Interview with Osterholm
Bird Flu: "This thing just continues to march"
The CP interview: Dr. Michael Osterholm of the University of Minnesota's Center for Infectious Disease Research and Policy talks about the flu bug that could bring the world to its knees
by Steve Perry
Scare headlines about the possibility of a deadly flu pandemic have been with us for a few years now, ever since the H5N1 bird-flu virus that first appeared in Hong Kong in 1997 resurfaced in the region in 2003. But in the past month the drumbeat of such stories has grown faster and louder: Avian Flu Arrives in Poland. Turkey. Azerbajian. Germany. Denmark. And, just last Friday, Israel. The good news, according to Dr. Michael Osterholm, the director of the University of Minnesota's Center for Infectious Disease Research and Policy, is that the arrival of infected birds in North America—sometime this year, in the estimation of most experts—is not likely to result in large numbers of human infections with the virus, because most domestic poultry in this part of the world is raised in factory-farm isolation units that prevent contact with wild birds.
The bad news is that that's pretty much the only good news.
What matters in judging the prospects for a human pandemic version of H5N1 (the name is shorthand for the chemical structure of two of the virus's key components, hemagglutinin and neuraminidase) is not so much the global reach of the bird version, but the question of if or when the virus mutates to a form that's easily passed from human to human. If that happens anywhere in the world, says Osterholm, the virus would likely start hitching rides with travelers and seed itself around the globe in a matter of days or weeks.
Of the hundred-plus human cases of H5N1 flu recorded so far, the vast majority have involved bird-to-human transmission, mostly among open-air poultry handlers in Asia. In addition, there are confirmed clusters in which it has passed from person to person, though none of those has yet resulted in a breakout of the virus. One thing is clear, however: In its present form, H5N1 has killed over half of the people it's infected. The great flu pandemic of 1918-19, by contrast, killed about 5 percent of its victims.
Will it cross over? If it does, can it possibly remain as deadly? Though Osterholm notes that viruses usually do lose strength as they spread—it's not really in their own evolutionary interest to kill the majority of their hosts—he believes the only responsible answer on both counts is we don't know. But it's not just the characteristics of the virus that worry him.
One of the things that sets the former Minnesota state epidemiologist apart from other public health officials is his attention to the fate of the medical and social infrastructure in any serious contagious outbreak. With respect to bird flu, his outlook recapitulates in many ways what he had to say in his 2001 book about bioterrorism preparedness, Living Terrors—much of the human toll in death, hysteria, and anarchy would be exacted not by infection but by the wide-scale breakdown of global supply chains and just-in-time delivery systems for vital goods and services. "I think [Health and Human Services] Secretary [Mike] Leavitt has been brutally honest in telling American communities, you're going to be on your own," says Osterholm. "And I think he's right."
City Pages: Let me start with the question of the likelihood of a global flu pandemic. Yesterday alone, I saw two wire service headlines with radically different-seeming implications, one indicating that the H5N1 avian flu is likely to go global within six months, and another speculating that the threat of human transmission may be passing as we speak. Is it possible to say, based on the epidemiological evidence, how likely a human flu pandemic is in the next six months, the next year, the next two years?
Michael Osterholm: First of all, let me take a step back and say that's a typical question we hear from the public. But it's almost like asking, how many apples can you harvest from the ocean? Because flu pandemics are like earthquakes, hurricanes, and tsunamis—they occur. There's going to be another one. There have been 10 in the last 300 years. What you're really asking, I think, is will it be H5N1? It's important to understand that when you're preparing for pandemic influenza, you're preparing for something that will happen. To say anything to the contrary would be like saying, now that Katrina's happened, we'll never have another hurricane like it.
