Anaesthetic Mechanism

Speculative If the molecular basis of general anaesthesia were known — which it is not, and section 7 says what would establish it — the prize is an agent designed for one component of the state instead of a molecule that produces all of them and several side effects at once. Every drug in clinical use today was found by looking, not by design, and the conditional form here is the honest one.

Established The magnitude is best stated as the size of the population exposed and the width of the dose window, because those are the two quantities the field actually reports. The Lancet Commission on Global Surgery counted 312.9 million surgical procedures worldwide in 2012, and between 4.8 and 5.3 billion people without access to safe, timely surgical and anaesthesia care. The dose window is narrow and the narrowness is the cost: the standard deviation of the minimum alveolar concentration is about 10%, so producing immobility in roughly 95% of patients takes 1.2 MAC, and the same excess depresses myocardium and respiration in everyone. A target-selective immobiliser would recover that margin; nobody can design one because nobody knows which target to aim at.

Frontier Who gains is unusually broad and unusually cheap, because the drugs are off patent. Isoflurane, sevoflurane, propofol and ketamine are generics; a mechanistic advance would not create a monopoly so much as redistribute safety toward the settings that cannot currently staff an anaesthetist. Speculative The counterweight is that the commercial incentive to find the mechanism is correspondingly weak, which is why the field's funding is discussed in section 10 as a structural constraint rather than an aside.

Handwave The condition, stated plainly: the upside requires that anaesthesia have a target, and the best current evidence is that it has several hundred. A 2019 review estimates that 10 to 15% of detectable proteins qualify as anaesthetic binding targets — over 300 proteins, possibly tenfold more — and a 2026 survey of 310 in vitro studies finds propofol more potent at a neuronal nicotinic receptor and at HCN1 than at the GABA-A receptor it is named for. Section 5 is about what follows if promiscuity is the answer rather than an artefact of incomplete searching.

General anaesthesia is not one state. It is a bundle of at least four separately produced effects — unconsciousness, amnesia, immobility in response to noxious stimulus, and suppression of autonomic responses — delivered by a single drug because a single drug happens to produce all of them. Each component has a different dose threshold, and at least two of them are produced in different parts of the nervous system.

Established The field's central measured quantity is the minimum alveolar concentration, and it measures immobility rather than consciousness. MAC is the alveolar concentration at which 50% of subjects do not move in response to a surgical stimulus. MAC-awake, the concentration suppressing response to verbal command, is roughly one third of MAC; MAC-BAR, which blocks the autonomic response to incision, is roughly 1.5 MAC. Established Human values: isoflurane 1.2%, sevoflurane 2.6%, desflurane 6.6%, halothane 0.75%, nitrous oxide 104% — a spread of more than two orders of magnitude in concentration for drugs producing the same endpoint.

Established The oldest quantitative fact in the subject is the Meyer-Overton correlation, and it is still the strongest single regularity anyone has found. Anaesthetic potency tracks solubility in olive oil across four to five orders of magnitude of lipid solubility, for molecules with no chemical feature in common — a noble gas, a halogenated ether, an alkyl phenol, a steroid. Speculative A correlation that strong across chemistry that diverse is what makes this a physics problem rather than a pharmacology problem, and it is also what has made it unsolvable for a century: it tells you the drug goes somewhere hydrophobic and says nothing about what it does there.

Frontier The correlation fails in four specific places, and the failures are what killed the lipid theory. The cutoff effect: in homologous series, potency disappears above a chain length of about 13 for n-alcohols and between 6 and 10 for n-alkanes, though lipid solubility keeps rising. Stereoisomers with the same oil/gas partition coefficient differ in potency. Non-immobilisers such as F6 are lipid-soluble, produce amnesia and do not produce immobility. And a temperature change of 1 °C perturbs the membrane as much as a clinical anaesthetic dose without producing unconsciousness. Established Those four facts moved the field from lipids to proteins around 1984 and it has not found the protein.

Speculative This brief takes the position that anaesthetic mechanism is pre-paradigmatic because the field has an excellent measurement of the wrong quantity. MAC is reproducible enough to run an operating theatre on and it indexes a spinal reflex; the endpoint patients and ethicists care about is cortical. Handwave Every molecular result below is anchored to one or the other, the literature routinely reports them as though they were the same, and that conflation is doing more damage than any single wrong hypothesis.

Established The measured record starts with what is genuinely solid, and the commission that produced this brief named three anchors, of which one survives intact and two need correcting. The corrections are below with the numbers, because in this subject the number is the argument.

Established Anchor one, verified with one qualification: the Meyer-Overton correlation. The standard statement is that potency tracks lipid solubility over four to five orders of magnitude when olive oil is the reference phase, so the commission's “five orders of magnitude” is the top of the conventional range rather than its centre. Speculative The correlation is real and it is also the single most misleading fact in the subject, because it is equally well explained by a lipid mechanism and by protein binding sites that happen to be hydrophobic — and the second explanation is now the majority view.

Established Anchor two needs correcting twice over: MAC is not reproducible to a few per cent, either within a population or across species. The standard deviation of MAC in a human population is about 10%, which is why 1.2 MAC — two standard deviations — is the concentration that immobilises roughly 95% of patients and 1.1 MAC immobilises about 68%. MAC also declines by roughly 6% per decade of age after a peak at six months, falls by 4 to 5% per °C of hypothermia and by up to 30% in pregnancy, and is raised by chronic alcohol use and by red hair. Frontier Across species the absolute value moves too: isoflurane MAC is about 1.2% in humans, about 1.3% in dogs and about 1.6% in cats — a spread of roughly a third between two domestic carnivores, not a few per cent. Frontier A 2026 British Journal of Anaesthesia study of isoflurane in male mice goes further and argues that MAC is not even a static quantity within one animal over time; this session could not retrieve its figures, so it is cited by name and no number from it is used here.

