Chapter 38: Peptide Regulation — Approval, Off-Label, Compounding, and Doping

"The life of the law has not been logic: it has been experience." — Oliver Wendell Holmes Jr., The Common Law (1881)

Overview

Almost everything in this book is about evidence. This chapter is about a different thing that constantly gets mistaken for evidence: regulatory status.

"FDA approved." "Not FDA approved." "It's legal in Europe." "It's banned." "It's just a research chemical." "It's a supplement." Every one of those phrases arrives sounding like a verdict on whether a molecule works, and not one of them is. They are statements about paperwork — about what was submitted, by whom, to which agency, in which country, under which category, and how that agency judged it. Sometimes the paperwork tracks the evidence closely. Sometimes it tracks nothing but the economics of who could afford to file.

Learning to tell those apart is the last piece of the reading skill this book has been assembling.

Two warnings before we start, and the first one is important enough that it will be repeated several times.

This is not legal advice, and it is not a description of the law where you live. Regulatory rules differ enormously by country, and they change — sometimes annually, sometimes in the middle of litigation, sometimes because an agency issued guidance on a Tuesday. Every specific rule in this chapter carries a jurisdiction and a date, because a rule without those two things is not a fact, it is a rumor. What you should take from this chapter is the structure and the reasoning — why categories exist, what pressures they respond to, what they can and cannot tell you. If you have an actual question about your own situation, the answer is not in a book. It is in the current rules of your own jurisdiction, and a professional who works with them.

And this chapter will not tell you what you are allowed to buy, import, or possess. Not because the question is uninteresting, but because the answer is different in every country, changes without notice, and is not something a general-audience textbook can responsibly assert. This book explains how the system is built. It does not tell anyone how to move through it.

The second warning is subtler. Regulation invites a particular kind of lazy thinking, in which approval becomes a synonym for "works" and non-approval becomes a synonym for either "doesn't work" or "they don't want you to have it," depending on who is talking. Both readings are wrong. The whole last section of this chapter exists to dismantle them, and the dismantling runs in both directions — against the credulous and against the cynical equally.

In this chapter, you will learn to:

• State what an approval is actually a judgment about, and what it leaves entirely open
• Apply the forty-amino-acid line and explain why a number in a definition decides who gets to compete on price a decade later
• Trace the approval pathway end to end and say where compounds die and why
• Break "not FDA approved" into its five distinct meanings and assign the right one
• Explain the deliberate asymmetry between off-label prescribing and off-label promotion
• Distinguish compounding from manufacturing, and 503A from 503B
• Recognize "research use only" and "sold as a supplement" as the classification claims they are
• Explain why WADA's S0 makes "it isn't on the banned list" a bad argument by construction
• Read a regulatory status without letting it move an evidence rating in either direction

Learning Paths

All five paths should read §38.4 and §38.10. They are the two sections that change how you read a sentence, and they are short enough to justify the detour from any track.

💊 GLP-1 — §38.2 (the forty-amino-acid line) and §38.6 (compounding and the shortage mechanism) are the two sections that explain why the compounded-semaglutide market appeared, what it was, and why it contracted. Read them with Chapters 12 and 19 in mind. 🏋️ Performance — §38.8 is the one that matters most, and it matters more than most athletes think. If you are tested, S0 is the rule that catches people. §38.7 explains the vial label. 🔬 Science — read straight through. §38.3 and §38.10 connect directly to Chapters 10 and 16, and §38.2 is the cleanest example in the book of a definitional line with enormous downstream consequences. 💄 Cosmetic — §38.4, case (5). A cosmetic ingredient is not an unapproved drug; it is not a drug at all in the regulatory sense, and that distinction does most of the work in Chapter 30. 🏥 Clinical — §38.5 and §38.6 are the daily-practice sections. Much of what patients believe about "approved" is fixable in one careful sentence, and §38.5 supplies it.

⚠️ Read this before the rest of the chapter

Nothing in this chapter is legal advice. It describes regulatory structures — mostly those of the United States, because they are the ones most often referenced in English-language peptide marketing — for the purpose of teaching you how to read a claim. Those structures differ by country and they change.

Every specific rule below is tagged with a jurisdiction and, where the timing matters, a date. If you are reading this well after 2026, assume the specifics have moved and the reasoning has not. Verify anything that matters to you against the current rules of your own jurisdiction, with someone qualified to interpret them.

38.1 What "approved" actually means

Start with the definition, because almost every error in this chapter is a failure to hold it steady.

An approval is a regulator's judgment that, for a specified indication and population, the evidence submitted shows the benefits outweigh the risks — together with an approved label describing that use, the population studied, the dosing evaluated, the adverse events observed, and the warnings the agency considers necessary.

Read that sentence twice. Every clause is load-bearing, and three consequences follow that readers routinely miss.

Approval is indication-specific. It is not a certificate about a molecule. It is a judgment about a molecule used a particular way, in a particular kind of patient, to achieve a particular result. Semaglutide is approved for type 2 diabetes. Semaglutide is approved for chronic weight management. Those are two separate approvals, resting on two separate development programs, with two separate labels and two different brand names for the same active molecule (Chapter 12). A regulator saying yes to one says nothing about the other until the second application is filed and judged. The sentence "semaglutide is approved" is therefore incomplete in the same way "the answer is 4" is incomplete — approved for what?

Approval is based on the evidence submitted. This is the clause that does the most work later in this chapter. A regulator does not survey the literature and issue a verdict on a molecule. It evaluates a dossier that a sponsor assembled and filed. Which means that an unapproved use is not necessarily a rejected use. It may simply be a use that nobody ever built a dossier for. Those two situations look identical from outside — the label says nothing about it either way — and they carry completely opposite evidentiary weight. §38.4 is entirely about that distinction.

Approval is jurisdiction-specific. Agencies are national or regional. The United States Food and Drug Administration, the European Medicines Agency, the United Kingdom's Medicines and Healthcare products Regulatory Agency, Japan's Pharmaceuticals and Medical Devices Agency, Australia's Therapeutic Goods Administration, and dozens of national authorities each make their own decisions under their own statutes. They cooperate, they read each other's reviews, they often converge — and they routinely reach different conclusions, on different timelines, for reasons that range from genuine scientific disagreement to a sponsor simply deciding a market was too small to file in (§38.9).

What approval is not

It will save time to state the negatives explicitly, because each one is a claim somebody has actually made.

Approval is not a statement that the drug works for everyone in the approved population. Trials report averages and distributions; a drug can be genuinely approved and genuinely fail to help a substantial minority of the people who take it.

Approval is not a statement that all long-term effects are known. Phase 3 trials typically run for months to a few years in a few thousand people. Rare adverse events and long-horizon effects are frequently discovered afterward, which is why post-marketing surveillance exists (§38.3).

Approval is not a statement that the drug is better than the alternatives. Many approvals rest on placebo-controlled trials. Superiority over an existing treatment is a separate question that sometimes nobody has asked.

Approval is not permanent. Drugs are withdrawn. Labels are revised. Boxed warnings are added. Indications are removed. An approval is a judgment at a point in time on a body of evidence, and evidence accumulates.

🔍 Check Your Understanding

1. Someone says "semaglutide is FDA approved." What is missing from that sentence, and why does the omission matter for insurance coverage (Chapter 12)?
2. A use is not on a drug's label. Name the two very different situations that could produce that, and explain why they are not evidentially equivalent.
3. Why does "approval is based on the evidence submitted" mean that a label is a better guide to what a sponsor filed than to what is known about a molecule?

38.2 The forty-amino-acid line: drug or biologic

Here is the most concrete regulatory fact in this book, and it is a number.

In the United States, an alpha-amino-acid polymer of more than 40 amino acids is regulated as a biological product. Forty or fewer, and it is regulated as a drug.

That is not a metaphor or a rule of thumb. It is a definitional line, drawn in United States regulation, and it sorts the entire subject matter of this book into two legal categories with different application types, different review divisions, different exclusivity rules, and — this is the part that matters — completely different pathways for competition after the original exclusivity ends.

