Clinical paperThe accuracy of various neuro-prognostication algorithms and the added value of neurofilament light chain dosage for patients resuscitated from shockable cardiac arrest: An ancillary analysis of the ISOCRATE study
Introduction
Neurological prognostication after cardiopulmonary resuscitation has been widely explored over the past decades, with an abundance of literature.1., 2. It still remains challenging with 2 main goals: avoiding the self-fulfilling prophecy of a poor functional outcome; and adequate detection of patients with a poor prognosis to avoid futile medical care.
Neurological prognostication is based on a multimodal approach, synthetized as algorithms including clinical examination, imaging, electrophysiology and biology.3 The first guideline, published by ESICM in 2015,4 was updated in 2021.5 However, scientific evaluations of these algorithms is poorly documented in clinical practice and they lack sensitivity.6 Most of the components are subject to human observation and interpretation, with a significant risk of bias considering that the observer is often also the care-deliver.
Biological parameters are of particular interest because they are unaffected by level of consciousness and interpretation bias. However, only Neuron-Specific Enolase (NSE) has been included in guidelines.5., 7. Current research focuses on various biomarkers, such as S100 calcium-binding protein b (S100b), tau protein and Glial fibrillary acidic protein (GFAP). Although those markers provide excellent specificity, they have nevertheless failed to show better sensitivity than tests that are already available.1 Neurofilament light chain (NF-L) plasma dosage is at the research stage, but it is a promising biological marker in post-cardiac arrest prognostication.8., 9. It is one of the 3 primary neurofilament isotypes that have been shown to increase in both the cerebrospinal fluid and blood in the presence of axonal damage and neurodegeneration.10 It is a neuronal cytoskeleton-specific protein, and its release in plasma has been described in several situations of axonal damage, such as multiple sclerosis, Parkinson’s disease, and other degenerative and/or specific axonal diseases.
We performed a post-hoc analysis of the ISOCRATE study to assess the predictive value of NF-L in patients with a shockable rhythm who received cardiopulmonary resuscitation, and compared it with the ESICM guideline algorithms. We also evaluated the predictive value of a modified algorithm when NFL dosage is included with previously described cut-off values.8
Section snippets
Study population and research approval
The ISOCRATE trial (NCT02555254) was a prospective, randomized study of patients resuscitated from shockable cardiac arrest who had had targeted temperature management at 33 °C. It aimed to compare high-speed rewarming (0.5 °C/h) to low-speed rewarming (0.25 °C/h) in comatose patients admitted to intensive care after shockable cardiac arrest. It was conducted in the La Roche Sur Yon District Hospital Center ICU, France, between February 12, 2016, and May 15, 2020. The primary outcome was the
Outcome
A favorable neurological outcome on day 90 was defined as a Cerebral Performance Category (CPC) score of 1 or 2. The CPC was assessed during a semi-structured telephone interview by a single psychologist specifically trained for the study in a similar way to the HYPERION trial.11 The CPC scale ranges from 1 to 5, with 1 indicating good cerebral performance or minor disability, 2 moderate disability, 3 severe disability, 4 coma or a vegetative state, and 5 brain death.12
Neurofilament Light Chain (NF-L) assay
Blood samples for NF-L measurement was obtained 48 h after cardiac arrest. All samples were pre-treated, aliquoted, and frozen at −80 °C at the collection site. Plasma NF-L concentration was measured using the Ella™ microfluidic platform (ProteinSimple kit, Bio-Techne, Rennes, France) according to the manufacturer’s instructions at CIMNA (Center for Immunologic Monitoring Nantes Atlantic; Nantes, France). Previous literature has demonstrated a strong correlation with the SIMOA, Quanterix©
Electroencephalography
We classified electroencephalographic information from reports according to 4 items: reactivity, the presence of a highly malignant pattern (i.e., suppressed background +/− periodic discharges or burst suppression), the presence of other unfavorable patterns (electroencephalographic status epilepticus), and benign patterns.14
Algorithm assessment
Each patient’s medical record was reviewed and the clinical history, imaging, electrophysiology and biology results were collected in order to evaluate prognosis in 4 ways:
- –
Neurological prognosis according to the 2015 European Resuscitation Council/ European Society of Intensive Care Medicine (ERC/ESICM) guidelines4
- –
Neurological prognosis according to the 2021 ERC/ESICM guidelines5
- –
Modified ERC/ESICM 2021 guideline algorithm including NF-L with a 500 pg/ml threshold (poor outcome if above, better
Statistical analysis
Qualitative variables were expressed as frequencies and percentages and compared using the chi-squared test. Quantitative variables were expressed as mean ± SD if normally distributed and median [interquartile range] otherwise. Quantitative variables were compared between the groups using the Wilcoxon test. Sensitivity and specificity were also estimated as percentages with their 95% confidence intervals (95 %CIs).
