What’s Actually First-Line for Antipsychotic-Induced Akathisia?
What's Actually First-Line for Antipsychotic-Induced Akathisia — and Why Isn't It Propranolol Anymore?
Your patient is more restless a week after you raised the antipsychotic. The reflex answer — read it as worsening agitation and raise the dose again — is the single move that turns a treatable side effect into a spiral. And when clinicians do recognise it as akathisia, most reach for propranolol, which the current evidence has quietly overtaken. Here is the short version of what to do, and the step almost everyone skips before prescribing anything at all.
The 30-second version
- Work the drug before you add one. Dose is often the single strongest predictor after adjustment — reduce it, stop antipsychotic polypharmacy, or switch to a lower-liability agent first.
- Three add-ons sit at the top with comparable efficacy — vitamin B6, mirtazapine and biperiden. Their confidence intervals overlap, so treat the order as soft and choose by what else the patient has.
- Vitamin B6 600–1200 mg is the best risk-benefit default — that is the trial authors' own bottom line, on tolerability and acceptability.
- Mirtazapine 15 mg carries the largest point estimate and is the pick when depression or insomnia coexist; biperiden when drug-induced parkinsonism does.
- Propranolol still works, but ranks behind all three — and roughly 17% discontinued it for hypotension or bradycardia in a head-to-head trial.
- Benzodiazepine evidence genuinely conflicts — three analyses, three different answers. A bridge for distress at most, not a treatment.
- Hold all of it loosely. The add-on evidence comes from small trials at low confidence. The dose is the part you can be confident about.
So what should I reach for first?
Not a drug — the dose. Reduce the offending antipsychotic, stop any antipsychotic polypharmacy, then consider switching agents, before you add anything.
Akathisia is dose-related, and after adjusting for confounders, antipsychotic dose is often the single strongest predictor of whether a patient develops it. Some agents keep climbing with dose rather than plateauing.
That makes the first three moves pharmacological housekeeping rather than prescribing, and the guideline sequence is deliberate: lower the dose, stop polypharmacy, switch to a lower-liability agent — and only then reach for an adjuvant. This is the lever most often skipped, which is why so many patients end up on an antipsychotic plus an add-on for a problem the dose was driving.
Which add-on should I actually pick?
Vitamin B6, mirtazapine or biperiden — they perform comparably, so choose by what else the patient has. All three rank ahead of propranolol.
A 2024 network meta-analysis in JAMA Network Open ranked the available options, and the order is not the one most of us trained on. But read the ranking carefully: the top three have overlapping confidence intervals, and the authors describe them as comparable rather than sequential. Mirtazapine carries the largest point estimate (SMD −1.20), ahead of biperiden (−1.01) and vitamin B6 (−0.92) — yet the authors' own bottom line elevates B6 as the best combined efficacy-and-tolerability option.
So the demotion of propranolol is firm. A clean 1-2-3 among the three above it is not. In practice that makes this a selection problem rather than a ranking one.
| Agent | Where the evidence puts it | Practical dose | Verdict |
|---|---|---|---|
| Vitamin B6 | Comparable efficacy (SMD −0.92); best tolerability and acceptability of the three | 600–1200 mg/day | Best risk-benefit default |
| Mirtazapine | Largest point estimate (SMD −1.20); better tolerated than propranolol head-to-head | 15 mg at night | Pick if depression or insomnia coexist |
| Biperiden | Comparable (SMD −1.01), but from a small number of trials | Best case when parkinsonism coexists, or after the two above fail | Best alternative |
| Propranolol | Effective, but ranks behind all three above | 20 mg BID, titrated across ~30–120 mg/day | Alternative |
| Benzodiazepine | Three analyses, three answers — null, possible benefit at very low confidence, and top-ranked | Short term only | Bridge, not treatment |
Ordered by practical priority, following the source authors. By point estimate alone the order is mirtazapine → biperiden → B6 — but the intervals overlap.
Isn't biperiden an anticholinergic — and don't those fail in akathisia?
Yes, it is one. The class does perform poorly for isolated akathisia — biperiden specifically separated from the others in this analysis, on a small number of trials.
The teaching most of us carry is well supported: anticholinergics as a group are not effective for isolated akathisia, and guidelines say so. Biperiden is an anticholinergic too, so the honest version isn't "the class fails and this one works" — it's that the class performs poorly overall while this agent showed a significant effect in one network meta-analysis, drawn from few trials with limited sample sizes.
That's enough to keep it on the list, not enough to call it a settled exception. Its strongest practical case is the patient who has drug-induced parkinsonism sitting alongside the akathisia — one agent, two problems. The reasons to think twice are cognitive burden and constipation.
Then why does vitamin B6 get recommended so often?
Because efficacy across the top three is comparable, and B6 has the best tolerability — which is exactly what the source authors conclude.
