What Counts? Establishing a Bioburden Strategy for a New Pharmaceutical Product

Dr. Tim Sandle, Ph.D. (CBiol, FIScT)-Bio Products Laboratory Limited

Introduction

For both sterile and non-sterile manufacturing, bioburden control as evidenced by microbial counts is especially important in relation to the quality of the finished product and as an indicator of process control. Samples are drawn from intermediate product at defined stages (ideally based on risk assessment) and these allow for the microbial levels to be tracked from upstream processing to downstream processing (with an expectation that the microbial levels decrease, or at least remain unchanged provided they are below an acceptable action level). Bioburden assessment informs the manufacturer about both the expected microbial load of the product and the presence or absence of specific microorganisms, some of which might be classed as objectionable.¹

For new product development, consideration must be given to the samples to be selected for bioburden testing, the appropriate limits to set, and the validation strategy to ensure that the obtained test results are free from inhibition or enhancement.

Contamination Control Strategy

Pharmaceutical preparations, especially biologic products, are at risk from microbial contamination in many stages. Such risks exist because biopharmaceuticals often include the types of carbon sources and other growth factors that favor microbial growth. Moreover, many of the types of microorganisms found within the environment including process areas can adapt and survive under a variety of conditions. Where microorganisms are capable of growth in conditions that favor cellular division, microbial contamination poses a significant risk to biologic products.² There are many variables where contamination can occur. For example, open processing presents a greater contamination risk than closed processing. Open processing may be an individual event or may occur when a vessel is opened several times for mixing or addition of chemicals. The room environment and operator aseptic practices may also impact microbial risk.

Although microbial contamination may occur, this does not necessarily mean that microbial growth and proliferation will take place. Product, process, time, and temperature present important variables as to whether microbial growth will occur. These should not be viewed as discrete factors. These factors often need to be combined since one factor in conjunction with another may lead to a different risk outcome. For example, one type of growth promoting product held at 2-8°C would be at a lower risk due to this temperature inhibiting the growth of most microorganisms than the same product held for the same time period at 30-35°C.³

Hence, bioburden control forms an essential part of the biocontamination control strategy. This strategy should center on controlling the source of microorganisms and ensuring that conditions that promote microorganism survival, growth, and persistence are minimized. The three main risk areas are arguably:

1. Starting materials and added materials (such as water during manufacturing).
2. Improperly cleaned or contaminated equipment, which can add microorganisms into the product stream.
3. Hold times. Hold conditions including time and temperature for a process should be validated to demonstrate control and prevention of potential microbial growth.⁴

Test methods for bioburden, whether they be conventional or rapid microbiological methods, should be verified or validated as being acceptable.

New Product Development

When developing a new product consideration must be given to the points in the process where sampling for bioburden will be appropriate. A smaller number of samples may be tested during the development stage prior to scale-up. While each manufacturing process will be different, sampling points for consideration in relation to process stages are:

1. Starting materials (as might be assessed through the Microbial Limits Test).
2. Points where additional materials are added (this may include tests on additional materials where microbial growth is possible, for example manufacturing a buffer that contains a sugar).
3. The addition of water (this should include a sample of the ingredient water taken at the same time).
4. Post-mixing and after any hold time of significance (defining ‘significance’ is somewhat arbitrary but given the rates of microbial growth, taking four hours to be ‘significant’ is a good yardstick). With mixing processes like centrifugation, consideration needs to be given as to whether the precipitate (pellet) or the supernatant constitutes the material required for the next process stage. Centrifugation typically leads to bacteria being compacted into the pellet,meaning the bioburden levels between precipitate and supernatant will vary considerably.
5. Pre-purification (such as prior to ultrafiltration, chromatographic column etc.)
6. Post-purification.
7. Final bulk.
8. Prior to final sterilizing filtration (for sterile products).

With the above there will be other samples to consider based on the specifics of the manufacturing process (the total number of samples taken will depend on the complexity of the manufacturing process in relation to manufacturing stages and processing time. Making an infusion fluid, for example, will have few process stages whereas some biotechnology products will have over a dozen). Such samples, and the suggestions above, should be determined by risk assessment.⁶ In addition, it is product undergoes any unexpected delays that lead to increased hold times then it is prudent to take a sample for bioburden determination immediately prior to moving into the next process stage, if it is technically feasible to do so without compromising the product.

