Why this poison is interesting

Paracetamol and ethylene glycol are poisons because a metabolite appears. Salicylate is the opposite case, and it is the one that best justifies this library's organising sentence: nothing about the molecule changes at all. Salicylate uncouples oxidative phosphorylation and stimulates the respiratory centre at 300 mg exactly as it does at 30 g. What changes is that three separate kinetic ceilings are crossed at once.

Elimination saturates. The conjugation pathways that clear salicylate are capacity-limited, so first-order elimination becomes effectively zero-order and the half-life lengthens with the dose.
Protein binding saturates. Salicylate is highly and saturably bound to albumin; once binding is full, the free fraction rises, and free drug is the drug that distributes and acts.²
Distribution becomes pH-dependent. Salicylate is a weak acid. As blood pH falls, a larger proportion is non-ionised, and non-ionised salicylate crosses membranes — into the brain.²

The third of those is what makes salicylate frightening rather than merely difficult. Acidaemia does not merely accompany severe poisoning; it moves the poison into the organ it kills people through. That is a positive feedback loop with a patient inside it.

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

The toxic principle

The poison is salicylate, the anion of salicylic acid. Aspirin — acetylsalicylic acid — is a prodrug in the sense that matters here: it is rapidly hydrolysed, and its own plasma half-life is only 15–20 minutes against 2–3 hours for salicylic acid at therapeutic doses.³ By the time a poisoned patient reaches you, there is essentially no aspirin left; the assay is measuring salicylate and so is the illness. Established

Salicylate does at least four things, and they do not all pull in the same direction — which is why the acid–base picture is a moving target rather than a diagnosis.

Direct stimulation of the medullary respiratory centre — a primary respiratory alkalosis that appears early and is not compensation for anything. Established
Uncoupling of oxidative phosphorylation — the proton gradient is dissipated, so substrate is burned without making ATP. Heat, lactate and an increased oxygen and glucose demand follow. Inferred
Inhibition of Krebs-cycle enzymes and stimulation of lipolysis — organic acids and ketones accumulate, adding an anion-gap metabolic acidosis underneath the alkalosis.
A greatly increased cerebral glucose demand. The ACMT guidance is explicit that CNS glucose utilisation is increased and that CNS glucose concentration may be lower than serum glucose.² A normal blood sugar does not mean a normal brain sugar. Established

Toxicokinetics

Salicylate — three ceilings crossed at once
Parameter	Therapeutic	In overdose	Why it changes
Absorption	Dispersible aspirin is absorbed from the stomach and upper gut, peaking at 20–30 min; bioavailability ~70% after first-pass conversion to salicylic acid³	Gastro-resistant tablets may not peak for up to 12 hours.³ Pylorospasm and concretion formation can extend it further	The formulation, not the molecule, sets the timeline — and the enteric coating that protects the stomach in therapeutic use is the thing that makes a single early concentration meaningless in overdose. A gastro-resistant preparation converts a four-hour decision into a twelve-hour one.
Protein binding	High, and saturable — bound to albumin	Binding saturates; the free fraction rises^1,2	This is the row that explains why dialysis works so well. It is also why total concentration understates severity as the dose climbs: the assay measures bound plus free, but only free salicylate crosses into tissue. Two patients with the same number are not in the same trouble if one is hypoalbuminaemic.
Volume of distribution	Small — salicylate is largely confined to plasma and extracellular fluid	Rises, because unbound drug redistributes into tissue — and rises further as pH falls	Vd is not a property of the molecule here; it is a property of the patient's albumin and their pH. The same poison in an acidaemic patient occupies a bigger space, and most of the newly occupied space is intracellular.
Half-life	Aspirin 15–20 min; salicylic acid 2–3 h³	Prolonged, and lengthens as the dose rises — the defining feature of capacity-limited elimination	A half-life that depends on the concentration is the signature of a saturated pathway. It means you cannot extrapolate from a therapeutic figure, and it means the decline you predict from two concentrations will be too optimistic.
Elimination route	Hepatic conjugation (salicyluric acid and glucuronides) with a minor renal route for unchanged drug; conjugates excreted renally³	Conjugation saturates, so the proportion cleared as unchanged drug in urine rises — which is precisely the route that alkalinisation exploits	The therapy works because the overdose shifts elimination onto a renal route that is sensitive to urine pH. In therapeutic use, alkalinising the urine would achieve very little; in overdose it becomes a major exit.
Order of kinetics	First order	Effectively zero-order once the conjugation enzymes are saturated	This is the archetype. Salicylate, phenytoin, ethanol and theophylline are all in this family, and in every one of them a modest increase in dose produces a disproportionate increase in concentration — because the clearance mechanism has no headroom left.
Dialysability	—	Dialysable by haemodialysis and haemoperfusion, level of evidence B.¹ Recommended if [salicylate] >7.2 mmol/L (100 mg/dL), or >6.5 mmol/L (90 mg/dL) with impaired kidney function, or with altered mental status, or new hypoxaemia needing oxygen, or if standard therapy fails (all 1D)¹	Small, water-soluble, and — critically — increasingly unbound as the dose rises.² The ACMT guidance makes the mechanism explicit: haemodialysis is very effective because an increased free fraction appears once protein binding saturates.² The property that makes the poisoning worse is the property that makes it treatable.

