Aches and Pains

Learning lessons from the influenza vaccine shortage 

By Sheri Fink, MD, PhD

As the US supply of influenza vaccine see-sawed from shortage to surplus this past year and flu experts again confronted warning signs of the next flu pandemic, AIDS vaccine experts might have considered taking notes. Experts say these unfortunate episodes provide valuable case studies highlighting the precarious nature of vaccine manufacturing, the difficulty of forecasting demand for biological products, and the challenges of ensuring an adequate supply. Many of the lessons flu experts are learning and the solutions they are proposing could apply to HIV/AIDS should an efficacious vaccine be developed.

Shortages — and then surpluses

Flu season is awaited anxiously every year for good reason. Influenza kills 36,000 Americans annually and hospitalizes 200,000 more. An estimated 500,000 deaths occur worldwide each year. So when US influenza vaccine demand exceeded supply in the 2000-1 and 2003-4 flu seasons, the industry primed itself not to be caught off guard again. The 2004-5 season was supposed to have the largest vaccine supply yet—100 million doses.

Chiron Corporation, one of only two inactivated flu vaccine producers for the US, was expected to supply 48 million of those doses. But the company’s Liverpool, UK vaccine manufacturing facility had a history of Food and Drug Administration (FDA) inspection deficiencies. On August 25th, 2004, Chiron informed the FDA that it had found Serratia marcescens bacterial contamination in eight lots of vaccine. During September, the company retested its lots and investigated its manufacturing processes. It informed the FDA that retesting results were negative and that the company would fulfill nearly all of its planned supply to the US.

Thus the October 5th announcement by the UK’s regulatory authority, the MHRA, of a suspension in Chiron’s vaccine licensing came as a surprise to the FDA which had conducted its oversight by conference call since the first reports of contamination at the facility,. After the announcement, the FDA sent a team to inspect the manufacturing facility. It found that some deficiencies noted during the previous inspection in June 2003 had not been remedied, including at least one sterility problem. In mid-October, 2004, the FDA determined that it could not assure that Chiron’s finished lots of vaccine met US safety standards, and none of Chiron’s vaccine was allowed to enter the US commercial market. To add to the confusion, this announcement led to contradictory calls—for more lenient regulations to allow other suppliers to rapidly license their vaccines for distribution in an emergency situation, and at the same time for tougher regulation in order to avoid future manufacturing contamination.

The rejection of Chiron’s entire contribution to flu vaccine supply instantly sliced the number of vaccine doses available to US consumers to nearly half. In response, the US Centers for Disease Control and Prevention (CDC) recommended restricting influenza vaccination to adults and children most at risk for severe complications or death from influenza, and others with a high potential of spreading influenza to these vulnerable populations such as health care providers and household contacts of infants under six months of age. In coordination with state and local health authorities, Sanofi Pasteur (previously known as Aventis Pasteur)—the only other US supplier of inactivated influenza vaccine—redirected its remaining shipments to target providers serving high-priority patients. Still, providers were free to dispense shots as they saw fit, and early in the vaccination season there were reports of price-gouging. As during previous shortages, those willing and able to pay these elevated prices had a better chance of receiving their immunizations.

But many Americans heeded the CDC’s advice and if they weren’t in a high risk category went without a flu shot. As a result of this civil obedience, an early vaccine shortage later turned into a surplus. As of January 2005, more than five million shots were still unsold.

Lessons learned

Chiron’s particular problems were triggered by bacterial contamination during the laborious embryonated chicken egg-based manufacturing process used to grow influenza virus. However, the resulting vaccine shortage stemmed from the fact that such a large proportion of the US influenza vaccine supply was produced at a single Liverpool facility. “I suppose the final point is not to put all your eggs in the same basket with HIV and distribute production in a variety of centers. It was a serious misjudgment for the USA to have 40% or so production in one unit,” says John Oxford of the University of London.

“In any situation where a large component of the supply of something is being produced at a single location there is the potential for natural or man-made events to disrupt that supply,” says John Treanor of the University of Rochester Medical Center. A general lesson that vaccine experts draw from this debacle is the need for a larger, more diverse group of vaccine suppliers.

