More Patients With High Viral Loads Experience Virologic Failure When Starting Therapy With an Abacavir-Based Regimen Than a Tenofovir-Based Regimen

An Interview With Paul Sax, M.D.

My name is Paul Sax. I'm the clinical director of the HIV program at Brigham and Women's Hospital in Boston, and an associate professor of medicine at Harvard Medical School. Today I presented the results of the ACTG 5202 study comparing tenofovir/FTC [tenofovir/emtricitabine, Truvada] and abacavir/3TC [abacavir/lamivudine, Epzicom, Kivexa].1 It was a comparison of those two drugs in patients with viral loads greater than 100,000 copies/mL. Here are the basics of the study.

It was a four-arm study. Over 1,800 patients were randomized to one of four regimens: They either got abacavir/3TC plus tenofovir/FTC placebo, or tenofovir/FTC plus abacavir/3TC placebo. That was the nucleoside component. Then the third drug was either efavirenz [EFV, Sustiva, Stocrin] or boosted atazanavir [ATV, Reyataz]. The study started in 2005 and completed enrollment in November of 2007, and we expected to complete follow-up of the patients in the fall of 2009.

Paul Sax, M.D.
Paul Sax, M.D.

In January of 2008, the Data Safety Monitoring Board [DSMB] reviewed this study and found an increased number of virologic failures in the abacavir/3TC arm.2 They said to us, "Instead of doing a four-arm comparison, why don't you combine it into two arms, tenofovir/FTC versus abacavir/3TC? Also, we would like you to unblind the nucleosides." So we did that.

The results that I presented today were really that analysis of the nucleoside comparison in the greater than 100,000 viral load stratum. Patients were stratified at baseline by their screening viral loads. At baseline, the groups are pretty well matched. The one thing I want to mention is that only about 43% of patients were genotyped at entry. They had to be genotyped if they had recently acquired infection, or they got genotyped because it was standard of care at their site, but it otherwise wasn't specifically required.

Our primary result -- which was time to virologic failure -- found that there was an excess number of virologic failures in the abacavir/3TC arm versus the tenofovir/FTC arm: 57 failures in the abacavir/3TC arm; 26 in the tenofovir/FTC arm and that was highly statistically significant at P = .0003. There was a hazard ratio for failure of over 2, for abacavir/3TC. We did a bunch of other analyses and they were all pretty much consistent.

A couple of things to mention was that virologic failure happened regardless of whether it was early in the course of the treatment, or later in the course of treatment. Also, importantly, it did not seem to be caused by suspected abacavir [ABC, Ziagen] hypersensitivity. There was 7% suspected hypersensitivity in both study arms, and most of the patients who had suspected abacavir hypersensitivity did not have virologic failure.

The other finding in the study, which was unanticipated, was that there was a shorter time to the first adverse event in the people who were receiving abacavir/3TC. Again, this was highly statistically significant. Most of these adverse events were metabolic in nature, with increases in lipids. It's unclear whether these are clinically significant, but nonetheless, there were also other side effects, such as body aches and pains, etc., that were seen more with abacavir/3TC than with tenofovir/FTC.

There was one study death that was possibly related to abacavir/3TC after a suspected HSR [hypersensitivity reaction] and then rechallenged. Clearly, the study site did not realize that it was suspected HSR, and that's how that ended up happening. We did do extensive education, both before the trial started, during the trial, and of course, after this event, about how to manage abacavir hypersensitivity. It was not standard of care to do HLA-B*5701 testing when the study started, and we only started doing it towards the end of the study, after the presentation of the PREDICT data last year.

That's in essence what the study showed: that there was a shorter time to virologic failure for these patients with high viral loads if they got abacavir/3TC versus tenofovir/FTC; that the hypersensitivity issue did not seem to explain the difference in the results; and finally, that more side effects were going on in the abacavir/3TC arm than in the tenofovir/FTC arm, and they happened sooner.

The rest of the study keeps going. We're still going with the lower viral load stratum. We were assured by the Data Safety Monitoring Board that there didn't appear to be any kind of toxicity signal or efficacy signal that should make us stop. We're also continuing with the efavirenz versus the ritonavir [RTV, Norvir]-boosted atazanavir comparison. As I said, the study will be completed in the fall of 2009.

What's the take-home? What's a practicing physician to do? And what's a patient to do, if he or she is on Epzicom and doing fine?

One post hoc analysis we did, which addresses that issue, showed that if patients have undetectable viral loads, measured at less than 50 copies/mL in our study, they did not have, at least at this point, an increased risk of virologic failure, whether they were on abacavir/3TC or tenofovir/FTC.

