Akathisia atlas

First hand

What worked for me

One person's account, not a trial and not a recommendation. This is the combination that made mine better.

Baclofen

A moderately high dose: 60 to 100 mg a day.

Guanfacine

1 to 5 mg a day, most likely 3 to 5 mg.

Lower limb rehabilitation

Ankle and lower limb rehabilitation work.

Spine and shoulders

Gentle spinal and shoulder movement rehabilitation.

Open the daily exercise tracker

How this sits against the evidence

Baclofen acts on GABA-B receptors. It is not in the akathisia guidelines or the 2024 meta-analysis, so this is an off-label, anecdotal use. It fits the GABA arm of the mechanism table below.
Guanfacine is an alpha-2 agonist, the same class as clonidine, which has a small open trial in akathisia. Guanfacine itself has not been trialled for it as far as this page's searches found.
Movement rehabilitation has no trials in akathisia. It sits alongside the patient reports that structured movement gives relief.

Both drugs are prescription only and need a prescriber. 60 to 100 mg of baclofen reaches or passes the usual labelled daily maximum of about 80 mg, and has to be built up gradually. Neither drug can be stopped suddenly: abrupt baclofen withdrawal can cause seizures and hallucinations, and abrupt guanfacine withdrawal can cause rebound high blood pressure. Together they add to sedation and can lower blood pressure and heart rate.

Orient

What akathisia is

Akathisia is a medication-induced movement disorder. It has two parts that must both be looked for: a subjective inner restlessness with a compulsion to move, and an objective pattern of restless movement, mostly in the legs (shifting weight, tramping on the spot, crossing and uncrossing the legs, pacing, being unable to stay seated).

It is classified with the extrapyramidal side effects of dopamine-blocking drugs. DSM-5-TR lists it under medication-induced movement disorders (medication-induced acute akathisia and tardive akathisia, both coded G25.71), which matters: it is a recognised diagnosis in its own right, not a variety of anxiety.

What it feels like

An urge to move that is hard or impossible to suppress
Tension or "crawling out of the skin" felt in the body, often the legs
Brief relief while moving, returning within seconds of stopping
Dread or dysphoria that follows the physical state, not a worry

What an observer sees

Rocking from foot to foot or walking on the spot when standing
Leg swinging, shuffling or repeated crossing when seated
Pacing, inability to sit through a meal or a consultation
Worse when asked to sit or lie still

Safety. Severe or prolonged akathisia is associated with self-harm and suicidal behaviour in clinical sources. If the restlessness comes with thoughts of ending your life, treat it as an emergency: in Australia call 000, or Lifeline on 13 11 14, and tell them it began with a medication.

Mechanism

How it works

The pathophysiology is not settled. The working model is an imbalance between dopamine and the systems that normally sit in tension with it. Each treatment class maps onto one arm of that model.

Pathway	Proposed role	Treatment that follows from it
Dopamine D2 blockade	Reduced dopamine signalling in mesocortical and ventral striatal circuits is the primary trigger. Risk rises with dose and with speed of dose increase.	Reduce, stop or switch the blocking drug
Noradrenaline	Compensatory noradrenergic overactivity from the locus coeruleus.	Propranolol, clonidine
Serotonin 5-HT2A	Serotonin restrains dopamine release. Blocking 5-HT2A releases that brake.	Low-dose mirtazapine, trazodone, mianserin, cyproheptadine
Acetylcholine	Dopamine and acetylcholine imbalance in the striatum, as in drug-induced parkinsonism.	Biperiden and other anticholinergics
GABA	Non-specific dampening of the restless drive.	Clonazepam, gabapentin, pregabalin
Iron, opioid, dopamine link	Shared ground with restless legs syndrome: low brain iron impairs dopamine function, and the endogenous opioid system modulates both.	Iron when ferritin is low, low-dose opioids (small trials only)

Olanzapine specifically. It has a lower akathisia rate than haloperidol, aripiprazole or lurasidone, which is one reason it gets missed, but case reports describe severe akathisia on it, including in people who had never taken an antipsychotic before. In one reported case it resolved on stopping olanzapine and came back when the drug was restarted, which is the strongest form of causal evidence available for a single patient.

Timeline

Types and timing

Hours to weeksAcuteStarts soon after a drug is begun or the dose is raised. Usually within the first 6 weeks.

After stopping or cuttingWithdrawalEmerges within about 2 weeks of a dose cut or stop. Typically settles within about 6 weeks.

