Achieving Balance in Sterile Product Manufacturing

Peer-Reviewed:
Submitted: Feb. 9, 2015. Accepted: May 29, 2015.

AbstractValidating sterile product manufacturing processes requires the use of biological indicators to demonstrate the lethality of sterilization techniques. The indicator organisms are there merely to measure the effect of the sterilization process. Today, too much focus on lethality to biological markers results in use of extreme sterilization conditions; insisting on excessive lethality requirements in sterile process validation threatens patient safety, the authors believe. Not only does the practice subject process equipment and raw materials to damaging conditions, it can lead drug manufacturers to opt for aseptic processing in cases where terminal sterilization would be the best and safest choice. Companies should be free to choose the best way to ensure a safe process, whether that is aseptic processing, terminal sterilization, or a combination of the two.

Since the mid-1970s, validation has been required for all processes that directly impact the microbiological safety of the materials being produced in sterile product manufacturing (1). The authors are concerned that in validation in general, and particularly in the area of sterilization, rigor is often applied to the wrong things. As a result, extreme requirements are dictated with the assumption that doing more will always provide a greater margin of safety. For example, if the probability of a non-sterile unit (PNSU) being present is 10^-9, that figure is “better” than 10^-6 but, from the patient’s safety perspective, the extra rigor is not significant and hence meaningless in a practical sense. In fact, a compulsive adherence to rigor for its own sake may actually increase patient risk by making terminal sterilization or other risk-management opportunities less accessible to the manufacturer, as will be explained in this article.

Compliance expectation is not properly aligned with science and engineering. As equipment, process control, and technical capabilities have been steadily improving, conventional process control and validation expectations have fallen behind. Conventional practices in sterilization and environmental monitoring have thus resulted in errors (2–4).

Over the past two decades, rules and requirements have been added in the mistaken belief that they will result in more rigorous safety outcomes. In fact, all that they have done is to draw focus away from the actual process, increase the costs of goods produced, and inflate the probability that perfectly safe product will be considered unsafe.

Comparing sterilization philosophies
Since the advent of validation in the healthcare industry, sterilization process design has focused on increasing the reliability of lethal sterilization processes. This focus was appropriate in the 1970s, when the concept of validation itself was new. However, once process capability, robustness, and reliability have been established and validated, further gains cannot be achieved by adding more requirements and imposing greater regulatory rigidity. The goal of “greater” confidence in sterilization processes has been pursued without evidence that it is necessary, resulting in unnecessary constraints that have not improved patient safety.

For example, process dwell cycles have been lengthened, minimum lethality requirements increased, and additional (and sometimes questionable) controls imposed on virtually all aspects of sterilization processes (5, 6). These changes have resulted in unnecessary increases in process lethality without considering whether the materials being sterilized will remain acceptable for their intended use.

With the exception of stainless steel and a few others, materials that are sterilized are subject to some adverse effects during the sterilization process. For example, excessive heat can be detrimental to the chemical/physical properties of pharmaceutical formulations, decreasing their shelf lives and increasing degradation. Heat can also adversely impact glass containers, creating lamelli, etching, and increased numbers of particles. Extreme conditions can alter the physical properties of elastomeric materials (e.g., closures, gaskets, hoses, filters) with potential changes in durometer. Excessive radiation causes substantial physical and chemical changes in nearly all materials. Extended gas, liquid, or vapor sterilization processes can result in corrosion, excessive adsorption, and chemical reaction in the materials and equipment. Increasingly restrictive filter pore size reduces throughput, increases operating pressure, and can increase extractable/leachable materials in the fluid.

A realistic safety margin can be added to a process, but requiring lethality is useless beyond a certain point. The food industry takes this approach to sterilization, by making it a priority to maintain the organoleptic properties of the food materials. Within the healthcare industry, however, this concern is largely ignored, and the effects of harsh process conditions on materials are rarely considered, except in terminal sterilization by moist heat or radiation.

