Clinical trials

More drugs are being tested for Alzheimer’s than ever before. Many will fail. For some families, joining a trial is still worth considering. It is free, it gives access to drugs years before approval, and it helps the people who come next.

The pipeline in 2026

Every year, Jeffrey Cummings and colleagues at the University of Nevada, Las Vegas count every registered Alzheimer’s drug trial. Their 2026 pipeline report in Alzheimer’s & Dementia: Translational Research & Clinical Interventions found:

158 drugs in 192 trials, up from 138 drugs in 182 trials in 2025.
About 73 in 100 drugs aim to change the disease itself: 39% small-molecule drugs (usually pills) and 34% biologics such as antibodies and vaccines. That total is our addition of their two figures.
28 in 100 treat symptoms: 18% for thinking and 10% for behaviour.
35 in 100 (56 drugs) are repurposed, meaning already approved for something else.
18 in 100 target amyloid, and another 18 in 100 target inflammation and the immune system, up from 6 in 100 in 2016.
54,728 volunteers are needed, 38,417 of them for final-stage (Phase 3) trials.
Industry sponsors 59% of trials, including 72% of Phase 3 trials.

The report was funded by National Institute on Aging grants. We did not check the authors’ personal conflict-of-interest statements.

How trials are staged

A Phase 1 trial tests safety in a few dozen people. Phase 2 looks for signs that the drug works in a few hundred. Phase 3 tests it against placebo in hundreds to thousands of people and is what regulators usually need for approval.

Late-stage candidates to watch

Status as of September 2026. Many of these updates come from the companies themselves. We say so where that is the only source.

Notable drugs in late-stage testingStatus as of 26 September 2026. “Company” means a company announcement not yet independently checked.
Drug	What it does	Stage	Who pays	Latest
Trontinemab	Anti-amyloid antibody with a “shuttle” to cross into the brain	Phase 3	Industry Roche	In an earlier trial, 92% on the top dose were below the amyloid-positive level by week 28, with little brain swelling. A prevention trial (PrevenTRON) was announced July 2026. No Phase 3 results yet.
Remternetug	Donanemab’s successor; an injection people can give themselves	Phase 3	Industry Eli Lilly	Results were expected around March 2026. We could not find them published. About 30% of trial participants self-injected at least once.
Buntanetap	Pill that lowers production of several toxic proteins	Phase 3, 850 people	Industry Annovis Bio	Fully enrolled July 2026. 6-month results expected early 2027, 18-month results early 2028 (company).
AR1001 (mirodenafil)	Pill that blocks the enzyme PDE5	Phase 3, 1,535 people in 13 countries	Industry AriBio (Korea)	Last patient visit July 2026. First results due at the CTAD conference in Boston, 16 November 2026.
Hydromethylthionine (HMTM)	Pill meant to stop tau from clumping	Phase 3 done	Industry TauRx	Submitted to the UK regulator in 2024; the regulator asked for more information in October 2025. The company expects a decision in 2026. Not licensed anywhere.
Masitinib	Pill aimed at inflammation, added to standard drugs	Phase 3 confirmatory, 600 people	Industry AB Science	An earlier trial was positive on a thinking test (company). Confirmatory trial ongoing; timing not confirmed.
Xanomeline–trospium (Cobenfy)	Treats psychosis in Alzheimer’s	Phase 3	Industry Bristol Myers Squibb	The company found “irregularities” at a few sites and excluded them (December 2025). Results expected by the end of 2026.
Etalanetug (E2814)	Antibody meant to stop tau spreading	Phase 1b/2	Industry Eisai	In 7 people with inherited Alzheimer’s, a tangle marker fell 62–89% in spinal fluid (company). Now tested with lecanemab in the DIAN-TU study.
Diranersen (BIIB080)	Lowers tau production; injected into spinal fluid	Phase 2, moving to 3	Industry Biogen, Ionis	Missed its main goal in May 2026. Tau fell, with hints of slower decline, oddly strongest at the lowest dose. Biogen is going ahead to Phase 3.

Percentages here are the companies’ figures. Small early results often shrink or vanish in larger trials.

Other ideas are being tested in smaller trials, including the diabetes drug metformin in 370 people with mild memory problems, vaccines against amyloid and tau, gene therapy, stem cells, and drugs aimed at ageing cells. We could not confirm the current status or results of most of these, so we do not report them here. For the most speculative approaches, see our frontier page.

Recent failures

A failed trial is still useful. It tells families what not to wait for, and it saves the next patients from a drug that doesn’t work.

Semaglutide (Ozempic family)

Many people hoped the popular weight-loss and diabetes drugs would protect the brain. In the largest test so far, a pill form did not slow Alzheimer’s.

Didn’t workIndustry

evoke and evoke+: semaglutide (the Ozempic family) as a pill

Novo Nordisk topline announcement, 24 November 2025 (company release) Read the study

Design: Two randomized, placebo-controlled trials
People: 3,808 people with early Alzheimer’s
Length: 2 years

Result No benefit. Decline on the CDR-SB dementia scale at 2 years was no slower than on placebo. Some blood markers of Alzheimer’s improved, but symptoms did not. The company cancelled the extension phase.

Who paid: Novo Nordisk, which sells the drug

Conflicts: Company-run trial; the company reported its own negative result

Valiltramiprosate (ALZ-801)

A pill tested only in people with two copies of the APOE4 gene. It missed its main goal. The company points to a subgroup that did better; subgroup findings like this need a new trial to confirm them.

