AIDS vaccine trial fades in closer look at the data

WASHINGTON — The "modest" success of an HIV vaccine clinical trial in Thailand, trumpeted by researchers last month as a milestone in the AIDS pandemic, is even more modest than originally advertised.

Full details of the study, released Tuesday at a scientific meeting in Paris, show the vaccine provides no protection to people at the highest risk of HIV infection. In people at lower risk, the benefits may start to wane after a year. Furthermore, when the results of the three-year experiment are analyzed using alternative methods, the protective effect falls short of formal statistical significance.

Despite the new caveats, many AIDS researchers say the findings are still important.

"This is a modest effect at best, but I believe it has relevance and is a real effect that needs to be built upon," said Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, which paid for much of the $105 million study.

Some other observers, however, think the results have been oversold in a potentially damaging way.

"When this was rolled out a couple of weeks ago, it was terribly hyped by the investigators," said Gregg Gonsalves, an AIDS activist for nearly two decades. "Some people think you have to dangle the slimmest morsels of hope in front of the general public in order to keep them interested in an AIDS vaccine. But I think that damages the credibility of the effort."

Initial results of the RV144 trial were revealed by a news release and interview in late September. The presentation in Paris provided much more detail, as does a paper released simultaneously by the New England Journal of Medicine, where it eventually will be published.

The effort to make an HIV vaccine has been dogged by failure for two decades. The Thai trial, which employed a two-component vaccine and a total of six injections, was the first one to show protection. It was conducted principally by Thailand's Ministry of Public Health, the U.S. Army and the Royal Thai Army.

The vaccine was tested in 16,402 Thai men and women ages 18 to 30 who were randomly assigned to receive either the vaccine or placebo injections. Only 132 of the volunteers became infected.

When the analysis included everyone, there was a trend suggesting that the vaccine provided protection, but the results were not statistically significant by the definition used in most medical research.

When counting only the people who received all six injections on schedule — a condition that reduced the total sample by 30 percent — there was also a trend toward protection, although not a significant one.

However, when the researchers examined the experience of everyone, except for seven people who were discovered to have already been infected when the trial started, the results significantly favored the vaccine. It reduced the chance of becoming infected by 31 percent.

Nelson Michael, an Army physician and virologist with the U.S. Military HIV Research Program, defended the decision to present only the third method in last month's announcement, saying it best reflected a real-world use of the vaccine.

Not everyone can be expected to follow a complicated shot schedule perfectly, so it was legitimate to include people who didn't receive all six injections, he told reporters in Paris. It was also reasonable to exclude people who became HIV-positive very early in the trial and whose initial blood samples, when further analyzed, showed they were infected before the study began.

"You can't protect a person from a disease if they already have it," he said.

The vaccine was 40 percent to 50 percent effective in people who identified themselves at low or medium risk for becoming infected, and in people who lived with their sexual partners. It offered little or no protection to people at high risk because they shared needles, worked in the sex trade or had multiple partners. There was also some evidence that protection declined during the first year after vaccination.

Information in this article, originally published October 20, 2009, was corrected October 22, 2009. A previous version of this story incorrectly stated that said that the vaccine was 40 to 50 percent effective in people who identified themselves as being at high or medium risk for becoming infected. That was the success rate for those who said they were at low or medium risk.