Extreme sterilization processes may deliver massive spore log-reduction levels of a highly resistant biological indicator microorganism, but they create two problems:

• Adverse impact on product quality, shelf life, and other properties when an overly rigorous process damages materials

• Greater risk for patients when a manufacturer opts for an aseptic process when terminal sterilization would be better.

The proper approach to sterilization must balance the needs to deliver lethality and preserve material quality attributes. Attaining this balance requires greater knowledge of sterilization processes, specifically, their effect on microorganisms and on the materials to be sterilized.

Recently, the United States Pharmacopeial Convention (USP) has taken steps to rectify the negative consequences of over-processing during sterilization by focusing greater attention on pre-sterilization bioburden (7, 8). Routine sterilization processes should aim to destroy bioburden to a probability of at most one non-sterile unit in 1,000,000. The identity, population, and resistance of the biological indicator are less important, because the organism does not define the sterilization process, but, rather, measures it. Sterilization processes should be no more lethal than necessary to reproducibly destroy the bioburden to safe levels, while maximizing the preservation of the material’s essential quality attributes. After all, one can only kill a microorganism once. We can speak of “overkill,” but in the real world, can one “overkill” a microbe that is already dead? The stacking of “worst-case” conditions on top of existing “worst-case” considerations benefits no one. In the authors’ opinion, any sterilization process that exceeds an F₀ of 10–12 minutes is excessive, and has been developed with excessive attention to factors (often, real or perceived regulatory expectations) that are unrelated to patient safety. Some practitioners and regulators impose unscientific regulatory requirements without understanding the microbiological principles at work in sterilization and patient safety or, perhaps, without fully considering the potential adverse consequences of over-processing.

It is possible that materials and products can be safely sterilized using methods that fall well outside the current expectation of lethality. It is essential that the industry remain open to scientific and technical advancements, such as lower thermal-input approaches, to ensure the best possible patient outcomes.

Aseptic processing: proven safe but not sterile
An aseptic process, which achieves safety by excluding microorganisms from products, relies heavily on the ability to sterilize and depyrogenate the materials being assembled into a sterile dosage form. Subsequent to sterilization, the now sterile items must be used under carefully controlled conditions. The introduction of new technologies in aseptic processing has paralleled that in other industries, and the operational improvements of the past 20–30 years, such as automation and separation of personnel from the process, will no doubt continue (9).

Although advanced designs are increasingly prevalent, present-day aseptic processing is performed using approaches ranging from completely manual to highly automated. With comparatively few exceptions, aseptic processing systems have consistently produced sterile products. Questions raised regarding the acceptability of an aseptic production facility have been based upon inferences drawn from monitoring or documentation practices. Contamination in aseptic products, or, perhaps more correctly, the perception of contamination, is often the result of well-intended, but scientifically unproven inferences on the part of regulatory investigators (2–4). It is the authors’ concern that unreasonable expectations may lead to the over-interpretation of limited, subjective, or analytically ambiguous data, resulting in an unwarranted assertion of a “lack of sterility assurance.”

At the core of these expectations lies a basic error-that “aseptic” means “sterile” or that the environment must be sterile in order for acceptable aseptic processing to be executed. While actually being sterile in an aseptic operation would be ideal, the industry lacks the means to prove that sterility exists because it can’t objectively demonstrate a negative absolute. To impugn aseptic operations that are not “sterile” is to ignore reality.

No amount of monitoring, process simulation, or sterility testing can assure sterility in an aseptic environment. The reverse is also true: the detection of microbial contamination in an environmental sample does not establish that contamination will be present in units filled in that environment. The opportunities for adventitious contamination are ever present, and sampling of aseptic environments is both analytically and statistically limited. Further uncertainty arises from the ever-present possibility that any given positive sample may be a false positive result.

Although the authors respect the concept of erring on the side of safety, the practice can be taken too far. The objective in aseptic processing must be safety to the end-user; safety and sterility are not the same thing. The greater risk to a patient population may accrue from product unavailability as opposed to the surmised, but generally impossible to prove, “lack of sterility assurance.”

