Elsevier

Resuscitation

Volume 143, October 2019, Pages 115-123
Resuscitation

Clinical paper
Neurophysiology and neuroimaging accurately predict poor neurological outcome within 24 hours after cardiac arrest: The ProNeCA prospective multicentre prognostication study

https://doi.org/10.1016/j.resuscitation.2019.07.032Get rights and content

Abstract

Aims

To investigate the ability of 30-min electroencephalogram (EEG), short-latency somatosensory evoked potentials (SEPs) and brain computed tomography (CT) to predict poor neurological outcome (persistent vegetative state or death) at 6 months in comatose survivors of cardiac arrest within 24 h from the event.

Methods

Prospective multicentre prognostication study in seven hospitals. SEPs were graded according to the presence and amplitude of their cortical responses, EEG patterns were classified according to the American Clinical Neurophysiology Society terminology and brain oedema on brain CT was measured as grey/white matter (GM/WM) density ratio. Sensitivity for poor outcome prediction at 100% specificity was calculated for the three tests individually and in combination. None of the patients underwent withdrawal of life-sustaining treatments before the index event occurred.

Results

A total of 346/396 patients were included in the analysis. At 6 months, 223(64%) had poor neurological outcome; of these, 68 were alive in PVS. Bilaterally absent/absent-pathological amplitude cortical SEP patterns, a GM/WM ratio<1.21 on brain CT and isoelectric/burst-suppression EEG patterns predicted poor outcome with 100% specificity and sensitivities of 57.4%, 48.8% and 34.5%, respectively. At least one of these unfavourable patterns was present in 166/223 patients (74.4% sensitivity). Two unfavourable patterns were simultaneously present in 111/223 patients (49.7% sensitivity), and three patterns in 38/223 patients (17% sensitivity).

Conclusions

In comatose resuscitated patients, a multimodal approach based on results of SEPs, EEG and brain CT accurately predicts poor neurological outcome at 6 months within the first 24 h after cardiac arrest.

Introduction

Prognostication in comatose survivors of cardiac arrest (CA) is challenging. Although some indices such as short-latency somatosensory evoked potentials (SEPs) or pupillary reflexes are highly specific for predicting poor neurological outcome, their sensitivity does not attain 50%.1, 2, 3 As a result, patients destined to a poor outcome are often not detected by these tests, and their prognosis remains indeterminate.4 Moreover, in most prognostication studies, the predictors under investigation have been used as criteria for withdrawal of life-sustaining treatment (WLST), creating a self-fulfilling prophecy bias.5, 6

In patients who are comatose after resuscitation from CA, the current guidelines for Post-Resuscitation Care7 co-issued by the European Resuscitation Council (ERC) and the European Society of Intensive Care Medicine (ESICM) suggest a multimodal prognostication algorithm based on clinical assessment performed at 72 h or later after CA and on a combination of indices, including SEPs, electroencephalogram (EEG), serum biomarkers, and neuroimaging. These guidelines identified pupillary reflexes and SEPs as more robust predictors to be used first, whereas early myoclonus, EEG, biomarkers and neuroimaging are identified as less robust predictors, to be used only in combination and when results of more robust predictors are indeterminate. At the time the ERC-ESCIM guidelines were written the evidence supporting EEG was limited to relatively few studies in which the definitions of the EEG patterns associated with unfavourable outcome were inconsistent. However, in the last few years a series of studies1, 8, 9, 10, 11, 12 adopting the 2012 American Clinical Neurophysiology Society (ACNS) standardized terminology for EEG in the critical care setting,13 has shown that specific EEG patterns predict poor neurological outcome after CA both early and accurately. In particular, in a preliminary retrospective study performed at our Institution,14, 15 we demonstrated that malignant EEG patterns, in combination with the presence of abnormal SEPs or oedema on brain computed tomography (CT) predicted poor neurological outcome at six months with high sensitivity and specificity as early as 24 h after CA. A special strength of this investigation was that WLST was not performed in the study population, limiting the risk of a self-fulfilling prophecy bias.16 In order to validate these results and assess their reproducibility, we conducted the present investigation, based on a predefined subset of patients enrolled in the multicentre prospective ProNeCA study.

