Introduction
Selective serotonin reuptake inhibitors (SSRIs), are among the most frequently prescribed medicines in modern psychopharmacology1,2. Their popularity is not accidental. Compared with tricyclic antidepressants and monoamine oxidase inhibitors, SSRIs generally have better tolerability, fewer anticholinergic and cardiovascular effects, and substantially safer profiles in overdose3–5. Yet the clinical mistake is to treat them as interchangeable. Fluoxetine, sertraline, paroxetine, citalopram, escitalopram and fluvoxamine all inhibit serotonin reuptake, but they differ meaningfully in half-life, CYP450 inhibition, QT liability, discontinuation risk, adverse-effect profile and suitability for special populations6,7
The central clinical argument is therefore simple:
SSRI prescribing should not be “class-based” alone, but pharmacologically individualized1,2
Individual SSRI personalities
Fluoxetine
Fluoxetine is the prototypical long-acting SSRI, and its clinical behaviour is shaped by both its parent compound and its active metabolite, norfluoxetine. Chemically, fluoxetine is administered as a racemic mixture of R- and S-fluoxetine; both enantiomers inhibit serotonin uptake, while S-fluoxetine is eliminated more slowly and predominates at steady state. Its major active metabolite, norfluoxetine, is formed by hepatic demethylation and contributes substantially to the drug’s prolonged serotonergic activity, with S-norfluoxetine showing particularly potent serotonin uptake inhibition16,17.This long pharmacological tail is clinically important: fluoxetine and norfluoxetine accumulate during chronic treatment and may persist in the body for weeks after discontinuation, making missed doses less likely to cause abrupt withdrawal symptoms but making drug interactions, adverse effects and switching slower to resolve. For this reason, at least five weeks should generally elapse after stopping fluoxetine before starting a monoamine oxidase inhibitor18. Fluoxetine also has lower affinity for muscarinic, histaminergic and α₁-adrenergic receptors than tricyclic antidepressants, helping explain its more favourable anticholinergic, sedative and cardiovascular tolerability profile9. Although its principal mechanism remains serotonin reuptake inhibition, preclinical evidence suggests that fluoxetine may increase extracellular norepinephrine and dopamine in the prefrontal cortex at higher exposures, which may partly explain why some patients experience it as relatively “activating”19,20. Clinically, this activating quality may be useful when depression is accompanied by fatigue, hypersomnia or concern about discontinuation symptoms, but it can also worsen insomnia, nervousness or anxiety early in treatment.¹ Fluoxetine also has an important role in younger populations, being FDA-approved for major depressive disorder from age 8 years, although treatment in children and adolescents requires careful monitoring for activation, behavioural change and suicidality16.
Sertraline
Sertraline is often the practical middle ground. It has a half-life of about 26 hours, reaches steady state in roughly one week, has a relatively modest CYP interaction burden and is widely used in patients with cardiovascular disease, older adults and pregnancy when clinically appropriate21–23. Sertraline reaches peak plasma concentrations approximately 4.5–8.4 hours after dosing and is metabolized to desmethylsertraline, a longer-lived metabolite with substantially weaker pharmacological activity. Its early gastrointestinal adverse effects, including nausea or diarrhea, are common clinical limitations, but it remains one of the most balanced choices when polypharmacy matters21,23,24
Paroxetine
Paroxetine is pharmacologically powerful but clinically “messy.” It is a potent SERT inhibitor, a strong CYP2D6 inhibitor, has nonlinear metabolism and is strongly associated with discontinuation symptoms25,26. Paroxetine has strong CYP2D6 inhibitory activity and an elimination half-life of approximately 21 hours after repeated dosing. Its nonlinear pharmacokinetics reflect a saturable metabolic pathway, which partly explains why exposure can rise disproportionately at higher doses26. It is also more anticholinergic than other SSRIs, making it less attractive in older adults, cognitive impairment, constipation, urinary retention, falls risk or polypharmacy15. Its use is not forbidden, but it should rarely be the casual first choice15,25.
