Recurrent serotonin syndrome during low-dose venlafaxine treatment in a patient with reduced CYP2D6 activity

Summary

Serotonin syndrome is a potentially life-threatening toxicity diagnosed clinically and characterised by a variable clinical presentation.

Spontaneous clonus in a patient receiving a serotonergic antidepressant should prompt immediate consideration of serotonin syndrome, even when the prescribed dose and measured drug concentrations are not high.

Unexpected toxicity at low doses should also prompt a systematic review of pharmacokinetic interactions, organ function and pharmacogenetic susceptibility.

A man in his late 70s with transthyretin amyloidosis experienced at least five episodes of serotonin syndrome (four on venlafaxine immediate release (IR) ≤75 mg/day), all meeting Hunter criteria.

Trough plasma concentrations revealed markedly elevated exposure relative to the administered venlafaxine dose. At the highest measured dose, venlafaxine and O-desmethylvenlafaxine (ODV) concentrations were 377 nmol/L and 1330 nmol/L, respectively, yielding an active moiety concentration of 1707 nmol/L. All trough concentrations exceeded the expected values for a 75 mg dose (venlafaxine ~38 nmol/L, ODV ~330 nmol/L, active moiety ~365 nmol/L), while remaining within the therapeutic reference range of approximately 518–2220 nmol/L. The metabolite to parent ratio was low, suggestive of reduced CYP2D6 activity.

Background

Venlafaxine is a serotonin-norepinephrine reuptake inhibitor¹ whose pharmacological effects are mediated primarily by the inhibition of serotonin reuptake, with additional noradrenergic activity at higher doses. Like other serotonergic agents, it can precipitate serotonin syndrome, through excessive stimulation of central and peripheral serotonergic pathways, typically within 1 to 2 hours after administration.

The diagnosis is primarily clinical and classically manifests as a triad of neuromuscular abnormalities, autonomic hyperactivity and altered mental status,² from mild to life-threatening.³

It is based on the Hunter Serotonin Toxicity Criteria (table 1), which demonstrate high sensitivity (84%) and specificity (97%).² These criteria require exposure to one or more serotonergic agents in combination with features such as clonus, hyperreflexia, tremor, hypertonia, agitation or hyperthermia.⁴

Table 1. The Hunter Serotonin Toxicity Criteria⁴, in the presence of recent exposure to a serotonergic agent.

1	Spontaneous clonus
2	Inducible clonus and agitation or diaphoresis
3	Ocular clonus and agitation or diaphoresis
4	Tremor and hyperreflexia
5	Hypertonia and temperature >38°C and ocular clonus or inducible clonus

There is also a scoring system proposed by Radomski that is based on the presence of major and minor symptoms.⁵ However, more than half of serotonin syndrome cases do not strictly meet the Radomski criteria,⁶ hence a lower sensitivity.

Although the true prevalence of serotonin syndrome remains unknown, risk is increased in the setting of polypharmacy, particularly among older adults, due to both pharmacokinetic and pharmacodynamic drug–drug interactions.² In addition to extrinsic factors, intrinsic determinants, most notably genetic polymorphisms affecting drug metabolism and transport, may influence both antidepressant response and toxicity.

Venlafaxine is extensively metabolised in the liver, by CYP2D6 with a minor contribution from CYP2C19. CYP2D6 catalyses O-demethylation to O-desmethylvenlafaxine (ODV), the principal active metabolite. ODV has comparable pharmacological activity and contributes to the antidepressant effect. CYP2C19 and CYP3A4 contribute to alternative metabolic pathways, including N-demethylation to N-desmethylvenlafaxine,⁷ which has weaker pharmacological activity and is not routinely measured. Venlafaxine and ODV are subsequently conjugated and eliminated predominantly by the kidneys. Consequently, genetic variation in these enzymes may alter venlafaxine exposure and influence both efficacy and toxicity.¹

Several polymorphisms have been implicated in venlafaxine pharmacokinetics and pharmacodynamics, including cytochrome P450 enzymes, particularly CYP2D6 and CYP2C19, variants in the serotonin transporter SLC6A4 and serotonin receptors (eg, HTR2A).⁸ P-glycoprotein-mediated efflux may also influence systemic and central nervous system exposure.⁹ Because the present case focuses on altered venlafaxine exposure, the discussion primarily addresses CYP2D6 and CYP2C19-mediated pharmacokinetic variability.

