The Flu Hunters

By Erik Larson/Hong Kong

It was a subtle warp in an otherwise routine day. Dr. Wilina Lim, chief virologist with the Hong Kong Department of Health, was sorting through the usual load of blood and tissue specimens sent to her laboratory from nearby hospitals, typically about 80 a day. On this particular day--Tuesday, May 20, 1997--one specimen came from Queen Elizabeth Hospital in Kowloon, at the far side of Victoria Harbor, where a three-year-old boy had been admitted with what turned out to be a fatal respiratory illness. Her lab quickly determined that the infectious agent was some type of Influenza A, one of two broad classes of flu virus that commonly affect humans. To identify the specific strain or subtype, the lab tested the sample, using reagents distributed by the World Health Organization. The test kits triggered no response.

Lim was intrigued but not terribly concerned. While she did not often receive flu viruses that resisted identification, it did happen. She retested the virus and again got no reaction. A month later, she forwarded samples to the Centers for Disease Control and Prevention in Atlanta and to England's Mill Hill, two laboratories in the top tier of a quiet but elaborate global surveillance network that tracks changes in the world's flu viruses. Almost as an afterthought, Lim sent a sample to Jan De Jong, a virologist at the Dutch National Institute of Health and the Environment who liked to collect unusual strains of influenza. She had never met De Jong, but over the years they had developed a rapport.

For more than a month, she heard nothing. Then suddenly, on Friday, Aug. 8, De Jong called. He was coming to Hong Kong. He had booked a flight that day. He would arrive Sunday. It seemed, at first, just a friendly visit--a chance, at last, to meet face to face.

Lim picked him up at the Kowloon Ramada on Monday morning. As she drove back to her laboratory, high in Hong Kong's craggy western hills, De Jong turned to her and asked mildly, "Do you have any idea what virus you sent me?"

THE HONG KONG INCIDENT

By now most of the world has heard of the "bird flu" that emerged in Hong Kong last year, infecting 18 people and killing six. One patient, a young woman, remains on a ventilator under intensive care. Although no new cases have been discovered since Dec. 28, virologists consider the emergence of this new virus one of the most significant and worrisome medical events of the day. And they don't think the danger has passed. In fact, the critical period could just now be arriving in Hong Kong. This is the start of the traditional flu season, when the new virus could, in theory, combine with ordinary human strains to create a supervirus that is both lethal and highly contagious.

While the outbreak highlighted the success of the surveillance network, it also showed how dangerously mutable influenza viruses can be and that, in their most sinister forms, they can be as deadly as any other disease known to man, more akin to Ebola than to the fevers and aches most people associate with flu. Virologists say the decision to kill all the chickens in Hong Kong--widely derided at the time--was in fact the smartest thing that could be done and that it might have prevented a more widespread disaster. "The question is," says Robert Webster, chairman of the virology department at St. Jude Children's Research Hospital in Memphis, Tenn., and a key actor in the quiet drama played out in Hong Kong, "did they close the stable door before or after the horses had gone?"

The CDC grabbed most of the headlines with its usual Ghostbusters aplomb, and even dispatched a public relations officer to accompany the agency's team of epidemiologists. But at least three investigations coalesced in Hong Kong. Only by following all three does the true significance of the outbreak become clear. Taken together, these threads weave a story that begins 80 years ago and winds forward through venues as varied as a high-security lab in Ames, Iowa, the ancient tissue collections of the Armed Forces Institute of Pathology in Washington and a frozen mass grave on Alaska's Seward Peninsula.

The Hong Kong Incident, as Webster calls it, arrived with cinematic timing--an almost supernatural confluence of event and inquiry. It occurred amid heightened sensitivity to the dangers of newly emerging viruses and just as several teams of researchers were closing in on the mysterious 1918 "Spanish flu," which killed more than 20 million people. At the same time, it turns out, public-health officials were quietly intensifying plans for the next great global epidemic, or pandemic.

While the rest of the world was wringing its hands over the remote threat from such exotics as Ebola and hantavirus, the health officials were busy staring down a far more likely global disaster and produced a closely held Pandemic Planning Document. In the course of their meetings, the planners are said to have wrestled with such issues as what to do if the President dies and how to deal with masses of dead or severely ill citizens, considerations reminiscent of civil-defense planning for nuclear war. The planners are so certain that another worldwide epidemic will occur that they refer to the present as the "interpandemic period."

