Why this poison is interesting

Tramadol has its own page rather than a paragraph on the opioid page for a reason that its manufacturer states more plainly than most textbooks do. The UK prolonged-release label describes naloxone as a specific antidote against the opioid symptoms and then, in the same sentence, says what it will not do: though it will not antagonise tramadol's inhibitory effects on MAO reuptake or serotonin releasing effects.² The immediate-release label adds: in animal experiments naloxone had no effect on convulsions. In such cases diazepam should be given intravenously.¹

A licensed antidote whose own label enumerates the parts of the poisoning it does not reach is unusual, and it is the reason this page exists. The patient whose breathing is restored by naloxone is not the patient who has been made safe: the seizures and the serotonergic effects belong to a different pharmacology within the same tablet, and they are unaffected.

The second reason is genetic. Tramadol's opioid activity depends almost entirely on O-desmethyltramadol, formed by CYP2D6, which the label notes is more potent than the parent substance by the factor 2–4 in animal experiments.¹ CYP2D6 activity varies enormously between people. A poor metaboliser makes little M1 and gets a serotonergic, seizure-prone poisoning with less respiratory depression; an ultra-rapid metaboliser makes a great deal and gets an opioid one. The dose is the same; the poisoning is not.

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

The toxic principle

Tramadol is a single molecule doing three separable things: weak mu-opioid agonism, inhibition of noradrenaline reuptake, and inhibition of serotonin reuptake with an additional releasing action. The label refers to the last two together as inhibitory effects on MAO reuptake or serotonin releasing effects.² Each contributes a distinct part of the clinical picture, and each responds to a different intervention or to none.

The seizures are the fourth thing and the least well explained. Both labels list convulsions among the features of overdose¹², and the immediate-release label's specific instruction to use intravenous diazepam because naloxone had no effect on convulsions in animal experiments¹ tells you the manufacturer regards them as non-opioid in origin. What it does not tell you is what causes them.

The seizure mechanism is asserted more confidently in teaching than on any label

Tramadol's proconvulsant effect is usually attributed to inhibition of GABA-A receptor function, sometimes to the monoamine effects, and sometimes to both. Neither UK label consulted for this page offers any mechanism at all — they state that convulsions occur and that naloxone does not help.¹² The clinical fact is secure and the explanation taught alongside it is not demonstrated, so this page does not print one as though it were. But note what this is not: the labels are silent, and silence is an absence of support rather than a contest. That is why the claim is marked as inference rather than downgraded further — the lithium page's badge survives because two labels positively assert the mechanism is unknown, and these labels do no such thing. Inferred

Naloxone's effect on tramadol seizures is worth separating from its effect on opioid-induced seizures generally. The label's statement is narrow and animal-derived: in animal experiments naloxone had no effect on convulsions.¹ It is not a claim that naloxone provokes seizures in tramadol overdose, which is a widely repeated assertion this page deliberately does not make — it is a claim that naloxone does not stop them. The label's instruction that follows is unambiguous: diazepam, intravenously.¹

Toxicokinetics

Tramadol is unusually well characterised for a drug this old, and the label gives figures for both the parent and the active metabolite in health, in cirrhosis and in renal failure — a rare completeness that makes the accumulation argument checkable rather than assumed.