Citalopram
Citalopram and escitalopram are pharmacologically “cleaner” in terms of CYP inhibition, but their clinical identity is shaped by QT prolongation27,28. Citalopram is racemic, while escitalopram is the S-enantiomer responsible for most serotonergic activity29. Both can be effective and well tolerated, but dose, age, electrolytes, hepatic impairment, CYP2C19 status and co-prescribed QT-prolonging drugs matter27,28. MHRA guidance states that both medicines have dose-dependent QT effects and lower maximum recommended doses in older adults30. Citalopram exposure is increased in CYP2C19 poor metabolizers, which is clinically important because higher plasma concentrations can increase QT-prolongation risk; therefore, lower maximum dosing is recommended in this group27,28.
Fluvoxamine
Fluvoxamine is different again. It is an SSRI, but its clinical fingerprint is strong inhibition of CYP1A2 and CYP2C19, with additional effects on other CYP enzymes. This makes it useful in some obsessive-compulsive disorder contexts but challenging in polypharmacy31,32. Fluvoxamine is a potent inhibitor of CYP1A2 and can markedly increase tizanidine exposure when the two drugs are co-administered, creating a clinically significant risk of hypotension, sedation and related adverse effects32. In a patient using clozapine33, olanzapine34 theophylline35, warfarin36 caffeine-heavy intake or tizanidine, fluvoxamine should trigger careful interaction review31,37.
These individual differences provide the foundation for safer SSRI prescribing. The next sections examine how shared class effects and drug-specific pharmacology translate into practical clinical concerns, from serotonin toxicity and withdrawal symptoms to CYP450 interactions, QT prolongation, sexual dysfunction, weight change and use in older adults or pregnancy.
Serotonin toxicity: a class risk amplified by combinations
Serotonin syndrome is not caused by SSRIs alone in most ordinary prescribing situations; it usually emerges when serotonergic load increases through combinations, overdose or rapid dose escalation. The classic clinical triad is mental-status change, autonomic instability and neuromuscular hyperactivity38,39. High-risk combinations include SSRIs with MAOIs, linezolid, methylene blue, tramadol, pethidine, fentanyl, lithium, MDMA, St John’s wort and other serotonergic antidepressants16,38. A recent focused review describes serotonin syndrome as a drug-induced state of excessive serotonergic receptor stimulation requiring early recognition and prompt management38.
Drugs associated with serotonin syndrome
| Serotonin-reuptake inhibitors | Monoamine oxidase inhibitors | Serotonin-releasing agents | Others |
|---|---|---|---|
| Selective serotonin-reuptake inhibitors (SSRIs) e.g., fluoxetine, sertraline |
Phenelzine, tranylcypromine | Amphetamines | Lithium |
| Serotonin–noradrenaline reuptake inhibitors (SNRIs) e.g., venlafaxine, duloxetine |
Moclobemide | Methylphenidate | Tryptophan |
| Tricyclic antidepressants (TCAs) e.g., clomipramine, amitriptyline |
Linezolid | Synthetic stimulants e.g., ecstasy, cathinones |
Buspirone |
| St John’s wort | Parkinson’s treatment e.g., selegiline, rasagiline |
Vortioxetine | |
| Opioid analgesics e.g., pethidine, tramadol, fentanyl |
Methylene blue |
Symptoms of serotonin syndrome
| Alterations in mental state | Neuromuscular abnormalities | Autonomic hyperactivity |
|---|---|---|
| Agitation | Tremor | Hypertension |
| Anxiety | Clonus | Tachycardia |
| Disorientation | Hyperreflexia | Tachypnoea |
| Restlessness | Muscle rigidity | Hyperthermia |
| Excitement | Bilateral Babinski signs | Mydriasis |
| Diaphoresis | ||
| Flushed skin | ||
| Shivering | ||
| Vomiting | ||
| Diarrhoea | ||
| Hyperactive bowel sounds | ||
| Arrhythmias |
The practical risk differs between SSRIs. Fluoxetine is concerning because its long half-life allows serotonergic activity to persist after discontinuation which is why extended monitoring and washout are required when switching to strongly serotonergic or MAOI-like agents16,18. Fluvoxamine is concerning because it can raise concentrations of co-medications while paroxetine is relevant because strong CYP2D6 inhibition may alter exposure to, or activation of, other drugs40. Importantly, not every theoretical interaction translates into frequent severe toxicity. In a population-based cohort study of older adults prescribed linezolid, serotonin syndrome occurred in fewer than 0.5% of patients, and concomitant antidepressant use did not significantly increase the risk; nevertheless, the authors still recommended clinical vigilance41.
