Serotonin toxicity (ST) (aka serotonin syndrome) is a cluster of symptoms that develops with very high serotonin levels or serotonergic activity from direct agonists at certain serotonin receptors. It is most often seen with a combination of drugs, such as a monoamine oxidase inhibitor (MAOI) with a serotonin reuptake inhibitor (SRI), but it can also occur with a single drug, usually when a substantial overdose is used.
The effects of overly high serotonin exist on a spectrum. The mildest cases often won’t meet the diagnostic criteria for serotonin toxicity even though the observed symptoms are technically the result of high serotonin. This means a threshold of sorts must be crossed for serotonin toxicity to be diagnosed. Classically the condition has been defined as including symptoms from three categories: neuromuscular (arguably the most important), autonomic, and mental.
It may be an underdiagnosed condition given how many other conditions can present similarly to it, especially with mild cases, and due to clinicians simply being unaware of it. One report found over 85% of clinicians are unaware of the diagnosis itself (Sun-Edelstein, 2008). At the same time, the literature contains many reports of alleged serotonin toxicity that may not actually be ST. Much of the time the potentially inaccurate reports are attributable to the use of the first diagnostic criteria for serotonin toxicity described by Sternbach. Those criteria allow for a diagnosis to be made with primarily mental symptoms and without some of the more serotonin-specific symptoms like neuromuscular changes, thereby making it easier for misdiagnosis.
The clearest reports of ST are those that have significant neuromuscular symptoms like clonus, hyperreflexia, or rigidity. When relying on other mental and physical symptoms it’s possible for different conditions (e.g. neuroleptic malignant syndrome, anticholinergic toxicity, malignant hyperthermia, and simple drug overdose/adverse drug effects) to be diagnosed as ST. While not perfect, the Hunter diagnostic criteria described by Dunkley (2003) may be preferable because they emphasize neuromuscular effects.
Historically the more common term has been serotonin syndrome, but because “syndrome” may imply an idiosyncratic reaction, as with neuroleptic malignant syndrome, the term “serotonin toxicity” is preferable.
The list of drugs that can cause or significantly contribute to ST includes:
- Serotonin releasers (e.g. MDMA, MDA, 6-APB)
- SSRIs (e.g. paroxetine, fluoxetine)
- SNRIs (e.g. venlafaxine, tramadol)
- Some opioids (e.g. meperidine/pethidine, methadone)
- MAOIs (e.g. tranylcypromine, phenelzine, linezolid)
- Some psychedelics. There are few reports of this occurring with certain drugs, such as 2C-I and 25I-NBOMe. Overall it’s usually not a risk with psychedelics, especially LSD, psilocin, DMT, and other common ones.
- Enzyme inhibitors that inhibit the metabolism of a serotonergic drug.
Mostly serotonergic substances are the primary causal drugs, including MAOIs, SSRIs, and SNRIs. Tricyclic antidepressants (TCAs) aren’t always a major concern due to them generally being less potent serotonergics, as is seen with amitriptyline, though some, like clomipramine and imipramine, carry a notable risk of ST.
SRIs alone (in overdose) or in combination have been associated with ST. This includes all SSRIs and SNRIs like venlafaxine, which may actually be more concerning than SSRIs on average.
- SRIs (selective and nonselective): paroxetine, sertraline, fluoxetine, fluvoxamine, citalopram/escitalopram, venlafaxine, milnacipram, duloxetine, sibutramine, tramadol, dextropropoxyphene, pentazocine, meperidine, fentanyl, methadone, dextromethorphan, clomimpramine, imipramine. ST occurs in 14-16% of patients who overdose on SSRIs, according to data from the UK’s National Health Service (Sun-Edelstein, 2008).
Older, irreversible MAOIs like tranylcypromine are some of the most concerning drugs and fatal ST can easily arise from those drugs being taken with SRIs or serotonin releasers. All nonselective MAOIs and selective MAO-A inhibitors come with some potential for ST, including moclobemide, despite some reviews claiming it can be used with other serotonergics without risk of ST (Sun-Edelstein, 2008). Moclobemide has been implicated in multiple cases of ST (Neuvonen, 1993) when used with an SSRI. When used alone, moclobemide is relatively safe even in overdose. It can take weeks for enzyme activity to return to baseline after an irreversible MAOI is used, so 4-6 weeks should be given between ending an irreversible MAOI and starting a new serotonergic drug.
