An experimental drug first developed to treat erectile dysfunction slowed the worsening of Alzheimer’s symptoms by 31% over a year in one group of patients — but the benefit did not hold across the full trial, leaving researchers with a promising lead and a difficult question.
By The BNN Media
The Best News Network
News Beyond the Borders
Boston, United States — An experimental drug called AR1001 slowed clinical decline by 31% over one year in a group of 604 people with early Alzheimer’s disease who were not taking other Alzheimer’s medicines, according to results presented at a major Alzheimer’s conference in Boston. But across the full trial of 1,535 people, the drug did not produce a statistically reliable benefit, and no peer-reviewed scientific paper on the results has yet been published.

What the Trial Found
The results come from the phase three POLARIS-AD trial, the final phase of testing in humans, run by AriBio, a South Korean biotechnology company. The findings were announced at the 19th Clinical Trials on Alzheimer’s Disease (CTAD) conference in Boston, held from November 16 to 19, 2026.
The trial enrolled 1,535 people with early Alzheimer’s disease at about 230 sites in 13 countries. Participants took either 30mg of AR1001 or a placebo once a day for a year. Researchers measured changes in memory, thinking and everyday activities using a scale called the Clinical Dementia Rating–Sum of Boxes, or CDR-SB.
The trial looked at two groups set out in advance: everyone who took part, and a smaller group of 604 people who were not taking other Alzheimer’s medicines to ease their symptoms, such as donepezil or memantine. Only this smaller group — the monotherapy group — saw the 31% slowing of worsening symptoms.
In that group, the difference between the drug and the placebo on the CDR-SB scale was 0.44 points, a result the company described as a 31% slowing of clinical decline, with a statistical value of P = .0266. In the overall population, the difference was only 0.11 points, with P = .4015 — a result that was not statistically significant, meaning it could plausibly have arisen by chance. BioSpace, reporting the same results, put the slowing across the full trial at 7%.
The headline figure is real, but it applies to only one part of the trial — and that distinction matters more than the number itself.
How the Drug Is Supposed to Work
AR1001 is a once-daily oral drug that blocks an enzyme called PDE5 — the same family of drugs as Viagra, which is used to treat erectile dysfunction. Blocking PDE5 allows a chemical messenger called cGMP to build up in the body, helping the muscles around blood vessels to relax so that more blood can flow through them.
The same chemical messenger also helps brain cells communicate with each other. Researchers think that increasing its levels could help brain cells survive and maintain the connections they need to share information — connections that are essential for learning and memory, and that Alzheimer’s gradually damages.
However, this is still a theory. As Scientific American, reporting on the trial, made clear, the clinical trial does not show that these biological changes explain the reported benefits in people. Some laboratory and animal studies support the idea: in a 2022 study, mice with features of Alzheimer’s disease performed better in memory tests after receiving AR1001, and the treatment also reduced levels of amyloid-beta and abnormal tau, two proteins linked to Alzheimer’s. But results in animals do not always translate into benefits for humans.
Interest in this family of drugs has also grown from research in people. A 2024 study followed more than 269,000 men with erectile dysfunction and found that men prescribed these drugs were 18% less likely to be diagnosed with Alzheimer’s than men who were not prescribed them. That was an observational study, however, which means it cannot show that the drugs themselves reduced the risk — the two groups may have differed in other ways that influenced the results.

