What the Empower Warning Letter Means for Prescribers

Most coverage of compounded GLP-1s discusses a single question: is the preparation "essentially a copy" of a commercially available drug? That is one of two independent restrictions, and satisfying it does nothing for the other.

The copy restriction. A compounded preparation is essentially a copy when it contains the same active pharmaceutical ingredient as a commercially available product, by the same route of administration, at the same, similar or easily substitutable strength. This is cured by the prescriber determining and documenting that a change produces a significant difference for an identified individual patient.

The volume restriction. Separately, section 503A limits compounding a product regularly or in inordinate amounts. Documentation does not cure this one. No quantity of well-written significant-difference determinations makes a high-volume operation compliant.

The Empower letter is best understood as an enforcement action under the second restriction that uses evidence drawn from the first. FDA looked at the volume, concluded the determinations could not plausibly be individual, and then found supporting evidence in how those determinations were written.

FDA has said it does not currently intend to act against a compounder filling four or fewer prescriptions of the same compounded product in a calendar month. The figure circulates widely and is routinely misread as a per-patient allowance. It is not.

Because refills count individually, the practical ceiling for a maintenance therapy is roughly four patients on ongoing treatment for a given formulation at a given pharmacy. A practice sending fifty patients to one pharmacy for the same preparation has exhausted the threshold twelve times over, regardless of how carefully each determination was written.

Two things about the figure are worth stating plainly. It is enforcement discretion, not a safe harbour — FDA describing where it currently intends to act, published as policy and revisable without notice. And the restriction beneath it is statutory, which is not revisable at all. The number is soft; the law under it is not.

FDA did not assert that niacinamide or cyanocobalamin are pharmacologically meaningless. It assembled three categories of evidence that the additions were there to defeat the copy restriction rather than to treat anyone.

Volume. The monthly order counts were inconsistent with individualized determination. This is the primary finding and the others support it. An operation producing thousands of units of one formulation is not making thousands of separate clinical judgments that happen to converge.

Templated language. Significant-difference justifications appeared verbatim across patient records. Identical reasoning for different patients is self-refuting: a determination specific to one person cannot be word-for-word identical to a determination specific to another.

Pre-selected options. Third-party platforms presented the combinations as menu choices. Where a patient or an intake flow selects a formulation before a prescriber evaluates anyone, the determination is not driving the prescription — the order is.

The second and third findings are the ones prescribers should sit with. Both are generated by prescriber workflow, not by the pharmacy. A pharmacy can be scrupulous about everything within its own walls and still receive a warning letter because of how the prescriptions reaching it were produced.

The instinctive response to the templated-language finding is to write better templates. That response fails, and understanding why is the most useful thing to take from the letter.

FDA's objection was not that the justifications were poorly worded. It was that they were identical in substance. Rephrasing an identical rationale into fifty different sentences produces fifty records that still describe the same clinical situation, which is the thing that cannot be true across fifty different patients. Varying the prose makes the pattern harder to spot; it does not make the determination real.

A determination that holds up has four properties:

1. It rests on a documented clinical fact about that patient. A recorded excipient reaction, an allergy noted in the chart, a prescribed strength genuinely not available commercially. Not a stated preference, and not an inference.
2. The change is tied to the fact. The specific modification addresses the specific problem. A patient with a documented polysorbate reaction gets a preparation without it — not a combination product whose second ingredient is unrelated to the reaction.
3. It is recorded when the prescription is written, in the patient's chart, retrievable. Not reconstructed later, and not held at the pharmacy.
4. It differs from other patients' determinations, because their charts differ. This is the test. If your determinations converge, the honest conclusion is that you are not making individual ones.

Cost does not qualify. Price is not a clinical difference and the standard does not recognise it as one. Neither does patient preference for a formulation, a dosing schedule or a delivery format, however reasonable the preference.

The uncomfortable implication: if a practice cannot produce patients whose charts genuinely differ, the problem is not the documentation. It is that there is nothing to document.

