CRYOSTAT-2

Davenport R, JAMA: Published online October 12th 2023 doi:10.1001/jama.2023.21019

Clinical Question

• In patients with trauma and critical bleeding who require activation of a major haemorrhage protocol, does the empiric administration of 3 pools of cryoprecipitate (6g fibrinogen) within 90 minutes of randomisation (and no more than 3 hours after injury) improve survival, compared to standard care?

Background

• The global burden of trauma is huge, with 4.4 million deaths worldwide per annum. It remains the leading cause of death in young people
• Deaths often occur early – > 50% of patients who die, do so in the first 4 hours
• The most common mechanism of death is haemorrhage, which is exacerbated by coagulopathy and fibrinolysis
• Fibrinogen acts to stabilise clot formation. Trauma causes activation of coagulation pathways and fibrinogen levels drop rapidly due to consumption, fibrinolysis and dilution
• There is an association between low fibrinogen levels and poor outcomes, and it is hypothesised that cryoprecipitate transfusion may ameliorate this state
• The ideal approach to transfusion in traumatic haemorrhage has been extensively studied and refined over the years
• The PAMPer trial demonstrated that for trauma patients at risk for hemorrhagic shock, the administration of thawed plasma during prehospital air medical transport was safe and resulted in lower 30-day mortality
• The CRYOSTAT pilot trial conducted at two UK trauma centres showed that administration of cryoprecipitate (10 units) in addition to the empiric massive haemorrhage pack (6u PRBC, 4u FFP +/- TXA) raised fibrinogen levels, and was feasible, with a mean time to administration of 60 minutes

Design

• Multi-centre, phase 3, interventional RCT
• Open-label, parallel-groups
• Allocation concealment achieved using opaque sealed envelopes
• The randomisation sequence was computer generated with variable block sizes and stratified by centre
• A waiver of consent was allowed in order to randomise patients
  • Informed consent to continue data collection was obtained from the patient or legally authorised representative as soon as was practically possible
• The trial was designed to detect a 7% absolute mortality difference with 90% power and a 5% level of significance
• A sample size of 1600 patients was calculated adjusted after 4% drop out
• Analysis was by intention to treat
• The trial was overseen by an ethics committee and a data monitoring committee to review interim data and monitor patient safety

Setting

• 26 major trauma centres in the UK and USA.
• Patients were enrolled from August 2017 to November 2021

Population

• Inclusion:
  • Trauma with evidence of active haemorrhage requiring activation of the local Major Haemorrhage Protocol (MHP)
    • MHP activation at all centres included SBP < 90 mmHg at any point
  • Started or received at least 1 unit of any blood component
• Exclusion:
  • Transferred from another hospital
  • Injuries deemed incompatible with life by trauma team leader
  • >3 hours elapsed since the time of injury

• Baseline characteristics were well matched:
  • Comparing cryoprecipitate group vs standard care:
  • Median Injury Severity Score: 29 vs 29
  • Men: 79 vs 80%
  • Median Age: 38 vs 40
    • 10% > 70 years old
  • Median time from injury to ED arrival: 75 vs 77 minutes
  • Blunt injury: 63 vs 65%
  • Penetrating injury: 37 vs 35%
  • Head injury (AIS >4): 24% vs 29%
  • Prehospital Tranexamic Acid: 79 vs 80%
  • Prehospital Blood Products: 43% across both groups
  • Median time to randomisation: 15 minutes across both groups

Intervention

• 3 pools of cryoprecipitate (6g fibrinogen) given as early as possible (ideally within 90 minutes) after hospital admission AND MHP as per control group
  • 85% of patients got cryoprecipitate in the intervention group
  • The main reasons for not receiving it were no evidence of active bleeding, haemostasis was achieved or the patient died
  • Median time to cryoprecipitate was 68 minutes (68% received the first dose within 90 minutes)

Control

• As per local MHP – typically patients were transfused RBC:FFP:Platelets in a ratio of 1:1:1 with 2 pools cryoprecipitate with second and subsequent pack (4g fibrinogen)
  • 32% of patients received cryoprecipitate in the control group
  • Median time to cryoprecipitate was 120 minutes (9% received the first dose within 90 minutes)

