This Cleaning Memo is focused on limits for indirect product contact surfaces. It first covers my recommendations, and then specifically addresses the considerations in setting limits for isolators as recommended in a November 2017 PDA Letter (“Isolator Surfaces and Contamination Risks to Personnel”, by Richard Denk et al).
I’ll start with a clarification of the three types of surfaces where limits might be set after cleaning. I will utilize the example of residues of the drug active (API, or drug substance) in discussing setting limits. For setting limits, I will focus on what I call the L3 value, which is the limit per surface area, in units such as mcg/cm2 (see the September 2012 Cleaning Memo for my shorthand notations for different ways to express limits). If you prefer a different way to discuss limits, such as mcg/swab, you should be easily be able to convert the value I will give for L3 to an amount per swab (L4a in my notation) by multiplying my L3 value by the area swabbed (such as 25 cm2 or 100 cm2).
Okay. The three types of surfaces, along which acronyms I will use to make this Cleaning Memo shorter, are:
The DPCS is what we generally focus on when we set limits. These are the equipment surfaces where manufactured product will directly touch the surface, such that there is a high probability of residues left on the surface transferring to the next product. These surfaces are usually the interiors of the manufacturing equipment. Note that transfer may not be 100%. For liquid processing the amount (or percent) of transfer will depend on a variety of factors including the type/nature of the residue, the agitation/turbulence, the time of processing, the temperature of processing, and the solubility of the residue in the next drug product. For solids processing the amount (or percent) of transfer will depend on a variety of factors including the type/nature of the residue, the “frictional” forces of the processed product with the surface, the time of processing, and the temperature of processing. While it is possible that not all the residue will transfer directly to the next product, for a typical carryover calculation it is usually standard practice to assume (as a worst case) 100% carryover of residue. [For those of you who might be worried about what happens if carryover to the next product is less than 100%, the good news is that (hopefully) any residue from a prior product left after processing a second product will be further reduced by the cleaning of that second product; therefore any possible transfer to the third product will be at a very low level. And “No”, I am not aware of any studies to support this, but it is a reasonable assumption. If it’s not a reasonable assumption, we are in real trouble.] Therefore for DPCS a typical carryover calculation is done.
Which brings us to NPCS. These are surfaces which don’t directly contact manufactured product (or at least manufactured product that will be sold and/or used by patients), and this is the key, that are remote from open product and where there is not a reasonable probability of transfer from the surface to the product (usually by an airborne route). Examples of NPCS might include walls, floors, ceilings, and the outside of process equipment. These typically have not been covered in a cleaning validation program for process equipment, but may be handled as part of area or facility cleaning program. Typically the requirement for cleaning is that the surface be visually clean immediately after cleaning. However, it is recognized that is not necessarily an expectation that the surfaces will be visually clean at a later time. Certainly if spilled product is seen on the floor, this should be addressed. But it is entirely possible that the floor will not be visually clean during a subsequent manufacturing process.
Now we get to the issue of IPCS. Like NPCS an IPCS is a surface which doesn’t directly contact manufactured product, but this is the key, it is a surface in close proximity to open product where there is a reasonable probability of transfer from the IPCS to manufactured product (again, usually by the airborne route). Examples of IPCS include the interiors of lyophilizers (freeze dryers) and the interiors of isolation chambers (such as those used for filling product into vials). It is not practical to perform carryover calculations for IPCS as we ordinarily do DPCS. If we assume 100% transfer from the surface to manufactured product, in most cases the L3 limit for the IPCS would be lower (more stringent) that the L3 limit for a DPCS in the same equipment train. From a product risk assessment perspective, it intuitively does not make sense that the residue limit for an IPCS should be more stringent than the limit for a DPCS.
We might think about various ways to posit a transfer for IPCS that would be somewhat less than 100%, such as what I have outlined for lyophilizer chambers in the August 2014 Cleaning Memo (with more detail in my training webinars on IPCS). But even in those situations, there are a variety of unknowns such as whether the transfer of residues will be uniformly carried over to each vial processed in a lyophilizer or in a filling isolator. My typical recommendation has been to select one or more of these three criteria for IPCS situations:
For the third option (which is my preference), what I mean is to select a L3 value for the IPCS the same as for equipment processing the same product immediately prior to the IPCS equipment. For example, for vial lyophilization, calculate the L3 for the filling equipment (a DPCS) used to fill the vials, and then apply that same L3 value for the lyophilizer surfaces. In this situation, I would still require the lyophilizer surfaces to be visually clean.
I realize that some may set limits for a lyophilizer based on process capability. The only issue with doing this on true process capability data is that the limit depends on the cleaning process chosen; a general principle in cleaning validation has been that the limit should determine how stringent the cleaning process must be (and not vice versa).
All this has been setting the stage for a discussion of the PDA Letter recommendation for the interior surfaces of isolators. I should clarify that the PDA Letter recommendations cover both operator safety as well as product (or patient) safety. I will just cover their recommendation for patient safety. For the interior of isolators, here is what the authors first state (before they provide a “worst case” scenario and the associated recommendation):
“Compared to the correlation between surface contamination and risk to the worker, it is even more cumbersome to establish a relationship between the contamination of a nonproduct contact surface and a cross-contamination risk. It is therefore very difficult if not impossible to establish science-based non-product contact surface limits in the GMP context.”
This essentially is an acknowledgement of the fact that no good models exist for transfer from such surfaces (whether IPCS or NPCS) to manufactured product. That said, the scenario used in the publication is that as a worst case the residue from 100 cm2 would never transfer into a single vial in a filling isolator. Therefore, assuming only one vial per day per patient, a limit of one PDE per 100 cm2 of the interior surface of the isolator should be more than protective. For example, an active with a PDE of 10 mcg/day would have a L3 limit of 0.1 mcg/cm2.
Note that in the PDE Letter article, the surface area units are generally expressed as dm2 (rather than the cm2 which I have used). Furthermore, in that article the authors go on to relate different levels to what might be visually clean, using a value of 4 mcg/cm2 as a “frequently quoted quantitative number for visual cleanliness”. Finally, they also suggest a “surrogate” study whereby residues on surfaces are measured both inside the isolator as well as outside the isolator to assist in a risk assessment.
I highly recommend this PDA Letter article be considered by all involved in setting limits for IPCS.