Orent Is at It Again

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 lethal second wave. Orent, in The LA Times:

As with any new outbreak, unraveling all of this flu's mysteries will take time. But, using the lens of Darwinian evolution, certain aspects are starting to come into focus. For one thing, it's clear that the virus, which originated in Mexico, is most virulent in that country. The 1,000 or so reported Mexican cases have been either fatal or severe enough to require hospitalization. But because of natural selection, the strains spreading across the world are milder.

According to evolutionary biologist Paul W. Ewald of the University of Louisville, human influenza is usually a mild to moderate disease because it depends on host mobility to spread. The U.S., Canadian and New Zealand teenagers on their spring breaks did not sit in hospitals with the very sick and dying; they mingled with people who were sneezing and coughing but walking around, riding subways, perhaps going to the beach or dancing in nightclubs. People don't start being really infectious until they show symptoms, and whatever symptoms those people had must have been mild enough to remain out in public. The strains sent out around the world were, by definition and necessity, milder than the most lethal strains....

Influenzas that have their origins in huge, crowded animal farms are often more virulent than other flu strains. Germs that kill their hosts quickly tend not to thrive; their hosts die before there is time to pass the virus on. But on crowded farms, the next snout is an inch away, and even virulent strains can gain a foothold. It is the same type of conditions that produced deadly avian influenza in giant poultry farms in Asia over the last 10 years.

I've dealt with the false linkage between the asymptomatic phase and virulence before--a virus with a long latent time that doesn't result in noticeable symptoms can still be very deadly. But what is equally frustrating is that Orent ignores the work that's been done about the non-optimality of infections when they enter new hosts.

Ewald (and Orent) assume that a virus is at an optimal state. If the density of hosts increases, then the virus can evolve to become more deadly (hosts don't need to survive as long for successful transmission). If the density of hosts decreases, then the virus evolves to become less deadly (the hosts need to survive long enough to transmit the virus).

But this argument assumes that the virus has already landed on this optimal relationship. There is no reason to think a virus that evolved in a different host has done so. Essentially, the relationship between transmission and deadliness has been sundered. This could mean that the virus could become more deadly and better at transmission to other people. Alternatively, the virus could be less deadly. We simply don't know.

Despite all of the advances we've made in monitoring the spread of influenza, we still have seen much phenotypic data--in a controlled setting, how dangerous is the virus? The first weeks of an outbreak make the proverbial fog of war look downright transparent. We simply don't know right now how the virus will evolve--again, it was the second wave of the 1918 pandemic that killed the most people. To claim that we've adequately parameterized this is ludicrous--after all, most people didn't die from the 1918 pandemic if they didn't contract it.

Personally, I think this will burn itself out, particularly if we add the swine flu to the annual vaccine. But I freely admit that's a guess based on what we've seen so far, and every disastrous epidemic starts with a single 'good' genotype and 'bad' circumstances (for us). A lot depends on luck. I certainly wouldn't claim that evolutionary biology tells me that we're out of the woods. We simply don't know.

It's irresponsible to claim evolutionary biology tells us otherwise, and it doesn't do evolutionary biology any favors either.

More like this

A little knowledge is dangerous, and that explains the optimism of the journalist, but one has to wonder what excuse Mr. Ewald's has for his optimism?

By hat_eater (not verified) on 04 May 2009 #permalink

You don't have to assume the virus is optimized for Ewald/Orent's argument to be valid. The point is that to the degree the virus is in fact virulent (whether or not it is optimized for transmission/growth), it is less likely to spread, because those who get it will be quickly knocked on their asses and not mingling with others.

Their argument presupposes a short incubation period and/or minimal contagiousness during incubation, and I don't know if that assumption is valid or not. But to the degree it is, there's no logical flaw with their argument.

I was about to post an argument much as Dana H, albeit less eloquent.

"Optimized" means what, exactly? According to Ewald/Orent, would seem factors affecting virus "optimization" include behaviour of its hosts.
By taking this into account, can explain several things re flu that would otherwise seem baffling, or require just arm waving (may suffice for mad biologists, but not most folk into science). For instance:
Occurrence of a lethal flu in tandem with WW1.
Bird flus being mild in wild, but tending to evolve towards being highly pathogenic in farms.
Human flu likewise being mostly mild, and more severe forms as some major changes in human society.

Even fact swine flu looking milder as it moves away from Mexico looks explainable this way. Here in Hong Kong, the guy w flu felt unwell, but was in good enough shape to take taxi.

So, would seem to me flu constantly able to evolve; change circumstances and it can do so. Stick chickens or pigs together, or young men into trench warfare conditions, and you may just get something you don't like.