Will H5N1 be the pandemic strain, and will it occur in the next six to twelve months? The answer is, we don't know. What is troubling about this virus is that this thing has continued to mutate from its earliest days, in Hong Kong in 1997. And what is very, very troubling to us is that it's mutating in very similar fashion to the way the 1918 virus did. We went back with the 1918 virus and found all eight genes of that virus in tissue samples—five from soldiers' pathology slides that had been stored away, three from the recovered corpse in Alaska. They didn't have any live virus, but they've been able to make the virus from those eight genes. And by studying that, they could determine how it actually mutated and jumped directly to humans from birds. It didn't go through other species as the 1957 and 1968 viruses did, where a bird and a human virus got together, most likely in a pig lot, because pigs happen to be the universal recipients for both [birds and humans].
They combined to make a third, dumbed-down virus that caused mild pandemics.
The 1918 virus jumped right from birds to people. There was no combining with other viruses. One of the problems we've had is, if you look at the 1918 virus and this one, they're in essence kissing cousins. Genetically, these things look very similar. Frank Obenauer and colleagues just published a paper the last week of January in Science, and they actually have gone back and looked at the full genetic codes for 169 avian virus genomes, dating way back. They looked at 2,169 distinct avian virus genes. There were two viruses that showed a protein tag at the end of one of the nonstructural genes that actually looks to help cause the cytokine storm that makes this a unique illness.* And guess which two viruses they were: 1918 H1N1, and the current H5N1.
Then, when you look at the Turkey virus—that thing mutated. This is the case of the young girl in Turkey who died from her infection, and so did her uncle. We definitely have clusters where it's not just bird contact [spreading the virus]. The uncle's only exposure to this virus was riding in the ambulance with her from hospital one to hospital two. He became ill three days later and died. Her virus has now been fully sequenced, and there were three mutations that occurred in that virus, between the bird version and hers. One was the substitution of a glutamic acid with lysine at the 223-hemagglutinin position. That is what changes it from a bird-receptor virus to a human-receptor virus. The second thing was two other substitutions that served to make it look more and more like a human virus.
So this thing just continues to march. Changes are occurring in it all the time. [Human-to-human transmission] could happen tonight. Or it may never happen. But I don't know what will keep it from happening, because when you have this kind of worldwide bird population as we do now—China's a good example. In 1969, during the last pandemic, China only had about 12 million chickens. Now it's got over 15 billion.
CP: Do you think the rise of poultry farms of vast scale has contributed to the viral soup that influenza viruses grow in?
Osterholm: Not really, and I'll tell you why. When you look at the rise of the really big bird operations, they are actually raised in these bio-security barns, which people have all kinds of problems with for entirely different reasons—humaneness and that kind of thing. They actually are very safe, generally speaking, because they keep the wild birds and the domestic birds separate. It's in Asia where you have all these small 20-, 40-, 50-chicken operations where the birds are living in open space with you—that's where the vast majority of the chicken population is at in the developing world. A good example is Turkey, where we're seeing the first cases outside of Asia now. This is taking the virus out of a tropical area and putting it in a temperate area that gets cold. Every night, those people bring their chickens into the house. It's just a very different mindset.
And for as much as this is going to come here someday, [bird-to-human transmission] is not going to be a big risk factor to humans on this continent, because other than free-ranging organic birds that are out there, domestic birds aren't going to be at big risk.
CP: Can you explain in lay terms what makes a strain like H5N1 novel, and so potentially deadly?
Osterholm: Well, there are three things that make a strain of influenza virus potentially capable of causing a pandemic. First of all, you have to have a situation where you've got a novel or a new strain, meaning you don't have any antibody protection against it. Then you have to have one that is able to go from human to human. That's what we don't have yet. The third thing is, it has some virulence characteristics that make it cause severe illness.
This virus is quite different from what we see with the standard annual flu, and what we saw in 1957 and 1968, because of the cytokine storm it causes. In 1918, the vast majority of the people who died were healthy young people, 20 to 40 years of age. And that was in large part because they had the strongest immune systems.
CP: You're saying that the symptoms that cause fatalities, aside from secondary bacterial infections, are actually a function of the immune system working overtime.
