Enlisting Darwin in the bird flu debate

The panic over H5N1 reveals that even experts are failing to grasp how natural selection works. Paul Ewald breaks the good news.

The general public are not the only ones confused about the threat posed by bird flu - so, it would appear, are most experts. The first step toward identifying both hyperbole and complacency is to acknowledge a simple truth. The threat posed by this disease is evolutionary - it is about a bird virus evolving into a human virus. But so far, the discussion has focused on mutation as though it were equivalent to evolution. Mutation is just the beginning. It generates the variation from which natural selection generates evolutionary change.

Nevertheless, the news is filled with claims that the virus may be just a mutation away from causing a terrible flu pandemic like the one that killed 1 per cent of the world's population in 1918. Some suggest that it will be even worse because about half of the people infected with bird flu have died, compared with less than 5 per cent of those with the 1918 virus. But such conclusions are not justifiable. The evidence tells us that the 1918 pandemic was not simply due to the sudden escape of a monstrous mutant into the human population. During the early months of 1918, the influenza was similar to the influenza that has recurred each year since.

But the situation on the Western Front at the end of the First World War was far from normal. There a virus could be transmitted prolifically from people who were entirely incapacitated by infection, as they were slowly moved from overcrowded trenches. Under such unusual conditions, predator-like variants of the influenza virus could exploit the infected without paying the price usually incurred from incapacitation, constantly exposed as they were to new potential hosts. These variants then overwhelmed their milder competitors and spread globally.

This is the context in which today's H5N1 virus should be evaluated.

Everyone agrees that increased contagiousness from person to person is the critical change. H5N1 is still a bird virus and has not yet been transmitted between humans. Experts talk as though containment of a pandemic is a lost cause if it evolves such a capability. But the evidence and evolutionary principles indicate that evolution towards a runaway virus will occur gradually. Organisms that are poorly adapted to a new environment - such as H5N1 in humans - are much more likely to acquire one of the many mutations that confer marginal improvements than those rare ones that confer optimal capabilities. We can therefore expect many more instances of crudely increased contagiousness than a quantum leap to a virus well-adapted for transmission between humans. H5N1 has not yet even acquired the crude increases. And if it does it is likely to fizzle, because the feeble transmission made possible by this first step will not be sufficient to generate an average of one new infection from each existing infection. If it did, it would still have more challenges to overcome before it could cause a pandemic. It would, for example, have to generate new infections sufficiently above the usual replacement rate to compensate for human activities such as isolation and quarantine. That is what has held back severe acute respiratory syndrome.

The other evolutionary process important here is the drop in harmfulness as the virus evolves increased transmissibility. H5N1 in humans infects the delicate air sacs in the lungs. This results in the terrible damage responsible for the lethality of the disease. But to be transmitted from person to person, the virus must infect the nose and throat, which are far hardier than the air sacs.J So as the virus evolves increased contagiousness, it will simultaneously evolve great reductions in its harmfulness.

Even after it evolves the ability to infect the nose and throat, newly humanised H5N1 would need other modifications to perfect transmission from its new home in the upper respiratory tract. Natural selection would inevitably tailor these characteristics towards a close fit to the existing circumstances, all the time favouring those variants that are sufficiently mild to allow the host to move into contagious contact with susceptible individuals. In 1918, a highly contagious virus evolved increased harmfulness, allowing it to spread globally. H5N1, in contrast, is a virus that needs to evolve increased transmissibility but currently harms its human host in a way that negates its chances of transmission. It will not spread effectively until it evolves reductions in harmfulness that befit transmission of an influenza virus between humans.

All of this bodes well for the future. Of course, we need vigilant surveillance for any hot spots where the virus may be evolving increased contagiousness in humans. But to alleviate concerns over a 1918-style pandemic we need only a dose of Darwinian insight.

Paul Ewald is professor of biology at the University of Louisville, in the US.