Q&As Cleaning Validation - Part I

In September the ECA offered the first Cleaning Validation Online Training Course. Many questions were asked to the speaker. Below you will find a number of questions which have been answered by Robert Schwarz from the FH Campus in Vienna.

Selection of Q&As during Live Online Training Cleaning Validation. All answers reflect the opinion of the speaker based on his experience.

1. Question: Regarding visual inspection of surfaces in Annex 15: Is during operator qualification a training on the relevant SOP sufficient or should a training, comparable to those for visual inspection operators, be established (incl. periodic eye sight test)?
Answer: A solely theoretical SOP-training is not sufficient. Practical training with surface samples spiked with different concentration of the respective product needs to be performed. This needs to be performed by collecting data to demonstrate down to which level of residues visual cleanliness can be detected.

2. Question: Are direct sampling methods necessary for transfer lines or other inaccessible equipment parts? Those parts can neither be tested visually nor via swab testing. Would final rinse testing be sufficient?
Answer: This clearly depends on factors like pipe length, flow conditions, percentage of surface area that can't be inspected and should be addressed in a risk assessment. Additional measurements like endoscopic inspection as part of maintenance could also be beneficial. Taking the solubility of residues into consideration final rinse sampling could be sufficient.

3. Question: Are PDE values applicable in products that are only applied once or twice (e.g. vaccines)?
Answer: Yes, PDE values derived from HBEL are also applicable. They need to be established for every medicinal product. Only if inactivation during the cleaning process could be shown a risk based strategy could be developed where applying the PDE is not needed:
"In view of this, the determination of health-based exposure limits using PDE limits of the active and intact product may not be required."

4. Question: If you introduce a new product in a multi-purpose facility and identify this product as worst case; do you have to revalidate the cleaning process for existing equipment?
Answer: Yes, a revalidation is necessary. The extent needs to be defined in a risk assessment.

5. Question: Is there a reason why a CHT and Sterile Hold Time (SHT) could not be shown within one single validation protocol? In the way that sampling will just be done at the end of SHT?
Answer: Yes, testing for CHT is first performed before sterilization having data within the accepted limit for bioburden (rinsing, swabbing or contact plate sampling) prior sterilization. This also is the minimum bioburden to inactivate via sterilization (excl. buffer ? SAL) - basis for sterilization validation.
SHT has nothing to do with cleaning validation. It is linked to sterilization validation, aseptic processing and aseptic process simulation (media fill). Validity can either be shown via sampling or can be covered, based on a profound risk assessment, in an aseptic process simulation.
If you sample only after SHT you have no data for bioburden prior sterilization and therefore no data for CHT.
The only scenario I could imagine to skip CHT is a CIP/SIP process where sterilization is automatically directly started after the cleaning process. Therefore there is no CHT, so only SHT needs to be validated. But at least you need data for endotoxins after cleaning.

6. Question: What about sampling of wire-mesh filters of fabrics filters?
Answer: According to my experience sampling wire mesh filters of fabrics filters is only possible via rinse sampling.

7. Question: Can the worst-case substance regarding to cleaning also be determined by only performing theoretical calculations and w/o performing any practical small scale studies in the lab?
Answer: Based on my experience with several research projects predicting cleanability of a product or component solely based on theoretical data like physico-chemical parameters and model assumptions are insufficient and lead to a "trial &error"-like cleaning validation.