Remlifanserin Phase 2 Miss Forces High-Stakes Phase 3 Bet on Secondary Signal Alone

Acadia's Remlifanserin Misses Primary Endpoint in Phase 2 ADP Trial

Acadia Pharmaceuticals' investigational drug, remlifanserin, a 5HT2A receptor inverse agonist, failed to meet its primary endpoint in the Phase 2 portion of the RADIANT trial for Alzheimer’s disease psychosis (ADP). The 60-mg dose showed a 12.6-point drop in hallucinations and delusions compared to a 10.4-point reduction for placebo, with a p-value of 0.0603, which the company called a "narrow miss." Despite this, Acadia plans to advance remlifanserin into Phase 3, citing statistically significant improvement on a secondary endpoint (clinician-rated severity scale for ADP) and a favorable safety profile. The 30-mg dose will be removed from the Phase 3 program.

• The Phase 2 RADIANT trial for remlifanserin in Alzheimer’s disease psychosis (ADP) did not achieve statistical significance on its primary endpoint, which measured improvement in hallucinations and delusions. The 60-mg dose resulted in a 12.6-point reduction compared to a 10.4-point reduction for placebo, yielding a p-value of 0.0603, just above the desired 0.05 threshold. Acadia characterized this outcome as a "narrow miss," suggesting clinical relevance despite the statistical shortfall.
• Despite the primary endpoint miss, remlifanserin demonstrated statistical significance on a key secondary endpoint: a clinician-rated severity scale for ADP, specifically for the 60-mg cohort. This positive secondary data supports Acadia's decision to continue development. Consequently, the company plans to remove the less effective 30-mg dosage arm from the ongoing Phase 3 RADIANT program, focusing on the 60-mg dose.
• Remlifanserin's development continues, bolstered by its Fast Track status from the FDA for ADP, highlighting the significant unmet need for patients experiencing psychosis in Alzheimer's disease. The drug also showed a favorable safety profile, with adverse event rates similar to placebo and no signals of QT prolongation or motor symptoms. This positive safety data supports its ongoing mid-stage development for psychosis associated with Lewy body dementia (DLB), with Phase 2 DLB results expected in early 2028.

Remlifanserin's Phase 2 Results in Alzheimer's Psychosis

One completed study in Alzheimer's disease psychosis (ADP) is ACP-103-019, a 12-week, randomized, double-blind, placebo-controlled trial evaluating pimavanserin (34 mg) in patients with ADP. The primary endpoint was change from baseline in the Neuropsychiatric Inventory-Nursing Home Version-Psychosis Score (NPI-NH-PS) at week six. A post hoc analysis of this study examined agitation and aggression outcomes, finding that pimavanserin-treated patients who achieved ≥50% reduction from baseline in NPI-NH-PS (psychosis responders, n = 44) showed significantly greater improvement in agitation and aggression on the NPI-NH domain C (week six, least squares mean difference = -3.64, t = -4.69, P < .0001) and the Cohen-Mansfield Agitation Inventory-Short Form (week six, LSM difference = -3.71, t = -2.01, P = .0483) compared with nonresponders (n = 32). Differences between responders and nonresponders were also observed in patients with more severe agitation and aggression at baseline on the NPI-NH domain C (responders, n = 26; nonresponders, n = 13; week six, LSM difference = -3.03, t = -2.44, P = .019).

A second study, the Lit-AD Randomized Clinical Trial, evaluated low-dose lithium carbonate (150–600 mg daily) versus placebo over 12 weeks across four sites in patients with Alzheimer's disease and an agitation/aggression score ≥4 on the Neuropsychiatric Inventory (NPI). Of 77 patients enrolled, 58 (75.3%) completed the trial. Lithium was not significantly superior to placebo on the primary efficacy outcome of change in NPI agitation/aggression. The proportion of responders (defined as ≥30% reduction in NPI score for agitation/aggression plus psychosis and a Clinical Global Impression score of much or very much improved) was 31.6% on lithium versus 17.9% on placebo (χ²=1.26, p = 0.26). However, moderate or marked improvement on the Clinical Global Impression scale was greater on lithium (10/38 = 36.8%) than placebo (0/39 = 0%, Fisher's exact test p < 0.001). Exploratory analyses indicated greater improvement on lithium versus placebo on NPI delusions and irritability/lability (p's < 0.05), and lithium showed greater reduction than placebo in patients with high Young Mania Rating Scale scores (β = 5.06; 95% CI, 1.18 to 8.94, p = 0.01). Lithium did not differ significantly from placebo on safety outcomes.