State boards of pharmacy retain independent authority over compounding standards regardless of FDA's determinations. A practice operating entirely inside FDA's enforcement-discretion threshold can still face a state board finding, and several states are already stricter than the federal standard — California tightened its compounding rules effective June 17, 2025.

For a telehealth practice this is not a footnote. A prescriber licensed in a single state has one additional regulator to satisfy. A practice licensed in all fifty has up to fifty, and both the patient's state and the filling pharmacy's state are in play for any given prescription.

The practical consequence is that a nationwide practice cannot realistically operate fifty compounding policies. It converges on whatever is defensible in the strictest state it treats patients in — which means the binding constraint is rarely the federal threshold everyone is discussing.

On October 7, 2026, the Washington Post reported that HHS is preparing a stopgap allowing certain previously restricted peptides to be sold temporarily while new regulations are written. It has been read in some quarters as a general loosening of peptide compounding. For GLP-1s it is not, and the two tracks are worth separating carefully.

These are different provisions addressing different problems. The peptides in the stopgap conversation have no approved commercial equivalent, so the question is whether a pharmacy may work with the substance at all. Semaglutide and tirzepatide have approved products on the market and no shortage in effect, so the question is entirely different: whether a compounded version duplicates something already available.

Nothing in the reporting touches the copy restriction, the volume restriction, or the Empower findings. The reporting is also thin by its own account: unannounced, attributed to two unnamed people, with HHS declining to comment, no named substances and no timeline. The Alliance for Pharmacy Compounding has requested interim enforcement discretion with conditions, but that is the trade body's ask rather than a confirmed element of any plan.

A prescriber waiting for this to resolve the GLP-1 question is waiting for the wrong thing.

The narrow reading of the Empower letter is that one pharmacy got caught. The useful reading is that FDA published its method, and most of the evidence that method relies on is generated in the prescriber's office.

Four things follow.

Volume is assessed at the formulation, not the practice. A clinic's billing arrangement is irrelevant to it. Whether patients pay per shipment or through a membership, the restriction counts how many units of one preparation a pharmacy compounds in a month. A subscription model does not soften this, and a recurring monthly shipment cadence arguably sits closer to the standing-schedule pattern FDA named than ad-hoc prescribing does.

A patient-specific exception cannot be operated as a product line. The threshold is roughly four patients on maintenance therapy per formulation per pharmacy. That is a clinical accommodation. Anything built to scale past it is relying on the exception for something it was not written to carry.

Intake design is now a compliance surface. If a formulation can be selected before a prescriber evaluates anyone, the order is originating the prescription rather than the other way round. That is the pattern FDA cited, and it lives in the platform, not the pharmacy.

The approved products remain available. Semaglutide and tirzepatide are on the market with no shortage in effect. For most patients for whom GLP-1 therapy is appropriate, the approved product is the straightforward answer, and the compounding question does not arise.

None of this makes patient-specific compounding improper. Section 503A provides for it deliberately, and a patient with a documented excipient reaction has a real clinical need that the exception exists to meet. The distinction FDA is drawing is between a practice that occasionally encounters such a patient and one that has discovered a great many of them at once.

• FDA, Clarifies Policies for Compounders as National GLP-1 Supply Begins to Stabilize — April 1, 2026
• FDA, Proposes to Exclude Semaglutide, Tirzepatide, and Liraglutide on 503B Bulks List — April 30, 2026
• Frier Levitt, FDA Warning Letter Targets Compounded GLP-1 Products — on the Empower Warning Letter of September 18, 2026
• Foley & Lardner, FDA Clarifies Policies for Pharmacy Compounders of GLP-1 Products
• Health Law Alliance, Compounding Copies of Commercially Available Drugs and Compounding Semaglutide After the Shortage Delisting
• Rachel Roubein, Trump administration paving way for temporary sale of some unapproved peptides, Washington Post — October 7, 2026
• FDA, Declaratory Order: Resolution of Shortages of Tirzepatide — December 19, 2024

This article is commentary on published regulatory materials. It is not legal advice, and prescribers should consult healthcare counsel on how these provisions apply in the states where they practise.