Established What IS conserved, and what the commission's anchor was probably reaching for, is the rank order of potency rather than its absolute value. The relative potency ordering of volatile agents is nearly identical from humans down to Caenorhabditis elegans — while the absolute concentrations that immobilise a human do not immobilise a worm at all. Speculative A rank order conserved across 600 million years of divergence with a scale factor that is not conserved is a much stranger and more interesting fact than a reproducible number, and it is the fact a mechanism has to explain.

Established Anchor three, on the exposed population, is too small and names a quantity nobody counts. The commission gave “a hundred million people a year”; this session could not retrieve any primary source for that figure. The best-documented denominator is the Lancet Commission on Global Surgery's 312.9 million surgical procedures worldwide in 2012, against an earlier modelling estimate of 234.2 million major procedures in 2004, of which the poorest third of the world's population received about 3.5%. Frontier Neither source separates general anaesthesia from regional and local technique, so the number of general anaesthetics per year is not a measured quantity in any source found here — and a field whose central justification is the size of its exposed population does not know that size to better than a factor of three.

Established The most consequential established result in the subject is that the endpoint defining MAC is produced in the spinal cord, not the brain. Delivering isoflurane or halothane selectively to the brain of a goat required 2.5-fold to 4-fold higher concentrations to produce immobility than whole-CNS delivery. Removing the forebrain of a rat does not alter isoflurane MAC at all. Frontier MAC therefore measures a spinal reflex threshold and is silent about consciousness, which is the quantity Consciousness Research is about and the quantity every depth-of-anaesthesia monitor claims to track.

Established On molecular targets, the pattern is that intravenous agents have one and volatile agents do not. The GABA-A receptor β3(N265M) knock-in mouse strongly attenuates the immobilising and hypnotic actions of etomidate and propofol, which is the cleanest genetic attribution in the field. Established For volatile agents the same receptor is not the answer: the standard summary is that GABA-A receptors appear not to be important for the endpoint of immobility where inhalational agents are concerned, and NMDA receptor block has been excluded pharmacologically as the immobilising action of volatiles.

Frontier The two-pore potassium channel story is the field's clearest example of a result that did not replicate. TREK-1 knockout mice were reported in 2004 to be less sensitive to volatile anaesthetics, and global knockouts lacking TASK-1, TASK-3 and TREK-1 are described in reviews as less sensitive to volatile but not intravenous agents. Frontier In 2023 a study in Anesthesiology reported that TREK-1 and TREK-2 knockout mice are not resistant to halothane or isoflurane. Speculative This matters directly to the commission's proposed decisive experiment, which set a lipid-raft/PLD2 mechanism against a GABA-A/TREK-1 one: one arm of that contrast rests on a knockout phenotype that a later group could not reproduce. This session could not retrieve the 2023 paper's MAC values, so it is cited by name.

Established Meanwhile a germline mutation in a mammal that shifts volatile-anaesthetic potency about twofold already exists, and it is in none of the candidate pathways. Mice lacking Ndufs4, a subunit of mitochondrial complex I, have an isoflurane MAC of 0.81% against 1.55% in wild-type at 22 to 25 days, and 0.65% against 1.68% at 31 to 34 days — roughly 1.9-fold and 2.6-fold hypersensitivity, in seven knockout, five heterozygous and five wild-type animals. Frontier The commission asked for a germline knock-in in a mammal that shifts MAC for a volatile agent as the decisive experiment. A germline knockout that does exactly that has been published, the target is a mitochondrial electron-transport subunit, and neither the lipid-raft camp nor the receptor camp predicted it.

Established The invertebrate genetics points the same way and is older. In C. elegans, syntaxin alleles span a 33-fold range of isoflurane sensitivity; gas-1, a complex I subunit, confers extreme hypersensitivity to every volatile agent tested; unc-1, a stomatin homologue, gives resistance or hypersensitivity depending on allele. Frontier The field's own summary of forty years of unbiased screening is that the genetic data do not identify a single protein or channel uniquely responsible for anaesthetic effects. Speculative That is a negative result of unusual quality: the screens worked, they were not underpowered, and what they found was a list.

Frontier The 2026 pharmacology survey is the strongest recent evidence and it is a refutation of the field's own vocabulary. Reviewing 310 in vitro studies, Kale and colleagues report propofol most potent at the neuronal α4β2 nicotinic acetylcholine receptor (IC₅₀ 3.13 µM) and at HCN1 (IC₅₀ 4.19 µM), both below its GABA-A potency (EC₅₀ 5.44 µM); halothane and isoflurane peak at the same nicotinic receptor (EC₅₀ 40.9 and 59.9 µM) rather than at GABA-A; only etomidate is relatively selective (EC₅₀ 1.08 µM at GABA-A). Established Their conclusion is that calling agents “GABAergic” or “NMDAergic” is unhelpful and misleading, and that the multi-site hypothesis is the more likely one.

Established The scale of the promiscuity has been estimated and it is the number that should govern expectations. A 2019 review in Trends in Pharmacological Sciences puts anaesthetic binding targets at 10 to 15% of detectable proteins — over 300 proteins, with the true number possibly tenfold greater — and states plainly that the mechanisms producing anterograde amnesia are still unclear. Speculative If a tenth of the proteome binds the drug at relevant concentrations, “find the target” may be the wrong research programme rather than an unfinished one.

