February 25, 2009
Hello and welcome. This is Bonnie Goldman, editorial director of TheBody.com. I'm here today with Dr. Joel Gallant. Dr. Gallant has been treating people with HIV since the beginning of the pandemic. He's also a leading HIV researcher and a professor at Johns Hopkins University School of Medicine in Baltimore, Md.
Welcome, Dr. Gallant.
Thanks, Bonnie. Good to be here again.
Before we begin, I think we should say that CROI is such a rich conference, with so many tracks, that your take on the highlights will focus on just one small part of the conference, specifically studies on antiretroviral therapy strategies.
Yes, there was a lot at this conference. There's always a lot at CROI. It's by far the most important of the HIV meetings, but it was interesting to note the relative absence of clinical trials on antiretroviral therapy.
But there's some good news to that. It means we're sort of satisfied at this moment with where we are.
Clearly we're in a much better place than we've been in a long time, and most patients are doing great on treatment with the drugs we have. But we all worry a little bit that the honeymoon will end and that, in a few years, we could have people with a lot of resistance to some of the newer drugs. The concerning thing is, if there's not a lot of drug development and we need new agents for those people, what will we do?
But isn't it true that there's less drug resistance seen in HIV-infected patients?
There is right now. However, I have to say that a lot of these drugs, like Isentress [raltegravir, MK-0518], just came out recently. I've seen a lot of misuse of some of these agents by less experienced clinicians. I've seen resistance developing already and it's concerning, because the same mistakes that we made in the '80s and '90s are still being made now. With these newer drugs, we worry whether they'll continue to last for all patients.
Could you be more specific? What kind of mistakes are being made do you think?
I'll give you an example. A patient who had very extensive drug resistance had been treated for many years with lots of failing regimens. His doctor put him on Truvada [tenofovir/emtricitabine, TDF/FTC] and Isentress. Of course, he was fully resistant to Truvada by this time, so it was only a matter of months before he was fully resistant to Isentress. That means he'll be resistant to other integrase inhibitors as well.
This actually is a good segue to the SWITCHMRK study that was presented at this conference, which really showed exactly that issue.1 I'll just jump ahead and talk about that study and that'll I think bring home my point. This was a study done by Merck looking at Isentress, the new integrase inhibitor. They took patients who were on a Kaletra [lopinavir/ritonavir, LPV/r]-based regimen and doing well, and randomized them to either continue Kaletra or switch to Isentress.
The problem with the study -- and this was what led to Isentress not performing well in the study -- was that they allowed pretty much anybody into the study, regardless of how much resistance they had or how many prior regimens they had failed. There were some people who got into the study, who had been treated for 22 years with antiretrovirals -- so that dates back to the AZT [zidovudine, Retrovir] monotherapy era -- and who had been on as many as 16 different antiretroviral agents. So, the fact that they were doing well on just Kaletra and some nukes [nucleoside reverse transcriptase inhibitors or NRTIs] was surprising, but to expect them to do equally well by switching them to Isentress is naive.
We know that boosted protease inhibitors [PIs] have a much stronger barrier against resistance than integrase inhibitors, so for all intents and purposes these patients were essentially taking Kaletra monotherapy. By switching to Isentress monotherapy in the SWITCHMRK study, they failed because it's easier to become resistant to an integrase inhibitor.
Most expert clinicians would have known better than to do this to a patient. I'm sure that that same thing is happening in clinical practice as well, and it's based on this belief that these new drugs are so great that you can do anything with them.
I don't mean to paint a bleak picture. Obviously, most people are doing great, but it's so important when using these new drugs to use them wisely and carefully, and not just assume that any use of them will be fine.
I noticed that there were few new drugs presented. I was looking for vicriviroc [SCH 417690; SCH-D], for instance, but I didn't see that.
There was a little bit about elvitegravir [GS 9137, JTK-303], but it was in the context of the new Gilead booster.
