Can science save us from the second wave?

• 06 May 2009 by Debora MacKenzie
• For similar stories, visit the Epidemics and Pandemics Topic Guide

Read more: Special report on swine flu

IS THE world ready for a killer, autumn wave of swine flu? Infections may be winding down in Mexico and causing fewer deaths worldwide than feared, but what is now keeping health officials awake at night is the possibility that a more virulent version will come roaring back and kill millions.

Flu pandemics historically come in waves, often getting worse as they go. The deadly one of 1918, which was also the last H1N1 pandemic, did just that. "So did pandemics in 1890, 1847, 1781 and others," says Lone Simonsen of George Washington University in Washington DC, who has studied the progress of the 1918 pandemic.

The 1918 flu started with a mild wave in March, followed by a deadly second wave later in the year. For the 2009 virus to follow the same path, two things need to happen: the virus has to spread readily enough in humans that it does not fizzle out, and it needs to mutate to a nastier form capable of killing more people.

It seems unlikely that this virus will simply fizzle out, as it has already managed to persist for months. Mike Worobey of the University of Arizona at Tucson has analysed the genetic divergence of virus samples from different cases to estimate their most recent common ancestor, a minimum estimate of when the virus jumped to people. He says that it probably appeared sometime between June and November last year, and at the end of November at the latest.

Assuming it stays, will it get nastier? There are no signs the virus is undergoing unusually rapid evolution in humans, Worobey says, which is good news, although that may yet change.

"You can't predict what this virus will do," he says. Hopes that we might use our knowledge about the 1918 virus to predict when this strain might turn really dangerous were recently dashed. Jeffrey Taubenberger of the US National Institutes of Health, who helped reconstruct the 1918 flu virus, found that although the 1918 virus seemed to need certain mutations when it adapted to humans, those mutations did not signal danger in other flu viruses (Journal of Virology, DOI: 10.1128/jvi.02565-08).

One difference between 1918 and now is the possibility of a vaccine. In theory, if this mild outbreak of H1N1 is followed by a lethal wave later in the year, we could make vaccines against what is circulating now and hope they still work against whatever the virus turns into.

Can we make enough vaccine in time? Production of ordinary flu vaccine for the next northern winter is well under way, and the plan, says the World Health Organization, is to finish this run and then turn the factories over to H1N1 vaccine, which could happen as early as June.

However, an unpublished study of global vaccine manufacturing capacity commissioned by the International Federation of Pharmaceutical Manufacturers and Associations suggests this isn't enough time. If production begins in June under the best possible conditions, we could have 190 million doses by September, jumping to a billion by October, not nearly enough for the world's people.

Another option is antiviral drugs, but supplies are limited and the virus is expected to develop resistance to Tamiflu (see "The problem with antivirals").

There were no drugs or vaccines in 1918, but two things did work in people's favour and may offer some hope now. One was the protective effect of the first, mild wave. "Where there were more cases during the first wave, fewer people died in the autumn," says Simonsen, who has analysed worldwide health records for 1918. "On average you had 70 per cent protection if you had flu in spring," she says. If the current virus, in its apparently mild form, continues to spread, the same could happen this time round. But while infection now may protect some people later, it is still not a good idea to catch it as even the current H1N1 strain can be deadly. However, the fact that the spring virus in 1918 conferred immunity to the deadlier later version bodes well for the efficacy of a swine flu vaccine.

In 1918, where there were more cases during the first, mild wave of flu, fewer people died in the autumn

The other lesson from 1918 is the potentially powerful impact of "social distancing" once a virus has turned really nasty. Measures such as isolating patients and their contacts, closing schools, banning large gatherings, closing some businesses and staggering working hours to thin out rush hour kept infection levels down in many places, says Howard Markel, a medical historian at the University of Michigan at Ann Arbor, who analysed how 60 cities across the US dealt with the virus in 1918 (see graphs).

"Cities that applied several measures, early and long enough, delayed the peak in cases and actually cut the number of deaths," says Markel. "That surprised us." St Louis, which implemented these measures most successfully, had half the deaths of Pittsburgh, the least successful. Denver called its measures off, reapplied them as case numbers rebounded and saw infection levels fall once more (The Journal of the American Medical Association, DOI: 10.1001/jama.298.6.6440).

How can we apply this to swine flu? At the very least, says Markel, who is advising US authorities on how to react, if the virus does come back as an autumn monster, good social distancing might buy time until vaccines can be made.

He stresses the importance of such planning: "This H1N1 may not turn into 1918, but we need to be ready," he says. "I've never had a patient who thanked me for not preparing."

Read more: Special report on swine flu