The GMP requirements for equipment and facilities for the production of medicinal products are constantly increasing. Many areas are affected by this: plant qualification, calibration, hygienic design, ventilation and water systems or supply systems, ageing plants. The technologies are also increasingly intervening in the manufacturing and control processes. The technical department and suppliers deal with this on a daily basis. What questions arise from this? Who is in what responsibility? Does technology need a greater say - also in regulatory terms?
The lively discussion among the DIALOGUE participants showed that this topic does not only concern the questioner. The results can be summarised as follows:
Maintenance, calibration, deviation and change control ensure continuous monitoring of the system status. The prerequisite, however, is that the data is documented, combined and evaluated. This assessment is also referred to as a "review". The results of the review should be summarised in writing in a separate document.
The review was introduced to avoid the need for physical testing in the sense of re-qualification. Annex 15 states in 4.1: "Equipment, facilities, supplies and systems should be evaluated with sufficient frequency to confirm that they remain in a controlled condition". A pure document review as an alternative to requalification is therefore possible, but requires that meaningful data be regularly collected and evaluated to the extent necessary.
But when is a re-qualification with physical tests necessary at all? This should be set out in a risk-based SOP for re-qualification. Possible reasons for a physical re-qualification could be critical changes, frequent deviations or technical faults. In general, the more that is continuously monitored and documented, the longer the period of time for the review can be chosen and the more superfluous a physical re-qualification becomes. The overall concept of review/re-qualification should be risk-based.
There was a clear recommendation from the experts on this question: Determine in the design qualification (DQ) on a risk-based basis which documents are required and which information must at least be contained in them. You cannot expect documents from different manufacturers to be all the same, but you can (and should) clarify in advance what is required and available. In this way, it is already possible to identify in the DQ phase where additional documents may need to be created or tests scheduled. If the documentation then later complies with the previously defined requirements (keyword specification sheet!), there is no reason for complaint.
The further discussion showed that in practice the subject of FAT/SAT is handled quite differently. Some participants were of the opinion that FAT/SAT was a matter of engineering and that FAT/SAT should not used for “pre-qualification“. After all, the purpose of the FAT is to discover and correct errors on site. It is sufficient to check afterwards which tests and documents can be used for the qualification.
Others countered that it makes more sense and is also quite practicable if QS and engineering work together in advance to identify GMP-relevant tests and to define requirements for GMP-compliant documentation. FAT tests could thus be used for qualification purposes, saving time and money.
It is certainly not possible to define a generally valid procedure here, because each qualification is an individual project with regard to the respective plant and the general conditions.
The GMP regulations do not provide a definition for this. According to Rainer Gnibl, this should be seen as an opportunity to determine for yourself what is meant by "like for like" - risk-based, of course. For example, it must be considered whether the part to be replaced is in contact with the product or not, whether it is a critical process step or not, etc. It is important to define the criteria for the risk assessment in advance and to use this assessment scheme for the application. Rainer Gnibl concluded by pointing out that the topic of "like-for-like qualification" is addressed for the first time in the new aide-mémoire on qualification and validation written by the German GMP inspectorates.
Rainer Gnibl answered this question with an example from his inspection practice: the monitoring system had not "noticed" that the server of the building management system (BMS) had failed because there was no alarm. What does that mean in concrete terms? If a monitoring system is connected to a BMS, which is frequently seen in practice, then at least the interface must be qualified and it must be ensured that an alarm is passed on to the monitoring system in the event of failure of the BMS. For the BMS, a technical acceptance is usually sufficient.
What can be done if an aging plant - in this case a plant for the production of purified water - cannot be completely replaced for cost reasons, but is gradually modernised? Is the resulting "material mix" still GMP compliant?
Rainer Gnibl's answer was as clear as it was pragmatic: as long as the water produced by the plant was of the required quality, there was no reason to complain about the plant. Especially as there are no concrete requirements for materials for water installations in the GMP regulations. Ruven Brandes added that the replacement of plant components should in principle be covered by a qualification.
Rainer Gnibl reported on inspection situations in which packaging lines and filling stations were obviously standing longer than they were running - without any indication of a plant stop or breakdown in the batch documentation. Nor was anything documented about the whereabouts of the products that were on the equipment during the downtime. The fact that these incidents took place in faraway India is of little consolation in view of the fact that such batches are later released on the basis of a flawless batch record and are possibly also sold by us. However, it allows many a complaint or recall action to appear in a different light.
He therefore urged the participants not to sweep such incidents under the carpet and to document both plant stops and breakdowns in the batch report. This naturally presupposes that all persons involved in the situation are appropriately trained and have the necessary problem awareness. He also recommended defining in an SOP what is meant by a plant stop or a breakdown, from when a plant shutdown must be documented and how to handle the product that is on the plant during the shutdown.
Particularly in the case of complicated technical systems, a lack of technical know-how on the part of employees is often evident. A widespread shortcoming is also incomplete documentation. In the laboratory area, expensive maintenance contracts with the reference to the execution of system tests are often waived for cost reasons. If the systems are then operated without maintenance until faults occur frequently or the system fails completely, the consequential costs are often higher than if a maintenance contract had been concluded. If no maintenance contract is concluded, the company must prove that it carries out maintenance work itself.
Device lists for qualification in the validation master plan, device lists for calibration, device lists for maintenance - the parallel maintenance of different lists is time-consuming and error-prone. But how can all relevant information be brought together? Ruven Brandes cited SAP as one option and pointed to a wide range of software solutions that can be customised. If such solutions are used exclusively online, however, a computer system validation must be carried out. If the system is documented on paper, this hurdle can be avoided. However, Excel is also suitable for such a central data collection. For this purpose, however, it is necessary to deal with the required functions.
Risk-based inspection planning first considers the type and number of products and processes. Are there cross-contamination risks? Are sterile, aseptic or non-sterile products manufactured? The inspection history also plays a role. During the inspection itself, attention is focused on situations where problems have occurred, such as OOS results, deviations or complaints. The inspection then refers to the documentation of the problem and the underlying systems and processes.
Rainer Gnibl was not aware of any registration from his area of responsibility. However, he reported that the cold WFI production process has been widely used in the USA for several years and that the plants are running stable. In Germany, he expects manufacturers to switch to cold WFI production only for new buildings. Manufacturers whose distilleries are running smoothly see no reason to change the technology. After all, distillation is a very robust and virtually autosterile process.
Ruven Brandes reported on a pilot project in the German federal state Lower Saxony. There, manufacturers ("I can do it!"), operators ("I dare!") and authorities ("I'll go with you!") have joined forces. Despite this positive approach, there were considerable difficulties, but the plant is now in operation. Ruven Brandes stressed that cold WFI production was more cost-effective. Rainer Gnibl pointed out that this is in contrast to the increased effort for monitoring, which must also be taken into account in the total cost calculation.
Here’s what you’ll learn: