HGH Fragment 176-191: One Atom Away From AOD-9604, and Why Every Human Trial Belongs to the Other Peptide

HGH Fragment 176-191: One Atom Away From AOD-9604, and Why Every Human Trial Belongs to the Other Peptide

Somewhere in the last twenty years, two different peptides ended up under one product name. One is the last sixteen amino acids of human growth hormone, exactly as they sit in the hormone. The other is those same sixteen amino acids with a single tyrosine swapped in at the front, which an Australian company called AOD-9604, put through six clinical trials in 925 people, and abandoned in 2007 when a 536-patient study found it did not beat placebo. Almost every claim you will read about 'HGH fragment 176-191' was generated with the second peptide. The first has never been given to a human being in a published study. This reference sets out what each one is, what was actually shown, how to tell which one is in your vial, and where both stand with WADA and the FDA in 2026.

Michael Phelps

Founder & Peptide Research Specialist, PrymaLab

Research reference · Last updated September 23, 2026 · ~24 min read

Two peptides share one name

Human growth hormone is a single chain of 191 amino acids. Count from the N-terminus and the last sixteen residues, positions 176 to 191, read Phe-Leu-Arg-Ile-Val-Gln-Cys-Arg-Ser-Val-Glu-Gly-Ser-Cys-Gly-Phe, with the two cysteines at 182 and 189 joined by a disulfide bridge in the folded hormone. Cut those sixteen out and you have HGH fragment 176-191: molecular formula C₇₈H₁₂₃N₂₃O₂₂S₂, average mass 1,799.1 Da, CAS number 66004-57-7.^[1]

AOD-9604 is that fragment with one change. Its developer replaced the phenylalanine at position 176 with a tyrosine, which adds a single hydroxyl group to the first side chain and raises the formula to C₇₈H₁₂₃N₂₃O₂₃S₂, about 1,815 Da, with the disulfide formed.^[2] The sponsor's own safety paper defines it as 'a peptide fragment of the C-terminus of hGH (Tyr-hGH177-191)'.^[3] Read that notation carefully: it is residues 177 to 191 with a tyrosine bolted on the front, which is the same thing as 176-191 with position 176 swapped. One oxygen atom separates the two peptides.

Why does this matter to someone typing 'hgh fragment 176-191' into a search box? Because the research-peptide market uses that name for both molecules, and sometimes for a linear version with the disulfide left open, and the evidence base for the two is not remotely the same. Wikipedia's entry for the native fragment says it plainly: it 'has erroneously been presented as a lipolytic peptide fragment based on extrapolations of clinical data pertaining to AOD9604'.^[1] Every human trial, the GRAS opinion, the mouse mechanism work and the pharmacokinetics belong to the tyrosine version. The native fragment has one published study, and it is about breast cancer cells in a dish.

By the definition in the what counts as a peptide reference, both are unambiguously peptides: sixteen amino acids, fifteen amide bonds, made by solid-phase synthesis, destroyed within minutes by serum proteases. That last property turns out to be the most important pharmacological fact about either of them.

Where the lipolytic fragment idea came from

The idea that a piece of growth hormone could carry its fat-burning effect without its other effects came out of Monash University in Melbourne in the 1990s, from the laboratory of Frank Ng. Growth hormone raises lipolysis, but it also raises blood glucose, lowers insulin sensitivity and drives IGF-1, which is why nobody uses it as an obesity drug. The hypothesis was that the hormone behaves as a pro-hormone, with different regions responsible for different effects, and that the C-terminal region was the lipolytic one. Metabolic Pharmaceuticals was formed to develop a synthetic version, and it chose the tyrosine-substituted fragment it called AOD-9604.^[2][3]

The clearest preclinical demonstration is Heffernan and colleagues, 2001. Obese ob/ob mice and lean C57BL/6J mice were treated for 14 days with human growth hormone, AOD-9604 or saline through mini-osmotic pumps. 'Both hGH and AOD significantly reduced body weight gain in obese mice', with increased in-vivo fat oxidation and raised plasma glycerol, an index of lipolysis. 'Unlike hGH, however, AOD9604 did not induce hyperglycaemia or reduce insulin secretion.' In cells transfected with the growth hormone receptor, 'AOD9604 does not compete for the hGH receptor and nor does it induce cell proliferation, unlike hGH'.^[4] The authors concluded that fragments of hGH 'can act in a manner novel to traditional hGH-stimulated pathways'.

