Issue link: https://www.e-digitaleditions.com/i/1394390

www.biopharminternational.com July 2021 BioPharm International eBook 11 sizes are recommended to meet ICH, USP, FDA, and other health author- ity guidance. A determination is one r un of the assay star ting from the unknown to the reportable result and includes sample preparation. SYSTEMS SUITABILITY AND VALIDITY Sy s te m s s u it a bi l it y a nd v a l id it y criteria are important elements of all assays and biological assays. Systems suitabi l it y is t y pica l ly based on a standard or on a well-characterized p o s it i v e c o nt r o l a n d a s s o c i a t e d l i m its. Va l id it y c r iter ia t y pic a l ly include a signal control, variation or repeatabilit y control, detection and elimination of outliers, measures of curve fitting, and confidence interval ranges for reportable values. Systems suitabilit y and validit y criteria are not part of the validation of an assay, however, and must be defined prior to conducting the validation. Finalization of all systems suitability and validity criteria limits may be completed and f inalized post validation. Only valid runs of an assay should be included in the validation report. COEFFICIENT OF VARIATION AND RECOVERY Coefficient of variation (CV)β(stan- dard deviation [SD]/mean * 100)β or relative standard deviation (RSD) a n d r e c o v e r y β (m e a n m e a s u r e d concent rat ion /t heoret ica l concen- tration)*100 βare not recommended measures of assay performance. They are commonly used but do not cor- rectly ref lect the inf luence of assay per for ma nce compa red to c r it ica l qua l it y at t r ibutes (CQA) or mea- sures of productivity. CV and RSD are not useful and may be mislead- ing because they are sca led to the concentration and will likely indicate that high variation at high concentra- tions is acceptable and low variation at low concentrations is unaccept- able. Relative drug product/substance concentrations or impurities are not processed, manufactured, nor tested; thus CV is irrelevant when compared to the product CQA. USP <1033> r e c o m m e n d s a c c e p t a n c e c r it e r i a should be determined relative to the tolerance of product specif ications and not relative to a target dilution nor concentration. If CV or recovery is included in a validation report, it should be used as a report only and should not form part of the assay val- idation acceptance criteria. STUDY DESIGN 1: LINEARITY 5X6 Study design 1 (Table I) consists of five concentrations of a reference standard by six independent determinations for each concentration covering 80% of the lower specification limits (LSLs) and 120% of the upper specif ication limit (USL) for a two-sided product specification limit. To evaluate accu- racy, a reference standard must be cre- ated or purchased. ICH Q2 does not recommend how to set the acceptance criteria; however, USP <1033> states: "When there is an existing product specif ication, acceptance criteria can be justified on the basis of the risk that measurements may fall outside of the product specification" (3). For a two-sided specif ication, the tolerance is used, and, for a one-sided speci f icat ion, t he ma rg in may be used (process average minus LSL) or (USL minus process average), and the process average may be used in the denominator when there is no speci- fication stated. The number 5.15 is a commonly used risk factor to include 99% of the analytical error. Assay precision includes repeat- ability (intra-assay error), intermedi- ate precision (intra- and inter-assay error), and reproducibility (inter-lab). Reproducibilit y is the inter-lab or locational error and is not part of a normal method validation; it is typ- ica l ly eva luated during technolog y transfer if needed. The study design is a main effects- only design of experiments with 30 runs. Factors to be included are the five con- centrations, the name of the analyst, the day the experiments were run, and the instrument used at two or more levels. Software programs like SAS or JMP are useful in designing the study. Table II is Biopharmaceutical Analysis Method Validation Reportable result Method of analysis Acceptance criterion Accuracy in units and 95% CI Mean Difference from standard ((Difference)/(USL-LSL))*100 <10% of tolerance for an analytical method <20% of tolerance for a bioassay Repeatability in units and 95% CI Variance components (within) POV or REML is the recommended method of calculating variance components. ((SD/sqrt(n))*5.15)/(USL-LSL)*100 <25% of tolerance for an analytical method <60% of tolerance for an analytical method Intermediate precision in units and 95% CI Variance components (within and between variation) POV or RMEL is the recommended method of analysis ((SD/sqrt(n))*5.15)/(USL-LSL)*100 <30% of tolerance for an analytical method <70% of tolerance for a bioassay Linearity, lowest concentration, and highest concentration where the response is linear Studentized residuals of the linear fit of the measured concentration Concentration where the 95% CI of the quadratic fit of the studentized residuals crosses +- 1.96 Range of the concentration where the assay is accurate, repeatable, and linear Comparison of concentration and all acceptance criteria Range where all of the above concentrations are acceptable Table I. Study design 1: Linearity 5x6. CI is confidence inter val. POV is partition of variation. REML is restricted expected maximum likelihood. SD is standard deviation.

- BPeBook0721_Ads
- BPebook0721_001_Cover
- BPebook0721_002_TOCMasthead
- BPeBook0721_Ads (1)
- BPeBook0721_004-008_CoverStory_BioPharma Analysis_Mirasol
- BPeBook0721_Ads (2)
- BPebook0721_010-014_Analytics_BestPractices_Little
- BPebook0721_015-019_DataIntegrity_IQConsortium
- BPebook0721_020-024_HCPControlGeneTherapy_CharlesRiver
- BPeBook0721_Ads (3)
- BPebook0721_020-024_HCPControlGeneTherapy_CharlesRiver (1)
- BPeBook0721_Ads (4)
- BPebook0721_026-030_HostCellProteinAnalysis_Bruker
- BPebook0721_031-034_IRSpectroscopy_RedShiftBio
- BPebook0721_035-037_CleaningValidation_Shanley
- BPeBook0721_Ads (5)