Serotonin Syndrome Hunter Criteria Calculator
Serotonin Syndrome Hunter Criteria Calculator
Apply the Hunter Serotonin Toxicity Criteria as the branching decision rule Dunkley published — five independent branches, no points to add — with its 84% sensitivity against a clinical toxicologist’s diagnosis and the rigidity pattern that separates it from neuroleptic malignant syndrome.
Hunter serotonin toxicity criteria
Decision rule, not a scoreA serotonergic agent taken, inducible clonus present, diaphoresis present, no spontaneous or ocular clonus, no agitation, no tremor, no hyperreflexia, no hypertonia, maximum temperature 37.6 °C
The rule, all five branches, exactly as published
1. spontaneous clonus
2. inducible clonus and (agitation or diaphoresis)
3. ocular clonus and (agitation or diaphoresis)
4. tremor and hyperreflexia
5. hypertonia and temperature > 38 °C and (ocular or inducible clonus)
- not a score
- there is nothing to add up. Dunkley’s criteria were derived by classification and regression tree analysis, so they are a branching rule in which each branch is independently sufficient and no branch is partially satisfied. A patient with four features spread across branches without completing any of them does not meet the criteria; a patient with one — spontaneous clonus — does
- 84% and 97%
- sensitivity and specificity against the diagnosis of serotonin toxicity by a clinical toxicologist, in 2,222 serotonergic drug overdose admissions with a learning dataset of 473 SSRI-alone overdoses. Sternbach’s earlier criteria scored 75% and 96% on the same data, so Hunter is both simpler and better on both axes
- clonus
- the load-bearing sign, and it appears in four of the five branches. Inducible clonus is elicited by rapid dorsiflexion of the ankle; ocular clonus is slow continuous horizontal eye movement. Both are lower-limb and eye predominant, reflecting the spinal and brainstem serotonin receptors involved
- > 38 °C
- strictly above 38, as published. Dunkley reports hypertonicity and a maximum temperature above 38 °C as universal in life-threatening cases, which is why branch 5 is the one this page flags as an emergency rather than a diagnosis
- what is NOT in the rule
- no serum concentration, no drug level, no laboratory test. There is no assay for serotonin toxicity — the serotonin unit converter handles plasma or platelet serotonin, which is a carcinoid investigation and has nothing to do with this diagnosis
Worked example
A serotonergic agent taken, inducible clonus present, diaphoresis present, no spontaneous or ocular clonus, no agitation, no tremor, no hyperreflexia, no hypertonia, maximum temperature 37.6 °C
A serotonergic agent has been taken, so the criteria apply
Branch 5 — hypertonia with a temperature above 38 °C and clonus: hypertonia absent and 37.6 is not above 38, so no
Branch 1 — spontaneous clonus: absent, so no
Branch 2 — inducible clonus and (agitation or diaphoresis): inducible clonus present, diaphoresis present → met
The remaining branches are not evaluated, because one sufficient branch is all the rule needs
Note what would happen with a point total instead. This patient has two of the nine features. A patient with agitation, diaphoresis, tremor and a temperature of 37.9 has four — and meets no branch at all, because tremor needs hyperreflexia beside it and agitation and diaphoresis need clonus. A score would rank the second patient higher; the rule correctly diagnoses the first
And the escalation to watch for: add hypertonia and a temperature of 38.2 to this same patient and the verdict becomes the life-threatening pattern, which is a critical care referral rather than an observation
Serotonin toxicity against neuroleptic malignant syndrome
| Serotonin toxicity | Neuroleptic malignant syndrome | |
|---|---|---|
| Onset | Rapid — “within a few hours of a serotonergic agent change” | Subacute — “within days to weeks” |
| Neuromuscular pattern | Hyperreactivity — hyperreflexia, clonus and tremor, lower-limb predominant | Slowed response — lead-pipe rigidity, generalised, with bradyreflexia |
| Reflexes | Increased, often strikingly | Reduced |
| Resolution | “Within 1 day generally” | “1 to 2 weeks” |
| Drug class | Serotonergic — usually two agents in combination | Dopamine antagonist started or increased, or a dopamine agonist withdrawn |
| Pupils | Mydriasis common | Not a feature |
| Bowel sounds | Increased; diarrhoea is common | Normal or reduced |
