41.6% of 6,285 tests failed a yardstick modelled on compounded drugs; 71.1% failed one modelled on approved drugs 1.
Median purity was 99.80%, but the amount in the vial was the common failure: 32.8% failed on amount alone under the looser yardstick 1.
TB-500 tested worst (median purity 97.29%, 22.5% pass, 10% wrong or missing peptide). The study is a preprint of voluntarily submitted samples 1.
On this page
- The study
- The two yardsticks
- What the tests found
- Compound by compound
- BPC-157 and TB-500 under the HHS policy
- What a course costs, three ways
- What the study cannot tell you
- How this fits with our test-results database
- Frequently asked questions
- Related on Grey Peptides
- Sources
The study
In April 2026 Christopher Mendias and Tariq Awan of the Performance Medicine Institute in Phoenix posted the largest analysis yet of test results on peptides sold as 'research use only' 1. It is a preprint: it has not been peer reviewed, and no journal version had appeared when we checked on November 18, 2026. The authors report no external funding and no conflict of interest 1.
They did not buy or test anything themselves. They analysed a public dataset, compiled by Finnrick Analytics, of results from vials that consumers and peptide companies chose to send for testing between December 17, 2024 and April 3, 2026 1. Of 6,487 results screened, 46 were incomplete and 156 found the wrong peptide or none; the remaining 6,285 tests cover 14 compounds from 203 companies 1. Purity and amount were measured by HPLC, with ultraviolet or mass-spectrometry detection; endotoxin by LAL or a recombinant assay 1. The paper names no seller and no laboratory, and neither do we.
The two yardsticks
| Compounding yardstick | Approved-drug yardstick | |
|---|---|---|
| Modelled on | Drugs compounded under FDA's 503A or 503B standards | FDA-approved drugs and their monographs |
| Amount in the vial | 90.0% to 110.0% of the label | 95.0% to 105.0% of the label |
| Purity by HPLC | 98.0% or higher | 99.5% or higher |
| Identity | Wrong or missing peptide fails | Wrong or missing peptide fails |
The authors call these pragmatic benchmarks taken from guidance and the literature, not regulatory determinations 1. Endotoxin was reported separately and was not part of either pass mark 1.
What the tests found
- 41.6% failed the compounding yardstick and 71.1% failed the approved-drug yardstick 1.
- Purity was mostly high. The median was 99.80% (interquartile range 99.50% to 99.90%) 1.
- The amount was the problem. The median vial held 101.80% of its label, but the middle half ranged from 95.00% to 109.00% 1. Under the compounding yardstick, 32.8% of all tests failed on amount alone, against 3.0% on purity alone and 3.3% on both 1. Under the approved-drug yardstick: 44.3% amount alone, 8.4% purity alone, 15.9% both 1.
- Wrong or missing peptide: 156 results, 2.4% 1.
- Endotoxin was measured in only 243 samples, under 4% of the dataset: none detected in 74 (30%), below the limit of quantification in 133 (55%), and a measurable amount, 0.5 to 40, in 36 (15%) 1. The paper gives the unit both as EU/mL and as EU per vial.
- Not tested: sterility, heavy metals, residual solvents, particles and stability, so the authors call their failure rates a lower bound 1.
The practical reading: a certificate printing 99% purity says little about whether the vial holds the dose its label claims, and the dose is where these vials most often went wrong. Our COA decoder and vial checker walk through what a certificate and a vial can and cannot tell you.
Compound by compound
The paper prints no per-compound table. These are the figures its text states; a blank means the text does not give one, and we do not read values off its charts 1. Pass rates are the share passing both purity and amount.
| Compound | Identity failures | Median purity (IQR) | Median amount vs label (IQR) | Pass rate, compounding yardstick | Pass rate, approved-drug yardstick |
|---|---|---|---|---|---|
| BPC-157 | 17 (3.5%) | not reported | 103.50% (94.20–115.10), the highest | not reported | not reported |
| Cagrilintide | not reported | not reported | not reported | not reported | not reported |
| CJC-1295 | 22 (9.1%) | 98.49% (91.59–99.60) | not reported | 29.2% | 10.7% |
| GHK-Cu | 1 | 99.90% (99.85–99.95), the highest | not reported | not reported | not reported |
| Ipamorelin | 6 | not reported | not reported | not reported | 34.2% |
| PT-141 (bremelanotide) | not reported | not reported | not reported | not reported | 29.7% |
| Retatrutide | 59 | not reported | not reported | 67.8%, the highest | 47.0%, the highest |
| Semaglutide | 6 | not reported | not reported | not reported | not reported |
| Sermorelin | not reported | not reported | 97.45% (89.30–105.00), the lowest | not reported | not reported |
| Survodutide | not reported | not reported | not reported | 64.0% | not reported |
| TB-500 | 4 (10.0%) | 97.29% (93.83–99.59), the lowest | not reported | 22.5%, the lowest | 7.5% |
| Tesamorelin | 11 (3.6%) | not reported | not reported | not reported | not reported |
| Thymosin beta-4 | not reported | not reported | not reported | 64.5% | not reported |
| Tirzepatide | 30 | not reported | not reported | not reported | 13.8% |
Retatrutide had the highest pass rates on both yardsticks, and the most identity failures of any compound the text gives a count for (59); TB-500 had the lowest median purity, the lowest pass rate on the compounding yardstick and a 10% identity-failure rate; BPC-157 vials held the most above their label, a median 103.50% 1. Tirzepatide shows the amount problem at its plainest: 58.8% met the approved-drug purity mark but failed on amount 1.
