You are leaving Medscape Education
Cancel Continue
Log in to save activities Your saved activities will show here so that you can easily access them whenever you're ready. Log in here CME & Education Log in to keep track of your credits.


Important New Data on Cardiovascular Risk and Metabolic Issues in HIV Disease From the 14th Conference on Retroviruses and Opportunistic Infections: An Expert Interview With Judith A. Aberg, MD

Authors: Judith A. Aberg, MDFaculty and Disclosures


Editor's Note:
The antiretroviral medications (ARVs) used to treat HIV disease have been associated with a variety of changes in metabolic parameters, such as dyslipidemia, among others. For many years, attention has focused on the potential role of ARVs in changes in such diverse manifestations as appearance (eg, lipoatrophy) and cardiac events. Much of the seminal data in this area are presented at the annual Conference on Retroviruses and Opportunistic Infections (CROI), and this year was no exception. Attendees heard presentations on new analyses of results of the SMART Study, which surprised many last year, and new results from the class-sparing strategy trial, ACTG 5142, which held some important surprises this year. In order to gain more insight into the most important data on cardiovascular risk and metabolic complications in HIV disease as presented at CROI 2007, Mary Anderson, PhD, Clinical Editor of Medscape HIV/AIDS, spoke with Judith A. Aberg, MD, Associate Professor of Medicine in the Division of Infectious Diseases at New York University School of Medicine, and Director of the NYU AIDS Clinical Trials Unit and Medical Director of HIV Services at Bellevue Hospital.

Medscape: Both antiretroviral therapy (ART) and HIV infection itself have been associated with metabolic changes in people with HIV infection. In order to frame the rest of our discussion, what are the highlights of these changes?

Dr. Aberg: HIV itself has been associated with dyslipidemias, in particular having high triglycerides and having a low high-density lipoprotein cholesterol (HDL-C). Those have been the most prominent lipid abnormalities. Before antiretroviral therapy patients were having a lot of complications, and some of it included wasting; you would see the manifestations of fat loss that were associated with wasting. Then after starting antiretroviral therapy, the things that we were seeing were initially called "fat redistribution." Now we know that it is really not that the fat itself is being redistributed, but that people are either experiencing fat loss or fat gain, and in some individuals, there is a mixed picture. The changes in body fat include fat loss, which is more likely to be found in the extremities and face, and fat gain, which can occur as either visceral or even subcutaneous gain and can result in the occurrence of "buffalo humps." We began seeing insulin resistance and development of diabetes. And then we saw a different type of lipid abnormalities with some antiretrovirals (ARVs). Particularly with the protease inhibitors (PIs), there were increases in triglycerides, but we were also sometimes seeing HDL-C going up a little bit depending on what the [ARV] combinations were. There were also changes in low-density lipoprotein cholesterol (LDL-C), with it going up in these individuals. People were developing what's called the metabolic syndrome, which is a combination of fat changes, insulin resistance, hypertension, and hypertriglyceridemia.

Medscape: The evidence has been mounting over the years that highly active antiretroviral therapy (HAART) and certain ARVs in particular are associated with dyslipidemia and even cardiac events. Data from the D:A:D cohort have been particularly instrumental in identifying these associations. What are the key relationships and what new data in this area from D:A:D and other surveillance studies were presented at the 14th CROI?

Dr. Aberg: I think one of the things that has come out through the years is how much traditional risk factors really do play a role in cardiovascular disease in HIV-infected people. Where it's gotten more complex is because the questions still remain: Is it HIV? Is it antiretroviral therapy? Or is it traditional risk factors, and by that I mean host genetic factors. And, in fact, all of these factors do play a major role.

The D:A:D study really was the first study where people recognized how much the traditional risk factors really played a role. Although combination antiretroviral therapy plays a role in the development of these abnormalities, not everybody gets these side effects. So it's not an all-or-none phenomena. In certain individuals, certain combinations lead to these abnormalities. Genomic studies are pointing towards the idea that certain people are genetically predisposed to developing insulin resistance or these dyslipidemias. In particular, the D:A:D study has shown that there is an increase [in cardiovascular events] associated with combination antiretroviral therapy and pointing a little bit more towards long-term PI use. The biggest contributing risk factor for cardiovascular disease is smoking.

