Scientists said on Wednesday they had made a breakthrough in the race to develop a drug for the H5N1 bird flu virus if it mutates into a form that can jump from human to human.

But they warned that it could take five years or longer to convert their discovery of a potential weak point in the N1 part of the virus into an effective drug.

So far the 238 cases of human infection have been from direct contact with infected birds, and scientists have said there is no evidence the virus is mutating toward making the leap between humans, though this could happen at any time.

Nearly 60 percent of those infected have died, and the best known drugs to tackle H5N1 infection in humans are oseltamivir known as Tamiflu and zanamivir known as Relenza, both originally developed to fight other forms of human ´flu.

Now a team of scientists lead by John Skehel of London´s National Institute of Medical Research say they have found a cavity in the N1 or neuraminidase part of the H5N1 virus that could be exploited as a potential weak point.

"The hope is that any new information like this which shows something specific for the N1 neuraminidase will be able to be used to develop a drug against the H5N1 virus," he told Reuters.

The team presented their detailed findings in the science journal Nature, concluding: "Our analysis suggests that it might be possible to exploit the size and location of the group-1 cavity to develop new anti-influenza drugs."

Any new drug is still years away, Skehel said. "This is not just round the corner. It could easily be more than five years to develop a new drug," he said.

Some countries have begun stockpiling Tamiflu and Relenza, but because the H5N1 virus mutates regularly it is not known how effective they would be against an H5N1 human pandemic.

"The Achilles heel of the neuraminidases has already been identified in the development of Tamiflu and Relenza. They are good drugs and they work," Skehel said.

"What this potentially will do is make the use of those drugs better by preventing the development of resistance."

The idea is that an array of drugs -- some specific to H5N1 and some like Tamiflu and Relenza to other N variants but effective against N1 -- would not stop the virus mutating but would stop it developing drug resistance, he said.

"It is a race. You have got to try to keep ahead of variation -- and in the case of H5N1 particularly the emergence of transmission from human to human," Skehel said.

"The hope would be that you would develop drugs which did not respond to the mutations that might come up and make neurominidases resistant to Tamiflu and Relenza," he added.