Opioid analgesic drugs and serotonin toxicity (syndrome): mechanisms, animal models, and links to clinical effects
Drugs may cause serotonin toxicity by a number of different mechanisms including inhibition of serotonin uptake and metabolism, increased serotonin synthesis and release, activation of serotonin receptors, and inhibition of cytochrome P450 oxidases. Some drug interactions involving opioids can increase intrasynaptic levels of serotonin, and opioid analgesic drugs are now recognized as being involved in some cases of serotonin toxicity especially if administered in conjunction with other serotonergic medications including monoamine oxidase inhibitors, selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, and tricyclic antidepressants. In March 2016, the FDA issued a Drug Safety Communication concerning the association of the entire class of opioid pain medicines with serotonin toxicity. Reports of the involvement of individual opioids particularly tramadol, tapentadol, meperidine, methadone, oxycodone, fentanyl, and dextromethorphan are reviewed. While relevance to human serotonin toxicity of animal models, including many studies on rat brain synaptosomes, is questionable, important insights have recently been forthcoming from research utilizing 5-HT receptors, serotonin transporter (SERT), and knockout mice. In studies with human SERT-transfected human HEK293 cells, the synthetic opioids tramadol, meperidine, methadone, tapentadol, and dextromethorphan inhibited SERT, but fentanyl and a number of phenanthrenes including morphine and hydromorphone did not. Receptor ligand-binding assays revealed interaction of fentanyl with 5-HT1A receptors and interaction of meperidine, methadone, and fentanyl with 5-HT2A receptors. Although the opioids most often associated with serotonin toxicity in humans inhibit human SERT in vitro, fentanyl and oxycodone are not inhibitory even though their clinical involvement has been reported. This suggests some SERT-independent effects on the serotonin system in vivo. Heightened clinician awareness of the possibility of serotonin toxicity among patients taking opioids and serotonergic antidepressants is called for.
KeywordsSerotonin toxicity Serotonin syndrome Opioid analgesics Opioids and serotonin toxicity Opioids as serotonergic drugs Diagnosis of serotonin toxicity Mechanisms of serotonin toxicity Tramadol and serotonin toxicity
Introduction to the Adverse Effects of Opioid Analgesic Drugs
Opioid analgesic drugs (OADs) are amongst the most commonly administered drugs in hospitals and widely used for pain relief outside the hospital environment (Duthie and Nimmo 1987). OADs provoke a wide variety of adverse events in patients ranging in severity from mild to life threatening (Schug et al. 1992). More commonly seen, and less hazardous, adverse reactions include nausea, vomiting, sedation, dizziness, constipation, withdrawal symptoms, hypotension, an increased risk of seizures, sweating, dysphoria, and euphoric mood. Some cardiovascular disorders may occur including tachycardia, bradycardia, and palpitations; cutaneous reactions include pruritus, urticaria, and rash. Serious events listed in US Food and Drug Administration (FDA) boxed warnings cover addiction, abuse, and misuse; life-threatening respiratory depression; neonatal opioid withdrawal; risks for concomitant use with benzodiazepines and other central nervous system (CNS) depressants; interaction with alcohol; and accidental ingestion [Morphine sulfate. Highlights of prescribing information (2016); Nucynta® ER (Tapentadol). Highlights of prescribing information (2016)].
Apart from adverse clinical effects resulting from opioid ligands binding specifically to opioid receptors, principally manifestations of the primary general effects of central depression and sedation (Freye 2008), a number of different mechanisms are known to underlie a variety of adverse reactions to OADs. One such ‘non-specific’ mechanism is the drugs’ capacity for releasing histamine (Baldo and Pham 2012; Nasmyth and Stewart 1950). Although OADs as a group are often assumed to be potent releasers of histamine, individual opioids show differences in the amount of histamine they release and in the anatomical sites where release occurs (Baldo and Pham 2012). Despite the heavy usage of OADs, especially the analgesics morphine, codeine, fentanyl, meperidine (pethidine), methadone, and tramadol, and the misuse of heroin and methadone, hypersensitive reactions (types I–IV) to the drugs are uncommon (Baldo 2017; Baldo and Pham 2013; Levy 1986).
A more recently studied adverse effect of OADs which, at first site, is not mediated via specific opioid receptor interaction, is toxicity resulting from elevated intrasynaptic serotonin concentrations commonly referred to as serotonin syndrome (Boyer and Shannon 2005; Gillman 2006a, b; Gillman and Whyte 2004; Isbister and Buckley 2005; Sternbach 1991; Volpi-Abadie et al. 2013), or more accurately, serotonin toxicity (see below). The former term has a number of different meanings, not always differentiated, and this can lead to misdiagnosis of a number of conditions showing a similar spectrum of symptoms. One consequence of this can be subsequent ineffectual treatments (Isbister and Buckley 2005).
In March 2016, the FDA issued a Drug Safety Communication containing three safety announcements for the entire class of opioid pain medicines. Required changes to the labels of all opioid drugs were called for, warning of risks for serotonin “syndrome”, adrenal insufficiency, and androgen insufficiency (FDA Drug safety communications 2016). This announcement came some years after earlier FDA safety labeling revisions for tramadol (as early as 2009), tapentadol, and meperidine, which drew attention to the potential risk for serotonin “syndrome” as a result of drug interactions.
Here, I review the increasingly recognized role of OADs in serotonin toxicity which is largely a consequence of the many different combinations of serotonergic drugs already in use and being introduced. Non-opioid drugs commonly implicated and different mechanisms underlying toxicity are also examined. Treatment of serotonin toxicity and supportive care are beyond the scope of this review and are not covered.
Serotonin ‘syndrome’, as it was originally called (Insel et al. 1982; Sternbach 1991), is a variable condition with symptoms ranging from mild to life threatening. The condition, an iatrogenic drug-induced toxidrome (Gillman 2006a, b), was first reported by Mitchell (1955) and mechanistically linked with raised intrasynaptic serotonin concentrations (Oates and Sjoerdsma 1960) associated with increased serotonergic activity in the central nervous system (CNS). Serotonin ‘syndrome’ suggests an idiosyncratic reaction, but ‘toxicity’ is preferred since the condition is a concentration-dependent toxicity that can occur in any individual and produce a spectrum of toxicity (Dunkley et al. 2003).
Symptomatology of serotonin toxicity (Boyer and Shannon 2005; Dunkley et al. 2003; Gillman and Whyte 2004; Isbister and Buckley 2005; Isbister et al. 2007; Sternbach 1991; Volpi-Abadie et al. 2013; Whyte 2004) is often viewed as a triad of changes in mental status, for example, agitation, anxiety, delirium, restlessness, confusion, disorientation, and excitement; autonomic stimulation, such as hyperthermia, tachycardia, tremor, diaphoresis, mydriasis, and flushing; and neuromuscular excitation, including clonus, hyperreflexia, myoclonus, and rigidity. In mild cases, patients are usually afebrile with mild hypertension, tachycardia, dilated pupils, shivering, sweating, myoclonus, and showing hyperreflexia. In moderate cases, patients also experience other symptoms including a rise in temperature (~ 40 °C), mild agitation, and clonus, while in severe cases the above symptoms are accompanied by a higher temperature (> ~ 41 °C), delirium, muscle rigidity, and variations in pulse and blood pressure. Occurring in all age groups including the newborn, the incidence of the toxicity is unknown but it is probably not always diagnosed in its mild form. The incidence is also thought to be increasing, probably because of the increasing use of serotonergic medicines and heightened awareness of the condition (Boyer and Shannon 2005; Gayle et al. 2015).
