The first door opened in Arizona. In June 2025, Cognition Therapeutics announced that an anonymous gift from the family of a person who had been treated with zervimesine in its Phase 2 SHIMMER study would substantially fund an expanded access programme in dementia with Lewy bodies, and that Banner Sun Health Research Institute was the first of eight sites to be activated, with David Shprecher as its principal investigator. That is the whole shape of the thing in one sentence: no approval, no reimbursement, no market. One family paid, and a small number of other families got a place. For DLB, where there is still no approved disease-modifying treatment anywhere, this is what access looks like.
The public record is unusually plain about the mechanics. The ClinicalTrials.gov entry NCT06961760, sponsor protocol COG1202, first posted on 8 May 2025 and last updated on 28 May 2026, describes a single-arm, open-label expanded access programme providing 100 mg of oral zervimesine and collecting long-term safety and efficacy data in mild-to-moderate DLB. Screening runs up to seven days, treatment for 360 days give or take four, with a follow-up visit at day 390, a maximum of 397 days in all. There are two doors in: you were randomised in the Phase 2 SHIMMER protocol (COG1201) and completed it, or your treating physician referred you to a participating site with a confirmed DLB diagnosis, an MRI or CT scan from the past three months, a Mini-Mental State Examination score between 18 and 27 inclusive, and the sponsor's approval. Ages 50 to 86. A care partner who sees you often enough to answer for you, oversees the tablets, comes to visits, and signs their own consent form. That last requirement is easy to skim past and hard to meet if you live alone.
It is worth being blunt about what this is not. Expanded access, as the FDA describes it, is the use of an investigational drug outside a clinical trial for a serious condition when there is no satisfactory alternative. There is no placebo group here and no randomisation, so nothing that happens to these 32 people can tell anyone whether the drug works. Places are capped because the supply of drug, sites, staff and donated money is finite, not because a regulator rationed compassion. And the sponsor holds a veto: the eligibility criteria say in as many words that a referred patient must "be approved by Sponsor".
Eight sites across a large country creates a travel problem that dementia makes worse. Cognition's answer, announced on 5 February 2026 when it extended the programme past its original twelve months, was to push routine health monitoring out to participants' own doctors and allow local laboratory testing, reserving the specialist sites for the more demanding assessments. Shprecher, who advised on the design, said the team "devised a system that would enable patients to see their local doctors for routine visits". For a family, that is the difference between a day of driving every few weeks and a morning at a familiar surgery. It also devolves a share of the safety watch to clinicians who are not trial investigators, which is a trade the programme has made openly and which the registry record does not itself capture.
What the company gets back is in the study's own title: long-term safety and efficacy data. The underlying evidence is thinner than the enthusiasm around it. SHIMMER, published in Alzheimer's & Dementia in December 2025, screened 293 people, randomised 130 with mild-to-moderate DLB to 100 mg zervimesine, 300 mg or placebo for 26 weeks, and saw 109 (83.8 per cent) complete. It met its primary endpoint, which was safety and tolerability; every efficacy measure was exploratory, and the widely quoted 86 per cent difference on the Neuropsychiatric Inventory carried a p value of 0.0545, which does not clear the conventional threshold. On 24 June 2026 the company said the FDA had agreed that DLB psychosis could be an approvable outcome, with a pivotal trial of 100 mg against placebo for nine months expected to start in mid-2027.
The question every family in an expanded access programme eventually asks is what happens when the money runs out. Here the answer arrived on 14 May 2026, in the form of a second donation, from Jeffrey Pechter of Mindful Capital, a participant in the programme, extending treatment to two years; the company's second-quarter update on 6 August 2026 says the programme now runs through 2027. Chief executive Lisa Ricciardi framed the need as the lack of approved treatments and limited research leaving people "with few options". That is generous and it is also precarious. Twice now, continuity of an investigational medicine for 32 people has depended on someone choosing to write a cheque.