HGH Fragment 176-191: Almost Every Trial You Have Been Shown Is a Different Molecule

Three molecules and one name

Human growth hormone is 191 amino acids. The last sixteen, Phe-Leu-Arg-Ile-Val-Gln-Cys-Arg-Ser-Val-Glu-Gly-Ser-Cys-Gly-Phe, closed into a loop by a disulfide bond between the cysteines at fragment positions 7 and 14, are sold as Fragment 176-191: CAS 66004-57-7, molecular weight 1799.1.

AOD-9604 is a different molecule: residues 177 to 191, fifteen amino acids, with a tyrosine added at the front. That swaps phenylalanine for tyrosine at position 1 and adds one oxygen atom, leaving the carbon, hydrogen, nitrogen and sulfur counts identical, 1815.09 against 1799.1, sixteen daltons apart, two CAS numbers. The name sticks because AOD-9604 is also sixteen residues long, so it gets called 176-191 by count despite lacking the phenylalanine hGH has at 176, and a 2026 Frontiers in Endocrinology review explains that convention, then prints the two as one compound in its own abstract (PMID 42395176).

I found no study putting the two side by side, in any system, in any species. Two loaves from one recipe with an ingredient swapped, never baked on the same day.

This site has run this story twice, with TB-500 sold on thymosin beta-4's research and with KPV, which did nothing its parent alpha-MSH did once the two met in the same cells. Here the borrowed evidence does not even belong to the parent. It belongs to a modified copy of the fragment.

Everything published on the fragment itself

Go to PubMed, filter to clinical trials, search for hGH 176-191. Zero. Filter to humans instead and you get three records, two reviews and one in vitro cell paper. Six searches, four databases, 2026-08-05. I am not asserting no such study exists somewhere. These searches found none.

Ng and Bornstein 1978 is the paper vendors cite as the foundation, and it reports close to the opposite of what they say. Six synthetic C-terminal hGH peptides in normal rats, a single dose of 5 nmol/kg body weight, about 9 mcg/kg in a rat, by a route the abstract never states. Four of the six, 176-191 among them, produced a short-lived rise in blood glucose and a more sustained rise in plasma insulin, and the peptides containing 178-191 significantly reduced insulin sensitivity (PMID 645904). The title is Hyperglycemic action of synthetic C-terminal fragments of human growth hormone. peptides.org, the page ranking highest for this compound, tells you the fragment lowers blood sugar without altering insulin sensitivity, and cites that paper for it. Its applications list prints reduced insulin sensitivity as a benefit.

Then almost nothing. Ma and colleagues 1982 followed up in rats, with no abstract published and a full text I did not obtain (PMID 6810951). Forty years pass. Habibullah and colleagues ran the peptide against MCF-7 breast cancer cells and through docking simulations (PMID 35783198). No animal, no person.

The AOD-9604 record, which is the one you have been shown

The fat loss claim breaks on the earliest work, which used neither of them but hGH 177-191, the bare fifteen residue chain with no tyrosine on the front. Wu and Ng 1993 found it antilipogenic, meaning it blocks fat being made, and found no significant lipolytic effect in treated rats (PMID 8358331). The same laboratory group stated the reverse seven years later, from adipose tissue in a dish (PMID 11152298), and I found no paper reconciling the two.

The rest is the modified compound. Ng and colleagues gave obese Zucker rats AOD9604 by mouth at 500 mcg/kg body weight daily for 19 days and more than halved their weight gain, 15.8 plus or minus 0.6 g against 35.6 plus or minus 0.8 g (PMID 11146367), a rat oral dose and not a human or subcutaneous one. Every joint claim traces to 0.25 mg of AOD9604 injected weekly into the knees of 32 rabbits (PMID 26275694). Of the animal work routinely cited for Fragment 176-191, one paper actually used it, and that is the hyperglycaemia paper. FDA, having read the whole nonclinical package, concluded the molecular targets and mechanisms of action remain unknown.

