1mg/kg/hr
  • Achieves sufficient plasma level to inhibit fibrinolysis
  • Larger dose does NOT provide additional haemostatic benefit

Comments

  • Acute traumatic coagulopathy is a hyperacute process in which systemic fibrinolysis releases D-dimers that are detectable within 30 min of injury