AOD-9604: The Growth Hormone Fragment That Failed Its Own Trial

Full-length human growth hormone is 191 amino acids long. AOD-9604 is fifteen of them, plus one extra tacked onto the front. That’s the whole idea behind it: cut growth hormone down to the piece that seems to burn fat, and leave behind the piece that makes you grow, spikes blood sugar, and messes with insulin.

On paper, it’s a clean piece of chemistry. What happened when that chemistry met a human body is a longer story, and not the one the design predicted.

What AOD-9604 actually is

Growth hormone does two jobs at once. One end of the molecule tells your body to grow, through a hormone called IGF-1. The other end, way down at the tail, appears to be where the fat-burning signal lives. AOD-9604 takes that tail: residues 177 to 191 of the hormone, with one tyrosine added to the front end. Nothing else.

The logic is straightforward. Isolate the fat-burning piece, throw away the growth-and-blood-sugar piece, and you get a fragment that burns fat without the downsides of taking full growth hormone: no unwanted growth in an adult, no insulin resistance, none of the reasons a doctor would think twice before prescribing it. That’s the pitch. It’s a rational piece of drug design, and structurally, it does exactly what it says on the label.

Does AOD-9604 actually burn fat, or does it just work in rodents?

In animal studies, AOD-9604 does what the design predicted. Rodents given the fragment showed increased fat burning, and researchers reported none of the growth-promoting or insulin-resistant effects that come with full growth hormone. The idea is that the fragment tells fat cells to release and burn their stored energy, at least in animals. That’s the mechanism as proposed, not as proven. Nobody knows exactly how it works at a biological level, and almost all the evidence comes from animals and lab dishes, not from humans.

That gap, rodent to human, is where a lot of promising compounds go to die. AOD-9604 is one of them.

What happened to the human trials

Metabolic Pharmaceuticals, the Australian company that developed AOD-9604, ran it through Phase II trials for obesity in the early 2000s. Early results looked promising enough to greenlight the next, larger trial, the one that would need to prove the drug actually worked.

It didn’t. The trial came back negative, the drug didn’t work well enough to count as working. Metabolic Pharmaceuticals discontinued the obesity program not long after.

A company doesn’t usually walk away from a compound that’s working. The animal data held up. The mechanism made sense. The human trial built to confirm both of those things came back negative, and the program never made it to Phase III.

Is AOD-9604 legal to use, compound, or buy?

AOD-9604 isn’t an FDA-approved drug, and no controlled human trial has confirmed it works for what people are using it for now. It exists in a gray zone, sold and used without the safety testing the FDA normally requires before a drug reaches patients, and most of what is known about it comes from a program that failed its own trial two decades ago.

The regulatory record adds a wrinkle most coverage skips. AOD-9604 already went in front of the FDA’s Pharmacy Compounding Advisory Committee, the panel that decides which peptides pharmacies are allowed to legally compound. At the December 4, 2024 meeting, the committee voted against including it on the 503A bulk drug substances list. The FDA’s reasoning was the same it applies to any unproven compound: not enough safety data in humans, and no reliable evidence it actually works. For AOD-9604, that last part isn’t abstract. It has a name: the Phase IIb trial that failed.

So AOD-9604 isn’t sitting in a regulatory waiting room, still to be decided. It already had its hearing, and it lost. Compounding pharmacies selling it today are operating in the space left after that vote, not ahead of it.

One more fact worth knowing if you’re using it anyway: AOD-9604 is outright banned in sport, not just during competition, but year-round, for any athlete subject to drug testing.

Does skipping the IGF-1 pathway actually matter?

The pitch for AOD-9604 was never just fat loss. It was fat loss without the tax that comes with full growth hormone: no growth-plate effects, no insulin resistance, none of what makes hGH a drug you’d only take under supervision. That part of the design story is solid. The fragment really is built to leave the IGF-1 pathway alone.

But a clean mechanism only matters if it produces a clean result, and the one trial built to test that came back empty. Whether a different dose, a different injection schedule, or some other tweak could have rescued the effect is a real question. Nobody has run that trial.

The compounding market, and what the safety data does and doesn’t cover

Compounding pharmacies and online sellers offer AOD-9604 today, usually injected under the skin. Sometimes it’s packaged for a different use entirely: one form sold is meant to go straight into a joint, chasing joint relief instead of fat loss. That use has no supporting evidence behind it at all, not even the animal-and-mechanism case built for fat loss, and no data showing what the fragment does once it’s inside a joint instead of under the skin. It’s a guess built on a guess.

What compounded AOD-9604 actually contains, in terms of purity and consistent dosing compared to the material used in the original trials, isn’t something anyone outside the manufacturer can verify. That’s a genuine unknown, not a minor caveat.

Who has a real reason to consider this right now?

Take the record at face value: a peptide engineered well enough to pass every preclinical test it was given, that then failed the one trial built to confirm it works in a person, and that has since been reviewed and rejected by the regulator that decides what’s allowed into legitimate pharmacy compounding. Fifteen amino acids, cut clean from a hormone that does more. On paper, the smartest possible shortcut. In the one human trial it got, not enough of one.

If you’re thinking about AOD-9604 anyway, that’s the bet you’re making: that the trial missed something, that the dose was wrong, that your body will do what the rodents’ did. Maybe. But the evidence that would tell you so doesn’t exist yet.