To facilitate comprehension, we have provided a pharmacogenetics glossary (table 2).

Table 2. Pharmacogenetic glossary.

Allele	One version of a gene defined by the association of several single-nucleotide polymorphisms: SNPs (eg, c.506-1G>A rs3892097+c0.100C>T rs1065852+c0.408G>C rs1058164+c0.1457G>C rs1135840 --> CYP2D6 *4)
Diplotype	The combination of the two inherited alleles (eg, *4/*9)
Star allele nomenclature (*)	Standardised nomenclature used to designate specific genetic variants of drug-metabolising enzymes (eg, CYP2D6 *1, *4, *9)
Activity score/allelic function	Numerical estimate of the combined enzymatic activity predicted by the diplotype
Metaboliser phenotype	Clinical interpretation of the activity score (poor, intermediate, normal or ultrarapid metaboliser, depending on the gene)
CYP2D6	Highly polymorphic: more than 180 alleles described *1: reference (“wild-type”) allele with normal enzyme function *4: non-functional allele *9: decreased-function allele
CYP2C19	Around 35 alleles described
CYP3A4	Around 50 alleles described
ABCB1 (P-gp)	Does not use the star allele nomenclature

SNPs, single-nucleotide polymorphisms.

Case presentation

This case involved a man in his late 70s, with a body mass index of 23 kg/m², transthyretin amyloidosis complicated by heart failure with reduced ejection fraction, chronic kidney disease and atrial fibrillation. He was a non-smoker and did not consume alcohol. His long-term treatment included amiodarone, dapagliflozin, dutasteride, eplerenone, gabapentin, tafamidis, torasemide and metoclopramide (as needed). His estimated glomerular filtration rate was between 25 and 40 mL/min/1.73 m².

He had a history of recurrent depressive disorder and experienced at least five episodes of serotonin syndrome (four on venlafaxine ≤75 mg/day IR), all meeting Hunter criteria.

Psychiatric treatment consisted of escitalopram from 2018 (taken intermittently), followed by mirtazapine in 2023 and a switch to immediate-release venlafaxine in March 2024, initiated at 37.5 mg/day and titrated to 75 mg/day.

The clinical course is summarised chronologically below (figure 1):

Figure 1. Different episodes of serotonin syndrome fulfilling the Hunter Serotonin Toxicity Criteria. The patient experienced at least five episodes of serotonin syndrome: one during escitalopram treatment and four during low-dose immediate-release venlafaxine. All episodes occurred in the setting of reduced renal function (eGFR 25–40 mL/min/1.73 m²) and concomitant cholestasis. The episodes were temporally associated with venlafaxine initiation or dose escalation and fulfilled the Hunter Serotonin Toxicity Criteria. Clinical features included tremor, hyperreflexia, spontaneous or inducible clonus, rigidity, agitation, anxiety, akathisia, hallucinations, confusion, diaphoresis and during the most severe episode, hypertensive crisis. Serial therapeutic drug monitoring revealed disproportionately high venlafaxine and ODV concentrations relative to the prescribed dose. Concomitant bupropion may have further increased venlafaxine exposure through CYP2D6 inhibition. Symptoms resolved after complete discontinuation of antidepressant therapy. ODV, O-desmethylvenlafaxine.

• 2022: Following reinitiation of escitalopram in 2022, the patient developed spontaneous clonus.

