Loading...
Each year, the HIV Drug Resistance Workshop manages to highlight the scientific complexity of the field while also providing clinically compelling information. This year's workshop, held in Florida in June, was no exception. Here are the most clinically relevant findings discussed at the meeting. (The entire abstract book is available at http://www.informedhorizons.com/resistance2009.)
Does persistent viremia result from long-lived, chronically infected cells or ongoing viral replication? Jason Dinoso and colleagues attempted to answer this question by assessing whether antiretroviral intensification can reduce persistent viremia [Abstract 10]. Thirteen patients with virologic control on standard antiretroviral regimens underwent intensified therapy for 4 weeks with efavirenz, lopinavir/ritonavir, or raltegravir; an additional 5 patients underwent 8-week intensification with atazanavir. Use of a highly sensitive single-copy virologic assay revealed no change in persistent viremia after intensification. The researchers concluded that ongoing viral replication might not be the source of persistent viremia.
Sarah Palmer and colleagues also evaluated the dynamics of persistent viremia, by comparing HIV RNA levels in 84 “elite controllers” (patients who have achieved long-term virologic suppression without using antiretrovirals) and 163 patients on suppressive antiretroviral therapy [Abstract 4]. The two groups had similar mean levels of viremia (based on a real-time RT-PCR assay with single-copy sensitivity), but, in a longitudinal analysis, the elite-controller group had greater intra- and interpatient variability in HIV RNA levels. These data suggest that ongoing viral replication is likely the mechanism underlying persistent viremia in elite controllers, but that a different mechanism exists for patients on suppressive therapy.
Finally, Constance Delaugerre and colleagues measured 2-LTR circles — a marker of ongoing viral replication — in 60 patients who were randomized to either switch from enfuvirtide to raltegravir or continue enfuvirtide [Abstract 9]. Neither group experienced a detectable increase in the number of 2-LTR circles or a change in HIV DNA reservoirs. Researchers concluded that raltegravir does not alter HIV DNA levels in cell reservoirs in highly treatment-experienced patients in 24 weeks. The lack of effect on 2-LTR circles suggests that persistent low-level viremia is not due to ongoing cycles of HIV replication.
Patients taking NNRTIs often have minority viral populations that are potentially predictive of treatment failure, but at what threshold do these populations foretell clinical trouble? Derrick Goodman and colleagues from Gilead Sciences used a sensitive assay to detect low levels of K103N (assay cutoff, 0.5% of the overall HIV population) in samples from treatment-naive patients randomized to receive FTC + tenofovir + efavirenz or AZT + 3TC + efavirenz [Abstract 41]. By week 144, K103N was detected in 16 of 476 patients. The circulating percentage of K103N was higher in patients who experienced virologic failure (range, 0.8%–15%) than in those who did not (range, 0.6%–3.2%). A threshold of 2000 copies/mL of mutant virus was strongly associated with virologic failure: 5 of 6 patients who experienced failure had mutant virus above this threshold, compared with only 1 of 10 patients who did not experience failure.
The Trofile assay (Monogram Biosciences, Inc.) is currently the gold standard for determining viral tropism and potential eligibility for receipt of CCR5 antagonists. However, investigators are now exploring whether genotype testing is an effective alternative [Abstract 15]. Richard Harrigan and colleagues retrospectively sequenced the V3 loop in stored plasma samples from 1200 patients who were screened with the first-generation Trofile assay for entry into the MOTIVATE 1 study of maraviroc (for patients found to have CCR5-tropic virus) or the Maraviroc 1029 study (for a subset of patients found to have CXCR4-using virus). Compared with Trofile, genotyping had only modest sensitivity for predicting tropism (around 60%, depending on the precise genotyping method) but very good specificity (around 90%). Median viral-load decreases at 8 weeks were similar between patients who were determined to have CCR5-tropic virus by genotyping and those who tested CCR5-tropic by Trofile. Although the researchers concluded that genotyping might effectively predict clinical outcomes, the sample size of patients with discordant results between the two assays might have been too small to make such judgments: Most patients identified as having dual-mixed virus at screening did not have the chance to receive maraviroc because Maraviroc 1029 was only a small phase II trial.
Previous research has shown that the resistance mutations associated with raltegravir emerge at codon 155 first, often accompanied by a second mutation with a higher level of resistance (92Q). These mutations are then ultimately replaced by the double mutant 148H/140S. Two studies now detail the fitness loss associated with these varying degrees of raltegravir resistance. In the first study, Zixin Hu and I found that in the presence of raltegravir, both double mutants (155H/92Q and 148H/140S) are fitter than the single mutants (155H and 148H, respectively), but that 148H/140S is fitter than 155H/92Q [Abstract 77]. In the second study, Elisabeth Dam and colleagues compared the relative infectivity of different raltegravir-resistant mutants in the presence and absence of drug [Abstract 26] and came to the same conclusions that we did. Whether the fitness loss associated with raltegravir resistance has any clinical benefit remains to be seen.
Mutations occur in the connection and RNase H domains of HIV reverse transcriptase (RT), but they are not assessed by current genotypic or phenotypic assays, and their clinical effect on NRTI and NNRTI resistance is unknown. To address this latter issue, Jessica Brehm and colleagues evaluated paired stored samples (pretherapy and at time of virologic failure) from 53 patients in ACTG 5142 [Abstract 32]. Overall, the mutations did not predict treatment failure and were not associated with outcome. Several mutations showed associations with treatment failure that were of borderline significance, but no multivariate analysis was conducted. In this case, what we can't see doesn't seem to be hurting us.