Isolator technology for aseptic manufacturing was introduced in Europe over 20 years ago. Since that introduction, isolators have become a widely accepted environmental control system for aseptic processing. Certainly, isolators have run into a certain amount of turbulence in their trajectory to a well-accepted technological solution in 2010. The purpose of this communication is not to revisit the well-known story of Isolator technology in aseptic manufacturing, but rather to consider what we have learned over these two decades and to suggest how we might most productively and safely apply these lessons.
One of the interesting aspects of Isolator technology is that as we enter the third decade of isolator usage in aseptic processing industry, the regulatory authorities to some extent continue to treat these systems as though they are “new”. Certainly, isolators are newer then the manned clean rooms, which continue to be the most common controlled environment for aseptic processing. Realistically though, we can’t reasonably consider technology that has been around this long and has an installed base numbering in the hundreds as “new”.
Isolators in 2010 are really a rather mature technology and choosing to use them is not a brave step into the future. In fact, by most any reasonable technical measure, isolators at this stage of development can, believe it or not, be considered a mature technology. In contract, the human-scale clean room can most accurately be considered at this point in time to be an aging technology, perhaps even one whose retirement is on the horizon. Certainly, the sun is still a long way from setting on the human scale aseptic processing clean room, but it is safe to say that the shadows are starting to lengthen.
Validation and In-Process Control Standards for Isolators
At the beginning of isolator usage for aseptic processing, manned clean room process control concepts were applied virtually across the board for isolator technology. This did not even then seem logical to those of us involved with isolator technology. However, it is easy to understand why a simple importation of clean room design criteria, validation expectation and process control considerations occurred. There is, unquestionably, a superficial resemblance between an isolator and a manned clean room. The following is a short list of obvious similarities:
1. Air supplied through HEPA grade air filters.
2. Vertical unidirectional airflow (UAF) in many isolator systems.
3. Identical particulate air classification requirements applied (ISO 5 now, then Class 100).
4. Aseptic conditions claimed in both isolators and clean rooms.
5. Chemical disinfection utilized in both isolators and clean rooms including sporicidal treatment.
The first production isolators were actually flexible wall systems with turbulent air flow, however, not surprisingly it didn’t take long for rigid wall UAF isolators to become the aseptic processing norm. So, by the early 1990’s, the often spoken and written description of isolators as a “small clean room” seemed perfectly logical as did a direct importation of manned clean room standards.
Isolators are Really Different
Those of us who worked with early isolator systems in aseptic processing, and also in sterility testing, observed from the outset that there were truly profound differences between these systems and manned clean rooms. The most obvious of these differences was isolator’s whole reason for being namely the complete separation of human operators from the aseptic workspace. This separation removed the gowned operator as a mobile generator of microbial contamination and as a result in even rather primitive isolator systems by today’s standards, a very substantial reduction in environmental contamination was noticeable, and consistently reported. Strangely though, while the vastly improved environmental control was often mentioned and even led to the unfortunate suggestion that isolators could serve as a direct alternative to terminal sterilization, clean room design, validation and operational controls continued to serve as a reference point.
Isolators Need Their Own Performance Standards
It seems to this author that it is past time for the introduction of standards intended for isolators. These do not need to be rigid standards taking away flexibility in design or use. Doing so would be unfortunate and would perhaps result in the use of inferior contamination control technologies in some applications. However, the isolator is not a miniaturized clean room and there is no reason for the continued application of certain clean room standards. There are obvious design requirements for ISO 5 clean rooms that need not apply among them are:
1. 0,45m/s mean air velocity measured at any location within the enclosure.