Serotonin, or 5-hydroxytryptamine, acts centrally and peripherally, modulating attention, aggression, thermoregulation, appetite, and behavior in the CNS and stimulates gastrointestinal motility, vasoconstriction, bronchoconstriction, uterine contraction, and platelet aggregation in the peripheral nervous system, where it is produced primarily in the intestinal enterochromaffin cells (Mason et al. 2000). The many actions of serotonin are mediated via at least seven receptors and although no single receptor appears to be responsible for the toxicity, the postsynaptic 5-HT1A and 5-HT2A receptors, particularly the latter, seem to be the most important ones involved (Boyer and Shannon 2005; Isbister and Buckley 2005; Mills 1997; Nisijima et al. 2001). Serotonin toxicity most commonly occurs after the administration of two serotonergic drugs, but it can also occur after a single drug or by increasing the dose of a serotonergic drug. Drug-induced serotonin toxicity may occur via a number of different mechanisms including inhibition of serotonin reuptake and metabolism, increasing serotonin synthesis and presynaptic release, and activation of serotonin receptors (Volpi-Abadie et al. 2013). Inhibitors of serotonin uptake include the selective serotonin reuptake inhibitors (SSRIs), for example, citalopram, fluoxetine, fluvoxamine, paroxetine, and sertraline; serotonin-norepinephrine reuptake inhibitors (SNRIs), e.g., desvenlafaxine, duloxetine, and venlafaxine; tricyclic antidepressants (TCAs), e.g., amitriptyline, clomipramine, imipramine, and nortriptyline; opioids, including tramadol; meperidine, methadone, and fentanyl; some drugs of abuse, e.g., cocaine and 3,4-methylenedioxymethamphetamine (MDMA; commonly known as ecstasy); some antidepressants such as bupropion; the antihistamine chlorpheniramine; the herb St John’s Wort; and the over-the-counter opioid cold treatment dextromethorphan. Inhibitors of serotonin metabolism include the monoamine oxidase inhibitors (MAOIs), such as isocarboxazid, linezolid, methylene blue, phenelzine and selegiline; triptans, e.g., almotriptan, frovatriptan, rizatriptan, sumatriptan, and zolmitriptan. It seems that cases of serotonin toxicity involving a MAOI are often more severe, leading to death in a few instances (Gillman 2006a, b; Vuori et al. 2003). Some drugs that increase the synthesis of serotonin are phentermine, the amphetamines used for weight loss, the dietary supplement L-tryptophan, and cocaine. Drugs that increase the release of serotonin include dihydroergotamine and triptans used for migraine; some opioids particularly meperidine and fentanyl; antidepressants mirtazapine and trazodone; the anxiolytic buspirone; LSD; and lithium. In addition to the above mechanisms, SSRIs inhibit some cytochrome P450 enzymes (CYPs) that catalyze the oxidative biotransformation of many drugs (Zanger and Schwab 2013), in particular CYP2D6 and CYP3A4, leading to an accumulation of serotonergic drugs (as substrates) that would otherwise be metabolized, thus increasing serotonergic activity. Note that serotonin toxicity occurs in 15% of SSRI overdoses (Dunkley et al. 2003; Isbister et al. 2007). Other drugs besides SSRIs inhibit CYPs, for example inhibition of CYP3A4 by ciprofloxacin (see sect. “Methadone” below) and the antifungal fluconazole which inhibits CYP2C19 resulting in accumulation of its substrate citalopram (Levin et al. 2008). In relation to triptans, there is no universal agreement that these agents are a significant risk for serotonin toxicity when used with SSRIs and SNRIs (Evans et al. 2010; Orlova et al. 2018; Soldin and Tonning 2008). Gillman (2010) has pointed out that triptans, by themselves or with other drugs, do not show serotonergic activity and do not possess the pharmacological properties to mediate serotonin toxicity.
Diagnosis of serotonin toxicity
Out of a number of different criteria advanced for the diagnosis of serotonin toxicity, the clinical features put forward by Sternbach (1991) and based on ten case reports and two case series, remained the most commonly used clinical diagnostic definition until the early 2000s. After ruling out other aetiologies such as infections, substance abuse, and withdrawal, the addition or increase of a serotonergic agent, and no addition/increase of a neuroleptic agent, Sternbach suggested that at least three of the following ten symptoms were diagnostic for serotonin toxicity: mental status changes (confusion, hypomania), agitation, myoclonus, hyperreflexia, diaphoresis, shivering, tremor, diarrhea, lack of coordination, fever. The Sternbach criteria, however, are non-specific (recognized by Sternbach himself), have never been validated, are weighted to patients with an abnormal mental state, and have many features in common with some other adverse drug reactions particularly neuroleptic malignant syndrome but also malignant hypothermia, anticholinergic toxicity, serotonergic discontinuation syndrome, sympathomimetic drug intoxication, encephalitis, meningitis, central hyperthermia, and heat stroke (Isbister et al. 2007; Volpi-Abadie et al. 2013). Like the Sternbach criteria, the so-called Radomski diagnostic criteria (Radomski et al. 2000) are based on cognitive, neuromuscular, and autonomous symptoms, each divided into major and mild symptoms. The Radomski criteria specify coincidence with a serotonergic agent and the development of at least three major plus two minor symptoms or at least four minor symptoms. The specified symptoms are: major—consciousness impairment, elevated mood, semicoma/coma, myoclonus, tremor, shivering, rigidity, hyperreflexia, fever, and sweating; and minor—restlessness, insomnia, uncoordination, dilated pupils, akathisia, tachycardia, dyspnea, diarrhea, and hyper/hypotension.
Concerns have been expressed about the Sternbach criteria, not least of which is the inclusion of four criteria related to mental status that weight the diagnosis toward patients with abnormal mental status (Dunkley et al. 2003). Other perceived deficiencies in the Sternbach and Radomski criteria, especially the former, led to the development of specific diagnostic criteria for serotonin toxicity, viz., the Hunter Serotonin Toxicity Criteria (HSTC) (Dunkley et al. 2003). This was done by constructing a decision tree via a series of ‘if-then’ decision rules. Validated by the Hunter Area Toxicity Service which deals with admitted consecutive patients frequently with other drug-induced syndromes, these criteria comprise a workable set of clinical features that distinguish serotonin toxicity. For accurately predicting the toxicity, only the following symptoms were required: spontaneous clonus, inducible clonus, ocular clonus, agitation, diaphoresis, tremor, and hyperreflexia. Clonus, spontaneous, inducible, and ocular, is strongly associated with serotonin toxicity and is the most important sign in the HSTC. In patients with serotonin toxicity due to an overdose of an SSRI alone, all types of clonus were seen and, if in the presence of a serotonergic agent, spontaneous clonus is the only sign; serotonin syndrome may be reliably diagnosed. However, since clonus and hyperreflexia are often not demonstrable in patients with severe rigidity, an extra decision rule was included for patients with severe life-threatening serotonin toxicity with a temperature > 38 °C and hypertonicity/rigidity (Dunkley et al. 2003). Employing the HSTC, clinical assessment of patients with suspected serotonin toxicity should include observation for tremor, myoclonic jerks, ocular clonus, diaphoresis, and agitation as well as examination of vital signs (temperature, blood pressure, heart rate) and a neurological examination focusing on mental state, limb tone, clonus and reflexes, pupillary size, and opsoclonus (Dunkley et al. 2003; Isbister et al. 2007). Diagnostic application of the HSTC shows a sensitivity of 84% and specificity of 97% and compared to Sternbach’s criteria, the HSTC are less likely to detect false positives.
Although frequently accepted as the criterion standard, it has been claimed that the Hunter criteria may not be superior to the other diagnostic criteria. Werneke et al. (2016), point out that the HSTC were derived exclusively from SSRI overdoses, that some cases used to derive the criteria were also used for validation, not all cases are of rapid onset, only relatively few cases present with hyperthermia, and application of the criteria may not be reliable in patients with underlying neurological pathologies, for example, reflexes or clonus may not occur in patients with peripheral neuropathy. For diagnosis, the authors propose instead a meta-analysis of cases focussing on etiology rather than relying on symptoms and symptom constellations.
Having diagnosed serotonin toxicity in a patient, it is important to assess the severity of the toxicity. Three groups of severity can be distinguished—those patients exhibiting mild serotonergic features, patients showing moderate toxicity, and those experiencing severe serotonin toxicity, a medical emergency also known as a serotonin crisis. Most often, this crisis appears to be strongly associated with a MAOI and an SSRI acting at different sites. Mild serotonergic features may occur after taking many serotonergic agents and generally do not greatly inconvenience patients; moderate toxicity induces significant distress and requires treatment (Dunkley et al. 2003; Gillman 2006a, b; Isbister et al. 2007; Boyer and Shannon 2005). Some deaths have been recorded in severe cases (Otte et al. 2003; Gillman 2006a, b).
Opioids and serotonin toxicity
Despite limited available clinical and experimental data, attention has been drawn to a number of severe cases of meperidine-induced serotonin toxicity, including some fatalities Gillman (2005). The borderline to weak SRIs synthetic OADS, methadone, fentanyl and congeners, tramadol, dextromethorphan, and propoxyphene have been implicated in many likely cases of serotonin toxicity. Morphine and its analogs, however, which are not SRIs, do not induce serotonin toxicity with MAOIs (Gillman 2005) and there are few, if any, reports of the toxicity with these drugs (but see oxycodone and codeine below). According to Boyer and Shannon (2005) and Rastogi et al. (2011), opioids in the presence of other serotonergic agents increase intrasynaptic concentrations of serotonin. In a retrospective analysis of 203 cases of serotonin toxicity registered in the French pharmacovigilance database in the period 1985–2013 (Abadie et al. 2015), the Sternbach (1991), HSTC (Dunkley et al. 2003; Isbister et al. 2007), and Radomski et al. (2000) clinical diagnostic criteria were applied to identify 125 patients who the authors described as having “indisputable serotonin syndrome”. Of the 125 cases, 86 fulfilled the Sternbach diagnostic criteria, 70 the HSTC, and 68 the Radomski criteria, meaning that some of the individuals were probably positive to only 1, or perhaps 2, of the three different criteria and therefore the diagnosis of some of the patients was likely to be on the basis of the Sternbach criteria alone. The fact that 55 more cases of serotonin toxicity were assigned according to the Sternbach criteria than the HSTC is a concern since the latter identifies a more specific set of clinical features with an emphasis on clonus which has been shown to be more specific for serotonin toxicity (Dunkley et al. 2003; Isbister et al. 2007; Whyte 2004; Gillman and Whyte 2004). Of the patients exposed to serotonergic drugs, 42.1% were given SSRIs, 14.8% received opioids (tramadol 20, dextropropoxyphene 5, methadone 3, fentanyl 2, remifentanil 1), 11.5% MAOIs, 9.1% nonimipraminic SNRIs, 8.6% TCAs and a total of 13.9% other drugs including some other antidepressants (including mirtazapine; c.f. Gillman 2006), amphetamines, lithium, buspirone, metoclopramide, aripiprazole, and tryptophan. Looking at the drug associations, of 74 cases, the association between SSRIs and opioids occurred most often (20.3%), followed by the SSRI-MAOI combination; the MAOI-opioid association was seen in 5.4% of patients. For individual drugs, opioids, mainly tramadol, were the second most implicated drug group while tramadol plus the potent CYP2D6 inhibitor paroxetine was the most frequently found combination (Abadie et al. 2015). Some OADs are now recognized by the FDA as serotonergic drugs sometimes associated with serotonin toxicity by different mechanisms, especially if administered in conjunction with other serotonergic medications (Gillman 2005; Gnanadesigan et al. 2005; FDA Drug safety communications 2016).
