A Study Said Up to 71% of Grey-Market Peptides Failed. We Checked 64,787 Lab Records.

Bottom line: the study combined four different problems—wrong identity, low purity, too little peptide and too much peptide—into one failure number. It also relied on reports voluntarily submitted to one testing program. Its math checks out; the claim that the number represents the whole market does not.

TitrateLab checked 64,787 usable lab records from multiple sources, extracted August 8, 2026. Serious failures were real but uncommon, and a small number of vendors accounted for a large share of them. Some measurements improved over time, although these records cannot prove that public testing caused the improvement.

Public testing makes problems easier to see. It does not make unapproved products safe or replace pharmaceutical manufacturing controls.

Bottom line

Why the 41.6%–71.1% figure is misleading

The Mendias and Awan preprint screened 6,487 Finnrick reports and analyzed 6,441 samples covering fourteen peptides. The authors explicitly acknowledge that their dataset is “not a random or representative sample.” They note that vendors confident in their products may be more likely to submit samples and that poor batches may be underrepresented.

Four limits matter:

• One source: every report came from Finnrick.
• Voluntary submissions: vendors and buyers chose what to send. Quality-focused vendors may test frequently, while complaints may send suspicious samples. Either pattern can skew the result.
• Report count is not market share: the twenty most-tested companies supplied 49.3% of the study’s reports. A heavily tested vendor therefore affects the result more than a lightly tested one, regardless of sales volume.
• One score combined different problems: wrong identity, low purity, underfill and overfill all counted as failures.

We also reproduced the math. The paper used two rule sets: one required 90%–110% of the labeled amount and at least 98% purity; the stricter version required 95%–105% of the labeled amount and at least 99.5% purity. The reported failure rates were 41.6% and 71.1%.

Applying similar rules to 9,212 current Finnrick records produced 42.2% and 70.8%. The calculation works. The sample and the all-in-one failure score are the problem.

What the lab records actually show

Under the paper’s first model, a vial measuring 111% of its label fails. Under the stricter model, a result 6% above label fails. The label “compounding” does not make these intervals universal standards. The paper states that its thresholds “do not correspond to formally codified specifications for any specific product.” It describes the 98% purity cutoff as pragmatic and the 99.5% cutoff as a conservative comparator.

A result containing the identified compound at 111% of label and a result in which the labeled compound was not identified both receive the same binary failure outcome. Both indicate quality-control concerns, but their meaning differs.

Finnrick’s amount field compares the lab result with a submitted batch claim, not necessarily the vial label. We therefore excluded Finnrick from the label-accuracy figures above. Finnrick remains in the identity and purity analysis.

Only 48.23% of the remaining amount tests were within ±10% of the label. Most results outside that range were overfilled, not underfilled. Overfill still signals weak control over how much goes into each vial, but it is not the same as low purity, wrong identity or contamination.

Results differ sharply by peptide

The study combined fourteen peptides. Retatrutide and tirzepatide made up about 64% of its reports, so their results strongly influenced the total. Individual peptides showed very different patterns.

Observed failure profiles differ materially by compound.

¹ Amount columns exclude Finnrick because its comparison is against a batch claim rather than the vial label.

CJC-1295’s below-95% purity rate was about eighteen times retatrutide’s, and its underfill rate was about four times as high. One combined percentage does not describe either peptide well.

Retatrutide generally showed high purity and comparatively little underfill, but frequent overfill. That is still inconsistent manufacturing; it is simply a different problem from wrong identity or low purity.

Did quality improve?

We compared the twelve months ending August 6, 2026 with the previous twelve months.

For non-Finnrick records, results more than 10% below the label fell from 5.47% to 4.02%, while results at least 40% below fell from 0.91% to 0.62%. Among 44 vendors with enough tests in both periods, the underfill rate fell from 4.79% to 2.56%; 24 improved, 15 worsened and five were unchanged.

But not every measure improved. Finnrick’s wrong-identity rate rose from 3.28% to 4.05%, and overfills above 10% rose from 38.73% to 49.00% in the non-Finnrick data.

The mix of testing sources also changed. For CJC-1295, Freedom supplied 11.58% of reports in the earlier period and 70.19% in the later one, while Finnrick’s share fell from 72.59% to 14.47%. Some apparent improvement therefore came from comparing a different mix of reports.

These changes happened while public testing expanded, but correlation is not causation. The records do not show whether testing caused vendors to improve.

A few vendors accounted for many serious failures

When we separated the problems, many serious failures traced back to a small number of vendors.

Each row shows how many vendors accounted for half of that problem:

• 19 of 3,583 vendors (0.53%) accounted for half of the tests where the named compound was not identified.
• 57 of 3,581 vendors (1.59%) accounted for half of the below-95% purity results.
• 69 of 2,840 vendors (2.43%) accounted for half of the tests at least 40% below the label.
• 140 of 2,840 vendors (4.93%) accounted for half of the tests more than 10% below the label.

This is not proof of fraud or intent. It describes only the submitted tests, and 369 of the 372 wrong-identity results came from Finnrick, which records identity outcomes differently from other sources.

What public testing can—and cannot—prove

A lab may report excellent purity for the samples it identified and separately report that it could not identify the named compound in another sample. Those are two different facts. Averaging a wrong-identity result as “0% purity” corrupts both measurements.

Endotoxin is another separate question. The study had endotoxin results for only 243 of 6,441 samples. Thirty-six had measurable endotoxin, or 14.8% of the tested subset—not 14.8% of the whole dataset or market. A measurable result is not automatically above an allowable limit; the FDA’s calculation depends on route and maximum dose.

TitrateLab has 4,846 records containing some endotoxin information, but labs report it in different units and formats. That is not enough for one defensible market-wide contamination rate. Purity testing also cannot substitute for endotoxin or sterility testing.

Approved manufacturers operate under mandatory batch-release, process, stability and recall requirements, including 21 CFR 211.165. A public community test usually tells us what was in one submitted sample. It does not prove that every vial in the batch matches it, and it does not replace sterility testing, stability programs or a recall system.

What this means

The study found a real problem, but its headline number is too broad. It combines unlike failures from a voluntary sample and presents them as one range. The more useful approach is to separate wrong identity, low purity, underfill, overfill and endotoxin evidence.

That separation changes the picture. Serious failures existed, but a small number of vendors accounted for many of them. Several measures improved in the largest testing sources, while some—including Finnrick’s wrong-identity rate and non-Finnrick overfill—worsened.

Public testing cannot make an unapproved market equivalent to approved manufacturing. It can make quality claims easier to check, problems harder to hide and changes easier to measure. TitrateLab’s role is to organize that evidence while keeping the source, date and limitations visible.

Sources

• Mendias and Awan preprint
• TitrateLab: Pharma’s Quality Bar Is Lower Than You Think
• FDA batch-testing requirement: 21 CFR 211.165
• FDA explanation of dose-based endotoxin limits
• FDA guidance on allowable injectable excess volume

How we selected records: Snapshot extracted August 8, 2026 UTC. We included published records tied to a visible, unmerged manufacturer and carrying a verification link or approved public source. A zero purity value was treated as “named compound not identified,” not averaged as a purity result. Label comparisons excluded Finnrick because its amount field is relative to a batch claim. Blends were excluded from peptide-by-peptide comparisons, and vendor counts used merged manufacturer identities.

This is an analysis of published laboratory evidence, not medical advice, a clinical-safety determination or a vendor endorsement.