GLP-1s for Weight Loss: What You Are Actually Buying

Who this is for

You want to lose weight. You have seen three names, probably in this order: semaglutide, tirzepatide, retatrutide. You are deciding between a prescription and a vial from a website, and you have likely already priced both, which is the real reason you are still reading.

Nothing here tells you which to do. What follows is what the trials found, what a lab found in the online product, and where the risk actually sits, which is not where the argument usually happens.

What people actually run

Reporting what circulates. This is observation, and none of these numbers came from a trial unless I say so.

The names come up in that order because that is the order of the headline weight loss figures. What differs underneath is not the biology. It is the supply route.

Route one is a prescription, which semaglutide and tirzepatide both have, and which has narrowed since the compounding window opened by the shortages closed.

Route two is a clinic or telehealth service selling a compounded version. Among 75 weight loss clinics and medspas in West Virginia and Oklahoma trading in late 2025, 56% offered GLP-1 products combined with B vitamins. (PMID 42467450)

Route three is a vial of powder and a bottle of bacteriostatic water. For retatrutide it is the only route that exists, and an unreviewed preprint analysing Reddit discussion to December 2025 counted 13,589 people reporting current use.

The conventions along that route are worth naming precisely, because they are conventions and nothing more. For tirzepatide, a microdose is any weekly dose below the 2.5 mg label starter, typically 0.25 to 1.5 mg, and those numbers appear on vendor and telehealth pages, which are marketing rather than evidence. For retatrutide, the reconstitution taught across vendor charts is 2 mL of bacteriostatic water into a 10 mg vial, giving 5 mg per mL, at which 1 mg is 20 units on a U-100 insulin syringe. The circulating start is 0.25 to 1.5 mg weekly and the circulating maintenance target is 4 mg.

Notice what 10 mg is. It is a vial size. The circulating numbers are vial arithmetic, not trial arithmetic.

What each one is actually built on

Semaglutide. The human evidence runs past 25,000 people in outcome trials alone. In STEP 1, 1,961 people lost 14.9% of body weight against 2.4% on placebo, and SELECT followed 17,604 people with obesity and heart disease and cut heart attacks, strokes and cardiovascular deaths by 20%. (PMID 37952131) Two things get left out: the effect is about a third smaller with type 2 diabetes, and in the withdrawal trial the group that switched to placebo gained back 6.9%. No trial has ever used research grade or compounded material. The full entry is here.

Tirzepatide. Approved as Mounjaro for diabetes and Zepbound for weight and sleep apnoea, on a numbered series of phase 3 trials. SURMOUNT-1 ran 72 weeks and produced 20.9% weight loss at 15 mg against 3.1% on placebo (PMID 35658024), and SURMOUNT-5 put it head to head against semaglutide and got 20.2% against 13.7%. The lowest dose ever carried through a randomised trial with a sustained weight endpoint was 1 mg, and it produced 0.9 kg over 26 weeks in people with type 2 diabetes. The full entry is here.

Retatrutide. A weekly injection acting on three receptor systems instead of one, and on the numbers the strongest of the three. Phase 2 gave 24.2% weight loss at 12 mg over 48 weeks against 2.1% on placebo (PMID 37366315), and the phase 3 TRIUMPH-1 readout in 2,335 participants reported 28.3% at 80 weeks, or 25.0% on the more conservative estimand. Those phase 3 results were announced by press release and have not been peer reviewed. It is approved in no country, and Lilly says it cannot legally be sold or marketed for human use. The full entry is here.

The honest read

Say it plainly, because everywhere else on this site the sentence runs the other way. The evidence here is not thin, and pretending otherwise would be its own kind of dishonesty. These drugs do what they claim. That moves the risk somewhere else, and three distinctions carry it.

Approved is not a formality, and retatrutide is approved nowhere. For semaglutide and tirzepatide, going grey market is a substitution: a characterised, inspected, assayed version of the same molecule exists, and a doctor can hand it to you. For retatrutide no such version exists at any price. It is the difference between ordering a part that is out of stock and ordering one that never went into production. In the first case a genuine article exists somewhere to hold yours against. In the second, whatever arrives is the only one you will ever see, and no reference sample exists to tell you it is wrong.

The strength of retatrutide's trial numbers makes this feel safer, and that is backwards. Good results describe a molecule. They say nothing about a vial.

