March 15, 2009
CROI [Conference on Retroviruses and Opportunistic Infections] is always such an exciting conference with a wide-ranging mix of research -- from basic HIV science to the latest antiretrovirals. Today we're going to focus on an issue that HIV researchers have been struggling to understand for a while now: It's the association between cardiovascular disease [CVD] and HIV. I have with me clinician researcher Dr. David Wohl from the University of North Carolina at Chapel Hill who has been following the research; he'll share with us some of what he learned at CROI.
Welcome, Dr. Wohl. I know it takes some time to digest all the great science that was presented at CROI. What now stands out for you?
What I'll remember most about this conference was the beginning of a shift in my own mind away from the issues we've been batting around, CROI after CROI, conference after conference.
|David Wohl, M.D.|
By that, I mean I now think more about the big picture -- not about drug one versus drug two, or what's the best way to lower someone's LDL [low-density lipoprotein] cholesterol who has HIV infection, but the bigger picture of:
Was most of the data that was presented descriptive?
There was a mix. For anyone who has not been to this conference, it's pretty unique in that it's a pretty small conference. The science is generally very good.
But these are conferences, and so presentations are made looking back at established data. People present their preliminary findings, and people also present pilot data. So you get a mixture of information, some of which is much more thoroughly backed up by numbers and by analyses, and others that are nascent; we have to connect the dots over time.
So you have to be careful here. A poster presentation may have a study of 10,000 people, but not really have the data, while next to it is a study of 30 people that is really, really, really important. You have to be careful when you're looking through this. There are a lot of data, and you get inundated. The trick is -- and, I think, the value of doing something like this is -- trying to pick out the pearls that you find along the beach.
So, what were the pearls?
That's the other beauty about going to a conference. My pearls may be different than someone else's pearls; it's very individual. It depends on what you stand in front of and take note of, and what you care about.
Vis-à-vis cardiovascular disease, some of the things that I really cared about were some of the bigger studies that quantify just how much cardiovascular disease we're seeing, because it's nice to have the numbers. We can say, "Well, there's an increased risk of cardiovascular disease among HIV-infected people."
What are the references for that? D:A:D is going to be a study that we're going to talk about time and time again, because it's such an important look at what happens to people living with HIV infection. It is a huge meta-cohort of people from mostly European, but also some North American, sites; we're talking about over 30,000 people. It generated a lot of data.
Adapted from Colette Smith and D:A:D Study Group. CROI 2009; abstract 145.
One of the things that I really liked was that they presented data on why people died -- what are the causes of death -- in this prospective study.1 Over 2,000 deaths have occurred while they have been following people, and the death rate is pretty remarkable. It's 13.8 per 1,000 person-years and they were able to categorize why people died. Still, this is in some ways a retro-scope, because it's looking at people who died over the last several years. This is not what people are dying of in 2009, necessarily. So you have to take this with a grain of salt.
It tells us what people with HIV infection were dying from over the last several years and that age-related causes were responsible for one third of the deaths. That's not a surprise to any of us in clinics, especially here in the United States, where we continue to see people who are really challenged trying to deal with their HIV infection.
But that means that two thirds were not necessarily AIDS-related, or classically AIDS-related, in that they didn't die of PCP [Pneumocystis carinii pneumonia], cryptococcal meningitis or the AIDS-defining conditions. They died of other things.
Of the entire group, one out of 10 died from a cardiovascular disease-related event. That to me seems pretty remarkable. There were other things, of course, such as liver-related diseases and cancers, that are not classically associated with AIDS that also contributed quite a bit.
But that cardiovascular disease is responsible for one out of every 10 deaths in this cohort is, I think, striking -- especially since it's a European cohort, and maybe here in the U.S. it would be a little bit worse.
There was another study from Europe that overlaps a little bit with D:A:D, the EuroSIDA study.2 It too is a large meta-cohort study following people over time in which the researchers were able to break down the causes of death by AIDS-related versus non-AIDS-related. If you just look at the non-AIDS-related deaths, that is over 680 people in this cohort who didn't die from AIDS-related causes. An almost equal amount did die from AIDS. But of the people who didn't die from AIDS-related causes, again, about one third died from cardiovascular disease; malignancy was responsible for most of the rest.