Diagnostic algorithm for serotonin toxicity
Severity spectrum of serotonin syndrome
Discontinuation syndrome
SSRI discontinuation is one of the most clinically important differences within the class. Symptoms may include dizziness, nausea, insomnia, vivid dreams, irritability, anxiety, flu-like symptoms and electric-shock sensations18,25. A 2024 systematic review and meta-analysis in The Lancet Psychiatry estimated that antidepressant discontinuation symptoms affect approximately one in six to seven patients after accounting for nonspecific symptoms in placebo groups, with severe symptoms occurring in a smaller minority25. A 2025 JAMA Psychiatry systematic review similarly found measurable discontinuation symptoms shortly after stopping, while emphasizing the importance of distinguishing withdrawal from depressive relapse42.
Paroxetine is the most problematic SSRI for discontinuation because of its shorter half-life, potent SERT binding and self-inhibition of metabolism26. Fluoxetine is the least problematic because of its long half-life and active metabolite16. Sertraline, citalopram, escitalopram and fluvoxamine sit between these extremes18. The clinical lesson is not that SSRIs are addictive in the classic substance-use sense, but that abrupt cessation can produce real physiological symptoms25. Tapering should be planned, gradual and patient-specific18,42.
CYP450 interactions
CYP450 inhibition is where SSRIs most clearly diverge. Fluoxetine and paroxetine are strong CYP2D6 inhibitors43,44. This matters for drugs such as metoprolol, certain antipsychotics, atomoxetine, tamoxifen, codeine and tramadol. For tamoxifen, CYP2D6 inhibition may reduce formation of endoxifen, its active metabolite45. For codeine and tramadol, CYP2D6 inhibition may reduce analgesic activation while tramadol also adds serotonergic toxicity risk46,47. Fluvoxamine has the highest interaction burden overall because of CYP1A2 and CYP2C19 inhibition43,44. By contrast, sertraline, citalopram and escitalopram generally have lower CYP inhibition, although citalopram and escitalopram exposure can rise with CYP2C19 inhibitors such as omeprazole, increasing QT relevance30,48
This is why medication history should drive SSRI choice. In a young adult taking no regular medicines, several SSRIs may be reasonable. In an older adult taking beta-blockers, anticoagulants, antipsychotics, proton-pump inhibitors and analgesics, the safest SSRI may be the one with the least interaction potential43,48,49.
QT prolongation
QT prolongation is not equal across the class. Citalopram and escitalopram have the strongest regulatory warnings for dose-dependent QT prolongation27,50. Risk increases with higher doses, older age, hypokalaemia, hypomagnesaemia, congenital long QT syndrome, bradycardia, structural heart disease and co-prescribed QT-prolonging medicines27,30. MHRA guidance restricts maximum doses in adults older than 65 years and advises caution or monitoring in patients with cardiac risk factors30.
