Study challenges notion of 'pandemic' flu

[Shiloh]

Source: http://web.mit.edu/newsoffice/2008/pandemic-0411.html

Study challenges notion of 'pandemic' flu
CDC-labeled pandemics no deadlier than seasonal influenza

April 11, 2008

The widespread assumption that pandemic influenza is an exceptionally deadly form of seasonal, or nonpandemic, flu is hard to support, according to a new study in the May issue of the American Journal of Public Health.

The study challenges common beliefs about the flu--in particular the Centers for Disease Control and Prevention (CDC) claim that "the hallmark of pandemic influenza is excess mortality."

Peter Doshi, a graduate student in the History, Anthropology, and Science, Technology and Society Program at MIT, based his study on an analysis of more than a century of influenza mortality data. He found that the peak monthly death rates in the 1957-1958 and 1968-1969 pandemic seasons were no higher than--and were sometimes exceeded by--those for severe nonpandemic seasons.

Doshi says the pandemic-equals-extreme-mortality concept appears to be a generalization of a single data point: the 1918 season, a period in which "doctors lacked intensive care units, respirators, antiviral agents and antibiotics." He argues that "had no other aspect of modern medicine but antibiotics been available in 1918, there seems good reason to believe that the severity of this pandemic would have been far reduced."

As may be expected given improvements in living conditions, nutrition and other public health measures, influenza death rates substantially declined across the 20th century. Doshi calculates an 18-fold decrease in influenza deaths between the 1940s and 1990s, a trend that began far before the introduction of widespread vaccination.

Noting the gap between evidence and fear, Doshi identifies possible reasons that pandemic flu might be so misunderstood, including the possibility that commercial interests may be playing a role in inflating the perceived impact of pandemics. With public policies such as universal vaccination being discussed and more than $5 billion of federal money spent on preparing for the next pandemic, the study raises many important questions of public policy.

"Should the trends observed over the 20th century continue to hold in the 21st, the next influenza pandemic may be far from a catastrophic event," he concludes.

 

[Laidback Al]

Re: Study challenges notion of 'pandemic' flu

American Journal of Public Health, 10.2105/AJPH.2007.119933

Trends in Recorded Influenza Mortality: United States, 1900?2004
Peter Doshi 1*
1 Massachusetts Institute of Technology
* To whom correspondence should be addressed. E-mail: pnd@mit.edu.

Abstract
Objectives. I sought to describe trends in historical influenza mortality data in the United States since 1900 and compare pandemic with nonpandemic influenza seasons.
Methods. I compiled a database of monthly influenza-classed death rates from official US mortality tables for the years 1900 to 2004 (1905?1909 excluded), from which I calculated adjusted influenza season (July 1?June 30) mortality rates.
Results. An overall and substantial decline in influenza-classed mortality was observed during the 20th century, from an average seasonal rate of 10.2 deaths per 100 000 population in the 1940s to 0.56 per 100 000 by the 1990s. The 1918 to 1919 pandemic stands out as an exceptional outlier. The 1957 to 1958 and 1968 to 1969 influenza pandemic seasons, by contrast, displayed substantial overlap in both degree of mortality and timing compared with nonpandemic seasons.
Conclusions. The considerable similarity in mortality seen in pandemic and nonpandemic influenza seasons challenges common beliefs about the severity of pandemic influenza. The historical decline in influenza-classed mortality rates suggests that public health and ecological factors may play a role in influenza mortality risk. Nevertheless, the actual number of influenza-attributable deaths remains in doubt.
Key Words: Epidemiology, Immunization/Vaccines, Infections, Public Health Practice, Mortality, Statistics/Evaluation/Research
http://www.ajph.org/cgi/content/abstract/AJPH.2007.119933v1

Access to complete article requires a fee or subscription.

