
AriBio’s oral drug AR1001 met the primary endpoint of its Phase 3 POLARIS-AD trial in early Alzheimer’s disease only in a prespecified monotherapy population, and did not separate from placebo in the full trial population, the Korean drug developer reported on 17 November 2026 at the Clinical Trials on Alzheimer’s Disease (CTAD) conference in Boston. The 1,535-participant, industry-sponsored study gives trial sites, sponsors and research offices a closely watched example of how a dual-population design is read.
What happened
AriBio announced topline results from POLARIS-AD on 17 November 2026, and the global principal investigator, Sharon Sha, presented them at CTAD 2026, according to Psychiatric Times and Clinical Trials Arena. AR1001 is a once-daily, brain-penetrant phosphodiesterase 5 (PDE-5) inhibitor given at 30 mg. The trial was randomised, double-blind and placebo-controlled, and the treatment period lasted 52 weeks.
The primary endpoint was the change from baseline to week 52 on the Clinical Dementia Rating Scale Sum of Boxes (CDR-SB), assessed in two prespecified populations: all randomised participants, and a monotherapy population of people not taking an acetylcholinesterase inhibitor or memantine at baseline. In the monotherapy population (604 participants), CDR-SB was 0.44 points better than placebo (P = .0266), which AriBio described as a 31% slowing of clinical decline. In the overall population, the difference was 0.11 points, numerically favouring AR1001 but not statistically significant (P = .4015), as reported by Psychiatric Times and by the Korean outlet MoneyToday.
Key details
The figures below come from the topline results as reported by Psychiatric Times and MoneyToday. They have not yet appeared in a peer-reviewed publication.
| Measure | Monotherapy population | Overall population |
|---|---|---|
| CDR-SB difference vs placebo (primary endpoint) | 0.44 points, P = .0266 | 0.11 points, P = .4015 |
| MMSE difference | 0.7 points, P = .0232 | Numerically favoured AR1001 |
| ADAS-Cog13 difference | 1.16 points, P = .0798 (not significant) | Numerically favoured AR1001 |
| CSF p-tau217 lower than placebo | 5.34 pg/mL, P = .0402 | 3.52 pg/mL, P = .1784 (not significant) |
| CSF cyclic GMP vs placebo | P = .0058 | P below .0001 |
The amyloid-beta 42/40 ratio showed no difference at 52 weeks, according to Psychiatric Times. The cyclic GMP result is described as consistent with the drug reaching its biological target. Of the participants, 1,349 (87.9%) completed the double-blind period, and 1,284 of those (95.2%) entered the open-label extension, in which everyone receives AR1001 30 mg.
On safety, treatment-emergent adverse events occurred in 69.0% of the AR1001 group and 66.4% of the placebo group. Serious events were reported in 10.0% and 9.8%. Fatal events numbered four (0.5%) on AR1001 and two (0.3%) on placebo. Diarrhoea (7.7% vs 4.4%), falls (7.3% vs 5.9%) and dizziness (6.1% vs 3.1%) were more common on AR1001, and 5.2% of the AR1001 group stopped treatment because of an adverse event, against 1.7% on placebo. Clinical Trials Arena reports that most adverse events were mild to moderate.
Trial design and registry record
The ClinicalTrials.gov record NCT05531526 lists AriBio Co., Ltd. as an industry sponsor and the study as Phase 3 with an actual enrolment of 1,535. It shows an actual start on 23 December 2022 and an actual primary completion date of 29 June 2026, and it estimates study completion in October 2027. The registered primary outcome is the change in CDR-SB from baseline to week 52 in both the overall and the monotherapy populations, which matches what was reported. The record lists sites in 13 countries, including the United States, Canada, China, South Korea, the United Kingdom and several European states.
An earlier NeurologyLive report from July 2026, on data presented at the Alzheimer’s Association International Conference in London, described participants randomised 1:1 to AR1001 or placebo, with confirmed amyloid pathology and baseline MMSE scores of 20 to 30. It listed enrolment by region as 658 in North America, 502 in the European Union, 200 in South Korea, 126 in China and 49 in the UK, and said investigators flagged 24 ocular adverse events for continued monitoring because of historical concerns with PDE-5 inhibitors.
