Registered is not approved: the 503B outsourcing facility and the unit its status attaches to

Regulatory vocabulary tends to migrate toward whichever term sounds most like approval. FDA registered, cGMP facility, outsourcing facility and 503B all circulate in descriptions of peptide material, often interchangeably and often at some distance from the statute that produced them. Section 503B of the Federal Food, Drug, and Cosmetic Act is a real and well-specified regime: it imposes substantial obligations on the entities that enter it, and it confers a defined status. What it does not do is approve a drug product. That distinction is not a technicality — the statute is built around it, and nearly every misreading of a 503B claim comes from collapsing it. This post sets out what the section created, what unit its status attaches to, what constrains the material an outsourcing facility may start from, and what any of it bears on research-grade material.

What the Compounding Quality Act created

Section 503B was added to the Act by the Compounding Quality Act, Title I of the Drug Quality and Security Act, signed 27 November 2013 in the wake of the 2012 New England Compounding Center contamination outbreak. Title II of the same law is the Drug Supply Chain Security Act, covered separately in the post on serialized traceability; the two titles are routinely discussed together and address quite different things. Title II asks how a package moves and how that movement is recorded. Title I asks who may compound, under what conditions, and with what starting material.

The problem Title I addressed was structural. Traditional pharmacy compounding under section 503A is patient-specific: a licensed pharmacist or physician compounds a preparation for an identified individual, generally against a prescription. A category of entity had emerged that compounded at industrial scale and shipped without patient-specific prescriptions — functionally manufacturing, while claiming the pharmacy exemptions. The Compounding Quality Act created a voluntary registration category to occupy that space explicitly: an entity that registers as an outsourcing facility may compound without patient-specific prescriptions, and in exchange accepts obligations a 503A pharmacy does not carry.

The central exchange is worth stating precisely, because it is where the misreading starts. Drugs compounded by a registered outsourcing facility can qualify for exemption from the new drug approval requirements and from the requirement that labeling bear adequate directions for use. They are not exempt from current good manufacturing practice requirements. Registration buys a path around approval, not around manufacturing quality obligations, and it does not convert the compounded product into an approved drug. The product remains an unapproved drug, lawfully distributed because a statutory exemption applies to it.

The unit the status attaches to

Approval, when it exists, attaches to a product: a specific formulation, made by a specified process, at specified sites, with a specified label, supported by review of specific data. Nothing in that description is true of 503B registration.

Outsourcing facility status attaches to an establishment. A facility registers with FDA, registers annually thereafter, pays an establishment fee, and appears on FDA’s published register of outsourcing facilities. The registration says this establishment has entered the 503B regime and is subject to its conditions. It says nothing about any individual product made there, because no product-level review occurs — no premarket evaluation of a formulation, no review of its stability data, no assessment of its analytical methods against a dossier.

What substitutes for product review is facility oversight, and the obligations are not trivial. Registered outsourcing facilities are subject to risk-based FDA inspection rather than the largely state-led oversight applying to traditional pharmacies; they must report adverse events to FDA; they must submit a report twice yearly identifying the products compounded over the preceding six months, including the source of ingredients used; and they must compound in compliance with cGMP, by or under the direct supervision of a licensed pharmacist.

That is a demanding compliance surface. It is also, structurally, a statement about a system operating inside a building over a period of time. An inspection finding, a warning letter, or a clean inspection history is evidence about that system — not about the composition of a particular vial, and it was never designed to be. The gap is the one the ICH Q7 and “GMP-grade” discussion turns on: a quality system constrains how a lot was made, while lot-level analytical documentation speaks to what the lot contains.

The bulks list and the clinical-need finding

The constraint doing the most work under 503B concerns starting material, and it is stricter than most summaries suggest. A registered outsourcing facility may compound from a bulk drug substance only if that substance appears on an FDA-established list of bulk substances for which the agency has affirmatively found a clinical need, or if the drug appears on FDA’s drug shortage list. This is the 503B bulks list, and it is not a list of substances FDA has declined to object to — it is a list for which the agency has made a positive determination.

