Resuscitation 85 (2014) 1674–1680
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical Paper
Bispectral Index to Predict Neurological Outcome Early After Cardiac Arrest夽 Pascal Stammet a,∗ , Olivier Collignon b , Christophe Werer a , Claude Sertznig a , Yvan Devaux c a
Department of Anaesthesia and Intensive Care Medicine, Centre Hospitalier de Luxembourg, 4, rue Barblé, L-1210 Luxembourg, Luxembourg Competence Centre for Methodology and Statistics, Centre de Recherche Public de la Santé (CRP-Santé), 1A-B, rue Thomas Edison, L-1445 Strassen, Luxembourg c Laboratory of Cardiovascular Research, Centre de Recherche Public de la Santé (CRP-Santé), 84, Val Fleuri, L-1526 Luxembourg, Luxembourg b
a r t i c l e
i n f o
Article history: Received 8 July 2014 Received in revised form 2 September 2014 Accepted 10 September 2014 Keywords: Cardiac arrest Electroencephalogram Prognosis Brain injury Prediction models.
a b s t r a c t Aim of the study: To address the value of continuous monitoring of bispectral index (BIS) to predict neurological outcome after cardiac arrest. Methods: In this prospective observational study in adult comatose patients treated by therapeutic hypothermia after cardiac arrest we measured bispectral index (BIS) during the first 24 hours of intensive care unit stay. A blinded neurological outcome assessment by cerebral performance category (CPC) was done 6 months after cardiac arrest. Results: Forty-six patients (48%) had a good neurological outcome at 6-month, as defined by a cerebral performance category (CPC) 1-2, and 50 patients (52%) had a poor neurological outcome (CPC 3-5). Over the 24 h of monitoring, mean BIS values over time were higher in the good outcome group (38 ± 9) compared to the poor outcome group (17 ± 12) (p < 0.001). Analysis of BIS recorded every 30 minutes provided an optimal prediction after 12.5 h, with an area under the receiver operating characteristic curve (AUC) of 0.89, a specificity of 89% and a sensitivity of 86% using a cut-off value of 23. With a specificity fixed at 100% (sensitivity 26%) the cut-off BIS value was 2.4 over the first 271 minutes. In multivariable analyses including clinical characteristics, mean BIS value over the first 12.5 h was a predictor of neurological outcome (p = 6E-6) and provided a continuous net reclassification index of 1.28% (p = 4E-10) and an integrated discrimination improvement of 0.31 (p = 1E-10). Conclusions: Mean BIS value calculated over the first 12.5 h after ICU admission potentially predicts 6months neurological outcome after cardiac arrest. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Successfully resuscitated cardiac arrest remains a condition with a high mortality rate, even when applying the best medical care to the patients.1–4 This high mortality of about fifty percent of the patients admitted to the hospital after successful pre-hospital resuscitation is mainly due to neurological impairment consecutive to the anoxic period during cardiac arrest and possibly during reperfusion.5,6 The ability to predict outcome early after cardiac arrest would represent a major breakthrough towards
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2014.09.009. ∗ Corresponding author. Department of Anaesthesia and Intensive Care Medicine, Centre Hospitalier de Luxembourg, 4, rue Barblé, L-1210 Luxembourg, Luxembourg. E-mail addresses: [email protected], [email protected] (P. Stammet). http://dx.doi.org/10.1016/j.resuscitation.2014.09.009 0300-9572/© 2014 Elsevier Ireland Ltd. All rights reserved.