Tramadol and O-desmethyltramadol — the parent and the opioid, tracked separately
Parameter	Therapeutic	In overdose	Why it changes
Absorption	More than 90% absorbed; absolute bioavailability approximately 70%, irrespective of food; first-pass effect a maximum of 30%¹	Detectable within 15 to 45 minutes, Tmax 1.6 to 2 h after 100 mg¹	Fast, and near-complete. There is no absorption bottleneck to buy time — which is why the immediate-release label puts decontamination within 2 hours¹ rather than the single hour more common elsewhere.
Prolonged-release absorption	Zydol XL 200 mg: Cmax 192 ng/mL at a median Tmax of 6 h (range 4–8); no evidence of dose dumping with food²	Elimination half-life extended to about 16 hours by the formulation²	The formulation, not the molecule, sets the duration. This is the same trap the calcium-channel blockers page describes: a prolonged-release tablet turns a six-hour drug into a sixteen-hour one, and the label says decontamination at a later time point may be useful for such preparations.¹
Protein binding	About 20%¹	Unchanged	Strikingly low, and it is the row that makes the dialysis conclusion counter-intuitive. A 20%-bound drug ought to be dialysable. It is not — and the reason is the next row.
Volume of distribution	203 ± 40 L — high tissue affinity¹	Unchanged	This is where the dialysis argument is actually won. Two hundred litres against about five litres of blood means the drug is almost entirely outside the compartment a dialyser can reach, exactly as for digoxin — low binding cannot compensate for that.
Metabolism	N- and O-demethylation with glucuronidation; only O-desmethyltramadol is pharmacologically active, and is 2–4 times more potent than the parent in animals. Eleven metabolites found in urine¹	CYP2D6 and CYP3A4 inhibition or induction alters the parent-to-M1 ratio¹²	The single most consequential row on the page. CYP2D6 decides how much opioid the patient makes from a given dose — and therefore whether they present with respiratory depression or with agitation and seizures. Compare codeine on the opioid page, where the same enzyme performs the same job.
Half-life — parent	Approximately 6 h, irrespective of route; prolonged by a factor of about 1.4 above age 75¹	Cirrhosis 13.3 ± 4.9 h, extreme case 22.3 h; renal failure (creatinine clearance below 5 mL/min) 11 ± 3.2 h, extreme case 19.5 h¹	A two- to fourfold prolongation in organ failure, from a label that gives the standard deviations. The extreme cases are quoted because they, not the means, are what a long observation period has to cover.
Half-life — M1	7.9 h (range 5.4–9.6) in six healthy volunteers¹	Cirrhosis 18.5 ± 9.4 h, extreme case 36 h; renal failure 16.9 ± 3 h, extreme case 43.2 h¹	The opioid metabolite outlives the parent in every condition, and the gap widens as organs fail. A patient in renal failure may carry the active opioid for well over a day — and naloxone's duration is measured in tens of minutes.
Elimination	Tramadol and metabolites almost completely excreted via the kidneys; cumulative urinary excretion 90% of the total radioactivity of the administered dose¹	Slower with renal impairment¹	A renally cleared active metabolite in a patient who may be hypotensive. The accumulation risk is M1's, not the parent's.
Linearity	Linear within the therapeutic dosage range¹. But immediate-release tramadol 50 mg six-hourly gave concentrations greater than would have been anticipated, consistent with non-linear elimination²	Not characterised above therapeutic doses	The two labels are not in conflict but they are not identical either, and the second is the more interesting: a formulation study detected non-linearity that a single-dose study did not. Nothing in either supports extrapolating a concentration–effect relationship into overdose.
Dialysability	—	No. Tramadol is minimally eliminated from the serum by haemodialysis or haemofiltration. Therefore treatment of acute intoxication ... with haemodialysis or haemofiltration alone is not suitable for detoxification¹². EXTRIP has never addressed tramadol³	Two independent UK labels state the same conclusion in near-identical words, and the volume of distribution explains it.¹ Note the label's careful alone — it declines to say dialysis is useless in every circumstance, only that it is not a detoxification strategy.

Metabolism and the metabolites

Tramadol is the library's clearest prodrug-and-drug: the molecule swallowed is fully active in one set of mechanisms and must be converted to become active in the other. The consequence is that the same tablet produces different poisonings in different people, and the branch point is a single polymorphic enzyme.

Tramadol — one molecule, two active species, two syndromes

TramadolAlready active. Inhibits serotonin and noradrenaline reuptake and releases serotonin²; lowers the seizure threshold¹²
CYP2D6 — O-demethylationO-desmethyltramadol (M1)The opioid. More potent than the parent substance by the factor 2–4 in animal experiments¹. Half-life 7.9 h¹
CYP3A4 — N-demethylation, plus glucuronidationInactive metabolitesEleven metabolites found in urine; only O-desmethyltramadol is pharmacologically active¹
Miosis, sedation, respiratory depression, apnoeaReversed by naloxone¹²
Serotonin toxicity, agitation, tachycardia, seizuresNot reversed by naloxone — the label says so explicitly², and animal data show no effect on convulsions¹

One paediatric note from the label deserves recording because it is a mechanism rather than a caution. CYP2D6 activity increases continuously through the neonatal period and reaches adult levels at about one year of age, while glucuronidation and renal function are simultaneously immature — so an infant under one may both form M1 progressively faster and clear it more slowly, and the label states this may result in slow elimination and accumulation of O-desmethyltramadol.¹

Elimination and accumulation

The accumulation problem in tramadol poisoning belongs to the metabolite, and the label's figures make the point without needing any interpretation: in every condition it lists, M1's half-life exceeds the parent's, and the difference grows as organ function falls.

