Study of the cerebral effects of sevoflurane, propofol and remifentanil as measured by the spontaneous electro-encephalogram

Multiple electroencephalographically derived indices have been developed to
measure the cerebral hypnotic drug effect during anesthesia, using a variety of
mathematical algorithms such as bispectral index, spectral entropy and spectral
edge frequency. The complexity of the raw EEG is reduced to -an easy to
interpret- number. It varies generally between 100 (fully awake patient) to 0
(an excessively sedated patient). The anesthesiologist adjusts his dosing
scheme to target a number between a predefined range. (e.g. between 40 and 60)

These monitors are currently solidly integreted in clinical practice although
they keep being hampered by several limitations. The most important problem is
that they are not extracted from a direct neuro-physiological phenomenon that
is known to be closely related to loss and return of consciousness, rather they
have a probabilistic nature, indicating whether your probability of
responsivenessis is high or low. You are never sure which EEG phenomenon
relates to the index. Also, the number that relates to loss of consciousness
rarely is the same as the number that indicates return of consciousness, which
decreases the predictive value during recovery of anesthesia. Additionally,
the dose response relationship differs on multiple parameters between each
monitors. As such the performance of one monitor cannot be extrapolated to
another. Finally, although the detection capacity of responses to verbal
command is fairly good, EEG extracted numbers perform worse on correlation with
movement after a noxious stimulus.

Most problems are related to the limited understanding of the relationship
between the changes in EEG and the underlying (insufficiently understood)
neurophysiological mechanism that evokes (un)consciousness during anesthesia.
Secondly, at the time of development of most monitors, the methods of drug
administration were less reproducible to allow a more rational drug titration
in a population of different demographics. The last two decades, major progress
has been made on both issues. Therefore, new insights in neuro-physiology and
better drug titration systems opens new perspectives to improve EEG derived
data extraction.

Recently, Mashuire et al found a typical EEG characteristic that is correlated
consistently with the loss and return of responsiveness, independent of the
anesthetic used (inhalation or intravenous) and independent of the species
tested (rats and humans). These findings do suggest that new information can be
extracted from the raw EEG that has a much closer connection with the essential
neurophysiological processes involved to evoke consciousness/unconsciousness or
responsiveness/unresponsiveness. In this study we want to collect data that
allows us to recognize these patterns on the EEG as a better measurement of
consciousness/unconsciousness.

Additionally, we have the ability to use (clinically availlable) target
controlled infusion techniques for propofol (plasma- and/or effect-site
controlled) and end-tidal titrated sevoflurane (through a Zeus Ventilator
(Draeger)) If we titrate our hypnotics through these more advanced and
pharmacologically more rational titration methods, we may detect a more
relevant correlation between the dose given and the EEG behavior. By adding
remifentanil, we will also explore the alteration of performance of EEG derived
information during interaction with opioids.

The obtained data may result in a breakthrough in the methodology to titrate
anesthesia in a more predictable and reproducible way, because the extracted
information relates closer to a neuro-physiological process related to
(un)responsiveness. Moreover, the index may produce more consistent results
whether inhaled anesthetics or intravenous drugs are given.

