This article is educational and does not replace medical advice. Prescription medication requires review by a licensed clinician and, when appropriate, a valid prescription. Compounded medications are not FDA-approved, and the FDA does not verify their safety, effectiveness or quality before marketing. Treatment eligibility is an individual clinical decision.
How to verify a compounding pharmacy
Quick answer
To verify a compounding pharmacy: confirm it is named, check its state license and any FDA 503B registration, confirm the prescribing clinician is named, look for disclosed concentration and formulation, and reject any 'research use only' product. Unnamed pharmacy equals red flag.
The checklist
- The pharmacy partner is named on the provider's site.
- Its state license (503A) or FDA registration (503B) can be independently verified.
- A licensed prescribing clinician is named.
- A real telehealth visit and medical review are required.
- Formulation and concentration are disclosed.
- There is no 'research use only' or 'not for human consumption' disclaimer.
Red flags
Monitoring and laboratory work
Questions to ask your clinician
- Given my history, is a GLP-1 appropriate for me at all — and is there a reason it might not be?
- What baseline laboratory work will you order before I start?
- What is the target dose, and how quickly will we escalate to it?
- What side effects should make me call you rather than wait?
- What is the plan for maintenance, and what happens if I stop?
- Will I see the same clinician at follow-up, or a different one each time?
Questions to ask about the pharmacy
- Which specific pharmacy will fill my prescription? Not "our network" — the name of the facility.
- Is it a 503A state-licensed pharmacy or a 503B FDA-registered outsourcing facility? These are different regulatory categories with different oversight, and a company can use both for different products.
- In which state is it licensed, and can I look up the licence? State boards of pharmacy publish licensee databases.
- What is the exact salt form and concentration? Semaglutide sodium and semaglutide acetate are not the same active ingredient as the semaglutide base in approved products, and the FDA has said they are not appropriate for compounding.
- Is the vial single-dose or multi-dose? A multi-dose vial requires you to measure each dose yourself, which is the most common source of the dosing errors behind reported adverse events.
- Will you provide a certificate of analysis?
- Has the pharmacy received any FDA warning letter or state board action?
What happens when you stop
Storage and handling
How to verify any of this yourself
- Go to the provider's own pricing page. Not a comparison site — the provider's. Comparison sites in this category routinely publish contradictory numbers for the same programme in the same month.
- Find the ongoing price, not the headline. Look for the words "first month", "intro", "starting at" or "new patients". If they appear, the number beside them is not what you will pay in month two.
- Add the membership. If the medication and the membership are billed separately, add them. That sum is your real monthly cost.
- Ask what the highest dose costs. By email or chat, so you have it in writing.
- Ask about early cancellation before you commit to a plan longer than a month.
- Check the manufacturer. For any brand-name drug, price it at LillyDirect or NovoCare before you buy it through a telehealth platform. Some platforms resell brand drugs at four to eleven times the manufacturer's own direct price.
Who is actually who: the entities in this transaction
| Entity | What it is | Regulated by | What it is NOT |
|---|---|---|---|
| Telehealth company | The website you sign up on. Arranges the consultation, handles billing and logistics. | State corporate practice rules; FTC for advertising | Not a pharmacy. Does not make your medicine. |
| Prescribing clinician | The licensed physician, NP or PA who evaluates you and writes the prescription. | Their state medical or nursing board | Not employed by the pharmacy. Must exercise independent judgement. |
| 503A compounding pharmacy | A state-licensed pharmacy compounding for an individual patient against a specific prescription. | State board of pharmacy; FDA for some provisions | Not FDA-approved. Products are not reviewed before marketing. |
| 503B outsourcing facility | An FDA-registered facility that may compound in bulk without patient-specific prescriptions. | FDA, including cGMP inspection | Still not making FDA-approved products. |
| Manufacturer | Eli Lilly, Novo Nordisk. Makes the FDA-approved branded drug. | FDA — full premarket approval | Not involved in compounded products at all. |
Two phrases to distrust immediatelyThere is no such thing as an 'FDA-approved pharmacy'. That phrase appears in marketing and it is meaningless. A pharmacy can be state-licensed (503A) or FDA-registered (503B). Neither makes its compounded products FDA-approved — approval is something that happens to a drug, after clinical trials, not to a facility.
Equally: a provider's statement about which pharmacy it uses is a provider-reported relationship until someone verifies it. We label it that way, and so should you when you read it.
Equally: a provider's statement about which pharmacy it uses is a provider-reported relationship until someone verifies it. We label it that way, and so should you when you read it.