THE FORTY-AMINO-ACID LINE            (United States; as of this writing, 2026)

    ≤ 40 amino acids                 │       > 40 amino acids
    ───────────────────────────      │       ───────────────────────────
    regulated as a DRUG              │       regulated as a BIOLOGICAL PRODUCT
    approved via an NDA              │       licensed via a BLA
    (New Drug Application)           │       (Biologics License Application)
                                     │
    follow-on competition:           │       follow-on competition:
      GENERICS                       │         BIOSIMILARS
      abbreviated application        │         abbreviated but far more demanding
      resting on demonstrated        │         analytical + clinical comparability
      BIOEQUIVALENCE                 │         + a separate INTERCHANGEABILITY
                                     │           determination
    cost to bring a follow-on:       │       cost to bring a follow-on:
      relatively low                 │         high — often orders of magnitude higher
                                     │
    ── examples ──                   │       ── examples ──
    semaglutide      31 residues     │       insulin       51 (two chains)
    tirzepatide      39 residues     │       growth hormone  191
    teriparatide     34 residues     │       erythropoietin  ~165
    oxytocin          9 residues     │
    BPC-157          15 residues     │
                                     │
              ↑ the line sits here, between 40 and 41 ↑

Nothing chemical happens at residue 41. Chapter 1 made this point about the peptide/protein convention and it is the same point here, except that in this case the arbitrary line has legal force. A 40-residue chain and a 41-residue chain are the same kind of object, made of the same bonds, and one of them will be competed against by generics and the other by biosimilars.

Why the two pathways are not equivalent

A generic drug application is abbreviated because it does not have to re-prove clinical efficacy. The applicant demonstrates that its product contains the same active ingredient in the same form and achieves bioequivalence — essentially, that the same amount of drug shows up in the bloodstream on the same time course. If that is shown, the agency accepts that the original efficacy data applies. This is why generic small molecules and generic short peptides can be cheap: the expensive part of drug development, the clinical program, is not repeated.

A biosimilar cannot work that way, because a large biological molecule made in living cells is not reproducible atom for atom by a second manufacturer using a different cell line and a different process. So the biosimilar pathway asks for extensive analytical characterization, usually pharmacokinetic and immunogenicity studies, and often a comparative clinical study — and then, if the sponsor wants pharmacy-level substitution, a further interchangeability determination with its own requirements. The result is a route that is genuinely abbreviated relative to a full original application and still vastly more demanding, slower, and more expensive than a generic filing.

The practical consequence: the more-than-40 side of the line has fewer competitors, later, at higher prices.

The worked example: insulin crossed the line

Insulin is the cleanest case available, because it did not stay on one side.

Insulin is 51 amino acids across two chains (Chapter 11) — comfortably over the line. But insulin was approved in the United States long before the modern biologics framework existed, and it was historically regulated as a drug, under NDAs. That legacy classification had a strange effect: it sat in the drug system, where the generic pathway theoretically applied, but the generic pathway does not actually work for a molecule like insulin, because bioequivalence for a recombinant protein made in living cells is not a straightforward demonstration. So insulin got the worst of both worlds — a drug-side classification whose competition pathway was, in practice, unusable, and no access to the biosimilar pathway that was designed for exactly this kind of molecule.

Under a statutory transition that took effect in March 2020, insulin and a set of other protein products were deemed to be biological products, moving them from NDAs to the biologics framework. That transition is what opened the biosimilar pathway for insulin. Whether the resulting price effects have been large or modest is a separate argument — and a live one — but the mechanism is not in dispute: a reclassification changed which follow-on pathway existed.

Now the contrast. Semaglutide's peptide backbone is 31 residues (Chapter 33 covers the modifications hung off it). Thirty-one is on the drug side. Tirzepatide, at 39, is also on the drug side, with two residues to spare. These are the most commercially consequential peptides of the decade, and when their exclusivity ends, the follow-on route available to competitors will be the generic route, not the biosimilar route.

Same molecular class, split by a number, with completely different competitive consequences.

This connects directly to two arguments elsewhere in the book. Chapter 12 laid out the access problem — what these drugs cost, who can get them, and what happens to people who stop. Section 36.9 covered the patent estate that sustains exclusivity in the first place. The forty-amino-acid line is the third leg: how a molecule is classified shapes what happens after exclusivity ends, and therefore shapes the price trajectory that determines access a decade out. A definitional threshold written into regulation is doing part of the work that people usually attribute entirely to patents.

🧬 The Molecule — counting residues, and where the count gets slippery

If a number decides the regulatory category, the number had better be well defined. Mostly it is, and occasionally it is not.

The definition in United States regulation, as of this writing in 2026, concerns an alpha amino acid polymer with a specific, defined sequence exceeding 40 amino acids in size. Read the qualifiers. Alpha amino acid — the standard backbone from Chapter 1, not the exotic variants. Specific, defined sequence — a characterized molecule, not a heterogeneous mixture. And size measured in amino acids, not in daltons.

That last point produces a counterintuitive result. Semaglutide is roughly 4,100 Da and insulin is roughly 5,800 Da — not an enormous gap in mass. But semaglutide is 31 residues and insulin is 51, and it is the residue count that decides. A heavily modified peptide can carry a great deal of non-amino-acid mass, in fatty acid chains and linkers and conjugates, without moving toward the line at all, because those atoms are not amino acids.

There are genuine edge cases. Chains joined by disulfide bonds — is a two-chain molecule counted as one polymer or two? Peptides with substantial non-peptide components. Molecules where the "specific, defined sequence" requirement is arguable. These get resolved case by case, and the resolutions are the kind of thing that changes. What does not change is the shape of the question: how many alpha amino acids, in a defined sequence, in the polymer?

38.3 The approval pathway, end to end

The pathway is worth walking once as a whole, because most people know the phase names without knowing what each phase is actually asking.

THE DEVELOPMENT PATHWAY          (US framing; other agencies are structurally similar)

  PRECLINICAL          in vitro work, animal pharmacology, toxicology, manufacturing
                       Nothing here has been given to a human being.
        │
        ▼
  IND                  Investigational New Drug application. Permission to administer
                       to humans. Not an approval of anything — a permission to test.
        │
        ▼
  PHASE 1              Safety, tolerability, pharmacokinetics. Small — often tens of
                       people — and frequently in healthy volunteers.
                       Question: what does this do to a human body at all?
        │
        ▼
  PHASE 2              Dose-finding and preliminary efficacy in people with the
                       condition. Hundreds. Often the first honest look at whether
                       the drug does the thing.
                       Question: does it plausibly work, and at what exposure?
        │
        ▼
  PHASE 3              Adequate and well-controlled efficacy trials in the target
                       population. Hundreds to many thousands. Randomized,
                       usually blinded, powered for a prespecified endpoint.
                       Question: does it work, compared to what, by how much?
        │
        ▼
  APPLICATION REVIEW   NDA or BLA (§38.2). The agency reads the whole dossier,
                       including manufacturing and the proposed label.
        │
        ▼
  PHASE 4 / POST-      Surveillance for rare and long-latency effects, plus any
  MARKETING            confirmatory trials the agency required as a condition.
                       Never truly finished.

Two things Chapter 10 established are worth restating here in regulatory terms.

Most compounds entering this pipeline do not finish it. The attrition is severe, and it is not evenly distributed. Compounds are lost at every stage, but the losses concentrate at the point where efficacy in humans is actually tested — late Phase 2 and Phase 3. That is exactly where you would expect them to concentrate, because that is the first place the question "does this help a person" gets a real answer. Everything before that stage is, in a precise sense, a prediction.

This matters for how you read Part III. A compound with encouraging animal data and no human efficacy trials is not "nearly there." It is sitting immediately before the stage where most encouraging compounds die. The base rate is the whole story, and Chapter 10 gives it.

Regulators also run accelerated pathways. Under defined circumstances — usually serious conditions with unmet need — an agency may approve a drug on the basis of a surrogate endpoint reasonably likely to predict clinical benefit, rather than on the clinical benefit itself, and require confirmatory trials afterward.

Now connect this straight to Chapter 16, because the connection is not decorative. Chapter 16 explained why a surrogate endpoint can move convincingly while the outcome patients care about does not. The surrogate-endpoint problem is not a hypothetical criticism of regulation. It is a designed feature of it. Accelerated approval deliberately trades certainty for speed, and the trade is explicit: confirmatory evidence is deferred, not waived. The system's integrity depends entirely on those confirmatory trials being run, being adequate, and having consequences when they fail.

Which is why the phrase "approved on an accelerated pathway" should change how you read an approval. It does not mean the drug does not work. It means the agency judged that a surrogate was reasonably likely to predict benefit and that waiting for the outcome data was worse than not waiting. That is a defensible judgment. It is also a different epistemic object from a full approval on a hard clinical endpoint, and treating the two identically discards real information.

💊 In the Clinic — what "in trials" tells a patient, and what it doesn't

Patients and clinicians constantly encounter the sentence "it's in clinical trials," usually offered as reassurance. Here is how to unpack it in one question: which phase, and for what indication?