Section snippets

Study design and patient selection

The ProNeCA (Prognostication of Neurological outcome after CA; Clinicaltrials.gov:NCT03849911) is a prospective multicentre study conducted in 13 mixed medical-surgical intensive care units(ICUs) from eight university-affiliated and five non-university affiliated Italian hospitals and coordinated by the Careggi University Hospital in Florence, Italy.17 The study included all consecutive adult (≥18years) patients who were admitted to participating ICUs in a coma (Glasgow Coma Scale ≤8) following

Ethical approval

The study protocol was approved by the Regional Ethics Committee of Tuscany (Ref OSS.15.009). Written informed consent was obtained from the patient’s authorized representative prior to the subject enrolment.

Statistical analysis

Continuous variables were reported as median and inter-quartile range (IQR), whereas categorical variables were reported as numbers and percentages. For the Glasgow Coma Scale the range was reported. Normality of baseline distribution was tested using the Shapiro–Wilk test. The Pearson's chi-square and the Mann–Whitney U tests were used for comparing categorical and continuous variables, respectively. For each of the instrumental outcome predictors a receiver operating characteristics (ROC)

Results

Among the 396 patients who were screened for inclusion, 362 fulfilled the inclusion criteria and had all the three tests performed. In six of them the tests could not be analysed, whereas ten patients were lost to follow-up, leaving 346 patients with measured primary outcome at 6 months (Fig.1). Of these, 232 (67%) survived to hospital discharge, of whom 96 (41%) had favourable outcome. At 6-month follow-up, 191 patients were alive, of whom 123 (64.3%) had favourable outcome. Among the 223

Multimodal prediction

At 100% specificity, the grade 2 (AA/AP) SEP pattern was the best single predictor in terms of sensitivity, occurring in 128/223 (57.4%) patients with poor outcome. Among the remaining 93 patients with poor outcome, 23 had a GM/WM ratio <1.21 on brain CT, raising the cumulative sensitivity to 70.4%. Finally, when a malignant EEG pattern was added, 15 additional patients with poor prognosis were identified and the sensitivity for poor outcome prediction raised to 74.4% [68.1–80], with 0% [0–3]

Discussion

This is the first multicentre prospective study evaluating the accuracy of a combination of EEG, SEPs and brain CT for early neurological prognostication after CA. Its results confirmed those of our preliminary study14 and showed that in the majority of comatose resuscitated adults poor neurological outcome can be accurately predicted within 24 h after CA, an early phase when results of clinical examination are unreliable, as demonstrated by previous literature6, 27 and also shown by the high

Conclusions

This is the first prospective multicentre study evaluating the reliability of a multimodal prognostic strategy based on standard EEG, SEPs and brain CT performed within 24 h after CA. Its results showed that using this approach a poor neurological outcome can be predicted early, accurately, and with a very high sensitivity using tests available in most clinical settings. The robustness of these results is reinforced by the absence of a WLST as a standard of care in the population under

Conflict of interest statement

Claudio Sandroni is member of the Editorial Board of Resuscitation, co-author of the 2015 ERC-ESICM Guidelines on Post-Resuscitation Care and lead author of the 2013 ERC-ESICM Advisory Statement on Prognostication in comatose survivors of cardiac arrest.

The remaining authors have no conflict of interest to disclose.

Funding

None.

References (36)

  • J.H. Kim et al.

    Multimodal approach for neurologic prognostication of out-of-hospital cardiac arrest patients undergoing targeted temperature management

    Resuscitation

    (2019)
  • J. Hofmeijer et al.

    Burst-suppression with identical bursts: a distinct EEG pattern with poor outcome in postanoxic coma

    Clin Neurophysiol

    (2014)
  • D. San-juan et al.

    Propofol and the electroencephalogram

    Clin Neurophysiol

    (2010)
  • R.B. Metter et al.

    Association between a quantitative CT scan measure of brain edema and outcome after cardiac arrest

    Resuscitation

    (2011)
  • A. Grippo et al.

    Neurophysiological prediction of neurological good and poor outcome in post-anoxic coma

    Acta Neurol Scand

    (2017)
  • R.G. Geocadin et al.

    Poor survival after cardiac arrest resuscitation: a self-fulfilling prophecy or biologic destiny?

    Crit Care Med

    (2012)
  • C. Sandroni et al.

    Prognostication in comatose survivors of cardiac arrest: an advisory statement from the European Resuscitation Council and the European Society of Intensive Care Medicine

    Intensive Care Med

    (2014)
  • A.O. Rossetti et al.

    Electroencephalography predicts poor and good outcomes after cardiac arrest: a two-center study

    Crit Care Med

    (2017)
  • Cited by (0)

    1

    Collaborating authors are listed in Appendix A.

    View full text