The full story of the Hong Kong Incident begins in 1918 with the most lethal epidemic in human history, one that eclipsed even the medieval Black Death. "It's why we do what we do every year," says Roland A. Levandowski, the Food and Drug Administration's chief flu expert and a member of the pandemic planning group. "This experience in Hong Kong, even if it doesn't go anywhere, is a reminder that these things can happen."

1918

The pandemic of 1918 remains a mystery. It began with a relatively mild initial assault on March 4, when the first reported case occurred at Camp Funston, Kans. Within four months, the virus had traversed the globe. The flu sickened millions but killed relatively few, and in the tumult of World War I, the first wave seemed pretty mundane.

No one knew it at the time, of course, but flu viruses are notoriously unstable-- "genetically labile," as one researcher puts it. Set one flu virus beside another, and the two may trade genes, a process called reassortment. If this reassortment produces a virus that closely resembles one of its parents, it is said to have undergone antigenic drift. On rare occasions, this scrambling can be dramatic. The virus becomes a kind of Frankenstein virus so different from existing strains that the human population has no immunity to it.

In August 1918, the mild virus apparently reassorted into something positively deadly. Outbreaks caused by the new variant exploded almost simultaneously in three far-flung locations: France, Sierra Leone and Boston. The flu struck with a ferocity that shocked doctors, who feared this strange new pathogen might be an airborne version of the Black Death. Patients died awash in blood and gore, literally drowning as fluid filled their lungs.

The virus rocketed to the farthest points of the globe. From September 1918 through March 1919, it killed 33,387 people in New York City, just over 1% of the city's population. In some Alaskan villages, the death toll topped 50%; in one, Teller Mission (now Brevig Mission), 85% were dead within a week.

One of the great mysteries of 1918 centers on who was killed by the virus. Even ordinary flu will cause deaths among the very young, the very old and people with a weakened immune system. The 1918 virus did kill within these groups, but it seemed to have a special passion for the young and hardy, ages 25 to 34, those typically most able to weather the flu.

Rumors flew of strange influenza-like diseases affecting animals, even moose, according to the pandemic's chronicler, Alfred W. Crosby Jr. One rumor turned out to be true--disturbingly so for anyone familiar with the subsequent history of influenza research and the recent Hong Kong outbreak. Farmers in 1918 discovered that something was making their pigs very sick, with high fevers and bad coughs. No such pig flu had ever been noticed before 1918, but every fall thereafter an influenza-like illness attacked the nation's hog population. In 1928 a researcher from the Rockefeller Institute, Richard E. Shope, went to Iowa to investigate the phenomenon, and in 1930 he became the first scientist to isolate an influenza virus. Copies of it are stored today in laboratories around the world.

The 1918 strain of influenza persisted into the '20s, then disappeared, or lost its virulence and faded into the great jigsaw of constantly reassorting viruses. Until lately, the epidemic had almost disappeared from our collective memory as well, prompting Crosby to title his history The Forgotten Epidemic. Among flu experts, however, its mysteries are still current and utterly significant. It has always stood as a vivid warning of what the next pandemic could be like. What made the virus so lethal? Why was it able to kill so quickly? And where in nature did it originate?

Last year flu researchers found themselves asking the same questions once again, but this time because of the strange events in Hong Kong.

THE NEW TERRITORIES

It was March of 1997 when the chickens began to die--6,800 on three farms in Hong Kong's rural New Territories. Because poultry is a vital part of Hong Kong's diet, agricultural authorities got concerned and quickly consulted Kennedy Shortridge, a microbiologist at the University of Hong Kong. He in turn contacted his friend and fellow flu specialist Robert Webster of St. Jude. For decades both men had studied influenza viruses in chickens and other birds in the belief that these viruses were more than just an agricultural problem and might hold the key to the origins of human influenza, possibly even the virus of 1918.

Shortridge and Webster immediately recognized the gravity of the chicken-flu outbreak in Hong Kong, at least for the region's chicken industry. They knew that while avian influenza did not ordinarily make its host sick, a benign virus could reassort to produce a pathogen of almost inconceivable lethality. Webster's Memphis lab had observed such a transformation in the wild on two occasions, the first in April 1983, when a relatively mild influenza struck chickens on the vast chicken farms of Pennsylvania. The birds got visibly sick, some died and egg production fell, but overall the outbreak remained only a vexing economic problem.