Established Structural biology has delivered binding sites without delivering a mechanism. A GABA-A receptor–propofol complex has been resolved at about 3.3 Å with propofol at the β+/α− subunit interface; isoflurane has been located in hydrophobic transmembrane pockets; ketamine has been resolved occluding the GluN2B channel pore with an M3 helix conformational change. Frontier Knowing where a molecule sits does not establish that the site is the one producing the clinical endpoint, and for the volatile agents no solved structure has been tied to a MAC shift in an animal.

Frontier The lipid theory has a modern descendant that makes a testable claim. The Hansen laboratory proposes that inhaled anaesthetics activate TREK-1 by disrupting ordered lipid domains, releasing phospholipase D2 to produce an anionic lipid that opens the channel, and reports that chloroform, isoflurane, diethyl ether, xenon and propofol all activate PLD2 in cellular membranes. Speculative This session could retrieve only the abstract, so no quantitative result from it is asserted here. Speculative The claim's value is that it is mechanistically specific and therefore falsifiable, and its exposure is that its downstream effector is the channel whose knockout phenotype failed to replicate in 2023.

Established Two results are usually filed as curiosities and belong in the measured record. Anaesthetics immobilise organisms with no nervous system: 15% diethyl ether abolished the Venus flytrap's action potentials in response to trigger-hair stimulation with recovery about 15 minutes after removal, stopped Mimosa pudica leaf movement within 1 hour with recovery over 7 hours, and halted pea-tendril circumnutation immediately; 80% xenon in oxygen and 1% lidocaine did likewise. Frontier Whatever general anaesthetics do, they do a version of it to excitable cells that have never had a neuron, and no receptor-centred account addresses that.

Speculative And the result at the far edge, reported honestly because it is real data attached to a fringe theory. Rats given 0.75 mg/kg of the microtubule-stabilising drug epothilone B took on average 69 seconds longer to lose the righting reflex under isoflurane (Cohen's d = 1.9, P = 0.0016) in eight male animals. Handwave The authors read this as support for anaesthetics acting on microtubules and for a quantum account of consciousness; a 69-second delay in eight male rats supports the first reading weakly and the second not at all, and the authors themselves list replication as required.

Established The clinical record is where the mechanism gap becomes measurable in patients. Accidental awareness under general anaesthesia occurs at rates quoted up to 2 per 1,000 patients. Processed-EEG depth monitors were built to reduce it, and the trial record is split: the B-Aware trial found a benefit at a number-needed-to-treat of 138 and about $2,200 per case prevented; B-Unaware found no reduction against end-tidal concentration monitoring; BAG-RECALL found fewer awareness episodes in the end-tidal arm than in the bispectral-index arm. Frontier A health technology assessment modelled 45 awareness cases per 10,000 high-risk intravenous-anaesthesia patients under standard monitoring against 10.8 with the index, pooled an odds ratio of 0.45 (95% CI 0.25 to 0.81) across all techniques, and reported confidence intervals crossing 1 in the inhalational subgroup. Speculative The monitors reflect a probability function of clinical state rather than a physiological quantity, which is what you build when you do not know the mechanism.

Established One clinical worry has been resolved cleanly and it is worth recording as a success. The GAS trial randomised 722 infants to awake-regional or general anaesthesia for hernia repair, median general-anaesthesia duration 54 minutes (IQR 41 to 70), and found full-scale IQ at five years of 99.08 against 98.97, a difference of 0.23 (95% CI −2.59 to 3.06) — strong evidence of equivalence. Frontier An hour of general anaesthesia in infancy does not measurably damage the five-year brain, and that is a properly powered negative result the field obtained without knowing the mechanism at all.

Frontier Question one: is promiscuity the mechanism or the noise? If a tenth of the proteome binds anaesthetics at clinical concentrations, then either the clinical state is the summed effect of small perturbations at hundreds of sites, or a handful of those sites matter and the rest are pharmacologically irrelevant binding. Speculative The field's own reviews say the question remains undetermined, and the two answers imply completely different research programmes: one says keep screening for the target, the other says stop and build a systems-level theory of distributed perturbation. Handwave Nobody has proposed an experiment that distinguishes them, which is the most pre-paradigmatic fact on this page.

Frontier Question two: why is the rank order of potency conserved when the scale factor is not? Volatile agents order identically in potency from human to nematode, yet the concentrations that immobilise a human leave a worm moving. Speculative A conserved order with an unconserved scale implies a conserved physical interaction whose downstream amplification differs — which is closer to the lipid theory's spirit than to a receptor account, and closer to neither than to a thermodynamic one. Speculative This is the observation a successful mechanism would explain first and no current candidate addresses.

Frontier Question three: what does a mechanism owe to organisms without neurons? Diethyl ether abolishes Venus flytrap action potentials and stops pea-tendril movement, reversibly, on a timescale comparable to a mammal's. Speculative Either the plant effect is a different phenomenon sharing a name, or the mechanism is at the level of excitable membranes and vesicle trafficking rather than at synaptic receptors. Handwave The literature almost universally takes the first option without testing it, and the test — does the plant effect obey Meyer-Overton, the cutoff and pressure reversal? — is cheap.

Frontier Question four: is MAC worth keeping? It is spinal, it has a 10% population standard deviation, it moves with age, temperature, pregnancy, alcohol and hair colour, and a 2026 study argues it is not even static within an animal. Speculative It is nonetheless the only endpoint in the subject with a century of comparative data behind it, and abandoning it would orphan every cross-species comparison in section 3. Speculative The likely resolution is that MAC survives as an operational dosing rule and stops being treated as a measurement of anaesthetic depth, which is a demotion the textbooks have not yet made.

