H5N1 influenza virus--does it live up to the hype?

Posted 19 September 2005 by Tara Smith

↗ The current version of this post is on the live site: https://pandasthumb.org/archives/2005/09/h5n1-influenza.html

I’m sure many of you have been following the media circus related to the tracking of H5N1 influenza viruses (for example, this article, which contains the following quote):

“Right now in human beings, it kills 55 percent of the people it infects,” says Laurie Garrett, a senior fellow on global health policy at the Council on Foreign Relations. “That makes it the most lethal flu we know of that has ever been on planet Earth affecting human beings.”

Now, I’m a big fan of Laurie Garrett, and I obviously have no idea of knowing if this is all the said on the H5N1 mortality rate or if she elaborated further, but it’s quotes like this that lead people to stress more than they should about the H5N1 situation. A new paper just out in the Journal of Infectious Diseases (summary here) demonstrates one reason why the public shouldn’t start freaking out just yet.

The problem with the figure Garrett cites is that it kills ~55% of the cases we know about. This is a classic case of sample bias. Those who are most sick (and thus, most likely to die) are also most likely to go to a hospital or clinic to be examined–and therefore, are also the most likely to have a clinically-confirmed case of influenza due to the H5N1 strain. Hence, the mortality data we have for H5N1 only comes from this sickest segment of the population–artificially raising the mortality rate. Puzelli et al.’s study, then, is timely due to the fact that it shows that sub-clinical infection with avian-type influenza viruses does occur (in almost 4% of their cohort of poultry workers).

For someone versed in influenza evolution and epidemiology, this is both disturbing and unsurprising. First, a bit of background on influenza virus. The virus has an RNA genome, which is segmented into 8 parts. These parts can re-combine and make a novel virus, containing a few segments of each of the parent viruses. In some cases, this progeny virus may be better adapted for a new host than one (or both) of the parent viruses.

An example. Say a human currently is infected with a (human) influenza virus. Let’s say it’s one of these folks in the Italian cohort of poultry workers. Maybe he’s not sick yet, so he goes into work, where he’s exposed to one or more types of avian influenza virus. If he becomes co-infected, the two types of virus can mix, possibly producing a “humanized” avian influenza virus, which may be transmissible between humans while the parent avian virus was not. This is the stuff of pandemics, and the kind of thing that keeps influenza researchers awake at night.

And this isn’t far-fetched. Other groups (such as Chris Olsen’s lab) have shown that workers in close contact with swine can show serological evidence of prior infection with swine influenza viruses, even though the clinical influenza infection is rarely documented (that is, it’s either subclinical infection, or mild enough so that the patient doesn’t see a medical professional). Therefore, these epidemics may very well be percolating as a series of mild infections, allowing the virus to become more adapted to humans (and hence, possibly more transmissible *between* humans), while the attention is focused on those rare severe or fatal cases. This again shows why we need good surveillance for not only influenza, but a number of different infectious diseases. We focus on the acute diseases, but it’s difficult to say how many other pathogens are out there, infecting us, and possibly triggering other types of disease down the road.

ETA: the new issue of Emerging Infectious Diseases has an article on the evolution of Asian H5N1 viruses, here. Of the ones they examined, they don’t see any reassortment with human viruses yet. Additionally, neuraminidase inhibitors work against the isolates they tested.

28 Comments

RBH · 19 September 2005

Two quotations, with some added emphasis:

A new paper just out in the Journal of Infectious Diseases (summary here) demonstrates one reason why the public shouldn't start freaking out just yet.

and

Of the ones they examined, they don't see any reassortment with human viruses yet.

Yeah, but soon, I'll bet. RBH

Mike Walker · 19 September 2005

What's the mortality rate for a more typical flu outbreak? Much less than 1%? 0.1%? Even if the claimed mortality rate for H5N1 of 55% is out by a factor of 10, that should be more than enough to make people sit up and take notice, if not freak them out.

I agree there is a danger of appearing to cry wolf over the bird flu by inflating the mortality rate, but, like Y2K, if it succeeds in getting us much better prepared should the worst ever happen, I'm all for it.

