In This Issue
Engineering and Vaccine Production for an Influenza Pandemic
September 1, 2006 Volume 36 Issue 3

Pharmaceutical Preparedness for a Pandemic

Friday, September 1, 2006

Author: Patrick J. Scannon

H5N1 may be the first real test of a rapid pharmaceutical response to a pandemic.

As H5N1 spreads around the world, the requirements for pandemic preparedness are being discussed on an unprecedented international scale. Despite an increasing awareness of need, the question remains: is the pharmaceutical industry prepared to meet this challenge and others like it? From a pharmaceutical perspective, the answer to the question is a resounding NO. The follow-on question should then be, why aren’t we ready? Or, more constructively, what is the baseline from which we can build toward readiness?

A central purpose of the NAE/Institute of Medicine (IOM) conference was to bring together engineers, scientists, and global health care leaders to discuss opportunities for improving pharmaceutical preparedness. This is not a theoretical exercise—H5N1 influenza virus may become the first real challenge to the innovations and interventions catalyzed by this initiative. And if the H5N1 challenge should be averted, it is only a matter of time before another virus with pandemic potential arises.

The Threat of Pandemic Influenza: Are We Ready? Workshop Summary, an IOM report published in 2005, addressed the question of pharmaceutical preparedness and summarized strategies for readiness in the United States and around the world (IOM, 2005):

  1. Vaccines are the single most important intervention for preventing associated morbidity and mortality during both seasonal epidemics and pandemics.
  2. No country will have adequate supplies of vaccine at the start of a pandemic. At least 4 to 6 months will be needed to produce the first doses of vaccine following isolation of a new pandemic virus. The subsequent augmentation of supplies will be progressive. Stockpiling in advance is not an option.
  3. Antivirals are expected to be effective against human illness caused by avian influenza and human pandemic strains. Pending the availability of vaccines, they will be the only influenza-specific medical intervention for use in a pandemic.
  4. Inadequate supplies are a major constraint. Supplies are presently extremely limited, and manufacturing capacity could not be augmented during the course of a pandemic. At current capacity, several years would be needed to increase supplies appreciably.

One approach to assessing the state of pharmaceutical preparedness is to understand how the seasonal influenza vaccine is made and to determine if such a process is applicable for an H5N1 (pandemic) vaccine. After all, H5N1, a strain of avian influenza, and seasonal influenza, are both influenza viruses (even though they are potentially very different diseases), and the seasonal influenza vaccine is made by a proven production process that uses fertilized chicken eggs.

Once a year (for each hemisphere), a group of scientists, epidemiologists, and other public health specialists assemble to decide upon the virus composition for that year. Prior to that meeting, egg production in approved sites has already begun, so that when the vaccine strains are selected hundreds of millions of eggs are ready. Over a period of nine to ten months, 60 to 100 million doses of vaccine are produced for the U.S. market (~300 million doses worldwide). Some of the vaccine is released as early as six or seven months after production begins (Figure 1 - see PDF version for figures).

With that as a “standard” for vaccine production, one of the most significant pandemic-related concerns is the life-and-death competition between the time it takes to make and release a useful vaccine and the time it takes a contagious virus to spread around the world. In “The Next Killer Flu,” an article in the October 2005 issue of National Geographic by Tim Appenzeller, the rate of spread of a pandemic (such as H5N1) is compared with vaccine responsiveness. The mathematical model used in the article predicts that a human pandemic would circle the globe within 180 days (roughly half the time it took the last pandemic infection in 1968), but the first vaccine would become available only after day 250 (i.e., about nine months after the start of production). This model assumes that scientists would identify the exact H5N1 isolate only after the pandemic had begun and that current egg-production methods would be used to produce a vaccine. These same assumptions were used in the dire IOM conclusions.

Drug-Development Process
To understand current pharmaceutical responsiveness, we first need a basic understanding of the drug-development process in general and, more specifically, the development process for vaccines. A common lay perception of drug development is that discovery research and product identification are the “gating” steps for drug approval. Looked at another way, the common misperception is that critical post-research processes, such as manufacturing, somehow just happen. In fact, the manufacture of drugs and vaccines involves very complex scientific and engineering challenges, quite separate from discovery research.

Another misconception is that the pharmaceutical industry is a single industry. In fact, it is a series of industries with many regulatory steps in common but also with independent technologies, depending on the type of product being made: biological drugs resulting from the biotech revolution include both recombinant proteins and antibodies; the targets of vaccine development, a separate business, are toxins and viruses; synthetic small molecules, a very important third class of drugs, are used across the spectrum of human diseases.

