Why this poison is interesting

Lamotrigine is the entry in this library where the therapeutic mechanism and the toxic mechanism are the same mechanism, applied to two different organs. The UK label states that lamotrigine is a use- and voltage-dependent blocker of voltage gated sodium channels, that it inhibits sustained repetitive firing of neurones and inhibits release of glutamate, and that these effects contribute to the anticonvulsant properties.¹ Blocking the neuronal sodium channel is how it stops seizures. Blocking the cardiac sodium channel is how it kills — and the drug does not distinguish between them.

This makes lamotrigine the prescribed, swallowed-on-purpose member of the sodium-channel blockade family, alongside the tricyclics and the class I antiarrhythmics. In overdose it produces exactly what that family produces: QRS broadening (intraventricular conduction delay) and QT prolongation, with the label warning that broadening of QRS duration to more than 100 msec may be associated with more severe toxicity.¹ The bedside ECG of a serious lamotrigine overdose is a tricyclic tracing, produced by an anticonvulsant.

The second surprise is the seizure. A systematic review of 51 published overdoses found that the commonest potentially life-threatening effect was seizures (55%), ahead of coma, hypotension, and wide-complex tachycardia.² An anticonvulsant is, in overdose, one of the more reliably convulsant drugs in the library. That is the paradox the page exists to explain, and it is a genuine one: the mechanism that suppresses pathological repetitive firing at therapeutic concentration becomes proconvulsant at toxic concentration, a pattern shared with the other sodium-channel-blocking antiepileptics and with the tricyclics.

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

Which of the four patterns is this?

None of the four — because nothing escapes. The mechanism is unchanged; only the tissue is wrong. There is no bioactivation (the glucuronide is inactive), no saturable pathway that flips, no cofactor depletion, no protein-binding trap — the label says displacement from its ~55% binding is very unlikely to cause toxicity.¹ What lamotrigine does in overdose is the identical use- and voltage-dependent sodium-channel block it does in therapy, now expressed in the myocardium where it is dangerous and, above a threshold, in a brain where the same block paradoxically lowers the seizure floor. This is a poisoning of selectivity lost, not of kinetics escaped — the closest structural cousin in the library is the way baclofen's danger is a matter of compartment rather than metabolism.

The toxic principle

The toxic principle is the parent drug and its block of the fast voltage-gated sodium channel. The label's description of the therapeutic action is, word for word, a description of the toxic one: a use- and voltage-dependent blocker of voltage gated sodium channels.¹ Use-dependence means the block deepens the faster the channel is cycling — which is why it preferentially silences a rapidly, repetitively firing epileptic focus, and why, in the heart, it bites harder at higher rates.

In the myocardium the consequence is the classic sodium-channel-blockade picture: slowed phase-0 depolarisation, a widening QRS, and — because the block is rate- and use-dependent — a tendency to wide-complex tachycardia. The label reports QRS broadening (intraventricular conduction delay) and QT prolongation in overdose and treats a QRS beyond 100 ms as a marker of severity.¹ The systematic review found wide-complex tachycardia or cardiac arrest in 6% of cases and located the cardiovascular toxicity primarily in adult patients above a serum concentration of about 25 mg/L.² The mechanism is established for the therapeutic action; its extension to cardiac conduction in overdose is well supported by the ECG findings, though the human channel-level demonstration in poisoning rests on the phenotype rather than on electrophysiology. Inferred

Why an anticonvulsant convulses

It is not a contradiction so much as a dose effect. At therapeutic concentration the sodium-channel block preferentially damps the sustained high-frequency firing of a seizure focus while leaving normal transmission intact — the label shows 240 mg produces no… difference from placebo on CNS testing.¹ At toxic concentration the drug's actions broaden and the net effect on cortical excitability reverses: seizures were the single commonest life-threatening feature (55%) in the systematic review, and in children the lowest seizure-associated concentration was 3.8 mg/L against 25.6 mg/L in adults, suggesting children may be more susceptible to CNS toxicity.² The proconvulsant reversal is shared with the other sodium-channel-blocking antiepileptics and with the tricyclics; the precise mechanism is not established, and this page badges the relationship, not a molecular cause. Inferred

Toxicokinetics

The kinetics are simple to state and consequential: rapid complete absorption, a small volume of distribution, moderate protein binding, glucuronidation as the only real clearance route, and a half-life of about 33 hours. The last figure is why lamotrigine toxicity is a problem measured in days, and the glucuronidation route is why co-medication changes the half-life by a factor of five.

