The Great Influenza: The Epic Story of the Deadliest Plague in History - John M. Barry (2004)


I STARTED THIS BOOK intending to explore not only the 1918 pandemic itself, but also several questions that did not involve influenza per se. One involved how the larger society reacted to an immense challenge. Another confronts anyone making a decision: What process do you follow to collect information that most likely leads to a good one? In short, how do you know when you know?

More narrowly, I also wanted to explore how an investigator should do science, even under the most stressful conditions. William Park, Oswald Avery, and Paul Lewis speak especially to this last point. They were very different people. Each approached science in his own way.

Park saw it as a means to a larger end. To him, a man who almost became a medical missionary, it was a tool to relieve suffering. Disciplined and methodical, he was interested chiefly in immediate results useful for that purpose. His contributions, particularly those made with Anna Williams, were enormous; their improvement of diphtheria antitoxin alone doubtless saved hundreds of thousands of lives over the past century. But his purpose also limited him, and limited the kind of findings he and those under him would make.

Avery was driving and obsessive. Part artist and part hunter, he had vision, patience, and persistence. His artist’s eye let him see a landscape from a new perspective and in exquisite detail, the hunter in him told him when something no matter how seemingly trivial was out of place, and he wondered. The wonder moved him to sacrifice all else. He had no choice but to pursue it. Cutting a Gordian knot gave him no satisfaction. He wanted to unfold and understand such things, not destroy them. So he tugged at a thread and kept tugging, untangling it, following where it led, until he had unraveled an entire fabric. Then others wove a new fabric for a different world. T. S. Eliot said any new work of art alters slightly the existing order. Avery accomplished more than that.

Paul Lewis was a romantic, and a lover. He wanted. He wanted more and loved more passionately than Park or Avery. But like many romantics, it was the idea of the thing as much or more than the thing itself that he loved. He loved science, and he loved the laboratory. But it did not yield to him. The deepest secrets of the laboratory showed themselves to Lewis when he was guided by others, when others opened a crack for him, but that crack closed. When he came alone the laboratory presented a stone face, unyielding to his pleadings. He could not find the key, the way to ask the question. Of the three, only he could not penetrate it. And, whether his death was a suicide or a true accident, it killed him.

But one cannot leave this subject without speaking to other questions: the likelihood and potential danger of another influenza pandemic, what we can learn from the one of 1918–1919, and how we can apply those lessons to the emergence of a new pathogen, whether that pathogen is a weapon of terror or a new natural menace—such as Severe Acute Respiratory Syndrome, SARS, the disease which spread from animals to man in the spring of 2003 and threatened to become a major pandemic.

The answer to the first question—the likelihood and potential danger of another influenza pandemic—is not reassuring. Every expert on influenza agrees that the ability of the influenza virus to reassort genes means that another pandemic not only can happen. It almost certainly will happen.

For influenza is not like SARS, which was contained and—as this book goes to press—may have been completely eliminated. SARS, although more lethal even than the 1918 influenza virus, is less dangerous for several reasons.

First, SARS requires fairly close contact to spread, while influenza is among the most contagious of all diseases. Also, in SARS, the virus reaches maximum concentration in the upper respiratory tract, where coughs and sneezes are most likely to spread the virus, a week or longer after symptoms develop. This gives public health officials time to find, identify, and isolate cases. By contrast, the influenza virus can spread from person to person before any symptoms develop, before a victim knows he or she is sick.

If a new influenza virus does emerge, given modern travel patterns it will likely spread even more rapidly than it did in 1918. It will infect at least several hundred million, and probably more than a billion, people. In the United States alone, the Centers for Disease Control estimates that a new pandemic would make between 40 and 100 million people sick. So the prospect is threatening indeed.

If one compares the 1918–1919 pandemic to AIDS, one sees how threatening.

Today the world population exceeds 6 billion. Worldwide, in the twenty-four years since AIDS emerged as a disease, the total death toll is estimated at 24,800,000; at this writing, an estimated 42 million people are currently infected with the HIV virus. In the United States the cumulative death toll from AIDS is 467,910 people.

In 1918 the world’s population was 1.8 billion, less than one-third today’s. Yet the 1918 influenza virus killed a likely 50 million and possibly as many as 100 million. The AIDS deaths occurred over twenty-four years; most of the influenza deaths occurred in less than twenty-four weeks.