"In Phase 1" means human beings have received it and the study is measuring whether they tolerate it. That is a genuine milestone over having never been given to a human at all — but a Phase 1 study in healthy volunteers is not designed to detect efficacy, is usually far too small to detect uncommon harms, and says nothing about the condition anyone hopes to treat.

"In Phase 3 for X" is a substantially stronger statement, because it means a sponsor spent a great deal of money on the belief that Phase 2 justified it — and that a prespecified efficacy question in the target population is being answered under randomization.

Two cautions. First, a registered trial is not a result. Trial registries list studies that were planned; a registration number is not evidence, and marketing that cites registry entries as though they were findings is doing something dishonest. Second, "in trials" often refers to a different indication, a different formulation, or a different route than the one being sold. Chapter 5's discipline applies: name the population, the intervention, the comparison, and the endpoint before you accept the sentence.

38.4 The five meanings of "not FDA approved"

This is the most useful section in the chapter for a general reader, and it fixes a single confusion that runs through the entire consumer peptide conversation.

The phrase "not FDA approved" is used as though it means one thing. It means at least five, and they carry completely different evidentiary weight.

"NOT FDA APPROVED" — FIVE DIFFERENT SITUATIONS

  (1) NEVER SUBMITTED          No application was ever filed.
                               → says nothing about the molecule.
                               → says everything about the absence of a sponsor.

  (2) SUBMITTED AND REJECTED   Filed, reviewed, and refused — or withdrawn after a
      (or withdrawn)           negative review.
                               → a genuinely negative signal.

  (3) APPROVED ELSEWHERE,      A registered medicine in another jurisdiction.
      NOT HERE                 → could be a different evidence standard.
                               → could be that no sponsor filed in this market.
                               → these are NOT equivalent and you have to ask which.

  (4) APPROVED, BUT NOT        Off-label use of an approved drug (§38.5).
      FOR THIS USE             → the drug is approved; this use is not on the label.

  (5) NOT A DRUG AT ALL        Cosmetic ingredient, research reagent, or a product
      (in the regulatory       marketed as a supplement.
       sense)                  → "approved" was never the applicable category.

  Same four words. Five situations. One of them is a negative finding and the
  other four are not.

Take them one at a time.

(1) Never submitted. Nobody filed an application. This is the situation for the great majority of the compounds in Part III of this book — the gray-market peptides that generate the most enthusiastic claims and the least data. There is no rejection on file because there is no file. This says nothing about the molecule and everything about the absence of anyone willing to fund a trial.

Understand what that funding requires. A modern development program is an enormous, sustained expenditure with a high probability of failure (§38.3). Someone has to be willing to make that bet, and the only entities that make it are those who expect to recover the cost — which requires a defensible commercial position, which usually requires patent protection, which is difficult to obtain for a short, well-known, long-published sequence. Chapter 36 laid out that economics. The result is a structural gap: compounds that are cheap to make and impossible to own do not get developed, whether or not they work.

(2) Submitted and rejected. An application was filed and refused, or a sponsor withdrew it after a review that was going badly. This is a genuinely negative signal — a regulator looked at the dossier the sponsor considered its best case and was not persuaded that benefits outweighed risks for the proposed use. It is not proof the molecule is inert; sponsors sometimes file with inadequate evidence, and sometimes a later program with a better design succeeds. But of the five cases, this is the only one where non-approval is itself evidence about the molecule.

(3) Approved elsewhere, not here. The compound is a registered medicine in some jurisdiction and not in the one being discussed. Thymosin alpha-1 (Chapter 18) is the standing example in this book: registered and used in a number of countries, not approved in others. Several compounds in the nootropic and neuropeptide space (Chapter 23) have the same shape — approved somewhere, unapproved elsewhere, described by sellers in the unapproved market as though the foreign registration settled the evidence question.

It does not, and the reason is that this case has at least two very different underlying stories. Either the two agencies applied different evidence standards to the same dossier and reached different conclusions, or the sponsor simply never filed in the second market — because the market was small, or the cost of a second regulatory program was not worth it, or the company that owned the molecule in one region had no counterpart in the other. Both explanations are common. They are not equivalent, and you cannot tell which applies without looking. §38.9 develops this.

(4) Approved, but not for this use. The molecule is an approved drug; the use under discussion is simply not on its label. This is off-label use, it is extremely common, and it is lawful for prescribers in the United States and many other jurisdictions. It gets its own section (§38.5) because it is the case most often confused with case (1), and the confusion runs in both directions — people treat off-label use as though it were unapproved-and-unstudied, and people treat unapproved compounds as though they were merely off-label.

(5) Not a drug at all in the regulatory sense. Some products are not seeking drug approval because "drug" is not the category they sit in. A cosmetic ingredient (Chapter 30) is regulated under a different framework entirely, one that generally does not involve premarket approval and that comes with a hard constraint on what may be claimed — a product making a therapeutic claim risks being treated as an unapproved drug precisely because it made the claim. A research reagent sold for laboratory use is not a drug because it is not offered for human use, at least on paper (§38.7). A product marketed as a dietary supplement sits in a category that, in the United States, has no premarket approval requirement at all — which is not the same as having passed one.

The lesson, stated plainly

"Not FDA approved" is a statement about a regulatory file, not about a molecule. The five cases above carry completely different evidentiary weight, and the phrase flattens them into one syllable of disapproval that fits none of them well.

Now apply it, using the compound this book has spent the most time on.

For BPC-157, the answer is case (1). No application has been submitted for approval as a drug in the United States. There is no rejection, no negative review, no adverse regulatory finding on the merits of the molecule as a therapeutic. The file is empty.

And here is the discipline: Chapter 17's ❌ rating for BPC-157 does not rest on that regulatory status at all. It rests on the published human literature, which is — as of this writing, in 2026 — essentially absent for the claims being made. If a well-designed randomized human trial were published tomorrow showing benefit, the rating would move, and the regulatory status would not have changed by one comma. The two facts are independent, they are being conflated constantly by people arguing in both directions, and keeping them separate is most of what this chapter is teaching.

📊 Evidence Rating

Claim form — "it's not FDA approved, which means it's being suppressed" Rating: ❌ Hype outpaces evidence Why: For most compounds in this book, no application was ever filed — the reason is the absence of a commercial sponsor willing to fund a development program, not the presence of an obstacle blocking one. What would change it: Documentation that an application was submitted for a specific compound and indication and was refused for reasons unrelated to the evidence, or evidence of an agency preventing a sponsor from filing. Note that a refusal on the merits would be case (2), not suppression.

🔍 Check Your Understanding

1. Assign one of the five cases to each: a compound with an empty regulatory file; a peptide registered as a medicine in one country and absent from another; a drug prescribed for a condition not on its label; a copper peptide in a face serum.
2. Which of the five is a genuinely negative signal about the molecule, and why are the other four not?
3. Explain why BPC-157's ❌ rating in Chapter 17 would be unchanged if it were approved tomorrow in some jurisdiction — and what would change it.

38.5 Off-label prescribing

Here is a structure that surprises people, and it is worth understanding because the surprise is usually the beginning of a misunderstanding.

In the United States and many other jurisdictions, a licensed prescriber may lawfully prescribe an approved drug for an unapproved indication, exercising clinical judgment about an individual patient. Manufacturers, by contrast, generally may not promote off-label use.

That is an asymmetry, and it is deliberate.

The reasoning runs like this. Medical practice moves faster than regulatory filings, and it should. Evidence accumulates in the literature long before — and sometimes instead of — a sponsor assembling it into an application. Whole fields depend on this: substantial portions of pediatric prescribing, of oncology, and of psychiatry involve uses that were never formally submitted, because the trials that support them were run by academic investigators rather than by companies seeking a label expansion. If prescribing were restricted to labeled indications, a great deal of well-supported medicine would become unavailable, and it would become unavailable for administrative reasons.

At the same time, the party that profits from a claim is the party with the strongest incentive to overstate it. So the burden is placed there: if a manufacturer wants to make a claim about what its drug does, it must submit the evidence and have it judged. The asymmetry preserves clinical freedom while keeping the evidentiary burden on the party that profits from the claim. That is a coherent design, whatever one thinks of how it plays out in practice, and the enforcement history around off-label promotion is one of the larger bodies of pharmaceutical litigation in existence.

Off-label is not a synonym for unsupported

This is the part that gets lost. "Off-label" describes the relationship between a use and a label. It describes nothing about the state of the evidence.

