These days, the phrase "blue flu" usually is applied to walkouts by police officers, but it just as readily could be used to describe the frightful influenza pandemic of 1918–1919. The adjective is apt because many victims of that virus turned blue before they died, slowly suffocating as their lungs filled with blood or fluid. During this pandemic, at least 40 million and possibly 100 million people died worldwide (the original estimate of 20 million now is thought to be far too low), including 675,000 in the United States. Possibly 1 billion people — half the world's population — became sick from the disease dubbed Spanish flu (owing to the propensity of neutral Spain to talk about it, unlike combatant nations in the Great War).
Spanish flu struck in three waves. The first time around, in the spring of 1918, many people were sickened but relatively few died. The second and third time, in the fall of 1918 and early 1919, it reappeared in a highly virulent form that proved far more deadly. The death rate was greatest among young adults — people between the ages 17–30 who were otherwise healthy. Notably, people in this age range succumbed even during the initial mild wave. It's speculated that older people had acquired some immunity from two suspected pandemics in the 19th century, while younger people's immune systems hadn't matured to the point to trigger a fatal immune response to this particular virus.
In other words, no one really knows why the pandemic of 1918–1919 was so deadly. Flu-related mortality often stems from secondary bacterial infections, but in the case of the Spanish flu, a substantial number were killed outright, with some succumbing in a matter of hours.
"We don't know why people died from the Spanish flu, but we suspect that the cause of death wasn't bacterial," says John Shanley, MD, director and professor of infectious diseases at the University of Connecticut School of Medicine.
Rapid and violent
From a pathologist's perspective, people killed directly by Spanish flu died rather spectacularly: Damage to their lungs was so violent that they appeared to be victims of poison gas. Either way, the mortality rate was about 2.5 percent, compared with less than 0.1 percent during the routine influenza epidemics to which we have become accustomed. In other words, the death rate was 25 times as great.
When the next pandemic struck, the Asian flu of 1957–1958, more people probably — were sickened than in 1918–1919, but fewer died (about 1 million, including 60,000 Americans), partly owing to the availability of antibiotics by then. The last influenza pandemic, the Hong Kong flu of 1968–1969, also killed about 1 million people, including 40,000 in the United States. That's close to the average number of excess deaths (the difference between the number of deaths observed in a group and the number of deaths that would have occurred if the group had the same death rate as a comparison population) attributed to influenza in recent years from U.S. epidemics, though.
Influenza and pneumonia together accounted for 66,000 deaths in the United States in 2002 (the most recent year for which data are available), ranking as the seventh leading cause of death but claiming less than one tenth the number of people killed by heart disease. Influenza and pneumonia are lumped together because it's hard to pinpoint deaths directly attributable to influenza, owing to the lack of virological confirmation of the disease or the listing of influenza on hospital discharge forms or death certificates.
As might be expected, deaths from influenza/pneumonia were distributed unevenly geographically, reflecting countless variables. For example, influenza and pneumonia death rates of 14.8 and 15.6 per 100,000, adjusted for age, were reported for Florida and Washington, respectively, while in Tennessee and Kentucky the respective rates were 30.2 and 30.8. The national rate was 22.6.
Looking at flu seasons, which span calendar years, NIH epidemiologists recently estimated that during the decade ending in 1998–1999, the mean number of influenza-associated deaths in the United States was 51,203. However, the range of deaths was broad, from a low of 25,570 in 1990–1991 to a high of 71,416 in 1997–1998. In general, though, the trend since 1976–1977 has been for influenza-associated deaths to increase, which is attributed in part to the aging of the population. Mortality rates rose sharply with increasing age.
Likewise, NIH epidemiologists found that the aging of the U.S. population was partly responsible for a substantial increase in the number of influenza-associated hospitalizations during the last two decades. Between the 1979–1980 and the 2000–2001 flu seasons, pneumonia and influenza were the primary reason for an average of 94,735 hospitalizations. Again the range was wide, from 18,908 in 1986–1987 to 193,561 in 1997–1998. However, the authors pointed out that influenza-related morbidity is considerably more extensive, because influenza also leads to hospitalization for an array of cardiorespiratory diagnoses — a number they place at more than 200,000 annually.
$12 billion annual cost
Altogether, according to the National Coalition for Adult Immunization, an average flu season costs the United States at least $12 billion, including $4.6 billion in direct medical costs and twice that amount for indirect costs, mostly from lost work days and school days.
