AOD-9604: The Compound That Skipped the Appetite Circuit

The short version

AOD-9604 is a 16-amino-acid piece of human growth hormone — specifically the tail end of the hormone, the part associated with fat metabolism, rebuilt as a standalone molecule. The design goal was to keep the fat-burning behaviour of growth hormone while leaving behind the growth-promoting and blood-sugar effects of the whole hormone. It does not bind the growth hormone receptor.

It also does not act on appetite. Nothing in its described mechanism reaches the hypothalamus or the brainstem. Its target is the fat cell: less fat synthesis, more fat oxidation.

In mice this worked. Chronic treatment reduced body weight and fat ^[16] and increased fat oxidation ^[17]. In humans it did not translate. Across roughly six trials totalling about 900 adults with obesity, tolerability was essentially indistinguishable from placebo ^[15] — and the obesity development programme was discontinued after the pivotal trial failed to show a significant weight reduction against placebo.

That is the honest summary: a well-tolerated compound with a real preclinical mechanism and no demonstrated human weight-loss effect. It is included here as the instructive contrast, not as a third option.

What it is

Class. A growth-hormone-fragment lipolytic peptide — investigational, and used as research material rather than as a medicine.

Structure. A synthetic hexadecapeptide analogue of the C-terminal lipolytic domain of human growth hormone, modelled on residues 177-191 with an N-terminal tyrosine substituted for the native phenylalanine. Two cysteines form an intramolecular disulfide bridge that reproduces the cystine loop of the parent hormone. The reported sequence is Tyr-Leu-Arg-Ile-Val-Gln-Cys-Arg-Ser-Val-Glu-Gly-Ser-Cys-Gly-Phe.

Regulatory status. Not approved by the FDA, EMA or TGA for any indication. The original anti-obesity development programme was discontinued after the pivotal Phase IIb trial failed to meet its primary endpoint. AOD-9604 was reviewed by the FDA's Pharmacy Compounding Advisory Committee in December 2024 regarding eligibility for 503A compounding; current status should be verified independently rather than assumed from this page.

Anti-doping status. Prohibited at all times in sport under Section S2 of the WADA Prohibited List as a growth-hormone fragment, and detectable by dedicated assays even though it does not interfere with the standard growth-hormone isoform immunoassay. This is a hard difference from the two incretin compounds on this site.

Human-consumption status. Sold and described for research purposes only. Not approved for human consumption or any human therapeutic indication, and not a controlled substance.

How it works: adipose tissue, not appetite

The mechanism is genuinely different from everything else in this briefing, and reading it carefully explains the outcome.

Antilipogenic before lipolytic. The foundational literature characterises the growth-hormone C-terminal fragment as primarily antilipogenic — inhibiting the synthesis of new fat via acetyl-CoA carboxylase — rather than directly lipolytic. Popular framing as a fat-burning peptide overstates the direct-lipolysis reading of the same data.

Beta-3 adrenergic signalling. In obese mice, 14 days of chronic treatment with either human growth hormone or AOD-9604 reduced body weight and fat and increased beta-3 adrenergic receptor RNA expression in adipose tissue. In beta-3 receptor knockout mice the chronic weight and lipolytic response was abolished, while acute increases in energy expenditure and fat oxidation persisted ^[16]. The chronic effect therefore depends on functional beta-3 adrenergic signalling; the acute one does not.

Fat oxidation. Chronic treatment with growth hormone or the modified C-terminal fragment increased fat oxidation and produced weight loss in obese mice ^[17].

Receptor-sparing by design. It does not bind the growth hormone receptor, which is why the elevated IGF-1, fluid retention and diabetogenic effects of the intact hormone are not expected — and why community reports consistently describe none of them.

Pharmacokinetics. Non-clinical work found an intravenous half-life of roughly 3 minutes, with degradation proceeding by sequential removal of N-terminal amino acids, plus demonstrated oral absorption and no genotoxic or toxicological concerns after chronic administration in rats and primates ^[14]. Subcutaneous human pharmacokinetics have not been published in the peer-reviewed literature.

What is absent from this list. There is no arcuate nucleus, no area postrema, no vagal afferent, no gastric emptying, and no meal-termination circuit. Whatever AOD-9604 does, it does not do it by making a meal end sooner.

What the research shows — and at what grade

The evidence here is graded very differently from the two incretin compounds, and collapsing the grades would misrepresent it.

Rodent efficacy — the strongest positive signal. Chronic intraperitoneal treatment in obese mice reduced body weight and fat and raised beta-3 adrenergic receptor expression, with the chronic response abolished in beta-3 knockout animals ^[16]. A companion study reported increased fat oxidation and weight loss in obese mice after chronic treatment ^[17]. Both are animal studies. Neither is human evidence.