There are now drugs that can contain the HIV virus; the difficulty lies in getting those drugs to the poorest parts of the world as well as in educating people there and in countries, such as China, that continue to minimize the disease. In the United States, those drugs limited AIDS deaths to 8,998 people in the most recent year for which statistics are available.

The U.S. Centers for Disease Control (CDC) estimates that the annual death toll in the United States from influenza now averages 36,000 in a nonepidemic year. The 1918 virus killed 675,000 people in the United States, out of a population not much more than one-third the size of today’s.

In 1999 the Centers for Disease Control produced a study of what would likely happen if a new pandemic virus struck the United States. It took into account modern medical advances.

Antibiotics would of course significantly cut 1918’s mortality rate for secondary bacterial infections following influenza. And several antiviral drugs have demonstrated some effectiveness against influenza. Amantadine and its more recent derivative, rimantadine, block the ability of the virus to build an ion channel between itself and the cell—in effect a tunnel into the cell—it attaches to. When these drugs work, the virus cannot get inside the cell, cannot invade it.

Two other drugs, zanamivir (Relenza), which is inhaled, and oseltamivir (Tamiflu), a pill, take a different approach. Both bind to the viral neuraminidase, so when new viruses try to escape the dead cell they get trapped on the cell surface as if on fly paper. They can’t infect other cells. (See the discussion of neuraminidase on page 104.)

All these drugs can reduce the severity and duration of an attack, but only if taken within forty-eight hours after symptoms appear. Taken prophylactically the drugs can also prevent an attack, although the preventative effect does not last long and at this writing the Food and Drug Administration has approved only oseltamivir for this purpose. The virus has also shown some ability to develop resistance to them. So, although antiviral drugs do show progress and promise, they are not an answer.

A vaccine offers far better protection, especially for the elderly. But to make the vaccine, investigators have to aim at a moving target. Every year they try to predict which virus strains will dominate and the direction of antigen drift. Then they design a vaccine for these antigens. When the investigators are right, when they hit their target, the vaccine protects very well for an entire flu season, preventing many attacks and reducing the severity of others. But the vaccine needs to be produced in huge quantities, which takes months, and in that time the virus can mutate in a direction different from the one anticipated. And even if the vaccine includes the right antigens, given the “mutant swarm” nature of the virus, some viral strains will escape it. Vaccines using killed viruses are injected, but in 2003 a new vaccine (FluMist) was introduced that uses live virus and is inhaled.

The real danger, though, is that it may not be possible to develop and distribute a vaccine in time to protect against a new virus. Influenza viruses for vaccines are grown in chicken eggs. When scientists tried to prepare a vaccine to the H5N1 Hong Kong virus of 1997, the virus initially proved too lethal: the virus killed the eggs in which it was being grown. Ultimately the problem was solved, but developing this vaccine took more than a year. If another lethal virus jumps to humans and it takes that long to develop a vaccine, by then the virus will have done its damage.

So even with all the medical advances since 1918, the CDC estimates that if a new pandemic virus strikes, then the U.S. death toll will most likely fall between 89,000 and 300,000. It also estimates a best case scenario of 75,000 deaths and a worst case scenario in which 422,000 Americans would die.

The CDC based that range, however, on different estimates of the effectiveness and availability of a vaccine and of the age groups most vulnerable to the virus. It did not factor in the most important determinant of deaths: the lethality of the virus itself. The CDC simply figured virulence by computing an average from the last three pandemics, those in 1918, 1957, and 1968. Yet two of those three real pandemics fall outside the range of the statistical model. The 1968 pandemic was less lethal than the best case scenario, and the 1918 pandemic was more lethal than the worst case scenario. After adjusting for population growth, the 1918 virus killed four times as many as the CDC’s worst case scenario, and medical advances cannot now significantly mitigate the killing impact of a virus that lethal.

If a new pandemic struck, people suffering from ARDS would quickly overwhelm intensive care units; those with ARDS who did not get true intensive care would have a mortality rate approaching that in 1918. A new virus would also feast on a population that did not exist in 1918—those with compromised immune systems, including people undergoing radiation or chemotherapy for cancer and transplant recipients, not to mention anyone with HIV.

