Nor did the shots reduce the amount of the AIDS virus in the blood when people who'd been vaccinated later became infected, the National Institutes of Health said.
"It's disappointing," said Dr. Anthony Fauci, head of NIH's National Institute of Allergy and Infectious Diseases. But, "there was important information gained from this" study that will help determine what to try next.
The study had enrolled 2,504 volunteers, mostly gay men, in 19 cities since 2009. Half received dummy shots, and half received a two-part experimental vaccine developed by the NIH. All were provided free condoms and given extensive counseling about the risks for HIV.
It's a strategy known as "prime-boost." A DNA-based vaccine made with genetically engineered HIV material is given to prime the immune system to attack the AIDS virus. Then a different vaccine, encasing the same material inside a shell made of a disabled cold virus, acts as a booster shot to strengthen that response. Neither vaccine could cause HIV.
The idea: Train immune cells known as T cells to spot and attack the very earliest HIV-infected cells in someone's body. The hope was that the vaccine could either prevent HIV infection, or help those infected anyway to fight it.
A safety review this week found that slightly more study participants who had received the vaccine later became infected with HIV. It's not clear why. But the difference wasn't statistically significant, meaning it may be due to chance. Overall, there were 41 HIV infections in the vaccinated group and 30 among placebo recipients. When researchers examined only participants diagnosed after being in the study for at least 28 weeks - long enough for the shots to have done their job - there were 27 HIV infections among the vaccinated and 21 among the placebo recipients.
The NIH said Thursday that it is stopping vaccinations in the study, known as HVTN 505, but that researchers will continue to study the volunteers' health.
Josh Robbins, 30, of Nashville, Tenn., is among the participants who became infected. He said he's glad he was in the study, because its close monitoring meant he was diagnosed and treated much sooner than most people - and he's feeling great - and because the findings help science.
"We've got to keep moving forward," Robbins said. The study "certainly can lead us down a new direction to hopefully find something that might work."
Multiple attempts at creating an AIDS vaccine have failed over the years. A 2009 study in Thailand is the only one ever to show a modest success, using a somewhat different prime-boost approach. Newer research suggests another approach - to try creating powerful antibodies that could work a step earlier than the T-cell attack, before HIV gets inside the first cell.
Both approaches need continued research funding, said Mitchell Warren of the international AIDS Vaccine Advocacy Coalition.
"Clearly an AIDS vaccine remains critical," he said.