Environmental monitoring (EM) trending programs are an integral tool for use in making sound, scientifically justified cGMP decisions. Moreover, the usefulness of this tool is directly proportional to its quality, and that quality is likewise proportional to the trending program design and implementation. There are a number of considerations to be made when designing and implementing meaningful, applicable and effective trending programs.
Determining How EM Data is to be Collected and Grouped for Analysis
Careful consideration must be given to how collected data is grouped for analysis as it has a major impact on the output of the trending program. Consider, for example, the trending of personnel monitoring data. Evaluation of this data can be performed by sorting the data either by the individual operator or across all operators performing similar operations or functions. The determination of which data set is appropriate depends on what information is sought.
For example, evaluation of the trended EM results collected from individual operators will demonstrate how that individual operator is performing against a firm's established parameters. It can reveal whether or not additional training may be required due to a lack of individual understanding of proper aseptic cleanroom behavior. Conversely, trending of results across all of the operators can reveal system-level deficiencies such as improper gowning stringency or lack of sufficiently detailed and/or inappropriate operation procedures. It can also indicate the overall impact that operators, through manual intervention(s), may have on the process and whether or not additional contamination controls are needed.
Ideally, both methods should be utilized. It is useful to be able to benchmark the performance of an individual operator against other operators performing similar tasks, in order to discern whether or not poor monitoring data generated by an individual is aberrant or commonplace. Similarly, the manner in which processing rooms, support areas and even individual EM sites are grouped for trending is critical to obtaining a clear picture of the microbiological state of the overall aseptic processing area and broader facility. For example, the grouping of multiple areas by classification may be appropriate, (e.g., all Grade C areas or all Grade D areas grouped together) but only if those areas are all utilized in a similar manner.
Consider a facility that utilizes Grade C areas for the preparation of non-sterile supplies for subsequent transfer into Grade B or Grade A areas. This same facility also utilizes other comparably classified Grade C areas as the background environment for aseptic processes conducted in Grade A biological safety cabinets (BSCs). Those Grade C surrounding areas that are designated as the background for the BSCs employ additional contamination control measures to protect the integrity of the aseptic process. These additional controls are not employed in the Grade C area used for non-sterile material preparation. Under this scenario, it is more appropriate to trend these two types of areas separately for two reasons. First, the counts are likely to be consistently lower in the Grade C rooms with the BSCs because of the stricter contamination controls. Second, the potential impact of an excursion in these tightly controlled areas is higher due to the potential impact a contamination event would have on the aseptic process being conducted in the BSCs.
Setting and Maintaining Appropriate Alert and Action Limits
How alert levels and action limits for EM parameters are set and managed is also critical to successful trending. Current regulatory guidance documents, including the FDA Aseptic Guidance document1 and EU Annex 12 and informational chapters such as USP <1116>3 suggest maximum allowable action limits for parameters such as airborne viable and non-viable particulate levels, surface monitoring, and personnel monitoring. These environmental monitoring limits are generally set based on the activities that are expected to be performed in the corresponding classified area. Many firms traditionally adopt the regulatory action limits indicated in the guidance documents as their own internal limit. Initially, in new facilities or newly constructed areas, alert levels are often established at 50% of the recommended action limits. Regardless, once a reasonable amount of EM data is collected, alert levels, should be set using a statistical calculation based on that collected data. Parenteral Drug Association (PDA) Technical Report #13 (Revised Edition)4 recommends a few statistical methods that may be used to calculate a cut-off value or normal distribution approach. Similarly, a Poisson distribution may be used in areas with low recoveries such as Grade A or B environments.
Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special offers
from American Pharmaceutical Review – all delivered right to your inbox! Sign up now!
It is crucial that both alert levels and action limits be established at setpoints appropriate for the specific process being conducted or the intended usage of a classified area. The values must be set so that they will accurately indicate drift from normal operating conditions or other baseline levels. In some instances, the actual operating levels are much lower than the recommended industry action limit. In such cases, it is feasible that the action limit will never be reached or exceeded unless there is a catastrophic system failure. If there is a large gap or range between the alert level and action limit, aberrant results that would normally indicate the beginnings of such a failure may be inadvertently addressed simply as an alert level excursion or insignificant “isolated event” unless the corresponding procedures specifically state criteria for what is to be considered “aberrant” and whether an investigation into the root cause is warranted. Both alert levels and action limits should be reviewed on an annual basis, at a minimum, to ensure that they still reflect the normal operating levels. If at any time an alert level or action limit is found to be clearly inappropriate or inadequate, there should be a procedure in place to allow for adjustment of that limit outside of the routine annual review.