Some off-label uses rest on substantial published evidence — multiple randomized trials, guidelines from professional societies, decades of accumulated clinical experience — that no sponsor had commercial reason to submit for a label change. This happens most often with older drugs whose exclusivity has expired: there is no economic return on funding a label expansion for a molecule that anyone can now manufacture. The evidence exists; the filing does not.

Some off-label uses rest on very little — a mechanistic argument, a small open-label series, and enthusiasm. The label is silent on both, in exactly the same way.

So the rule to carry forward: the label tells you what was submitted and approved. It does not tell you what is known. To find out what is known, you do what Chapter 5 taught, which is to go look at the evidence for the specific claim — the population, the intervention, the comparison, the endpoint, the effect size, and the quality of the design.

This also cuts against a rhetorical move you will meet in clinics that market peptide services. "It's prescribed by a licensed physician" is offered as though it were a statement about evidence. It is not. It is a statement about who wrote the prescription. A prescription documents that a licensed professional exercised judgment; it does not document what that judgment was based on, and the range of what it can be based on is enormous.

⚠️ Hype Check — "a doctor prescribed it, so it must be proven"

The claim, in its usual form:

"This isn't some internet thing — it's prescribed by a board-certified physician at a licensed clinic. If it weren't legitimate, they couldn't prescribe it."

What's true in it. A prescriber is a trained professional with obligations to the patient, and a clinical relationship is genuinely better than no clinical relationship. Someone is taking a history, is in principle watching for adverse effects, and is accountable to a licensing board. That is not nothing, and this book has argued repeatedly (Chapter 39) that having a clinician involved is better than not having one.

Where it fails. The argument conflates three separate things: that a use is lawful, that a professional judged it appropriate for this patient, and that the use is supported by evidence. Off-label prescribing is lawful in many jurisdictions specifically so that clinical judgment can run ahead of filings — which necessarily means it can run ahead of evidence too. The permission is broad by design, and breadth is exactly what makes it a weak signal.

There is also a category confusion buried in the phrasing. Prescribing an approved drug off-label is case (4) of §38.4. Prescribing something that is not an approved drug anywhere — obtained through compounding (§38.6), or through channels this book does not discuss — is a different situation entirely, and it is frequently described to patients using the same reassuring sentence.

Verdict: as a statement about legality and about clinical involvement, it may well be accurate. As a statement about evidence, it is a non sequitur. The question "what is the evidence for this use, in people like me, for the outcome I care about?" is still sitting there unanswered, and it is a reasonable thing to ask the prescriber directly.

38.6 Compounding: 503A, 503B, and shortages

A compounding pharmacy prepares a medication for an individual patient — combining, mixing, or altering ingredients to produce something not commercially available in the needed form. The classic cases are genuinely necessary: a patient allergic to a dye in the commercial tablet, a child who needs a liquid version of a drug that is only made as a capsule, a formulation without a preservative someone reacts to.

Compounding is old, it is legitimate, and it is not manufacturing. That last distinction is the one this section is about.

503A and 503B

In the United States, the distinction that matters is between two sections of the federal food and drug law, and the shorthand names come from those section numbers.

503A — traditional pharmacy compounding. A licensed pharmacy compounds for an identified individual patient, on the basis of a patient-specific prescription. Products compounded under these conditions are exempt from certain requirements that apply to manufactured drugs — including premarket approval, and including certain manufacturing and labeling requirements — on the theory that this is the practice of pharmacy rather than the production of a drug product for the general market.

503B — outsourcing facilities. This category was created by federal legislation in 2013, after a fungal meningitis outbreak traced to contaminated compounded injections made clear that some "compounders" were operating at industrial scale with pharmacy-level oversight. An outsourcing facility registers with the federal agency, may compound without patient-specific prescriptions, may distribute to healthcare facilities, and is subject to current good manufacturing practice requirements and to federal inspection.

COMPOUNDING IN THE UNITED STATES        (structure as of this writing, 2026)

                        503A                    │  503B OUTSOURCING FACILITY
  ──────────────────────────────────────────────┼──────────────────────────────────
  scale                 individual patient      │  batch
  prescription          patient-specific        │  not required
  registration          state pharmacy board    │  registers federally
  GMP requirements      no                      │  yes
  federal inspection    limited                 │  yes
  ──────────────────────────────────────────────┼──────────────────────────────────
  PREMARKET REVIEW OF                           │
  SAFETY OR EFFICACY    NO                      │  NO
  ──────────────────────────────────────────────┴──────────────────────────────────

  The bottom row is the one people miss. It is identical on both sides.
  A compounded drug is not an FDA-approved product, in either category.

That bottom row is the teaching point. Compounded drugs are not FDA-approved products, and neither category involves premarket review of safety or efficacy. 503B brings real manufacturing oversight — sterility, quality systems, inspection — which is a substantial improvement over none. It does not bring a regulator's judgment that the product works or that its benefits outweigh its risks for a stated indication. Those are different questions answered by different processes.

The shortage mechanism

There is a specific circumstance that makes compounding of something resembling an approved drug permissible, and it is the mechanism behind the largest peptide story of the mid-2020s.

Ordinarily, compounding a product that is essentially a copy of a commercially available approved drug is restricted — the point of compounding is to serve needs the approved product cannot, not to produce a parallel supply of it. But when a drug appears on the official shortage list, compounding of what would otherwise be an essentially-a-copy product is permitted in defined circumstances, on the reasoning that a patient who cannot obtain the approved drug is worse served by a rule that also blocks the compounded version.

And when the shortage is declared resolved, that permission narrows, and compounders must wind down. The permission was contingent on the shortage; removing the shortage removes the basis.

This is exactly what happened with the GLP-1 drugs. During the mid-2020s, semaglutide and tirzepatide were in shortage in the United States, and a very large compounded market grew up in that window — telehealth prescribing, 503A and 503B supply, and a great deal of consumer marketing. The shortages were subsequently declared resolved during that period, the permission narrowed, and the compounded market faced restriction. The details and the timing were contested, including in litigation, and this book is not the place to adjudicate them. What matters for your reading is the structure: a temporary legal permission created a market, and the removal of the permission removed the market's basis.

The teaching point

A compounded version of a drug is not the approved drug. It may contain the same active molecule. It may not. And even where it does, Chapter 19's entire argument about quality applies with full force:

• Different salt or ester forms. A compounded product may use a salt form of the active molecule that was never the subject of the approval, with different solubility and potentially different pharmacokinetics.
• Different concentrations. Approved products come at defined concentrations in defined devices. A compounded vial may be at a different concentration, and concentration errors are a documented source of harm.
• Different delivery devices. An approved pen is a metering device that was designed, tested, and reviewed alongside the drug. A vial and a syringe is a different system with different failure modes.
• Different additives. Compounded formulations sometimes include additional ingredients that were not part of any approved product and were not evaluated with it.

None of that makes every compounded product bad. It makes every compounded product a different product from the approved one, which means the clinical evidence for the approved product does not transfer to it automatically — and it is transferred automatically, constantly, in marketing.

🩺 Safety and Risk — what changes and what doesn't when a drug is compounded

What may be preserved. If a competent facility uses genuine active pharmaceutical ingredient at the intended concentration in a sterile preparation, the pharmacology of the molecule is the pharmacology of the molecule. Receptors do not know who filled the vial.

What is not preserved, automatically:

Identity and purity. The approved product's supply chain, specifications, and testing regime were reviewed as part of the application. A compounded product's are governed by whatever requirements apply to that facility — substantial under 503B, more limited under 503A, and in the gray market outside both, nonexistent (Chapters 19 and 34).

Concentration. This is the documented failure mode that has caused real harm with compounded GLP-1 products: a person accustomed to thinking in units or in fractions of a pen encounters a vial at a different concentration, and the arithmetic goes wrong. Dosing errors of this kind have generated poison-center reports.

The device. Injection devices are engineered products. Substituting a syringe for a metered pen transfers the metering job to a human being, at home, sometimes at night.

The label. An approved label is a reviewed document listing contraindications, interactions, and warnings. Whatever accompanies a compounded product is not that document and has not been through that review.

As always, decisions here belong with a clinician who knows your history — and, on the legal questions, with someone qualified in your own jurisdiction. This book describes the categories; it does not tell you what you may obtain.

38.7 The "research chemical" fiction and the supplement question

Two classification claims dominate the gray market, and both are doing work that they cannot actually do.

"For research use only — not for human consumption"

Chapter 6 named this fiction and Chapter 19 dissected it. The regulatory framing is quick.

Genuine research reagents exist. Laboratories need compounds, those compounds are sold, and they are not drugs because they are not offered for administration to human beings. The category is real and uncontroversial.

