Modern Trends in Non‐Viable Particle Monitoring during Aseptic Processing

Introduction

This paper describes changes and improvements to non‐viable particle monitoring (NVP), sometimes referred to as total particulate monitoring, which is a regulatory requirement during aseptic processing. Aseptic processing is becoming more automated and increasingly important to future products of the biopharmaceutical industry. In the same manner, NVP monitoring is also becoming more automated and increasingly important for contamination risk management during aseptic processing.

Growth in Aseptic Processing

Historically, vaccine production has accounted for a large portion of aseptic production. Recent years have seen increased focus on new biologic therapies, most of which are administered via injection or IV. Major pharmaceutical companies are investing to drive the development and release of these new biologics. Since most of these will be produced with aseptic processing, consider the following trends:

• More than a thousand biologics are currently either approved or in clinical development¹
• Biologics are expected to comprise half of the top 100 selling drugs by 2016¹
• The molecular complexity of biologics means each product can have specific delivery needs¹ (i.e. specific sterile packaging, such as custom syringes or injection technologies.)
• Sales growth for small molecule drugs (i.e. pills) is stalling as sales grow strongly for biologics and vaccines²

Components for many biologic therapies come from cell line production activities. Gene therapies are developed and produced with specific cell lines requiring aseptic processing. Vaccine production continues to increase in the developing markets, and through new vaccine-based therapies, such as new cancer prevention treatments (e.g. HPV). New biologic therapies have arrived, or are in the approval process, for cardiovascular, musculoskeletal, gastrointestinal, central nervous system, hematological, respiratory, oncological, immune system, inflammation, and urological conditions. The approval rate for biologics is much higher than for new small molecule treatments. Due to the profitability of these new biologic therapies, investment is also increasing in biosimilars. All of these factors point to an increasing need for aseptic processing into the foreseeable future. Many of these aseptic processing facilities will need custom handling and packaging equipment, along with an increased use of automation.

^{Aseptic Processing Improvements and the Need for Event Detection}

It is widely accepted that human contact with sterile products and components is the leading cause of microbiological contamination in aseptic processing. To combat this, as new aseptic processing facilities are built, there is an increased use of automated equipment, isolators and Restricted Access Barrier Systems (RABS) to reduce human contact with sterile product. Reducing human contact is the contamination control priority, along with improving unidirectional filtered (HEPA) air flow coverage, such as 100% HEPA coverage for exposed sterile products and components.³ 

Figure 1. Modern aseptic liquid fill operation 

“The greatest single advancement in aseptic processing has been the elimination of interventions using various types of automation. Examples include, depyrogenation tunnels, automatic lyophilizer loading, clean-in-place/sterilize-in-place systems and automated weight check/adjustment.“³

From Akers’ and Allacogo’s article, “Clean Rooms, RABS and Isolators: Validation and Monitoring in the Diverse World of Aseptic Processing” These aseptic technology improvements have begun to appear in regulatory guidance, such as the EU GMP requirement, starting in 2011, to extend continuous Grade A air flows over stoppered, but uncapped, containers.

Figure 2. Modern isolated aseptic lyophilization operation

The result of these improvements in modern manufacturing environments is the virtual elimination of viable counts during processing. It is now difficult, if not impossible, to establish a statistically meaningful baseline of viable count levels during processing activities.

“Increasing monitoring intensity (i.e. viable EM [environmental monitoring]) in nearly all instances is only going to result in more zeroes, a phenomenon we’ve seen over and over again. Too often the most common intervention in aseptic processing is environmental monitoring, which is completely illogical. What this means is that the method erroneously relied upon to measure risk is itself the greatest source of human contamination risk!” ³

Due to isolation and improved laminar airflows, the focus of contamination monitoring has now moved from daily processing activities to uncommon events during processing, such as momentary disruptions of the environment or equipment, human interventions, failure to follow SOPs, etc. Event detection requires continuous sampling and non-zero results during events. To limit the exposure of sterile products to events, real-time event alarms are important. Everything is getting cleaner; the risks are now more random and less consistent.

