icon-folder.gif   Conference Reports for NATAP  
  7th European HIV Drug Resistance Workshop
March 25-27, 2009
Back grey_arrow_rt.gif
Weighted Genotyping Scores Work
Equally Well in Predicting Response to Etravirine

  7th European HIV Drug Resistance Workshop, March 25-27, 2009, Stockholm
Mark Mascolini
Two weighted genotypic scores for predicting response to the nonnucleoside etravirine yielded highly similar results in a 4000-sample analysis and in a review of more than 300 DUET trial participants [1]. Both scores--one devised by Tibotec (etravirine's maker) and the other by Monogram Biosciences--closely reflected phenotypically measured viral susceptibility to etravirine in these viral isolates.
Johan Vingerhoets, who crafted the Tibotec weighted genotypic score [2], compared it with a score independently developed by Monogram scientists [3]. Both systems weight mutations to reflect greater or lesser impact on phenotypically measured susceptibility to etravirine. The comparison tested the two tools for their ability to predict response in 4248 nonnucleoside-resistant viral isolates and in 323 isolates from DUET trial participants. The DUET isolates came from people who did not use enfuvirtide for the first time in their salvage regimen and did not drop out of the study for reasons other than virologic failure. Vingerhoets used a single cutoff specified by each system's designers to distinguish sensitive from resistant virus--2 for Tibotec and 3 for Monogram.
Monogram's response predictor considers an array of 30 resistance mutations, while the Tibotec calculator weights 17 mutations. The systems diverge considerably in ranking the importance of certain mutations, such as L100I (Tibotec weight 2.5, Monogram weight 4), K101P (Tibotec 2.5, Monogram 4), V179E (Tibotec 0, Monogram 3), and G190Q (Tibotec 0, Monogram 3). Tibotec weights 5 mutations that Monogram ignores, while Monogram factors in 18 mutations absent from Tibotec's scheme.
In the panel of 4248 isolates, phenotyping rated 95.9% resistant to nevirapine, 87.6% resistant to efavirenz, and 79.8% resistant to delavirdine. All isolates were resistant to at least one of these nonnucleosides. Tibotec's score called 49.4% of these isolates resistant to etravirine, while Monogram's score called 48.8% resistant. Phenotyping rated 41.2% of these isolates resistant to etravirine (with a fold-change in susceptibility of 3 or less).
Tested in isolates from 323 DUET trial participants, Tibotec's system ranked 39.9% resistant, Monogram called 38.7% resistant, and phenotyping called 32.8% resistant. Looking at all the DUET isolates tagged sensitive by Monogram, Tibotec's tool found 96.5% sensitive and phenotyping found 93.4% sensitive. Looking at all DUET isolates considered sensitive by Tibotec, Monogram's method called 98.5% sensitive and phenotyping called 92.8% sensitive.
What about the DUET isolates on which Tibotec and Monogram disagreed? Monogram's system assigned resistance to 3 isolates that Tibotec called sensitive. In all cases, Vingerhoets traced the discrepancy to mutations with high weights in the Monogram score but absent from the Tibotec score: V179E, Y188L, and G190Q. Two of these 3 people reached a viral load below 50 copies by DUET week 24. Tibotec called 7 isolates resistant when Monogram called them sensitive. Five of these 7 had mutations in the Tibotec list but not in the Monogram score: A98G, V179T, G190A, and G190S. Three of these 7 people had a sub-50 load after 24 weeks of treatment.
Next, with the DUET patients as a model, Vingerhoets figured how well Tibotec and Monogram tools would predict a viral load under 50 copies after 24 weeks of therapy. Almost three quarters of patients with virus rated sensitive by either genotypic score reached a sub-50-copy load by week 24. A slightly lower proportion of patients with virus called sensitive by phenotyping reached sub-50 territory by week 24. All three approaches proved less accurate in correctly forecasting nonresponse to etravirine (the proportion of nonresponders among people with virus rated resistant): 53.9% with the Tibotec system, 53.5% with the Monogram system, and 56.0% with phenotyping.
Finally, Vingerhoets looked at how well the Stanford University online database would predict responses of DUET participants, and it matched results with the Tibotec and Monogram tools.
1. Tambuyzer L, Vingerhoets J, Azijn H, et al. Comparison of two ETR weighted genotypic scores with phenotypic susceptibility and virologic response data. 7th European HIV Drug Resistance Workshop, March 25-27, 2009, Stockholm. Abstract 114.
2. Vingerhoets J, Peeters M, Azijn H, et al. An update of the list of NNRTI mutations associated with decreased virological response to etravirine: multivariate analyses on the pooled DUET-1 and DUET-2 clinical trial data. Antivir Ther. 2008;13(suppl 3):A26.
3. Benhamida J, Chappey C, Coakley E, Parkin NT. HIV-1 genotype algorithms for prediction of etravirine susceptibility: novel mutations and weighting factors identified through correlations to phenotype. Antivir Ther 2008;13(suppl 3):A142.