Results
In the ISOCRATE trial, 58 patients were included and 50 were randomized. NF-L dosage 48 h after cardiac arrest was measured for 49 patients who were included in this analysis. Table 1 shows the baseline characteristics of the patients included. The outcome was qualified as poor (CPC 3 to 5 or death) for 23 patients (47.0%). Among them, 7 patients (14.3%) died in the ICU for non-neurological causes (3 for multiple organ failure under 72 hours after cardiac arrest, 2 for early cardiac arrest
Discussion
In our study, a high NF-L plasma level 48 h after cardiac arrest was significantly associated with a poor outcome. Our study also confirms the excellent specificity of NF-L and/or guideline algorithms, with no false positives recorded.
The 2015 ERC/ESICM guideline has chosen to focus on specificity rather than sensitivity regarding the decision to withdraw life sustaining therapies. Concurrently specificity is always perfect in the 3 evaluations of these guidelines.6., 15., 16. However, low
Conclusion
A high NF-L plasma level 48 h after cardiac arrest was significantly associated with a poor outcome. The adjunction of an NF-L assay 48 h after cardiac arrest with a 500 pg/ml threshold to the current guideline algorithm provided the best sensitivity compared to the algorithm alone, while specificity remained excellent.
Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Dr. Lascarrou has received reimbursement for travel expenses and lecture fees from Zoll (Voisin Le Bretonneux, France) and BD (Le Pont de Claix, France). None of the other authors has any conflict of interest to declare.
Acknowledgement
We would like to thank the healthcare staff and research nurses at the trial site; A. Wolfe, MD, for assistance in preparing and reviewing the manuscript and S. Britt for managing the database. We are grateful to I. Gralepois for performing the day-90 neurological evaluations.
References (23)
- et al.
Blood biomarkers of brain injury after cardiac arrest – a dynamic field
Resuscitation
(2020) - et al.
European Resuscitation Council and European Society of Intensive Care Medicine Guidelines for Post-resuscitation Care 2015
Resuscitation
(2015) - et al.
European Resuscitation Council and European Society of Intensive Care Medicine Guidelines 2021: Post-resuscitation care
Resuscitation
(2021) - et al.
NSE as a predictor of death or poor neurological outcome after non-shockable cardiac arrest due to any cause: ancillary study of HYPERION trial data
Resuscitation
(2021) - et al.
Distinct predictive values of current neuroprognostic guidelines in post-cardiac arrest patients
Resuscitation
(2019) - et al.
Protocol-driven neurological prognostication and withdrawal of life-sustaining therapy after cardiac arrest and targeted temperature management
Resuscitation
(2017) - et al.
A survey on general and temperature management of post cardiac arrest patients in large teaching and university hospitals in 14 European countries—The SPAME trial results
Resuscitation
(2017) - et al.
Brain injury after cardiac arrest: pathophysiology, treatment, and prognosis
Intensive Care Med
(2021) - et al.
How to assess prognosis after cardiac arrest and therapeutic hypothermia
Crit Care Lond Engl
(2014) - et al.
Performance of a guideline-recommended algorithm for prognostication of poor neurological outcome after cardiac arrest
Intensive Care Med
(2020)