Why this poison is interesting

Baclofen is a GABA-B agonist that works in the spinal cord, and the central fact of its pharmacology is that swallowing it is a poor way to get it there. The UK label states plainly that in cerebrospinal fluid active substance concentrations are approximately 8.5 times lower than in the plasma.¹ An oral dose therefore exposes the whole body to a concentration nearly nine times higher than the one reaching the tissue it is aimed at. Every unwanted effect of oral baclofen follows from that ratio.

The intrathecal pump exists to invert it. Deliver the drug directly into the CSF and the same spinal effect is achieved at a systemic exposure the label describes as substantially lower than after oral administration.² The engineering solves the pharmacological problem completely — and creates a second one, because it makes the patient dependent on a mechanical device sitting inside them. The commonest reasons the label gives for abrupt interruption are not clinical at all: catheter disconnection, a low reservoir, end of pump battery life, device malfunction.²

That is why this page has two poisonings in it rather than one, and why the second is the dangerous one. Abrupt discontinuation of intrathecal baclofen has resulted in sequelae including a hyperactive state with rapid uncontrolled spasms, hyperthermia, tachycardia and symptoms consistent with neuroleptic malignant syndrome... in rare cases this has advanced to seizures/status epilepticus, rhabdomyolysis, coagulopathy, multiple organ failure and death.² Overdose of this drug is usually survived. Withdrawal of it is what kills.

A poison is a drug whose kinetics have escaped its pharmacology.

The toxic principle

Baclofen is a GABA derivative that is chemically unrelated to other antispastic agents and acts at the spinal level, depressing monosynaptic and polysynaptic reflex transmission probably by stimulating the GABA-B receptors, which in turn inhibits the release of the excitatory amino acids glutamate and aspartate.¹ Note the label's own hedge — probably — on a mechanism that is otherwise universally taught.

The GABA-B receptor is metabotropic rather than a chloride channel, which distinguishes this poisoning from the benzodiazepine one at the level of the receptor. Presynaptic GABA-B activation reduces calcium entry into the terminal and so reduces transmitter release; postsynaptic activation opens potassium channels and hyperpolarises the cell. Both actions reduce excitability without touching the neuromuscular junction — the label is explicit that neuromuscular transmission is unaffected.¹ A flaccid, areflexic patient in baclofen poisoning has an intact junction and an intact muscle; the lesion is entirely central.

Why the pupils and the reflexes lie in this poisoning

The overdose label lists impaired pupillary reflex, generalised muscular hypotonia, hyporeflexia or areflexia, coma, respiratory depression and an abnormal EEG showing burst suppression pattern and triphasic waves.¹ That is, item for item, a brainstem examination consistent with catastrophic brain injury, produced reversibly by a drug. A published case describes exactly this: a 37-year-old man with an intrathecal pump who developed fixed dilated pupils and a low Glasgow Coma Scale intraoperatively, whose CT showed no acute intracranial pathology, and whose conscious level improved after the pump was aspirated.⁴ A single case report, and it is treated as one — but the label's independent list of the same findings is what makes it credible. Established

The overdose picture is otherwise a long and rather undifferentiated list: somnolence, coma, respiratory depression, tinnitus, confusion, hallucination, agitation, convulsion, accommodation disorder, myoclonus, peripheral vasodilatation, hypotension or hypertension, bradycardia or tachycardia, hypothermia, salivary hypersecretion, raised hepatic enzymes and rhabdomyolysis.¹ The paired opposites are worth noticing. A poison that can produce either direction of blood pressure and either direction of heart rate is not acting at one autonomic site, and no bedside vital-sign pattern will identify it.

Toxicokinetics

Oral baclofen has kinetics that look, on paper, ideal for extracorporeal removal — and EXTRIP recommends against it for the acute overdose anyway. The gap between those two statements is the most instructive thing in this section.