Lamotrigine — a small volume and a long half-life, cleared entirely by glucuronidation
Parameter	Therapeutic	In overdose	Why it changes
Absorption	Rapidly and completely absorbed… with no significant first-pass metabolism; Cmax at ~2.5 h; food delays the peak but not the extent¹	Complete and predictable — no absorptive ceiling	There is no saturable transporter to blunt a large ingestion, and no first-pass loss to soften it. What is swallowed is absorbed, which is why the severity tracks the dose and the serum concentration so closely.²
Distribution	Vd 0.92–1.22 L/kg; protein binding ~55%, and displacement… very unlikely to cause toxicity¹	A relatively small volume — most of the drug is accessible, not sequestered	A small volume of distribution is the one kinetic feature that would ordinarily favour dialysis — and it is the reason haemodialysis has been attempted. The moderate protein binding and the long half-life pull the other way.
Metabolism	Glucuronidation by UDP-glucuronyl transferases; induces its own metabolism modestly; no significant cytochrome-P450 interactions¹	The glucuronide is inactive — no toxic metabolite is made	This single fact controls the half-life and the interactions. Because clearance is one conjugation route, anything that induces or inhibits glucuronidation moves the half-life dramatically — see the interaction callout below.
Elimination	Clearance primarily metabolic; <10% excreted unchanged in urine, ~2% in faeces; clearance ~30 mL/min¹	Metabolic, not renal — renal failure alone does not cause accumulation	Because elimination is metabolic rather than renal, the accumulation trap here is a drug interaction, not kidney impairment — the mirror image of the renally cleared drugs elsewhere in this band.
Half-life	~33 h (range 14–103 h); ~14 h with carbamazepine or phenytoin; ~70 h with valproate¹	Long, and swung five-fold by co-medication	A 33-hour half-life means a serious overdose is a multi-day problem, and the observation and monitoring period is correspondingly long — the conduction abnormality does not resolve in an afternoon.
Dialysability	—	Doubtful. There is no experience with haemodialysis as treatment of overdose. In six volunteers with kidney failure, 20% of the lamotrigine was removed… during a 4-hour haemodialysis session¹	Only a fifth removed in four hours, against endogenous clearance, is a weak return for an invasive procedure. There is no EXTRIP recommendation for lamotrigine; haemodialysis has been used in individual severe cases⁴ but the label's own figure is why it is not routine.

Metabolism and the metabolites

There is almost nothing toxic to say about the metabolism, and that is itself the point: lamotrigine is not bioactivated. It is conjugated by UDP-glucuronyl transferases to an inactive glucuronide that is excreted in urine, with less than 10% of the parent leaving unchanged.¹ No reactive intermediate, no oxon, no NAPQI — the toxicity is entirely the parent drug's, and the metabolic step only removes it.

Lamotrigine — one conjugation route, no toxic metabolite, and a half-life set by whatever else is on board

LamotrigineThe active and the toxic species; blocks voltage-gated sodium channels in neurone and myocardium¹
UDP-glucuronyl transferase (UGT) — the single clearance routeInduced by carbamazepine / phenytoin (half-life → ~14 h); inhibited by valproate (→ ~70 h)¹ — the interaction that governs everything
Lamotrigine glucuronideInactive — no toxic metabolite is produced
Urine (<10% as unchanged drug)¹Metabolic clearance dominates; renal failure alone does not cause accumulation — a drug interaction does

Elimination and accumulation

Elimination is metabolic and slow. The apparent plasma clearance is only about 30 mL/min and the half-life about 33 hours¹, so lamotrigine leaves the body over days rather than hours, and a large ingestion produces a plateau of toxicity that does not resolve within a single emergency-department stay. The accumulation risks are two: the interaction that slows glucuronidation, and, for a subset, a modified-release preparation that prolongs absorption.

There is no enterohepatic recirculation of consequence and no active metabolite to prolong the effect — the persistence is simply the parent drug's own slow metabolic clearance. Where a patient is co-medicated with valproate, that clearance is halved again, and the plateau lasts correspondingly longer.¹ Established

Target organs — and why those

Heart — the His–Purkinje system and ventricular myocardium

TargetFast voltage-gated (Naᵥ1.5) sodium channels; phase-0 depolarisation

Why hereThe organ that makes lamotrigine dangerous rather than merely sedating. The same use- and voltage-dependent block that quiets a neurone slows cardiac depolarisation, widening the QRS and — because the block is rate-dependent — predisposing to wide-complex tachycardia.¹² The label treats QRS >100 ms as a severity marker¹; the systematic review found WCT or arrest in 6%, concentrated in adults above ~25 mg/L.² The heart is targeted because Naᵥ1.5 is the same molecular family the drug was designed to block in the brain. Inferred

At the bedsideQRS broadening, QT prolongation, wide-complex tachycardia, hypotension and, at the extreme, cardiac arrest.¹² This is the tricyclic ECG produced by an anticonvulsant, and it is treated the same way.

