icon- folder.gif   Conference Reports for NATAP  
  XVII International AIDS Conference
Mexico City
3-8 August 2008
Back grey_arrow_rt.gif
NATAP Summary of IAC 2008 Clinical Late Breaker Abstracts
  Andrew Zolopa, MD
Stanford University
Attendees of the late breaker session at this year's International AIDS Conference in Mexico City saw several new studies with results that will likely have a major impact on clinical practice. I will review the most important of the abstracts presented and provide some commentary, which reflects my own personal opinion of the data presented.
Another drug fails to provide relief from painful HIV-related peripheral neuropathy (PN)
Neurologist David Simpson, presented the results of a randomized placebo-controlled double blind study of the drug -pregabalin. Pregabalin has previously been shown in to improve both painful post-herpetic neuralgia as well as diabetic peripheral neuropathy and so there was hope that the drug would prove useful in the treatment of HIV-related PN. The study randomized HIV-1 infected patients with PN equally into a treatment group (N=151) that received an increasing dose of pregablin (increased to tolerance) versus placebo (N=151). Standardized pain scores were measured at baseline and over the course of the 14 week study. The treatment group was able to achieve a median dosage of nearly 400mg day which is comparable to doses used in earlier studies. Overall there was significant improvement in subjective pain over the 14 weeks of the study -however there was no difference between the treatment arm and the placebo arm at study end. In general, the pregablin was well tolerated although patients did have increased CNS-related adverse events compared to the placebo group.
Dr. Simpson put these trial results in context of the previous randomized studies using pregabalin and the overall result was that pain scores improved around the amount in the treated groups from the earlier studies. The difference in this study was that the placebo group had shown much better improvements in pain than in previous studies resulting in the negative over all result.
Comment: These results are obviously disappointing. Patients with HIV-related PN have very few effective non-opioid treatment options to control pain. Cannabis has been shown to significant reduce painful PN in one study from San Francisco but most other treatments have failed to show efficacy in randomized trials. Fortunately, fewer patients are developing PN in the modern treatment era that avoids the "D-drugs" like D4T, DDI and DDC. However, there are still significant numbers of patients with chronic debilitating painful PN that are desperate for a more targeted treatment than cannabis or opioids.
Is ABC/3TC less effective than TDF/FTC?
In the much anticipated presentation of ACTG A5202 early results from a DSMB-mandated interim analysis conducted earlier this year, Paul Sax from Harvard Medical School showed that fixed dose combination ABC/3TC was associated with a significant increased risk of early virologic failure when compared head-to-head with the fixed dose combination of TDF/FTC. A5202 is a Phase IIIB, randomized study of treatment naive HIV-infected adults. Over 1800 individuals were randomized to one of 4 treatment arms: ABC/3TC with either ATV/r or EFV compared to TDF/FTC with either ATV/r or EFV. The fixed dose combination nRTI are double blinded in this study while the 3rd drug is open label. This early analysis was the result of a DSMB review conducted in January of this year which showed that individuals entering the study with high viral loads (ie plasma HIV RNA >100,000 copies/cc) had a significant increased risk for early virologic failure. Although the primary study endpoints are at 96 weeks and the trial is still on-going - the DSMB mandated that patients in the high viral load strata have their nucleoside combination unblinded- the results based on the 3rd drug (ie ATV/r vs. EFV) and those of the patients entering with viral loads <100,000 remains unknown to Dr. Sax and ACTG colleagues.
Based on a time to virologic failure analysis, patients with high base line viral loads assigned to ABC/3TC had 57 events compared to compare to 26 events in those patients receiving TDF/FTC [ HR=2.33, 95%CI = 1.46-3.72, p <0.01]. Virologic failure was defined as having a confirmed viral load >1,000 copies at week 16-24, or viral load >200 at week 24 or rebounding viral load >200 after being <200 copies. The ABC group appeared to have higher rates of failure by all 3 of these definitions. Dr. Sax also analyzed the impact of suspected hypersensitivity reactions to ABC on the virologic failure rates and found no association with HSR to explain this outcome. There were however, higher rates of grade 3 or 4 adverse events in the ABC arms - predominately lipid abnormalities and generalized body aches. Interestingly, the investigators did not find a high rate of suspected HSR in the ABC arms - for unclear reasons.
In a planned secondary analysis of the viral load response by % <50 copies at 48 weeks, the investigators found no significant difference between the ABC and TDF arms when subjects who had early virologic failure were included and were allowed to change their ART regimens [75% (69-80%) in the ABC arms compared to 80% (74-85%) in the TDF arms, p = 0.20].