Reports of individual opioids and serotonin toxicity
Prior to the safety announcement by the FDA concerning the association of opioids with serotonin syndrome (FDA Drug Safety Communication 2016), the opioids meperidine, tramadol, and tapentadol were already subject to FDA warnings concerning serotonin syndrome, but with other clinically important members of the opioid group now implicated, the important and commonly used OADs will be considered individually. Before doing so, however, it should be pointed out that case reports of serotonin toxicity have come in for a number of criticisms including the inconsistent reporting of toxic events, failure to report important positive and negative findings, failure to adequately review and quote the relevant literature, and the presentation of incomplete or erroneous information on drugs involved, symptoms, and treatments (Gillman 2006).
The centrally active analgesic tramadol is the prodrug for its active metabolite O-desmethyltramadol. Both compounds are weak-intermediate agonists of the µ opioid receptor. Involvement of a second mechanism of action was suggested by the failure to fully inhibit the analgesic action by the µ receptor antagonist naloxone and this was found to be due to monoaminergic activity inhibiting reuptakes of serotonin and norepinephrine. Although tramadol exists as four stereoisomers, the marketed product is a racemic mixture of the (1R,2R) or (+)-enantiomer and the (1S,2S) or (−)-enantiomer. (+)-Tramadol is approximately four times as potent as (−)-tramadol in inhibiting serotonin uptake (Driessen and Reimann 1992) while (−)-tramadol is about ten times more potent than the (+)-enantiomer in inhibiting norepinephrine reuptake (Driessen et al. 1993). As an analgesic, tramadol elicits fewer serious opioid adverse reactions such as respiratory depression and drug dependence than other OADs (Cicero et al. 2005; Houmes et al. 1992; Richter et al. 1985) although it can induce a degree of physical and psychological dependence and subjects with a history of substance abuse are higher risks for dependence (Tjäderborn et al. 2007).
Tramadol overdosage and intoxications
Tramadol’s low potential for abuse, its analgesic effectiveness and widespread use have been factors in it becoming one of the main causes of drug poisoning due to overdosage, especially in young males with a history of drug abuse and mental disorders (Shadnia et al. 2008). This is reflected in an increasing number of case and post-marketing reports of tramadol-related intoxications and fatalities (Cicero et al. 2005; Clarot et al. 2003; Daubin et al. 2007; Houmes et al. 1992; McDiarmid et al. 2005; Moore et al. 1999; Raffa et al. 1992; Richter et al. 1985; Tjäderborn et al. 2007). A prospective evaluation of tramadol exposure reported to seven poison centers over a 1-year period 1995–1996 revealed 126 cases of which 87 were tramadol alone. Symptoms of overdose were, in order of frequency, lethargy, nausea, tachycardia, agitation, seizures, coma, hypertension, and respiratory depression. All seizures were brief. Tramadol 500 mg was the lowest dose associated with seizure, tachycardia, hypertension, or agitation; 800 mg was the lowest dose associated with coma and respiratory depression. The authors concluded that monoamine uptake inhibition rather than the drug’s opioid effects accounted for much of the toxicity of tramadol overdose (Spiller et al. 1997). In a review of tramadol intoxication in 114 cases, cardiopulmonary arrest was found to be the cause of death in patients who had ingested more than 5000 mg of the drug (Shadnia et al. 2008).
There are a number of reports implicating tramadol as a cause of seizures, especially in cases of overdose intoxication (Gasse et al. 2000; Jick et al. 1998; Kahn et al. 1997; Labate and Newton 2005; Mehrpour 2005; Tobias 1997). Seizures have been reported to be more frequent in patients taking tramadol only and in tramadol overdose; mydriasis or tachycardia may indicate a higher risk for seizure (Tashakori and Afshari 2010). A 3-year prospective study found tonic/clonic seizures in 31 (54.4%) of 57 patients. Generalized tonic-clonic seizures usually occurred within 24 h of drug administration and seizures were more often seen in young abusers with longer exposure to tramadol and a combined intake of alcohol (Jovanović-Čupić et al. 2006). In the study by Labate and Newton (2005), seizures were generalized tonic-clonic, none of the patients had a prior history of seizures and no recurrences occurred after ceasing tramadol therapy. Others believe that tramadol rarely provokes seizures (Jick et al. 1998) and the risk of seizures is not greater for tramadol than for other analgesics (Gasse at al. 2000). Of 71 cases of tramadol overdose (> 400 mg, median ingested dose 1000 mg), Ryan and Isbister (2015) found tachycardia in 27 patients, mild hypertension in 32, respiratory depression in 13, and seizures, which were dose related, occurred in 8 patients. There were no cases of serotonin toxicity meeting the Hunter Serotonin Toxicity Criteria (Dunkley et al. 2003). The authors concluded that tramadol overdose is associated with seizures and respiratory depression but is unlikely to cause serotonin toxicity.
Tramadol and serotonin toxicity
Tramadol can contribute to the development of serotonin toxicity by inhibiting serotonin reuptake, and this risk is increased when tramadol is used together with, for example, another reuptake inhibitor such as an SSRI, with an inhibitor of serotonin metabolism (e.g., a MAOI) or drugs that increase the synthesis or release of serotonin. Analysis of pharmacovigilance data notified to Swissmedic in the period 1998–2009 identified 22 cases of serotonin toxicity, 19 of which involved tramadol medication. One case each was ascribed to methadone, fentanyl, and codeine (Chassot et al. 2012). Apart from its use with SSRIs (Isbister et al. 2004), reports of serotonin toxicity resulting from coadministration of tramadol with other serotonergic drugs include fluoxetine (Gonzalez-Pinto et al. 2001; Kesavan and Sobala 1999; Lange-Asschenfeldt et al. 2002); citalopram (Mahlberg et al. 2004; Shakoor et al. 2014); sertraline (Mason and Blackburn 1997; Mittino et al. 2004); escitalopram (Caamano et al. 2016), and paroxetine (Egberts et al. 1997; Lantz et al. 1998). Serotonin toxicity has also been reported when tramadol was administered with the SNRI venlafaxine and mirtazapine (an antidepressant with noradrenergic and serotonergic activity) (Houlihan 2004) and when it was given with the antidepressant bupropion, oxycodone and trazodone, an antidepressant of the serotonin antagonist and reuptake inhibition (SARI) class (Falls and Gurrera 2014). Besides oxycodone, other OADs in combination with tramadol have provoked toxicity; other opioids involved include morphine (Vizcaychipi et al. 2007), hydrocodone (Takeshita and Litzinger 2009) and dextromethorphan (Kung and Ng 2007).
Tramadol alone was claimed to cause serotonin toxicity when given as an overdose or in high doses to the elderly (Pothiawala and Ponampalam 2011; Garrett 2004) and to infants (Maréchal et al. 2011). An example of the latter occurred in an 8-month-old girl after ingesting 200 mg of tramadol. Symptoms included sinus tachycardia, reactive pupils, agitation, drowsiness, increased lower limb reflexes, hyperthermia, and high blood pressure.
