Deprescribing antipsychotics is its own beast.

And it often takes more work than starting them.

This is a bonus post in my A-to-Z series from the upcoming 8th edition of the Med Fact Book.

We covered switching antipsychotics last week.

But what about stopping them?

I posted on this a few months ago, and the response made one thing clear:

There’s a hunger for better tools—and better language—around antipsychotic withdrawal.

Relapse rates are high—65–80% within 1–2 years off meds.

But some patients ask. And sometimes, we wonder if less is more.

So here’s how I approach discontinuation: 

▸ Taper ultra-slowly—5–10% every 1–3 months 

▸ Build a relapse prevention plan with clear warning signs 

▸ Educate about supersensitivity psychosis 

→ (Too many dopamine receptors, not enough blockers) 

▸ Watch for withdrawal dyskinesia—not just TD, but movements that emerge during the taper 

▸ For anticholinergic rebound (SLUD symptoms), taper clozapine/olanzapine slowly or cover briefly with benztropine

These cases take time. And nuance.

But the payoff is real when done thoughtfully—and collaboratively.

Not every return of symptoms is a relapse.

And not every smooth taper means the diagnosis was wrong.

Next up: Benzos—and why “as needed” isn’t as simple as it sounds.

▸ What’s your approach when a patient wants off their antipsychotic? 

▸ Share if you’ve navigated this tightrope—or want to help reimagine it 

▸ Follow me (Daniel Carlat, MD) for grounded psychopharmacology from the Med Fact Book

Join the conversation on LinkedIn with Dr. Carlat.