B6 carries the smallest of the three point estimates but the best tolerability and acceptability profile, and with the efficacy differences inside the noise, that combination is what makes it the sensible default. It is why B6 gets elevated from "adjunct" to a primary choice in practice.
What about benzodiazepines?
Three network meta-analyses give three different answers — which is itself the finding.
In the 2024 JAMA Network Open analysis, clonazepam failed to separate from placebo. A 2026 analysis found benzodiazepines might be beneficial, with an effect size that looks large but confidence rated very low. And a separate 2024 analysis in CNS Spectrums ranked benzodiazepines highest of all agents, ahead of the beta-blockers and mirtazapine.
That spread is not a reason to reach for one. It is a reason to distrust any confident claim in either direction.
The practical position: a short-term bridge for a distressed patient, not a treatment. Don't build the plan around it, and don't leave a patient on one because it seemed to help in week one.
Is there a catch?
Two. Late-onset restlessness may be tardive akathisia — where reducing the dose backfires. And akathisia is a safety problem, not a comfort problem.
Acute and tardive akathisia look identical but respond in opposite directions to the same move. Acute akathisia improves when you reduce or stop the antipsychotic; tardive akathisia can worsen. So for restlessness that appears months into stable treatment rather than weeks after a change, the reflexive dose reduction that helps in the acute case is the wrong instinct.
The second catch is the one that changes urgency. Ask directly about dysphoria and suicidal thoughts. A causal link to suicidality is not established — the systematic review evidence is small, low-quality and inconclusive — but it is plausible and concerning enough that proactive screening is the right call. Treat akathisia as a safety issue.
And before any of it: anchor the diagnosis with a number. The Barnes Akathisia Rating Scale takes about two minutes and rates objective movement, the patient's own awareness of restlessness, the distress it causes, and your global judgment. Re-score after each intervention — a falling Barnes is how you know the ladder is working.
How much weight should I put on this ranking at all?
Less than the decimal places suggest. Hold the order loosely and the dose principle firmly.
Every add-on recommendation here rests on small trials at low to very low confidence — the 2024 network meta-analysis pooled 492 patients, the 2026 one 661. And no randomised trial has tested dose reduction or switching head-to-head against any adjuvant, so the sequence itself is guideline consensus rather than trial evidence.
None of that argues for doing nothing. It argues for where to place your confidence: firmly on optimising the antipsychotic, loosely on which of the three add-ons you reach for after that.
Adverse Events A Rapid-Decision Series
Course 1: Managing Antipsychotic Adverse Events — table of contents
Nineteen chapters, releasing in three parts. Part 1 is live now, and the akathisia chapter is free to read — no membership needed.
- Acute Dystonia Members
- Akathisia Free · this chapter
- Drug-Induced Parkinsonism Members
- Tardive Dyskinesia Members
- Neuroleptic Malignant Syndrome Members
- Weight Gain Soon
- Diabetes & Glucose Dysregulation Soon
- Dyslipidemia Soon
- Hyperprolactinemia Soon
- QTc Prolongation Soon
- Orthostatic Hypotension Soon
- Agranulocytosis Soon
- Myocarditis Soon
- Constipation & Ileus Soon
- Sialorrhea Soon
- Sedation Soon
- Sexual Dysfunction Soon
- Antipsychotic-Induced Edema Soon
- Quick-Reference Grid Soon
Part of Managing Psychopharmacology Adverse Events — A Rapid-Decision Series. Companion courses on antidepressants, mood stabilizers, anxiolytics & hypnotics, and stimulants follow.
Want every chapter?
Full access to all 19 chapters of this course, plus 500+ clinically relevant chapters across the Academy.
View Membership Options →
This is a pretty high dose of B6 — if used long term at this dose, can increase risk of peripheral neuropathy (has been seen after even 100mg daily) — this would seem to me a reason why putting it lower on the algorithm. What am I msising?
You’re not missing anything — you’ve identified a genuine limitation in how that recommendation is usually presented, including in my post.
The tolerability argument for B6 comes from the akathisia RCTs, and those trials are short: days to a few weeks. “Best tolerated” in that context means best tolerated over a brief course, which is not the same claim as safe at 600–1200 mg/day sustained. Pyridoxine sensory neuropathy is well described with chronic use, and reported at doses well below the ones trialled here — which is why several regulators have restricted lower-dose supplements. So the trial evidence and the long-term safety question are answering different things, and the ranking quietly imports the first into the second.
How I’d square it in practice: treat B6 as a short course, not maintenance. Review at around four weeks, stop if there’s no benefit, and counsel the patient to report paresthesia or numbness. If you find yourself needing months of cover, that’s a signal the antipsychotic dose or agent is the thing to fix — or that mirtazapine or biperiden is the better fit for that patient.
Full chapter, with the ranking and its limits: https://psychiatryeducationforum.com/lesson/akathisia/