Tapentadol is a centrally active synthetic OAD with a dual mechanism of action approved as a potent Schedule II analgesic by the FDA in 2008 and indicated for long-term opioid treatment of moderate to severe pain and neuropathic pain associated with peripheral neuropathy when alternative treatment options are inadequate. Tapentadol is a weak opioid agonist with about 2% of morphine’s affinity for the µ receptor, but also with high selectivity for the norepinephrine transporter protein resulting in inhibition of the reuptake of norepinephrine (Guay 2009; Ramaswamy et al. 2015; Tzschentke et al. 2007, 2009). Unlike tramadol, tapentadol has a minimal effect on serotonin reuptake suggesting that the risk of serotonin toxicity should be less with tapentadol. Also unlike tramadol, tapentadol exists as a single active enantiomer [(−)-(1R,2R)], being converted to inactive metabolites mainly by O-glucuronidation (Tzschentke et al. 2007, 2009; Vadivelu et al. 2011). Adverse events reported for tapentadol include effects commonly associated with opioids, in particular and most frequently, gastrointestinal intolerance, and nausea. Other adverse effects include anxiety, headache, somnolence, dizziness, vomiting, and diarrhea, but the drug is a very weak serotonin reuptake inhibitor (Raffa et al. 2012) suggesting that it is unlikely to cause significant serotonin toxicity. Post-marketing surveillance, however, has identified occasional psychiatric disorders, adrenal, and androgen deficiencies, anaphylaxis to ingredients in Nucynta ER®, and rare cases of serotonin toxicity to both tapentadol alone and especially when it is given concomitantly with other serotonergic drugs [Nucynta® ER (Tapentadol). Full prescribing information (2016)]. Diagnosis of a possible case of serotonin toxicity following fatal intoxication by Nucynta® was based on a tapentadol blood concentration of more than 20 times the upper limit of the drug’s therapeutic range (Franco et al. 2014). Other possible mechanisms of death considered were respiratory depression and central nervous system depression. Observations to support serotonin toxicity were lacking and drug overdoses were the most likely explanation for the death. The manufacturer of tapentadol, Janssen Pharmaceuticals, has stated that some post-marketing reports have implicated the drug in serotonin toxicity when administered together with other serotonergic drugs (Walczyk et al. 2016). What was assessed as a probable case of tapentadol-associated serotonin toxicity after tapentadol overdose was recently reported in a patient who had also ingested the serotonergic agents amitriptyline and duloxetine within the past 5 weeks (Walczyk et al. 2016). The conclusion was claimed to be based on the patient fulfilling the Hunter criteria plus the presence of tremor and hyperreflexia and a positive response to benzodiazepines. Others have questioned the interpretation. Russo et al. (2017) point out that the clinical picture was more consistent with opioid overdose followed by withdrawal and tapentadol’s weak serotonin reuptake inhibitory activity makes it highly unlikely to cause serotonin toxicity; Mullins and Dribben (2017) believe that the patient clearly had an opioid toxidrome and there was faulty use of the Naranjo score and insufficient evidence for the claim of serotonin toxicity. A more convincing possible case of serotonin toxicity involving tapentadol was reported by Mancano (2013) after its concomitant use with the SNRI venlafaxine. Admitted twice to hospital in 12 days, the patient on each occasion showed symptoms of total body tremors, anxiety, and excessive sedation. At the first admission, the patient had recently been switched from oxycodone/acetaminophen to tapentadol.
So far then, there is little clear-cut evidence of an association between tapentadol and serotonin toxicity in the literature. In a systematic review of original research, Gressler et al. (2017) attempted to identify rates of serotonin toxicity associated with tapentadol use. From a total of 22 studies that met inclusion criteria (13 randomized trials, 7 open-label trials, and 2 observational studies) none of the studies reported the development of serotonin toxicity. However, none of the trials differentiated between the development of adverse events in patients taking, and those not taking, serotonergic drugs. The authors concluded that “the current tapentadol literature has important limitations that prevent the adequate characterization of the potential association between tapentadol and serotonin syndrome”.
Meperidine is an OAD used for the short-term management of pain, often by emergency clinicians (Clark et al. 1995; Latta et al. 2002). It is a potent inhibitor of serotonin reuptake and for at least the last 36 years; toxic reactions have been known to occur when meperidine is given to patients on MAOI or tricyclic antidepressant therapy (Meyer and Halfin 1981; Zornberg et al. 1991). In fact, as early as 1962 what was called an “alarming” reaction to pethidine (meperidine) in patients on the MAOI, phenelzine, was reported in the Lancet (Taylor 1962). Later experiments in mice showed that elevated levels of serotonin generated by the combination of meperidine and a MAOI provoked a serious toxic reaction that correlated with a critical concentration of serotonin ~ 60% above normal levels (Rogers and Thornton 1969). Then, in 1973, p-chlorophenylalanine, an inhibitor of serotonin synthesis, was shown to prevent the toxic interaction between meperidine and phenelzine in rabbits (Mattila and Jounela 1973). The finding in a retrospective study in 1999 that 26 of 262 patients (10%) receiving meperidine in an emergency department were also taking one or more serotonergic drugs, led to the criticism that the routine use of meperidine in emergency departments places at risk a number of these patients taking other serotonergic drugs (Weiner 1999). In what appears to be two first time demonstrations, induction of serotonin toxicity by meperidine alone and an effective emergency treatment by the histamine H2 receptor antagonist famotidine have recently been reported (Joe et al. 2017). In another unusual case, probable meperidine-induced serotonin toxicity was diagnosed in a patient with a history of fluoxetine who was taking the insulin sensitizer rosiglitazone and the cholesterol lowering drug fenofibrate at the time but who had not taken fluoxetine for two weeks prior to the adverse reaction (Tissot 2003). Of possible relevance to this finding is the relatively long half-lives of SSRIs, meaning that patients may remain at risk for a significantly long period after therapy is discontinued. The 2-week half-life of norfluoxetine, the active metabolite of fluoxetine (Coplan and Gorman 1993), may account for the above finding which, in any case, serves as a warning of possible unexpected extended risk periods for some patients. In a more extreme example of the possible risk of serotonin toxicity following extended previous contact with a serotonergic agent, signs and symptoms of the toxicity began to appear in an adult male within 1 min of receiving 30 mg IV of meperidine. Five years earlier, the patient experienced a life-threatening episode of serotonin toxicity after taking the TCA clomipramine (Guo et al. 2009). The SSRI citalopram has also been associated with serotonin toxicity in a patient taking meperidine (Altman and Manos 2007).
The synthetic opioid methadone which has a diphenylpropylamine structure like morphine and meperidine, is a racemic mixture of levomethadone (l-methadone; R-(−)-methadone) and dextromethadone (d-methadone; S-(+)-methadone). The R-enantiomer, levomethadone, is the active form of the opioid with µ-opioid receptor selectivity 50 times the opioid potency of the S-enantiomer and twice the potency of methadone by weight (Verthein et al. 2005). Levomethadone is also a weak competitive antagonist of the glutamatergic N-methyl-D-aspartate (NMDA) receptor complex and a potent noncompetitive antagonist of the α3β4 nicotinic acetylcholine (nACh) receptor. The S-enantiomer dextromethadone has virtually no opioid activity at normal therapeutic doses, but is a NMDA receptor antagonist. Levomethadone, used itself as a drug in some European countries, has antitussive effects but is mainly given for pain management and opioid maintenance therapy. Dextromethadone also has antitussive properties. Methadone is slowly metabolized by enzymes CYP3A4, CYP2B6 and CYP2D6 and has an elimination half-life of 15–60 h (mean ~ 22 h), although metabolism rates can vary greatly between individuals.
Methadone therapy has been shown to be associated with serotonin toxicity when given in high dosage with the SNRI duloxetine and the TCA desipramine (Rastogi et al. 2011); as a 200 mg overdose with the SSRI and SNRI antidepressants sertraline and venlafaxine (Martinez and Martinez 2008); with sertraline and the antidepressant mirtazapine (Martin-Lazaro et al. 2017); and in a case described as a first report associated with fentanyl and without coadministration of other serotonergic agents (Hillman et al. 2015), although this claim has been called into question (Atkinson and Fudin 2015). An interaction between methadone and the quinolone antibacterial ciprofloxacin may also contribute to serotonin toxicity as well as to some other adverse reactions during methadone therapy. Development of the toxicity in a chronic pain patient receiving concurrent methadone, ciprofloxacin, and venlafaxine (Lee et al. 2009) might be explained by involvement of the SNRI venlafaxine but other cases involving the opioid and the quinolone suggest a possible role for methadone metabolism. An interaction between methadone and ciprofloxacin that caused severe life-threatening respiratory depression might have been caused by an accumulation and toxicity of the opioid due to inhibition of methadone metabolism by ciprofloxacin, a potent inhibitor of the cytochrome CYP450 isoenzyme system (Samoy and Shalansky 2010). Two other published reports detail a reaction between methadone and ciprofloxacin, one resulting, again, in respiratory depression (Herrlin et al. 2000) and the other resulting in torsades de pointes (Nair et al. 2008). Methadone was associated with serotonin toxicity 5 times in the FDA Adverse Event Reporting System (FAERS) database for the period 1 January, 1969 to 12 June, 2013 (FDA Drug safety communications 2016).