Osterholm: That's it. And that's what we're trying to understand at this point, in terms of how to best prevent this [immune reaction]. And right now it doesn't look like there's much you can do. I mentioned this "kissing cousins" phenomenon. If you put 1918 H1N1 into animal models at very, very low doses, it basically kills all of them in 24 hours. The lab science people had never seen that. At 16 to 24 hours, that virus was different from anything they'd ever seen in killing these animals. The only virus that was similar was H5N1, and it was fatal at much lower doses. H5N1 is the most powerful influenza virus we've seen in modern human history.
What makes them so similar is that they both cause this cytokine storm phenomenon.
CP: Which essentially results in a person's drowning in his or her own blood as it fills the lungs, right?
Osterholm: It's even worse than that. You get that kind of leakage, yes, but it also goes in and
begins to shut down all your vital organs. It's a domino effect. Your kidneys go down, then your liver goes down, you have all this destruction through necrosis of your lungs and your internal organs. Everything goes.
CP: In the limited human sample we've seen so far, this influenza has exacted a much higher mortality rate than the 1918 flu. Are there mechanisms that tend to dilute the virulence of a strain as it spreads?
Osterholm: That's a really critical question. We can only anticipate that this will attenuate. Meaning that once it starts spreading in humans, it will lose some of its punch in order to better adapt to humans. That's traditional with virtually all agents you see like this. The thing that is very difficult to talk about is, we don't know how much. If this were to go human-to-human—we talk about a worst-case scenario in terms of what happened in 1918, when roughly 2.5 percent of the world's population died. Of those who contracted it, roughly 5 to 6 percent of populations died, varying by age.
The mortality rate so far for this virus is around 55 percent, so this virus would have to attenuate a lot to get down to that level. And we do have good data. There are not a lot of mild, asymptomatic infections out there [with H5N1]. We're now aware of six studies involving over 5,000 close contacts of H5N1-infected people, in Indonesia, Vietnam, and Hong Kong, in which less than one person per thousand contacts had evidence of an H5N1 infection that was missed—that is, a mild infection.
This [virus] is not causing a lot of asymptomatic infections right now. Some people are saying there's a lot of mild [H5N1-related] illness all over out there, but it's just not true. That means we're not artificially inflating the mortality rate by missing a lot of infections. I'm actually pretty confident that the real mortality is almost that high.
So for that number to drop all the way down to a couple percent is a pretty big drop. Which says to me that when people talk about 1918 as a worst-case scenario, well, maybe that isn't the worst-case scenario. That's hard for people to hear, because then they think you're really trying to scare the hell out of people. But you know what? It's just the data.
If this virus were to ultimately go human-to-human, none of us know what the human mortality would be.
CP: Does the fact that it seems to be gaining more currency in other mammal species augur one way or the other for its becoming transmissible from person to person?
Osterholm: None of us know. In 1918, for instance, we don't know whether it infected cats and dogs. We've been trying to find that out. Nobody's got that data. There just weren't good reports. It surely can't be good that it's adapting to more species. It says that the lung receptors of chickens aren't the only ones that will take this virus. And we know humans surely will take it, on the off chance that they're exposed [to infected birds].
The bird-to-cat thing is not new. Some people have made a lot out of the German situation. That's not new. The Bengal tigers at the Bangkok Zoo died two years ago. They got fed H5N1-infected chickens, and 50 of the Bengal tigers died. They also transmitted to each other—there were cats there that did not eat the chickens. Even Albert Osterhaus's work of the past couple of weeks, which has been really important to confirm it, was not a surprise.
CP: I wanted to ask you about a scenario you described in your New England Journal of Medicine article from last year, "Preparing for the Next Pandemic." If a flu strain transmitted from human to human did break through in some part of the world, how would you expect events to unfold over the first weeks?
Osterholm: Well, look at what's already happening with the bird situation. You've got countries like South Korea saying, don't go to Egypt. You've got a lot of bird embargoes already taking place. If you saw this morning's Wall Street Journal, the travel industry in Europe is tanking. And this is a non-threat to a vast majority of humans. What we're concerned about is that if this takes off in a given area, it's going to move around the world quickly, just like SARS. Last year 750 million people crossed a national border somewhere in the world, either by plane, automobile, or on foot.
