November 11, 2009
The other comment I would make is along the lines of what Trevor talked about with this issue of sequencing, because I hear this discussed a lot.
Historically, we thought about sequencing protease inhibitors when we used non-boosted PIs. For example, in the past we used nelfinavir [NFV, Viracept] first in patients, because it selected for a unique resistance pattern, which meant that there would be a variety of other options left to patients should they experience virologic failure due to nelfinavir resistance.
People have sometimes tried to extrapolate from that experience to current boosted PIs; for example, saying, "Darunavir/ritonavir is currently approved as a first-line regimen for treatment-naive patients as a once-a-day therapy. But, why would I want to use darunavir/ritonavir as my first PI, when it was developed as a salvage drug?"
The POWER studies looked at the efficacy and safety of darunavir/ritonavir in treatment-experienced patients.15,16 Some providers still believe because darunavir was used in treatment-experienced patients that if they use darunavir first, and the patient develops resistance, they will not have other PIs to use. This is a faulty argument, because the experience with all of these boosted PIs is that unless you really don't follow patients at all, and allow them to have ongoing viremia for long periods of time, they are not going to select for resistance.
The reason they are likely to experience virologic failure is if they don't take the meds consistently -- not because of emerging resistant mutations to protease. Again, this focus on the decision needs to be on tolerability and specific, unique issues towards the patient.
Bonnie Goldman: Are there particular patients that get special attention when crafting a regimen? You mentioned homeless people. But can you discuss patients with kidney problems, heart disease risks or mental health problems?
Trevor Hawkins: Clearly, those things are very important. So we can take a few of those and look at them.
The preferred nucleoside backbone in the current guidelines is tenofovir [TDF, Viread] and emtricitabine [FTC, Emtriva]. Clearly, tenofovir can and does reduce creatinine clearance a small amount in most patients. If patients have normal creatinine clearance and normal kidney function, generally, that small (8 to 10 mL/min) reduction in creatinine clearance is probably not significant.
However, if patients have preexisting renal disease, then it may, indeed, be significant. It may be a contributing factor to making the kidney disease worse. You can reduce the dose, or adjust the dose. I do that very liberally, because I think, with its very long intracellular half-life, tenofovir can quite easily be dosed every other day.
Nevertheless, if you have patients with preexisting renal disease, then it may be better to avoid tenofovir and use something else, which is why the debate over abacavir [ABC, Ziagen] becomes important. Because abacavir doesn't have any toxicity to the kidney at all, and was clearly a go-to choice for patients with preexisting renal disease.
Now we have some markers that suggest that abacavir may be related to an increased risk of myocardial infarction. Of course, heart disease is the leading cause of death of people with renal disease. So that has confounded the issue a little bit.
We're waiting to try and clarify this issue about abacavir and heart disease. The D:A:D data, data from Quebec, and other cohort datasets have suggested there is a relationship.17-20 While other cohorts have not found that relationship -- most recently, the VA [U.S. Department of Veterans Affairs] cohort.21
We can discuss those in detail, and argue which are the better cohorts and the better studies, but I think the bottom line is -- particularly with the lack of an understanding of the pathogenesis of why abacavir would cause an increased risk for myocardial infarction -- the debate is still raging and we don't really know the final answer. I think it's important that we do, partly because of this renal issue.
At the current time, my practice is to avoid tenofovir in people with renal disease, and to avoid abacavir in people with very high risks for heart attack. Because I think that the data suggests that if there is a risk, then it's in those patients.
Eric Daar: I think those are important issues, and they are difficult for us to deal with, both amongst us clinicians, as well as with our patients, because the data is imperfect. I agree with Trevor: We can debate forever how to interpret these cohort studies, but the bottom line is --with at least one very large cohort study, the D:A:D -- there is this signal that's been seen with abacavir and 3TC [lamivudine, Epivir]. And then we have the other datasets that haven't shown it.
So we're forced, I think, to share this information with our patients, and probably use it as best we can in clinical practice, much like Trevor suggested.
If there's a compelling reason to use abacavir/3TC, such as in somebody who is not a great candidate for tenofovir, I would freely use it. I might have a little more angst in doing so in somebody who has multiple cardiovascular risk factors. But even in that setting, ongoing renal failure in the face of a potentially nephrotoxic drug is a cardiovascular risk factor, as well.
