Compounded vs branded peptides: the regulatory architecture, explained

The short version: a compounded drug is not an FDA-approved drug. It has not been reviewed for safety, efficacy, or quality before it reaches a patient. It's prepared, on a prescription basis, by a pharmacy operating under a separate set of federal exemptions - exemptions that exist for narrow clinical reasons and that the FDA polices when they get stretched.

For most of pharmacy history, this distinction was technical. A patient allergic to a dye in a commercial tablet got a custom version from the local compounder. Volumes were small. Risk was contained.

Then came the GLP-1 wave. Between 2022 and early 2025, compounders filled the gap when Mounjaro, Ozempic, Wegovy, and Zepbound went onto the FDA's drug shortage list, and the legal exemption that lets compounders make a near-copy of a branded drug during a shortage was suddenly carrying tens of millions of doses. By the time the shortages resolved in late 2024 and early 2025, the question of which pharmacy was operating inside the lines and which was outside had become a multi-billion-dollar policy fight - one now playing out in a federal court in the Northern District of Texas.

The rules matter again. Here is how they got built.

On November 21, 1997, President Clinton signed the Food and Drug Administration Modernization Act. Buried inside it was Section 503A, the first time Congress had written down, in statute, what a compounding pharmacy was allowed to do without being treated as a drug manufacturer ^[1].

The logic of 503A goes like this. A pharmacist, working from a prescription written by a licensed physician for an identified patient, prepares a drug tailored to that patient's clinical need. Because the patient, the prescriber, and the pharmacist are all named - what the statute calls the triad - the drug doesn't have to pass through the FDA's pre-market approval process. It doesn't have to be made under Current Good Manufacturing Practice. Its labeling doesn't have to match the federal labeling rules for commercial drugs.

In exchange for these exemptions, compounders had to accept restrictions. They couldn't compound a drug that was "essentially a copy" of a commercially available product. They couldn't compound from bulk ingredients unless those ingredients appeared on a specific FDA-approved list, or were components of an FDA-approved drug, or had a USP or NF monograph. And - in the original 1997 text - they couldn't advertise the particular compounded drugs they made.

That advertising restriction is what cracked the statute open. In 2002, the Supreme Court struck it down as a First Amendment violation in Thompson v. Western States Medical Center. Some federal circuits read the ruling as invalidating all of 503A, since Congress had said the provisions were non-severable. For roughly a decade, the compounding industry operated under inconsistent federal authority depending on which circuit you were standing in.

So the statutory ground was already unstable when, in the fall of 2012, an outbreak in Massachusetts forced Congress to rebuild it.

A 503A pharmacy is what most people picture when they hear "compounding pharmacy." It's licensed by a state board of pharmacy. It dispenses on patient-specific prescriptions. It's exempt from FDA pre-market approval, from CGMP, and from federal labeling requirements - but only when it stays inside the 503A box.

The box has hard walls.

The triad. The physician writes for an identified patient. The pharmacist prepares the drug for that patient. No bulk production for inventory, no anticipatory compounding beyond limited quantities, no shipping wholesale to clinics.

Bulk ingredient rules. A 503A pharmacy can only compound from an active ingredient that's either (a) the subject of an applicable USP or NF monograph, (b) a component of an FDA-approved drug, or (c) on the FDA's 503A bulks list. Semaglutide has no USP monograph. Its only legal pathway into a 503A compound was (b) - being a component of an FDA-approved drug, which means the base form of semaglutide, not the sodium or acetate salts. More on that in a moment.

Interstate limits. A 503A pharmacy distributing real volumes across state lines, or to prescribers who haven't yet identified a patient, drifts out of the exemption. State boards police some of this. The FDA polices the rest.

No copies. The "essentially a copy" prohibition is the wall that matters most for GLP-1s. A compounder can't legally make a near-duplicate of a commercially available branded drug - unless the branded drug is on the FDA shortage list, in which case the exemption opens ^[2].

That last sentence is the doorway the entire 2022-2025 compounding wave walked through.

A 503B outsourcing facility looks more like a small manufacturer than a pharmacy. It registers with the FDA voluntarily. It accepts CGMP requirements. It submits to FDA inspection on a regular cadence. In return, it gets two things 503A doesn't: it can compound without a patient-specific prescription, and it can ship in volume across state lines to hospitals, clinics, and physician offices.

503B didn't exist before 2013. It was invented by Congress in direct response to a single mass-casualty event.

The 503B model fits a particular use case: a hospital that needs a standard sterile preparation in bulk, ready for the next patient who walks in, rather than waiting for a prescription-by-prescription preparation. Pre-2013, this kind of bulk preparation was happening anyway, in a regulatory gray zone where state-licensed pharmacies were operating at industrial scale without federal manufacturing oversight. One of those operations was the New England Compounding Center.

