Bristol Myers Squibb’s hopeful late-year phase 3 rebound is no longer in the cards. The company said that the long-awaited Adept-2 readout will slip past year-end after BMS found execution-related irregularities at a small group of trial sites.
Adept-2 is one of three late-stage studies evaluating Cobenfy in Alzheimer’s-related psychosis.
Because of the delay, BMS appears set to close out 2025 without the turnaround moment investors were counting on for Cobenfy, especially after the drug failed earlier this year as an add-on therapy for schizophrenia. More specifically, of the six pivotal studies BMS has unveiled in 2025, only one delivered a win, meeting one of two primary endpoints in a multiple myeloma trial of iberdomide.
BMS views Alzheimer’s-related psychosis as a major commercial opportunity, noting that roughly half of the 6 million Americans living with Alzheimer’s experience hallucinations or delusions.
According to a note from William Blair analysts, investor expectations for Adept-2 are elevated because of the size of this potential market, the need to build positive traction for Cobenfy’s late-stage program, and the recent disappointments across other pivotal Bristol Myers studies.
Investor concerns around Adept-2 were first triggered by remarks made during BMS’ second-quarter earnings call in July. At the time, then–CMO Samit Hirawat noted that the company was reviewing trial sites to increase the likelihood of success across its broader clinical portfolio.
Those site checks have now uncovered “irregularities,” prompting BMS to remove data from the affected locations. After aligning with the FDA, the company brought in an independent group to run an interim evaluation of both efficacy and safety while BMS stayed blinded. Based on guidance from the data monitoring committee, BMS will begin enrolling new participants in Adept-2, which is now slated for a readout at the end of the upcoming year.
BMS has maintained that it needs at least two successful studies before seeking approval, so the Adept-2 setback does not change its initial regulatory timeline.
Although site problems and delays can raise red flags, William Blair analysts viewed the data monitoring committee’s decision to keep the study going as encouraging—arguing that if no meaningful signal existed, there would be little reason to expand enrollment.
According to William Blair, the problematic sites are in the U.S. after the analysts spoke with BMS’ investor relations team. The company isn’t sharing how many patients were affected or the specifics of the issues, saying only that the situation involved several different problems. ClinicalTrials.gov indicates the trial had already closed enrollment after hitting 400 participants.
Laura Gault, who oversees neuroscience drug development at BMS, said in a statement that the choice to remove patient data from sites where concerns arose underscores the company’s firm commitment to preserving study integrity. She emphasized that psychosis linked to Alzheimer’s disease remains a significant unmet need and that upholding strict research standards is crucial as they pursue new therapies for patients and families facing this challenging condition.
