FDA Aligns With Cognition Therapeutics on Phase 3 Trial Design for Zervimesine in DLB-Associated Psychosis

The FDA has indicated that psychosis associated with dementia with Lewy bodies (DLB) could serve as an approvable outcome for a future new drug application (NDA), following a formal meeting with Cognition Therapeutics. The agency reached alignment with the company on key design elements of a registrational phase 3 trial of zervimesine, an investigational oral sigma-2 receptor modulator, with the pivotal program expected to launch in mid-2027.¹

“We reached an important agreement with the FDA that DLB psychosis could be an approvable outcome and that key aspects of our registrational trial design are appropriate and supportive of an NDA,” Anthony Caggiano, MD, PhD, chief medical officer at Cognition, said in a statement.¹ “To date, few drugs have been researched for DLB and none have been approved. The only recourse for DLB patients experiencing psychosis is the off-label use of potentially dangerous antipsychotics.”

Phase 3 Trial Design

As aligned with the FDA, the proposed phase 3 study will enroll adults with DLB who experience psychosis, defined as hallucinations and delusions. Patients currently receiving stable background treatment with off-label antipsychotic medications will be eligible to participate, a design that reflects real-world clinical practice. Following a screening period, participants will be randomly assigned to receive either zervimesine 100 mg once daily or placebo for 9 months.

The company proposed using the Neuropsychiatric Inventory (NPI), a validated instrument used to assess behavioral disturbances in dementia, as the primary end point. Cognition and the FDA have agreed to work collaboratively on the analytical and statistical framework for applying the NPI as a novel primary end point in the DLB psychosis indication. Further details of the end point strategy, including analytical modeling, are still under development.

Phase 2 SHIMMER Trial

The phase 3 program builds on the phase 2 SHIMMER trial (NCT05225415), a multicenter, double-blind, placebo-controlled study that enrolled 130 adults aged 50 to 85 years with probable mild to moderate DLB and Mini-Mental State Examination scores of 18 to 27.² In the study, participants were randomly assigned 1:1:1 to once-daily oral zervimesine 100 mg, 300 mg, or placebo for 26 weeks. The primary end point was safety and tolerability, which the trial met. Secondary measures encompassed the NPI, the Alzheimer’s Disease Cooperative Study–Activities of Daily Living (ADCS-ADL) scale, and the Unified Parkinson’s Disease Rating Scale Part III.

READ MORE: Donanemab Shows Durable Amyloid Clearance and Consistent Safety in Early-Start Patients

In the study, those in the zervimesine arms progressed more slowly than those on placebo on all exploratory efficacy assessments. Neuropsychiatric outcomes measured by the NPI-12 showed 86% slowing of decline relative to placebo (P = .0545) for the full NPI-12 scale. Core DLB symptoms that improved relative to placebo included hallucinations (P = .0207), anxiety (P = .0082), and delusions (P = .0609). NPI Caregiver Distress scores showed a 115% improvement vs placebo on the full NPI-12 scale (P = .0254), with significant between-group differences noted for hallucinations (P = .0048) and anxiety (P = .0174).

Presented at the 2026 International Conference on Alzheimer’s and Parkinson’s Diseases, held March 17 to 21, 2026, in Copenhagen, Denmark, additional findings showed that assessments using the Cognitive Drug Research (CDR) battery had statistically significant signals on immediate word recall (P = .0279), word recognition (P = .0081), and picture recognition (P = .0148). On the ADCS-ADL, results showed that zervimesine demonstrated 52% preservation of daily functioning relative to placebo (P = .0503).

A more recent post hoc analysis specifically focused on hallucinations and delusions reported an 89% slowing in progression of those symptoms with zervimesine vs placebo. The company also noted that treatment-related adverse events were predominantly mild or moderate in severity. Cognition Therapeutics noted that it plans to present these results at the 2026 Alzheimer's Association International Conference (AAIC), scheduled July 12 to 15, 2026, in London, England.¹

“Working with the FDA, we intend to find a path forward for a patient population that has waited too long,” Lisa Ricciardi, president and CEO, said in a statement.¹ “Our ultimate goal is to provide patients and their families with a durable treatment option for DLB psychosis that actually slows the progression of hallucinations and delusions.”

Mechanism of Action

Zervimesine, formally known as CT1812, is a small-molecule sigma-2 receptor modulator. Beyond DLB, zervimesine has been studied in other phase 2 studies including for mild to moderate AD in the SHINE study (NCT03507790) and for geographic atrophy secondary to dry age-related macular degeneration in the phase 2 MAGNIFY trial (NCT05893537). Furthermore, the agent is currently being held under evaluation in the ongoing phase 2 START study (NCT05531656) among patients living with mild cognitive impairment and early AD.

REFERENCES
1. Cognition Therapeutics aligns on key aspects of pivotal study for zervimesine (CT1812) in DLB psychosis following receipt of FDA meeting minutes. News release. Cognition Therapeutics. June 24, 2026. Accessed June 30, 2026. https://ir.cogrx.com/press_releases/cognition-therapeutics-aligns-on-key-aspects-of-pivotal-study-for-zervimesine-ct1812-in-dlb-psychosis-following-receipt-of-fda-meeting-minutes/
2. Galvin JE, Tolea MI, Scharre DW, et al. Phase 2 study of zervimesine (CT1812) in participants with mild-to-moderate dementia with Lewy bodies (DLB). Alzheimers Dement. 2025;21(12):e71004. doi:10.1002/alz.71004
3. Grundman M, Galvin J, Iaci J,Caggiano A. Zervimesine slows progression of symptoms in mild to moderate DLB. Alzheimer's Dement. 2025;21:e71004. Presented at: 2026 International Conference on Alzheimer’s and Parkinson’s Diseases, March 17-21, 2026; Copenhagen, Denmark. Abstract 2640.