Oral baclofen — a molecule that looks dialysable, in a poisoning that mostly is not worth dialysing
Parameter	Therapeutic	In overdose	Why it changes
Absorption	Rapidly and completely absorbed; peak at 0.5 to 1.5 hours after 10–30 mg; AUC proportional to dose¹	No absorptive ceiling — exposure rises with the amount swallowed	Unlike gabapentin, there is no saturable transporter to blunt a large ingestion. Dose-proportional absorption is a therapeutic convenience that removes a toxicological safety net.
Distribution	Vd 0.7 L/kg; protein binding about 30%, constant from 10 ng/mL to 300 µg/mL¹	Unchanged — binding does not saturate across a 30,000-fold concentration range	A small volume and low, concentration-independent binding is the textbook profile of a dialysable poison. The label's own binding statement spans the whole plausible overdose range, so no free-fraction surprise is available.
CSF penetration	CSF concentrations approximately 8.5 times lower than plasma¹	Measured at therapeutic oral doses; the label does not state whether the ratio holds at overdose concentrations	This single figure explains both halves of the page. It is why oral dosing needs so much drug for a spinal effect, why the systemic burden in oral overdose is disproportionate to the central one, and why intrathecal delivery — which bypasses the ratio entirely — needs microgram quantities.
Metabolism	Metabolised to only a minor extent; deamination to β-(p-chlorophenyl)-4-hydroxybutyric acid, pharmacologically inactive¹	Irrelevant — there is no pathway to saturate	No bioactivation, no active metabolite, nothing to block. The label adds that hepatic impairment is unlikely to alter kinetics significantly because the liver does not play a significant role in the disposition of baclofen.¹
Elimination	Largely unchanged in urine — about 75% of the dose within 72 hours, only about 5% of that as metabolites¹	Collapses with renal function	Patients with renal impairment can develop signs of overdose even on low doses of oral Lioresal.¹ That sentence is the entire justification for EXTRIP's split recommendation and describes a poisoning that occurs at a licensed dose.
Half-life	3 to 4 hours in adults¹; about 5.1 hours reported in children aged 2 to 12¹	Prolonged in renal impairment	Short enough that intact kidneys resolve an oral overdose without help — which is exactly the argument against dialysing the patient who has them.
Dialysability	—	Physically removable, but recommended against in acute overdose. The label suggests generous fluids and a diuretic, and that haemodialysis (sometimes unscheduled) may be useful in severe poisoning associated with renal failure¹	EXTRIP suggests against ECTR in severe acute baclofen poisoning, and for it in severe toxicity from therapeutic baclofen in kidney impairment — specifically where there is associated coma requiring mechanical ventilation. Both weak recommendations on very low quality evidence. Where ECTR is performed, intermittent haemodialysis is recommended over other modalities and cessation is on clinical improvement (both strong recommendations).³ The label and the workgroup agree, and both make renal failure the condition.

Metabolism and the metabolites

There is almost nothing here, and as with the gabapentinoids the emptiness is the finding. Baclofen is metabolised to only a minor extent, by deamination, to a single named product that the label states is pharmacologically inactive.¹ About 75% of a dose is recovered in urine within 72 hours, of which only about 5% is metabolite.¹

Baclofen — two routes to the same receptor, separated by a factor of 8.5

Swallowed — the oral routeCompletely absorbed, peak 0.5–1.5 h, AUC proportional to dose¹
Infused — the intrathecal routePlaced directly in CSF; systemic availability substantially lower than after oral administration²
The blood-brain barrier — the step the pump exists to skipCSF concentrations about 8.5 times lower than plasma after oral dosing¹. The oral route pays for a spinal effect with a systemic one
Spinal GABA-B receptorsReduced release of glutamate and aspartate; monosynaptic and polysynaptic reflexes depressed¹
β-(p-chlorophenyl)-4-hydroxybutyric acidPharmacologically inactive¹ — about 5% of what is excreted
Urine — about 75% of the dose in 72 hours¹The only elimination route that matters, and the one that fails in the patients who poison themselves at licensed doses¹

Elimination and accumulation

For the oral drug, elimination is a single renal route with a short half-life, and the accumulation trap is correspondingly simple and correspondingly common: give an unchanged dose to a patient whose glomerular filtration rate has fallen. The label states the consequence without hedging — signs of overdose even on low doses¹ — and the intrathecal label repeats the warning for its own route, noting that severe neurological outcomes have been reported with oral dosing in renal impairment.²