Metabolism and the metabolites

This is the section where, for most poisons in this library, the story happens. For salicylate it is where the story conspicuously fails to happen — and that absence is the point.

Salicylate — a metabolic map with no toxic branch

Aspirin (acetylsalicylic acid)Plasma half-life 15–20 min; essentially gone by the time of presentation³
Rapid hydrolysis by plasma and hepatic esterasesFirst-pass; bioavailability ~70%³
SalicylateThis is the poison. No further activation is required and none occurs
Conjugation with glycineSalicyluric acidThe main route — and the one that saturates. Non-toxic, renally excreted³
GlucuronidationSalicyl phenolic and acyl glucuronidesAlso capacity-limited. Non-toxic, renally excreted³
No metabolismSalicylate excreted unchanged in urineA minor route therapeutically that becomes major in overdose — and it is pH-sensitive

What changes the answer

Chronic ingestion behaves worse than acute at the same concentration. Steady-state exposure has had time to distribute into tissue, so a given plasma number represents a much larger body burden. The SmPC records that chronic toxicity can arise from >100 mg/kg/day over two days and warns that it is insidious because the signs are non-specific.³ Established
Methyl salicylate is a different problem in the same family. Oil of wintergreen is concentrated enough that small volumes carry large salicylate doses, and it is absorbed rapidly. The mechanism is identical; the arithmetic is not.
Hypoalbuminaemia raises the free fraction at any total concentration, which means the number on the report means something different in a malnourished or chronically unwell patient. Inferred

Elimination and accumulation

Salicylate is the library's canonical ion-trapping poison, and it traps in two places at once — one of which helps you and one of which kills the patient.

The trap that helps: urine

Salicylate is a weak acid. In alkaline tubular fluid it is predominantly ionised, and an ionised molecule cannot diffuse back across the tubular membrane. Raising urine pH therefore converts filtered salicylate into salicylate that must leave. This is why the therapy targets urine pH rather than blood pH, and why it fails in a hypovolaemic or hypokalaemic patient: the ACMT guidance notes that hypovolaemia impairs alkalinisation through enhanced renal sodium and bicarbonate reabsorption and potassium excretion, and that a normal to high-normal serum potassium is believed to facilitate urinary alkalinisation.² Inferred

The trap that kills: brain

The same chemistry runs in reverse across the blood–brain barrier. As blood pH falls, an increased proportion of salicylate is non-ionised and more readily distributes into the cerebrospinal fluid and other tissues.² So acidaemia does not merely mark severity; it relocates the poison. Established

The EXTRIP position

Eighty-four articles, one controlled clinical trial, three animal studies and eighty case reports or series; clinical data on 143 patients including 14 fatalities. Very low quality of evidence for every recommendation.¹