Anticipating Bioburden Levels

When developing a new product, bioburden considerations must form part of the strategy. This involves considering:

1. What do we expect the starting bioburden to be?
2. Where in the manufacturing process do we expect to see a reduction?
3. What should the final bioburden be?
4. Are there any specific objectionable microorganisms that need to be assessed? (for example, screening for the biovars of the Burkholderia cepacia complex).

As an example, imagine a sterile product manufacturing stream, with bioburden samples tested using conventional culture-based methods. With bioburden levels (or limits), it might be planned for:

1. Starting material / first stage of the process: Maximum 1000 CFU/mL.
2. Initial manufacturing: 100 CFU/mL.
3. Purification: 10 CFU/mL.
4. Final bulk / pre-sterile filtration: 10 CFU/100mL (equivalent to <1 CFU/mL).

Once planned indicative alert and action levels need to be set. The alert level is a signal to begin an investigation into an atypical data trend, although no action is necessarily required; the action level being met or exceeded requires an investigation and a corrective action. These levels will require modifying once sufficient process data has been obtained using an appropriate method for what will probably be skewed data (that is data not showing a normal distribution pattern). New alert and action levels should be set after a minimum of 100 sample results for each stage have been generated.

Some specifications may also need to be set. By specification, this is taken to mean a limit that enters a product license and where exceeding this limit could lead to batch rejection or some form of license variation upon discussion with the applicable regulatory authority. It is important to minimize such entries into the product license for bioburden samples since occasional bioburden fluctuations can occur only to be removed due to a facet of the process or by dilution further downstream.⁷ For example, for sterile manufacturing, it might be appropriate only to specify a license limit for the pre-sterile filtration sample.

Method Suitability Testing

At the development stage, consideration needs to be given to the method of testing and to conduct method suitability testing (to show the recovery of microorganisms in the presence of product for the selected test method, what is sometimes, if erroneously, referred to as ‘method validation’). Three factors will influence microbial recovery: the material may require no modification; or, either the material contains an inhibitory substance that can be overcome with a modification to the method or the material is antimicrobial.

The method selected will also impact upon, or be shaped by, the sample size. Considering conventional methods, we have:

Membrane filtration

1. Pour plate
2. Spread plate
3. Most Probable Number

Of these, membrane filtration is typically the method of choice. This is because a larger sample size can be tested (100mL as opposed to 1mL) and inhibitory substances that might interfere with microbial growth can be more easily overcome through rinsing or using an alternative filter (some drug compounds have an affinity to the nitrocellulose membranes commonly used leading to a concentration of the antimicrobial on the surface of the membrane. The use of polyvinyl difluoride membranes may alleviate this problem).

As for the method design, the pharmacopeial chapters for the Microbial Limits Test can either be followed or adapted (as intermediate bioburden samples are not ‘finished products’ unless they are required to be sold as intermediates to another manufacturer the compendial methods do not need to be followed). In terms of adapting, it might be possible through method development, for example, to show that one culture medium can recovery both bacteria and fungi rendering the dual media step unnecessary.⁸

To develop a suitable method to assess bioburden, the first consideration is with whether the intermediate product is in liquid form or in a form that is soluble. If the material is not, for example being in a paste form, it might be possible to dissolve a portion of the sample but care needs to be taken that microbial recovery can be demonstrated or that any extraction process (such as heating) does not lead to microbial death. It may be prudent to take a sample of a subsequent stage code once the product material has been processed into a liquid state.

The next consideration is whether the sample needs to be neutralized? This may or may not be known initially (the chemical nature of the material or the pH will provide an indication). The common methodological considerations for neutralization are:

• Chemical neutralizers,
• Dilution,
• Or by filtration and rinsing.

With chemical neutralizers, some examples are:  

• Polysorbate 80/20
• Bisulfate
• Thiosulfate 
• Sodium thioglycolate (DE Broth)
• Lecithin and Polysorbate 80/20
• Glycine
• Thioglycolic acid (DE Broth)

Many diluents, agars, and broths can be purchased with these inactivators already incorporated into them. For example, Polysorbate 20 and lecithin are commonly added to: Soybean Casein Digest Broth and Soybean Casein Digest Agar.