Oxycodone (Kalso 2005; Pöyhiä et al. 1993) is a semisynthetic opioid derived from the phenanthrene alkaloid thebaine present in the opium poppy. Like other important OADs, oxycodone contains important groups for good analgesic action at positions C3 (a methoxy group), C6 (a keto group) and attached to the N at position 17 (a methyl group) of the phenanthrene nucleus. Oxycodone is a selective agonist of the µ opioid receptor and has lower affinity for the δ and κ receptors but the binding affinity for the µ receptor is less than that of morphine and methadone (Chen et al. 1991). An important metabolic pathway of oxycodone is O-demethylation via CYP2D6 (Zwisler et al. 2009) to form the active metabolite oxymorphone which has higher binding affinity for the µ receptor than oxycodone (Davis et al. 2003).
There is a small number of claimed cases of serotonin toxicity involving oxycodone and the opioid was associated with the condition seven times in the FAERS database for the period 1 Jan 1969—12 June 2013. This included four reports of serotonin toxicity occurring when oxycodone was used with fentanyl (FDA Drug safety communications 2016). In a published case report, a male adult taking a dramatically increased dose of oxycodone and a normal dose of sertraline experienced visual hallucinations and severe tremor. This was assessed as a case of probable serotonin toxicity after symptom resolution when sertraline was discontinued (Rosebraugh et al. 2001). Diagnosis of serotonin toxicity in a 70-year-old woman following the addition of 40 mg of oxycodone twice daily to her daily dose of the SSRI fluvoxamine (Karunatilake and Buckley 2006) and symptoms of the toxicity in another elderly patient taking citalopram resolved within 48 h after substituting morphine for oxycodone (Walter et al. 2012). In another report of serotonin toxicity in a patient given oxycodone, the drugs responsible were not identified with certainty and the possible mechanism involved remains ill-defined. Postoperative pain was controlled using patient-controlled analgesia (fentanyl, ketorolac, ramosetron) followed by daily pregabalin, oxycodone, and celecoxib (Song 2013).
The short-acting synthetic opioid fentanyl, perhaps the most medically widely used opioid, is a potent agonist of the µ opioid receptor with a binding affinity ~ 50 to 100 times that of morphine. The drug’s high lipophilicity allows it to penetrate the CNS and it is often administered as a transdermal patch for long-term pain management.
The difficulty of diagnosing some cases of serotonin toxicity are well illustrated in a case of polypharmacy in a patient with a transdermal fentanyl patch taking oral oxycodone/acetaminophen, celecoxib, citalopram, mirtazapine, doxazosin, and zolpidem and later, haloperidol and lorazepam. After missed diagnosis and misdiagnosis, it was eventually concluded that an increase in the frequency of replacement of the fentanyl patch was the precipitating factor in the onset of serotonin toxicity. Discontinuation of the fentanyl patch, oxycodone/acetaminophen, celecoxib, citalopram and mirtazapine resulted in complete resolution of symptoms within 24 h (Rastogi et al. 2011). There are reports of cases of serotonin toxicity associated with fentanyl administration during esophagogastroduodenoscopy (Alkhatib et al. 2010); during procedural sedation of patients taking SSRIs (Kirschner and Donovan 2010) including citalopram (Allawadhi et al. 2007) and paroxetine (Rang et al. 2008); when used together with ondansetron, paroxetine, the SNRI duloxetine and the antidepressant bupropion (Gollapudy et al. 2012); when administered intrathecally in a patient taking multiple drugs including ephedrine, ergot alkaloids, marijuana, and methylamphetamine (Ozkardesler et al. 2008); and together with methadone in a burn injury patient (Hillman et al. 2015). Following administration of fentanyl, hydromorphone, and ondansetron during the perioperative period, diagnosis of serotonin toxicity on the basis of the Hunter criteria was made for a patient receiving multiple psychiatric medications including the SSRI duloxetine, lithium, and the serotonin receptor blocker quetiapine (Altman and Jahangiri 2010). Fentanyl administered by nasal spray was also implicated in a case of serotonin toxicity when given together with oxycodone and escitalopram (Reich and Lefebvre-Kuntz 2010). All of these cases illustrate the difficulty of identifying the drug(s) responsible for provoking the observed toxicities. Fentanyl, with 28 cases, was the most commonly recorded opioid associated with serotonin toxicity in the FAERS database for the period 1 Jan 1969–12 June 2013. This included five cases of serotonin toxicity when fentanyl was used together with at least one other opioid. Oxycodone was used in four cases while morphine, hydromorphone, and hydrocodone were each used in one case (FDA Drug safety communications 2016).
The morphinan, dextromethorphan, is the dextrorotatory enantiomer of the OAD levomethorphan. Dextromethorphan is a noncompetitive NMDA receptor antagonist with low receptor affinity. It also shows very low affinity for the µ, δ and κ opioid receptors (Codd et al. 1995) even when abused at excessive doses. Dextromethorphan acts centrally, raising the coughing threshold and is therefore used as an antitussive, often in the form of over-the-counter remedies. It is converted to the tenfold more potent active metabolite dextrorphan by the cytochrome P450 enzyme CYP2D6 and the drug’s therapeutic activity is due to this metabolite as well as the parent compound. Dextrorphan is further metabolized to 3-methoxymorphinan and on to 3-hydroxymorphinan (Kerry et al. 1994).
Dextromethorphan-related problems include abuse, attempted suicide, and poisonings with moderate-to-severe toxicities including serotonin toxicity resulting from drug interactions (Chyka et al. 2007). In some cases when used at excessively high doses as a recreational drug, dextromethorphan and its active metabolite dextrorphan exhibit sufficient activity at the NMDA receptor to induce dissociative hallucinogenic states. In one published report, visual hallucinations were precipitated by the combination of dextromethorphan and fluoxetine (Achamallah 1992). Well before the description of serotonin toxicity in 1991, a possible lethal reaction between dextromethorphan and the MAOI phenelzine was suggested as a “toxic reaction” causing the death of a young woman following the ingestion of approximately two ounces of a dextromethorphan cough preparation 6 h after a double normal dose of phenelzine (Rivers 1970). A case of serotonin toxicity was the presumptive diagnosis in a patient taking the so-called ‘night and day’ capsules containing dextromethorphan for 2 or 3 days for a head cold. The patient’s usual medications were methadone, citalopram and gabapentin and although methadone and citalopram had been taken for 2 years without ill-effect, a full recovery resulted after its withdrawal (Cameron 2006). There are at least three other reports of serotonin toxicity in patients taking an over-the-counter source of dextromethorphan, the first in the form of a cold medicine in association with paroxetine (Skop et al. 1994). The second report concerned an 18-year-old male after recreational use of Coricidin HBP (chlorpheniramine 4 mg, dextromethorphan 30 mg) who had a dextromethorphan serum concentration of 930 ng/ml. Propofol infusion rapidly reversed the patient’s agitation, neuromuscular hyperactivity and autonomic instability. The authors concluded that dextromethorphan alone can produce serotonin toxicity in the absence of another drug (Ganetsky et al. 2007). In the third report (Schwartz et al. 2008), supratherapeutic doses of dextromethorphan were claimed to be the cause of two cases of serotonin toxicity, the authors arguing that therapeutic doses are not enough to provoke reactions. In the first of these two cases, a 20 year-old man on aripiprazole, benztropine, and escitalopram had a dextromethorphan serum concentration of 950 ng/ml after overdosing on a dextromethorphan-containing cough medicine. GC-MS urine drug tests also revealed the presence of chlorpheniramine. In the second case, a 6 year-old boy on sertraline had a dextromethorphan serum concentration of 2820 ng/ml. In comparison, a 20 mg dose of dextromethorphan produces a peak level of ~ 1.8 ng/ml in 2.5 h. It was concluded that because of the widespread use of dextromethorphan cough medications and SSRIs, serotonin toxicity due to this combination therapy should be much more common and for the toxicity to occur with SSRI therapy, supratherapeutic doses are necessary. The fact that chlorpheniramine was being taken in two of the cases reviewed briefly above (Ganetsky et al. 2007; Schwartz et al. 2008), raises the question of its possible contribution to the development of serotonin toxicity in the relevant patients.
Attention has been drawn to the proserotonergic effects of chlorpheniramine and its strong serotonin uptake inhibitor effect (Karamanakos and Panteli 2008; Monte et al. 2010), factors not taken into account in the Skop et al. (1994) and Schwartz et al. (2008) reports.
In a report from Japan, serotonin toxicity was diagnosed in a patient 4 days after commencement of dextromethorphan tablets at a conventional dosage of 15 mg three times a day to relieve cough. Symptoms were restlessness, tremulous voice, tremor of the head, neck, shoulders, trunk, and extremities, jerky involuntary movements that worsened with movement and negative clonus. The patient reported the same symptoms after taking dextromethorphan 5 years earlier. There appears to be at least one other report of serotonin toxicity to dextromethorphan in Japan, thought to be from an overdose (Kinoshita et al. 2011).
Apart from fentanyl, oxycodone, and methadone, in the FAERS database for the period January 1, 1969–June 12, 2013, there were reports of serotonin toxicity associated with alfentanil, remifentanil, sufentanil, hydrocodone, hydromorphone, morphine, naltrexone, and pentazocine, but no reports of toxicity with an opioid used alone (FDA Drug safety communications 2016).
