Schrödinger's cat: anaesthetised and not!
2017; Elsevier BV; Volume: 120; Issue: 3 Linguagem: Inglês
10.1016/j.bja.2017.11.068
ISSN1471-6771
Autores Tópico(s)Neural dynamics and brain function
ResumoHow does the brain reassemble consciousness and cognition upon exiting general anaesthesia? Traditionally, emergence from anaesthesia has been considered to be a mirror image of anaesthetic induction. Indeed, textbooks contain a single dose-response curve relating anaesthetic concentration to probability of wakefulness.1Miller R.D. Miller's Anesthesia.8th ed. Elsevier/Saunders, Philadelphia, PA2015Google Scholar, 2Longnecker D.E. Newman M.F. Mackey S. Sandberg W.S. Zapol W.M. Anesthesiology.3rd ed. McGraw-Hill Education, New York2017Google Scholar Nevertheless, several studies suggest that dose-response curves for induction and emergence are distinct.3McKay I.D. Voss L.J. Sleigh J.W. Barnard J.P. Johannsen E.K. Pharmacokinetic-pharmacodynamic modeling the hypnotic effect of sevoflurane using the spectral entropy of the electroencephalogram.Anesth Analg. 2006; 102: 91-97Crossref PubMed Scopus (66) Google Scholar, 4Kelz M.B. Sun Y. Chen J. et al.An essential role for orexins in emergence from general anesthesia.Proc Natl Acad Sci USA. 2008; 105: 1309-1314Crossref PubMed Scopus (251) Google Scholar, 5Friedman E.B. Sun Y. Moore J.T. et al.A conserved behavioral state barrier impedes transitions between anesthetic-induced unconsciousness and wakefulness: evidence for neural inertia.PLoS One. 2010; 5e11903Crossref PubMed Scopus (156) Google Scholar, 6Joiner W.J. Crocker A. White B.H. Sehgal A. Sleep in Drosophila is regulated by adult mushroom bodies.Nature. 2006; 441: 757-760Crossref PubMed Scopus (340) Google Scholar As with Schrödinger's cat that is simultaneously both alive and dead, this dissociation of anaesthetic induction and emergence curves implies that at the same anaesthetic concentration individuals can be either awake or anesthetized (Fig. 1A). In other words, entry into and exit from the anaesthetic state is characterized by hysteresis. Two recent studies test the existence of hysteresis in humans with mixed results.7Kuizenga M.H. Colin P.J. Reyntjens K.M. et al.Neural inertia in humans during general anaesthesia: fact or fiction?.Br J Anaesth. 2018; 120: 525-536Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar, 8Warnaby C.E. Sleigh J.W. Hight D. Jbabdi S. Tracey I. Investigation of slow-wave activity saturation during surgical anesthesia reveals a signature of neural inertia in humans.Anesthesiology. 2017; 127: 645-657Crossref PubMed Scopus (44) Google Scholar Hysteresis describes history-dependent phenomena and is commonly observed in nature. It is familiar to physicians plotting pulmonary pressure-volume loops, engineers examining stretching and recoiling of rubber bands, and physicists magnetizing and demagnetizing ferromagnetic materials. The area bracketed by a hysteresis loop often denotes a meaningful quantity such as the work of breathing, energy spent, or heat lost. Hysteresis is intentionally incorporated into modern control systems such as thermostats and Schmitt triggers to prevent unwanted and otherwise frequent state switching. In short, hysteresis can and does occur in a variety of systems. In the setting of anaesthesia, hysteresis might imply neural inertia–an intrinsic resistance to transitions between unconscious and conscious states. If resistance to state transitions were too small, frequent state switching would predispose individuals with collapsed neural inertia to awareness under anaesthesia. Conversely, if neural inertia were amplified, individuals could get stuck in an anesthetized state and exhibit delayed emergence. Despite its theoretical significance, evidence supporting or refuting neural inertia in humans has been lacking. Hence, we were very gratified to see two recent studies investigating neural inertia in humans, one by Kuizenga and colleagues in this issue of the British Journal of Anaesthesia7Kuizenga M.H. Colin P.J. Reyntjens K.M. et al.Neural inertia in humans during general anaesthesia: fact or fiction?.Br J Anaesth. 