Clinical Utility of Resistance Testing

• Phenotypic Susceptibility and Virologic Responses in Nucleoside Reverse Transcriptase Inhibitor (NRTI)-Experienced Subjects Receiving NRTIs + Efavirenz (EFV), Nelfinavir (NFV), or Both in ACTG 364 (Poster 435)
  Authored by D. Katzenstein, N. Hellmann, S. Liou, R. Bosch, M. Albrecht, and the ACTG 364 Study Team
  View the original abstract

Using drug susceptibility testing is suggested for patients with virologic failure or suboptimal suppression after the initiation of therapy. The optimal interpretation of phenotypic susceptibility data remains under investigation. One complicating factor in using phenotypes for treatment-experienced patients relates to the fact that patients in the U.S. now all receive multiple drug therapy; how can one weigh the contribution of susceptibility to all the components of a cocktail? Is resistance all or nothing? Are there degrees of resistance, or partial resistance?

In this presentation, Dr. David Katzenstein (Stanford University) presented the results of his analysis of the predictive value of phenotypic analysis in ACTG protocol 364. This study was a placebo-controlled trial of nelfinavir (NFV) vs. efavirenz (EFV) vs. NFV/EFV, in combination with dual nucleosides (NRTI), in treatment-experienced patients. Previously, results of this study showed that viral load responses were greater among patients who received quadruple therapy with NFV/EFV/dual NRTI, with 67% of these patients achieving viral loads of less than 50 copies at 40-48 weeks of therapy. Phenotypic analyses were performed using Virologic's PhenoSense assay. Dr. Katzenstein calculated a Phenotypic Susceptibility Score (PSS). The PSS was calculated in one of two ways. The first, used a dichotomous cut-off point (at 2.5-fold change from wild type) to call a virus either susceptible or resistant. The second, used a modified continuous assignment of resistance (PSS-continuous), calling viruses with greater than 10-fold changes "resistant," while viruses with 2.5- to 10-fold changes as "partially resistant." Logistic regression analysis was done to see what factors -- either PSS score, CD4 count or baseline viral load predicted the viral response to therapy. The only factors significantly associated with a week 40-48 viral load less than 50 were the number of new drugs used and the PSS-continuous. There was no association with CD4 count or the PSS-dichotomous. Significantly, the PSS-continuous was not associated with viral responses at week 16.

These data are of limited immediate relevance for the treatment community. The use of drug susceptibility testing is being used with increasing frequency to assist with treatment management. It is clear that a compound analysis of the susceptibility of HIV to all drugs in a salvage regimen is relevant to patients, but how to best define this parameter remains unclear. Some simple features emerge from several analyses: it is better to have more active ("susceptible") drugs than fewer, the breakpoint for resistance and susceptible may be different for different drugs. As this study illustrates, resistance is not all or nothing; analyses that incorporate the concept of partial resistance may offer a more complete picture of salvage treatment options.