Select Guide Rating

Gases in Medicine Anaesthesia

By E.B. Smith, S. Daniels

The Royal Society of Chemistry

Copyright © 1998 The Royal Society of Chemistry
All rights reserved.
ISBN: 978-0-85404-718-5

Contents

BOC Centenary Lecture 1997 Anaesthesia in the 21st Century C. Prys-Roberts, 1,
New Anaesthetics Intravenous anaesthetics: the alternative to gases John W Sear, 11,
Entonox and its development J.A.S. Ross, I. L. Marr & M.E. Tunstall, 27,
PET scanning – What can it tell us about anaesthesia? P. Hartvig, J. Andersson, A. Wessen, M. Enlund, L. Wiklund, S. Valind & B. Langström, 42,
effects of nitrous oxide and halothane on somatosensory transmission A. Angel, R H. Arnott & S. Wolstenholme, 53,
Anaesthetic actions on fast synaptic transmission C.D. Richards, 86,
The actions of anaesthetics on voltage-gated and voltage-dependent ion channels K.T. Wann, 105,
The GABAA receptor: an important locus for intravenous anaesthetic action J.J. Lambert, D. Belelli, S. Shepherd, A.-L. Muntoni, M. Pistis & J.A. Peters, 121,
Anaesthetic effects on the spinal cord Joan J. Kendig, 138,
Historical Session Humphry Davy, Thomas Beddoes and the introduction of nitrous oxide anaesthesia E.B. Smith, 155,
William Morton and the early work on anaesthesia in the USA N.G. Coley, 163,
Objections to anaesthesia: the case of James Young Simpson C.A. Russell, 173,
The manufacture of anaesthetic nitrous oxide N2O – a study in technology blending W A. Campbell, 188,
Priestley Lecture 1997 Nitric oxide S. Moncada, 195,
Anaesthetics and Other Medical Gases Other gases used medically M.E. Garrett, 209,
Non-hypnotic effects of general anaesthesia C.S. Reilly, 217,
Interaction between general anaesthesia and high pressure S. Daniels, 225,
A genetic approach to understanding anesthesia S. Rajaram, B. Kayser, P.G. Morgan & M Sedensky, 234,
Closing Plenary Lecture Do we need new anaesthetic drugs? R M Jones, 251,
Subject Index, 255,

CHAPTER 1

BOC Centenary Lecture 1997 Anaesthesia in the 21st Century

Cedric Prys-Roberts

SIR HUMPHRY DAVY DEPARTMENT OF ANAESTHESIA, BRISTOL ROYAL INFIRMARY, BRISTOL, UK

Applying the premise that an oarsman steers by his wake, I intend to tackle my impossible task of predicting what might happen to anaesthesia in the whole of the next century, on the basis of forward extrapolation from our present state of knowledge. Given that extrapolation is potentially a statistically weaker process than interpolation, I shall limit that degree of extrapolation to no more than ten years into the 21st century. To extend my prediction beyond that limit would be to court ridicule let alone downright rejection. Our ability to interpret phenomena is no greater than the sum of our knowledge of the natural sciences which pertain to anaesthesia, surgery and other branches of medicine which we practice. Lord Kelvin expressed it perfectly [1]:

‘… when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it may be the beginning of knowledge, but you have scarcely, in your thoughts, advanced to the stage of Science whatever the matter may be.’

Thus restricting my crystal ball gazing, I shall attempt to use numbers such as we know them at present to justify any planned leaps into the dark. A reliable paradigm for such an extrapolation is what is known in control theory as ‘two step ahead adaptive control with exponential forgetting’ [2]. This wonderful piece of jargon gave me the key to presenting a scientifically valid prediction of where the science and chemistry of anaesthesia may be by the year 2010.