I think we tend to lean towards tenofovir/FTC [Truvada], if all things are equal, because we don't have these concerns. But we need to recognize it's not because we know unequivocally that abacavir/3TC is a problem. We don't. We know it's a highly efficacious drug in people who are B*5701 negative -- which is a prerequisite for using it, since we can define those people as not being at risk for hypersensitivity reaction -- and that this drug has a long track record as being highly efficacious and very, very well tolerated.
It comes down to sharing all the information with our patients and individualizing based on unique factors for the patients, as to which nucleoside we start with. And always being prepared to make changes, as we monitor people carefully for the development of new toxicities or increasing toxicities -- such as in someone, perhaps, who has chronic renal disease.
I think it's worth mentioning that this whole debate, based on data from cohort studies, has extended beyond nucleosides to also some of the PIs. There are several cohort studies, including the D:A:D study, that have suggested that select PIs, such as lopinavir/ritonavir, may also be a risk factor for cardiovascular disease.22
Now, the data is not as strong, at least within the cohort studies. The actual relative risk isn't as great. But it is there.
One of the limitations we have in how to deal with that data is that we have not seen data yet telling us that the alternatives, such as atazanavir or darunavir/ritonavir, are better, just because they simply have not been able to do those analyses as carefully with those drugs as they have with lopinavir/ritonavir.
We've got a lot of things we're dealing with, and that we're struggling with. I continue to say the best thing we can do in our practice, and with our patients, is make sure we have a careful distinction between the things that we really know and what we think, so that we can be clear with our patients, and be consistent in making our recommendations, based on the various concerns that are unique to the individual. I'm sure more data will be coming from the existing cohorts and new cohorts over the coming years that will hopefully provide some additional clarity.
Trevor Hawkins: It might be worth mentioning pregnancy. We've got rather limited choices in pregnancy. Current recommendations are, usually, to use AZT/3TC [zidovudine/lamivudine, Combivir] and boosted lopinavir, which is really not the best-tolerated regimen these days. Pregnant women are already prone to experiencing nausea and other symptoms from pregnancy. We really do, I think fairly urgently, need approval of some newer and perhaps better-tolerated regimens for women who are pregnant. I think that's coming, but currently, AZT/3TC and boosted lopinavir is the preferred choice.
People with a history of mental health disorders might not be good candidates for efavirenz-based regimens, based on the CNS [central nervous system] side effects of efavirenz, particularly in the first four weeks. There are not actually a lot of hard data that say that you can't treat people with psychiatric disease with efavirenz; you can. But nevertheless, it's a cautionary issue and I think many people, particularly with patients with high levels of anxiety, prefer to use other regimens.
Then I think we should also mention patients who you have to treat right away -- for instance, those who present very late or present with an opportunistic infection [OI], such as Pneumocystis, where the data from ACTG 516423 clearly indicates that starting treatment concurrently with the treatment for this OI is preferred even though you haven't had time to really get to know these patients, and work out issues around their likelihood to adhere, or not to adhere, to a regimen. Those patients often get started on boosted PIs simply because of the higher barrier to resistance.
Eric Daar: There is another factor that I'm not sure we've mentioned yet, and that is the role of baseline drug resistance testing, which has been the standard of care now for some time. The frequency of baseline drug resistance ranges from 5% to 25% or 30%, depending on the community of transmitted drug-resistant virus.24,25 Ideally, that information should be available and incorporated into the thinking behind which regimen to start.
If you don't have that data available (such as in the situation Trevor was talking about, where you may want to start someone on treatment acutely along with an opportunistic infection such as Pneumocystis), you may choose drugs for which the likelihood of transmitted resistance is lower, such as choosing protease inhibitors versus NNRTIs, until drug resistance testing information becomes available.
Regardless of what we do -- and I've mentioned this before, and I think it's important to keep in mind -- this is not a one-step process. Often we do, and can, make adjustments. But we need to do it carefully, and patients need to be aware that all decisions about stopping therapy and changing therapy should be made after careful discussion with their provider.
Bonnie Goldman: Many providers are going off the guidelines and putting together novel regimens for treatment-naive patients. Do you think it's a good idea or a bad one? Dr. Hawkins?
Trevor Hawkins: The short answer is that, in general, I think it's a bad one. There may be good reasons to choose novel regimens. For instance, if you really don't want to give someone tenofovir because of preexisting renal disease, or if the patient has high-risk factors for cardiovascular disease, and you decide you want to do a nucleoside-sparing regimen, then I think that's legitimate.