Comment: This trial is likely to change the way many clinicians view the two fixed dose combination (FDC) non-thymidine nucleoside backbones - ABC/3TC and TDF/FTC for treatment -naive HIV-infected individuals - especially for those with high baseline viral loads. To date most clinicians have largely viewed these two FDC as roughly equivalent choices - at least in terms of virological efficacy. Both were seen as potent, easy to take and safe choices for most treatment naive HIV-infected patients. The results of this rigorously conducted large randomized trial is the first adequately powered study of the 2 FDCs which now calls that generally held view into question. Treatment guidelines are responding to these results in different ways: the European guidelines are now calling TDF/FTC as the only preferred first choice while the recently published IAS-USA guidelines take a more cautious approach keeping both ABC and TDF as part of preferred choices but noting these early results in the guidelines. One cannot be too definitive at this juncture. We need to await the final (and planned) analysis of A5202 -expected late 2009 and further evaluation of other study results before making a strong recommendation. None-the-less, many clinicians may now feel that for those treatment naive patients with high viral loads that TDF/FTC is the preferred choice over ABC/3TC all other things being equal.
GSK's counter to A5202: Meta-analysis of nearly 3000 treatment naive patients
The presentation by GSK investigators was given right after Paul Sax's presentation of A5202 at the late breaker session. Dr. Keith Pappa and colleagues presented a meta-analysis of nearly 3000 treatment naive patients on ABC/3TC-containing regimens from several different GSK-sponsored clinical trials. The criteria for early virologic failure that was used in A5202 was not employed in any of these studies - so a direct comparison with A5202 is not possible. But when evaluating the results in the more widely accepted standard of virologic response (ie %<50 copies at 48 weeks) the investigators found no significant difference in response rates between subjects with high viral loads (ie >100,000 copies) compared to those with lower viral loads (<100,000). Responses were high in all studies and ranged from 87-95% achieving <50 copies by week 48. These trials included several different boosted-PI and EFV based regimens. Furthermore, Dr. Pappa emphasized that the HEAT trial which is a head to head comparison of ABC/3TC versus TDF/FTC in combination with LPV/r established that ABC/3TC was non-inferior to TDF/FTC by a variety of statistical analyses endpoints at week 48.
Comment: Why are these results different than those of A5202? Is ABC/3TC inferior or non-inferior to TDF/FTC? It is difficult to completely reconcile the differences seen across these different studies but there are several possible explanations as to why there appears to be a difference. ACTG A5202 used early time points (week16-24) and time to virologic failure to define efficacy where as the GSK-sponsored studies used the more widely accepted % patients with <50 copies at week 48 result to define efficacy. It is possible that ABC/3TC may simply take longer to achieve response and that had the patients in A5202 simply waited - they too may have achieved a viral load of <50 copies at week 48. In fact at least some did - as patients meeting early virologic failure endpoints were allowed to stay on or change ART. Some chose to stay on and were able to achieve <50 copies by week 48. Another difference is the size of the studies, A5202 is much larger than the HEAT trial and therefore better powered to detect small differences in efficacy. If the HEAT study had the same rate of early virologic failure as was seen in A5202, the expected number of events would have been 21 in the ABC/3TC arm and just under 10 events in the TDF/FTC arm - a difference that would have been more difficult to find statistical significance. Finally we have yet to analyze the impact of the 3rd drug in the A5202 and of course the 3rd drug used was different than that used in the HEAT study. Despite these differences and incomplete data the results of A5202 calls into question the relatively effectiveness of ABC/3TC dosed daily (compared to TDF/FTC) for patients with high viral loads at treatment initiation.
SMART Trial Analysis: Supports the D:A:D study results of higher Cardiovascular risks Associated with ABC use
Dr. Jens Lundgren of the University of Copenhagen presented an analysis of the SMART trial that seems to support the finding reported earlier this year from the D:A:D cohort, that ABC is associated with a higher rate of cardiovascular events compared to other nucleoside reverse transcriptase inhibitors. Focusing on those patients who were randomized to the "viral suppression" arm - that is those patients who stayed on continuous ART - were the focus of this analysis. Patients came into the SMART trial on a variety of ART regimens- Lundgren and colleagues categorized patients into 3 groups based on their nRTI's backbone: ABC (no DDI) N=1019, DDI (with or without ABC) N=643, other nRTI (no DDI or ABC) N=2882. Baseline risk factors for CV disease were similar across the 3 arms, however, the ABC group was more likely to be on a triple nucleoside regimen (namely ABC/AZT/3TC) compared to the other groups.
Cardiovascular events were defined in a number of different ways but included the SMART pre-study specified events that were adjudicated by an independent panel. In comparison to the other groups the group on ABC-containing ART had a statistically higher rate of MI's, as well as other cardiovascular events. The hazard ratios varied from 2- 4 for the various endpoints, suggesting that use of ABC in this study was associated with a 2 - 4 fold increase in risk of a cardiovascular event. This finding is consistent with the findings in the D:A:D results published in Lancet this year. Unlike the D:A:D result, the DDI group in this analysis did not appear to be at increased risk compared to the other nucleoside group. Again like in D:A:D the increased risk associated with ABC appears to be only in those patients with other CV risk factors - in this analysis, those patients with 5 or more other risk factors.
