KarXT for Alzheimer’s psychosis (ADEPT-4)

KarXT (xanomeline-trospium) has mechanistic rationale for psychosis via M1/M4 agonism, and ADEPT-2 in AD psychosis read out positive. However, AD psychosis trials historically fail (pimavanserin mixed; many antipsychotics negative), NPI-C H+D is noisy with high placebo response in dementia, and tolerability in elderly AD patients is a risk. Primary completion Dec 2026 with readout likely 2027; execution risk and base rate for CNS Ph3 success (~40-50%) weigh against. Slight lean negative.

Phase 3 and a focused 14-week psychosis endpoint support modest success odds, but Alzheimer’s severity heterogeneity, subjective clinician scoring, placebo response, and recruitment still ongoing only 86 days before estimated completion create substantial efficacy and execution uncertainty.

Phase 3 placebo-controlled design and a focused clinician-rated psychosis endpoint support interpretability, but Alzheimer’s patients with moderate-to-severe psychosis are heterogeneous and difficult to treat. A 14-week change endpoint can be placebo-sensitive, and recruiting status only 86 days before estimated completion adds execution and disclosure risk.

Phase 3 drug-versus-placebo design and a clinically targeted NPI-C hallucinations/delusions endpoint support a modestly favorable prior. However, the mild-to-severe Alzheimer population is heterogeneous, the clinician-rated endpoint can be noisy, and the fields provide no sample size, blinding, effect-size history, or operational detail. Recruiting near completion also leaves execution risk, so confidence is limited.

Phase 3 recruiting trial; standard 14-week NPI-C H+D endpoint in AD psychosis population; muscarinic agonist with prior positive schizophrenia data; no design or operational issues noted in fields.

KarXT (xanomeline-trospium) has strong biological plausibility for Alzheimer's psychosis. Historical Phase 2 data for xanomeline alone in the 1990s demonstrated significant reductions in psychotic symptoms in AD patients, but was limited by peripheral side effects which trospium now mitigates. The primary risk is the high placebo response rate typical in AD psychosis trials using the NPI-C scale.

The trial is Phase 3 and still recruiting, so no efficacy data exist. Alzheimer’s psychosis endpoints are historically difficult, and KarXT’s novel mechanism lacks prior confirmatory results. Although the sponsor is strong, early‑stage operational risk and limited prior data keep intrinsic success modest, around 25%.

ADEPT-4 is the third Phase 3 in KarXT's AD psychosis program after ADEPT-2 missed its primary endpoint in late 2024 and ADEPT-3 also failed to meet significance, sharply reducing conditional success odds. NPI-C: H+D is sensitive and showed Phase 2 effect, but AD psychosis trials suffer high placebo response and enrollment heterogeneity. Still recruiting with 86 days to primary completion adds operational risk. Mechanism validated in schizophrenia (EMERGENT), but translation to AD psychosis remains unproven at Phase 3. Prior program failures dominate the prior.