A close examination of the genetic evolution of the three major influenza epidemics of the 20th century concludes that all of the viruses involved evolved slowly, through interspecies genetic exchange, and that genes from the catastrophic 1918 pandemic may have been circulating as many as seven years earlier. If true, this means that widespread genetic surveillance methods should have ample time to detect the next pandemic strain, and possibly even vaccinate against it before it gets out of control.

Prior research suggested that the 1918 influenza strain was the result of an avian virus introduced into humans just before the epidemic began. But the latest study, published today in the Proceedings of the National Academy of Sciences, suggests that all three influenza pandemics--1918, 1957, and 1968--were the result of stepwise genetic integrations of both avian and mammalian genes over a number of years, ultimately creating the more virulent virus strains.

And although the research was done before the emergence of the current H1N1 "swine flu" strain, the scientists' conclusions are relevant, showing that the current virus follows the same historical pattern. For each pandemic, "our results argued that there was at least one intermediate host that was most likely to be pigs, and that they're involved in the emergence of these pandemic strains," says Gavin Smith, the paper's lead author and a viral-evolution researcher at the State Key Laboratory of Emerging Infectious Diseases, at the University of Hong Kong.

The researchers collected all available genetic sequences of the influenza virus--human, bird, and pig variants--then plugged the data into a computer program that uses genetic information to build evolutionary trees, dating species' divergence back to their most recent common ancestor. But there are no known precursor viruses to the 1918 strain, so the computational results can only infer the time of interspecies transmission, based on known patterns of genetic evolution. The genetic data itself was derived from virus strains that have evolved since 1918.

Such studies have only become possible in the past few years, with the advancement of computational techniques that can incorporate known rates of various species' evolution--techniques that are proving to be quite accurate when tested against known relationships. But the results are still, as Smith notes, "all just inference," working backward from known relationships and based on estimated dates.

According to the virus's updated family tree, the 1918 strain was not newly minted but actually a slightly modified version of a mild flu strain already in the human population. In fact, according to the new analysis, some genes of the virus may have been circulating as early as 1911. "It was certainly different in terms of severity of the actual pandemic," Smith says. "But our results show that, in terms of how the virus emerged, it looks like much the same mechanism of the 1957 and 1968 pandemics, where the virus gets introduced into the human population over a period of time and reassorts with the previous human strain."