Frontier Question five: does mitochondrial complex I sit upstream of everything else? The Ndufs4 knockout, the gas-1 nematode allele, and the paediatric observation that children with complex I defects show altered anaesthetic sensitivity all point at electron transport. Speculative A metabolic mechanism would also connect this brief directly to Sleep Function, where the strongest 2025 result locates sleep pressure in an electron-supply mismatch in mitochondria. Handwave Whether the two states share a substrate or merely share a fashionable organelle is exactly the kind of question this category exists to hold open rather than answer.

Frontier Question six: what is the relationship between immobility and unconsciousness at the circuit level? MAC-awake is about a third of MAC, so patients are unconscious long before they are immobile, and the two thresholds are produced in different structures. Speculative A drug that produced cortical unconsciousness without spinal immobility, or the reverse, would be clinically valuable and would also be the sharpest possible demonstration that the bundle is separable. Speculative Etomidate approaches the first half through a known receptor; nothing approaches the second.

Speculative Question seven, the one the field least likes: does the mechanism question have a well-posed answer at all? A state produced by a noble gas, a steroid and an alkyl phenol, in a nematode, a flytrap and a human, may not have a mechanism in the sense the word usually carries. Handwave The honest possibility is that “general anaesthesia” names a convergent behavioural outcome of many sufficient perturbations rather than a natural kind — in which case a century of searching has been looking for a thing that is not there, and the correct successor programme is to ask which perturbations are sufficient and why.

Frontier The bottleneck is not instrumentation. It is that the field's measurable endpoint and its question of interest are in different parts of the nervous system. Everything below follows from that and from the promiscuity of the ligands.

Frontier Bottleneck one: the endpoint problem. MAC is cheap, comparative and spinal; unconsciousness is what matters and has no species-portable measure. Speculative A genetic manipulation that shifted MAC would be reported as a discovery about anaesthesia and would strictly be a discovery about a spinal reflex threshold. Handwave Every mouse genetics result in section 3 carries this ambiguity and almost none of them state it.

Established Bottleneck two: the ligands are promiscuous by construction. Isoflurane has a molecular mass around 184 Da, no charge, no hydrogen-bond donor and a partition behaviour that puts it into every hydrophobic pocket in the cell. Speculative Affinity-based target identification assumes a target with distinguishing affinity, and a drug that binds 300 proteins with millimolar-to-micromolar promiscuity defeats the assumption rather than the technique.

Frontier Bottleneck three: no animal model of the component that matters. Immobility is measurable in a goat, a mouse and a worm; amnesia and unconsciousness are inferred. Speculative Depth monitors reflect a probability function of clinical state rather than a physiological parameter, which means the clinical instrument cannot serve as the laboratory readout either. Speculative This is the same calibration trap that Consciousness Research documents from the other side.

Frontier Bottleneck four: knockouts are developmental, and the compensations are invisible. A germline deletion of a potassium channel or a complex I subunit reorganises a nervous system over development; the resulting shift in anaesthetic sensitivity may report the compensation rather than the target. Speculative The failure to replicate the TREK-1 resistance phenotype is exactly the shape of result that strain background and developmental compensation produce, and it is why conditional and inducible manipulations matter more here than in most pharmacology.

Speculative Bottleneck five: the concentrations are awkward for every biophysical method. Clinical isoflurane is a few hundred micromolar in tissue; structural and spectroscopic methods that resolve binding generally want higher occupancy, and results obtained above clinical concentration are routinely reported without that caveat. Handwave A binding site demonstrated at ten times MAC is a chemistry result, not a pharmacology one, and the literature does not consistently separate the two.

Speculative Bottleneck six is commercial and it is binding. Every agent in wide use is off patent. Handwave There is no company with a revenue reason to fund the decade of mechanism work that would let somebody design a better one, which is why the requirement in section 10 is a funder rather than an instrument.

Frontier This subject waits on results from fields that are not studying anaesthesia, which is the usual shape of a problem that has resisted direct attack.

Frontier Dependency one: a species-portable measure of unconsciousness that is not behavioural responsiveness. Without it, no genetic or pharmacological manipulation can be scored against the endpoint that matters. Speculative Consciousness Research owns this problem and its own verdict is that every marker in the field was calibrated against verbal report, so the dependency is currently unmet and known to be unmet.

Speculative Dependency two: chemical proteomics that works for a small, weak, promiscuous ligand. Photoaffinity analogues of anaesthetics exist; a method that returns ranked occupancy across the proteome at clinical concentration, in intact tissue, does not. Speculative This would be built by a chemical-biology group for its own reasons and inherited here.

Established Dependency three: mitochondrial bioenergetics. The strongest germline result in a mammal is a complex I subunit knockout, and the nematode result that matches it is also complex I. Speculative Longevity Therapies and Sleep Function both depend on the same organelle's physiology, and a measurement made for either would be read here.

Speculative Dependency four: conditional and inducible mammalian genetics at adult stages. Every ambiguity in section 5's fourth bottleneck is removed by a manipulation made after development, and the tools exist. Handwave What does not exist is a systematic programme applying them to the candidate target list, and the list is now over three hundred proteins long.

Frontier The experiments below rank by how much a result would change this brief's assessment. The commission proposed one, and the first thing to say is that a version of it has already been done.

