Cross-validated stepwise regression for identification of novel non-nucleoside reverse transcriptase inhibitor resistance associated mutations.

BACKGROUND: Linear regression models are used to quantitatively predict drug resistance, the phenotype, from the HIV-1 viral genotype. As new antiretroviral drugs become available, new resistance pathways emerge and the number of resistance associated mutations continues to increase. To accurately identify which drug options are left, the main goal of the modeling has been to maximize predictivity and not interpretability. However, we originally selected linear regression as the preferred method for its transparency as opposed to other techniques such as neural networks. Here, we apply a method to lower the complexity of these phenotype prediction models using a 3-fold cross-validated selection of mutations. RESULTS: Compared to standard stepwise regression we were able to reduce the number of mutations in the reverse transcriptase (RT) inhibitor models as well as the number of interaction terms accounting for synergistic and antagonistic effects. This reduction in complexity was most significant for the non-nucleoside reverse transcriptase inhibitor (NNRTI) models, while maintaining prediction accuracy and retaining virtually all known resistance associated mutations as first order terms in the models. Furthermore, for etravirine (ETR) a better performance was seen on two years of unseen data. By analyzing the phenotype prediction models we identified a list of forty novel NNRTI mutations, putatively associated with resistance. The resistance association of novel variants at known NNRTI resistance positions: 100, 101, 181, 190, 221 and of mutations at positions not previously linked with NNRTI resistance: 102, 139, 219, 241, 376 and 382 was confirmed by phenotyping site-directed mutants. CONCLUSIONS: We successfully identified and validated novel NNRTI resistance associated mutations by developing parsimonious resistance prediction models in which repeated cross-validation within the stepwise regression was applied. Our model selection technique is computationally feasible for large data sets and provides an approach to the continued identification of resistance-causing mutations.

HIV-1 resistance testing influences treatment decision-making

OBJECTIVE: To investigates how the use of HIV-1 resistance tests influences physician decision-making. METHODS: Ten experienced reference physicians from the Brazilian Network for Drug Resistance each received ten patients' case histories. The selected patients had experienced at least two virological failures. First, reference physicians were asked to empirically select a new regimen for each patient. Second, after genotype report (ViroSeq 2.6) was provided, and physicians were again asked to select a new regimen considering this additional information. Finally, they were asked to select a regimen after receiving a virtual phenotype result (vircoTYPE 3.9.00). RESULTS: In 79 percent of the cases, physicians changed their empirical choice of regimen after receiving the genotype report, resulting in an increase in the mean number of active drugs from 1.8 to 2.2 (p = 0.0003), while the average number of drugs/regimen remained at 4.0. After receipt of the virtual phenotype report, additional changes were made in 75 percent of the patient cases, resulting in an increase in the number of active drugs to 2.8 (p < 0.0001), while the average number of drugs/regimen remained at 4.0. After receipt of the genotype report, 48 percent of the changes were in NRTIs, 29 percent were in NNRTIs and 60 percent were in PIs; after consideration of the virtual phenotype, 61 percent, 10 percent and 49 percent of the changes, respectively, were in these categories of drugs. Fourteen percent of the physicians rated the genotype report as "extremely useful", whereas 34 percent rated the subsequent virtual phenotype report as "extremely useful" (p = 0.0003). CONCLUSIONS: Resistance testing has a significant impact on physicians' choices of antiretroviral salvage therapies, and it promotes the selection of more active drugs.

HIV-1 resistance testing influences treatment decision-making.

OBJECTIVE: To investigates how the use of HIV-1 resistance tests influences physician decision-making. METHODS: Ten experienced reference physicians from the Brazilian Network for Drug Resistance each received ten patients' case histories. The selected patients had experienced at least two virological failures. First, reference physicians were asked to empirically select a new regimen for each patient. Second, after genotype report (ViroSeq 2.6) was provided, and physicians were again asked to select a new regimen considering this additional information. Finally, they were asked to select a regimen after receiving a virtual phenotype result (vircoTYPE 3.9.00). RESULTS: In 79% of the cases, physicians changed their empirical choice of regimen after receiving the genotype report, resulting in an increase in the mean number of active drugs from 1.8 to 2.2 (p = 0.0003), while the average number of drugs/regimen remained at 4.0. After receipt of the virtual phenotype report, additional changes were made in 75% of the patient cases, resulting in an increase in the number of active drugs to 2.8 (p < 0.0001), while the average number of drugs/regimen remained at 4.0. After receipt of the genotype report, 48% of the changes were in NRTIs, 29% were in NNRTIs and 60% were in PIs; after consideration of the virtual phenotype, 61%, 10% and 49% of the changes, respectively, were in these categories of drugs. Fourteen percent of the physicians rated the genotype report as "extremely useful", whereas 34% rated the subsequent virtual phenotype report as "extremely useful" (p = 0.0003). CONCLUSIONS: Resistance testing has a significant impact on physicians' choices of antiretroviral salvage therapies, and it promotes the selection of more active drugs.