2018; 120: 525-536Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar and one by Warnaby and colleagues in Anesthesiology.8Warnaby C.E. Sleigh J.W. Hight D. Jbabdi S. Tracey I. Investigation of slow-wave activity saturation during surgical anesthesia reveals a signature of neural inertia in humans.Anesthesiology. 2017; 127: 645-657Crossref PubMed Scopus (44) Google Scholar In a seemingly disappointing blow, evidence for neural inertia was underwhelming—observed only under some circumstances. In this issue, Kuizenga and colleagues conducted a formidable trial in 36 healthy subjects who each volunteered to receive four different anaesthetic exposures in randomized order: propofol, sevoflurane, propofol with remifentanil, or sevoflurane with remifentanil. Depth of anaesthesia was assessed using a sedation scale and responses to noxious electrical stimuli together with electroencephalogram (EEG)-based parameters such as spectral edge and Patient State Index (PSI). A major methodological strength of the Kuizenga study is its actual measurement of plasma propofol and remifentanil concentrations. Nevertheless, as is true for all target-controlled infusion studies, the authors must assume that plasma and effect-site concentrations are in equilibrium in order to build a meaningful pharmacodynamic model. Evidence supporting neural inertia was found for both behavioural and EEG measures with some sevoflurane-based regimens but intriguingly, not with propofol. Using a measure of EEG power in the 0.5–1.5 Hz bandwidth, termed slow wave activity saturation (SWAS), Warnaby and colleagues analysed data from four different anaesthetic trials involving nearly 400 subjects.8Warnaby C.E. Sleigh J.W. Hight D. Jbabdi S. Tracey I. Investigation of slow-wave activity saturation during surgical anesthesia reveals a signature of neural inertia in humans.Anesthesiology. 2017; 127: 645-657Crossref PubMed Scopus (44) Google Scholar Warnaby's cohorts spanned both volunteers exposed only to propofol or sevoflurane as well as patients receiving real world mixtures of volatile anaesthetics, propofol, opioids, and muscle relaxants. Volunteers in the Warnaby study who received only propofol also failed to show overt behavioural evidence of neural inertia. Nevertheless, SWAS was asymmetric during induction and emergence, leading Warnaby and colleagues to suggest that SWAS is a signature of neural inertia in humans. What do these studies mean for the neural inertia hypothesis? In every human experiment, the issue of resistance to state transitions is conflated with pharmacokinetic-pharmacodynamic (PK/PD) modelling.9Louizos C. Yanez J.A. Forrest M.L. Davies N.M. Understanding the hysteresis loop conundrum in pharmacokinetic/pharmacodynamic relationships.J Pharm Pharm Sci. 2014; 17: 34-91Crossref PubMed Google Scholar As direct brain measurement of anaesthetic concentration is not feasible in humans, a hypothetical effect-site compartment has been incorporated into PK/PD models to account for the delay between changes in plasma drug concentration and observed anaesthetic effects such as changes in the EEG.10Schnider T.W. Minto C.F. Shafer S.L. et al.The influence of age on propofol pharmacodynamics.Anesthesiology. 1999; 90: 1502-1516Crossref PubMed Scopus (833) Google Scholar Because the effect-site concentration cannot be measured directly, the rate of equilibration of the effect-site is typically adjusted such that the observed hysteresis is minimized. This tautological argument has been used to assert that hysteresis is merely a consequence of the equilibration time constant between the plasma and effect-site concentrations. While slow equilibration between plasma and effect-site concentration could and likely does contribute to hysteresis, several observations argue that hysteresis is not likely due to pharmacokinetic mechanisms alone. First, in animals, anaesthetic concentration in the brain—the target organ for hypnosis—has been measured using robust analytical chemistry techniques. The second line of reasoning comes from the observation of concentration-response curves themselves. Indeed, the consistent rightward shift in EC50 for induction relative to emergence could in principle be explained by equilibration with the effect-site. It is more challenging to explain why across species and anaesthetic agents the Hill slopes for emergence are consistently smaller than those for induction3McKay I.D. Voss L.J. Sleigh J.W. Barnard J.P. Johannsen E.K. Pharmacokinetic-pharmacodynamic modeling the hypnotic effect of sevoflurane using the spectral entropy of the electroencephalogram.Anesth Analg. 