The Role of Non-Viable Particle Monitoring

The emphasis in aseptic process monitoring is on viable particle EM, as this is the critical metric for optimizing and validating aseptic environments and processes during media fills. Viable particle monitoring during production provides a reference to our original validated aseptic conditions established during media fill studies.

In daily operation of modern aseptic environments most particles produced are sterile foreign material. These submicron particles are not detected by viable particle EM methods. Viable samples from operations typically produce a series of zero counts (CFU’s) and these samples only “spot check” aseptic conditions due to their infrequent (non-continuous) nature. Although this viable sampling is a recommendation of regulators, it provides no immediate information to capture environmental events.

Continuous non-viable particle monitoring (NVP), when properly configured, provides useful information on changes in the aseptic environment with minute-by-minute updates of count populations and trends, which catch contamination events and thereby identify and quantify risks (i.e. a non-zero result). If root cause investigations for these non-viable particle events identify human involvement in the event, biological contamination may be a possibility. If the event was caused by automated equipment or components, non-biological foreign material contamination may be a possibility. Therefore, assurance of minimum risk starts with continuous non-viable particle monitoring. Although viable particle EM is essential, particularly to qualify aseptic environments, it is a less reliable indicator of risk during sterile product processing, and it does not identify the risk of non-biological foreign matter contamination.

Foreign matter testing in liquid sterile products (via USP 788 testing of injectables) tests a small number of final products (product is consumed during testing). This is essential to qualify the process for large particle contamination from the process and packaging, but it is unlikely to catch foreign matter affecting a small number of products at risk during an event in a large batch. It is the combination of all test methods: viable, non-viable and USP, which creates an umbrella of protection during the production of sterile products. Each test method adds a layer of protection, but it is the unique role of non-viable air sampling to detect real-time changes and events that may compromise portions of aseptic lots.

“Where advanced aseptic technologies are used we should be relying almost exclusively on evaluation of physical parameters, and electronic particulate monitoring, which produce data that can be reviewed in real time rather than requiring several days of incubation.”³ Akers and Allacogo

“We believe that in advanced aseptic systems electronic total particulate analysis is largely sufficient to provide the information required to ensure the maintenance of a validated state of environmental control. If a microbiological sampling component is retained by regulators it should operate at a much lower sampling intensity than that currently expected in manned clean rooms.”3 Akers and Allacogo

Setting NVP Alarms to Capture Particle Events

Regulatory guidance recommends monitoring non-viable particle counts in two size ranges: particles 0.5 micron and larger, and particles 5.0 micron and larger. The FDA focuses on the 0.5 micron size range and the EU recommends monitoring both size ranges. Both the FDA and EU reference continuous monitoring of NVP during aseptic processing for the purpose of event detection.

Figure 3: EU Annex 11 Particle Limits
(Tighter than ISO 14644 standard for 5 um in Grade A/B)

“The Grade A zone should be monitored at such a frequency and with suitable sample size that all interventions, transient events and any system deterioration would be captured and alarms triggered if alert limits are exceeded.” 2009 EU Annex 11. This statement creates a requirement for continuous monitoring and alarms.

NVP monitoring systems update once per minute to quickly capture and alarm on particle events. Both the 0.5 and 5.0 micron size ranges are monitored every minute, and count results are provided for a cubic foot air volume, which is collected over the last minute, and a cubic meter air volume, collected over the last 36 minutes. The cubic foot volume is useful for event detection; the cubic meter volume is a comparable reference to the ISO cleanroom standard (ISO 14644) and EU non-viable count limits.