These things move fast. With SARS, we had one physician from China who came to Hong Kong, stayed in the Metropole Hotel on one floor where there were nine other individuals he infected just through breathing the air. They then took it to four different continents in the next two days. That gives you kind of a model, though influenza is much, much more infectious.
I was very critical of those models that came out last fall that suggested you could put a blanket over this with Tamiflu. My whole criticism was practical—that you would never find this quick enough and confirm it. Despite the fact we now know Turkey was going on for weeks before we understood what was happening over there, it was only last week that the uncle's isolates were confirmed, almost eight weeks after the fact.
CP: There's been a lot of covering up by some governments, hasn't there? It's been repeatedly alleged of the Chinese.
Osterholm: Exactly. So my whole point is that before anyone will have even figured out that this is going on with any certainty, it'll be gone—the cow will be out the barn door. With influenza, that's just...it's something you can't pull back in.
CP: Periodically we see stories about the race to derive vaccines, but that's not conceivable, is it, until we know which viral subtype will break through?
Osterholm: Well, even that's a misconception, I think. The concern we have is that people all want to know if some new vaccine is "the answer." The problem is, we do need a new vaccine. We're dealing with a 1950s-technology vaccine with only one update. Right now, given the amount of virus needed to make vaccines for H5N1—it needs a lot more antigen [than typical flu strains]—our total worldwide capacity right now, in one year's time, is only enough vaccine to protect 100-200 million people worldwide. That's in one year after a pandemic starts. And that's it. You can't make any more, given the limited capacity we have.
So you know what?
It doesn't matter if we invent vaccines if we can't manufacture them. This is a point I've tried to make over and over again. We also are approaching this from a very American-centric point of view, which in the end will be the death of us. What's going to happen is, even if we could produce vaccines for our country in a timely manner, this global just-in-time economy we live in today is going to see the rest of the world shut down. Eighty percent of all the drugs we use in this country—all the childhood vaccines, everything—come from offshore. Your cardio drugs, your cancer drugs, your diabetes drugs, 80 percent of the raw ingredients come from offshore. I could go through a whole laundry list of other critical and essential products and services that come from offshore. If the rest of the world experiences a pandemic, we're still screwed. That's what people don't understand. Somehow they have this attitude that we can wall ourselves off in the Eighth District of Minneapolis and be okay.
The bottom line is, it will be years, even at the accelerated rate we're going now, before we even get the right candidate-vaccines. Then they still have to be approved. No company's going to embark on building manufacturing capacity without the certainty of a market. And second of all, they have to know exactly what vaccine will be used and how they need to develop their plant. All this means that an influenza plant, from start to finish, would probably take three to five years to build.
CP: Why don't we have the capacity to produce more in the way of basic vaccines here in the U.S.?
Osterholm: Because it's all about the market. Today, many of the anti-infective antibiotics and vaccines are not considered blockbuster products for any pharmaceutical company. We have a number of our childhood vaccines that are down to a single manufacturer now. If you go to the Society for Health Care Pharmacists' website, you'll see a list of 42 drugs today that are in short supply because of the just-in-time supply chain. Now, you put on top of that products that are marginal in their [profit] return, and it becomes a loss leader for these companies [to make vaccines]. They want out of it.
This is not true just for vaccines, but also for antibiotics. There, we tell people to hardly use them, and only for specific purposes and limited periods. It's not like the lifestyle drugs, where you're taking your statins for the rest of your life. So they just don't see the profitability [in antibiotics]. We're in real trouble. And it's not just with respect to the flu.