Acknowledging the inherent difficulty in constructing a mathematical model of pandemic influenza — we don't know anything about the characteristics of the virus, or how it would be spread, or who would be most likely to be affected — Martin Meltzer and colleagues at the CDC nevertheless took a stab at it a few years ago. Their intention was to show the economic effect of various vaccination strategies. Using illness and death rates reported in previous influenza pandemics and epidemics, they estimated that during the next pandemic, U.S. deaths might range from 89,000 to 207,000; hospitalizations, 314,000 to 734,000; outpatient visits, 18 million to 34 million. The lower end of these ranges represent a pandemic in which 15 percent of the population becomes ill; the upper end, 35 percent. Direct and indirect costs were placed between $71 billion and $166 billion (1995 dollars). Deaths accounted for 83 percent of economic losses; outpatients, 8 percent; inpatients, 3 percent; and ill persons not seeking medical care, 6 percent (from lost work).
The model makers considered four options for determining who would be vaccinated — assuming vaccine is readily available. Given our unhappy experience with vaccine supplies during the current flu season, that is not a safe assumption. The first option applied the usual standards for identifying the high-risk populations during routine, interpandemic years, some 77 million people. The next option added to that group 5 million health care workers and 15 million providers of other services, for a total of 99 million vaccinees. A third option sought 40 percent effective coverage of the U.S. population — 106 million — and a fourth option considered coverage for 60 percent of the population, or 159 million people.
In the event of a pandemic involving a new influenza subtype, effective immunization might require two doses of vaccine per person. In other words, to achieve 60 percent coverage, 320 million doses would have to be produced, delivered, and administered in a couple of months. Again, the experience during 2004, when sufficient vaccine couldn't be delivered to meet the demands of a normal flu season, is sobering.
"The last few years have been a dress rehearsal for a pandemic, and it hasn't gone well," says Steve Black, MD, director of the Kaiser Permanente Vaccine Study Center in Oakland, Calif., and a member of the National Vaccine Advisory Committee, which counsels the secretary of health and human services. "The current system doesn't work for routine influenza," he says. "In a pandemic, there could be a panic." Indeed, he wonders whether there might be a role for the National Guard in distributing and safeguarding scarce vaccine.
Results of the Meltzer model, which was published in 1999, were among the many issues discussed last June by a panel of influenza experts assembled by the Institute of Medicine's Forum on Microbial Threats to assess the threat posed by the next pandemic. Many infectious disease experts worry that when the next pandemic emerges — and they say it is a question of when, not whether — it could extract a mortality toll more akin to that of Spanish flu than to Asian or Hong Kong flu. That's because the next pandemic might arise in southeast Asia, lurking amidst domestic and wild birds.
So far, the chief victims of the avian influenza have been the fowl that have died outright or been slaughtered by the millions in an attempt to contain the virus. What has caught the attention of epidemiologists is that, among the 44 cases of infection confirmed in humans, the mortality rate has been 73 percent. Except in one case, the route of transmission appears to have been from bird to human, rather than human to human. However, bird-to-mammal infection has been observed (mostly among felines that were fed raw chicken), and that is worrisome indeed.
"Everything that has been touched by this avian influenza has been affected," says University of Connecticut's Shanley. "We normally don't think of cats as a target of influenza. This is a different kind of virus."
Once the virus develops the capability to spread efficiently from human to human, the next pandemic could be upon us. Compared with any cell, an influenza virus is a simple piece of machinery. It's designed to do just a few things: get into a respiratory cell, get the goods, and then get out and see the world. Toward that end, eight wisps of RNA are enclosed within a lipid membrane, which is studded with two kinds of protein — hemagglutinin and neuraminidase. Different subtypes of hemagglutinin and neuraminidase have been identified. In viral strains, they're identified by the letter H or N, followed by a numeral. The 1918 virus is known as H1N1. Strains of that virus served as the foundation for the pandemic viruses that emerged in 1957 and 1968 (See "Antigenic Shifts Leading to the Pandemics of 1957 and 1968" below).
