The STEP study

The third large scale efficacy trial, after the two AIDSVAX trials and before RV144, was the STEP study¹ (HVTN 502). This was an international multicentre, randomised, double-blind, placebo-controlled Phase II test-of-concept clinical trial. The trial enrolled 3,000 HIV-negative volunteers from diverse populations at high risk of HIV infection, including men who have sex with men and female sex workers, between 18 and 45 years of age.

The vaccine candidate, the adenovirus Ad5 vaccine, developed by Merck, was a mixture of three components, each made with a weakened version of a common virus that serves as a carrier, or delivery vector, containing three synthetically produced HIV genes known as gag, pol and nef.

The first volunteer enrolled in the study in December 2004, and enrolment was completed in March 2007.

The study evaluated two primary efficacy endpoints: whether the vaccine prevented HIV infection and whether the vaccine reduced the amount of virus in those who developed infection. As planned, an interim efficacy analysis was conducted in the approximately 1500 volunteers expected to have the best response to the vaccine because they had low levels of pre-existing immunity to adenovirus 5.

The trial was stopped in September 2007 when an independent panel of experts (the Data Safety Monitoring Board or DSMB) reviewed safety data and results of the interim efficacy analysis, and recommended that vaccination be discontinued because the trial would not meet its efficacy endpoints.

Hopes had been high that the STEP trial would produce a significant positive result, so this result was a considerable disappointment. The size of the trial had been doubled in August 2005 after CD8 responses seen in volunteers appeared to be promisingly high and relatively independent of the degree of pre-existing immunity to the adenovirus 5 vector. This latter consideration was particularly heartening as it had been thought that pre-existing immunity to the vector would be the biggest limiting factor on the success of the vaccine.² The trial protocol was therefore relaxed to take in a second wave of recruits with higher titres of anti-Ad5 antibodies.

For example, we reported from the AIDS Vaccine Conference in September 2006 that “the gloom that has dominated the AIDS vaccine field since the pessimistic AIDS Vaccine ’04 conference in Lausanne began to lift further last week with the confident statement by a leading researcher that 2006 has been ‘a vintage year for vaccine development.’” After years of fruitless attempts to develop vaccines that stimulated broadly protective neutralising antibodies which prevent HIV infection, vaccine researchers had begun to concentrate on vaccines that would stimulate T-cell responses – and thought they were seeing some very promising results.

We commented at the time: “These promising results suggest that the success of the Merck adenovirus-based HIV vaccine is not a flash in the pan, and that the strategy of using a strongly immunogenic vector to stimulate T-cell responses is working. What’s less clear is what the end result of vaccination will be: prevention of infection or control of HIV disease progression? We should know by 2008 or 2009 whether this vaccine has some protective effect against HIV transmission, at which point it could move into much larger Phase III trials required for licensing.”

However, in a warning that was remembered a year later, Jerald Sadoff of the Aereas Global TB Vaccine Foundation told the conference that determining which immune responses corresponded with true immune protection was still a puzzle for researchers. While correlates of protection help a lot in proving that a vaccine works, he said, “often you don’t have your correlate until you prove it works – you assume it works until you prove it.”

There was shock then, when the HIV vaccine community heard of the trial closure.

STEP evaluated two primary efficacy endpoints: whether the vaccine prevented HIV infection and whether it reduced the amount of virus in those who developed infection. As planned, an interim efficacy analysis was conducted in the approximately 1500 volunteers expected to have the best response to the vaccine because they had low levels of pre-existing immunity to adenovirus 5.

Study volunteers were followed for a mean of 13 months. Overall adverse event rates were generally similar among the two groups, except for a higher rate of local injection-site related reactions in the vaccine group.

The vaccine did not prevent infection: in the 741 volunteers who received at least one dose of the three-dose vaccine series, 24 cases of HIV infection were observed and 21 cases of HIV infection were observed in the 762 participants in the placebo group.

In the subgroup who had received at least two vaccinations and who were HIV-negative for at least the first twelve weeks of the trial, 19 cases of HIV infection were observed in the 672 volunteers who received vaccine and 11 cases were observed in the 691 volunteers who received placebo.

These were interim figures: later on, when all figures were collected, it was found that there were 49 infections in patients receiving the vaccine and 33 in those receiving placebo. This 48% higher rate of infection in vaccine recipients was not statistically significant and could have been a random result.

The result was additionally disappointing in that the vaccine did not reduce the amount of virus in the bloodstream of those who became infected; HIV RNA levels approximately eight to twelve weeks after diagnosis of infection were similar in the vaccine and the placebo arms. The geometric means of the HIV RNA levels in the blood of infected individuals, the standard measure of ongoing HIV replication, were approximately 40,000 copies/ml in the vaccine group and approximately 37,000 copies/ml in the placebo group.

