Uncommon Sense in Execution of Process Simulations

The aseptic process simulation (or media fill test) has been a compliance expectation for aseptic processing operations since the 1980s. Industry guidance from the Parenteral Drug Association (PDA), Pharmaceutical and Healthcare Sciences Society (PHSS), International Standards Organization (ISO), Food and Drug Administration (FDA), European Medicines Agency (EMA) and Pharmaceutical Inspection Cooperation/Scheme [PIC/S] regulatory outlines have endeavored to establish the “what” and “how” of their execution. The PDA’s latest guidance outlined the subject in a comprehensive manner addressing all of the major elements of the simulation design and execution.7 It might be expected that given the extent of the available content, that no controversies or confusion would exist regarding process simulations. Unfortunately, that is not the case. There are a number of areas where interpreting the published information and applying it in the real world has proven to be problematic.
 

The sterility test uses two media, Soybean-Casein Digest Medium (SCDM) and Fluid Thioglycollate Medium (FTM), to mimic the potential growth conditions present in the human anatomy. Early guidance documents on process simulation indicated that both media were to be used in evaluating an aseptic process. More recent guidance has largely eliminated FTM from consideration recognizing that true anaerobic conditions cannot be attained in the vast majority of conventional aseptic filling operations — even those with inert gas purges and/or lyophilization. Anaerobic conditions (oxygen levels NMT 0.2%) can only be attained in isolator systems with a total inert gas environment. For all other installations, SCDM should be the only media used, and oxygen should be substituted for inert gas. This approach will detect facultative anaerobes, which are the only microbes capable of growth in FTM.

Standard practice in the design of process simulation is to bracket the vial sizes normally processed on an individual filling line. Thus, a filling line used for 1, 2, 3, 5, 6, 10, 12 and 15 mL would ordinarily use the 1 and 15 mL vials in the simulation. Consider, however, that several of the containers might differ only in height, thus the 3 mL vial might have a higher center of gravity than either the 1 or 2 mL containers and be more susceptible to tipping over requiring added human intervention. Its inclusion in the media fill program design is essential to support the added activity necessary. The smaller 1 mL vial might present its own set of unique difficulties and also warrant inclusion. The preferred approach would include both the 1 and 3 mL vials representing the lower end of the range. Size alone should not dictate the simulation design.  

Container/Closure Handling

As noted, the selection of process simulation test components is commonly driven by container size. Handling considerations have recently been given mention to ensure that those components requiring added handling for proper feeding are considered. In order to make such a determination, data must be gathered on the performance of components on the filling line. Filling batch records should require the collection of intervention data in a manner that allows for the identification of those components that cause the highest frequency of corrective interventions on the filling line. This must extend to both containers and closures, and perhaps their combination to ensure that the program design incorporates the materials that create the need for greater operator intervention. 
 

The core concerns in all manned aseptic processes are the activities performed by the operators. As the gowned operators are universally acknowledged to be the predominant source of microbial contamination in the entire activity, it stands to reason that everything the operator comes into contact with is placed at risk. There are three categories of operator activity in aseptic processing:

•  set-up — the preparation of the line from individually sterilized components 
•  inherent interventions — the activities required to operate the line, i.e., component replenishment, environmental monitoring, etc.
•  corrective interventions — the steps taken in response to a system failure, i.e., vial breakage, stopper jam, etc. 10,11,12  

There are important considerations to be taken in each of these areas that would benefit from clarification. The specifics of each activity should be reviewed from a microbial contamination potential, operator safety and cGMP conformance perspective to provide a “best practice” approach to their execution. 

The task of assembling/positioning and adjusting the individual sterilized filling parts should be considered a specialized and singular inherent intervention. Its execution by an experienced operator is the first aseptic step in the process and often entails extensive interaction with sterilized materials. The steps taken to ready the line for use should be detailed and the operator proficient in the entire sequence. The last portion of the set-up includes: introduction of container/closures; confirmation of their proper feed through the line; adjustment of fill weights/volumes, closure placement/seal force; and removal of the set-up units. Preference should be given in equipment design and selection to filling equipment that simplifies any aspect of the set-up to reduce the interventional activity. Environmental monitoring should be performed during the set-up, and personnel monitoring upon completion. 

There are tasks involved in operating a filling line that cannot ordinarily be eliminated in which the operator must interact within the critical zone. These tasks are ordinarily repetitive and/or periodic in nature and include: feeding components, taking environmental samples, operator breaks, etc. The execution of inherent interventions should be performed in a singular manner by all operators. The operators’ location, hand placement, speed/sequence of execution and all other aspects of them should be nearly identical on each execution regardless of the operator. A parallel can be drawn to the first ball thrown in each frame in bowling. Regardless of who is throwing the ball, it must arrive in the 1-2 or 1-3 pocket to achieve the ‘best practice’ result of a strike. Allowing operators to perform inherent interventions in different ways invites contamination as compared to the “best practice” procedure they have been instructed on. 
 

