Library Volume 2, Issue 1: Resistance Reporter© 2008 Conference on Retroviruses and Opportunistic Infections (CROI)


Selections from the 2008 Conference on Retroviruses and Opportunistic Infections (CROI)

  • Section 1: Select presentations concerning the virologic success/failure of second-line therapy and further analyses of adherence from the CPCRA FIRST Study with a focus on class-specific resistance
  • Section 2: Select presentations on the relevance of detecting minority variants and the comparative utility of 2 tropism tests

Section 1 (Section 2 Also Available)

Select presentations concerning the virologic success/failure of second-line therapy and further analyses of adherence from the CPCRA FIRST Study with a focus on class-specific resistance


Long-Term Consequences of the Delay Between Virologic Failure of HAART and Regimen Modification: A Prospective Cohort Study

CROI Abstract #798, Petersen M et al.

Although current treatment guidelines advocate modifying HAART as soon as virologic failure is detected, the opportunity to change therapy can often be delayed. To determine the consequences of such lags in therapy optimization, this study pooled data from 2 prospective cohorts (Johns Hopkins, University of North Carolina) to determine the impact of delayed regimen modification following failure of first-line HAART on mortality and immunologic failure and compared the effect of delayed modification between PI-based and non-PI-based regimens. Virologic failure was defined as 2 consecutive HIV RNA measurements above a time-dependent threshold: greater than 1000 copies/mL during Weeks 12-24 and/or greater than 500 copies/mL thereafter. Of the 982 patients included in the study, 76% failed a PI-based, first-line HAART regimen; 32% of these individuals also received ritonavir boosting. At the time of first HAART failure, the individuals in each of the cohorts, segregated by first-line PI use versus non-PI use, had a median age of 39-41 years, 62%-73% were male, the median most recent viral load measurement ranged from 3.9-4.5 log10 copies/mL, and the median most recent CD4+ cell count ranged from 182-332 cells/mm3. Delayed treatment modification was associated with an increased risk of death (HR: 1.23; 95% CI: 1.08-1.40; P = .002 for every 3 months delay) and immunologic failure or death (HR: 1.21; 95% CI: 1.07-1.36; P = .002 for every 3 months delay) among individuals who initially received a non-PI-based regimen. In contrast, delayed regimen modification showed a slight protective effect with regard to death (HR: 0.93; 95% CI: 0.87-0.99; P = .03) and no association with immunologic failure or death (HR: 0.98; 95% CI: 0.94-1.03; P = .45) among patients who initially failed a PI-based regimen. Trends were similar when separate analyses were conducted for the 2 independent cohorts, following 2nd HAART failure, and after excluding subjects on nonboosted PI regimens or with no history of antiretroviral use before HAART. Weighted data-adaptive regression analyses showed that the longer the delay after failure of a non-PI-based regimen, the sharper the increase in the mortality hazard, which then remained consistently elevated following regimen modification. In contrast, the mortality hazard increased gradually among patients failing a PI-based regimen, with little apparent short-term differences between individuals who had their therapy modified early versus late. The authors concluded that delaying treatment modification after failure of NNRTI-based HAART is associated with an increased risk of disease progression, whereas PI-based regimens are less dependent on early versus delayed switching strategies.


Back to Top


Analysis of the Relationship Between Antiretroviral Medication Adherence and Class-Specific Resistance in a Large Prospective Randomized Clinical Trial

CROI Abstract #777, Gardner E et al.

The CPCRA 058 FIRST study was designed to compare long-term clinical outcomes among 1397 treatment-naive individuals randomized to begin therapy with a regimen containing an NNRTI, a PI, or both. The current study was a subanalysis of 903 patients in the NNRTI arm (n = 446) and the PI arm (n = 457) that focused on the relationship between adherence and class-specific genotypic resistance at the time of first virologic failure (HIV-1 RNA > 1000 copies/mL at Month 4 or thereafter). Adherence was assessed using the 7-day CPCRA Adherence Self-Report Form at Months 1, 4, and every 4 months thereafter. At baseline, patients had a median age of 38 years, 78% were male, the median viral load was 5.1 log10 copies/mL, and the median CD4+ cell count was 166 cells/mm3. Among patients in the PI arm, 74% initiated treatment with an unboosted PI regimen, whereas only 26% received a boosted PI regimen, which is the current standard. Virologic failure was significantly more common among individuals on PI-based therapy versus NNRTI-based therapy (71% vs 59%; P < .01). The median time to virologic failure was also significantly shorter for individuals on PI-based therapy versus NNRTI-based therapy (1.2 vs 3.0 years; P < .01). Despite these differences, the median cumulative adherence score at the time of virologic failure or censoring was comparable between the 2 respective groups (92% vs 93%; P = .98). In addition, similar proportions of PI- and NNRTI-treated individuals had evidence of no resistance (70% vs 72%), resistance to 1 drug class (21% vs 17%), or resistance to 2-3 drug classes (9% vs 11%; P = .19 overall) at the time of virologic failure. The NNRTI arm showed an inverse relationship between adherence and the risk of virologic failure with the presence of NNRTI resistance (HR per 10% decrease in cumulative mean adherence score: 1.24; 95% CI: 1.14-1.34), whereas no association was found between adherence and the risk of virologic failure with the presence of PI resistance in the PI arm (HR per 10% decrease in cumulative mean adherence score: 1.09; 95% CI: 0.88-1.36). However, poorer adherence was associated with an increased risk of NRTI resistance at virologic failure in both the NNRTI (HR per 10% decrease in cumulative mean adherence score: 1.19; 95% CI: 1.06-1.34) and PI arms (HR per 10% decrease in cumulative mean adherence score: 1.28; 95% CI: 1.18-1.38). Patients who initiated treatment with a boosted PI versus an unboosted PI were less likely to develop PI resistance (2% vs 10%; P < .01) or NRTI resistance (19% vs 29%; P = .03). On-treatment findings were similar to those of the intent-to-treat analysis.