Speculative Tier one, and the decisive experiment for this brief: an adult-inducible, tissue-restricted mammalian manipulation that shifts MAC for a volatile agent, with the target chosen in advance from competing theories and the predicted direction registered. Established The germline version already exists and it did not go to any of the predicted targets: Ndufs4 knockout mice show isoflurane MAC of 0.81% against 1.55% wild-type, and 0.65% against 1.68% a week later — roughly two- to 2.6-fold hypersensitivity from a mitochondrial complex I subunit. Frontier What that result cannot do is separate target from developmental compensation, which is precisely what an adult-inducible, spinally restricted version would settle. Speculative Run it against three arms — complex I, the lipid-raft/PLD2 pathway, and a two-pore potassium channel — in the same animals, same rig, same operator, and the field gets a ranked answer instead of three laboratories' separate ones.

Frontier Tier two: settle the two-pore potassium channel question, because one arm of the commission's proposed contrast rests on it. A 2004 report of volatile-anaesthetic resistance in TREK-1 knockouts was followed in 2023 by a report that TREK-1 and TREK-2 knockouts are not resistant to halothane or isoflurane. Speculative The arbitration is a co-designed replication on matched genetic backgrounds, with MAC determined by the same up-and-down protocol in both laboratories and analysis code fixed in advance. Speculative It is one figure, it costs a few dozen mice, and until it is run every downstream claim about lipid rafts acting through TREK-1 has a broken link in it.

Speculative Tier three: the promiscuity discriminator. Take the ranked target list and ask whether blocking the top five simultaneously shifts MAC more than the sum of blocking them individually. Frontier Additivity favours the distributed-perturbation account; a disproportionate shift from one combination favours a small set of privileged sites. Speculative This is the only design on this page that tests the question in section 4's first paragraph, and it requires no new technology — only the willingness to run a factorial experiment in an area that publishes one target at a time.

Speculative Tier four: separate immobility from unconsciousness by construction. Deliver an agent selectively to the cord or selectively to the brain in a large mammal, as the goat experiments did, and measure a cortical marker and the immobility threshold simultaneously. Frontier The prediction is already published and quantitative: brain-only delivery took 2.5-fold to 4-fold more agent to immobilise. Speculative Repeating it with a modern cortical readout would convert a forty-year-old anatomical result into a dissociation the depth-monitor industry would have to answer.

Speculative Tier five: test whether the plant effect is the same phenomenon. Does anaesthetic action on Dionaea action potentials obey the Meyer-Overton correlation, show the alcohol cutoff, and reverse under hydrostatic pressure? Frontier All three are the classical diagnostics, the preparation is cheap, and a positive answer would make every receptor-centred account incomplete rather than wrong. Handwave A negative answer would retire the comparison, which would also be worth having and would cost about the same.

Speculative Tier six: the epothilone replication, done properly. Eight male rats and a 69-second effect is a pilot. Frontier A pre-registered replication in both sexes with a full MAC determination rather than a righting-reflex latency, and an inactive microtubule-binding control, would establish whether the effect is real independently of the quantum interpretation attached to it. Speculative The interpretation is the reason the result is dismissed and the result deserves testing on its own terms.

Frontier Tier seven: the human natural experiment already running. Patients with mitochondrial disease undergo anaesthesia routinely, and the observation that children with complex I defects show altered sensitivity is the only human evidence bearing on the strongest animal result here. Speculative A registry recording end-tidal concentration at loss of response, genotype and complex I activity across such patients is an observational study using data anaesthetists already record. Handwave Nobody has built it, and it would cost a fraction of any experiment above.

Established And the experiments that are not tests of mechanism and are regularly reported as though they were. Demonstrating that an anaesthetic modulates a receptor in a heterologous expression system, or resolving a binding site in a purified channel, establishes that a molecule fits a pocket. Speculative Neither establishes that the pocket produces the clinical endpoint, and for the volatile agents no solved structure has yet been connected to a MAC shift in an animal.

Frontier There is no engineering payoff here yet and the page will not manufacture one. Every requirement below is for an instrument or a preparation that would answer a question.

Speculative The specification for the decisive test, stated from the protocols it has to combine. An adult-inducible, tissue-restricted mammalian line for each candidate target; an up-and-down MAC determination with a standardised noxious stimulus and controlled core temperature, since MAC falls 4 to 5% per °C; end-tidal gas monitoring calibrated across sites; a simultaneous cortical readout so immobility and unresponsiveness are scored separately; and pre-registered analysis. Frontier Each component is routine. The combination is not built because MAC determination and cortical electrophysiology belong to different laboratories.

Established The clinical engineering that exists is instructive about what a mechanism would buy. Processed-EEG depth monitors are deployed at scale and the health technology assessment puts their incremental cost-effectiveness between about £5,681 and £49,198 per quality-adjusted life year depending on device and population, with the pooled awareness odds ratio of 0.45 not separable from standard monitoring in the inhalational subgroup. Frontier That is the price of a monitor built on a correlate. A monitor built on a mechanism would report drug effect at a target rather than a probability of a clinical state, and no such device has been specified. Brain-Computer Interfaces supplies the recording hardware this would inherit.

Speculative One preparation is missing and it is cheap. A standardised non-neural test organism — Dionaea or a comparable electrically excitable plant — with published concentration-response curves for the classical volatile series would let any laboratory test whether an account of anaesthesia survives outside animals. Handwave Nobody maintains one, and its absence is why the plant results are quoted as a curiosity rather than used as a control.

