compound
Fosgonimeton
The clinical version of an idea sold as a research chemical, covered trial by trial: six registrations, one Phase 2/3 with posted results, and a primary endpoint the registry records as not met.
- Primary sources Registry records and papers are read at the source, not through summaries. Read the methodology
- Every claim graded Each claim about what a compound does carries an evidence tier and a numbered citation. How grades derive
- Sells nothing No cart, no price, and no product page anywhere on this site. Conflict disclosure
- Corrections logged Errors are fixed and recorded in a permanent public log, never edited away. Corrections log
Fosgonimeton is on this site for one reason: it is the properly conducted version of a bet that is also sold as a research chemical. It targets the hepatocyte growth factor and c-Met system, the same system dihexa is sold for, and unlike dihexa it went into registered trials with a sponsor, a protocol and posted results.
That makes it unusually useful to read. Most compounds in this catalogue have no clinical counterpart, so the question of what a proper test would show stays open. Here it does not stay open. A company ran a 554 participant Phase 2/3 in Alzheimer's disease and the registry records what happened.
It is not sold as a research chemical and this site is not aware of it being available to buy. The page exists as evidence rather than as a product guide.
| Property | Value | Source |
|---|---|---|
| Category | Small-molecule prodrug given by subcutaneous injection, developed as a treatment for Alzheimer's disease | NCT04488419 |
| Also known as | ATH-1017, NDX-1017, fosgonimeton acetate | NCT03298672 |
| Sponsor of record | LeonaBio, named as lead sponsor on all six registrations as read on 23 August 2026 | NCT04488419 |
| Largest trial | 554 participants, Phase 2/3, mild to moderate Alzheimer's disease, completed July 2024, results posted to the registry | LIFT-AD, NCT04488419 |
| Primary endpoint | Not met. The registry records this in the termination reason of the open-label extension trial | NCT04886063 |
| Approval status | Not an approved drug in any indication | ClinicalTrials.gov, read 23 August 2026 |
Not medical advice
This page reports what has been documented about a compound. It is not medical advice, not a diagnosis, not a dosing protocol, and not a recommendation to obtain or use anything described here. Talk to a licensed clinician about anything concerning your health. Read the full disclaimer.
Reviewed against editorial standards Updated
Key takeaways
- The Phase 2/3 trial in mild to moderate Alzheimer's disease enrolled 554 participants and did not meet its primary endpoint. The registry states this in the open-label extension's own termination reason, which is a primary source rather than a company summary. [Human RCT] [1] [2]
- The trial posted its results, which is uncommon. The primary outcome can therefore be read directly from the registry: on the Global Statistical Test at week 26, placebo returned a least-squares mean change of -0.126 across 144 participants and the 40 mg arm -0.208 across 143. No statistical analysis and no p-value is posted alongside them. [Human RCT] [1]
- The pre-specified primary analysis population was restricted to participants not taking acetylcholinesterase inhibitors, the standard drugs for the condition. That restriction belongs in any accurate description of the endpoint.
- A second trial, in Parkinson's disease dementia and dementia with Lewy bodies, was terminated at 28 participants. The registry gives the reason as enrollment ending early due to study design limitations. [Human RCT] [5]
- This compound matters to readers of this site mainly as a control. It is the only case where the hepatocyte growth factor and c-Met premise has been examined at a scale capable of settling anything, and a research chemical sold on that premise inherits the answer.
- PARTICIPANTS IN THE PHASE 2/3
- 554
- REGISTERED TRIALS
- 6
- NUMBERED SOURCES
- 12
- EVIDENCE GRADE
- C
Who researches Fosgonimeton?
Most people reach this page from somewhere else on the site rather than by searching for the compound, and there are two routes in.
The first is from dihexa, which is sold on the same mechanism and has no trials of its own. This page is the evidence that page does not have.
The second is a reader following Alzheimer's drug development, for whom the useful content is the posted results section: a Phase 2/3 that reported its primary outcome to the registry rather than only through a press release.
What is Fosgonimeton?
Plain-English version: a small molecule given by injection that is converted in the body into an active form aimed at a growth-factor system in the brain.