A comparison of HIV-1 drug susceptibility as provided by conventional phenotyping and by a phenotype prediction tool based on viral genotype.

Concordance between the conventional HIV-1 phenotypic drug resistance assay, PhenoSense (PS), and vircoTYPE HIV-1 (vT), a drug resistance assay based on prediction of the phenotype, was investigated in a data set from the Stanford HIV Resistance database (hivdb). Depending on the drug, between 287 and 902 genotype-phenotype data pairs were available for comparisons. Test results (fold-change values) in the two assays were highly correlated, with an overall mean correlation coefficient of 0.90 using single PS measurements. This coefficient rose to 0.94 when the vT results were compared to the mean of repeat PS measurements. These results are comparable with the corresponding correlation coefficients of 0.87 and 0.95, calculated using single measurements, and the mean of repeat measurements, respectively, as obtained in the Antivirogram assay, the conventional HIV-1 phenotypic drug resistance test on which vT is based. The proportion of resistance calls resulting in a "major" discordance (fully susceptible or maximal response by one assay but fully resistant or minimal response by the other) ranged from 0% to 8.1% for drugs for which two clinical test cut-offs were available in both assays (didanosine, abacavir, tenofovir, saquinavir/r, fosamprenavir/r, and lopinavir/r), from 2.4% to 8.1% for the drugs for which two clinical test cut-offs were available in the vT assay and one clinical test cut-off in the PS assay (lamivudine, stavudine, indinavir/r, and atazanavir/r) and from 3.1% to 10.3% for drugs for which biological test cut-offs were used (zidovudine, nevirapine, delavirdine, efavirenz, indinavir, ritonavir, nelfinavir, saquinavir, and fosamprenavir). Our analyses suggest that these assays provide comparable resistance information, which will be of value to physicians who may be presented with either or both types of test report in their practice.

Clinical cut-offs for HIV-1 phenotypic resistance estimates: update based on recent pivotal clinical trial data and a revised approach to viral mixtures.

The clinical utility of HIV-1 resistance testing is dependent upon accurate interpretation and application of results. The development of clinical cut-offs (CCOs) for most HIV antiretroviral drugs assessed by the vircoTYPE HIV-1 resistance test has been described previously. Updated CCOs based on new methodology and new data from clinical cohorts and pivotal clinical studies are presented in this communication. Data for analysis included the original records for CCO derivation from eight clinical trials and two cohort studies plus new records from the clinical cohorts and from the TITAN, POWER, and DUET clinical studies. Drug-specific linear regression models were developed to describe the relationship between baseline characteristics (phenotypic resistance as estimated by virtualPhenotype-LM using methods revised recently for handling mixed viral sequences; viral load; and treatment history), new treatment regimen, and 8-week virologic outcome. The clinical cut-offs were defined as the estimated phenotypic resistance levels (fold change, FC) associated with a 20% and 80% loss of drug activity. The development dataset included 6550 records with an additional 2299 reserved for validation. The updated, v.4.2 CCOs were generally close to the v4.1 values, with a trend observed toward marginally higher cut-offs for the NRTIs. These results suggest that the updated CCOs provide a relevant tool for estimating the contribution to virological response of individual antiviral drugs in antiretroviral drug combinations as used currently in clinical practice.

Advantages of predicted phenotypes and statistical learning models in inferring virological response to antiretroviral therapy from HIV genotype.