For the intrathecal drug, the accumulation question is replaced by a delivery question, and the important direction is loss rather than excess. Withdrawal is not a rebound of spasticity; it is a distinct hypermetabolic syndrome. The label describes increased spasticity, pruritus, paraesthesia and hypotension progressing to a hyperactive state with rapid uncontrolled spasms, hyperthermia, tachycardia and symptoms consistent with neuroleptic malignant syndrome, e.g. altered mental status and muscle rigidity, and in rare cases to seizures/status epilepticus, rhabdomyolysis, coagulopathy, multiple organ failure and death.²

A syndrome that is designed to be misdiagnosed

The label lists what intrathecal baclofen withdrawal resembles, and the list is a differential diagnosis of critical illness: autonomic dysreflexia, infection (sepsis), malignant hyperthermia, neuroleptic-malignant syndrome, or other conditions associated with a hypermetabolic state or widespread rhabdomyolysis.² The patients at risk are frequently those with spinal cord injury, in whom autonomic dysreflexia is independently plausible and the label warns that abrupt withdrawal may itself precipitate it.² In most cases, symptoms of withdrawal appeared within hours to a few days following interruption of baclofen therapy.² The specific treatment is restoration of intrathecal baclofen at or near the previous dose; where that is delayed, oral or enteral baclofen or benzodiazepines may prevent potentially fatal sequelae, and the label adds that oral baclofen alone should not be relied upon to halt the progression.²

That last clause is a direct consequence of the 8.5-fold gradient in the kinetics table. Oral baclofen cannot reproduce an intrathecal concentration at any tolerable systemic dose, which is precisely why the pump was implanted. The label is not being cautious; it is stating an arithmetic fact about the route. Inferred

Target organs — and why those

Spinal cord

TargetGABA-B receptors in the dorsal horn and on interneurons; monosynaptic and polysynaptic reflex arcs

Why hereThis is the intended target and the reason the drug exists. The label locates the action at the spinal level, depressing reflex transmission and inhibiting release of glutamate and aspartate.¹ The organ is selected by the therapy rather than by the poison — the intrathecal route was engineered specifically to concentrate drug here and nowhere else. In oral overdose the cord is reached last and least, because of the 8.5-fold CSF gradient.¹ Established

At the bedsideGeneralised muscular hypotonia, hyporeflexia or areflexia in overdose¹; uncontrolled spasms and rigidity in withdrawal². The same structure produces opposite signs depending on the direction of the error.

Brainstem

TargetRespiratory centres and pupillary reflex pathways

Why hereBaclofen crosses into the brain as well as the cord, and at high concentration the depression is not confined to spinal reflexes. This card exists because of what it can imitate: the label lists respiratory depression, coma and impaired pupillary reflex¹, and a published case describes fixed dilated pupils with a normal CT resolving after pump aspiration.⁴ The brainstem is targeted because GABA-B receptors are there too, and because nothing in this drug's distribution keeps it out. Established

At the bedsideApnoea, coma and non-reactive pupils. Potentially fully reversible, which is the reason the finding matters more here than almost anywhere else in the library.

Cerebral cortex

TargetCortical excitability; the EEG

Why hereIncluded because the label supplies an objective measurement rather than a symptom: overdose produces an abnormal electroencephalogram (burst suppression pattern and triphasic waves, generalised slowing on EEG).¹ Burst suppression is the electrical signature of profound cortical depression, and triphasic waves are more usually associated with metabolic encephalopathy. A drug that reproduces the EEG of hepatic or uraemic encephalopathy will also reproduce the diagnosis. Established

At the bedsideComa, confusion, hallucination and convulsion¹. Seizures may also occur during recovery from an overdose of the intrathecal drug².

Skeletal muscle — in withdrawal, not in overdose

TargetNot a receptor; the consequence of uncontrolled spasm and hyperthermia

Why hereThe only organ on this page injured by the absence of the drug. Loss of GABA-B mediated inhibition releases the spinal reflex arcs the drug was suppressing, producing rapid uncontrolled spasms and rigidity; sustained contraction and hyperthermia then produce muscle breakdown. The label's own progression is spasms, then hyperthermia and rigidity, then rhabdomyolysis, coagulopathy, multiple organ failure and death.² Neuromuscular transmission itself is never involved — the label states it is unaffected by baclofen¹ — so this is a central lesion with a peripheral casualty. Inferred

At the bedsideRigidity, hyperthermia, raised creatine kinase, acute kidney injury and coagulopathy². Rhabdomyolysis is also listed among the features of oral overdose¹, where immobility rather than spasm is the likelier route to it.

