HEAT: Virologic Failure Rates Similar (and Low), Emerging Resistance Uncommon on Abacavir/Lamivudine vs. Tenofovir/Emtricitabine
An Interview With Benjamin Young, M.D., Ph.D.
October 26, 2008
There's nothing like hearing the results of studies directly from those who actually conducted the research. In this interview, you'll meet one of these impressive HIV researchers and read his explanation of the study he presented at ICAAC/IDSA 2008.
This is Ben Young from the University of Colorado and the Rose Medical Center in Denver. I'm presenting a poster here today on the 96-week drug resistance and virologic data from the HEAT study.1 The HEAT study was a study in therapy-naive patients, all of whom received once-daily lopinavir/ritonavir [LPV/r, Kaletra]. They were randomized in a double-blind, placebo-controlled fashion to receive either abacavir/lamivudine [ABC/3TC, Epzicom, Kivexa] or tenofovir/FTC [TDF/emtricitabine, Truvada].
The primary data from these studies at 48 and 96 weeks were presented by Kim Smith, most recently in the Mexico City meetings [i.e., at the XVII International AIDS Conference].2 In brief, her data showed that at 96 weeks, the two arms of the study, namely abacavir/3TC and tenofovir/FTC, produced very similar and statistically non-inferior results. The side effect profile and toxicity profile were also very similar between the two groups.
This analysis focuses on the virologic failure characteristics of the patients. Approximately 343 patients were randomized in each arm of the study. Forty-nine [patients on abacavir/lamivudine] and 48 [patients on tenofovir/emtricitabine], or 14% of patients overall, had protocol-defined virologic failure and there were similar proportions of patients in both study arms who had confirmed virologic rebound and failure to achieve virologic suppression.
When we looked at virologic failure characteristics stratified by baseline viral load, we saw that 41% of abacavir patients failed with baseline viral loads less than 1,000, compared to 63% in the tenofovir arm. Conversely, there were more virologic failures among high viral load patients with abacavir and fewer with tenofovir.
Was that unexpected?
I had very little in terms of pre-study hypotheses about what would happen. These data, in part, are consistent with the ACTG [AIDS Clinical Trials Group] 5202 dataset that observed more virologic failures among high [baseline] viral load abacavir patients.3 What needs to be stated for balance, though, is that there are more virologic failures [among patients] on the tenofovir arm who had a low viral load.
To me, what this means is that virologic failures were infrequent overall. We're really talking about differences between the two arms that represent small numbers of patients, so trying to ascribe either statistical or clinical significance to these things in a study of 700 patients is challenging. I think we'd be overreaching and over-interpreting the data to draw those conclusions.
What we can say that I think is reasonable, based on those patients who did have virologic failure, is that the overall number of patients who had treatment-emergent mutations in the study was very low. About half of the patients who had virologic failure had treatment-emergent mutations. Most of those mutations were observed in the nucleosides, which was not surprising. There was only one patient who had a major protease inhibitor resistance mutation emerge. We'll talk about that patient anon, because he's also very interesting.
We saw that 11 patients in the abacavir arm and 17 patients in the tenofovir arm developed mutations and substitutions at [codon] 184. This runs counter to other expectations that the Truvada arm would actually have fewer 184 mutations. In fact, we actually found the opposite: There were fewer in the abacavir arm. But again, the important caveat is that these were small numbers. Nevertheless, the vast majority [of treatment-emergent mutations] were associated with phenotypic resistance, confirming that these changes in the genes were significant.
There was [a finding that] I think makes some very important points, both in terms of the study management and clinical management. There was a single patient who developed major protease mutations at the time of virologic failure. This patient was in the abacavir/3TC arm. What has emerged from talking to the patient is that this patient was re-exposed to HIV at week 48 in the study, and he emerged with very high levels of multidrug resistance -- data that we'll present hopefully in the near future. This represents a clear case of superinfection with a multidrug-resistant HIV strain. This emphasizes the idea that, even when a patient is receiving effective therapy, safe sex and safe drug use practices should be continued.
The last part of the analysis looked at variables that predicted virologic failure. These variables include gender, race, drug use, mode of HIV acquisition and so on. We found a number of predictors in univariate analyses, but of course, univariate analyses should not be used to fully interpret effect -- in part because there are a lot of covariables. When all these factors are assembled into multivariable regression models, we found that black race, drug use and baseline CD4 counts were predictors of virologic failure.
I see baseline CD4 counts of less than 200.
Yes. It's an interesting contrast to the aggregate primary data that looked at the same virologic and CD4 count strata -- in which, using a dichotomous analysis, we did not see this outcome. I have difficulty at this moment reconciling this, because this is seen in some studies and not others. I will point out that more recent studies using Kaletra, namely the CASTLE study,4 have shown a decrement in virologic response in patients with a lower CD4 count -- again, using these kinds of nucleoside backbones, so maybe this is a more contemporaneous view of the effect of CD4 count on Kaletra-based therapies.
When you say drug users, you're talking about injection drug users?
Yes, this is injection drug use. Overall, what we found was that the two drugs performed very similarly in terms of aggregate virologic response. Virologic failures were very infrequent overall. There were subtle differences in the characteristics of those patients who had virologic failure; in the types of failures they experienced, meaning confirmed rebound versus never suppressed. Mutations emerged, with treatment failure mostly related to nucleoside mutations, consistent with previous observations with boosted PIs [protease inhibitors] -- we really did not see any protease resistance. The presence or the emergence of 184V mutations was similar overall, but more frequent among the Truvada-exposed patients.
What's the take-home?
The take-home message for me is that this is one of the largest, head-to-head studies of the two currently recommended nucleosides [according to U.S. Department of Health and Human Services HIV treatment guidelines], using a currently recommended third agent, showing very similar overall virologic response rates, virologic failure reasons and resistance mutations. It gives us, as physicians, the opportunity to individualize care to suit the characteristics of individuals, rather than whole populations of people at one time.
You talked about ACTG 5202. What should one make of the results from that study versus this study? Who should people believe?
First, I should cite a potential conflict of interest. This is a study that was sponsored by GSK [GlaxoSmithKline] and I'm an investigator in that study. That potentially imparts bias.
That said, I think different studies use different methodologies and have different patient populations. That's not to discredit one study over the other, but rather to say that there is non-agreement between different studies. Until we see the full display of the 5202 data, it's difficult to really drill down on that dataset.
Overall, this suggests to me that virologic response seems to be pretty good when using either nucleoside. The 5202 dataset is large. It's a credible organization. It needs to be thought through, but until I see confirmatory data that is consistent with 5202, I'm reluctant to draw a broad, sweeping conclusion about these two different drugs.
Would you say there's an exception when it comes to patients who have a very low CD4 count?
Well, this predictor that we noticed applies to both -- to all failures -- so it doesn't segregate between one drug or another. There's little data that impugns abacavir versus tenofovir in [terms of baseline] CD4 count. The question was largely raised through 5202 and the baseline viral load characteristics.
This transcript has been lightly edited for clarity.
This article was provided by TheBodyPRO.com. It is a part of the publication The 48th Annual ICAAC/IDSA 46th Annual Meeting.
Add Your Comment:
(Please note: Your name and comment will be public, and may even show up in
Internet search results. Be careful when providing personal information! Before
adding your comment, please read TheBody.com's Comment Policy.)