Challenges Encountered in the Implementation of Bio-Fluorescent Particle Counting Systems as a Routine Microbial Monitoring Tool

Why Is This a Challenge?

The CFU is considered a gold standard within the industry; however, the sensitivity limitations of this method were highlighted as early as 1898 (14). The CFU provides an estimate of microorganisms in a controlled environment that is based on the ability of an organism to grow on general microbiological growth media under standard incubation conditions. Therefore, traditional sampling with the culture-based method will only report culturable organisms, underestimating total bioburden due to the inability of this method to detect viable but nonculturable (VBNC) organisms in the environment. In United States Pharmacopeia (USP) <1223>, it is mentioned that “studies on the recovery of microorganisms from potable and environmental waters have demonstrated that traditional plate-count methods reporting cell count estimates as CFU may recover 0.1%-1% of the actual microbial cells present in a sample” (15). Despite the limitations of the traditional method, current guidance documents provide acceptance criteria in units of CFU. For example, in the EU Guideline to Good Manufacturing Practice Annex 1, the recommended air sample limits for microbiological monitoring of clean areas during operation are specified in terms of CFU/m³ (16). The Grade A (ISO 5) classification has a presence/absence limit of <1 CFU/m³ whereas Grade B–D (ISO 6-8) classifications have a numerical limit of 10, 100, and 200 CFU/m³, respectively, which has consequences in terms of demonstrating equivalency. The industry’s history with the CFU and its use as the unit of measure in established environmental monitoring acceptance criteria can make the transition to a new, nonequivalent unit of measure more of a challenge.

What Can Be Done to Overcome This Challenge?

An early understanding of a BFPC’s advantages and limitations, the basis for the technology’s unit of measure, and regulatory support for use of new technologies that report in a nonequivalent unit to the CFU is important. This information can be used to introduce the technology to internal stakeholders early on, to minimize this challenge during implementation of the technology. Users of new technologies and regulatory authorities also have stressed the importance of early communication with regulators on the planned use of the technology. The creation of new technology groups within regulatory agencies, like the U.S. Food and Drug Administration (FDA) Center for Drug Evaluation and Research (CDER) Emerging Technology Team (ETT), the FDA Center for Biologics Evaluation and Research (CBER) Advanced Technologies Team (CATT), the European Medicines Agency’s (EMA’s) Innovation Task force (ITF), and the Medicines and Healthcare products Regulatory Agency (MHRA) Innovation Office, supports this communication.

More Information:

Despite the nonequivalency of the traditional growth-based and BFPC methods, it should be remembered that the ability to detect a loss of process control is more important in maintaining overall environmental control than the absolute microbial count. Both methods provide an estimate of the microbial count in an environment, and the overall ability of each system to support an assessment of control should be considered and communicated by microbiologists to internal stakeholders. For example, rapid and continuous methods like BFPC can provide a more sensitive and continuous assessment of microbial presence, with results reported in real time. This contrasts with the intermittent spot sampling performed with the traditional growth-based method, in which data are gathered, and results are verified only after a significant delay.

Major regulatory authorities have communicated an understanding and expectation of modern methods reporting in new units of measure and potentially requiring different limits than those used with the traditional growth-based method. The latest Annex 1 draft revision states in the notes to Table II: Limits for microbial contamination during qualification and Table VII: Maximum action limits for viable particle contamination that “if different or new technologies are used that present results in a manner different from CFU, the manufacturer should scientifically justify the limits applied and where possible correlate them to CFU” (17). More specific guidance is required on the quantity of data needed to set and support established limits, and on the necessity of side-by-side testing with the traditional method. Annex 1 acknowledges that technologies reporting in a different unit of measure may be used and communicates a direction to be taken. As early as 2006, regulators were addressing the emergence of new technologies that may require changing acceptance criteria, with an article from David Hussong, Ph.D. and Robert Mello, Ph.D. of the U.S. FDA mentioning that “implementation of newly developed, or more rapid, microbiology methods may also require the establishment of new acceptance criteria” (18). In another example, in a 2020 PEMM working group meeting with the FDA CDER ETT, it was stated that “the Agency agrees that BFPCs and traditional methods that provide results in units that may not directly correlate could be acceptable as an alternative method” (8). Health authorities are encouraging dialog through such groups, as modern methods like BFPC are adopted. This collaboration of four industry working groups has a manuscript underway with more information on the nonequivalence of autofluorescent particle counts (AFUs) and CFUs.

