Last update: 25-Oct-2024.
Peptide users should consider the effect of sample size on understanding overall safety profile of the product lot, especially as the community begins to pool funds to reduce the cost of testing to a nominal amount, often less than the cost of the vials themselves.
Manufacturing Defects Happen Randomly
All manufactured products vary throughout production. Companies worldwide spend billions of dollars to try and control this variation to prevent product failures from reaching customers. For peptides specifically, there are errors in weighing of materials, errors in mixing the materials, and errors in loading the final product into vials, and errors during labeling and packaging.
As an example of this, peptides vials are produced by mixing the peptide and excipients in a solution, and then workers pipette the solution into vials before they are freeze-dried (lypholized). As you can probably imagine, this process may be more or less accurate – the workers pipettes vary and may fail, they may be set incorrectly, they may be not paying attention and double-pipette into vials, etc.
Even if the process is automated, there can be errors. The machine could be set to the incorrect volume, a nozzle could be clogged, or the cleaning solution wasn’t fully purged with the liquid from the batch before running the vials resulting in the initial vials being loaded with the cleaning solution and the remaining vials containing the proper amount of peptide.
One issue that came up recently is tirzepatide contaminated with semaglutide. One potential way for this to happen is due to a lack of cleaning validation to prevent cross-contamination between production lots. If the semaglutide solution is in the bottom of the vessel during filling, only the first section of the lot will be contaminated until the leftover semaglutide solution is fully replaced by the tirzepatide solution.
Assuming all of this is done perfectly, line clearance for labeling can be an issue. If labels are left in the production area from the last lot, they could inadvertently be used on a portion of the new lot, which may be an entirely different peptide.
Once the vials are produced and warehoused, they can be mixed up by lack of controls. If vials fall out of an labeled container and are unlabeled themselves, a worker may put them back in the wrong box, mixing them with a different peptide. If they share the same size, cap color, and powder color, it is unlikely that it will be detected during packaging and shipment.
How likely is it to catch any of these errors testing a single vial out of 10,000 vials produced in a lot?
Statistical Sampling
For this reason, the manufacturing industry uses statistical sampling techniques to make sure they sample and test enough product from the lot to understand the quality of the lot and catch minor and major production errors before they release it.
A quick history lesson: In the early 1960s, the military got tired of buying lots of product that failed, so they enlisted my personal heroes Shewhart, Romig, and Dodge to develop MIL-STD-105, which started the basis of Acceptable Quality Limit (AQL) sampling or acceptance sampling.
Now these standards have been replaced by standards like ANSI Z1.4 or ISO 2859 to provide guidance to industry in how to sample products to determine lot quality using the power of statistics.
These standards tell manufacturers how many vials to sample and test based on the size of the lot and the acceptable number of failures. In practice, a larger lot requires more samples, and a 1% AQL (allowable failure rate) would allow less failed tests than a 4% AQL.
Done correctly, the manufacturer then generates a “representative sample” of the lot by selecting samples from the beginning, middle, and end of the lot. The more representative the sample, the better the power of the testing. Representative sampling ensures that all the samples aren’t taken from a section of the lot that meets specification while there were major issues in another portion lot.
These samples are then taken to a lab and tested, and as long as there are fewer failures than allowed the lot is deemed acceptable for release.
Risk Mitigation
AQL testing sample sizes may range from 10 – 500 samples, but just testing three vials from the beginning, middle, and end of the lot, or across boxes received from the manufacturer is better risk mitigation than testing a single vial.
While this testing may be more expensive, it is best practice in industry because it mitigates the risk of testing the one good vial in the lot while many of the other vials in the lot may be sub-potent, hyper-potent, cross-contaminated, non-sterile, etc.
More Work to Do
TrustPointe is dedicated to providing education that drives improvements in safety and reduce risks for this community. There is a lot more work to do, and we look forward to partnering with the community as we move the industry ahead.