February 9, 2009
There's nothing like hearing the results of studies directly from those who actually conducted the research. In this interview, you'll meet one of these impressive HIV researchers and read his explanation of a study he presented at CROI 2009. After his explanation, he then answers several questions from the audience.
Salim Abdool Karim: I'm sitting here in front of you today to share with you very briefly the results of the HPTN 035 trial, which is a trial undertaken under the auspices of the microbicide trials network funded by the NIAID [U.S. National Institute of Allergy and Infectious Disease] of the National Institutes of Health [NIH].1 This study was conducted in five countries and enrolled 3,099 women in these countries.
Salim Abdool Karim, M.D., Ph.D.
This study looked at two particular microbicidal gels, BufferGel and PRO 2000 -- and, in a four-arm study, compared HIV incidence rates, pregnancy rates and sexually transmitted infection rates. Two of the arms were control arms: One was an arm that had a placebo gel, and the other was an arm that had no gel.
In short, the main finding is that PRO 2000 gel, which is a naphthalene sulfonate polymer, an anionic polymer, was shown to have a 30% reduction in HIV incidence when compared to the placebo arm. A similar level of effect, at 33%, was observed when compared to the no gel arm. This finding was not statistically significant because the study was not powered to answer the question of 30% efficacy.
However, when we look at the corroborating evidence in the subgroup analysis, what we find is further evidence that supports our overall conclusion that this is a promising candidate as a microbicide. One of those analyses looks at whether the effect is stronger and more prominent in those who have used the gel more often, and particularly in those women who used condoms to a lesser extent, but used high levels of gel.
In the study, BufferGel showed no effect on STDs [sexually transmitted diseases], contraception and HIV.
On that note, I think for the microbicide field -- which has over the last several years had studies that were stopped either for harm or for difficulty insuring adequate incidence rates, or have shown no effect -- this is the first study that now shows that we have a promising candidate. Because the result is not statistically significant, we do not regard it as a definitive conclusion that PRO 2000 is a microbicide, but we certainly view it as very promising, particularly in light of the fact that there is another study of PRO 2000 whose results will be released sometime later this year.
Reporter #4: Professor, how did you measure adherence?
Salim Abdool Karim: Thank you for the question. Within the HPTN 035 trial, we measured adherence by the participant's recall of product use and condom use at the last sexual act and over the last seven days before they attended their clinic visit. There was strong correlation between the two measures. It is self-reported data, so it has all the caveats of self-report, but that's the one measure we have.
Reporter #4: In the Carraguard study, self-reported adherence was a great deal higher than adherence measured by dye.
Salim Abdool Karim: When we look at adherence within the trial, what we have is just about 81% of last sexual acts involved gel use. If they were not using the gel, we would not see the level of effect we did, so I think that's the one positive thing -- that what we've shown is that use of the gel is associated with some measure of protection. Ideally, we'd like to have a study where everybody used gel and nobody used any condoms. That's not the real world, unfortunately, so we, within the constraints of the real world, I think what we've shown is -- it's amazing that we can even show a 30% effected. It means quite a lot within the broad prevention arena.
Reporter #5: Dr. Karim, can you describe the mechanism of action of this compound? It doesn't contain an ARV [antiretroviral], I don't believe. And how does it differ from Carraguard and other formulations?
Salim Abdool Karim: You're absolutely right: PRO 2000 is not an antiretroviral-containing compound. Instead, the way it works is that it has an anionic charge. It's negatively charged. When it's inserted in the vagina, the virus, in its variable loops on the envelope, has a positive charge. PRO 2000 attaches to the variable loops of the virus and as a result, the virus is no longer able to attach to the host cell. So it prevents attachment of the virus to the cell. Its effect therefore is through its presence in the vagina at the time of the sexual act.
Reporter #6: When you say 30% reduction, what numbers are we talking about?
Salim Abdool Karim: We had 194 cases of HIV and the distribution differed across the arms; that's how we show the effect. The incidence rate was just over four per hundred person-years within the trial, and that gave us enough power to show an effect that was in the region of around 50%. We had much lower power to show an effect at the level of 30%. If we look at the other measures we've used -- of pregnancy and so on -- those effect sizes were similar across the four arms; we didn't show any effect. If you want any more specific data: Retention rate was 94% within the trial -- in other words, only 6% of the women were lost to follow-up. We had about 73% condom use within the trial and despite that level of condom use and 81% gel use, we were able to show a 30% benefit.
Reporter #6: How many pregnancies did you have during the trial?
Salim Abdool Karim: I'll have to check that. I can't remember offhand. It was 11.1% within the trial overall.
Reporter #6: Was there distribution of pregnancy among the groups or were they similar?
Salim Abdool Karim: They were similar. They ranged between 10 and a half and 11 and a half.
Reporter #8: Thirty percent does not sound very impressive. Is there any thinking about how can it be improved?
Salim Abdool Karim: We would ideally have liked to see a bigger effect within the trial. However, it's 30% against a backdrop where we've had difficulty in showing any level of efficacy in many of the other products that have gone before. I think among the challenges that we face is to try and improve adherence, and in particular to focus on those populations within the countries involved where condom use levels are not as high and where women will be more adherent to the gel, because we find a bigger effect among those women who've used more gel and fewer condoms. I think that's among the challenges that we face. There's also a new generation of products that's also moving forward and we would hope that those products, which have different strategies of use and different levels of efficacy in animal studies, would also show promise.
To answer your question in terms of this particular gel and the future trials that are being undertaken, I think it remains for us to see what comes out of the British trial. If that trial shows a level of effectiveness in this order of around 30% to 40%, then I think it's very compelling, even though we do not show statistical significance [in our trial], that that is a level of effectiveness that we can see within the trial.
To contextualize that: For a woman who has no other option -- in other words, a woman who's trying to fall pregnant, who has a migrant husband who is at high risk of HIV, who's faithful to her husband -- what other option is there? We are looking at a situation where this product could be a niche product for a group of women who have no other options to protect themselves against HIV.
Reporter #9: Professor Karim, I had one more follow-up question about the British study. You suggest, I think, that it's not powered to show a result of 30%, so it won't be statistically significant if it reaches the same findings. Can they change the study at this point? Is there anything that they can do to try and improve, to show that kind of effect?
Salim Abdool Karim: There are designs that are available, called adaptive designs, where you can change the trial midway, so that you can allocate people differently to obtain this. Our trial wasn't designed in that way. Our trial had preset endpoints. We predetermined what we were looking for within the study and so when that was achieved, the study was stopped. We were stopped at a time when we were blinded. We didn't know what this result was.
That's part of the risks we take in designing trials: We aim for a particular level and we don't reach it. It's part of the challenges we face. There's no specific way in which we could have looked at the data and say, "Well, actually, if we do a few more HIV infections, that will get us a statistically significant result." Unfortunately, then that would not be acceptable.
Reporter #9: So for the MDP [Microbicides Development Programme] trial, that's also true?
Salim Abdool Karim: The MDP trial [the "British study" referred to earlier] is a much larger trial. It has about three times the sample size and about three times as many endpoints as well. So it is powered to answer this question at 30%.
This transcript has been edited for clarity.