- MAOIs: Tranylcypromine, phenelzine, nialamide, isoniazid, iproniazid, paragyline, selegiline, clorgyline, moclobemide, toloxatone, furazolidone, procarbazine, linezolid.
Some opioids have enough serotonergic activity to be implicated in ST when combined with other drugs. This includes the phenylpiperidine opioids meperidine, tramadol, methadone, fentanyl, and propoxyphene. Fentanyl should be used cautiously but more evidence is needed to confirm it is problematic. Fentanyl derivatives like remifentanil appear to be safer due to having shorter half-lives and/or preferable pharmacodynamic properties.
Morphine and its analogs codeine, oxycodone, hydrocodone, and buprenorphine, among others, do not have known SRI properties and don’t appear to be associated with ST alone or in combination with serotonergics (Sun-Edelstein, 2008).
Serotonin releasers, particularly highly serotonergic releasers like the entactogens (e.g. MDMA, 6-APB, 5-MAPB, and MDA), can cause ST. Most often this occurs when combining them with other drugs, but overdoses can cause ST as well. When combined with SRIs the risk of ST is usually reduced due to the SRI blocking the releaser’s entrance to presynaptic terminals, thereby blocking the releaser’s activity.
Serotonin precursors carry some risk of ST, though essentially just in combination. Among the implicated precursors are 5-HTP and tryptophan. One of the earliest models of ST-like effects in animals used tryptophan combined with an MAOI, which has been shown to increase brain serotonin more than an MAOI used by itself (Sun-Edelstein, 2008). Tryptophan with MAOIs was also a cause of ST in humans when it used to be prescribed for depression.
Despite a warning from the FDA and other health agencies stating triptans can cause ST when combined with SRIs, there is little supporting evidence for this. The triptans are 5-HT1B, 5-HT1D, and 5-HT1F agonists; those receptors have not been implicated in the pathophysiology of ST. An alert about their connection to ST was issued by the FDA in 2006 based on 29 reports of apparent ST stemming from triptans combined with SSRIs, yet a subsequent analysis found the FDA-cited cases largely had questionable diagnoses and even if every FDA reported case was legitimate and 10% of all US cases of ST stemming from triptans/SSRIs were reported, the annual incidence of ST would be under 0.03% for patients exposed to both drugs, with an annual incidence of life-threatening events under 0.002% (Sun-Edelstein, 2008).
Other implicated drugs are chlorpheniramine and brompheniramine (due to SRI activity), dextromethorphan, tramadol, and lithium, which appears to potentiate serotonergics and ST (Goodwin, 1986). Because some of the causal drugs are found in cough/cold medications, the potential interaction with existing serotonergic drugs should be considered before taking those over-the-counter preparations. That being said, the normal use of an SSRI with an OTC preparation containing dextromethorphan, for example, typically won’t be dangerous.
Though rare, 5-HT2A agonists like psychedelics may cause serotonin toxicity on their own or in combination with other serotonergic drugs. A few case reports have shown this in humans, such as with 2C-I (Bosak, 2013) and animal studies have shown ST-like symptoms can be generated by LSD and 5-MeO-DMT (Goodwin, 1986 ; Lucki, 1982). If 5-HT2A is indeed the primary receptor responsible for ST it makes sense that 2A agonists can, at least in overdose, cause the condition.
Mirtazapine, triptans, ondansetron, olanzapine, buspirone, lithium (alone), trazodone, cyclobenzaprine, and olanzapine do not appear to cause ST (Foong, 2018). Rather than causing ST, mirtazapine has been shown in animals to inhibit hyperthermia and behavioral effects when rats are given an MAOI/SSRI combo (Shioda, 2010).
The three main effect categories are mental, autonomic, and neuromuscular and the most accurate diagnoses will come when a patient is showing symptoms from each category:
- Mental: Agitation, delirium, coma, mutism, consciousness impairment, restlessness, hypervigilance, pressured speech, and insomnia.
