Traumatic Cardiac Arrest – Can we Find Prognostic Factors that Predict Survival?

Background: Trauma remains the leading cause of death in the United States for those aged less than 45 years old. Those who arrest from hemorrhage or other traumatic mechanism often carry a very poor prognosis. Various studies have placed the survival from blunt traumatic arrest at <10%. Much is dependent however on the systems approach to managing these patients – for example those patients who have very rapid access to surgical resuscitative techniques may have better outcomes. Nevertheless, given the typical young age of these victims, a significant effort is often made at resuscitation. This must be balanced with the potential risks to clinical staff, appropriate use of limited resources, and expected quality of life of survivors. To be able to predict better outcomes would be of use both in the prehospital and in-hospital environments.

Paper: Tran A et al. Pre-arrest and intra-arrest prognostic factors associated with survival following traumatic out-of-hospital cardiac arrest – a systematic review and meta-analysis. Resuscitation 2020. PMID: 32531405

Clinical Question: What are the prognostic associations of various pre- and intra-arrest factors associated with return of spontaneous circulation (ROSC) and survival in adult patients following a traumatic out-of-hospital cardiac arrest (OHCA)?

What They Did

• Systematic review and meta-analysis
• Search of six databases (Medline, Embase, Web of Science, Scopus, Cochrane Database of Systematic Reviews, and Pubmed) from inception to December 2019
• Inclusion of observational (prospective and retrospective) and randomized controlled trials (RCTs)
• Two authors performed data abstraction of key variables
• For each prognostic factor, two authors collected or calculated adjusted odd ratios (aOR) or unadjusted odds ratios (uOR) for ROSC and survival (in-hospital or 30-day)
• Two authors assessed the risk-of-bias for each study
• Meta-analysis was performed on the available uOR using a random-effects model
• GRADE methodology was used to assess overall certainty in pooled estimates, categorizing into high, moderate, low, or very low

 Outcomes

• ROSC or survival (in-hospital or 30-day)

Inclusion

• English language
• All study types
• Majority (>80%) adult patients ≥16 years of age
• Reporting of either ROSC or survival (in-hospital or 30-day)

Exclusion

• Non-English language papers
• Those papers evaluating only hemodynamically unstable patients
• Those papers evaluating only a specific patient intervention e.g. thoracotomy for reasons other than cardiac arrest
• Those papers that did not provide either aOR or uOR with confidence intervals (CI) or those that did not allow authors to calculate uOR

Results

• Search Results
  • 1545 citations were found on their search
    • 1086 were screened after removing duplicates
    • 87 underwent full-text review
    • 53 studies were included with 37,528 patients
  • Of the included studies
    • 45.3% from North America, 28.3% from Europe, 20.8% from Asia, and 5.7% from Australia
    • Study design was observational in all
    • 90.6% retrospective
    • 90.6% studies included a mixed mechanism of injury (blunt and penetrating) and 9.4% included only blunt mechanism
  • Risk of bias
    • Most studies showed low risk of bias in domains of participation, attrition, prognostic factor measurement and outcome measurement
    • Most studies were judged to be at moderate risk of bias for confounding and statistical reporting

• Outcomes (Predictors of ROSC)
  • Pre-arrest factors with low odds of ROSC
    • Male sex (pooled uOR 0.76, 95% CI 0.61-0.93, low certainty)
  • Pre-arrest factors with high odds of ROSC
    • Head injury (pooled uOR 1.84, 95% CI 1.22-2.77, very low certainty)
  • Pre-arrest factors with no association with ROSC
    • Penetrating mechanism – not associated with ROSC (pooled uOR 0.97, 95% CI 0.51-1.85, very low certainty)
  • Intra-arrest factors with higher odds of ROSC
    • Presence of cardiac motion on ultrasound (pooled uOR 15.72, 95%CI 4.62-53.54, low certainty)
    • Shockable initial cardiac rhythm (pooled uOR 2.11, 95% CI 1.58-2.81, low certainty)
    • witnessed arrest (pooled uOR 1.65, 95% CI 1.36-2.00, moderate certainty)
    • use of prehospital or in-hospital epinephrine (pooled uOR 2.69, 95% CI 2.10-3.44, low certainty)
    • prehospital tracheal intubation (pooled uOR 2.77, 95% CI 1.58-4.86, low certainty)
  • Intra-arrest factors not associated with ROSC
    • Bystander CPR (pooled uOR 1.04, 95%CI 0.81-1.34, low certainty)

