Comparison of genotyping and phenotyping methods for determining susceptibility of HIV-1 to antiretroviral drugs

Objective(s): To compare antiretroviral resistance susceptibility testing o f patient HIV-1 strains using genotype and phenotype methods. Design: Eighteen plasma samples with viral load > 2000 HIV-1 RNA copies/ml were randomly selected for testing by both methods. Disease and treatment d ata were available for all patients. Methods: Samples were analysed genotypically using a kit assay (HIV-1 Genot yping Systems, Applied Biosystems), performed by the Clinical Research Labo ratory at Macfarlane Burnet Centre for Medical Research. Samples were analy sed phenotypically using a rapid phenotypic assay (PhenoSense (TM) HIV, Vir oLogic), performed by the manufacturer. Results from both methods were inte rpreted using a defined protocol. Each susceptibility assay was performed a nd interpreted by individuals unaware of either the clinical data or the re sults of the other susceptibility assay. Concordance was defined categorica lly as either the presence of reduced susceptibility (> 2.5-fold change) in the phenotypic assay and resistance associated mutations in the genotypic assay, or the absence of these findings in both assays. Results: Concordance between phenotypic and genotypic susceptibility testin g was 81% for nucleoside reverse transcriptase inhibitors, 91% for non-nucl eoside reverse transcriptase inhibitors and 90% for protease inhibitors. Co mplete concordance between phenotype and genotype for all 14 drugs evaluate d was observed in three (17%) patient samples. Conclusions: Phenotypic and genotypic susceptibility appear to provide simi lar results. However, interpretation of genotypic results can be complicate d, and both methods still require clinical validation. (C) 2001 Lippincott Williams & Wilkins.