The finding that there was no difference whatsoever in the viral loads of seroconverters who took vaccine against those who took placebo was not only disappointing, it was an indication that some of the assumptions and methodology behind the trial and indeed all trials of CD8-stimulating vaccines might need revising. It had been hoped that, at the very least, the CD8 response created by the vaccine – a real and measurable effect that had already been seen in previous studies – would moderate the course of infection. The fact that it did not do so caused researchers to ask whether the concept of a T-cell-based HIV-1 vaccine was clinically not valid, or whether the lack of efficacy was specific to the Ad5 vaccine candidate.

Worse was to come, however. A hastily conducted post-hoc analysis of the trial was done, which reported at the HIV Vaccines Trials Network Conference in Seattle in November 2007, and was then published in The Lancet in 2008.¹ This uncovered a more alarming possibility: the vaccine, though it could not possibly have caused HIV infection in itself, may have made some participants more vulnerable to HIV.

Overall, the higher rate of HIV infection in vaccine recipients was not statistically significant. But further analysis suggested that the rate in a subgroup of male volunteers were more likely to be real than random.

There was only one infection in female volunteers during the trial, though there were twelve (six in each arm) during the two-year follow-up period. While this looks like a startling finding at first, it was not highlighted in the analysis, and reflects three facts; firstly, although 40% of all trial participants were women, they lagged behind men in trial recruitment and did not have so much time to become infected with HIV; secondly, HIV prevalence amongst heterosexual male contacts of female trial volunteers was considerably lower than it was in the partners of gay male volunteers; thirdly, anal sex transmits HIV more easily. The HIV infection figures below are for men only.

An apparently slight excess of HIV infection in those receiving vaccine started looking more significant when the results were stratified according to the level of pre-existing immunity volunteers had to the Ad5 virus.

Originally, people with high levels of pre-existing immunity to the Ad5 adenovirus had been excluded from the STEP trial, but, as we said above, a decision to double the trial’s size had been taken in August 2005 when reasonable immune responses were observed in this group. The trial had reached its target of 3000 volunteers by March 2007, and the analyses that follow are based on 1503 male volunteers who received at least one shot of vaccine or placebo and 1363 male volunteers who received at least two shots (the full schedule was three injections at zero, one and six months).

Volunteers were stratified into four groups according to Ad5 antibody levels in their blood:

• In those with the lowest antibody levels, 4% of volunteers caught HIV regardless of whether they had vaccine or placebo.
• In those with medium-low levels, 4.4% of those given vaccine caught HIV but only 2.2% given placebo. In those with medium-high levels, 6.1% given vaccine caught HIV and 3.0% given placebo, so twice as many given the vaccine caught HIV in both of these groups.
• In those with the highest levels of Ad5 antibodies, 4.4% given vaccine versus 1.2% given placebo became infected. This amounts to 3.5 times as many, although in this case numbers were very small (seven on vaccine and two on placebo).

In all cases the confidence intervals overlapped, and no result in any category was near statistical significance. But the direction of change was consistent.

The classical definition of ‘statistical significance’ is that there is a less than one-in-20 probability that the effects observed could be due to chance. The round table heard that there was a one-in-17 probability that the higher rates of HIV infection seen in male volunteers who had high levels of pre-existing Ad5 immunity could have been a chance effect, as opposed to an effect of the vaccine.

One possibility to be considered, however, is that high Ad5 immunity conferred some kind of pre-existing resistance to HIV infection in the placebo group – which was undermined by the vaccine. It could mean that there is some characteristic in both the vaccine and the placebo groups that may have to be taken into account, and this would then create what is called a “double sided” possibility – which changes the probability that observed effects are due to chance.

This would mean that there was one chance in 35 that the difference in infection rates observed in male volunteers with high Ad5 immunity was a random effect – well within the boundary of statistical significance.

What differences were there between the volunteer groups? In line with this being a randomised double-blind trial, there were no behavioural differences between those receiving placebo and those receiving vaccine.

There were significant differences between volunteers with low and high levels of Ad5 immunity. People with higher levels of Ad5 immunity, recruited in the second wave of enrolment, were twice as likely to be female, nearly half as likely to be white and gay, and somewhat younger. We know there is already a gender bias in the infections, but was this due to gender, because more women were black, or for some other reason?

