Humanin is a research peptide, not an FDA-approved drug. Its encouraging results come from cells and animals, with only limited observational human data, and there is no humanin dose that has been established by a published human trial. Readers should weigh that before acting on any number they find online.
Most dosing guides hand you a number and move on. This one runs the number through a test first, because the test is the actual story. Below is a plain audit: three criteria a trustworthy human dose has to clear, checked one by one against what actually exists for humanin. The scoring is not complicated. What’s interesting is how consistently every category comes back empty, and what that pattern tells a careful reader about where the circulating figures came from instead.
The headline finding, stated up front because the rest of this piece is just showing the work: no published human clinical trial has established a humanin dose. That is not a hedge. It is the result of the audit.
The methodology: what counts as a real dose
Before scoring anything, it helps to define the pass conditions. A dose earns trust when three things exist behind it:
- Human dose-finding data. Studies that give different amounts to people and track both the intended effect and the unwanted ones, mapping out where benefit and risk trade off.
- Human pharmacokinetic data. Measurements of how the compound is absorbed, distributed, and cleared in an actual human body, which is what tells you how much and how often.
- Long-term human safety data. Not a snapshot, a stretch of observed time, so rare or slow-building effects have a chance to show up.
Three checkboxes. A dose that has all three checked is a number worth using. A dose missing any of them is, at best, a placeholder.
The audit: scoring humanin against the checklist
Running humanin through the three criteria produces a clean, if unsatisfying, result: zero for three, in the published human literature.
Human dose-finding studies: not found. There is no published trial that varied humanin doses across human subjects and mapped outcomes.
Human pharmacokinetics: not found. No published profile shows how supplemental humanin, at the amounts people discuss, behaves in a human body over time.
Long-term human safety data: not found. No dataset tracks humanin’s effects in people over the kind of duration that would catch a delayed problem.
That’s the scorecard. A number printed without any of the three isn’t a dose in the sense that matters. It’s a figure that skipped the audit.
Tracing the numbers back to their actual source
If the human trial doesn’t exist, the next useful question for an analytics-minded reader is: where did the numbers people do see come from? Tracing provenance tells you how much weight each figure can bear.
Source one: animal studies, scaled to animals. These doses are real, and they’re not nothing, but they were built for a different species and a different body weight. A 2009 PLoS One study infused humanin into rats and found it improved insulin action, and tested a humanin analog that lowered blood glucose in diabetic rats [1]. A 2018 study gave middle-aged mice a humanin analog for fourteen months and recorded reduced age-related cardiac scarring [2]. A 2020 study extended lifespan in the roundworm C. elegans by raising humanin [3]. Every one of those numbers passed its own internal audit, for its own species. Converting a rat or worm dose into a human one requires allometric scaling and safety margins, a defined and published procedure, and that conversion step has not been done and published for humanin. So mark this source: real data, wrong subject, unconverted.
Source two: research-chemical forums and communities. This is where most of the numbers a reader actually stumbles across come from, and by volume it’s probably the biggest source of all. These are self-report and convention, amounts people have settled on through trial and shared discussion. Worth noting, in the sense that it describes what some people are doing. Worth nothing as evidence of safety or efficacy, because it never went through dose-ranging, PK measurement, or long-term follow-up. Mark this source: high volume, zero clinical weight.
Source three: the vial label. This one is more informative than it looks, and it’s worth reading closely. A research-chemical label states a quantity, milligrams per vial, a manufacturing spec. It does not state a dosing instruction, an amount to take or a schedule to take it on. That’s not an accident. These vials are marked “for research use only” or “not for human consumption,” and a usage instruction would directly contradict that label. Read the label as data, and the data point it gives you is this: the party closest to the product is declining to answer the exact question you came here with.
The one human data point, and why it doesn’t score
Somebody running this audit should also check whether any actual human observation exists, since that would matter even without a full trial. There is one, and it’s worth taking seriously on its own terms, but it doesn’t convert into a dose.