That paper is the origin of every 'burns fat without the side effects of growth hormone' sentence on the internet. It is a fourteen-day mouse study of the tyrosine peptide delivered by a surgically implanted pump. The mechanism proposed for the effect, upregulation of beta-3 adrenergic receptors in fat tissue, is also a mouse finding.^[2] No receptor for AOD-9604 has been identified in humans, and the native fragment was not in the experiment.

What the beta-3 receptor knockout experiment showed

AOD-9604's effect in mice is often put down to upregulation of the beta-3 adrenergic receptor. The paper behind that statement found something narrower. Heffernan and colleagues treated obese mice with human growth hormone or AOD-9604 for 14 days and recorded lower body weight and body fat, together with beta-3 receptor RNA restored from the depressed levels of obese mice to those of lean mice (Heffernan et al. 2001). They then repeated the long-term treatment in beta-3 receptor knockout mice, where neither compound produced the change in body weight or the rise in lipolysis seen in wild-type controls (Heffernan et al. 2001). In an acute experiment, though, AOD-9604 still raised energy expenditure and fat oxidation in the knockouts (Heffernan et al. 2001). The authors concluded that the lipolytic actions of both compounds were not mediated directly through the beta-3 receptor, even though both raised its expression and that rise might contribute (Heffernan et al. 2001).

An earlier rat study from Frank Ng's laboratory is also relevant. In an oral study lasting 19 days in obese Zucker rats, Ng and colleagues reported body weight gain less than half that of controls, higher lipolytic activity in adipose tissue, and no adverse effect on insulin sensitivity by euglycaemic clamp, in contrast to chronic treatment with intact growth hormone (Ng et al. 2000). It is the animal precedent for the oral route that the later human trials used. Both studies used AOD-9604; neither tested the native phenylalanine fragment.

The human record: six trials, 925 people, one negative Phase 2b

Metabolic Pharmaceuticals took AOD-9604 into people as an oral tablet, which is itself a fact worth pausing on: the peptide sold for injection was tested clinically by mouth. The company began compiling safety data in 1998 and reported spending more than 50 million dollars on the programme.^[5]

The result that made the news came in December 2004. In a 12-week trial of about 300 men and women aged 30 to 65, the reported tablet arm produced 'an average weight loss of 6.2 pounds over 12 weeks, more than triple the weight loss achieved by study participants who took a placebo pill', with small improvements in cholesterol and fewer people with impaired glucose tolerance.^[6] Other accounts of the same trial give 2.6 kg against 0.8 kg on placebo, which is the source of the '1.8 kg more than placebo' figure that circulates.^[2] Call it 1.8 to 2 kg over placebo in three months.

The confirmatory study was called OPTIONS. It enrolled 536 adults with obesity, roughly half men and half women, at 16 sites across Australia, randomised to placebo or one of three oral tablet doses for 24 weeks inside a 32-week study. Enrolment finished in late April 2006 and results were expected in March 2007.^[7] The result was that AOD-9604 did not produce a statistically significant weight loss against placebo, and development was terminated in 2007.^[2] The sponsor's own 2014 paper records it in one sentence: 'Effects of AOD9604 on weight loss were seen in initial trials but were not seen in the last study in which an intensive diet and exercise regime was incorporated.'^[8]

What survived the failure was the safety database. Stier, Vos and Kenley pooled six randomised, double-blind, placebo-controlled trials and reported that AOD-9604 'displayed a very good safety and tolerability profile indistinguishable from placebo', with no detectable anti-AOD9604 antibodies, no treatment-related serious adverse events or withdrawals, 'no negative effect on carbohydrate metabolism' and 'no effect on serum IGF-1 levels'.^[3] The company's 2012 announcement put the database at 'six human clinical trials involving 925 humans'.^[5] Both papers were written by the sponsor; Kenley was its chief executive.^[8] That does not make them wrong, and a 925-person safety record with no IGF-1 signal is a real finding, but the reader should know whose pen it came from.

What AOD-9604 does and does not do in the body

The 2014 metabolism paper by Moré and Kenley is the best single source on what happens to the peptide after it is given, and it explains why the clinical programme was built on tablets rather than injections. After intravenous injection the plasma half-life was 'approximately 3 min'; in serum the half-life was about 4 minutes 'with intact AOD9604 undetectable 56 min' later, and the principal fragments observed in vivo were the peptide minus its last two and last three amino acids.^[8] A sixteen-residue peptide with no protection at either end is food for carboxypeptidases, and the data show exactly that.

