Prescribing in the Dark: Peptide Liability, the Regulatory Gray Zone, and the Genetics Problem No One Has Solved
On April 1, 2026, a federal grand jury in Utah indicted Justin Bradley Watkins, D.O., on eight criminal counts arising from the alleged sale of BPC-157, TB-500, tirzepatide, semaglutide, retatrutide, cagrilintide, ipamorelin, CJC-1295, GHK, GHK-Cu, and NAD+ to patients. Watkins was not a supplement vendor. He was a licensed physician operating a clinic. According to the indictment, the products were sourced from China and relabeled before being provided to more than 200 patients. The charges are receipt and delivery for pay of misbranded drugs with intent to defraud, under 21 U.S.C. § 331. An indictment is only an allegation, and Watkins is presumed innocent unless and until proven guilty. United States v. Watkins, No. 1:26-cr-00015 (D. Utah, filed Apr. 1, 2026).
That case is the clearest signal yet that federal enforcement has moved from manufacturers and distributors to the clinical level. The question for Oregon practitioners operating peptide clinics is not whether they recognize the name Justin Watkins. It is whether they recognize the fact pattern.
The Regulatory Landscape by Category
Peptides do not occupy a single regulatory status. They exist across a spectrum, and where a specific compound sits on that spectrum determines the nature of the legal exposure.
GLP-1 agonists: compounding authority largely gone. The FDA declared the semaglutide shortage resolved in February 2025 and the tirzepatide shortage resolved earlier. Under 503A and 503B compounding framework, pharmacies may only compound drugs that appear on an active shortage list. Once a shortage is delisted, the authority to compound evaporates. Novo Nordisk has filed more than 130 lawsuits in 40 states against telehealth companies, medical spas, and clinics that continued selling compounded semaglutide after the delisting. In Oregon, Novo Nordisk A/S v. Gorin Plastic Surgery Investments, LLC, No. 3:25-cv-01373 (D. Or., filed Aug. 4, 2025), targeted a Tualatin clinic. Consent-to-judgment documents were filed February 9, 2026. The clinic agreed to stop. Eli Lilly filed parallel suits for tirzepatide. The Fifth Circuit is considering whether FDA's shortage delisting required notice-and-comment rulemaking, and that appeal remains pending, but the civil enforcement wave does not wait for appellate resolution.
BPC-157 and TB-500: removed from prohibition, not authorized. In April 2026, FDA removed twelve peptides, including BPC-157 and TB-500, from its Category 2 "do not compound" list. That removal is frequently mischaracterized as authorization. It is not. It means only that the explicit prohibition was lifted while the substances await formal evaluation. The FDA's Pharmacy Compounding Advisory Committee meets July 23 and 24, 2026, to evaluate seven of them, including BPC-157 and TB-500, for possible addition to the 503A list of bulk drug substances that may be used in compounding. Clinics prescribing BPC-157 or TB-500 right now are operating in a window whose legal character will be defined by that meeting's outcome.
Sermorelin, ipamorelin, CJC-1295, and related growth hormone secretagogues. These stimulate the pituitary to release growth hormone rather than supplying exogenous GH directly. They have been prescribed off-label for anti-aging, body composition, and recovery purposes. None have FDA approval for those indications. Compounding authority depends on each compound's specific regulatory status, which varies. Several remain in a working gray zone, but the Watkins indictment included ipamorelin and CJC-1295 among the charged substances.
PT-141 (bremelanotide). An FDA-approved version (Vyleesi) exists for hypoactive sexual desire disorder in premenopausal women. Compounded versions are sold for a wider range of indications and patient populations. The existence of an approved version complicates the compounding analysis: 503A pharmacies generally cannot compound a drug that is commercially available in an approved form unless the patient has a documented allergy or intolerance to an ingredient in the commercial product.
Dihexa and other experimental neuroactive peptides. The primary supporting preclinical study for Dihexa was retracted in April 2025 for data fabrication. There are zero FDA-recognized human exposure data for Dihexa. It has no approved form, no clinical trials, and now no credible preclinical foundation. Prescribing it in any formulation presents acute informed consent problems.
Semax and Selank. These Russian-developed neuropeptides, marketed for cognition and anxiety respectively, have no FDA approval for any indication and rest on a largely Russian-language evidence base with no US clinical trials. Semax is one of the seven substances on the July 2026 PCAC agenda. Selank is not on that list, which means it was never nominated for the 503A bulks list at all, leaving its compounding status more precarious rather than less.
