As the nation struggles with a historic shortage of flu vaccine, the last batches of this year's supply are being bottled at a slickly run, modern factory in the Pocono Mountains.
A little to the west, in Amish country, next year's doses are already in the works. At the moment, they have two legs and soft, downy feathers. By January, they will be laying millions of eggs for flu shots.
This system based on seasonal egg-laying has harsh implications for what would happen if a new vaccine had to be made in a hurry to fight a super-strain flu pandemic. At best, it would take half a year.
And because chicken flocks for next year's vaccine are already established and plants already run at full capacity, it is unclear how much Aventis Pasteur or any single company can increase production to cover a shortfall such as this year's loss of Chiron Corp.'s vaccine.
As Chiron executive Kevin Bryett said in an interview two weeks before its British factory was shut down: "If the market was to change dramatically, it is almost impossible to turn up production. The primary issue is access to the eggs. Large numbers of birds cannot be obtained immediately. It's not something you can just go down to the shop and buy eggs off the shelf."
A visit to Aventis Pasteur, the country's only flu shot maker, reveals how much Americans depend on a fragile system to protect them from a killer disease.
Vaccines are biological products, not chemicals that can be cranked out in times of need.
The viruses they are made from need cells in which to grow. Yellow fever and flu vaccines are the only vaccines that use eggs for this -- not the kind that comes from the store but those produced under strict pharmaceutical conditions.
Flu vaccine is also the only vaccine made fresh every year. In late February or early March, the World Health Organization picks its three strains based on the viruses going around. If it acts too soon, newly emerging strains can be missed, as Fujian was last year.
If it waits too long, vaccine makers must race against the biological clocks of hens, which lay eggs only for nine months until they are too old, then rush through a several-months-long process to make enough vaccine before the flu season starts.
"We're constrained on the front end and the back end. No other vaccine has that kind of time pressure," Aventis Pasteur spokesman Len Lavenda said.
Aventis Pasteur starts making the vaccine more than a year in advance, around August, on nearly 50 farms throughout Pennsylvania.
"They're fairly small operations," many with 10,000 birds, said Sam Lee, 40, a chemical engineer who is the company's operations team leader.
White leghorn hens are used. The exact type is a company secret. The breeder holding the patent supplies the eggs, which take 21 days to hatch and become chicks.
They are moved in late September into buildings where they can move freely as opposed to cages and coops, and spend three months maturing into hens.
Egg-laying starts in late December, typically one a day. How many eggs it takes to make a flu shot is another Aventis Pasteur secret, but Chiron's Bryett said: "If you're very lucky, you'll get three doses per egg." That is for a single flu strain; three strains go into each dose of vaccine.
The fertilized eggs are collected by two large egg producers, who incubate them for seven to 12 days and then bring them to Aventis Pasteur. Eggs delivered in January would hatch into chicks if not used for vaccine, so manufacturers often gamble and start making whichever of the three flu strains WHO seems most likely to choose.
"Any production before February is done at the company's risk," Lee said.
A machine punches a tiny hole in each shell, and a needle inoculates the chick embryo with a single flu strain. The virus is allowed to multiply for about three days.
Then the eggs are broken, and the fluid around the embryo containing the virus is collected and purified. Formaldehyde is added to inactivate the flu virus, and a machine spins the mixture to separate out the part containing the virus.
Once again, eggs are needed: A sample of the spun solution is put back in the eggs to see whether any virus grows -- a test to ensure that the germ has been inactivated. A few more processing steps turn the solution into a lot, or batch, of several hundred thousand doses of a single strain.
Next comes sterility testing in which vaccine is spread onto lab dishes and checked to see whether it contains contaminating bacteria.
"We haven't had a contamination in years," Lee said.
It is not known whether it was at this step or at a later one that Chiron's vaccine was discovered to have been tainted with serratia bacteria.
As Aventis Pasteur tests for sterility, samples go to the Food and Drug Administration, which does not start its testing until late May or early June because it has to brew specific chemicals each year to test specific strains.
After that, three viral strains are combined to form the trivalent vaccine. Four weeks of potency and sterility testing follow, then packaging into single-dose syringes and 10-dose vials and quality checks for potency.
The last doses are made by the end of September. Chicken farms are free to sell old birds and must do a mandatory cleanout and disinfection to get ready to start the whole process all over again.
As cumbersome as this egg-based vaccine recipe may seem, "it has been, over the years, a reliable, time-tested and reasonably efficient way to get virus grown," said Anthony S. Fauci, director of the National Institute of Allergy and Infectious Diseases at the National Institutes of Health.
"Changing that requires genuinely a very, very large investment" to make a product sold only once a year and for a pittance compared with expensive cholesterol pills and other drugs that people take every day, said William Schaffner, a Vanderbilt University flu expert who advises the government on vaccines.
"Vaccines in general and flu in particular are undervalued. Prices are not at the level they should be to attract new investments," Aventis Pasteur's Lavenda said.
Aventis Pasteur has partnered with a company in the Netherlands, Crucell, to research the use of human and animal cells in place of eggs. These cell cultures could be maintained indefinitely and ramped up on demand whenever vaccine is needed.
It is pricey technology, but cost is not the only obstacle. Some worry that the genetic material of these cells might interact with the flu virus, creating dangerous hybrids and undercutting the purpose of the vaccine.
"If it changes the virus, obviously, that would not be desirable," Chiron's Bryett said.
Such technology is at least a couple of years away, said Fauci, whose office is funding many studies on it.
"There are safety issues, consistency issues," he said. "That's the reason why you do research on it. If it was just sticking the virus in a cell culture and go, we would have done it a long time ago."
The government has plans in the works to encourage chicken farmers to maintain year-round flocks so eggs would be available anytime in case a new single-strain vaccine has to be made to fight a pandemic.
"Right now, we don't have the capacity to reliably create all the vaccine we would need in that situation," Julie L. Gerberding, director of the federal Centers for Disease Control and Prevention, said in a news conference earlier this year.
"It is clear that we need to substantially expand our options," including looking at whether the vaccine really needs to be updated every year, John Treanor, a government vaccine adviser from the University of Rochester, said in a recent New England Journal of Medicine article.
Public health officials have worried much about bioterrorism, a threat of unknown proportions. Instead, Treanor said, the nation has been caught off guard and threatened by a long familiar foe, "a virus that predictably -- in each and every year -- causes major mortality and morbidity."