By October, the virus had changed. Before, it attacked the respiratory and intestinal tracts of chickens; now, suddenly, it assaulted every tissue in the chickens, including the brain. It caused all their blood vessels to leak and killed them within days, turning the birds, as one researcher put it, into "bloody Jell-O." Federal inspectors arriving at Pennsylvania farms found themselves walking through factory-size chicken coops struck eerily silent, with thousands of dead or hemorrhaging chickens at their feet. The U.S. Department of Agriculture ordered the extermination of 20 million chickens in Pennsylvania, more than 10 times the number that would be killed in the Hong Kong chicken slaughter.

Webster assigned a young scientist, Yoshihiro Kawaoka, to try to figure out how the virus transformed itself into such a "hot" pathogen. Kawaoka, now a professor of virology at the University of Wisconsin, Madison, compared the genetic structure of viruses from the first and second waves and found only a single, extremely subtle change in the H gene. The two viruses differed by just one nucleotide--one of 1,700 nucleotides that made up the gene.

Last year, on two of the three farms stricken in Hong Kong, mortality was 100%. The scientists knew the virus had a variation of the H gene known as H5--one that is notoriously lethal to chickens. Shortridge did briefly wonder if the virus might eventually cause problems for humans. In an earlier study, conducted with great discretion, his lab had found that residents of rural Hong Kong had antibodies to all the known bird-flu viruses. What that suggested, says Shortridge, was that "any virus could cross the species barrier to humans. But whether it could set up an infection, be established as an infection and maintained as an infection is, of course, another matter."

Shortridge, in Hong Kong, asked Webster, in Memphis, if he could help him arrange to ship a sample of the deadly virus for in-depth analysis to the U.S. Department of Agriculture's high-security laboratory in Ames, Iowa. When the package arrived, it was sent to a P3+ containment laboratory--one notch below the P4 level required for studies of Ebola virus--where Dennis Senne inoculated the virus into chicken eggs and chickens to gauge its pathogenicity. It killed 10 out of 10 chickens; each died within one or two days.

Senne then subjected the virus to detailed genetic analysis, a process known as gene sequencing. On the H gene at a point called the cleavage site, he found a telltale mutation, the same kind of mutation found in other highly pathogenic avian viruses. Senne shipped his findings and samples of the virus to Webster, who analyzed its viral heritage. The virus, he discovered, had regions that were identical to portions of the avian virus that struck Pennsylvania in 1983.

The outbreak in Hong Kong was quickly contained. All birds on the three farms were destroyed. And that's where Webster and Shortridge left it. "At that point," Webster remembers, "it was merely interesting."

THE WASHINGTON CONNECTION

In March, even as the chickens were dying, a molecular pathologist at the Armed Forces Institute of Pathology in Washington, Dr. Jeffery Taubenberger, startled the flu research community with a paper in the prestigious journal Science in which his team claimed to have at least partly penetrated the fog surrounding the 1918 pandemic. The coincidence was striking: just as a new virus was emerging in Hong Kong, here was fresh news about the mother of all epidemics.

Taubenberger's work began not out of some great passion to plumb the mysteries of 1918 but rather a desire to showcase two of the Pathology Institute's crown jewels: its vast collection of tissue specimens gathered over the past century, and its new technique for extracting RNA from biological materials fixed in Formalin and paraffin. Even he, however, wondered if the institute's tissue repository, "the annex," would be able to locate such old specimens. He had never seen the place and pictured it as a forlorn vault like the vast warehouse in the closing scene of Raiders of the Lost Ark. Taubenberger and colleague Ann Reid put in a request to the annex for tissue samples of three dozen soldiers who had died in the 1918 pandemic. They then settled in for what they assumed would be a long wait.

They assumed wrong. The repository, housed in a nondescript building 10 minutes away in Forest Glen, Md., is not the gloomy storehouse they imagined. A few seconds after receiving Taubenberger's request, the annex's robotic retrievers had located the laboratory slides associated with his cases, rising on quiet greased chains to retrieve them from the upper reaches of a 10-ft.-tall, room-length revolving carousel. A few days later, a collection of small brown lunch bags turned up at Taubenberger's office, each marked with a case number, each containing flecks of tissue taken from a young soldier killed by the flu nearly a century earlier, by doctors struggling to cope with a lethal epidemic they did not understand. For Taubenberger and Reid, it was a strangely haunting moment.

