This is a ready-to-use matrix justifying a bracketed (matrix) filter validation approach across a portfolio of related formulations sharing a platform, rather than a fresh product-specific bacterial retention study for every construct. Replace every <<FILL: ...>> placeholder with your own specifics and route it through your normal document control, review, and approval. A worked filled specimen follows. Verify each cited reference against the current source before you rely on it.
Document control header
| Field | Entry |
|---|---|
| Document title | Filter Bracketing Justification for <<FILL: formulation platform>> |
| Document number | <<FILL: e.g. BR-2026-01>> |
| Version | <<FILL>> |
| Effective date | <<FILL>> |
| Platform / product family | <<FILL>> |
| Document owner | <<FILL: e.g. Validation Lead>> |
1. Purpose
This matrix documents why a single bacterial retention study, run against the worst-case member of a formulation group, is scientifically justified to represent every named member of the group, rather than requiring an independent study for each.
2. Scope
| Dimension | Decision |
|---|---|
| Formulation platform / family | <<FILL>> |
| Constructs covered by this bracket | <<FILL: list>> |
| Filter model(s) covered | <<FILL>> |
| Out of scope | <<FILL: e.g. any construct outside the bracketed parameter ranges below>> |
3. Bracketed parameters
| Parameter | Range across the group | Worst-case member and value |
|---|---|---|
| pH | <<FILL>> | <<FILL>> |
| Ionic strength | <<FILL>> | <<FILL>> |
| Viscosity | <<FILL>> | <<FILL>> |
| Surfactant content | <<FILL>> | <<FILL>> |
| Bacteriostasis/bactericidal potential | <<FILL>> | <<FILL>> |
The worst-case member for each parameter must genuinely sit at the demanding end of that parameter’s range across the group; a member chosen for convenience rather than demonstrated severity is not a valid bracket.
4. Worst-case selection rationale
State explicitly why the member(s) selected for testing represent the demanding end of every bracketed parameter, not just one. Where no single member is worst-case on every parameter, state whether a combined worst-case (hybrid) test article was used and how it was constructed.
5. Retention study reference
| Field | Entry |
|---|---|
| Study protocol / report reference | <<FILL>> |
| Test article used | <<FILL: which member, or hybrid, and why>> |
| Result | <<FILL: LRV achieved>> |
6. Equivalence table: bracket coverage by member
| Construct | pH | Ionic strength | Viscosity | Surfactant | Within bracket? | Covered by this study? |
|---|---|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | Yes/No | Yes/No |
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | Yes/No | Yes/No |
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | Yes/No | Yes/No |
Any construct marked “No” on any parameter falls outside this bracket and needs its own product-specific study or a new bracket built to cover it.
7. Conclusion
State which constructs are covered by this bracket and which are not, and the process for adding a new construct to the platform in the future, whether it can be evaluated against this existing bracket or requires a new one.
8. References
ASTM F838, Standard Test Method for Determining Bacterial Retention of Membrane Filters Utilized for Liquid Filtration. PDA Technical Report No. 26, Sterilizing Filtration of Liquids. ICH Q9(R1), Quality Risk Management.
Confirm the current version of each reference before issue.
9. Revision history
| Version | Date | Author | Summary of change |
|---|---|---|---|
<<FILL: 1.0>> | <<FILL>> | <<FILL>> | Initial issue. |
10. Approvals
| Role | Name | Signature | Date |
|---|---|---|---|
| Author (Validation lead) | <<FILL>> | ||
| Reviewer (Formulation/Process Dev) | <<FILL>> | ||
| Approver (QA) | <<FILL>> |
Filled specimen
Extract for a portfolio of four antisense oligonucleotide constructs sharing one buffered aqueous formulation platform. Company, product, and numbers are illustrative.
| Parameter | Range across the group | Worst-case member and value |
|---|---|---|
| pH | 6.8 to 7.6 | ATX-315, pH 6.8 (lowest, closest to the range where filter material compatibility narrows) |
| Ionic strength | 100 to 180 mM | ATX-322, 180 mM (highest, most demanding for adsorptive retention effects) |
| Viscosity | 1.1 to 1.4 cP | ATX-322, 1.4 cP (highest) |
| Bacteriostasis | None of the four constructs showed inhibition in screening | N/A, not a differentiator |
Worst-case selection rationale: no single construct is worst-case on every parameter, so a hybrid test article was formulated combining ATX-315’s pH with ATX-322’s ionic strength and viscosity, representing a condition more demanding than any single marketed construct on every bracketed axis simultaneously.
Retention study RA-BR-2026-004 achieved LRV greater than or equal to 7 with the hybrid test article at the full 10^7 CFU/cm2 challenge level.
| Construct | pH | Ionic strength | Viscosity | Within bracket? | Covered? |
|---|---|---|---|---|---|
| ATX-315 | 6.8 | 140 mM | 1.2 cP | Yes | Yes |
| ATX-318 | 7.2 | 160 mM | 1.3 cP | Yes | Yes |
| ATX-322 | 7.6 | 180 mM | 1.4 cP | Yes | Yes |
| ATX-330 (new, in development) | 6.5 | 200 mM | 1.5 cP | No (pH and ionic strength both outside the tested range) | No |
Conclusion: the bracket covers ATX-315, ATX-318, and ATX-322. ATX-330 falls outside the bracketed ranges on two parameters and requires its own product-specific study or a newly defined bracket before it can rely on this justification.
Common inspection findings this matrix prevents
- A bracketing claim naming “similar formulations” with no table showing the actual parameters compared.
- A worst-case member selected on one parameter but not verified as worst-case, or a defensible hybrid, across all bracketed parameters.
- A new construct added to a product line and assumed covered by an existing bracket without checking it against the bracket’s actual ranges.
- No defined process for what happens when a new construct falls outside the existing bracket.
- Bacteriostasis screening omitted from the bracketing comparison even though it can differ across formulation variants.
How to adapt this matrix
- Set your document number and platform identifier in the header.
- Populate section 3 with real measured or specified values for every construct in the group, not estimates.
- Justify the worst-case selection in section 4 explicitly; use a hybrid test article where no single member is worst-case on every parameter.
- Check every construct against the bracket in section 6 before relying on this justification, including any new construct added later.
- Confirm every reference in section 8 against the current published version before issue.