Tara Smith · 19 September 2005

What's the mortality rate for a more typical flu outbreak? Much less than 1%? 0.1%? Even if the claimed mortality rate for H5N1 of 55% is out by a factor of 10, that should be more than enough to make people sit up and take notice, if not freak them out.
— Mike Walker

Estimates in the US are that between ~5-20% of the population are infected with the flu every year: so between 15 and 60 million cases, using a population of 300 million. In the US, mortality usually is at about 36,000 per year; so a fraction of 1%. (Of course, there are error rates both in diagnosis of cases and deaths attributable to influenza, so the numbers won't be exact). There are problems inherent in comparing the two viral mortalities, however. With "typical" flu, we've watched it enough years to have a pretty decent handle on the approximate percentages of cases and deaths, so while our numbers certainly aren't exact, they're likely to be decently close. Additionally, we're working with very large numbers, so a few dozen cases missed won't matter. With H5N1, we're talking about much smaller numerators and denominators. Since late 2003, WHO has recorded around 60 deaths, out of ~115 diagnosed cases. If we take the numbers from the Italian study and estimate that 3% of people in close contact with poultry in the affected countries (Thailand, Vietnam, Indonesia, and Cambodia) may have been exposed, it doesn't take much to get above 1000 or even 10,000 individuals (keeping in mind that there are many more people who can have an effective exposure in those countries, where many people raise their own birds for meat or buy them whole in open markets).

I agree there is a danger of appearing to cry wolf over the bird flu by inflating the mortality rate, but, like Y2K, if it succeeds in getting us much better prepared should the worst ever happen, I'm all for it.

I somewhat agree, but it can come back to bite you in the ass. Were something to happen on a global scale, the fatality rate certainly would not be 50% when the final analyses were published, which throws doubt on all of these predictions of doom and gloom, a billion dead worldwide, etc. The 1918 epidemic only had ~20% fatality rate (depending on whose estimates you use), and that was bad enough. There's no need to overstate the danger, and indeed, good reasons not to.

Bob Davis · 19 September 2005

The fear associated with flu is (should be) based on the failures of the Bush Admin. Whatever the worst case scenario is, that's the one we'll get hit with without an effective gov't.

Mike Walker · 19 September 2005

Thanks for the information, Tara - very interesting. I agree that it's best to get an accurate picture of what the mortality rate would be if the bird flu really does hit. I would say that it's best to err on the side of caution (i.e. don't underestimate the impact) but I admit it must be hard to keep sensationalism out of the news headlines when talking about this stuff.

Ken Shackleton · 19 September 2005

The fear associated with flu is (should be) based on the failures of the Bush Admin. Whatever the worst case scenario is, that's the one we'll get hit with without an effective gov't.
— Bob Davis

Too bad that it won't be selective enough to just make the neo-cons ill.

mike syvanen · 20 September 2005

tara wrote:

The 1918 epidemic only had ~20% fatality rate (depending on whose estimates you use),

It was closer to 1%. No serious estimate as high as 20% has been measured. What made the 1918 flu so deadly was its infectivity. It spread through communities and infected up to 75% of the population. The very high mortality rate of avian flu in humans is not impossible (though the bias mentioned above is quite real, death is much easier to count than someone who sickens and survives). Thus the danger of this flu will come down to whether it increases its infectivity rate without attenuating its mortality rate. As is well understood, very high mortality rates tend to attenuate infectivity rates (ie living infecteds are much more efficient in spreading disease than dead infecteds). NO one really knows what is going to happen. I suspect most likely that nothing worse than what we see today, but the potential for a world wide disaster is sufficiently large that all of the attention we have been giving it is justified.

Steviepinhead · 20 September 2005

Mike Sylvanen said:

As is well understood, very high mortality rates tend to attenuate infectivity rates (ie living infecteds are much more efficient in spreading disease than dead infecteds).

--and this relates to the discussion we were having on another thread about why disease-causing viruses often (but not always) evolve from "parasitic" toward "symbiotic" (or, at least, less virulent). Well put.

Tara Smith · 20 September 2005

It was closer to 1%. No serious estimate as high as 20% has been measured. What made the 1918 flu so deadly was its infectivity. It spread through communities and infected up to 75% of the population. The very high mortality rate of avian flu in humans is not impossible (though the bias mentioned above is quite real, death is much easier to count than someone who sickens and survives).
— mike syvanen

You're right; I misspoke. I was thinking of the "~20% of the population was infected" figure.

NO one really knows what is going to happen. I suspect most likely that nothing worse than what we see today, but the potential for a world wide disaster is sufficiently large that all of the attention we have been giving it is justified.

And just to clarify, I'm in no way writing H5N1 off or anything; I definitely take it seriously as a threat and am watching all the ProMed and other updates closely. I just see so many in the general public becoming fixated on the high (and IMO misleading) mortality rates, and getting so involved in the hype without understanding the underlying issues and the science behind them. As any regular PT reader knows, this seems to be a common theme here in the US...

Russell · 20 September 2005

NO one really knows what is going to happen. I suspect most likely that nothing worse than what we see today, but the potential for a world wide disaster is sufficiently large that all of the attention we have been giving it is justified.

