UPDATE: THE VISUAL INSPECTION OF PARENTERALS
There is some movement in the area of the 100% visual inspection of injectables. The EU GMP Guide Annex 1 is in revision, and the USP published the long-awaited chapter <1790> as a first draft.
Due to the intensive discussions on this subject, the ECA Foundation had decided already in 2013 to establish a new Working Group on the visual inspection of parenterals - which started in December the same year. Since then the founding members from industry and authorities have been working on a document about the requirements and the implementation of the 100% visual inspection. This document is rather considered as a best practice paper than a guide, though. It comprises both the manual as well as the fully automatic inspection, and it also deals with the issues of training, qualification, validation and batch evaluation with regard to the data of the visual inspection. The document was first presented at the conference "Particles in Parenterals" in Copenhagen in September 2014. As expected, the pragmatic and practical document received a lot of attention. In late 2014/early 2015 the Working Group was transformed to an Interest Group, allowing interested persons to become member, to exchange experience in the discussion forum and to obtain the best practice paper. Within a few weeks, the number of members grew to nearly 600. At the beginning of the year the document was further submitted to various official bodies (among them FDA, USP, the EDQM and the EMA) for comments. The first contacts with the FDA showed a positive feedback, but also the need for discussion - specifically with regard to re-inspection of a batch when the AQL was not passed and the dealing with a gray fraction of the inspected objects. This means containers that could not definitely or clearly be inspected in the fully automated inspection. The discussions will continue with the goal to reach at least a partial harmonisation within industry and the various authorities. The results will be incorporated in the revision of the best practice paper, and the revised version will then be presented at the next group meeting in Berlin in September 2015.
End of January 2015 - and thus almost a year later than planned - the long-awaited USP Chapter <1790> Visual inspection of injectable products was issued in the Pharmacopeial Forum 41(1) for comments. Chapter <790>, which had come out the year before, already answered the question "what is practically free of particles" for the first time. The industry has been uncertain for a long time how this requirement has to be interpreted and fulfilled. For one, it is clear that it is impossible to produce and to fill a batch of sterile product 100% free of particles. On the other hand, in the 100% visual inspection of the batch there are supposed to be no or "almost no" particles. For that reason the USP has established the concept of AQL testing as a quality control step after the actual 100% inspection. If the required AQL limits are met, the batch is considered 'practically particle-free' and thus compliant with the requirements. This idea is continued in the new chapter <1790>. However, the chapter reads more like an overview article on the subject of visual inspection than a pharmacopoeia monograph. In contrast to the ECA document the new chapter <1790> covers more the manual visual inspection and less the increasingly important fully automatic inspection - at least in the current draft version. It aims in particular at the control of particles, but also references other defects such as cracks in primary containers or ill-fitting stoppers. There are also general statements about patient risk due to particulate contamination, which is determined by its size, type and origin and the state or age of the patient. The prevention of particles is also covered, for example by avoiding glass delamination through selecting appropriate formulations in the development. Another issue covered is the removal of particles, for example through the washing of primary containers and through associated particle depletion studies. But, obviously, the actual visual inspection is treated as well. The new USP chapter comprises some practical advice. For instance, when assessing poorly visible product containers (brown glass or plastic containers) it recommends to increase the lighting strength up to 10,000 lux and to examine the containers preferably from the rear. Also interesting is the note with regard to the re-inspection of the batch when it doesn't pass the AQL. The USP holds a "retesting" admissible, even a multiple repeat of the check, if the procedure for non-compliance with the limits is previously described. This also includes, among other things, the number of allowed repetitions and inspection parameters to be selected. As is known, for the FDA re-testing is rather critical, and the authority warns of "testing into compliance". Thus it will be interesting to read the wording in the final version of the new chapter. A maximum number of retries is not mentioned here, but a maximum of two tests are currently common industry practice, which is also described in the ECA paper.
As already described in USP chapter <790> an AQL testing is supposed to be part of the evaluation of a batch. For this purpose samples from the good portion of the tested batch are drawn and tested again according to the defined sampling plan to assess the quality of both the process and the 100% inspection. The AQL limits exemplarily mentioned in chapter <1790> are slightly more stringent than the values specified for the visual inspection in the first version of the ECA best practice paper. The section on the visual inspection of lyophilised products is quite helpful, making the requirements of chapter <790> easier to understand. For example, lyophilised products are also supposed to be 100% visually inspected after closing. However, as the lyo cake could cover particles, a small number of samples has to be reconstituted in addition. These samples have then to be tested for particles again, analogous to the routine-to 100% inspection (sampling for example according to S-3 and S-4 in ANSI/AQS Z1.4 or to DIN ISO 2859).
It is also interesting that the new USP chapter implies in several sections, that particles are not critical defects - the opinions on this issue have been varying for years, though. For instance, the USP states that one of the requirements for the qualification of employees is the 100% detection of critical defects. As the particle detection for both manual and automatic inspection is probabilistic, particles cannot be critical defects. The USP here also calls for a limit for the false reject rate and specifies 5% as an example. For the requalification of employees the USP indicates at least an annual cycle.
This year the basic guideline for the area of sterile medicinal products will be revised. In early February the EMA announced the revision of EU GMP Guide Annex 1. Technological developments as well as the adaptation to the ICH Guidelines Q9 and Q10 are the reasons for the review. The visual inspection is covered here as well, however in little detail. In the currently valid version from 2009 item 124 in the chapter on Finishing of sterile products merely describes the most rudimentary requirements for premises, eye examinations and breaks in the manual inspection. It also defines the need for the validation with regard to non-manual methods as well as their regular review - but without going into details. It is quite possible that further and detailed requirements with regard to the visual inspection will be included in a new version as well, and that it will also discuss a 100% control of the container-/ closure integrity.
For more information on the subject "Visual Inspection" please also see the website of the ECA Foundation Working Group at www.visual-inspection.org.
Author:
Dr Robert Eicher
CONCEPT HEIDELBERG
