Jens Lundgren, M.D., Discusses New Findings Regarding MI Risk of Specific Antiretrovirals

An Interview With Jens Lundgren, M.D.

While at CROI 2009, I had the chance to talk with Jens Lundgren, M.D., the chief physician and director of the Copenhagen HIV Programme. Dr. Lundgren presented new data at this conference showing that specific antiretrovirals may be associated with a significantly increased risk of myocardial infarction (MI). I asked him to summarize his data and discuss the ramifications of the results. With me is Jeff Berry, editor of the Chicago-based HIV magazine Positively Aware.

Jens Lundgren: The D:A:D study is a prospective cohort study of around 33,000 patients whom we have followed now for the last nine years. The purpose of the study is to assess whether there is an association between the drugs we're using to treat HIV and cardiovascular disease. The primary outcome to assess is myocardial infarctions (heart attacks).

Jens Lundgren, M.D.
Jens Lundgren, M.D.

We have been observing for several years that there's one drug class -- namely the protease inhibitors [PIs] -- that is linked to an excess risk of myocardial infarctions. And then last year, much to our surprise, we also found that a drug from another drug class -- the nucleoside abacavir [ABC, Ziagen] -- was also linked to an excess risk.

This year, we can report on the further follow-up of those many patients.1 The main conclusion is that we can now identify that the abacavir signal is clearly still there; it appears specific to abacavir and not to other drugs that are used in the same way, specifically tenofovir [TDF, Viread]. With tenofovir there was no signal. We didn't have enough data on that last year, but we can present that now. It is reassuring that it is specific for abacavir and not for the whole class. That's one important outcome, I think.

NRTIs and Risk of MI: Recent and Cumulative Exposure

Click to enlarge

Jens Lundgren et al. CROI 2009; abstract 44LB. Reprinted with permission. Download the full slide presentation.

The other important outcome is that we had, for the protease inhibitors, only looked at protease inhibitors as a class. But now we have enough power to look at some of the protease inhibitors by themselves, and we can see that some of them clearly are linked [to myocardial infarction risk], as we would expect.

Specifically, Kaletra [lopinavir/ritonavir, LPV/r] is linked to an excess risk of myocardial infarction. We don't have enough information for some of the newer protease inhibitors, in particular atazanavir [ATV, Reyataz], so we'll need to wait some years to accrue more information on that. But we are confirming that Kaletra is associated with an excess risk.

The third main finding is that we have now been able to look at the non-nucleosides, efavirenz [EFV, Sustiva, Stocrin] and nevirapine [NVP, Viramune], and can confirm what we've seen before for the class: If we look at the two drugs individually, neither of them is associated with an altered risk of myocardial infarction. So those are the three main findings that we're reporting this year.

Bonnie Goldman: Can you quantify the risk?

PIs/NNRTIs and Risk of MI: Cumulative Exposure to Each Drug

Click to enlarge

Jens Lundgren et al. CROI 2009; abstract 44LB. Reprinted with permission. Download the full slide presentation.

Jens Lundgren: The protease inhibitor risk is expressed in how many years you've taken the drug. The longer you've taken the drug, the more important this risk becomes, on a relative term. It's around 13% per year, but it's like interest in your bank account: longer you've been on the drug, the bigger this problem becomes. So after, for example, five years, it's around an 80% to 90% increased risk. On a relative scale.

Putting it in the context of a patient's underlying risk is very, very important, because if your underlying cardiovascular risk is very low, then even if you were to double that risk, it would remain very low. On the other hand, if your underlying risk is elevated -- because you're older, you have diabetes, you have hypertension, you smoke, you have elevated cholesterol levels, for example -- then a doubling of that risk becomes clinically important. That's the protease inhibitor signal.

The abacavir signal is essentially what we've been reporting earlier. It's a little different with abacavir because the abacavir problem that we're seeing emerges very quickly: Whereas the protease inhibitor signal takes years before it really becomes a problem, the abacavir signal is more immediate, and then it sustains as long as you stay on the drug, in the same magnitude -- around a doubling [of risk] after five years.

Bonnie Goldman: So the risk doesn't dissolve over time.

Jens Lundgren: No, unfortunately. That was the hope, that if you're OK after a year, that you'd stay OK. We know for hypersensitivity reaction, if you don't develop hypersensitivity reaction within three months after you've started on abacavir, you're OK -- you won't develop that. But that doesn't seem to be the case, unfortunately, for myocardial infarction.

Bonnie Goldman: What are the patient characteristics? Did you determine who had a particularly high risk?

Characteristics of Patients at Time of MI/Last Follow-Up

Click to enlarge

Jens Lundgren et al. CROI 2009; abstract 44LB. Reprinted with permission. Download the full slide presentation.