No one has attempted to estimate the worldwide death toll of another influenza pandemic, but one could easily imagine a lethal virus—even one less virulent than that of 1918—killing tens of millions. No disease, including AIDS, poses the long-term threat of a violent explosion that influenza does.

Investigators and public health officials are not simply sitting back waiting for the next pandemic. In 1948 the World Health Organization established a formal monitoring system for influenza viruses. Currently 110 laboratories in eighty-two countries participate. Four collaborating WHO influenza centers—the CDC in Atlanta and laboratories in London, Tokyo, and Melbourne—provide detailed analysis.

The surveillance has two purposes: first, to track mutations of existing viruses to adjust each year’s vaccine, and second, to search for any sign of the emergence of a new strain—a strain that might cause another pandemic. To know where to look matters. Therefore it matters where the 1918 virus crossed into man.

This book hypothesizes that the 1918 virus emerged in rural Kansas. There are, however, other theories. Since influenza is an endemic disease, not simply an epidemic one, and since investigators at that time lacked modern technology’s ability to distinguish one influenza virus from another, the only real evidence is epidemiologic. Therefore it is impossible to state with absolute certainty which theory, if any of them, is correct.

Some medical historians and epidemiologists have hypothesized that the 1918 pandemic began in China. Most pandemics whose origin is known did begin in Asia or Russia. There is no scientific reason for this; it is only a question of probabilities. There large numbers of people live in close contact with pigs and birds, so more opportunities exist for a virus to cross over from animals to humans.

British scientist J. S. Oxford believes the 1918 pandemic originated in a British army post in France, where a disease British physicians called “purulent bronchitis” erupted in 1916. Autopsy reports of soldiers killed by this outbreak—today we would classify the deaths as ARDS—do bear a striking resemblance to those killed by influenza in 1918.

But these alternative hypotheses have problems. After the 1918–1919 pandemic, many scientists searched for the source of the disease. The American Medical Association sponsored what is generally considered the best of several comprehensive international studies of the pandemic, conducted by Dr. Edwin Jordan, editor of the Journal of Infectious Disease. He spent years reviewing evidence from all over the world and the AMA published his work in 1927.

Jordan first considered China as the possible source. Influenza did surface in early 1918 in China, but the outbreaks seemed minor and did not spread. Chinese scientists, trained by the Rockefeller Institute, themselves believed there was no evidence connecting any outbreak to the pandemic. Hong Kong had only twenty-two influenza hospital admissions in the first five months of 1918, and in Canton the first case of influenza did not surface until June 4. Recently some medical historians have suggested that one particular outbreak of deadly pulmonary disease in China in 1918 was actually influenza, but contemporary scientists diagnosed it as pneumonic plague and by 1918 the plague bacillus could be easily and conclusively identified in the laboratory. Also, one could not confuse pneumonic plague, with its then nearly 100 percent mortality rate, with even the most lethal influenza. So after tracing all known outbreaks in China, Jordan concluded that none of them “could be reasonably regarded as the true forerunner of the European epidemic.”

Jordan also considered Oxford’s hypothesis of the 1916 “purulent bronchitis” as a possible source. He rejected it for several reasons. At least some members of the British medical corps did not consider the infection contagious. No evidence suggested that it spread rapidly or widely, and a new influenza virus almost always does both. In fact, the outbreak did not seem to spread at all.

Also, we now know a sudden mutation in an existing influenza virus can account for a sudden virulent outbreak. In the summer of 2002, for example, an influenza epidemic with an extremely high death rate erupted in parts of Madagascar and in some towns it sickened an outright majority—in one instance 67 percent—of the population. But the virus causing this lethal epidemic was an old one that normally caused mild disease. (Technically, it was an H3N2 virus of a subtype isolated in 1999 in Panama.) It had simply mutated in a violent direction, then reverted to its normal mild status. The epidemic did not even spread to the whole island before fading away; it affected only thirteen of 111 health districts in Madagascar. Something similar may have happened in the British base.

Jordan also considered as possible sources other eruptions of influenza in early 1918 in France as well as some in India. He concluded that it was highly unlikely that the pandemic began in any of them. They too behaved like local eruptions of endemic influenza.

That left the United States. Jordan looked at a series of spring outbreaks there. The evidence seemed far stronger. One could see influenza jumping from army camp to camp, then into cities, and traveling with troops to Europe. His conclusion: the United States was the site of origin.