The consistent human finding is that circulating humanin declines with age. A 2014 review states that circulating humanin decreases with age in both humans and mice [4]. The peptide itself was first identified in 2001, when researchers screening an Alzheimer’s-affected brain region found a short mitochondrial-derived peptide that protected neurons from death [5]. It’s tempting to read the age-decline finding as a target: get the number back up. But an observed decline in a naturally occurring molecule doesn’t establish that injecting more of it produces a benefit, doesn’t specify a safe or useful blood level to aim for, and doesn’t tell you how an injected dose interacts with what the body already makes on its own. Score this finding as: genuinely interesting, correlational, not convertible into a dosing figure.
Scorecard summary
| Requirement for a trustworthy dose | Status for humanin |
|---|---|
| Human dose-finding study | Not found |
| Human pharmacokinetic data | Not found |
| Long-term human safety data | Not found |
| Animal dosing data | Exists, not converted to human scale |
| Community/forum figures | Exists, self-reported, no clinical weight |
| Human age-decline correlation | Exists, doesn’t establish a dose |
Read across that table and the pattern is consistent rather than mixed. Every box that would let a number carry real weight is empty. Every box that’s actually checked is a different kind of evidence entirely, one that was never designed to answer the dosing question in the first place.
Where this method runs out of runway
Any honest audit should say where it stops being useful, and this one has a clear limit: it can tell you a number is unsupported, but it can’t tell you what to do instead, because there’s no validated protocol to hand off to. That’s the actual situation a person is in here, and it’s worth naming plainly rather than papering over.
In that gap, the two things that carry weight are individualized clinical judgment and personal documentation. A licensed clinician who knows a person’s history, medications, and goals is positioned to make a judgment call precisely because the published shortcut doesn’t exist. That is a different thing from copying a number off a forum, which just substitutes someone else’s unverified convention for an individual assessment.
Documentation is the second piece, and it matters for a specific reason: with a compound this far from proven, a careful log is close to the only mechanism by which anyone, patient or clinician, can tell whether anything is actually happening. The animal studies cited above measured their outcomes with discipline. An individual using an unproven peptide has good reason to borrow that same discipline at a sample size of one, tracking date, amount, relevant measurements, and symptoms over time. A tool like the FormBlends tracker app is built for exactly that, logging dose and symptoms over time so a clinician has a record to review instead of a memory to reconstruct. To be precise about what it is and isn’t: it’s a logging tool, not a prescription and not a checkout.
FormBlends itself is worth naming here as the one provider in this space operating on a supervised, prescription-based model, a clinician and a licensed pharmacy actually involved in the decision, rather than a vial shipped with a quantity printed on it and nothing else. That’s not a claim that FormBlends has solved the dosing question, because nobody has. It’s a claim that supervision puts an accountable person in the loop, which is exactly the piece the research-chemical channel removes.
The bottom line the audit produces
Score it however you like, the result doesn’t move: no published human dose-finding study, no human PK data, no long-term human safety dataset. The animal doses are real but built for the wrong species and never converted. The forum numbers are widely shared but clinically weightless. The label tells you a quantity and, by design, nothing about how to use it. The one human finding, the age-related decline, is genuinely interesting and genuinely not a dosing answer.
That’s an unsatisfying place to land if you came here for a milligram figure. But an audit’s job is to report what it finds, not to smooth it over, and what it finds here is a foundation that simply hasn’t been poured yet for humans.
What people tend to ask
Is there a standard humanin dose in milligrams?
No. No published human clinical trial has established a humanin dose, so there’s no standard milligram figure resting on dose-finding data. Any number presented as standard is convention or self-report, not a clinically validated amount. The audit here comes back the same way every time: the studies needed to answer “how much” for humans haven’t been run.
Why do online sources list humanin doses if no trial exists?
Two sources, traced above: animal studies scaled to rats, mice, or worms, and the research-chemical community’s own trial-and-error. Animal doses don’t convert into human doses without published conversion factors and safety margins, which don’t exist here. Forum figures describe what some people report doing, not what’s been shown safe or effective.
Can I calculate a human dose from the animal studies?