That's going to be, hopefully, part of a new four-in-one drug?
Well, yes. There was quite a feeling of surprise among the audience when Gilead announced that they had already co-formulated a combination drug containing all four Gilead drugs.2 This coformulation has the equivalent of Truvada (which is tenofovir and FTC), elvitegravir (which is their new investigational integrase inhibitor) and GS-9350 (which is a drug that they are developing to replace Norvir [ritonavir, RTV] as a booster).
We use Norvir pretty much only for its ability to boost levels of other drugs. Elvitegravir needs a booster. Obviously, drug companies don't want to roll out a new drug that's dependent on ritonavir. They want to be able to co-formulate, so they are developing this booster in tandem with elvitegravir.
What this does -- in addition to allowing co-formulation -- is that it allows them to study elvitegravir in treatment-naive patients. If they have to use ritonavir, they're not going to be allowed by the FDA [U.S. Food and Drug Administration] to study elvitegravir in treatment-naive patients. The FDA is concerned about giving low-dose ritonavir to a patient who's not taking a protease inhibitor since it might create resistance. So that was leaving them stuck with a treatment-experienced patient population and a head-to-head comparison of elvitegravir with raltegravir, which just seemed like it was not likely to go quickly or go well. Now they can study this four-drug pill head-to-head in comparison with the three-drug combination Atripla [efavirenz/tenofovir/emtricitabine, EFV/TDF/FTC] in treatment-naive patients, and it could lead to a much more rapid development of both elvitegravir and GS-9350.
The booster will be used separately for a lot of other reasons.
There'll be two purposes for GS-9350. One will be as a booster for elvitegravir. Gilead says it will also be a stand-alone product that can be used to boost other protease inhibitors. They'll be doing some studies looking at their booster boosting other PIs.
Of course, some of the other companies will probably be interested in potentially using it to boost their drugs -- for example, BMS [Bristol-Myers Squibb] may want to use it with Reyataz [atazanavir, ATV] and Tibotec may want to use it with Prezista [darunavir, TMC114]. Lots of other drugs, including drugs outside of the HIV field, might be boostable with a drug like this.
Isn't it true that GS-9350 has no activity against HIV?
Right. Of course, the hope is that GS-9350 causes less gastrointestinal [GI] side effects and less lipid toxicity than ritonavir. All we have now is very early data, but so far, it looks like it's well tolerated.
Wasn't there another booster presented?
Yes, a very small company called Sequoia Pharmaceuticals has a booster that they're studying with protease inhibitors.3 So far it looks like it's doing the job to boost those levels as well. My concern with that drug is that although they call it well tolerated, which is what all new drugs are always said to be, it seemed like there was quite a bit of nausea, headache and vomiting in the study -- even ritonavir doesn't do all that. So we'll see what the tolerability profile looks like in further studies.
Was there anything else presented about new antiretrovirals?
There was a study presented about PRO 140, which is a monoclonal antibody that's used by infusion,4 although I didn't get anything out of the presentation that was newer than what we've heard before. Most of the new drug presentations were in the very early stages of development and I sort of glossed over those.
So there was no news on vicriviroc?
The thing to remember about vicriviroc is that although it's a once-a-day CCR5 inhibitor that's boostable with ritonavir or some other booster, other than that, it doesn't have a lot of advantages over the existing CCR5 inhibitor maraviroc [MVC, Selzentry, Celsentri]. And maraviroc is not exactly a big seller. There are not a whole lot of people taking it and so it calls to question the future role of these drugs.
Most of the reasons for their lack of use are not because of problems with the drugs themselves, but because of the need for the expensive Trofile assay first. So there are a lot of CCR5 inhibitors in development, but it seems like it's a pretty small market right now.
Let's now switch subjects and talk about the adverse effects of HIV, which look to be the new focus of conferences like this one. There were a lot of data about cardiovascular issues. Can you talk about that a little bit?
