Evolution, Tradeoffs, Ignoring Biology, and Influenza

Inspired by this excellent post by Revere about the evolution of influenza, I've delved deep into the archives of the Mad Biologist, and summoned up some evolutionary thoughts of my own about influenza:

I meant to post something about evolution and influenza before my travels up north, but I was swamped by work and couldn't get to it. Thankfully, two colleagues, Carl Bergstrom and Marc Lipsitch, have decided to deal with Wendy Orent's faith-based virology. Orent writes (italics mine):

Indeed, a strictly enforced quarantine could do more harm than good. Herding large numbers of possibly infected people together makes it likely that any influenza strain passed among them would actually increase in virulence. Usually, in order to spread, human flu germs need hosts mobile enough to walk around and sneeze on other people. Those flu strains so deadly that they kill or disable their hosts won't get the chance to spread and will die off. This keeps human flu virulence within bounds.

The signal exception is the 1918 flu, which acquired its extreme lethality, according to University of Louisville evolutionary biologist Paul W. Ewald, in the crowded and terrible conditions on the Western Front during World War I. Troops by the train and truckload were constantly being moved in and out of this petri dish, meaning a severely flu-stricken soldier didn't have to move much to infect others.

Suppose that a government official today decided to round up exposed people and move them to a space like the Superdome in New Orleans. It's unlikely that even a crowded Superdome could replicate the conditions on the Western Front. But, depending on how densely packed people were, you could expect the flu strain trapped among them to increase in virulence. You'd be breeding a deadlier flu.

If you let people walk around freely, only those strains mild enough to allow people to stay on their feet would spread easily.

In the field of microbial population biology, we've been discussing Ewald's ideas for a while, and most disagree (including the Mad Biologist). Essentially, Orent's statement requires a tradeoff between transmissibility (i.e., how well the virus moves from host to host) and virulence (how bad an illness the virus causes). Obviously, there's some merit to this idea: a virus that kills a new host instantly couldn't spread, and a virus that spreads easily but causes no symptoms doesn't matter.

But casting about for a magic percentage of mortality that tells us when we should really start to panic is silly. That's because, even if the next influenza were H5:N1 (and some think it might be another avian serotype), the virus probably wouldn't be identical to the 1918 pandemic virus. The parameters of any tradeoff would probably be different due to what one of my dissertation committee members called "those stupid fucking natural history facts" (in this case, we're dealing with molecular natural history).

Even if transmission between hosts were to select for less virulent viruses, a great body of work has shown that this can be counteracted, and even superceeded, by selection within a host for more virulent viruses: the more virulent the bug is, the greater number of viral particles produced. Natural selection is very short-sighted, which is why, in historical retrospect, it often looks very 'random.'

But there's another problem with Orent's argument: influenza can be spread by those who are asymptomatic-a point she alludes to when she wants to argue against quarantine, but forgets when she wants to downplay the possibility of a pandemic. This demonstrates a problem I've always had with Ewald's tradeoff argument: it neglects the possibility of an asymptomatic but contagious phase.

If a virus requires a low infectious dose (i.e., a very small number of viral particles) to establish an infection, but requires a much higher dose to cause significant disease symptoms, this can lead to an asymptomatic but contagious phase. In other words, a person could have enough virus to spread it ('adequate viral load'), but not enough to show symptoms (although as the virus continues to grow within that person, ultimately, he will become ill).

Then there are those stupid natural history facts. There is no evidence that pandemic strains cause noticeable symptoms at a lower viral load than 'regular' influenza (which 'only' kills ~36,000 U.S. residents per year). What makes avian influenza so dangerous is that it is much harder for human immune systems to mount a response against it. This means it is harder to stop an avian influenza infection, but the time to a symptomatic infection doesn't necessarily differ between the two forms of influenza. In other words, the train continues to speed up, but it doesn't come out of the station any faster initially. This would have little effect on the asymptomatic but contagious phase; the link between spread and virulence is rather small or non-existent.

(note: for those of you who are really thinking this through, you might wonder why the time to a symptomatic infection changes very little if the immune system can't stop it. After all, it should be able to grow faster and cause damage more rapidly. The reason is a 'ramp up' time. Unless you have been vaccinated, your body doesn't recognize influenza as trouble right away; there's a gap in time where there is essentially no immune response. That's when influenza can gain a foothold. That's also why vaccination is so important-your immune system knows what to look for right away.)

Avian influenza also may infect other tissues that human influenzas don't; this would cause much more systemic damage. However, this happens much farther along in the illness, decoupling the 'nastiness' of the disease from its ability to spread. Again, where's the tradeoff?

To suggest that we should not worry about an influenza pandemic due to this notion of tradeoffs when the scientific evidence to support that position is lacking is not only irresponsible, but potentially lethal.

my apologies to immunologists for vastly simplifying a very complex and interesting system.

More like this

Wendy Orent, having decided that Paul Ewald is the end-all and be-all of evolutionary epidemiology, is again repeating the mantra that pandemics will evolve to become less deadly. Never mind that, as ScienceBlogling Greg Laden reminds us, the first wave of the 1918 pandemic was far milder than the…
[Warning: this post is fairly long and has a reasonable geek factor. It explores the question whether the virulence of H5N1 "must" moderate as the virus evolves.] The high case fatality ratio of H5N1 (currently around 60%) is a reflection of how virulent this virus appears to be at the moment.…
CDC's Advisory Committee on Immunization Practices (ACIP) recently rolled out their 2009 Recommendations. It's for seasonal flu, for which a vaccine exists, not for swine flu, for which there is (as yet) no vaccine. There is a lot to say on the subject of vaccines (see what we've said over the…
by revere, cross-posted from Effect Measure CDC's Advisory Committee on Immunization Practices (ACIP) recently rolled out their 2009 Recommendations. It's for seasonal flu, for which a vaccine exists, not for swine flu, for which there is (as yet) no vaccine. There is a lot to say on the subject of…

I found this very interesting. One point I'd like to add is that speed of travel has developed massively since 1918 so a similarly transmissible and virulent strain could potentially be a lot worse in todays society, particularly if it does have a long period without syptoms developing.

Diccon

www.dnapropertyltd.com