Timeline of effects

Two clocks — an oral overdose that declares itself quickly, and a pump failure that does not

Time

What you seeWhat is happening

0–1.5 h (oral)Absorption
What you seeIncreasing drowsiness, dizziness, sometimes agitation or hallucination.¹
What is happeningRapid and complete absorption, peak plasma concentration at 0.5 to 1.5 hours, AUC proportional to dose.¹ No manufacturing step and no latent phase on this route.
1–12 h (oral)The full picture
What you seeComa, respiratory depression, areflexia, impaired pupillary reflex, hypothermia — an examination that reads as catastrophic.¹
What is happeningPeak central concentration. The paired opposites appear here: hypotension or hypertension, bradycardia or tachycardia.¹ The EEG may show burst suppression and triphasic waves.¹
12–72 h (oral)Elimination — if the kidneys work
What you seeGradual return of tone, reflexes, pupils and consciousness.
What is happeningHalf-life 3 to 4 hours¹ and about 75% of the dose excreted unchanged within 72 hours.¹ In renal impairment this phase may not happen at all, and it is the arm in which EXTRIP suggests dialysis.³
Hours to days after a pump stopsThe silent interval
What you seeNothing, or only escalating spasticity, itch, pins and needles — and priapism.² Easy to attribute to the underlying condition.
What is happeningThe device has already failed — disconnection, empty reservoir, exhausted battery, malfunction.² In most cases, symptoms of withdrawal appeared within hours to a few days following interruption.² The gap is not pharmacological. It is the interval between a machine stopping and a person noticing, and no measurement of the patient will close it.
After the intervalWithdrawal
What you seeUncontrolled spasms, hyperthermia, tachycardia, rigidity, altered mental state; then status epilepticus, rhabdomyolysis, coagulopathy, multi-organ failure, death.²
What is happeningLoss of GABA-B inhibition on reflex arcs adapted to its presence. Restoration of intrathecal baclofen at or near the previous dose is the specific treatment; oral baclofen alone cannot be relied on to stop the progression.²

A gap made by a machine

Baclofen — not the poisoning but its withdrawal — an implanted pump that has silently stopped delivering, with hours to days before anything appears

The other 28 kinds of latent phase in this library

Amphetamines and MDMA — a hormone acting normally on a kidney behaving normally, while the patient supplies the water
Anticoagulants — clotting factors that were already made, still working — the poison stops production and nothing happens until the existing stock decays
Antipsychotics — a physical object in the stomach — extended-release quetiapine forming a pharmacobezoar
Arsenic — a tissue declaring on its own timetable rather than the poison's — the arsenic is excreted within days, but the nail that was growing while it circulated does not show its white transverse line for several weeks
Arsine and stibine — a red cell mass haemolysing faster than a kidney can cope with — the exposure is over, the haemolysis is silent until the urine changes colour, and the renal failure that follows is the cause of death
Beta-blockers — a repolarisation lesion waiting for an ectopic beat to fall inside it — sotalol prolongs the QT and then, for hours, nothing happens
Calcium-channel blockers — a tablet that has not yet dissolved
Carbon monoxide — an inflammatory process continuing after the poison itself has gone
Chlorine and ammonia — paraquat's cause at its shortest — the time an injured alveolar epithelium takes to leak
Cocaine — a package that has not yet failed — the calcium-channel blocker's cause moved outside the tablet, where wrapping integrity cannot be measured
Digoxin — distribution of the drug to its target — which that page argues is not a latent phase in the paracetamol sense at all
Ethylene glycol — paracetamol's cause with a different enzyme — glycolate accumulating behind glycolate oxidase while the patient looks merely drunk
GHB and GBL — not the poisoning but its withdrawal — a half-life under an hour means admission for any reason interrupts round-the-clock redosing
Hydrofluoric acid — an ion diffusing far enough to reach a nerve ending — and the thinner the solution, the further it travels before anybody feels it
Iron — a true remission — the corrosive injury settling while absorbed iron moves into the cells it will poison, so the patient improves on their way to the dangerous phase
Ketamine — nitrous oxide's cause on a longer clock — urothelial damage accumulating over months to years of repeated exposure
Lead — bone giving lead back — a store with a half-life of 10 to 30 years releasing its contents when pregnancy, lactation, menopause or osteoporosis resorbs it, so the blood concentration rises after the exposure has ended
Lithium — transport across cell membranes
Mercury — distribution on two clocks — tissue concentrations peaking within 24 hours everywhere except the brain, which is not reached until 2 to 3 days, and which then cannot let the poison out again
Methaemoglobin inducers — paracetamol's cause at its shortest — one to four hours manufacturing the toxic species
Methanol — paracetamol's cause again — formate accumulating behind a folate-dependent disposal step that primates perform poorly
Nitrous oxide — damage accumulating to a threshold
Opioids — an antidote wearing off before the poison does — renarcotisation, the only gap in this set that treatment creates rather than reveals
Organophosphate insecticides — a fat store emptying — and, separately, a second and unexplained lesion declaring itself at a neuromuscular junction the first phase had already left
Paracetamol — time spent manufacturing a toxic metabolite
Paraquat — the body responding to an injury that is already complete
Sodium-channel blockade — a gap that cannot be shortened
Thallium — hair on its own clock rather than the poison's — sensory symptoms come first and the alopecia that makes the diagnosis obvious follows them, well after the interval in which treatment is recommended