The EuroSIDA study was really very helpful, striking and quantified the cardiovascular risk for me.
Adapted from Daniel Klein et al. CROI 2009; abstract 710.
Similarly -- and this is the last one of these big database studies -- the Kaiser Permanente researchers in California, every conference or so, look at their MI [myocardial infarction] rates.3 They are able to do this best by looking at hospitalization rates for MI. You have to be a little careful, because we're not looking necessarily at every event; rather, we're looking at who gets admitted to a hospital in this large, prepaid care organization in California.
What the researchers were able to show was that, for the last several years, clearly there has been excess hospitalization for MI among people who are HIV infected when compared to people who are not HIV infected, but it's been going down over the last couple of years.
In fact, the hospitalization rates for MI among HIV-positive people became no different than for the HIV-negative people who were in the HMO [health maintenance organization]. So that may be a little bright spot.
Again, caveats galore: These are hospitalization rates; these are not exactly events; and they weren't able to correct for a lot of the things we care about, such as smoking.
But again, I think it helps us understand a little bit about where we're at. Generally, in this cohort, as well, if you look at the last few years, people with HIV seem to have more MIs -- at least the MIs that get you into a hospital.
Isn't the Kaiser Permanente group a particular population (i.e., patients with private health insurance)? Aren't they a minority of the HIV-infected patients in the United States?
It's an HMO-like group. Although they would probably argue that they have a pretty diverse group of people -- maybe not as diverse socioeconomically as the epidemic in the U.S., but certainly racially and ethnically diverse, given that they are in California. But you're right. There are certain things we have to be careful about.
This is just one view of the bigger picture. But again, these are helpful data, especially if the decreases that were seen in the rates of MIs -- at least serious MIs -- over time are seen in other cohorts. That would be a possible indicator that we're doing a better job of preventing this ... or missing it, I guess you could also say. But I think that's unlikely.
Mostly this sounds descriptive, and not an investigation into why this is happening.
Right. So, moving into sort of the idea of, well, here's the magnitude. Is there more cardiovascular disease among HIV-infected people, relative to HIV-negative people? Then you start getting into what's going on. There were other studies that I didn't mention, smaller studies.
Priscilla Hsue did a very nice analysis of coronary calcium scores, showing that people with HIV had higher scores than people without HIV, beyond what you would expect.4 The analysis included people who were not on HIV therapy, so HIV itself seems to be doing something.
I think that's where we get into some of the really interesting effects of the virus itself.
Whether we're looking at inflammatory markers, immune activation markers or coagulation markers, there were some really important data that seem to indicate that people with HIV, just due to the virus itself, may be at higher risk for cardiovascular disease.
So, again, moving beyond observational studies, there were data from the FRAM [Study of Fat Distribution and Metabolic Change in HIV Infection] cohort.5 The beauty of FRAM, unlike D:A:D, is that it has a control group. D:A:D only involves people with HIV infection. FRAM has people who are HIV infected, but also people in a control group who don't have HIV. D:A:D is prospective; FRAM is cross-sectional.
Adapted from Carl Grunfeld et al and the FRAM Study. CROI 2009; abstract 146.
What I think is probably the final word, perhaps, on carotid intimal thickness studies is the really nice analysis the FRAM people did looking at carotid intimal thickness in people with HIV and those without HIV, while controlling for a whole bunch of different things.6 Lo and behold, using a really rigorous way of measuring intimal thickness of their carotids, people with HIV infection were shown to have increased needle thickness -- a marker of atherosclerosis.
We're seeing that there's an effect, perhaps, of the virus itself. One of the most striking, I think, pieces of data along this line came from Colorado.7 There was a very interesting study, very small, where 10 HIV-infected men, age 28 to 44, underwent a rigorous evaluation of their endovascular health. If your endothelium (the cells lining your blood vessels) is not healthy, you're not healthy, nor is your circulatory system.
