Like the course of a pandemic itself, the great avian influenza scare came in waves. In 1997, the first case of the now infamous H5N1 strain of bird flu emerged in Hong Kong, infecting 18 people and killing six. In late 2003 it struck again—more than 120 people fell ill and at least 60 died over the next two years. Then this fall, David Nabarro, MD, the UN coordinator for avian and human influenza, shocked everyone when he said a bird flu pandemic could claim up to 150 million lives—a figure the World Health Organization (WHO) quickly retracted.

Meanwhile, Americans witnessed nature's destructive capacity at home. Hurricane Katrina struck the Gulf Coast on August 29, and as the waters ebbed over the next few weeks, the nation mourned hundreds of victims. The storm, and the broken levees, left thousands homeless and sparked outrage at the government's failure to prepare for the disaster. The message was clear: The catastrophe may have been inevitable, but better government preparation should have softened the blow.

As media frenzy over Katrina subsided, influenza assumed the front page. A pandemic outbreak was no longer a question of if, said Jai P. Narain, MD, director of WHO's Department of Communicable Diseases. It was a question of when.

Lines of Defense

Dominick Iacuzio, Roche

More important, no one knows what viral structure the avian flu will take. The virus must mutate before it can jump from person to person, and since it travels in fast-moving swarms, pinpointing its future molecular structure can be tricky. H5N1 is the prime pandemic suspect, but even so, "you're guessing it's a specific subtype of H5N1, which you never know," says Norbert Bischofberger, executive vice president of Gilead Sciences. "Your vaccine may not produce antibodies that effectively neutralize the [mutant] virus."

Vaccines are also too slow. Even after a subtype of the pandemic strain is isolated, manufacturing a vaccine will take six to eight months. The 1918 Spanish influenza virus killed as many as 25 million people worldwide in 25 weeks, so six to eight months may be too long to wait.

Antivirals

Norbert Bischofberger, Gilead

Oseltamivir was discovered at Gilead in the mid-90s, shortly after scientists in Australia-based Biota learned that the flu virus's neuraminidase protein could be inhibited. Gilead researchers closed in on a compound around Christmas in 1994. "People were so excited," Bischofberger recalls. "Once you're so close, you really can't sit at home and eat Christmas cookies." Soon after New Year's, Bischofberger and his team had a compound.

Gilead licensed the product to Roche in 1996. The two companies settled a long-running dispute about terms of the deal on November 16, when Roche agreed to pay Gilead $62.5 million in retroactive royalties, along with $18.2 million it had already paid under protest. Gilead's royalty remains 14 to 22 percent, depending on sales volume.