Back to Top


RELATED ARTICLES:


Back to Top



Section 2 (Section 1 Also Available)

Select presentations on the relevance of detecting minority variants and the comparative utility of 2 tropism tests


Virologic Success of Different Strategies for Initial ART Regimens Is Predicted by the Type and Detection Level of Minor Drug-Resistant Variant Detected by Ultra Deep Sequencing: The CPCRA 058 FIRST Study

CROI Abstract #878, Huppler Hullsiek K et al.

Commercially available bulk sequencing assays are unable to detect minority resistant variants with a prevalence below 20% of the total viral population. Ultra-deep sequencing uses parallel pyrosequencing technology to generate thousands of single-genome sequences from each sample assessed, thereby allowing for rapid and sensitive identification of minority resistant variants. The current study assessed a random subset of 504 patients who participated in the CPCRA 058 FIRST study to: (1)determine the baseline prevalence of low-abundance drug-resistant variants using ultra-deep sequencing, (2) compare the prevalence of baseline mutations identified by standard versus ultra-deep sequencing, and (3) determine the influence of baseline minority variants identified using ultra-deep sequencing on the response to NNRTI-, PI- or NNRTI+PI-based therapy. Of 258 samples with both standard and ultra-deep sequencing results, ultra-deepsequencing identified twice as many patients harboring minority resistant variants with at least 1 mutation in any class as standard sequencing (28.3% vs 13.6%; P < .001). This trend held true when broken down by the presence of at least 1 resistance mutation within the NRTI (14.0% vs 6.2%; P < .001), NNRTI (15.1% vs 6.6%; P < .001), and PI (4.7% vs 2.3%; P = .03) classes. The majority of the variants identified by ultra-deep sequencing comprised only 1%-20% of the total viral population. Among patients who initiated treatment with NNRTI-based therapy, individuals with a minority variant harboring at least 1 NNRTI resistance mutation at baseline had a 3.1-fold higher risk of virologic failure versus those without (P = .0001), while individuals with a minority variant with an NNRTI or NRTI resistance mutation at baseline had a 1.8-fold higher risk of virologic failure versus those without (P = .04). No significant correlation between the presence of minority resistant variants with any mutation in any class at baseline and virologic failure was identified for patients who initiated PI-based therapy. Among patients who initiated treatment with NNRTI+PI-based therapy, individuals with a minority variant with a PI mutation at baseline had a 7.2-fold higher risk of virologic failure versus those without (P = .003), although only 3 individuals comprised this group.  Overall, the rate of virologic failure was highest for individuals harboring minority resistant variants (with any mutation) present at a prevalence of 1%-20% as opposed to individuals with no detectable mutations or with variants present at a prevalence greater than 20% (38.9% vs 26.5% and 18.1%, respectively; P = .09). This trend was observed within each individual treatment arm as well. According to the investigators, use of ultra-deep sequencing to determine baseline mutation patterns may be useful for guiding regimen selection to reduce the risk of treatment failure.


Back to Top


Presence of Minor Populations of Y181C Mutants Detected by Allele-Specific PCR and Risk of Efavirenz Failure in Treatment-Naïve Patients: Results of an ACTG 5095 Case-Cohort Study

CROI Abstract #83, Paredes R et al.