With the symptoms combination of agitation, diaphoresis, tremor, and hyperreflexia, serotonin toxicity was diagnosed in an elderly patient given venlafaxine, codeine, acetaminophen, and diazepam for migraine and taking, without medical supervision, rizatriptan. 30–36 h after the first codeine dose, reported symptoms were nervousness, irritability, agitation, mania, confusion, tremor, diaphoresis, and nausea. After considering opioid intoxication and withdrawal to be unlikely, it was concluded that the reaction was triggered by the codeine metabolite morphine, generated by a high CYP2D6 activity in the patient (Milano et al. 2017).
A single dose of buprenorphine as Suboxone®, a combination of the OAD with the opioid antagonist naloxone used for opiate dependence, triggered serotonin toxicity in a patient on tricyclic antidepressants. Symptoms were not controlled by benzodiazepines but improvement resulted after administration of the serotonin receptor antagonist, cyproheptadine (Isenberg et al. 2008).
To minimize the possibility of adverse reactions such as relapse of depressive illness and perioperative hypotension in a surgery patient receiving the MAOI phenelzine, remifentanil was selected for intraoperative use because of its short elimination half-life of 5–8 min. In light of the report of a postoperative death after high-dose fentanyl in a patient treated with an MAOI, the absence of data suggesting that remifentanil interacts adversely with MAOIs was an additional reason for its selection. Using remifentanil, no clinical signs indicating adverse interaction with the MAOI were seen (Ure et al. 2000).
Serotonin receptors, animal models, and effects of different opioids on the regulation of serotonergic neurotransmission
Whereas there is a dearth of research on the pathophysiology of serotonin toxicity in humans, there is an extensive animal literature on the effects of serotonin in the CNS (for reviews see Haberzettl et al. 2013; Isbister and Buckley 2005; Sternbach 1991), but the relevance and value of much of this research to human studies is questionable. Responses to serotonin excess in animals (in particular rodents), termed serotonin behavioral syndrome, are, in general, different from responses seen in human serotonin toxicity and these dissimilarities, until recently, did not generally contribute significantly to the accumulation of clinical and mechanistic insights into the human condition (Isbister and Buckley 2005). Acute administration of serotonergic drugs to rats and mice induces a set of distinctive behavioral and autonomic responses, the major ones being forepaw treading, hindlimb abduction, head weaving, head twitching, back muscle contraction, and hyperthermia (Haberzettl et al. 2013; Isbister and Buckley 2005; Kalueff et al. 2007; Sternbach 1991). Despite the differences between the responses to serotonin in animals and humans, an extensive list of animal model studies have been pursued over many years in attempts to better understand, prevent, and advance the diagnosis and treatment of serotonin toxicity patients. Some insights derived from animal model have resulted from studies of 5-HT receptors and the 5-HT transporter (serotonin transporter; SERT; 5-HTT), both components of the 5-HT system. 5-HT receptors, made up of 7 families with 14 known subtypes, occur in the central, enteric and peripheral nervous systems. All, except receptor 5-HT3, are G-protein-coupled receptors (Hoyer et al. 2002; Nichols and Nichols 2008). The diverse adverse drug effects induced in rodents models and human patients indicate that no one receptor mediates serotonin toxicity. The most prominent effects in the rodent serotonin behavioral syndrome have been shown to be mediated by postsynaptic 5-HT1A receptors (Darmani and Ahmad 1999; Forster et al. 1995; Goodwin and Green 1985; Smith and Peroutka 1986; Tricklebank et al. 1984). 5-HT2A receptor agonists can potentiate the syndrome but cannot induce it (Arnt and Hytell 1989; Eison and Wright 1992; Isbister and Buckley 2005). One feature, however, head twitching in rats and mice, has been shown to be mediated by the 5-HT2 receptor; in mice, it is the 5-HT2A receptor subtype that is involved (Koshikawa et al. 1985; Schreiber et al. 1995; Yap and Taylor 1983). In a rat model of the toxicity induced by the MAOI antidepressant clorgyline and 5-hydroxy-L-tryptophan, pretreatment with potent 5-HT2A receptor antagonists prevented hyperthermia and death of the animals suggesting that such antagonists might be effective drugs for treatment of the toxicity in humans (Nisijima et al. 2001).
For obvious reasons, experimental evaluation of toxic responses to serotonergic and serotonin enhancing drugs, and assessment of the risk potential of individual drugs and drug combinations, can only be undertaken with the greatest of caution in humans. Human studies with specific 5-HT receptor agonists and antagonists are therefore few. In particular, there appears to be no data on the administration of 5-HT2A receptor agonists to human subjects although one controlled clinical pharmacology human study of 5-HT1A receptor and 5-HT transporter blockade has been reported. Designed to evaluate the safety of the 5-HT1A antagonist lecozotan and the SSRI citalopram in a clinically controlled setting using two double-blind treatment groups treated with citalopram/lecozotan or citalopram/placebo, the study demonstrated the utility of the HSTC in detecting the symptoms of serotonin toxicity (Parks et al. 2012).
The SERT, found in neurons, platelets and enterochromaffin cells, is involved in the control of 5-HT plasma levels and reuptake of the neurotransmitter into presynaptic nerve terminals. SSRIs block the SERT protein resulting in an increase in concentration of the transmitter in the synaptic cleft (Schloss and Williams 1998). Barann et al. (2006) speculated that SERT inhibitors such as some opioids may elevate 5-HT plasma and synaptic concentrations that in turn activate 5-HT receptors, a mechanism that may explain how opioids such as tramadol, which has a low affinity for opioid receptors and lacks a direct effect on 5-HT receptors, may induce opioid-induced emesis. A number of studies have demonstrated interaction of some opioids with the SERT. In most, rat brain synaptosomes were used (Codd et al. 1995; Driessen et al. 1993; Giusti et al. 1997; Larsen and Hyttel 1985; Tzschentke et al. 2007) but a few investigations with opioids employed human SERT. Barann et al. (2006), using HEK-293 cells stably transfected with human SERT or 5-HT3A receptor DNA, demonstrated potent concentration-dependent inhibition of SERT by (+) and (−) tramadol at concentrations corresponding to tramadol bolus plasma concentrations (~ 10 µM); the corresponding enantiomers of the tramadol active metabolite, O-demethyltramadol, were weakly inhibitory. At substantially higher concentrations, both enantiomers of tramadol and O-demethyltramadol were only very slightly inhibitory of receptor 5-HT3A. Since OADs, including tramadol with its known opioid µ receptor affinity, are known to induce nausea and emesis, particularly after rapid titration at high doses (Ruoff 1999), and the fact that human 5-HT3A receptors are known to mediate emesis, the authors suggested that the findings were compatible with their speculated mechanism for tramadol-induced emesis. This conclusion fits with successful treatment of tramadol-induced emesis by IV metoclopramide, an inhibitor of receptor 5-HT3 (Walkembach et al. 2005).
SERT knockout (−/−) mice (Bengel et al. 1998) display some symptoms of toxicity to serotonin (Kalueff et al. 2007). These mice have fewer 5-HT1A receptors with decreased function and diminished 5-HT2 receptor numbers and function. In a study of receptor-mediated responses to serotonergic drugs in SERT-deficient mice, Fox et al. (2007) compared drug-induced behavioral responses in SERT knockout, wildtype (+/+), and heterozygous (+/−) animals using the serotonin precursor 5-HTP, MAOIs and 5-HT1A, 5-HT2A, and 5-HT7 antagonists before administration of 5-HTP. Pretreatment with the specific inhibitors for receptor 5-HT1A significantly reduced the 5-HTP-induced enhanced abnormal behavior in SERT (−/−) mice but results with inhibitors for 5-HT2A, and 5-HT7 did not produce this effect leading to the conclusion that the responses in SERT (−/−) and (+/−) mice were mediated by postsynaptic 5-HT1A receptors. The results also indicated that 5-HT7 receptors may not be involved in serotonin toxicity in mice. In a similar but smaller study using 5-HT receptor agonists in 5-HTP-dosed NMRI mice, Haberzettl et al. (2014) found a range of behavioral and autonomic responses after treatment with 5-HT1A and 5-HT2A agonists. Some responses, for example Straub tail, appeared to be associated with postsynaptic 5-HT1A receptor activation but the 5-HT2A receptor appeared to have more effect on the mouse model of serotonin toxicity than expected.
As well as SERT knockout mice, mutations in MAO genes in mice have also been investigated as another possible rodent model for the study of serotonin toxicity. Fox et al. (2013) examined what they described (by the misnomer) as “hypersensitivity” to 5-HTP and tramadol in MAOA/B knockout mice, that is mice lacking genes for both MAOs. The mice displayed baseline serotonin toxicity effects and serotonin levels that were increased compared to wild type mice. These levels were further accentuated following administration of 5-HTP or tramadol.