What is not real is the use of that label as a shield on a product that is quite obviously being sold for human use — priced by the milligram in injectable-friendly vial sizes, marketed alongside reconstitution accessories, discussed in customer forums entirely in terms of human effects, and accompanied by a disclaimer nobody involved believes.

The disclaimer is a liability posture, not a regulatory category. It is a sentence written by a seller for the seller's benefit. It does not change what is in the vial, does not change what happens when it enters a person, and does not confer any of the assurances that a regulatory category would confer. The seller's disclaimer transfers risk without changing what the product is.

Notice also what it transfers risk away from. A product sold as a research reagent is not accompanied by manufacturing standards for human use, by a reviewed label, by adverse-event reporting obligations, or by any mechanism through which a purchaser can find out what actually happened to other purchasers. Every one of those absences is a benefit to the seller and a cost to the buyer. Chapter 34's testing data is the empirical follow-up: when independent analyses have looked at what is actually in gray-market vials, the answers have included wrong identity, wrong quantity, and contaminants.

The supplement question

The other classification claim is "it's a supplement," and it requires a bit of structure.

In the United States, the Dietary Supplement Health and Education Act of 1994 (DSHEA) created a regulatory category for dietary supplements with no premarket approval requirement. A supplement does not have to be shown safe and effective before it is sold. Enforcement is largely after the fact, and the category is enormous.

But — and this is the part that consumer marketing skips — DSHEA did not make the category unlimited. A substance must actually qualify as a dietary ingredient to be sold in it. The definition covers things like vitamins, minerals, herbs and other botanicals, amino acids, and certain substances used in the food supply, plus a pathway for new dietary ingredients that requires notification. A synthetic peptide that has never been part of the food supply and does not fit the statutory categories is not automatically a dietary ingredient because someone put it in a capsule.

Regulators have taken exactly that position for certain peptides. BPC-157 among them: as of this writing in 2026, the United States regulator has taken the position that it does not qualify as a dietary ingredient, and has therefore treated products marketed as supplements containing it as outside the category. Separately, compounds including BPC-157 have been placed on lists identifying bulk substances that present significant safety risks for use in compounding — a different mechanism, aimed at a different channel, pointing the same direction.

So, plainly: "sold as a supplement" is not a regulatory endorsement, and for several compounds in this book it is not even a lawful classification. The first half of that sentence is true of the entire supplement category — no premarket approval means no premarket approval, for fish oil and for everything else. The second half is specific and more serious: some products are in a category they do not qualify for, which means the classification claim on the label is not merely uninformative but incorrect.

⚠️ Hype Check — "it's just a supplement" and "it's just a research chemical"

The claims, in their usual forms:

"Relax, it's a supplement — it's not some pharmaceutical."

"It's sold for research use only. That's a completely legitimate, regulated category."

What's true in them. Both categories are real. Dietary supplements are a lawful category with a genuine statutory basis, and research reagents are a real thing that laboratories genuinely buy. Neither phrase is a fabrication.

Where they fail. Each phrase is being used to imply a form of oversight that the category does not provide, and in some cases to claim membership in a category the product does not belong to.

"It's a supplement" implies review. There is no premarket review of supplements in the United States — that is the defining feature of the category, not a loophole in it. And for several peptides, regulators have said the products do not qualify as supplements at all.

"Research use only" implies a regulated pathway. It is a disclaimer, written by the seller. It is not a category that any agency conferred, and it does not survive contact with a product that is obviously configured for human use.

Notice the shared structure: both phrases substitute a classification claim for an evidence claim. Neither one says anything at all about whether the compound does what the seller says. That question is untouched, and it is the only one that matters.

Verdict: both are category claims dressed as reassurance. One of them is frequently also inaccurate as a category claim.

38.8 Doping: WADA, S0, and S2

If you compete in tested sport, this section is the one with immediate consequences — and if you do not, it is still the cleanest illustration in the chapter of regulatory status and evidence being completely independent variables.

The World Anti-Doping Agency (WADA) publishes a Prohibited List, updated annually, that national anti-doping organizations and international federations adopt. Substances and methods are grouped into categories, and two of them cover most of this book.

S2 — Peptide Hormones, Growth Factors, Related Substances and Mimetics. This is the category built for exactly the molecules in Part II. It covers growth hormone and its releasing factors — including GH secretagogues and GHRH analogs (Chapters 14 and 15) — insulin-like growth factor 1 and its analogs (Chapter 16), erythropoietin and agents affecting erythropoiesis, and a range of related compounds and mimetics. If you have been reading Part II and wondering where the sporting authorities stand on it, they stand here, and they have for a long time.

S0 — Non-Approved Substances. This is the catch-all, and it is the one most athletes misunderstand. In substance: any pharmacological substance that is not currently approved by any governmental regulatory health authority for human therapeutic use is prohibited at all times. That covers drugs in preclinical or clinical development, discontinued drugs, designer compounds, and substances approved only for veterinary use.

BPC-157 was added to the Prohibited List under S0, effective from the 2022 List.

Why S0 breaks the most common argument

Athletes reason about the Prohibited List as though it were an enumeration: a roster of named banned substances, and if your compound is not on the roster, you are fine. That model is wrong, and S0 is why.

S0 is not a list of substances. It is a rule about a property. The property is "not approved by any governmental regulatory health authority for human therapeutic use," and it applies whether or not anyone at WADA has ever heard of your compound. A novel peptide synthesized last month and sold on a website is prohibited under S0 the moment it exists, because it satisfies the description.

Which produces the sharp point of this section: "it isn't on the banned list" is usually wrong by construction. For a tested athlete, an unapproved experimental compound is prohibited precisely because it is unapproved. The very feature that gray-market marketing presents as an advantage — that it is outside the pharmaceutical system, unencumbered, not yet caught up in regulation — is the feature that places it squarely inside S0.

There is a second-order irony worth stating. Under S0's logic, a compound moving toward approval becomes less prohibited, not more. The route out of S0 is approval by a governmental health authority for human therapeutic use — after which the substance may or may not be prohibited under some other category on its own merits, but S0 no longer reaches it.

The sanction does not depend on whether it works

This is the part that connects to the rest of the book, and it is worth being explicit about.

An anti-doping violation is established by the presence of a prohibited substance or its metabolites, or by use or attempted use, or by a handful of other defined violations. None of those definitions contains a clause about efficacy. An athlete who takes an unapproved compound that does absolutely nothing has committed the same violation as one who takes a compound that works.

So consider BPC-157 from two directions at once. Chapter 17 rates the healing claims ❌ — hype outpaces the evidence, because the human literature is essentially empty. And it is prohibited in tested sport under S0, effective from the 2022 List. Both statements are true simultaneously, and they are entirely independent. The ❌ is a judgment about published human evidence. The prohibition is a judgment about regulatory status. Neither one caused the other, and neither one would move if the other changed.

That independence is this chapter's whole thesis, stated in the crispest available case. If you can hold those two facts about one molecule at the same time without either one pulling on the other, you have the skill.

Scope, and who this binds

These rules bind athletes in tested sport — those subject to an anti-doping code through a federation, a national organization, an event, or an employer that has adopted one. Many readers of this book are not subject to them at all, and for those readers the Prohibited List is informational rather than binding.

For those who are subject to them: the List changes annually, therapeutic use exemption processes exist and have their own requirements and deadlines, and the rules are administered by organizations with their own procedures. This book is a poor substitute for the current List and for the compliance resources of the relevant organization. Check the current documents; do not rely on a textbook chapter, including this one.

📊 Evidence Rating

Claim form — "it isn't on the banned list" Rating: ❌ Hype outpaces evidence Why: WADA's S0 category prohibits any pharmacological substance not currently approved by a governmental health authority for human therapeutic use, so unapproved compounds are prohibited by default rather than by enumeration — being absent from a named list is not evidence of being permitted. What would change it: For a specific compound, approval by a governmental regulatory health authority for human therapeutic use, removing it from S0's scope — after which its status would depend on the other categories on their own terms. Nothing about the compound's efficacy is relevant either way.

38.9 The international patchwork

Everything above has been mostly United States framing, because that is the framing peptide marketing in English most often invokes. Widen the lens and the picture gets more complicated in a way that is easy to misread.

Different agencies — the FDA in the United States, the EMA for the European Union, the MHRA in the United Kingdom, the PMDA in Japan, the TGA in Australia, and national authorities across the rest of the world — apply broadly similar principles. They all ask whether the evidence shows benefits outweighing risks for a defined use. They read each other's assessments, participate in international harmonization efforts, and often converge.