A later, equally comprehensive, multivolume British study of the pandemic agreed with Jordan. It too found no evidence for the influenza’s origin in the Orient; it too rejected the 1916 outbreak of “purulent bronchitis” among British troops; and it too concluded, “The disease was probably carried from the United States to Europe.”

Australian Macfarlane Burnet, quoted earlier on this point, also studied the pandemic closely. He too found the evidence “strongly suggestive” that the disease started in the United States and spread with “the arrival of American troops in France.”

More evidence against the 1916 origin comes from scientists Jeffrey Taubenberger and Peter Palese. Taubenberger is sequencing the genome of the 1918 virus after extracting samples of it from Alaska and the army’s pathology “museum.” Based on rates of mutation of the genome, he concludes that the virus emerged a few months prior to the pandemic. Peter Palese states, “The evidence that the virus was around before 1918 is very flimsy. It’s much more likely from all the data I’m aware of that the virus developed in 1918, or no more than six months earlier.”

If the disease did emerge a few months prior to the pandemic, and if the judgments of Jordan and other contemporaries were correct in thinking it started in the United States, then Haskell County, Kansas, seems the most likely origin. First, the outbreak in January and February 1918 was so unusual and so dangerous that even though influenza was not a reportable disease, Loring Miner reported it to the U.S. Public Health Service.

Second, if the virus did not originate in Haskell, there is no explanation for how it arrived there. Someone infected with the virus would have had to travel from an infected area somewhere else while leaving absolutely no trace of the disease in the country through which he or she passed. Given the length of time people with influenza can infect others, without air travel it would be physically impossible for the Haskell virus to have come from Europe. Nor are there other known outbreaks in the United States where someone could have become infected and carried it to Haskell. This strongly suggests that a new virus did emerge in Haskell.

And unlike the 1916 outbreak in France, which did not seem to spread, one can trace with perfect definiteness the route of the virus from Haskell to the outside world. The local paper listed by name people exposed to the disease who traveled to Camp Funston only a few days before the first reported case there; others the paper did not name may well also have gone there. Other than Haskell, Camp Funston was the first known outbreak of epidemic influenza in the United States. Several histories of the pandemic have begun their story there. And, one can easily track the disease from Funston outward—to other cantonments, to Europe, and to the U.S. civilian population.

The fact that the 1918 pandemic likely began in the United States makes a difference because it warns investigators where to look for a new virus. They must look everywhere.

The World Health Organization tries to do just that. Its surveillance system quickly identified a new H7N7 virus that appeared in the spring of 2003 in European poultry farms. This virus infected eighty-three people and killed one, a veterinarian. To prevent it from adapting to people, public health authorities in the Netherlands, Belgium, and Germany slaughtered nearly thirty million animals—most of them poultry but some swine. (The simultaneous SARS outbreak buried information on this occurrence in American news media.) WHO also quickly jumped on the 1997 Hong Kong outbreak. But the 1997 virus still survives in chickens and in 2003 killed one of two people it infected.

This same surveillance system also helped lead to the quick identification and containment of SARS, which was initially thought to be, and feared as, a new influenza virus. SARS offers both a historic public health success story and a warning. The success is obvious. Once WHO officials learned of it, it brought enormous resources to bear. Investigators around the world collaborated—entirely unlike the French and Germans in their search for the causes of cholera and plague a century earlier—and quickly identified the virus. At the same time world and national public health officials, except in China, moved rapidly and ruthlessly to quarantine and isolate anyone with or exposed to the disease. What once threatened to become a worldwide scourge was contained and may have been eliminated entirely. Even if it reemerges, close monitoring should keep it in check.

But before the first notification of WHO, the disease existed for months in China. For political and commercial reasons mainland Chinese authorities kept the disease secret and then initially lied about it. Once they did recognize the threat they moved aggressively and successfully to contain it, but had it been a new influenza virus, the months of silence would have made it impossible for public health authorities to have any chance either to contain the virus or develop a vaccine before a pandemic exploded across the world. Possibly the Chinese government—and other governments—learned a lesson they will not forget; possibly they will be both open and aggressive in the future whenever any indication of a new disease surfaces. One hopes so.

