Millions of older adults with dementia are prescribed antipsychotics well beyond recommended treatment windows, often because clinicians and families fear what stopping them might cause. This study directly addresses that fear — and its findings should meaningfully shift how long-term care decisions are made for one of medicine's most vulnerable populations.
Using a target trial emulation framework applied to real-world data, investigators compared outcomes among dementia patients who discontinued antipsychotics — either abruptly or through gradual tapering — against those who continued treatment beyond guideline-recommended durations. Both discontinuation strategies reduced the risk of delirium and fracture. Critically, neither abrupt cessation nor tapering increased mortality, stroke risk, or pneumonia incidence, countering the prevailing clinical anxiety that stopping these medications could precipitate dangerous neuropsychiatric or physical deterioration.
This finding carries significant weight in the broader context of antipsychotic prescribing in dementia care. Antipsychotics such as haloperidol, risperidone, and quetiapine carry FDA black-box warnings for use in elderly dementia patients due to elevated cardiovascular and mortality risks associated with ongoing use — yet deprescribing rates remain low. The emulation methodology used here is noteworthy: by mimicking a randomized trial design within observational data, it partially addresses the confounding that plagues routine prescribing studies, though residual confounding cannot be fully excluded. The study also does not capture granular data on dementia severity or the specific behavioral symptoms that originally prompted prescribing, which limits generalizability to the most complex cases. Still, this is among the more methodologically rigorous real-world analyses to date on antipsychotic discontinuation in this cohort. For clinicians and caregivers navigating deprescribing conversations, the evidence now leans clearly toward attempting discontinuation when treatment has already exceeded recommended timelines — a genuinely actionable and potentially paradigm-shifting conclusion.