Every other gap in this set belongs to the poison. This one belongs to a device: the pump has stopped, the drug is already falling, and nothing about the patient announces it until the syndrome arrives hours to days later.² It is drawn as a hole because during it there is genuinely nothing to find — which is why the label puts the emphasis on refill scheduling, alarms and monitoring rather than on clinical vigilance.

What the mechanism predicts at the bedside

Fixed dilated pupils and areflexic coma can be baclofen, and can be fully reversible. The label lists impaired pupillary reflex, areflexia, coma and burst suppression among ordinary overdose features¹, and a published case of intrathecal toxicity presented with fixed dilated pupils and a normal CT before recovering after pump aspiration.⁴ Any brainstem-death assessment requires exclusion of drug effect, and this is the drug that makes the point.
Ask whether there is a pump, and if there is, find out when it was last refilled. The label names low reservoir volume, battery exhaustion, catheter disconnection and device malfunction as the common causes of abrupt interruption.²
Priapism, new itch, paraesthesia and escalating spasticity are early withdrawal signs, not incidental findings — the label singles out priapism as one to educate patients and carers about.²
A hyperthermic, rigid patient with a spinal cord injury and a pump is withdrawal until proven otherwise, even though sepsis, autonomic dysreflexia, malignant hyperthermia and neuroleptic malignant syndrome are all on the label's own list of resemblances.²
Oral baclofen will not rescue intrathecal withdrawal on its own², because the CSF-to-plasma gradient the pump was implanted to defeat is still there.¹
Measure renal function. About 75% leaves unchanged in the urine¹ and toxicity occurs on low doses when that route narrows¹ — the presentation is a confused older patient on a stable prescription, not an overdose.
Do not expect dialysis to be offered for an ordinary acute overdose. EXTRIP suggests against it with intact kidneys and for it in therapeutic toxicity with kidney impairment, especially with coma requiring ventilation.³
Hypotension and hypertension, bradycardia and tachycardia are all listed¹ — no vital-sign pattern rules this poisoning in or out.
Seizures may occur during recovery, not only during the poisoning.² Improvement is not the end of the monitoring period.

The antidote, from the poison's side

Both labels state that there is none. No specific antidote is known¹; there is no specific antidote for treating overdoses of intrathecal baclofen.² What the intrathecal label offers instead is something no other page in this library contains: a physical instruction for removing the poison from the compartment it is acting in.