But even if Chinese authorities do change their approach to epidemic disease, even if SARS taught them and other governments around the world the same lesson, the fact that SARS killed people for several months before it attracted WHO’s attention demonstrates the vulnerability of the influenza surveillance system. If the 1918 virus crossed into humans in Haskell County, influenza can cross into man anywhere. Although eighty-two countries participate in WHO’s surveillance effort, more than one hundred do not. One Latin American physician at Tulane University involved in public health warns that at least as late as 1985—and probably later than that—the national medical school of Honduras taught its students that influenza was a bad cold. Those former students now practice medicine with that attitude.

It takes time to manufacture and distribute vaccines, and vaccines are the most effective defense. Early warning can make an enormous difference.

In the meantime the World Health Organization and individual countries continue to monitor influenza viruses, and continue to refine plans on how to respond to another epidemic or pandemic.

If one erupts, whether we want the knowledge or not, we will learn how good a job these planners have done.

Finally comes the question of how to apply lessons from 1918 to a new pandemic, and how these lessons relate to bioterrorism.

The use of biological weapons has a history going back at least to the Romans, who catapulted sick animals into enclaves of their enemies. The British and Americans likely used smallpox against Native Americans, and in 1777 British Major Robert Donkin recommended using smallpox against “American rebels” in a book on military strategy—but his recommendation was physically removed, the pages referring to it torn out of, nearly every copy of his book.

Yet in only three verified modern instances has disease been used as a weapon. During World War II Japan spread bubonic plague in China, and Japanese scientists also infected prisoners of war with other pathogens in experiments. In 1984 in Oregon a cult infected salad bars with salmonella (no deaths, 751 became ill). And in 2001 an unknown terrorist sent anthrax through the United States mail.

The threat of bioterror is nonetheless real. The World Health Organization believes forty-three different infectious organisms could be used as weapons. It considers the three most serious infectious threats anthrax, plague, and smallpox. It also considers botulinum toxin, a pure poison that can paralyze and kill, a bioterror threat.

All can be countered. Vaccines can prevent smallpox, anthrax, and plague—antibiotics also work against anthrax and plague—and antitoxin can neutralize botulinum. Also, neither anthrax nor botulinum toxin can spread from person to person. The ability to counter these weapons, however, does not mean their use would not cause mass terror even if their use was isolated. The reaction across the country to the anthrax attacks demonstrates that. And more than isolated use is possible.

The WHO has studied what it called a “worst case” scenario of an attack with pneumonic plague, the most lethal and contagious incarnation of bubonic plague, on a city of 5 million, and concluded it would make 150,000 ill and kill 36,000. Adjusted for population, these numbers represent considerably less than what influenza did to Philadelphia in 1918.

The 1918 pandemic, then, provides a case study of the public health and government response to a major bioterrorism attack, and it teaches two main lessons. The first involves threat assessment, planning, and allocating resources. It applies to both epidemics and large-scale bioterror attacks.

In 1999 the CDC issued a formal call for each of the fifty states to prepare plans for pandemic influenza and laid out suggested guidelines. The same plans would apply to an outbreak of nearly any epidemic disease or use of biological weapons. Since then, and more importantly since September 11, 2001, most states have begun to develop plans. But clearly epidemiologists, scientists, public health officials, and ethicists will have to join with the professionals who handle disasters to have sets of alternative recommendations in place—actual decisions will likely be up to elected officials—and ready to implement.

Some of the issues are obvious and simple, such as making sure health care workers are the first to get vaccinated. If they become sick, they can care for no one else. Emergency rooms need to recognize symptoms that can raise red flags, although the best clue will probably be a rush of cases. Investigators must be prepared to identify a pathogen, and epidemiologists must know the best ways to contain each likely pathogen. Legislation has to be in place to indemnify manufacturers and health care providers in the event of well-defined emergency circumstances. Production facilities have to be ready to manufacture vaccines and drugs; others should be stockpiled and distributed around the country, conceivably even in a form that individuals can administer to themselves to lessen the strain on professionals. (A study published in 2003 drives home how important logistics can be. It warned that under existing plans to distribute antibiotics, a small plane spraying anthrax spores over New York City could, under theoretically perfect conditions, kill 120,000 people, while improving distribution of antibiotics alone would slash the death toll from an identical attack to 1,000.)