Grouping of EM data can also have a significant impact on the appropriateness of calculated alert levels and action limits, especially if that method is based on average or mean counts (e.g., the mean plus or minus one or two standard deviations) If EM data is inappropriately grouped, the mean can be artificially skewed in one direction or another. If an area with consistency higher counts is grouped with areas which typically yield lower counts, the mean may be skewed toward the lower counts, possibly resulting in an artificially low alert level or action imit. There is also the possibility of the reverse scenario, resulting in alert levels and action limits that are set too high, rendering them less effective as a means of monitoring the effectiveness of contamination control strategies. Other deleterious effects such as masking of atypical counts or adverse trends (such as increasing microbial counts) through dilution can also occur. If the alert level is set too high, there is not sufficient opportunity to implement an appropriate corrective action to offset a potential action limit excursion.
What is the Correct Trending Frequency?
How often trending is performed is another critical consideration, and should be appropriate to the type, amount and frequency of both production activities and environmental monitoring. Many firms perform trending on a quarterly basis based on the frequency of exceeded alert levels or action limits. USP <1116>3 specifically recommends that aseptic facilities tabulate recovery rates on a rolling monthly basis for classified areas. However, whether these frequencies are appropriate can depend on the amount of collected data. If the amount of collected data is too small, it can result in artificially inflated indicators that must be addressed. For example, if a firm has only one Grade A room, and zero colony forming units (CFU) have been recovered in one month and one CFU is recovered the next, that represents a 50% increase in counts. It may also encourage improper grouping of areas to increase the amount of data available for analysis. The selected trending interval should allow for a proper statistical evaluation to be performed on a routine basis. Should a significant event or adverse trend occur, it is always possible to increase the frequency of trending either temporarily or permanently, as appropriate, to address such issues.
For example, it is not unusual to evaluate upwards of two years’ worth of data when trending sterility failures because of the infrequency of the event. Conversely, if the collected data indicates that microbial counts are increasing in a particular area, the trending interval may need to be increased (e.g., from quarterly to monthly) to allow for more proactive approach to be taken and ascertain the effectiveness of the corrective action.
The Design of the Trending Report
It is not enough to simply report the trending results from one interval to the next. The data needs to be analyzed and discrepancies must be investigated and explained. If there is an increase in microbial counts from one quarter to the next, why did that occur?
Any alterations to procedures or change controls that could potentially impact the state of control of the facility in any way should be noted in the trending report along with any excursions or aberrant results that occurred during that particular interval. Examples include, but are not limited to, the presence of visitors or contractors in the facility, significant repairs that were made, the presence of additional or new personnel, power failures, changes in cleaning procedures, frequencies or agents used, and any identified training issues or lapses in technique. If any corrective and preventative actions (CAPA) were implemented, there should also be a review of CAPA efficacy, as appropriate.
Data should be trended in a consistent manner at each trending interval. This allows the user to monitor the existing state of the facility and to see evolving changes in the environment including elements such as seasonal variations, occurrence of new floras, changes in the amounts of common floras, etc.
The data should be presented in a clear, concise and logical manner that makes it easy for the user to draw conclusions and to scientifically support and justify any decisions made. Tables and graphs, pareto charts, pie charts, etc. are all very useful in accomplishing this. It is fundamental to present all of the data as it actually exists, so that the state of control can be completely and accurately assessed. If graphic or tabular presentations of the data are used, each event should be accompanied by a thorough explanation of what the data demonstrates and its significance in terms of the facility state of control and the efficacy of the contamination control program (or lack thereof).
This clarity is also critical when considering who will be responsible for reviewing the report. It is important that those responsible for reviewing trending reports possess the knowledge and experience necessary to understand the overall significance of the reported results, although they may not necessarily be microbiological subject matter experts. Reviewers must be able to clearly understand what the data demonstrates, how conclusions were drawn and supported, and why any decisions made or corresponding CAPAs implemented were necessary, appropriate and effective.
Conclusion
A trending report can be a truly useful tool in assessing the state of control in the facility, the efficacy of the contamination control program, and making sound data-driven and scientifically justified cGMP decisions. Developing such a report requires careful thought and consideration during the design of the report. Failure to do so will result in a report that will take a great deal of time and eff ort to prepare but will yield little useful information. Considering the points presented in this article will provide the microbiologist with a good starting base for developing this truly useful tool.
References
- United States Food and Drug Administration, “Guidance for Industry. Sterile Drug Products Produced by Aseptic Processing – Current Good Manufacturing Practice,” (FDA, Rockville, MD, August 2004).
- EudraLex. (2008). The Rules Governing Medicinal Products in the European Union, Volume 4, EU Guidelines to Good Manufacturing Practice Medicinal Products for Human and Veterinary Use; Annex 1–Manufacture of Sterile Medicinal Products. https://www.gmp-compliance.org. Accessed July 23, 2020
- U.S. Pharmacopoeia, USP42-NF37 2S (2020). < 1116> Microbiological Evaluation of Clean Rooms and Other Controlled Environments. https://online.uspnf.com/uspnf. Accessed July 23, 2020
- Parenteral Drug Association Technical Report No. 13. (2014). Fundamentals of an Environmental Monitoring Program (Revised Edition). Parenteral Drug Association, Bethesda, MD.