The researchers provoked the endothelial cells. They provoked the vasculature to dilate using two different chemicals. They used acetylcholine as a challenge. Acetylcholine is an endothelium-dependent vasodilator. The other chemical used was nitroprusside, which is independent of the endothelium and its vasodilatory properties. They administered these to these men, and did brachial reactivity.
Interestingly, the acetylcholine response was a little bit less than they thought it would be. They compared these results to that of HIV-negative men who were in the same age range, and then to that of 10 HIV-negative men who were much older (59 to 68 years).
What they found was that the patterns of reactivity with provocation with acetylcholine were dramatically different across the HIV-positive to HIV-negative men -- such that the HIV-positive men's vasculature responded like that of the 60-year-old HIV-negative men, and was much more suppressed compared to that of the HIV-negative men of the same age. It was only with the acetylcholine provocation that you saw the differences, indicating this is an endothelial problem.
Men with HIV have an endothelial dysfunction that is analogous to what we see in older men, suggesting that there's premature aging of the vasculature and the vasculature reactivity, with HIV. And these are all treatment-naive men. These are not people on any sort of HIV drug that you may be worried about. These are relatively young treatment-naive men.
What was the average length of HIV infection for these men?
They didn't put that in. It's a small study and we have to see more on this. But, boy, these were significant results.
This is one of those tiny studies that you mentioned earlier that really made an impression.
This is a tiny study, but this measure, when done well, can be very, very reliable and reproducible. It was all done at one center by a group that's pretty familiar with how to do this, and I think were really able to generate some thought-provoking information that we now have to act on.
It's in lockstep with the data from the calcium coronary scores and other data that were seen that indicate that people with HIV have got a problem on their hands.
Even way back when the Framingham Risk Score in D:A:D under-predicted actual MI events in that cohort, indicating that maybe there's another factor -- call it the HIV factor -- that has to be entered into the equation.
It's just pure HIV? It's not length of time infected, maybe?
Length of time is hard, because we never really know length of time. We can guess it from length of time from diagnosis, but we don't generally know. And I think in these smaller studies there's going to be quite a bit of variability. But you're right. That has to play into it.
I would think that if someone's just been infected this year you wouldn't see any changes yet. But it would be interesting to find out.
I think you're right that there probably is a duration of infection where this starts to mount and accumulate. Perhaps early on, we don't see as much. We don't know this. But generally, most of the people who were involved in any of these studies will have had HIV for years. You would expect that over that kind of period of time -- given what we know about other risk factors of cardiovascular disease -- that may be sufficient to cause the damage we're talking about.
Was there any analysis of any correlation with smoking, genetic issues or environmental issues? Were these patients overweight? Was anything like that going on, as well?
In this particular study, there was no difference across the arms in terms of anything that would be a cause for concern like that.
So this is not explained by smoking or major differences in baseline cardiovascular disease risk. But you certainly, in larger studies, would need to think a lot about that. This group was pretty careful about selecting people to make sure that there was none of that confounding. You can do that in a smaller study that's a little bit more pure.
These were all people who had normal blood pressure, who had no medication and who were non-diabetic. Their lipids were all normal. They didn't have any evidence of cardiovascular disease. I think they were really well matched, as far as that was concerned. They weren't smokers.
That's a pretty big discovery, if it turns out to be true.
I think it fits into this whole thing that we have been talking about over the last few months, which is that there does seem to be an excess of not just cardiovascular disease, but other diseases that we associate with the aging process. It seems to be happening in people with HIV either more so, or at an earlier age.
We're talking about not only cardiovascular disease, but even malignancies. We know immunosuppression could lead to malignancies, but we're seeing malignancies we don't necessarily associate with immunosuppression.
We also know that bone mineral density is lower in people with HIV infection. We know that there may be more diabetes. We know that there may be more neurocognitive decline. Things that may occur in people as they get older seem to be happening earlier in people with HIV infection. I think cardiovascular disease is the most important one of those.
There were some interesting studies about HDL [high-density lipoprotein] levels. Can you talk a little bit about that?