|Antigenic shifts leading to the pandemics of 1957 and 1968|
|Gene||Molecule encoded||Function of molecule||1957 (H2N2)
"Hong Kong flu"
|HA||Hemagglutinin||Facilitates penetration of target cells||H1 subtype is replaced by H2 gene related to those in avian virus strains||H2 subtype is replaced by avian-derived H3 subtype|
|NA||Neuraminidase||Facilitates escape from target cells; inhibited by zanamivir (Relenza) and oseltamivir (Tamiflu)||N1 subtype is replaced by N2 gene related to those in avian virus strains||N2 subtype is retained|
|NP||Nucleoprotein||Interacts with RNA polymerases||NP gene from prior H1N1 strains is retained||NP gene from prior H1N1 strains is retained|
|M||Matrix proteins (2)||M1 provides structure. M2 is ion channel that acidifies the virus when it is within an endosome (inhibited by amantadine and rimantadine)||M gene from H1N1 strains is retained||M gene from H1N1 strains is retained|
|NS||Nonstructural protein||Uncertain||NS gene from H1N1 strains is retained||NS gene from H1N1 strains is retained|
|PA||Polymerase acidic protein||Promotes transcription and replication of viral RNA||PA gene from H1N1 strains is retained||PA gene from H1N1 strains is retained|
|PB1||Polymerase basic protein 1||Promotes transcription and replication of viral RNA||Replaced by avian-derived PB1 gene||Replaced by avian-derived PB1 gene|
|PB2||Polymerase basic protein 2||Promotes transcription and replication of viral RNA||PB2 gene from H1N1 strains is retained||PB2 gene from H1N1 strains is retained|
|SOURCE: Taubenberger JK. Chasing the elusive 1918 virus: preparing for the future by examining the past. In: Knobler SL, Mack A, Mahmoud A, Lemon SM, eds. The Threat of Pandemic Influenza: Are We Ready? Washington, DC: National Academies Press; 2004.|
The hemagglutinin helps the virus gain entry into a target cell. First it binds with sialic acid on the cell surface. As soon as that happens, a cellular defensive mechanism is triggered and the virus is engulfed by an endosome, or endocytic vesicle — a little sphere formed from part of the cell's lipid membrane. But the virus has a way to escape. The interior of the endosome is acidic, the better to destroy an invader. In the case of the influenza virus, it does the opposite. The virus moves protons, through ion channels, to its interior (drugs like rimantadine and amantadine block these channels), where the resulting increase in acidity causes the hemagglutinin to change its shape and perform its second task: fusing the membrane of the virus to the vesicle. The viral RNA then spills into the interior of the cell.
Next, the viral genes hijack the cell's internal apparatus to synthesize the components required to make numerous copies of the virus; material needed for additional lipid bilayers is simply appropriated from the host. The new viruses use the neuraminidase to cut their ties with sialic acid and break through the cell membrane, infecting more cells. If activity of the neuraminidase can be blocked (for example, by zanamivir or oseltamivir), the viruses remain trapped inside the host cell.
Drift or shift?
Hemagglutinin and neuraminidase are the chief targets of the immune system. If the gene that directs the synthesis of hemagglutinin or neuraminidase is modified slightly, in a process known as antigenic drift, an old virus may be transformed into a new strain that has acquired the ability to evade antibodies. A point mutation affecting a single nucleic acid may be sufficient to lead to the synthesis of a different amino acid and hence a protein with a different shape, which may be recognized partially or not at all by existing antibodies. Drift is why new vaccines must be developed each year to cope with routine influenza epidemics.
Antigenic shift is believed to be behind the emergence of pandemic viruses. A cell can be invaded by more than one kind of virus at a time, and when that happens, the pieces of RNA that are spilled into the cell's interior can be rearranged. Take two snippets of avian influenza virus and six snippets of human influenza virus, mix well inside a pig (or a person), and you create a novel virus. That means no human immune system has ever encountered it before. "Hello, pandemic."
So let's say that the mixing bowl of southeast Asia — being mindful that rural Kansas may have served the same purpose in 1918 — gives rise to an H5N1 virus adept at spreading from human to human. What then?
Public health authorities can be expected to detect an outbreak quickly. After that, things get trickier. Here are some of the outstanding problems:
The incubation period for influenza is so short (one to four days) that it's virtually impossible to identify and quarantine the contacts of people with confirmed infection before the contacts spread the virus. Besides that, clinical symptoms of influenza aren't distinctive and many people remain asymptomatic.
