icon_folder.gif   Conference Reports for NATAP  
 
  13th CROI
Conference on Retroviruses and Opportunistic Infections
Denver, Colorado
Feb 5-8, 2006
Back grey_arrow_rt.gif
 
 
 
Sensitive Genotypic Testing Uncovers Drug Resistance Associated with Viral Failure
 
 
  Reported by Jules Levin
June 13-17, 2006
Sitges, Spain
 
"Baseline Detection of Low-Frequency Drug-resistant HIV in Drug-Naive Individuals is Strongly Associated with Poor Treatment Responses"
 
Jeffrey Johnson, Laboratory Branch, Division of HIV/AIDS Prevention, Centers for Disease Control and Prevention
 
STUDY QUESTION
Do viruses with drug resistance-associated mutations at levels below the detection of (standard genotypic resistance test) sequence analysis impact virologic responses to HAART?
 
AUTHOR SUMMARY OF STUDY RESULTS

Sensitive genotype testing found that 7.2% of the participants had treatment-relevant mutations at baseline as compared to 2.9% detected by sequencing by standard genotype test.
 
Conventional sequencing (standard genotype test) identified only 4 of the 12 (33%) relevant resistance mutations detected by real-time PCR (sensitive resistance test).
 
All persons with detectable K103N, Y181C, and/or M184V mutations experienced virologic failure.
 
AUTHOR CONCLUSIONS

Sensitive PCR-based testing found that the number of persons with relevant baseline drug resistance mutations was 2.5-times the number identified by conventional sequencing.
 
The clinical significance of drug-resistant viruses at frequencies <20% was demonstrated by their strong association with virologic failure independent of viral load and CD4 cell counts.
 
These data imply that sensitive baseline testing can improve ART management of HIV-1-infected persons.
 
NOTE from Jules Levin: This study is not the first to find such results. Several studies over the course of the past two years have found the sensitive resistance tests uncover mutations not found with standard testing and this has been found with NNRTI mutations, tenofovir-related mutations (K65R, 74), and PI-associated mutations. This subject received much attention & discussion at the workshop. The question is should sensitive testing become more standard? The sensitive tests are not commercially available. And the cost is high and quality control is difficult for such testing. In addition, a question of clinical significance lingered. For example, is there a cut-off the amount of virus found by sensitive testing that would ampunt to the development of viral failure; perhaps not all patients identified with key mutations would develop viral failure. How much impact & how significant and relevant is this testing overall? So at this meeting it was generally agreed by the field of experts that this is a matter of concern but no general consensus was reached about what should be done about this. I think researchers & other attendees at this meeting will continue to discuss this matter.
 
BACKGROUND
Developed (sensitive resistance tests) real-time PCR-based assays for HIV drug resistance mutations with detection sensitivities up to 1-2.5 logs lower than conventional sequencing.
 
Earlier screening found a cumulative 24% increase in the number of transmitted resistance mutations .
 
Additional mutations implied transmitted multi-drug resistance is underestimated by at least 20%.
 
Over 2/3 (~70%) of newly detected low-frequency mutations were associated with resistance to another drug.
 
THE STUDY
This_ was _a_ retrospective analysis of 138drug-naive persons who participated in a 2001-2002 GSK ABC/3TC/EFZ clinical trial (CNA30021)
 
Using sensitive assays, they retested blinded baseline RNA samples for three treatment-relevant mutations (K103N, Y181C, M184V)
 
About 1/2 study population who were identified to have resistant HIV with the sensitive test but did not have resistance with the standard genotype test were failures = VL>50 copies/mL within 48 weeks
 
Participant clinical information was unmasked after results were reported
 

result-1.gif

Frequencies of mutations detected by sensitive vs standard genotype testing:
The standard genotype test missed 50% of the mutations. Standard genotype test detected 2/138 (1.4%) with K103N, the NNRTI mutation; 1/138 (0.7%) Y181C, also a NNRTI mutation; 1/138 (0.7%) mutation with the M184V, the 3TC/FTC mutation). BUT, the sensitive genotype test detected 8 resistance mutations: 4/136(2.9%) with the K103N, 1/138 with the Y181C (0.7%), and 3/137 (2.2%) patients had the M184V.
 

PCR-2.gif

Treatment-relevant mutations were detected in 10/138 (7.2%) persons at baseline with sensitive resistance test (real-time PCR test).
 
Conventional (genotype test) sequencing detected only
- 2 of 6(33%) K103Nmutants
- 1 of 4(25%) M184V
- 1 of 2(50%)Y181C
 
Viral Failures: 4 patients identified with the K103N with the sensitive test but not picked by with the standard test were viral failures (>50 copies/ml). All 3 identified with the M184V mutation but not picked up by standard genotype test had viral failure). In the table below the red results were the sensitive test.
 

Viro-3.gif

Sensitive testing identified 9 patients with K70R mutation at baseline whilr standard genotype test identified only 1 patient with K70R.
 

sample-4.gif

All persons with detectable K103N, Y181C, M184V mutations experienced virologic failure.
 
Of those with treatment-relevant mutations at baseline:
2/10 never suppressed
4/10 failed within 2 months
2/10 failed by month 4 ¥ ¥ the remaining two failed by month 6
 
The person with dual-class mutations failed by 2 months.
 
10/69 (14.5%) failures had detectable relevant resistance mutations.
 
Three of four persons with available genotypes at failure had the same mutations as low-frequency variants at baseline.
 
Fisher's exact test revealed the association between baseline low-frequencyresistance and virologic failure was significant (p=0.012)
 

Loads-5.gif

Baseline CD4 Counts
Wilcoxon rank sum found no difference in CD4 counts between the failures with mutations and the treatment successes (p=0.68).
 
Spearman rank identified in failures with baseline resistance a negative correlation between viral loads and CD4 counts that approached significance (p=0.088).