Fosgonimeton is a prodrug, meaning it is given in one form and converted in the body into the molecule that does the work. It is administered by subcutaneous injection, which is one of the reasons it is a clinical candidate rather than something sold as a capsule.
Its target is the hepatocyte growth factor and c-Met system. That system is involved in cell survival and in the formation and maintenance of connections between neurons, which is the reason it attracted interest as a route to treating dementia.
It appears in this catalogue under a peptide heading because of the company it keeps rather than its chemistry. It is here as the clinical counterpart to a compound sold as a peptide on the same mechanism [11].
Not a peptide of defined residue sequence; a small-molecule prodrug · Not stated here
How strong is the evidence, claim by claim?
Every claim area this compound is discussed for gets a row, including the ones with the weakest support. The tier records what the row rests on, and the last column states what that work can and cannot show. Each tier links to its definition on the methodology page.
| Evidence Area | What Has Been Studied | Evidence Level | What It Can and Cannot Show |
|---|---|---|---|
| Cognition and function in Alzheimer's disease | LIFT-AD, a Phase 2/3 trial in mild to moderate Alzheimer's disease that enrolled 554 participants and completed in July 2024. The primary outcome was the Global Statistical Test score, a composite of cognition and function built as the average of two change-from-baseline z-scores, measured from baseline to week 26. The pre-specified primary analysis population was participants receiving placebo or 40 mg, in the modified intention-to-treat set, restricted to those not taking acetylcholinesterase inhibitors. Results are posted to the registry [1] [2] [3] | [Human RCT] | The primary endpoint was not met, which the registry states in the termination reason recorded on the open-label extension trial. The posted figures are a least-squares mean change of -0.126 with a standard error of 0.0683 for placebo across 144 participants, and -0.208 with a standard error of 0.0707 for the 40 mg arm across 143. No statistical analysis and no p-value accompanies the primary outcome in the posted results. This site has not established the sign convention used to build the composite and therefore states no direction of effect for either arm on that measure. Two posted secondary outcomes at week 26 point differently from a naive reading of the primary: ADAS-Cog11 change from baseline was -0.39 for placebo and -1.09 for the drug, and ADCS-ADL23 change was -0.02 for placebo and 0.65 for the drug |
| Parkinson's disease dementia and dementia with Lewy bodies | SHAPE, a Phase 2 trial in Parkinson's disease dementia and dementia with Lewy bodies, with the same Global Statistical Test primary outcome. It enrolled 28 participants and completed in April 2023 [5] | [Human RCT] | The trial was terminated. The registry gives the reason as enrollment ending early due to study design limitations, which is the sponsor's own wording. A trial stopped at 28 participants for design reasons does not report on whether the drug did anything, and no conclusion about the indication follows from it |
| Safety and tolerability | A Phase 1 safety, tolerability and pharmacokinetics study in 88 participants, a Phase 2 study in 77 participants with mild to moderate Alzheimer's disease, and a Phase 1 absorption, metabolism and excretion study in 8 participants. All three completed [4] [6] [7] | [Human RCT] | None of the three has posted results, so what they found is not on the public record. The open-label extension that did post results was terminated for a reason unrelated to safety, following the parent trial's endpoint result. Exposure in these trials ran to weeks and months rather than years, so nothing here characterises long-term use |
Rolled up, that puts Fosgonimeton at evidence grade C Human data, wrong question . Completed human evidence exists for this compound, and none of it tested the use the compound is sold for. A grade describes the quality of the research, not whether something works and not whether anyone should use it. The derivation is published.
How Fosgonimeton is thought to work
The mechanism here is a hypothesis that was taken seriously enough to fund a clinical programme, which is a higher bar than most entries in this catalogue clear. That does not make it correct, and the trial section below is the reason to say so.
1. A prodrug given by injection, given in one form and turned by the body into the form that acts (established as the design)
The compound is administered subcutaneously and converted in the body to its active form. A Phase 1 absorption, metabolism and excretion study was run to characterise that conversion, and a separate Phase 1 covered safety, tolerability and pharmacokinetics.