Our first goal is collecting high quality raw EEG waves, - measured
simultaneously on multiple locations of the brain - during a pharmacological
reproducible anesthesia. The goal is to observe EEG patterns that allow the
development of technology to monitor anesthesia drug effect more effectively.
As we use the same neural network as a benchmark (all healthy volunteers will
receive all 4 anesthesia regimens, we can detect typical EEG characteristics
that are evoked by either sevoflurane (inhaled anesthesia) or propofol
(intravenous anesthesia) with or without the addition of remifentanil.
(powerful pain killer)
Our methodology allows to describe the following dose response relationships:
1) The dose response relationship in an identical population of healthy
volunteers between raw EEG effects at one hand and the responsiveness to
noxious and non-noxious stimuli at the other hand evoked by steady state
propofol concentrations.
2) The dose response relationship in an identical population of healthy
volunteers between raw EEG effects at one hand and the responsiveness to
noxious and non-noxious stimuli at the other hand evoked by steady state
sevoflurane concentrations.
3) The dose response relationship in an identical population of healthy
volunteers between raw EEG effects at one hand and the responsiveness to
noxious and non-noxious stimuli at the other hand evoked by steady state
propofol concentrations with addition of a high or low dose of remifentanil.
4) The dose response relationship in an identical population of healthy
volunteers between raw EEG effects at one hand and the responsiveness to
noxious and non-noxious stimuli at the other hand evoked by steady state
sevoflurane concentrations with addition of a high or low dose of remifentanil.
5) In all upper situations the typical EEG characteristics compatible with loss
and return of consciousness will be explored. (We hypothesize that loss and
return of consciousness occur at very different drug concentrations but may
have comparable features in EEG behavior)
6) After a step up steady state induction towards unconsciousness and a step
down steady state recovery, a non steady-state bolus dose will be administered
to evaluate the hysteresis between bolus dose and EEG effect. Both steady
state and non steady state conditions need to be tested as they are a
reflection of respectively the maintenance and induction of anesthesia.
7) Based on the observed dose response phenomena, we want to develop an
optimized EEG derived index that meets the demands of the anesthesiologist
better compared to other contemporary monitoring systems
These seven goals imply the need for some interventions:
1) Use of adjusted drug titration technology for the administered molecules.
2) Blood sampling at multiple intervals (only for the sessions with propofol
and/or remifentanil) to adapt the predicted population concentration according
to the individual measured pharmacokinetic-dynamic behavior. For sevoflurane
this is not necessary as the end tidal concentration is already an
individualized measurement.
3) Simultaneous high quality EEG monitoring on multiple locations of the brain
4) At every steady state, we need to determine whether the patient responds to
the observers assessment of alertness and sedation scale.
5) At every steady state, we need to determine whether the patient responds to
a standardized noxious stimulus (tetanic electrical current for short
duration). This test is not performed when the patient is responsive to verbal
stimuli.
6) The availability of a pharmacometricist who knows how to use NONMEM
software. (Department of anesthesiology, UMCG)
7) The availability of signal processing engineers to develop and test multiple
algorithms for EEG data extraction. (Masimo research and development
department)
During all measurements, all hemodynamic and respiratory parameters (ECG, SpO2,
NIBP and ETCO2), as well as the non invasive Rainbow Acoustic Monitor (Masimo,
Irvine, CA, USA)) are stored on a computer with Automated Data Collection
software (ADC, Masimo Inc) en RUGLOOPII software (Demed, Temse, Belgium) for
further posthoc analysis.
All data (inclusive residual bloodsamples) will be stored in order to allow
future research that relates anesthetic effect with EEG response and humoral or
endocrine stress responses.

Prospective randomized and stratified observational study with healthy
volunteers.

ll volunteers receive 4 standardized anesthesia sessions with at least one week
recovery between each session. EEG characteristics are collected using a Masimo
multichannel EEG recorder and compared post hoc for drug dependent
characteristics. No off label drug use is applied in this study. All monitors
used are CE approved and clinically applied in common anesthesia practice.

In this study, we will administer anesthesia in conditions that are comparable
to clinical practice. The volunteers will be under the supervision of a
certified anesthesiologist at all times and all techniques and methods
described in this study are performed according to current state of the art
daily practice. The risks of canulations may be haematoma, infiltrations of
drug, embolisation of air and trombi, and flebitis. These are very rare
complications in a healthy volunteer population.
The arteriel line will be placed in the arteria radialis of the non dominant
hand under local anesthesia. It will be used for blood pressure measurements
and blood sampling. The total volume of blood taken has no clinical
consequences.
The small intravenous canulae will be introduced in a vein of the back of the
hand or forearm, which may cause a slight temporary pain. Patients receive
cristaloids through that canula. (Ringer Lactate 500ml).
All sensors used for vitale parameters are used routinely in clinical practice
and are CE certified. They do not evoke any risk for the patient.
If the study takes longer than three hours a bladder scan will be performed at
the end of the study and if deemed necessary, a single bladder catheterization
will be performed.