 |
|
|
| |
Early Response to Maraviroc Equivalent With
Coreceptor Genotyping and Trofile Phenotyping
|
| |
| |
Mark Mascolini
XVIII International Drug Resistance Workshop, June 9-13, 2009, Fort Myers, Florida
Although genotyping to infer coreceptor use by HIV remains an insensitive predictor of non-R5 virus compared with the original phenotypic Trofile assay, using either approach to assign maraviroc therapy in two maraviroc trials would have yielded highly similar virologic response rates [1]. The findings may have clinical implications because genotyping has certain advantages over phenotyping, including speed, cost, and ability to predict coreceptor use in a wider array of viral samples. (from Jules: several additional posters at this meeting also found promising results using genotyping and the sensitive Roche 454 assay to identify X4 and R5 and NATAP will be reporting on these.)
Richard Harrigan (BC Centre for Excellence in HIV/AIDS, Vancouver) and coworkers from other centers sequenced the V3 loop region of HIV's envelope gene in stored plasma samples from patients enrolled in MOTIVATE 1 and the A4001029 trial of the CCR5 antagonist maraviroc, which has no activity against CXCR4-using HIV. The investigators used two algorithms to predict coreceptor use from V3 sequence data: PSSM with a -2.96 cutoff and geno2pheno with a 5% false-positive rate setting. They used a weighted sensitivity score to estimate the activity of each patient's background regimen.
Among 1230 people screened by Trofile for MOTIVATE 1, that assay determined that 553 had virus that could use CXCR4. People with Trofile-determined CXCR4-using virus were eligible for the 1029 study.
Both PSSM and geno2pheno had high specificity in detecting virus that Trofile called X4 or R5/X4 (dual/mixed)--91% and 90%. But sensitivity in matching the Trofile call was "pretty miserable," Harrigan reported--only 56% with PSSM and 63% with geno2pheno. The genotypic systems often had a different readout of virus that Trofile called dual/mixed, meaning the person's viral population could use CCR5 or CXCR4. PSSM matched Trofile in dual/mixed calls in only 238 of 285 samples (83.5%), while geno2pheno matched Trofile in only 202 of 321 samples (63%).
But when Harrigan looked at 12-week virologic response curves in three groups--people who took maraviroc twice daily, people who took maraviroc once daily, and people who took placebo--patients judged to have CXCR4-using virus by genotyping and by Trofile had virtually overlapping failure rates. In the same way, people whose HIV got rated R5 by genotyping or Trofile had nearly overlapping response curves.
Next Harrigan figured the sensitivity and specificity of Trofile, geno2pheno, and PSSM in predicting week-8 response defined as a viral load below 50 copies or a viral load reduction of 100-fold or more in people taking maraviroc twice daily. Sensitivity for predicting response was 90% with Trofile, 85% with geno2pheno, and 87% with PSSM. Specificity was uniformly low with all three tests because a maraviroc regimen may fail for several reasons besides CXCR4-using virus.
Sensitivity of the three coreceptor-calling methods remained indistinguishable when the investigators defined virologic success as (1) a load below 400 copies or at least a 100-fold decline at week 8, (2) a load below 400 copies or at least a 10-fold drop at week 8, (3) a load below 400 copies or at least a 1000-fold plunge at week 8, or (4) a load below 50 copies at week 24.
Harrigan concluded that "all methods discriminated between short-term responders and non-responders to maraviroc with broadly similar performance in this treatment-experienced population."
Reference
1. Harrigan PR, McGovern R, Dong W, et al. Screening for HIV tropism using population-based V3 genotypic analysis: a retrospective virological outcome analysis using stored plasma screening samples from MOTIVATE-1. XVIII International Drug Resistance Workshop. June 9-13, 2009. Fort Myers, Florida. Abstract 15.
|
| |
|
|
|
|
|
