October 26, 2008
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 the study he presented at ICAAC/IDSA 2008.
I'm Joe Eron, from the University of North Carolina. This is a subanalysis of the two BENCHMRK trials, which were studies comparing raltegravir [MK-0518, Isentress] plus optimized background versus placebo plus optimized background.1
What we did here was look at clinical events. We haven't had studies for 10 years that have shown clinical benefit. We have had great therapies. But this was an attempt to see if there was actually a clinical benefit. So we looked at AIDS-defining conditions and deaths in patients randomized either to raltegravir plus optimized background or to placebo [plus optimized background].
What we found is that, for the most part, looking at any of these endpoints, raltegravir was more active, had better antiviral activity and had fewer events, whether it was new AIDS-defining events or the combination of AIDS-defining events or death. The decrease in risk was about twofold; the relative risk was about .5. About half as many patients had events [as those who received placebo].
The important thing is, it's actually a very small number of events. None of the comparisons reached statistical significance. They all were in the same direction, but they all were borderline in terms of significance. They did not reach statistical significance.
How many people were in the study?
There were about 700 patients total in the two studies combined. It was a 2-to-1 randomization; 460 patients got raltegravir and 237 got placebo -- remember, all with optimized background. That's really important.
The number of events was actually small. There were only 17 events in the raltegravir arm and 11 events in the placebo arm. But remember: 2-to-1 randomization. Also, people got raltegravir for a much longer period of time. The analyses are all adjusted for the amount of treatment that was received, and the fact that it was a 2-to-1 randomization.
The actual number of clinical events overall was small, but in the adjusted analysis, there were fewer events over time in the people treated with raltegravir, compared to placebo. That includes new AIDS events, death -- there were very few deaths -- or time to AIDS or death. We looked at this with the Kaplan-Meier plots that are on the poster.
What I think is really important about the analysis is that, if you actually look at all the people who had an event, regardless of whether they got placebo or raltegravir, the predictors of a clinical event are not quite what you might think. For example, you might think, "Well, the people who had clinical events had more resistant virus," but in fact, there wasn't a difference in the GSS [genotypic sensitivity score] between people who had clinical events and people who didn't have clinical events. You might have thought they would be more treatment experienced, but the number of drugs previously exposed and the years of therapy were no different. The big difference is that people who had clinical events were clinically advanced, which makes perfect sense. The median CD4 was 9 in the people that had clinical events, and 36 out of the 37 actually had a previous AIDS-defining event.
I think this is telling us that, if we're going to use these new therapies -- whether it's raltegravir, etravirine [TMC125, Intelence] or a combination of multiple new drugs -- we want to do it before patients reach these more advanced clinical states. It was in people with a very low CD4 and previous AIDS-defining illnesses that we saw most of the events occur.
When you were thinking of doing this study, did you believe that patients would have fewer AIDS-defining conditions on raltegravir?
That's a really good question. We weren't sure, because there have been a lot of studies done in the last eight years that have shown more effective therapies. But those studies haven't shown a difference in clinical events. TORO is a really good example; those T-20 [enfuvirtide, Fuzeon] studies.4,5 Very effective, obviously, but they didn't see a difference in clinical events. This was an exploratory analysis.
But we did anticipate that the activity of raltegravir would be such that perhaps we would see a difference in clinical events, and we knew that at least a certain proportion of patients would have advanced enough disease that they would be at risk for clinical events. There's another poster, from the DUET studies with etravirine, that shows very similar results: a clinical advantage to adding a fully active agent, compared to optimized background, where they also saw fewer clinical events.6
I don't think we were sure ahead of time. The time to first AIDS event, new or recurrent, was a preplanned analysis. The rest of these analyses that I've talked about are post hoc. But we did think there would potentially be a difference.
This transcript has been lightly edited for clarity.