Full data from Thai HIV vaccine study show no significant protective effect

BMJ 2009; 339 doi: (Published 23 October 2009) Cite this as: BMJ 2009;339:b4370
  1. Bob Roehr
  1. 1Washington, DC

    Disclosure of the full data from the trial of an HIV vaccine in Thailand shows that although there was a trend towards the prevention of HIV among vaccine recipients this did not reach statistical significance.

    The study sparked controversy in September when the authors released incomplete trial results, in which one type of analysis showed the vaccine reduced infection rate by a third (BMJ 2009;339:b3963, doi:10.1136/bmj.b3963). Critics said that they should have waited until complete data were available for release.

    The additional data were presented at the AIDS Vaccine 2009 conference in Paris on 20 October and simultaneously published in the New England Journal of Medicine (2009 Oct 20, doi:10.1056/NEJMoa0908492).

    The study involved 16 402 participants aged 18 to 30, who were broadly representative of the local population. They were randomised to receive a series of six shots of either placebo or two different vaccines administered on a strict schedule over six months and followed for three years.

    Among those who received the vaccine, 51 became infected with HIV compared with 74 who received placebo. There was no suggestion that the vaccine provided clinical benefit, such as reducing the level of viraemia or the viral setpoint (steady viral load), to people who became infected

    The two vaccines used in the study did not show efficacy when used separately. Some vaccinologists argued that there was little chance that they would show a better outcome when used together. They had tried to stop the $105m (£63m; €70m) study several years ago.

    On 24 September the study authors released a modified intention to treat analysis, which showed a modest protective effect of 31.2%, which barely reached statistical significance (P=0.04). This was not part of the original protocol design but was added six months before the data were analysed.

    The trial protocol, which was presented at the conference, called for intention to treat and per protocol analyses. This showed more modest protection (26.4% and 26.2%) but did not achieve statistical significance (P=0.08 and P=0.16).

    “Although the merits of each type of analysis can be debated, all three yielded a possible, albeit modest, effect of the vaccine in preventing HIV infection,” wrote Raphael Dolin, a researcher from Harvard medical school, who was not affiliated with the trial, in an accompanying editorial in the New England Journal of Medicine (2009 Oct 20, doi:10.1056/NEJMe0909972).

    “What we have is preliminary, it is hypothesis generating,” said Jerome Kim, a senior US army researcher involved with the trial. “They point us in a direction that maybe we should be looking.”

    “Having a positive signal [on the possibility of a protective vaccine], even if it is weak and we are still debating whether it is real or not, is important,” said Nicole Frahm, from the US HIV Vaccine Trials Network, based in Seattle at the University of Washington. “It is important that there is hope; that has been missing for two years since the STEP trial that failed.”

    The research team is gathering suggestions as to how best to use the remaining blood samples from the study to better understand the immune responses and possible protection generated by the course of vaccination.


    Cite this as: BMJ 2009;339:b4370

    View Abstract