The broad mathematical principle of this method of adaptive control is that the variable to be predicted is measured at a given sample interval, determined by the precision of control required, and sequential values are submitted to a standard least squares regression, using the same sample interval, the next two values in the sequence are determined by extrapolation, and the accuracy of the prediction is governed by the residuals in the least squares calculation. As this prediction is made on-line, usually at intervals of about 30 seconds for human cardiovascular variables, in order to minimise the effects of distant historical values, these are ‘forgotten’ in an exponential manner, so that only relatively recent values influence the prediction [3,4]. We applied this process to the closed loop control of blood pressure using a stepping motor connected to the concentration control bezel of an isoflurane vaporiser This allowed the computer to determine the required concentration of isoflurane to maintain a predetermined blood pressure level, and allowed us to determine, in a completely unbiased way, the interaction of pre-dosing with clonidine, an α2- adrenoceptor agonist, in decreasing the required concentration of isoflurane [5]. This is a wonderful tool which could be applied to investigating the interaction of other drugs in decreasing the dose of an inhalational or intravenous anaesthetic required to maintain a given level of anaesthesia. As we can apply such control theory to the regulation of a physiological variable such as blood pressure or heart rate, both of which can be measured with a remarkable degree of precision (direct blood pressure to ±1 mm Hg during quiet spontaneous breathing; ± 2 mm Hg during IPPV) it is reasonable to ask whether we could control the delivery of other drugs, to produce a given state of anaesthesia. The existing control algorithm is certainly robust enough, and new ones controlled by artificial neural networks may solve any non-linearity problems. The problem does not lie in the control algorithms but, precisely as Lord Kelvin suggested, in that we cannot adequately measure anaesthesia; we find it difficult to express it in numbers. Will that be the major breakthrough by or at the start of the next century? As I look at the programme for this meeting I see considerable evidence of progress towards understanding how the gases and vapours which produce clinical anaesthesia act at membrane and synaptic sites, and in hippocampal slices, but I detect no attempt to express clinical anaesthesia in numbers.

General anaesthesia is an unnatural but reversible state of depression of the central nervous system of such a degree that consciousness is lost and that on recovery nothing is recalled relating to the period of anaesthesia [6]. In humans there is also depression of spontaneous ventilation, motor hypotonia [7], and suppression of physiological responses to noxious stimulation. But which of these functions are effects of anaesthetics on the brain, as opposed to the spinal cord or the motor neurons? Most of the evidence now suggests that the suppression of somatic motor responses to noxious stimulation is a function which is suppressed at a motor neuron level, in that increasing concentrations of anaesthetics hyperpolarize the motor neurones at a spinal level. The index of equipotency of inhalational anaesthetics, the minimum alveolar concentration (MAC) is independent of cerebral function [7,8]. A new set of correlations between the human MAC and various (lecithin, olive oil, benzene) solvent/gas partition coefficients [9] emphasise that the site of action may be related to neuronal cell membranes. It is now clear that to explain cerebral actions of anaesthetics (loss of consciousness and new learning) in such terms we need new MAC values for suppression of consciousness rather than for suppression of somatic motor action. Such data are difficult to come by and will require studies of large numbers of patients to provide complete concentration/effect curves, rather than relying on the rather small numbers on which traditional MAC values have been based [10]. It is fortuitous that the concentrations of both inhalational or intravenous anaesthetics required to suppress movement in response to noxious stimuli are considerably higher than those required to suppress consciousness and short term memory. Nevertheless, if we are to make any progress in the next century in the direction of automatic control of anaesthesia we must be able to correlate some measurable index of the state or level of anaesthesia with varying degrees of wakefulness, through sedation, to varying degrees of the anaesthetic state. Such correlations were achieved to a remarkable degree during induction of anaesthesia with five intravenous anaesthetics in volunteers by Schiittler and colleagues [11,12]. They infused the relevant drug at a constantly increasing rate until the subject lost consciousness and certain reflexes, at which point the infusion was turned off and the subject allowed to awaken. This process was repeated twice. They then correlated an index of neurophysiological activity, the median power frequency (MPF) of the electroencephalogram, with blood concentrations of the relevant anaesthetic. For propofol, for instance, the Cp50 (the blood concentration at which the MPF is decreased by 50%) was found to be 2.3 µg ml-1. Those studies were based on transient states of anaesthesia, with the inevitable hysteresis between the pharmacokinetic and dynamic responses of the subjects, both during increasing and decreasing drug concentration. What happens during quasi-steady state anaesthesia, in which the drug is infused according to a pharmacokinetic model to try and reach a stable state of anaesthesia?