The unexpected results of D:A:D caused many in the field to begin searching for a biological explanation. In a substudy of 791 patients from the SMART trial, Lundgren showed results of "biomarkers" of inflammation. Patients in the ABC group, appeared to have higher baseline plasma levels of hs-CRP, IL-6 and amyloid A when compared to the DDI group. These results suggest that ABC may be associated with an underlying inflammatory state that contributes to CV morbidity in patients with vascular disease and may provide a clue as to an underlying biological mechanism that would help support the verisimilitude (something that only appears to be true or real, for example, a statement that is not supported by evidence) of the findings from D:A:D and SMART.
Comment: Reported elsewhere by NATAP [see ref] investigators from GSK had in an earlier session showed data from a meta-analysis looking at CV risk for patients enrolled into ABC-containing ART clinical trials. Although the data was from randomized controlled trials and would therefore be a potentially better design to find a CV risk that was independent of the biases inherent in the D:A:D and SMART trials, the GSK studies showed no increased risk of MI's with ABC. Again how do we reconcile these differences? Is ABC associated with an increase risk of cardiovascular complications or not? One has to look at why there may be different outcomes from these studies. First, none of the GSK trials that were subject to the meta-analysis from GSK were intended to find CV events, so were not powered to find these kind of outcomes. Secondly, the patients in the GSK trials were on average about a decade younger than the patients followed in the D:A:D and SMART studies. Patients entering clinical trials of ART are also likely to be healthier and those with high risk factors for CV disease would likely have been screened out. It is this group of patients that the risk for CV events is seen with ABC. So the study populations differ despite the very large number of patients in the GSK analysis there is simply not enough patients in those studies that are like the patients in the D:A:D cohort or SMART. On the other hand, both D:A:D and SMART are subject to a kind of selection bias, in that patients with high risk factors for CV disease tend to be on ABC, probably because ABC was used preferentially in more treatment experienced subjects and therefore subjects who were older, HIV-infected longer etc. In both analysis the investigators control for known risk factors but one can never be absolutely certain that those statistical maneuvers completely correct the selection bias.
As Jens Lundgren discussed with me after his presentation "the science is evolving and we need to let it continue to evolve before we can be more clear about what is going on". In the meanwhile what are patients and clinicians to do? It is probably prudent that in patients with high risk factors for CV disease who happen to be on ABC-containing ART regimens that at least a discussion about switching their ABC for some other antiretroviral take place. Of course this discussion should also provide the clinician an opportunity to discuss reduction of other CV risk factors as well - like stopping tobacco use, weight loss, diet modification, exercise etc. However, if there is a reasonable and safe alternative to ABC then changing the ART regimen in those patients with high underlying risk factors seems appropriate at this time.
A randomized, double-blind, placebo-controlled, multicenter trial of pregabalin vs placebo in the treatment of neuropathic pain associated with HIV neuropathy Presented by David Simpson, United States (IAC Mexico City, Aug 2-8 2008; THAB0301) D.M. Simpson et al. Mount Sinai Medical Center, New York, United States, 2Pfizer Global Pharmaceuticals, New York, United States
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 Presented by Paul Sax, United States (IAC Mexico City, Aug 3-8 2008; THAB0303) P. Sax et al. Division of Infectious Diseases, Brigham and Women's Hospital, Harvard Medical School, Boston, United States
Abacavir/lamivudine (ABC/3TC) shows robust virologic responses in ART-naive patients for baseline (BL) viral loads (VL) of 100,000c/mL and <100,000c/mL by endpoint used in ACTG5202 Presented by Keith Pappa, United States (IAC Mexico City, Aug -8 2008; THAB0304) K. Pappa United States et al. GlaxoSmithKline, Infectious Diseases MDC, U.S. HIV Collaborative Studies, Research Triangle Park, NC, United States
Use of nucleoside reverse transcriptase inhibitors and risk of myocardial infarction in HIV-infected patients enrolled in the SMART study Presented by Jens Lundgren, Denmark (IAC Mexico City, Aug 3-8 2008; THAB0305) J. Lundgren et al. Rigshospitalet & University of Copenhagen, Copenhagen,
Denmark Is abacavir (ABC)-containing combination antiretroviral therapy (CART) associated with myocardial infarction (MI)? No association identified in pooled summary of 54 clinical trials Presented by Jaime Hernandez, United States (IAC Mexico City, Aug 3-8 2008; WEAB0106) GlaxoSmithKline Research & Development, HIV Medicine Development Center, Research Triangle Park, United States
D:A:D. Lancet April 2008