Brain — cortical excitability

TargetNeuronal sodium channels; the balance between suppressed repetitive firing and lowered seizure threshold

Why hereThe paradoxical organ. At therapeutic concentration the sodium-channel block is anticonvulsant; at toxic concentration the net effect reverses and seizures become the commonest life-threatening feature (55%).² Children are more susceptible, seizing at concentrations that would be sub-toxic in an adult.² The reversal is real and shared across the sodium-channel-blocking antiepileptics; the molecular explanation is not established. The brain is targeted twice over — sedated and then convulsed — by the same channel block. Inferred

At the bedsideNystagmus, ataxia, impaired consciousness progressing to coma, and grand mal convulsions.¹ Status epilepticus is described in severe cases. Benzodiazepines are the treatment.

The whole patient — through the long half-life

TargetNot a tissue; the duration of exposure

Why hereIncluded because the temporal profile is itself a hazard. A 33-hour half-life¹, swung to ~70 hours by co-prescribed valproate¹, turns a single overdose into a multi-day exposure of the heart and brain to a concentration that falls slowly. The danger of lamotrigine is partly that its toxicity does not go away quickly, which changes the admission, the monitoring and the point at which a patient can be considered safe. Established

At the bedsideA plateau of conduction abnormality and CNS effect that persists over days rather than hours; the observation period is set by the half-life, not by symptom resolution.

Timeline of effects

Lamotrigine — a fast onset, a cardiac and convulsive peak, and a tail measured in days

Time

What you seeWhat is happening

0–3 hAbsorption
What you seeNystagmus, ataxia, drowsiness; vomiting. A large ingestion may seize early.¹²
What is happeningRapid, complete absorption with Cmax at ~2.5 h and no first-pass loss.¹ Concentration — and therefore severity — is set by the dose swallowed.
3–12 hPeak toxicity
What you seeGrand mal convulsions, impaired consciousness to coma, QRS broadening, QT prolongation, wide-complex tachycardia, hypotension.¹²
What is happeningPeak sodium-channel block in brain and heart. Seizures are the commonest life-threatening feature (55%); WCT or arrest occurs in 6%, chiefly in adults above ~25 mg/L.²
12–48 hSlow decline
What you seePersisting conduction abnormality and CNS depression; the patient is not clear at the end of the first day.
What is happeningHalf-life ~33 h¹ means the concentration falls slowly. A patient above the ~25 mg/L adult severity threshold² stays there for an extended period.
DaysProlonged tail
What you seeGradual resolution over 2–4 days; longer if co-medicated with valproate.
What is happeningClearance is a single glucuronidation route at ~30 mL/min.¹ Co-prescribed valproate roughly doubles the half-life to ~70 h¹, stretching the tail further.

What the mechanism predicts at the bedside

Get an ECG and watch the QRS. Lamotrigine is a sodium-channel blocker; the label makes QRS broadening beyond 100 ms a severity marker¹, and wide-complex tachycardia is the dangerous cardiac event.² Treat the tracing as you would a tricyclic overdose.
Expect seizures, and expect them even though this is an anticonvulsant. Seizures were the commonest life-threatening feature (55%)²; benzodiazepines are first-line, and status epilepticus is described.
Sodium bicarbonate is first-line for the QRS — but do not assume it will work. In the systematic review, bicarbonate produced no response in four of nine cases with conduction delays, and two of the four cases subsequently responded with lipid therapy.² The label names lipid emulsion explicitly for cardiotoxicity that responds insufficiently to sodium bicarbonate.¹ See sodium bicarbonate and lipid emulsion.
Take the paediatric threshold seriously. A toddler can seize after ~525 mg and at a serum concentration of 3.8 mg/L² — concentrations that are sub-toxic in an adult.
Ask what else the patient takes. Co-prescribed valproate roughly doubles the half-life¹, so a valproate–lamotrigine patient clears the poison at half speed; carbamazepine or phenytoin shorten it.¹
Plan for a long stay. A 33-hour half-life¹ and a concentration-linked conduction lesion² mean the monitoring period is measured against the half-life, not against an afternoon of observation.
Do not expect much from dialysis. Only 20% was removed in a 4-hour session in the label's volunteer study¹; there is no EXTRIP recommendation, and haemodialysis is a case-by-case measure for the extreme.⁴
Activated charcoal for a recent significant ingestion, on the label's recommendation.¹ See activated charcoal.