Neither study has posted results, so the specifics of the conversion and the exposure it produces are not on the public record even though the studies were completed. [Human RCT] [4] [7]
2. Activating the hepatocyte growth factor and c-Met system, switching on a growth-factor system involved in keeping brain cells alive and connected (the programme's premise, tested once at scale)
The hepatocyte growth factor and c-Met system supports cell survival and the formation of connections between neurons. Reviews from the middle of the last decade proposed it as a target in Alzheimer's disease on that basis.
What separates this compound from others aimed at the same idea is that the hypothesis was carried through to a trial with an enrolment large enough to detect a clinically meaningful effect if one were there. The result of that trial is in the claims table. [In-vitro] [8] [9] [10] [11] [12]
What we do not know
What the three completed trials without posted results found. Two Phase 1 studies and a Phase 2 study completed and none has reported.
The sign convention of the Global Statistical Test composite, without which no direction of effect can be read from the posted primary figures.
Whether a peer-reviewed paper reporting the Phase 2/3 exists. This site did not search for one.
The nature and date of the sponsor change. The registry names LeonaBio, and this site has not verified the corporate history behind that name.
What this evidence can and cannot show
These results come from Human trials for the clinical work; the target literature is animal and cell culture. in Mild to moderate Alzheimer's disease; Parkinson's disease dementia and dementia with Lewy bodies., not from people. An animal model is a deliberate simplification: the injury is created on purpose, the animal is young and healthy, the dose is scaled to body weight in a way that does not translate directly, and the outcome is measured at a fixed point rather than lived with. Cell and tissue work is a further step removed, because the concentrations that produce an effect in a dish are often far above anything a body reaches. Mechanism is a reason to run a trial. It is not a substitute for one, and compounds that looked mechanistically convincing have failed in people many times.
Has Fosgonimeton been tested in humans?
Every registration for this compound is below, pulled from the ClinicalTrials.gov v2 API on 23 August 2026 and read from the returned records. The sweep queried the compound name and the development codes ATH-1017 and NDX-1017 across both the intervention and free-text fields.
Six records returned and all six are this compound. Two have posted results.
| Study | People | What was done | Result | Evidence level |
|---|---|---|---|---|
| LIFT-AD, NCT04488419 | 554 | Phase 2/3 in mild to moderate Alzheimer's disease. Primary outcome the Global Statistical Test at week 26. Results posted | Completed July 2024. Primary endpoint not met | [Human RCT] |
| NCT04886063 | 423 | Open-label extension of the trial above. Results posted | Terminated. The registry gives the reason as the parent trial not meeting its primary endpoint | [Human RCT] |
| NCT04491006 | 77 | Phase 2 in mild to moderate Alzheimer's disease | Completed. No posted results | [Human RCT] |
| NCT03298672 | 88 | Phase 1 safety, tolerability and pharmacokinetics | Completed. No posted results | [Human RCT] |
| SHAPE, NCT04831281 | 28 | Phase 2 in Parkinson's disease dementia and dementia with Lewy bodies | Terminated. Enrollment ended early due to study design limitations | [Human RCT] |
| NCT05511558 | 8 | Phase 1 absorption, metabolism and excretion | Completed. No posted results | [Human RCT] |
Why we are not calling this proof
Six registrations is a real programme and it is still six registrations rather than six answers. Three completed trials have posted nothing, so what they found is not public.
The two that did post are the pivotal trial and its extension, which is the useful direction for that to run in. It means the headline result can be read from the registry rather than taken from a summary.
Benefits: what the research shows
The claims table carries what was tested and what was found. This section is the shape of the record as a whole.
The honest bottom line on benefits
A pharmaceutical company took a specific hypothesis about dementia into a properly sized trial, posted the results, and the primary endpoint was not met. That is a complete and honest outcome, and it is more than almost any compound on this site can show in either direction.
For a reader who arrived from a compound sold on the same mechanism, the relevant sentence is that the premise has been tested and the test did not succeed. That is not the same as proving the mechanism is wrong, and it is the strongest evidence anyone has.
How long Fosgonimeton stays in the body
This is the rare case on this site where the right studies exist. A Phase 1 covered safety, tolerability and pharmacokinetics in 88 participants, and a separate Phase 1 in 8 participants covered absorption, metabolism and excretion, which is the study type designed to answer where a drug goes and what becomes of it.
