A vaccine to protect people from the H5N1 bird flu virus has been shown to be partially effective, but is unlikely to be suitable for widespread use in preparing for a pandemic.
The results, published in this week's issue of the New England Journal of Medicine, are from a US trial that gave more than 400 participants different doses of the vaccine.
It triggered an immune response in half of those given two 90-microgram doses, suggesting that they would be protected from severe disease if H5N1 sparked a pandemic.
But this dose contains 12 times the amount of viral particles (or 'antigen') contained in the similarly-designed US seasonal flu vaccine, so it would be a challenge to produce enough of the H5N1 vaccine to protect large populations.
The results are disappointing, says John Wood of the UK National Institute for Biological Standards and Control, which works as an intermediary between the World Health Organization and pharmaceutical companies developing vaccines.
Wood says that using this vaccine to fight a pandemic would be "impractical" because of the large amount of antigen needed to mass-produce it.
Flu experts unanimously agree that the global capacity to produce vaccines is already insufficient.
John Treanor of the University of Rochester, who led the trial, says the next step is to add a chemical to the vaccine to increase its efficiency.
Such chemicals, known as adjuvants, are commonly used in vaccines to reduce the amount of antigen needed to trigger an immune reaction.
Last year, companies in Australia and France carried out trials of H5N1 vaccines containing alum, a commonly used adjuvant. Their results suggest that this could help increase production capacity.
In the Australian trial, researchers found that two shots of a vaccine containing alum and 15 micrograms of antigen triggered a sufficient immune response in half of those who received it.
China, Thailand and Vietnam are also testing H5N1 vaccines.
Reference: New England Journal of Medicine 354,1343 (2006)