BPC-157 and TB-500 under the HHS policy
Two of the 14 are among the six peptides HHS's interim policy lets a licensed 503A pharmacy compound for a named patient. That policy requires the pharmacy's raw peptide to come from an FDA-registered maker with a certificate of analysis covering identity, purity, potency and contaminants including endotoxin 2. The study's TB-500 and BPC-157 figures are the gap those conditions are meant to close; see the TB-500 and BPC-157 legal-access pages. The study did not include KPV, MOTS-c, semax or epitalon 1.
What a course costs, three ways
The authors priced a full trial-dose course of four peptides from grey-market price-tracking websites, compounding pharmacies found by web search and the makers' cash prices, with all prices as of April 13, 2026 1. Medians:
| Course | Grey market | 503A compounded | FDA-approved |
|---|---|---|---|
| PT-141 (24 weeks) | $122 | $266 | $4,819 |
| Tesamorelin (26 weeks) | $1,605 | $3,928 | Not compared (limited-access programme at no cost) |
| Semaglutide (68 weeks) | $973 | $5,004 | $5,633 (Wegovy, one price) |
| Tirzepatide (72 weeks) | $4,561 | $9,512 | $7,882 (Zepbound, one price) |
Per milligram, the grey market was cheapest and compounded prices were significantly higher (P<0.05) for all four 1:
| Peptide | Grey market, $/mg | 503A compounded, $/mg | FDA-approved, $/mg |
|---|---|---|---|
| PT-141 | $3.88 (n=86) | $8.50 (n=34) | $153.00 (n=47) |
| Semaglutide | $7.00 (n=53) | $36.00 (n=64) | $36.35 (n=1) |
| Tesamorelin | $6.30 (n=81) | $15.42 (n=33) | Not compared |
| Tirzepatide | $4.90 (n=51) | $10.25 (n=47) | $7.48 (n=1) |
These are one-day prices from April 2026. Branded GLP-1 cash prices have changed since; our cost guide carries the makers' current figures. The paper has no prices for BPC-157 or TB-500.
What the study cannot tell you
- It is not a random sample. Vials were sent in voluntarily, so sellers confident in a batch may be over-represented and bad batches under-represented 1.
- Endotoxin data are thin: under 4% of samples 1.
- The pass marks are benchmarks, not legal specifications for products that are not drugs 1.
- Prices are a snapshot from one day 1.
- Several hazards were not measured (sterility above all), so a pass here is not a clean bill of health 1.
- It is a preprint, and a few of its own numbers disagree: the sample total is 6,487 in one place and 6,441 in another, and it ranks survodutide's 64.0% above thymosin beta-4's 64.5% 1.
How this fits with our test-results database
Our test-results database records individual products tested by named, published studies and regulators, one row per product, each with its method and source 3. This study reports aggregates only, with no per-product results, so it is summarised here rather than added row by row. The two agree on the main point: what is on the label is often not what is in the vial.
Frequently asked questions
How many grey-market peptides fail lab testing?
In the largest analysis so far, 6,285 tests of 14 peptides from 203 companies, 41.6% failed a yardstick modelled on compounded drugs and 71.1% failed one modelled on approved drugs. It is a preprint, not peer reviewed, and the samples were sent in voluntarily.1
Are research peptides pure?
Mostly, by HPLC: the median purity was 99.80%. The common failure was the amount: under the compounding yardstick 32.8% of tests failed on amount alone, against 3.0% on purity alone.1
Does grey-market TB-500 test worse than other peptides?
In this study, yes: TB-500 had the lowest median purity (97.29%), the lowest pass rate on the compounding yardstick (22.5%) and a 10% identity-failure rate.1
Do grey-market peptides contain endotoxin?
Of the 243 samples tested for it, 36 (15%) had a measurable amount, between 0.5 and 40; 74 had none detected and 133 were below the limit of quantification. Fewer than 4% of samples were tested.1
Related on Grey Peptides
Sources
The preprint is the source of every figure on this page; it is not peer reviewed. Read November 18, 2026.
- Mendias CL, Awan TM. Evaluation of Research Grade Peptides Marketed Directly to Consumers Reveals Extensive Variability in Purity and Measured Abundance. Preprints.org 2026; 202604.1748, version 1, posted April 24, 2026. doi:10.20944/preprints202604.1748.v1. Preprint, not peer reviewed; CC BY 4.0. Read November 18, 2026. Source
- HHS: Enforcement Discretion Policy Regarding 503A Compounding Using Certain Peptide Bulk Drug Substances (policy text, undated PDF). Read November 18, 2026. Source
- Grey Peptides lab-results database (/verify/test-results/): independent test results on individual products from journals and regulators, each row sourced.
This page reports a published analysis; it is not a guide to buying and rates no seller or laboratory. Grey Peptides sells nothing. If you have taken a product and feel unwell, contact a clinician or Poison Help (1-800-222-1222 in the US).