Last month we covered some possible sampling methods that could be used for sieves/screens. This month is a follow-on Cleaning Memo that discusses sampling recovery studies for sieves/screens. Of course, if my suggestion made last month that sieves/screens be dedicated to one product (or at least to one active) were followed, then it would not be necessary to conduct sampling recovery studies for the active (we’ll cover sampling recovery for detergent later). But, if you choose to use one of the options discussed last month for sieves/screens in a cleaning validation protocol to use the analytical data to compare to an acceptance limit, then it can be expected that regulatory inspectors might want to see that you have performed a sampling recovery study for this unique sampling method.
So, how can that sampling recovery done? The best answer is to replicate the sampling method on a clean screen which has been spiked with a fixed known amount of the active (the API). As with any sampling recovery study, the amount (in µg) analyzed in the sampling solvent can be compared to the amount (also in µg) spiked onto the area of the sieve/screen that was sampled. The sampling method used for the recovery study should be the identical sampling method to be used in an actual cleaning validation protocol. However, below are some considerations in doing such a recovery study for sieve/screen sampling.
Area to be spiked
Assuming the area sampled as covered last month will be a circular shape (such as a circle with a diameter of 11.3 cm, which will give a “surface area” of about 100 cm2), then the spiking material should be spread as evenly as possible with that circle with a diameter of 11.3 cm. Let’s assume that the calculated acceptance limit is 1 µg/cm2; in that case a solution containing 100 µg of active should be prepared so that it can be spiked onto the sieve/screen. Here’s the tricky part: I can’t spike a very large amount of solution onto the sieve/screen without having part of the spiking solution fall through (in which case I would inevitably get a low recovery value). So I need to consider the “openness” of the sieve/screen and by experimental trial and error determine what is an appropriate amount of spiking solution to place on the screen. This might mean using a micropipette to dispense very small amounts of the spiking solution on different portions of the sieve/screen. Realize that the application consistency across the surface will not be even; the fact that there will be some areas with greater amounts of active only makes the sampling recovery study a worst case (which is essentially the same situation as in swab sampling recovery studies).
Cleanliness of sieve/screen and collection vessel
It is preferable to spike a “virgin” (new, never used) sieve/screen for the recovery study. The sieve/screen should be effectively cleaned with organic solvents to remove any organic materials, and then be cleaned with a detergent, thoroughly rinsed, and then dried by use of isopropyl alcohol. This helps assure there are no extraneous materials on the sieve/screen which would interfere with the sampling recovery procedure for the active. Similar care should be used for the cleanliness of the collection vessel (or in the case of using multiple passes of the same solvent, the cleanliness of the vessel or pipette used to transfer the solvent for multiple passes through the screen).
Different sizes of sieves/screens
It can be expected that the larger the number of opening per standard area (such as openings per square inch), the lower the recovery will be based on the greater difficulty of access. Therefore, it may be possible to perform a recovery study on that worst case, and apply that percent recovery to all sieve/screen sizes.
Recovery for detergent
The above considerations related to cleanliness of the sieve/screen and the collection vessel apply for recovery studies for the active (API). If a recovery study is needed for a detergent, then it is possible to use a different detergent (that is, different from the detergent used for cleaning of a “soiled” screen in a cleaning validation protocol) for cleaning of the sieve/screen and collection vessel to be used in the recovery study. On the other hand, the recovery of the detergent should generally be >90% provided that (a) the detergent is a readily water soluble detergent (why would you use anything other than a readily water soluble detergent), (b) thorough rinsing is done, and (c) the water rinsing is followed by an alcohol application (further reducing any detergent that might have been left). Realize in this cleaning of the sieve/screen and collection vessel, it generally is the case that the detergent has no opportunity to dry onto the surfaces, further assuring that the sieve/screen and collection vessel are essentially “free” of any detergent. So, when the sieve/screen is spiked with a known quantity of detergent as the 100% value, the resulting recovery sampling provides an accurate number.
With all of that said, my preference for a recovery of detergent would be to not do a formal recovery study. In place of a formal, I suggest a “rationale” document as to why a formal study is not needed. One option would just be to say the percent recovery for sieve/screen sampling for detergents would be no worse than a typical swab recovery for detergents on stainless steel, and use the recovery percentage obtained in that swab study. Another option would be to say that the recovery of the detergent on the sieve/screen should be no worst (based on relative solubility, assuming the same solvent were used for sampling each) than the percent recovery for the active on the sieve/screen; therefore the recovery percentage for the active could be used for the detergent. The rationale for such approaches should be clearly documented.
The issue of recovery studies further complicates what is needed for sampling sieves/screens. Certainly the costs of these additional design studies and recovery studies should be balanced against the cost of just dedication of sieves/screens.
Copyright © 2023 by Cleaning Validation Technologies