Frontier The nearest page on this map is Consciousness Research, and the two divide the subject cleanly. That brief uses anaesthesia as an instrument: a reversible, titratable way to move a brain between states so that markers of consciousness can be calibrated against it. Established This brief treats the drug as the unknown. Speculative The seam is sharp and consequential: that brief's perturbational complexity index was calibrated on propofol, midazolam and xenon transitions, and this brief's central finding is that the concentration scale used to define those transitions indexes a spinal reflex rather than a cortical state. Neither brief can fix the other's problem and both inherit it.

Frontier The second adjacency is Sleep Function, and it is closer than the clinical framing suggests. Anaesthesia is routinely used as a model of sleep in fluid dynamics and circuit work, and the 2024 result that brain clearance falls under three structurally unrelated anaesthetics is evidence about anaesthesia as much as about sleep. Speculative Both briefs converge on mitochondrial electron transport from opposite directions, which is either a real shared substrate or a shared fashion, and no experiment currently distinguishes them.

Speculative Three neighbours inherit something concrete. Precision Neuropsychiatry is downstream of ketamine, which went from anaesthetic to rapid-acting antidepressant without its mechanism being settled in either role. Longevity Therapies shares the complex I biology. Aging Societies carries the demography behind postoperative delirium, which is the clinical endpoint a mechanism would most plausibly improve and which the 2026 pharmacology literature names as an open problem.

Frontier One adjacency runs outward from medicine and is usually missed: plant electrophysiology. The organisms that respond to diethyl ether and xenon with no nervous system belong to a community with no contact with anaesthesiology. Speculative That is an adjacency the field could exploit for the price of a growth chamber, and section 8 says what it would take.

Frontier The institutional problem is that the question is clinically enormous and commercially worthless. Isoflurane, sevoflurane, propofol and ketamine are generic; the exposed population runs to hundreds of millions of procedures a year; and no firm captures the return on discovering why they work. Speculative Mechanism research is therefore funded as basic neuroscience competing against neuroscience, rather than as drug development competing against drug development, and it loses.

Frontier The field also lacks the arbitration structure its disagreements need. A 2004 knockout phenotype and a 2023 failure to reproduce it sit in the literature side by side; a 2026 survey says the field's own vocabulary for its drugs is misleading. Speculative What is missing is not data but a convention: a pre-registered, multi-laboratory MAC determination protocol with fixed analysis, of the kind clinical trials have had for thirty years and preclinical pharmacology has not.

Established A second requirement is a reporting standard, and it is nearly free. A convention requiring every mechanism paper to state which component of the anaesthetic state it measured — immobility, unresponsiveness, amnesia or autonomic suppression — and at what multiple of clinical concentration its in vitro work was done would remove most of the ambiguity documented in sections 3 and 5. Speculative Journals impose comparable conventions routinely elsewhere.

Frontier Third, and largest: the exposed population is not counted. The best available figures are surgical-procedure counts — 234.2 million in 2004, 312.9 million in 2012 — with no breakdown by anaesthetic technique. Speculative A registry recording technique and agent alongside procedure would cost almost nothing at the point of care and would give the field a denominator it currently guesses at by a factor of three. The same registry is the substrate for the mitochondrial natural experiment in section 7.

Speculative Fourth, an equity requirement that belongs here rather than in section 11. Between 4.8 and 5.3 billion people lack access to safe, timely surgical and anaesthesia care, and the poorest third of the world received about 3.5% of operations in the 2004 estimate. Handwave A mechanism that permitted a safer agent with a wider therapeutic index would matter most where there is no anaesthetist to compensate for a narrow one — which is an argument for public rather than commercial funding and is not currently made by anyone with a budget.

Established The live ethical fact is that a hundred and eighty years of practice rests on a correlation. Consent for general anaesthesia describes a state whose mechanism is unknown, delivered by titration against a threshold that measures a spinal reflex. Speculative That is not an argument against anaesthesia — the practice is safe by every outcome measure — but the routine assurance that medicine understands what it is doing here is not accurate, and it is the kind of inaccuracy that erodes trust when it surfaces.

Frontier Accidental awareness is the sharpest case and the numbers are contested. Rates quoted run up to 2 per 1,000 patients; the monitors marketed to prevent it produced a benefit in one trial at a number-needed-to-treat of 138, no benefit in a second, and a worse result than end-tidal monitoring in a third. Speculative A patient consenting to depth monitoring is consenting to a device whose trial record is split and which reports a probability function rather than a physiological state.

Frontier The paediatric question has been answered well and the answer deserves stating because the fear outlived it. The GAS trial found full-scale IQ at five years of 99.08 against 98.97 between awake-regional and general anaesthesia, a difference of 0.23 with a confidence interval from −2.59 to 3.06. Speculative An hour of anaesthesia in infancy is not measurably harmful, the study was designed as an equivalence trial rather than left to a null result, and it is the best-run piece of clinical science in this brief. Handwave It says nothing about repeated or prolonged exposure, which is where the remaining concern honestly sits.

Speculative Animal use is a real cost here and the mechanism gap inflates it. MAC determination requires a noxious stimulus in a living animal by definition, and the field's habit of testing one candidate target per laboratory multiplies the total. Handwave A factorial, multi-target, multi-laboratory design of the kind section 7 describes would almost certainly use fewer animals than the sequence of single-target studies that is the alternative, which makes the institutional failure a welfare question too.

Frontier And the distributional point, which is the largest. Safe anaesthesia is mostly a staffing achievement rather than a pharmacological one, and between 4.8 and 5.3 billion people lack access to it. Speculative A wider therapeutic index is worth most exactly where the skilled titration that currently substitutes for it is unavailable, and the research that would produce one is funded where the titration already exists. Aging Societies carries the demographic half of this.