Jens Lundgren: Yes. The median age when they develop the MI is around 48, 49 years. It's mostly men. There's certainly a predominance of people who have diabetes, hypertension or dyslipidemia, who develop the MI. It's the traditional, elevated-cardiovascular-risk patients who get this problem, as we would expect.

It's not very young people, and it's certainly not people when they start on therapy in their 20s and 30s. It is the aging HIV population. This is the challenge we're faced with as physicians: How are we going to deal with people as they enter an age range where we know, as with the general population, that there are issues around cardiovascular disease, cancers, renal disease and liver disease? We're not used to that, and we don't have enough information on that, and that's what we're trying to accrue. What will happen as we allow people to live longer, because we have effective drugs? That's what we're trying to do.

Bonnie Goldman: What percentage of the population smoked?

Jens Lundgren: It's around 40% of the population. To give you a little bit of an impression about this: Within a population of patients, you can calculate their underlying risk, and their underlying risk can be low, moderate or high. At the moment, around 25% of the population is at moderate or high risk. Of that 25%, only around 5% to 8% are at high risk. At the moment, in cardiovascular terms, HIV patients are still young. But of course, in five or 10 years, many will have aged into their 50s or 60s, which we know is the age group where you get myocardial infarctions even if you're not HIV infected. So we will see many more cardiovascular problems in the next 10 years compared to what we've seen so far. Again, the challenge for physicians is to find a way to deal with this in a reasonable and proper manner.

Jeff Berry
Jeff Berry

Jeff Berry: Your findings suggest that indinavir [IDV, Crixivan] and lopinavir are associated with a relative risk of MI. Any theories as to why those two?

Jens Lundgren: Over the years, we and others have speculated that maybe the reason why protease inhibitors were linked to an excess risk was because they can induce elevated cholesterol levels. We know that if you elevate your cholesterol level, then you increase your risk of myocardial infarction. The way to assess whether the indinavir and lopinavir signal was also explained by that is to take into account the lipid effect of the drugs. Once you do that, if the explanation is the lipids, you would expect the drug effect to pretty much disappear. Once you take the lipids into account in your analysis, there's no more drug effect.

LPV/r and IDV: Effect of Adjustment for Latest Metabolic Factors on Estimates of Risk

Click to enlarge

Jens Lundgren et al. CROI 2009; abstract 44LB. Reprinted with permission. Download the full slide presentation.

We tried to do that, but we couldn't get the drug effect to go away -- which is actually rather surprising. What it suggests is that the PI effect is not necessarily driven by lipids. It is something else.

You can then ask, "What is that something else?" I don't know, but we had a little bit of a clue in our presentation today. For two of the protease inhibitors, namely saquinavir [SQV, Invirase] and indinavir, you can either use those drugs without ritonavir [RTV, Norvir] or with ritonavir. We assessed the risk with and without. What ritonavir does is elevate drug levels. At the same time, it also induces more dyslipidemia. So if dyslipidemia was all of the problem, you would expect to see those who were using those drugs together with ritonavir to have a bigger problem than those using them without ritonavir. But we couldn't see that. This suggests that there is something else that this drug class is doing -- at least some of the drugs in this drug class are doing -- that affects and aggrevates cardiovascular disease. But not via lipids.

Unfortunately, our data do not tell us what that other mechanism is. But there's a lot of research going on at the moment that suggests protease inhibitors perturb the function of the cells, and that is the reason why these plugs in the arterial wall are building. Maybe that's the explanation. I won't speculate on that further today, but I think that's where the research needs to go, and where it has been going the last couple of years. We did present this the first time in a paper in the New England Journal of Medicine in early 2007, that we couldn't get the PI effect to go away -- we could reduce it, but couldn't get it to go away. I know that many people have tried to come up with explanations for why that is, so that will be interesting and important to follow.

Jeff Berry: I know you cautioned people not to draw any conclusions because of the type of study and other confounding factors, but what's the take away for physicians?

Jens Lundgren: Our position is that this is an observational study, it's not a randomized trial, so we're not able to claim that the drugs are the reason why people get excess risk. We can certainly try to address whether there are other explanations, but we haven't found other good explanations. That leaves us in a situation where we're not completely sure. But on the other hand, there's now pretty substantive, circumstantial evidence to suggest that this is a problem.

The correct way -- in my opinion, as a physician -- to deal with this is: As long as we can provide other drugs that do not seem to have this problem, we should probably try to do that if patients have elevated underlying risk. This comes back to the discussion we had earlier. If your underlying risk is low, I would say, "This is not really solid evidence, so we shouldn't react to that, really." But on the other hand, if you have a high underlying risk, where this is a really important health concern, and there are other drugs that are safer to give to a patient, then I think serious consideration should go into taking those other drugs instead.

You can then ask me, "What are the safer drugs?" That's a debate that, really, should be between the patient and the physician; it depends on previous drug history, resistance and other issues. I don't want to go into that. All the drugs have their own problems. So that's really a case-by-case discussion, rather than something I would want to speculate on.