| Shared | Hyperthermia, diaphoresis, tachycardia, altered mental state, raised creatine kinase, rhabdomyolysis | The same — which is why the pattern of tone and the onset speed are the discriminators and the temperature is not |
Hunter against Sternbach, on the same 2,222 admissions
| Hunter | Sternbach | |
|---|---|---|
| Sensitivity against a clinical toxicologist’s diagnosis | 84% | 75% |
| Specificity | 97% | 96% |
| Structure | Branching decision rule from a CART analysis; each branch independently sufficient | A list of features with a minimum count, plus exclusion criteria |
| Features required | Nine, of which clonus in three forms does most of the work | Thirteen, weighted equally |
| Practical consequence | Better on both axes and shorter; the clonus examination is the whole skill | Over-diagnoses mild cases and under-diagnoses severe ones relative to Hunter |
Serotonergic agents, and the combinations that cause trouble
| Class | Examples | Note |
|---|---|---|
| SSRIs and SNRIs | Sertraline, fluoxetine, citalopram, venlafaxine, duloxetine | The commonest single class involved. Fluoxetine’s long half-life means its interaction window outlasts the last dose by weeks |
| MAO inhibitors | Phenelzine, tranylcypromine, moclobemide, selegiline, rasagiline, linezolid, methylene blue | The most dangerous combinations involve these. Linezolid and methylene blue are monoamine oxidase inhibitors and are routinely not recognised as such — the methylene blue dose calculator covers the same interaction from the other side |
| Opioids | Tramadol, pethidine, fentanyl, methadone, oxycodone, dextromethorphan | Tramadol and pethidine are the classic culprits, and both are prescribed by people not thinking about serotonin |
| Others | Lithium, triptans, ondansetron, St John’s wort, MDMA, cocaine, amphetamines, LSD | Lithium contributes and also has its own toxicity — the lithium level interpreter covers that, and the two pictures overlap in tremor, hyperreflexia and myoclonus |
A decision tree, not a tally — and why that changes the answer
The Hunter Serotonin Toxicity Criteria are not a score, and treating them as one gives wrong answers. Dunkley and colleagues derived them by classification and regression tree analysis from 2,222 serotonergic drug overdose admissions, using a learning dataset of 473 overdoses of an SSRI alone and taking a clinical toxicologist’s diagnosis as the reference standard. What came out was a branching rule with five branches, each independently sufficient: in a patient exposed to a serotonergic agent, spontaneous clonus alone makes the diagnosis; so does inducible clonus with agitation or diaphoresis; so does ocular clonus with agitation or diaphoresis; so does tremor with hyperreflexia; and so does hypertonia with a temperature above 38 °C together with ocular or inducible clonus. There is nothing to add up and no threshold to reach.
That structure carries real information, which a point total would throw away. A patient with agitation, diaphoresis, tremor and a temperature of 37.9 °C has four of the nine features and meets no branch at all, because tremor needs hyperreflexia beside it and agitation and diaphoresis are only meaningful in the rule when clonus accompanies them. A patient with spontaneous clonus and nothing else has one feature and meets the criteria. Clonus is doing almost all of the diagnostic work — it appears in four of the five branches — because it is the sign that reflects the spinal and brainstem serotonin receptor activity the syndrome consists of. The practical consequence is that the examination matters more than the history: rapid dorsiflexion of the ankle for inducible clonus, and a deliberate look for slow continuous horizontal eye movements for ocular clonus. Neither is found by accident.
The performance figures are worth holding precisely. Against a clinical toxicologist’s diagnosis the Hunter criteria are 84% sensitive and 97% specific, where Sternbach’s earlier criteria managed 75% and 96% on the same data. Better on both axes with fewer features is an unusual result and it is why these displaced Sternbach in practice. But 84% means that roughly one patient in six whom an expert would diagnose does not meet the rule, and the missed patients are not randomly distributed: clonus is hard to elicit in a limb that is already rigid, or in a patient who has been sedated, which is exactly the patient in whom the diagnosis matters most. A negative result in someone on a new serotonergic combination is a reason to re-examine in an hour, not a reason to stop.