Half-lives from the immediate-release label — parent against active metabolite¹
Condition	Tramadol	O-desmethyltramadol (M1)
Healthy adults	About 6 h	7.9 h (range 5.4–9.6)
Age above 75	Prolonged by a factor of about 1.4	Not separately stated
Cirrhosis	13.3 ± 4.9 h; extreme case 22.3 h	18.5 ± 9.4 h; extreme case 36 h
Renal failure (CrCl below 5 mL/min)	11 ± 3.2 h; extreme case 19.5 h	16.9 ± 3 h; extreme case 43.2 h

The prolonged-release preparations create the same problem from the other direction. Zydol XL's label gives a Tmax of 6 hours, range 4 to 8, and an elimination half-life extended to about 16 hours by the formulation.² A patient who has taken prolonged-release tablets has not reached peak concentration when most immediate-release patients would be improving, and the immediate-release label acknowledges that gastrointestinal decontamination at a later time point may be useful in case of intoxication with exceptionally large quantities or prolonged-release formulation.¹

Target organs — and why those

Brainstem respiratory centres

TargetMu-opioid receptors, reached mainly by M1

Why hereThe opioid half of the poisoning, and the only half naloxone reverses. What makes this card different from the opioid page's equivalent is that the agonist here is manufactured by the patient: a poor CYP2D6 metaboliser produces little M1 and correspondingly less respiratory depression from the same ingested dose, while an ultra-rapid metaboliser produces a great deal of a substance 2–4 times more potent than the parent.¹ Established

At the bedsideMiosis, sedation, respiratory depression up to respiratory arrest.¹ Naloxone works here and is named as the antidote on both labels.¹² Its duration is the problem — M1's half-life is 7.9 hours in health¹ and up to 43.2 hours in severe renal failure.¹

Cerebral cortex — seizure threshold

TargetNot established on either UK label

Why hereThe most clinically important unexplained finding in this poisoning. Both labels list convulsions among the features of overdose¹², and the immediate-release label directs intravenous diazepam because naloxone had no effect on convulsions in animal experiments.¹ The GABA-A account taught alongside it appears on neither label. The phenomenon is established; the mechanism printed in textbooks is not, and this card declines to print it — though neither label contests it either, which is why this is marked as inference rather than as teaching that has been refuted. Inferred

At the bedsideSeizures, which may be the presenting feature and may occur without marked respiratory depression — particularly in a poor CYP2D6 metaboliser. Diazepam intravenously, on the label's own instruction.¹

Serotonergic synapses

TargetSerotonin reuptake transporter, plus a releasing action

Why hereThe mechanism tramadol shares with the SSRIs, plus one they do not have. The label describes both limbs: tramadol's inhibitory effects on MAO reuptake or serotonin releasing effects.² Blocking reuptake has a ceiling; adding release does not — which is the pharmacological reason tramadol appears so disproportionately in serotonin toxicity reports relative to how weak an antidepressant it would make. Established

At the bedsideSerotonin syndrome has also been reported, on both labels.¹² Examine for clonus — see the SSRI page for why that finding rather than the fever carries the diagnosis.

Sympathetic nervous system

TargetNoradrenaline reuptake transporter

Why hereThe third mechanism, and the one that produces the discordant vital signs. Blocking noradrenaline reuptake raises synaptic noradrenaline, giving tachycardia, hypertension and agitation — findings that sit uncomfortably beside miosis and sedation. A patient with pinpoint pupils and a heart rate of 120 has a poisoning that is not a pure opioid one, and tramadol is high on that short list along with clonidine's biphasic behaviour and the mixed pictures of polydrug overdose. Inferred

At the bedsideTachycardia, hypertension, agitation, diaphoresis. Useful diagnostically precisely because it contradicts the expected opioid picture.

Cardiovascular system

TargetCirculatory collapse in severe poisoning

Why hereBoth labels list cardiovascular or circulatory collapse among overdose features.¹² Neither attributes it to a mechanism, and the honest reading is that it is a consequence of profound respiratory depression, seizures and hyperthermia rather than a direct cardiac action — tramadol is not a sodium- or potassium-channel blocker in the sense the sodium-channel blockade page describes. Inferred

At the bedsideHypotension and collapse in severe cases. Its appearance should prompt a search for the cause — hypoxia, seizure, hyperthermia — rather than treatment as a primary cardiac event.