And they routinely reach different conclusions. Understanding why is the point of this section, because "approved in [country]" is used as an argument constantly and it means several different things.

Genuinely different evidence standards. Registration requirements are not uniform. Some national systems require a full dossier equivalent to what a major agency demands; some accept approval by a reference agency as a basis for local registration; some have historically operated with substantially lighter requirements, particularly for products with long local use. A registration in a system with lighter requirements is a real registration and a weaker signal about the underlying evidence than a reader typically assumes.

Different filings by sponsors. This is probably the most common reason and the least often considered. A company decides which markets to file in based on market size, pricing environment, regulatory cost, patent position, and the availability of a local partner. A molecule can be absent from a major market because nobody thought the return justified the filing. That is a business decision, and it leaves no trace that looks any different from a scientific rejection.

Different medical practice and different unmet need. Agencies weigh benefit against risk in the context of what else is available. A drug with a modest benefit and real risks may be approvable where alternatives are poor and unapprovable where they are good. Neither agency is wrong; they are answering slightly different questions.

Different timing. Approvals do not happen simultaneously. A molecule approved in one region and not another may simply be earlier in a queue. This resolves, often quietly, and a claim built on the gap goes stale.

On import and possession

Rules governing personal import and possession of medicines differ enormously between countries and within them, and this book does not advise on them. They involve customs law, controlled-substances schedules, prescription requirements, quantity thresholds, and enforcement practice, and they change. If you have a question in this area, it is a question for someone qualified in your own jurisdiction, not for a textbook. This chapter describes how approval systems are structured. It does not tell you what you may bring across a border, and no sentence in it should be read that way.

The honest summary

"Approved in [country]" is a fact worth knowing and a poor substitute for looking at the evidence that approval rested on.

It is worth knowing because it tells you that somebody's regulator saw a dossier and said yes. That is more than an empty file. But it does not tell you how large the dossier was, what standard it was judged against, whether the endpoint was clinical or surrogate, how many patients were studied, or whether the approval is decades old and rests on evidence nobody would accept for a new filing today. The evidence behind an approval may have been strong, weak, or — in a market with a lighter registration standard — considerably lighter than a reader assumes.

The move is the same one this book always makes: take the regulatory fact as a pointer, then go read what it points at.

📊 Evidence Rating

Claim form — "it's approved in [another country], so the evidence is there" Rating: ❌ Hype outpaces evidence, as usually stated Why: Approval elsewhere may reflect a different evidence standard, a different benefit-risk context, a lighter registration route, or simply a sponsor's filing decision — it locates a dossier without telling you what is in it. What would change it: For a specific compound and indication, the underlying assessment: which trials, what design, what population, what endpoint, judged against what standard. A foreign approval resting on adequate randomized trials with clinical endpoints is strong evidence — because of the trials, not because of the approval.

38.10 Why regulatory status is not evidence, in both directions

This is the chapter's spine, and everything above has been assembling the parts.

Two errors are common. They come from opposite political and commercial directions, they are made by people who despise each other, and they are the same error. This chapter refuses both.

THE TWO ERRORS

   ERROR 1 — approval as proof                ERROR 2 — non-approval as proof
   ───────────────────────────                ────────────────────────────────
   "It's approved, so it works."              "It's not approved, so it doesn't work."
   "It's approved, so it's safe."             "It's not approved, so it's suppressed."

   Made by: marketers of approved drugs,      Made by: sellers of unapproved compounds
   and by patients reading a label as         (as a conspiracy claim), and by skeptics
   a guarantee.                               (as a refutation).

   Both errors treat a REGULATORY FILE as though it were a BODY OF EVIDENCE.

   ┌──────────────────┬────────────────────────┬────────────────────────────┐
   │                  │  APPROVED SOMEWHERE    │  APPROVED NOWHERE          │
   ├──────────────────┼────────────────────────┼────────────────────────────┤
   │ WELL EVIDENCED   │  the common case       │  possible — evidence       │
   │                  │  (semaglutide, insulin)│  exists, nobody filed      │
   ├──────────────────┼────────────────────────┼────────────────────────────┤
   │ THINLY EVIDENCED │  possible — old        │  the common case for       │
   │                  │  approvals, surrogate  │  most of Part III          │
   │                  │  endpoints, small      │                            │
   │                  │  trials                │                            │
   └──────────────────┴────────────────────────┴────────────────────────────┘

   All four cells are occupied. That is the whole argument.

The first error: treating approval as proof

Approval means an application was submitted and judged favorable for a specific indication, on the evidence in the file. §38.1 listed what it does not mean, and this is where those negatives cash out.

Approval does not mean the drug works for other uses. The approval is bounded by the indication. An approved GLP-1 receptor agonist has been judged for the indications on its label; every other proposed benefit is a separate question with its own separate evidence, and some of those questions have good answers while others have almost none.

Approval does not mean it works for everyone. Trials produce distributions. Approvals are granted on averages that clear a threshold.

Approval does not mean the long-term effects are known. They are known to the extent they were studied, over the horizon that was studied, in the population that was studied.

Approval does not mean it is better than the alternatives. Frequently nobody ran that trial.

And the sharpest case: approval does not guarantee that later, larger, better-designed evidence will agree.

Nesiritide (Chapter 28) is the example this book keeps for exactly this purpose. It is a recombinant form of B-type natriuretic peptide, approved in the United States in 2001 for acutely decompensated heart failure, on the basis of an improvement in a symptom measure — dyspnea — against placebo. It was used widely. Concerns accumulated. And a large randomized outcome trial, published in 2011 and enrolling roughly seven thousand patients, found no benefit on the endpoints that mattered most: it did not reduce death or rehospitalization, and the effect on dyspnea was small and did not meet the prespecified threshold for significance.

That is not a scandal. It is the system working slowly. But it is a decisive demonstration that an approval is not a permanent guarantee, and that a drug can be approved, widely used, and then shown by better evidence to be doing much less than everyone thought.

Add to that the structural point from §38.3: accelerated approvals rest on surrogate endpoints by design. Chapter 16 explained why a surrogate can move while the outcome does not. An approval on a surrogate is a bet that the surrogate predicts benefit, taken deliberately, with the confirmation deferred. Some of those bets lose.

The correct formulation: approval is strong evidence about a narrow claim. Narrow, because it is bounded by indication, population, and endpoint. Strong, because a regulator with access to the full dataset — including unpublished analyses that never reach the literature — read it and was persuaded. Both words matter. Drop "strong" and you become the cynic who thinks approvals are meaningless. Drop "narrow" and you become the person who thinks a label is a certificate.

🔬 Read the Study — nesiritide, before and after

The approval (United States, 2001). Nesiritide was approved for acutely decompensated heart failure. The supporting evidence centered on improvement in dyspnea — how short of breath patients reported being — compared with placebo. That is a symptom endpoint, it is meaningful to patients, and it is not the same as living longer or staying out of the hospital.

The concern (mid-2000s). Pooled analyses raised questions about renal function and about mortality. These were analyses of existing trials rather than new trials, with the limitations that implies (Chapter 5), and they did not settle the question. They did make it urgent.

The outcome trial (published 2011). A large randomized, double-blind, placebo-controlled trial enrolling roughly seven thousand patients with acute decompensated heart failure. Results: no significant reduction in death or rehospitalization at 30 days; a small change in dyspnea that did not meet the prespecified significance threshold; no evidence of the feared worsening of renal function; more hypotension in the treated group.

What to take from it. Three things. First, the reassuring finding and the disappointing finding arrived together — the trial exonerated the drug on the safety concern and deflated it on efficacy, which is what an adequately powered trial is for. Second, the approval was not fraudulent or negligent; it was granted on a symptom endpoint that turned out not to carry the weight people loaded onto it. Third, it took a decade and thousands of patients to find this out, which is the honest reason regulators approve on surrogates only under defined circumstances and with confirmation attached.

The second error: treating non-approval as proof of absence

This one is used dishonestly in both directions, which is what makes it interesting.

Sellers present non-approval as suppression. The story is that the compound works too well, or is too cheap, or threatens someone's revenue, and that this is why it has not been approved. The story is emotionally satisfying and, for the compounds in this book, essentially always false — because §38.4 case (1) applies. There is no rejection. There is no adverse finding. There is an empty file.

Skeptics present non-approval as refutation. The story is that if it worked, it would be approved, therefore it does not work. This has the same defect from the other side. An empty file is an empty file. It contains no verdict, because nobody asked for one.

The accurate statement, for most gray-market peptides, is a chain — and the chain is worth reading slowly because each link is a fact about economics rather than biology:

No one submitted an application. No one submitted because no one funded the trials. No one funded the trials because there is no patentable commercial position in a well-known short sequence.