The frequency present in inherent intervention should be considered flexible. Where a filling machine has demonstrated outstanding weight/volume control, the interval between weight sampling can be lengthened. Similarly, superior environmental control might provide an opportunity for less frequent invasive environmental monitoring. Automated equipment such as in-line weight checkers and electronic adjustment can replace operator interventions for both sampling and adjustment of fill weight, in much the same way the introduction of the dry heat tunnel can eliminate the need for operator transfer of vials depyrogenated in a dry heat oven.

The need for corrective interventions is largely driven by faults in the system. These can be the result of several factors including: inadequate equipment, deficiencies in the set-up, and component handling (resulting from lax quality standards). Corrective interventions are never desirable and their incidence rate in any lot can vary substantially given the number of potential causes for them; they are also variable with respect to the location and severity.  

Consider the following scenarios relating to a fallen vial and the expected operator corrective intervention: 

•  Vial falls over — Remove the vial, leave everything else on the line alone to the extent possible.
•  Vial falls over and breaks — Remove the vial and glass fragments; remove any other potentially impacted vials.
• Vial falls over, doesn’t break, but spills — Remove the vial, any other potentially impacted vials and remove liquid.
• Vial falls over, breaks and spills liquid — Remove the vial, glass fragments, any other potentially impacted vials, and remove the liquid.

Each of these interventions might be performed differently on different parts of the fill line. It should be immediately evident that corrective interventions have to be approached with greater flexibility in mind. The operators will need flexibility in dealing with corrective activities, and their training/execution must provide for more general guidance. The procedures for corrective interventions should be more principle-based to allow adaptation to varying circumstances.  
 

One aspect of wrong-headed corrective interventions is a mandate for the removal of a specific number of units or clearance of a specific zone. Blanket statements of this nature should be avoided, because they can result in increasing the duration/risk of an intervention by mandating more operator activity in the critical zone. The direction should be to remove the least number of units possible focusing only on those that are potentially impacted. If a single vial falls over or has a misaligned stopper, its removal may be sufficient to correct the fault. Removing adjacent units only serves to increase the risk of contamination to additional units. Holding to that practice suggests that the entire line be cleared one unit at a time, which is clearly an absurd response to a minor failure. The incidence rate for all corrective interventions should be recorded for each lot with continuous improvement practices defined to reduce the rate over time. 
 

The last consideration of corrective interventions that needs to be considered is whether a particular corrective intervention should be simulated at all. There may be instances where the execution procedure represents a contamination potential that is excessive. Just because something can be done, doesn’t mandate that it should. It might be preferable to terminate any filling process when an extraordinary situation presents itself rather than undergo some heroic measures to execute it once, and then endeavor to support it over time.   
 

Addressing interventions in media fill programs is actually quite simple. The set-up and inherent interventions associated with the media fill should be virtually identical to those in routine filling. The only differences that would arise with these should be the result of the process simulation itself: media introduction, replacement of inert gases with compressed air, etc. With respect to corrective interventions, the process simulation should incorporate these (real or simulated) at a rate that matches that observed in routine filling. This can be supported in interventions/number of units, intervention/time or interventions/batch. If the corrective interventions are properly recorded in routine operation, their inclusion at the proper frequency is straightforward. 
 

The criticality of interventions in aseptic processing and their simulation is such that the following bear repeating. 

• Interventions always mean increased risk to the patient.
•  There is no truly safe intervention
•  Interventions are to be eliminated if at all possible; if they can’t be eliminated, then their frequency should be reduced as far as possible; simplified with respect to their execution in all instances.
• The “perfect” intervention is the one that is designed out or eliminated from the process!

The length or duration of a media fill has always been a point of regulatory concern. In the 1970s, media fills rarely exceeded 3,000 units.1 When the statistical implications of using an acceptance criterion of a maximum 0.1% contamination rate were understood, media fills started to exceed 4,700 units. The FDA’s 2002 draft Aseptic Processing Guideline brought an increase to 5,000 units.13  More recently, there have been EMA and FDA expectations for media fills that exceed the maximum batch size allowed on the line. This premise behind these increasing size simulations is apparently rooted in the regulatory belief that the microbial conditions on the line slowly deteriorate over the course of the fill such that units filled at the end of the batch have a higher potential for contamination. There is ample evidence that this is not the case, and if one considers the fundamentals of aseptic process system design why such a concern is incorrect.14 The typical manned aseptic filling line whether conventional or RABS is designed to continuously eliminate contamination by the unidirectional downward flow of air through the critical zone. There are no reports of microbial contamination trending upward during an operational day, and clear support to the contrary. The appropriate duration for a process simulation should be longer than the operators’ continuous presence in the fill room, or 5,000 units — whichever comes first.  
 

One of the more puzzling aspects of process simulation execution is the continued use of 2 different (20-25°C and 30-35°C) incubation temperatures for seven days each for the filled units. The trouble lies in the dichotomy of the temperature sequence. Industry surveys beginning in 1980s indicated that the industry was split almost equally among firms that incubate first at 20-25°C and then at 30-35°C, and those that began at 30-35°C and then switched to 20-25°C. The arguments posed for doing this in either fashion are largely theoretical. Noting this, PDA suggested and FDA agreed to the use of a single temperature between 20-35°C with an allowable range of ±2.5°C.7 Using a single temperature simplifies the execution, eliminates the need for large incubators and more closely mimics what occurs with many products anyway. The persistence of the antiquated practices across the industry is puzzling given the operational advantages and regulatory acceptance of a single temperature.
 