Intranasal oxytocin. A 2025 systematic review of 16 randomized controlled trials found mixed efficacy for PTSD and anxiety indications, with opposite neural connectivity effects in male and female patients and cardiovascular signals including elevated heart rate. No professional society has issued prescribing guidelines for psychiatric oxytocin. The evidence base does not support a defined standard of care in either direction.
The Civil Litigation Wave
Criminal prosecution is not the only exposure vector. The civil litigation framework for peptide and compounded GLP-1 prescribing is developing rapidly across three distinct theories.
Manufacturer suits under Lanham Act and state UTPA. The Novo Nordisk and Eli Lilly suits allege false advertising: clinics represented compounded products as equivalent to FDA-approved drugs. Oregon's Unlawful Trade Practices Act, ORS 646.608, provides a state-law vehicle for these claims that does not require federal preemption analysis. The Oregon Tualatin consent judgment demonstrates this theory reaches clinical operators directly, not only manufacturers or large telehealth platforms.
RICO class actions from patient plaintiffs. In Day v. OpenLoop Health Inc., No. 1:25-cv-01418 (D. Del., filed Nov. 20, 2025), patient plaintiffs sued a telehealth prescribing platform and its affiliated compounding pharmacy under civil RICO, alleging they sold compounded oral tirzepatide that was pharmacologically worthless. The theory: wire fraud through marketing and prescribing a product with no credible evidence of efficacy. A motion to dismiss was pending as of July 2026. If the motion is denied, RICO becomes a viable vehicle for patient plaintiffs suing any peptide clinic that prescribes unvalidated compounded formulations for pay.
Consumer class actions. Donoho v. Hims & Hers Health, Inc., No. 1:26-cv-01954 (N.D. Ill., filed Feb. 20, 2026), is one of several consumer class actions pending against telehealth GLP-1 prescribers. These cases are largely stayed pending arbitration motions, suggesting that arbitration clauses are the primary defense strategy for larger platforms. Smaller independent clinics are less likely to have enforceable arbitration agreements with their patients.
Vendor supply chain liability. In United States v. Kawa, No. 3:25-cr-00091 (N.D. Ind.), the owner of Paradigm Peptides pleaded guilty in December 2025 to introducing unapproved new drugs into interstate commerce and importing merchandise contrary to law. Sentencing is scheduled for July 30, 2026. Clinics that purchased peptides from gray-market research chemical suppliers face potential aiding and abetting exposure if the DOJ traces the supply chain.
A pending SCOTUS case, Wells Pharma of Houston, LLC v. Zyla Life Sciences, LLC, No. 25-257, will determine whether the FDCA preempts private state-law unfair competition claims against compounders. If the Fifth Circuit's ruling stands, state UTPA claims become available not only to brand manufacturers but potentially to patient plaintiffs as well.
Malpractice Insurance: The Gap That Matters
Malpractice coverage is another exposure. Professional liability policies commonly exclude claims arising from unapproved or off-label drug use, and a prescriber who relies on such coverage may find that a claim tied to compounded BPC-157 or Dihexa falls outside the policy. Whether a given policy responds turns on its specific exclusions, which prescribers should read closely rather than assume. Where the coverage does not reach, the prescriber is personally exposed.
Where Genetic Testing Intersects: A Different Question
For pharmaceutical drugs, the genetic testing liability question is primarily about metabolism: CYP2D6 and CYP2C19 variants determine how quickly a patient processes a medication, affecting dose, efficacy, and side-effect risk. That framework does not transfer directly to peptides.
Peptides are not small molecules. They are broken down by peptidases and proteases, not by CYP enzymes. A standard pharmacogenomics panel would not tell a prescriber much about how a patient will metabolize BPC-157. The CYP testing conversation that is relevant to psychiatric prescribing is largely irrelevant here.
But the receptor side is a different matter. For several peptides in active clinical use, established genetic variants affect receptor function in ways that would change the risk-benefit calculation for a specific patient, if anyone were studying them.
For GLP-1 agonists, GLP1R variants are documented to affect both weight loss response and gastrointestinal side effect severity. TCF7L2 variants, the primary type 2 diabetes susceptibility gene, affect downstream GLP-1 signaling. A prescriber who orders semaglutide or tirzepatide without knowing a patient's GLP1R status is making a personalized medicine decision without the relevant personalized data.
