It takes 12 times the dose of a regular winter flu shot for the nation's first vaccine against bird flu to work — and even then only about half the people who get it are protected.

Scientists viewing the cup half full say at least the vaccine is safe, and that efforts are under way to boost its effectiveness.

But the anxiously awaited results of the first human testing underscore how difficult it could be to protect a panicking population if bird flu began to spread among people any time soon.

"We have a long way to go," acknowledged Dr. Anthony Fauci, infectious disease chief at the National Institutes of Health, which funded the research being published Thursday in The New England Journal of Medicine.

The government had signaled that the vaccine had serious flaws even as it ordered $162 million worth of shots last summer.

But Wednesday's result showed the extent of the problem: The vaccine sparked a protective immune response in just 54 percent of healthy volunteers who got two mega-doses, 28 days apart.

Regular winter flu shots, in contrast, protect 75 percent to 90 percent of young healthy people, the same group that first tested the experimental bird-flu vaccine. The elderly typically fare worse; how they respond to the bird flu shots still is being analyzed.

The results weren't too surprising, said lead researcher Dr. John Treanor of the University of Rochester. Humans have never been exposed to the deadly bird-flu strain called H5N1, and it takes the immune system awhile to ramp up to fight unique types of influenza.

But the vaccine's weak effect means that the nation's stockpile is, in Fauci's words, "a stopgap measure." So far it contains enough doses for just 4 million people, far below the goal of 20 million.

Those shots would be reserved for health care providers and workers in flu vaccine factories if a human epidemic of H5N1 began in the next year or so, Fauci said — leaving most Americans to rely on old-fashioned methods such as closing schools, avoiding crowds and wearing masks in hopes of staying well.

The world's vaccine factories are now brewing regular flu shots for next winter. They would make only bird-flu vaccine if a pandemic began. But at the high doses, the maximum that could be produced would fully immunize just 75 million people — 1.25 percent of the world's population — half of whom wouldn't be adequately protected, Mayo Clinic flu specialist Dr. Gregory Poland wrote in an accompanying editorial.

Still, "my impression is we are better off having stockpiled this vaccine than none," said Dr. William Schaffner of Vanderbilt University, a member of an independent panel that closely monitored the shots' safety during this first human testing.

"My concern is the public will see (the findings) as the jar half empty," he said, adding that he preferred that it be viewed as "the first strong step in a long journey."

More than 180 people worldwide, mostly in Asia, are known to have been infected with the H5N1 virus since 2003; more than 100 of them have died. Virtually all were infected by close contact with sick poultry. But flu viruses are prone to genetic mutations, and as H5N1 is now rapidly continent-hopping via migrating birds, there is increasing fear that it may eventually become easily spread from person-to-person, sparking a global epidemic.

Scientists don't know how much of an immune response — the creation of infection-fighting antibodies — a vaccine must prompt to protect people against bird flu. So in this first human study, Treanor and colleagues tested whether the H5N1 vaccine would prompt as much antibody protection as do regular winter flu shots.

Those annual flu shots contain 15 micrograms of antigen, the key element. For the H5N1 vaccine, it took two shots that each contained 90 micrograms of antigen to spur a protective immune response in slightly over half of recipients, Treanor found.

However, 70 percent of recipients had a slightly lower immune response — and scientists couldn't say whether they might have some protection against bird flu.

"Our guess about what will be a protective response may be very conservative," Treanor cautioned.

The main side effect was pain at the site of the injection. So researchers are giving the study's 451 volunteers a third dose, to see if that spurs more protection.

More promising, vaccine manufacturers Sanofi-Pasteur and Chiron Corp. now are adding immune-enhancing compounds — called alum and MF59, respectively — to the experimental vaccine in hopes they would spark protection with doses closer to 15 micrograms, thus stretching limited supplies. Pilot studies are optimistic; Fauci said results may come in the fall.

Further complicating matters: This first H5N1 vaccine is already outdated, based on a version of the virus culled in Vietnam in 2004. Scientists now are creating a vaccine based on a slightly different Indonesian version that emerged last year; they don't yet know how much protection the older vaccine would spur against the newer virus.