This month’s Cleaning Memo covers bioburden limits (microbial) for initial cleaning validation protocols. The focus will be bacteria and yeasts/mold, and will not include endotoxins. Subsequent Memos will cover bioburden for Clean Hold Time and for routine monitoring.
The first thing to remember that is that microbiological control is important for regulatory bodies. While chemical residues (such as the API/active) have relatively standard approaches (such as carryover calculations based on activity or toxicity), that is not the case for bioburden.
The second thing to remember is that analytical methods for chemicals residues are generally accurate and precise (such as plus or minus 5%). Furthermore, recovery of chemical residues from equipment surfaces is generally high (such as 70% or more). For typical bioburden methods involving growth on suitable media, this is not the case. The reason for this is that these techniques that involve incubation on growth media measure CFUs (colony forming units). During the sampling and application to the growth medium, one colony taken from a surface may be “disrupted” to form two or more colonies. As a general rule, a difference of one log (a factor of 10 times) is required to see a significant difference in bioburden. For example, I might have a bacterial inoculum and ask two microbiologists to measure the bioburden level at the same time, using equivalent equipment, with same lot of media and with incubation side by side in an oven. If one person gets a result of 5 CFU/mL, I don’t necessarily expect the other person to get 5 CFU/ml also. If the result of the second person is 2 CFU/mL, 8 CFU/ml or even 15 CFU/ml, I am not surprised. On the other hand, if the second person gets a value of 50 CFU/mL or more (a one log difference), then something is wrong. Furthermore, most microbiologists will state that one cannot specify with any confidence the accuracy and precision of any result below 20 CFU per 25 square centimeters in a traditional growth method (which doesn’t – and shouldn’t – stop us from reporting out values such as 1 CFU/plate or 5 CFU/plate).
That hypothetical example involves taking a sample from a standard preparation. The situation is more complex if the evaluation involves either swab sampling or rinse sampling. In both cases, the action involved in either can disrupt the colonies on the surface resulting in higher CFU values in the test results. For example, I might prepare three surfaces (three coupons) with the same (low) level of bacteria, and then sample one by swabbing, one by rinsing, and one by direct contact plate. What I will find? Generally the contact plate will have the lowest CFUs, and the swab and rinse will have higher CFUs due to colony disruption that occurs in the swab and rinse sampling.
This discussion so far has been so you can appreciate the difference between analytical results for chemical residues as compared to analytical results for microbiological residues. So you might now ask “Okay, how are limits set, which should be the topic of this Cleaning Memo?” Well, they are generally set by relatively common “industry standard practices”. These industry practices are based on what has commonly been used and achieved in different facilities. They may be slightly different in sterile manufacturing as compared to non-sterile manufacturing. They also may be different in wet versus dry processing. What follows is a discussion of those differences.
For sterile manufacturing for finished product, it is not expected that the equipment be sterile (zero CFU) following cleaning. Why? Generally the equipment will be sterilized by steam following cleaning and before product is processed in the equipment. However, that doesn’t mean that there should be no testing for bioburden in the cleaning validation protocol. There are two reasons for setting relatively low bioburden limits in this situation. One is that any sterilization is typically validated with a specified bioburden level which should not be exceeded in processing. Yes, that sterilization validation is done with a very high level of bacterial spores, levels which would never be that high following a cleaning process. Furthermore, the bacteria left after the cleaning process are generally in the vegetative state, and not in the spore state. Secondly, if the bacteria left behind after cleaning are gram negatives, while they may be killed during the sterilization process, the killing process may release endotoxins which may compromise the safety of the next processed product. While some might want to require the final rinse to meet the WFI specification if WFI is used for the final rinse, I generally recommend higher level such as between 1 CFU/mL and 10 CFU/mL. A common industry value for sterile processing is 5 CFU/mL for final rinse sampling. The reason for these higher values is that there should be no expectation that the final rinse continues to meet the WFI specification after passing through cleaned equipment. Yes, I might still meet the WFI specification, but is that necessary (particularly considering the equipment may be SIPed before seeing product. Swab sampling is generally not done for sterile processing because the equipment is a closed system, and opening it for swabbing should be avoided.
For non-sterile manufacturing, there is no expectation that the manufactured product be sterile. However, there are industry expectations for bioburden content of finished wet products (100 CFU/ml) and dry (1000 CFU/g) products. Individually, companies may set more stringent specifications below those levels. It is possible to do “reverse” calculations to see what would be allowed on equipment surfaces based on product specifications. Such a calculation is given in a 2002 paper I published (“Equipment Cleaning Validation: Microbial Control Issues”, Journal of Validation Technology 8:4, pages 40-46). Such a reverse calculation will generally result in equipment surface limits well above what should be achieved with a robust cleaning process. For this reason, bioburden limits for non-sterile manufacturing are typically set on what I call “industry standard practices” of 1 or 2 CFU per square centimeter (equivalent to 25 or 50 CFU per swabbed area of 25 square centimeters) for aerobic bacteria, with levels for fungi set at a level 10 times lower. For rinse sampling, the bioburden level is generally set at the final rinse water specification, such as 100 CFU/mL for Purified Water PW). In most cases where the final rinse is with hot PW, actual values will be below 5 CFU/mL.
While it might be desirable to have a more scientific way to set limits for bioburden, none have been proposed that are even somewhat practical and achievable. This idea of “industry standard practices” is probably the best to use at this time for bioburden limits, Of course, a company may choose to use more strict limits. However, care should be used in selecting limits for protocols based on historical data such a “mean plus three standard deviations”. That approach would statistically allow for exceeding that limit a small percentage of times, which would cause havoc for the success of cleaning validation protocols (even though patient safety or product quality would not be affected).
Note that this discussion covers limits following completion of the cleaning process. Next month we will cover bioburden limits in clean hold time protocols.
Copyright © 2025 by Cleaning Validation Technologies