The Results in Detail
The trial’s design allowed it to succeed if either of its two pre-set groups met the main measure, so the monotherapy result counts as a success under the trial’s own rules — but the failure in the overall population leaves the picture mixed.
Among the monotherapy group, the drug also showed a difference of 0.7 points on the Mini-Mental State Examination, a standard test of thinking and memory, with P = .0232. On another test, the Alzheimer’s Disease Assessment Scale–Cognitive Subscale 13, the difference was 1.16 points, with P = .0798, which was not statistically significant.
Of the 1,535 participants, 1,349 — or 87.9% — completed the 52-week double-blind period of the trial. Of those, 1,284 entered an optional 52-week open-label extension, in which all participants receive AR1001 30mg daily. That extension is expected to conclude in the third quarter of 2027.
Sharon J. Sha, the trial’s global principal investigator, presented the results at the CTAD conference. She said the CDR-SB result and the safety profile were encouraging, and that the secondary and biomarker findings warranted further evaluation alongside the overall result, according to Psychiatric Times.
AriBio’s chief executive, Jai-Jun Choung, said in a prepared statement: “The 31% slowing of clinical decline at one year in the monotherapy population supports our hypothesis.” The company says it plans to discuss the complete findings with the US Food and Drug Administration (FDA).
Safety: What Was Reported
Treatment-emergent adverse events — health problems that occurred during the trial, whether or not they were caused by the drug — were reported in 69.0% of people taking AR1001 and 66.4% of those taking a placebo. Serious events occurred in 10.0% of the AR1001 group and 9.8% of the placebo group, figures that were very close.
Diarrhoea was more common on AR1001 than on placebo, affecting 7.7% of treated participants compared with 4.4% of those on placebo, according to BioSpace.
The closeness of the overall safety figures is one of the more reassuring parts of the report, but the full safety picture will only be clear when the complete trial data are published and reviewed by independent researchers.
On safety, the drug and the placebo looked broadly similar — but the full data have not yet been published.
Why the Benefit Appeared in Only One Group
The central puzzle of the trial is why the drug appeared to help people who were not taking other Alzheimer’s medicines, but not the trial population as a whole. The researchers have not yet given a settled answer, and the full data may shed light on it.
One possibility raised by the reporting is that other Alzheimer’s medicines may interact with AR1001, or may already be doing some of the work the new drug is meant to do, leaving less room for an additional benefit to show up. Another is that the monotherapy group differed from the rest of the trial population in ways that made a benefit easier to detect. Until the full results are published, these remain open questions rather than answers.
For comparison, BioSpace noted that the result in the monotherapy group is in line with existing Alzheimer’s drugs: Biogen and Eisai’s Leqembi achieved a 27% slowing of decline in its phase three trial at 18 months, and Eli Lilly’s Kisunla achieved a 29% slowing at 76 weeks.
The relationship between erectile dysfunction and Alzheimer’s is also complicated. Erectile dysfunction can be associated with conditions such as diabetes, high blood pressure and heart disease, which can damage blood vessels. Those same conditions are also linked to poorer brain health and a higher risk of dementia — so some of the apparent link between the drugs and dementia could come down to shared risk factors affecting blood vessels, rather than a direct effect of the drugs.

What This Does — and Does Not — Mean for Patients
The most important point for readers is a cautious one. Erectile dysfunction drugs are not an established treatment for Alzheimer’s, and nothing in these results suggests people should take them to protect their memory. AR1001 is an investigational drug: it is not approved for Alzheimer’s, and regulators could require more trials or decline to approve it.
What the trial does keep alive is an interesting possibility — that drugs which improve blood flow and help brain cells communicate might also slow the damage caused by Alzheimer’s. Whether that is the case will depend on the full trial results and further research in people, including studies looking at why the apparent benefit was seen only among those not taking other Alzheimer’s medicines.
- AR1001, an experimental once-daily oral PDE5 inhibitor from AriBio, was tested in the phase three POLARIS-AD trial.
- 1,535 people with early Alzheimer’s in 13 countries took 30mg of AR1001 or a placebo daily for 52 weeks.
- In the 604-person monotherapy group, decline slowed by 31% on the CDR-SB scale (difference 0.44 points, P = .0266).
- Across the whole trial, the difference was 0.11 points (P = .4015) — not statistically significant.
- No peer-reviewed paper has yet been published; AriBio plans to discuss the findings with the FDA, and the open-label extension runs until the third quarter of 2027.
A 31% slowing in one group is a genuine reason for interest — but until the full data are published and reviewed, it is a lead to follow, not a treatment to rely on.
Reporting based on Scientific American (via The Conversation), NeurologyLive, Psychiatric Times and BioSpace, 16–22 November 2026.
By The BNN Media
The Best News Network
News Beyond the Borders