• Outcomes (Predictors of survival)
  • Pre-arrest factors with no association with survival
    • male sex (pooled uOR 1.19, 95% CI 0.89-1.59, low certainty)
    • penetrating mechanism of injury (pooled uOR 1.40, 95% CI 0.79-2.48, very low certainty)
    • head injury (pooled uOR 0.94, 95% CI 0.56 to 1.58, low certainty)
  • Intra-arrest factors with higher odds of survival
    • presence of cardiac motion on ultrasound (pooled uOR 33.91 95% CI 1.87 – 613.42, low certainty)
    • shockable initial cardiac rhythm (pooled uOR 7.29, 95% CI 5.09-10.44, moderate certainty)
    • witnessed arrest (pooled uOR 1.76, 95% CI 1.19-2.60, low certainty)
    • bystander CPR (pooled uOR 1.89, 95% CI 1.29-2.79, low certainty)
  • Intra-arrest factors with lower odds of survival
    • use of prehospital or in-hospital epinephrine (pooled uOR 0.62, 95% CI 0.44-0.88, very low certainty)
    • prehospital intubation (pooled uOR 0.70, 95% CI 0.53-0.93, low certainty)

• Additional Findings
  • Ultrasound for assessment of cardiac motion
    • 448 patients evaluated
    • No survivors out of 340 patients without cardiac motion
  • Shockable rhythm
    • 397 patients evaluated with shockable rhythm vs 4825 patients without
    • Survivors with shockable rhythm: 14.4%
    • Survivors with non-shockable rhythm: 2.4%
  • Long-term neurologic outcomes
    • 11 studies describing this outcome
    • Utilizing Glasgow Outcome Scale:
      • 7% long-term survivors
      • 7% had good neurologic recovery

Strengths

• Multiple databases searched resulting in a high number of included patients
• Adherence to guideline recommendations for systematic reviews/meta-analyses
• Assessment of multiple factors related to various stages of traumatic OHCA
• Assessment of bias

Limitations

• A systematic review is only as good as the papers it includes – unfortunately, the majority of included studies were observational and retrospective in design
• Majority of studies only allowed meta-analyses of unadjusted odds ratios
• Some studies prevented meta-analysis because of significant variability of included covariates
• All predictors were at serious risk of bias using GRADE methodology
  • Specifically, there is confounding bias, patient selection bias, and resuscitation time bias which can all impact outcomes
• The lack of adjustment for confounders also could bias results, resulting in hypothesis-generating associations rather than true causality
• Few included papers assessed long-term neurologic outcomes