The hard work was just beginning for Reid. She would spend the next year searching through the samples for a snip of 1918 RNA.

HONG KONG, AUG. 11

As Wilina Lim drove back to her laboratory with Jan De Jong, the Dutch researcher, she considered his question about the nature of the virus she had sent him. Clearly he already knew what it was. She thought a moment, then guessed the virus was probably an H3, common in humans, that had changed sufficiently to evade detection.

"No," De Jong told her. "It was H5."

Lim was startled. "I'm not a vet," she says. "I don't know much about influenza in animals." But she had never heard of H5 infecting humans. For it to do so now was surprising. Even impossible.

And suddenly she understood why De Jong had felt it necessary to come in person to Hong Kong, why he had waited until now to tell her about the virus. He suspected that the H5 had not really come from human patients but was the result of laboratory contamination. Everyone knew that her lab was situated close to Shortridge's and that Shortridge worked with avian viruses. Moreover, this was Hong Kong, where poultry stalls with live chickens could be found in the same neighborhoods as five-star hotels. "I think he came to Hong Kong to have a look-see if it was a sloppy laboratory," says Lim. She knew his concern was justified, but still it offended her. She is known for her buoyancy, but at this moment her expression hardened. "I knew it was not a contaminant," she says, "because I know my lab."

And soon De Jong was also convinced. That night he spoke with Albert Osterhaus, chairman of the virology department at Erasmus University in Rotterdam, where virologist Eric Claas had analyzed the suspect virus using a panel of reagents derived from flu strains isolated and maintained by Webster. Claas had first determined that the virus was H5N1, well before the CDC and Mill Hill. At the outset even he did not believe it. An H5 infection in humans was unheard of. He too assumed the H5 was a contaminant.

Meanwhile Osterhaus had called Webster in Memphis to learn more about H5. Only then, in that phone call, did the human-flu research community at last learn of the earlier outbreak of chicken flu on the three Hong Kong farms; and only then did Webster and Shortridge learn of the first human case--even though Shortridge's laboratory and Lim's are housed in adjacent buildings.

Webster already had the virus in his collection, its genetic structure detailed, its heritage mapped. He recalls, with obvious delight, how he told Osterhaus, "Abe, I have the precursor of this virus in my laboratory."

For Webster, it was an exciting moment. "The situation in Hong Kong is what I've been predicting throughout my career," he says. For years, he contends, people have dismissed avian flu "as a problem of chickens--who cares?" He revels in his newfound credibility. "Finally," he says, laughing, "at the end of my career, the chickens have come home to roost."

He concedes, however, that he was startled when Osterhaus told him about the three-year-old boy who had died on May 21, the day after Lim received his specimen. Webster also wondered whether the H5 was merely a contaminant. Osterhaus assured him it was not. After the call, Webster taped a note to the wall over his desk: H5 IN A CHILD!

THE NEWS SPREADS

Like Webster, virologists around the world were galvanized. The CDC, alerted by Claas, quickly tested its own copy of Lim's virus and confirmed the finding. In San Francisco, Dr. Keiji Fukuda, chief epidemiologist for the CDC's influenza section, was doing a clinical rotation at Mount Zion Hospital when he received an urgent call from the agency's head of surveillance. "Whenever you get a call like that," he says, "you know it's probably not great news." Shortridge was vacationing in England when his phone went wild. "The first thing that crossed my mind was, 'Is this the start of a new pandemic?' " he recalls.

To anyone who knew influenza, the news instantly raised the specter of 1918. Or worse, as this was a purely avian virus against which most humans would have no defense. The world, moreover, was far more densely populated, and high-speed travel now linked all the major cities. In 1918, when transportation was still painfully slow, the pandemic circled the globe in a matter of months. Traveling by jet, a new killer virus could reach Tokyo in three hours and New York City within a day.

The fact that the new virus did not seem readily transmittable from person to person was a consolation, but flu experts know that influenza viruses are utterly unpredictable. In Hong Kong the big question was this: Would the H5 reassort with a common human strain to produce a new virus that was as lethal as H5 but could be passed along by a human sneeze? Or would this new H5 virus, through repeated exposure, find some other way to adapt to human hosts? "That's an interesting point," says Shortridge, "because it raises questions about the 1918 pandemic. Did a similar sort of thing happen?"