Timeline of effects

Tramadol overdose — two syndromes on different clocks, and a formulation that changes both

Time

What you seeWhat is happening

15–45 minDetectable
What you seeNausea, vomiting, drowsiness.
What is happeningPlasma concentrations detectable within 15 to 45 minutes of a 100 mg dose¹. More than 90% is absorbed and bioavailability is about 70%¹ — there is very little absorption barrier.
1.6–2 hPeak — immediate release
What you seeMiosis, sedation, respiratory depression; or agitation, tachycardia and clonus; or both.
What is happeningTmax 1.6 to 2 h¹. Which syndrome dominates depends on how much M1 the patient's CYP2D6 has made¹ — the branch point is metabolic and precedes the presentation.
Within 2 hThe decontamination window
What you see—
What is happeningThe immediate-release label puts gastrointestinal decontamination within 2 hours of tramadol intake¹ — twice the usual single hour, reflecting how completely and how fast the drug is absorbed.
Any timeSeizures
What you seeConvulsions, which may occur without severe respiratory depression.¹²
What is happeningMechanism not stated on either label. Naloxone had no effect on convulsions in animal experiments; the label directs intravenous diazepam.¹ This phase does not sit at a predictable point on the curve, which is part of what makes it dangerous.
4–8 hPeak — prolonged release
What you seeThe same features, arriving when an immediate-release patient would be improving.
What is happeningZydol XL 200 mg: median Tmax 6 h (range 4–8), Cmax 192 ng/mL, no dose dumping with food.² Late decontamination may be useful for prolonged-release preparations, though the label notes there is no clinical trial evidence for it.¹²
6–16 hThe naloxone gap
What you seeRecurrent respiratory depression as each naloxone dose wears off. The gap drawn here is not a latent phase — it is the interval in which the antidote has gone and the poison has not.
What is happeningParent half-life about 6 h, M1 7.9 h¹, extended to about 16 h by prolonged-release formulation². Naloxone's duration is a fraction of an hour. The mismatch is arithmetic, not clinical judgement.
Up to 43 hOrgan failure
What you seeProlonged opioid effects, far beyond any standard observation period.
What is happeningM1 half-life 18.5 ± 9.4 h in cirrhosis, extreme case 36 h; 16.9 ± 3 h in renal failure, extreme case 43.2 h.¹ Tramadol and its metabolites are almost completely renally excreted.¹

What the mechanism predicts at the bedside

Naloxone restoring the breathing does not mean the patient is safe. The label states it will not antagonise the monoamine reuptake or serotonin-releasing effects², and animal data show no effect on convulsions.¹
Miosis with tachycardia is the tell. The noradrenaline reuptake limb produces vital signs that contradict the opioid picture, and few other poisons do that.
Examine for clonus in any tramadol overdose, exactly as for an SSRI — both labels report serotonin syndrome.¹²
Ask which preparation. Immediate release peaks at under two hours; prolonged release at four to eight, with a half-life extended to about sixteen.¹² The tablets look nothing alike and the histories often do.
Ask what else is prescribed, with CYP2D6 in mind. An SSRI both inhibits M1 formation and adds a second serotonergic mechanism¹² — the interaction pushes in one direction on both counts.
Seizures may come before, or instead of, respiratory depression — particularly in a poor CYP2D6 metaboliser who makes little opioid from the dose.
Renal or hepatic impairment changes the observation period, not just the dose. M1's extreme-case half-life is 36 hours in cirrhosis and 43.2 hours in severe renal failure.¹
Charcoal has a two-hour window here, not the usual one, on the immediate-release label's own statement.¹
Do not request dialysis. Two labels state independently that tramadol is minimally removed and that dialysis alone is not suitable for detoxification¹²; the 203 L volume of distribution explains why despite only 20% protein binding.¹
In a child under one, M1 accumulates — CYP2D6 matures towards adult activity over the first year while glucuronidation and renal clearance lag.¹

The antidote, from the poison's side

Tramadol has a genuine, licensed, effective antidote for part of what it does. That is a rarer and more interesting situation than either having an antidote or not having one, and both UK labels handle it with unusual candour.

