Incidence of Slow Delta EEG Frequencies (0.1-1 Hz) Before Burst Suppression

1. BACKGROUND The anaesthetic drugs given to induce general anaesthesia act on the brain and produce characteristic changes in the electroencephalogram (EEG) signal recorded non-invasively from the frontal scalp area. In particular, oscillations in the delta (around 1 - 4 Hz) and alpha (8 -12 Hz) frequency bands are generally considered as markers of adequate anaesthesia dosing and hence unconsciousness, and disappear as the patient returns to consciousness. A further easily recognisable pattern in the EEG is where periods of high-amplitude bursts alternate with periods of low amplitude suppression (a nearly flat EEG signal) - so-called burst suppression. This pattern occurs at deep stages of general anaesthesia and is thought to represent a brain state response to excessive anaesthesia doses. Burst suppression occurs in around one third of all patients undergoing general anaesthesia, but can also reach much higher incidence rates such as 70% in older patients, and will depend largely on clinical dosing practice and cohort age. In research led by a group from Harvard University, there have been claims that prior to the onset of burst suppression, slow wave activity in the EEG (which occurs at frequencies between 0.1 and 1 Hz) increases in amplitude (and in relationship to higher frequency oscillations), and that 'down states' in the slow EEG signal can act as indicators of upcoming burst suppression. Unfortunately, the two most commonly used commercial EEG anaesthesia devices used in Switzerland and worldwide, the Patient State Index (PSI by SedLine® Masimo) and the Bispectral Index™ (BIS™ by Medtronic), both remove these slower waveforms by high pass filtering so that they cannot be observed in the EEG signal visually on the monitor, and are absent in the EEG signal extracted for subsequent analysis. It is currently not known how consistently the slow wave occurs during general anaesthesia, and if this waveform is consistently present prior to the onset of burst suppression, and if it can act as advance warning for impending burst suppression. This observational study will test the feasibility of measuring these slow frequency waveforms and examining their relationship to burst suppression. For recording these slow waves between 0.1 and 1 Hz, that cannot be detected with a standard clinical anesthesia EEG monitor, the investigators plan to place an ULTEEMNite device (by CSEM Neuchâtel, Switzerland) immediately beside the electrodes strip of the commonly used Sedline EEG monitor on the foreheads of older patients undergoing general anaesthesia and to record the EEG signals from both devices simultaneously. The placement of the medically approved commercial Sedline device is standard clinical practice at the Cantonal Hospital of Aarau and the Hirslanden Clinic Aarau, and any clinically decision referring to titration of the depth of anesthesia will be made according to this device. The placement of the ULTEEMNite device will allow the investigators to record EEG signals that include the slow wave activity between 0.1 and 1 Hz. To note is that this study is independent from the ULTEEMNite device - the investigators do not wish to evaluate this device, but only use it as a recording device. The investigators are only interested in the information contained in slow wave activity between 0.1 and 1 Hz.
2. PROJECT OBJECTIVES AND DESIGN 2.1 Hypothesis and primary objective The hypothesis of this study is that EEG slow wave amplitudes and slow frequency spectral power will become maximal in the period immediately prior to the onset of burst suppression in the EEG recorded from the ULTEEMNite device. More specifically, power will be analysed in the EEG frequency spectra in the ranges from 0.1 to 1 Hz to confirm or disconfirm the presence of slow wave EEG activity. The investigators will use the clinically validated burst suppression algorithm of the Sedline monitor to provide a reproducible and objective measure of exact burst suppression onset. 2.2 Primary and secondary endpoints The primary variables of interest are peak amplitudes in the slow wave signal and absolute power in the EEG frequency spectra when recorded from immediately prior to the onset of burst suppression (as measured by the Sedline monitor). 2.3 Project design This project is an observational study and requires EEG recordings from 30 patients. All measurement and analysis of EEG signals will occur retrospectively using Matlab software (MathWorks Inc.). In clinical practice at the Cantonal hospital Aarau and the Hirslanden Clinic Aarau it is standard to have an EEG monitor (such as the Sedline) in place, and this monitor also allows for simple extraction of the raw EEG signal after the operation. EEG analysis will be retrospective, and the EEG from both devices will be displayed visually to check for an adequate connection and generally reasonable signal quality. Any clinical decision making will be solely based on the EEG signal shown by the standardly used commercial Sedline device - as it is done in the standard of care. The ULTEEMNite device records and stores the EEG data directly and can be easily downloaded for subsequent analysis. Because patient age has an impact on EEG amplitudes (with older patients having smaller EEG amplitudes), age will be recorded and be part of the final analysis.
3. PROJECT POPULATION AND STUDY PROCEDURES 3.1 Project population, inclusion and exclusion criteria In this project the investigators plan to recruit 30 patients undergoing elective surgery requiring general anaesthesia. General anaesthesia will be provided by either volatile based ethers (such as sevoflurane) or by intravenous anaesthesia (propofol) as these lead to known and similar EEG signals. The choice of thirty patients will allow the investigators to account for some minor differences in EEG characteristics between patients. In general, patients over the age of 60 are much more likely to show burst suppression in their EEG signal during anaesthesia at some point compared to younger patients, which is why we plan to only include older (aged 60 or above) patients

3.2 Recruitment, screening and informed consent procedure The recruitment will take place at the Cantonal Hospital of Aarau and the Hirslanden Clinic Aarau. Each patient will be informed that the participation in the project is voluntary and that the patient may withdraw from the study at any time and that withdrawal of consent will not affect his or her subsequent medical assistance and treatment. The patients will be informed that their medical records may be examined by authorized individuals other than their treating physician. The formal consent of a patient, using the approved consent form, will be obtained before the patient is submitted to any study procedure.

The consent form will be signed and dated by the investigator or their designee at the same time as the patient signs.

3.3 Study procedures It is expected that the project recruitment will require a duration of around six months. For each patient who chooses to participate there are only two 'visits' (Preoperative Assessment Clinic and general anaesthesia). Those 'visits' are part of standard clinical practice and are independent from this research project.

When a participant chooses to take part in the project, the EEG electrodes will be attached to the forehead while the patient is awake prior to induction of anaesthesia. The Sedline electrodes (wet electrodes) are placed on the forehead as per standard clinical practice after the skin is cleaned with an alcohol-based wipe. The ULTEEMNite dry electrodes will also be placed directly beside the Sedline electrode strip, and the dry electrodes that simply sit on the skin (see figure 1) are held in place using a stretchable fabric band. The EEG is displayed on the Sedline monitor for clinical use as per standard clinical practice.

Following the end of general anaesthesia, the EEG devices will be removed when the patient has become responsive to anaesthetist command.

4 STATISTICS AND METHODOLOGY The statistical plan will be descriptive in nature, focusing on the amplitudes and power spectral power values of the signal at different frequencies over time within patients. The investigators will show basic population averages of differences and their variation (in the form of 95% confidence intervals, or interquartile ranges as appropriate). Although there are no power analyses completed for this study, 30 patients should be an adequate number to feasibly gauge the presence or absence of slow wave activity prior to burst suppression.