Three things follow. First, whatever either peptide does after injection, it does within the hour, after which the intact molecule is gone. Second, the oral trials imply the sponsor believed a fragment survived digestion in some active form, since the human trials gave it by mouth after an oral study in obese rats; the 2014 paper describes the -2 and -3 residue fragments as the main circulating species, which raises the possibility that a shorter piece carries whatever activity exists.^[8] Third, the two claims that actually matter for safety held up across the whole programme: 'AOD9604 did not stimulate the production of IGF-1 in any of the systems studied', and it 'elicited the fat reducing effect without inducing insulin resistance'.^[8] Chronic toxicology ran 26 weeks in rats and 39 weeks in cynomolgus monkeys.^[8]

What the pharmacology does not include is a human receptor, a human dose-response for lipolysis, or any measurement of fat oxidation in people. The 12-week weight result stands alone as the human efficacy signal, and the 24-week trial did not reproduce it.

What has actually been studied with the native fragment

Strip away everything that belongs to AOD-9604 and the literature on the Phe-176 fragment sold as 'HGH fragment 176-191' comes down to a single paper. Habibullah and colleagues, writing in Drug Design, Development and Therapy in 2022, loaded chitosan nanoparticles with doxorubicin and with hGH fragment 176-191, ran molecular docking simulations against breast-cancer protein targets, and tested the particles on MCF-7 breast cancer cells. The dual-loaded particles showed 'greater anti-proliferative activity against a breast cancer cell line (MCF-7) than doxorubicin-loaded Chitosan', and the authors propose the fragment as a tumour-targeting ligand that might let chemotherapy hit cancer cells harder.^[9]

That is the study. It has nothing to do with fat, it was done in a dish, and the peptide's role in it is as a targeting handle on a nanoparticle. There is no animal lipolysis study of the native sequence, no pharmacokinetic study, no receptor study and no human study. Wikipedia's search of PubMed in December 2023 found the same absence.^[1]

Is the native fragment likely to behave like AOD-9604? Probably, in most respects: a tyrosine-for-phenylalanine swap at a terminus is a small change, and the sponsor chose it for reasons of synthesis and stability rather than because the native residue was inactive. But 'probably similar' is an inference, not a finding, and a researcher who wants the molecule behind the published trials should work with AOD-9604 and read the AOD-9604 reference, which covers that molecule in its own right.

Frag 176-191 vs AOD-9604: which one is in the vial

Because the two names are used interchangeably, the honest question is not 'which is better' but 'which one do I have'. Three physical facts settle it.

The purity testing guide explains why an HPLC trace alone cannot answer an identity question: two peptides 16 Da apart elute close together and both give a clean single peak. Mass spectrometry answers it in one number. When I document our own HGH Fragment 176-191 autoinjector and AOD-9604 vials, the sequence and the observed mass are the two lines I would want any reader to check first, and the same standard applies to any other supplier.

One more practical difference. The disulfide bridge between the cysteines at 182 and 189 forms a loop that the folded hormone has and that AOD-9604 was made with. A linear synthesis that skips the oxidation step yields a peptide with two free thiols, which is more prone to aggregation and to forming dimers on storage. If a certificate says nothing about the disulfide, that is a question worth asking before you buy.

GRAS, WADA and the 503A list: three different answers

Fragment 176-191 gets described online as everything from 'FDA approved' to 'banned', and the confusion comes from three separate regulatory facts that concern different things.

The GRAS opinion. On 25 June 2012, Calzada Limited, the successor to Metabolic Pharmaceuticals, announced that an expert scientific panel convened through GRAS Associates LLC had issued 'a self-affirmed conditional GRAS status' for AOD-9604 'to be added to foods, drinks & dietary supplements in the US', resting on the six-trial, 925-person safety database.^[5] The 2014 paper describes it as 'conditional on publication of pre-existing safety data', which the 2013 and 2014 papers then supplied.^[8] Self-affirmed GRAS means a company's own panel concluded the ingredient is generally recognised as safe for a stated food use; no notification was filed with FDA, and FDA has made no finding. It says nothing about efficacy, nothing about injection, and nothing about the native fragment.

The FDA compounding list. On 4 October 2023 FDA revised its interim 503A bulk drug substances list and placed AOD-9604 in Category 2, the list of substances that raise significant safety concerns and may not be used in compounding, alongside kisspeptin-10, BPC-157 and ipamorelin.^[10] The native fragment is not on any FDA list. Neither peptide is an approved drug anywhere, and no application has been made since development ended in 2007. The legal status guide covers what that means for research-grade material: legal to sell and hold for laboratory research, not for human use.

The WADA list. The 2026 Prohibited List names both peptides in section S2, under growth hormone, its analogues and fragments: 'growth hormone fragments, e.g. AOD-9604 and hGH 176-191', prohibited at all times as non-specified substances.^[11] WADA does not care which of the two you have. An athlete holding either is holding a listed substance.

