The Stickiest SSRI · Hell to Quit
IUPAC: (3S,4R)-3-[(2H-1,3-benzodioxol-5-yloxy)methyl]-4-(4-fluorophenyl)piperidine · C₁₉H₂₀FNO₃ · MW 329.37 g/mol · CAS 61869-08-7 (free base) · marketed as the HCl hemihydrate / mesylate
Paroxetine (Paxil · Seroxat · Aropax · Brisdelle). A selective serotonin reuptake inhibitor — but the most extreme one. It binds SERT with sub-nanomolar affinity (tighter than any other SSRI), is the most anticholinergic of the class, and famously destroys its own metabolizing enzyme (CYP2D6). Short half-life + no active metabolite + brutal receptor adaptation = the worst discontinuation syndrome in psychiatry. Trade names: Paxil, Seroxat, Aropax, Deroxat, Pexeva.
Paroxetine is a competitive inhibitor of the serotonin transporter (SERT / SLC6A4). Like cocaine, it binds the orthosteric central (S1) site and occludes the pore without being transported — but where cocaine is a promiscuous, micromolar monoamine blocker, paroxetine hits SERT with sub-nanomolar affinity (Ki ≈ 0.07 nM), the tightest of any marketed SSRI. Blocking reuptake raises synaptic 5-HT; the therapeutic effect, however, emerges over weeks as somatodendritic 5-HT1A autoreceptors desensitize and postsynaptic signaling remodels.
Paroxetine wedges into the S1 pocket between TM1/TM3/TM6/TM8, its 4-fluorophenyl ring in subsite B and benzodioxole in subsite C, salt-bridging Asp98. PDB 5I6X resolves paroxetine here in human SERT. Ki ≈ 0.04–0.13 nM — a near-irreversible clamp on an alternating-access carrier.
Like escitalopram, SSRIs can occupy a shallow allosteric (S2) vestibule that slows dissociation from S1. Paroxetine's extremely low Ki reflects a very slow off-rate — the molecule stays put, which matters both for efficacy and for the abruptness of withdrawal when levels fall.
Paroxetine is the most anticholinergic SSRI (M1 Ki ≈ 35 nM). Muscarinic blockade drives dry mouth, constipation, blurred vision, sedation, cognitive dulling and weight gain — a TCA-like tail bolted onto an SSRI. It also fuels cholinergic rebound on discontinuation.
Paroxetine has the highest noradrenaline-transporter affinity of the SSRIs (NET Ki ≈ 86 nM). Clinically minor at low doses, but at higher doses a genuine noradrenergic component appears — nudging paroxetine toward SNRI-like territory.
Paroxetine inhibits neuronal nitric oxide synthase (nNOS). Reduced NO signaling in spinal and peripheral pathways contributes to its notably high burden of sexual dysfunction (delayed orgasm, anorgasmia) relative to other SSRIs.
Paroxetine is oxidized by CYP2D6 to a reactive carbene that forms a stable adduct with the heme iron — mechanism-based (irreversible) inactivation. It kills the very enzyme that clears it, producing saturable, nonlinear kinetics (see §03).
Paroxetine is well absorbed orally but undergoes extensive first-pass metabolism by CYP2D6. Because it irreversibly inactivates CYP2D6, its clearance falls as dosing continues: plasma exposure rises disproportionately (nonlinearly) with dose and time. A modest dose increase can more than double steady-state concentrations. Half-life averages ~21 h and — critically — there is no active metabolite to cushion a missed dose. Short T½ + no active metabolite is exactly the PK signature that makes paroxetine the worst SSRI to stop.
Metabolism cascade: CYP2D6 oxidizes the methylenedioxy bridge to an unstable catechol intermediate, which is methylated by COMT and then conjugated (glucuronide / sulfate) for renal and fecal excretion. The oxidative step also generates the reactive species that inactivates CYP2D6 itself.
CYP2D6 poor metabolizers (~7–10% of Europeans) start with high exposure; everyone else effectively becomes a poor metabolizer within days as the drug knocks out the enzyme. This is why paroxetine is also a potent perpetrator of drug interactions: by shutting down CYP2D6 it raises levels of TCAs, antipsychotics, metoprolol and atomoxetine, blocks codeine/tramadol activation (reduced analgesia), and — clinically important — cripples conversion of tamoxifen to active endoxifen in breast-cancer patients.
Most drugs are cleared by enzymes at a rate proportional to plasma concentration — clean first-order kinetics. Paroxetine breaks that rule. Its oxidation by CYP2D6 produces a reactive carbene that coordinates the heme iron and forms a metabolic-intermediate complex, permanently disabling the enzyme molecule. New CYP2D6 must be synthesized (days) to recover activity. The practical consequence: as you keep dosing, the clearance pathway shrinks, and concentration climbs faster than dose.
Dose 20 → 40 mg does not double exposure — it can more than triple it, because at 40 mg CYP2D6 is more completely saturated and inactivated. This is dose- and time-dependent nonlinearity: the AUC/dose ratio rises with both. It makes paroxetine harder to titrate predictably than SSRIs with linear PK (e.g. citalopram), and it means small dose changes near the top of the range can produce outsized shifts in concentration, side-effect burden, and — on the way down — withdrawal severity.
Perpetrator, not just victim: the same inactivation makes paroxetine one of the strongest CYP2D6 inhibitors in the pharmacopeia. Co-medications cleared by CYP2D6 accumulate; prodrugs activated by CYP2D6 fail. The tamoxifen → endoxifen block is the textbook cautionary tale — avoid paroxetine in patients on tamoxifen.
