The prognostic value of bispectral index and suppression ratio monitoring after out-of-hospital cardiac arrest: a prospective observational study

A prospective, observational study was performed between March 2011 and May 2015 in a Belgian tertiary care hospital (Ziekenhuis Oost-Limburg, Genk). All adult comatose survivors successfully resuscitated from OHCA with a presumed cardiac cause were admitted to the Coronary Care Unit (CCU) and were considered as eligible for this study. According to the institutional post-CA care protocol, all patients were treated uniformly and BIS monitoring was started immediately after admission to CCU [19]. Approval from the local Committee for Medical Ethics was obtained prior to study onset (11/06) and written informed consent was obtained from patient’s next of kin.

The institutional post-resuscitation protocol has been described previously [19, 20]. To summarize, TTM at 33 °C was initiated immediately after arrival at the emergency department by administering cold saline intravenously (4 °C—15–30 ml/kg). Urgent coronary angiography was performed followed by a percutaneous coronary intervention when indicated. TTM was further mechanically induced after CCU admission and maintained at 33 °C for 24 h using either endovascular (Icy cathether, CoolGard^® 3000, Alsius, Irvine, CA, USA) or surface (ArcticGel™ pads, Arctic Sun^® 5000, Medivance, Louisville, Colorado, USA) cooling systems. Subsequently, patients were rewarmed over the following 12 h (0.3 °C/h). Both cooling systems were equipped with a feedback loop system to control target temperature using an oesophageal temperature probe. All patients were intubated, mechanically ventilated and sedated with propofol, midazolam and remifentanil. Within the period of TTM, sedation doses were titrated to obtain values between − 3 and − 5 on the Richmond Agitation-Sedation scale. According to current guidelines, cisatracurium was only administered in case of shivering [18]. After the return to normothermia, sedation was reduced to evaluate patients’ neurological status properly. Patients not ready for extubation owing to circulatory or respiratory issues or due to persisting coma were kept sedated under the lowest dose needed to tolerate the endotracheal tube. EEGs were performed on clinical indication, and antiepileptic drugs were given in case of epileptic activity. Every EEG was characterized by a description of the posterior dominant rhythm (or absence thereof) and amplitude as well as the presence of non-dominant rhythms. Lateralization and the presence of artefacts was described where applicable. If present, epileptic activity was described as interictal, ictal or as status epilepticus. A status epilepticus was defined as continuous epileptic activity including rhythmic focal or generalized spikes lasting for more than 5 min and was considered as therapy-refractory in case of persistent epileptic activity in spite of at least two lines of antiepileptic drugs. Once the neurological, hemodynamic and respiratory status had been recovered sufficiently, patients were extubated.

Bilateral BIS monitoring was initiated as soon as possible after CCU admission using the BIS VISTA™ (Aspect Medical Systems, Inc. Norwood, USA). A six-electrode frontotemporal bilateral sensor was applied to the patient’s forehead. The BIS monitor is a simplified EEG monitor that uses fast Fourier transformation to convert raw frontal EEG signals into simple and real-time BIS numbers ranging from 0 (iso-electric EEG) to 100 (normal electrical activity in awake subjects). Additionally, a SR is calculated, representing the percentage of each 63-s period that is iso-electric [21]. Finally, EMG activity is another parameter calculated by the BIS monitor describing the electromyographic content of the EEG signal and ranges from 0 (no EMG activity) to 100 (large EMG activity). BIS, SR and EMG values were continuously recorded and stored during the hypothermic and rewarming phase (i.e. 36 h). Although treating physicians (cardiologists) were not blinded to values displayed on the BIS monitor, decisions to withdraw life support were never based on these parameters and the dosage of sedatives was not titrated based on BIS values.

In patients not regaining consciousness despite complete cessation of sedation, maximal supportive treatment was provided until at least 72 h after normothermia was reached. Together with clinical neurological examination, malignant EEG patterns (i.e. status epilepticus, persisting burst suppression rhythms or long-lasting cerebral inactivity) were considered as the first-line support for the decision to withdraw therapy, and in specific, when seizures remained therapy-refractory or if other EEG rhythms with a poor prognosis remained present on subsequent EEG assessments. In case subsequent EEGs were inconclusive or when patients failed to wake up after the end of active temperature control, SSEPs and/or brain CT were performed. In line with international guidelines, absent corneal and pupillary reflexes, bilateral absence of the N20 component of SSEPs and refractory epileptic activity was considered to support this decision to withdraw therapy [22].

The primary endpoint was neurological outcome defined by the Cerebral Performance Category (CPC) at 180 days post-CA. According to the scale classification, CPC1 is indicative for good cerebral performance; CPC2 implies moderate disability with sufficient cerebral functioning for independent daily-life activity; CPC3 indicates severe neurological sequelae; CPC4 implies coma or vegetative state; and CPC5 stands for death [23]. In this study, a CPC1–2 and CPC3–5 was considered as good and poor neurological outcome, respectively.

Statistical analysis was performed using SPSS version 22.0 (SPSS Inc, Chicago, USA). Equal distribution was tested with a Kolmogorov–Smirnov test. Depending on normality, categorical data were compared between patients with a good and poor neurological outcomes using Fisher exact or Chi-Square tests, while unpaired t tests or Mann–Whitney U tests were used to compare continuous data. BIS, SR and EMG values were stored per second and left and right values were averaged. Mean BIS, SR and EMG values were then calculated per hour from initiation of TTM onwards and were used for data analysis. To assess the predictive ability of BIS and SR, receiver operating characteristic (ROC) curves were constructed at each hour using mean BIS and SR values and the Youden index was calculated for each ROC curve (Youden index = sensitivity + specificity – 1). The optimal time point and threshold to predict poor neurological outcome was determined based on the highest Youden index across all ROC curves. Relative risk ratios of poor neurological outcome were computed for the presence of single BIS and SR values between given intervals on the respective optimal time points. In addition, regression curves were fitted (including the calculation of the Pearson correlation coefficient) to describe the relationship between mean EMG and BIS values below and above our calculated (optimal) BIS threshold. Survival analyses were executed using Kaplan–Meier curves and log-rank statistic. p values < 0.05 were considered as significant.