Delayed or persistingTardiveOnset after months of exposure, or persistence after the drug is gone. Can last months or years.

By durationChronicAny akathisia lasting beyond 3 months (some sources say 6).

The cut-offs differ between sources, so quote the definition with its source when it matters. The practical reading for symptoms that are 5 to 6 weeks old:

If the causing drug is still being taken, this is acute akathisia that has not been dealt with, and the first-line step (reduce, stop or switch) has not yet been done.
If the drug has already been stopped, acute akathisia usually eases over days to a few weeks and withdrawal akathisia is expected to settle by about week 6. Symptoms continuing at that point are leaving the "wait and see" window and heading toward the persistent category, where outcomes are worse.

That second point is the hinge of the whole argument: "it will settle once the drug is out" is a time-limited prediction. When the time has passed and it has not settled, the plan needs to change.

Differential diagnosis

Akathisia versus anxiety

They overlap enough that no single feature is decisive, and they can coexist. The discrimination comes from the pattern across several features, anchored to the timeline of drug exposure.

Feature	Akathisia	Anxiety
Core experience	A physical urge or compulsion to move	Worry, apprehension, fear with content ("what if")
Where it is felt	In the body, most often the legs	Chest, stomach, throat, head, or diffuse
Effect of moving	Brief relief, returns on stopping	Little or no specific relief
Effect of rest	Worse when made to sit or lie still	Often better with rest, reassurance or distraction
Thought content	Little rumination. Distress is about the restlessness itself	Prominent rumination and preoccupation
Movements	Stereotyped, repetitive, lower limb: rocking, tramping, pacing	Non-specific fidgeting, hand wringing, tremor
Autonomic signs	Usually not prominent	Palpitations, sweating, breathlessness, especially in panic
Onset	Tied to starting, raising, cutting or stopping a drug	Tied to stressors, or longstanding before the drug
Past history	A new state the person has never had before	Recognisable as their usual anxiety
Response to more antipsychotic or an SSRI	Worse	May improve

Other conditions to exclude

Mimic	How it differs
Restless legs syndrome	Strong evening and night pattern, limited to the legs, disturbs sleep onset, often an unpleasant sensation in the limb. Akathisia is present through the day and is a whole-body restlessness. Check ferritin for both.
Agitated depression or psychotic agitation	Driven by mood or psychotic content, not relieved by moving, not leg-predominant.
Tardive dyskinesia	Involuntary movements (mouth, tongue, fingers) without the urge to move. Rated with the AIMS. Can coexist with akathisia.
Substance or drug withdrawal	Alcohol, opioid, benzodiazepine or antidepressant withdrawal each have their own timeline and autonomic signs.
Serotonin syndrome	Tremor, clonus, fever, sweating, diarrhoea after a serotonergic drug change.
Medical causes	Thyroid overactivity, low blood sugar, iron deficiency, renal failure, excess caffeine or stimulants.

Instruments

Every rating scale that touches the question

Akathisia-specific scales measure the condition. General side-effect scales screen for it. Anxiety and restless-legs scales are run alongside to show what the picture is not.

Akathisia-specific

Scale	What it rates	Scoring	Use
Barnes Akathisia Rating Scale (BARS), 1989	Objective movements, subjective awareness of restlessness, distress, and a global clinical assessment. Observed seated then standing, at least 2 minutes each, during neutral conversation.	Three items 0 to 3, global 0 to 5. Global 2 or more is generally taken as akathisia present. Inter-rater reliability about 0.82 to 0.90.	The standard in trials and clinics. This is the one to ask for.
Prince Henry Hospital Akathisia Rating Scale (Sachdev), 1994	Separate objective items for sitting and standing, plus subjective items, plus a global rating.	Items 0 to 3. Validated against the BARS in 50 patients.	More granular than the BARS. Useful for serial tracking.
Hillside Akathisia Scale (Fleischhacker), 1989	2 subjective items and 3 objective items (head and trunk, hands and arms, feet and legs), rated sitting, standing and lying.	Items 0 to 4, plus a clinical global impression. Reliability 0.86 to 0.92 subjective, 0.51 to 0.89 objective.	Trials. The lying-down rating helps separate it from restless legs.
Braude criteria, 1983	Early operational criteria, with the inability to keep the legs still as the defining complaint.	Criteria, not a score.	Historical. Underlies later scales.
Sachdev research diagnostic criteria, 1994	Formal criteria for acute, tardive, withdrawal and chronic akathisia.	Criteria, not a score.	Subtyping and research definitions.
DSM-5-TR medication-induced movement disorders	Subjective restlessness with observed movement, after exposure to a causative drug.	Diagnostic category.	The formal diagnosis to be recorded.

