Discontinuation syndromes in psychiatric medications

SSRI/SNRI discontinuation syndrome

Recognition

The FINISH mnemonic (Berber 1998) still holds up:

• Flu-like symptoms (fatigue, myalgia, chills, headache)
• Insomnia (often with vivid dreams)
• Nausea
• Imbalance (dizziness, vertigo)
• Sensory disturbances ("brain zaps" or electric-shock sensations, paresthesias)
• Hyperarousal (anxiety, agitation, irritability)

Onset is usually 1 to 3 days after last dose (later for long half-life drugs). Duration is typically 1 to 3 weeks in mild cases, but can be much longer for high-risk drugs or protracted use. Some patients experience a protracted discontinuation syndrome lasting months, which is real but still debated in how it should be characterized.

Risk by drug

Distinguishing discontinuation from relapse

• Timing. Discontinuation appears within days; relapse takes weeks.
• Physical symptoms. Brain zaps, dizziness, GI symptoms, flu-like features point to discontinuation. Pure mood or thought symptoms point to relapse.
• Response to reinstatement. Discontinuation resolves within 24 to 72 hours of restarting the drug. Relapse doesn't.

If unsure, reinstate. If it resolves quickly, it was discontinuation.

Management

Reinstate at the previous dose (or last well-tolerated dose) and taper more slowly. Standard taper: reduce by 10% of the current dose every 2 to 4 weeks. For sensitive patients or high-risk drugs, use a hyperbolic taper (Horowitz and Taylor): reduce by a proportional amount of the current dose, so the absolute cut gets smaller as the dose gets smaller. Liquid formulations or compounded low-strength capsules make this feasible at the very low end.

Example hyperbolic taper from paroxetine 20:

• 20 to 15 (25% cut)
• 15 to 11 (about 27%)
• 11 to 8
• 8 to 6
• 6 to 4
• 4 to 3
• 3 to 2
• 2 to 1
• 1 to 0.5
• 0.5 to 0

Each step held 2 to 4 weeks, longer if symptoms emerge. Total taper time: often 4 to 12 months for high-risk drugs after prolonged use. That's the trade-off. Rushing produces symptoms that make patients afraid to try again.

Benzodiazepine withdrawal

Recognition

Symptoms: anxiety, insomnia, tremor, sweating, tinnitus, perceptual disturbances (hyperacusis, visual distortion), muscle tension, GI upset, depersonalization, seizures. Severe withdrawal produces delirium, hallucinations, and seizures. This is the one psychiatric discontinuation syndrome that can be a medical emergency.

Timeline

• Short-acting (alprazolam, lorazepam, oxazepam): onset 1 to 2 days after last dose, peaks day 2 to 4, resolves over 1 to 4 weeks (acute phase).
• Long-acting (clonazepam, diazepam, chlordiazepoxide): onset 3 to 7 days after last dose, peaks day 5 to 10, resolves over 2 to 6 weeks (acute phase).
• Protracted withdrawal: anxiety, insomnia, cognitive symptoms lasting months to years, especially after high-dose or long-duration use.

Management

Never stop chronic benzodiazepines abruptly (seizure risk). Stabilize on a long-acting equivalent (usually diazepam or clonazepam) at the current equivalent dose. Then taper 5 to 10% of the current dose every 2 to 4 weeks. For high-dose or protracted-use patients, slower is better (2 to 5% cuts, or hyperbolic tapering). Total taper time: often 6 to 18 months for high-dose users. See the benzodiazepine equivalents guide for specific conversion doses.

If a patient is in acute severe withdrawal (seizures, delirium): treat as alcohol withdrawal syndrome would be treated (long-acting benzodiazepine loading, monitored setting).

Stimulant hold or discontinuation

Stimulants don't produce classic physiological withdrawal in therapeutic use, but abrupt discontinuation causes:

• Rebound fatigue and hypersomnia (peaks day 1 to 3)
• Low mood, sometimes marked
• Increased appetite
• Difficulty concentrating (of course; ADHD symptoms are unmasked)
• Irritability

Timeline: symptoms peak 1 to 4 days after stopping, resolve over 1 to 2 weeks. In patients on high doses or with heavier catecholamine dependence, symptoms can be more marked.