2. Recirculation ratios of approximately 80% recirculated to 20% fresh air.
3. Any need to measure 5um particulate matter.
4. Airflow visualization in the complex manner expected in clean room aseptic processing.
Taking these items one at a time, air velocity is not as critical a factor in isolators and the user may wish to create a special environment for their product in which minimal air movement and exchange is required, for example inert gas, or very low humidity. High air exchange rates are important, in fact absolutely vital in manned clean rooms where continuous operator generated microbial contamination is in fact THE risk. Because isolators are unmanned, they do not have the same contamination risk as a clean room, and the HEPA filter in spite of speculation to the contrary is quite efficient at removing airborne contamination down to the sub-micrometer range. 20% fresh input air is vital in a clean room where human occupancy could lead to dangerous increases in CO2 levels. This is simply not an issue in isolators and higher recirculation rates can in fact be disadvantageous, in fact single pass air has proven beneficial in some applications. Thus, a recirculation rate of zero accompanied by an air velocity of ~0,2m/s or even less may be perfectly reasonable, or even superior. Word is already out that the 5um particulate size class will be removed as a classification option for ISO 5 clean rooms in ISO 14644, as well it should be. Far better counting statistics are available at for example the 0.3um particulate size class. 5um requirements, are still found in some standards, most notably EU Annex 1 and it must be said that at this point in time they are completely without value. It has been suggested over the years that the 5um particulate size class was likely close to the human skin cell upon which bacteria might be passengers. This idea always was completely hypothetical and in the isolator it is irrelevant because people are not in the aseptic processing environment. Finally, we should all know by now that “laminar” airflow exists neither in clean rooms nor in isolators. Some turbulence and eddy currents are a fact of life, which makes scoring of airflow visualization highly subjective. An argument can be made that air flow visualization has some value in clean rooms where fine tuning of air flow around obstacles and at positions at risk from human contamination may be of benefit, however even in this setting “smoke” studies are highly subjective in character. In isolators the value of such studies is greatly diminished and probably rendered insignificant. Contamination clearance studies may be of more value and even here the assignment of a fixed acceptance criterion is not possible since microbial contamination is a far lower risk issue.
Environmental Monitoring
Environmental monitoring (EM) is a staple in manned clean room process control. EM methods and philosophy have been imported from clean room use into isolators with only minor modification. Also, so-called acceptance criteria for EM in isolators have been applied more or less wholesale from the clean room standards. Application of standard clean room EM practices must have seemed reasonable to most regulators and many industry technologists 20 years ago, but in reality in never had the benefit of scientific merit.
Thirty or forty years ago, when clean room design was comparatively primitive and gowning materials several generations inferior to the current supply contamination was present in clean rooms at much higher levels than seen today. It must be said that in those days of limited processing technology automation a higher operator population in clean rooms was also a significant factor in the comparatively high contamination recovery rates observed in those days. Recently, in looking at some older files in my office, I came across notes taken nearly two decades ago on ISO 5 (then Class 100) aseptic areas and found that even then recovery rates were often > 2% of plates exposed, which is to say less than 98% of the air, and surface plates exposed were free of contamination, and personnel contamination rates were quite a bit higher. At present, the better clean rooms are, in my experience, consistently able to attain contamination recovery rates as a percentage of exposed samples of <0.5% and personnel rates are only slightly higher. Also, noticeable in comparing the old data with the newer was the much more frequent occurrence of significant excursions of 15 or more CFU in the “old days”.
Now let us consider the modern isolator, or for that matter a first generation isolator. Contamination in isolators was from their first days in sterility testing always considerably less than manned clean rooms. Isolators of even 20 years ago would easily outperform the best clean rooms of today in terms of contamination control as crudely measured by EM recovery rates. Airborne microbial contamination in isolators has always been an exceptionally rare event, and this is true even of the flexible wall turbulent airflow isolators used in sterility testing.
The greatest risk of contamination in isolators has been thought to arise from glove tears, separations or pinhole leaks. However, the switch from the Neoprene gloves used in the earliest isolators to Hypalon gloves has resulted in a dramatic reduction in glove-associated contamination. Also, the near universal use of a sterilized under glove and increased use of single piece glove/sleeve systems which eliminate separations of the glove and sleeve and the wrist, have further reduced glove related contamination introduction. At the present time, glove introduced contamination is approaching the level of airborne contamination recovery in isolators, which means it is rare enough that recoveries are exceptional.
What the EM Data Mean
One of the early reactions to the very low contamination incidence rates in isolators was the regulatory recommendation to increase monitoring intensity. It may seem quite logical to conclude that if recovery is very rare more sampling is a reasonable way to go. This is the old “if you can’t find something you should look harder” notion. There is a fly in the ointment though regarding this logic. We know very well in both chemistry and biology that if we attempt to use an analytical method below its limit of detection (LOD) we are exceedingly unlikely to recover anything.
This raises the question of what we should consider the limit of detection of growth and recovery based microbiological assays to be. Actually, the LOD of the methods we use in EM is not clearly known and logically is quite variable. Among the technical issues that could affect recovery are the selection of media type, lot to lot variability, sterilization method, incubation length, incubation temperatures, oxygen tension (organisms may grow best in the presence of standard, low or no O2) and last but far from least the condition of organisms likely to be encountered during sampling. By some estimates up to 50% of the organisms that may be present in a typical environment are viable but not culturable. This means that an organism might have measurable metabolic activity but is not normally culturable. Some environmental organisms depending upon their stress level may be culturable but not on the media or growth conditions provided them in EM sampling.