PRIVATE ROSCOE VAUGHN

Back in Washington, Taubenberger and Reid had decided to concentrate on the seven cases in which the victim had died most quickly, figuring that these specimens would be most likely to retain the genetic remains of the virus. They found plenty of RNA, but none of it looked like flu--until, after a full year's work, they came to Private Roscoe Vaughn.

Vaughn was a 21-year-old soldier at Camp Jackson, S.C., who reported for sick call on Sept. 19, 1918, at the peak of the pandemic. He complained of chills, fever, headache and a bad cough. He had trouble breathing. A week later, at 6:30 a.m., he died. At 2 p.m., his body was autopsied, and specimens were extracted, preserved and sent to Washington.

Using an array of powerful if arcane gene-hunting tools, Taubenberger and Reid slowly picked their way through the shattered genetic landscape of Private Vaughn's cells. This time they got lucky. They found small pieces of flulike RNA. Their subsequent analysis showed that the virus was an H1N1 influenza unlike any flu virus identified during the past 80 years. The closest known strain was Swine Iowa 30--the pig flu isolated by Richard Shope in 1930 and kept alive at various culture repositories ever since. Their findings suggest that the 1918 virus came to people from pigs, not from birds--although Taubenberger cites studies by Webster and others indicating that human viruses and the pig flu of the 1930s may share a common avian ancestor. This suggests that sometime before 1918, a bird virus could have entered the mammalian population and, through reassortment, produced the pathogenic flu virus known to man.

HONG KONG, AUG. 20

The CDC's Fukuda arrived in Hong Kong on Wednesday, Aug. 20. The next day he and a team of CDC investigators joined an intensive investigation already being conducted by the Hong Kong Department of Health. Working with health-department officers, Fukuda and his colleagues conducted scores of interviews and collected hundreds of blood samples, trying to figure out how the first victim, the three-year-old boy, could have contracted a virus that infects only birds.

The CDC took a hard look at the boy's preschool, in particular a corner of his classroom set aside as a kind of nature corner, with live chicks and ducklings. Fukuda knew that the birds had died before the boy got sick, but no one knew what killed them. The team swabbed the classroom floor to try to capture some of the virus, but found none. Although press reports suggested a close tie between the death of the classroom birds and the boy's illness, Fukuda says the source of the boy's infection is by no means certain. "It was unclear then," he says. "It is unclear now."

The CDC's investigation of the boy's illness lasted 2 1/2 weeks. By the time Fukuda left Hong Kong, his team had collected 2,000 blood samples. Antibodies indicating previous exposure to H5N1 were found in only nine samples, including one of the boy's classmates and one of his doctors. None of the nine recalled being ill. The fact that so few showed signs of exposure was concrete evidence that the virus was not particularly contagious.

For the moment, there appeared little reason to fear that this first case, however tragic, represented the start of a pandemic. Says Fukuda: "I left thinking, 'You know, this is probably some odd, sporadic thing.'"

He expected no more cases.

ALASKA'S "LUCY"

In another odd coincidence, that same August, as Fukuda investigated the new virus in Hong Kong, the quest to understand the 1918 epidemic suddenly gained momentum, with help from a surprising quarter. Out of the blue, Taubenberger got a letter from a retired San Francisco pathologist, Johan Hultin, who had read Taubenberger's paper in Science and saw at last an opportunity for which he had been waiting for nearly a half-century.

In 1951 Hultin took part in an expedition to Alaska to try to extract live virus from long-frozen victims of the 1918 flu in what is now Brevig Mission, Alaska. Now he was ready to try again. He knew from hard experience that no live virus had survived under the permafrost. But Taubenberger's paper convinced him that technology had advanced to the point where even a dead virus could be of immense value. The moment he saw the Science paper, he told himself, "There. This is it."

Hultin asked Taubenberger whether he would accept and analyze samples of lung tissue from frozen graves, if he, Hultin, went to Alaska to get them. "When are you planning to leave?" Taubenberger asked. He knew firsthand that such ventures take a lot of advance planning. "I can't go this week," Hultin told him. "But I can go next week." Taubenberger got really quiet. "I don't know what was going through his mind," Hultin says, chuckling. "He probably thought I was some kind of a nut."

