Longer STEP Follow-up Suggests Adenovirus HIV Vaccine Did Not Increase Infection Risk, May Lower Viral Set-point

		SUMMARY: Over a longer follow-up period, the apparent link between adenovirus type 5 immunity and increased risk of HIV infection in the STEP vaccine trial diminished, researchers reported at the AIDS Vaccine 2009 conference last week in Paris. In the companion Phambili trial, women who became infected with HIV despite receiving the vaccine had trend toward a lower viral load set-point than infected placebo recipients.

The STEP study evaluated an HIV vaccine candidate known as V520 (aka MRKAd5), Merck's trivalent vaccine containing 3 recombinant HIV genes (gag, pol, and nef) carried by an attenuated adenovirus type 5 vector. V520 was being studied as both a preventive and a therapeutic vaccine (i.e., both to prevent initial infection and to slow viral replication in people who became infected).

The Phase 2 study enrolled 3000 high-risk HIV negative volunteers in North and South America, the Caribbean, and Australia, including men who have sex with men and female sex workers. Participants were randomly assigned to receive 3 doses of the active vaccine or inactive placebo injections.

Merck and the National Institutes of Health discontinued the trial in September 2007 after interim data showed that V520 did not lower the risk of HIV infection, nor did it reduce viral load among individuals who became infected.

Further analysis of the data suggested that participants in the vaccine arm appeared to actually have a higher rate of HIV infection than those in the placebo arm. A subgroup analysis indicated that the increase occurred among volunteers with pre-existing immunity to adenovirus type 5.

At the 2008 Conference on Retroviruses and Opportunistic Infections, researchers reported that the risk of HIV infection was elevated among men who were not circumcised, and that circumcision status might in fact account for regional differences in response and the apparent effect of adenovirus 5 immunity. Furthermore, only 1 woman became infected with HIV, even though women made up more than one-third of the total study population.

Although the STEP trial was halted, researchers continued to monitor the enrolled participants. At last week's conference in Paris, Susan Buchbinder from the San Francisco Department of Public Health reported that the V520 vaccine did not actually enhance the risk of HIV infection, including among uncircumcised men.

"With ongoing follow-up, the trend in the wrong direction is diminishing," she said. "Either they were at risk, and that has gone away, or they were never at increased risk. It was never significant."

Results

	After the study data were unblinded, through January 2009, 12 new HIV infections occurred among women, 6 each in the vaccine and placebo groups.
	48 new infections occurred among men, 26 in the vaccine group and 22 in the placebo group (not a significant difference).
	Infection rates demonstrated "rough equivalence" in the proportion of vaccine and placebo recipients who were adenovirus 5 seropositive (54% vs 59%) and uncircumcised (50% in both groups).
	Hazard ratios (HRs) for HIV infection, calculated using combined data from before and after unblinding, were not significantly different between adenovirus 5 seropositive versus seronegative men (HR 1.7 vs 1.1; P = 0.3).
	Hazard ratios remained significantly greater among uncircumcised compared with circumcised men (HR 2.2 vs 1.1; P = 0.045).
	The early HR > 1.0 -- suggesting an elevation in HIV infection rates among vaccine recipients -- began to decline approximately 12 months after study enrollment, and reached 1.0 or less (indicating no difference in risk) beginning at approximately 20 months.
	There were no apparent overall effects of the vaccine on HIV disease progression among infected participants.
	Some vaccinated participants who became infected showed a slight decline in HIV viral load compared with unvaccinated volunteers, but this was only temporary.

San Francisco Department of Public Health, San Francisco, CA; Merck & Co., North Wales, PA; HIV Vaccine Trials Network.

Phambili Study

Phambili (HVTN 503) was a Phase 2b trial of the same Merck adenovirus type 5 HIV vaccine. This study was designed to include 3000 high-risk participants in South Africa. Administration of the vaccine was suspended after the STEP interim results were released, but follow-up of already vaccinated participants continued.

At the Paris meeting, researchers presented data from an interim analysis through the end of December 2008. A total of 801 participants at 5 sites were enrolled at the time of study suspension; only 7% had received all 3 vaccine or placebo injections, 66% had received 2 injections, and 27% had received 1 injection.

Results

	In a modified intent-to-treat analysis, 41 participants were newly infected with HIV, 24 in the vaccine group and 17 in the placebo group (not a significant difference).
	32 of these infections occurred within 1 year of initial vaccination, 18 in the vaccine group and 14 in the placebo group.
	The HIV infection rate was 4.69 per 100 person-years in the vaccine group versus 3.80 per 100 person-years in the placebo group, for a HR of 1.26 (P = 0.37, not statistically significant).
	Among participants who received 2 or 3 doses, 22 vaccine recipients and 19 placebo recipients became infected.
	Women had higher rates of HIV infection compared with men in both the vaccine and placebo groups (HR 2.8, or 2.8 times more likely to be infected).
	HIV infection rates also varied by age (HR 2.9 for age 21-30 and HR 3.7 for age 31-35, compared with age 18-20).
	Baseline adenovirus 5 antibody titer was not a predictor of HIV infection risk.
	The mean HIV viral load set-point was about 19,000 copies/mL among vaccine recipients compared with about 44,000 copies/mL among placebo recipients.
	There was a trend toward a significantly lower viral load set-point for vaccine versus placebo recipients among women (P = 0.09), but not among men.
	Overall, CD4 cell counts did not differ between the vaccine and placebo groups.

The researchers concluded that, "Vaccination with the MRKAd5 gag/pol/nef vaccine and baseline adenovirus 5 titer were not associated with increased HIV-1 acquisition in Phambili" -- a result that differed from the initially reported STEP findings, but is in agreement with the more recent data described above.

Perinatal HIV Research Unit, University of Witwatersrand, Johannesburg, South Africa; National Institute of Allergy and Infectious Diseases, NIH, Bethesda, MD; Aurum Institute for Health Research, Cape Town, South Africa; University of Cape Town, Cape Town, South Africa; Desmond Tutu HIV Centre, Cape Town, South Africa; Medunsa HIV Clinical Research Unit, Medunsa, South Africa; Centre for the AIDS Programme for Research in South Africa, Durban, South Africa; Statistical Center for HIV/AIDS Research and Prevention, FHCRC, Seattle, WA; Merck & Co, North Wales, PA; HIV Vaccine Trials Network, Fred Hutchinson Cancer Research Center, Seattle, WA.

11/3/09

References

S Buchbinder, M Robertson, and A Duerr. Clinical outcomes from the STEP study. AIDS Vaccine 2009. Paris. October 19-22, 2009. Abstract SS01-02.

G Gray, M Allen, G Churchyard, and others. 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. October 19-22, 2009. Abstract SS01-04.