 

[Laidback Al]

Re: Study challenges notion of 'pandemic' flu

American Journal of Public Health, 10.2105/AJPH.2007.119933

Trends in Recorded Influenza Mortality: United States, 1900–2004
Peter Doshi 1*
---------------------------------------------------------------------

Critique --
<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>
By definition, an epidemic is a regionally localized event, while a pandemic is a worldwide event. Doshi is not studying mortality rates of worldwide pandemics but is simply taking a localized epidemic perspective on <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-comffice:smarttags" /><st1lace w:st="on"><st1:country-region w:st="on">United States</st1:country-region></st1lace> influenza mortality data from 1900 to 2004. He also chooses to restrictively define a pandemic as an influenza outbreak with "excess mortality".
<o></o>
As readers of FluTrackers know there are other metrics to identifying a pandemic beyond excess mortality. For example, from WHO....<o></o>

"A pandemic can start when three conditions have been met: a new influenza virus subtype emerges; it infects humans, causing serious illness; and it spreads easily and sustainably among humans." <o></o>
http://www.who.int/csr/disease/avian_influenza/avian_faqs/en/

<o></o>
Note that the WHO definition of a pandemic does not require "excess mortality" as a necessary or sufficient condition for a pandemic. Using a restrictive definition of a pandemic based on "excess mortality", Doshi can claim that 1957 and 1968 don't meet his definition of a pandemic -- but only for the United States because he doesn't study worldwide mortality rates during these periods.
<o></o>
He "calculates an 18-fold decrease in influenza deaths between the 1940s and 1990s, a trend that began far before the introduction of widespread vaccination." With a narrow focus on the United States, a first world country, it is not surprising that he concludes that "improvements in living conditions, nutrition and other public health measures, influenza death rates substantially declined across the 20th century."
<o></o>
In the news release Doshi is reported to claim that commercial interests may be fanning the flames of pandemic fear. He concludes "the next influenza pandemic may be far from a catastrophic event."
<o></o>
This is a rather myopic view of the potential impacts, both direct and indirect, of a deadly pandemic. Assuming that the most likely current candidate for a novel pandemic virus is H5N1, with a current overall CFR of 63%, I think he needs to reconsider his data.
<o></o>
He also needs to reconsider his tacit assumptions about the stability of the health care system in the <st1lace w:st="on"><st1:country-region w:st="on">United States</st1:country-region></st1lace>. Many anecdotal posts here at FluTrackers indicate that with a 30% CAR the <st1:country-region w:st="on"><st1lace w:st="on">U.S.</st1lace></st1:country-region> health care system will be brought to its knees. Any first world health care advantages developed during the 20th century i.e. "intensive care units, respirators, antiviral agents and antibiotics" will evaporate very quickly, and primary as well as secondary deaths attributable to a pandemic will numerically explode.
<o></o>
This article allows Doshi to add a publication to his graduate student vita, but the article, with a naive view of a pandemic, does little to further our understanding of the nature of world-wide influenza pandemics. It is unfortunate that the peer reviewers at the American Journal of Public Health didn't scrutinize this article in more detail.<o></o>

 

[gsgs]

Re: Study challenges notion of 'pandemic' flu

a pandemic typically infects 20-40% of the population,
while seasonal flu only infects 5-20% and that is
for 3 different serotypes and also different types
within the same serotype whose common anchestor is
more than 1 season ago in some segment.

 

[cartski]

Re: Study challenges notion of 'pandemic' flu

Excellent response, Laidback Al.

I hope it was submitted to the journal as a publishable commentary.

J.

 

[sharon sanders]

Re: Study challenges notion of 'pandemic' flu

Excellent response, Laidback Al.

I hope it was submitted to the journal as a publishable commentary.

J.

I agree with Cartski.

btw - We are a publishable commentary. lol

 

[HenryN]

Re: Study challenges notion of 'pandemic' flu

a pandemic typically infects 20-40% of the population,
while seasonal flu only infects 5-20% and that is
for 3 different serotypes and also different types
within the same serotype whose common anchestor is
more than 1 season ago in some segment.

Please. Any links for the above "definition"?

 

[cartski]

Re: Study challenges notion of 'pandemic' flu

I agree with Cartski.

btw - We are a publishable commentary. lol

Yes, I agree.