Two weeks later, Hultin was on his way--one 73-year-old man with a sleeping bag, a carry-on bag and two duffels full of equipment. He traveled solo and avoided publicity so as not to raise too much fuss among Brevig's villagers. By the afternoon of Aug. 20, he and a local crew had begun digging, and they eventually produced a trench 6 ft. wide, 27 ft. long and 7 ft. deep. Hultin came across several bare skeletons before he hit pay dirt: the well-preserved body of a 30-year-old woman so obese that her fat had insulated her organs from the effects of decades of frost and thaw. He took both lungs, sliced them into thin strips and carefully packed them for shipment. Hultin named the body "Lucy," a nod to the prehistoric Lucy who shed so much light on human origins.

Hultin expected to wait months to hear of any results, but Taubenberger called within a few weeks. He had found fragments of the 1918 virus in Hultin's Lucy. Taubenberger and Reid had meanwhile recovered yet another sample of 1918 virus from tissues in the Armed Forces annex. Taken together, the three samples put to rest any doubt that Taubenberger's lab had indeed found and sequenced key portions of the original Spanish-flu virus.

Hultin says he was struck by the uncanny timing of his journey, which took place just as a strange virus with great pandemic potential was emerging in Hong Kong. "I was very apprehensive," he says. "I was waiting for it to come--and it didn't." But another pandemic, he believes, is inevitable. He has given his wife instructions on what to do to survive it: retreat to their mountain cabin until the onslaught passes. It was a tactic, he knows, that was successfully used in 1918 by a village just 30 miles from Brevig. Its elders, after learning of the advancing plague, stationed armed guards at the village perimeter with orders to shoot anyone who tried to enter. The village survived unscathed.

"THE VIRUS IS MOVING"

Hong Kong, in the meantime, had begun to relax. From August into November, nothing happened. No new cases appeared. In postmortems on the first case, researchers congratulated themselves on how well the global flu-surveillance system had worked. Some even suggested that it worked too well, that the avian flu had been discovered only because the surveillance network was looking for such events and that isolated bird-to-human infections had probably happened before and gone undetected.

On Nov. 8, Lim's virology lab got its usual load of new specimens to analyze, including one from a two-year-old boy admitted the day before to Queen Mary Hospital. Her lab applied the ordinary WHO reagents for H3 and H1, but just as in May, got no reaction. This time Lim tried an H5 reagent supplied by the CDC. And got a positive reading.

By now, however, the patient had already been discharged, well on his way to full recovery. In fact, he had been only mildly ill and was admitted because of a heart condition that made him vulnerable to even routine infections.

"So now I think, 'This cannot be,'" says Lim. Perhaps it was contamination, after all; maybe this H5 reading had been caused by the presence of the H5 she had grown and tested in May. She asked the hospital to send over anything that remained of the material originally swabbed from the boy. This too tested positive for H5. "Now I'm worried," she says, "because after six months it came out again."

Webster was in Memphis driving home from a Saturday at work when his wife told him he had received a call from the CDC. He called back, waited, called again, and this time got the news: "The virus is moving."

OUTBREAK

On Dec. 1, Queen Elizabeth Hospital sent Lim a specimen from a 54-year-old male who had developed a fever and a cough and soon had to be admitted, apparently suffering from pneumonia. Four days later, Lim's lab succeeded in growing a virus extracted from his specimens. The next day, the patient died. Lim tested the virus with her H5 reagents. Again, a positive.

Lim knew that she had another specimen in her lab, taken from a 13-year-old girl admitted to Prince of Wales Hospital so sick that she had been placed on a respirator. The hospital had identified the underlying virus as Influenza A but wanted Lim to determine the subtype. Lim asked her lab technicians to come in early the next morning, Saturday, Dec. 6, to test specimens from the two patients. Both again reacted to the H5 reagents.

By then the CDC's Fukuda was already in the air, aware only of the initial two cases--not Lim's most recent discoveries. That weekend Lim's daughter complained of a sore throat. Instantly the dispassionate virologist became a frightened mother. She barred her daughter from all sports and canceled her piano lessons.

As his plane crossed the Pacific, Fukuda consoled himself with the fact that in six months only two cases of H5 flu had occurred. Upon his arrival, a medical officer with the Hong Kong Department of Health greeted him warmly, then gently told him of the latest discoveries. "The good news," the officer said, "is we will have a nice dinner. The bad news: there are two more cases."

This brought the total to four. Two patients had died. One was on a respirator. Fukuda, a member of the pandemic planning committee, suddenly had a glimpse of what it might be like to confront an explosive outbreak. It is one thing to plan rationally, he says. "It's another thing all of a sudden to be struck with a sense that, my God, what will happen if there are a thousand cases like this? What will happen to all of those people? How will hospitals cope? How will any of us cope?"