Oh, I know what is going to happen. I just don't know when. It may or may not be this strain of avian flu, but there's no shortage of others out there.

Steve LaBonne · 21 September 2005

Something as lethal as the Spanish Flu virus, as has been pointed out many times, could only have evolved in the particular conditions existing at the end of WWI, and no such incubator exosts for avian flu. WHEN will epidemidiologists start reading and understanding Paul Ewald's book, published lo these many years ago?

Tara Smith · 21 September 2005

Something as lethal as the Spanish Flu virus, as has been pointed out many times, could only have evolved in the particular conditions existing at the end of WWI, and no such incubator exosts for avian flu. WHEN will epidemidiologists start reading and understanding Paul Ewald's book, published lo these many years ago?
— Steve LaBonne

Many have indeed read and understand it. Not all of us agree with all parts of it. :) While I definitely agree WWI could have been a factor in the 1918 epidemic, I disagree that "only the conditions existing at the end of WWI" could lead to an outbreak of that magnitude (and did he really say "only" those conditions could? My copy is in my other office cross-town at this moment, so I can't check it right now).

Steve LaBonne · 21 September 2005

The unusual thing about 1918 was immobilized sick people in large numbers, packed into close quarter in trenches, encampments and ships and other military transports in close proximity to large numbers of uninfected people. How do you propose that a virus with the very high mortality rates some alarmists have been predicting for bird flu, could arise without those special conditions? (Normally flu needs victims who can walk around in order to spread.) And where do those conditions exist in Asia now?

SteveF · 21 September 2005

Another Tsunami maybe or some other natural distaster......

Tara Smith · 21 September 2005

The unusual thing about 1918 was immobilized sick people in large numbers, packed into close quarter in trenches, encampments and ships and other military transports in close proximity to large numbers of uninfected people. How do you propose that a virus with the very high mortality rates some alarmists have been predicting for bird flu, could arise without those special conditions? (Normally flu needs victims who can walk around in order to spread.) And where do those conditions exist in Asia now?
— Steve LaBonne

First, I don't know what you're considering "very high mortality rates." As I said above, one of my beefs is with those very "alarmists" who are (IMO) suggesting artificially high mortality rates (50-60% or higher) for the H5N1 viruses. Second, I don't think we have enough evidence to accept as fact that the conditions you mention are necessary to produce a highly transmissible influenza virus with a higher-than-usual mortality rate, even if they indeed were responsible for that in 1918. I'm a big fan of Ewald's work, but I think he does a bit too much extrapolation and generalization from too few data points. There have already been some models looking at spread of avian influenza. In Ferguson et al.'s model here, they suggest that "containment is unlikely if R₀ exceeds 1.8 for the new pandemic strain." ((Longini et al. cite similar figures using their model). In these papers, even with containment, vaccination, antivirals, etc., the virus does not have to spread to many people in order for a pandemic to take off. And since mortality would still be a rare occurrence, and mild or sub-clinical infections would also occur and could spread the virus to naive individuals, I don't think it's a stretch at all for a virus with 4 or 5, or maybe even 10, times the lethality of an "everyday" flu virus to have an R₀ of greater than 1.8.

Steve LaBonne · 21 September 2005

Where do your large lethality numbers come from if not "extrapolation and generalization", based on neither data nor arguments? We've heard these alarmist predictions (from people like Garrett among others) repeatedly for other diseases- where's the beef? I see no reason at this point to suspect that a worst-case svain flu scenario will lead to more than 2 - 3x the deaths from a "normal" flu pandemic- if that. Bad, but not a best-seller-list-worthy catastrophe. So, it'll be on to the next exotic diseaese for the alarmists...

Steve LaBonne · 21 September 2005

That should be "avian"- don't ask me what "svain flu" is. Must remember that preview is my friend... ;)

Tara Smith · 21 September 2005

Where do your large lethality numbers come from if not "extrapolation and generalization", based on neither data nor arguments?
— Steve LaBonne

There are hard numbers out there in regard to current mortality rates. As of today, that roughly 55% mortality rate is indeed correct--for cases which have been clinically diagnosed and laboratory-verified. I discussed above why I think this is too high and inaccurate, but I don't know how you go from that to saying that it's not based on any data. IMO the data is just incomplete.

We've heard these alarmist predictions (from people like Garrett among others) repeatedly for other diseases- where's the beef? I see no reason at this point to suspect that a worst-case svain flu scenario will lead to more than 2 - 3x the deaths from a "normal" flu pandemic- if that. Bad, but not a best-seller-list-worthy catastrophe. So, it'll be on to the next exotic diseaese for the alarmists...