The antidote, from the poison's side

There is no specific antidote. What exists is a two-target supportive strategy that follows directly from the mechanism: treat the conduction lesion as sodium-channel blockade, and treat the seizure as a seizure. The unusual feature is that the label itself sequences the cardiac treatment — bicarbonate first, lipid if that is insufficient.¹

Sodium bicarbonate: First-line for QRS broadening, borrowed from the sodium-channel blockade approach. Effective in some cases and not others — it failed in four of nine conduction-delay cases in the systematic review.² See sodium bicarbonate.
Intravenous lipid emulsion: Named on the UK label for cardiotoxicity that responds insufficiently to sodium bicarbonate¹; a case report describes QRS and left bundle branch block recovering after 20% lipid when bicarbonate had failed.³ Lamotrigine is lipophilic, consistent with a lipid-sink effect. See lipid emulsion.
Benzodiazepines: The treatment for the seizure, which is the commonest life-threatening feature.² A GABA-ergic brake against a cortex the sodium-channel block has paradoxically made hyperexcitable; propofol or barbiturates were used in a minority of cases.²
Activated charcoal: For a recent significant ingestion, on the label's recommendation.¹ See activated charcoal.
Haemodialysis: A weak lever. The label reports only 20% removed in a 4-hour session¹; there is no EXTRIP recommendation, and it is reserved for individual severe cases rather than used routinely.⁴

Critical appraisal

No EXTRIP recommendation exists for lamotrigine, and this page does not invent one. The dialysability statement rests on the label's own figure — 20% removed in 4 hours in renal-failure volunteers¹ — and on individual case reports of haemodialysis in severe poisoning.⁴ The absence of a formal recommendation is stated as an absence.
The sodium-channel and proconvulsant mechanisms are badged Inferred, not Established. The therapeutic sodium-channel block is the label's established claim¹; its extension to cardiac conduction and its paradoxical reversal to a proconvulsant state in overdose are strongly supported by the clinical phenotype² but are not demonstrated at the channel level in poisoned humans, and the page badges them accordingly.
The severity thresholds are review-derived and population-specific. The ~25 mg/L adult and 3.8 mg/L paediatric figures come from a case-based systematic review²; they indicate susceptibility and are not diagnostic cut-offs, and serum concentrations are not routinely available in time to guide acute care.
The lipid-emulsion evidence is case-level and partly confounded. The clearest single-agent case is Castanares-Zapatero 2012, in which QRS and LBBB recovered after lipid when bicarbonate had failed³; the widely cited Sirianni 2008 resuscitation was a combined bupropion and lamotrigine overdose, so its dramatic response cannot be attributed to lamotrigine alone, and the page flags the confound rather than borrowing the drama.
The paediatric susceptibility is real but rests on eight cases. It is stated prominently because the direction of the effect matters clinically, but eight single-agent paediatric seizures² cannot fix a precise threshold, and the page claims a direction, not a number.
No lethal dose and no concentration incompatible with life appears here. The label's 10 to 20 times the maximum therapeutic dose¹ is quoted as a reported-ingestion range, not a lethal dose, and risk assessment belongs to TOXBASE and NPIS.

References

1
Lamictal Tablets (lamotrigine) — Summary of Product Characteristics. electronic medicines compendium, product 8052 (GlaxoSmithKline UK). Sections 4.9 (Overdose), 5.1 (Pharmacodynamic properties, mechanism of action and cardiac-conduction study) and 5.2 (Pharmacokinetic properties). medicines.org.uk/emc/product/8052
2
Alyahya B, Friesen M, Nauche B, Laliberté M. Acute lamotrigine overdose: a systematic review of published adult and pediatric cases. Clinical Toxicology (Phila) 2018 Feb;56(2):81–89. PMID 28862044. 51 cases; seizures 55%, GCS ≤8 20%, hypotension 12%, wide-complex tachycardia/arrest 6%; adult severity threshold ~25 mg/L, paediatric seizures from 3.8 mg/L; bicarbonate failed in 4/9 conduction-delay cases with lipid succeeding in 2 of those.
3
Castanares-Zapatero D, Wittebole X, Huberlant V, Morunglav M, Hantson P. Lipid emulsion as rescue therapy in lamotrigine overdose. Journal of Emergency Medicine 2012 Jan;42(1):48–51. PMID 21621362. Single-agent lamotrigine overdose with QRS widening and left bundle branch block unresponsive to sodium bicarbonate, recovering after 20% lipid emulsion.
4
Agrawal A, Nogar JN, Koenig S. Management of lamotrigine overdose using hemodialysis. American Journal of Emergency Medicine 2019 Aug;37(8):1603.e1–1603.e2. PMID 31109780. Case report of haemodialysis used in a severe lamotrigine overdose, cited only for the fact that it has been attempted.

Lamotrigine

At a glance