The men with high pre-existing immunity to Ad5 were also half as likely to be circumcised. This was a potentially significant finding. Nearly two-thirds of the men in the trial (64%) were circumcised. Excluding men with uncertain circumcision status, there were 26 infections each in circumcised vaccine and placebo recipients, but in uncircumcised recipients only six receiving placebo caught HIV versus 26 receiving vaccine – the latter figure representing a fourfold greater risk of HIV acquisition for vaccine over placebo.

More information on the STEP trial emerged at the CROI conference in February 2008.³ Further data were presented from the whole group suggesting that the Ad5 vaccine specifically increased the vulnerability of uncircumcised men to infection through insertive anal sex.

Over the whole trial group, the difference in infection rates between vaccine and placebo recipients was marginally statistically significant, with a p-value of 0.044 if the effect observed was caused by the vaccine alone (so-called ‘one-tailed’ probability), and a p-value of 0.077 if the effect was caused by some combination of a direct vaccine effect and some kind of protective characteristic of the placebo group that was removed (‘two-tailed’ probability).

In the group that had high levels of adenovirus immunity, however, there were 21 infections in vaccine recipients and only nine in placebo recipients, and this was statistically significant (p = 0.02 one-tailed, 0.029 two-tailed).

Merck’s Mark Robertson, presenting preliminary data on immune responses in the trial,⁴ showed that volunteers with high Ad5 immunity levels had higher levels of generalised CD4 cell activation. Susan Buchbinder of UCSF said that the incidence of HIV infection, at about 4 to 5% a year, was the same in vaccine recipients regardless of their Ad5 immunity level, but the incidence of infection was lower in placebo recipients who had higher levels of Ad5 immunity, ranging from 4% a year in those with the lowest level to 1.2% a year in those with the highest level. The vaccine may therefore somehow have removed a protective effect of Ad5 immunity.

Both in a univariate model and in a multivariate model that progressively eliminated other possible confounding factors, high Ad5 immunity conferred a threefold higher risk of infection in vaccine recipients than in placebo recipients, compared to no increased risk in people with low Ad5 immunity when compared to the placebo group.

However, Buchbinder added, the real reason for the additional risk in vaccine recipients may have had little to do with Ad5 immunity, or it may only have had an accessory role.

There was one more important risk factor for HIV acquisition in vaccine recipients when compared to the placebo group: circumcision, or rather, the lack of it.

Uncircumcised vaccine recipients were, in univariate and multivariate analyses, four times more likely to become infected with HIV if they received vaccine than if they received placebo. In contrast there was no difference at all in infection rates between circumcised vaccine versus placebo recipients.

In questions after the presentation, Buchbinder commented that the risk to uncircumcised men was greater than the risk to men with high Ad5 immunity, and that the latter may be a passive marker for the former. This was because the men with high Ad5 immunity, who were largely recruited in the second wave, also tended to come from countries and communities with lower rates of circumcision; they were younger and more likely to come from Latin America, for instance.

She also commented that preliminary data hinted that the enhancement of infection risk was specifically seen in uncircumcised gay men who largely or exclusively had insertive anal sex. This would imply that the vaccine was undermining some immune mechanism that normally protected uncircumcised ‘tops’ from infection through the mucosa of the foreskin.

However, this would not explain in itself why placebo recipients were less likely to get HIV if they had high Ad5 immunity, so the two effects may be synergistic.

The big concern arising from the STEP trial results was that the vaccine might have stimulated some recipients’ immune systems in a way that sensitised, rather than immunised, people to HIV. If that were the case, it could mean that recipients of the V520 vaccine could be unusually vulnerable to HIV for a long time – possibly for the rest of their lives. Immune responses to the vaccine did appear to be long-lasting and had faded only slightly in a group of early vaccinees two years after injection.

However by October 2009, when more follow-up data were presented at the AIDS Vaccine 2009 conference in Paris, it was beginning to seem that the excess risk to adenovirus-experienced male recipients was fading over time.⁵ Susan Buchbinder told the conference that an excess risk was still evident in Ad5-seropositive men, but it was no longer trending towards significance. The risk was still of borderline statistical significance in uncircumcised men, though it had declined too.

She told the conference that 60 new infections had occurred in study participants between unblinding in October 2007 and January 2009. Twelve occurred in women (six in each group) and 48 in men (26 in the vaccine group and 22 in the placebo group).

There are still many questions left unanswered by the STEP trial. Buchbinder said that the trial participants would need to continue to be followed to find out if the apparent extra vulnerability to HIV conferred by the vaccine was long-lasting, and Mark Robertson outlined a whole range of other factors that would be analysed, including other immune activation markers and CD4 subsets, markers of genetic vulnerability to infection such as HLA genotypes, herpes (HSV-2) status, seminal immune-cell subsets and seminal HIV viral load in seroconverters, and so on.