Why Is This a Challenge?

The validation of modern microbial methods can require a different test methodology than is used with the traditional method, and the interpretation of available guidance can seem daunting when trying to apply it to continuous, real-time detection methods. Furthermore, it can be assumed that extensive validation is required before any use of the system can occur.

What Can Be Done to Overcome This Challenge?

Use of rapid method validation guidance documents, industry working groups, experts in the field, and regulatory authority offices that support use of modern technology can help minimize this challenge (5, 15, 19⇓⇓–22). Also, depending on the intended application of the system, varying degrees of validation may be required. For example, the use of a BFPC system as a process monitoring tool may require a lower level of validation, appropriate to intended use, than a reduction or replacement of traditional environmental monitoring methods (8). Also, as a new system is evaluated, it has been communicated that U.S. FDA safe harbor principles may apply and a research exemption may be used (7, 8). It should be determined if other regulatory authorities permit or require use of something like a research exemption. Overall, the extent of testing should meet internal requirements and be reviewed with relevant regulatory authorities.

More Information:

Industry guidance documents such as USP <1223>, European Pharmacopeia 5.1.6 and Parenteral Drug Association (PDA) Technical Report 33 have been available from 2000 onward to support the validation and implementation of rapid and alternative methods (15, 20, 21, 23⇓⇓–26). Each document has been updated since the first publication, and guidance on the validation of modern methods that report in a unit other than CFU has improved. However, such guidance is still not clear for these types of continuous, real-time detection methods and often results in questions regarding the most applicable approach for validation. Because BFPC systems use the detection of intrinsic fluorescence instead of colony growth, BFPC systems may be considered nonequivalent to traditional microbial monitoring methods. Therefore, when validating BFPC systems, the application of a well-established method validation strategy using parameters like limit of detection, limit of quantification, and equivalence may not be required or recommended. As an example, there is an ongoing discussion on the applicability of the limit of detection assessment for the BFPC technology as it is diagnostic in nature and interrogates every particle that passes through the detection laser, assuming the concentration is low enough to prevent coincidence loss (i.e., overlapping particles). The primary vendor validation should address many of the parameters listed previously, whereas the secondary or user validation need not re-execute this work. Furthermore, establishing noninferiority to the traditional method may be more appropriate than demonstrating equivalence given that AFU and CFU are not equivalent. Validation is another topic that will be discussed in more detail by this working group collaboration in a future publication.

USP <1223> offers four validation options, namely the acceptable procedures, performance equivalence, results equivalence, and decision equivalence options (15, 22). Of these options, the decision equivalence would appear most applicable to in-process monitoring systems like BFPCs. Decision equivalence can be assessed by looking at the ability of the BFPC system, as compared with the traditional method, to identify an out-of-limit event. The salient decisions made in response to in-process monitoring would include the decisions on when to take out of service and sanitize a pharmaceutical-grade water system, when to clear the aseptic filling line after a human intervention, and when to cease manufacturing when there was a loss of environmental control in a clean room.

Why Is This a Challenge?

Establishing alert and action levels based on a technology that uses a different unit of measure (i.e., AFU), that has the ability to provide a more sensitive estimate of bioburden, and that provides data continuously can be a challenge. In the case of water-based BFPC monitoring in particular, the level of baseline AFU counts will differ for each water system, which may result in water system specific alert and action levels. Historically, the traditional growth-based method has been used to establish a state of control. Therefore, even if a new method reports in a unit other than and not equivalent to the CFU, side-by-side testing may be needed to establish what a control state looks like with the new method. Questions on the amount of side-by-side data required and how to implement alert and action levels on a continuous data stream, for example, frequency or time-based limits, are common. Furthermore, answers to these questions will likely differ depending on the use of an air- or water-based BFPC system and the degree of control within the environment.

