Elsevier

Resuscitation

Volume 139, June 2019, Pages 343-350
Resuscitation

Clinical paper
Distinct predictive values of current neuroprognostic guidelines in post-cardiac arrest patients

https://doi.org/10.1016/j.resuscitation.2019.03.035Get rights and content

Abstract

Purpose

To assess the performance of neuroprognostic guidelines proposed by the American Academy of Neurology (AAN), European Resuscitation Council/European Society of Intensive Care Medicine (ERC/ESICM), and American Heart Association (AHA) in predicting outcomes of patients who remain unconscious after cardiac arrest.

Methods

We retrospectively identified a cohort of unconscious post-cardiac arrest patients at a single tertiary care centre from 2011 to 2017 and reviewed hospital records for clinical, radiographic, electrophysiologic, and biochemical findings. Outcomes at discharge and 6 months post-arrest were abstracted and dichotomized as good (Cerebral Performance Category (CPC) scores of 1–2) versus poor (CPC 3–5). Outcomes predicted by current guidelines were compared to actual outcomes, with false positive rate (FPR) used as a measure of predictive value.

Results

Of 226 patients, 36% survived to discharge, including 24 with good outcomes; 52% had withdrawal of life-sustaining therapies (WLST) during hospitalization. The AAN guideline yielded discharge and 6-month FPR of 8% and 15%, respectively. In contrast, the ERC/ESICM had a FPR of 0% at both discharge and 6 months. The AHA predictors had variable specificities, with diffuse hypoxic-ischaemic injury on MRI performing especially poorly (FPR 12%) at both discharge and 6 months.

Conclusions

Though each guideline had components that performed well, only the ERC/ESICM guideline yielded a 0% FPR. Amongst the AAN and AHA guidelines, false positives emerged more readily at 6 months, reflective of continuing recovery after discharge, even in a cohort inevitably biased by WLST. Further assessment of predictive modalities is needed to improve neuroprognostic accuracy.

Introduction

Neurologic prognosis is frequently uncertain in individuals who are unconscious following cardiac arrest, as the degree of hypoxic-ischaemic brain injury may be difficult to assess early on. For those with return of spontaneous circulation (ROSC) after out-of-hospital cardiac arrest, between 50% and 90% fail to survive to hospital discharge.1, 2 Regardless of the aetiology of arrest, the majority of comatose post-arrest patients die after withdrawal of life-sustaining therapies (WLST) due to a perceived poor neurologic prognosis.3, 4, 5 During neuroprognostication, clinicians must balance two competing goals: 1) avoiding premature WLST in patients who may achieve a good neurologic outcome, and 2) avoiding prolonging care in patients destined for a poor outcome. Adopting a multimodal approach to neuroprognostication is recommended,6, 7, 8, 9, 10, 11 as no individual modality is infallible.

In the 2006 American Academy of Neurology (AAN) practice parameters for cardiac arrest survivors,12 myoclonic status epilepticus (MSE) on post-arrest day 1, bilaterally absent N20 somatosensory evoked potentials (SSEP), elevated serum neuron specific enolase (NSE) levels, and absent pupillary or corneal reflexes with extensor or absent motor response on day 3 are regarded as poor outcome predictors. This algorithm, however, is derived from studies pre-dating the widespread use of targeted temperature management (TTM), which alters cellular metabolism13 and delays clearance of sedatives and paralytics,14, 15, 16 thus also delaying clinical signs of recovery.17, 18, 19, 20

In contrast, the 2014 European Resuscitation Council/European Society of Intensive Care Medicine (ERC/ESICM) guideline21 comprises recent data on TTM-treated patients, with acknowledgement of varying levels of prediction confidence, and reaffirm the complete abolishment of pupillary and corneal reflexes and N20 potentials as robust poor outcome predictors. The 2015 American Heart Association (AHA) guideline22 assesses current prognostic modalities separately, including their timing in relationship to TTM, and establishes the absence of pupillary reflexes at 72 h in TTM-treated patients as the only poor prognostic parameter with Class I evidence.