What I meant was putting the commentary in the journal's letters or replies box so that a thorough reader of the journal article will immediately see the contrary view, with reference to FT and other research publication services.

J.

 

[Mamabird]

Re: Study challenges notion of 'pandemic' flu

a pandemic typically infects 20-40% of the population,
while seasonal flu only infects 5-20% and that is
for 3 different serotypes and also different types
within the same serotype whose common anchestor is
more than 1 season ago in some segment.

As I understand available literature on the subject, the clinical infection rate for the 1918 Pandemic in the US was 29%, for the 1957 Pandemic it was 24%, and for the 1968 Pandemic 15%. In talking with US business people about 1968 and the absentee rates experienced, 1968 was just another flu season. Not much to talk about.

 

[Laidback Al]

Re: Study challenges notion of 'pandemic' flu

Excellent response, Laidback Al.

I hope it was submitted to the journal as a publishable commentary.

J.

Thanks for the comment. No, it was not submitted to the journal.

<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o>I was unable to access the full article so I must give Peter Doshi the benefit of the doubt. </o>Most of my comments responded to the press release (post 1). Without access to the full article it is hard to know what his position specifically is.
<o></o>
I hope that the news office at an august institution such as MIT has not sunk to the depths of manipulating author?s quote to creating sensationalistic, tabloid-like headlines for an official press release.

 

[Oracle]

1968-70 Pandemic seasons were no walk in the park.

1968-70 Pandemic seasons were no walk in the park.

Trends for Influenza-related Deaths during Pandemic and Epidemic Seasons, Italy, 1969–2001. http://www.cdc.gov/EID/content/13/5/694.htm

Italy's second year was 3-fold greater than the outbreak in the US, and twice as high as the rest of Europe. I don't think anybody would typify the 1969-70 season as mild in Italy, Europe, or parts of Asia/South Pacific.

Multinational Impact of the 1968 Hong Kong Influenza Pandemic: Evidence for a Smoldering Pandemic. http://www.journals.uchicago.edu/doi/abs/10.1086/431150

See Fig. 2, P&I mortality trends for Seasons 1 and 2 for H3N2, 6 countries.

H3N2 is a butt-kicking human influenza strain, with much higher case rate, harsher symptoms and more complications/deaths than other strains.

 

[cartski]

Re: Study challenges notion of 'pandemic' flu

On re-reading the news article and abstract of this study, I noted the reference to CDC. The article goes to the credibility of the scientists at CDC, in effect saying that they errred in determining when a pandemic actually occurs.

I wonder if CDC plans an official response.

J.

 

[JJackson]

Going off on a tangent - again.

Going off on a tangent - again.

Firstly let me add my thanks to Al for his typically Al like intervention.

If we agree that a flu pandemic is an epidemic with extended global scope the reason is normally because it is new, in the sense its hosts immune systems’ are not primed to react to it, and so it spreads more easily as it does not just spread until it hits a population that had it last year so dropping Ro < 1. The virulence of the strain in question is largely unrelated - as long as it does not kill its host before they have a chance to infect someone else.
But that is not really what I want to say. I would like to bring up a rather obvious point which I am not sure has been aired enough. It is to do with our current top pandemic candidate HPAI H5N1 and it’s the fact that it is HP. While we know the serotypes for the last century’s pandemics we can not go much further back and they have all been LP. This is not an academic distinction while HP stands for Highly Pathogenic - and that refers to its effect on poultry not us - it is more technically a description of the number - and type - of bases at the Hemagglutinin cleavage site. While this may seem a rather odd way to split Type A influenza into two groups what it is actually measuring is very important. For the virus to infect a cell it needs to bind to the cell surface and then be cleaved to get into the cell. The cleaving is done by a protease enzyme supplied, not by the virus, but by the host. Both seasonal type A flus and all know past pandemic flus have had short cleavage sites which can only be cleaved by a limited range of proteases, normally tryptase clara which is localised to clara cells in the respiratory tract. HP H5N1’s polybasic cleavage site can be snipped by many proteases which can be found all over the body. Once H5N1 (or any HP flu) infects you it has the ability to get in to a wide range of cell type and manifests it self not just as a respiratory tract infection. Although 1918 H1N1 was significantly more virulent that ’57, ’68 or our seasonal flus it was ‘only’ an LP flu. My point is H5N1 is very different not because our immune systems have not met it before but because – to the best of my knowledge – it is the first time man has met a not exclusively respiratory form of flu.

