Charting Process Improvement in Sterile Product Manufacturing

I am honored to have been asked to contribute an essay on the occasion of Pharmaceutical Technology's 30th anniversary. I became a regular reader of this fine publication in 1981, only four years after PharmTech's beginning. So, I can honestly say that I've grown up in this industry alongside Pharmaceutical Technology. It has been a distinct pleasure to read the magazine and to make the occasional contribution to its contents over the years.

An evolution in testing

My primary field of interest has been well-covered in this periodical during its 30 years of service to the industry and has developed in ways that neither I nor, I suspect, those responsible for the technical content of Pharmaceutical Technology could have envisioned fully back in 1977. Incredible as it may seem, media-fill tests were not a universal requirement. In fact, these process-simulation tests were not extended to all types of aseptically produced dosage forms until several years later. The first international standard that mentioned an acceptance criterion for media-fill tests, to the best of my knowledge, is the WHO Technical Bulletin No. 28 that suggested a 0.3% contamination rate (1). When I started in this business, many firms considered 9 positives out of 3000 units filled to be a successful media-fill test. Just writing the previous sentence today leaves me shaking my head in amazement at how far capability and expectations have evolved in the past 30 years.

Milestone. 30 years of Pharmaceutical Technology

Process validation emerges

Not only was the media-fill test only starting on the pathway to becoming a mandatory requirement, in 1977 the word validation was only starting to enter our lexicon. At that point, firms were validating autoclaves, and that was by and large the full extent of process validation. The appropriate approach for validation of moist-heat sterilization processes would have been a topic of considerable debate in 1977, although Irving Pflug, PhD, Carl Bruch, PhD, and other sterilization scientists were hard at work attempting to educate what was already an audience eager for knowledge. The seminal industry document about moist-heat sterilization, PDA Technical Monograph No. 1 (2) had not yet been issued at the time Pharmaceutical Technology began publication. It is with some disappointment that I must report that in some ways the art and science of sterilization validation has actually regressed over these three long decades. The clear thinking and analytical approach of Dr. Pflug has been replaced by a mishmash of requirements, some of which seem to have been borne out of worst-case thinking run amok or by efforts to force parochial approaches on industry, ironically enough in the alleged interest of harmonization. What has not changed is that sterilization, in 2007 as in 1977, is still a process best measured by microbiological analysis. The statement that will forever be associated with Dr. Pflug still rings true today: "The bugs don't lie." It is important to remind ourselves that those spores are just as honest and reliable evaluators of sterilization now as they were in 1977. None of the foregoing is intended to diminish the importance of thermometric and pressure data; it is to remind the reader that the purpose of sterilization is to destroy microbial contamination. Thankfully, there are ample data on the destruction of microorganisms by moist heat, heat, radiation of various types, and chemicals, so let us not reinvent the wheel.

I am sure that many readers under 40 years of age will find it astonishing that prefilled syringes were a real rarity in 1977, and ampuls were still a widely manufactured dosage form. Today, as prefilled syringes gain market share in most of the world, the ampul is as extinct as the Tyrannosaurus. Most significantly, though, many aseptic processing lines in those days lacked the now-common depyrogenation tunnel. Instead, gowned operators, who typically wore nothing more than undergarments or street clothes under their gowns and were never required to "double glove" manually scraped vials off of the trays that had been depyrogenated in an oven that might or might not have contained a high-efficiency particulate air filter so that the oven could be operated at a positive pressure relative to the surrounding environment. The complete elimination of manual glassware transfer is very close to being a reality today, although I cannot honestly say how close with absolute certainty. In my opinion, the introduction of in-line glassware washing and depyrogenation would rank very near the top of the list of developments most responsible for increased aseptic-product safety during the last 30 years. Of course, it has proven difficult to implement the direct sterilization in conjunction with some product types, notably plastic containers with volumes of 100–1000 mL. Cold sterilization technology, however, has evolved tremendously during the last 30 years, and the means of achieving direct in-line sterilization of most products appears be within our grasp.