These three sub-industries are not at parity. The biggest revenue generator, by far, is synthetic small molecules. Vaccine revenues are a distant third. The three classes of pharmaceuticals, by and large, also use different manufacturing technologies.

In recent years, new discovery technologies have accelerated the pace of vaccine discovery, but vaccine manufacturing processes have essentially remained unchanged, in large measure because there has been little or no incentive to change post-discovery downstream processes. Thus, even though a vaccine might be selected much more rapidly today than ever before, the impact on time-to-patient availability (especially in a pandemic setting) is minimal, because the manufacturing and testing steps have not kept pace. In contrast, the processes to make a synthetic molecule like Tamiflu?, an important antiviral drug against H5N1, are completely different from those used to make vaccines, even though the regulatory (e.g., Food and Drug Administration [FDA]) steps are similar.

To ensure drug safety and effectiveness, the pharmaceutical industry is highly regulated. Drug development and regulatory requirements must be integrated with each technology to create an acceptable pathway to marketing approval. Even within the vaccine world, technologies besides eggs have been developed for making human vaccines—and each of these is separately regulated.

“Knowns” and “Unknowns” in Drug Development
During the drug-development process, some components are unavoidably “unknown” (meaning new discoveries are expected to be made); other components are kept, as much as possible, as “knowns,” to minimize complexity, time, and expense. Figure 3 shows examples of new-product “unknowns,” such as human safety, efficacy, potency, dosage, drug kinetics, drug metabolism, and drug interactions. In both drug and vaccine development, it is highly desirable to limit the “unknown” variables to the product itself, because new downstream manufacturing processes are very costly, both in dollars (tens to hundreds of millions) and time (months to years). To the extent that downstream manufacturing processes can be kept as “knowns,” the overall development program is greatly simplified.

Changing even minor manufacturing methods requires validating the new methods to meet FDA regulatory guidelines, such as the current Good Manufacturing Practices (cGMPs). New manufacturing methods, which may offer major improvements, are useless until they are validated (i.e., meet cGMPs) and are thus carefully considered before changes are made. The risk of a new manufacturing process—which could set a program back by years if it fails—provides considerable motivation for retaining the “old way” of doing things. In the pharmaceutical world, one must not only be highly motivated, but also willing to take multimillion dollar risks and be well financed to change “known” manufacturing processes that work, however inefficiently by modern standards. Hence, most of the time, drug-development risks fall primarily on human testing of the product (“unknowns”). Only if absolutely necessary, is manufacturing risk included.

Egg Production for Pandemic Vaccine Production
Egg production certainly has some advantages. First, this decades-old technology is an established, validated process, and regulated egg production facilities are in place. The only significant parameter that changes from year to year is the influenza virus selected for insertion into the eggs. However, this stable, predictable process may not work in a pandemic setting for many reasons:
  1. The pandemic could occur off cycle when not enough eggs are available.
  2. The virus could mutate sufficiently to nullify the potency (immunogenicity) of a vaccine already prepared and stockpiled (i.e., start over).
  3. H5N1 may not be the ultimate cause of the pandemic (i.e., start over).
  4. If the pandemic is rapidly contagious, there may not be enough time to produce an adequate supply of vaccine (e.g., difficult to start over).
  5. Increased worldwide demand from a pandemic, combined with the requirement for seasonal influenza vaccine, would completely overwhelm the programmed seasonal influenza vaccine manufacturing capacity.
  6. If an avian influenza infected the population of chickens that produce eggs for seasonal influenza vaccine, egg production would cease.

Considering these plausible possibilities, what can be done to ensure vaccine manufacturing preparedness? And how fast can it be done?

The Regulatory Process
A good place to start is with the FDA, the federal body that regulates all pharmaceutical manufacturing for the United States. A tendency in discussions of preparedness is to assume that the FDA imposes excessive, burdensome regulations that interfere with rapid change. From this author’s perspective, despite the strict regulations for manufacturing, the FDA is not the problem.

cGMPs, which have been in place for 30 years, establish compliance for drug products, require reproducibility of manufacturing, and ensure drug safety for the American public. Although expensive and time consuming for companies to implement, cGMPs have established a gold standard for the entire world for the production of pharmaceutical products. In fact, cGMPs are essential.