And this is exactly the reason why I think scientists need to be cautious about discussing issues like these--because it leads to attitudes like this, which makes it harder to do our job and to get the public to take a threat seriously. I also think you're beating a bit of a strawman here. Few scientists would claim the threat of avian influenza is a "best-seller-list-worthy catastrophe." But even if it does cause, as you say, 2-3 times the mortality than a "normal" flu pandemic--say, comparable to 1957 or 1968--that's still going to overwhelm resources. From the CDC's site:

Many scientists believe it is only a matter of time until the next influenza pandemic occurs. The severity of the next pandemic cannot be predicted, but modeling studies suggest that its effect in the United States could be severe. In the absence of any control measures (vaccination or drugs), it has been estimated that in the United States a "medium---level" pandemic could cause 89,000 to 207,000 deaths, between 314,000 and 734,000 hospitalizations, 18 to 42 million outpatient visits, and another 20 to 47 million people being sick. Between 15% and 35% of the U.S. population could be affected by an influenza pandemic, and the economic impact could range between $71.3 and $166.5 billion.

For this "medium-level pandemic," that's still around 3-6 times the usual number of deaths from influenza in a year, and ~1.5-4 times the number of hospitalizations, with a huge economic impact. While I think it's a mistake to be support all the doom 'n' gloom, "the world's coming to an end and The Stand scenario is here" type of prophecies, I think it's equally wrong to dismiss the large impact that any influenza pandemic will have; and obviously, the higher the mortality, the worse it will be.

Steve LaBonne · 21 September 2005

I would go futher and say that the current numbers are likely to be extremely unreliable, again for reasons you discussed yourself- the denominator of the fraction, the true number of people who have actually been infected, is really not accurately known at all. I don't see how anything of value can be extrapolated from such numbers. As you note, the problem of a major flu pandemic of a "normal" kind is very real and very serious, and don't you think the media hype about superviruses only serves to distract attention from that sober reality and the steps needed to prepare for it?

Tara Smith · 21 September 2005

I would go futher and say that the current numbers are likely to be extremely unreliable, again for reasons you discussed yourself- the denominator of the fraction, the true number of people who have actually been infected, is really not accurately known at all. I don't see how anything of value can be extrapolated from such numbers.
— Steve LaBonne

I don't recommend throwing the baby out with the bath water. We do know 1) how many people are presenting to hospitals with "regular" influenza; 2) how many are presenting with avian influenza; and 3) the relative differences in mortality in the two populations. Again, these are all subject to the drawbacks and flaws I discussed above, but we're certainly seeing more deaths from H5N1 than from other strains circulating in the same area. IMO the quantitative analysis (55% mortality) isn't as valuable as the qualitative data ("this virus is worse than your typical flu--this is bad news.")

As you note, the problem of a major flu pandemic of a "normal" kind is very real and very serious, and don't you think the media hype about superviruses only serves to distract attention from that sober reality and the steps needed to prepare for it?

Yes and no. I've already discussed some of the cons of all the publicity and hype, but on the "pro" side, I think it heightens awareness about the notion of influenza pandemics, period, and can serve to get people to actually get out there and get their flu vaccine. The reports of avian influenza have led a lot of people to get interested in influenza virus history and epidemiology in general as well, and as such, people are increasingly realizing that even your garden-variety flu year can be a major threat, particularly if you're in one of the high-risk groups. (My parents and grandparents finally broke down and got their vaccinations last year). So I think a bit of a kick in the pants isn't always a bad thing.

mike syvanen · 21 September 2005

steve wrote:

The unusual thing about 1918 was immobilized sick people in large numbers, packed into close quarter in trenches, encampments and ships and other military transports in close proximity to large numbers of uninfected people. How do you propose that a virus with the very high mortality rates some alarmists have been predicting for bird flu, could arise without those special conditions?

First sentence. Not true. The first cases occurred outside of the front lines. You just described where the epidemic was worse. There were many small towns in the US (well beyond the front) where up to 80% of the people became infected. This was a flu of unprecedented infectivity. Second sentence. We really do not know what the conditions were when the 1918 flu first emerged. As above, once it emerged, it spread to the front.

Bayesian Bouffant, FCD · 21 September 2005

Meanwhile, back at the ranch: Old Flu Drugs Bite the Dust

... Paying the price? Widespread use of an over-the-counter product containing amantadine may have caused the H3N2 influenza strain to lose its sensitivity to the drug. ...