Aseptic processing of sterile materials is a complex task, and many influences must be taken into account to advance the concept of sterility by design (SbD) (10). Consider the following:

• Environmental monitoring, process simulations, and sterility testing are not control measures, and do nothing more than provide a limited means of assessing process and, to a lesser extent, facility capability.

• Success in an aseptic process is not about zero counts. There are more important and direct considerations to address for a satisfactory outcome.

• Environmental monitoring is not an accurate way to measure either product safety or sterility assurance, and further increases in monitoring intensity will not alter this fact.

• Success is understood to rely on maintaining operator separation and intervention control.

• Where aseptic processing can be conducted with the greatest safety, it is because of equipment improvement and operational controls. Sterility assurance should be considered an engineering challenge, not an analytical microbiology problem.

• At the commercial scale, the contamination capability of aseptic processing is currently estimated to be 1/10,000 units or better. The preponderance of evidence indicates this is sufficiently safe for public health.

The industry should not create the impression that aseptic processing, when it uses current state-of-the-art practices, is an unsafe or safety-limited technology. There may be legitimate concern about contamination risks when a process includes a series of open and rather complex operations, or when final dosage forms are produced manually without the use of separation technology, such as isolators. Performed in accordance with current good manufacturing practices (cGMPs), the risk to the patient from commercial aseptic processing activities should be understood as low. 

Evaluating terminal sterilization
The divide between aseptic processing and terminal sterilization has always been considered substantial (11). Rigid dogma that define the acceptability of terminal sterilization processes are often based on concepts that are irrationally conservative and inappropriate today. Similarly, today it is obvious that not all aseptic processing technologies are equivalent in process capability. The industry is on the cusp of a new understanding with respect to these core production methods. Improvements in equipment capability and a more rational understanding of microbiology are bringing forth opportunities for increasingly safe and cost-effective means for production of sterile products.

Industry should have the flexibility to design terminal sterilization processes that make greater sense from an engineering and microbiological safety perspective. Regulations or decision trees that restrict industry scientists from establishing safe processes beneficial to the patient are examples of bad regulatory or standard-setting practice. The notion that effective terminal sterilization begins at a F₀ value of 8 or 15 minutes is both completely wrong and unnecessarily restrictive (12). The most likely cause for this rigidity is a false belief that biological indicator resistance (and, hence, destruction) correlates directly to product safety; in other words, one must kill at least a million of the most resistant spores available to deliver a PNSU of 10^-6. This misunderstanding, combined with a belief that industry process design scientists and engineers are incapable of making the appropriate choices for both process reliability and patient safety, support the flawed logic that supports rigid and dogmatic sterilization requirements.

The prevalent belief that G. stearothermophilus, owing to its very high resistance to moist heat, is a requirement for the demonstration of process safety in terminal sterilization is quite clearly wrong (2, 3). Insistence on the use of G. stearothermophilus and the high F₀ cycles that inevitably arise from its use actually increases risk to the patient, because excess requirements for minimum lethality of sterilization processes drive users to select an aseptic process. This decision runs counter to the near universal preference to using terminal sterilization rather than aseptic processing. A proposed terminal sterilization process that might be unable to kill huge numbers of nonpathogenic, G. stearothermophilus (a highly improbable bioburden isolate) but can reproducibly destroy less resistant spores and all of the medically important pathogens is by any reasonable measure a safe process. Therefore, risk arises from standards and regulations that actually discourage the scientific selection of sterilization conditions. It should be noted that, in Japan, a high level of product safety has been observed in medicines terminally sterilized at F₀s of two minutes or less (13). A sterilization process need be no more lethal than necessary to reproducibly destroy any bioburden that might be present. Figure 1 depicts newer thinking with respect to terminal sterilization process selection.