Setting Alarm Limits for 5.0 Micron Particles

Much has been written about how 5.0 micron and larger particles can be a food source for bacteria and how most bacteria are detected in this size range. However, particles 5.0 micron and larger are much less mobile in air than the smaller 0.5 micron particles. The 5.0 micron particles are therefore much more difficult to detect. It should be understood that airborne particle counters all need the particle to pass through the instrument to be detected and therefore particle ”suspension” and “transport” are important. In modern isolators and RABS, 5.0 micron particles are rare, even during events. The regulatory limit (EU Annex 11) for 5.0 micron in Grade A is 20 particles per cubic meter, which is less than one particle per minute in real-time NVP monitoring systems (e.g. 36 one-minute samples = one cubic meter of air).

A count alarm limit set at one count for a single sample is not statistically significant but due to the EU limit, many systems are set for Action alarms on one 5.0 micron particle. Ironically, some viable particle limits will accept one CFU per sample. A more meaningful Action alarm limit for NVP would be to alarm on three consecutive one-minute samples of one particle at 5.0 micron. This indicates a particle generation source that is significant. Any investigation of root cause should pay special attention to whether the event involved a human intervention. Also of note is that mechanical failures due to friction, such as a bad bearing on an automated robot, tend to generate large particles.

Setting Alarm Limits for 0.5 Micron Particles

Usually, particle generation sources tend to create more small particles than large ones, and small particles are easier to collect and transport in an airstream. Thus a particle generation event is more likely to be detected at smaller sizes. A particle event that shows 0.5 particles, but does not register 5.0 micron particle counts, may still produce particles of a larger size. The larger particles may not be counted because there are less of them and they do not travel as well in the air sample. Thus, more attention should be paid to the smaller 0.5 micron particles in NVP monitoring.

A common misconception is that Action alarm limits for 0.5 micron can also be set at the Grade A limit, which is 3520 particles per cubic meter. This translates to an Action limit of 98 particles per minute (i.e 1 cubic foot sample). This particle level is well above the average particle level for modern aseptic processing environments that are in operation (i.e. not “at rest”), which can average somewhere between zero and 10 particles per minute of the 0.5 micron size. If an aseptic process is averaging 5 particles per minute at a given sample location, a sample that jumps to 90 particles is an 18 times increase – but it may not trip an Action alarm. Such a large increase indicates a particle generation event; an event that may have also generated larger particles that may not have been detected.

The best practice for setting 0.5 micron NVP count alarm levels is to use the same practices we used for setting viable count limits in older environments containing humans:

• Use non-viable particle data from the media fill to determine baseline count level.
• Set the limit to capture events that are statistically significant deviations from the baseline.

In conclusion, use these two principles to detect potential NVP contamination events during aseptic processing (i.e. “in operation” alarm limits):

1. Place emphasis on 0.5 micron count alarms and set the limit low enough to detect deviations from the normal count baseline based on the results of the media fill study.
2. Set the 5.0 micron Action alarm levels to catch large one-sample events or multiple consecutive samples with single counts.

Modern NVP Monitoring Systems

Just as aseptic processing environments are improving, so also are the non-viable monitoring systems used to monitor them. The following trends are beginning to emerge in NVP monitoring of aseptic processes:

• Increased regulatory focus on reducing use of portable, human-operated, particle counters in favor of permanently installed, online, computer-controlled particle counters.

• Reduces need for operator presence in Grade A to place probes and take samples.
• Reduces need for operators in Grade B areas.
• Removes equipment from aseptic Grade B (i.e. large portables/carts) which can interfere with airflow and introduce additional surface areas to the sterile core.
• Improves consistency of monitoring to SOPs
• Removes operators from managing data (e.g. paper tapes, manual data uploads, etc.) to reduce transcription or collection errors.

Figure 4. Permanent installation of online particle counter for continuous sampling during filling. Sensor is installed outside of Grade A/B – inside the filling equipment

• Increase in the number particle sensors, particularly in automated fill/finish operations. Modern aseptic lines, especially those involved in lyophilization, involve many robotic devices. Two or three filling lines may supply a common bank of lyophilizers. Automation includes automatic tray loading devices, mobile Grade A containment systems for moving partially stoppered product to lyophilizers, lyophilizer loading robots, automatic stopper inspection/rejection systems, automatic vial weighing devices, etc. Such an operation increases the size of the ISO5 Grade A and B areas, with many points for human intervention into isolators to service automation equipment during an event. Sterile product and components move greater distances and are exposed to more equipment (but less human contact). In this type of workflow, a risk analysis of potential contamination events produces more points that need to be continuously monitored.