Forrest and colleagues [13] maintained 72 patients in states varying from awake, through light to heavy sedation, to light and deep anaesthesia, by predetermined infusion schemes designed to achieve and maintain steady blood propofol concentrations over a range from 1 to 9 µg ml-1. At 30 minutes into such infusions there was a clear and strong correlation between the blood propofol concentration and the patient’s state of consciousness, and the eye-lash reflex; and also with the suppression of the EEG median power frequency. The minimum MPF during stable anaesthesia was 6 Hz, substantially higher than that recorded by Schüttler and his colleagues [12] although the Cp50 for EEG suppression (2.3 µg ml-1) was identical in both studies. A subsequent study [14] using a different approach in volunteers found a Cp50(awake) for propofol of 2.69 ± 0.56 µg ml-1.

Based on their foregoing studies the Bonn group [15] chose an MPF of 3 Hz as the set point for closed-loop of a propofol infusion to maintain surgical anaesthesia. Based on the stable infusion studies [13] it would seem that this would ensure a more than adequate level of anaesthesia, indeed they would be overdosing the patients by at least 30%. You simply do not need a computer controlled infusion system to overdose patients by that margin. It can be done just as effectively and much cheaper with a simple manually controlled step infusion scheme [16].

This is a good example of progress being held up by the application of information derived from inappropriate studies. One cannot fault the original studies [11,12], indeed they were innovative and brought considerable new insights into the relationships between the pharmacokinetics and dynamics of intravenous anaesthetics. The many subsequent studies of closed-loop control by the same group [15 is a good example] were also innovative, and the control programme was properly applied, but the chosen set point resulted in systematic overdosing and a potentially excessive level (depth) of anaesthesia. One has to ask why the market has not been inundated with commercially available devices to control anaesthesia.

Another example of this problem concerns the application of mid-latency auditory evoked potentials (MLAEP) as an indicator of so-called depth of anaesthesia [17-19] and as a controlled variable for closed-loop control of anaesthesia. There is no doubt that the latency of the MLAEP increases and the amplitude decreases, but there have been few studies which adopt critical statistical estimates of the selectivity and sensitivity of the various parameters derived from the MLAEP as a means of classifying correlations between such parameters and the state of the patient’s consciousness or, for that matter, indices of suppression of the physiological response to noxious stimulation during anaesthesia. There has also been a good deal of publicity, but regrettably few published data, about the control of anaesthesia using very simple computers to control drug infusions on some derived index from the auditory evoked potential. Based on figures in published articles one might imagine that it should be easy to derive a value for the shift of the latency, but therein lies the fallacy. It is very easy to publish a figure showing an almost perfect evoked potential, whatever that may be, but it is a quite different order of difficulty to persuade such a potential to repeat itself, from one period to the next, and one minute to the next with any consistency. It can be done by very complex digital filtering of the original signal – the main problem is that we are dealing with a microvolt signal in a forest of millivolt bushes and larger trees.

Tooley and colleagues [20] applying such digital filtering techniques to evoked potentials derived from groups of patients similar to those in the Forrest study [13], showed a ‘best correlation’ between the latency of the Nb wave and blood propofol concentrations during stable anaesthesia, with a threshold for loss of consciousness of 53 ms, a specificity of 96% and a sensitivity of 100%. Clearly there is a correlation, but how much confidence can we have in the values that we measure? In other words we have to modify Kelvin’s dictum and introduce the concept that it simply isn’t good enough to express anaesthesia in numbers – we must be able to do so with confidence, and that is the beginning of statistical science.