CP: The one easy-to-use drug that's been shown to arrest H5N1 in humans is Tamiflu—
Osterholm: Well, that's not really true. There's no evidence that it makes a difference in H5N1 infection. I have a slide here showing that the case mortality rate in Vietnam was identical for those who got Tamiflu and those who didn't. I actually believe Tamiflu could work, but the problem is, we're applying it in the H3N2 [average seasonal flu] model. Seasonal flu grows up much slower in a human, so that if you get the drug into somebody two days after their onset, you can still have a pretty measurable impact on the severity of their illness and the likelihood they'll have complications. With the H5N1 virus, the virus storm that precedes the cytokine storm is so remarkable in those first 24 hours that if you don't have the drug onboard in those first 24 hours, it may only have limited impact.
The second thing is, we're using the same dosage levels we use for H3N2 [seasonal flu]. This is a very different virus. We may have to double the dosage for twice the length of time. We have anecdotal data on people who got the drug early and appeared to do better, but then, after the typical five-day course was stopped, they died on day ten.
The problem we have is not whether Tamiflu is available and will work, but can you get it soon enough and in high enough amounts? How do you get the drug to somebody in the first 12 hours of their illness? The second thing is, when people talk about having 20 million or 40 million protective doses, what does that mean? If we're going to have to use twice the dose for twice as long, that means maybe you've got one-fourth as much of the drug you think you have.
That's the story on H5N1 and Tamiflu. I think it can work, but it won't work the way we're approaching it.
In the next five years, with [Tamiflu manufacturer] Roche outsourcing all of the production they can, our best guesstimate is that—using the old, low dosage standards—we could probably produce enough Tamiflu to treat about 7 percent of the world's population in the next five years, because of the precursor chemicals needed and the complicated nature of making this stuff. That's the sum total of production capacity. It's no panacea.
CP: How would you rate the public health infrastructure in the U.S. for its ability to deal with massive numbers of very ill people? You wrote in the NEJM last year that our system lacks "surge capacity" to deal with a lot of sick people at once—why is that?
Osterholm: It's much deeper and bigger than the public health system. It's really the total care system. For example, I talked [in that article] about the 105,000 mechanical ventilators? On any given day, 70,000-80,000 of them are in use, and in a normal flu season we butt up against the 100,000 mark. We have no excess capacity there whatsoever. Just right here in our own Twin Cities, we've reduced intensive care beds by 20 percent in the last four years, as a matter of cost containment.
We'll run out of masks and respirators overnight, because it's a global just-in-time supply chain. There are two manufacturers who have the largest share of the market there, but with virtually no surge capacity.
We'll run out of IV needles. We'll run out of IV bags. We'll run out of drugs very quickly. Remember I talked about the 80 percent offshore figure? Go to the Society for Health Care Pharmacists' website and you'll see what I'm talking about.
So the whole medical system will collapse, at a time when we still need drugs for heart attacks, cancer, and everything else. We'll be in freefall. That may sound scary, but it's a reality. And unlike Katrina, where the hurricane did some of the destruction and separated people from health care through evacuation or otherwise, the same thing's going to happen here in every city, town, and village in this country as well. We're all going to need things at the same time, and there won't be any products.
At this Business Continuity Summit we just had, we actually had a major discussion in which a lot of voices expressed concern about the internet—that it doesn't have the backbone and elasticity that everyone thinks it has, and just a limited surge could bring down the internet.
Right after Katrina, when FEMA was trying to rescue itself, they put out a call for anyone who had a refrigerated truck unit to come and sit in one of several parking lots in the Gulf states down there, in case they had 10,000 bodies, etc. A contingent of them went. Not all of them, by any stretch of the imagination.
Within 72 hours, major food manufacturers throughout the United States reported that they couldn't ship their goods. They had no trucks. We have a razor-thin capacity in this country right now on virtually everything. They had to get FEMA to release the trucks.
Cities like Seattle have already come to the conclusion they won't be able to have refrigerated trucks, because of that issue. For their work with corpse management, for example, they've already mapped out where every one of the ice arenas in Seattle is. Because you won't be able to bury people, either. In 1969, during the last pandemic, the average time from when a casket was made until it was in the ground was about six months. Today it's a little over a month, and in some areas of the country it's less than that. There's no surge capacity in casket manufacturing. Today, crematorium space in this country is a just-in-time situation too. This means we would quickly run out of the ability to handle bodies in a timely, respectful way. That's in a moderate pandemic, not at all like 1918.