Figure 5. Screen capture from monitoring system for Lyo batch operation. (22 sensors controlled in three groups: Filling, Lyo Loading, and Capping)

• Use of recipe-driven monitoring systems based on workflow. Modern particle monitoring systems control sensors based on the workflow state. Monitoring software prequalifies sensor cleanliness (zero counting) and prequalifies the aseptic environment for At Rest cleanliness levels prior to exposing sterile product. During product processing, the right sensors are monitored based on which part of the operation is active (e.g. filling, capping). During Action level alarms, equipment is automatically paused until aseptic conditions are reviewed and approved.

• Batch-driven operation and reporting for non-viable/viable particle monitoring. Particle monitoring systems now automatically label data with batch/lot identifications and immediately produce comprehensive batch reports when the aseptic process is complete. These reports include sample data, statistics and alarm histories with assignable causes. If production is paused to perform an intervention, sample data is recorded for quality purposes but this data can be excluded from batch reports as long as sterile product was not exposed during the intervention. Viable samples can also be controlled from the non-viable monitoring system, including sample plate start/stop times, air sample volumes, operator per plate, flow alarms, and resamples, with all data recorded to ensure a complete viable/non viable EM program for the batch.

Figure 6. Screen capture showing batch monitoring process. (The right sensors and alarm limits are automatically used for each process step. Results are reported per batch.)

• Alarm notification and acknowledgment that captures root cause at the time of the event. Modern systems now require operators to identify the event associated with the particle event and this information is automatically stored and included in batch reports for review and approval. Batch reports also show non-viable particle trends, which help to determine if an event is an isolated, short duration event that is not statistically significant, or the result of slowly-increasing particle levels which eventually tripped an Alert or Action level alarm.

Modern continuous NVP monitoring systems automatically produce the data reports and alarm history to ensure continuous control of aseptic environments so product batches can be safely released. Secondarily, the comprehensive nature of the monitoring enables meaningful analysis of particle events to drive continuous reduction of risks to sterile product.

Advantages of NVP Monitoring in Modern Aseptic Areas

As the biopharmaceutical industry becomes more dependent on aseptic processing for new therapies, the careful implementation of non-viable particle monitoring uniquely offers the following advantages for automated and isolated aseptic areas.

1. Detects all particles at defined risk points
   • Detects both viable & non-viable (foreign material) particles
   • Detects smaller particles which are more populous during events
   • Provides useful non-zero data during events
2. Does not require human intervention to sample isolated Grade A areas.
3. Continuously monitors to capture all events
4. Provides immediate results for real-time alarms to protect sterile product
5. Can automate monitoring and batch report generation to reduce errors and save labor

Summary

Modern non-viable particle monitoring systems can provide assurance of aseptic processing while reducing human involvement in aseptic areas for monitoring purposes. These systems can provide data to improve aseptic conditions over time, simplify root cause alarm investigations, and produce instant compliance reports for batch release—eliminating the labor associated with creating reports.

References

1. Excerpts from Unilife’s description of biologic drugs. http://www.unilife.com/blog/biologic-drugs
2. Summary of sales data from National Center for Biotechnology Information, a reprinted article from Biotechnology Healthcare: With Injectable Biologic Therapies on the Rise, Payers Face Tough Reimbursement Issues. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2899801/
3. Quotes from article on American Pharmaceutical Review: Clean Rooms, RABS and Isolators: Validation and Monitoring in the Diverse World of Aseptic Processing.
   http://www.americanpharmaceuticalreview.com/Featured-Articles/36878-Clean-Rooms-RABS-and-
   Isolators-Validation-and-Monitoring-in-the-Diverse-World-of-Aseptic-Processing/