NEJM Offers Sobering Peer Review of Thai HIV Vaccine Study

One month ago, results of the mammoth HIV-vaccine trial in Thailand were called "promising" by more than one media outlet. But peer review of the study, published yesterday in the NEJM, shows that the outcome was largely a statistical bust. Only a modified intent-to-treat (ITT) analysis, which excluded 7 individuals who were discovered to have HIV infection at baseline, showed a statistical benefit with the vaccine (reducing the risk of HIV infection by ~31%).

Nevertheless, the general sentiment: One tweaked outcome is better than nothing—at least when it comes to HIV prevention.

Comprehensive results of the trial, including those for the primary endpoints of HIV infection and early HIV viremia, are tabulated here. The ITT analysis accounts for all subjects who underwent randomization; the per-protocol analysis includes subjects who received the vaccination series within the defined time period and who were not found to be infected with HIV at baseline.

Outcome	ITT (N = 16,402)	Per-Protocol (n = 12,452)	Modified ITT (n = 16,395)
Vaccine efficacy, %	26.4 (P = .08)	26.2 (P = .16)	31.2 (P = .04)
HIV-1 infection	132	86	125
Vaccine	56	36	51
Placebo	76	50	74
Mean viral load, log10/mL
Vaccine	4.36*	4.24	4.30
Placebo	4.21*	4.19	4.20
Mean early postinfection CD4+T cells/mL
Vaccine	541	572	555
Placebo	568	532	568

* P = .09, vaccine vs placebo. 

The trial was randomized, double-blind, and placebo controlled. Four "priming" injections of a recombinant canarypox vector vaccine (Alvac-HIV; sanofi-pasteur) were followed by 2 booster shots of a recombinant glycoprotein-120 subunit vaccine (AIDSVAX; Global Solutions for Infectious Diseases) in adults aged 18-30 years. The prime-boost vaccination series was chosen on the basis of previous trials (Belshe et al, 1994; Russell et al, 2007; and AIDS Vaccine Evaluation Group 022 Protocol Team, 2001). Subjects were primarily at risk of HIV infection through heterosexual contact. After the 6-month vaccination series, enrollees were monitored for HIV infection and viremia every 6 months for 3 years.

Data regarding the potentially protective benefit of male circumcision were evidently not collected, but expect a good number of post-hoc subgroup analyses.

HIV = human immunodeficiency virus.