This does not mean citalopram or escitalopram are “bad” drugs. It means they are less forgiving when exposure rises. Escitalopram can be an elegant and well-tolerated SSRI in many patients, but dose ceilings and ECG/electrolyte considerations should be respected especially in older adults and patients receiving CYP2C19 inhibitors such as omeprazole or other QT-prolonging medicines51. Sertraline is often preferred when QT risk, coronary heart disease or polypharmacy are prominent, because contemporary NHS Specialist Pharmacy Service guidance identifies sertraline as the first-line SSRI option in coronary heart disease and notes its few known cardiac adverse effects and minimal interaction with cardiac medicines24.
Sexual dysfunction and weight
Sexual dysfunction is a major SSRI class effect and can include reduced libido, delayed orgasm, anorgasmia, erectile dysfunction, ejaculatory delay and genital sensory changes52. Paroxetine is often considered among the highest-risk SSRIs, though sexual adverse effects can occur with any agent53. The Australian therapeutic goods administration (TGA) updated product information in 2024 to warn that sexual dysfunction may persist after SSRI/SNRI discontinuation, while noting that persistent cases appear rare, underreported and of uncertain prevalence54.
Weight effects are also clinically important. SSRIs may be relatively weight-neutral early in treatment, but long-term weight gain can occur55,56. A 2024 target-trial emulation study in Annals of Internal Medicine found small but clinically relevant differences in weight change across first-line antidepressants; escitalopram and paroxetine were associated with a higher 6-month risk of gaining at least 5% of baseline body weight than sertraline, whereas fluoxetine did not differ meaningfully from sertraline at 6 months55. Fluoxetine can also produce appetite suppression or weight loss in some patients, especially early in treatment, while paroxetine has historically been more associated with weight gain and sedation16,55. For many patients, sexual function and weight are not “minor” tolerability issues; they determine adherence, self-image, relationships and willingness to continue treatment52,54.
Special populations: older adults and pregnancy
In older adults, SSRI selection should begin with vulnerability rather than dose alone. Hyponatremia, falls, bleeding risk with antiplatelets or anticoagulants, QT prolongation, renal/hepatic impairment and drug interactions all become more relevant15,24.The 2023 AGS Beers Criteria identifies paroxetine as potentially inappropriate in older adults because of strong anticholinergic properties15. Sertraline is often a pragmatic first choice when cardiac safety and polypharmacy matter. Escitalopram can be reasonable, but lower maximum doses and QT considerations are important29,30.
In pregnancy, the decision is not “SSRI versus no risk.” Untreated depression and anxiety are themselves exposures, associated with impaired functioning, poor antenatal care, substance use, suicidality, preterm birth and postpartum deterioration57,58. ACOG’s 2023 guideline recommends SSRIs as first-line pharmacotherapy for perinatal depression and anxiety, with sertraline or escitalopram reasonable first-line options when there is no previous medication-response history. UKTIS emphasizes that SSRI pregnancy data are sometimes conflicting and that treatment should balance fetal/neonatal risks against maternal relapse risk. Therefore, a patient stable on an effective SSRI should not be switched automatically solely because of pregnancy.
Conclusion
SSRIs transformed depression and anxiety treatment because they are effective, scalable and generally safer than older antidepressants. Their familiarity, however, can create therapeutic complacency. The best SSRI is not always the newest, the most prescribed or the one a clinician uses most often. It is the one whose pharmacokinetics, pharmacodynamics, adverse-effect profile and interaction risks fit the person in front of the prescriber.
For a typical adult without major comorbidity, sertraline or escitalopram often offer a favorable balance of tolerability and practicality. For patients at high risk of withdrawal, fluoxetine may be useful. For complex polypharmacy, sertraline, citalopram or escitalopram may be preferable over fluvoxamine, fluoxetine or paroxetine, though QT risk modifies that choice. For older adults, paroxetine should usually be avoided. For pregnancy, sertraline and escitalopram are commonly reasonable new-start options, but stability on an existing SSRI matter. Ultimately, SSRI prescribing is precision pharmacology in everyday clinical clothing: small molecular differences become large clinical differences when applied to real patients