Recommended in severe salicylate poisoning (1D), and specifically for altered mental status (1D), new hypoxaemia requiring supplemental oxygen (1D), or failure of standard therapy (1D) — all regardless of the concentration.¹
Recommended if [salicylate] >7.2 mmol/L (100 mg/dL), or >6.5 mmol/L (90 mg/dL) with impaired kidney function (both 1D). Suggested at >6.5 mmol/L (90 mg/dL) normally, or >5.8 mmol/L (80 mg/dL) with impaired kidney function (both 2D).¹
Suggested if systemic pH ≤7.20 in the absence of other indications (2D).¹
Stop when clinical improvement is apparent and [salicylate] <1.4 mmol/L (19 mg/dL) (1D), or after at least 4–6 hours if concentrations are not readily available (2D).¹
Intermittent haemodialysis preferred (1D); haemoperfusion (1D), CRRT (3D) and exchange transfusion in neonates (1D) are alternatives. Continue intravenous bicarbonate between sessions (1D).¹

Target organs — and why those

Brain

TargetNon-ionised salicylate crossing into the CSF, plus a glucose supply that cannot meet demand

Why hereTwo mechanisms compound. Entry is pH-gated — falling blood pH raises the non-ionised fraction that crosses into CSF² — and demand is raised because uncoupled oxidative phosphorylation makes the brain burn more glucose for less ATP, so that CNS glucose may be lower than serum glucose.² More poison arrives at the same time as less fuel. Established

At the bedsideTinnitus and deafness early; agitation, confusion, then coma and seizures. A euglycaemic patient can still have a hypoglycaemic brain, which is why glucose is given on the picture rather than on the reading.

Lungs

TargetThe alveolar–capillary membrane

Why hereNon-cardiogenic pulmonary oedema is a recognised feature of severe salicylate poisoning and is one of EXTRIP's strong dialysis indications in the form of new hypoxaemia requiring supplemental oxygen.¹ The mechanism is usually attributed to increased pulmonary capillary permeability rather than to fluid overload — which matters, because the two call for opposite fluid strategies. Inferred

At the bedsideNew oxygen requirement in salicylate poisoning is a dialysis trigger in its own right, not a respiratory problem to be managed separately.¹

The whole-body energy economy

TargetMitochondrial oxidative phosphorylation, uncoupled

Why hereThis is not organ selectivity so much as its absence — uncoupling happens wherever mitochondria are. The consequences show up first where metabolic rate is highest: hyperthermia, tachypnoea beyond what the respiratory centre alone would produce, lactataemia, and rapid depletion of glycogen. Inferred

At the bedsideSweating, hyperthermia, and a patient who looks septic. In a young person with fever, tachypnoea and confusion, salicylate belongs on the differential with meningitis and sepsis.

Timeline of effects

Salicylate has no latent phase in the paracetamol sense. What it has instead is a sequence of acid–base states, and the danger is that the middle one looks like recovery on a blood gas.

Salicylate — the acid–base sequence, and what it is hiding

Time

What you seeWhat is happening

0–4 hEarly
What you seeNausea, vomiting, sweating, tinnitus and mild deafness. Tachypnoea and hyperpnoea. The patient is uncomfortable, alert and often not obviously unwell.
What is happeningDirect stimulation of the medullary respiratory centre produces a primary respiratory alkalosis. Uncoupling has begun but has not yet generated enough organic acid to be visible. Absorption is still ongoing — and with a gastro-resistant preparation it may not have peaked for another eight hours.³
4–12 hThe mixed picture
What you seeHyperventilation persists. Vomiting and sweating produce volume depletion and hypokalaemia. Arterial pH is normal or high, which reads as a patient who is compensating well.
What is happeningTwo processes are now running against each other: respiratory alkalosis and an accumulating anion-gap metabolic acidosis from uncoupling, lipolysis and Krebs-cycle inhibition. The SmPC describes this mixed pattern with a normal or high pH as the usual adult finding.³ The normal pH is the product of two abnormalities, not of one normality.
12 h onwardsDecompensation
What you seeConfusion, agitation, hyperthermia, non-cardiogenic pulmonary oedema, then coma and seizures. Acidaemia (pH <7.35) marks severe poisoning.²
What is happeningThe respiratory compensation fails — through exhaustion, through CNS depression, or because somebody intubated the patient. As pH falls, the non-ionised fraction rises and salicylate moves into the CSF.² The plasma concentration may be falling at exactly this moment, because the drug is leaving plasma for tissue.
Any time
What you seeIn chronic or therapeutic-excess poisoning, the whole sequence is compressed and the presenting picture is non-specific: confusion in an older person, breathlessness, a metabolic acidosis of unclear cause.
What is happeningSteady-state exposure means tissue and plasma have equilibrated, so a given plasma number represents a far larger body burden than the same number after an acute ingestion. The SmPC notes that chronic salicylate poisoning "can be insidious as signs and symptoms are non-specific".³ Established

