10/11 March 2020
After many postponements, the long awaited Annex 1 draft of the EU GMP Guideline has finally been published on 20 December, 2017. While the chapters on integrity testing and visual inspection of injectables used to be rather short, the draft now contains several updates. Besides the passages on crimping and the handling of stoppered vials before crimping, the currently applicable Annex 1 only contains two sections on the important issues of integrity testing and visual inspection: paragraph 117 calls for 100% integrity testing for containers sealed by fusion, and paragraph 124 describes the requirements for 100% visual control of parenterals in a total of seven lines.
The new document specifically states which containers are to be 100% tested for integrity: glass or plastic ampoules (this also includes BFS containers) as well as Large Volume Parenteral (LVP) or Small Volume Parenteral (SVP) bags closed by fusion. For other containers, a statistically valid sampling plan should be implemented. Visual inspection alone is explicitly ruled out as an acceptable integrity testing method.
Containers sealed under vacuum are to be tested for the maintenance of the vacuum after a previously defined, appropriate period of time and during shelf life. The "and" is unclear in this instance. Also, there is no indication as to the extent of this testing. There is also a new notice stating that transportation validation studies are to be considered for the validation of container closure systems.
The topic visual inspection of parenterals has been expanded to four paragraphs in the draft (8.26 - 8.29). The requirement for 100% visual inspection is unchanged (8.26). There is, however, reference to the use of quality risk management principles for determining the criticality of defects.
Moreover, there has to be a defect library containing all typical defects of a process. Different defect types are to be defined and the defect count in the individual categories is to be monitored by trending. Deviations from the typical level of defects ought to influence the release process for the batch involved. Critical defects shall "not be identified during any subsequent sampling of acceptable containers" after 100% visual control; this would put the previous inspection process in doubt. This is the only reference to possible AQL testing as part of visual control as is required by the American Pharmacopoeia USP (Chapter 790 and 1790), even though there is no explicit mentioning of AQL testing.
The paragraph on manual visual inspection focuses mainly on the training and qualification of staff in visual inspection. The qualification should include worst case conditions like line speed, inspection time or fatigue for example at the end of a shift. As for break times, the draft only says: frequent.
Automated visual inspection received its own paragraph. It demands the validation of the system with known defects, with a sensitivity that is on par with or better than manual inspection. Human inspection therefore remains the gold standard when it comes to validating a system. The inspection system's performance is to be tested before "start up" (meaning, presumably, before the inspection of a batch) and in frequent intervals as well. In this case, the use of the word "and" is not self-explanatory, either.
Comments can be submitted until 20 March 2018. On the EU Commission's website the original text of the Annex 1 draft can be found.