• March 2024: Immediate-release venlafaxine was initiated at 37.5 mg/day and titrated to 75 mg/day while he was receiving mirtazapine. He developed rigidity, spontaneous clonus and anxiety. Trough concentrations were venlafaxine 463 nmol/L (expected ≈38 nmol/L), ODV 1095 nmol/L (expected ≈330 nmol/L) and active moiety 1558 nmol/L (expected ≈365 nmol/L), with a metabolite to parent ratio (MPR) of 2.4 (expected 2.7–7.76). Venlafaxine was reduced to 37.5 mg/day.

• May 2024: Venlafaxine was again increased to 75 mg/day, resulting in recurrent tremors and anxiety. Trough plasma concentrations showed venlafaxine 196 nmol/L, ODV 635 nmol/L (active moiety 831 nmol/L), with an MPR of 3.2.

• July 2024: The patient was rehospitalised with severe tremor, hyperreflexia, clonus, visual, auditory and cenesthetic hallucinations, as well as akathisia. Trough plasma concentrations were venlafaxine 377 nmol/L and ODV 1330 nmol/L (active moiety 1707 nmol/L), with an MPR of 3.5.

• May–September 2025: Clonus persisted despite progressive dose reductions and was initially attributed to gabapentin accumulation in the setting of impaired renal function (creatinine clearance, 26 mL/min). Reducing venlafaxine to 18.75 mg/day deteriorated the depressive symptoms, prompting dose escalation on several occasions and then consideration of an alternative antidepressant to bupropion.

• September 2025: Venlafaxine was partially replaced by bupropion (150 mg/day). Shortly thereafter, the patient developed confusion, worsening clonus, chills and a hypertensive crisis, fulfilling four major Radomski criteria for serotonin syndrome.⁵ This episode was exacerbated by bupropion, a potent CYP2D6 inhibitor, which may have further reduced CYP2D6-mediated metabolism and thereby increased venlafaxine exposure, resulting in phenoconversion superimposed on genetically reduced CYP2D6 activity.

The patient also reported spontaneous haematomas since 2023, related to apixaban therapy, although concomitant serotonergic treatment may have further increased the bleeding risk in the absence of other identifiable risk factors such as thrombocytopenia.

The MPR, calculated as the plasma concentration of ODV divided by that of venlafaxine, provides an indirect measure of CYP2D6 activity, as CYP2D6 is the predominant enzyme responsible for the conversion of venlafaxine to ODV. A low MPR may therefore indicate reduced CYP2D6-mediated metabolism.⁷ However, the ratio is not exclusively determined by CYP2D6 and may also be influenced by interacting drugs, age, organ function, sampling time and contributions from other metabolic pathways.

Therapeutic drug monitoring is performed using trough concentration at steady state and is based on the active moiety, defined as the combined concentration of venlafaxine and ODV. Proposed therapeutic reference ranges vary, from 370 to 1480 nmol/L in earlier guidelines from the Association for Neuropsychopharmacology and Pharmacopsychiatry AGNP¹⁰ to 518–2220 nmol/L in a more recent systematic review, with an alert threshold of 2960 nmol/L.⁷ These population-based ranges should not be interpreted as toxicity thresholds. In this patient, active-moiety concentrations were 1558, 831 and 1707 nmol/L on serial measurements. Although none exceeded the alert threshold, the concentrations were relatively high for the prescribed doses and occurred in the context of recurrent episodes fulfilling the Hunter Serotonin Toxicity Criteria.

Overall, the patient experienced at least five episodes of serotonin syndrome, four during venlafaxine treatment. Considering the repeated positive rechallenge with venlafaxine, the causal relationship between venlafaxine exposure and the occurrence of serotonin syndrome was considered certain.

Investigations

All trough concentrations were above expected values (75 mg: venlafaxine ~38 nmol/L, ODV ~330 nmol/L, active moiety ~365 nmol/L).

Low MPRs indicate CYP2D6 impairment.