What the mechanism predicts at the bedside

Why a single concentration cannot be trusted

Three separate reasons, all mechanical. Absorption may still be ongoing — up to twelve hours with a gastro-resistant preparation.³ Elimination is zero-order, so extrapolating a decline from two points systematically underestimates the time course. And the free fraction, not the total, determines effect. The ACMT guidance's practical instruction follows: take serial concentrations until they are clearly declining, and be guided by the patient rather than the number.²

Why urine pH, not blood pH, is the target of alkalinisation

Because the therapeutic trap is in the renal tubule. Raising blood pH is desirable for a separate reason — it keeps salicylate out of the brain — but the elimination benefit depends on the urine. Which is why volume state and potassium matter so much: a hypovolaemic, hypokalaemic patient reabsorbs bicarbonate and excretes potassium, and their urine will not alkalinise however much bicarbonate is given.² Inferred

Why the Done nomogram should not be used

The Done nomogram plots concentration against time to grade severity, in the manner of the paracetamol treatment line. It was evaluated against the judgement of three experienced emergency physicians across 55 acute adult intoxications, and its calculated predictive index was 0.42 — the highest value, 0.79, being in the mild category. It tends to overpredict severity in the moderate and severe categories.⁴ Traditional teaching

That result should not be surprising given everything above. A nomogram assumes first-order elimination from a single compartment with fixed protein binding — and salicylate violates all three assumptions simultaneously, which is precisely what makes it interesting. The authors' conclusion was that management should rest on clinical presentation and judgement as well as the concentration,⁴ which is also EXTRIP's position thirty years later.¹

Why intubation is the intervention to fear

Covered in the organ section above, but it is the single highest-value thing on this page. The patient's hyperventilation is not a symptom to be relieved; it is the only thing holding their pH up, and their pH is what is holding the salicylate out of their brain. Hyperventilation is not an indication for intubation, and where intubation is genuinely unavoidable, the pH must not be allowed to fall around it.² Established

Why hypoglycaemia should be treated on the picture, not the reading

Because CNS glucose utilisation is increased and CNS glucose may be lower than serum glucose.² A confused, agitated salicylate-poisoned patient with a normal capillary glucose may still have a brain that is short of substrate — and glucose is cheap, immediate and harmless. Inferred

The antidote, from the poison's side

There is no antidote in the strict sense — no enzyme to inhibit, no cofactor to replace. What exists is a chemical manipulation of where the poison sits, and it is one of the most mechanistically satisfying treatments in toxicology precisely because it does not touch the drug at all.

Bicarbonate does two different jobs with one drug. Raising blood pH keeps salicylate ionised in plasma and therefore out of the brain; raising urine pH keeps it ionised in the tubule and therefore out of the body. Same chemistry, two compartments, both beneficial. Established
It fails predictably when the kidney cannot cooperate. Volume depletion and hypokalaemia both defeat urinary alkalinisation.² Giving bicarbonate to a dry, hypokalaemic patient and concluding that alkalinisation does not work is a common and avoidable error.
Haemodialysis is the definitive treatment and works better the sicker the patient is — because the free fraction, which is what dialysis removes efficiently, rises with the dose.^1,2 Most antidotes get less effective as the poisoning worsens; this one gets more so.
EXTRIP recommends continuing intravenous bicarbonate between dialysis sessions (1D),¹ which makes sense only if you hold both mechanisms in mind: dialysis addresses body burden, bicarbonate addresses distribution, and stopping the second while running the first would let salicylate move into the brain during the treatment.