Yes, so there was a lot of attention trying to further understand the possible connection between abacavir [ABC, Ziagen] and heart attack, or cardiovascular, risk. Up until this conference, we had the D:A:D study that showed a risk,5 the SMART study that showed a risk and then we had the GSK [GlaxoSmithKline] analysis of their data that did not show a risk. At this meeting there was some more information.
First, there was another presentation from the D:A:D.6 They now have enough data on tenofovir [TDF, Viread] to look at the risk with tenofovir and they did not show a risk with tenofovir. They have argued that this strongly suggests that the abacavir risk that they're seeing is not what we call a "channeling bias," that is to say that people who are at high risk of heart disease would be put on abacavir preferentially and therefore it would make abacavir look like it was associated with risk.
The argument is that had that been true, they should have seen the same thing with tenofovir and they did not. That argument has some problems, but I think they've demonstrated in many other ways that channeling bias does not appear to be a big issue here.
In this study, they also looked at the risk of heart disease with protease inhibitor use. Of course, we've always known -- especially with the D:A:D presentations in the past -- that protease inhibitors have been associated with an increased risk in heart disease. They've helped further define that and showed that indinavir [IDV, Crixivan], which we rarely use anymore, is associated with a higher risk, as is Kaletra. What was interesting was that the risk in the case of indinavir was not associated with the use of ritonavir. In other words, you couldn't explain the increased risk of heart disease based on whether or not people were boosting indinavir with ritonavir. Also it was interesting to learn that the risk with indinavir and Kaletra was not dependent on their effect on lipids. It was an independent risk.
Can you explain this further?
A lot of people have just assumed that the protease inhibitors increase heart disease risks because they increase cholesterol and if you control the cholesterol, you'll eliminate the risk with protease inhibitors. The D:A:D study suggested that while cholesterol is part of the problem, it may not be the only problem, and that there may be an independent association that's not just explained by cholesterol effects. I'm not trying to say that this is an established fact. This is an observational study, but that is what they found.
So they are suggesting that it may be something in the medication itself?
Yes. For example, indinavir is well known to cause insulin resistance and affect glucose metabolism, and that could affect heart disease risk. That is not as true with Kaletra as it is with indinavir, but there may be some role there. Of course the problem is that they don't yet have data on the PIs that we're using most commonly, namely Reyataz and Prezista. So we've assumed that those drugs may be a little bit safer, but we don't have data from the D:A:D study yet on those two.
There was a new study from the French National Hospital database.7 It's a case-controlled study where they, again, found a risk associated with abacavir for myocardial infarction [MI].
In this study, the heart attacks were confirmed by cardiologists, so it was well done in that way. When they did their first pass at the data, they did not find a risk. However, they found a risk when they did a post-hoc analysis looking specifically at patients who have taken abacavir recently -- that is within the last six months -- and during their first year of abacavir therapy.
So their finding was consistent with the D:A:D in that it's early and active use of abacavir that is associated with a risk. It is not a cumulative risk as you would see with a protease inhibitor. Rather it's this on-and-off phenomenon that was described with D:A:D, a different kind of a risk than with a protease inhibitor.
People have argued that that probably suggests that the risk with abacavir is not a metabolic risk. It's not from increased cholesterol and buildup of plaque on your coronary arteries, but if anything, it could be some kind of a risk due to inflammation that occurs while you're on abacavir and then goes away when you stop it. That's certainly a logical explanation, except that the data on looking for this inflammatory process has been inconsistent -- some studies show this and some don't.
There were a couple of other interesting antiretroviral studies presented at this conference. One was the STEAL study from Australia.8 It was a randomized trial comparing people who switched to Epzicom [abacavir/lamivudine, ABC/3TC], or Kivexa as they call it in Australia, versus switching to Truvada once they were suppressed. They found eight cardiovascular events in patients who switched to Epzicom versus one among the Truvada patients, and that was a significant difference. It's noteworthy that these were not all heart attacks though. There was a mishmash of heart events.