personalized medicine by adapting the treatment strategy individually to the patient. This early prediction would allow avoiding futile healthcare to patients with irreversible neurological damage while maintaining resources in patients most likely to benefit. However, neurological prognostication of comatose patients after successful resuscitation and admission to the intensive care unit (ICU) remains a challenge.7 Several studies have focused on the prognostic value of biomarkers and electrophysiological parameters, but none could demonstrate an accurate prediction of outcome within the first 24 hours after cardiac arrest in hypothermia-treated patients.7–10 Thus, there is an unmet need for early prediction tools. The actual lapse of time to reliably predict outcome, generally 24, 48 or even 72 hours after cardiac arrest, might be too lengthy if therapeutic decisions on highly invasive acute cardiac assistance have to be taken.11,12 In previous reports, bispectral index (BIS), a processed electroencephalogram, initially designed to assess the depth of
P. Stammet et al. / Resuscitation 85 (2014) 1674–1680
anesthesia, has potential to predict outcome early after cardiac arrest.13–17 In these studies, either BIS value monitored at a single time-point or the lowest recorded BIS value were considered in prediction analyses. The aim of this study was to refine the value of BIS as an early predictor of outcome after cardiac arrest. Using serial measurements of BIS over the first 24 hours after admission to the ICU, we determined BIS cut-off values and time of recording providing optimal prediction of outcome. 2. Methods 2.1. Patients We included all successfully resuscitated adult cardiac arrest patients enrolled in a prospective local registry admitted from February 2009 to June 2013 to our general ICU. Part of the patients have been involved in previous studies, although none of these addressed the prognostic value of continuous BIS monitoring.16,18–20 All patients were older than 18 years, unconscious (Glasgow coma score below or equal to 8) and received induced hypothermia at 33 ◦ C with sedation and neuromuscular blockade according to our standard protocol (midazolam, max. 0.25 mg/kg/h; fentanyl, max. 2.5 g/kg/h and cisatracurium 0.15 mg/kg/h).14,16 This sedation regimen allows suitable sedation with minimal electromyogram (EMG) artefacts on the electroencephalogram (EEG) signal. After 24 h, patients were rewarmed to 36 ◦ C at a maximum rate of 0.5 ◦ C/h and sedation was tapered. Patients’ relatives were asked for informed consent according to the requirements of the National Committee for Ethics in Research (which approved the protocol). Patients regaining consciousness were also re-consented a posteriori. Only data sets with approved consents were considered in this study. Decisions to withdraw life support or to limit care were never taken by considering BIS values. Only clinically relevant parameters like absence of awakening after complete cessation of sedation, signs of brain death or early myoclonus or status epilepticus in combination with bilaterally absent N20 peak on somato-sensory evoked potentials or imaging findings compatible with irreversible brain damage were taken into account.21,22 2.2. Bispectral index monitoring The BIS (BIS XP Quatro, COVIDIEN, Mansfield, Massachusetts, USA) monitor was routinely applied to every patient by the nursing staff: application of the electrodes on the dry skin of the forehead according to the manufacturer’s instructions. Signal quality was immediately assessed by the treating nursing personnel, using the build-in tools, to obtain a stable and high signal quality index (SQI) and low EMG artefacts. BIS values were continuously registered via the patient monitoring system (IntelliVue, Philips, Böblingen, Germany), recorded in the patient data management system (Metavision, IMDsoft, Tel Aviv, Israel) and retrospectively exported as excel files into R. Throughout the 24 hour study period, only BIS values that fulfilled our quality criteria of low EMG artefacts (80) were retained for the analysis. Patients without a BIS monitoring or without BIS data meeting these quality criteria were excluded from the study. 2.3. Neurological evaluation The endpoint of this study was neurological outcome as defined by the cerebral performance category (CPC) score at 6 months,23 CPC 1-2 being considered as good outcome and CPC 3-5 being considered as poor outcome. CPC score 1 or 2 indicates no or minor neurological sequelae, 3-4 indicates severe neurological sequelae
1675
Table 1 Study population. Characteristic
Good outcome (CPC 1-2, n = 46)
Poor outcome (CPC 3-5, n = 50)
P value
Age, years Gender Male SAPS II Time to ROSC, min Initial rhythm VF/VT PEA Asystole Associated factors Shock AMI EEG Status epilepticus Seizures Medical history Tobacco Hypertension Heart failure Coronary disease Pulmonary disease Diabetes
57 (21-81)
67 (24-83)
0.023
42 (91%) 61 (43-83) 19 (4-60)
36 (72%) 71 (48-100) 30 (3-104)
0.016
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical Paper
Bispectral Index to Predict Neurological Outcome Early After Cardiac Arrest夽 Pascal Stammet a,∗ , Olivier Collignon b , Christophe Werer a , Claude Sertznig a , Yvan Devaux c a
Department of Anaesthesia and Intensive Care Medicine, Centre Hospitalier de Luxembourg, 4, rue Barblé, L-1210 Luxembourg, Luxembourg Competence Centre for Methodology and Statistics, Centre de Recherche Public de la Santé (CRP-Santé), 1A-B, rue Thomas Edison, L-1445 Strassen, Luxembourg c Laboratory of Cardiovascular Research, Centre de Recherche Public de la Santé (CRP-Santé), 84, Val Fleuri, L-1526 Luxembourg, Luxembourg b
a r t i c l e
i n f o
Article history: Received 8 July 2014 Received in revised form 2 September 2014 Accepted 10 September 2014 Keywords: Cardiac arrest Electroencephalogram Prognosis Brain injury Prediction models.
