#3832
Martin W
Participant

Had lengthy correspondence re natural selection and H5N1, with “a correspondent”; also led to comments from Wendy Orent.
> Indicates quoted text within the chunks of quoted text – gets a bit complex like this I’m afraid.

Lest of interest, here goes:

Again, the paper by Ewald, with predictions re evolution of pathogens including flu:

Quote:
Yes, this is well-known. This is a famous piece in learning
ecology in terms of natural selection. This is a reason why experts
more fear a long-range transport (either by humans or birds) than
gradual geographical invasion.

But we can’t predict exactly, particularly when various species
are involved. We don’t even know why LPAIs are so “evolutionary
static” in wild ducks, while they can be so pathogenic to humans
when they happen to enter the human world. The only truth is
“natural selection works”, but we may not know or deal with all
factors of natural selection.

But again: Ewald makes predictions re flus becoming pathogenic entering human world.

Takes special conditions – very sick people able to readily transmit – to evolve a dangerous flu.
Most extreme in 1918: First World War.

Mao maybe helped cause 1957 and 1968 flus.

No such special conditions today; so Ewald argues that we won’t get a highly pathogenic human flu today.

His theory predicts avian flus will be mild in wild birds. Need to have birds flying to carry the flus, so evolution to mild strains. So, to me, we do know why LPAIs are “evolutionary static” in wild birds.
High path strains into wild birds, and quickly to low path. Or extinction.

Quoting Ewald directly:

“With regard to the future I am predicting that such a highly lethal pandemic (i.e., 1 death per 50 infections) will not occur, not from H5N1 and not from any other influenza virus that will arise unless regional conditions allow transmission from immobile hosts, as they did on the Western Front in 1918. This is not “speculation” as is claimed by our hooded critic with the self-aggrandizing name. It is a prediction based on careful consideration of theory and evidence. The future will demonstrate whether it is accurate.”
http://blog.sciam.com/index.php?title=bird_reaper_pt_iii_paul_ewald_replies&more=1&c=1&tb=1&pb=1

Makes sense to me.

Can make analagous predictions for birds (Ewald does so for poultry):
– crowd together, indefinitely, so sick birds can readily transmit: and evolve dangerous flus
– wild situations, need birds to fly to transmit, and equilibrium when flus are mild

That is, predictions fit what we observe. Which to me is science; and not speculation.

Only mystery to me is why this is so widely ignored.

Quote:
> No such special conditions today; so Ewald argues that we won’t get a
highly pathogenic human flu today.

As I have (indirectly) heard from flu experts, some argue the virus
will not enter the human world in the HP form, but others’ claim
is different — we (even virologists) don’t know how HPAI will behave.

“Most extreme in 1918: First World War”

> No such special conditions today

Wouldn’t airplanes, locomotion, population density be special?
We have never met a pandemic strain in such extremely globalized
world — we don’t really have an experience.

> His theory predicts avian flus will be mild in wild birds. Need to have
birds flying to carry the flus, so evolution to mild strains. So, to
me, we do know why LPAIs are “evolutionary static” in wild birds.

Yes, this explains “why”. We don’t exactly know “how”. This
means we don’t know exactly how selection pressure works. (Also
we don’t know how Zq strain retained high path to natural hosts).
As we haven’t seen LPAIs arising from Zq strain, we don’t know the
time-scale this process would require (may not be “very quickly”).

> High path strains into wild birds, and quickly to low path. Or
extinction.

Most look like to be going extinction (i.e. R0 But planes, crowded conditions etc not enough to him; not so special.
Need to have very sick people – immobile with disease – able to readily
transmit the virus.
Crowding doesn’t matter here, if very sick people stay
home/hospitalised.

But flu is already contagious during the incubation period.
Less traffic than in WW I? Less packed people? (imagine Tokyo
trains) Though I can’t figure out the effect, all present factors
seem to increase the risk of a more virulent pandemic.

> I’d figure that with wild birds, there’s always potential for virulent
flus to evolve.
Spectrum of virulence it seems to me (this from chemistry background,
not viruses): get some higher path, others lower path. Get an
equilibrium, depending on prevailing conditions.
As need flying birds to transmit flu in the wild, the equilibrium is
greatly towards non virulent forms. High path forms stay rare. (This
again from chemistry; some memories from when I did this re systems
reaching equilibrium.)

Thanks! This is much easier to understand. If “random” distribution
of mutated form is close to Gaussian, natural selection would work
in this way (for a specific species). If it is very far from Gaussian,
we can’t be sure (because there is no effective average — this might
explain some of social phenomena like Zipf’s law). What if some
populations (due to genetic diversity) are more resistant (not all
infected individuals die, but can excrete substantial amount of the
virus) — we probably need a more complex view. Natural selection on
incubation period may also occur.

> Could well be that doesn’t matter what bird species is: if cram into
captivity, infect with flu, and have substantial chance that birds with
high path forms can transmit flu, then will get evolution towards
higher pathogenic forms.