BACKGROUND: Inferring response to antiretroviral therapy from the viral genotype alone is challenging. The utility of an intermediate step of predicting in vitro drug susceptibility is currently controversial. Here, we provide a retrospective comparison of approaches using either genotype or predicted phenotypes alone, or in combination. METHODS: Treatment change episodes were extracted from two large databases from the USA (Stanford-California) and Europe (EuResistDB) comprising data from 6,706 and 13,811 patients, respectively. Response to antiretroviral treatment was dichotomized according to two definitions. Using the viral sequence and the treatment regimen as input, three expert algorithms (ANRS, Rega and HIVdb) were used to generate genotype-based encodings and VircoTYPE() 4.0 (Virco BVBA, Mechelen, Belgium) was used to generate a predicted -phenotype-based encoding. Single drug classifications were combined into a treatment score via simple summation and statistical learning using random forests. Classification performance was studied on Stanford-California data using cross-validation and, in addition, on the independent EuResistDB data. RESULTS: In all experiments, predicted phenotype was among the most sensitive approaches. Combining single drug classifications by statistical learning was significantly superior to unweighted summation (P<2.2x10(-16)). Classification performance could be increased further by combining predicted phenotypes and expert encodings but not by combinations of expert encodings alone. These results were confirmed on an independent test set comprising data solely from EuResistDB. CONCLUSIONS: This study demonstrates consistent performance advantages in utilizing predicted phenotype in most scenarios over methods based on genotype alone in inferring virological response. Moreover, all approaches under study benefit significantly from statistical learning for merging single drug classifications into treatment scores.

Additional HIV-1 mutation patterns associated with reduced phenotypic susceptibility to etravirine in clinical samples.

We investigated the phenotypic impact of a number of uncommon amino acid substitutions at HIV-1 reverse transcriptase positions 103 and 138, which are not part of the etravirine resistance score and were found in combination with the high-impact mutation K101P. Etravirine phenotypic fold changes were 380-1400 for K101P + E138A/G/Q + K103N/S/T + V179I and 12-130 for K101P + (K103S +/- V179I) in the absence of E138A/G/Q. Although the effect of K103S is unclear, additional position 138 substitutions seem important for etravirine susceptibility.

Resistance and virological response analyses in a three initial treatment strategy trial: a substudy of the INITIO trial.

PURPOSE: To study the impact of baseline resistance mutations and HIV-1 subtypes on virological response to first-line antiretroviral therapy and to analyse the concordance of the results of two antiretroviral resistance interpretation tools in the INITIO trial. METHOD: Genotype and virco TYPE resistance analyses were studied at baseline, Year 2, Year 3, and at first therapeutic failure on plasma specimens stored at -80 degrees C. Relations between resistance mutations at baseline, subtype, initial virological response, and virological outcome after Week 24 were studied. RESULTS: 781 participants had genotypic results available at baseline. Therapeutic failure occurred for 112 participants. Initial virological response as well as virological outcome after Week 24 were not associated with HIV subtype. Before Week 24, the proportion of participants remaining under strict initial regimen was lower in patients with resistance mutations at baseline than in those without any resistance mutations. Presenceof resistance mutations at baseline also impacted negatively long-term virological outcome. Few discrepancies were observed between genotypic and virco TYPE for resistance interpretation. CONCLUSION: These data showed that presence of resistance mutations at baseline was associated with a poorer long-term virological outcome in the INITIO trial.

A randomized trial of therapeutic drug monitoring of protease inhibitors in antiretroviral-experienced, HIV-1-infected patients.

OBJECTIVE: Whether therapeutic drug monitoring of protease inhibitors improves outcomes in HIV-infected patients is controversial. We evaluated this strategy in a randomized, open-label clinical trial, using a normalized inhibitory quotient (NIQ), which incorporates drug exposure and viral drug resistance. NIQs < or = 1 may predict poor outcome and identify patients who could benefit from dose escalation. DESIGN/METHODS: Eligible patients had a viral load > or =1000 copies/ml on a failing regimen, and began a new protease inhibitor containing regimen at entry. All FDA-approved protease inhibitors available during the study recruitment (June 2002-May 2006) were allowed. One hundred and eighty-three participants with NIQ < or = 1, on the basis of their week 2 protease inhibitor trough concentration and pre-entry drug resistance test, were randomized at week 4 to standard of care (SOC) or protease inhibitor dose escalation (TDM). The primary endpoint was change in log10 plasma HIV-1 RNA concentration from randomization to 20 weeks later. RESULTS: Ninety-one patients were randomized to SOC and 92 to TDM. NIQs increased more in the TDM arm compared to SOC (+69 versus +25%, P = 0.01). Despite this, TDM and SOC arms showed no difference in outcome (+0.09 versus +0.02 log10, P = 0.17). In retrospective subgroup analyses, patients with less HIV resistance to their protease inhibitors benefited from TDM (P = 0.002), as did black and Hispanic patients (P = 0.035 and 0.05, respectively). Differences between black and white patients persisted when accounting for protease inhibitor susceptibility. CONCLUSIONS: There was no overall benefit of TDM. In post hoc subgroup analyses, TDM appeared beneficial in black and Hispanic patients, and in patients whose virus retained some susceptibility to the protease inhibitors in their regimen.