In September 2012, a Tennessee neurosurgery clinic noticed a cluster of patients developing fungal meningitis after epidural steroid injections. The lots traced back to the New England Compounding Center in Framingham, Massachusetts. By the time the CDC closed the investigation, 793 patients across 20 states had been infected. 64 died.

The contaminant was Exserohilum rostratum, a fungus that doesn't belong anywhere near a sterile injection. Investigators found it in vials of preservative-free methylprednisolone acetate that NECC had prepared in bulk and shipped to pain clinics and hospitals nationwide. The Massachusetts state pharmacy board had inspected NECC. So had the FDA. Neither had stopped what was happening.

NECC was operating, in practice, as a drug manufacturer - preparing thousands of vials of a standard product and shipping them to identified clinics rather than identified patients - while holding only a state pharmacy license. It was running outside the 503A triad and inside no other federal framework, because no other federal framework existed for what it was doing. The cleanroom conditions described in subsequent federal indictments were, plainly, not sterile.

14 NECC employees were eventually indicted ^[3]. The pharmacy's co-founder was convicted on racketeering and mail fraud charges. The president, charged with second-degree murder, was acquitted on the murder counts but convicted on others ^[4].

The outbreak forced a question Congress had been ducking since 1997: what about the pharmacies that have outgrown the 503A model? What rules apply to them?

Representative Fred Upton introduced H.R. 3204 - the Drug Quality and Security Act - in September 2013. It passed the House by voice vote on the 28th. The Senate cleared it on November 18. President Obama signed it on November 27, just over a year after the NECC outbreak began.

The DQSA did three structural things.

It created Section 503B - the outsourcing facility designation - giving large-volume compounders a federal pathway with CGMP and FDA inspection in exchange for the right to compound without patient-specific prescriptions ^[5]. It cleaned up 503A by removing the advertising restriction the Supreme Court had already struck down, putting the statute on stable constitutional footing. And it built the federal drug-supply track-and-trace system that's now the Drug Supply Chain Security Act.

The intent was clear. Large-volume bulk compounding would have a federal home, with federal oversight. Traditional pharmacy compounding would stay where it had always been, under state boards, with the 503A triad as the federal exemption.

What Congress did not anticipate is how the shortage list would interact with this architecture once a blockbuster drug went short.

A 503A or 503B compounder can't, ordinarily, make a near-copy of a commercially available branded drug. The FD&C Act explicitly prohibits it. But the prohibition has a carve-out: if the FDA places the branded drug on its drug shortage list, the "essentially a copy" bar effectively lifts. Compounders can step in and prepare the same active ingredient, in roughly the same form, to fill the gap.

This is sensible policy in the abstract. A cancer drug goes short; compounders bridge the supply gap for hospitals; patients keep getting treated.

It becomes more complicated when the drug in shortage is a once-weekly injection used by tens of millions of people for weight loss. The shortage doorway, designed for narrow clinical bridging, becomes a wide commercial channel - and that's what happened starting in 2022.

Tirzepatide (Mounjaro) was approved by FDA in May 2022 for type 2 diabetes. Demand outran supply almost immediately. Semaglutide products - Ozempic, then Wegovy in 2021, then expanded Wegovy uptake through 2022 and 2023 - followed the same arc. By late 2022 the FDA had placed both molecules on the drug shortage list.

That listing flipped the compounding switch.

503A pharmacies began compounding semaglutide and tirzepatide from bulk active ingredient sourced from FDA-registered API manufacturers. 503B outsourcing facilities began compounding the same molecules at scale and shipping to physician offices and telehealth platforms. New telehealth brands appeared, built almost entirely around compounded GLP-1 supply. Volume estimates from the period are imprecise, but by the back half of 2024 compounded GLP-1s were almost certainly being prescribed in the millions of doses per month.

Then the shortage resolved.

Tirzepatide came off the shortage list on December 19, 2024 ^[6]. Semaglutide came off in March 2025 ^[7]. The doorway began to close.

FDA enforcement had been ramping all along. The agency was particularly focused on a specific technical question: which form of semaglutide could legally be compounded at all.

The base form of semaglutide is a component of Ozempic, Wegovy, and Rybelsus. Under 503A, it's eligible because it's a component of an approved drug.

But some compounders, and some bulk API suppliers serving them, were working with salt forms - semaglutide sodium and semaglutide acetate. These are chemically distinct from the base. They're not components of any FDA-approved drug. They have no USP monograph. Under any reasonable reading of 503A, they can't be compounded.

On July 26, 2024, the FDA issued a formal alert flagging dosing errors with compounded semaglutide and explicitly calling out the illegal use of the salt forms ^[8]. The Washington State Department of Health followed a month later with a regulatory bulletin saying the same thing in plainer language: only the base is permitted ^[9].