Detectable NNRTI resistance at baseline can dramatically increase the risk of virologic failure to efavirenz-based HAART, but the clinical significance of minority NNRTI-resistant variants is unknown. The current case-cohort substudy of ACTG 5095 assessed the relationship between pre-existing minor NNRTI-resistant variants and the efficacy of efavirenz-based HAART in treatment-naïve patients who were previously determined to be resistance free by bulk sequencing before treatment initiation. Previous findings from ACTG 5095 established that the 2 efavirenz-based regimens (zidovudine/lamivudine + efavirenz, zidovudine/lamivudine/abacavir + efavirenz) were superior to the non-efavirenz-based regimen (zidovudine/lamivudine/abacavir). A total of 322 efavirenz-treated patients comprised the case-cohort sample. A total of 178 of these individuals had experienced virologic failure (HIV-1 RNA ≥ 200 copies/mL at Week 16 or thereafter); 144 had not. Among a random cohort of 220 (195 evaluable) efavirenz-treated patients from the total case-cohort sample, 12 individuals (5%) were determined to harbor NNRTI resistance at baseline by bulk sequencing. The remaining 183 patients were determined to be resistance free and underwent more intensive allele-specific PCR to identify the presence of minor variants with K103N (detection threshold: 0.001%-0.003%) and/or Y181C (detection threshold: 0.03%). Allele-specific PCR revealed minority variants with K103N, Y181C, and K103N + Y181C in 4.4%, 29.5%, and 6%, respectively. These minority variants represented less than 1% of all viral quasispecies. The presence of minority Y181C variants was significantly more common in patients with virologic failure versus without (58% vs 29%; P = .001); this held true for subjects with a baseline viral load below 100,000 copies/mL as well as for those with a baseline viral load above 100,000 copies/mL. The presence of minority Y181C variants was associated with a 3.57 (95% CI: 1.98-6.42) increased risk of virologic failure to efavirenz-based therapy in patients with recent treatment adherence; the hazard ratio increased slightly to 6.09 (95% CI: 2.83-13.1) for those with poor recent adherence. Despite this, some individuals with minority Y181C variants still achieved long-term virologic control with efavirenz-based therapy. The presence of pre-existing minority K103N variants was not significantly associated with an increased risk of virologic failure (HR: 1.58; 95% CI: 0.76-3.28), but the investigators speculated that this may have been due to the fact that a very small number of individuals harbored minority K103N variants.


Back to Top


Comparison of Results From the SensiTrop™ vs Trofile™ Assays on 100 Samples From the Maraviroc Expanded Access Program

CROI Abstract #920a, Tressler R et al.

Trofile™ is a phenotypic tropism assay that has been used to determine candidacy for CCR5 antagonists in several major clinical studies performed to date. In the MOTIVATE 1 and 2 clinical trials, 8%-10% of treatment-experienced patients had discordant tropism results when the Trofile™ assay was performed at screening and then again at baseline 4-6 weeks later. The new, commercially available SensiTrop™ assay uses molecular heteroduplex tracking to determine viral tropism (assignment as R5 or dual/mixed only) and reportedly has a high degree of sensitivity. The current study compared the performance of the Trofile™ and SensiTrop™ assays using 100 stored samples from the multinational maraviroc expanded access program. The SensiTrop™ assay characterized a higher proportion of the screening samples as having R5 virus and a lower proportion as having dual/mixed virus in comparison with the Trofile™ assay. At screening, Trofile™ identified 52 individuals with R5 virus, 39 with dual/mixed or X4 virus, and 9 with nonreportable samples. In comparison, SensiTrop™ identified 69 individuals with R5 virus at screening, 20 with dual/mixed virus, and 11 with nonreportable samples. Among the 39 screening samples identified as dual/mixed or X4 according to the Trofile™ assay, the SensiTrop™ assay classified only 14 samples as such; 19 were scored as R5, and 6 were scored as unreportable. Similarly, among the 8 screening samples identified as X4 according to the Trofile™ assay, the SensiTrop™ assay classified only 4 of the samples as dual/mixed; 2 were scored as R5, and 2 were scored as unreportable. The ability of the SensiTrop™ assay to identify dual/mixed virus did not appear to be influenced by viral load at screening. Based on those samples determined to be dual/mixed or X4 according to the Trofile™ assay, the SensiTrop™ assay demonstrated high specificity (92.5%; 95% CI: 84.3%-100.7%) but low sensitivity (42.5%; 95% CI: 25.6%-59.3%) for also classifying those samples as dual/mixed, yielding positive and negative predictive values of 82.2% and 66.4%, respectively. The authors concluded that relative to the results obtained with the Trofile™ assay, SensiTrop™ lacks adequate sensitivity, failing to detect dual/mixed HIV in more than 50% of samples.


Back to Top



RELATED ARTICLES:



To order new educational programs that give unbiased, independent information and hope to patients running low on treatment options, view or download (FREE) Visionary Health Concepts' educational booklet, "Are You Experienced? HIV and You©"at www.freehivinfo.com.