Thirty-eight million people around the globe are infected with HIV, and 8,000 of them die each day from AIDS. Although anti-HIV drugs can extend lives, they have serious limitations, and the vast majority of infected people still do not have access to them. So it is hard to overstate the need for an AIDS vaccine that c an slow the virus’s spread and ultimately safeguard the world’s population.

Scientists in academia, government, and the pharmaceutical industry have spent the last 18 years testing three dozen different AIDS vaccines in human studies. Time and again, high hopes have given way to crushing disappointments, and the field has been roiled repeatedly by bitter disputes about the best way to move forward. If the different players worked in isolation, as private companies often do, the tensions might not matter much. But in the world of international vaccine research, there’s a constant tussle for resources and influence among government agencies, universities, drug companies, health ministries, networks of clinics, and the communities that agree to participate.

Keeping this tussle from devolving into self-defeating conflict depends on consultation and coordination among all parties. But that’s easier said than done, as was dramatically illustrated by an AIDS-vaccine trial launched in Thailand in 2004 with U.S. government backing. The study – the largest, most expensive AIDS vaccine trial in history – has turned out to be a model of bad communication, scientific disagreement, international face-saving, and clashing bureaucratic interests. Many leading AIDS researchers have publicly decried the trial as a near-certain failure and a waste of scarce resources.

The trial combines two different vaccines, and although critics don’t particularly like either of them, they dislike one more than the other – and that has become the main point of contention. The controversial vaccine, called gp120, was once the darling of the field, but it fell out of favor more than a decade ago, after disappointing test results. It survives because indefatigable corporate and government scientists have brought it back time and again, arguing that it deserves yet another study to assess its worth. General scientific opinion would have dictated otherwise.

The saga began in the summer of 1993, when a protein called gp120 enjoyed frontrunner status in a spirited race to develop an AIDS vaccine. The protein sticks out from the surface of the HIV virus like the hooks in Velcro, allowing it to latch onto and infect human cells. Biotech stars Genentech and Chiron believed that injecting the surface protein into people would stimulate their immune systems to create antibodies against it, theoretically protecting them if they contracted the virus. The U.S. National Institutes of Health (NIH) planned to bankroll efficacy trials.

Then gp120’s fate suddenly turned. In test-tube experiments reported in October 1993, antibodies taken from vaccinated people proved unable to stop HIV from infecting human cells. After consulting widely with critics of the approach, NIH decided in June 1994 to abandon plans for the multimillion-dollar efficacy trials.

The decision spelled the end of the gp120 programs at both companies, but out of the ashes of the Genentech effort rose VaxGen, which raised millions of private dollars to stage two new efficacy trials. Though no less a personage than Nobel laureate David Baltimore, who in 1998 headed NIH’s AIDS Vaccine Research Committee, expressed serious doubts about gp120’s utility, VaxGen charged that its critics put too much faith in their theories and not enough in the good old trial-and-error empiricism that had guided research on other vaccines.

As the VaxGen studies progressed, AIDS vaccine researchers moved on to the next big thing: the “prime boost.” A “prime” shot would produce killer cells, which target and destroy the cells invaded by viruses, while a second “boost” shot would trigger an antibody response.

The lead candidate for the priming shot came from Aventis Pasteur, where scientists stitched HIV genes into a harmless bird virus called canarypox. For the antibody boost, researchers turned once again to gp120. Although this struck some as illogical, no other potential antibody-boosting vaccine had yet proven its safety in hundreds of humans, and gp120’s proponents further reasoned that the two vaccines might somehow synergize with each other. The HIV Vaccine Trials Network, an NIH-funded group of academic researchers, and the U.S. Military HIV Research Program, run from the Walter Reed Army Institute of Research in Silver Spring, MD, drew up plans for two separate efficacy trials of the canarypox/gp120 strategy.

Backroom grumbling about these two new trials grew audible in January 2002, when AIDS immunologist John Moore of Cornell University’s Weill Medical College in New York City wrote a stinging commentary for Nature magazine. Moore had no quarrel with the prime-boost idea, but he questioned most every other aspect of the proposed trials. Moore charged that NIH and the military behaved like rivals and felt compelled to “shadow” each other. He further argued that it “would surely be prudent” to wait for results from the VaxGen gp120 efficacy trials then under way before launching new studies.

The next month, the HIV Vaccine Trials Network pulled the plug on its prime-boost study after small human trials of the canarypox vaccine showed that it wasn’t very good at stimulating killer cells. The military, however, stood by its decision to conduct its own $119-million prime-boost study in collaboration with Thailand’s Ministry of Public Health, Mahidol University, and the Thai Royal Army. In fact, the leading proponents of the Thai trial, including army colonel John McNeil, met quietly with outside consultants in Geneva in November 2002 to consider whether the design of the study should change if gp120 also failed as a solo agent. No, the group concluded, reasoning that the prime-boost strategy deserved to be tested anyway.

Predictably, the results of VaxGen’s efficacy trials, released in 2003, showed that gp120 conferred no more protection against HIV than a placebo. With solid evidence that their skepticism was warranted, gp120’s detractors – including Robert Gallo, whose lab first proved that HIV caused AIDS, and Neal Nathanson, the former head of the Office of AIDS Research at NIH – cranked up their objections to including the vaccine in the prime-boost study. In a Science magazine opinion piece published in January 2004, Gallo, Nathanson, Moore, and 19 other prominent AIDS researchers assailed the Thai trial and specifically complained about their exclusion from the Geneva meeting. The process “lacked input from independent immunologists and virologists who could have judged whether the trial was scientifically meritorious,” they wrote.

McNeil and three other gp120 proponents replied in another Science opinion piece that the trial couldn’t be stopped, since the United States had already made commitments to industry, the Thai government, and the people who had started to receive the vaccines. In a separate letter, an official at the Thai Ministry of Public Health added a revealing detail: the government of Thailand felt that the country would benefit from the study regardless of the results, since its scientists were gaining experience with HIV, its laboratories were being modernized, and general awareness of HIV/AIDS was rising.

And that, more or less, is where the controversy stands today. Last September, Thai and U.S. military scientists reported that they had enrolled more than 5,500 Thai volunteers in the study and that they would meet their goal of 16,000 participants within another year, with the entire study planned to conclude in 2009.

When scientific institutions seek robust criticism up front – as NIH did in 1994 – it doesn’t guarantee that everyone will walk away happy. But permitting this kind of dialogue does bow to a truth that seems to have escaped the Defense Department researchers in charge of the Thai trial: the later objections surface, the more they hurt. Had the prime-boost study’s organizers invited a few well-known, sharp-edged basic researchers to take part in the study’s design, they might still have decided not to change their basic plan. But they wouldn’t have had nearly two dozen leading scientists in the field publicly accusing them of playing ostrich.

Indeed, when the stakes are as high as they are in the search for an AIDS vaccine, it’s worth striving to keep the peace in constituent communities, even if it means inviting your enemies to the table. But that’s a battle plan that simply runs counter to the culture at the Department of Defense.