Genotyping revealed CYP2D6 *4/*9 diplotype (activity score 0.25) predicting an intermediate metaboliser phenotype, supported by dextromethorphan phenotyping (dextromethorphan to metabolite ratio, 0.63). The patient also carried a CYP2C19 *1/*17 genotype consistent with an ultrarapid metaboliser phenotype, supported by omeprazole phenotyping (omeprazole to metabolite ratio 0.33).

CYP2D6 activity was further reduced by concomitant use of amiodarone, metoclopramide and bupropion, all reported to inhibit CYP2D6. Additionally, the patient carried the ABCB1 rs9282564 c.61A>G variant encoding P-glycoprotein (table 3).

Table 3. Genotype and phenotype of cytochrome P450 and ABCB1 (P-gp).

	Genotype	Predicted phenotype	Metabolic ratio	Measured phenotype
CYP 2C19 *1/*17	c.-806C>T rs12248560	Predicted increased activity (AS ≥2)	0.33	Ultrarapid metaboliser phenotype
CYP 2D6 *4/*9	CNV: 2 *9 c.841_843delAAG rs5030656 *4 c.506-1G>A rs3892097 c.100C>T rs1065852 c.408G>C rs1058164 c.1457G>C rs1135840	Predicted reduced activity (AS=0.25)	0.63	Intermediate metaboliser phenotype
ABCB1 (P-gp)	c.61A>G rs9282564	Predicted unknown activity	–	–

AS, activity score.

Outcome and follow-up

Several months after stopping his antidepressant therapy, the patient attended a pharmacogenetic consultation to review and clarify the findings. By that time, discontinuation of all antidepressants had resulted in complete resolution of tremor and clonus, along with a significant improvement in the underlying depressive symptoms. This progress appeared to be associated with an overall better quality of life and a restored ability to carry out daily activities without these adverse effects.

Discussion

This case demonstrates recurrent serotonin toxicity during low-dose venlafaxine treatment in a patient with reduced CYP2D6 activity. The CYP2D6 *4/*9 diplotype, low or borderline MPR, reduced ODV concentrations, persistently increased venlafaxine exposure and concordant dextromethorphan phenotyping provide mutually supportive evidence of reduced CYP2D6-mediated metabolism.

However, a low MPR or reduced ODV concentration alone is insufficient to establish genetically determined impaired CYP2D6 activity, as these measures can be affected by multiple non-genetic factors, including renal function, age, sampling conditions and concomitant medications. In this case, amiodarone, metoclopramide and later bupropion may have further inhibited CYP2D6 activity,¹¹ thereby reducing venlafaxine metabolism.

The patient also carries a CYP2C19 ultrarapid metaboliser genotype. Its contribution was probably limited given the relatively minor role of CYP2C19 in the formation of ODV.¹²

A previously reported patient with a similar metabolic profile (CYP2D6 *1/*4 and CYP2C19 *1/*17) developed venlafaxine-associated cardiotoxicity, providing additional support for the clinical relevance of altered venlafaxine metabolism.¹³

Other pharmacogenetic factors may have influenced individual susceptibility to venlafaxine toxicity and therapeutic effects, although their clinical significance remains uncertain. The ABCB1 rs9282564 variant may affect P-glycoprotein activity, but its effect appears substrate-dependent, and no published data specifically link this variant to venlafaxine pharmacokinetics or toxicity.

Variants in HTR2A, HTR2C, and SLC6A4 have been associated with variability in antidepressant response and adverse effects, but their role in serotonin toxicity remains unclear. These variants were not investigated in this case and therefore cannot be considered established contributors.^{8 14 15}

Current CPIC recommendations support considering an alternative antidepressant that is not predominantly metabolised by CYP2D6 in CYP2D6 poor metabolisers because of the potential for increased exposure and adverse effects. Recommendations are less prescriptive for intermediate metabolisers but emphasise the importance of concomitant CYP2D6 inhibitors.¹⁵

In this case, the distinction between genotype-predicted and functional CYP2D6 activity was particularly relevant, as pharmacological inhibition likely further reduced an already impaired metabolic capacity.