Finally, there was a study looking at a variety of ACTG [AIDS Clinical Trials Group] studies with long-term follow-up on abacavir that showed no risk of MI.9 As Peter Reiss pointed out in a really nice review of this subject, the score's now 4 to 2 -- four studies showing an increased risk and two not showing an increased risk.10 Reiss pointed out, first of all, that in the studies that don't show a risk, the average patient age is about seven to 10 years younger than in the studies that do show risk. Of course, age is a risk factor for heart disease, so you could argue that it's going to only be in older people that you're going to see a significant difference.
Reiss also made the point, which I thought was interesting, that the studies that show a risk tend to have more people who were already suppressed when they started abacavir -- their viral load was undetectable. Whereas in the studies that don't show a risk, everybody had a detectable viral load before starting abacavir.
The reason why that could be important is that when you treat anybody for HIV with any combination, you're going to see an immediate decline in risk because inflammation, immune activation and risk of clotting all decrease dramatically when you start antiretroviral therapy.
What Reiss said was that if you had minor differences between drug A and drug B, you wouldn't really see them in a study that looked at suppressing patients who were treatment naive because any difference between the drugs would be completely overwhelmed by the benefit patients are getting from treatment. It's only in studying people who are already on treatment and then go on abacavir versus another drug, that you're going to see a difference because it's a much more subtle difference. This is an interesting hypothesis that deserves to be looked at further.
It's an interesting theory. It just seems counterintuitive that someone suppressed would be at more risk for heart disease on abacavir.
Remember, they're not more at risk than if they weren't suppressed. They're clearly at lower risk than if they weren't suppressed, but there may be a difference between drug A and drug B once you're suppressed.
What does all this mean for the clinic, and to a patient today?
I think that the current DHHS [U.S. Department of Health and Human Services] guidelines have it right.11 The guidelines say that for patients who have no risk factors for anything, then Truvada is the best nucleoside backbone to use. But there are patients who are not good candidates for Truvada, mostly people who have kidney issues. For those people, Epzicom may be the appropriate choice in many cases. I think Epzicom is the most appropriate nucleoside backbone for people who have kidney issues, but who don't have a high risk for heart disease.
The problem we get into is with the people who have both risks. For example, a diabetic with high blood pressure is at risk for both kidney disease and heart disease. So what do you do with that person? This may be where we start getting into nucleoside-sparing options, but very few of those regimens have been well studied, and we have not studied them for this particular population. But that's what a lot of us are doing for people who are not good candidates for either tenofovir or abacavir. I would just stress again that I don't worry much about abacavir in people who have no significant cardiovascular risk. It's the people who smoke and have a family history of heart disease, diabetes or high blood pressure, and have multiple risk factors that I would worry about.
Do you think clinicians are doing more risk assessment, in terms of assessing the cardiovascular health of their patients?
I hope so. I hope that was happening even before the abacavir issue, because we know that people with HIV are at higher risk for heart disease and that drugs like protease inhibitors and, to some degree, efavirenz [EFV, Sustiva, Stocrin] can increase the risk. Of course, a lot of us are doing some degree of primary care anyway and in addition our patients are getting older. We should be looking at this just from a primary care angle. This only increases the importance of using the online calculators, calculating Framingham risk and talking to patients, especially about modifying the correctable risk factors.
I'm sure that if you were to have a choice between smoking cessation versus switching from abacavir to another agent, you'd get much, much more benefit from smoking cessation. The benefit you might get out of an abacavir switch would be much more modest. So it's really important for us all to be trying to correct the risk factors that we can correct.
Have you seen as a result of this news being disseminated a lot of patients coming into your office frightened and not really knowing what they can do to reduce their risk?
