VaxGen's AidsVax: "shot works in blacks" claim hung on one Chicago heroin abuser

The failure of AidsVax to prevent infection with HIV - in clinical trial results published in 2003 - triggered an intense debate about the controversial product and its manufacturer, VaxGen Inc of Brisbane, California. Mail to this website, maintained by Brian Deer, shows that existing material on a VaxGen-AidsVax index is read by significant numbers. This page seeks to further inform the discussion

Below is material from two slides released by VaxGen Inc on February 24 as part the company's initial results presentation from the phase III AidsVax trial. Both slides - but particularly the second, relating to black, Asian and other subjects - suggest a sudden rise in the number of infections towards the end of the trial (in the first slide apparently after the trial) among subjects who received placebo.

Although the vaccine showed no statistically significant efficacy (<4%), in the full cohort of 5,009 subjects, the company claimed that it appeared to have astounding efficacy among blacks (>78%), Asians (>68%) and "others" (not white or Hispanic) (>46%), who together totalled 498 subjects. If the numbers of infections among these minorities were statistically significant, as VaxGen claimed, why would they temporally bunch in the placebo group? Critics say that, among other curiosities in these graphs, the bunching is suggestive of some confounding factor.

It was later revealed that three and possibly four of the infections among placebo recipients were black women who, extraordinarily, contracted HIV from the same man in Chicago - and that the company's widely-publicised claim of efficacy among blacks rested on this wholly bizarre circumstance. The Chicago trial site attended by the women recruited one third of the trial's females, and most were African-American. The site's co-ordinator, Parrie Graham, was reported as saying that the male was involved in the local drug scene.

Discussion of the company's claims, such as on the Yahoo message board, further raised the question of whether the category "black/Asian/other" even represented any credible association between such ethnic classifications, or whether it had been devised by VaxGen purely for the purpose of presenting the statistics in the most commercially-favorable light.

Three versions of AidsVax trial results, released in February and March 2003, are also available.