When we come to correlations between neurophysiological signals and conscious or anaesthetised states, the problem which arises is with boundaries. When we define good or adequate anaesthesia, we must be able to differentiate such a state from bad anaesthesia, during which the patient may be aware. Can we use current neurophysiological signals to aid such differentiation? Is there further information that can be derived from signals such as the electroencephalogram which can enable us to better classify the patient’s descriptive state?

The EEG is basically chaotic. It is a stochastic process, a set of random variables which unfold in time, strictly stationary only if all its statistical properties are invariable with time. Fourier analysis, widely used to analyse the complex frequency spectrum, is not strictly appropriate for a stochastic or truly random variable. There are three popular models for analysis of stochastic data in the time domain, autoregressive, or moving average, or a combination of both. Unlike Thomsen and colleagues [21] who also used autoregressive modelling, we allowed the EEG data from [13] to be analysed by a multi-layer neural network [22]. Bayes’s theorem enables the classification of events according to conditional probabilities. Such classifications can be determined by computer cells called perceptrons which are programmed to classify events by forming decision boundaries based on statistical probabilities [23]. Thus a modern computer perceptron can be considered as a discriminant function for a two class problem. Multi-dimensional or non-linear problems can be classified by using multiple layers of perceptrons – a so called neural network [24].

The EEGs from our digitised database [13] were analysed by a Multi Layer Perceptron developed by Holt and Tarassenko in the Department of Engineering Science in Oxford. Their multi-layered perceptron was trained on examples of the awake EEG, and was subsequently able to classify three stages of natural sleep in another database of sleep EEGs [22]. Based on the same training samples, the perceptron was able to classify the EEG of our sedated or anaesthetised patients, with a high degree of probability on to quite different mappings to those of natural sleep. Thus our ability to use statistical processes to better predict the state of anaesthesia in our patients may lead to safer, more efficacious, and economical anaesthetic practice in the early years of the next century.

Do these predicitive methods allow us any glimpses into the future of new anaesthetics? Do we need any new anaesthetics or adjuvant drugs? We could manage quite effectively with what we have at present. We can give almost perfectly satisfactory anaesthesia and post-operative pain relief with the drugs that we have – the main adverse outcome for the majority of patients is the prevalence of nauusea and vomiting in the post-operative period. It must be clear from what I have said up to now that I believe that any major developments in anaesthesia during the next ten to twenty years will be the result of new computer technology. This will not only improve the standards by better controlling the delivery of drugs to achieve improved efficacy, greater safety, and above all greater economy; but it will also give us all better insights into the mechanisms of anaesthesia, which in turn may also improve our practice of anaesthesia.

CHAPTER 2

Intravenous Anaesthetics: the Alternative to Gases

John W. Sear

JOHN RADCLIFFE HOSPITAL, OXFORD, UK

The provision of clinically adequate general anaesthesia can be attained by either inhaled agents or intravenous hypnotic agents given alone or in combination with nitrous oxide.

The onset of action of the inhaled agents depends on their inspired concentration, alveolar ventilation and cardiac output; while offset depends on the alveolar concentration and the latter two factors as well as the extent of metabolism (particularly so in the case of halothane). One of the problems with all anaesthetics is that when used alone those concentrations needed to maintain hypnosis may cause significant cardiorespiratory depression; as well as other organ and drug specific effects such as altered cerebral blood flow regulation; altered hepatocyte integrity; fluoride-induced nephrotoxicity; uterine relaxation; risk of seizures; malignant hyperpyrexia; and impaired cellular immunity. Because of these concerns, there has been a trend towards the development and evaluation of new agents having improved pharmacological profiles. One obvious possibility is intravenously administered drugs.

(Continues…)Excerpted from Gases in Medicine Anaesthesia by E.B. Smith, S. Daniels. Copyright © 1998 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.