Paroxetine's clinical profile is not just "more serotonin." Its off-target promiscuity — muscarinic, noradrenergic, nitrergic — plus its adaptive receptor changes give it a texture unlike the cleaner SSRIs. It is a strong anxiolytic and antidepressant (approved for MDD, OCD, panic, social anxiety, GAD, PTSD, and menopausal hot flushes as Brisdelle), with a side-effect load that reads halfway to a tricyclic.
SERT blockade elevates 5-HT across limbic and cortical circuits. Acute effect is often increased anxiety/activation (presynaptic 5-HT1A feedback dampens firing); therapeutic benefit tracks the slow desensitization of somatodendritic 5-HT1A autoreceptors in the raphe over 2–6 weeks. High SERT occupancy at standard doses explains strong efficacy in OCD and panic.
Muscarinic M1 antagonism (Ki ~35 nM) makes paroxetine the most "TCA-like" SSRI: dry mouth, constipation, blurred vision, urinary hesitancy, sedation, and the greatest weight gain of the class. Anticholinergic load is also a cognitive concern in older adults. When the drug is stopped, loss of chronic muscarinic blockade produces cholinergic rebound — a major driver of discontinuation misery.
Elevated 5-HT (via 5-HT2) suppresses dopaminergic and nitrergic sexual signaling; paroxetine adds direct nNOS inhibition and mild prolactin elevation. The result is the highest reported rate of delayed orgasm, anorgasmia, and reduced libido among SSRIs — and paroxetine features prominently in reports of post-SSRI sexual dysfunction (PSSD) that can persist after discontinuation.
When a system adapted to constant SERT blockade suddenly loses it — and paroxetine's short T½ and absent active metabolite guarantee a sharp drop — downregulated serotonergic tone is unmasked. The syndrome (FINISH: Flu-like, Insomnia, Nausea, Imbalance, Sensory disturbance, Hyperarousal), including the classic electric-shock "brain zaps," is more frequent and severe with paroxetine than any other SSRI. Anticholinergic rebound layers on top.
Paroxetine is the SSRI most consistently flagged for teratogenic risk. On the strength of epidemiological signals, the FDA moved it to Pregnancy Category D and issued a specific warning: first-trimester exposure is associated with an increased risk of congenital cardiac malformations — principally atrial and ventricular septal defects — with most analyses estimating roughly a 1.5–2× relative increase over baseline. The absolute risk remains low, but paroxetine is singled out where the class as a whole is not.
Third-trimester exposure carries risk of a neonatal adaptation syndrome (jitteriness, feeding and respiratory difficulty, irritability) and a small increase in persistent pulmonary hypertension of the newborn (PPHN). Paroxetine's short half-life may make neonatal withdrawal comparatively abrupt.
Guidance (ACOG / FDA) is to avoid initiating paroxetine in people who are pregnant or planning pregnancy where an alternative is reasonable — but never to stop an effective antidepressant abruptly, given both maternal relapse risk and paroxetine's severe discontinuation syndrome. Decisions are individualized; abrupt cessation is its own hazard.
Paroxetine is a conformationally constrained ligand: a rigid piperidine ring bridging a
4-fluorophenyl group and a fused benzodioxole, with only a handful of rotatable bonds. It arrives at the SERT
S1 pocket largely pre-organized, so the ΔS_conf (conformational entropy) penalty it pays on binding
is small — one of the structural reasons its affinity is sub-nanomolar rather than merely nanomolar.
The transporter is the flexible partner. Unliganded SERT samples outward-open, occluded and inward-open states
as part of its alternating-access cycle — a high H_pocket (Shannon entropy over accessible
conformers). Paroxetine wedges into the central site (resolved in PDB 5I6X), salt-bridging Asp98
and packing its aryl groups into subsites B and C, and traps a single outward-oriented conformation.
That is entropy collapse: H_pocket falls sharply, the ensemble is quenched to essentially one
state, and — with the drug's slow off-rate — it stays quenched.
Theoretical binding free energy from the measured affinity: ΔG = −RT·ln(1/Ki) ≈
−13.9 kcal/mol at 298 K for Ki ≈ 0.07 nM — roughly 5 kcal/mol tighter than cocaine's
micromolar-to-nanomolar grip on DAT (~−8.7 kcal/mol). In FlexAID∆S / tENCoM decomposition, that gap is dominated by
deep enthalpic burial of a rigid ligand plus minimal conformational entropy loss: the near-ideal thermodynamic
fingerprint of a very high-affinity, slow-dissociating inhibitor. The same slow off-rate that makes it potent is
what makes withdrawal so abrupt once plasma levels finally fall.
Paroxetine is a prescription antidepressant, not a recreational drug — the real hazards are stopping it, combining it, and starting it in the wrong body. Evidence-based, non-alarmist.
| Target | Affinity | Rel. | Mechanism |
|---|---|---|---|
|
SERT
Serotonin transporter (SLC6A4)
|
Ki ≈ 0.07 nM
range 0.04 – 0.13 nM · primary
|
Reuptake inhibitor | |
|
M1
Muscarinic ACh receptor M1 (CHRM1)
|
Ki ≈ 35 nM
most anticholinergic SSRI
|
Antagonist | |
|
NET
Noradrenaline transporter (SLC6A2)
|
Ki ≈ 86 nM
highest NET of the SSRIs
|
Reuptake inhibitor | |
|
CYP2D6
Cytochrome P450 2D6 (metabolic)
|
Mechanism-based
irreversible self-inactivation
|
Suicide inhibitor |