General side-effect scales with an akathisia item

Scale	Akathisia content	Note
Extrapyramidal Symptom Rating Scale (ESRS, Chouinard)	A subjective questionnaire item and an objective akathisia item, each rated on a wide range, plus global impressions.	Covers parkinsonism, dystonia and dyskinesia in the same sitting.
St Hans Rating Scale	Separate psychic and motor akathisia ratings.	Splits subjective from objective, which is the distinction that matters here.
UKU Side Effect Rating Scale	One clinician-rated akathisia item, 0 to 3.	Broad side-effect interview.
LUNSERS	Self-rated restlessness within a 51-item side-effect checklist.	Self-report, so it can be completed before an appointment.
Simpson-Angus Scale (modified versions)	Mainly parkinsonism. Some versions add an akathisia item.	A normal score does not rule akathisia out.
AIMS	None. It rates dyskinesia.	Used to exclude tardive dyskinesia, not to detect akathisia.

Comparison scales run alongside

Scale	Why include it	Trap
Hamilton Anxiety Rating Scale (HAM-A)	Splits psychic anxiety (worry, fear, tension) from somatic anxiety. Akathisia tends to load on restlessness with low psychic scores.	The tension item includes restlessness, so the total alone misleads.
GAD-7	Quick self-report of worry-based anxiety.	Item 5 is "so restless it is hard to sit still". Akathisia can score 3 on that item while the worry items stay low. Report item by item.
Beck Anxiety Inventory	Somatic anxiety symptoms.	"Unable to relax" and "nervous" overlap with akathisia.
IRLS and the restless legs diagnostic criteria	Tests the evening-only, sleep-linked pattern of restless legs.	Both conditions share the urge to move.
Naranjo adverse drug reaction scale	Scores how probable it is that a drug caused an effect (timing, dechallenge, rechallenge, alternatives).	It rates causality, not severity.

Objective measurement. Leg-worn actigraphy (actometry) has been compared with the BARS. The BARS was better at screening for the diagnosis, and actigraphy was judged a useful objective add-on for severity. So movement data supports a clinical rating but does not replace it.

Method

The most rigorous way to tell them apart

No blood test or scan confirms akathisia. Diagnosis is clinical, so rigour comes from combining independent lines of evidence that each point the same way.

Build the exposure timeline

One page: every drug, with start dates, dose changes and stop dates, plotted against the day the restlessness began and how it changed. A new state appearing within days to weeks of a dopamine-blocking drug is the single strongest piece of evidence. Score it with the Naranjo scale.

Take the subjective history with the right questions

Is there an urge to move? Where is it felt? What happens when you move, and when you stop? What happens when you sit or lie still? Is there a worry driving it, or is the distress about the restlessness? Have you ever felt this before the drug?

Have a clinician do a formal BARS

Seated at least 2 minutes, standing at least 2 minutes, scored and written in the notes. Repeat at each visit at a similar time of day. A second rater or a video strengthens it.

Run anxiety scales in parallel and read them item by item

HAM-A or GAD-7. A profile of high restlessness with low worry, low fear and few autonomic symptoms argues against primary anxiety. High scores on both mean both may be present, which still requires the akathisia to be treated.

Exclude the mimics

Ferritin and iron studies, thyroid function, glucose, kidney function, B12 and magnesium. Caffeine and stimulant intake. Alcohol, opioid and benzodiazepine history. AIMS for dyskinesia. Restless legs criteria for the night-only pattern.

Add objective movement data

A short phone video of 2 minutes seated and 2 minutes standing on a bad day, and step counts or wearable movement data across the day. Supporting evidence, not proof.

Audit every other medication

Work through the list below. A second causative or perpetuating drug explains persistence after the first was removed.

Use treatment response as evidence

A documented fall in the BARS after a specific anti-akathisia drug supports the diagnosis. It is weaker with propranolol and benzodiazepines, which also ease anxiety, and stronger with low-dose mirtazapine, vitamin B6 or an anticholinergic, which are not anxiety treatments at those doses.