Background and context
AR1001 works differently from the anti-amyloid antibodies lecanemab (Leqembi) and donanemab (Kisunla). Clinical Trials Arena describes it as acting by relaxing blood vessels to improve blood flow to the brain, rather than by clearing plaques, and notes that it belongs to the PDE-5 inhibitor class that includes sildenafil. It has not received regulatory approval for Alzheimer’s disease.
Two prespecified primary populations meant the trial carried two separate statistical tests, each with its own significance bar. That is why the sponsor describes the primary efficacy objective as met in the monotherapy group while the overall result was not significant. How regulators weigh a win in a subgroup that was named in advance is a question for the FDA review, not for the topline announcement.
Reaction
Sharon Sha described the CDR-SB result and safety profile as “encouraging,” according to Psychiatric Times, and said the secondary and biomarker findings “warrant further evaluation alongside the overall population outcome,” as quoted by Clinical Trials Arena. Richard Oakley of the Alzheimer’s Society called the results an “exciting sign” for a new treatment that addresses a cause of the disease, adding that “the drug doesn’t seem to be as effective for people taking other drugs for Alzheimer’s symptoms, so this needs further investigation.”
According to MoneyToday, which notes that its English version is machine-translated, AriBio chief executive Jung Jae-jun said the company would “quickly proceed to file for FDA approval,” and that it also plans further analyses, consultations with global regulators and a long-term extension study to test durability and safety. Clinical Trials Arena puts it more narrowly: AriBio plans to share the complete POLARIS-AD findings with the US Food and Drug Administration to inform next steps.
What it means for researchers, universities and research offices
For clinical research offices, the result is a working example of why the analysis populations in a registered protocol matter. Both populations were named as primary outcomes in the public record before the readout, so the subgroup finding is a prespecified analysis and not a post hoc one. Offices that support investigator-initiated or sponsor-funded trials can use the record as a model when they review registration and results-reporting obligations and the wording of endpoints and statistical analysis plans.
The split between a monotherapy subgroup and the full population also bears on enrolment rules. Trials in early Alzheimer’s disease increasingly enrol people who already take symptomatic drugs or anti-amyloid antibodies, and the POLARIS-AD pattern suggests that background therapy can change how a treatment effect appears. Sites and institutional review boards reviewing similar protocols may ask sponsors how concomitant medication is handled in randomisation and analysis.
The imbalance in fatal events, four against two, and the higher rate of discontinuation on AR1001 are the kind of detail that a data safety monitoring board and local safety reporting processes track, and that participants should hear about through updated consent materials where a study continues in an extension. Researchers following the field should wait for the full dataset: the figures reported so far are topline, and a Phase 3 trial read out by its sponsor is not a regulatory decision. The design is also a placebo-controlled comparison of one dose over 52 weeks, which limits what it can say about longer-term benefit.
What happens next
AriBio says it will bring the full data to the FDA and plans additional analyses. The open-label extension, in which all participants receive AR1001, is expected to conclude in the third quarter of 2027, according to Psychiatric Times, while the registry estimates study completion in October 2027. AriBio has not announced a filing date. Details such as the full statistical analysis, subgroup tables and a peer-reviewed paper will be the basis for judging the monotherapy result. Records of registered trials, including ClinicalTrials.gov entries, will show any change to the status or outcomes.
Sources
- ClinicalTrials.gov, NCT05531526: Phase 3 trial of AR1001 in early Alzheimer’s disease (POLARIS-AD), U.S. National Library of Medicine, accessed 19 November 2026.
- Mixed results from Phase 3 data for AR1001 in Alzheimer disease, Psychiatric Times, 17 November 2026.
- AriBio’s oral Alzheimer’s drug nets Phase III win, Clinical Trials Arena, 17 November 2026.
- AriBio reports Phase 3 POLARIS-AD results for AR1001, MoneyToday, 17 November 2026.
- POLARIS-AD Phase 3 trial of PDE5 inhibitor AR1001 completes treatment phase, NeurologyLive, 16 July 2026.