Determination of clinically relevant cutoffs for HIV-1 phenotypic resistance estimates through a combined analysis of clinical trial and cohort data.

BACKGROUND: Clinically relevant cutoffs are needed for the interpretation of HIV-1 phenotypic resistance estimates as predicted by "virtual" phenotype HIV resistance analysis. METHODS: Using a clinical data set containing 2596 treatment change episodes in 2217 patients in 8 clinical trials and 2 population-based cohorts, drug-specific linear regression models were developed to describe the relation between baseline characteristics (resistance, viral load, and treatment history), new treatment regimen selected, and 8-week virologic outcome. RESULTS: These models were used to derive clinical cutoffs (CCOs) for 6 nucleoside/nucleotide reverse transcriptase inhibitors (zidovudine, lamivudine, stavudine, didanosine, abacavir, and tenofovir), 3 unboosted protease inhibitors (PIs; indinavir, amprenavir, and nelfinavir), and 4 ritonavir-boosted PIs (indinavir/ritonavir, amprenavir/ritonavir, saquinavir/ritonavir, lopinavir/ritonavir). The CCOs were defined as the phenotypic resistance levels (fold change [FC]) associated with a 20% and 80% loss of predicted wild-type drug effect and depended on the drug-specific dynamic range of the assay. CONCLUSIONS: The proposed CCOs were better correlated with virologic response than were biological cutoffs and provide a relevant tool for estimating the resistance to antiretroviral drug combinations used in clinical practice. They can be applied to diverse patient populations and are based on a consistent methodologic approach to interpreting phenotypic drug resistance.

A 21-base pair insertion/duplication at codon 69 of the HIV type 1 reverse transcriptase in a patient undergoing multiple nucleoside therapy.

We describe the selection of a previously unreported 21-base pair insertion following codon 69 of the HIV-1 reverse transcriptase (RT) from a patient undergoing multiple nucleoside analogue therapy. This insertion was a direct duplication of the preceding 21 bases of HIV-RT, and was selected in a background of NRTI-resistance mutations including substitutions at RT codons 41, 67, 184, 210, and 215. Longitudinal genotypic and phenotypic resistance tests performed before and after selection of the insertion suggested that the insertion conferred an additional decrease in susceptibility to some nucleoside analogues, most notably didanosine, stavudine, abacavir, and tenofovir. However, phenotypic analysis of an insertion-containing site-directed mutant constructed in an HIV-1 HXB2 background revealed no direct association between the 21-base pair insertion and decreased susceptibility to NRTIs, suggesting that the insert requires the context of the patients' virus in order to confer resistance. These observations may offer new insight into the relative contribution of HIV-RT codon 69 insertion mutations to antiretroviral resistance.

Investigation of baseline susceptibility to protease inhibitors in HIV-1 subtypes C, F, G and CRF02_AG.

OBJECTIVE: To compare baseline susceptibility to protease inhibitors among HIV-1 isolates of subtypes C, F, G and CRF02_AG, and to identify polymorphisms that determine the differences in susceptibility. METHODS: A total of 42 samples of drug-naive patients infected with subtypes G (n=19), CRF02_AG (n = 10), F (n = 6) and C (n = 7) were phenotyped and genotyped with the Antivirogram and the ViroSeq 2.0 genotyping system, respectively. A Bayesian network approach was used for a preliminary analysis of the collected data and the dependencies indicated by the network were statistically confirmed. RESULTS: CRF02_AG samples were found to be more susceptible to nelfinavir and ritonavir than other subtypes. Hypersusceptibility to these drugs was associated with the 70R polymorphism. 37D/S/T was associated with reduced susceptibility to indinavir and 89M with reduced susceptibility to lopinavir. Susceptibility to tipranavir was the lowest among the subtype F samples and the highest for subtype G samples, with samples carrying 57R being more susceptible than samples carrying 57K. CONCLUSIONS: Our study suggests that there are baseline susceptibility differences between subtypes and these differences are due to naturally occurring polymorphisms in these subtypes. The predictive value for phenotype of these polymorphisms was even valid in subtypes where these polymorphisms are less prevalent. Taking into account such polymorphisms should improve current algorithms for interpretation of genotyping results in a subtype-independent way.

