Last month we focused on limits for indirect product contact surfaces (IPCSs). This month we will focus on studies that could be performed as part of a risk assessment related to cleaning validation for such surfaces. In particular, the examples I use will involve the interiors of isolators, although the principles of the studies should be readily translated to other IPCSs, such as lyophilizers and air supply/exhaust vents for process equipment. I have divided the assessment into three parts: (a) a baseline of clean equipment, (b) data on residues after processing, and (c) data after “routine” cleaning.
The first step is to start off with “absolute” clean equipment for the internal surfaces of the isolator. Unless this is done, any conclusions about data from the next step would be tenuous. I realize there is no such thing as equipment being absolutely clean. The point is to clean it (perhaps repeatedly) and test surfaces so that you are confident the results are as clean as practical. I would prefer to see results below the LOD (or non-detectable) for results for actives measured by HPLC. If you measure cleanliness using TOC, you would also prefer to have non-detectable results (just make sure you are subtracting out a blank and that you are using a reasonable estimate of what is truly detectable). My actual preference in such a situation is use a specific method and TOC, realizing if this is a biotech protein TOC (or perhaps a total protein method) may be your only practical option. A visual assessment of the equipment should also be done. Once you are confident that the equipment is clean, then you can start with the next step.
This second step involves processing a product. The product selected should be representative of the other products processed, and hopefully represents a possible worst case. By worst case, I mean things like having a dried residue which is “lighter” and more likely to become airborne, or a liquid product with lower viscosity (more likely to splash), or a liquid product with higher solid content (more likely to dry out and become airborne). This list of what might make a product worst case is not exhaustive.
Once you have selected a product, you begin processing that product for the longest processing time (either as amount of processed product or elapsed time) that you ordinarily encounter. The rationale here is that the longer the processing time or the greater there amount of product processed, the more likely you are to leave residue on the various internal surfaces of the isolator. Realize further that you should also consider interventions that might occur during process. This might be similar to interventions that normally occur or that are planned in an operation such as a media fill.
At the end of processing, now is the time to sample and analyze the interior surfaces of the isolator. That “sampling” and “analysis” should start with a visual assessment. As you move to swab sampling, the sampling locations are not necessarily the same worst-case locations you might select for sampling after cleaning of process equipment.
The reason is that in this step you should be looking to see where processed product can transfer to the interior surfaces of the isolator. For example, this might include anywhere product might spill or splash, and anywhere product might be transferred by airborne currents. However, in addition you should also sample locations that are worst case for cleaning, because it is possible that product not present in these locations before cleaning may be transferred to these locations during the cleaning process itself.
Collecting this data not only provide a baseline for comparison to the data collected in the third step, but it also provides data that could enable you to make changes in the product processing to reduce transfer of product to isolator surfaces. Finally, such data may also help in making sure that your cleaning processes are not only effective, but also that the subsequent cleaning process can be done in a manner that minimizes operator exposure to product residues.
The third step starts with cleaning the surfaces with the established (or proposed) cleaning SOP by a trained operator. I recommend that this cleaning be observed by an independent person. The purpose of that independent assessment is to confirm that the cleaning process was carried out correctly (see the Cleaning Memo of January 2019). Remember that your objective is not to find out what happens if cleaning is done incorrectly, but rather to determine cleanliness after correct execution of the cleaning process. After successful completion of the cleaning process, the first assessment should be a visual assessment, which should be followed by sampling of the same locations you swabbed in the second step. Again, the analytical procedures should be the same analytical procedure utilized in the second step.
These sampling results should meet (that is, be below) the residue limits established for the isolator surfaces (see last month’s Cleaning Memo). If not, the cleaning process should be modified and improved. The data at this step may also suggest that certain sampling locations, while having acceptable residue results, are more likely to have higher residue values than other locations; these locations are what I sometimes call the “worst cases of the worst cases”. The possible value in identifying these locations is that this data might assist in the selection of more critical sampling locations for any routine monitoring activities. Preferably the identification of those more critical locations is not based on just one study (one run), but is best established based on multiple runs.
Such a study should be documented as part of a risk assessment for your company’s approach to cleaning and routine monitoring for such indirect product contact surfaces.