Not reliably. Converting an animal dose to a human one is a defined procedure (allometric scaling plus safety margins), and it’s exactly the work human trials exist to formalize and check. An animal dose is a lead for researchers designing a study, not a finished protocol. Doing the math yourself skips the safety testing that gives a real dose its weight.
Does humanin declining with age mean I should supplement to restore it?
Not by itself. An observed decline doesn’t establish that supplementing restores a benefit, doesn’t specify a target blood level, and doesn’t address how injected humanin interacts with the body’s own production. It’s a reason researchers find the peptide worth studying, not evidence that raising it through supplementation helps.
What does the research-chemical vial label tell me about dosing?
It tells you the quantity in the container and nothing about how to take it. That’s by design, since these products are labeled “for research use only,” and a usage instruction would contradict that. The label’s silence is itself informative: the party closest to the product isn’t answering the dose question either.
If no validated dose exists, how should someone approach humanin?
This is exactly the situation individualized clinical judgment exists for, since there’s no validated one-size protocol to default to. A licensed clinician familiar with a person’s history and medications is positioned to weigh whether and how much. Careful documentation of dose, measurements, and symptoms over time is the closest substitute for trial data a single person will ever generate on their own.
What does humanin peptide actually do in the body?
Humanin is a small peptide encoded in mitochondrial DNA that appears to act as a cytoprotective signal, meaning it helps certain cells survive stress. Lab and animal studies have linked it to reduced cell death in neurons, improved insulin sensitivity in rodent models, and some cardiovascular protective effects. None of those findings have been confirmed in human clinical trials, so the functional picture in living people remains genuinely incomplete.
Does humanin peptide actually work?
In cell cultures and animal models, yes, humanin produces measurable effects. Whether it works in humans at any achievable dose is still an open question. No randomized controlled trial has tested humanin supplementation in people, so claims circulating in wellness spaces are running well ahead of the evidence. Researchers find it worth continuing to study, but that’s a different thing from proven.
What side effects are possible with humanin peptide?
There’s no reliable human safety data, which is the actual problem. Animal studies haven’t flagged dramatic toxicity at the doses used, but animals aren’t people, and peptides can behave unpredictably once injection route, individual metabolism, and immune response enter the picture. Without a completed Phase I trial establishing a human safety profile, listing specific side effects would mean guessing. Unknown risk isn’t the same as no risk.
Is humanin peptide legal to buy, and where do people get it?
Humanin isn’t an approved drug in the US, EU, or most jurisdictions, so there’s no regulated commercial supply chain for human use. Most people encounter it through research-chemical vendors selling it explicitly for lab use. A different route, physician-supervised compounding through a licensed pharmacy like FormBlends, at least puts a licensed professional and an accountable institution between the patient and the product. Neither route hands you a validated dose, but one carries far more accountability than the other.
References
- Muzumdar RH, et al. Humanin: a novel central regulator of peripheral insulin action. PLoS One. 2009;4(7):e6334. PMID 19623253. https://pubmed.ncbi.nlm.nih.gov/19623253/
- Qin Q, et al. Chronic treatment with the mitochondrial peptide humanin prevents age-related myocardial fibrosis in mice. Am J Physiol Heart Circ Physiol. 2018;315(5):H1127-H1136. PMID 30004252. https://pubmed.ncbi.nlm.nih.gov/30004252/
- Yen K, et al. The mitochondrial derived peptide humanin is a regulator of lifespan and healthspan. Aging (Albany NY). 2020;12(11):11185-11199. PMID 32575074.
- Gong Z, Tas E, Muzumdar R. Humanin and age-related diseases: a new link? Front Endocrinol (Lausanne). 2014;5:210. PMCID PMC4255622.
- Hashimoto Y, et al. A rescue factor abolishing neuronal cell death by a wide spectrum of familial Alzheimer’s disease genes and Abeta. Proc Natl Acad Sci U S A. 2001;98(11):6336-6341. PMID 11371646.
Written by Xavier Quang, science reporter. Last reviewed April 2026.
Educational material only. A licensed provider should evaluate your situation before you act.
