The big news at the 6th IAS Conference in Rome this past July was Treatment as Prevention (TasP), and specifically the findings from the large HPTN 052 study of over 1,700 serodiscordant couples which found a 96% reduction in HIV transmission when the positive partner was on suppressive therapy. Immediately following lead investigator Myron Cohen's presentation, I sat down with him to talk about some of the key findings in the study, and what they meant to him, not only personally, but for the field of HIV prevention research as well.
Jeff Berry: Could you tell us what one or two of the most important take-away messages from the study are?
Myron Cohen: There has been compelling biological evidence that antivirals would prove able to reduce sexual transmission of HIV. This large study provides definitive results of that very compelling hypothesis, so in the manuscript we say, "This proves the hypothesis that's been available for some long period of time." Take-away point one.
Take-away point two -- the design of the study allowed the first randomized, controlled trial investigating the benefits of CD4 count differences. The field was driven by observational data. That's not bad, but it's subjected to a lot of different biases. This is randomized, so when we see a benefit, whether it's small or large, in earlier ART [antiretroviral therapy], we have great confidence in the benefit. That's very straightforward.
Now, with Dr. [Beatriz] Grinsztejn's data [in Study 052 on the clinical benefits of starting therapy immediately], we are indicating [that] we need to really look carefully at the benefit that's accrued to tuberculosis prevention, because we have some options there. We can treat people earlier or we can give them prophylaxis and we have to kind of weigh those two options. So a person asking the question said, "Look, this is a triple-whammy. You're preventing tuberculosis, you're preventing transmission, and you're benefitting the CD4 count of the infected person. What are we debating?" And I could kind of accept that, even given the resource constraints. But I'm not a guideline person -- I'm the investigator contributing the data. The guidelines people will look at our data and make their own decisions.
JB: Could you talk about the design of the study, the reasoning behind choosing an immediate vs. a deferred treatment arm?
MC: Sure. First of all, a study like this has a huge number of ethicists involved. I can't tell you the number of ethical considerations from the beginning to the end. But you need to find an ethical study design. You certainly don't want to delay therapy to such a level that you might harm an infected person. And when we started, in an ironic sense, the criticism of our study was, "Everyone 'knows' that you can't treat people early." The criticism was not starting people late, but that no one should be started [with CD4 counts] above 200. We said, "Wait -- we can start people at higher CD4 counts." But the upper boundary became the tension, so we found an upper boundary that the countries, or the IRBs [Institutional Review Boards], could live with. Now, with the upper boundary of the study, 350-550, we needed people who would be healthy and not require treatment immediately, because they needed to have time to have an opportunity to see the benefit to prevent transmission. That required years, right? So we needed high CD4 counts, slowly lowering, no requirement from the country or the person for ART, as we look at this prevention question. That's how we got that upper boundary of 550. We never want to go below 200, we thought 200 was too low. So we gave ourselves 250 and you saw a tightening if it wasn't around 250. Investigators are fanatics about no one falling below 250, so we had a range; we had 100 CD4 cells. How long does it take to lose 100 CD4 cells as long as the person is untreated? Probably a couple of years. But then we saw regional differences. Faster loss of CD4 count in Asia than in Africa and America, which is interesting. So anyway, the design had two things -- it needed to have time for CD4 counts to fall where people didn't emergently need therapy for their own health, if that makes sense. And then built into the design was the opportunity to compare early vs. late ART, which we did. So there are really two RCTs [randomized, controlled trials] -- what we presented was actually two randomized controlled trials, one for prevention, one for treatment, but they happened to be concomitant.
JB: What about the START trial?
MC: That's stuff for the NIH [National Institutes of Health]. They'll look at our data and they'll decide what they're going to do. Of course, it might have implications for the NIH. The beauty of my job is to do this study, which is ongoing. The NIH and the START group will look at the data we generated and decide what they think.
JB: So, in your opinion, based on the study results, what would have the greatest impact on reducing HIV transmission -- getting people on treatment as they come in to care, or getting people into treatment earlier?
MC: Well, you're trying to dichotomize. I think we're not going to go wrong doing a better job of testing people. The current situation worldwide is such that people are way late in getting the benefits of treatment. So this discussion we're having about 350 vs. 550, is not very relevant, because what we need to do is to get all the people under 350, most of whom are not getting treatment. They're going to have personal benefit unequivocally and they're going to have prevention benefit unequivocally. So this is not a point of discussion really. The discussion starts with, "Should we ration drugs?" And where we ration drugs. We all agree on the under 350. We've got to bring in all these people who are under 350 who are presenting with 100 CD4 cells and sick. That's unacceptable. There's nothing worse to me as a physician than when I'm in the hospital and somebody comes in sick -- I saw this just last month. A woman came in from a community in rural North Carolina. She knew she was HIV-infected. She came in with a terrible infection, a low CD4 count. I said, "How is it possible that you were lost to follow-up? How is it possible that you're not on treatment?" She was so sick! I wasn't pejorative, I just thought what an incredible failure we, the health care system, made letting this woman become so ill. I felt terrible about it. So I think the first issue is the low CD4 count people, getting them earlier. And then the second issue is like a "tennis racket argument," as in what's the "sweet spot" on a tennis racket that you hit the ball to get it to go in the right direction -- okay, we see the benefit of earlier treatment but with the resources available as they are now, where's the very best group that we should treat to get the treatment benefit and the prevention benefit and the tuberculosis benefit all at once? That's the question. I don't know the answer. It takes thinking and a lot of consultation. I have data, I have 052 data, but I don't know the best answer. Somebody else asked a question about the efficiency of transmission -- the answer is very high in acute infection, it's very high in very advanced disease.