Most of them know what they can do in terms of diet, exercise and smoking cessation. It would be an unusual patient who hasn't been told a few times by a doctor about that kind of thing. But it never hurts to talk about it again. With any substance abuse counseling or smoking cessation counseling, people eventually may get it. It may take 20 conversations and the 21st is the one that does it, but we should never give up.
Were there any other presentations that you found interesting?
I think a big study was the NA-ACCORD [North American AIDS Cohort Collaboration on Research and Design] study about when to start; that was a very important one.12
There were two trials presented back to back. One was NA-ACCORD. This is a big observational study in North America, combining data from lots of different centers. You and I talked after ICAAC/IDSA about their first study that compared people who started therapy with CD4s between 350 and 500 versus those who waited until after their CD4 was below 350, and found a 70% better survival in those who started early. [To read or listen to this interview with Dr. Gallant, click here.]
At CROI, they presented the same kind of analysis, but now they looked at people who started above 500 versus those who waited and they found a 60% improvement in survival. Keep in mind that for most of the people who started at CD4 counts of above 500 that was back in 1998 when people were using the "hit early, hit hard approach." In their cohort, there has been a lot fewer people in more recent years who have done that.
Yet, even with the drugs available in 1998 -- which were difficult toxic combinations -- they still found this survival benefit. So if they're right, then the benefit would be even greater today than it was back then.
There are all sorts of potential problems in a study like this. The main one is the potential for what's called an unmeasured confounder. For example, the reason some people started on treatment early was because they were highly motivated, well-informed, health-seeking type people who were eager to treat their HIV as aggressively as possible and, at the same time, those same behaviors and characteristics also caused them to eat right, exercise, not use illicit drugs, not drink to excess and not smoke. In other words, they were doing all sorts of things that improve their survival. So the very same factor that could affect their decision to treat would also affect the endpoint, and that's what we call a confounder.
Now, if that were true, that could invalidate the results of the study. However, the investigators -- who are well aware of this potential pitfall -- have done as much as they can to try to control for that. They point out that for a confounder to invalidate the results, it would have to be a very, very large confounder. They showed data to support that. They also pointed out that people in the group that waited to start treatment tended to do as well virologically as people who started early. They tended to have the same rate of suppression. So it's a very strong powerful study, but it's not a randomized trial. We're going to have to wait and see whether this is enough to change the way we treat or whether we need to go ahead with a randomized trial.
The other study related to treatment initiation was from the ART Cohort Collaboration, another large conglomeration of cohorts.13 They presented data that again showed that it's better to start at CD4s above 350 than below 350. They, however, did not find a benefit from therapy in patients who had a CD4 count of above 500. Their methods are a little different. They don't have data on what happened to people before they started therapy. They had to impute that information from older historical studies.
Can you explain how they did that?
They take cohorts from the pre-HAART era and look at prognosis at various CD4 counts and estimate that in their population of people who waited up to a certain point to start therapy that those same data would apply to them.
It's a way of trying to reduce lead-time bias. Lead-time bias can occur if you have people start therapy at different time points, but you don't have the data about what happened to them while they were waiting to start therapy, in which case you could be missing things.
For example, in a study comparing patients who start therapy at a CD4 count of 200 versus 350, how do you know there weren't people who died while they were waiting to start at 200 if you don't have that data? The advantage of the NA-ACCORD study is that they do have the data.12 They can look at everybody from the same CD4 point, whereas in the ART Cohort Collaboration they had to impute that from other sources. They're both perfectly acceptable techniques in statistics, but you could argue that having the data is better than having to guess at the data.
Do you think we still need to spend millions of dollars to discover the answer to this in a randomized study?
I don't think so. I have to say that, and others have made this point before me, when you think about what we're really arguing about, it's kind of akin to arguing about how many angels can dance on the head of a pin. We all agree now I think that we should be starting therapy before someone gets to a CD4 count of 350. When arguing about 500 versus 350, it may sound like a big difference, but in terms of how long it takes the average person to get from 500 to 350, we're talking about only a year or two.