Within the United States, only one facility (Aventis Pasteur in Swiftwater, Pa.) is capable of producing influenza vaccine. In 2004, it produced about 50 million doses of trivalent vaccine (a vaccine that contains three influenza subtypes). In a pandemic, a monovalent vaccine would be used, but each person might need two doses. In that case, existing domestic capacity would serve 75 million Americans. To vaccinate 50 to 60 percent of the population, manufacturing capacity would have to be tripled.
Eight other countries produce nearly all of the remainder of influenza vaccine. In the event of a pandemic, they're probably not going to be exporting any of it — not that they have much to export, anyway.
If manufacturing facilities concentrate their efforts on production of influenza vaccine during a pandemic, shortages of other vaccines could arise. Those shortages might include vaccine against H3N2 influenza strains, which have proved more deadly than H1 strains during recent epidemics.
Only one facility in Switzerland manufactures oseltamivir, a neuraminidase inhibitor. It could be useful for slowing a pandemic and reducing morbidity and mortality in its early stages, buying time for vaccine producers to rev up. Once a new pandemic virus has been isolated, it takes several months to produce the first vaccine, using the traditional egg-based method.
Did we mention that the H5N1 avian flu kills eggs? Reverse genetics may provide a work-around (research is under way), but this production method isn't here yet, and there are unresolved patent issues.
Stockpiles of antiviral agents are slim to none, worldwide. For one thing, they're expensive. Even so, stockpiles might be nice to have. That's especially true because, according to the U.S. Department of Health and Human Services, oseltamivir is the only antiviral drug shown in vitro to be active against Asian bird flu. With limited production capacity, it would take years to build up supplies.
But, even oseltamivir may turn out to be a disappointing first-line defense — a recent study in the Lancet showed that, in Japanese children with influenza, resistance to this drug develops more frequently than had been thought, and treated patients continue to shed infectious viruses for up to five days after treatment.
Owing to legal issues, including product liability, intellectual property rights, and tax credits, industry lacks incentive to pursue new drugs such as a vaccine based on the conserved portions of the influenza genome, which wouldn't be affected by the vagaries of shift and drift.
Depending on managed care
A lot has changed since then. Whereas the vast majority of swine flu vaccinations were provided by the public sector, today the task of administering vaccinations during a pandemic presumably would fall to the private sector — notably, MCOs. During the June workshop assembled by the Institute of Medicine's Forum on Microbial Threats, a representative of Aetna pointed out that Aetna and many other health plans have successfully conducted immunization campaigns, sometimes using retail establishments to reach large numbers of members quickly.
Health plans also possess data that might permit identification of high-risk patients, who are most likely to benefit from vaccination. But as Kaiser Permanente's Black notes, "In a pandemic, we won't be distinguishing high-risk patients from low-risk patients, but rather high risk from higher risk."
But health plans can't serve the millions of Americans who lack health insurance. Neither can health plans establish national priorities for determining which people receive vaccine (or antiviral drugs) if supplies are scarce, as they assuredly will be. If you want to reduce overall mortality, the elderly would be high on the list. If you want to keep society running, you give priority to health care workers and providers of essential services. If you want to reduce economic losses, you target children and working adults. But at the current levels of (un)preparedness, hard decisions — ultimately political — will have to be made.
Through October, HHS was soliciting comments on its Draft Pandemic Influenza Preparedness and Response Plan. The Infectious Diseases Society of America, representing 8,000 physicians and scientists working in infectious diseases, obliged with a lengthy response. While generally supportive of the endeavor, it put the impending influenza pandemic in perspective thusly:
The United States has recently committed substantial resources toward research and control of bioterrorism agents. However, in terms of its public health implications, pandemic influenza represents a far greater threat than bioterrorism agents to citizens in the U.S. and other countries. The administration's request of $100 million for pandemic influenza activities for fiscal years 2004–05 seriously underestimates the amount of funds realistically needed to effectively respond to the next pandemic.
Shanley agrees with this grim assessment. "A pandemic will eclipse bioterrorism enormously. It will overburden our health care resources," he says.
What health plans can do, says Kaiser Permanente's Black, is lobby federal and state officials for increased vaccine manufacturing capacity and a decent distribution system to make sure vaccine is allocated fairly and efficiently, without hording or profiteering.
"The difficulty in producing and distributing vaccine is not widely appreciated," adds Shanley.
So if you'd like to take some of the sting out of the impending influenza pandemic, perhaps it's time to talk with your friendly, local members of Congress. Tell them to think in terms of tsunami.