The differential that matters is neuroleptic malignant syndrome, and it turns on two things: how fast it came on, and what the muscles are doing. Serotonin toxicity develops within a few hours of a change in a serotonergic agent and produces neuromuscular hyperreactivity — hyperreflexia, clonus, tremor, lower-limb predominant. Neuroleptic malignant syndrome develops over days to weeks and produces a slowed neuromuscular response — generalised lead-pipe rigidity with reduced reflexes. Serotonin toxicity usually resolves within a day once the agent is stopped; neuroleptic malignant syndrome takes one to two weeks. Everything the two share — hyperthermia, diaphoresis, tachycardia, altered mental state, a raised creatine kinase, rhabdomyolysis — is useless for telling them apart, and the temperature in particular separates nothing.
Treatment follows from the mechanism rather than from the branch that was met. Stop every serotonergic agent. Sedate with benzodiazepines, which address the agitation and the muscle hyperactivity at once. Understand that the hyperthermia is produced by muscle activity and not by a raised hypothalamic set point, so antipyretics are useless and cooling plus abolition of the muscle activity is the whole of the treatment — up to intubation and neuromuscular blockade in a severely hyperthermic patient, with care over suxamethonium where rhabdomyolysis and hyperkalaemia are likely. Do not use physical restraint, which increases isometric contraction and makes the hyperthermia worse. Cyproheptadine is used by some services but the evidence is weak and it is oral only. And in every case, review the drug chart for the second serotonergic agent, because the syndrome is nearly always an interaction and the interacting drug is usually one nobody thought of as serotonergic.
Frequently asked questions
Are the Hunter criteria a score?
No, and this is the commonest way they are misused. They came out of a classification and regression tree analysis, so they are a branching decision rule with five independently sufficient branches. Spontaneous clonus alone satisfies the rule. Four features spread across branches without completing any of them satisfies nothing. A patient with agitation, diaphoresis, tremor and a temperature of 37.9 °C has four of the nine features and does not meet the criteria, because tremor requires hyperreflexia and agitation and diaphoresis only count alongside clonus. Any implementation that adds points changes which patients it diagnoses.
How accurate are the Hunter criteria?
Sensitivity 84% and specificity 97%, measured against the diagnosis of serotonin toxicity by a clinical toxicologist in 2,222 serotonergic drug overdose admissions, with a learning dataset of 473 SSRI-alone overdoses. On the same data Sternbach’s criteria gave 75% sensitivity and 96% specificity. The practical reading is that a positive result is reliable and a negative result is not a clearance: about one in six patients an expert would diagnose does not meet the rule, and those are disproportionately patients in whom clonus is hard to elicit because the limb is rigid or the patient is sedated.
Is the ocular clonus branch “agitation or diaphoresis” or “agitation and diaphoresis”?
Sources disagree, and it is worth knowing that rather than being surprised by it. The American Family Physician review of serotonin syndrome gives it as “ocular clonus with agitation or diaphoresis”; StatPearls renders the same branch with “and”; the Merck Manual merges the inducible and ocular branches into one condition reading “ocular or inducible clonus, plus agitation or diaphoresis”. Dunkley’s published abstract does not settle it. This page implements “or”, for three reasons: two of the three secondary renderings use it, it makes the branch symmetrical with the inducible clonus branch immediately above it, and it is the more sensitive reading — which is the safer direction of error for a rule used to catch a condition that kills through hyperthermia.
How do I tell serotonin toxicity from neuroleptic malignant syndrome?
By the onset speed and the pattern of tone, not by the temperature. Serotonin toxicity develops within a few hours of a change in a serotonergic agent and produces neuromuscular hyperreactivity — hyperreflexia, clonus and tremor, lower-limb predominant — and usually resolves within a day. Neuroleptic malignant syndrome develops subacutely over days to weeks and produces a slowed neuromuscular response — generalised lead-pipe rigidity with reduced reflexes — and takes one to two weeks to resolve. Hyperthermia, diaphoresis, tachycardia, altered mental state and a raised creatine kinase occur in both and discriminate nothing. The two are also not mutually exclusive, since antipsychotic and serotonergic drugs are commonly prescribed together.
