Antipsychotic Dose Reduction in Chronic, Stable Schizophrenia
Two long-stable patients both ask about lowering their antipsychotic dose. The real research question — who can actually tolerate this — plays out in opposite directions once their individual illness histories are weighed honestly.
D.R., a 47-year-old woman, has worked as a bookkeeper at the same small accounting firm for eight years, and her manager, who wrote a note of support for this visit at D.R.'s request, describes her as one of the most reliable people on staff. She has been stable on a moderate dose of risperidone for eleven years, with a single psychotic episode at illness onset followed by complete, sustained remission and no relapses since — a course her care team describes as unusually mild for a chronic schizophrenia diagnosis.
She has maintained steady employment throughout and asks directly about lowering her dose, citing bothersome mild EPS symptoms and a general desire to be on "the lowest dose that still works" now that she's been stable for over a decade. Dose reduction in chronic, stable schizophrenia is a genuine, actively studied question, not a settled one: some patients maintain stability at meaningfully lower doses than initially required, particularly those with a milder overall illness course, while others relapse even after years of stability once the dose drops below a threshold that varies unpredictably by patient. D.R.'s history — a single episode, complete remission, over a decade of unbroken stability, strong psychosocial function — places her among the more favorable candidates the research on this question has actually identified, though even in that favorable group relapse risk with reduction is real, not eliminated.
I'd support a careful, gradual dose reduction for her. Her illness course — a single episode, complete remission, over a decade of stability — is close to the profile the dose-reduction research has actually identified as more favorable, and her own EPS concern gives us a real clinical reason to try, not just a preference in isolation.
I'd agree, and I'd add a specific number to how we do it, not just that we do it. Leucht and colleagues' 2021 meta-analysis in JAMA Psychiatry found something like a protective end-dose threshold — above a certain point, roughly in the range of five milligrams risperidone-equivalent, relapse risk from a careful reduction looks comparable to staying at full maintenance dose. Below that threshold, risk rises meaningfully. That's a real caveat worth stating to her directly: even in a favorable-profile patient like her, relapse risk with reduction isn't eliminated by a good history, it's just better-positioned to stay low if we taper toward that threshold and stop, not past it. A gradual taper with close monitoring, targeting that range, is what the more favorable outcomes in this research are actually associated with.
Whatever threshold we're targeting, I'd still set explicit, specific early-warning criteria with her before we start — concrete signs she and her family should watch for, not vague "let us know if anything changes" — so a subtle prodromal change gets caught quickly rather than only after clear relapse, regardless of how favorable her statistical profile is.
Agreed: begin a slow, incremental risperidone dose reduction with explicit relapse early-warning criteria reviewed with D.R. and her family, and follow-up visits scheduled more frequently through the taper than her usual maintenance interval.
P.K., a 47-year-old man — matched to D.R.'s age for this comparison — runs a small landscaping business with his brother, work he says he's grateful to have been able to keep doing even through the harder years of his illness, thanks largely to his brother covering for him during the worst stretches. He has also been stable on a moderate antipsychotic dose, olanzapine in his case, for eleven years, and raises the same request as D.R.: lowering his dose, motivated by concern about long-term metabolic effects rather than EPS.
Unlike D.R., his illness course prior to reaching stability involved four psychotic episodes over three years, two requiring hospitalization, before his current regimen finally achieved sustained remission — a course his records describe as having taken considerably longer, and more attempts, to control. The pivot is not that his request is less reasonable or his metabolic concern less real — both are genuine — it's that the dose-reduction research pointing toward more favorable outcomes has consistently centered on patients with milder illness courses like D.R.'s, not patients whose stability was hard-won after multiple relapses. P.K.'s eleven years of stability are just as real as D.R.'s, but they were reached by a harder path, and what predicts tolerance of dose reduction in someone with his specific relapse history is far less established than it is for someone with D.R.'s.
I'm far more cautious here than I was with Case A. His stability took four relapses and two hospitalizations to achieve, and I don't think his eleven stable years since then tell us he'd tolerate a lower dose the way her eleven years do — a harder-won stability isn't automatically a more fragile one, but we don't have the evidence to assume it's equally robust either.
I'd agree that the same reasoning we used for Case A doesn't transfer cleanly, and I'd be specific about why. The dose-threshold data I cited for her comes from Leucht and colleagues' research on relapse prevention broadly. Højlund and colleagues' 2021 meta-analysis, specific to patients with his actual history — multiple episodes before stabilization — found low-dose reduction increased relapse risk by roughly forty percent compared with standard dosing, and a very low dose increased it further still. That's not the same evidence base D.R.'s plan rests on; it's a meaningfully starker risk in exactly his population. That said, his metabolic concern is genuine and not something to dismiss just because his illness history is harder — if we're not reducing the antipsychotic dose, we owe him a real answer to the metabolic problem some other way, not just a no.
If dose reduction isn't the right lever for his relapse risk, then his metabolic concern should be addressed directly — baseline labs, dietary counseling, consideration of a metabolic-focused adjunct — rather than leaving him with a declined request and no alternative plan for the actual problem he came in with.
Agreed: continue olanzapine at the current dose unchanged, and start active metabolic monitoring and counseling to address his stated concern through a different, more evidence-supported route.
P.K. was told directly why his situation was handled differently from what he may have heard about dose reduction being tried for other stable long-term patients — the team was explicit that this reflects his specific relapse history, not a dismissal of how genuinely stable he has been for over a decade.