Both completed. Neither has posted results to the registry, and this site did not locate a publication reporting either. So the figures exist somewhere and they are not on the public record, which is a different situation from the compounds elsewhere in this catalogue where the studies were never run at all.
No half-life, clearance or exposure figure is quoted on this page, because none is available from a source this site can check.
Two Phase 1 studies were run and neither has posted results. [Human RCT] [4] [7]
What is documented about dosing
Nothing here is a dosing recommendation. Statements about dosing reflect community and research reports only.
Commonly cited protocols (extrapolated, not validated)
Nothing here is a dosing recommendation. Statements about dosing reflect community and research reports only.
| Study | What was given | Frequency | Duration | Notes |
|---|---|---|---|---|
| LIFT-AD, NCT04488419 | Two dose arms against placebo, reported on the hub | Not restated here | 26 weeks to the primary endpoint | 554 participants. Subcutaneous injection. The primary analysis compared placebo with the lower of the two arms |
| NCT04886063 | Open-label extension dosing | As per the extension protocol | Terminated before completion | 423 participants. Results posted |
| NCT03298672 | Phase 1 dose escalation | As per protocol | Short-term | 88 participants. No posted results |
| NCT05511558 | Single dose for absorption, metabolism and excretion | Single dose | Single dose | 8 participants. No posted results |
The figures above describe an investigational medicine given by subcutaneous injection to people with a diagnosis, in a hospital-supervised trial, with monitoring and withdrawal criteria attached. They are a record of what a protocol specified.
Commonly cited protocols (extrapolated, not validated) do not exist for this compound in the way they do for the research chemicals elsewhere in this catalogue, because it is not sold that way. Where a figure from these trials appears attached to a different compound sold on the same mechanism, it has been moved from one molecule to another and it does not carry across.
Nothing here is a dosing recommendation. Statements about dosing reflect community and research reports only.
Main routes people compare, and what each source class actually reported, are set out on the documented protocols page.
Reported effects and what has been measured
What the programme was designed to measure
A Phase 1 study in 88 participants was run specifically for safety and tolerability, and the open-label extension listed incidence of treatment-emergent adverse events as its own primary outcome. Those are the right studies to answer a safety question.
The Phase 1 has posted no results. The extension posted results and was terminated for a reason the registry ties to the parent trial's endpoint rather than to a safety finding. [Human RCT] [4] [2]
What is not established
Long-term exposure has been observed without being characterised. The open-label extension followed 423 participants for up to 173 weeks before being terminated, in a single arm with no comparator, and its adverse event table is set out on the safety record. The controlled record stops at the pivotal trial's 26-week primary endpoint.
The one posted adverse event table belongs to the open-label extension and is reported on the safety record, with the single-arm design that bounds it. The Phase 1 built for safety and tolerability has posted nothing, so the controlled comparison that would sit beside that table is not public. [Human RCT] [1]
What has been measured about Fosgonimeton, and what has never been studied
Talk to a licensed clinician about anything concerning your health.
Sourcing and quality
What a credible product should show
This compound is not sold as a research chemical and this site is not aware of a consumer supply. Anything offered under this name or under the code ATH-1017 is not the investigational product used in the registered trials.
The trial material was a manufactured investigational medicine with a regulatory file behind it. That is a different product class from a synthesised research chemical, and the two are not interchangeable because they share a name.
Red flags
This compound's trials cited on a page selling a different molecule. The trials belong here and the primary endpoint was not met.
The phrase Phase 3 used without the result. The Phase 2/3 completed and did not meet its primary endpoint.
A figure from these dose arms attached to an orally taken research chemical. The route, the molecule and the setting are all different.
Will it show on a drug test?
This compound is not named on the WADA 2026 Prohibited List. That is not a clearance: the growth-factor class at S2.3 closes with a catch-all covering other growth factors and growth factor modulators, and this site has not established how it applies to a c-Met-directed small molecule.
An athlete should treat the absence of a name as an open question and take it to the relevant anti-doping authority.