There was one, in particular, that caught my eye. I actually had to go back and re-listen to the presentation. I think that's something worthwhile doing. The good thing about this particular conference is that you can download most of the posters, and many of the presentations. Since most of the presentations are only 10 minutes, it's not that difficult. The quality is sometimes not as good as you'd like. It depends upon your computer, and your connection. But I really do think it's good to do that. As I said, I did that again for this presentation that I think you're referring to. [To watch this presentation, click here.]
These are data that came out of the SMART study. The SMART study randomized people to continue on their HIV medicines -- if they were on HIV medicines -- or stop their HIV medicines -- if they were on HIV medicines -- or defer starting them if they were ready to start medicines, until their T cell count got to 250, and then start back again.8 The idea being to try to reduce someone's exposure to HIV medicines, and thus their risk for bad things happening to them, like cardiovascular disease, hepatic disease or renal disease.
At the time the study was designed, the assumption was that HIV therapies are metabolically toxic and we should limit people's exposure and do that in a way that keeps someone's CD4 cell count in the right range. Almost like sliding scale HIV therapy, for those of you who are familiar with sliding-scale insulin.
That study, of course, famously was halted early, 16 months into the study, because there were excess cardiovascular disease, renal disease and hepatic disease. But these illnesses were seen in the people who discontinued their HIV therapy, not the people who continued on HIV therapy.
Study subgroup analyses have been coming out over the last few years and producing really very, very important data for our understanding of this issue we're talking about now: HIV and its direct effects on metabolic complications.
One of the presentations from the SMART data set that were presented at CROI looked at HDL, as you mentioned.9 We know that there were predictors of bad things happening to patients in the study. One was discontinuing HIV therapy. But the researchers have also looked at other factors that may be across both arms.
The presentation looked specifically at HDL and LDL particle size, because we know that there are different types of lipoproteins, but they also come in different sizes. Generally the smaller the particle, the more atherogenic it is.
This holds not only for LDL, but also for HDL. If a person has HDLs that are smaller, that's not as good as having larger HDLs. In a really nice demonstration of this, the presenter showed the lipid profiles of two individuals, which were very, very similar when you just look at total LDL, total HDL, triglycerides, etc. However, when you looked at the subfractions of their HDL and LDL, there were major differences between these two individuals who, clinically, we would say were no different, but that their risk profile could be different, looking at the particle size.
What they did do was a case-control study of cardiovascular disease cases and controls who did not have cardiovascular disease. They looked at the particle sizes to see if they were different. The HDL particle concentration was significantly different between those who had cardiovascular disease and those who did not have cardiovascular disease.
In addition, when you look at people who were in the treatment-discontinuation arm -- the arm that did not do so well when it comes to these events -- there was a significant drop in HDL particle concentration. So not only were you more likely to have cardiovascular disease at baseline, if you had a lower HDL particle concentration, but if you got into the discontinuation arm, your particle size was more likely to drop than if you had been in the other study arm.
They did a really nice presentation in this talk about how CVD risk rises when you add lower HDL particle concentration to the other risk factors that they have identified -- such as high IL-6 [interleukin-6], which has been associated with cardiovascular disease outside of HIV. In the SMART study, high IL-6, high CRP [C-reactive protein] and high D-dimer were all associated with bad things happening to patients in the study.
If you overlay IL-6 levels being high or low, with total HDL particle concentration being high or low, they were at least additive, such that when you had high IL-6 and you had low HDL particle concentration, your odds for getting cardiovascular disease went up substantially. So they seem to build on each other, and it's very helpful for us to understand this.
There is a whole industry that is trying to sell us particle-size determination tests, and some cardiologists have a more favorable opinion of them than others. The question, of course, for all the SMART data -- and I think this is a really important question -- is: How do we integrate this into our clinical care?
They have already showed us, as I said, that IL-6 is important. They showed us D-dimers are important. D-dimers are really easy to measure. We get it in our hospitals and clinics all the time. CRP? Doctors get CRP all the time. Should we be doing CRP? Should we be doing D-dimer? IL-6 may be a little bit out there because it's more researchy, and it's not as easy to get. And doing the particle size?
