What We Learned from H1N1. Part 1: MedImmune
Now that more than a year has passed since the outbreak of the H1N1 influenza pandemic, manufacturers and regulators alike are reviewing what was learned from the process. As part of a news series on this topic, BioPharm International interviewed Ben Machielse, Drs, executive vice president of operations at MedImmune, to ask about his company's experience manufacturing vaccines against H1N1. Below are several of the lessons he feels we should apply to preparations for future pandemic outbreaks.
A Live, Attenuated Influenza Vaccine Is a Useful Part of a Pandemic Arsenal
One of the major difficulties that delayed the production of vaccines against the H1N1 virus was that in most cases, the H1N1 virus did not grow well. MedImmune was the exception to the rule, and was able to produce ~80 doses per egg of its live, attenuated influenza vaccine (LAIV), compared to 2 to 3 doses per egg for the inactivated vaccines that comprise the bulk of the flu vaccine supply.
The reason, Machielse says, is that MedImmune has a lot of experience developing flu virus strains. Unlike the manufacturers of the inactivated flu vaccines, who receive a feed strain directly from the Centers for Disease Control (CDC), MedImmune, the only company with a licensed LAIV in the US, develops its own strains for its influenza vaccine production; this is true for MedImmune's vaccines against seasonal flu as well as those against pandemic strains like H1N1.
"We have a lot of institutional knowledge about how to screen strains for productivity," says Machielse. "Based on that, we predicted to BARDA [The Biomedical Advanced Research and Development Authority] that we could produce a lot of vaccine quickly, and in fact, we did," he says.
MedImmune received the wild-type strain of the H1N1 virus from the CDC at the end of April, and after screening about 25 strains, began manufacturing bulk vaccine by July 3, in about 10 weeks. And if the company, like other vaccine manufacturers, had not been completing production of the seasonal flu vaccine at the same time, it could have begun making H1N1 vaccine even faster.
"We could have done it in five to six weeks," Machielse says.
Part of the speed can be attributed to MedImmune's use of plasmid rescue to develop its strains. Unlike the classical reassortment method, in which the genes of the wild-type strain and the strain developed for attenuation are mixed in an uncontrolled fashion, plasmid rescue allows developers to selectively clone two genes for the surface antigens of the current wild-type virus with six genes from the attenuated strain, and combine them precisely.
"Not only is plasmid rescue faster, but our predictability also is higher," says Machielse.
FDA and Industry Should Collaborate Now to Prepare for the Next Pandemic
Machielse says that the regulatory process went well during the H1N1 response, but the experience can be used to accelerate the process for future pandemics.
For example, a key decision facing regulators in this pandemic was whether to evaluate H1N1 vaccines as new products or, as was ultimately decided, to license them as strain changes, the way seasonal flu vaccines are licensed every year. The time needed to make that decision delayed the approval of product inserts by two to three weeks.
"Now that we know that a pandemic flu vaccine can be approved as a strain change, we manufacturers should propose a template product insert for use with future pandemic strains," says Machielse. The wording of the insert could be based on language used for seasonal flu vaccines; the product name and other key information then could be inserted and reviewed quickly, he says.
"This would take the product insert off the critical path," he says.
Machielse believes that now is a good time for the FDA, BARDA, and industry to talk about this proposal and other ways to accelerate pandemic response, such as using new technologies to accelerate stability or potency testing. "I believe there is openness at the agency and within industry, and we could develop a meaningful dialogue," he concludes.
A Hub and Spoke System Could Improve Worldwide Distribution
Another problem with the H1N1 pandemic was uneven worldwide distribution of vaccine, with excess supply in developed countries and shortages in the developing world. Machielse says a hub-and-spoke model could help address this challenge.
In such a model, bulk product would be manufactured in a central location and distributed worldwide for local filling. "This would simplify distribution, because it is easier to ship bulk than finished dosage forms," he says.
The key would be to develop a cost-efficient method for carrying out fill–finish on the local level. Setting up such a system would require a lot of work, and public–private partnerships, Machielse says, but is worth pursuing.
"We need to commoditize the fill–finish component," he says. "That's our target."
Even in the United States, fill–finish operations are the bottleneck, Machielse says. "We need to continue to evaluate alternatives to accelerate fill–finish operations," says Machielse. "We want to be able to ramp up fill–finish to match our speed in manufacturing bulk product."
Public Perception Is Still a Problem
Apart from manufacturing issues, public perception is critical, Machielse said. "We had long lines for H1N1 vaccines early on, before the large volumes of vaccines were ready, and then when we had them, the public was suspicious," Machielse notes. "Because we produced it so quickly, they thought we must have cut corners."
"We need to explain to the public that this was not a new product, just a different strain, just as we do with the season flu vaccine every year," he says. "Public messaging is important around flu vaccines."