Aseptic processing techniques
There are a many different aseptic processing technologies in current use. Processes that are largely manual are still used in investigational new drugs and in early-stage clinical settings, but really have no place in commercial-scale operations except for orphan drugs. Isolation technology should be used for all manual aseptic processing, but clearly the transition to isolators has been faster in commercial-scale filling. Until the mid-1980s, when isolators first appeared, aseptic filling was accomplished in classified environments with gowned personnel performing the activities associated with set-up and inherent and corrective interventions (14–17). Isolators provide near absolute separation of personnel from sterile surfaces and materials and allow automated decontamina- tion of the enclosure. Introduction of restricted access bar- rier systems (RABS) sought to match superior capabilities of isolators while retaining the simplicity of manned cleanrooms and avoiding the need for decontamination with vapor phase hydrogen peroxide or an equivalent sporicide. RABS vary in their sophistication with the best designs approaching isolators in capability, while some less-suitable RABS designs require significant human intervention within the enclosure. New technologies for aseptic processing continue to emerge, with gloveless isolators and completely closed systems representing the current and future pinnacles of aseptic processing performance. Many of these advanced aseptic technologies have already been adopted in the global food industry for aseptic filling of foods and dairy products. Their increased adoption within the healthcare sector in the future can be anticipated. This hierarchy of aseptic technologies is depicted in Figure 2.

Defining sterile manufacturing
Sterile product manufacturing should not be viewed as a stark choice between either terminal sterilization or aseptic filling. Rather, sterile product manufacturing should be understood as a range of process technologies, and process designers should have the latitude to choose a technology or a combination of technologies that best satisfies both patient safety and commercial objectives. Standards and regulations that narrow processing choices without real scientific justification do nothing but harm both the producer and the consumer. The authors see no reason for a decision tree that implies an arbitrary thermal (lethality) baseline in terminal sterilization (12).

The overwhelming majority of the recognized families of bacteria, mold, yeast, and viruses associated with human disease are effectively killed at temperatures well below 100 °C. The list of organisms killed at temperatures less than 100 °C include all Gram positive and Gram negative nonspore formers, all viruses, and all yeast and molds. Thus, the organisms responsible for more than 99% of human and animal infectious diseases are killed effectively and efficiently at temperatures well below those in common use for sterilization. Any reasonable microbiologist or sterilization engineer would consider this as clear evidence that considerable process safety can be added at conditions considerably lower than an F₀ ≥ 8 minutes.

There are spore-forming organisms that cause disease, and the patient risk from these organisms should be considered in the development of any sterilization process. The spore-forming pathogen most well-known to the general public is bacillus anthracis, which is the etiological agent of the disease known as Anthrax. Anthrax has been in the public consciousness because of its use as a bioterrorism agent. This rare pathogenic spore provides a suitable example of the moist heat resistance of endospore-forming organisms of medical concern. An extensive study regarding the thermal resistance of b. anthracis compared to other strains of the genus bacillus was undertaken by Montville, et al. (18).

The reported results (18) demonstrate a clear disconnect between a stipulation of a minimal acceptable temperature of 121 °C and a minimum cycle time of 15 min. and the lethality actually required to achieve a real-world probability of non-sterility of 10^-6 or better. There is ample evidence that pathogenic spores can be readily and reproducibly killed at temperatures considerably lower than those given in the Pharmaceutical Inspection Convention and Pharmaceutical Inspection Co-operation Scheme decision tree. The cycles given in the decision trees are excessively conservative and, taken literally, preclude the use of safe process conditions. A more scientifically appropriate approach would be to encourage sterilization of formulations to be examined freely, without restriction to only those temperatures reported effective within pharmaceutical related literature.