Discussion

• The authors note a significant predictive value of the presence of cardiac motion on ultrasound on both ROSC and survival
  • This falls in line with recent evidence from other centers showing the predictive value of cardiac motion and pericardial fluid in predicting survivors from penetrating traumatic arrest mechanism
  • However, there are only a few studies that have looked at this (hence the wide confidence intervals) – this could be because the role of cardiac POCUS for this purpose has been a relatively newer development and hence has not been studied enough yet
• The authors findings to support a shockable cardiac rhythm as a predictor of survival was interesting
  • Asystole (absent cardiac motion) is clearly a poor prognostic indicator
  • Traumatic arrest occurs primarily from hypovolemia or obstruction – the typical rhythm is more likely therefore to be sinus (often tachycardia) as, unlike with medical arrest, an arrhythmia is less likely the cause of arrest
  • Perhaps the signal is more to the presence of a rhythm in general as opposed to specifically a shockable rhythm
• Mechanism of injury (blunt or penetrating) was not found to be associated with ROSC or survival
  • This is in conflict with available guidance where penetrating injuries may actually be more likely to survive their traumatic arrest
  • Nevertheless, it is likely problematic to lump all mechanisms together, as specifics of injury (for example a single stab wound to the heart versus multiple gunshot wounds to the abdomen or major blunt thoracic injury) are important contributors to survival.
• Head injury was associated with an increased likelihood of ROSC
  • This is challenging to interpret – even if ROSC is more likely in the presence of a head injury, what is the chance of meaningful neurologic function on recovery?
  • The information regarding this is varied and not clearly defined.
• Some variables demonstrate a mixed signal
  • The notable ones are intubation and the use of epinephrine
  • The medical cardiac arrest literature has had variable outcomes with epinephrine, with the PARAMEDIC2 study showing higher 30-day survival compared to placebo but with higher rates of poor neurologic outcome. ROSC alone is therefore not the ultimate goal here – especially in a patient population that is typically younger than the medical cardiac arrest population.
  • The AIRWAYS2 study suggested the a supraglottic airway (SGA) was non-inferior to intubation but a SGA may allow more focus to be placed on high-quality CPR
  • This study suggested epinephrine and prehospital intubation predicted higher ROSC but lower overall survival. The role of vasoactive agents in hemorrhagic shock has long been questioned and is often not utilized and this study seems to support not utilizing it. Intubation however addresses a key contributor to arrest, hypoxia, and so its role is likely still important. What may be difficult to address in this study are other factors associated with intubated including the role of adding positive-pressure ventilation in an under-resuscitated, hypovolemic patient
• Bystander CPR was not associated with ROSC
  • Traumatic arrest is not medical arrest. While there were inconsistencies in the results in this study, it is more likely that rapid identification and management of arrest are better predictors of outcome than performing CPR in this population. Recall the CPR may delay the use of life-saving surgical interventions that can address some of the causes of traumatic arrest (e.g. decompression of a tension pneumothorax).
• Response times and duration of arrest were not accounted for in any of the included studies
  • This is critical as it may well have been the case that those that did well had rapid access to intervention and very short arrest times – however, this is not examined in this review.
  • This plays in to the role of CPR – there is confounding as witnessed arrests of short duration may have had better outcomes with or without CPR. Hence the role of CPR itself is unclear.
• Overall this is an important addition to the traumatic arrest literature, addressing some key gaps and strengthening findings from other trials. Keeping in mind the overall poor outcomes of this pathology as well as the need to best utilize limited resources and protect staff, the findings of this study can assist staff in tailoring their traumatic arrest protocols with an evidence-based approach to better target those individuals who will most benefit from resuscitative efforts.

Author(s) Conclusion: “This review provides very low to moderate certainty evidence that pre- and intra-arrest prognostic factors following penetrating or blunt traumatic OHCA predict ROSC and survival. This evidence is primarily based on unadjusted data. Further well-designed studies with larger cohorts are warranted to test the adjusted prognostic ability of pre- and intra-arrest factors and guide therapeutic decision-making”

Clinical Take Home Point: In this systematic review and meta-analysis, there was a signal to improved survival from traumatic out-of-hospital arrest with the presence of cardiac motion on ultrasound and a shockable initial rhythm. The findings do need to be taken in the context of the overall poor statistical quality (primarily retrospective and observational) and the significant risk of bias of included studies. While this study, within these limitations, is a valiant attempt to clarify gaps in clinical care around an important pathophysiologic process, it is also a call to arms to improve future research efforts in identifying the best predictors of good outcomes in traumatic arrest.

References

1. Tran A et al. Pre-arrest and intra-arrest prognostic factors associated with survival following traumatic out-of-hospital cardiac arrest – a systematic review and meta-analysis. Resuscitation 2020. PMID: 32531405
2. Evans CC et al. Prehospital traumatic cardiac arrest: management and outcomes from the resuscitation outcomes consortium epistry-trauma and PROPHET registries. J Trauma Acute Care Surg 2016. PMID: 27070438
3. Lockey D et al. Traumatic cardiac arrest: who are the survivors? Ann Emerg Med 2006. PMID: 16934644
4. Inaba K et al. FAST ultrasound examination as a predictor of outcomes after resuscitative thoracotomy: a prospective evaluation. Ann Surg 2015. PMID: 26258320
5. Perkins GD. A randomized trial of epinephrine in out-of-hospital cardiac arrest. NEJM 2018. PMID: 30021076
6. Benger J et al. Effect of a strategy of a supraglottic airway device vs tracheal intubation during out-of-hospital cardiac arrest on functional outcome: the AIRWAYS-2 randomized clinical trial. JAMA 2018. PMID: 30167701

Post Peer Reviewed By: Salim R. Rezaie, MD (Twitter: @srrezaie)