With dilution the antimicrobial properties can be diluted to a level where it is no longer active. When considering dilution, the resolution or sensitivity for the test in relation to intended action level will dictate the actual level of dilution possible.

For microbial challenges, these should be selected with consideration as to the nature of the starting materials, product environmental risks, the degree of personnel involvement, the addition of water and so on. The organisms will typically consist of: a Gram-positive coccus (such as a species of the genera Staphylococcus or Micrococcus); a Grampositive rod, preferably a spore former (such as a Bacillus species); a Gram-negative rod (something representative of a water system); and a fungus. The list of organisms under the compendial Microbial Limits Test is a useful starting point, if these are representative. An additional consideration is whether these organisms will be sourced from a culture collection or sourced from the manufacturing facility (environmental isolates) or some combination thereof. The regulatory trend is towards the inclusion of facility isolates, irrespective of whether these organisms are any more difficult to recover.

Method development should also demonstrate:

• Reproducibility (closeness of the agreement between the results of a sufficient number of replicates runs): Testing should be performed on at least three lots.

• Repeatability (closeness of agreement between the result of successive measurements): Testing should be performed at least in triplicate during the method development stage.

• Accuracy (closeness of agreement between measured value and accepted value) and precision (closeness of agreement among repeated measurements): These are assessed in terms of microbial recovery. This is typically set as a better than 50% recovery although some facilities will try to reduce the margin of error by setting a better than 70% recovery as the test criterion.

The microbial challenge process involves: 

• Test Group: The samples are subjected to the neutralization method followed by the addition of a low level of microorganisms (<100 CFU).

• Viability Group: The microbial inoculum is plated without exposure to the neutralization step or the product.

• Negative Control Group: The neutralization method is used as in Test Group above with the exception of the neutralization fluid is used in place of the product.

The test organism can be added in the final rinse of the membrane filtration step, directly to the Petri dish prior to the addition of the agar or to the final dilution of the dilution step.

Post-incubation, the recoveries between the Test Group and the Viability Group are compared to determine whether the test group’s recovery is within the target recovery (that is not less than 50% or 70% recovery when compared to the Viability Group). In addition, the Peptone Control Group should not show less than a 50% or 70% recovery when compared to the Viability Group. This shows that the neutralization step was not toxic.

When method development is not successful, the following can be applied:

1. Increasing the volume of diluent (with the quantity of test material remaining the same).
2. Incorporating a sufficient quantity of suitable inactivating agent(s) in the diluents.
3. A combination of modifications (1) and (2) above.
4. Trying alternative filters, such as low binding filters (for the membrane filtration method).
5. Adding suitable inactivators to the rinses (for membrane filtration).
6. Increasing the number of rinses (for membrane filtration).

The above list further emphasizes why membrane filtration is the method of choice, based on the extent that it can be modified compared with other plate counting methods.

At times, some or all of the test microorganisms may not be recovered in spite of the incorporation of suitable inactivating agents and a substantial increase in the volume of diluent. It might be that the failure to recover these test organisms is due to the antimicrobial nature of the sample and that the sample is unlikely to be contaminated with that species (or strain) of microorganism. Where there is partial recovery of some organisms, it is logical to continue to test the stage code in routine production. If no organisms are recovered, and the stage is not the final stage in the process, then a justification can be made not to test a sample at this manufacturing juncture.

Rapid Microbiological Methods

The pharmaceutical sector should be gravitating away from conventional methods and towards rapid microbiological methods at as faster pace. Many technologies offer improvements to bioburden testing, both in improving the time to result (reduced incubation times given the substantial lag time that exists between sampling and reporting) and with the accuracy of the result (be that a direct count, early detection of colony formation, or indication of biologic activity.⁹ Real-time capture of data also addresses data integrity concerns.

Several alternative microbiological technologies are automated variants of long-established growth-based compendial microbiological methods. The qualification of these methods is generally based on the considerations above, in terms of microbial recovery (absence of inhibition) and the demonstration of accuracy and precision assessments should be performed.¹⁰

Data Examination

Sampling at appropriate stages and ensuring that the methods used are suitable are important considerations. To these we need to add data considerations. The results of bioburden monitoring should be assessed for each batch. The data review may lead to:

1. Batch rejection (for example, if the pre-final filtration sample is out-of-limits, suggesting a contaminated bulk).
2. Confirmation of an adulterated batch should a finished product fail and this can be attributed to contamination entering the process.
3. Opportunity for process improvement, should a particular stage code be outside of limits.