Established The civilizational fact is already banked and it is easy to forget: reversible, controllable unconsciousness is what made modern surgery possible, and it arrived in 1846 without a theory. Speculative That sequence — technique first, mechanism never — is the single most interesting thing about this subject, and it is a standing counterexample to the belief that useful intervention requires understanding.

Speculative The consequence for how a civilization allocates research effort is uncomfortable. A century of well-funded, competent work by serious laboratories has produced a target list of three hundred proteins and no mechanism, while the clinical practice improved anyway through monitoring, training and checklists. Handwave If mechanism is not on the critical path to safety, then the honest case for funding it is curiosity and the possibility of a designed agent, and that case should be made in those terms rather than dressed as a patient-safety emergency.

Speculative If the promiscuity answer is right, the implication reaches past anaesthesia. A clinical state produced by summed small perturbations across hundreds of proteins is a category of drug action that pharmacology has no framework for and regulators have no vocabulary for. Speculative Most of medicine is built on the assumption that a drug has a target; anaesthesia would be the standing demonstration that the assumption is optional, and several other old drug classes might turn out to belong to the same category.

Handwave And the far edge, flagged as such. If anaesthetics act on something as general as membrane order or electron transport — the thing that stops a flytrap and a human alike — then the capacity to be switched off is a property of excitable matter rather than of nervous systems. Handwave That would be a real and strange fact about biology, it is not established by anything on this page, and the plant results are the only evidence pointing at it.

These horizons track arbitration and target identification rather than product development, because there is no product programme here to track:

• 10 yr: Frontier Expect the vocabulary change to land first: the 2026 promiscuity survey's recommendation that agents stop being labelled by a single receptor is cheap to adopt and hard to argue with. Speculative Expect the two-pore potassium channel discrepancy to be resolved or quietly dropped, and expect mitochondrial complex I to accumulate evidence faster than any receptor candidate. Speculative Expect no designed agent.
• 25 yr: Speculative This is the window in which an adult-inducible, tissue-restricted manipulation ranking three candidate mechanisms in the same animals is plausible, and it is the only development that would convert the target list into a hierarchy. Speculative The most likely outcome is not that one target wins but that the field formally adopts a distributed-perturbation account and stops asking the question in its current form. Speculative A depth monitor reporting drug effect at a named target rather than a probability of clinical state is possible on this horizon and nobody has specified one.
• 50 yr: Handwave The defensible statement is about shape rather than date. Either the mechanism question acquires a well-posed form and is answered, or it dissolves into separate mechanisms for immobility, unresponsiveness and amnesia that are studied independently and never reunited. Speculative The second is more likely, and it would count as progress even though it would look like abandonment.
• 100 / 250+ yr: Handwave Beyond useful forecasting. Handwave Any specific claim that anaesthesia will be replaced by targeted circuit interruption, or that its mechanism will turn out to be the key to consciousness, is an assertion, and this page does not make one — the subject's own history is of a technique that worked for a century and three-quarters while every theory offered for it failed.

• Depends on One brief on this map produces a result this one is genuinely blocked on. Consciousness Research owes this page a species-portable measure of unresponsiveness that is not behavioural, and its own verdict is that no such measure exists — so the dependency is unmet and known to be unmet rather than merely unaddressed. Sleep Function is not blocking but shares both the anaesthesia-as-model confound and the mitochondrial candidate. Longevity Therapies develops the complex I biology for its own reasons and this page would inherit it.
• Requires (not on this map) Six conditions, in the order of the tokens above. The endpoint. Without a portable measure of unresponsiveness, every genetic result is scored against a spinal reflex and reported as a discovery about consciousness. The proteomics. A method returning ranked target occupancy in intact tissue at a few hundred micromolar isoflurane, because affinity methods assume distinguishing affinity and this ligand has none. The mouse lines. Adult-inducible and tissue-restricted, so that a MAC shift reports a target rather than a developmental compensation — the ambiguity that the 2004 and 2023 potassium-channel results cannot currently be separated by. The protocol. One up-and-down MAC determination method, one noxious stimulus, controlled core temperature, fixed analysis, agreed between laboratories before the next disagreement rather than after it. The registry. Technique and agent recorded alongside procedure, which would give the field a denominator it currently misses by a factor of three and would supply the mitochondrial-disease cohort section 7 asks for. The funder. Every agent in wide use is generic, so no commercial sponsor captures the return; this is the binding constraint and it is the cheapest one to state and the hardest to remove.
• Enables A ranked mechanism would enable a component-selective agent — immobility without cardiovascular depression, or unconsciousness without either — and a depth monitor reporting drug effect at a target instead of a probability of clinical state. It would give Consciousness Research a pharmacological handle with a known mode of action rather than a calibrated correlate, and would sharpen Precision Neuropsychiatry's use of ketamine. No typed enabling edge is claimed, because a mechanism with three hundred candidate targets enables nothing on a schedule.
• Adjacent Membrane biophysics, structural biology, mitochondrial bioenergetics, invertebrate genetics and plant electrophysiology supply most of the numbers on this page, and only the first two think of themselves as contributing to anaesthesia. Within this map: Consciousness Research for the endpoint problem, Sleep Function for the shared state confound, Human Hibernation for the induced-state pharmacology, Precision Neuropsychiatry and Brain-Computer Interfaces for the downstream clinical instruments, and Aging Societies for the delirium demography.