Critical appraisal

The Done nomogram is still taught and still reproduced, and it does not work. A predictive index of 0.42, overpredicting severity in exactly the categories where the decision matters.⁴ Its persistence is a good example of the drug-series lesson that a memorable graphic outlives its evidence — and of why this site puts an evidence tier on things everyone already knows.
"Salicylate is a saturation poison" is secure; the numbers attached to it are much less so. The direction of every change in the kinetics table is well established and mechanistically necessary. Specific figures for the volume of distribution, the percentage bound, and the half-life at a given concentration vary widely between sources and are not stated in the UK labels, so they are described here by direction rather than quoted. That is a deliberate omission, not an oversight.
All the EXTRIP salicylate recommendations rest on very low quality evidence — 84 articles, one controlled trial, 143 patients.¹ They are the best available synthesis and should be used; the concentration cut-points are consensus positions, not measured thresholds.
The pulmonary oedema mechanism is inferred. Inferred Increased pulmonary capillary permeability is the standard explanation and is consistent with the clinical picture, but it is not demonstrated in humans in a way that excludes contributions from fluid administration or from the acidaemia itself. The clinical implication — that this is not a fluid-overload problem — is the part worth acting on.
Uncoupling of oxidative phosphorylation is a textbook mechanism carried over largely intact from mid-twentieth-century work. Inferred It explains the hyperthermia, the lactate and the raised glucose demand coherently, and nothing contradicts it, but the quantitative contribution of uncoupling versus Krebs-cycle inhibition versus lipolysis to the anion gap in a human overdose has not been separated.
Chronic salicylate poisoning is under-diagnosed for a structural reason. Its presentation — confusion, breathlessness, metabolic acidosis in an older patient — sits in the differential of half a dozen commoner things, and nobody sends a salicylate concentration unless they think of it. The SmPC's own word for it is insidious.³ The mechanism explains why the plasma number will look unimpressive when they are very unwell.

References

1
Juurlink DN, Gosselin S, Kielstein JT, Ghannoum M, Lavergne V, Nolin TD, Hoffman RS; EXTRIP Workgroup. Extracorporeal treatment for salicylate poisoning: systematic review and recommendations from the EXTRIP workgroup. Ann Emerg Med 2015;66(2):165–81. PubMed 25986310 · Recommendation set also published at extrip-workgroup.org/salicylates. Source of every dialysability grading, indication and cessation threshold quoted above, and of the 84-article / 143-patient evidence base with its very low quality grading. Verified 31 Aug 2026 from the abstract and the workgroup's own published recommendation page.
2
American College of Medical Toxicology. Guidance document: management priorities in salicylate toxicity. J Med Toxicol 2015;11(1):149–52. PMC4371029 Open access. Source of the quoted statements on pH-dependent CNS distribution, on clinical deterioration with a falling concentration, on CNS glucose being lower than serum glucose, on hypovolaemia and potassium defeating urinary alkalinisation, and on the danger of intubation and mechanical ventilation. Verified 31 Aug 2026.
3
Aspirin 300mg Gastro-resistant Tablets — Summary of Product Characteristics, Alliance Pharmaceuticals, text revised March 2025. emc product 8627 — and Disprin (dispersible aspirin 300 mg), Reckitt Benckiser. emc product 614. §4.9 and §5.2 of both were fetched and read separately. Source of the >350 mg/L and >700 mg/L concentrations, the <100 mg/kg and >100 mg/kg/day figures, the 15–20 min aspirin and 2–3 h salicylic acid half-lives, the ~70% bioavailability and 20–30 min dispersible peak, the up-to-12-hour peak with the gastro-resistant formulation, and the mixed acid–base description. The 12-hour figure is specific to the gastro-resistant product and does not appear in the dispersible label. Verified 31 Aug 2026.
4
Dugandzic RM, Tierney MG, Dickinson GE, Dolan MC, McKnight DR. Evaluation of the validity of the Done nomogram in the management of acute salicylate intoxication. Ann Emerg Med 1989;18(11):1186–90. PubMed 2817562 Source of the predictive index of 0.42, the 0.79 figure in the mild category, and the finding that the nomogram overpredicts severity in the moderate and severe categories. This is the citation for the doubt behind the traditional-teaching badge. Verified 31 Aug 2026.
5
Palmer BF, Clegg DJ. Salicylate toxicity. N Engl J Med 2020;382(26):2544–55. PubMed 32579814 Cited as the current comprehensive review. Paywalled; not used as the source of any figure on this page. Citation verified 31 Aug 2026.
6
TOXBASE — salicylate; aspirin. National Poisons Information Service. toxbase.org (NHS login required. NPIS: 0344 892 0111. The authoritative UK source for treatment thresholds, alkalinisation regimens and dialysis criteria. Login-gated, so not quoted here.)

Salicylate

At a glance