Prevalence of the HIV-1 protease mutation I47A in clinical practice and association with lopinavir resistance.

Mutation proI47A has recently been associated with lopinavir/ritonavir (LPV/r) resistance. Only four out of 1859 specimens (0.2%) sent for drug resistance testing (219 drug-naive and 1650 antiretroviral-experienced) showed I47A. All belonged to patients failing LPV/r. The prevalence among protease inhibitor-experienced patients was 0.6%. Phenotypic testing showed that proI47A caused high-level lopinavir resistance (> 100-fold) and cross-resistance to amprenavir, whereas it caused hypersusceptibility to saquinavir. ProI47A should thus be considered the primary lopinavir resistance mutation.

The K65R mutation in human immunodeficiency virus type 1 reverse transcriptase exhibits bidirectional phenotypic antagonism with thymidine analog mutations.

The K65R mutation in human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) is selected in vitro by many D-nucleoside analog RT inhibitors (NRTI) but has been rarely detected in treated patients. In recent clinical trials, the K65R mutation has emerged frequently in patients experiencing virologic failure on antiretroviral combinations that do not include 3'-azidothymidine (AZT). The reason for this change is uncertain. To gain insight, we examined trends in the frequency of K65R in a large genotype database, the association of K65R with thymidine analog mutations (TAMs) and other NRTI mutations, and the viral susceptibility profile of HIV-1 with K65R alone and in combination with TAMs. Among >60,000 clinical samples submitted for genotype analysis that contained one or more NRTI resistance mutations, the frequency of K65R increased from 0.4% in 1998 to 3.6% in 2003. Among samples with K65R, a strong negative association was evident with the TAMs M41L, D67N, L210W, T215Y/F, and K219Q/E (P<0.005) but not with other NRTI mutations, including the Q151M complex. This suggested that K65R and TAMs are antagonistic. To test this possibility, we generated recombinant HIV-1 encoding K65R in two different TAM backgrounds: M41L/L210W/T215Y and D67N/K70R/T215F/K219Q. K65R reduced AZT resistance from >50-fold to <2.5-fold in both backgrounds. In addition, TAMs antagonized the phenotypic effect of K65R, reducing resistance to tenofovir, abacavir, 2',3'-dideoxycytidine, dideoxyinosine, and stavudine. In conclusion, K65R and TAMs exhibit bidirectional phenotypic antagonism. This antagonism likely explains the negative association of these mutations in genotype databases, the rare emergence of K65R with antiretroviral therapies that contain AZT, and its more frequent emergence with combinations that exclude AZT.

A randomized controlled trial of genotypic HIV drug resistance testing in HIV-1-infected children: the PERA (PENTA 8) trial.

OBJECTIVE: To evaluate the longer-term utility of genotypic resistance testing in HIV-1-infected children with virological failure. METHODS: Children aged 3 months-18 years switching antiretroviral therapy (ART) with HIV-1 RNA > 2,000 copies/ml were randomized between genotypic testing (Virtual Phenotype) and no testing at baseline and subsequent virological failures. Children were followed to at least 96 weeks. RESULTS: One hundred and seventy eligible children, from 24 clinical centres in six countries, were randomized to resistance testing (n = 87) or no testing (n = 83) between June 2000-July 2003. At baseline, mean HIV-1 RNA and CD4+ T-cell percentage were 4.7 log10 copies/ml and 20%, respectively. Children had taken ART for a mean of 5 years; 24% had received all three classes, 53% nucleoside reverse transcriptase inhibitors (NRTIs)+protease inhibitors (PIs), 9% NRTIs+non-nucleoside reverse transcriptase inhibitors (NNRTIs) and 14% NRTIs only. There was no difference between the arms in the drug classes or the individual PIs/NNRTIs prescribed. However, 49% in the resistance test arm (RT) versus 19% in the no-test arm (NT) continued at least one NRTI from their failing regimen; 56% versus 19% were prescribed didanosine+stavudine as their NRTI backbone. Adjusting for baseline HIV-1 RNA, mean reductions in HIV-1 RNA at 48 weeks were 1.51 log10 copies/ml in the RT arm and 1.23 in the NT arm (P = 0.3); the difference between the arms was smaller at week 96 (RT: 1.50, NT: 1.47; P = 0.9). CONCLUSION: In this first paediatric trial of resistance testing, we observed a substantial difference in NRTI-prescribing behaviour across arms. However statistically significant evidence of a long-term virological or immunological benefit was not observed. This trial is registered as an International Standard Randomised Controlled Trial, number ISRCTN14367816.