A sequence published decades ago, cheap to synthesize (Chapter 32), with no composition-of-matter protection available, generates no way for a sponsor to recover the cost of a development program. A competitor could sell the identical molecule the day after approval. So the program does not happen, and the compound sits in permanent evidentiary limbo — not refuted, not supported, just unexamined.

This is a structural fact about drug economics, not a fact about biology. It is genuinely a flaw in how medical evidence gets funded, and it is fair to say so. What it is not is evidence that any particular unexamined compound works. The economics predict that good and bad compounds alike go unstudied. An empty file is exactly as consistent with an inert molecule as with a useful one, which is precisely why it carries no information.

The conclusion the whole book has been building toward

State it once, cleanly:

Regulatory status tells you what has been submitted and judged. Evidence tells you what is known. They correlate, and they are not the same thing.

They correlate because a well-supported compound with a commercial sponsor usually does get filed and usually does get approved, and because a filed application is scrutinized more intensely than almost any published paper. So approval is genuine information.

They diverge because filing is a business decision, because agencies differ, because categories like "supplement" and "cosmetic ingredient" and "research reagent" route products around the drug system entirely, and because the funding of clinical evidence is organized around ownership rather than around what would be useful to know.

So when you meet a regulatory fact, do the thing this book has been training you to do since Chapter 5: treat it as a pointer, not a verdict. Ask which of §38.4's five cases applies. Ask what jurisdiction and what date. Ask what indication and what population. Then go and read the evidence, because Chapter 5's method is the one you apply — not the label.

📊 Evidence Rating

Claim form — "it's FDA approved, so it's safe and effective" Rating: ❌ Hype outpaces evidence, as usually stated Why: Approval is indication-specific and population-specific, rests on the evidence a sponsor submitted, and does not certify a molecule — it certifies a benefit-risk judgment about a defined use, which is a narrower and more perishable object than the claim implies. What would change it: A restatement that names the indication, the population, and the endpoint — "approved for chronic weight management in adults with obesity, on the basis of randomized trials showing X" — which is not a rating change so much as the claim becoming a different, defensible claim. Nesiritide's 2011 outcome trial is the standing demonstration that the broad version fails.

🔍 Check Your Understanding

1. Fill in the four cells of the grid above with a compound or claim from this book, one per cell. Which cell was hardest, and what does that difficulty tell you?
2. A seller says non-approval proves suppression; a skeptic says non-approval proves the compound is worthless. Name the single shared error and state the accurate alternative.
3. Nesiritide was approved in 2001 and a large trial in 2011 found no benefit on the endpoints that mattered. Does that make the 2001 approval a mistake? Defend your answer using §38.1's definition of what an approval is.

📋 Your Evidence Dossier

This chapter completes Field 10 — status: regulatory and sport-doping. It is the last piece of a field that has been assembled slowly, and the assembly is worth naming, because it explains something about the question itself.

Field 10 has taken five chapters to fill, because "what is the status of this compound?" turns out to be five different questions.

FIELD 10 — HOW IT WAS BUILT

  Ch. 6   What does the sales channel claim about itself?     → the "research
          chemical" fiction, named
  Ch. 12  Who can actually get it, at what cost, through       → access, cost,
          what supply, and who is selling it?                    supply, sellers
  Ch. 19  What is actually in the vial?                        → the quality-risk
                                                                 line
  Ch. 34  What can a certificate of analysis establish          → and what it
          about that — and what can it not?                      cannot
  Ch. 38  What is its status with regulators, and with          → THIS CHAPTER
          sporting authorities?                                  closes the field

Five questions, five chapters, one field — and notice that no two of them can substitute for each other. A compound can have an immaculate certificate of analysis and no regulatory standing anywhere. It can be widely sold and prohibited in tested sport. It can be approved in one jurisdiction and sold in another under a classification it does not qualify for. "Status" was never one fact, and a dossier that recorded it as one fact would have been hiding four questions inside a single word.

What to record

Field 10's final two lines are this chapter's contribution, and each has a required companion detail:

FIELD 10 — STATUS (completing lines)          (record the date you filled this in)

  APPROVED ANYWHERE?     Which jurisdictions, and FOR WHAT INDICATION.
                         "Approved" without an indication is not an answer (§38.1).

  IF NOT APPROVED HERE,  Which of §38.4's five cases applies, in which jurisdiction:
  WHICH OF THE FIVE?       (1) never submitted   (2) submitted and rejected
                           (3) approved elsewhere, not here
                           (4) approved, but not for this use
                           (5) not a drug at all in the regulatory sense
                         If you cannot tell which, write "cannot determine" — that is
                         a finding, not a failure.

  PROHIBITED IN TESTED   WADA category if any (S0, S2, other), and from which List
  SPORT?                 year. The List changes annually (§38.8), so an undated
                         entry here goes stale without telling you.

  HOW IS IT ACTUALLY     Prescription product / compounded / marketed as a supplement /
  BEING SOLD?            cosmetic ingredient / "research use only" / not sold.
                         This is often different from the category it legally belongs
                         in, and the gap is the interesting part (Ch. 6, Ch. 12).

Worked demonstration — two performance peptides

Both of these are Part III compounds, both are unapproved, and both are prohibited in tested sport. Those two facts have nothing to do with each other, which is exactly why these entries are the right ones to work through.

FIELD 10 — BPC-157                       [worked demonstration, completed 2026]
  Approved anywhere      No approval as a drug in any major jurisdiction that I have
                         been able to verify.
  Which of the five      (1) — never submitted. No rejection on file. Empty file.
                         → the reason is the absence of a sponsor, not the presence
                           of an obstacle (§38.10).
  Prohibited in tested   Yes — WADA Prohibited List, category S0 (Non-Approved
  sport                  Substances), effective from the 2022 List. Prohibited
                         BECAUSE it is unapproved, not because it was judged to
                         work (§38.8). Check the current List before relying on this.
  How it's actually      Gray market, largely under "research use only" labeling
  being sold             (Ch. 6, Ch. 19); also marketed as a supplement, a
                         classification US regulators have said it does not qualify
                         for (§38.7). Certificate-of-analysis caveats: Ch. 34.
  ────────────────────── SEPARATE COLUMN ──────────────────────────────────────
  EVIDENCE RATING        ❌ for the healing claims (Ch. 17). Rests on the empty
                         human literature — NOT on any of the four lines above.

FIELD 10 — TB-500 / thymosin beta-4      [worked demonstration, completed 2026]
  Approved anywhere      No approval as a drug for the marketed uses in any major
                         jurisdiction that I have been able to verify.
  Which of the five      (1) — never submitted for these uses. Note the wrinkle:
                         related thymosin beta-4 material has been through early
                         clinical work for other indications, which is a fact about
                         a development program, not an approval.
  Prohibited in tested   Yes — treated as a non-approved substance under S0. Check
  sport                  the current List for exactly how it is named, and from
                         which year, before recording anything here as settled.
  How it's actually      Gray market, "research use only" labeling; sold as a
  being sold             fragment/peptide whose identity relative to the parent
                         molecule is often unclear on the label (Ch. 1, Ch. 34).
  ────────────────────── SEPARATE COLUMN ──────────────────────────────────────
  EVIDENCE RATING        Recorded from the human literature for the specific claim
                         being made — see the Part III chapter for the compound.
                         Whatever it is, nothing above produced it.

For contrast, run the same four lines on an approved compound — semaglutide, say — and watch what changes. Approved in multiple jurisdictions, for named indications, each resting on its own application; not swept up by S0 precisely because it is approved; sold as a prescription product, and for a period also as a compounded product under the shortage mechanism (§38.6). Four lines, all different, and the evidence column still has to be filled in separately from the trial literature. The form of the field is identical whether the answer is yes or no. That is the point of having a form.

The discipline

Record Field 10 in a column entirely separate from the evidence rating, and never let one adjust the other.

This is not a formatting preference. It is the operational version of §38.10, and the physical separation is what makes the discipline enforceable. When the two live in the same cell, they blend. When they live in separate columns, you can look across the row and notice the mismatches — and the mismatches are where the learning is.

Two combinations to watch for specifically:

Approved somewhere and thinly evidenced. This happens. An old approval on a small trial, an approval on a surrogate endpoint with confirmation still pending, a registration in a market with a lighter standard. The regulatory column says yes; the evidence column should say what the evidence actually supports, which may be much less.