2006; 102: 91-97Crossref PubMed Scopus (66) Google Scholar, 5Friedman E.B. Sun Y. Moore J.T. et al.A conserved behavioral state barrier impedes transitions between anesthetic-induced unconsciousness and wakefulness: evidence for neural inertia.PLoS One. 2010; 5e11903Crossref PubMed Scopus (156) Google Scholar, 11Joiner W.J. Friedman E.B. Hung H.T. et al.Genetic and anatomical basis of the barrier separating wakefulness and anesthetic-induced unresponsiveness.PLoS Genet. 2013; 9e1003605Crossref PubMed Scopus (56) Google Scholar (Fig. 1A). Yet, the strongest evidence that anaesthetic hysteresis is not explained by pharmacokinetics arises from genetic studies. Several single gene mutations in fruit flies dramatically modify the hysteresis loop without affecting pharmacokinetics.5Friedman E.B. Sun Y. Moore J.T. et al.A conserved behavioral state barrier impedes transitions between anesthetic-induced unconsciousness and wakefulness: evidence for neural inertia.PLoS One. 2010; 5e11903Crossref PubMed Scopus (156) Google Scholar, 11Joiner W.J. Friedman E.B. Hung H.T. et al.Genetic and anatomical basis of the barrier separating wakefulness and anesthetic-induced unresponsiveness.PLoS Genet. 2013; 9e1003605Crossref PubMed Scopus (56) Google Scholar A straightforward pharmacodynamic change that globally altered anaesthetic sensitivity would result in parallel changes to both induction and emergence curves, but would leave the hysteresis loop unchanged. Thus, mutations affecting hysteresis must differentially affect induction or emergence. In this regard, a specific mutant, sssP1, is particularly intriguing. Deletions of the SLEEPLESS protein, encoded by the sss gene, cause a dramatic collapse in the hysteresis. Remarkably, selective restoration the SLEEPLESS protein in neurones (but not in glia) restores the anaesthetic hysteresis loop (Fig. 1B). There is no plausible mechanism through which SLEEPLESS protein specifically restored in neurones affects equilibration kinetics with the hypothetical effect-site. Single gene mutations have also been shown to affect anaesthetic hysteresis in mammals. Genetic disruption of orexin/hypocretin signalling in the brain does not affect induction of anaesthesia but impairs emergence.4Kelz M.B. Sun Y. Chen J. et al.An essential role for orexins in emergence from general anesthesia.Proc Natl Acad Sci USA. 2008; 105: 1309-1314Crossref PubMed Scopus (251) Google Scholar Moreover, genetic deletion of adrenergic ligands specifically from the brain exaggerates the anaesthetic hysteresis loop in mice.5Friedman E.B. Sun Y. Moore J.T. et al.A conserved behavioral state barrier impedes transitions between anesthetic-induced unconsciousness and wakefulness: evidence for neural inertia.PLoS One. 2010; 5e11903Crossref PubMed Scopus (156) Google Scholar Hence, to distinguish from the term hysteresis that has been historically reserved to describe effect site equilibration kinetics, the notion of "neural inertia" as an intrinsic behavioural state barrier that retards transitions to and from states of unconsciousness was conceived.5Friedman E.B. Sun Y. Moore J.T. et al.A conserved behavioral state barrier impedes transitions between anesthetic-induced unconsciousness and wakefulness: evidence for neural inertia.PLoS One. 2010; 5e11903Crossref PubMed Scopus (156) Google Scholar Neurophysiological experiments in mammals strongly support the notion of neural inertia. Hudson and colleagues12Hudson A.E. Calderon D.P. Pfaff D.W. Proekt A. Recovery of consciousness is mediated by a network of discrete metastable activity states.Proc Natl Acad Sci USA. 2014; 111: 9283-9288Crossref PubMed Scopus (107) Google Scholar recorded local field potentials in the cortex and thalamus of rats anaesthetized with a fixed concentration of isoflurane (Fig. 