J Clin Psychiatry. 2026 Apr 15;87(2):25m16022. doi: 10.4088/JCP.25m16022.
ABSTRACT
Objective: Cognitive impairment associated with schizophrenia (CIAS) is a key driver of functional deficits in patients with this disorder. CIAS may involve deficient NMDA receptor-dependent neurotransmission. Negative allosteric modulators (NAMs) of γ-aminobutyric acid (GABA) type A (GABAA) α5 receptors enhance their activity and improve cognition in rodents and nonhuman primates. We tested whether prolonged treatment with basmisanil-a selective GABAA α5 NAM-improved CIAS.
Methods: This 24-week placebo-controlled phase 2b study in patients with CIAS diagnosed according to DSM-5 was conducted between November 2016 and December 2019. Two hundred thirteen patients were randomized to 80 mg of basmisanil, 240 mg of basmisanil, or placebo twice a day. Recruitment into the 80 mg group was stopped after a planned futility analysis. The primary outcome was the MATRICS Consensus Cognitive Battery (MCCB) neurocognitive composite score evaluated in the efficacy analysis population (EAP) in the placebo (n=76) and basmisanil 240 mg arms only (n=77, total EAP n=153). Secondary outcome measures included specific hippocampal and prefrontal-dependent cognitive tasks, functional outcomes, and symptoms. Novel study design features addressing potential sources of heterogeneity in drug response and learning and practice effects included stratification by cognitive trajectories and age as well as repeated administration of the MCCB test battery during screening.
Results: Twenty-four weeks of treatment with basmisanil was well tolerated without improving cognitive or functional measures overall or in any of the stratified subgroups. Practice and learning effects were not observed at week 12.
Conclusion: Basmisanil did not improve cognitive deficits. Although unsuccessful, some results offer insight into important factors that should facilitate the design of more informative trials.
Trial Registration: ClinicalTrials.gov identifier: NCT02953639.
PMID:41983730 | DOI:10.4088/JCP.25m16022