The knowledge base does not have sufficient information on this aspect.

Addressing Unmet Needs in Alzheimer's Disease Psychosis

The treatment of Alzheimer's disease psychosis (AD+P) remains a significant clinical challenge, with no pharmacologic therapy currently approved by the FDA specifically for this indication. Atypical antipsychotics (AAPs) are used off-label despite a well-documented burden of safety risks and only marginal efficacy, underscoring a persistent unmet need for disease-specific therapeutic options.

• Lack of approved pharmacotherapy: No drug has received FDA approval for the treatment of psychosis in patients with dementia, including AD+P. Atypical antipsychotics are used off-label in severe cases where non-pharmacological interventions have failed, without a regulatory framework specific to this population.

• Marginal and inconsistent efficacy: AAPs — including quetiapine, risperidone, olanzapine, and aripiprazole — demonstrate only numerically small improvements in psychotic symptoms among patients with dementia-related psychosis. In network meta-analysis, quetiapine showed no improvement over placebo, while olanzapine (SMD −0.17; 95% CI −0.04, 0.02) and aripiprazole (SMD −0.12; 95% CI −0.31, 0.06) produced small, non-significant numerical improvements on the NPI-NH psychosis subscale.

• Serious safety risks, including mortality: All AAPs carry an FDA black box warning for elevated risk of mortality in elderly patients with dementia-related psychosis. Network meta-analysis data show higher odds of mortality versus placebo for aripiprazole (OR 1.58), quetiapine (OR 1.68), risperidone (OR 1.63), olanzapine (OR 2.21), and brexpiprazole (OR 2.22). Risperidone (OR 3.68; 95% CI 1.68, 8.95) and olanzapine (OR 4.47; 95% CI 1.36, 14.69) demonstrated significantly greater odds of cerebrovascular adverse events compared to placebo.

• Broad adverse event profile compounding dementia burden: AAP use in this population is associated with motor function disorders, cognitive impairment, somnolence, extrapyramidal symptoms, falls, fractures, and weight gain — adverse effects that directly worsen the underlying disease burden. Olanzapine carries significantly higher odds of discontinuation due to adverse events (OR 2.62; 95% CI 1.75, 3.92) compared to placebo.

• AD+P as a distinct, undercharacterized disease subtype: Psychosis in Alzheimer's disease has been identified as an independent predictor of more-rapid cognitive decline and is now recognized as a heritable disease subtype with neuropathological specificity. Current treatments were not developed for this syndrome and are approved for schizophrenia, not AD+P, reflecting a fundamental mismatch between available therapies and the underlying neurobiology of the condition.

A Calculated Bet on 5-HT2A Agonism in ADP

The recent Phase 2 results for remlifanserin in Alzheimer's disease psychosis (ADP) present a nuanced picture, yet Acadia's decision to push forward into Phase 3 is a calculated strategic move. While the primary endpoint was narrowly missed, the observed trend towards improvement and a statistically significant secondary endpoint, coupled with a favorable safety profile, provide a foundation for this advancement. This situation echoes the development path of pimavanserin, another 5-HT2A inverse agonist from the same company, which ultimately gained approval for Parkinson's disease psychosis (PDP) after navigating mixed trial outcomes and refining its primary endpoint.

The company's confidence likely stems from the established mechanism of action. Selective 5-HT2A inverse agonism offers a crucial advantage over traditional antipsychotics by avoiding dopamine receptor blockade, thereby mitigating motoric worsening and other severe side effects in vulnerable elderly populations. This differentiation is paramount in a landscape where current treatments for ADP are largely off-label and carry significant safety concerns, including the class-wide black box warning for increased mortality in elderly patients with dementia-related psychosis.

However, the path ahead is not without its challenges. The 'narrow miss' in Phase 2 highlights the inherent difficulty in designing clinical trials for neuropsychiatric conditions in neurodegenerative diseases. Future Phase 3 trials will need to demonstrate robust efficacy, potentially requiring a highly refined primary endpoint or a more targeted patient selection strategy, similar to how the SAPS-PD scale was instrumental for pimavanserin's success. The removal of the 30-mg dose from the Phase 3 program suggests a data-driven optimization of the therapeutic window. If successful, remlifanserin could offer a much-needed, safer alternative for ADP, further validating the 5-HT2A inverse agonist class and solidifying the company's position in addressing psychosis in neurodegenerative disorders. This strategic persistence could ultimately redefine the standard of care for a patient population desperately in need of effective and well-tolerated treatments.