For PT-141, MC4R variants are well-established in the pharmacology literature. Melanocortin 4 receptor loss-of-function variants cause obesity and affect drug response. A prescriber using PT-141 in a patient with an unidentified MC4R variant is prescribing to an unknown pharmacological target.
For growth hormone secretagogues including sermorelin, ipamorelin, and CJC-1295, the GHR exon 3 deletion polymorphism is one of the most studied variants in the growth hormone literature. Patients with the d3-GHR variant respond differently to GH-axis stimulation. This is not theoretical: the d3-GHR variant is why some patients respond robustly to GH therapy and others do not at equivalent doses.
For intranasal oxytocin, OXTR rs53576 is one of the most studied receptor polymorphisms in behavioral psychiatry. CD38 variants affect endogenous oxytocin release. The 2025 systematic review found opposite neural connectivity effects in male and female PTSD patients, a finding that suggests sex-specific biological factors interact with oxytocin response in ways that are not yet understood.
Semax and Selank sit in a revealing spot on this spectrum. Both are proposed to work by increasing BDNF signaling, and BDNF carries one of the best-characterized common variants in psychiatry: Val66Met (rs6265), present in roughly one in five people. The Met allele reduces activity-dependent BDNF secretion, which is the very process these peptides are marketed to enhance. Semax, as an ACTH fragment, also retains affinity for the same MC4R receptor implicated in PT-141 response. No human study has tested whether Val66Met or MC4R status predicts response to either peptide; the evidence is confined to rodent and cell-culture gene-expression work. But the biological plausibility is higher here than for BPC-157, where no candidate variant exists at all. That makes the disclosure problem sharper, not softer: a clinic selling a BDNF-mediated benefit is selling something a common, testable variant may blunt in a substantial minority of patients, a fact that appears nowhere in the consent conversation.
The Informed Consent Problem
Here is where the genetic testing question cuts differently for peptides than for traditional drugs. For citalopram and CYP2C19, the genetic data exists, the clinical guidelines exist, and the liability question is whether the prescriber had a duty to test. The answer is contested but the tools are available.
For BPC-157 and CYP metabolism, the genetic data does not exist because no one has studied it in humans. A prescriber who wanted to tell a patient exactly how their genetics would affect their response to BPC-157 cannot do so, not because they failed to order a test, but because the test has never been developed.
That is a harder informed consent problem. The standard for informed consent requires disclosure of material risks that a reasonable patient would want to know. If a material risk is that the patient's individual biological profile, including their receptor genetics, may make the drug more or less effective or more or less dangerous, and no data exists to quantify that risk, then truly informed consent is structurally unavailable. The prescriber is asking the patient to accept unknown pharmacogenomic variability with no data to frame it.
In malpractice terms, the informed consent failure and the prescribing-without-evidence failure may be the same claim. The argument is not: you should have tested for BPC-157 receptor genetics. The argument is: you prescribed a substance with no human clinical data for the indication, no established dosing standard, no professional society guidelines, no insurance coverage, and no ability to quantify individual genetic risk. Whatever you told the patient about what to expect, you did not have the information necessary to tell them accurately.
The Decision Window
The PCAC meeting on July 23 and 24 will produce recommendations that may resolve some of this uncertainty for BPC-157 and TB-500. If the committee recommends adding these compounds to the 503A bulks list, the regulatory authorization question moves toward resolution in the clinics' favor. If the committee recommends against adding them, clinics that continued prescribing after April 2026 will have done so during a window that is now clearly defined on both ends.
Clinics operating right now, between the April 2026 prohibition removal and the July 2026 PCAC outcome, are making a decision about that window. The legal question that will eventually be asked is what they knew about the regulatory status, what they told patients, and whether the documentation in the chart reflects a considered clinical judgment or a commercial practice that operated without examining the legal environment it occupied.
David Brunk is a civil litigation attorney. newmanbrunk.com · david@newmanbrunk.com
The allegations described here are taken from the filing and remain unproven; no responsive pleading is reflected in the source document.
David Brunk is a civil litigation attorney. He can be reached at david@newmanbrunk.com.
Questions about this topic: david@newmanbrunk.com