Storage
No storage guidance is given here. Storage belongs to the manufactured investigational formulation, which is not available outside the trials.
Regulatory status, as of 23 August 2026
| Body | Position | Date | Document |
|---|---|---|---|
| No agency | Not an approved drug in any indication | Read 23 August 2026 | ClinicalTrials.gov |
| ClinicalTrials.gov | Six registrations, all this compound. Two carry posted results | Swept 23 August 2026 | ClinicalTrials.gov v2 API |
| ClinicalTrials.gov | The Phase 2/3 primary endpoint was not met, recorded in the open-label extension's termination reason | Extension completed October 2024 | NCT04886063 |
| Sponsor of record | LeonaBio on all six registrations as read. The corporate history behind that name has not been verified by this site and is not stated | Read 23 August 2026 | ClinicalTrials.gov |
There is no agency approval to report and no compounding position, because this compound has never been sold and has not been nominated for compounding. Its regulatory record is its trial registrations.
The most useful regulatory fact for a reader is the one in the third row. A registry field recording why a trial stopped is a primary source, and it says the parent trial did not meet its primary endpoint in the registry's own words rather than a summary of them.
This section is dated and re-checked on review. It records actions and their documents, not evidence about an effect, and carries no evidence tier.
Fosgonimeton compared with other cognitive peptides
| Compound | Evidence grade | What the human record covers | Read more |
|---|---|---|---|
| Dihexa | Graded on its own page | The same target sold as a research chemical, with no registered trial of any kind and a retracted mechanism paper | /peptides/dihexa/ |
| Cerebrolysin | Graded on the cognitive group page | A different approach to the same clinical problem, with its own separate evidence base | /peptides/groups/cognitive/ |
| Acetylcholinesterase inhibitors | Not covered as compounds on this site | The approved standard drugs for Alzheimer's disease. The Phase 2/3 primary analysis population deliberately excluded people taking them | Described in the claims table above |
The comparison this page exists for is with dihexa. Two molecules, one biological premise, and two records that could hardly be more different: six registrations and posted results against none at all.
The third row is not a comparison so much as a caution about reading the trial. Excluding people on the standard drugs changes what the primary endpoint is measuring, and any description of that endpoint that leaves the exclusion out is incomplete.
What Fosgonimeton typically costs
Frequently asked questions
What is fosgonimeton?
A small-molecule prodrug given by subcutaneous injection that targets the hepatocyte growth factor and c-Met system, developed as a treatment for Alzheimer's disease and tested in six registered trials.
What did the largest trial find?
It did not meet its primary endpoint. The Phase 2/3 in mild to moderate Alzheimer's disease enrolled 554 participants and did not meet its primary endpoint. The registry records this in the termination reason of the open-label extension trial.
What were the actual numbers?
On the Global Statistical Test at week 26, placebo returned a least-squares mean change of -0.126 with a standard error of 0.0683 across 144 participants, and the 40 mg arm -0.208 with a standard error of 0.0707 across 143. No p-value is posted alongside them.
Does that mean the drug was worse than placebo?
This site does not say so. The measure is a composite of two z-scores and the sign convention used to build it has not been established here, so no direction of effect is stated. Two posted secondary outcomes point differently from a naive reading of the primary, which is a further reason not to narrate a direction.
Why did the trial exclude people on standard Alzheimer's drugs?
The registry records the pre-specified primary analysis population as those not taking acetylcholinesterase inhibitors. This page reports that restriction because it changes what the endpoint measures; it does not speculate about why it was chosen.
What happened to the Parkinson's trial?
It was terminated at 28 participants. The registry gives the reason as enrollment ending early due to study design limitations, which is the sponsor's own wording.
How many of the trials reported results?
Two of six. The pivotal Phase 2/3 and its open-label extension both posted results to the registry. The two Phase 1 studies and the smaller Phase 2 completed without posting.
Is it the same thing as dihexa?
No. They are different molecules aimed at the same system. Dihexa is sold as a research chemical and has no registered trial; this compound has six and posted results from the largest.
Does this trial result carry over to dihexa?