With ‘c’ above, more meaningful data will be drawn out from trend analysis. This may indicate a deterioration with part of the process. For example, consider the data (Figure 1) for an ultrafilter final rinse sample.

The graph shows a deterioration in the data and the need to investigate. With the above example, a protein residue had built up, a microbial biofilm developed, and this could be effectively cleaned by the standard ultrafilter cleaning cycle and a more aggressive sanitization agent was required.

Assessing the data of a sufficient period of time also aids process understanding. For example, it is possible to learn from data reviews:

• If levels of bioburden are higher at the start of the process (upstream samples) compared with later in the process (downstream samples).

• If parts of the process expected to lead to bioburden reduction are effective.

• Whether additional process steps, such as water rinses, contribute to the bioburden.

• Where applicable, if additives to the process, such as formulation buffers, contribute to the bioburden. This can lead to process improvements, such as the filtration of buffers.

Trending is important to differentiate between isolated events and a systemic breakdown of control. Where conventional test methods are used, given the limitations of the colony forming unit,¹¹ assessing data over a longer time frame contributes to greater understanding of the biocontamination levels.

Conclusion

This article has presented some considerations for developing a bioburden strategy for a new product. Elements of the article may also be of use for those who wish to go back and reconsider current processes, in terms of whether the bioburden samples selected are appropriate for informing as to the state of manufacturing biocontamination controls. As well as presenting advice on appropriate stages at which to test product, the article has also presented considerations for method development. The article also includes considerations for setting alert and action levels and the importance of data trending.

These aspects come together for providing a suitable bioburden control strategy for product processing.

References

1. Sandle, T. Bioburden determination. In Pharmaceutical Microbiology: Essentials for Quality Assurance and Quality Control, Woodhead Publishing, 2016, pages 81-91
2. Sandle, T. Contamination Control Risk Assessment in Masden, R.E. and Moldenhauer, J. (Eds.) Contamination Control in Healthcare Product Manufacturing, Volume 1, DHI Publishing, 2013, River Grove: USA, pp423-474
3. He, Y., Darou, S., Henn, S., Walter, R., Cundell, T., Yerden, R., Henn, A. BioPharm International, Temperature and Relative Humidity Control to Reduce Bioburden in a Closed Cell Processing and Production System without Disinfectants, BioPharm International, 2021, 34 (6): 36–41
4. Clontz, L. Microbial Limit and Bioburden Tests: Validation Approaches and Global Requirements, CRC Press, Boca Raton, USA, 2013, p49
5. Peterson BW, Sharma PK, van der Mei HC, Busscher HJ. Bacterial cell surface damage due to centrifugal compaction. Appl Environ Microbiol. 2012;78(1):120-125
6. Tidswell, E. C. (2004) Risk profiling pharmaceutical manufacturing processes. Eur. J. Par. Pharma Sci., 9 (2), 49-55
7. Gervais, D. Quality Control and Downstream Processing of Therapeutic Enzymes, Therapeutic Enzymes: Function and Clinical Implications, 2019, 9 (3): 55-80
8. Sandle, T.,Skinner,K. and Yeandle, E. Optimal conditions for the recovery of bioburden from pharmaceutical processes: a case study, European Journal of Parenteral & Pharmaceutical Sciences 2013; 18(3): 84-90
9. London R, Schwedock J, Sage A, Valley H, Meadows J, Waddington M, Straus D. An automated system for rapid non-destructive enumeration of growing microbes. PLos ONE. 2010;5(1):e8609
10. Jones, D. and Cundell, T. Method Verification Requirements for an Advanced Imaging System for Microbial Plate Count Enumeration, PDA Journal of Pharmaceutical Science and Technology, 2021, 75 (6): DOI: https://doi.org/10.5731/pdajpst.2017.007955
11. Cundell,T. The limitations of the colony-forming unit in microbiology, European Pharmaceutical Review, Issue 5, 2015 https://www.europeanpharmaceuticalreview.com/ article/37416/the-limitations-of-the-colony-forming-unit-in-microbiology/

Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special
offers from American Pharmaceutical Review delivered to your inbox!
Sign up now!