The one published test of what is actually sold found two errors pointing opposite ways. Researchers bought semaglutide from three online sellers in 2024 and had it analysed. Purity came back at 14.37, 8.97 and 7.7 percent against 99 percent advertised. And the amount of actual semaglutide in those vials ran 28.6 to 38.7 percent higher than the label claimed. (PMID 39509151)

Most people picture the grey market failure as a watered down drink. You paid for a double and got a single: annoying, but safe. What that lab found was a glass mostly made of something other than alcohol that still held more alcohol than the bottle promised. The one who poured carefully to the line got hit hardest. The researchers' own stated concern was overdose, not underdosing.

The only vial level result published for retatrutide points the same way. A vendor commissioned test on a BulkGLP 10 mg vial reported content of 12.05 mg against the 10 mg label, 121% of stated. I could find no peer reviewed analysis of grey market retatrutide.

Two of these three have a fabricated trial registration. On 31 July 2026, Morgan McSweeney published an investigation into a cluster of eight ClinicalTrials.gov registrations under one sponsor, Hudson Biotech, all marked recruiting, all at a single site. (Dr Noc) Three of the eight say in their own text that they are fictional examples. The two that concern this page do not. NCT07481747 reproduces SURMOUNT-1's official title character for character and carries Eli Lilly's own internal protocol code for that trial with "(b)" appended. NCT07467447 copies the title of Lilly's phase 2 retatrutide study word for word and calls itself recruiting, when the study it copied completed on 22 November 2022. Semaglutide has nothing of the kind attached to it.

And one thing is true of all three. The weight comes back when you stop. That is what the withdrawal arms of the manufacturers' own trials found, so it is not a course you finish.

If you are going to do this anyway

This is the part I actually care about, and almost none of it is about the molecules.

The real risk is what is in the vial. Everything above says the same thing: the label is not evidence. In April 2025 the United States International Trade Commission barred importation of products containing tirzepatide and of "products purporting to contain tirzepatide." That last phrase is a federal agency recording, in an enforcement order, that some of what ships may contain no drug at all.

What you can do is demand a batch specific certificate of analysis and know how to read it. The line almost everyone skips is content, sometimes called assay, which is how many milligrams are actually in the vial. It is a different question from purity, and it is the one that determines your dose. Both published results above failed on content, in the same direction, upward. Here is how to read one, walked through on real certificates including an annotated retatrutide one.

The second risk is arithmetic, and this category is worse for it than most. The circulating retatrutide numbers are built around vial sizes rather than trial doses, and those vary between vendors: 10, 12, 24, 30 and 60 mg all circulate. The same reconstitution instruction poured into a different vial gives a different concentration, so the same units on the syringe deliver a different dose. The one published case report of harm is a man who bought retatrutide online, escalated from 10 mg to 20 mg weekly, then injected again the next day after forgetting the first, roughly 40 mg across two days. He had a week of intractable diarrhoea, dehydration and acute kidney injury, and recovered. The full walkthrough of the math is here.

What would make you stop. The gastrointestinal effects are dose related rather than rare. In the retatrutide phase 3, nausea ran 42.4% at the top dose against 14.8% on placebo, and more than one in nine people stopped that arm for adverse events. Both approved drugs carry a boxed warning for thyroid C-cell tumours seen in rats, human relevance undetermined. The semaglutide safety comparison found raised signals for pancreatitis, bowel obstruction and gastroparesis, the pancreatitis estimate carrying a margin of error wide enough to be genuinely uncertain.

Sterility and endotoxin you cannot assess at home at any price. All three of those semaglutide samples carried bacterial toxin, and nothing about their appearance would have told you. If you compete under an anti-doping code, semaglutide is banned by WADA, and retatrutide falls into the category covering substances approved nowhere, prohibited at all times.

My take

I have spent most of this project arguing that the confidence around these compounds is manufactured. Here it is not. Somebody ran the trials, they were large, and they published the unflattering results too.

Which is why the failure mode is so specific. The evidence is real and it is being used as cover for something it does not cover. A clinic quotes SURMOUNT-1's 20.9% while selling a compounded product no trial ever used. A vendor page quotes TRIUMPH-1 while shipping a vial no regulator has ever seen. The number is true. It is just not a number about the thing in the box.

Retatrutide is the sharpest version, holding the best results and the worst position at once. Being approved nowhere is not a delay in the paperwork. It is the absence of any characterised product to hold a vial against.

The question I would ask myself is not which molecule. It is whether I am buying the drug that was tested, or something that shares its name.

---

You have priced both, so I would like the two numbers: what the prescription actually costs you per month after your insurance, and what the vial costs delivered. The gap people quote in forums is wider than the gap most people face, and nobody publishes the real one.

---

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.