Milestone. 30 years of Pharmaceutical Technology

A new regulatory approach

There is one year that stands out in my mind in the 30 years that Pharmaceutical Technology has been published. That year is 1984, and it is memorable to me because in that year the pharmaceutical industry was made aware that the US Food and Drug Administration was working on a draft guideline, which was eventually published in 1987 as the Guideline on Sterile Drug Products Produced by Aseptic Processing (3). Word began to spread that this guideline would force industry to do some things very differently. The nature of one important change emerged when the pharmaceutical industry learned through podium presentations by key FDA personnel and inspectional activity that the industry practice of automatic sterility test "retests" was no longer considered acceptable.

Before that time, the formal sterility-test failure investigation did not exist; rather, sterility-test positives were considered more like a minor glitch on the road to release than as a potential processing failure. Actually, my firm did take the occurrence of a sterility positive seriously, and we performed a limited version of the investigation required today. My firm also immediately did a retest while conducting the investigation. Some firms, however, skipped most of the steps involved in today's investigation and just went right to the retest and, if it passed, released the product.

Improving contamination control

Industry was forced to face rather quickly that this change really meant that the 0.25–0.5% retest rates that were common in sterility testing operations were no longer acceptable. The industry needed a more reliable sterility test. At the same time, the industry needed to improve the contamination control in aseptic processing. Out of this situation came word that a new technology had been used for sterility testing in Europe and that it seemed to effectively eliminate microbial contamination. This technology was, of course, the isolator. Later in 1984, I saw a flexible wall isolator for the first time at a trade show, and as I walked away from that exhibit, I mentioned to a colleague that we had just seen not only the future of sterility testing, but also the future of aseptic processing.

Having been involved in aseptic processing for three years by 1984, and having spent over a decade before that engaged in cell culture and virology, I knew very well that microbial contamination was almost exclusively a problem associated with humans performing interventions. Our firm had already considered using robots to reduce reliance on personnel, and we had a robotics laboratory to evaluate concepts. In the isolator, we saw a system that could separate the human from the operation and, that when used in concert with machine automation and robotics, would enable us to achieve levels of control over microbial contamination that would never be attainable in conventional cleanrooms.

This year at the International Society for Pharmaceutical Engineering's Barrier Isolation Technology Forum, Rick Friedman, director of the division of manufacturing and product quality in FDA's Center for Drug Evaluation and Research, commented that nearly all submissions for new facilities at the agency are isolators or restricted access barrier systems (4).

So, the revolution that started in 1984 is now in the end game. Although the end of the manned-cleanroom era in aseptic processing has not come, the industry can certainly see it from here. Look for machine automation and robotic applications to provide the next phase of this contamination-control revolution as the industry moves from the isolator era to the gloveless-isolator era.

If I am still around to write an essay on the occasion of Pharmaceutical Technology's 50th anniversary, I predict the human cleanroom will be gone forever, and it will be possible to run plants with the lights turned off, which will be very good for the carbon footprint and for contamination control.

Areas for improvement

The improvements in aseptic processing equipment and facilities during the last three decades have resulted in a circumstance that is still not widely recognized. Although microbiological assays remain the way to best analyze moist-heat sterilization processes, the industry has reached a point in the evolution of aseptic processing where the utility of microbiological assays is now questionable in many ways. Two or three decades ago, the pharmaceutical industry did a lot less environmental monitoring in support of aseptic processing than is the current norm. Back then, the industry actually recovered organisms with some frequency, and media fills were still something of a cross-your-fingers adventure. Today, in ISO 5 environments, microbial recoveries are rare, and in isolators they are so rare that some operations have reported having nothing but zero recovery rates for years. I believe what this should mean to any pragmatic microbiologist is that we may have reached the limits of growth and recovery microbiology as a means of evaluating process control. Really, this has been an issue in the cleanest environments longer than the industry might imagine.

Now the reader might quite reasonably ask why it is that I can endorse the value of microbiological analysis in the validation of moist heat while at the same time asserting that microbiological analysis may no longer be particularly useful when applied to contamination assessment of the cleanest environments. The answer is simple. Sterilization processes begin with a very large population of microorganisms and use the well-known thermal-death kinetics of spores to determine the effectiveness of the sterilization cycle. The statistics of starting with large populations of biological indicators with known resistance properties allows operators to work in a quantitative range that is suitable for growth and recovery assays. This is to say that the biological variation that might be as much as ±0.5 log in the worst case is actually of only limited significance when we have a starting spore population of >10⁵ . The biological effectiveness of the sterilization process can be estimated well enough to establish correlations with what would be expected from thermometric data.