At this CROI, one of the things that the D:A:D study looked at was whether the combination of abnormalities that constitute the metabolic syndrome is a greater predictor of cardiovascular disease than each individual risk factor for coronary heart disease. What they found was that it wasn't -- that it was just like in the general population. The INTERHEART study showed that it's really the summation of multiple risk factors. The INTERHEART study in the general population showed that if you smoke and you have insulin resistance or diabetes plus abnormal cholesterols, then you're more likely to have cardiovascular disease. The D:A:D study essentially agrees with what's been found in the general population: that having multiple risk factors increases your risk of having a coronary event.

There was also a study that was looking at the utilization of lipid-lowering drugs [reported at CROI]. There are 2 ways to look at this: One, I think that HIV experts actually have done a really great job of recognizing the metabolic complications of HIV and its therapies, and have been using a lot of preventive strategies; nevertheless, the D:A:D study showed that we're still not doing enough and that there were still a significant number of individuals at high risk for cardiovascular mortality. What they found was that only one fifth of individuals who have diabetes were receiving primary prevention for lipid-lowering therapies. So we still have a ways to go, and I think that's important. On the other hand, we have made significant progress. The utilization [of preventive modalities] overall has increased tremendously, but we still need to keep going and recognizing these individuals, and getting them on appropriate cardioprotective medications.

Medscape: One possible approach to limiting the cumulative, or long-term, adverse effects of ARVs is to use strategic treatment interruptions to limit exposure to these agents. At last year's CROI, many were surprised by the results of one such study, the SMART study. What did that study show with respect to clinical outcomes and what were the key data and most important take-home points regarding cardiovascular disease in particular from this year's CROI?

Dr. Aberg: Let me just go back a little bit to explain what SMART is: The SMART study was really a CD4 cell-guided intervention looking at whether you could discontinue therapy when individuals had higher CD4 counts, with a plan to restart antiretrovirals when their CD4+ T cells declined to less than 250[/microliter(mcL)]. The control arm remained on therapy and were virologically suppressed, and the other group discontinued therapy when their CD4 counts were greater than 350[/(mcL)], with plans to restart when they were less than 250[/(mcL)].

Last year's news was that in the subjects who had discontinued therapy, there were higher cardiac-, renal-, and liver-disease-related events -- in other words, not HIV-related events, which people were expecting. The original hypothesis of this study was that it was the combination antiretroviral therapy that's been associated with development of cardiac events, and they expected that it would be the people who remained on antiretroviral therapy who would have more events, not the group that discontinued it. This has led to lots of controversy, with people asking: "Well, what's going on in this group?" Some people have thought that part of it might be that when you stop the antiretroviral therapy you have viral rebound, and that leads to inflammation.

At this year's CROI, Andrew Phillips, who is one of the SMART study statisticians, presented data on the SMART study analysis. Although there were trends, it actually wasn't statistically significant; each parameter bordered on the discontinuation arm having more events. I think we're starting to get a little bit clearer picture, and it's not as obvious as last year when the conclusion was that discontinuation was really a bad thing, and people were having more events. This time, everything was borderline significant. But they did look to see whether it was the higher viral. If so, as I just suggested, maybe it's viral rebound, and it would be the virus that's leading to these events. They didn't find that a higher viral load was associated with increased risk. They didn't find that being off HAART for a timed duration was associated with increased risk either; it was of borderline significance. They found that among those who were on HAART at baseline, the hazard ratio was higher in those who were on nucleosides [nucleoside analog reverse transcriptase inhibitors, NRTIs] only or nonnucleosides [nonnucleoside analog reverse transcriptase inhibitors, NNRTIs] compared with those on PIs. I think we're still in a very unclear zone with what was going on. Dr. Phillips then looked at total cholesterol and HDL ratios to see whether that could explain the results. And again, there was a hint that there were alterations that led to an unfavorable change in that ratio for those who stopped antiretroviral therapy. Again, it seemed that it was more significant in those that were on [NRTIs] only. I think we had a little bit more of a borderline result presented this year when it was really looked at more closely compared with last year.

The other thing to remember about these results is looking at what types of events there are, and we're looking at a small number of events: There were 48 events in the discontinuation arm vs 31 in the control group. Overall, I think we need to remember that it's a small number of events. To get an increased number of events, they looked at individuals that had EKG changes, and they called this group the silent MI [myocardial infarction] group, meaning that the patients were asymptomatic, and it was based only on these EKG changes. Some of us have questioned whether those EKG changes really are significant or not, because we don't have patients who are symptomatic, and I don't really know about hospitalizations or true [clinical] events in these individuals. So my question throughout the year has been: What do those EKG changes really mean? Many individuals will have nonspecific ST segment changes or small Q waves, and does this mean anything or not? I think we still don't know the answer to that, and until we do, I'm a little reluctant to say that these EKG changes are significant.

