A clinical test of Merck’s adenovirus-vectored subtype B HIV vaccine that ran alongside the halted STEP vaccine trial has reported that women who became infected despite vaccination had a lower viral load set point than women who became infected in the placebo group, Dr Glenda Gray reported on Tuesday at the AIDS Vaccine 2009 conference in Paris.
However, the numbers involved were small, and the difference in viral load set points was not matched by a sustained difference in CD4 cell counts.
The Phambili study was a companion to the STEP trial, a large international study of an HIV vaccine designed to produce cell-mediated immune responses that might protect against infection, or in those infected despite vaccination, slow disease progression, therefore providing a clinical benefit.
The STEP study was halted in 2007 after the trial’s Data and Safety Monitoring Board concluded that no matter how long the study lasted, it would never demonstrate a difference between vaccine and placebo groups (a statistical test called futility).
Subsequent tests showed that uncircumcised men were at higher risk of HIV acquisition if they received the vaccine compared to the placebo group, as were individuals in the vaccine group who had higher levels of adenovirus-5 antibodies due to previous exposure.
The Phambili study tested the vaccine in high-risk men and women in South Africa, and was intended as a test of the vaccine’s potential for cross-clade protection. The Merck vaccine was based on HIV-1 subtype B, but the dominant subtype of HIV circulating in South Africa is subtype C.
The study was designed to recruit 3000 high-risk men and women at five sites across South Africa, but was halted when the decision was reached to close the STEP study. By the this point the study had recruited 801 participants.
Dr Glenda Gray reported interim results to December 2008. At this point 90% of participants were still undergoing follow-up, and 80% of participants had received at least two vaccinations before study vaccinations were halted.
By December 2008 a modified intent-to-treat analysis that included all participants who were randomised showed that 24 infections had occurred in the vaccine recipients and 17 in the placebo group, a non-significant difference. A per-protocol analysis, which included all participants who had received at least two vaccinations, showed 22 infections in the vaccine group and 19 in the placebo group.
The infection rate per 100 person-years of follow-up was 4.69 in the vaccine group and 3.8 in the placebo group, with a hazard ratio of 1.26 (95% confidence interval 0.76 to 2.10, p=0.37).
Although none of the differences in infections or infection rates between the two arms was statistically significant, there were always more infections in the vaccine arm, no matter how the data were compared.
Dr Nelson Michael of the US Military HIV Research Program noted that the infection rates seen in the Phambili study were at least tenfold higher than those seen in the Thai RV144 vaccine study, conducted predominantly in a population at low and medium risk for HIV infection, and the rates reflect the very high HIV prevalence at trial sites in South Africa.
People in their early 30s (31 to 35) were at higher risk of HIV acquisition: compared to 18 to 20 year olds. They were 3.7 times more likely to become infected (HR 3.7, 95% CI 1.2 to 11.5). 21 to 30-year-olds were also at higher risk (HR 2.9, 95% CI 1.1 to 7.5).
Women were significantly more likely to become infected in both vaccine and placebo group. After adjustment for age the hazard ratio was 2.8 (95% confidence interval), and there was no discernable effect of prior adenovirus exposure or circumcision status on the risk of HIV acquisition, unlike in the STEP study.
Women in the vaccine group who became infected showed a trend towards a lower viral load set point than women in the placebo group (-0.57 log lower) in the modified intent-to-treat analysis (p=0.09), but the same trend was not seen in men. An early statistically significant trend at three months towards a lower number of individuals in the vaccine group with a CD4 count below 350 cells/mm3 had disappeared by month 12.
Overall, the viral load and CD4 trends showed no significant difference, said Dr Gray, but the viral load setpoint effect in women requires further study.
Even through the vaccine was based on subtype B HIV-1 sequences, trial participants generated strong T-cell responses to HIV-1 subtype sequences as measured by ELISPOT assay.
Immunogenicity data from the first 186 participants to be vaccinated, measured at week 8 showed that among those uninfected at week 12, 77% produced a response to at least one subtype C epitope, and these were strongly correlated with the development of subtype B ELISPOT responses. Responses tended to be better in those with no pre-existing adenovirus 5 immunity.
Gray G et al. Interim efficacy analysis of HVTN 503 / Phambili: a phase IIb test of concept trial of the MRK Ad5 HIV-1 gag/pol/nef vaccine conducted in HIV-1 uninfected adults in South Africa. AIDS Vaccine 2009, Paris, abstract SS01-04, 2009.