In short order, more cases began turning up throughout Hong Kong. On Dec. 4, a 24-year-old woman developed a fever, sore throat and cough, and complained of dizziness. Five days later, she was in the intensive-care unit on a respirator with a confirmed case of H5 influenza. On Dec. 7, a five year-old girl began vomiting and developed other flu symptoms. H5 again. On Dec. 12, another child, a cousin of the five-year-old, came down with a fever and was hospitalized with H5. And a new outbreak of H5 had turned up on a fourth chicken farm in the New Territories.

In Fukuda's war room, Room 58 of the health department's downtown headquarters, a large whiteboard listed all the cases and tracked their medical progress. A big downward arrow meant death. With new urgency, Fukuda and the CDC hunted the sources of infection, collecting 3,000 blood samples and helping question some 2,500 people.

THE MARKETS

To Webster and Shortridge, Hong Kong's many outdoor markets held the key to why the confirmed cases of H5 were spread in such haphazard fashion throughout Hong Kong. In some cases, the CDC team and health-department investigators were unable to prove direct contact with poultry, which suggested that some of the victims caught the virus through contact so casual they simply weren't aware of it. Says Shortridge: "It suggested to me there was a hell of a lot of virus in the environment that we weren't aware of."

Webster and Shortridge quickly arranged an ad hoc task force to begin testing poultry in the city's "wet" markets, so named because retailers use water to clean their stalls and adjacent sidewalks. The group began its probe on Dec. 22 and worked 18 hours a day right through Dec. 28, the day Hong Kong authorities began their territory-wide slaughter. The research showed that 10% of chickens in the markets carried the virus. Ducks and geese in the markets carried it too--especially worrisome, given their ability to carry infections without outward sign of illness. In the markets, all poultry--ducks, geese, chickens--was killed. The slaughter, according to Shortridge and Webster, removed a substantial reservoir of H5 virus from contact with people.

Then suddenly, almost as soon as it started, the second outbreak seemed to be over. The last case occurred on Dec. 28, the day the slaughter began. By late January, Fukuda's whiteboard in Room 58 showed 18 confirmed cases, with six downward arrows.

EPILOGUE

A killer had come and gone, raising new mysteries even as old mysteries from 1918 were being solved. What allowed this avian virus to cross the species barrier and set up killing infections in man? Why did it strike the young and hardy with the most ferocity--just as the 1918 virus had? And, most important, has the virus really ceased to be a threat, or is it circulating more quietly, primed for a "reassortment event" that will set off the next global disaster?

So far, the new virus has shown no evidence of reassortment. The fact that the outbreak happened in December, before Hong Kong's regular flu season, reduced opportunities for reassortment, as did the prompt slaughter of the chickens. But the flu season is coming. It will peak in late February and early March, with a second peak this summer. What researchers fear most is that someone infected with a common flu strain will also become infected with H5, and thus become an inadvertent mixing chamber for the production of a wholly new virus.

Webster and Shortridge are convinced that the avian virus is still circulating in the environment. "I don't think we're out of the woods yet," says Shortridge. Fukuda agrees: "You would be a fool to predict what the virus is going to do next. I'm equally prepared for this thing to disappear as I am to hear one day when I walk into the office, 'Oh, did you hear? There's another 10 cases--or 100 cases.'"

It would be easy to dismiss the Hong Kong Incident as just a one-time quirk of blood and protein. But the U.S.'s leading flu experts seem unwilling to do so. This became particularly apparent at the annual meeting of the FDA Vaccine Advisory Panel, convened two weeks ago in Bethesda, Md., to decide what flu strains should be targeted for next year's flu vaccine. Ordinarily these meetings are routine, if not downright boring. But this year the committee devoted half the day to the Hong Kong outbreak.

For Webster, it was a striking moment--the first time he had ever been invited to the meeting, a point he made clear in the opening moments of his talk. Equally striking, no one on the panel tried to minimize the potential danger of the new avian virus. Far from it. In a vote the FDA had not even requested, the committee unanimously agreed to move ahead to develop a vaccine against H5, even take it through clinical trials.

Shortly before the vote, Webster was asked his opinion. He believes the Hong Kong Incident may have given the world early warning of more H5 outbreaks to come. "We have a window of opportunity," he told the assembled scientists. "Let's do it now."