Dihexa: What It Is, Where to Get It Prescribed, and What It Costs

Dihexa is a very small peptide built from a fragment of angiotensin IV, itself a breakdown product of angiotensin II, the hormone most people know from blood-pressure medicine. It circulates in the brain as well as the bloodstream. A group at Washington State University led by Joseph Harding and John Wright reported that angiotensin IV and close relatives improved memory-task performance in rats, and set out to build a version that would survive in the body long enough to be useful.

Dihexa is the result. Its full chemical name is N-hexanoic-Tyr-Ile-(6)-aminohexanoic amide: a tyrosine and an isoleucine from the parent hormone, capped with a six-carbon fatty chain at one end and a six-carbon linker at the other. Those additions make it "metabolically stabilised" — the naked fragment is destroyed by enzymes in minutes, the modified one is not. The 2013 paper in the Journal of Pharmacology and Experimental Therapeutics that introduced it describes the design, its stability in rat blood, and its ability to cross into the brain.

It is not a hormone in the usual sense, not a steroid, and not a stimulant. It does not act on dopamine or noradrenaline the way attention medicines do.

"Dihexa" is lab shorthand for the two hexanoic groups on either end of the molecule — di-hexa. It is sometimes written as PNB-0408, the code from the company the Washington State group set up to develop it. Compounding pharmacies label it dihexa acetate, the salt form. All three names refer to the same substance.

Dihexa is not an approved treatment for Alzheimer's disease, dementia, or any memory disorder, in any country, and it has not been through human trials for those conditions or anything else. It is not a retail nootropic supplement and not a stronger racetam — it works on a different system. And it is not a substitute for a medical evaluation of memory complaints, which can have treatable causes (thyroid, B12, sleep apnoea, depression, medication side effects) that need to be found first.

The Washington State group's starting point was the observation, replicated across several labs, that angiotensin IV and stabilised analogues of it improve spatial learning in rodents, apparently through the hippocampus. A 2018 systematic review in Neuroscience and Biobehavioral Reviews gathered the experimental studies on angiotensin IV and the related fragment angiotensin-(1-7) and concluded the procognitive effect in animals is consistent across models. That review is not about Dihexa specifically, but it is the independent foundation the compound rests on; Dihexa was built to be a longer-lasting, brain-penetrant version of that signal.

Harding and Wright's central claim was that angiotensin IV analogues, including Dihexa, work by binding hepatocyte growth factor (HGF) and making it a more effective activator of its receptor, c-Met. HGF and c-Met are best known from liver regeneration and cancer biology, but the receptor is also present on hippocampal neurons, and the group argued in a 2015 review in the Journal of Alzheimer's Disease that this system is a plausible cognitive target because it drives the growth of new synaptic connections.

The two laboratory papers carrying that claim — a 2012 paper describing Dihexa as an HGF/c-Met modifier, and a 2014 paper reporting that its procognitive and synapse-forming effects depend on c-Met activation, both in the Journal of Pharmacology and Experimental Therapeutics — received expressions of concern in 2021 and were retracted in April 2025. The 2013 paper that introduced the compound carries a 2021 expression of concern and has not been retracted. This does not mean the mechanism is wrong; it means the strongest evidence for it can no longer be cited, and HGF/c-Met should be treated as the developing lab's hypothesis rather than an established finding.

The most useful independent data comes from a separate group. A 2021 paper in Brain Sciences gave Dihexa to APP/PS1 mice — an engineered strain that accumulates amyloid plaques and develops memory deficits — and reported improved memory-task performance alongside increased signalling through PI3K/Akt, a downstream route shared by several growth-factor receptors including c-Met. It is one study, in mice, from one lab, but it was not produced by the compound's developers and it broadly points the same way.

The claim that Dihexa causes cultured hippocampal neurons to sprout new dendritic spines — the structures where synapses form — is why the compound gets described as helping brain cells "form new connections." That finding came from the retracted 2014 paper; the 2015 review in Progress in Neurobiology by the same authors summarises it alongside the rodent behavioural work. Until it is reproduced independently, the fair description is that spine formation was reported by the developing lab in cell culture and not confirmed elsewhere.

The 2013 introductory paper reports that Dihexa, by mouth or by injection, improved performance in the Morris water maze — a swimming task that measures spatial memory — in rats whose memory had been impaired with scopolamine, and in aged rats. Those are the standard first-line models for a candidate cognitive drug, and the reason the compound attracted attention. The paper carries an expression of concern, which readers should weigh.

