POLARIS-AD Phase 3 Trial of PDE5 Inhibitor AR1001 Completes Treatment Phase With Favorable Safety Profile

AriBio Co. presented updated data from the POLARIS-AD phase 3 trial of AR1001, an oral PDE5 inhibitor in early Alzheimer's disease (AD), at the 2026 Alzheimer’s Association International Conference (AAIC) in London.¹ All 1,535 enrolled participants have completed the 52-week double-blind treatment phase, with topline results planned for CTAD 2026.

Niels Prins, MD, PhD, of the Brain Research Center in Amsterdam, Netherlands, presented the data alongside co-investigators including Alireza Atri, MD, PhD, of Banner Sun Health Research and Banner Alzheimer's Institutes.

POLARIS-AD Trial Design

POLARIS-AD (NCT05531526) is a global, multicenter, double-blind, placebo-controlled trial with a 52-week treatment phase and a 52-week extension phase in people with early AD.¹ Participants were randomized 1:1 to AR1001 30 mg once daily or placebo. The primary endpoint is change from baseline on the Clinical Dementia Rating-Sum of Boxes (CDR-SB) at week 52, evaluated in both the intent-to-treat and monotherapy populations. Key secondary endpoints include the Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog13), Amsterdam Instrumental Activities of Daily Living Questionnaire-Short Version (A-IADL-Q-SV), Mini-Mental State Examination (MMSE), and Geriatric Depression Scale-15 (GDS-15). Biomarker endpoints assess plasma and CSF changes in p-Tau181, p-Tau217, p-Tau231, Abeta42/40 ratio, glial fibrillary acidic protein (GFAP), and neurofilament light chain (NfL).

The enrolled population required confirmed amyloid pathology via CSF Abeta42/40 ratio or PET, MMSE between 20 and 30, and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) Delayed Memory Index of 85 or below, targeting mild cognitive impairment (MCI) due to AD or mild AD dementia.

Enrollment and Baseline Characteristics

A total of 4,025 individuals were screened, of whom 1,535 were enrolled across 240 sites globally. Enrollment was distributed across North America (658), the European Union (502), South Korea (200), China (126), and the United Kingdom (49). Amyloid positivity was confirmed by CSF Abeta42/40 ratio in 964 participants (62.8%; mean 0.048) and by PET in 571 participants (37.2%; mean 89.1 Centiloids).¹

Baseline clinical characteristics were well-matched to other global early AD registration trials. Mean CDR-SB was 3.45 (SD 1.68) overall, 2.68 in the MCI subgroup (n = 1,139), and 5.68 in the mild dementia subgroup (n = 395). Mean MMSE was 23.99 and mean ADAS-Cog13 was 27.87 overall.¹

Screen fail rates varied substantially by region, ranging from 34.8% in Korea to 82.9% in the UK (largely driven by a single high-volume site). The most common screen fail reasons were RBANS score in the US and UK, and CSF/PET amyloid confirmation in China and Korea.¹

Safety and Completion Data

Of 1,535 enrolled participants, 961 (62.6%) completed the 52-week treatment phase and 919 of these (95.6%) enrolled in the extension phase. A total of 180 participants (11.7%) discontinued during the treatment phase.

Adverse events were predominantly mild to moderate. The most common were fall (n = 95; 3.1%), diarrhea (n = 94; 3.1%), urinary tract infection (n = 93; 3.1%), headache (n = 73; 2.4%), and dizziness (n = 70; 2.3%). No cases of ocular neuropathy were reported, though 24 ocular adverse events occurred, a finding the investigators flagged as requiring continued monitoring given historical concerns with PDE5 inhibitors. No unexpected safety signals were identified.¹

A subset of POLARIS-AD participants also received approved anti-amyloid therapies during the trial: 6 patients were on donanemab (Kisunla) and 5 on lecanemab (Leqembi). Among donanemab-treated patients, all 6 remained on study drug; one asymptomatic amyloid-related imaging abnormality with edema (ARIA-E) event and one infusion reaction were reported. Among lecanemab-treated patients, 3 of 5 remained on study drug, with one patient experiencing both ARIA-H and ARIA-E.¹ No unexpected safety signals emerged from these concomitant use subgroups.

Mechanism and Phase 2 Background

AR1001 operates through a polypharmacological mechanism distinct from amyloid-targeting antibodies. By inhibiting PDE5, the drug elevates cyclic GMP (cGMP) levels, activating protein kinase G (PKG) and downstream PI3K signaling. Preclinical data show this cascade produces neuroprotection via NGF and BDNF upregulation, reduces tau phosphorylation via GSK3-beta inhibition, and improves brain perfusion and blood-brain barrier integrity.¹

The phase 2 trial (NCT03625622) enrolled 210 patients with mild-to-moderate AD at 21 U.S. sites, randomized to AR1001 10 mg, 30 mg, or placebo for 26 weeks. The trial did not meet its co-primary endpoints of ADAS-Cog13 and Alzheimer's Disease Cooperative Study-Clinical Global Impression of Change (ADCS-CGIC) in the overall population. However, a pre-specified monotherapy subgroup showed statistically significant cognitive improvement on ADAS-Cog13, with the effect most pronounced in amyloid-positive patients.² Safety and tolerability were favorable throughout, with no serious drug-related adverse events.

This monotherapy subgroup signal, combined with a clean tolerability profile, formed the basis for advancing to phase 3 in a population with early AD where monotherapy use is more prevalent and disease stage more homogeneous. Topline efficacy results are expected at CTAD 2026.

Click here for more AAIC 2026 coverage.

REFERENCES
1. Prins N, Atri A, Kim SY, et al. POLARIS-AD: AR1001 additional phase 2 data and phase 3 trial update with topline results expected later this year. Abstract and presentation at: Alzheimer's Association International Conference (AAIC); July 2026; London, United Kingdom.
2. Greeley D, Atri A, Rock JA, et al. A phase 2 randomized, placebo-controlled study on the efficacy and safety of AR1001, a phosphodiesterase-5 inhibitor, in patients with mild-to-moderate Alzheimer's disease. J Prev Alzheimers Dis. 2025;12(9):100337. doi:10.14283/jpad.2025.100337. https://doi.org/10.14283/jpad.2025.100337