- Autonomic: Sweating, hyperthermia, tachycardia, tachypnea/dyspnea, diarrhea, hyper/hypotension, mydriasis.
- Neuromuscular: Hyperreflexia, clonus (e.g. myoclonus, ocular clonus, inducible clonus, or spontaneous clonus), hypertonia, tremor, shivering, and rigidity.
A lot of the same effects can be caused in a setting that is not concerning, such as the normal use of MDMA. Many recreational drug users become unnecessarily concerned about serotonergic drugs due to knowing too much serotonergic activity can be dangerous. In reality, it is a pretty uncommon occurrence when a single serotonergic is taken at a normal dose. And some combinations that are listed as having a risk of ST are actually not that concerning. As Buckley (2014) notes, a clinician searching drug interactions on a resource like Drugs.com may be warned of a thousand interacting drugs that are supposedly linked to “rare but serious” ST, yet ST is primarily linked to just a a couple dozen drugs from a few categories.
Lab findings include increases in white blood cell count (WBC) and creatine kinase (CK).
Someone experiencing ST will often initially be alert alongside less-risky symptoms like sweating, mydriasis, and increased heart rate. When neuromuscular symptoms do develop they are more pronounced in the lower limbs. In more severe instances the mental effects can deteriorate to agitated delirium (Sun-Edelstein, 2008).
The most common neuromuscular symptoms are tremor and hyperreflexia, followed by muscle rigidity and hyperetonia. Myoclonus is more common than clonus (Werneke, 2016). Ocular clonus can include a range of abnormal eye movements such as fine or coarse alterations of gaze in various directions. Those alterations can either be continuous or triggered by eye movement (Buckley, 2014).
While it has often been differentiated from neuroleptic malignant syndrome (NMS) based on it supposedly having an onset of a few hours after the initiation of a serotonergic drug or a dose increase, this doesn’t appear true. Symptoms can show up hours to days after a change in medication.
There’s little agreement between the three diagnostic criteria (Sternbach, Radomski, and Hunter), complicating diagnosis. Werneke (2016) demonstrated with a meta-analysis of cases published between 2004 and 2014 that only 48.8% of cases meet all three criteria and the Hunter criteria were found to identify fewer cases with overdose, rhabdomyolysis, and intensive care treatment compared to Sternbach and Radomski, making it useful for clinicians to consider all criteria when deciding if someone has ST, rather than exclusively using the Hunter criteria. 35.7% of rhabdomyolysis cases and 35.1% of cases requiring ICU treatment would not have been diagnosed as ST if strictly following the Hunter criteria and it’s unlikely all of those cases were misdiagnosed with the other criteria.
More severe cases, such as those requiring ICU treatment, produce significantly more clonus, rigidity, hypertonicity, hyperthermia, and tachypnea/dyspnea. Respiratory compromise may arise from rigidity affecting the truncal musculature and via separate serotonergic mechanisms.
Elevated temperature is very common, though a greater than 38°C temperature is only seen in 59.7% of cases and a temperature higher than 41.1°C is only seen in 9.2% (Werneke, 2016). Because of this, hyperthermia should not be relied on in diagnosis since diagnosing based on 38+°C temp will lead to missing ST in one-quarter of ICU cases and diagnosing based on 41.1+°C temperature will cause 4/5 of ICU cases to be missed. Hyperthermia is a primary cause of many of the worst effects of ST, including metabolic acidosis, rhabdomyolysis, and DIC.
There are many similarities between ST and NMS, which is why taking into account a patient’s medication/substance use history is vital during diagnosis. Altered mental state, autonomic instability, rigidity, increased temperature, and increased CK can occur in both. Though NMS can often be distinguished by the presence of lead-pipe rigidity rather than clonus/hyperreflexia, a lack of mydriasis, and exposure to neuroleptic agents or the withdrawal of dopamine agonists.
Diagnosis can be complicated by drugs having persistent activity (e.g. irreversible MAOIs) or long half-lives. Factors like those have sometimes led to drugs causing ST despite having been discontinued days to weeks prior. This also means it is best to discontinue serotonergic medications for at least a couple weeks before starting a new one.
