Initiatives for Developing and Comparing Genotype Interpretation Systems
Initiatives for Developing and Comparing Genotype Interpretation Systems
Objectives: To investigate the concordance between any of the results of nine HIV-1 drug-resistance interpretation systems (ISs) and their ability to predict week 8 and week 24 virological responses to abacavir-containing combination therapy.
Patients and Methods: A total of 1306 HIV-infected patients with a viral load >500 HIV-1 RNA copies/mL and a baseline genotypic resistance test were included in the study. Predicted abacavir susceptibilities according to each rule-based IS were compared. Linear and logistic regressions were used to assess the prognostic value of each IS for week 8 and week 24 responses, respectively.
Results: A median of three (interquartile range 1-5) abacavir mutations were detected at baseline. Comparing the IS predictions for abacavir susceptibility, 9% to 45% of patients were predicted to have resistant (R) virus, 9% to 53% virus with intermediate (I) resistance, and 23% to 74% susceptible (S) virus. Overall, the median week 8 viral load reduction was 1.61 log10 copies/mL (95% confidence interval 1.52-1.71) and 50% of patients experienced virological failure at 24 weeks. Most ISs showed better virological responses with S and I viruses than with R viruses.
Conclusions: Despite some degree of variability in predicted abacavir susceptibility among ISs, most ISs are useful to predict virological response.
The interpretation of HIV-1 genotypic resistance is a difficult task because a large number of drug resistance mutations interact and emerge in complex patterns. In order to help to interpret HIV-1 drug resistance genotyping results, a number of genotypic resistance interpretation systems (ISs) have been developed in recent years, and some are widely used in clinics. Most of these systems are rule-based ISs, in which the rules are developed by an expert (or a panel of experts), and some are freely accessible through a website.
Substantial differences in predicted drug susceptibility among these rule-based ISs have been reported, especially for drugs, such as abacavir, didanosine and lopinavir/ritonavir, requiring multiple mutations for resistance development. More recently, a number of new interpretation rules for single drugs have been developed by correlating genotypic resistance data with short-term virological response via statistical modelling. Some of these statistically based genotypic scores are, however, based on small studies and have not been formally cross-validated to establish whether their predictive values are retained in different study
populations. Other tools [e.g. the virco® TYPE HIV-1 test, details at http://www.vircolab.com] that link genotypes directly to decreased in vitro susceptibility do not suffer from the issue of small sample size, although a two-stage approach is needed to correlate genotypic data with viral load change in vivo.
Abacavir is a nucleoside reverse transcriptase inhibitor (NRTI) frequently used in antiretroviral combination therapies. Resistance to abacavir is relatively slow to develop in vitro and in vivo, requiring multiple mutations (such as K65R, L74V, Y115F, and/or M184V) before susceptibility is significantly reduced. Additionally, two mutational patterns associated with NRTI multi-drug resistance, the Q151 M complex and the family of amino acid insertions between codons 67 and 70 of reverse transcriptase (RT), both result in abacavir resistance. Moreover, the whole family of mutations selected by the thymidine NRTIs zidovudine and stavudine, called thymidine analogue mutations (TAMs), may influence the susceptibility to other NRTIs, including abacavir. Hence, the impact of mutations on the antiviral activity of abacavir in patients can be difficult to predict.
The aim of our study was to develop a standardized method to evaluate the concordance between predicted abacavir susceptibilities obtained using different rule-based expert ISs and to validate such predictions against virological response.
Objectives: To investigate the concordance between any of the results of nine HIV-1 drug-resistance interpretation systems (ISs) and their ability to predict week 8 and week 24 virological responses to abacavir-containing combination therapy.
Patients and Methods: A total of 1306 HIV-infected patients with a viral load >500 HIV-1 RNA copies/mL and a baseline genotypic resistance test were included in the study. Predicted abacavir susceptibilities according to each rule-based IS were compared. Linear and logistic regressions were used to assess the prognostic value of each IS for week 8 and week 24 responses, respectively.
Results: A median of three (interquartile range 1-5) abacavir mutations were detected at baseline. Comparing the IS predictions for abacavir susceptibility, 9% to 45% of patients were predicted to have resistant (R) virus, 9% to 53% virus with intermediate (I) resistance, and 23% to 74% susceptible (S) virus. Overall, the median week 8 viral load reduction was 1.61 log10 copies/mL (95% confidence interval 1.52-1.71) and 50% of patients experienced virological failure at 24 weeks. Most ISs showed better virological responses with S and I viruses than with R viruses.
Conclusions: Despite some degree of variability in predicted abacavir susceptibility among ISs, most ISs are useful to predict virological response.
The interpretation of HIV-1 genotypic resistance is a difficult task because a large number of drug resistance mutations interact and emerge in complex patterns. In order to help to interpret HIV-1 drug resistance genotyping results, a number of genotypic resistance interpretation systems (ISs) have been developed in recent years, and some are widely used in clinics. Most of these systems are rule-based ISs, in which the rules are developed by an expert (or a panel of experts), and some are freely accessible through a website.
Substantial differences in predicted drug susceptibility among these rule-based ISs have been reported, especially for drugs, such as abacavir, didanosine and lopinavir/ritonavir, requiring multiple mutations for resistance development. More recently, a number of new interpretation rules for single drugs have been developed by correlating genotypic resistance data with short-term virological response via statistical modelling. Some of these statistically based genotypic scores are, however, based on small studies and have not been formally cross-validated to establish whether their predictive values are retained in different study
populations. Other tools [e.g. the virco® TYPE HIV-1 test, details at http://www.vircolab.com] that link genotypes directly to decreased in vitro susceptibility do not suffer from the issue of small sample size, although a two-stage approach is needed to correlate genotypic data with viral load change in vivo.
Abacavir is a nucleoside reverse transcriptase inhibitor (NRTI) frequently used in antiretroviral combination therapies. Resistance to abacavir is relatively slow to develop in vitro and in vivo, requiring multiple mutations (such as K65R, L74V, Y115F, and/or M184V) before susceptibility is significantly reduced. Additionally, two mutational patterns associated with NRTI multi-drug resistance, the Q151 M complex and the family of amino acid insertions between codons 67 and 70 of reverse transcriptase (RT), both result in abacavir resistance. Moreover, the whole family of mutations selected by the thymidine NRTIs zidovudine and stavudine, called thymidine analogue mutations (TAMs), may influence the susceptibility to other NRTIs, including abacavir. Hence, the impact of mutations on the antiviral activity of abacavir in patients can be difficult to predict.
The aim of our study was to develop a standardized method to evaluate the concordance between predicted abacavir susceptibilities obtained using different rule-based expert ISs and to validate such predictions against virological response.
Source...