Naloxone: Reverses the mu-opioid effects, which are largely M1's.¹ Its duration is far shorter than M1's half-life of 7.9 h¹, and shorter still relative to the 36–43 h extremes in organ failure¹ — so the recurrence problem described on the opioid page applies here with more force, not less.
Benzodiazepines: Named on the label for the convulsions: in such cases diazepam should be given intravenously.¹ They also reduce the agitation and muscle activity of the serotonergic component, which is the same role they play on the SSRI page.
Activated charcoal: Immediate release: within 2 hours.¹ Prolonged release: within 1 hour, with the caveat that later charcoal for prolonged-release preparations may seem reasonable but has no clinical trial evidence to support it.² The two labels differ, and the difference is the formulation, not a disagreement.
Cyproheptadine: For serotonin toxicity, with the same weak evidential footing described on the SSRI page. Neither tramadol label mentions it.
Sodium bicarbonate: No role established. Tramadol is not described as a sodium-channel blocker on either label, and nothing in the overdose sections points to QRS widening.¹² Mentioned here only to say that the reflex from the tricyclics does not transfer.
Extracorporeal removal: Not suitable. Tramadol is minimally eliminated from the serum by haemodialysis or haemofiltration. Therefore treatment of acute intoxication ... with haemodialysis or haemofiltration alone is not suitable for detoxification.¹² EXTRIP has published no recommendation covering tramadol.³

Critical appraisal

The seizure mechanism carries an inferred badge, and it was downgraded from traditional teaching at audit. The original justification was that both labels describe convulsions and give management instructions for them while neither offers any mechanism. Two independent reviewers rejected that reasoning and they were right: silence is an absence of a citation for the claim, which the house rule explicitly refuses, and it is not the same as the lithium precedent, where two labels positively assert the mechanism is unknown. The GABA-A account taught in textbooks is therefore unsupported here rather than refuted, and no source contesting it was found. If one is found, the badge should go back on.
The naloxone-and-seizures statement is animal-derived and is quoted as such.¹ This page deliberately does not make the widely repeated claim that naloxone provokes seizures in tramadol overdose. The label says only that naloxone did not stop them in animals, and the stronger claim is not supported by either document consulted.
The M1 potency figure of 2–4 times the parent is from animal experiments, and the label says so.¹ It is the basis of the CYP2D6 argument on this page, and that argument is therefore inference from animal potency plus human metabolic variation — flagged as inferred on the organ card and in the pathway callout rather than asserted.
The CYP2D6 phenotype account is mechanistically sound and clinically unvalidated in tramadol overdose specifically. That poor and ultra-rapid metabolisers form different amounts of M1 follows directly from the label¹; that this produces two clinically distinguishable overdose syndromes is a prediction from that fact, not a finding from a cohort. It is badged accordingly.
The two labels differ on the decontamination window — 2 hours for immediate release¹ against 1 hour for prolonged release² — and this page reports the difference rather than reconciling it, because the difference is real and formulation-driven. The prolonged-release label's own caveat that there is no clinical trial evidence for late charcoal² is the more useful statement of the two.
The non-linearity observation is from a steady-state formulation study, not from overdose data², and is not evidence about what happens at overdose concentrations. Nothing on this page extrapolates from it.
No lethal or fatal dose appears here, and none should. Both labels note that overdose may be fatal¹²; neither figure nor threshold is reproduced, in line with this site's permanent boundary.
EXTRIP has never addressed tramadol³, but unlike clonidine the absence leaves no open question: two labels answer it directly and the 203 L volume of distribution¹ explains the answer.
The cardiovascular collapse card is the weakest on the page and is marked as inference. Both labels list it¹²; neither explains it, and attributing it to hypoxia, seizure and hyperthermia rather than to a direct cardiac action is a reading rather than a demonstration.

References

1
Zydol 50 mg Capsules (tramadol hydrochloride) — Summary of Product Characteristics. electronic medicines compendium, product 81. Sections 4.9 (Overdose) and 5.2 (Pharmacokinetic properties). Source of the M1 potency factor, the parent and metabolite half-lives in health, cirrhosis and renal failure, the 203 ± 40 L volume of distribution, the 20% protein binding, the 2-hour decontamination window, and the statement that naloxone had no effect on convulsions in animal experiments. medicines.org.uk/emc/product/81
2
Zydol XL 400 mg prolonged-release Tablets (tramadol hydrochloride) — Summary of Product Characteristics. electronic medicines compendium, product 88. Sections 4.9 (Overdose) and 5.2 (Pharmacokinetic properties). Source of the statement that naloxone will not antagonise tramadol's inhibitory effects on monoamine reuptake or its serotonin-releasing effects, the prolonged-release Tmax and 16-hour half-life, and the caveat that there is no clinical trial evidence for late activated charcoal. medicines.org.uk/emc/product/88
3
EXTRIP Workgroup — published recommendations index. Extracorporeal Treatments in Poisoning Workgroup. Cited for the absence of any recommendation covering tramadol or the opioids. extrip-workgroup.org/recommendations

Tramadol

At a glance