What Can Be Done to Overcome This Challenge?

A review of the methods used by industry colleagues and the intended test plan with regulatory authorities can help to ensure that the side-by-side testing locations and quantity of data obtained are reasonable for establishing alert and action levels. Seasonality and historical profiles of the utilities or environments may need to be considered when establishing the length of time or periods of the year during which side-by-side testing will be performed. Such consideration may not be warranted for closed systems and highly controlled environments, where the duration of side-by-side testing may be considerably shorter and based on an assessment of counts during interventions and process activities, for example, as opposed to differences in counts due to seasonality. Initial alert and action levels can be established and applied during this side-by-side testing period to gain more information on the sensitivity of these levels and on the need for a time- or frequency-based implementation.

More Information:

When the industry transitions from active air or water grab sample monitoring at a limited frequency, for example, once per shift, to semicontinuous or continuous monitoring, we must expect to see an occasional excursion that may initially confound our quality control colleagues and be viewed as a potential compliance liability during regulatory agency inspections. The best way to view these excursions is to determine if they are directly related to personnel interventions and/or whether they indicate an adverse monitoring trend representative of a loss of environmental control within the clean room or water system. Modern methods that provide continuous monitoring without the need for intervention due to equipment or sample preparation can help limit the number of such operator interventions during manufacturing.

Due to the sensitivity of the BFPC method and continuous monitoring, an isolated single excursion may not be an indication of loss of control. Depending on the environment monitored, it may no longer be necessary to assign importance to a single overlimit result, given the continuous assessment of the environment and a better understanding of the return to control, that is, baseline. Consideration should be given to multiple excursions appearing as a cluster or over a more extended period. Frequent excursions may be indicative of a loss of overall environmental control and would be subject to investigation and corrective action.

As initial alert and action levels are determined, one challenge is selecting a relevant time period for trending analysis. For example, what number of excursions during this period would indicate a change in trend? Would, for instance, three excursions in less than ten minutes signal an out-of-trend event? To set this time period and number of excursions, the pharmaceutical manufacturer would need to characterize the environment by running the BFPC system over an extended period of time under operating conditions to define the particle baseline and initial alert and action levels (e.g., average +2 and +3 standard deviations). Concurrent side-by-side testing with the traditional method and a comparison of over alert and action events obtained with the two methods can then occur. A time- and/or frequency-based application of the BFPC alert and action levels can be made so that the applied limits are not too sensitive, considering the continuous and potentially more sensitive stream of data provided by the BFPC system.

Another management tool to navigate this challenge and the application of alert and action levels is the creation of a decision tree (9). A decision tree can be used to establish a plan if an alert or action level is exceeded, as the response may be different to that currently performed with the traditional method.

Why Is This a Challenge?

Qualification of a new, non-growth-based technology can differ from that performed with growth-based methods and will require considerations specific to the technology. This includes testing of the intended environment to establish baseline count levels on the BFPC system and an assessment of potential interferent materials.

What Can Be Done to Overcome This Challenge?

Technology vendors often have Installation, Operation, and Performance Qualification documents available to support the qualification of their technology. It is recommended to consult with the vendor and available industry working groups as qualification of a new technology is undertaken. It is also advisable to evaluate the desired installation location(s) in advance of establishing a timeline for unit qualification. An understanding of the installation environment and its compatibility with the new technology is an important first step. Questions such as “Is the background fluorescence of water high?” and “Are false positive interferent materials too significant at the selected location?” need to be answered. Again, these topics will be discussed in more depth in future publications from this working group collaboration.

More Information:

In air-based BFPC applications, technology users are not likely to be expected to create test organism bioaerosols (e.g., due to safety and equipment requirements) and can make use of a vendor’s primary validation package, expertise, and data. However, in-situ feasibility testing may be important for the technology user to perform. In the case of Grade A environment side-by-side testing with the traditional method, it may be beneficial to also sample in Grade C and D environments to gather data in an environment with nonzero AFU and CFU counts and obtain natural microflora information as part of the testing. In a Grade A environment, the traditional method and the BFPC system may both predominantly report counts of zero. In the case of water-based BFPC testing, a fluorescent bead may be used in qualification testing, potentially reducing or eliminating the need for end user testing of microbes, if required.