Frontier Claim: we know how general anaesthetics work — they act on GABA-A receptors. Established For etomidate and propofol there is real genetic support: the β3(N265M) knock-in strongly attenuates their immobilising and hypnotic actions. Established For the volatile agents the standard summary is the opposite: GABA-A receptors appear not to be important for immobility where inhalational agents are concerned. Frontier And a 2026 survey of 310 in vitro studies finds propofol itself more potent at a neuronal nicotinic receptor (3.13 µM) and at HCN1 (4.19 µM) than at GABA-A (5.44 µM), concluding that the single-receptor labels are unhelpful and misleading.

Frontier Claim: the Meyer-Overton correlation shows anaesthetics work on the lipid membrane. Established The correlation is real across four to five orders of magnitude of lipid solubility. Frontier It is also satisfied by hydrophobic binding pockets on proteins, and four observations break the lipid reading: the cutoff in homologous series above about 13 carbons for alcohols, stereoisomers with identical partition coefficients and different potencies, lipid-soluble non-immobilisers such as F6, and the fact that a 1 °C temperature change perturbs a membrane as much as a clinical dose without causing unconsciousness.

Frontier Claim: MAC measures anaesthetic depth. Established MAC is the concentration at which 50% of subjects do not move to a surgical stimulus, and that endpoint is spinal: brain-only delivery in the goat needed 2.5-fold to 4-fold more agent, and removing a rat's forebrain does not change isoflurane MAC. Speculative MAC-awake is about a third of MAC, so a patient at 1 MAC is well past unconsciousness and the number being titrated is a reflex threshold. Speculative Treating MAC as depth is the single most consequential conflation in the clinical literature.

Frontier Claim: MAC is reproducible to within a few per cent. Established Its population standard deviation is about 10%, which is why 1.2 MAC rather than 1.0 MAC is what immobilises about 95% of patients. Established It declines about 6% per decade of age, falls 4 to 5% per °C of hypothermia and up to 30% in pregnancy, rises with chronic alcohol use, and differs between species: isoflurane is about 1.2% in humans, 1.3% in dogs and 1.6% in cats. Speculative What is genuinely conserved is the rank order of potency across species, not the number.

Speculative Claim: depth-of-anaesthesia monitors prevent awareness. Frontier The trial record is split three ways: a benefit at a number-needed-to-treat of 138 and about $2,200 per case prevented in one trial; no benefit against end-tidal monitoring in a second; fewer awareness episodes in the end-tidal arm than the monitored arm in a third. Frontier A health technology assessment pooled an odds ratio of 0.45 (95% CI 0.25 to 0.81) overall with confidence intervals crossing 1 in the inhalational subgroup, and the devices are described as reporting a probability function of clinical state rather than a physiological parameter.

Frontier Claim: anaesthesia in infancy damages the developing brain. Established The GAS trial randomised 722 infants and found five-year full-scale IQ of 99.08 against 98.97, a difference of 0.23 with a confidence interval of −2.59 to 3.06 — strong evidence of equivalence for a median 54 minutes of general anaesthesia. Speculative The honest residual is that this tests one short exposure and says nothing about repeated or prolonged anaesthesia, which is where the remaining uncertainty lives.

Speculative Claim: anaesthetics prove consciousness is quantum, because a microtubule drug delays them. Frontier The underlying measurement is real and modest: 0.75 mg/kg of epothilone B delayed loss of righting reflex by an average of 69 seconds in eight male rats, Cohen's d = 1.9, P = 0.0016. Handwave A 69-second shift in a righting-reflex latency in one sex of one species is evidence that microtubule stabilisation touches anaesthetic sensitivity; it is not evidence about the physical basis of consciousness. Speculative The authors say replication is needed, and the correct treatment is to test the effect on its own terms rather than to dismiss it because of the theory attached to it.

Speculative Claim: anaesthetics work on plants too, so plants are conscious. Established The first half is measured: 15% diethyl ether abolished Venus flytrap action potentials with recovery about 15 minutes after removal, stopped Mimosa leaf movement within an hour with recovery over seven, and stopped pea-tendril circumnutation immediately; 80% xenon did likewise. Handwave The second half does not follow at all. Speculative What the plant results establish is that a general anaesthetic can reversibly abolish electrical signalling and organ movement in a system with no neurons, which is a constraint on mechanism, not an argument about experience.

Frontier Claim: the target has nearly been found; it is a matter of better methods. Established A 2019 review estimates that 10 to 15% of detectable proteins are anaesthetic binding targets, over 300 proteins and possibly tenfold more, and states that the mechanism of anterograde amnesia is still unclear. Frontier Forty years of unbiased genetic screening in nematodes returned a list including syntaxin alleles spanning a 33-fold sensitivity range and a complex I subunit, and the field's own summary is that the genetic data do not identify a single protein or channel uniquely responsible. Speculative That is not a methods problem. It is a result.

Frontier Claim: a knockout that changes anaesthetic sensitivity identifies the target. Established TREK-1 knockout mice were reported resistant to volatile anaesthetics in 2004; a 2023 study reported that TREK-1 and TREK-2 knockouts are not resistant to halothane or isoflurane. Speculative Germline deletion reorganises a nervous system over development, so a sensitivity shift may report compensation rather than target — which is why section 7's decisive experiment is adult-inducible and tissue-restricted rather than a further knockout.

Handwave Claim: understanding the mechanism would make anaesthesia much safer. Speculative The historical record argues the other way. Anaesthetic mortality fell by orders of magnitude over the twentieth century through monitoring, training, pulse oximetry, capnography and checklists, none of which required knowing the mechanism, and the GAS trial answered the largest outstanding safety question without it. Speculative The honest case for mechanism research is a component-selective agent and a monitor that measures drug effect rather than a correlate — both real, neither a safety emergency.