Unfortunately, the trend in the vaccine business is exactly the opposite. A shrinking pool of manufacturers is becoming responsible for meeting a growing vaccine market. For flu vaccine, for instance, there were about 20 million doses per year distributed in the mid-1980s, compared with an expected 100 million doses for the 2004-5 season. Even with this increasing demand, many manufacturers dropped out of the market, leaving only Chiron (Fluvirin) and Sanofi Pasteur (Fluzone) to supply inactivated influenza vaccine. The production of FluMist, a live, attenuated flu vaccine manufactured by MedImmune Inc., was also scaled up for 2004-5 but was only scheduled to produce a relatively small 3 million doses.

“I think that all vaccines are potentially vulnerable to supply disruptions, and I expect that this will also be true of AIDS vaccines, when they are developed,” says Treanor. “Having multiple sources of supply can help to protect against such disasters.” The problem is finding those sources. The flu vaccine crisis serves to highlight a widespread problem. “On a global scale the experience points out the fragility of vaccine supply in general,” says Stanley Plotkin, emeritus professor at the University of Pennsylvania and an executive advisor to Sanofi Pasteur. From 1966 to 1977, half of all commercial vaccine manufacturers left the market. Now, only five manufacturers—GlaxoSmithKline, Merck, Sanofi Pasteur, Wyeth and Chiron—produce all vaccines for the US market that are recommended for routine child and adult immunizations. As of 2003, eight important vaccines for US consumers were each made by a single company—measles/mumps/rubella, tetanus toxoid, tetanus/diphtheria, inactivated poliovirus, varicella, pneumococcal conjugate (PCV-7), meningococcal, and pneumococcal polysaccharide (adult). The reason companies are either leaving or reluctant to enter the vaccine market is no secret: vaccines are seen as a risky business. Vice President for New Business and Scientific Affairs of Wyeth, Peter Paradiso, testified on this subject at a Senate Special Committee on Aging Hearing in November 2004. “It has become increasingly difficult to justify remaining in the vaccine business,” he said. Wyeth recently pulled out of both the influenza and DTaP (diphtheria, tetanus and acellular pertussis) markets. Wyeth’s departure from influenza vaccine manufacturing following the 2002-3 season was straightforward. The company’s influenza business had experienced four out of five years of financial losses associated with millions of unsold doses and rising costs related to meeting regulatory requirements.

The problem is that while the need for vaccines may be great, the demand for the product and the ease with which it can be brought to market can fluctuate, leaving companies with product they can’t sell.

Defining demand

The past year’s whipsaw in supply threw production for the following year’s influenza season into disarray. As of January 2005, a critical month for planning the next season’s influenza vaccine supply, it was unclear which manufacturers would supply the US vaccine market and how much vaccine they would be willing to manufacture. There was also the question of uncertain demand. Health care providers and vaccine distributors alike held off on placing orders. Some questioned whether low-risk patients would line up again for influenza shots after doing without vaccination during the 2004-5 season.

FluMist manufacturer MedImmune Inc. also held off on committing to a number of doses for next year’s season. The company had initially indicated it was capable of producing 10-40 million doses for the 2006-7 season. On the other hand, manufacturers GlaxoSmithKline PLC and ID Biomedical Corp were interested in supplying the influenza vaccine market, but had to wait for FDA approval. And it wasn’t until March 2, 2005, that British health officials gave Chiron permission to resume vaccine production in Liverpool.

This was not the first time that uncertain demand plagued the influenza vaccine market, as it could well plague the market for an AIDS vaccine [see Breaking the bottleneck]. Even in years when ample influenza vaccine existed, high-risk patients had poor coverage rates. Public campaigns to encourage immunization have contributed to some improvement in vaccination rates. Still, according to results of a National Health Interview Survey, only 43 million vaccine doses would have been required in 2004-5 to vaccinate high-risk patients at the same rate these groups were vaccinated in the 2002-3 season. That year, a mere 64% of the over-65 population was vaccinated (compared with a target rate of 90%), and rates for other high-risk groups were even lower. While the CDC recommended that 185 million Americans in at-risk populations and other target groups get vaccinated in 2004-5, the US had planned for a supply of only 100 million vaccine doses.