However, the FDA also has recognized that in times of emergency, routine requirements may not be implementable. In June 2005, the FDA put out a draft guidance document, “Emergency Use Authorization of Medical Products,” that addresses emergency situations. The FDA is committed to working with industry on preparedness and related problems and has addressed the regulatory spectrum for product availabilities according to the actual need. So, if the FDA is not the problem, then what (or who) is?

Engineering Challenges in a Pharmaceutical Context
Many pharmaceutical rapid-response issues are, in fact, engineering challenges. But engineering solutions, however potentially beneficial, must fit into pharmaceutical business practices. Like all other successful industries, the pharmaceutical business is driven by profitability. Unlike most other industries, however, manufacturing expenses (the cost of goods sold) are typically only a small component of the total selling price. In this setting, even major savings in manufacturing are likely to have a minimal impact on a drug’s profitability. Because impact on profitability is a key driver in evaluating change, pharmaceutical companies will have to be sold on the value (financial or otherwise) of novel engineering solutions that might replace regulated, established manufacturing processes.

Pandemics, however, are not routine, and pharmaceutical companies are beginning to address industry responses to emergencies. Examples of engineering challenges that might impact pharmaceutical pandemic responsiveness include: capacity, yield and potency, timing, and dosage.

Egg Industry
Increasing capacity in the egg industry might be feasible if chickens could be made to lay more (or much bigger) eggs. Short of that, the challenges of increasing the supply from 300 million eggs to one billion or more eggs on demand are staggering. The logistics alone of providing feed, facilities, and routine care would be monumental. Although the egg method may yet prove useful for production of a pandemic vaccine, we need new, more flexible manufacturing approaches.

Mammalian Cell-Culture Systems
Mammalian cell-culture systems (essentially fermentation systems that produce protein instead of alcohol) are being developed as alternatives to egg-based vaccine systems. Although cell-culture systems are not dependent on the vagaries of egg availability and have greater flexibilities, worldwide mammalian cell-fermentation capacity is limited.
Right now, approximately 1.5 million liters of pharmaceutical mammalian cell-culture capacity exist between contract manufacturers and pharmaceutical companies (additional capacity is also under construction). Contract manufacturers own a total installed capacity of ~400,000L; the rest is owned by product-development companies. However, only a very small percentage of this fermentation capacity is currently available or fully dedicated to producing vaccines; most of it is routinely used to make other essential biological drugs.

In a recent analysis, forecasted demand was compared with forecasted supplies across the biopharmaceutical industry for the next five years (Levine, 2005). Figure 5 shows that after 2006, forecasted demands will closely match forecasted supplies, which means there is no meaningful surge or idle capacity to meet a sudden demand for a pandemic vaccine. The reason is simple—it is too expensive to allow multimillion dollar facilities to sit idle. Therefore, even in the most extreme emergency, companies with existent facilities would have few alternatives to halting production of otherwise needed biological drugs.

Thus, the promise of cell-culture vaccine production appears to be limited, not by technical feasibility, but by the current lack of expansion capacity. Even if construction were begun today to make such capacity available, worldwide capabilities would remain essentially fixed for the next five years, because it takes that long (and upwards of a billion dollars) to construct and validate a single large biological production facility.

Even interim alternatives, such as upgrading or expanding existing facilities, would cause major interruptions until revalidation and recertification could be completed. In addition, lead times for large manufacturing components are very long; for example, it can take up to two years for the order and delivery of a large stainless steel fermenter. There are no easy answers.

Yield and Potency
For a pandemic, the more vaccine available, the better. By increasing the potency of the vaccine, if it can be accomplished safely, the dose per patient could be reduced, thus allowing more people to be vaccinated with the same amount of vaccine. Conversely, if the vaccine is less potent than expected, fewer people could be vaccinated, or more vaccine would have to be manufactured. Unfortunately, vaccine potency (and dosage) is not predictable until the vaccine is tested in humans—and in a true pandemic, it may not be feasible to look for a vaccine with optimal potency.

Maximizing the vaccine yield would help mitigate against this vagary. And because many variables can affect the final yield of a vaccine, there are many different kinds of engineering opportunities for increasing yield. Increasing yield would have the added advantage of increasing the productivity of existent facilities.
In the egg system, for example, increases in the vaccine virus yield per egg (sometimes, but not always, possible) and downstream virus recovery from the egg could greatly increase overall vaccine yield. In cell-culture systems, numerous process components are targets for increasing yield. These include, for example, cell-line genetics and selection, growth media, and other feed variables, such as temperature, gases, and pH. Exploring these areas to find meaningful yield improvements would require, at a minimum, input from cellular/genetic engineers, biochemical engineers, and mechanical engineers.

