Quick answer: A certificate for a finished vial order describes the units that were tested, never the whole batch, because release testing generally consumes the unit. What makes that sample informative is not simply that it was taken at random, but that the sampling plan — how many units, from where in the fill run, against which acceptance criteria — was defined before production and is not open to cherry-picking.
Two suppliers quote the same peptide, in the same fill, with the same documentation package. Both hand you a passing certificate before shipment. On paper you are looking at the same thing. Not quite: one detail rarely appears in a quotation and largely determines what that certificate supports.
This guide is written for procurement and QA teams specifying finished-vial peptide projects. It is a specification and process reference. It does not establish that any material is sterile, endotoxin-controlled, pharmaceutical grade, suitable for injection, or approved for human use, and it does not provide dosing or medical guidance.
1. A certificate describes the units that were tested
Many release tests — vial assay among them — consume the tested unit or leave it unfit to ship. So a certificate for a finished vial order describes a sample of the batch, never the whole of it. Whether that sample supports a conclusion about your goods depends on how it was chosen.
Two arrangements produce very different documents.
Selected from the finished batch under a plan agreed in advance. The run produces more units than the order requires. A defined number of vials is taken according to a sampling plan set before production, tested, and the remainder ships. The report describes units that were interchangeable with your goods until the moment they were drawn.
Prepared for the purpose. A few vials are made separately and carefully, submitted, and pass. The rest of the order is produced alongside. Nothing in the report is false — those units did pass — but it has quietly stopped being about the batch you receive.
Same document type. Different evidentiary value.
2. What makes a sample informative
It is tempting to reduce this to “take them at random”, and randomisation does real work: it removes the supplier’s ability to choose which units are examined. But randomisation alone does not make a sample representative. Several things do, together:
- How many units are tested, defined by the project rather than by habit.
- The sampling plan itself — written before production, not decided afterwards.
- Coverage across the fill run. If the concern is drift during filling, units defined in advance from the beginning, middle and end of the run can be more informative than the same number drawn purely at random.
- The batch definition. What counts as one batch — one bulk, one filling session, one day — determines what the result generalises to.
- Acceptance criteria agreed beforehand, so the result is judged against a standard rather than interpreted after the fact.
The principle worth writing into a specification is therefore not “random” but pre-defined and not open to cherry-picking. Random selection is one way to achieve that; a stratified plan across the fill is another, and sometimes a better one.
As an illustration of the arithmetic — not as a standard — a supplier working honestly on an order of twenty vials might run twenty-five, draw the agreed number of units from the finished twenty-five for testing, and ship twenty. The number that matters is the one your specification defines, not the one in this example.
3. What a batch COA can and cannot support
A passing result gives you evidence about the units that were tested. Batch acceptance is a broader claim, and it requires three things to have been fixed in advance: the sampling plan, the test method, and the acceptance criteria.
This matters most with content uniformity, which buyers reasonably ask about. No one assays every vial, because assay consumes the vial. What can be done is to assay a defined number of units from the batch and report the values against the range agreed in the specification. That is evidence for the sampled units under an agreed plan. A supplier who promises that every individual vial is identical is describing something that was not measured.
The same discipline applies to sterility and bacterial endotoxin testing: a result is a test result on the tested units of that batch. It is not a statement that a process has been validated, and the two should never be presented as the same claim. How each report type is read is covered in our guide to reading a COA, HPLC and MS data.
4. The quantity question
Units consumed by testing, reconstitution checks and retained samples have to be produced in addition to the order rather than taken out of it. If a supplier tests three units from a run of exactly twenty and ships seventeen, the shortfall is yours.
Buyers increasingly write this directly into the purchase order: the quantity on the pro forma invoice is the net delivered quantity. That is a reasonable clause. How many extra units a given test scope consumes is a separate calculation, worked through in our note on why finished vial orders require extra QC vials. Fill amount, packaging and MOQ interact with it, as set out in our lyophilized vial sourcing guide.
5. How to read the paperwork
- Is there a sampling statement? A report that states how many units were tested and how they were selected is a different document from one that only reports numbers.
- Does the lot number on the report match the lot number on the vial label? The cheapest check available, and it catches the most.
- Was the report issued for your production run, or does it predate it?
- Are retained samples held from the same run, for a defined period? Retains are what make a question raised months later answerable.
- Do the quantities reconcile — produced, consumed by testing, delivered?
6. Clauses worth putting in the specification
Quantities stated on the pro forma invoice are net delivered quantities. Units consumed by quality control, reconstitution verification or retained samples shall be produced in addition and shall not be deducted from the delivered quantity.
Units for release testing shall be selected according to a sampling plan defined before production. The certificate shall state the number of units tested, the basis of selection, the test methods and the acceptance criteria applied.
Retained samples shall be held for the agreed retention period and used for investigation under mutually agreed procedures.
A supplier may reasonably discuss cost, the number of units, retention periods and report formats — that is a negotiation, not a warning sign. The real red flag is a supplier who will not explain the sampling arrangement, or who cannot execute a plan they have already agreed to in writing.
7. Where we stand
For projects that include finished-vial release testing, we define the sampling plan, the number of units, the tests, the acceptance criteria and the retain-sample requirement in the specification before production begins, so that none of it is decided after results are known. Sterility or bacterial endotoxin testing is offered where it forms part of the agreed project scope, with the applicable method and reporting basis stated in the batch documentation — as a test result on the tested units, not as a claim about process validation.
Filling is carried out in cleanroom areas at our production site in Suzhou that were tested in October 2025 by a CNAS/CMA-accredited laboratory: 25 rooms including the filling room, covering airborne particles, air changes, pressure differentials and microbial counts, with all items passing. That report is available to qualified buyers under NDA.
If a batch does not test the way it should, everyone is better off finding out from the units drawn for testing than from the units you received. Project scope for filled vials is set out on our private label peptide vial service page.
Where a reconstitution verification is part of the release scope, our peptide reconstitution calculator covers the diluent-volume arithmetic for a given fill and target concentration.
Frequently Asked Questions
Why does it matter which vials were tested?
Because release testing generally consumes the unit, a certificate can only describe a sample. If those units were prepared separately rather than drawn from the batch under a plan set in advance, the result describes them and not your goods.
How many vials should be tested?
There is no universal number. It should be defined by the project alongside the test scope and acceptance criteria, and stated on the certificate together with how the units were selected.
Is random selection always best?
Random selection removes the ability to choose favourable units, which is the main point. Where drift during filling is the concern, a plan that takes units defined in advance from the start, middle and end of the run can be more informative. What matters is that the plan is fixed before production.
Should tested vials come out of my order quantity?
No. Units consumed by QC, reconstitution checks or retains should be produced in addition. Write it into the purchase order rather than leaving it to interpretation.
Can a supplier guarantee every vial contains the same amount?
Not by measurement, since assay consumes the unit. A defined number of units can be assayed and reported against the agreed range — evidence for the sampled units under an agreed plan, not proof about every vial.
What is a retain sample for?
It is the batch’s memory: a unit from the same run, held for an agreed period, so that a later question can be investigated rather than debated.
Does a passing sterility result mean the process is validated?
No. It is a test result on the tested units of that batch. Process validation is a separate exercise with separate evidence, and the two should not be presented as the same thing.