Didn’t workIndustry

APOLLOE4: valiltramiprosate (ALZ-801) pill in people with two copies of APOE4

Alzheon topline announcement, April 2025 (company release) Read the study

Design: Randomized, placebo-controlled Phase 3 trial
People: People with early Alzheimer’s and two copies of APOE4 (total not confirmed here)
Length: Not reported in the sources we read

Result Missed its main goal. The difference on a thinking test (ADAS-Cog13) was 0.50 points, which could easily be chance (p = 0.607). The company highlights a pre-planned subgroup of 125 people with mild cognitive impairment who did better (p = 0.041). Brain swelling was the same in both groups (3.4%). No FDA filing confirmed.

Who paid: Alzheon (sponsor). The company says a $51 million National Institute on Aging grant (R01AG065253) paid for the trial.

Conflicts: Company reports its own results

Blarcamesine (Anavex 2-73)

This pill was turned down by Europe’s drug regulator’s expert committee before the company withdrew its application.

Didn’t workIndustry

Blarcamesine (Anavex 2-73): the European regulator’s review

European Medicines Agency, blarcamesine Anavex assessment, Dec 2025–Mar 2026 Read the study

Design: Regulator’s review of a Phase 2b/3 trial
People: Not reported in the sources we read
Length: Not reported in the sources we read

Result The EMA’s expert committee recommended refusing approval on 11–12 December 2025. It cited failed main goals, concerns about the trial methods, dropouts driven by dizziness and a possible impurity. Anavex asked for a re-examination, then withdrew its application on 25 March 2026.

Who paid: Anavex Life Sciences, the maker

Conflicts: Maker-sponsored trial

Prevention trials in people without symptoms

Amyloid starts building up years before memory loss. So researchers are testing drugs in people who have amyloid but no symptoms yet. These trials are large and slow.

Didn’t workIndustry

A4: solanezumab in people with amyloid but no symptoms

Sperling RA et al., New England Journal of Medicine 2023 Read the study

Design: Randomized, placebo-controlled prevention trial
People: More than 1,100 cognitively normal people aged 65–85 with amyloid on a PET scan
Length: About 4.5 years

Result Failed. On a combined memory and thinking score (PACC), people on the antibody declined by 1.69 points vs 1.40 on placebo, a difference that could be chance (p = 0.26). The drug did not slow decline in people who had amyloid but no symptoms.

Who paid: Eli Lilly with the National Institute on Aging and philanthropy (public–private partnership); exact split not confirmed

Conflicts: Not reported in the sources we read

TRAILBLAZER-ALZ 3 (donanemab, Eli Lilly). About 2,996 people aged 55–80 with normal thinking but a positive blood test for Alzheimer’s changes. It measures time until symptoms appear. It is no longer recruiting; main results are expected around November 2027. See the TRAILBLAZER-ALZ 3 record on ClinicalTrials.gov.
AHEAD 3-45 (lecanemab, Eisai with the NIH-funded Alzheimer’s Clinical Trial Consortium). Two sub-studies, for people with intermediate and with high amyloid. We could not confirm its current status or when results are due.
TRAILRUNNER-ALZ 3 (remternetug, Eli Lilly). Screens people by blood test and follows them for up to 255 weeks.
PrevenTRON (trontinemab, Roche). Announced July 2026.
DIAN-TU (Washington University in St. Louis). Trials for families with inherited, early-onset Alzheimer’s, both before and after plaques appear. It is now testing etalanetug with lecanemab.

Drug trials are not the only prevention research. Lifestyle trials such as FINGER and US POINTER are covered on our page on lowering your risk.

What’s coming soon

16 November 2026: first results of the AR1001 (mirodenafil) Phase 3 trial, at the CTAD conference in Boston.
By end of 2026: results from Bristol Myers Squibb’s trials of Cobenfy for psychosis in Alzheimer’s.
2026: a UK decision on hydromethylthionine, according to the company. 29 Phase 2 trials were due to report during 2026.
Early 2027: 6-month buntanetap results (company timeline).
Around November 2027: main results of TRAILBLAZER-ALZ 3.

How to find and join a trial

Where to look

Alzheimer’s Association TrialMatch. Free and confidential. Answer about 10 questions and it matches you with drug and non-drug studies. Open to people with memory problems or dementia, caregivers and healthy volunteers. Call 1-800-272-3900 (24 hours a day) or visit TrialMatch on alz.org.
ClinicalTrials.gov. The NIH registry of every US trial. Search “Alzheimer” plus your ZIP code, and filter for “Recruiting.” Each listing shows who can join and a contact. Start at ClinicalTrials.gov.
Alzheimer’s Disease Research Centers. NIH-funded centres at universities across the country. Many run long-term memory studies, offer free assessments, and connect families to trials.
Families with inherited Alzheimer’s. Contact DIAN-TU at Washington University.

What joining usually involves

Screening. Blood tests, memory tests, and often MRI or PET scans. Many people are screened out. That is normal.
Consent. If your parent can no longer decide for herself, a legally authorised representative can consent for her.
A study partner. Usually a family member who comes to visits and reports on daily life.
Placebo. She may get a dummy treatment. Many trials offer the real drug to everyone afterwards.
Cost. The study drug and study tests are free, and travel is often paid back.
Leaving. She can withdraw at any time.
For antibody trials. Expect regular MRI scans to check for brain swelling.

Questions to ask a study team

Before you agree What is the chance of getting placebo? How many visits are there, and how long is each?Ask about scans, spinal taps and travel. Can she keep taking her current medicines? What happens when the trial ends? Is there an open-label extension? Will we be told the results? Who is paying for the trial?The answer is usually on the ClinicalTrials.gov listing under “Sponsor.”

Words used on this page

Phase 1, 2, 3: The stages of drug testing, from small safety studies (Phase 1) to large placebo-controlled trials (Phase 3) that regulators use to decide on approval.