Interestingly, Chiron had been running against the trend of companies leaving the vaccine business, when it made a recent entry into the influenza marketplace. In 2003 it purchased PowderJect Pharmaceuticals and its influenza vaccine Fluvirin. The company attributed its expansion of manufacturing investments to the broadening of influenza vaccine recommendations to include adults between 50 and 64 years of age and children between six and 23 months, to improved reimbursement rates for vaccine providers, and to increased influenza vaccine prices. This demonstrates that companies may be lured back into vaccine production if their executives think market conditions are improving.

For that reason, policy wonks are conjuring up incentives to lure industry back to the vaccine game. For example, in 2003 a committee of the Institute of Medicine, a respected health advisory body associated with the National Academy of Sciences, proposed a system of subsidies and insurance coverage mandates for vaccines.

Increasingly, governments, international financial institutions and other sponsors are proposing to counterbalance the uncertain demand for both influenza and, one day, HIV immunization by strengthening markets and ensuring sales. So-called “pull” mechanisms such as advance purchase commitments would guarantee that a definite quantity of vaccine will be bought at an agreed price, provided it meets pre-specified criteria. Legislation has also been proposed in the US Senate (S.2038) to offer a tax credit to companies constructing or renovating vaccine production facilities.

Problems encountered last year also underlined the need for improved cross-border vaccine regulatory cooperation—an issue that was raised in the recently published Global HIV/AIDS Vaccine Enterprise strategic plan since any AIDS vaccine will likely need to be manufactured and distributed regionally.

Caution on egg-based technology

Contamination is a constant worry with the egg-based influenza vaccine production process. Eggs are not sterile and because each hen typically can lay only one egg per day, ensuring adequate egg supply is challenging. At three eggs per influenza vaccine dose, producing for the US market requires hundreds of millions of eggs per year. In order to secure adequate supplies of eggs, manufacturers need to forecast yearly vaccine demand six to nine months in advance of each flu season. This makes it impossible to respond to emergencies.

New influenza vaccines must be produced each year because the influenza virus, like HIV, quickly changes its genetic stripes. The influenza surface proteins hemagglutinin (HA) and neuraminidase (NA)—antigens targeted by traditional influenza vaccines—are constantly changing, a process known as antigenic drift. More marked and sudden genetic reassortment leads to new virus subtypes and the risk of pandemics and is known as antigenic shift.

The complexity of the egg-based manufacturing process requires more stringent regulatory processes, including FDA review of test results on each lot of influenza vaccine prior to commercial release. To overcome some of these challenges, particularly ensuring sterility and the ability to respond quickly to changes in demand, the US Department of Health and Human Services is supporting research on cell-culture based influenza vaccines, though at only half the budgetary level proposed in 2004. An influenza vaccine composed entirely of HA proteins and manufactured using a protein expression system could also shorten the time between identification of new strains and vaccine production. A vaccine produced using a recombinant baculovirus expression system in insect cell lines (not requiring eggs or animal serum) has already reached Phase III trials.

Other researchers, including Andrew Pekosz of Washington University St. Louis, David Milich of the Vaccine Research Institute, San Diego, and Walter Gerhard at the Wistar Institute, Philadelphia, are experimenting with stimulating an immune response to conserved influenza antigens like M2 that are not subject to antigenic drift and shift, which might also induce protection across influenza strains. It is also thought that the FluMist live vaccine may induce a broader immune response than the traditional inactivated vaccine.

The hope is that these new approaches could obviate the need for fresh vaccine formulations each year and thus ease the vulnerability to supply shortages. “Influenza vaccines are uniquely vulnerable to this type of thing because of the need to produce a new vaccine every year,” says Treanor. This vulnerability may apply to some extent to future AIDS vaccines given that any candidate that proves effective enough to justify large-scale production will undoubtedly be a first-generation product. Subsequently the production process may well be optimized, or the vaccine immunogens may benefit from tweaking to follow any trends in HIV genetic evolution or the emigration of viral clades and recombinants into new geographic regions. Similarly, while early AIDS vaccines are likely to incorporate immunodominant epitopes, it might turn out that conserved and more immunogenic subdominant epitopes are subsequently identified. “If it were to turn out that it is necessary to do the same kind of frequent formulation changes for an AIDS vaccine as we need to do for flu, then one might see the same situation arising with an AIDS vaccine,” says Treanor.