In the authors’ view, the magnitude of process lethality as measured in F₀ should be understood as a choice and not an arbitrary requirement. In recent decades, however, there has been an inflexible belief in two factors that weigh on the choice of lethality. One of these factors is that 121.1 °C is a required moist heat sterilization condition, and the other is that an overwhelming margin of safety must be achieved in terms of the destruction of G. stearothermophilus. The data concerning the resistance of pathogenic organisms, including b. anthracis and others, highlight the unreasonableness of these dogmas. A temperature of 121.1 °C is actually nothing more than the conversion of 250 °F (a temperature equivalent to a steam pressure of 15 psi), which was historically used in treatment of canned foods in a retort. Because it is possible to achieve substantial lethality against pathogens, including sporeformers, at temperatures below 100 °C, it makes absolutely no sense to consider sterilization cycles operating between 121.1 °C and 100 °C (or potentially lower temperatures) as inadequate in terms of lethal input. Additionally, given the efficacy of temperatures between 70 °C and 100 °C for the control of non-spore formers, which include the vast majority of pathogens, cycles at these temperatures can be considered effective as an adjunct to aseptic or even tightly bioburden- controlled manufacturing. Figure 3 depicts some process alternatives, including combinations of processes.

Combining aseptic processing and terminal sterilization
Using aseptic processing as a precursor to terminal sterilization changes the paradigm dramatically. This approach provides superior control over the pre-sterilization bioburden, such that the subsequent sterilization process can be designed with lower thermal input (lower overall lethality), thereby making it possible to substantially broaden the use of terminal sterilization. From a safety perspective, this approach has distinct advantages, as follows:

• Any adventitious bioburden contaminant entering the container during aseptic processing is easily killed in the terminal sterilization step, minimizing or eliminating the need for extensive in-process environmental monitoring.

• Bioburden concerns for the terminal process are essentially eliminated because all units are aseptically filled.

• If the product were made using either process alone, the known limitations of each (no terminal lethal component in aseptic, more degradation in terminal sterilization) persist.

Conventional decision tree-based F₀ and time–temperature targets are arbitrarily selected targets intended to simplify process validation and execution, but they reduce flexibility, thereby reducing the opportunity of sterilization scientists to choose a safer, more easily controlled approach. The proper focus must be on killing bioburden microorganisms, not on physical data; numerical values are inherently conservative and ignore the impact of the process on the product. In addition, contamination recovery rates should be used for on-going evaluation of monitoring trends rather than alert and action levels, which often lead to over-interpretation or misinterpretation of results (19).

The goal in the production of sterile drugs should always be to assure patient access to safe and efficacious products. When using a combination of aseptic processing and terminal sterilization, everyone wins. The patient receives safer and more stable product with fewer degradation materials, and the producer has fewer operational issues. Regulators also win because patient safety is their primary mission, and the consumer is optimally protected from risk.

The implementation of these processes is straightforward, as existing sterilization equipment can be adapted or used almost directly. A combination process adds validation requirements, but reduces aseptic contamination control concerns significantly and reduces the need for extensive and costly environmental monitoring. Both FDA and the European Medicines Agency have documented their preference for the use of terminal sterilization (11, 20); they have also suggested flexibility in the sterilization process to be used.

Terminal sterilization processes in the range of 70–80 °C can be effective with well-controlled aseptic processes. Nearly all water-for-injection (WFI) systems are maintained at temperatures above 70 °C, and the microbial content is consistently near zero colony-forming units. However, aseptic processing should not be a mandatory prerequisite for terminal sterilization, and not all aseptic processes should or can be followed by a lethal process. Operational and regulatory flexibility should allow scientists and engineers to choose processing conditions optimal for the product in question.

A further advantage of wider-scale use of terminal sterilization should be the increased applicability of parametric release. The authors agree with Dr. Tsuguo Sasaki, formerly of the Japan Pharmaceuticals and Medical Devices Agency, who has suggested that parametric release should be applicable to any terminal process that yields a F₀ >2 minutes (13). Given the statistical and analytical limitations of the sterility test, liberalizing the use of parametric release makes sense.

It is said that rigid standards are required because firms do not have staff scientists with sufficient training or intellectual ability to make the proper processing choice on their own. The authors reject this as equal parts alarmist and disingenuous. It is an insult to the many well-trained and capable engineers and microbiologists who work in this field. It is at least possible that this assertion of universal incompetence as a justification for dogmatism is a self-serving excuse for regulatory overreach. Whatever the reason, it is hard to find any consistency in expressing a technical preference for terminal sterilization while at the same time restricting its use through unnecessarily rigid standard setting and regulation.