Management common to both groups

• All other therapies were as per local major trauma protocols

Outcome

• Comparing Cryoprecipitate vs Standard Care
• Primary outcome:
  • No difference in all-cause mortality at 28 days
  • 25.3% (192 of 760) vs 26.1% (201 of 771)
    • OR 0.96 [95% CI 0.75-1.23] p=0.74
    • The per-protocol analysis also showed no difference
• Secondary outcomes:
• No difference between groups in:
  • 6-hour mortality: 7.1% vs 8.6% (OR 0.82 [0.58-1.17] p=0.26)
  • 24-hour mortality: 11.2% vs 12.2% (OR 0.91 [0.63-1.31] p=0.61)
  • 6- or 12-month mortality
  • Transfusion Requirements, Ventilator Days, ICU LOS, Hospital LOS, GOS score at day 28
• Median time to death was 191 minutes in the cryoprecipitate group and 86 minutes in the control group
• An analysis of the impact of time to cryoprecipitate administration on 28-day mortality (eFigure2):
  • < 45 minutes: OR 1.45 [0.91 – 2.31]
  • 46-60 minutes: OR 1.16 [95% CI, 0.78-1.73]
  • 61-90 minutes: OR 0.57 [95% CI, 0.38-0.87]
  • > 90 minutes: OR 1.00 [95% CI, 0.62-1.60]
• Pre-defined subgroups:
  • No difference in results based on sex, age, head injury
  • The penetrating injury group fared worse with cryoprecipitate
    • 28-day mortality 10% vs 16.2% OR 1.74 [1.2-2.51]
  • The blunt injury group had 28-day mortality of 30.4% (Cryo) vs 34.8% (Standard Care) OR 0.82 [0.62-1.09]
• Safety:
  • No significant difference in
    • Thrombotic events (PE, DVT)
    • Myocardial infarction, Stroke

Authors’ Conclusions

• Early empiric high dose cryoprecipitate in patients with haemorrhage following trauma did not reduce 28-day mortality

Strengths

• Strong internal validity:
  • Allocation concealment
  • Intention-to-treat analysis
  • Well balanced baseline characteristics
  • Minimal loss for primary analysis
    • 4.6% (39 patients in cryoprecipitate group and 34 patients in control group)
• An important clinical question with outcomes that are clinically meaningful
• Patients had high injury severity scores thus increasing the chance of finding a treatment effect
• The results are likely applicable to other similar advanced trauma care systems
  • However, this was primarily a trial in UK hospitals, with only 49 US patients randomised so the external applicability outside the UK would need to be assumed cautiously

Weaknesses

• Unblinded – a potential source of bias
• There was some crossover between groups
  • 15% of cryoprecipitate group did not receive any
  • 9% of the control group got early cryoprecipitate (within 90 minutes)
  • This was not deemed to impact the results in the per-protocol analysis
• The intervention was empiric and not tailored to the individual’s needs such that some patients may not have had hypofibrinogenemia
  • Giving cryoprecipitate in these patients with normal fibrinogen may be pro-inflammatory and harmful
• There is a signal to benefit in the blunt trauma subgroup and harm in the penetrating subgroup. However, there is inadequate statistical power to interpret these subgroup findings

The Bottom Line

• The results from this trial do not support empiric cryoprecipitate therapy in trauma patients with major haemorrhage
• The hypothesis that there is a Goldilocks zone (i.e. 61-90 minutes) in which administration of cryoprecipitate may be beneficial is thought-provoking and warrants further exploration
• I will use ROTEM to guide product administration in my bleeding trauma patients

External Links

• Article CRYOSTAT-2
• Further reading PATCH trial
• Further reading Centre for Trauma Sciences
• Further reading St Emlyn’s review
• Critical Care Reviews Presentation

Metadata

Summary author: Celia Bradford @celiabradford
Summary date: October 14 2023
Peer-review editor: George Walker

Picture by: Clker-Free-Vector-Images / Pixabay