The most consequential fight over retatrutide is not about weight loss data, and it is not the six lawsuits Lilly filed against sellers on August 12. It is a two-year argument with the FDA about how to count amino acids — and the answer determines whether a compounding pharmacy will ever be allowed to make the molecule.
Lilly says retatrutide has 41 amino acids, which clears FDA's threshold for a protein and makes it a biologic. FDA says the count that matters is smaller than that and the molecule is an ordinary drug. In September 2025 a federal court split the difference, and in February 2026 Lilly appealed. With Lilly planning to file for approval in the first quarter of 2027, the question is now on the clock.
Research-context information only. This article reports on federal litigation, an agency classification decision and published regulatory guidance as they stand. Nothing here is medical or legal advice. Retatrutide is an investigational molecule in Phase 3 clinical trials and is not approved for any use in humans in any country; research-use-only material sold by peptide vendors is not FDA-approved for human use and has not been evaluated for safety, purity or potency as sold. Consult a licensed physician for personal medical decisions.
The 41st amino acid
FDA's definition of a protein lives in 21 C.F.R. § 600.3(h)(6): an alpha amino acid polymer with a specific, defined sequence that is greater than 40 amino acids in size. Anything at or under that line is regulated as a drug. Anything over it is a biological product, licensed under section 351 of the Public Health Service Act rather than approved under the Federal Food, Drug, and Cosmetic Act.
Retatrutide sits directly on the boundary, which is what makes it a test case rather than a footnote. Lilly's position, filed September 3, 2024 in Eli Lilly & Co. v. Becerra, No. 1:24-cv-01503 (S.D. Ind.), is that the molecule contains 41 amino acids in total and that 41 is plainly greater than 40. FDA's position is that only the alpha amino acids in the primary backbone count toward the threshold, and by that method retatrutide falls short — the court record puts the primary chain at 39 alpha amino acids, connected by an isopeptide bond to a second, shorter chain.
Both sides are describing the same molecule. The disagreement is entirely about whether a branched structure joined by an isopeptide bond is one polymer or two, and whether non-alpha residues in the sequence are countable. That is a genuinely hard question of chemistry-meets-drafting, not a case of either party being obviously wrong.
Lilly also pleaded a fallback. Even if retatrutide is not a protein, the statute defines a "biological product" to include anything "analogous to" a protein — a catch-all Congress added precisely because molecules would not always sort neatly. Lilly argued retatrutide qualifies under that clause regardless of the count.
What the court actually held
On September 30, 2025, the Southern District of Indiana issued a split decision, and the split is the whole story.
| Question | Outcome |
|---|---|
| Is retatrutide a "protein" under the 40-alpha-amino-acid test? | FDA won. The court agreed the molecule does not meet the definition as written. |
| Was FDA's reading of "analogous to a protein" lawful? | Lilly won. The court set the decision aside and remanded. |
On the second question the court was blunt about FDA's approach, finding that the agency's "bright line" treatment of the analogous-product category "flouts the statutory text and sidesteps congressional intent." Congress adopted that category, the court reasoned, to capture products that do not already fit the other listed buckets — so reading it as coextensive with the protein definition drains it of meaning. The court allowed that FDA "may have a degree of flexibility in determining the threshold for scientific similarity" but required that the threshold be clearly defined and consistently applied, and remanded with instructions for the agency to identify, in a uniform fashion, the fundamental defining feature of analogous proteins.
That is a remand, not a win on the merits. FDA now has to articulate a principled standard and apply it. It could do that and reach the same conclusion. Lilly did not wait to find out: in February 2026 it appealed to the Seventh Circuit, asking the appellate court to go further and direct that retatrutide be classified as a biologic outright.
Why classification decides compounding
This is the part that matters to anyone tracking access rather than share price.
Biological products licensed under section 351 of the PHS Act are not eligible for the compounded-drug exemptions in sections 503A and 503B of the FD&C Act. FDA has stated the position directly in its compounding guidance: biologics subject to licensure under section 351 fall outside those exemptions, and biological products subject to approval in a BLA will not be considered for the 503A bulks list. The precedent is recent and concrete — when insulin and several other protein products transitioned from NDAs to BLAs on March 23, 2020 under the Biologics Price Competition and Innovation Act, they left the compounding regime on that date. Compounders retained a narrow allowance to mix, dilute or repackage certain biological products under separate guidance, but that is not the same thing as compounding from bulk substance.
Apply that to retatrutide and the fork is stark:
- Approved as a drug under an NDA. A compounding pathway is at least structurally available. It would still require a shortage declaration or a place on a bulks list, and the last two years show how hard both are to get for a GLP-1-class compound — but the door exists.
- Licensed as a biologic under a BLA. The door does not exist. There is no shortage exemption, no bulks list, no route. Compounded retatrutide would be legally impossible after approval, permanently.
The second outcome is the one Lilly is fighting for, and it is worth being clear-eyed that the compounding lockout is not an incidental side effect of the exclusivity argument. It is a durable second benefit of winning.
For context on how narrow the compounding door already is: FDA proposed on April 30, 2026 to permanently exclude semaglutide, tirzepatide and liraglutide from the 503B bulks list, with the comment period closing July 30. And when the Pharmacy Compounding Advisory Committee voted in late July on adding peptides to the 503A bulks list, the six compounds it recommended were BPC-157, KPV, TB-500, MOTS-c, epitalon and semax — no GLP-1-class molecule was on the table at all.
What changes for buyers, and what does not
Nothing changes today. Retatrutide is not approved in any country. It remains investigational, in Phase 3, with the TRIUMPH programme reporting through 2026 and a filing expected in Q1 2027. The classification dispute concerns the status of a product that does not yet exist on the market. Current vendor pricing, coupon status and COA availability on the retatrutide comparison surface are entirely unaffected by it.
The dispute explains the enforcement, though. The six lawsuits Lilly filed on August 12 against sellers, the 14,000-plus listings it says it has reported and the 200-plus referrals it has made to regulators are not separate from this. A company arguing that its molecule deserves 12 years of protection has an obvious interest in demonstrating that it polices the molecule now. The classification fight and the enforcement campaign are two halves of the same posture.
The long-run implication runs the other way from what you might expect. Historically, the compounding channel is what absorbed demand when a branded GLP-1 was short or unaffordable — that is exactly what happened with semaglutide and tirzepatide from 2022 through 2025. If retatrutide is licensed as a biologic, that pressure valve is welded shut from day one. There would be no compounded version at any point in the product's life, only the branded product and whatever biosimilars eventually arrive after 12 years. Whatever your read on how that plays out, it is a structurally different market than the one the last generation of GLP-1s created.