A further oddity is that the “design space” between aseptic processing and terminal sterilization is allowed to go essentially unexplored. There is apparently little thought given to the possibility of aseptic processing followed by thermal, radiation, or other processes that would kill the overwhelming majority of pathogens. Knowing that a drug could be safely treated with no impact on efficacy for five minutes at 70 °C, for example, would be useful knowledge that could benefit the patient/user at all stages of the product’s lifecycle, including both generic production and pharmacy compounding. Ignoring the benefits of such an approach is short-sighted and contrary to the universal concerns for patient safety that consumers, producers, and regulators all share.

The use of aseptic processing or bioburden controlled processing in combination with terminal sterilization is not a new idea; there are numerous examples in current use. Given the advantages in these combinations, it should be evident that the restraints are self-imposed by industry. It is past the time to destroy the arbitrary and unnecessary silos around aseptic processing and terminal sterilization and to give organizations the right to choose the best possible solution for their products. 

References
1. FDA, 21 CFR 211, Proposed Good Manufacturing Principles for Large Volume Parenterals (Rockville, MD, June, 1976).
2. J. Agalloco, J. Akers, and R. Madsen, PDA J. Pharm. Sci. Tech. 52 (6) 346-350 (1998).
3. J. Akers and J. Agalloco, PDA J. Pharm. Sci. Tech. 55 (3) 176-184 (2001).
4. J. Agalloco, J. Akers, and R. Madsen, PDA J. Pharm. Sci. Tech. 63 (2) 89-102 (2009).
5. Her Majesty’s Stationary Office, Health Technical Memorandum 2010 (HTM- 2010) Part 3: Validation and Verification (London, 1994).
6. British Standards Institution, EN 285, Sterilization - Steam sterilizers - Large sterilizers; EN 285:2006+A2:2009 (London, June, 2006).
7. USP, USP General Chapter <1229>, “Sterilization of Compendia Articles” (US Pharmacopeial Convention, Rockvile, MD, 2013).
8. USP, USP In-process Draft <1229.5> “Biological Indicators,” Pharmacopeial Forum 41 (2) 2015.
9. J. Agalloco and J. Akers, Ed., Advanced Aseptic Processing Technology (InformaUSA, New York, 2010).
10. J. Akers and J. Agalloco, Pharm. Tech. 36 (5) s48-50 (2012).
11. EC, EudraLexVolume 4: Good manufacturing practice Guidelines, “Annex 1, Manufacture of Sterile Medicinal Products,” (Brussels, 2008).
12. PIC/S, Decision Trees for the Selection of Sterilization Methods (CPMP/QWP/155/96) (Geneva, 1999).
13. T. Sasaki, PDA J. GMP Validation Jpn. 4 (1) 7–10 (2002).
14. J. Agalloco, Pharm. Tech., 29 (3) 56-66 (2005).
15. J. Agalloco and J. Akers, Pharm. Tech., 31 (5) s8-11 (2007).
16. J. Agalloco and J. Akers, Pharm. Tech., 35 (4) 69-72 (2011).
17. J. Agalloco, Pharm. Manuf. 10 (3) 28-32 (2013).
18. T.J. Montville, et al., J. Food Protection 68 (11), 2362-2366 (2005).
19. USP, USP General Chapter <1116>, “Microbiological Control and Monitoring of Aseptic Processing Environments,” (US Pharmacopeial Convention, Rockville, MD, 2012.)
20. FDA, Guidance for Industry, Sterile Drug Products Produced by Aseptic Manufacturing (Rockville, MD, Sept. 2004).

About the AuthorsJames Agalloco is president of Agalloco & Associates, jagalloco@aol.com. James Akers, Phd, is president of Akers Kennedy & Associates.

Article DetailsPharmaceutical Technology
Vol. 39, No. 12
Pages: 36–41

Citation:
When referring to this article, please cite it as J. Agalloco and J. Akers, "Achieving Balance in Sterile Product Manufacturing," Pharmaceutical Technology 39 (12) 2015.