a b s t r a c t Aim of the study: To address the value of continuous monitoring of bispectral index (BIS) to predict neurological outcome after cardiac arrest. Methods: In this prospective observational study in adult comatose patients treated by therapeutic hypothermia after cardiac arrest we measured bispectral index (BIS) during the first 24 hours of intensive care unit stay. A blinded neurological outcome assessment by cerebral performance category (CPC) was done 6 months after cardiac arrest. Results: Forty-six patients (48%) had a good neurological outcome at 6-month, as defined by a cerebral performance category (CPC) 1-2, and 50 patients (52%) had a poor neurological outcome (CPC 3-5). Over the 24 h of monitoring, mean BIS values over time were higher in the good outcome group (38 ± 9) compared to the poor outcome group (17 ± 12) (p < 0.001). Analysis of BIS recorded every 30 minutes provided an optimal prediction after 12.5 h, with an area under the receiver operating characteristic curve (AUC) of 0.89, a specificity of 89% and a sensitivity of 86% using a cut-off value of 23. With a specificity fixed at 100% (sensitivity 26%) the cut-off BIS value was 2.4 over the first 271 minutes. In multivariable analyses including clinical characteristics, mean BIS value over the first 12.5 h was a predictor of neurological outcome (p = 6E-6) and provided a continuous net reclassification index of 1.28% (p = 4E-10) and an integrated discrimination improvement of 0.31 (p = 1E-10). Conclusions: Mean BIS value calculated over the first 12.5 h after ICU admission potentially predicts 6months neurological outcome after cardiac arrest. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Successfully resuscitated cardiac arrest remains a condition with a high mortality rate, even when applying the best medical care to the patients.1–4 This high mortality of about fifty percent of the patients admitted to the hospital after successful pre-hospital resuscitation is mainly due to neurological impairment consecutive to the anoxic period during cardiac arrest and possibly during reperfusion.5,6 The ability to predict outcome early after cardiac arrest would represent a major breakthrough towards
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2014.09.009. ∗ Corresponding author. Department of Anaesthesia and Intensive Care Medicine, Centre Hospitalier de Luxembourg, 4, rue Barblé, L-1210 Luxembourg, Luxembourg. E-mail addresses: [email protected], [email protected] (P. Stammet). http://dx.doi.org/10.1016/j.resuscitation.2014.09.009 0300-9572/© 2014 Elsevier Ireland Ltd. All rights reserved.