Now put that into the context of a person's lifetime. Say you're a 25-year-old with a CD4 of 500, and you're looking at being on treatment, hopefully, for 30 years or more. Do you start at 500 or do you start at 350? Well, that's one or two years out of, say, 40 years of treatment. Is that really an important thing to figure out? I mean, it certainly has cost implications, but from a medical point of view, it just doesn't seem all that important to me.
The other thing is that we're at the point now where we've got this disease that's quite easily treatable with very effective treatment. I'm trying to think of some other infectious disease where you would have to prove that it was OK to treat.
HIV is unique in that the burden of proof seems to rest on those who want to treat early. Whereas for another infectious disease, you'd say, "Show me that I can wait. Show me that it's safe to wait." The reason for that is historical. It's not medical.
It's because first we didn't have treatment. Then we had lousy, ineffective treatment. Then we sort of had lousy, effective treatment. Now we have good, effective treatment. And because it's gone in that direction, we've approached it from the opposite direction, where the default is not to treat, and what you have to prove is a reason to treat.
I just think the data -- and not just from these studies -- are showing that from the moment you're infected, your body is in a state of inflammation and immune activation. No matter how high your CD4 count is, it is not normal or healthy to have HIV in your blood. My own feeling is that for people who are highly motivated and who are highly confident in their ability to adhere to therapy, it probably makes sense to treat at any time. The question remains when you have people who may not be good at taking therapy. We certainly don't want to see people developing resistance when their CD4 is still 500.
That would be a bad thing. So we have to be a little careful about not applying these ideas across the board. Certainly, if you ask the average HIV expert, "What would you do if you were positive?" I can tell you, they all would say, "I would start treatment immediately."
I always tell patients that the most important thing you can ask a doctor is "What would you do if you were in my position or if it was your son or daughter or something like that?"
It's a useful question to ask a doctor, because a lot of times doctors are afraid to give an opinion that is clearly an opinion and not based on solid data, but it's sometimes useful to know what they really think down deep. [As a matter of fact, a new series at TheBody.com asks prominent doctors and activists what they would do if they were HIV positive.]
But isn't this question kind of moot? Most people in the United States get diagnosed at a CD4 around 260.14
Yes, and that's another reason why I worry about the feasibility of the trial.12 It relies on being able to enroll thousands of people with CD4s above 500 and not only that, but people who are willing to be randomized to what they might consider treatment too early or they might consider treatment too late. So I worry it's going to be hard to enroll this study.
I'm hoping that if we start following CDC [U.S. Centers for Disease Control and Prevention] recommendations and making HIV testing a routine matter where everybody's getting tested, then we will start to diagnose people at higher CD4 counts and this will become an important question.
But at the moment, you're absolutely right. By the time most people are diagnosed, they already meet unequivocal indications for therapy and there's no debate about whether they should be treated.
One final note. The trial that is being thought of is the "START [Strategic Timing of Antiretroviral Treatment] trial": How far along is it?
It's been funded. I don't think it's enrolling at this point.
So it seems like it's going through.
I don't know. There's a lot of pushback about this. In these difficult economic times, I think a lot of us are questioning whether this is the best way to spend NIH [U.S. National Institutes of Health] money.
But there are certainly a lot of people who would like to see this trial go forward and who would argue, and rightly so, that the data that we have so far are just not convincing enough for them, and that we need a randomized trial.
The great danger, of course, is that when you plan a trial that takes this long, you could end up six years from now with results that don't matter to anyone anymore because they've moved on to a new question. Or you could find yourself midway through the study with nobody being willing to enroll in the study because treatment standards have evolved during that time period. I think that's a big concern.
Was there anything at the conference that surprised you?
I was the most surprised by the fact that Gilead had this quadruple co-formulation already made before we'd even heard any data on the drugs.2 That was impressive.
Thank you very much for an excellent update and I'm sure we'll be talking to you soon.