The 2021 Brain Sciences study in APP/PS1 mice is the strongest independent result: treated mice performed better on memory testing than untreated littermates. The interpretation offered is protection of neuronal signalling rather than removal of plaque.

A 2024 paper in the Journal of Huntington's Disease examined an angiotensin IV analogue in rats given 3-nitropropionic acid, a toxin that produces movement and behavioural changes resembling Huntington's disease, and reports effects on some of those measures. This is early, single-study work and not evidence for any human neurological condition.

Reviews by Wright and Harding — in Frontiers in Endocrinology in 2013 and in Progress in Neurobiology and the Journal of Alzheimer's Disease in 2015 — set out the case that the brain's angiotensin system is a target for Alzheimer's and Parkinson's disease, with Dihexa as the lead compound. They are well argued, and written by the compound's inventors.

None. There are no randomised controlled trials of Dihexa in humans, no published open-label trials, and no published case series. The FDA's own summary states that it has not identified any human exposure data on drug products containing dihexa acetate by any route. What exists in people is prescriber experience, which is real but unpublished and unmeasured.

That is a thinner position than most peptides on this site. BPC-157 has early-phase human work and decades of animal literature from many labs; Dihexa has rodent studies, a few labs, and a retraction history. It is prescribed on mechanism plus a physician's judgement about a particular patient, and should be described that way — "studied in animals," not "shown to work."

• It does not establish that Dihexa improves memory, attention or thinking in humans. That trial has not been run.
• It does not establish a human dose. The doses used are extrapolated by prescribers, not derived from a dose-finding study.
• It does not establish safety in humans over any duration, and it does not answer the question raised by its proposed mechanism: HGF/c-Met is a growth pathway active in many cancers, and nobody has measured what long-term stimulation of it does in people.
• It does not establish that the c-Met mechanism is correct at all, now that the papers supporting it have been withdrawn.
• It does not establish efficacy for any condition as an FDA-approved treatment. It is not approved for anything.

Most people report nothing specific. Some describe feeling slightly sharper or sleeping differently; some describe mild headache or a wired feeling early on. None of this has been measured in a study, and noticing nothing in the first fortnight is not a sign the course is failing.

If a subjective change in recall or focus is going to be reported, prescribers generally describe it in this window, which is why a vial is sized at about four weeks of weekday dosing. Anyone expecting a dramatic effect is setting themselves up for disappointment; the honest framing is modest and uncertain.

People who start with a specific, checkable goal — names at work, a language, a task they had been struggling with — can usually say whether anything changed. "General brain fog" is much harder to judge, so agree with your physician beforehand what you will be watching for.

No withdrawal or rebound is described in the animal literature. Whether any subjective change persists once dosing stops is unknown. Because the proposed mechanism is a growth pathway and the long-term human safety picture is blank, most prescribers structure this as a defined course with a review rather than indefinite use.

The directions are written in millilitres and in insulin-syringe units because that is what you can read off the barrel. 0.25 mL is 25 units on a U-100 syringe. You are not calculating anything — the number is printed on your medication. A different volume is a physician decision, never a self-adjustment.

Subcutaneous means into the fat layer, not the muscle. Abdomen and thigh are the usual sites, rotated so the same spot is not used repeatedly. There is no reason to inject near the head — the compound is designed to cross into the brain from the bloodstream, and the animal work used ordinary systemic routes.

The reviewed directions are once daily on weekdays with weekends off. That is a prescriber convention, not something derived from a human study. The practical effect is 20 doses per vial covering about four weeks, which lines up with the 28-day refill cycle and gives a monthly checkpoint for you and your physician to decide whether to continue. One vial per fill, always.

Refrigerated, and it ships that way in insulated packaging with ice packs. Beyond-use dating runs from first puncture and is on the label. Keep it out of light and do not freeze it.

Different systems entirely. Semax is a fragment of ACTH described around BDNF and attention, with a history of human use in its country of origin; Dihexa is an angiotensin IV analogue described around growth-factor signalling with no human use on record. Semax is the more conservative choice, available as an injection or nasal spray, and a physician cautious about Dihexa's evidence base will often suggest it first.

Selank is a tuftsin analogue described around anxiety and stress tolerance rather than memory formation. People whose "brain fog" is driven by anxiety usually do better starting there. The two are not interchangeable and are occasionally prescribed together.

NAD+ is a coenzyme with human trial data and a well-understood biology, aimed at cellular energy rather than synapse formation. If the complaint is fatigue showing up as poor concentration, NAD+ is the more direct conversation. If it is specifically recall, Dihexa is the one built for that, with the caveats above.