Bonnie Goldman: Did you see an association between viral load levels and heart disease risk?

Jens Lundgren: No, we did not. I'd like to emphasize that we didn't see that, and the SMART study didn't see that. The SMART study suggested that untreated HIV led to an excess risk [of MI]. It wasn't a direct association between the virus and the outcome; it's something else. So that's a whole lot of discussion in terms of inflammation and HDL [high density lipoprotein] depletion; there's very interesting data from the SMART study on HDL depletion. But it's an indirect effect, so no, we did not.

Jeff Berry: You mentioned a theory you had, that there was potential for inflammatory reaction by a drug that could be exacerbating atherosclerosis.

Jens Lundgren: That's correct. After we came up last year with this abacavir finding, a legitimate question from many was, "So, this is surprising," -- and yes, it is -- "So, could you explain to us why this is?" We scratched our heads for quite awhile; we engaged in collaboration with the SMART study and analyzed that. And that's where we found the first clue that this could actually be the reason.

There are biomarkers reflecting inflammatory reactions that were, at least in SMART, higher in those on abacavir than on other drugs, which is circumstantial evidence. There'll be much more of that sort of data at this conference to inform that discussion. But if it was so, it shouldn't really be that surprising, because we know that abacavir is able to induce inflammation: In those that have specific genes, they develop hypersensitivity reaction to abacavir, which is clearly an inflammatory reaction.

I'm not suggesting [the MI risk on abacavir is] because of the hypersensitivity reaction, but if the drug can induce that type of inflammation in those who are genetically predisposed to that, maybe it can also induce other inflammatory reactions that may not show themselves as a hypersensitivity reaction, but may show themselves as just an elevated inflammatory response in the arterial wall. And we know if that occurs, if there are plugs in the wall, they will become more unstable, easier to burst, and result in what we know as a clinical myocardial infarction. At least, that's what our cardiology colleagues tell us. If that's how the drug works, that's a plausible way.

I'm not trying to close the book on that explanation, but that's where we are at the moment. I think this conference will be instructive, and I think the research field needs to work further on this. And we will.

Bonnie Goldman: Did you look into the impact of ritonavir [RTV, Norvir]?

Impact of Concomitant Use of Ritonavir on Association Between SQV and IDV on Risk of MI

Click to enlarge

Jens Lundgren et al. CROI 2009; abstract 44LB. Reprinted with permission. Download the full slide presentation.

Jens Lundgren: I've explained that the indinavir and saquinavir effects were not altered depending on whether the drug was used together with ritonavir or not. In my mind, that's really important information, because it suggests that there's something else these drugs are doing to the vascular wall that increases the risk of myocardial infarction, and that occurs in relatively low doses as well as high doses. So essentially you have saturated that system by overusing the drugs without ritonavir, and if you boost with ritonavir, they are saturated already, so it doesn't really matter -- what I usually say is, it's on the right side of the dose response curve. Again, what this is, I'm not sure I can give you a good explanation for. But that's what the data looks like.

Bonnie Goldman: Are you going to be doing something similar to what you did for the abacavir study: a question-and-answer document to address many of the concerns these findings may raise in the patient population?

Jens Lundgren: Actually, we don't have a lot of new information to bring to the table this year. We have confirmed the abacavir signal, so the statements we made last year in terms of abacavir still stand; we are now more comfortable that this is likely a true association.

In terms of the lopinavir signal, we did publish a paper in 2007 suggesting that the protease inhibitors are associated with an excess risk. In terms of the thinking around how to deal with that, it's exactly the same thinking as how to deal with abacavir. I'm not sure we could bring much that's new to the table, and we certainly don't want to be presumptuous and over-instructive, so I'll refer back to that position statement. That still holds.

Jeff Berry: In London at the lipodystrophy conference, I thought it was interesting that when you take away the abacavir, you take away the risk, so it's not cumulative. Would that hold true for PIs?

Jens Lundgren: For the PIs, we don't have enough information to inform that discussion. So I'm less comfortable about saying that. One would think that that would be the case, but we're not there yet in terms of the analysis and follow-up to address that very important question. So I can't give you a straight answer for that, sorry.

Jeff Berry: Thank you.

Bonnie Goldman: Thank you.


  1. Lundgren J, Reiss P, Worm S, et al, and Aquataine, AHOD, ATHENA, INSIGHT, EuroSIDA, ICONA, Nice, SHCS, St Pierre Cohorts & D:A:D Study Group. Risk of myocardial infarction with exposure to specific ARV from the PI, NNRTI, and NRTI drug classes: the D:A:D Study. In: Program and abstracts of the 16th Conference on Retroviruses and Opportunistic Infections; February 8-11, 2009; Montréal, Canada. Abstract 44LB.
    View slides: Download PDF

Download the audio file