Weekend or holiday drug holidays: reasonable for some patients (especially children with growth concerns or adults with insomnia), but expect rebound symptoms on the off days. Don't schedule high-stakes activities on drug-holiday days.

Not addiction physiology (in therapeutic-dose ADHD treatment), but real functional impact. Patients who have taken stimulants for years and then stop often describe weeks to months of subjective fatigue and cognitive dulling, some of which reflects unmasking of untreated ADHD.

For supratherapeutic use or misuse: withdrawal is more pronounced and can include dysphoria severe enough to raise suicide risk during the acute period.

Antipsychotic discontinuation

Withdrawal dyskinesia

New abnormal movements (orofacial, choreoathetoid, tongue) that appear during antipsychotic dose reduction and typically resolve over weeks to months. Reflects unmasking of D2 receptor upregulation. Not the same as tardive dyskinesia, though the mechanism overlaps and TD can be unmasked by taper.

If movements persist beyond 4 to 6 months after the taper completes, that's TD.

Cholinergic rebound

Most prominent with clozapine, less so with olanzapine, quetiapine at higher doses, and low-potency typicals (chlorpromazine, thioridazine). Symptoms: nausea, vomiting, diarrhea, sweating, hypersalivation, insomnia, agitation.

Management: taper slowly. If discontinuation is urgent (e.g., agranulocytosis on clozapine), consider anticholinergic cover (benztropine 1 to 2 mg BID) during and after the switch.

Rebound psychosis

Rapid discontinuation of high-D2-affinity antipsychotics (haloperidol, risperidone, olanzapine, paliperidone) can produce psychotic symptoms more severe than the patient's baseline, reflecting D2 receptor upregulation. Most reliably documented with typicals, but occurs with SGAs too.

Clinical implication: taper slowly (weeks to months for chronic users). If you must stop abruptly, expect a rough few weeks.

Supersensitivity psychosis

Contested but discussed concept: chronic dopamine blockade produces receptor upregulation manifesting as psychosis that's harder to treat than the original illness. Evidence is mixed. Whether or not you accept the specific label, the clinical implication is the same: slow tapers minimize risk.

Lithium discontinuation

Rebound mania

Lithium discontinuation, especially rapid discontinuation, carries a substantial risk of manic relapse in bipolar patients. Some data suggest post-discontinuation relapse rates exceed pre-treatment baseline rates (rebound above baseline).

Suominen and others have shown that abrupt or fast discontinuation produces markedly higher relapse rates than gradual discontinuation. The classic Baldessarini data: taper over less than 2 weeks produces the highest relapse rates; taper over 4+ weeks meaningfully reduces this.

Management

If lithium must be discontinued in a bipolar patient:

• Taper over at least 4 weeks, ideally 8 to 12 weeks.
• Have another mood stabilizer in place before starting the taper.
• Monitor closely for early manic signs during and for months after the taper.
• Warn the patient specifically about the elevated rebound risk.

If lithium is being stopped because of toxicity, you don't have that luxury, but the increased relapse risk still applies.

Other lithium discontinuation effects

Lithium withdrawal itself isn't dramatic (no seizure risk, no clear withdrawal syndrome), so the concerns are relapse-driven, not withdrawal-driven.

Other psychiatric medications

• Valproate: No withdrawal syndrome per se. Rebound mania possible in bipolar patients. Taper over weeks to months when discontinuing for maintenance indications.
• Lamotrigine: No withdrawal syndrome. Increased seizure risk if discontinued abruptly in epilepsy patients (taper over weeks).
• Gabapentin and pregabalin: Real discontinuation syndrome (anxiety, insomnia, sweating, GI symptoms) especially at high doses. Taper by 10 to 25% every 1 to 2 weeks.
• Prazosin: Rebound hypertension possible. Taper if used at high doses for extended periods.
• Clonidine: Rebound hypertension is a genuine risk. Taper over 1 to 2 weeks.