If this is population density-dependent, how can we be sure our
population density is below a threshold where high path strains can
be sustained for a meaningful (effective transmission to a next
cluster) time? What is the major difference from poultry chickens?
(Well, some of recent pandemic plans from various companies seem to
assume “forced working” of employees with milder symptoms — they
will mix healthy populations during movement or in taxis — we may
eventually be poultry chickens ;-)

One more on “equilibrium theory”, why we have never seen a human
pandemic strain eventually forming a non-pathogenic form (as in LPAIs
in wild ducks?) What would be the difference between ducks and humans?
(Why “evolutionary stasis” is never reached)

By the way, when considering selection pressure, won’t the extensive
use of Tamiflu in pandemic lead to a more virulent strain? Not
necessarily drug resistance, but won’t we be selecting a more neurotrophic
strain (since Tamiflu doesn’t effectively cross the blood-brain barrier),
I casted this question to a public health expert, but haven’t received
a reply. This might be another factor different from past pandemics.

Well, now you ask questions I wish I had all answers for! – should really go directly to Paul Ewald, as I have some understanding but relatively superficial (I’ll forward to science writer Wendy Orent, who has written several articles based on his ideas, and with whom I’ve had some correspondence; she’s now in aiwatch).

The packed people, Tokyo trains or Hong Kong malls, not so important – if people who get sick, v quickly go to bed/hospital.

Seen re flu becoming infectious before symptoms; queried Wendy re this.
How infectious, I wonder, if not coughing/sneezing? How do you transmit virus without doing these things?

Gaussian curve: not sure, but it’s my way of understanding things, as noted based on (physical) chemistry.

Main thing w poultry farms, to Ewald, is that can have (ready) transmission from even very sick chickens – so dangerous forms can transmit, even intensify.

Wendy notes that 1918 flu did become non-virulent, and still circulates.

I also don’t know re Tamiflu; hadn’t known this re brain.
Doesn’t seem wise, to me, to use it extensively; cf antibiotics and resistance.
Rather as I’m also sceptical re vaccinations, perhaps helping sustain h5n1 (when vaccinations and surveillance less than near perfect).

Quote:
> Seen re flu becoming infectious before symptoms; queried Wendy re this.
How infectious, I wonder, if not coughing/sneezing? How do you transmit
virus without doing these things?

If high path mechanism (replicate without trypsin) indeed works,
we don’t necessarily require respiratory organs. Virus replicates
everywhere in the body.

> Wendy notes that 1918 flu did become non-virulent, and still circulates.

The 1918 flu once disappeared (around 1950), reappeared later
(likely from a lab) and now circulating.

Quote:
What if some
populations (due to genetic diversity) are more resistant (not all
infected individuals die, but can excrete substantial amount of the
virus) — we probably need a more complex view.

I’ve noticed that this possibility is a real concern. If such
individuals (or individuals of different species) are sporadic, we
don’t need to worry. But if chains of such individuals are
established? — This corresponds to the “percolation theory”.
(You may have read Simon Levin’s “Fragile Dominion” or Kauffman’s
“At Home in the Universe” in relation to percolation leading to
phase transition and its role in ecosystem).

Quote:
> Well, not sure if water-borne disease should be more specialised to
this transmission route, between humans. Like cholera.
And like cholera, can’t imagine it becoming widespread, but more in few
places w bad sanitation.
Worst SARS outbreak outside hospital (that we know of) was evidently
from sewage (apparently from toilet, somehow reached people’s showers,
and several people infected in an apartment block). Looked scary, but
proved isolated.

Natural selection works as if a pathogen is seeking for a higher basic
reproduction number (not plainly necessarily less lethal). If a pathogen
has an ability to spread in a more efficient way, this would become
a primary route of transmission. If the virus replicates in intestines
or kidneys, sewage would be an efficient place for viral adaptation
(much resembling avian infections??).

Quote:
> As discussed, I don’t believe Osterholm is correct re predictions.

> What may happen though, is that if get pandemic – and no matter if it’s
relatively mild – panic will lead to problems.
Already too many problems (such as worry, Tamiflu stocking etc), even
in US – where no H5N1, just fear of the disease.

As we know, our existence is dependent on the present biodiversity
— a product of ecosystem evolution, to which we best adapt.
We don’t know when our present existence is threatened how much and
how rapidly biodiversity is degraded, but there should be some number
(not easily predicted). The same is true for our society; our life
is dependent on the present social system — a product of social
evolution, to which we best adapt. We don’t know when our present
world is threatened how much and how rapidly social system is degraded.
These two problems are alike, both arising from a complex adaptive
system. Complex social systems could amplify the effect of a minor
mortality.

Time-scales also play a role. If any change is slow enough, we
can, or ecosystem would adapt to a new form. If the change is rapid
enough, they may fail. There is a simple physical analogy; the
adaptation of gas is limited by the sound speed. If change goes
faster than the sound speed, the gas fails to adapt — the net result
is a well-known supersonic shock. The same would be true for our
society. If the spread of the pandemic is rapid enough, our system
would fail to adapt. Of course, with the advent of the internet,
we have a better chance of adaptation before the wave comes. But the
expected result of adaptation is so drastically different from the
current social system, the arrival of pandemic flu will trigger
a reaction that looks like to change the world overnight.
I’m skeptical about such a drastic change in social systems could be
done overnight (even officials declare “immediately”), since no one
is accustomed to the change, and expect the situations something
between adapted (with a drastic change) to less-adapted (little
reaction before the wave reaches, and an immediate panic is
triggered).

Post edited by: martin, at: 2006/02/02 00:49