Steve LaBonne · 21 September 2005

Mike, the problem with flu, and the whole basis for the bird flu scare, is the rapid evolution of flu viruses. The site of first infection in 1918 is not necessarily where the most virulent form of the Spanish Flu arose. (In those days of course they didn't even know the illness was caused by a vuris, let along have the ability to keep track of changes in the viral genome). And we do know that no other flu pandemic, occurring in more "normal" conditions, has carried anything like the mortality of that unique epidemic. Furthermore we know that all the alarms about "breakout" of thing like Marburg, Ebola and SARS, contrary to the scaremongers, have failed to materialize. To me this pretty strongly suggests that Ewald is right and the alarmists are wrong. It's just not so easy to generate a pandemic illness with very high mortality rates. Selection pressures for high mortality and rapid transmission of the illness normally work at cross purposes with each other, as Ewald argues, except under the kind of exceptional circumstances that existed at the time of the Spanish Flu. I remain more worried than Tara about the boy-who-cried-wolf effect of all the "Hot Zone"-type hype on rational public responses to realistic infectious disease threats.

Jim Harrison · 21 September 2005

The sad probability is that every threat will turn out to have been overhyped except the one that turns out to be for real. And then we'll all wonder why it wasn't taken seriously.

It isn't just in made-for-TV movies that the scientist character is trapped in a dilemma between warning the village about the volcano and screwing up the real estate deal for the locals. Has anybody got a good idea about how to deal with this perennial problem?

mike syvanen · 21 September 2005

Steve

I agree and disagree. The likelihood of any one of these viruses causing a world pandemic of biblical proportions is small. But the costs will be so high if the unlikely happens. Insurance companies deal with this all the time. In these cases we deal with probability of occurrence times the cost of an event. (In poker we call it pot odds, go ahead and bet even if it is unlikely that one will necessarily win).

Let us play this out. Consider a KT asteroid type event.
P (10^-8 per year) times cost (6*10^9
lives) = 60.

Consider rabies

P (close to 1.0 per year) times cost (100 lives per year) = 100.

Thus each of us face about the same danger per year from catastrophic asteroids or rabies.

I apply this kind of thinking to disease epidemics.

In the past 500 years we have had a number of major events. Let us say the chances are that in any given year P is .01. If cost is 1% of the world's population (6*10^7) = 6 x 10^5. The worst we can imagine, based on historical precedent is 30% of the world's population.

Thus we should worry about 600 times (or 1800 times in a worse case) more about this outcome than we should about rabies. If we accept that the amount of money and time we spend on protecting ourselves against rabies is reasonable, I think our concern about one of these catastrophic epidemics is not misplaced.

Steve LaBonne · 21 September 2005

What extraordinary measures do you recommend be taken against avian flu that wouldn't have been taken against an "ordinary" looming flu pandemic unaccompanied by "Hot Zone" hype?

mike syvanen · 21 September 2005

What extraordinary measures do you recommend be taken against avian flu that wouldn't have been taken against an "ordinary" looming flu pandemic unaccompanied by "Hot Zone" hype?

None that we arn't doing today. The WHO is actually doing quite a lot monitoring of what is going on in SE Asia. The press coverage is only to be expected. The hype is also the predictable hysteria whenever infectious disease is discussed. For example, do you realize how little danger we were ever exposed to by an anthrax spore attack. Even if the perps had succeeded in contaminating every PO in the US, the number of deaths would have been quite small. Anthrax, which does not spread from person to person, also happens to be quite suseptible to numerous antibiotics.

Russell · 21 September 2005

Anthrax is a reasonably known quantity, so even though the attack was lethal, and therefore not something to be taken lightly, I agree it was not a WMD style threat. The "news" media being what they are, though, it's very difficult to imagine how the facts (random people getting lethal letters, intentional contamination of congressional offices) could be reported in low-key, let's-not-panic sort of way.

SARS, on the other hand, was completely unknown. We didn't know where it came from, what its ancestors were, how to predict anything at all, except on the basis of some sweeping generalities. This virologist was unwilling, at the time, to make any predictions about whether the outbreak was going to balloon or fizzle the next season.

I think it would be a huge mistake, however, to assume - again on the basis of some sweeping generalities - that the extraordinary public health measures that were deployed were an over-reaction. Moreover, had there not been a lot of public alarm, I'm not at all confident that the Chinese government's reluctance to recognize a problem would not have led to the fulfillment the "alarmists'" concerns.

Likewise with avian flu. It's in the nature of emerging infectious diseases to be unpredictable. So while I endorse the general principle that extreme pathogenicity tends to be incompatible with extreme contagiousness, that represents a quasi-equilibrium view.

Catastrophic plagues have happened in the past. I doubt if we've seen the last one.