Several alternative diagnoses were considered throughout the clinical course, contributing to the delayed recognition of venlafaxine toxicity and the prolonged continuation of treatment, including parkinsonism, gabapentin accumulation related to chronic kidney disease, neuroleptic malignant syndrome in the context of metoclopramide exposure, and metabolic causes of clonus.

Chronic kidney disease may have increased exposure to venlafaxine, ODV or gabapentin, but it did not adequately explain the recurrent, dose-related episodes.

Neuroleptic malignant syndrome was considered unlikely because metoclopramide exposure was brief, the chronology was inconsistent, and symptoms resolved after venlafaxine withdrawal without specific treatment.

The other differential diagnoses did not adequately account for the temporal relationship with venlafaxine initiation or dose escalation, the recurrence after rechallenge, or the resolution following withdrawal of serotonergic antidepressants.

The repeated fulfilment of the Hunter criteria supported a serotonergic mechanism. Although the initial episodes were mild and incompletely documented, retrospective review indicated that they fulfilled the Hunter criteria from the outset. Radomski’s criteria were met only during the final episode, after bupropion was introduced, which may have made the diagnosis more clinically apparent.

The relatively low venlafaxine doses and active moiety concentrations below conventional laboratory alert thresholds may initially have provided false reassurance. However, serotonin toxicity remains a clinical diagnosis, and concentrations within population-based reference ranges do not exclude toxicity in a susceptible individual. In this case, the combination of pharmacogenetic susceptibility, pharmacological inhibition, advanced age, multimorbidity and impaired renal clearance probably contributed to toxicity at doses below the usual therapeutic range.^{16 17}

Earlier recognition of the Hunter criteria, together with a systematic review of pharmacokinetic and pharmacodynamic interactions, may have facilitated earlier recognition and prevented subsequent rechallenges. Therapeutic drug monitoring and pharmacogenetic testing may also have helped identify increased individual susceptibility.

Several cases of recurrent serotonin syndrome with positive antidepressant rechallenge have been reported. However, previous reports generally involved higher antidepressant doses or lacked pharmacogenetic characterisation and therapeutic drug monitoring.^{18 19} To our knowledge, this is the first reported case of repeated serotonin syndrome during low-dose venlafaxine treatment with multiple episodes of serotonin toxicity after venlafaxine rechallenge, combined CYP2D6/CYP2C19 genotyping, dextromethorphan phenotyping and serial therapeutic drug monitoring.

Patient’s perspective.

At presentation, the patient experienced severe and unexpected tremors associated with spontaneous clonus, which were highly disabling, limiting mobilisation and significantly restricting daily activities. These symptoms were a major source of anxiety; he initially feared a serious underlying condition, and each recurrence or worsening led to considerable distress and discouragement, further exacerbating his anxiety.

At follow-up, several months after discontinuation of antidepressant therapy, the patient reported complete resolution of tremor and clonus, along with a marked improvement in mood. He regained independence in daily activities and described a substantially improved quality of life. He expressed pride and satisfaction in contributing to this case report, was highly cooperative throughout the process, and hoped that his experience might help prevent similar adverse outcomes in other patients.

Learning points.

• Consider serotonin toxicity in any patient receiving a serotonergic agent who develops spontaneous clonus, tremor or hyper-reflexia.

• Perform therapeutic drug monitoring when available: blood sampling should be performed at steady state (ie, after at least four half-lives of regular dosing) and at trough level (ie, immediately before the next scheduled dose).

• When toxicity occurs at an unexpectedly low dose, review renal function, drug interactions and pharmacogenetic susceptibility.

• Consider pharmacogenetic testing in cases of recurrent toxicity or unexpected plasma concentrations.

• Reconsider further rechallenge with the same serotonergic agent.