Dechallenge and rechallenge (it stops when the drug stops and returns when it restarts) is the strongest causal evidence, but nobody should be re-exposed on purpose to prove a point. If it has already happened, record it.

Treatment

Evidence-based treatments

Ranked by strength of evidence. The best pooled data is a 2024 network meta-analysis of 15 double-blind trials with 492 patients, all in antipsychotic-induced akathisia. Every trial is small, so treat the order as provisional.

Treatment	Evidence	Typical trial dose	Notes
Reduce, stop or switch the causing drug	Guideline first step	Dose cut of 25 to 50%, or a lower-risk agent	2018 Canadian guideline. Also stop antipsychotic polypharmacy. Taper speed matters: abrupt changes can provoke withdrawal akathisia.
Mirtazapine, low dose	Top tier	15 mg at night	Most effective in the 2024 analysis but least well tolerated (sedation, appetite). Higher doses lose the benefit and can themselves cause restlessness.
Vitamin B6 (pyridoxine)	Top tier	600 to 1200 mg a day, short term	Best balance of efficacy and tolerability, but from 33 patients. Prolonged high doses cause nerve damage, so this is a time-limited, supervised trial.
Biperiden (anticholinergic)	Top tier	Oral or injected	Most useful when parkinsonism is also present. Not marketed in the US. Benztropine is the usual substitute, with weaker evidence.
Propranolol	Second tier	20 to 80 mg a day in divided doses	The traditional first add-on. Avoid in asthma, slow heart rate, low blood pressure.
Trazodone, mianserin	Second tier	Trazodone 50 to 100 mg	Same 5-HT2A mechanism as mirtazapine. Mianserin is not available in the US.
Cyproheptadine	Second tier, fragile	About 16 mg a day	Result did not hold in sensitivity analysis.
Clonazepam and other benzodiazepines	Two small trials	0.5 to 1 mg	Short term only because of dependence. Stopping them can itself produce restlessness.
Zolmitriptan	One trial	7.5 mg a day	Comparable to propranolol in one double-blind study.
Clonidine	Open trial, 6 patients	0.2 to 0.8 mg a day	Limited by low blood pressure and sedation.
Amantadine, gabapentin, pregabalin	Insufficient	Varies	Case-level support. Gabapentinoids appear on some cause lists too.

Persistent and tardive akathisia has no approved therapy and no substantial trials. One review cites under 5% full resolution, about 20% improved and about 75% with little benefit. That poor outlook is the reason to act while it is still early.

Beyond the guidelines

Reported to help, not medically established

These come from small studies, case reports and patient communities. Reddit and the largest patient forums could not be searched directly for this page, so the list rests on peer-support sites and the studies behind the common anecdotes.

Remedy	What backs it	Caution
Low-dose opioids	A 1986 trial in 5 patients: codeine and propoxyphene beat placebo, and naloxone brought symptoms back in one patient. Seem to work better for acute than tardive. Mentioned in recent reviews but in no guideline.	Dependence. Kratom reports probably share this mechanism and this risk.
Very slow taper, then time off the drug	The Akathisia Alliance reports this as the most effective long-term approach across thousands of support group members.	Fast tapers can sharply worsen it.
Iron	One study found lower ferritin and an inverse link between iron and severity. A later study found only a minor role. One case of intravenous iron helping in confirmed deficiency.	Only when ferritin is actually low. Test first.
Magnesium	Widely reported as partial relief in peer communities. No trials found.	Some people in withdrawal report supplements making things worse.
Nicotine patch	One single-blind study of 16 non-smokers showed less akathisia on patch days.	Small and never replicated. Smoking changes olanzapine levels (see the audit).
N-acetylcysteine	Akathisia improved as a secondary finding in a 24-week schizophrenia trial.	Unconfirmed.
Fluvoxamine	A case series of 5 patients.	SSRIs also cause akathisia, and fluvoxamine raises olanzapine levels.
Electroconvulsive therapy	Case reports of chronic and tardive akathisia resolving, including one caused by escitalopram plus olanzapine that had not responded to standard drugs.	Case-level evidence, last resort.
Exhaustive movement	Walking to exhaustion, pacing, rocking and a punching bag are commonly reported as temporary relief.	Relieves, does not treat.
Structured relaxation	A small case series in which all but one patient benefited.	Very limited.
Cutting caffeine, alcohol and sugar	Standard advice on patient sites.	No trials, low risk.
Cannabis and CBD	Often mentioned. No akathisia-specific studies found.	Unknown either way.