I find in my meetings with even experienced validation and compliance personnel in industry that there is a tendency to believe that we can actually come very close to measuring the sterility or sterility assurance of an environment. Actually, sterility is the complete absence of organisms with the ability to reproduce and sadly we have no method at our disposal, which has a sufficiently low LOD to measure the attribute of sterility. In fact, no assay biological or chemical has such ability because it would require a sample of near infinite volume to be processed by an assay with a LOD of one (1) viable cell. We simply don’t have that assay, so we have no choice but to conclude that EM can’t get us to “sterile”. In fact, it can’t come particularly close. In fact, although rarely considered and almost never discussed, there are entire classes of human pathogens that EM simply can’t detect, a classic example being viruses.
Therefore, although we might be tempted to conclude an investigation by writing that an aseptic processing ISO 5 environment was “sterile” because we recovered no microorganisms during EM, this is a temptation that should for utmost technical veracity be resisted. It may very well be possible in some clean rooms to conclude based upon EM to conclude that on a given occasion the room was operating outside a normal level of control, but no clean room is ever sterile.
Zero Doesn’t Mean Absent
It is an immutable truth of analytical science that a value that is below the limit of detection does not mean zero. This is not as has often been stated an inherent weakness of microbiology. A chemical assay may have a limit of detection of 50ppb, but if the objective is to reach an active pharmaceutical ingredient carry over level of <50ppb in a cleaning study you will not be able to prove success using an assay reliable “only” to a detection level of 100ppb. You would need a more sensitive assay to reach your acceptance criterion as we can clearly see. Fortunately though in cleaning validation we aren’t ask to reach a carryover level of zero (0). Instead, we set our acceptance criterion to a value known to be safe for a patient based upon their body weight, and what we know of general toxicity, teratogenicity, carcinogenicity or other health risk factors considered pertinent.
Only in the area of sterility do we imagine that somehow we can measure the existence of a sterile environment and use microbiological assays to prove “sterility assurance”. Thus, to the extent our regulations are based upon this absolutist perspective we are in effect being asked to do the impossible and although it may seem counterintuitive doing more sampling will not get us closer to the goal of proving sterility. In fact, in the clean room environment where human operators must do the sampling, we may, through our desire to prove the “sterility” of our processes, actually put them at greater risk regarding human released contamination. We can’t measure zero and this holds true no matter how much effort we expend in the form of sampling intensity. In recent years, there has been much talk about “continuous sampling”, but this takes us no closer to being able to measure sterility. Scores of continuous sampling systems running in parallel could be required to test even 5% of the air entering an aseptic critical zone and none of these sampling systems would have an LOD low enough to ensure that even the air they did sample was sterile.
What Does All This Mean?
What it must logically mean is that like most things in life there is a limit to the benefit accrued by EM sampling. It simply does not make sense to keep requiring higher and higher levels of EM intensity, because the ratio of samples that recover contamination to those that don’t is going to stay pretty constant and that constancy is related to the sensitivity of the method more than anything else. As is the case in all things in life going beyond this point of diminishing return will accomplish quite literally nothing, other than raise the cost of process control proportionate to the increasing intensity. Thus, the cost of goods produced is increased with no net benefit to society.
The Point of Diminishing Returns in Isolators
A reasonable question to ask then is what the point of diminishing returns is in isolator EM? Scientifically this is not a difficult question to answer in general terms. Given that we cannot measure zero contamination and given the limitations in LOD we should be doing a lot less monitoring in an isolator than in a conventional manned cleanroom. This is simply because the isolator environment is so clean that we are likely to get positives no more often than one in 10,000 samples. We should not consider the isolator sterile of course, but we should be able to conclude that our two decades of experience confirms that they are more than safe enough. In fact, product made in modern, well-designed and operated clean rooms is medically safe and isolator manufactured product is safer still.
So, Here is a Modest Proposal:
1. No more than two 1m3 air samples need be taken in a reasonably sized filling isolator over a four-hour period.
2. Indications are that active air sampling is more sensitive than settle plates so there is no good reason to do settle plates at all.
3. Microbiological sampling of surfaces that have been decontaminated with VPHP is unlikely to yield anything and should be minimized.