That is an important sort of practical message for physicians and for patients, that if they have an undetectable viral load, they are tolerating it well and there are no other compelling issues to make them want to change, then they should stay with what they are on.

The study, though, does address the issue of people who come into care with very high viral loads. I have to say, based on our prospective randomized double-blind study, which is kind of the best evidence you can get, it does appear that tenofovir/FTC is more effective virologically than abacavir/3TC, at least when combined with either efavirenz or boosted atazanavir.

How do we put this in context with respect to renal insufficiency, which seems to be quite common among the patient population?

Clearly, there's no drug without side effects, and every clinically active HIV provider knows that there are patients for whom tenofovir [TDF, Viread] nephrotoxicity can occur. In this particular study, we didn't see it at this point. There were only two cases of renal failure in both study arms. But in patients who have preexisting renal disease, who are taking other nephrotoxic drugs, tenofovir might not be the best option for them, and one might still want to choose to use abacavir/3TC, especially since we now can test for HLA-B*5701, and essentially eliminate the risk of severe hypersensitivity reactions.3

Were there hints before of abacavir's weaknesses, in the Trizivir [zidovudine/lamivudine/abacavir, AZT/3TC/ABC] study?4

You can't really compare the Trizivir study to this study, because the Trizivir study, of course, used an all-nucleoside approach. But this study did show something that the Trizivir study also showed, and that is that people with high viral loads tended to have less of a response than if they got other standard triple therapies.

What we didn't see, but was seen in the 5095 data,4 was that once patients got viral loads below 50 copies/mL, they had a greater risk of rebound. At least, we didn't see it yet. As I said, the study is still ongoing and we're collecting data even as we speak. One possible theoretical explanation for why tenofovir/FTC might be better than abacavir/3TC has to do with the 184V mutation. That mutation is, of course, very common. The 184V mutation does cause a loss of susceptibility to both 3TC [lamivudine, Epivir] and abacavir, but does not cause a loss of susceptibility to tenofovir.

Finally, doesn't this speak to the idea that we're going to be customizing therapy for patients so much more with every drug?

Absolutely. I think you have got to look at the patient, look at all the different characteristics. It might be that the studies going forward looking at nucleoside-sparing regimens might be the best for the kind of patient I described before.

Let's say you have a person with underlying renal disease and he or she needs to go on a regimen. It might be that a PI [protease inhibitor] plus an integrase inhibitor is the patient's best option. So I think you are clearly right. Every patient has to be considered individually. There's no one-size-fits-all. And we have such good medicines now that there's no reason why we can't choose the best thing for our patients.

One thing I do want to emphasize is that the response rates, even in the abacavir/3TC arm, were quite good overall, and that we were able to follow patients even after they had virologic failure or switched their regimens. What is encouraging to see is that, over time, all the curves sort of go together. So you get the sense that both first-line and second-line treatments now are so effective. So long as people are able to take their medicines, they should be able to be treated successfully.

Thank you very much.

You're welcome.

This transcript has been lightly edited for clarity.


  1. Sax P, Tierney C, Collier A, et al. ACTG 5202: shorter time to virologic failure (VF) with abacavir/lamivudine (ABC/3TC) than tenofovir/emtricitabine (TDF/FTC) as part of combination therapy in treatment-naive subjects with screening HIV RNA ≥ 100,000 c/mL. In: Program and abstracts of the XVII International AIDS Conference; August 3-8, 2008; Mexico City, Mexico. Abstract THAB0303.

  2. NIAID modifies HIV antiretroviral treatment study: combination therapy that includes ABC/3TC found less effective in subgroup of antiretroviral-naive individuals. Bethesda, Md: National Institute of Allergy and Infectious Diseases; February 28, 2008.

  3. Wannamaker P, Mallal S, Phillips E, et al. HLA-B*5701 screening reduces abacavir hypersensitivity reactions (ABC HSR); multiple lines of evidence to establish a new standard of care. In: Program and abstracts of the XVII International AIDS Conference; August 3-8, 2008; Mexico City, Mexico. Abstract TUPE0068.
    View slides: Download PDF

  4. Gulick RM, Ribaudo HJ, Shikuma CM, et al, for the AIDS Clinical Trials Group Study A5095 Team. Triple-nucleoside regimens versus efavirenz-containing regimens for the initial treatment of HIV-1 infection. N Engl J Med. April 29, 2004;350(18):1850-1861.

Download the audio file