To continue this train of thought to its logical conclusion should we be buying face masks? Just because all the forms of flu we have met before were via coughs and sneezes it does not hold true for HP H5N1. The natural reservoir for Type A flus is in birds and for them infection is via the intestinal tract, mammals are a bit of a side show in the flu story. We think of flu as being caught via the nasopharynx but that is because that is where the clara cells are, if H5N1 does not need them we may need to start thinking of it more in terms of intestinal transmission. Even in birds it is requiring some difference in sampling. Typically other strains are collected by cloacal swab but experience is showing for H5N1 better results are gained from nasal swabs. I am sure there is an expression in colloquial English that covers this something about having got things *** about face.

Wash your hands!

JJ<o></o>

 

[cartski]

Re: Study challenges notion of 'pandemic' flu

So, if I understand JJackson's excellent analysis, there's yet another huge gaping hole in Doshi's work; namely, he has assumed that all human pandemic flus have the same pathogenic potential. In other words, he's assumed that all human flu variants have the same power to produce disease.

His "challenge [of the] common beliefs about the severity of pandemic influenza" is no challenge at all because he fails to consider the biological differences between LP and HP viruses.

Has anyone read the full article? Is the assumption Doshi makes there also?

J.

 

[gsgs]

Re: Study challenges notion of 'pandemic' flu

hard to imagine that something can go pandemic quickly
as seasonal flu without respiratory transmission

 

[cartski]

Re: Study challenges notion of 'pandemic' flu

hard to imagine that something can go pandemic quickly
as seasonal flu without respiratory transmission

If it is intestingal transmission, should we not see outbreak in a similar distribution and places as other infections that require intestinal transmission?

And in the few h2h situations we've seen, these are close contact transmissions, likely (although this is without seeing the reports) with poor sanitary and protective measures. So maybe the air transmission was ineffective, but the ingesting of the virus was. So we have h2h, but not with pandemic potential.

Perhaps we should be looking for comparison not to 1918, but to other pandemics without respiratory transmission - if such a thing existed.

J.

 

[JJackson]

Dump the data

Dump the data

Gsgs:
You are quite right. It is not a good epidemic transmission method respiratory would be better. But speed is not a prerequisite. HIV is doing very nicely and is a text book pandemic but in slow motion compared to flu pandemic. I was really just trying to get people thinking more about the about the potential difference that HP?s extended tissue tropism allows for.
<o></o>
I didn?t really address Doshi?s paper ? or at least the press overview in post #1 as Al had picked over most of its weaknesses. He could have added that while antibiotics would have been very useful in most seasonal flus, as a treatment for secondary pneumonia, in 1918 many of the fatal case involved primary pneumonia ? for which it would not have helped.
<o></o>
If we completely ignore all the data (this is a new scientific method I am trialling) and look at what ought to happen from first principles Doshi should be wrong (I am also going to ignore the HP/LP complication covered before).
<o></o>
Firstly there isn?t enough data to say anything very definitive from. Anything before 1900 is sketchy (I am being polite) after that the pattern is there is only one Type A strain circulating at a time and in 1918, ?57 & ?68 a new strain emerged and supplanted the incumbent. There is one exception the H1N1 strain that reigned from 1918 to 1957 suddenly reappeared in the 1970s and since then we have had both H3N2 and H1N1 (I have never understood why ? of the 144 possible Type ?A? strains - birds seem to coexist with many in parallel but we typically take ours serially). So in the history we have strains for n=3.
<o></o>
So returning to first principles the pandemic strain becomes the seasonal strain.
Case zero in a flu pandemic is infected with avian influenza but infects with human flu, by definition he, or she, is the one who transmutes the virus.
Humans, up to this point, are a dead end host and transmissibility & virulence are evolutionarily irrelevant. The starting virulence is a bit of a lottery and could be almost anything as it is a function of how ill a pathogen, in equilibrium with another species, happens to make us ? and as we can see with H5N1 an asymptomatic Mallard shedding virus can make us very sick indeed. Once the human pandemic is underway the virus will undergo rapid genetic change under the severe selection pressures imposed by a new host species. Virulence ? in humans - becomes relevant to the virus as too severe and the host does not get out and about infecting others. You can have too mild also. As the virus is hijacking your cellular machinery to produce more virus if you are not feeling ill then that means it could be hijacking more cells and making more virus i.e. it is evidence the virus is not milking you hard enough. Over time the virus and host will adjust to find equilibrium.
Going back to Doshi then you should not be able to say anything about starting virulence only about where it ends up.<o></o>