Eligibility, and who is likely to be declined
State availability, and why it varies
Limitations of this analysis
Frequently asked questions
How do I know if an online pharmacy is legitimate?
Sources
- U.S. Food and Drug Administration — labels and safety communications.
- Peer-reviewed clinical trials cited above.
- Our methodology and medical review policy.
Registration is per-facility, not per-company: one company can operate both. Source: FDCA sections 503A and 503B; FDA compounding guidance.
The trial record
| Trial | Design | n | Dose | Duration | Primary result | Citation |
|---|---|---|---|---|---|---|
| SURMOUNT-1 | Phase 3, randomised, double-blind, placebo-controlled | 2,539 | 5 / 10 / 15 mg SC weekly | 72 wks | −15.0% / −19.5% / −20.9% vs −3.1% placebo | Jastreboff, NEJM 2022; NCT04184622 |
| SURMOUNT-2 | Phase 3, RCT, in type 2 diabetes | 938 | 10 / 15 mg SC weekly | 72 wks | −12.8% / −14.7% vs −3.2% placebo | Garvey, Lancet 2023; NCT04657003 |
| SURMOUNT-3 | Phase 3, RCT, after 12-wk intensive lifestyle lead-in | 806 | Max tolerated (10/15 mg) | 72 wks | −18.4% additional, vs +2.5% placebo | Wadden, Nat Med 2023; NCT04657016 |
| SURMOUNT-4 | Randomised WITHDRAWAL after 36-wk open-label lead-in | 670 | Max tolerated | 88 wks | Continue: −5.5% further. Withdraw to placebo: +14.0% REGAINED | Aronne, JAMA 2024; NCT04660643 |
| SURMOUNT-5 | Phase 3b, OPEN-LABEL, active-controlled head-to-head | 751 | Max tolerated vs semaglutide | 72 wks | −20.2% vs semaglutide −13.7%, p<0.001 | Aronne, NEJM 2025; NCT05822830 |
| SURPASS-2 | Phase 3, RCT, type 2 diabetes, active-controlled | 1,879 | 5 / 10 / 15 mg vs semaglutide 1 mg | 40 wks | HbA1c −2.01 to −2.30% vs −1.86% | Frías, NEJM 2021; NCT03987919 |
| SURPASS-CVOT | Phase 3, cardiovascular outcomes, vs dulaglutide | 13,299 | Max tolerated | ~4.5 yrs | Non-inferior for MACE; not superiority vs placebo | Nicholls, 2024; NCT04255433 |
The caveats that belong with the numbersThree things must travel with every one of those numbers.
1. They are means, not promises. A −20.9% mean in SURMOUNT-1 contains people who lost far more and people who lost almost nothing. A trial average tells you what happened to a population; it does not tell you what will happen to you.
2. Every one is an FDA-APPROVED SUBCUTANEOUS INJECTION. No trial in this table tested a compounded preparation, a microdose regimen, or an orally disintegrating tablet. When these figures appear on a page selling a compounded ODT, evidence has been moved across a dosage form without justification.
3. All were funded by Eli Lilly, which manufactures tirzepatide. That is normal in drug development and does not make the results false — these are large, peer-reviewed studies. It belongs in the citation anyway, and it matters most in SURMOUNT-5, where the funder made the winning drug and the trial was open-label.
1. They are means, not promises. A −20.9% mean in SURMOUNT-1 contains people who lost far more and people who lost almost nothing. A trial average tells you what happened to a population; it does not tell you what will happen to you.
2. Every one is an FDA-APPROVED SUBCUTANEOUS INJECTION. No trial in this table tested a compounded preparation, a microdose regimen, or an orally disintegrating tablet. When these figures appear on a page selling a compounded ODT, evidence has been moved across a dosage form without justification.
3. All were funded by Eli Lilly, which manufactures tirzepatide. That is normal in drug development and does not make the results false — these are large, peer-reviewed studies. It belongs in the citation anyway, and it matters most in SURMOUNT-5, where the funder made the winning drug and the trial was open-label.
Jastreboff AM et al., N Engl J Med 2022, n=2,539 (NCT04184622). The effect rises with dose — which is precisely why a ~1mg 'microdose' cannot be expected to produce the headline result. FDA-approved subcutaneous injection.
What the trials do and do not coverThe boundary of the evidence, for this treatment. Every efficacy figure on this page comes from a trial of an FDA-approved subcutaneous injection. None of it was collected on a compounded preparation, a microdose regimen, or an orally disintegrating tablet.
The evidence is strong exactly where it was gathered and silent everywhere else. The gap between those two things is where most of the marketing in this industry operates, and recognising it is the single most useful skill a patient in this market can have.