In an era of variable management strategies and evolving neuroprognostic tools, it is imperative that prognostic strategies are accurate. Measurements of accuracy, however, are inherently confounded by the self-fulfilling prophecy from WLST, in which treating physicians are not blinded to the results of a prognostic assessment and consequently use them to inform care decisions. The purpose of this study is to ascertain the neuroprognostic performance of the AAN, ERC/ESICM, and AHA guidelines when applied to a real-world cohort of patients. We hypothesized that the guidelines overestimate their predictive value, such that actual false positive rates (FPR) at both hospital discharge and 6 months are higher than reported.

Section snippets

Sample selection

The study was approved by the Yale University Human Investigation Committee (HIC# 2000021220), and the need for informed consent was waived. The cohort was retrospectively identified by querying the electronic medical record (EMR) for patients age 18 years or older with a cardiac arrest diagnosis code (ICD-9-427.5/ICD-10-I46.9) between January 2011 and June 2017. Additional inclusion criteria included successful resuscitation and unconsciousness for at least 24 h after ROSC. Patients who died

Demographic data

A total of 226 patients met the study criteria. The population was predominantly male (55%) and non-Hispanic white (58%), with an average age of 58 years (Table 1). Sixty-two percent suffered an out-of-hospital arrest. Non-perfusing rhythms included pulseless electrical activity (50%), asystole (23%), ventricular fibrillation or ventricular tachycardia (20%), and unknown rhythm (7%). Fifty-seven percent of patients underwent TTM, and 45 (30%) had a targeted temperature of 36 °C.

Outcomes at discharge and 6 months

WLST prior to

Discussion

Current neuroprognostication strategies after cardiac arrest are imprecise, stemming from studies marred by the bias of the self-fulfilling prophecy―unavoidable in this setting to date. While the AAN, ERC/ESICM, and AHA all propose discrete and nuancedly different guidelines for neuroprognostication, they place value in common tools used in a multimodal manner—namely, the clinical examination, neuroimaging, and electrophysiologic and biochemical findings. However, predictions of poor outcome

Conclusions

No gold standard currently exists for neuroprognostication of post-cardiac arrest survivors. Comparing the AAN, ERC/ESICM, and AHA guidelines, the ERC/ESICM guideline provides the best specificity for predicting poor outcome, though at the expense of significantly lower sensitivity. Future prospective initiatives should seek to better characterize the value of available neuroprognostic modalities, both independently and in a multimodal fashion, with respect to both discharge and long-term

Conflicts of interest

Ms. Sonya E. Zhou reports no disclosures.

Dr. Carolina B. Maciel reports no disclosures.

Ms. Cora H. Ormseth reports no disclosures.

Dr. Rachel Beekman reports no disclosures.

Dr. Emily J. Gilmore reports no disclosures.

Dr. David M. Greer serves as Editor-in-Chief of Seminars in Neurology and has received compensation for medico-legal consultation.

Authors’ declarations

We wish to confirm that there are no known conflicts of interest associated with this publication, and there has been no significant financial support for this work that could have influenced its outcome.

Acknowledgments

Research in this publication was supported by the National Heart, Lung and Blood Institute of the National Institutes of Health under Award Number T35HL007649. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute or the National Institutes of Health.

References (36)

  • E.J. Benjamin et al.

    Heart disease and stroke statistics-2018 update: a report from the American Heart Association

    Circulation

    (2018)
  • S. Laver et al.

    Mode of death after admission to an intensive care unit following cardiac arrest

    Intensive Care Med

    (2004)
  • R.G. Geocadin et al.

    Neurologic prognosis and withdrawal of life support after resuscitation from cardiac arrest

    Neurology

    (2006)
  • M. Mulder et al.

    Awakening and withdrawal of life-sustaining treatment in cardiac arrest survivors treated with therapeutic hypothermia

    Crit Care Med

    (2014)
  • D.M. Greer et al.

    Neuroprognostication of hypoxic-ischaemic coma in the therapeutic hypothermia era

    Nat Rev Neurol

    (2014)
  • M. Oddo et al.

    Early multimodal outcome prediction after cardiac arrest in patients treated with hypothermia

    Crit Care Med

    (2014)
  • F. Taccone et al.

    How to assess prognosis after cardiac arrest and therapeutic hypothermia

    Crit Care

    (2014)
  • M. Karapetkova et al.

    Early prognostication markers in cardiac arrest patients treated with hypothermia

    Eur J Neurol

    (2015)
  • Cited by (46)

    View all citing articles on Scopus
    View full text