Medscape: Information about the metabolic effects of selected classes or individual ARVs is one way to guide treatment choices when options are available. The ACTG 5142 study is a class-sparing strategy trial that provides data on questions such as these. Data from this trial were presented at CROI. What were the key features of this study design and the most important results?

Dr. Aberg: The ACTG 5142 study probably received most of the press, and people were very surprised by the results. Briefly, this was a study that randomized individuals to a boosted lopinavir regimen or an efavirenz-based regimen, both with lamivudine [3TC] plus d4T [stavudine extended-release], AZT [zidovudine], or tenofovir, or a [NRTI]-sparing arm, in which people received boosted lopinavir plus efavirenz. It's important retrospectively to remember that the worst lipid abnormalities we have ever seen were in combinations of efavirenz with a PI. In an initial study that Merck did years ago, indinavir/efavirenz was associated with lipid abnormalities; and in a study that BMS did, atazanavir with efavirenz was associated with significant lipid abnormalities. I think it's important to remember that sometimes it's not just about an individual drug, it's about drugs that they're combined with. If you think about the initial Gilead trials, there were significant differences in lipids in individuals who received d4T with 3TC and efavirenz vs those who received tenofovir, 3TC, and efavirenz. Over time you see that initially people were blaming the PIs for all the lipid abnormalities, and then, lo and behold, we started seeing that it wasn't just the PIs -- that it really could be dependent on the NNRTI or on which [NRTI]. Certainly, d4T has been the one that's been most associated with lipid abnormalities among the [NRTIs]. Taking that into consideration when you look at [ACTG] 5142, in fact, the worst lipids occurred among subjects who were randomized to receive boosted lopinavir and efavirenz. What everyone was startled about was that the efavirenz arm showed more lipoatrophy. I want to be cautious about that term, and I think we need to understand that the lipoatrophy was defined as having greater than a 20% limb fat loss. In this particular study, the subjects coming in actually had median limb fat of about 7 kg, which is actually normal size -- these individuals came in pretty healthy as far as limb fat. What was seen is what we've seen in other studies: At 48 weeks, everybody's limb fat increases. People always say, "Well, this is a return to health." People get better, they're eating more, they gain weight. What you see, though, is that, in the boosted lopinavir arm, the median percent change is an increase of 18% at 96 weeks; the boosted lopinavir with efavirenz arm plateaus at 48 weeks, with a slight curve upward, and ends with an increase of about 9.8%; and the efavirenz arm goes up then down, and it ends at 1.4% above baseline. Although people are saying that there's lipoatrophy, it's in a subset of individuals; and remember -- overall the [treatment] arm is still above baseline. In fact, they didn't really develop lipoatrophy; they still have normal limb sizes. I think that's an important message because I think people are just looking at these curves and saying that these people have lipoatrophy, and they don't. They have a statistical difference, and they have lipoatrophy by this definition that's used to determine statistical significance, but it wasn't clinically significant.

When you break down the data by [NRTIs] and you look among the [treatment] arms, you do see, as you would expect, a certain proportion of individuals who did have lipoatrophy (by that same definition) on the d4T plus efavirenz arm. It was less so with efavirenz plus AZT . Overall 12% of the individuals on efavirenz plus tenofovir had lipoatrophy compared with 6% of those on boosted lopinavir, so that result was not statistically different. There was no difference in development of lipoatrophy between boosted lopinavir and efavirenz in the individuals who were taking tenofovir. I think that's important because I think it actually confirms what we've seen before -- that it's really a matter of what you're combining things with. Previous studies, such as the Gilead study I mentioned before, had shown before that d4T with efavirenz was associated with a significant increase in triglycerides. In some ways, when you actually take a step away and you look at the results, it's really not as surprising as I think many people believed. The question still remains, though: Why do we see with d4T and efavirenz a significant decline or backing down [after the initial gain], whereas, in the boosted lopinavir arm, you are continuing to see fat gain? I think that's an intriguing question.