The human record is six trials run in Australia between roughly 2001 and 2007 by Metabolic Pharmaceuticals, surviving as one summary paper by that sponsor's own employees and consultants, in a journal PubMed does not index. FDA summarised it in December 2024, and everything below is FDA's summary of it. FDA's reading of the first four is that none showed statistically significant weight loss. The fifth put 300 patients with obesity on oral AOD-9604 at 1, 5, 10, 20 or 30 mg per day or placebo for 12 weeks, and exists only as a conference abstract whose methods FDA could not find published. Weight loss ran at minus 0.22 kg per week on 1 mg against minus 0.07 on placebo. The largest effect was at the smallest dose.

OPTIONS was the one built to settle it. 502 randomised out of 536 enrolled, adults with BMI 30 to 45, oral AOD-9604 at 0.25, 0.5 or 1 mg once daily for 24 weeks against placebo, powered at 80 percent to detect a 1.8 kg difference at 12 weeks. It did not separate, and on 21 February 2007 Metabolic Pharmaceuticals terminated development for obesity. Watch the doses across the last two trials, 1 to 30 mg a day, then 0.25 to 1 mg. The one trial powered to find anything ran at the bottom of every oral dose the compound had ever been given.

Where 250 to 500 mcg does not come from

Reported as convention, because that is all it is. peptides.org gives 200 to 500 mcg per day subcutaneously, fasted, over about four weeks, and the 2026 Frontiers review tabulates self-reported forum protocols at 250 to 500 mcg per day subcutaneously in 8 to 12 week cycles. A forum number reprinted in a journal table is still a forum number.

I could not establish where the 200 to 500 mcg convention came from, and I will not invent a lineage for it. The human doses on record in this family are AOD-9604 at 25 to 400 mcg/kg intravenously and 0.25 mg to 54 mg orally, an oral range spanning a factor of 216 from bottom to top. Nothing in that spread produces 250 to 500 mcg a day under the skin of a person.

The route is in worse shape than the number. FDA reported in two separate sections that it did not identify human exposure to AOD-9604 by the subcutaneous or transdermal route, and did not identify any clinical study of its pharmacokinetics or pharmacodynamics by any route, so the 15 to 20 minute half life everyone quotes has no published human source.

The safety record, and what regulators have said

I could find no human safety data for the unmodified fragment: no adverse event series, no case report, no tolerability study. FDA has never evaluated it, because nobody ever nominated it.

What exists belongs to AOD-9604, and it is not frictionless. In that 300-patient oral study, five subjects reported serious adverse events that were neoplasms, four malignancies and one benign lipoma, spread across the 5, 10 and 20 mg groups. The investigators called none of them related, and FDA declined to accept that. Five neoplasms among 300 people over 12 weeks is not by itself a signal.

WADA is the only regulator I found that names the unmodified fragment in its own right, listing AOD-9604 and hGH 176-191 side by side under S2.2.3 of the 2026 Prohibited List, prohibited at all times and non-Specified, the more serious tier. FDA's advisory committee voted AOD-9604 down on 4 December 2024, 0 yes and 12 no once the tally was corrected on the record. Search FDA's compounding pages for the string 176-191 and you get zero hits, against three for AOD-9604. Never having been in front of FDA is not the same as having been cleared by it.

My opinion

Missing human data is ordinary in this category. What decided this entry is that the foundation paper has said the opposite of the sales pitch since 1978, in its title, in a journal anyone can pull up. Raised blood glucose and significantly reduced insulin sensitivity in rats, printed as lowered blood sugar without altering insulin sensitivity on the page most people read first. Nobody had to falsify anything to get there. Somebody just did not open it. And once the two are treated as one, a failed development programme becomes the evidence base for a peptide I could not find one human study of.

If you run Fragment 176-191, dig out the certificate that came with the vial and read the mass off it. The two published weights are 1799.1 for the fragment and 1815.09 for AOD-9604, sixteen daltons apart. Which of those two numbers is printed on yours?

---

This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.