Stopping and draining the pump: The specific measure for intrathecal overdose, and the one with a mechanism — the drug is taken out of the CSF rather than out of the blood.² The published case above records improvement in conscious level after aspiration.⁴
Restoring intrathecal delivery: The specific measure for intrathecal withdrawal: restoration of intrathecal Lioresal at or near the same dosage as before therapy was interrupted.² It is the mirror image of the paragraph above, on the same catheter.
Benzodiazepines: Named on both labels — diazepam for convulsions in oral overdose¹ and cautiously intravenously in intrathecal overdose² — and, in withdrawal, as a GABA-ergic holding measure that may prevent potentially fatal sequelae while intrathecal delivery is restored.² Acting at a different receptor on the same inhibitory system; see benzodiazepines.
Fluids and diuresis: The oral label suggests generous quantities of fluid... possibly together with a diuretic¹, on the reasoning that elimination is renal. It is a supportive measure aimed at the route that already works.
Haemodialysis: Physically effective, and conditional. May be useful in severe poisoning associated with renal failure¹; EXTRIP suggests against it in acute overdose and for it in therapeutic toxicity with kidney impairment.³
Activated charcoal: Considered early after oral ingestion on general principles. Neither label makes a baclofen-specific claim for it and this page does not manufacture one.
Flumazenil: Not applicable and worth saying so. Baclofen acts at GABA-B; flumazenil is an antagonist at the benzodiazepine site of the GABA-A receptor. Two drugs on the same neurotransmitter and no pharmacological contact between them.

Critical appraisal

The GABA-B mechanism is the label's own hedged claim, not this page's confident one. Lioresal's section 5.1 says the action is probably by stimulating GABA-B receptors.¹ The page reproduces the hedge rather than smoothing it out. No evidence-tier downgrade is applied to it, because a single qualifying adverb in a summary of pharmacodynamic properties is not a source contesting the mechanism — the house rule requires a citation for the doubt, and none has been found.
The brainstem-death mimicry claim rests on a label list plus one case report, and is written to show that. The label independently lists impaired pupillary reflex, areflexia, coma and burst suppression as overdose features¹; the case supplies the clinical gestalt and the reversal after pump aspiration in one patient.⁴ A single case cannot establish frequency, and this page claims none.
The 8.5-fold CSF gradient is a single figure from one label and the page leans on it heavily. It is quoted for oral dosing in healthy pharmacokinetics; whether the same ratio holds at overdose concentrations or in the presence of an intrathecal catheter is not stated by the source and is not assumed here. An auditor should check that no argument on the page silently extends it beyond what ¹ says.
The statement that oral baclofen cannot substitute for intrathecal delivery is the label's instruction plus this page's explanation. The instruction is explicit²; the gradient-based reason for it is this page's inference and is badged as such.
No lethality figure, no toxic dose and no concentration threshold appears here. Both labels describe severity qualitatively, and the risk-assessment thresholds belong to TOXBASE and NPIS.
No evidence-tier downgrade appears anywhere on this page. The candidates were considered and each failed the house test: the GABA-B mechanism (a hedge, not a contest), the withdrawal syndrome's resemblance to neuroleptic malignant syndrome (the label asserts the resemblance, nothing contests it), and the value of dialysis (EXTRIP grades its own evidence as very low, which is a statement of quality rather than a citation refuting a claim). An absence of strong evidence is not a citation for the doubt — the same reasoning that withheld a badge from iron and arsine.
The paediatric kinetics quoted are from a 2.5 mg dose in children aged 2 to 12¹ and are given to show that the profile differs, not to support any dosing inference. Vd of 2.58 L/kg in that group against 0.7 L/kg in adults is a large difference and this page does not build on it.

References

1
Lioresal Tablets 10 mg (baclofen) — Summary of Product Characteristics. electronic medicines compendium, product 1279. Sections 4.9 (Overdose), 5.1 (Pharmacodynamic properties) and 5.2 (Pharmacokinetic properties). medicines.org.uk/emc/product/1279
2
Lioresal Intrathecal Infusion 10 mg/20 ml (baclofen) — Summary of Product Characteristics. electronic medicines compendium, product 1033. Sections 4.4 (Special warnings and precautions for use, including treatment withdrawal) and 4.9 (Overdose). medicines.org.uk/emc/product/1033
3
EXTRIP Workgroup. Baclofen — recommendations. Extracorporeal Treatments in Poisoning Workgroup. Quoted with the recommendation strength and evidence quality attached to each statement. extrip-workgroup.org/baclofen
4
Alamri A, Mostofi A, Aziz T, et al. Intrathecal baclofen overdose mimicking brainstem death during deep brain stimulation surgery for pain. Annals of the Royal College of Surgeons of England 2022 Sep;104(8):e232–e235. PMID 35616338. A single case report, cited for the clinical presentation and its reversal after pump aspiration, not for frequency.

Baclofen

At a glance