Unapproved everywhere and simply untested. This is the Part III default. The regulatory column says no; the evidence column should say ❌ or, where the work is genuinely early and being done properly, 🔬. Neither of those ratings is derived from the regulatory status. They are derived from the literature, and they would be the same if the regulatory status were different.

If you find yourself softening an evidence rating because a compound turned out to be approved somewhere, or hardening one because a compound turned out to be prohibited in sport, stop and write down what actually moved you. That sentence is the most valuable thing this project will produce, and Chapter 40 will ask you for it.

Conclusion

Regulation is not a scoreboard. It is a filing system with legal consequences, and reading it as a verdict on whether molecules work is the single most common error in the public conversation about peptides.

An approval is a regulator's judgment that, for a specified indication and population, the evidence submitted showed benefits outweighing risks — indication-specific, submission-dependent, jurisdiction-bound, and revisable. In the United States, a definitional line at forty amino acids sorts these molecules into drugs and biologics, and that line decides whether follow-on competition arrives as generics or as biosimilars: insulin at 51 residues was deemed a biological product under a transition effective in March 2020, which opened the biosimilar pathway for it, while semaglutide's 31-residue backbone keeps it on the drug side. The development pathway runs from preclinical work through an IND to Phases 1, 2, and 3, then review, then surveillance — with attrition concentrated exactly where human efficacy is first tested, and with accelerated pathways trading surrogate endpoints for speed by design.

"Not FDA approved" means five different things, and only one of them is a negative finding about the molecule. Off-label prescribing is lawful where off-label promotion is not, deliberately, and the label tells you what was submitted rather than what is known. Compounded drugs are not approved drugs in either the 503A or the 503B category, because neither involves premarket review of safety or efficacy. "Research use only" is a liability posture rather than a regulatory category, and "sold as a supplement" is not an endorsement and, for several compounds here, not even a lawful classification. In tested sport, S0 prohibits unapproved substances by default, which makes "it isn't on the banned list" wrong by construction — and the sanction has nothing to do with whether the compound works.

And none of that is evidence.

That is the sentence to carry out of this chapter. Regulatory status tells you what has been submitted and judged. Evidence tells you what is known. The two correlate because well-supported drugs with sponsors usually get filed — and they diverge because filing is a business decision, agencies differ, whole categories route around the drug system, and clinical evidence gets funded on the basis of ownership rather than usefulness. Nesiritide was approved and later shown to do much less than believed. BPC-157 is prohibited in sport and rated ❌ for entirely unrelated reasons. Both facts sit comfortably in the same book because they were never the same kind of fact.

One last time: nothing in this chapter is legal advice, the rules described here differ by country and change, and no passage in it tells you what you may lawfully buy, import, or possess. If that is your question, ask it of the current rules in your own jurisdiction and of someone qualified to read them.

Chapter 39 turns to the conversation this book cannot have for you: how to bring any of this to a clinician in a way that makes the appointment useful.

Key Terms

Approval — a regulator's judgment that, for a specified indication and population, the evidence submitted shows that benefits outweigh risks, together with an approved label describing that use. Indication-specific, submission-dependent, and jurisdiction-specific.

Indication — the specific condition and population a drug is approved to treat. An approval without an indication is an incomplete statement.

Label — the reviewed document accompanying an approved drug, describing the approved use, the population studied, dosing evaluated, adverse events observed, and required warnings.

Biological product — in United States law, a category including alpha-amino-acid polymers with a specific, defined sequence of more than 40 amino acids. Licensed through a Biologics License Application.

New Drug Application (NDA) — the United States application type for a drug, including peptides of 40 or fewer amino acids.

Biologics License Application (BLA) — the United States application type for a biological product.

Generic — a follow-on version of an approved drug, approved through an abbreviated application resting on demonstrated bioequivalence rather than repeated clinical trials.

Biosimilar — a follow-on version of a biological product, approved through a pathway requiring extensive analytical and usually clinical comparability. Substantially more demanding and expensive than a generic filing.

Interchangeability — an additional determination, beyond biosimilarity, relevant to whether a product may be substituted at the pharmacy level without prescriber involvement.

Investigational New Drug application (IND) — the filing that permits administration of an investigational compound to human beings. A permission to test, not an approval of anything.

Phase 1 — first-in-human studies of safety, tolerability, and pharmacokinetics; usually small and often in healthy volunteers.

Phase 2 — dose-finding and preliminary efficacy studies in people with the target condition.

Phase 3 — adequate and well-controlled efficacy trials in the target population, powered for a prespecified endpoint.

Phase 4 / post-marketing surveillance — study and monitoring after approval, including any confirmatory trials required as a condition of approval.

Accelerated approval — a pathway permitting approval on a surrogate endpoint reasonably likely to predict clinical benefit, with confirmatory trials required afterward. Confirmation is deferred, not waived.

Surrogate endpoint — a measured quantity used as a stand-in for the outcome that actually matters to patients. See Chapter 16 for why the substitution can fail.

Off-label prescribing — a licensed prescriber's use of an approved drug for an unapproved indication. Lawful in the United States and many other jurisdictions; says nothing about the state of the evidence.

Off-label promotion — a manufacturer's marketing of a drug for an unapproved use. Generally prohibited; the asymmetry with prescribing is deliberate.

Compounding — preparation of a medication for an individual patient by combining, mixing, or altering ingredients. Not manufacturing, and not subject to premarket approval.

503A — traditional pharmacy compounding in the United States, on patient-specific prescriptions, exempt from certain requirements including premarket approval.

503B outsourcing facility — a United States category created in 2013 for larger-scale compounders, which may produce without patient-specific prescriptions and are subject to good manufacturing practice requirements and federal inspection. Still no premarket review of safety or efficacy.

Drug shortage list — an official listing of drugs in shortage. Listing can permit compounding of products that would otherwise be restricted as essentially copies; delisting narrows that permission.

Essentially a copy — a compounded product that duplicates a commercially available approved drug. Generally restricted outside defined circumstances such as a shortage.

Research use only — a seller's disclaimer stating a product is not for human consumption. A liability posture, not a regulatory category conferred by any agency.

Dietary supplement — a United States product category created by DSHEA (1994) with no premarket approval requirement. Membership requires that the substance qualify as a dietary ingredient.

DSHEA — the Dietary Supplement Health and Education Act of 1994, which established the United States supplement category and its requirements.

Dietary ingredient — the statutory categories a substance must fall into to be lawfully sold as a dietary supplement. Regulators have taken the position that certain synthetic peptides do not qualify.

World Anti-Doping Agency (WADA) — the organization that publishes the Prohibited List adopted by anti-doping organizations and sporting federations.

Prohibited List — WADA's annually updated list of prohibited substances and methods, organized into categories.

S0 — Non-Approved Substances — the WADA category prohibiting, at all times, any pharmacological substance not currently approved by any governmental regulatory health authority for human therapeutic use. A rule about a property, not a list of names.

S2 — Peptide Hormones, Growth Factors, Related Substances and Mimetics — the WADA category covering growth hormone and its releasing factors, IGF-1 and analogs, erythropoiesis-affecting agents, and related compounds.

Therapeutic use exemption — a process by which an athlete may be permitted to use an otherwise prohibited substance for a documented medical need, subject to the relevant organization's requirements.

Marketing authorization — the term used in the European Union and elsewhere for what United States practice calls an approval.

Spaced Review

1. (Ch. 38 + Ch. 17) BPC-157 is prohibited in tested sport under WADA's S0 category and rated ❌ for its healing claims. State, in two sentences, what each of those facts is a judgment about — and explain why a well-designed positive human trial would change one of them and not the other.

2. (Ch. 38 + Ch. 19) A telehealth service offers a compounded version of an approved peptide drug. Using §38.6 and Chapter 19, list four specific ways the compounded product may differ from the approved one, and explain why the approved product's clinical evidence does not automatically transfer.

3. (Ch. 38 + Ch. 5) A seller's page says a compound is "approved in several countries and used safely for decades." Write the three questions you would ask before that sentence changed your assessment at all — one about jurisdiction, one about indication, and one about the underlying evidence.

4. (Ch. 38) Assign one of the five meanings of "not FDA approved" to each of the following, and say which one you cannot assign without more information: a peptide with no filing history anywhere; thymosin alpha-1; a copper peptide in a cosmetic serum; an approved drug used for a condition absent from its label; a drug whose application was withdrawn during review.

5. (Ch. 38 + Ch. 5 + Ch. 19) Someone argues: "The reason there are no trials of these peptides is that the pharmaceutical industry doesn't want them to exist." Steelman that argument as far as the evidence honestly allows, then state precisely where it stops being supportable — and say what the accurate version of the claim is.