The standard behind that determination is narrow. FDA’s position is that there is no clinical need to compound using a bulk drug substance unless the nomination identifies a drug product that must be produced from the bulk substance rather than from an FDA-approved drug product. The question is not whether a substance is interesting, but whether starting from bulk is necessary. A substance can be entirely uncontroversial and still fail the test, simply because an approved drug exists that a compounder could start from instead.

While the list was being built, FDA operated an interim policy sorting nominated substances into three categories: those under evaluation, those raising significant safety risks, and those nominated without adequate support. In January 2025 the agency issued final interim guidance ending that categorization for substances nominated on or after its publication date. Newly nominated substances now proceed through the advisory-committee and rulemaking route instead — the process described in the account of the July 2026 PCAC meeting, which concerned the parallel 503A list.

Two lists, two statutes, two standards. A substance’s position on the 503A list does not determine its position on the 503B list, and neither bears on whether material may be distributed for laboratory research.

What the statute puts on the label

Section 503B is unusually specific about labeling, and the required elements are themselves informative about what the regime takes itself to be doing.

The label must prominently identify the product as a compounded drug, by the statement “This is a compounded drug” or a comparable alternative. It must carry the lot or batch number, the established name, dosage form and strength, quantity or volume, the date of compounding, an expiration date, storage and handling instructions, and the National Drug Code number where one is available. It must bear “Not for resale”, and — where distributed other than against a prescription for an identified patient — “Office Use Only”.

The statute also restricts onward transfer: the drug is not to be sold or transferred by any entity other than the outsourcing facility that compounded it — a restriction that does not prevent administration in a health care setting or dispensing against a prescription. The combined effect is a product that announces its own compounded status, names its own lot, and is kept out of ordinary distribution channels. The labeling exists precisely so that no recipient mistakes the article for an approved drug, which tells you plainly what the drafters understood the status to mean.

A worked example: the 2026 GLP-1 proposal

The clinical-need standard is easier to read against a live case. In April 2026 FDA announced a proposal to leave semaglutide, tirzepatide and liraglutide off the 503B bulks list, finding no clinical need for outsourcing facilities to compound those three substances from bulk absent a drug shortage; a comment period ran into late June 2026.

The reasoning tracks the standard directly. Approved products containing each of those active ingredients exist and are commercially available, so a nomination has to explain why compounding must start from bulk substance rather than from the approved product — and shortage conditions, which had previously supplied that explanation, no longer obtained. The proposal is not a safety finding about the molecules; it is a finding about the necessity of a starting material, which is the only question the statutory test asks.

That distinction matters for reading coverage of such decisions, which frequently reports them as safety determinations or bans. They are neither. A negative clinical-need finding closes one compounding pathway for one category of entity, on grounds unrelated to whether the substance has been characterized as hazardous.

Where this leaves research-grade material

Research-use-only material sits outside all of it, for structural reasons rather than as a matter of degree. A 503B outsourcing facility compounds finished drug products intended for human administration — dispensed against prescriptions or supplied for office use in health care settings. Research-grade peptide material is not compounded, not dispensed, not labeled for administration, and not lawfully directed toward human use. No 503B determination — a clinical-need finding, an inspection outcome, a registration, or the absence of any of them — creates, expands or restricts any pathway for research-grade material, because the regime does not reach it. The import alert analysis and the Drug Master File discussion reach the same conclusion from different directions, for the same reason: each regime has a defined scope, and scope is the first question, before content.

Which means a 503B reference appearing in a research-material context is doing one of two things. It may be accurate and irrelevant — the supplier has some genuine relationship with a registered facility, governing a different line of business under a different regime. Or it may be inaccurate, borrowing the authority of the term without the substance behind it. Neither case supports an inference about the vial. The facility register is public, so the first question is whether the named establishment appears on it; the second is whether the material in hand has anything to do with that establishment’s 503B activity, which it usually does not.

The evidentiary load lands where it always has. What a specific lot contains is established by analytical documentation and the methods behind it — whether chromatographic purity was distinguished from peptide content, whether the methods were validated against recognized criteria, whether independent analysis supports the certificate where the documentary chain is thin. Registration status is a fact about an organization; composition is a fact about a lot. They are established by different evidence, and one has never stood in for the other.