The test system must be properly controlled to ensure reliable release-testing results. System suitability criteria should be established at the end of AMD, usually by running a set of control points. For each test, results are considered valid if all control points are within established limits. A test system must be able to reproduce measurable results of a homogeneous sample (control) to allow examination of differences among various product batches.

Sample suitability should be established during AMD and should ideally ensure that samples, controls, and standards are prepared identically and run simultaneously. In addition, sample suitability should include a statistical analysis of the number of replicates needed to generate significant release results. Single measurements may be acceptable if the production-process sampling can deliver truly batch-representative samples and the precision of assay repeatability is high compared to the product specifications (and therefore high compared to the batch-to-batch variation upon which these specifications are based).

Sometimes, data transformation (for example, logarithmic conversion) may lead to improved linearity for essentially non-linear assay responses. However, many biological assay response curves are not linear even after transformation.^12,13 These are particularly difficult and should only be handled by experienced statisticians. Just as different test methodologies have different biases, changing statistical models may significantly change the final results. Some models may be inappropriate or may not provide acceptable results over the entire assay range.

AMV Protocol Acceptance Why must we set protocol acceptance criteria? One answer is that we always must validate against limits or specifications, just as release testing results are compared to specifications. Only when all validation results are within established limits is an AMV considered acceptable (assuming that all limits were reasonably set). In addition, a validation document is a contract that not only defines in detail the test parameters but also the exact conditions and contingencies (acceptance criteria). All signing parties agree to the acceptance criteria before the protocol is executed. We must derive appropriate acceptance criteria so validations that should fail do fail and vice versa. Acceptance criteria should demonstrate to regulatory authorities (and ultimately patients) that quality systems and production processes are well designed and maintained to ensure consistent product quality. We also must satisfy acceptance criteria to remain in compliance.¹⁴

Deriving reasonable AMV protocol acceptance criteria is one of the most difficult AMV tasks. In general, the most critical AMV characteristic for the quantitation of excipients, product potency, and impurities in biopharmaceutical manufacturing is intermediate precision. It provides very valuable information concerning the overall contribution of assay variability to the observed process variability. Accuracy is often of lesser concern because we are mostly concerned with batch-to-batch consistency and how well release batches match the purity and impurity levels of the clinical reference batches.^6,7

When setting AMV intermediate precision acceptance criteria, one may only know the observed production-process variability. The best approach is to derive acceptance criteria for intermediate precision from historical process data (observed or expected batch-to-batch variability and content uniformity) and the product specifications (existing or target). The AMV protocol should not permit test method variability to be so high that we could face not being able to release product batches with results that should have fallen within specifications, or vice versa, release product batches with results outside of specifications.¹⁴ Establishing an efficient AMV execution matrix has been discussed elsewhere.^6,7

Conclusions In addition to satisfying regulatory requirements, AMV enables reasonable product specifications to be set and helps separate actual process variability from test method and sampling variability. AMV does not improve a test method but merely provides evidence for its suitability and the confidence in reported results. The quality of the development work determines the quality of the test method and, therefore, the quality of the production process and the product.