We may be able to come up with a much more insightful lipid profile for our patients, but that's hard to say. We haven't even gotten there outside of HIV medicine yet, let alone within HIV.
But this is where I think we need to move on to, and think about: Can we better assess our patients' risks beyond the lipid profiles we have right now, if our current risk assessments may be falling a little short?
Given all this, is there an understanding yet about whether treatment choice can accelerate or slow down this premature aging process?
It's a really good question. I think that on the one hand, you could talk about HIV treatments. I'll talk about that very briefly, because I don't want to get too bogged down with that, although that's the thing that people like to talk about.
But, you are right: Are there other treatments that we may have to think about? Are we using statins to the degree that we should? Are patients getting aspirin who really should get aspirin? Those at moderate to high risk of cardiovascular disease, for whom there's not a contraindication.
There are modalities out there that we may not be employing as robustly as we should. So I think those things that we think about for anti-aging in HIV-uninfected populations, we might have to think about for HIV-infected populations as well.
When you get to specific HIV drugs, though -- just as a process -- I think we should be careful not to get too distracted about this.
Over the last few months, we have allowed, understandably, discussion of abacavir [ABC, Zaigen] to dominate the metabolic complications arena, and I think we should be careful about that. It's an important issue. But remember, this is not a drug that was extremely popular. Most of the people who need abacavir are on it for a good reason. Many don't have many other choices. And the majority of people who probably get abacavir don't have extremely high risk for cardiovascular disease.
So I think it's important, but I just don't think we should drown out almost everything else to allow this conversation to dominate.
Jens Lundgren et al. CROI 2009; abstract 44LB. Reprinted with permission. Download the full slide presentation.
At this conference, data on abacavir were again presented. It's become sort of a cottage industry. The D:A:D folks who first allowed us to understand that there may be a problem with abacavir updated their data.10
Interestingly, they are showing us that not only recent exposure to abacavir is associated with risk of an MI, but it's looking now like cumulative exposure is as well.
I think that's an important finding, because one of the big criticisms of the D:A:D analysis with abacavir was that there could be channeling bias, such that patients with higher risk for cardiovascular disease could have been channeled to abacavir. That should be the case, as well, for tenofovir and that wasn't seen here. So I think that was helpful.
Jens Lundgren et al. CROI 2009; abstract 44LB. Reprinted with permission. Download the full slide presentation.
The other part of it is that the researchers also started looking at protease inhibitors. They found out that cumulative exposure to lopinavir/ritonavir [LPV/r, Kaletra] and indinavir [IDV, Crixivan] are associated with MIs.
Importantly, the way the analyses were done -- if someone was ever on the drug -- it didn't matter for how long, apparently, or how long someone had not been on the drug -- once someone was on the drug, even if it was discontinued, the subsequent follow-up continues to be attributable to the level of the exposure at discontinuation.
So if someone was on lopinavir/ritonavir for six months three years ago, he or she was considered to be in the lopinavir/ritonavir group. Whatever happened, we're thinking about someone who's been exposed to lopinavir/ritonavir.
There is a lot of discussion that I think will happen further with this.
In addition, there was another interesting study presented. It was a French hospital study with a large database.12 The investigators did a case-control study looking at the abacavir question. Again, complicated statistics.
Sylvie Lang et al. CROI 2009; abstract 43LB. Reprinted with permission. Download the full slide presentation.
They found something similar and not similar to the D:A:D study. They found that there may be an association with abacavir when they look at the cases of MI versus no MI. They found that recent exposure -- only within the first year of abacavir use -- was where you see the risk of MI. And that this risk goes away over time. That's not what D:A:D is finding.
These things don't exactly measure up. This is very complicated statistics, and it's just a matter of what you believe, almost. There may be a signal here.
In another study looking at abacavir risk, the ACTG [AIDS Clinical Trials Group] did a large analysis looking at people in their clinical trials.13 Importantly, this is nice, because these are not people who selected to be on abacavir due to any risk factor; it was randomized. So if you're on abacavir in these trials, it's because you were randomly assigned abacavir.
