ICH Guidelines Show
This article focuses on ICH Q6A for testing new drugs—Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and New Drug Products: Chemical Substances—and considers pertinent sections, in order of their appearance within the guidance document. Where necessary, I have provided my comments after each excerpt. In some specific cases, the words of the Guidance are highlighted and underlined, followed immediately after by my related comments.
2.1 Periodic or Skip Testing
Periodic or skip testing is the performance of specified tests at release on pre-selected batches and/or at predetermined intervals, rather than on a batch-to-batch basis with the understanding that those batches not being tested still must meet all acceptance criteria established for that product. This represents a less than full schedule of testing and should therefore be justified and presented to and approved by the regulatory authority prior to implementation.
Author Comment: In a GMP-controlled process, the burden of testing every lot is time-consuming and expensive. Since end-product testing rarely identified process problems, 100% final testing of batches was deemed unnecessary. Under QbD/PAT, where the criterion of a set “design space” is to be met through feedback and feed-forward (continuous) testing, skip testing is an anathema. Perhaps this may be addressed in later attachments.
2.2 Release vs. Shelf-life Acceptance Criteria
The concept of different acceptance criteria for release vs. shelf-life specifications applies to drug products only; it pertains to the establishment of more restrictive criteria for the release of a drug product than are applied to the shelf-life.
Author Comment: There is much to be learned about an “ideal” from following stability samples more, not less closely. If a PAT paradigm is used to produce a product, wherein thousands of units are examined, it does not add to knowledge of the process to only examine six units for dissolution and (up to) ten for content uniformity. While it is necessary to show that a product remains potent and safe until the expiry date, merely taking small, non-statistical samples for stability tests reveals little about lot-to-lot variations over time.
Perhaps a monitoring system, utilizing techniques similar to process controls, could be in place to ascertain what percentage of units “change” over time. These “changes” could lead to testing of those units and, depending on how many outliers occur per lot, process conditions could be correlated (using multivariate algorithms) to help determine a design space which not only makes superior product for release, but has a longer shelf life. [This would be a financial benefit, since, with a longer shelf-life, fewer lots would be recalled or returned.]
2.3 In-process Tests
In-process tests, as presented in this guideline, are tests which may be performed during the manufacture of either the drug substance or drug product, rather than as part of the formal battery of tests which are conducted prior to release. In-process tests which are only used for the purpose of adjusting process parameters within an operating range, e.g., hardness and friability of tablet cores which will be coated and individual tablet weights, are not included in the specification.
Certain tests conducted during the manufacturing process, where the acceptance criterion is identical to or tighter than the release requirement, (e.g., pH of a solution) may be sufficient to satisfy specification requirements when the test is included in the specification. However, this approach should be validated to show that test results or product performance characteristics do not change from the in-process stage to finished product.
Author Comment: While this Guidance is only ten years old, the tests cited are ancient: hardness and friability are seldom telling about modern products. Certainly, they cannot be considered PAT/QbD tests of the 21st century. The reference to tests that are more stringent than release tests is a foreshadowing of “real-time” release testing…true QbD/PAT!
2.4 Design and Development Considerations
The experience and data accumulated during the development of a new drug substance or product should form the basis for the setting of specifications. [Author Comment: The learning curve may flatten as the product is made, but changes will become apparent as experience mounts. The “data accumulated” after product development should also be utilized to make a better product.] It may be possible to propose excluding or replacing certain tests on this basis. Some examples are:
- microbiological testing for drug substances and solid dosage forms which have been shown during development not to support microbial viability or growth (see Decision Trees #6 and #8); [Author Comment: Sounds logical, but a RMM could be used occasionally to be certain.]
- extractables from product containers where it has been reproducibly shown that either no extractables are found in the drug product or the levels meet accepted standards for safety; [Author Comment: The burden then may be shifted to a packaging control assay or vendor validation of the packaging supplier. Assuming that polymers are unchanging is very far from the reality of packaging.]
- particle size testing may fall into this category, may be performed as an in-process test, or may be performed as a release test, depending on its relevance to product performance; [Author Comment: Particle size has been demonstrated to be quite critical, especially in delayed/sustained release dosage forms. It should be tested prior to use on every lot! In addition, as we see in QbD, physical parameters of excipients should also be controlled.]
- dissolution testing for immediate release solid oral drug products made from highly water soluble drug substances may be replaced by disintegration testing, if these products have been demonstrated during development to have consistently rapid drug release characteristics (see Decision Trees #7(1) through #7(2)). [Author Comment: In the case of immediate release, this may be acceptable, but smacks of 1950s methodology.]
2.5 Limited Data Available at Filing