Why Is This a Challenge?

Pharmaceutical companies are often reluctant to be the first to implement a new technology or approach regulatory authorities regarding their use of a modern method. This reluctance is exacerbated by the fact that other companies that have reached out to regulatory authorities are often unable or unwilling to share information about such a meeting, what their planned use of a system is, and what feedback, either positive or negative, they obtained.

What Can Be Done to Overcome This Challenge?

It is advisable to obtain feedback from regulatory authorities early on, preferably from emerging technology programs like the FDA CDER ETT, the FDA CBER CATT, the EMA ITF, and the MHRA Innovation Office. In addition, a preinspection meeting could be used as a way of introducing the technology and its planned use to local inspectorates. However, the communication with regulatory authorities is not just left to the technology users alone. The industry working groups that are part of this collaboration are also actively working on communication with these groups (8). Attending industry conferences like the PDA Pharmaceutical Microbiology Conference can be an excellent way to ask questions of a roundtable of regulators, see presentations on the use and implementation of new technologies, and connect with a support network of regulators, industry consultants, industry colleagues, and vendors. Lastly, it is recommended to make use of guidance documents, articles, and presentations from regulatory authorities encouraging and supporting the use of alternative methods (7, 15, 18, 20, 21).

Why Is This a Challenge?

Internal stakeholders often have different concerns about and needs from a new technology compared with a microbiologist. The pros and cons of implementing a new technology may not be adequately highlighted as related to different internal stakeholders’ needs, which can cause hesitancy when a new technology is introduced and implemented.

What Can Be Done to Overcome This Challenge?

It is important to bring internal stakeholders into the new technology assessment process as early as possible and set realistic expectations for the new technology up front. If a small team is in charge of initial technology evaluation, it can be helpful to look at pros and cons of the new technology from the viewpoint of different internal stakeholders and make clear what the technology is and is not capable of in terms of stakeholder needs. This can help address potential concerns before implementation of the technology is undertaken and allow internal stakeholders to participate in the evaluation process. Creation of a new technology checklist can ensure that information pertinent to these groups is gathered as a technology is evaluated.

More Information:

What are the internal stakeholder groups? A short list to consider includes maintenance and utility engineers, quality assurance, quality control, plant operators (i.e., those using the technology), automation teams, facility equipment manufacturers (e.g., isolator manufacturers if installing into an isolator), information technology/data support, regulatory teams, validation teams/metrology, site management and global functions. Because many stakeholders have relevant knowledge and input, knowledge about the new technology should not reside with one person or group alone. For global companies, it may be important to investigate whether there are others within the company already looking at the technology (for example, other sites or corporate centers focused on new technologies). Furthermore, if employee turnover occurs, knowledge management is critical to prevent delay and rework.

In the case of a BFPC system, many applications relevant to internal stakeholders are possible. It is envisioned that process control use of the technology will support a reduction and, in some applications, eventual replacement of the traditional method. BFPC systems can now support a better understanding of process control, accelerate risk assessments, and allow for faster decision-making through continuous monitoring and higher measurement sensitivity (27). As energy saving is important, a water-based BFPC system can provide information that supports a reduction in sanitization frequency, sanitization agent concentration or contact time, whereas an air-based BFPC system can support a reduction in air exchanges, especially when an aseptic process is not occurring. Initially, BFPC systems can also support bringing a water system or controlled environment back online more quickly, at risk, until the traditional method results are available (28). Air-based BFPC systems are also quite useful as gowning and general training tools due to the real-time feedback that they provide to the trainee. Industry movement toward robotic and gloveless applications in closed isolators, continuous manufacturing, and small batch manufacturing with short delivery requirements, like cell and gene therapies, are also well suited to the real-time process control provided by continuous BFPC monitoring. Highlighting the intended use of the BFPC system, initially and in the future, can help manage expectations for the technology.