Other questions also involve logistics and risk assessment. Influenza and most biological weapons attack the respiratory system. An outbreak would quickly fill beds in intensive care units, so resources need to be available to help huge numbers of people breathe. Public health officials also have to know the risks of side effects of vaccines, and based on the risk assessment they will have to know under what circumstances they would recommend vaccination and for whom.

Some elements of any plan, however, involve questions of power and ethics. Public health officials will need the authority to enforce decisions, including ruthless ones. If, for example, unvaccinated individuals threaten not only themselves but others by providing a reservoir in which pathogens can breed, officials might decide to order mandatory vaccination. Or, if there is any chance to limit the geographical spread of the disease, officials must have in place the legal power to take extreme quarantine measures. A centralized system should exist to allocate all resources including professionals as well. The utter waste of resources in 1918 in New York City—when doctors repeatedly crossed each other’s paths entering and leaving the same building because no centralized system was used to dispatch them—should not be tolerated.

Questions about who will have the authority to make and enforce such decisions, and under what circumstances, must be settled in advance. Neither an epidemic nor an attack will leave time for debate.

Some of the issues are almost purely ethical ones. If, say, containment of a pathogen is possible, but doing so requires isolating a building entirely, possibly saving many lives but at the cost of those in that building—what then? Medical ethics require physicians to do their best for each individual patient, but a military commander may ethically sacrifice a patrol, a platoon, a company to save a larger group. What ethic applies?

Another ethical question involves the free flow of scientific information. An investigator will probably at some point discover what made the 1918 virus so lethal. The influenza virus can be created to design in the laboratory, so publishing the information would give it to terrorists. A weaponized influenza virus could be the equivalent of a worldwide nuclear holocaust. But publishing would also give the information to researchers who could find a way to block whatever mechanism made the virus deadly, conceivably both countering any made-to-order killer virus and preventing any future natural outbreak on that scale. Should the information be published?

Scientific journals have already developed voluntary guidelines on what to publish, but these are not simple questions. Some go to the heart of medical or societal ethics, others to limits on freedom.

And some of these issues, such as stockpiling vaccines or training workers, simply cost enormous sums of money. (So does paying nurses enough to escape the current nursing shortage, which may soon approach that of 1918.)

What to do depends upon the assessment of the risk. Just as there was disagreement over the threat from the Soviet Union during the Cold War and how large the defense budget had to be to handle that threat, there will be disagreement over how real and how severe the threat from biological weapons is and how much must be spent—in money and in the erosion of values—to defend against it.

But there is another lesson from 1918 that is clear. It is also less tangible. It involves fear and the media and the way authorities deal with the public.

There was terror afoot in 1918, real terror. The randomness of death brought that terror home. So did its speed. And so did the fact that the healthiest and strongest seemed the most vulnerable.

The media and public officials helped create that terror—not by exaggerating the disease but by minimizing it, by trying to reassure.

Terror rises in the dark of the mind, in the unknown beast tracking us in the jungle. The fear of the dark is an almost physical manifestation of that. Horror movies build upon the fear of the unknown, the uncertain threat that we cannot see and do not know and can find no safe haven from. But in every horror movie, once the monster appears, terror condenses into the concrete and diminishes. Fear remains. But the edge of panic created by the unknown dissipates. The power of the imagination dissipates.

In 1918 the lies of officials and of the press never allowed the terror to condense into the concrete. The public could trust nothing and so they knew nothing. So a terror seeped into the society that prevented one woman from caring for her sister, that prevented volunteers from bringing food to families too ill to feed themselves and who starved to death because of it, that prevented trained nurses from responding to the most urgent calls for their services. The fear, not the disease, threatened to break the society apart. As Victor Vaughan—a careful man, a measured man, a man who did not overstate to make a point—warned, “Civilization could have disappeared within a few more weeks.”

So the final lesson, a simple one yet one most difficult to execute, is that those who occupy positions of authority must lessen the panic that can alienate all within a society. Society cannot function if it is every man for himself. By definition, civilization cannot survive that.

Those in authority must retain the public’s trust. The way to do that is to distort nothing, to put the best face on nothing, to try to manipulate no one. Lincoln said that first, and best.

Leadership must make whatever horror exists concrete. Only then will people be able to break it apart.