1C). At every individual steady-state concentration, the characteristics of local field potentials fluctuated between several distinct, individually-stabilized activity patterns. These fluctuations occurred despite fixed anaesthetic concentrations and persisted for more than 1 hr—an order of magnitude longer than any model for effect-site equilibration. In addition to the Hudson study, quasi-cyclic fluctuations in brain states resembling transitions between sleep stages have been described for anaesthesia induced with urethane in rodents.13Clement E.A. Richard A. Thwaites M. Ailon J. Peters S. Dickson C.T. Cyclic and sleep-like spontaneous alternations of brain state under urethane anaesthesia.PLoS One. 2008; 3e2004Crossref PubMed Scopus (248) Google Scholar Every pharmacokinetic/pharmacodynamics (PK/PD) model eventually settles into a steady state and thus cannot account for such oscillations. Furthermore Ishizawa and colleagues showed abrupt transitions in cortical activity patterns under steady-state propofol concentration in primates.14Ishizawa Y. Ahmed O.J. Patel S.R. et al.Dynamics of propofol-induced loss of consciousness across primate neocortex.J Neurosci. 2016; 36: 7718-7726Crossref PubMed Scopus (41) Google Scholar While Chander and colleagues used EEG acquired in a clinical setting and thus not steady-state, they nevertheless did observe abrupt fluctuations in EEG under anaesthesia.15Chander D. Garcia P.S. MacColl J.N. Illing S. Sleigh J.W. Electroencephalographic variation during end maintenance and emergence from surgical anesthesia.PLoS One. 2014; 9e106291Crossref PubMed Scopus (66) Google Scholar The PK/PD explanation of neural inertia makes a fundamental assumption—there is a one-to-one correspondence between drug concentration in the effect-site and depth of anaesthesia as measured by some feature of brain activity. In other words, if one knew the concentration at the effect-site, one would know the state of the brain unambiguously and vice versa. This assumption is necessary to justify the calculation of the concentration at the hypothetical effect-site based on brain activity measured by EEG for instance. However, even quite simplified mathematical models of cortical circuits are generically multi-stable (Fig. 1D). They exhibit one of several distinct activity patterns depending on the previous history without any changes in defining parameters. In fact, multi-stability is a generic property of nonlinear dynamical systems16Strogatz S.H. Nonlinear dynamics and chaos: with applications to physics, biology, chemistry, and engineering.2nd ed. Westview Press, Boulder, CO2015Google Scholar of which the brain is a quintessential example. Abrupt transitions in brain activity at a fixed anaesthetic concentration in the absence of acute stimuli, observed by us and by others, are consistent with multi-stability. All multi-stable dynamical systems exhibit resistance to state transitions. There is no fundamental reason to believe that the assumption of a one-to-one correspondence between anaesthetic concentration in any compartment of a PK/PD model and brain activity is valid generically. Furthermore, there is ample empirical evidence to the contrary acquired in different species, using different anaesthetics and recording modalities. In toto, the combination of behavioural, genetic, neurophysiological, and theoretical arguments strongly argue that PK/PD is not sufficient to explain anaesthetic hysteresis—some other non-pharmacological variables must intervene between the concentration of the drug and the observed behavioural or neurophysiological response. So, can we re-interpret Kuizenga's and Warnaby's human data in light of overwhelming theoretical and experimental support in animals across taxa? One honest assessment could be that the clinical significance of neural inertia might be modest at best. However, before settling upon this interpretation some important caveats must be addressed. Kuizenga and colleagues attempted to identify neural inertia by fitting their data with a standard sigmoidal dose-response curve and added additional variables to detect potential shifts in the EC50 between induction and emergence. Critically, they assume identical Hill slopes for their induction and emergence datasets. This assumption represents a departure from previous animal studies showing that Hill slope for emergence is consistently shallower than for induction.3McKay I.D. Voss L.J. Sleigh J.W. Barnard J.P. Johannsen E.K. Pharmacokinetic-pharmacodynamic modeling the hypnotic effect of sevoflurane using the spectral entropy of the electroencephalogram.Anesth Analg. 