The evidence is strong exactly where it was gathered and silent everywhere else. The gap between those two things is where most of the marketing in this industry operates, and recognising it is the single most useful skill a patient in this market can have.
Dosing, titration, and what it does to your bill
| Period | Dose | What it is for |
|---|---|---|
| Weeks 1–4 | 2.5 mg | Tolerance-building only. This dose is not intended to produce weight loss. If your provider's price is quoted at 2.5 mg, that is not the price of treatment. |
| Weeks 5–8 | 5 mg | First therapeutic dose (−15.0% in SURMOUNT-1). |
| Weeks 9–12 | 7.5 mg | Escalate only if tolerated. |
| Weeks 13–16 | 10 mg | A common maintenance dose (−19.5%). |
| Weeks 17–20 | 12.5 mg | Escalate only if tolerated. |
| Week 21+ | 15 mg | Maximum maintenance dose (−20.9%). |
Why titration decides your real priceTitration is where cost is actually decided, and almost no pricing page says so.
The advertised price is usually the 2.5 mg price. On a programme that escalates with dose, the rate you are quoted at signup is for a dose the label explicitly describes as a starting dose — not a treatment dose. Ask what you will pay at 10 mg, and compare that number instead.
A 'microdose' of ~1 mg/week sits below every dose in SURMOUNT. The trials used 5, 10 and 15 mg. A microdose is not a discounted route to the SURMOUNT result; it is a different product with a smaller expected effect and no equivalent trial evidence.
The advertised price is usually the 2.5 mg price. On a programme that escalates with dose, the rate you are quoted at signup is for a dose the label explicitly describes as a starting dose — not a treatment dose. Ask what you will pay at 10 mg, and compare that number instead.
A 'microdose' of ~1 mg/week sits below every dose in SURMOUNT. The trials used 5, 10 and 15 mg. A microdose is not a discounted route to the SURMOUNT result; it is a different product with a smaller expected effect and no equivalent trial evidence.
Safety, contraindications and monitoring
Percentage of participants reporting each event. Gastrointestinal effects dominate, are usually mild-to-moderate, and are most pronounced during dose escalation. Source: SURMOUNT-1, N Engl J Med 2022.
Discontinuation: what the withdrawal trial found
Aronne LJ et al., JAMA 2024, n=670 (NCT04660643). After a 36-week open-label lead-in, participants randomised to placebo regained ~14% of body weight over the following 52 weeks; those who continued lost a further ~5%. This is the single most important trial for understanding the true cost of treatment.
Questions to ask your clinician
- Given my history — specifically thyroid, pancreatic and gallbladder — is a GLP-1 appropriate for me at all?
- What baseline laboratory work will you order before I start?
- What is my target dose, and how quickly will we escalate?
- Which side effects should make me call you rather than wait it out?
- What is the plan for maintenance, and what happens if I stop?
- Will I see the same clinician at each follow-up, or a different one each time?
Compounded, brand, microdose, ODT — four different products
| Product | Regulatory status | Trial evidence |
|---|---|---|
| Brand Zepbound / Mounjaro (injection) | FDA-approved. Reviewed for safety, effectiveness and quality before marketing. | Direct. SURMOUNT and SURPASS tested exactly this product. |
| Brand Foundayo (oral, orforglipron) | FDA-approved. Its own trial programme. | Direct, for that product. |
| Compounded Semaglutide (injection, full dose) | NOT FDA-approved. No premarket review of safety, effectiveness or quality. | None for the compounded product itself. Same molecule, same route — but the product in your hand was never in a trial. |
| Microdose (~1 mg/wk) | NOT FDA-approved. | None. Sits BELOW every dose in SURMOUNT (5/10/15 mg). Expect a smaller effect. |
| ODT / oral compounded | NOT FDA-approved. | NONE. No trial has ever tested it. Oral bioavailability for these peptides is a real pharmacological problem and is unpublished for this product. |
What the trials do and do not coverThe boundary of the evidence, for this treatment. Every efficacy figure on this page comes from a trial of an FDA-approved subcutaneous injection. None of it was collected on a compounded preparation, a microdose regimen, or an orally disintegrating tablet.
The evidence is strong exactly where it was gathered and silent everywhere else. The gap between those two things is where most of the marketing in this industry operates, and recognising it is the single most useful skill a patient in this market can have.
The evidence is strong exactly where it was gathered and silent everywhere else. The gap between those two things is where most of the marketing in this industry operates, and recognising it is the single most useful skill a patient in this market can have.