Enforcement letters followed. A January 15, 2025 warning letter to a compounder in Exton, Pennsylvania. Another in April to a Phoenix facility. In late February the FDA announced it would use "all available compliance and enforcement tools" against unsubstantiated compounded GLP-1 marketing claims ^[10], and on June 26 it issued roughly 30 warning letters at once to online wellness clinics for misleading advertising ^[11].

The deadlines for the shortage-removal enforcement also took effect: April 22, 2025 for 503A pharmacies to stop compounding semaglutide, May 22 for 503B facilities. Both subject to the outcome of Outsourcing Facilities Association v. FDA, the OFA's challenge pending before Judge Pittman in the Northern District of Texas.

On December 2, 2024, the American Diabetes Association issued a formal statement recommending against the use of compounded GLP-1 and dual GIP/GLP-1 receptor agonists ^[12].

This matters for two reasons. First, the ADA is the professional society whose clinical practice guidelines define the standard of care for type 2 diabetes in the United States. When ADA tells endocrinologists not to use a category of product, that recommendation carries weight in malpractice analysis, in insurance coverage decisions, and in how individual physicians counsel patients.

Second, the statement was not framed as a regulatory opinion. It was framed as a patient-safety position, citing dosing errors, inconsistent potency, and uncertain sterility in the compounded supply. The Alliance for Pharmacy Compounding pushed back hard in January 2025, arguing the ADA had relied on selective evidence and accusing it of carrying water for the branded manufacturers ^[13]. That dispute is unresolved.

The bottom line is that the country's leading diabetes society is now on record against the compounded supply that millions of patients have been using.

Compounded semaglutide and tirzepatide are, as of mid-2025, no longer permitted under the shortage exemption. The doorway has closed. What's left are narrow patient-specific pathways: a 503A pharmacy can still prepare a customized version of a drug for a particular patient with a documented clinical need that the commercial product can't meet - a different dose, a removal of an excipient the patient can't tolerate, a route of administration the branded product doesn't offer. That's the original 503A use case, and it still applies.

What is no longer legal: mass production of standard-strength semaglutide or tirzepatide injectables on the theory that the branded supply is constrained. The supply is no longer constrained. The exemption is no longer available.

What was never legal: any compounded product using semaglutide salts. Sodium or acetate forms have always been outside the bulk ingredient rules. The FDA's July 2024 alert did not change that - it stated it.

What remains contested: the OFA litigation in Texas, which seeks to enjoin FDA enforcement on procedural and statutory grounds. The case has not been resolved. Some 503B facilities have continued compounding while the case proceeds, citing the litigation as authority.

The substantive difference between a branded GLP-1 and a compounded one isn't the molecule. The base semaglutide in a properly sourced compound is the same active ingredient as the base semaglutide in Wegovy.

What differs is everything around it. A branded manufacturer makes the drug under CGMP, in a process the FDA has reviewed, with each lot tested to specifications also reviewed by the FDA, in a fill-finish operation audited on a known cadence. The label tells you, with regulatory backing, what's in the vial.

A compounded preparation made by a competent 503B operating inside USP 797 may be quite close to that standard. A compounded preparation made by a 503A pharmacy with newer staff, intermittent state inspections, and API from a supplier whose certificate of analysis hasn't been independently verified is a different proposition. The FDA's reports on compounded GLP-1 adverse events through 2024 and into 2025 included dosing errors at multiples of the intended dose - patients receiving five or ten times what their prescription called for - alongside concerns about endotoxin, sterility, and potency variation.

The variance is the point. A branded product is a known quantity. A compounded product is as good as the specific facility that made it, and the patient generally has no way to evaluate that facility.

The regulatory trajectory now points toward tighter enforcement. The shortage doorway has closed for the two largest compounded categories. The FDA has signaled, through the February 2025 statement and the June warning letter wave, that it intends to police marketing claims as aggressively as it polices production. State boards are following the FDA's lead on salt forms. The ADA has aligned the clinical guideline community against the compounded supply.

Whether the OFA litigation produces a meaningful carve-out remains to be seen. Whether Congress revisits the shortage-list mechanism - perhaps to prevent a blockbuster drug from triggering years of bulk compounding the next time it happens - is the harder question, and not one this Congress seems eager to take up.

What's clear is that the architecture built in 1997, rebuilt in 2013, and stress-tested between 2022 and 2025 is no longer ambiguous about GLP-1s. The patient-specific exemption still exists. The shortage exemption is, for these molecules, gone.

For patients and prescribers trying to read the current landscape: the question to ask a compounder is not whether their product works. The question is which legal pathway they believe authorizes them to prepare it, and whether that pathway actually applies to your prescription.