2006; 102: 91-97Crossref PubMed Scopus (66) Google Scholar, 5Friedman E.B. Sun Y. Moore J.T. et al.A conserved behavioral state barrier impedes transitions between anesthetic-induced unconsciousness and wakefulness: evidence for neural inertia.PLoS One. 2010; 5e11903Crossref PubMed Scopus (156) Google Scholar, 11Joiner W.J. Friedman E.B. Hung H.T. et al.Genetic and anatomical basis of the barrier separating wakefulness and anesthetic-induced unresponsiveness.PLoS Genet. 2013; 9e1003605Crossref PubMed Scopus (56) Google Scholar Another potential methodological difficulty arises from their choices of targets for anaesthetic concentrations. Hysteresis is most apparent in the linear region of the concentration-response curve between the EC25 and EC75 (e.g. Fig. 1A). Thus, one has the best chance of detecting hysteresis by densely sampling this region.17Sonner J.M. Issues in the design and interpretation of minimum alveolar anesthetic concentration (MAC) studies.Anesth Analg. 2002; 95: 609-614PubMed Google Scholar Unfortunately, most sampling in the Kuizenga study occurs outside of this region. Concentration sampling for the propofol volunteer group in the Warnaby study was denser, which could be one reason why they detect evidence for hysteresis. On the other hand, in contrast to Kuizenga and colleagues, Warnaby and colleagues did not measure propofol concentrations to verify that they indeed correspond to the targeted concentrations. Differences between Warnaby and Kuizenga studies offer a potential explanation for the inconsistency of experimental evidence for hysteresis. Kuizenga and colleagues studied PSI and spectral edge as EEG markers of anaesthetic depth. In contrast, Warnaby and colleagues used SWAS—a measure that they have recently proposed as an improved indicator of depth of anaesthesia.18Ni Mhuircheartaigh R. Warnaby C. Rogers R. Jbabdi S. Tracey I. Slow-wave activity saturation and thalamocortical isolation during propofol anesthesia in humans.Sci Transl Med. 2013; 5 (208ra148)Crossref PubMed Scopus (119) Google Scholar With respect to SWAS, clear hysteresis was observed for propofol. Yet, with PSI and spectral edge no evidence for hysteresis was found. Methodological caveats notwithstanding, the differences between the findings suggest perhaps that distinct measures derived from the EEG do not necessarily change in parallel as a function of the anaesthetic concentration. This was raised as a possibility by Kuizenga and colleagues, and we concur wholeheartedly. Indeed, EEG signals are complex.19Buzsaki G. Anastassiou C.A. Koch C. The origin of extracellular fields and currents–EEG, ECoG, LFP and spikes.Nat Rev Neurosci. 2012; 13: 407-420Crossref PubMed Scopus (2201) Google Scholar The debate concerning proper methodology and theoretical framework in which EEG signals ought to be interpreted has been raging in neuroscience for decades. Spectral characteristics,20John E.R. Prichep L.S. Kox W. et al.Invariant reversible QEEG effects of anesthetics.Conscious Cogn. 2001; 10: 165-183Crossref PubMed Scopus (224) Google Scholar connectivity,21Imas O.A. Ropella K.M. Ward B.D. Wood J.D. Hudetz A.G. Volatile anesthetics disrupt frontal-posterior recurrent information transfer at gamma frequencies in rat.Neurosci Lett. 2005; 387: 145-150Crossref PubMed Scopus (144) Google Scholar, 22Lee U. Ku S. Noh G. Baek S. Choi B. Mashour G.A. Disruption of frontal-parietal communication by ketamine, propofol, and sevoflurane.Anesthesiology. 2013; 118: 1264-1275Crossref PubMed Scopus (271) Google Scholar phase lag index,23Blain-Moraes S. Lee U. Ku S. Noh G. Mashour G.A. Electroencephalographic effects of ketamine on power, cross-frequency coupling, and connectivity in the alpha bandwidth.Front Syst Neurosci. 