Mild ST can be hard to differentiate from the mere non-toxic effects of a drug, Therapeutic SSRI use, for example, may cause altered mental status and some hyperreflexia or ankle clonus without toxicity (Buckley, 2014), so it’s easiest to diagnose the condition when a range of overt symptoms are present.
The original diagnostic criteria were outlined by Sternbach (1991) and those criteria are still the most widely used:
- Recent addition or increase in dose of a known serotonergic drug.
- Absence of other possible causes like infection, substance abuse, withdrawal, etc.
- No recent addition or increase in dose of a neuroleptic drug.
- At least 3 of these: mental status changes (confusion, hypomania), agitation, myoclonus, hyperreflexia, diaphoresis, shivering, tremor, diarrhea, ataxia, incoordination, and fever.
Because the Sternbach criteria allow for diagnosis with primarily mental symptoms there may be a higher than ideal risk of misdiagnosis. Certain symptoms, namely ataxia/incoordination, should also generally be left out of the diagnostic criteria given a lack of relevance to ST.
Radomski’s criteria also require the clinical features not be the result of a different acute condition or an underlying disorder and that a neuroleptic has not been recently initiated. Beyond that, there should be at least four minor or three major plus two minor symptoms from the following list associated with the recent addition or increase in dose of a serotonergic drug:
- Major mental: Consciousness impairment, elevated mood, semicoma/coma
- Minor mental: Restlessness, insomnia
- Major neurological: Myoclonus, tremor, shivering, rigidity, hyperreflexia
- Minor neurological: Incoordination, pupil dilation, akathisia
- Major vegetative: Sweating, fever
- Minor vegetative: Tachycardia, tachypnea/dyspnea, diarrhea, hyper/hypotension
The Hunter criteria published by Dunkley in 2003 place much greater emphasis on a handful of neuromuscular symptoms along with agitation and diaphoresis. In the presence of a serotonergic drug, if any of the following are present then ST can be diagnosed:
- Spontaneous clonus
- Inducible clonus and either agitation or diaphoresis
- Ocular clonus and either agitation or diaphoresis
- Tremor and hyperreflexia
- Hypertonicity and hyperthermia (temp over 38°C) and ocular clonus or inducible clonus
Dunkley (2003) found severe cases had a temperature over 38.5°C and/or significxant hypertonia or rigidity (particularly truncal) and in those cases there was a high risk of respiratory compromise, necessitating fast treatment such as neuromuscular paralysis, intubation, and assisted ventilation.
Some of the conditions that overlap enough with ST to result in misdiagnosis are anticholinergic toxicity, malignant hyperthermia, NMS, drug overdose, benzodiazepine or alcohol withdrawal, and meningitis or encephalitis (Foong, 2018).
In animal models high serotonin is associated with a behavioral response featuring hyperactivity and reactivity, forepaw treading, head-weaving, hind-limb abduction, and the Straub or arched tail, sometimes along with tremor, rigidity, salivation, flushing, myoclonus, and seizure (Martin, 1996). This set of symptoms has been used to investigate the mechanisms of apparent ST in animals and the findings underlie much of what is known about ST in humans.
Animal models, though useful, are complicated by the fact that the generated symptoms aren’t identical to those in humans and there are also species differences, such as serotonin/5-HT2A agonists causing hyperthermia in rats and hypothermia in mice (Harberzettl, 2013).
Prognosis and Treatment
The prognosis is good when it’s recognized early, the causal drugs are stopped, and supportive or pharmacologic treatments are initiated.
ST has been fatal and in severe cases can easily be lethal if it’s not quickly treated; life-threatening ST occurs in up to 50% of those ingesting an adequate combo of an MAOI and SSRI, for example (Sun-Edelstein, 2008). With treatment, the condition usually resolves within a matter of days without lasting effects, though altered mental status and muscle pain can persist for a few days and occasionally longer.
When fatalities occur they’re usually associated with issues like renal failure, DIC, cardiac arrhythmia, hyperthermia, myoglobinuria, respiratory arrest, and rhabdomyolysis (Bodner, 1995).