Pandemic fears

Flu and AIDS health experts also share the worry of dealing with pandemic outbreaks during which production and distribution would need to be rapidly scaled for a global vaccination program. For AIDS, the pandemic is ongoing. For influenza, the emergence of a pandemic strain is considered inevitable and overdue. Avian H5N1 influenza virus first emerged in 1997 but has re-emerged in recent years and made its way through large regions of Asia. This viral strain is able sometimes to infect humans in contact with birds and cause a high mortality rate and, most worryingly, it seems to have been transmitted on rare occasion from human to human. Only the inefficiency of this transmission between humans seems to have prevented it from becoming a full-blown pandemic virus.

In November 2004, the World Health Organization convened a two-day meeting of all major vaccine manufacturers to assess the status of vaccine preparedness for an influenza pandemic. The troubling conclusion was that should a pandemic strain emerge, companies wouldn’t be able to quickly produce vaccine for the commercial market. Even at full production levels, worldwide influenza vaccine manufacturing capacity totals only an estimated 300 million doses per year. “A new pandemic would show up the inadequacy of current facilities to produce enough vaccine for billions of people, and poor countries would be the first to suffer,” says Plotkin. Experts predict a flu pandemic could cause in excess of 200,000 deaths in the US alone.

Effective vaccines have never been available to counter such outbreaks in the past. For pandemic influenza, various solutions are being explored, including the use of “mock-up” vaccines to practice production. Because much of the data required for licensure would be gathered with the mock vaccine, manufacturers would be able to more quickly ramp up production of the actual vaccine when needed.

Novel technologies and tools are also being explored for use in potential influenza pandemic vaccine distribution—including vaccine formulations that do not require injections and dose dilution. Also, researchers are using molecular biology techniques such as reverse genetics in an attempt to more rapidly produce vaccine candidates that match a particular pandemic virus strain, and to modify the DNA sequences of particularly pathogenic avian influenza viruses, making them more suitable for vaccine manufacturing. Similar approaches are being considered in the search for an effective AIDS vaccine, particularly as there is a need to develop a vaccine that can be distributed as quickly as possible.

Some of the challenges faced by companies, such as Sanofi Pasteur, Chiron, and Canada’s ID Biomedical, which are committed to developing and being able to quickly manufacture a pandemic influenza vaccine are similar to those faced by AIDS vaccine developers. These include funding shortfalls, potential low return on investment, limited production capacity, liability fears, and licensing and regulatory concerns. Both pandemic influenza vaccine developers and AIDS vaccine developers face risk—for influenza, early production of a pandemic vaccine prior to the knowledge of whether a pandemic would occur involves a significant degree of risk. For HIV, as for pandemic influenza, risk also centers on the need to build and validate manufacturing and process development capacity at least five years before commercial production, at a time when efficacy testing is incomplete. “Lessons for HIV vaccines are that production facilities will have to come on line years before a vaccine has been licensed, and that those facilities will have to be planned for a presently unknown demand,” says Plotkin. “The problem of who will pay the difference between cost of production and price in developing countries remains to be solved.”

The influenza vaccine shortages resulting from contamination at Chiron’s Liverpool plant and the specter of the next flu pandemic have served as a wake-up call, reminding the public and elected officials of the importance and fragility of vaccine production. With the potentially devastating influenza pandemic looming and millions of deaths expected from the existing pandemic of HIV, experts say the need to increase support for vaccine research and manufacturing is more urgent than ever.

*Sheri Fink, M.D., Ph.D., is a freelance writer whose work has appeared in such publications as the New York Times and Discover Magazine, and the author of "War Hospital: A True Story of Surgery and Survival.”