2014; 8: 114Crossref PubMed Scopus (91) Google Scholar and stability24Solovey G. Alonso L.M. Yanagawa T. et al.Loss of consciousness is associated with stabilization of cortical activity.J Neurosci. 2015; 35: 10866-10877Crossref PubMed Scopus (70) Google Scholar have all been used to describe the effect of anaesthetics on the brain. How can then one be sure that the EEG measure upon which propofol effect site concentrations is computed is correct? This question exposes a fundamental flaw in the current conception of the effect-site concentration that confounds most human anaesthetic studies. By its very construction, it is impossible to verify the effect-site concentration using any direct empirical measurements. As currently defined, calculations of the effect-site concentration assume that the brain is static and that relevant measures of brain activity are readily available. Neither of these assumptions is demonstrably true. Therefore, a reasonable scientific inquiry might rightly question the propofol effect-site concentrations (x-axis) with respect to which the response (y-axis) is plotted. To be clear, anaesthetics do not act on plasma proteins to cause unconsciousness. In order to produce unconsciousness, anaesthetics must distribute into the brain—their target organ. This distribution is not instantaneous. Thus, we do not object to the idea that there must be a delay between appearance of the drug in the plasma and the target organ. What we object to is the notion that one can infer the concentration in the target organ unambiguously, simply on the basis of a somewhat arbitrary measure of brain activity, such as the EEG. It seems undesirable from a modelling perspective to define a compartment in a PK/PD model in which drug concentration cannot be empirically measured. The assumptions implicit in the calculation of the effect-site concentration are reminiscent in their ontology to the cosmological constant that Einstein originally postulated to ensure that general relativity theory gave rise to a static universe. Of course, the universe was subsequently shown to be expanding. Einstein famously called the cosmological constant the biggest blunder of his life. In the same vein, we propose to move beyond the assumption of a fixed relationship between effect-site concentrations and brain activity. So doing will facilitate the exploration of interesting neuroscientific questions surrounding transitions between conscious and unconscious states. We applaud the efforts of Kuizenga and Warnaby and colleagues for undertaking these translational studies with both volatile anaesthetics and propofol. While tackling neural inertia may prove challenging in a human study, direct detection of neural inertia in humans would be more feasible using an anaesthetic whose concentration in human brain could be readily measured. Once such human trial using xenon as an anaesthetic is currently underway (NCT02768727). Xenon is unique amongst anaesthetics because its radio-opacity permits direct measurement of xenon in the brain using CT scanning.25Lee T.Y. Ellis R.J. Dunscombe P.B. et al.Quantitative computed tomography of the brain with xenon enhancement: a phantom study with the GE9800 scanner.Phys Med Biol. 1990; 35: 925-935Crossref PubMed Scopus (29) Google Scholar As such, xenon would be an ideal drug to settle the issue of neural inertia in humans. Results from xenon and future anaesthetic trials could help us peer into the black box of the brain to determine if at some critical anaesthetic concentration Schrödinger's cat, mouse, fly, and indeed, patient, is truly anesthetized or not. M. B. K. and A. P. drafted and revised the editorial for its intellectual content. M. B. K. approved the final version to be published. None declared. This work was supported by the National Institute of General Medical Studies (Bethesda MD, USA) 5K08GM106144 (A.P.), R01GM107117 (M.B.K.), and R01GM088156 (M.B.K.). Both A.P. and M.B.K. are supported by the Department of Anesthesiology and Critical Care at the University of Pennsylvania and by a grant from the James S. McDonnell Foundation for understanding human consciousness.
Referência(s)