Questions that have been asked include: What are the effects of PIs on adipocytes? What are the effects of efavirenz on adipocytes? We really haven't been looking at those things. We've been focusing on mitochondrial toxicities with the nucleosides, but nobody has really looked at these new questions. We do know that the effect on the mitochondria has to do with the DNA polymerase gamma, and efavirenz and the PIs don't interfere with that . It doesn't mean that there couldn't be a second mechanism that we're unaware of and we haven't explored. The other observation that is pretty consistent through most studies is that individuals do gain more CD4 cells on a PI [-based regimen] compared with nonnucleosides [in HAART regimens]. So one of the other questions could be: Is an inflammatory component involved here? Are there differences between efavirenz and PIs in the immune system response? Are there differences in cytokines that are produced? Why is it that there is more of a blunting, if you want to consider it that way, of CD4 count responses among nonnucleosides compared with what you see with PIs? I think these are things that we still don't understand about the pathogenesis of why people gain fat or lose fat that we still need to understand. I think the results are very interesting, but I think now that I've had a week away and I'm looking at this, it's not as unexpected as my initial reaction was when I first heard the results.

Medscape: Were there any other studies that showed significant findings regarding metabolic effects of individual ARVs-- particularly the newer, first-in-class ARVs?

Dr. Aberg: That includes the integrase inhibitors and maraviroc, which is a new CCR5 entry inhibitor. We haven't really seen any metabolic effects yet with these drugs, and I think time will tell; we're still young at exploring these. If you think back when we first released the other ARVs, we weren't so aware of metabolic side effects with these drugs. But so far we haven't been seeing lipid abnormalities with the integrase- or CCR5 inhibitors, so things are looking very favorable for both of those classes from a complications point of view.

Medscape: The demographics and epidemiology of HIV infection continue to evolve. What are some of the most significant changes that have occurred recently, and what are the implications with respect to metabolic changes and cardiovascular risk?

Dr. Aberg: We have been seeing more minorities with HIV infection in the United States and more infections occurring in women and in older age. I always think it's interesting because -- especially the older age population -- we have 2 things going on: One is that our population is living much longer thankfully due to these medications. And two, I do think there are older individuals who are at risk of acquiring HIV at older age, particularly women. Older women typically were looking at condoms as a form of birth control, and now that they're postmenopausal, they're no longer concerned about birth control, so they don't really see a need to use condoms. People have a tendency to think that sexually transmitted diseases occur among those who are young: "I'm never the one that's at risk." They don't really protect themselves, and they're not really as aware of HIV as younger individuals are. I think they, unfortunately, are placed at greater risk. I know in my own clinic, I'm starting to see older women coming in who have newly acquired HIV. They may have gone through divorces; they don't perceive themselves to be at risk for HIV and therefore aren't aware of any of the risk-prevention strategies. We're also having a higher incidence in African Americans and Hispanics. We also have, in many cities (such as New York where I am), many immigrants coming in to the clinic. We need to understand the risks for cardiovascular disease for each one within their own ethnicities or races.

One study that I thought was very interesting is the SUN study, which is the study to understand the natural history of HIV and AIDS. This study has been looking at carotid intima thickness and coronary artery calcium scores, and has been following these individuals for a couple of years. They found that a significant number of Hispanics have diabetes, which matches what we're seeing in the general Hispanic population, yet they're not seeing coronary events in that population. One of the questions I had for Kristin Mondy [MD, Washington University School of Medicine, St Louis, Missouri], who is one of the lead investigators for the SUN study, is: Why is that? Because diabetes, as we know, is traditionally considered a high-risk factor for coronary heart disease. In the Framingham risk scale, having diabetes gives you a greater than 20% risk of having an MI in the next 10 years, yet we're not seeing events nor were we seeing increased carotid intima thickness in the Hispanic population. I'm not sure how to interpret that because I would have assumed that Hispanics would be at increased risk for cardiovascular events. African Americans have a higher prevalence of hypertension as well as diabetes, and we're seeing more and more African Americans being infected with HIV. Certainly in this population, there is a concern that we're going to be seeing more cardiovascular events.

I think it comes back to what I was talking about before with the D:A:D study about lipid-lowering interventions. Although I think we're doing a fairly good job of preventive medicine and recognizing that there can be cardiovascular complications, we have to make sure that we actually implement these preventive measures. Probably the most important thing that we can do for our population is to get people to stop smoking, because study after study has shown that smoking is much more prevalent among those with HIV infection compared with those without HIV infection. We could achieve a tremendous benefit for our population by starting smoking-cessation programs -- not only encouraging people to stop smoking but also helping them get into the various support groups that we now have. There are also various medications that are available to help people stop smoking. I think that probably the biggest impact we could have is to get people to stop smoking and then prescribe lipid-lowering therapies for those who need it, maintaining good blood pressure control, and monitoring for diabetes.

Supported by an independent educational grant from Bristol-Myers Squibb.

  • Print