Which drugs count as serotonergic?
SSRIs and SNRIs, tricyclics, monoamine oxidase inhibitors, lithium, triptans, ondansetron, St John’s wort, MDMA, cocaine and amphetamines — and, importantly, several drugs that are not thought of as psychotropic at all: tramadol, pethidine, fentanyl, methadone, dextromethorphan, linezolid and methylene blue. Linezolid and methylene blue are monoamine oxidase inhibitors, which is the fact most often missed. Serotonin toxicity is nearly always an interaction rather than a single-drug effect, so a patient stable on an SSRI for years can become toxic on the day tramadol or linezolid is started, and the interaction window for fluoxetine or an irreversible MAO inhibitor outlasts the last dose by weeks.
Is there a blood test for serotonin toxicity?
No. It is a clinical diagnosis, which is exactly why a validated decision rule exists for it. Serum serotonin concentrations have no role — measuring plasma or platelet serotonin is a carcinoid investigation and bears no relation to this syndrome — and drug concentrations do not correlate with severity. What the laboratory contributes is the assessment of complications rather than the diagnosis: creatine kinase for rhabdomyolysis, renal function, clotting for disseminated intravascular coagulation, lactate, and a blood gas. A normal set of bloods does not exclude serotonin toxicity and a deranged set does not diagnose it.
Related calculators
References
- Dunkley EJC, Isbister GK, Sibbritt D, Dawson AH, Whyte IM. The Hunter Serotonin Toxicity Criteria: simple and accurate diagnostic decision rules for serotonin toxicity. QJM. 2003;96(9):635–642. 2,222 serotonergic drug overdose admissions with a learning dataset of 473 SSRI-alone overdoses; reference standard the diagnosis of serotonin toxicity by a clinical toxicologist; clonus (inducible, spontaneous or ocular), agitation, diaphoresis, tremor and hyperreflexia were needed for accurate prediction, and hypertonicity with a maximum temperature above 38 °C were universal in life-threatening cases; sensitivity 84% against Sternbach’s 75% and specificity 97% against 96%.
- Serotonin Syndrome. StatPearls. NCBI Bookshelf. Lists the criteria as recent serotonergic exposure plus at least one of spontaneous clonus, inducible clonus with agitation or diaphoresis, ocular clonus with agitation and diaphoresis, tremor and hyperreflexia, or hypertonia and temperature over 38 °C with ocular or inducible clonus; 84% sensitivity and 97% specificity against a medical toxicologist’s diagnosis; serotonin toxicity develops within a few hours of a serotonergic agent change while NMS develops within days to weeks, and shows neuromuscular hyperreactivity against the slowed neuromuscular response and rigidity of NMS; resolution within one day against one to two weeks.
- Ables AZ, Nagubilli R. Prevention, diagnosis, and management of serotonin syndrome. Am Fam Physician. 2010;81(9):1139–1142. Renders the ocular clonus branch as “ocular clonus with agitation or diaphoresis” and the fifth branch as hypertonia with a temperature above 38 °C and ocular or inducible clonus; gives NMS as lead-pipe rigidity with bradyreflexia against the hyperreflexia of serotonin toxicity.
- Serotonin Syndrome. Merck Manual Professional Edition. Combines the clonus branches as “ocular or inducible clonus, plus agitation or diaphoresis, or hypertonia and temperature > 38° C”; distinguishes serotonin syndrome from neuroleptic malignant syndrome by serotonergic drug use, rapid onset within about 24 hours, and hyperreflexia in contrast to the often decreased reflex responses of NMS.
- Isbister GK, Buckley NA, Whyte IM. Serotonin toxicity: a practical approach to diagnosis and treatment. Med J Aust. 2007;187(6):361–365. The authors of the Hunter criteria on their application at the bedside, including the escalation to intubation and neuromuscular blockade for severe hyperthermia and the limited evidence for cyproheptadine.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.