A patient I’ve been managing for three years — late-seventies, moderate Alzheimer’s, history of alcohol use disorder in remission — started showing escalating agitation about eight months ago. Sundowning, verbal outbursts, occasional physical aggression toward staff at his memory care facility. The chart from his previous provider already listed two failed antipsychotic trials. He had developed mild tardive dyskinesia on the second one. His family was asking me, explicitly, for anything that wasn’t another antipsychotic. At that point, I didn’t have a clean answer. That gap has now started to close.

Axsome Therapeutics recently secured FDA approval for Auvelity as the first non-antipsychotic treatment for agitation associated with Alzheimer’s disease, and the company revised its peak sales projection for the drug to at least $8 billion — roughly split between this new indication and the original major depressive disorder approval it received in August 2022. The previous projection had ranged between $2.5 billion and $6 billion. That revision isn’t just a financial signal. It reflects how severely under-addressed behavioral symptoms in Alzheimer’s have been, and how much clinical demand was waiting behind a regulatory door that finally opened.

The Mechanism the Field Underestimated

Dextromethorphan/bupropion works through a combination of NMDA receptor antagonism and sigma-1 receptor agonism layered on top of dopamine-norepinephrine reuptake inhibition — a dual mechanism that doesn’t carry the dopamine D2 blockade responsible for tardive dyskinesia risk with conventional and atypical antipsychotics. That distinction matters enormously in a geriatric population where motor vulnerability is already elevated and cumulative antipsychotic exposure has often been long. The GEMINI Phase 3 trial, which ran between June 2019 and December 2021, established the efficacy foundation for the MDD indication — but the mechanism always suggested broader applicability to states characterized by pathological excitatory activity, which is exactly what you see in Alzheimer’s agitation.

The counterintuitive reality here is that the field spent decades treating Alzheimer’s agitation primarily with tools designed for psychosis. Agitation in dementia shares surface features with psychotic agitation — behavioral dyscontrol, disinhibition, apparent perceptual disturbance — but the underlying neurobiology is distinct. NMDA receptor dysfunction and glutamatergic dysregulation are central to Alzheimer’s pathophysiology, not incidental to it. A drug addressing that pathway directly should arguably have been a first-line behavioral candidate years earlier. Instead, the atypicals dominated by default, accumulating a black-box warning for mortality in elderly dementia patients and a well-documented tardive dyskinesia burden in a population least equipped to tolerate it.

The polypharmacy context makes this even sharper. A study drawing on Medicare data from 1.2 million people with dementia found that nearly 70% were prescribed at least one CNS-active medication, with 13.9% meeting criteria for CNS-active polypharmacy. Adding an antipsychotic to that substrate — in a patient already on a cholinesterase inhibitor, a benzodiazepine PRN, and an antihypertensive — is a clinical decision made under duress, not genuine therapeutic choice. Auvelity’s approval creates an option that doesn’t require that calculus.

What This Demands From the Next Wave of Trials

Here is where I want the design community to pay close attention. The approval of a non-antipsychotic for Alzheimer’s agitation exposes a structural gap in how behavioral symptom trials in geriatric populations have been constructed. Virtually every agitation trial in dementia has enrolled patients without rigorous characterization of their prior antipsychotic exposure, tardive dyskinesia status, or concomitant CNS medication burden. Those variables aren’t just safety covariates — they’re likely moderators of treatment response for any agent working through dopaminergic or glutamatergic pathways.

In my experience running trials in populations with dual diagnoses and complex medication histories, the assumption that you can adequately randomize around this heterogeneity with standard stratification fails in practice. A patient with three years of prior antipsychotic exposure has a different receptor landscape than an antipsychotic-naive patient with the same MMSE score. Protocols that treat them as interchangeable are not measuring what they think they’re measuring. The CMAI — the Cohen-Mansfield Agitation Inventory, the standard endpoint in most of these trials — captures frequency and severity of behavioral symptoms but says nothing about whether improvement reflects genuine mechanistic action or caregiver accommodation effects that inflate response in both arms.

Spanish-speaking patients with dementia, a population I see regularly across my telehealth practice, are almost categorically absent from the trials that generated this evidence base. CMAI validation in Spanish-speaking populations is limited. Consent procedures in most behavioral dementia trials have no bilingual infrastructure. The result is that an $8 billion drug’s efficacy data rests on a sample that excludes a demographic carrying disproportionate dementia burden — and future trials building on this approval face the same enrollment failure if nothing structural changes.

The next signal I’m watching: how Axsome’s commercial rollout in memory care settings interacts with existing antipsychotic prescribing patterns, and whether payer coverage decisions create a de facto formulary barrier that pushes physicians back toward older agents regardless of what the label now permits. If coverage lags the indication by 18 months — which is common — the clinical gap this approval was supposed to close stays open.

References

  1. FiercePharma — “After Alzheimer’s agitation nod, Axsome jacks up Auvelity’s peak sales projection to $8B”
  2. Drugs.com — Auvelity FDA Approval History (MDD, August 18, 2022)
  3. Practical Neurology — “First Non-Antipsychotic Treatment for Agitation Associated with Alzheimer Disease Dementia Approved by FDA”
  4. PubMed — GEMINI Phase 3 Trial: Dextromethorphan-Bupropion for Major Depressive Disorder
  5. National Institute on Aging — “The Dangers of Polypharmacy and the Case for Deprescribing in Older Adults”
+ posts

Dr. Leonardo Vando, MD is Principal Investigator of Psychiatry at Vando Medical Services.