personalized medicine by adapting the treatment strategy individually to the patient. This early prediction would allow avoiding futile healthcare to patients with irreversible neurological damage while maintaining resources in patients most likely to benefit. However, neurological prognostication of comatose patients after successful resuscitation and admission to the intensive care unit (ICU) remains a challenge.7 Several studies have focused on the prognostic value of biomarkers and electrophysiological parameters, but none could demonstrate an accurate prediction of outcome within the first 24 hours after cardiac arrest in hypothermia-treated patients.7–10 Thus, there is an unmet need for early prediction tools. The actual lapse of time to reliably predict outcome, generally 24, 48 or even 72 hours after cardiac arrest, might be too lengthy if therapeutic decisions on highly invasive acute cardiac assistance have to be taken.11,12 In previous reports, bispectral index (BIS), a processed electroencephalogram, initially designed to assess the depth of
P. Stammet et al. / Resuscitation 85 (2014) 1674–1680
anesthesia, has potential to predict outcome early after cardiac arrest.13–17 In these studies, either BIS value monitored at a single time-point or the lowest recorded BIS value were considered in prediction analyses. The aim of this study was to refine the value of BIS as an early predictor of outcome after cardiac arrest. Using serial measurements of BIS over the first 24 hours after admission to the ICU, we determined BIS cut-off values and time of recording providing optimal prediction of outcome. 2. Methods 2.1. Patients We included all successfully resuscitated adult cardiac arrest patients enrolled in a prospective local registry admitted from February 2009 to June 2013 to our general ICU. Part of the patients have been involved in previous studies, although none of these addressed the prognostic value of continuous BIS monitoring.16,18–20 All patients were older than 18 years, unconscious (Glasgow coma score below or equal to 8) and received induced hypothermia at 33 ◦ C with sedation and neuromuscular blockade according to our standard protocol (midazolam, max. 0.25 mg/kg/h; fentanyl, max. 2.5 g/kg/h and cisatracurium 0.15 mg/kg/h).14,16 This sedation regimen allows suitable sedation with minimal electromyogram (EMG) artefacts on the electroencephalogram (EEG) signal. After 24 h, patients were rewarmed to 36 ◦ C at a maximum rate of 0.5 ◦ C/h and sedation was tapered. Patients’ relatives were asked for informed consent according to the requirements of the National Committee for Ethics in Research (which approved the protocol). Patients regaining consciousness were also re-consented a posteriori. Only data sets with approved consents were considered in this study. Decisions to withdraw life support or to limit care were never taken by considering BIS values. Only clinically relevant parameters like absence of awakening after complete cessation of sedation, signs of brain death or early myoclonus or status epilepticus in combination with bilaterally absent N20 peak on somato-sensory evoked potentials or imaging findings compatible with irreversible brain damage were taken into account.21,22 2.2. Bispectral index monitoring The BIS (BIS XP Quatro, COVIDIEN, Mansfield, Massachusetts, USA) monitor was routinely applied to every patient by the nursing staff: application of the electrodes on the dry skin of the forehead according to the manufacturer’s instructions. Signal quality was immediately assessed by the treating nursing personnel, using the build-in tools, to obtain a stable and high signal quality index (SQI) and low EMG artefacts. BIS values were continuously registered via the patient monitoring system (IntelliVue, Philips, Böblingen, Germany), recorded in the patient data management system (Metavision, IMDsoft, Tel Aviv, Israel) and retrospectively exported as excel files into R. Throughout the 24 hour study period, only BIS values that fulfilled our quality criteria of low EMG artefacts (80) were retained for the analysis. Patients without a BIS monitoring or without BIS data meeting these quality criteria were excluded from the study. 2.3. Neurological evaluation The endpoint of this study was neurological outcome as defined by the cerebral performance category (CPC) score at 6 months,23 CPC 1-2 being considered as good outcome and CPC 3-5 being considered as poor outcome. CPC score 1 or 2 indicates no or minor neurological sequelae, 3-4 indicates severe neurological sequelae
1675
Table 1 Study population. Characteristic
Good outcome (CPC 1-2, n = 46)
Poor outcome (CPC 3-5, n = 50)
P value
Age, years Gender Male SAPS II Time to ROSC, min Initial rhythm VF/VT PEA Asystole Associated factors Shock AMI EEG Status epilepticus Seizures Medical history Tobacco Hypertension Heart failure Coronary disease Pulmonary disease Diabetes
57 (21-81)
67 (24-83)
0.023
42 (91%) 61 (43-83) 19 (4-60)
36 (72%) 71 (48-100) 30 (3-104)
0.016