Non-clinical safety. Evaluation found the compound free of genotoxic and toxicological concerns and generally safe after chronic oral administration in rats and primates, with the pharmacokinetics described above ^[14].

Human tolerability — the clean result. Across approximately six clinical trials totalling about 900 adults with obesity, oral AOD-9604 administered at daily amounts ranging from 0.25 mg to 54 mg for durations of 7 days to 24 weeks showed a safety and tolerability profile indistinguishable from placebo, without the adverse effects associated with full-length growth hormone ^[15]. This is a real and reasonably large human dataset — and it is a safety dataset.

Human efficacy — the missing result. The pivotal Phase IIb obesity trial did not demonstrate statistically significant weight loss against placebo, and the development programme was discontinued in consequence. There is no published human trial showing that this compound reduces body weight. The longest published human exposure runs to about 24 weeks ^[15], so there is also no long-term human data of any kind.

Joint and cartilage work — preclinical only. In a collagenase-induced knee osteoarthritis model in 32 New Zealand white rabbits, weekly intra-articular injection of 0.25 mg AOD9604, with or without 6 mg hyaluronic acid, for 4 to 7 weeks reduced gross morphological and histopathological cartilage-degeneration scores against saline control ^[13]. This is a rabbit model. No human osteoarthritis trial of AOD-9604 has been published, and this line of work should not be read as a second indication in waiting.

The pattern. Strong mechanism, real animal effect, clean human safety, absent human efficacy. That is a recognisable shape in obesity pharmacology, where many mechanistically promising candidates have failed to translate.

Reported effects, cautions, and what to watch

Community reports. The following is drawn from fat-loss and biohacker community discussion. It is anecdotal, not clinical evidence, it carries no doses, and it is unusual in this briefing for pointing the same direction as the trial data.

The most common report is the absence of an effect: people describe seeing no meaningful reduction in body fat, which is consistent with the human obesity trials that did not beat placebo. Where any visible change is reported, it is almost always credited to a calorie deficit and training programme running at the same time, which makes attribution to the peptide impossible. Tolerability is described as easy, with few day-to-day complaints — again matching the published finding. People with growth hormone experience specifically note the absence of water retention, puffiness, joint swelling and carpal-tunnel-type tingling, and the absence of any change in strength, recovery or muscle, all of which fits a molecule that does not engage the growth hormone receptor. A subset describe a vague lift in energy or mood, and some report eating slightly less; both are easily explained by expectation or concurrent diet changes, and neither is an outcome the trials supported. Occasional self-resolving injection-site redness is reported.

Two community themes deserve to be named directly. The belief that injecting near a stubborn area produces fat loss at that spot is not supported by any human trial and is not biologically plausible; bodies do not lose fat by injection location. And experienced users repeatedly caution that research-grade vials vary widely in identity and purity, so any reported effect — or absence of one — may reflect the contents rather than the molecule.

Cautions from the literature. It is investigational and not approved for any use, so no approved indication, dosing standard or quality standard attaches to it. Human weight-loss efficacy was not demonstrated, so expectations of fat loss are not supported by the clinical evidence. The mechanism is largely preclinical and indirect: the acetyl-CoA carboxylase and beta-3 adrenergic findings were characterised chiefly in mouse, rat and cell models ^[16][17], and animal efficacy is not human benefit — this compound is itself the example of that gap. Long-term human safety data are limited; reported exposure came from a finite set of trials running to roughly 24 weeks ^[15], with no large-scale surveillance, so chronic and rare risks remain uncharacterised. And it is prohibited in sport at all times.

Where it sits in the satiety picture

AOD-9604 is the control case for this entire briefing. It is the metabolic peptide that never engages the satiety circuit, and it is also the one that did not move body weight in humans.

The comparison is not perfectly clean — different eras, different trial designs, mostly oral rather than injected administration, and a compound whose acute mechanism may simply be too brief to matter at an intravenous half-life measured in minutes ^[14]. It does not prove that adipose-tissue mechanisms cannot work. But the contrast is instructive all the same: semaglutide and retatrutide produce large, reproducible weight changes and both reach the brain, while the compound aimed only at the fat cell produced a tolerability profile and not much else.

Its most defensible current description is as an exceptionally well-tolerated peptide with an interesting preclinical mechanism, a discontinued obesity programme, and a preclinical joint-repair literature that has never been tested in humans ^[13]. Anyone reading it as a third weight-loss option alongside the two incretin agonists is reading a marketing claim rather than the evidence.