The most common treatments include benzodiazepines, other forms of sedation (e.g. opioids and barbiturates), cooling measures, and 5-HT2A antagonists like cyproheptadine and chlorpromazine. 5-HT1A agonism has often been implicated as the primary causal mechanism in ST, but a lot of animal and human evidence supports a greater role of 5-HT2A, particularly in moderate to severe cases, making 5-HT2A antagonists more useful than 1A antagonists (Martin, 1996). For severe ST, intravenous use of chlorpromazine appears to be the most effective 2A antagonist (it should be given alongside IV fluids to prevent hypotension), while in moderate cases oral cyproheptadine can be used, though it’s unclear if its general sedating effects or specific 2A antagonist property is responsible for the benefits (Buckley, 2014).
Hyperthermia is a common finding in the condition and it’s an important source of toxicity. Despite this, dantrolene and antipyretics are not recommended as they target temperature increases stemming from mechanisms unrelated to those seen in ST.
Benzodiazepines are a nonspecific way to treat some of the core symptoms, including increased muscle tone and agitation due to a reduction in sympathetic activity. They have been very effective in a number of cases, but sometimes alternative treatments are required to yield any response.
Toxicity is greatest when hyperthermia and increased muscle tone/clonus/hyperreflexia go inadequately treated, making aggressive treatment important in severe cases that show signs of increased muscle tone, increased temperature, and/or rhabdomyolysis. Rhabdomyolysis is a fairly common complication in severe cases, though the majority of ST cases don’t lead to it. Occasionally other severe conditions like disseminated intravascular coagulation (DIC) can develop.
Non-depolarizing neuromuscular blockers (e.g. vecuronium) are recommended in instances where muscle tone is not adequately controlled by benzodiazepines.
Drugs to avoid include pancuronium, succinylcholine, and bromocriptine are not recommended (Katus, 2016). Succinylcholine should be avoided because of the risk of arrhythmia from hyperkalemia caused by rhabdomyolysis. Bromocriptine, which is dopaminergic and a 5-HT1A agonist, has been found to worsen symptoms when patients are accidentally treated as if they have NMS.
Serotonin is synthesized from the dietary amino acid tryptophan, which undergoes hydroxylation (the rate-limiting step) inside the cytoplasm of neurons and other cells. Then it is decarboxylated to form serotonin and it’s stored in vesicles for release upon depolarization. Because serotonin itself doesn’t cross the blood-brain barrier (BBB), it must be independently produced peripherally and centrally. In the periphery it’s mostly made by intestinal enterochromaffin cells and centrally it’s produced by nuclei in the lower pons and upper brain stem, regions that have axonal projections to almost every part of the CNS (Mills, 1997).
The level of serotonin within the CNS and outside the CNS is usually tightly regulated by cellular uptake via the serotonin transporter (SERT) and via metabolism to 5-HIAA by the enzyme monoamine oxidase (MAO), which is present on the mitochondrial membrane, but enough serotonin to cause toxicity can accumulate, most often when multiple drugs are used or a single serotonergic is overdosed. More specifically, serotonin is metabolized by MAO-A.
Seven types of serotonin receptors and many subtypes exist, though 5-HT2A is the primary receptor implicated in ST, with possible (though inadequately confirmed) contributions from 5-HT1A and 5-HT3. It’s been proposed 5-HT1A, a more abundant receptor, may be involved in the minor-to-moderate symptoms of ST while overactivation of 2A is vital for all or most of the truly toxic effects (Ellahi, 2015). This is potentially supported by the finding that the combination of 5-HTP and clorgyline, which is known to cause ST, yields hypothermia at low doses and hyperthermia at higher doses (Harberzettl, 2013). 1A is both more common and a much higher affinity site for serotonin than 2A, so high serotonin levels will initially cause more effects via 1A.
Indirectly from increases in serotonergic activity there may be a rise in glutamatergic, noradrenergic, and dopaminergic function, which could play important roles in the physiology of ST. Dopamine activity might be important since dopamine antagonists like haloperidol can treat ST in animals (Martin, 1996). Some of the autonomic symptoms (e.g. diaphoresis and tachycardia) point to a role for hyperactive noradrenergic activity.
