Sixth International Congress Drug Therapy in HIV Infection 17-21 November, 2002 Glasgow, UK

The VirtualPhenotype™. is an HIV-1 resistance analysis tool that provides a quantitative, data-driven assessment of drug susceptibility for all currently approved HIV antivirals based on viral genotype. The basis of this analysis is a large dataset (G/P dataset, currently >28,000) of HIV-1 clinical isolate genotypes and matching drug susceptibility phenotypes. Each genotype to be analyzed is summarized as a mutational profile for each drug, composed of 8-14 logical expressions detailing the presence or absence of key mutations or groups of mutations with similar effects on drug susceptibility. Up to 16384 mutational profiles considering 31-54 mutations and their combinations are theoretically possible, but only a fraction of these (2-57%) are observed in clinical samples. To analyze a viral genotype, phenotypic results for isolates with the same mutational profile are retrieved from the G/P dataset. If µ10 results are available a quantitative prediction, the mean fold resistance (FR), is reported. Averaging FR across multiple test results on essentially equivalent viral strains enhances the accuracy of this quantitative assessment. If <10 results are available, a qualitative, rules based prediction of "resistance likely" or "resistance unlikely" is provided. The proportion of such "rules based calls" caused by rare mutational profiles is low and declines as the number of G/P pairs in the dataset increases (»2% for 28,000 G/P). Examination of the G/P dataset reveals that the dynamic range of drug susceptibility observed in clinical isolates is drug specific e.g. upper limit of »16 FR for d4T vs.»500 for EFV. For maximum clinical utility, phenotypic results should be set in a clinical context; however, preliminary "clinical cut-offs" have been described for only a few drugs. As an interim measure, "biological cut-offs" comparing the predicted FR for a viral genotype to the natural variation in susceptibility observed in wild type clinical isolates are provided. The virtual phenotype approach provides a rapid, cost effective, quantitative assessment of HIV-1 drug susceptibility.

Presenting author: L Bacheler

Download Conferences Abstracts

1 Tibotec-Virco, Durham, North Carolina, USA.

2 Mechelen, Belgium

2002-11-17
PL5-5

Copyright © 2002 - The Gardiner-Caldwell Group, Ltd.. All Rights Reserved. Reproduction of this abstract (other than one copy for personal reference) must be cleared through the Sixth International Congress on Drug Therapy in HIV Infection, c/o The Gardiner-Caldwell Group Ltd, part of The Thomson Corporation, Peakside House, Alder Court, Tytherington Business Park, Tytherington, Cheshire SK10 2XG, UK - Tel: +44 (0)1625 668000, Fax: +44 (0)1625 668121 Email: hiv6@gardiner-caldwell.com

This information is designed to support, not replace, the relationship that exists between you and your doctor. ©1980, 2005. AEGiS.