The animals were protected even though they did not show the usual antibody response -- a measure of immune system reaction often used to gauge vaccine effectiveness.

The findings suggest it may be possible to stockpile a vaccine ahead of a pandemic of H5N1 influenza, the researchers report in this week's issue of The Journal of Infectious Diseases, something that experts believed was not possible.

"Possibly the greatest significance of Govorkova et al.'s study is the demonstration of a significant cross-strain protective effect even in the presence of minimal antibody levels," Alan Hampson of the Australian Influenza Specialist Group at Australia's Monash University wrote in a commentary.

H5N1 avian flu outbreaks have been confirmed in more than 48 countries and territories, according to the World Organization for Animal Health.

The virus almost exclusively infects birds but it has killed 128 people in nine countries and has infected at least 225.

Experts say a pandemic of some kind of influenza is inevitable and that H5N1 looks closer than any other virus to causing such a global wave of disease.

A vaccine would provide the best protection. But flu vaccine technology is slow and unwieldy and a new vaccine has to be formulated every year to match the current circulating strains.

Vaccine experts fear that they would have to wait until H5N1 changes into a human pandemic strain before they can make a vaccine against it. By then it could have infected millions.

LABORATORY SHORTCUT

But Govorkova's team recreated the H5N1 virus in their lab, using a strain from Hong Kong. Unlike some other experimental H5N1 viruses, which use bits of its DNA, they made a whole virus.

It protected the ferrets against strains of the same virus.

"To determine the extent of cross-protection induced by (our) vaccine, we challenged vaccinated ferrets with H5N1 viruses that were antigenically and genetically distinct from the vaccine strain," they wrote.

"All 4 unvaccinated ferrets inoculated with (a different H5N1) virus survived but showed signs of disease," they added. These included fever.

Yet the ferrets did not produce large amounts of antibodies, they found. "H5N1 vaccines may stimulate an immune response that is more cross-protective than what might be predicted by (lab tests) and, thus, hold potential for being stockpiled as 'initial' pandemic vaccines," the researchers concluded.

This may be difficult to test, as people cannot be deliberately infected with a virus to see whether a vaccine works. Usually, scientists look at antibodies in the blood to see if their immune systems have been primed by the vaccine.