Many of the symptoms are primarily mediated by central effects, but peripheral activity could be important in neuromuscular activity and vomiting and diarrhea (Harberzettl, 2013).
Hyperthermia appears to be caused in large part by excess muscular activity; a change in the hypothalamic temperature set point from a pyrogen isn’t involved, which is why antipyretic drugs are not used in treatment. Because physical restraints can increase muscle contractions (raising the risk of lactic acidosis, hyperthermia, and rhabdomyolysis), they should be avoided. Though not the result of pyrogen activity, some reports have noted hyperthermia may be partly coming from activity at hypothalamic 5-HT2A receptors, which could impair temperature regulation (Steele, 2010).
Some researchers have proposed ST shares some of the same pathophysiology as NMS because high serotonin levels could potentially depress dopamine activity in some regions, producing a relatively hypodopaminergic state and therefore extrapyramidal symptoms (Martin, 1996), though the two conditions still have significant mechanistic and symptomatic differences. When ST was less well known, some cases of obvious ST were actually described as neuroleptic malignant syndrome “without neuroleptics.”
Individuals could have different susceptibilities to ST due to genetic differences at SERT, serotonin receptors, enzymes involved in drug metabolism (e.g. CYP2D6), and P-glycoprotein.
In the musculature, serotonin can increase the frequency and duration of firings, leading to clinical symptoms like tremor, clonus, and hyperreflexia. Stimulation of central 5-HT2A is linked to hyperthermia, while 1A causes hypothermia, implicating the former in the generation of ST. Similarly, animal models have found 1A agonism decreases HR and BP, while 5-HT2 agonism does the opposite and indeed, increased HR and BP is much more common than low BP/HR in ST (Brown, 1996).
Metabolic analysis in rats with ST induced by 5-HTP and the MAOI clorgyline showed hyperthermia could be coming from an increase in the expression of the mitochondrial protein UCP-3 (Zaitsu, 2018). Many metabolic changes were found in muscles, the liver, and plasma in ST rats. Along with a rise in UCP-3 (in the gastrocnemius muscle but not the trapezius) there was a significant increase in urea cycle metabolites and signs of anaerobic respiration.
Downregulation of serotonin receptors may be protective against ST. Chronic MAOI administration has been shown to protect against ST-like symptoms in rats, while TCAs given for the same length of time (7 days) did not protect against ST (Lucki, 1992). The protection from chronic MAOI administration was associated with a decline in serotonin binding sites in the brain stem and spinal cord; the TCAs failed to downregulate serotonin receptors.
Serotonin availability is required for the presentation of ST, just as serotonin receptors must be adequately available. The effects can be blocked by para-chlorophenylalanine, which inhibits serotonin synthesis (Sternbach, 1991). Decreasing serotonin synthesis by decreasing tryptamine is also an effective way to attenuate the condition (Martin, 1996).
Diaz (2011) showed 5-HT2B is protective against ST in mice. When it’s absent, such as with 2B knockout mice or with a 2B antagonist, serotonin toxicity is significantly enhanced.
It’s possible the reports of “convulsive ergotism” stretching back to the 11th century may have involved serotonin toxicity (Ganetsky, 2005). The symptoms reported in those cases included distortion of trunk and limbs, ankle flexion or extension, drowsiness, mania, hallucinations, diarrhea, sweating, fever, muscle stiffness, twitching, and opisthotonus. Ergot alkaloids are known to stimulate 5-HT1 and 5-HT2 receptors, giving a potential route to ST.
The modern history of ST goes back to the 1950s. It was first reported in humans treated with MAOIs in the early 1950s to 1960s, sometimes with it being prescribed alongside tryptophan (Martin, 1996). A dose-response effect was found with tryptophan combined with MAOIs:
- 12-18 mg/kg with phenelzine 60 mg: No hypertensive effect (Sternbach, 1991)
- 20 mg/kg: Altered mental status and an inebriated state.
- 30 mg/kg: Hyperreflexia and clonus
- 50+ mg/kg: Diaphoresis and myoclonus.