 

[JJackson]

n=3

n=3

On reflection I am not sure I explained the n=3 bit very well.
Al critiqued Doshi&#146;s conclusions but in my first post I rather ignored Doshi and talked about something else(HP/LP); the reason being I did not buy the premise. On first reading I took post #1 to mean Doshi was making predictions about future pandemics based on past pandemics. Seems like a good idea but, if you only have 3 past pandemics we know anything much about, how reliable can his statistical analysis be (leave the SPSS in its box).
To put it another way if you were E.T and had landed on this planet and only ever seen three dogs a Spaniel, a Collie and a Labrador would you be able to predict a Great Dane, Dachshund or Wolf?

 

[Laidback Al]

Re: n=3

Re: n=3

. . . To put it another way if you were E.T and had landed on this planet and only ever seen three dogs a Spaniel, a Collie and a Labrador would you be able to predict a Great Dane, Dachshund or Wolf?

That is a very understandable analogy.

Indeed, the main criticism I have of the structure of the article, as reported in the news release, revolves around the many hidden assumptions about pandemics.

First, he commingles the pandemic data with seasonal influenza data to study death rates. So it would seems that he has a large, statistically valid, sample for analysis. But as we know, a pandemic influenza is not just a seasonal influenza with a higher death rate. Removing seasonal influenza data from his study reduces his sample size to 3.

You are correct that a sample size of 3 (three past pandemics) is statistically meaningless, especially when trying to predict a trend. One hidden assumption in this article is that the CFR across the the past three pandemics has remained constant. He needs this hidden assumption to explain that the long term decline in associated death rates is a function of improvements in 20th century medicine rather than a variable CFR.

Doshi's notes his only outlier in excess mortality is the 1918-1919 pandemic. The estimated CFR for that pandemic is often reported to have been about 2.5%. The current CFR for H5N1 is about 63%, 25 times more fatal than the 1918 strain. If H5N1 becomes the next pandemic strain it will clearly be the Wolf in your analogy.

 

[Oracle]

Re: n=3

Re: n=3

Minor glitch in grand theory of HP vs LP:

1.The 1918 strains did not have the basic amino acid cleavage site.
2. H5N1 clades show variability in basic amino acid cleavage site (acquired, then lost in some strains).

Ergo, while it may effect upper respiratory tract receptor recognition, there are other factors at work that determine pathogenicity and transmissibility.

Doshi does indeed note that the 1918 pandemic is an outlier in his plot. It is also an outlier when a plot of genetic sequence vs protein sequence polymorphisms are evaluated for a variety of mammalian HP strains (including the sequences of the last 3 human pandemics). This, Taubenberg did in a recent article; indeed, the 1918 strains are outliers in this plot, FAR removed from the comparatively wimpy CFR of the more recent pandemics.

What is missing in the defined features of pandemic versus seasonal, is the degree to which global populations are affected, in total population affected and symptom severity (a function of both host nativity and strain virulence), infectious duration and strain adaptation, reflected in a second or third wave of continued high virulence, despite the expected drop due to host adaptation.

Yes, I read the paper.