But to do it willy-nilly -- in other words, to start using off-label regimens without a clear indication or need for them -- I think is not very responsible. Because the regimens that we have now are generally very well tolerated, and have been shown to be highly effective. I think to start experimenting with regimens at this stage of the game is generally not a very good idea.
Eric Daar: I completely agree. There are some historical disasters that nobody would have predicted. Perhaps the best example, of course, is the triple nuke [NRTI] regimen that included tenofovir, 3TC and abacavir, which resulted in extraordinarily high levels, not just of virologic failure, but of, essentially, pan-nucleoside resistance amongst those who experienced failure.
While people might say, "But that was triple nukes, and we wouldn't do that," at the time, there was no reason to believe that that triple nuke regimen would perform as poorly as it did. I think that's a legitimate concern any time you use a novel regimen that hasn't been thoroughly investigated in a clinical trial.
I completely agree with Trevor that if you get backed into a corner, sometimes you need to do what's best for your particular patient. But it should only be done in the setting where you have been backed into a corner, where all efforts have been made to work with the patient to start on a tried-and-true, well-tested regimen. Until that's happened, I think that we really should avoid these novel combinations for first-line therapy, when patients have so much at stake, and they are going to have such a long future with therapy.
Bonnie Goldman: What do you think about strategies like induction maintenance? Is that still popular?
Trevor Hawkins: In the past, the results with induction maintenance weren't very good, because partly, the maintenance regimens were not very effective. But I think it is a potential strategy in patients who develop a toxicity to part of the drug regimen.
There is a thing that I always used to do that I don't do anymore, because the most common resistance pattern we see now is K103N and 184. It's rare now to see patients with multiple drug resistance, although they do occur. In those patients I would treat with a boosted PI -- usually this is after an Atripla failure, an efavirenz failure. I would treat with a fixed-dose combination of tenofovir and the cytosine analog FTC, and add a thymidine analog until the patient was undetectable; then I would drop the thymidine analog, because of long-term toxicity issues.
In general, these days, I don't do that. I would treat that particular failure with just the boosted PI and tenofovir/FTC. I think there is not a lot of data -- in fact, there's really no randomized data -- on how to treat that particular set of mutations. But induction maintenance might still be a strategy for patients who fail an efavirenz-based regimen with a very high viral load, for instance.
Eric Daar: What Trevor is talking about, I think, is a group of patients who fail a first-line regimen with an NNRTI, who are then going to get a boosted PI. Then the question is: How many other drugs do you need to use with it? And I agree. I think the amount of experience that we have in doing this now with simply two nukes, like tenofovir/FTC, with a boosted PI, suggests that, for most patients, that is sufficient to fully suppress virus.
Certainly, the experience in small randomized control trials, even, with boosted PIs alone in PI-naive patients would suggest that these drugs are going to contribute a lot of activity, and that they may not need two other fully active drugs to completely suppress the virus.
There was recently a retrospective review of data from Canada, where they were actually able to show that patients who receive just the two nukes, like you propose with the boosted PI, actually did as well as those who received additional drugs, along with two nukes and a boosted PI.26 So I agree.
But I think where maintenance has been best tested, with the most promising results, has probably been with boosted PIs, for which there has now been numerous studies taking patients who start on a triple drug regimen that includes a boosted PI, and then dropping the nukes.
The one constant in these studies seems to be -- again, not in every single one of them -- that while the boosted PI alone in this type of strategy seems to be very effective, with a relatively low risk of selecting for PI resistance, there tends to be more people with low-level viremia.
It leaves us wondering what we need to worry about in those people, as far as selecting for resistance over long periods of follow-up, and whether there's enough data to suggest that that's a viable strategy, considering the limited concerns of current nucleoside combinations, which come as one pill, once a day. Certainly, it's lower cost and, for select patients, where there are unique toxicities that may push us away from nucleosides, it may still be an option.
But I'm not sure that the data is sufficient to justify this as a routine strategy in somebody where you don't have a really compelling reason to stop nucleosides.
Trevor Hawkins: Right. And also, I would add the ARIES study, which showed that maintenance with unboosted atazanavir, after patients had suppressed on a boosted-atazanavir regimen, seemed to work quite well.27 Particularly, that would be a reasonable strategy for patients who are having a tough time with their low-dose ritonavir.
Eric Daar: I agree. I think that was also a very important study, and does provide an option for people who are having difficulties with ritonavir -- which is not that common, but it is a situation that occurs. Then it's useful to have that kind of data to help us make decisions.