Adverse events: the figure almost every site gets wrong
Source: FDA GLP-1 webpage, reporting 1,700+ adverse events associated with compounded semaglutide and tirzepatide as of May 21, 2026 — against the 775 total, Feb 2025 figures from February 2025 that almost every comparison site is still quoting. Reports are voluntary and do not establish causation, but the trend is the point.
1,700+ — not 775. We were wrong too, and we have corrected it.The adverse-event figure you have read elsewhere is out of date. Nearly every comparison site — and this site, until we rechecked — quotes 1,700+ reports for compounded semaglutide and 320 for compounded tirzepatide. Those are February 2025 figures.
As of 21 May 2026, the FDA reports having received more than 1,700 adverse events associated with compounded semaglutide and tirzepatide. That is more than double the figure still in circulation, in roughly fifteen months.
Adverse-event reports are voluntary, are not adjudicated, and do not by themselves establish causation. That caveat is real and we will not drop it. But a site that quotes the 2025 number in mid-2026 is not being cautious — it is being out of date, and in a direction that flatters the product it is paid to sell.
As of 21 May 2026, the FDA reports having received more than 1,700 adverse events associated with compounded semaglutide and tirzepatide. That is more than double the figure still in circulation, in roughly fifteen months.
Adverse-event reports are voluntary, are not adjudicated, and do not by themselves establish causation. That caveat is real and we will not drop it. But a site that quotes the 2025 number in mid-2026 is not being cautious — it is being out of date, and in a direction that flatters the product it is paid to sell.
Tirzepatide + B12 is not tirzepatideA March 2026 study identified a previously unknown tirzepatide–B12 adduct in mass-compounded tirzepatide formulated with vitamin B12. An adduct is a new chemical entity formed when two molecules combine. This one does not exist in FDA-approved tirzepatide, and its safety has not been characterised.
This matters far beyond one study, because it exposes the flaw in the whole ‘personalized dosing’ defence. Adding B12 was one of the commonest ways compounders argued their product was not “essentially a copy” of the approved drug — a clinical difference that kept them inside the law. The finding shows that the additive did not merely differentiate the product on paper. It chemically changed it, into something nobody has tested in a human being.
What to do: if you are taking a compounded tirzepatide that contains B12 — and many do, often marketed as ‘tirzepatide + B12’ or ‘with methylcobalamin’ — ask your provider and your pharmacy, in writing, whether they have tested for adduct formation. Most will not have. That answer is itself information.
This matters far beyond one study, because it exposes the flaw in the whole ‘personalized dosing’ defence. Adding B12 was one of the commonest ways compounders argued their product was not “essentially a copy” of the approved drug — a clinical difference that kept them inside the law. The finding shows that the additive did not merely differentiate the product on paper. It chemically changed it, into something nobody has tested in a human being.
What to do: if you are taking a compounded tirzepatide that contains B12 — and many do, often marketed as ‘tirzepatide + B12’ or ‘with methylcobalamin’ — ask your provider and your pharmacy, in writing, whether they have tested for adduct formation. Most will not have. That answer is itself information.
FDA: affordability is not a clinical needThe FDA has explicitly rejected the argument that this entire industry rests on.
In the 30 April 2026 Federal Register notice (docket 2026-08552), the agency stated that there is no clinical need for outsourcing facilities to compound semaglutide, tirzepatide or liraglutide from bulk — and went out of its way to clarify that supply and affordability are not what the statute means by clinical need.
In plain terms: there are FDA-approved products; they work; patients can be treated with them. Whether a patient can afford them is a different problem, with a different set of policy tools.
That single sentence does enormous work. Every compounded-GLP-1 marketing page in America is, at bottom, an affordability argument. The agency has now said, on the record, that affordability is not a legal basis for compounding these drugs. If you are choosing a compounded programme because it is cheaper, you should know that the regulator has explicitly said that reason does not count.
In the 30 April 2026 Federal Register notice (docket 2026-08552), the agency stated that there is no clinical need for outsourcing facilities to compound semaglutide, tirzepatide or liraglutide from bulk — and went out of its way to clarify that supply and affordability are not what the statute means by clinical need.
In plain terms: there are FDA-approved products; they work; patients can be treated with them. Whether a patient can afford them is a different problem, with a different set of policy tools.
That single sentence does enormous work. Every compounded-GLP-1 marketing page in America is, at bottom, an affordability argument. The agency has now said, on the record, that affordability is not a legal basis for compounding these drugs. If you are choosing a compounded programme because it is cheaper, you should know that the regulator has explicitly said that reason does not count.