JB: So does the lower number of enrollees in the U.S. have any significance to translating the results?
MC: Yes, it had a lot of significance. With transmission studies done in the U.S. -- listen, prevalence is very low in the U.S., so in the screening, you can't find the couples. The only place you can get sufficient prevalence is among the MSM [men who have sex with men] population in the U.S. They were offered enrollment in the study, but I think that they declined because there are just too many options available to Americans, too many treatment options, so they have all kinds of different options that might not be available in other areas. Now, we never coerced, but it was a little bit easier where there were fewer options or people were more used to participating in trials. We of course did have a site in the U.S., in Boston, but it only participated briefly because the enrollment was not sufficient. They couldn't find MSM couples who really wanted to participate at a high level, so we ended up with very few couples and they were followed through the course of the study.
JB: My question is how could we not do the study [in gay men]?
MC: We don't do the study because people extrapolate from 052. At some point they say this is good enough. It depends on what society does. It's not clear to me how people are going to feel about this. I'll make no prediction. People might think this is plenty of evidence [and we'll stop], and they might say we have to do this stuff. I have no idea. It's the same with PrEP [pre-exposure prophylaxis]. It's one thing to applaud the investigations and studies, it's a whole other thing to build an infrastructure that lets all that stuff happen. We don't have such an infrastructure. And in the United States, at least, pre-exposure prophylaxis -- it's not been in great demand so far. I don't know about Europe, whether it's coming into demand. It's complicated. You have to be re-tested and re-tested and re-tested. It's not 100% effective. If you properly use a condom, if you protect anal intercourse, while not zero degree risk, it's way, way low. It's much better than 62% [as seen in PrEP research from a range of 42% to 92%, depending on adherence to medication]. It's not clear how people are going to perceive this. You have to think, is 62% enough? If I told you that if you get on a plane you would have a one in two chance of crashing, would you get on that plane?
[In 052] first of all, the biological plausibility is always there. Biology is always strong. But our results really are limited. When you say comparable [to gay men], our results reflect on couples first, established couples second, so they're a little special. Third, heterosexual couples, who denied [having] anal intercourse. The results show that, superimposed above condoms, you see a very strong protective effect from antiretroviral therapy. That said, there's another point: ART was managed. We could see the ART. We were only asking one question. All the PrEP studies you hear really are asking two questions. They didn't mean to ask two questions, but they did. One question was does PrEP prevent acquisition? But inadvertently they're also asking are people taking their pills? We didn't want to ask about pills. We wanted to measure virus in blood. From the beginning we could see the virus in blood. If the virus went up, we would call the person and say you're not suppressed. That's an important consideration here. So, a patient who's going to look at this data and be concerned about their partner has all of a sudden the obligation to be suppressed. Combination prevention is what we're going to do. Part of the combination is treating infected people. Part of it. Every combination now includes treatment.
I think condoms are imperfectly used and not used every time. I think we make a contribution at an individual couple level as they're making a decision to reduce dramatically the risk of transmitting HIV. At the population level, we hope to see a benefit. But that's more complicated. So all we can do is present our study results.
I don't think it's true that condoms are more effective than treatment. I think the one big study now says that treatment works better than condoms. However, condoms were part of the trial. So you can't be sure who's using condoms. But if you only use condoms, you don't get 96% prevention. The only time you get 96% prevention in this field is condoms plus treatment. Okay? But we don't know if people tell us the truth about condoms. How do we ever know if they really used a condom? In this study, the people who said they had a 100% condom use had less chance of getting HIV than the rest of the people in the study. So they might be telling us the truth. It's all self-reported. Study 052 is condoms plus ART.
We don't ever use the word "can't" [get HIV]. It's impossible. We say "reduces the probability dramatically."
In Europe there's a study called Partners, that's an observational study. They're enrolling gay men who are on therapy and they're trying to look over time at partners, gay partners. Are they transmitting? Only 16% enrolled, of 2,000 couples. Nevertheless, it's very hard to do this in gay men, there are a whole number of reasons. One of the questions at this meeting is should we do 052 in men? It is complicated. It's expensive. That's not a reason not to do it. Interpretation becomes very difficult because efficiency [transmission via anal intercourse] is greater. No one really, in my mind, has designed this properly. We're discussing this. Should we design a whole other study for gay couples and how do you randomize them? We can enroll in the U.S. and Europe, but because of stigma against homosexuality and laws against it, it might be difficult to enroll in Africa, in China, and elsewhere. We allowed gay men and had only 3% gay couples, from all over the world, so that's 24 couples.
It's not about being gay. It's about receptive anal intercourse.
JB: How did you feel about the standing ovation [at the end of your presentation]?
MC: It's quite remarkable, I think. This study has been going on a long time. I think the people in the audience, the 3,000 people, realized the tenacity of the investigative team and they've been talking about it (the study). Every time this came up -- does treatment work as prevention? -- they would say, "Yeah, but what about 052? It's not done yet." This audience has been waiting for these results for a long time, so they were recognizing the tenacity of the study team. And I appreciate that very, very much. But it's not like I invented something. I didn't make a vaccine. I didn't invent a new drug that you use only once a year. I proved an idea. I'm happy about that. But I don't hold myself out as Galileo.