Consolidated table

Common questions

Are brain zaps really a thing? Yes. Patients describe them as brief electric-shock sensations, often in the head, sometimes with lateral eye movement. They're a common feature of SSRI/SNRI discontinuation, most reported with paroxetine and venlafaxine. Mechanism isn't fully worked out (theories involve serotonergic effects on inhibitory interneurons, possibly cerebellar). They're not dangerous, but they're distressing and disruptive. They resolve with reinstatement, slower taper, or time.

How do I taper alprazolam safely? Convert to a long-acting equivalent first (diazepam or clonazepam), then taper the long-acting agent. Direct alprazolam tapers are hard because the inter-dose withdrawal makes each dose reduction feel like withdrawal itself. See the benzodiazepine equivalents guide for the conversion approach. Reduce the equivalent dose by 5 to 10% every 2 to 4 weeks. For high-dose users, expect 6 to 18 months. Don't rush.

Why is fluoxetine easier to stop than paroxetine? Half-life. Fluoxetine has a half-life of 4 to 6 days, and its active metabolite norfluoxetine has a half-life of 7 to 15 days. When you stop, plasma levels fall gradually over weeks. The drug tapers itself. Paroxetine has a half-life of 21 hours, plus anticholinergic effects that create their own rebound. When you stop paroxetine, plasma levels crash within days, and the brain is left in an abrupt state change. Half-life is the dominant variable in discontinuation risk.

Can rebound psychosis happen after just a few weeks off an antipsychotic? Yes, though it's more commonly seen in the first 1 to 4 weeks after discontinuation, especially with high-D2 blockers and abrupt discontinuation. In chronic schizophrenia patients, relapse rates after discontinuation are meaningful. Meta-analyses of antipsychotic discontinuation trials show relapse rates of about 60 to 80% within 1 year off medication in patients who had been stable on it. Some of that is underlying illness returning; some is likely rebound above baseline. The two are hard to disentangle clinically.

What is a hyperbolic taper? A taper strategy where each reduction is a proportional amount of the current dose, not a fixed absolute amount. So instead of cutting sertraline by 25 mg each step (100, 75, 50, 25, 0), you cut by a percentage of the current dose (100 to 75 to 55 to 40 to 30 to 22, and so on). The absolute cuts get smaller as the dose gets smaller. This mirrors the sigmoidal receptor-occupancy curve: at low doses, small absolute changes correspond to large changes in receptor occupancy, so proportional cuts produce more even physiological transitions. Horowitz and Taylor made the mathematical case in the Lancet Psychiatry paper. In practice: use liquid formulations or compounded low-strength doses to hit the small increments at the tail end.

Sources

• Berber MJ. FINISH: remembering the discontinuation syndrome. J Clin Psychiatry. 1998;59(5):255.
• Horowitz MA, Taylor D. Tapering of SSRI treatment to mitigate withdrawal symptoms. Lancet Psychiatry. 2019;6(6):538-546.
• Horowitz MA, Taylor D. Distinguishing relapse from antidepressant withdrawal: clinical practice and antidepressant discontinuation studies. BJPsych Advances. 2022.
• Ashton CH. The Ashton Manual: Benzodiazepines. 2002, revised 2011.
• NICE NG215: Medicines associated with dependence or withdrawal symptoms: safe prescribing and withdrawal management.
• Chouinard G, Chouinard VA. Antipsychotic-induced dopamine supersensitivity psychosis. Psychother Psychosom.
• Baldessarini RJ, Tondo L, Viguera AC. Discontinuing lithium maintenance treatment in bipolar disorders. Bipolar Disord. 1999;1(1):17-24.
• Suominen K, Mantere O, Valtonen H, et al. Early recovery and relapse rates in bipolar disorder. J Clin Psychiatry.
• Maudsley Prescribing Guidelines in Psychiatry, 14th edition.
• Fava GA, Gatti A, Belaise C, Guidi J, Offidani E. Withdrawal symptoms after selective serotonin reuptake inhibitor discontinuation: a systematic review. Psychother Psychosom. 2015;84(2):72-81.

Reviewed against current guidelines as of June 8, 2026. This is not medical advice.