Pointing out that the capacity of opioid serotonergic drugs such as tramadol to induce serotonin toxicity in rodents had not been evaluated experimentally, Fox et al. (2009) examined this drug together with meperidine and morphine in SERT knockout (−/−), wildtype (+/+), and heterozygous mice. The two synthetic OADs but not morphine elicited serotonin toxicity-like behavior in mice, particularly in the mutant animals lacking one or two copies of the SERT gene. The observed response to tramadol in the SERT (−/−) mice was blocked by pretreatment with the 5-HT1A receptor antagonist WAY 100,635. Furthermore, analgesia induced by all three OADs was markedly reduced in the knockout mice.
Following the demonstration that morphine, and to a lesser extent hydromorphone, interact with the 5-HT3 receptor (Wittmann et al. 2008), and early suggestions that some OADs may be weak serotonin reuptake inhibitors (Gillman 2005), Barann et al. (2015) extended the study of the effects of OADs on human SERT, examining, in addition to tramadol, morphine, hydromorphone, fentanyl, alfentanil, and meperidine. Tramadol and meperidine inhibited the human 5-HT transporter in the transfected HEK-293 cells and platelets but the other OADs examined did not.
Extending the work of Barann et al. (2015), and with the aim of investigating possible links to human serotonin toxicity, Rickli et al. (2018) determined the in vitro potencies of different OADs to inhibit the human SERT and norepinephrine transporter (NET) using transporter transfected HEK-293 cells. With results of previous serotonin toxicity studies involving animal models in mind, binding of opioids to the 5-HT receptors, 5-HT1A, 5-HT2A, and 5-HT2C were also examined. Consistent with the findings of Barann et al. (2015), the synthetic opioids dextromethorphan and l-methadone potently inhibited the SERT and tramadol, meperidine, tapentadol, and d-methadone inhibited at low molar concentrations whereas fentanyl and the phenanthrenes morphine, hydromorphone, oxymorphone, codeine, hydrocodone, dihydrocodeine, oxycodone, and buprenorphine were inactive. Dextromethorphan, l-methadone, tramadol, meperidine, tapentadol, and d-methadone inhibited human SERT at concentrations at, or close to, clinical drug plasma and brain concentrations. Whereas dextromethorphan showed greater SERT than NET inhibitory activity, this was reversed for tramadol and tapentadol (Rickli et al. 2018). Earlier studies with rat brain synaptosomes had shown that tramadol, meperidine, tapentadol, methadone, and dextromethorphan blocked rat SERT and NET in vitro, while morphine and codeine did not (Codd et al. 1995; Driessen et al. 1993; Giusti et al. 1997; Larsen and Hyttel 1985; Tzschentke et al. 2007). Receptor radioligand binding assays revealed direct binding of fentanyl to receptors 5-HT1A and 5-HT2A at concentrations higher than occur in human plasma and binding of meperidine and methadone to the 5-HT2A receptor at approximate therapeutic drug plasma concentrations (Rickli et al. 2018).
Our understanding of the detailed mechanisms underlying the involvement of opioids in serotonin toxicity is obviously still incomplete but there is some evidence of weak serotonin reuptake inhibition and, using opioids in conjunction with antidepressants, synergistic µ-opioid and 5-HT1A presynaptic inhibition of GABA release in rat periaqueductal gray neurons (Kishimoto et al. 2001). Preliminary evidence for the latter is based on inhibition of GABAergic miniature inhibitory postsynaptic currents in rat periaqueductal gray neurons by a specific µ-opioid receptor agonist and serotonin in a dose-dependent manner. The presynaptic opioid effect was blocked by a specific µ-opioid receptor antagonist and the presynaptic serotonergic effect was mimicked by a specific 5-HT1A receptor agonist and blocked by a specific antagonist. Both the opioid- and serotonergic-induced inhibitions of GABA release appear to proceed by the opening of 4-amidopyridine-sensitive potassium channels via G-proteins. Synergistic inhibition of GABA release by activated presynaptic µ-opioid and 5-HT1A receptors suggest still poorly defined cellular mechanisms for analgesic effectiveness within the periaqueductal gray whereby synaptic serotonin levels are increased when OADs are used in conjunction with antidepressants that increase synaptic serotonin levels (Kishimoto et al. 2001).
The recently issued FDA Drug Safety Communication warning (2016) of the possibility of potentially harmful interactions between “the entire class of opioid pain medicines” with numerous other medications, particularly antidepressants and migraine medicines, acknowledges the increasingly recognized possibility of the induction of serotonin toxicity in some opioid-treated patients. The increasing use of OADs in hospitals to produce procedural sedation and to treat acute, chronic, and postsurgical pain, as well as their wide usage in pain management outside the hospital setting, creates the opportunity for adverse drug interactions in some patients who are also taking, or have taken, other serotonergic medications. In particular, the often prescribed psychiatric medicines and other serotonergic agents increase the possibility of serotonin toxicity in the perioperative setting where fentanyl is widely used, in emergency departments where OADs such as meperidine and methadone may be routinely given, and in non-hospital settings where drugs such as antidepressants and MAOIs may be coadministered with OADs to an increasingly large number of patients including those with drug abuse, drug dependence and/or a psychiatric history. It has been suggested that the induction of serotonin toxicity correlates with the increasing use of psychiatric medicines in all types of patient populations (Gayle et al. 2015). Adding to the risks, are the long half-lives of some serotonergic agents, for example SSRIs, and the extended risk period after cessation of therapy before subsequent exposure to an opioid (Coplan and Gorman 1993; Guo et al. 2009; Tissot 2003).
Isbister and Buckley (2005) pointed out that the few studies so far carried out in humans, as well as many in animals, do not contradict a role for the 5-HT2A receptor in mediating hypothermic and muscle rigidity responses in humans. As discussed, some results with 5-HT2A receptor antagonists produce varying effects on individual behaviors in rodents (for example, head twitching and hyperthermia), pointing to a role for the 5-HT2A receptor in the toxicity. Other, somewhat less compelling, evidence of such a role in humans is the demonstration that serotonin toxicity can be reversed by high doses of the non-specific 5-HT2 receptor antagonists cyproheptadine and chlorpromazine (Chan et al. 1998; Gillman 1999). Of course, the 5-HT1A and 5-HT2A receptors should not be thought of as always acting alone; functional relationships between the different receptors, and even between other neurotransmitters, may exist (Isbister and Buckley 2005). An example of the former is seen in 5-HT1A-mediated rat forepaw treading potentiated by 5-HT2A receptor activation (Arnt and Hyttel 1989). The two different receptor types also have different affinities with the 5-HT1A receptor being the higher. Hence, these receptors are likely to be occupied and functionally dominant at lower serotonin concentrations whereas 5-HT2A receptor effects may not be apparent until higher serotonin concentrations are reached. An example of this is seen in the prevention of hyperthermia by 5-HT2A antagonists in an animal model (Nisijima et al. 2004). Mechanistically then, and taking into account the uncertainty of relevance of animal models, recognition of postsynaptic 5-HT1A (to at least some extent) and 5-HT2A receptors, and synergistic µ-opioid and 5-HT1A receptor presynaptic inhibition of GABA release, may be involved in opioid-induced serotonin toxicity in humans (Boyer and Shannon 2005; Haberzettl et al. 2013; Isbister and Buckley 2005; Kishimoto et al. 2001; Mills 1997; Nisijima et al. 2001).
Total avoidance of serotonin toxicity in patients will always be difficult but increased awareness of patients and physicians of the potential for toxicity, especially when a number of different drugs are being taken and progress in risk assessment would obviously be of benefit. The most important necessary measure is the education of physicians in anticipating the possibility of the toxicity and being well versed in its diagnosis. To achieve this and increase awareness of serotonin toxicity, post-marketing surveillance linked to physician education has been proposed. In a 1999 post-marketing surveillance survey of reactions to the antidepressant nefazodone prescribed in general practice, 19 cases from a cohort of 11,834 patients (an incidence of 0.4 cases/1000 patient months) met Sternbach’s serotonin syndrome diagnostic criteria. In this study, 85% of responding general practitioners reported that they were unaware of serotonin syndrome (Mackay et al. 1999). More recent assessments do not appear to be available but some physicians may still not be fully aware of the presentation and management of the toxicity, especially the less severe form. Since serotonin syndrome is a clinical diagnosis, heightened clinician awareness of the possibility of serotonin toxicity among patients taking opioids and serotonergic antidepressants has been called for (Gnanadesigan et al. 2005). For the prevention of serotonin toxicity, careful prescribing and diagnosis by informed clinicians and avoidance of multidrug regimens if possible are the most obvious and surest measures to take. For the future, increased understanding of underlying mechanisms and progress in risk assessment for serotonin toxicity are important aims.
No sources of funding were used to assist in the preparation of this study.
Compliance with ethical standards
Conflict of interest
The author, Brian A. Baldo, declares that he has no conflict of interest.