Contrast this with attempting to evaluate sterility in more or less the opposite manner by recovering the miniscule residual bioburden within a very clean environment. This is a task made all the more daunting by the fact that this miniscule bioburden is not homogeneously distributed. Making the situation more difficult still is that those organisms that may be present can be or are likely to be under significant nutritional and environmental stress. It is not without reason that the compendial growth promotion tests suggest a challenge population of 10–100 colony-forming units to support the fertility of microbiological media. This process is done because it is very difficult to achieve growth at low inoculum levels, even when working with healthy stock cultures. Of course, the same logic applies to the media-fill test for precisely the same reasons.

I have heard some regulators suggest that if the environmental-monitoring test results are all zeros, then there is not enough testing. The problem, however, is if the results are below the limit of detection of the assay testing, more testing will not help. Even if the possible problem was simply that there were microbes in the environment that could not be found at current sampling intensities at modern air flow rates, it is very hard to imagine how large the sample size would have to be. The answer may well be that there just is not a practical way to sample enough.

Risk analysis in aseptic processing

Late last year at the Parenteral Drug Association's Asian Symposium in Tokyo, Katayama et al. reported that they had used two risk-analysis models for aseptic processing to evaluate three different aseptic processing production lines spanning more human intervention-intensive technology to modern separative technology with a high level of line automation (5). They compared these quantitative risk analysis models (including one Jim Agalloco and I published in Pharmaceutical Technology) (6, 7) and microbiological methods and found that in the higher technology lines, risk- analysis may be a more effective way to evaluate process control than microbiological monitoring.

This finding means that less reliance on traditional microbiological monitoring and process-simulation testing and more reliance on risk analysis may be the way forward. With this in mind, I list the move toward risk- and science-based regulation and the greater emphasis on quality systems as major innovations of the last 30 years.

The pharmaceutical industry needs to be inspired enough to ask whether what it is doing makes sense, and if it does not, the industry needs to be on the lookout for better ways. Albert Einstein thought that in the case of creativity, inspiration was more important than knowledge. He just may have been right. It may also be true that if we are going to insist upon trying to measure something, we need to know how capable we are of measuring that something.

So let's hope that Pharmaceutical Technology will continue to inspire us in the future as it has it has in its first 30 years. I'll bet that it will. I also hope that as the industry moves forward, it keeps an open mind and is inspired enough not only to embrace the inevitability of change but to facilitate that change through its own creativity. The industry should demand that its standards and its harmonization efforts always be driven by evidentiary science rather than merely by someone's opinion. That is where Pharmaceutical Technology comes in, as it will continue to enhance our understanding during the next 30 years.

James E. Akers is the president of Akers Kennedy & Associates, PO Box 22562, Kansas City, MO 64113, akainckc@aol.com

Where were you 30 years ago?

"In 1977, I was a post-doctoral fellow and instructor at East Carolina School of Medicine, Department of Microbiology. I took my first position in the pharmaceutical industry in 1981 at Burroughs Wellcome, Co. (now part of GlaxoSmithKline) in 1981."

References

1. World Health Organization, World Health Technical Report No. 28 (Geneva, Switzerland), 1973.

2. Parenteral Drug Association, "Validation of Steam Sterilization Cycles," PDA Technical Monograph No. 1, (PDA, Bethesda, MD), 1978.

3. US Food and Drug Administration, Guideline on Sterile Drug Products Produced by Aseptic Processing, (Rockville, MD), 1987, www.fda.gov/cder/guidance/old027fn.pdf (accessed June 5, 2007).

4. R. Friedman, Presented at the 16th Annual Barrier Isolation Technology Forum, Arlington, VA, June. 5, 2007.

5. H. Katayama and A.Toda, "Risk Categorization of Aseptic Processing Facilities Based Upon Risk Assessment Scores," in Proceedings of the Parenteral Drug Association's Asian Symposium (Tokyo, Japan), 2006.

6. J. Akers and J. Agalloco, "Risk Analysis for Aseptic Processing: The Akers–Agalloco Method," Pharm. Technol. 29 (11), 74–88 (2005).

7. J.Akers and J. Agalloco, The Simplified Akers–Agalloco Method for Aseptic Processing Risk Analysis," Pharm. Technol. 30 (7), 60–72 (2005).