In this study, they didn't find any risk associated with abacavir, albeit there were something like only 27 MIs and 60 cardiovascular disease events in this cohort of 10,000 people. So I feel like if there is a risk, it's probably a relatively fairly small risk -- and that could be argued. And it probably occurs more likely in people with more risk factors.
I am not as blown away by this. I think we should not get too caught up in the politics and the arguments. I think the data are there, and clinicians will have to make decisions for themselves.
This conference was so different than previous conferences, at least in the last two years, where there was so much new data about new drugs, and a lot of excitement about new things. This conference seemed to be more about new problems, or deeper understanding of these problems, and not at all about new antiretroviral drugs and all the wonders of them. Almost more depressing, I think, than upbeat, in terms of a conference. Do you agree?
I don't know. I actually thought it was a great conference. It may just be because I like Montréal. But I thought that there was a lot going on. I think the science was really good. We're trying to hammer out what's going on with our patients in a bigger way, and I think we're taking a bigger picture approach. I found that very helpful.
Again, I did find this common thread throughout many of the presentations, of: Lo and behold, our patients are going to be surviving for decades. How are we going to prepare them for that and prevent devastating consequences from occurring -- whether it be neurocognitive, cardiovascular disease related, bone-related or what have you?
I think there were less new and exciting drugs because we have gone through a flush of the pipeline just recently, where we had some great drugs come out.
I wouldn't undersell the presentations about the ritonavir [RTV, Norvir]-like boosters that can pharmacologically boost other drugs.14,15 There are two products that were discussed at the conference, both of which are exciting. And relative to what we're talking about, if they turn out to be as cardiovascularly neutral as they seem to be -- at least the Gilead product does -- boy, that's going to be a tremendous advance, potentially.14
Conferences all come in their different flavors, and sometimes that's by accident, sometimes by design. But I think we are starting to get onto something here -- whether it be inflammation or paying more attention to these long-term living-with-HIV issues. Those are exciting to me. And so I kind of like that.
But Dr. Wohl wouldn't you agree that, from a patient point of view, much of the new data from CROI about cardiovascular disease were worrisome since nothing has been suggested to prevent these issues?
Agreed. But you know what? And this is a point I make constantly to people who ask about that. The first step in developing an intervention is to:
We're getting to the recognition. I think we all appreciate that there's more stuff going on. You only have to pick up The New York Times to see articles about aging and HIV.
Two, we're getting a better handle on what's going on. There do seem to be markers of this badness that we are able to identify. I think it's not going to take too long, given the community of researchers and advocates who are involved with HIV, for us to start seeing interventional studies coming along. That's something we do very well, and we're very quick, right off the blocks.
But the first thing is, we don't want to go down the wrong path. This study helped me think about: What's the preparation? What do we need to understand as we go forward so we don't go off on some tangent that doesn't matter?
There was a convergence at this conference. Inflammation is bad, in a whole bunch of different ways. We know that from outside of the field of HIV. In HIV, it's even worse. Let's see what we can do about reducing inflammation, and then let's see if it has a benefit. That's where I see this going. And I think you're going to see more of that over time.
Were there studies on cognitive issues related to this issue?
Allen McCutchan et al. CROI 2009; abstract 458. Reprinted with permission. Download the full poster.
There are. There was even one that indicated that people with metabolic syndrome had a higher risk of neurocognitive impairment with HIV.16 That's important. When asked, "Well, why is that?" The answer is going to continue to be, "Inflammation. Inflammation. Inflammation."
It's like in that movie The Graduate. Plastics, that's the future. For here, it's inflammation. I think that we are seeing more and more about this -- and coagulation, by extension. We're going to start seeing more about the bad humors that come about when you have HIV that is not controlled, or not adequately controlled. I think that trumps all these discussions about this nucleoside versus that nucleoside, or this protease inhibitor versus that protease inhibitor.
We've come to the end of our time. What a terrific discussion you have provided, Dr. Wohl. Thank you so much for taking the time to talk to me.
This transcript has been lightly edited for clarity.