Perhaps the most confusing aspect of all with respect to media fill execution is that of operator participation. The FDA’s 2004 aseptic guidance includes the following: 

“All personnel who are authorized to enter the aseptic processing room during manufacturing, including technicians and maintenance personnel, should participate in a media fill at least once a year. Participation should be consistent with the nature of each operator’s duties during routine production.” (7) 

On the surface, this seems completely clear. However, what exactly does “participation” in this context mean? For those individuals that do not perform interventions in the critical zone, the expectation is actually excessive. To mandate that supervisory personnel participate to the same extent they routinely do means they should stand to the side and little else. For other gown-qualified personnel sanitization staff, instrument mechanics and so on, that do not intervene on an operating line they should also be exempted from any direct involvement in the media fill program. That leaves the set-up personnel (if different from the operating staff), line operators and environmental monitoring staff (if present) as those whose participation is necessary to meet the regulatory expectation.  
 

The implementation of the regulatory desire is still not without some difficulty. In a single line facility, with a staff of four, two media fills would afford ample repetitive opportunity for the four operators to perform each of the permitted interventions without difficulty. Applying the guidance to a larger plant is far more complicated. Consider the following real world parenteral facility — eight filling lines, two operating shifts and approximately 50 line operators. Ignoring the participation concern, the facility would perform 32 media fills per year (8 lines X 2 shifts X 2 fills per year = 32). If there are 10 interventions to be included and they are performed on average five times each in the media fill, then 1,600 total interventions would be performed each year. Considering the 50 qualified operators, that converts to an average of 3.2 interventions per operator per year! Those 3.2 interventions would in theory support each operator’s participation on all lines for all of the different interventions. The confidence provided by this approach is clearly compromised and an alternative approach is necessary. Increasing the number of media fills to match the intervention numbers performed in the smaller facility would require 400 media-fills-per-year levels and is not a realistic alternative. 

The most appropriate means to address participation at larger sites is reliance on procedural conformance. If the means to execute the intervention is sufficiently detailed and the operators well trained in the procedure, then the aseptic process should be considered adequately qualified despite every operator not having performed every intervention on every line multiple times. As the individual operators can be gown-qualified and separately demonstrated to be proficient in the aseptic techniques necessary, their participation is assured through the procedural controls. 
 

The execution of process simulations is an essential part of the sterile product compliance demonstration. The details of their execution have been addressed in industry guidance documents, however the structure of those documents precludes detailed examination of the methods as afforded here.

1) PDA, "Technical Monograph No. 2, Validation of Aseptic Filling for Solution Drug Products," 1980.
2) Technical Report No. 6, Validation of Aseptic Drug Powder Filling Processes," 1984.
3) Process Simulation Testing for Aseptically Filled Products", PDA Technical Report #22, PDA Journal of Pharmaceutical Science and Technology, Vol. 50, No. 6, supplement, 1996.
4) Process Simulation for Aseptically Filled Products", PDA Technical Report #22, 2011 Revision, PDA, 2011.
5) PHSS, PS Technical Monograph No.4, The Use of Process Simulation Tests in the Evaluation of Processes for the Manufacture of Sterile Products, 1993.
6) ISO 13408-1:2008, Aseptic processing of health care products -- Part 1: General requirements, 2008.
7) FDA, Guideline on Sterile Drug Products Produced by Aseptic Processing, 2004.
8) EMA, Annex 1 - Manufacture of Sterile Medicinal Products, 2008.
9) PIC/S, PI 007-6, Validation Of Aseptic Processes, 2010.
10) Agalloco, J., “Management of Aseptic Interventions”, Pharmaceutical Technology, Vol. 29, No. 3, p. 56-66, 2005.
11) Agalloco, J., & Akers, J., “The Truth about Interventions in Aseptic Processing”, Pharmaceutical Technology, Vol. 31, No.5, p. S8-11, 2007.
12) Agalloco, J., & Akers, J., “Revisiting Interventions in Aseptic Processing”, Pharmaceutical Technology, Vol. 35, No.4, pp 69-72, 2011.
13) FDA, draft Guideline on Sterile Drug Products Produced by Aseptic Processing, 2002.
14) Cundell, A.M., Bean, R., Massimore, L. & Maier, C., “Statistical analysis of environmental monitoring data: does a worst case time for monitoring clean rooms exist?”, PDA J. Pharm Sci Technol. , Vol. 52, No. 6, pp 326-330, 1998.
15) Agalloco, J., Akers, J., & Madsen R., "Current Practices in the Validation of Aseptic Processing - 2001", PDA Technical Report #36, PDA Journal of Pharmaceutical Science and Technology, Vol. 56, No. 3, 2002.