The cartilage detour

After the obesity programme closed, the peptide found a second research life in joints. Kwon and Park, 2015, induced osteoarthritis in 32 New Zealand white rabbits with intra-articular collagenase and then gave intra-articular injections over 4 to 7 weeks of saline, hyaluronic acid, AOD-9604, or both. Days to normal walking were 25 (saline), 15 (hyaluronic acid), 16 (AOD-9604) and 11 (both), and gross and histological cartilage scores were best in the combination group.^[12] It is a small animal study with a local injection into the joint, and it has not been followed by a human trial that I could find. It does show that the research interest in the peptide has moved on from fat, which the fragment 176-191 marketing has not caught up with.

Research-grade fragment 176-191: what the certificate must show

Both peptides are legal to sell for laboratory research in the United States and neither is a controlled substance. Both are on the WADA list, and AOD-9604 is on the FDA's Category 2 list for compounders. Within that frame, the buyer's job is identity first and purity second.

Ask for the sequence line and the observed mass. 1,799 means the native Phe-176 fragment; 1,815 means the tyrosine-substituted AOD-9604; anything else means a question. Ask whether the disulfide was formed. Then ask for HPLC purity by lot, not by catalogue. PrymaLab's preloaded 10 mg fragment autoinjector and 5 mg AOD-9604 vials are documented on that basis, and the peptide calculator converts a vial and its diluent into a concentration for bench work. Lyophilised peptide keeps for months refrigerated or frozen; dissolved peptide should be refrigerated and used within weeks, because a peptide that lasts minutes in serum is not going to last indefinitely in solution either.

The version of the story the evidence supports, without the marketing: a C-terminal piece of growth hormone that does something to fat metabolism in obese mice without touching the GH receptor, IGF-1 or blood glucose; a tyrosine-substituted version of it that lost about two kilograms over placebo in twelve weeks and nothing in twenty-four; a clean 925-person safety record written by the sponsor; a food-ingredient opinion the company gave itself; a WADA entry that names both peptides; and a native fragment, the one most often sold, that has never been tested in a person. If you want a broader view of the class, the fat loss peptides overview and the peptides and weight loss guide put it beside the compounds that did clear a Phase 3.

What this article does not settle

Whether the native fragment shares AOD-9604's activity. Nobody has run the mouse experiment with the Phe-176 sequence, so the assumption that one atom makes no difference is untested.

Why the 24-week trial failed. The sponsor attributes it to the intensive diet and exercise regime in the protocol, which is one reading; a 12-week effect of 1.8 kg that disappears under lifestyle intervention is also consistent with a small effect that the larger study had the power to see for what it was. The full OPTIONS data have never been published.

The mechanism in people. Beta-3 adrenergic upregulation is a mouse finding; no human receptor or target has been identified for either peptide.

Anything about injection. Every human trial used oral tablets or an intravenous pilot dose. The subcutaneous injection that the research market assumes has no clinical record, and the 3-minute plasma half-life means whatever it does, it does quickly.

Long-term use. The longest human exposure was 24 weeks by mouth in a supervised trial. The toxicology ran 39 weeks in monkeys. Nothing longer exists.

What is well established: the two sequences, formulas and masses; that AOD-9604 does not bind the GH receptor and does not raise IGF-1; the mouse lipolysis result; the 12-week and 24-week trial outcomes; the pooled safety profile; the plasma half-life; the 2012 GRAS self-affirmation; the 2023 Category 2 placement; and the WADA listing of both names.

hGH 176-191 and AOD-9604: what the numbers and letters refer to

In 'hGH 176-191', hGH is human growth hormone, somatotropin, a 191-residue protein secreted by the pituitary, and 176-191 are the positions of the sixteen C-terminal amino acids counted from the N-terminus, so the fragment is the tail end of the hormone.^[1] 'Tyr-hGH 177-191' is the sponsor's notation for AOD-9604: residues 177 to 191 with a tyrosine added at the front, which occupies the position where the hormone has a phenylalanine.^[3] AOD is the developer's compound prefix, generally expanded as anti-obesity drug, and 9604 is the series number; neither carries chemical information. Other names for the same two peptides include Frag 176-191, HGH frag, lipolytic fragment, growth hormone fragment 176-191 and, for the tyrosine version, AOD9604 and AOD 9604. The WADA list uses 'AOD-9604 and hGH 176-191' and treats them as the same class of substance.^[11]

Frequently asked questions

References