The case

From "it is anxiety" to documented akathisia

The goal is not to win an argument. It is to put evidence in the notes that any clinician reading them later would have to deal with. Each step produces a document.

Write the one-page timeline

Olanzapine start date, dose, any changes, stop date if stopped. Date the restlessness began. Whether this state ever existed before. Everything else taken in the same period.

Describe it in akathisia terms

Use the discriminating features: urge to move, felt in the legs, relief on moving, worse at rest, no driving worry, new since the drug. Clinicians diagnose from the patient's account of the inner sensation, so the words chosen matter.

Ask for a formal BARS and for the score to be recorded

A specific, reasonable request. A recorded global score of 2 or more is a diagnosis in the notes. A recorded score also gives a baseline to measure any treatment against.

Hand over the item-level anxiety scale

A completed GAD-7 or HAM-A showing high restlessness with low worry is direct evidence against "just anxiety". If the worry items are high too, say so: both can be true, and the akathisia still needs treating.

Bring the video and the movement data

Two minutes seated and two minutes standing, plus daily step counts if they have jumped since the drug.

Request the exclusion bloods

Ferritin and iron studies, thyroid function, glucose, kidney function, B12, magnesium. Normal results remove the alternatives. A low ferritin is treatable.

Present the completed medication audit

Either a second contributing drug is found, or the list is clean and the persistence cannot be blamed on something else.

Ask for the diagnosis and the plan in writing

"Medication-induced akathisia" recorded, a named treatment trial with a review date, and a repeat BARS at that review. If the answer is still no, ask for the reasoning to be documented, then seek a second opinion from another psychiatrist or a movement disorder neurologist.

The argument

Why "just remove the olanzapine" is not enough

The waiting period has a limit

Acute akathisia usually eases within days to weeks of removing the cause, and withdrawal akathisia is expected to settle by about 6 weeks. Symptoms still present at 5 to 6 weeks are at or past the point where "it will pass" is a reasonable plan.

Early is treatable, late is not

Acute akathisia responds to several drugs. Persistent and tardive akathisia has no established treatment and a poor reported outcome. Delay moves a person from the first group toward the second.

The guideline lists more than removal

The 2018 Canadian guideline and the 2024 meta-analysis both set out add-on options: mirtazapine 15 mg, short-term vitamin B6, propranolol, an anticholinergic. Offering none of them is a choice, not a default.

The options are low risk compared with the condition

A time-limited trial of B6, propranolol or low-dose mirtazapine carries modest, well-known risks. Untreated akathisia carries severe distress and a documented association with self-harm.

The wrong label leads to the wrong drugs

Treating it as anxiety invites an SSRI, an SNRI or a "calming" antipsychotic, all of which are on the cause list and can make it worse.

A treatment trial is also a diagnostic test

If the BARS falls after a specific anti-akathisia drug, that confirms the diagnosis. If nothing changes after adequate trials, that is useful too and redirects the search. Either way the uncertainty shrinks.

A reasonable sequence to propose

Confirm the causing drug is fully out and nothing on the audit list is still in play.
Correct anything found on bloods, especially low ferritin.
First trial: low-dose mirtazapine (15 mg) or short-term vitamin B6, or propranolol if those do not suit. One at a time, with a BARS before and at 1 to 2 weeks.
If no response, switch class instead of raising the dose: 5-HT2A blocker, beta-blocker, anticholinergic, then a short benzodiazepine course.
If still present at 3 months, ask for a movement disorder neurology opinion and discuss the less established options with them.

This is a framework to take to a prescriber, not a prescription. Doses are those used in trials and are listed so the conversation can be specific.

Scratchpad

Notes for the appointment

Timeline, questions, what was said. Saved in this browser only, never sent to the server.

Reference

Limits of this page

Built from web search summaries on 14 November 2026, not from full-text reading of each paper.
The treatment trials are in antipsychotic-induced acute akathisia. Evidence for withdrawal and persistent akathisia is far thinner.
Exact item wording for the Prince Henry and Hillside scales was not retrieved. The ESRS, St Hans, UKU, LUNSERS and Simpson-Angus descriptions are from general knowledge.
The DSM-5-TR text was not retrieved directly. Timing cut-offs differ between sources.
No reliable figure was found for how often akathisia is missed, and the suicidality link rests on clinical commentary more than controlled studies.
Patient forums could not be searched directly.
This is general information, not medical advice about any individual.