Everything now is done on the basis of just-in-time delivery systems.
Now, think about Katrina. As horrible as it was seeing the Superdome, seeing shots of bodies rotting would kind of take you over the top. What really offends the sensibilities of most Americans is when you see the dead handled in a disrespectful manner. I worry that one of the tipping points in a panic/fear situation over pandemic flu would come when we can't, in a timely way, handle the bodies or get them into cold storage. Those are issues we haven't dealt with at all.
CP: Your answer to this is partly implicit in all you've said, but let me ask you to sum up how you'd rate the U.S.'s and Minnesota's state of preparedness for a pandemic.
Osterholm: I think the U.S. right now is leading the way in terms of the overall issue of pandemic influenza preparedness. The problem is that, for much of the developing world, they don't have a lot to prepare for. They don't have sophisticated supply-chain issues. In some ways I think they're better off than we are, because they already live a very simple, hand-to-mouth existence. For us, with our complicated supply networks, we have a long, long way to go. And I think that's as much a private-sector issue as a government issue. I think Secretary Leavitt has been brutally honest in telling American communities, you're going to be on your own. And I think he's right. You know how you creep, then you walk, then you run? We're still on our knees.
CP: What kind of measures might we in the Twin Cities expect to see undertaken in the short term if an outbreak did reach the U.S.?
Osterholm: I can't say. I don't know. I know that the governor today, in his State of the State address, said that he's just committed $10 million to influenza preparedness. What that means, I can't say. We don't have the details yet.
When the virus gets to this continent as a bird virus, it's important that we not overreact and misinterpret it to mean that now the pandemic's here. This will be an important issue for poultry, but as I pointed out, I think most of the poultry in this country—aside from the free-range organic poultry—is going to be pretty well-protected.
It's about the birds and it's not about the birds. It's about the birds now, but when real pandemic influenza occurs, the birds become inconsequential. It's human-to-human transmission that matters then. That's where we're not any better prepared here than anyone else in the world.
CP: How would you rate the media's performance in covering this story—too salacious, too understated, just right?
Osterholm: It's like talking about the weather, I guess—depends on which day, and where. There have been some outstanding reporters on this issue. Helen Branswell, from Canadian Press, covered SARS and has become the primary person on this. She's done it as well as anybody. On the whole, it has been poorly covered in the sense of distinguishing H5N1 in the bird population from pandemic influenza. That part has not been covered in depth. The issue about covering preparedness—no one has really gone into depth and explored the supply-chain issues to say, okay, what does this all mean? Let's take a community and say, what would happen here? We're so fixated on the news cycle that it's too often about, "what new country has just been infected with bird virus?" and not really get into the issues. Ted Koppel at Nightline did an excellent series of stories and asked good, hard questions. But now he's gone and Nightline hasn't picked it up since he left.
*Cytokines are a class of proteins produced by white blood cells whenever the body finds itself responding to an infection. They vary in function—some cytokines attack invading microbes directly, others relay chemical messages from cell to cell, still others bind with cells in the hypothalamus region of the brain to produce fevers. Cytokines are toxic not only to infectious agents in the body but to the body itself: Much of the pain and discomfort that accompany illnesses like the common flu, for example, are in effect hangover symptoms from the toxic effects of the body's own immune response. The term "cytokine storm" refers to the immune response that occurs when the body is confronted with an infectious agent that reproduces at great speed and in huge volume. This "viral storm" generates an equally huge immune response—the cytokine storm—that can take such a toll on lung tissue (the main battleground where the virus and the immune system face off) that it deprives vital organs of enough oxygen to function, and sets off cascading organ failure.
http://citypages.com/databank/27/1320/article14219.asp?page=2