Independent of a global pandemic, improving cell-culture manufacturing parameters to increase yield is currently an area of intense study throughout the biopharmaceutical industry, in large part to get more out of existing manufacturing facilities. Downstream improvements after fermentation, such as minimizing purification losses or increasing shelf life, are examples of engineering opportunities that could increase the availability of vaccine to the patient.

Time Scales
For each new product, the manufacturing development program includes both upfront times to develop the process and standardized cycle times thereafter. Even though a “known” manufacturing process may be used, each new product brought into an established process has specific variables (e.g., cell division time, temperature, media development requirements) that must be determined.

The typical ramp-up time for bringing a new product online in an established cell-culture system is 10 to 12 months. It often takes six months just to establish a new cell bank (clone) because mammalian cells only divide at a certain (slow) rate. In evaluating the potential of mammalian cell-culture systems for pandemics, it will be important to look for technological opportunities to compress this time frame. For example, if it were biologically possible to establish a cell bank more rapidly (not at all certain), this would speed the entire process. The incorporation of systems management into plant operations (a more certain type of improvement) has already been shown to reduce both initiation and cycle times in other settings; this would also probably be helpful in a pandemic response, especially with prior planning.

The smallest amount of vaccine per patient and the fewest injections per patient are highly desirable goals, especially for global administration of a pandemic vaccine. Today, vaccine dosages and schedules are empirically determined through careful human clinical trials that take many months. Might there be technologies to determine dosages more quickly or even to find ways of reducing effective dosages safely? Rapid screens to find more immunogenic (potent) vaccines and novel/safe adjuvants (given with vaccines to boost their immunogenicity) are two possibilities already being evaluated. Other possibilities are novel systems of administration to minimize losses and speed delivery in the field. In short, any technology that leads to fewer doses and/or smaller total doses per patient will make vaccine available to more people.

Take-Home Message
The pharmaceutical industry has never before been called upon to solve rapid response issues on a global scale. Although during World War II the production of penicillin was truly an unprecedented engineering feat, it cannot be compared with meeting worldwide pandemic vaccine requirements. Right now is the time for novel solutions. Yet, the pharmaceutical industry, even with its huge research capacity, cannot solve these rapid response issues alone.

Meeting the needs of the global community for vaccine during a pandemic outbreak will require exponential improvements in today’s capabilities. It is likely that such advances will have to be made across many fronts to meet the time and volume requirements. Depending on which estimates are used, up to a 1,000-fold composite increase in yield, shorter time, and smaller dosage, as well as other process innovations and enhancements, will all be necessary to have a meaningful impact on a truly global pandemic. We need unprecedented and as yet unknown solutions, which may only be possible through previously unheard of scientific and engineering collaborations.

The technological solutions necessary to meet the rapid-response requirements of the pharmaceutical industry will require input from many engineering disciplines—mechanical, structural, civil, chemical/biochemical, genetic, materials, and systems engineering. The sum of such innovations will require integration to meet both pharmaceutical and regulatory requirements, but (and importantly) these are technologically surmountable obstacles. With this conference, we have created a unique opportunity to examine, perhaps for the first time, “potential engineering approaches to a pandemic,” if not for H5N1, then for an unknown virus that may threaten our world in the future.

Appenzeller, T. 2005. The next killer flu. National Geographic. Available online at: html.
CFR (The Code of Federal Regulations), Title 21 – Food and Drugs Good Manufacturing Practices, Parts 11, 210 and 21, 820.
FDA (Food and Drug Administration). 2000. International Conference on Harmonization (ICH) of Technical Requirements for Registration of Pharmaceuticals for Human Use. Draft Consensus Guideline: Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients (Q7A). Washington, D.C.: Food and Drug Administration, Center for Drug Evaluation and Research.
FDA. 2005. Emergency Use Authorization of Medical Products. Draft Guidance. Available online at:
IOM (Institute of Medicine). 2005. The Threat of Pandemic Influenza: Are We Ready? Workshop Summary, edited by S.L. Knobler, A. Mack, A. Mahmoud, S.M. Lemon. Washington, D.C.: National Academies Press.
Levine, H. 2005. The State of Biomanufacturing Capacity—Do We Finally Have Enough? Presented at IBC BioProcess International Conference, Boston, Massachusetts, September 2005.
About the Author:Patrick J. Scannon is chief biotechnology officer at XOMA Ltd.