- 70-90 mg/kg: May cause ST alone. Also causes euphoria, drowsiness, nystagmus, hyperreflexia, ankle clonus, and clumsiness (Sternbach, 1991).
A report in 1955 described a patient with severe muscle twitching, ankle clonus, and Babinski signs who was given iproniazid (an MAOI) with the opioid meperidine (Bodner, 1995). The condition was further described by Oates and Sjoerdsma who reported mental changes, restlessness, lower-extremity hyperreflexia, and diaphoresis in four patients in the early 1960s (Bodner, 1995). They attributed the symptoms to excess serotonin activity caused by the use of high tryptophan doses with MAOIs. Oates and Sjoederma also gave single doses of tryptophan (20-50 mg/kg) with the MAOI beta-phenylisopropylhydrazine and reported it caused dizziness, drunken-like effects, clonus, restlessness, hyperactive reflexes, and diaphoresis without significant cardiovascular changes.
Hodge et al. reported giving five hypertensive patients a decarboxylase inhibitor before they received the MAOIs pargyline or isocarboxazid and this was shown to block nystagmus, jaw tremor, hyperreflexia, clonus, sweating, and intoxication, implicating serotonin in the condition (Sternbach, 1991).
Between the mid-1950s and early 1980s there were many case reports of serotonin toxicity. Even by 1962 there were already at least 20 separate reports of apparent ST in dozens of patients: 18 involved MAOIs with imipramine (7 fatalities) and 10 involved MAOIs with meperidine (3 fatalities). Nearly all were reported without the authors connecting the observed effects to the ST-like condition that had been reported in animals, so it would take decades for ST to be well-understood and widely known (Gillman, 1998).
Interactions between MAOIs and TCAs leading to sweating, muscle twitches, restlessness, rigidity, hyperthermia, and loss of consciousness were reported by Beaumont in 1973. Some of those cases were fatalities. All fatalities involved patients who were already on an MAOI (tranylcypromine or mebanazine) at the time the TCA was added.
The term “serotonin syndrome” was first used in animal research in the late 1970s and then in a human case report in the early 1980s.
The first ST-related death to receive substantial attention was that of Libby Zion in 1984. It is believed she died from an interaction between the MAOI phenelzine and meperidine. Her death eventually led to nationwide regulations in the US to limit working hours among medical staff (Martin, 1996).
By the 1980s various drugs had been shown to counteract ST in humans and animals, including chlorpromazine, thiothixene, clozapine, pimozide, methysergide, and cyproheptadine. Given what was known about these drugs at the time, their efficacy appeared to be evidence of a substantial role of 5-HT2A in the disorder since those drugs are generally much more potent at 2A than 1A.
Sternbach laid our diagnostic criteria for ST in 1991 based on a review of 38 case reports in which the symptoms were: restlessness (45%), confusion (42%), myoclonus (34%), hyperreflexia (29%), and sweating, shivering, and tremor (26% each). Nearly all of those cases came from the use of antidepressants with other substances.
In 2001 Radomski published an updated diagnostic criteria based on a larger review of 62 cases (including Sternbach’s original 38) and differentiated between major and minor symptoms, along with adding rigidity to the list of neuromuscular symptoms.
Another set of criteria was published in 2003 by Dunkley using 2222 cases of SSRI overdose in order to more precisely identify serotonin-specific effects, leading to the conclusion that an excess specifically of serotonin will yield clonus as the most important diagnostic symptom. Dunkley’s Hunter criteria limited the role of mental symptoms in the diagnosis, which has often been considered a good way to increase precision. However, the fact that it only used SSRI overdose cases could make the outlined symptomatology less generalizable to other types of serotonergic drugs and to combinations of substances, so strictly using the Hunter criteria may lead to false negatives. Also, the criteria were validated partially using the same cases used to derive the symptoms, which could have produced an overestimation of the criteria’s precision. Regardless, the Hunter criteria do seem superior to the more common Sternbach criteria because of a lower risk of false positives.
Because so many case reports have been published utilizing either no formal diagnostic criteria or using Sternbach’s criteria, some of the case reports outlined here may not be demonstrating actual ST, even though the authors have reported the presence of ST.