AOD-9604: the growth hormone fragment that missed its big trial

The fat-burning slice of a hormone

Growth hormone is a double-edged tool for fat. It releases stored fat, but long-term treatment also raises the risk of glucose intolerance, insulin resistance and fluid retention ^[5]. So in the 1990s Frank Ng, a biochemist at Monash University in Melbourne, asked a tidy question: could you keep only the fat-burning part ^[6]?

His group traced lipid effects to the hormone's last 15 amino acids, residues 177–191 ^[1,5]. AOD-9604, short for "anti-obesity drug" ^[2], is that stretch with one tyrosine added at the start, which the developers described as a stabilising change ^[5]. Unusually for a research peptide, it then went through real human trials, and they show how a clean idea in mice can come apart in people.

From Monash to the stock exchange

Between 2001 and 2006, Metabolic Pharmaceuticals ran six randomised, placebo-controlled trials in Australia and the UK, first with intravenous doses and then with oral capsules and tablets ^[5,11]. By its own account, the company invested more than $50 million in the molecule ^[10].

One naming note: natural growth hormone has phenylalanine at position 176 ^[7], and AOD-9604 has tyrosine there, so it is the natural 176–191 stretch with one amino acid swapped. Our HGH Fragment 176-191 article untangles the two names.

How it works: a key that skips the main lock

Full-length growth hormone works by gripping two copies of its receptor at once, through two separate binding sites. AOD-9604 contains part of one site and none of the other ^[5]. In receptor-transfected cells it did not compete with growth hormone for binding and did not drive cell growth, unlike the full hormone ^[3].

Think of it as a key cut from a broken piece of the original. It does not open the main lock, yet in obese mice it still increased fat burning and fat release, without growth hormone's rise in blood sugar ^[3].

Which door it does open is unclear. In obese mice it raised expression of the β3-adrenergic receptor, a fat-cell receptor that triggers fat release. In mice lacking that receptor, long-term treatment lost its weight effect, though a short-term test still showed higher fat burning ^[4]. The FDA's 2024 review concluded that its molecular target remains unknown ^[8]. One investigator has suggested the mouse result may not translate because the β3 receptor differs between mice and humans ^[6].

What the research shows

Rodents

• Obese Zucker rats: given orally for 19 days, AOD-9604 cut body-weight gain by more than half (15.8 g versus 35.6 g in controls), with no loss of insulin sensitivity on clamp testing ^[1].
• Obese mice: 14 days of treatment reduced weight gain and raised fat oxidation and blood glycerol, a marker of fat breakdown ^[3,4].

Six human trials, 2001–2006

Sources: ^[5,8].

In the 12-week trial, a conference abstract reported weight loss of 0.22 kg per week in the 1 mg arm versus 0.07 kg on placebo; FDA reviewers called the differences small and of unclear meaning ^[8].

OPTIONS was the decisive test. It enrolled 536 obese adults (502 randomised) at 16 Australian sites, added a supervised diet and exercise programme, and compared three daily tablet doses with placebo ^[5,8]. Weight loss was not significantly different from placebo at 12 weeks, the primary endpoint, or at 24 weeks. On 21 February 2007 the company ended development for obesity ^[8].

Professor Gary Wittert, who led five of the six human studies, later said the six-month trial "showed quite clearly there was no effect on weight" ^[6].

Safety in those trials

Across the trials, AOD-9604 did not raise IGF-1, the growth factor behind many of growth hormone's effects, did not worsen glucose tolerance, and triggered no detectable antibodies ^[5]. In the 12-week trial, five serious adverse events occurred in AOD-9604 arms, including four cancers (three skin cancers and one breast cancer); investigators judged none related to treatment ^[5]. In short-term studies, some obese volunteers given intravenous AOD-9604 reported headache, chest tightness, palpitations or euphoria that were considered drug-related ^[6].

Cartilage: a second act in rabbits

After the obesity programme ended, the company turned to cartilage and joint repair, at the preclinical stage only ^[11]. In 32 rabbits with chemically induced knee osteoarthritis, weekly joint injections of AOD-9604 improved cartilage scores against saline, and combining it with hyaluronic acid worked best ^[13]. No human osteoarthritis trial has been published.

What we don't know yet

• The full efficacy data. Neither phase 2b trial was published in full in a peer-reviewed journal; what exists are company filings and a conference abstract ^[8].
• Independence. The human safety summary was written by authors linked to the developer ^[5].
• Mechanism. No confirmed molecular target ^[8].
• Animal toxicology. Reading the developer's published toxicology studies, FDA reviewers flagged dose-related changes in a bone-turnover marker in rats, minimal liver-cell changes in monkeys, and equivocal genotoxicity signals ^[8].
• Injected, long-term use. Human trials used single intravenous infusions in crossover designs, or daily oral doses for up to six months ^[5]. The effects of repeated injections over longer periods are unknown ^[6].
• Product identity. FDA found conflicting CAS numbers in circulation and mismatches between suppliers' stated sequences and structures ^[8]. See our guide to reading a certificate of analysis.
• Joint repair. One rabbit study is not a basis for conclusions about people; our guide to levels of evidence explains why.

Regulatory status

• Approval: never approved as a medicine by the TGA, FDA or any other regulator ^[8,11].
• US food law (GRAS): in June 2012 the developer announced a conditional, self-affirmed GRAS determination by an expert panel it had hired, covering up to 1 mg per person per day in foods, on condition that its safety studies were published ^[10]. GRAS is a food-ingredient status, not a drug approval.
• US compounding: after the original nominations were withdrawn, FDA reviewed AOD-9604 on its own initiative, and on 4 December 2024 its advisory committee voted 12–0 against adding it to the list of substances pharmacies may compound ^[8,9]. FDA's April 2026 safety-risk page lists AOD-9604 under substances "nominated but withdrawn", citing immunogenicity risk, impurity concerns and limited safety information ^[12].
• Australia: the medicines-scheduling committee recommended listing AOD-9604 in the Poisons Standard among performance- and image-enhancing drugs ^[8].
• Sport: in April 2013, during a doping controversy in Australian football ^[6], WADA stated that AOD-9604 falls under S0, non-approved substances, prohibited at all times ^[14]. The current list names it explicitly under S2.2.3, growth hormone fragments ^[15].

In the lab: handling and storage

AOD-9604 is supplied as a lyophilised powder, either as the free peptide or as its acetate salt ^[8]. Long-term storage is usually at −20 °C, with 2–8 °C acceptable for short periods, protected from light and moisture. Its ring depends on a disulfide bond, so reducing agents will open it and change the molecule. Once reconstituted, solutions are usually kept cold and split into single-use aliquots to avoid repeated freeze–thaw cycles. A certificate of analysis for the AOD-9604 research peptide should show the sequence YLRIVQCRSVEGSCGF and a mass near 1815 Da.

FAQ