- Allawadhi S, Sung KW, Carlson LA et al (2007) Serotonin syndrome caused by interaction between citalopram and fentanyl. J Clin Pharm Ther 32:199–202Google Scholar
- Baldo BA (2017) Editorial. Opioid analgesic drugs: Misuse, toxicity, and hypersensitivity. J Allergy Clin Immunol 5:1607–1608Google Scholar
- Baldo BA, Pham NH (2013) Drug allergy: clinical aspects, diagnosis, mechanisms, structure-activity relationships. Springer, New York, pp 303–308Google Scholar
- Barann M, Stamer UM, Lyutenska M et al (2015) Effects of opioids on human serotonin transporters. Naunyn Schmiedeberg’s Arch Pharmacol 388:43–49Google Scholar
- Boyer EW, Shannon M (2005) The serotonin syndrome. N Eng J Med 352:1112–1120Google Scholar
- Caamano A, Din R, Eter A (2016) Serotonin syndrome induced by combined use of tramadol and escitalopram: a case report. J Med Cases 7:554–557Google Scholar
- Chyka PA, Erdman AR, Manoguerra AS et al (2007) Dextromethorphan poisoning: an evidence-based consensus guideline for out-of-hospital management. Clin Toxicol 45:662–677Google Scholar
- Daubin C, Quentin C, Goullé J-P et al (2007) Refractory shock and asystole related to tramadol overdose. Clin Toxicol 45:961–964Google Scholar
- FDA Drug safety communications (2016) FDA warns about several safety issues with opioid pain medicines; requires label changes. March 22, 2016. https://www.fda.gov/downloads/Drugs/DrugSafety/UCM491302.pdf. Accessed 30 January 2018
- Fox MA, Jensen CL, Gallagher PS et al (2007) Receptor mediation of exaggerated responses to serotonin-enhancing drugs in serotonin transporter (SERT)-deficient mice. Neuropharmacol 53:643–656Google Scholar
- Fox MA, Jensen CL, Murphy DL (2009) Tramadol and another atypical opioid meperidine have exaggerated serotonin syndrome behavioral effects, but decreased analgesic effects, in genetically-deficient serotonin transporter (SERT) mice. Int J Neuropsychopharmacol 12:1055–1065PubMedPubMedCentralGoogle Scholar
- Freye E (2008) Opioids in medicine. Springer, DordrechtGoogle Scholar
- Gayle JA, Volpi-Abadie J, Kaye AM et al (2015) Serotonin syndrome. In: Kaye AD, Kaye AM, Urman RD (eds) Essentials of pharmacology for anesthesia, pain medicine, and critical care. Springer, New York, pp 797–807Google Scholar
- Gillman PK, Whyte IM (2004) Serotonin syndrome. In: Haddad P, Dursun S, Deakin B (eds) Adverse syndromes and pschiatric drugs. Oxford University Press, Oxford, pp 37–49Google Scholar
- Insel TR, Roy BF, Cohen RM et al (1982) Possible development of the serotonin syndrome in man. Am J Psychol 139:954–955Google Scholar
- Isenberg D, Wong SC, Curtis JA (2008) Serotonin syndrome triggered by a single dose of suboxone. Am J Emerg Med 26:840.e3–840.e5Google Scholar
- Jovanović-Čupić V, Martinović Ž, Nešić N (2006) Seizures associated with intoxication and abuse of tramadol. Clin Toxicol 44:143–146Google Scholar
- Karamanakos PN, Panteli ES (2008) Comment on: “Dextromethorphan-induced serotonin syndrome”. Clin Toxicol 46:1101Google Scholar
- Kinoshita H, Ohkubo T, Yasuda M et al (2011) Serotonin syndrome induced by dextromethorphan (Medicon) administered at the conventional dose. Geniatr Gerontol Int 11:121–122Google Scholar
- Kishimoto K, Koyama S, Akaike N (2001) Synergistic µ-opioid and 5-HT1A presynaptic inhibition of GABA release in rat periaqueductal gray neurons. Neuropharmacol 5:529–538Google Scholar
- Kung SW, Ng MH (2007) Serotonin syndrome with tramadol and dextromethorphan. Hong Kong J Emerg Med 14:48–52Google Scholar
- Larsen JJ, Hyttel J (1985) 5-HT-uptake inhibition potentiates antinociception induced by morphine, pethidine, methadone and ketobemidone in rats. Acta Pharmacol Toxicol (Copenh) 57:214–218Google Scholar
- Levy JH (1986) Anaphylactic reactions in anesthesia and intensive care. Butterworths, BostonGoogle Scholar
- Mancano MA (2013) ISMP adverse drug reactions. Serotonin syndrome with concomitant use of tapentadol and venlafaxine. Hosp Pharm 48:542–549Google Scholar
- Martin-Lazaro JF, Hayde-West J, Chatzimichael S et al (2017) A dangerous triad: Sertraline, mirtazapine and methadone. Clin Med Rev Case Rep 4:154–155Google Scholar
- Mason PJ, Morris VA, Balcezak TJ (2000) Serotonin syndrome. Presentation of two cases and review of the literature. Medicine (Baltimore) 79:201–209Google Scholar
- Mehrpour M (2005) Intravenous tramadol-induced seizure: two case reports. Iran J Pharmacol Ther 42:146–147Google Scholar
- Morphine sulfate injection. Highlights of prescribing information (2016) Adverse reactions. https://www.accessdata.fda.gov/drugsatfda_docs/label/2016/204223s006lbl.pdf.Accessed 2 March 2018
- Nasmyth PA, Stewart HC (1950) The release of histamine by opium alkaloids. J Physiol 111:19PGoogle Scholar
- Nucynta® ER (Tapentadol). Highlights of prescribing information (2016) Warnings and precautions. Adverse reactions. https://www.accessdata.fda.gov/drugsatfda_docs/label/2016/200533s014lbl.pdf. Accessed 2 Mar 2018
- Orlova Y, Rizzoli P, Loder E (2018) Association of coprescription of triptan antimigraine drugs and selective serotonin reuptake inhibitors or selective norepinephrine reuptake inhibitor antidepressants with serotonin syndrome. JAMA Neurol. https://doi.org/10.1001/jamaneurol.2017.5144 PubMedGoogle Scholar
- Pothiawala S, Ponampalam R (2011) Tramadol overdose: a case report. Proc Singap Healthc 20:219–223Google Scholar
- Rogers KJ, Thornton JA (1969) The interaction between monoamine oxidase inhibitors and narcotic analgesics in mice. Br J Pharm 36:470–480Google Scholar
- Ryan NM, Isbister GK (2015) Tramadol overdose causes seizures and respiratory depression but serotonin toxicity appears unlikely. Clin Toxicol 53:545–550Google Scholar
- Schreiber R, Brocco M, Audinot V et al (1995) (1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane)-induced head-twitches in the rat are mediated by 5-hydroxytryptamine (5-HT) 2A receptors: modulation by novel 5-HT2A/2C antagonists, D1 antagonists and 5-HT1A agonists. J Pharmacol Exp Ther 273:101–112PubMedGoogle Scholar
- Schwartz AR, Pizon AF, Brooks DE (2008) Dextromethorphan-induced serotonin syndrome. Clin Toxicol 46:771–773Google Scholar
- Soldin OP, Tonning JM (2008) Serotonin syndrome associated with triptan monotherapy. N Eng J Med 358:2185–2186Google Scholar
- Tashakori A, Afshari R (2010) Tramadol overdose as a cause of serotonin syndrome: a case series. Clin Toxicol 48:337–341Google Scholar
- Taylor DC (1962) Alarming reaction to pethidine in patients on phenelzine. Lancet 280(7252):401–402Google Scholar
- Tzschentke TM, Christoph T, Kögel B et al (2007) (−)-(1R,2R)-3-(3-dimethylamino-1-ethyl-2-methyl-propyl)-phenol hydrochloride (tapentadol HCl): a novel µ-opioid receptor agonist/norepinephrine reuptake inhibitor with broad-spectrum analgesic properties. J Pharmacol Exp Ther 323:265–276PubMedGoogle Scholar
- Tzschentke TM, Jahnel U, Kogel B et al (2009) Tapentadol hydrochloride: a next-generation, centrally acting analgesic with two mechanisms of action in a single molecule. Drugs Today (Barc) 45:483–496Google Scholar
- Walczyk H, Liu CH, Alafris A et al (2016) Probable tapentadol-associated serotonin syndrome after overdose. Hosp Pharm 51:320–327Google Scholar
- Weiner AL (1999) Meperidine as a potential cause of serotonin syndrome in the emergency department. Acad Emerg Med 6:56–58Google Scholar
- Whyte IM (2004) Serotonin syndrome (toxicity). In: Dart RC (ed) Medical toxicology, 3rd edn., vol. 1. Lippincott Williams and Wilkens, Baltimore, pp 103–106Google Scholar
- Yap CY, Taylor DA (1983) Involvement of 5-HT2 receptors in the wet-dog shake behavior induced by 5-hydroxytryptamine in the rat. Neuropharmacol 22:801–804Google Scholar