4. Physical measures of isolator performance may be of more value than microbiological measures. In other words if an isolator air filtration system is working well, pressure differentials are properly maintained, gloves are integral contamination risk is low.
5. Total particulate counts using electronic devices are likely to give more direct and more immediate indications of changes in isolator performance status than microbiological sampling.
It follows then that we should be looking to minimize the logistical difficulties in moving EM samples into and out of isolators and perhaps we should use robots or automation to do the sampling we do perform.
Conclusion
We should as an industry engage in open discussions with regulators and standard setters regarding EM as applied to aseptic processing. These discussions should be based upon where we currently are in the evolution of technology. Separative technologies such as isolators are very low risk regarding microbial contamination and we should be able to right size the EM program based upon our experience and the realities of analytical performance. We should also start from the very reasonable perspective that we’ll not be able to measure the attribute of sterility using EM no matter how intensely we endeavor to monitor.
As scientists and technologists, we should reach at some point general agreement that it will eventually be logical to phase out microbiological EM in isolators and probably other advanced aseptic processing technologies as well. Paradigm shifts are a fact of technological life and just because we’ve always done EM and we tended to do more of it, as time progressed, doesn’t mean this is a pattern that must be followed now and forever more. More and more EM appears to be a regulatory compliance habit, and it is time to get off this merry-go-round.
I fully understand that we aren’t yet at the point where every stakeholder in this process is ready to agree that EM can be dramatically reduced or eliminated in isolators. We have ways to do things in aseptic processing that have evolved in many and varied manners over a long time, and it was natural to move these processes into newer technologies as they were introduced and continued to evolve. Logically though in this era of risk and science based process development, validation, in-process control and regulation, and with the increased interest in statistical process control and Quality by Design, we should be in a position to have a discussion centered on science and engineering principles.
The object should be to go where the science is taking us not merely to do what we have always done. If we had technology to measure all the air entering an isolator for “sterility” I’d be the first to support its implementation. However, at the same time, what good is there in attempting to measure a nanometer with a meter stick? The absolutist approach to aseptic processing validation and in-process control is a dead-end street, and a mighty expensive road to travel at that. So, lets replace the absolutist approach to aseptic processing with a pragmatic approach that carefully considers what we can measure and equally importantly what we can’t.
References
1. FDA, Guideline on Sterile Drug Products Produced by Aseptic Processing, 2004.
2. FDA, Pharmaceutical CGMPS for the 21st Century - A Riskbased Approach, 2004.
3. Design and Validation of Isolator Systems for the Manufacturing and Testing of Health Care Products. Technical Report No. 34 PDA, Inc. Bethesda, MD USA 2000.
4. Agalloco, J., Akers, J., “Simplified Risk Analysis for Aseptic Processing: The Akers-Agalloco Method”, Pharmaceutical Technology, Vol. 30, No.7, p. 60-76, 2006.
5. Katayama, H., et al, “Proposal for a New Categorization of Aseptic Processing Facilities Based on Risk Assessment Scores”, PDA J Pharm Sci and Tech, Vol. 62, No.4, p. 235-243, 2008.
6. EMEA, Annex 1, Sterile Medicinal Products, 2008. 7. Sigwarth,V., Gessler, A., and Stark,A., “Relevance of Physical Glove Integrity Testing to Microbial Contamination of Isolator Systems,” presented at ISPE meeting, Prague, Czech Republic, 2005.
James E. Akers, PhD, is President of Akers Kennedy & Associates, Inc., located in Kansas City, MO. Dr. Akers has over 25 years experience in the Pharmaceutical industry and has worked at various director level positions within the industry and for the last decade as a consultant. Dr. Akers served as President of the PDA from 1991 to 1993 and as a member of the PDA Board of Directors from 1986-1999. Currently, he is Chairman of the USP Committee of Experts Microbiology and Sterility Assurance, as well co-chairman of the PDA Isolator Technology Task Force, Aseptic Processing Task force and member of several recent program committees.
Dr. Akers has lectured world-wide, and taught numerous pharmaceutical technology courses world-wide including training for the US/FDA. Dr. Akers has also authored over eleven textbook chapters, edited a book on isolation technology, has a second book on isolator technology in preparation, and authored more than 100 technical and review articles on a variety of subjects including validation, aseptic processing, contamination control, environmental monitoring and control, biotechnology, isolator technology, sterilization and disinfection, sterility testing, media fill testing, HACCP analysis, pharmaceutical microbiology and regulatory compliance.