Not as evidence, because a trial of one molecule is not a trial of another. It does mean the shared premise has been examined at a scale capable of detecting a meaningful effect, and the examination did not succeed.
Can I buy it?
This site is not aware of any legitimate supply. It is unapproved and investigational, and anything sold under this name is not the trial material.
Who makes it?
The registry names LeonaBio as lead sponsor on all six registrations as read on 23 August 2026. This site has not verified the corporate history behind that name and does not describe it.
References
- LeonaBio. ATH-1017 for Treatment of Mild to Moderate Alzheimer's Disease (LIFT-AD). Phase 2/3, 554 participants, completed 15 July 2024. Results posted. Primary outcome the Global Statistical Test score at week 26; primary analysis population placebo or 40 mg, modified intention-to-treat, restricted to participants not taking acetylcholinesterase inhibitors. Posted least-squares means: placebo -0.126 (SE 0.0683, n=144), ATH-1017 40 mg -0.208 (SE 0.0707, n=143). No analysis or p-value posted for the primary outcome. Read through the v2 API on 23 August 2026. ClinicalTrials.gov. 2020. NCT04488419
- LeonaBio. Open Label Study of ATH-1017 for Treatment of Mild to Moderate Alzheimer's Disease. Phase 2/3, 423 participants, terminated, completed 23 October 2024. Results posted. The registry's whyStopped field reads: ATH-1017-AD-0203 was terminated after the Ph2/3 parent trial (LIFT-AD; NCT04488419) did not meet its primary endpoint, the GST score. ClinicalTrials.gov. 2021. NCT04886063
- LeonaBio. A Study of ATH-1017 in Mild to Moderate Alzheimer's Disease. Phase 2, 77 participants, completed. No posted results. ClinicalTrials.gov. 2020. NCT04491006
- LeonaBio. Safety, Tolerability, and Pharmacokinetics Study of ATH-1017. Phase 1, 88 participants, completed. No posted results. ClinicalTrials.gov. 2017. NCT03298672
- LeonaBio. ATH-1017 Treatment in Subjects With Parkinson's Disease Dementia or Dementia With Lewy Bodies (SHAPE Trial). Phase 2, 28 participants, terminated, completed 19 April 2023. The registry's whyStopped field reads: Enrollment ended early due to study design limitations. ClinicalTrials.gov. 2021. NCT04831281
- LeonaBio. A Study of the Absorption, Metabolism, and Excretion of ATH-1017. Phase 1, 8 participants, completed. No posted results. ClinicalTrials.gov. 2022. NCT05511558
- ClinicalTrials.gov. Registry sweep for fosgonimeton, ATH-1017, NDX-1017 and fosgonimeton acetate across the intervention and free-text fields, run through the v2 API on 23 August 2026. Six distinct records returned, all of them this compound. ClinicalTrials.gov. 2026. ClinicalTrials.gov
- Wright JW, Harding JW. The Brain Hepatocyte Growth Factor/c-Met Receptor System: A New Target for the Treatment of Alzheimer's Disease. J Alzheimers Dis. 2015. PMID 25649658
- Wright JW, Kawas LH, Harding JW. The development of small molecule angiotensin IV analogs to treat Alzheimer's and Parkinson's diseases. Prog Neurobiol. 2015. PMID 25455861
- Ho JK, Nation DA. Cognitive benefits of angiotensin IV and angiotensin-(1-7): A systematic review of experimental studies. Neurosci Biobehav Rev. 2018. PMID 29733881
- Benoist CC, Kawas LH, Zhu M, Tyson KA, Stillmaker L, Appleyard SM, Wright JW, Wayman GA, Harding JW. The procognitive and synaptogenic effects of angiotensin IV-derived peptides are dependent on activation of the hepatocyte growth factor/c-met system. J Pharmacol Exp Ther. 2014. PMID 25187433 DOI 10.1124/jpet.114.218735 Retracted April 2025. Cited here only to identify the mechanism literature shared with the research chemical sold on this target, and not as evidence.
- Siller R, Greenhough S, Naumovska E, Sullivan GJ. Small-molecule-driven hepatocyte differentiation of human pluripotent stem cells. Stem Cell Reports. 2015. PMID 25937370