Protease mutation M89I/V is linked to therapy failure in patients infected with the HIV-1 non-B subtypes C, F or G.

OBJECTIVE: To investigate whether and how mutations at position 89 of HIV-1 protease were associated with protease inhibitor (PI) failure, and what is the impact of the HIV-1 subtype. METHODS: In a database containing pol nucleotide sequences and treatment history, the correlation between PI experience and mutations at codon 89 was determined separately for subtype B and several non-B subtypes. A Bayesian network model was used to map the resistance pathways in which M89I/V is involved for subtype G. The phenotypic effect of M89I/V for several PIs was also measured. RESULTS: The analysis showed that for the subtypes C, F and G in which the wild-type codon at 89 was M compared to L for subtype B, M89I/V was significantly more frequently observed in PI-treated patients displaying major resistance mutations to PIs than in drug-naive patients. M89I/V was strongly associated with PI resistance mutations at codons 71, 74 and 90. Phenotypically, M89I/V alone did not confer a reduced susceptibility to PIs. However, when combined with L90M, a significantly reduced susceptibility to nelfinavir was observed (P < 0.05) in comparison with strains with L90M alone. CONCLUSIONS: The results of the present study show that M89I/V is associated with PI experience in subtypes C, F and G but not in subtype B. M89I/V should be considered a secondary PI mutation with an important effect on nelfinavir susceptibility in the presence of L90M.

Rare mutations at codon 103 of HIV-1 reverse transcriptase can confer resistance to non-nucleoside reverse transcriptase inhibitors.

BACKGROUND: The K103N mutation in HIV-1 reverse transcriptase (RT) confers high-level resistance to current non-nucleoside reverse transcriptase inhibitors (NNRTI). The prevalence and resistance profile of HIV-1 with other substitutions at RT codon 103 is less well documented. METHODS: K103 substitutions among over 70,000 clinical samples submitted for routine antiretroviral resistance testing at two independent centres were examined. Phenotypic resistance profiles of isolates harboring rare K103 variants in the absence of known NNRTI-associated resistance mutations were retrieved from Virco's correlative genotype/phenotype database. Genotyped samples with known treatment histories were retrieved from the British Columbia Centre for Excellence in HIV/AIDS database. Site-directed mutants containing K103 variants were constructed and phenotyped. RESULTS: K103N, R and S were observed in 29, 1.8, and 0.9% of Virco isolates and in 16, 1.5 and 0.4% of British Columbia isolates. K103T/Q/H substitutions were observed only rarely (<0.2%). The prevalence of unusual codon 103 substitutions remained stable over 5 years, except K103S, which increased over fourfold in both datasets. K103R/Q-containing clinical isolates remained phenotypically susceptible to NNRTI, whereas K103S/T/H-containing isolates showed over 10-fold decreased NNRTI susceptibility. Among patients with a known treatment history, K103S/T/H were observed primarily in individuals failing NNRTI-containing regimens. Site-directed mutants confirmed decreased susceptibility to NNRTI in K103S/T/H-containing recombinants. CONCLUSION: Variants at HIV RT codon 103 other than K103N are observed relatively rarely in clinical isolates, but K103 S, T and H confer decreased susceptibility to NNRTI. These data are relevant for interpretive genotype algorithms and in the design of assays specific to RT codon 103 mutations.

Characterization of mutations in CRF01_AE virus isolates from antiretroviral treatment-naive and -experienced patients in Singapore.

OBJECTIVE: To determine the frequency of drug resistance mutations among antiretroviral treatment-naive and -experienced patients infected with CRF01_AE virus. DESIGN: Prospective observational study. METHODS: We recruited antiretroviral-experienced HIV-infected patients with suspected drug resistance and consecutive newly diagnosed drug-naive patients. Viral sequencing was performed using standard methods. Frequencies of mutations in CRF01_AE virus isolates were compared with reference data for subtype B virus. RESULTS: Sequences were obtained for CRF01_AE virus isolates from 69 patients with treatment exposure and 35 treatment-naive patients. Treatment-naive patients had numerous polymorphisms but no major drug resistance mutations. There were differences between CRF01_AE and subtype B viruses in several drug resistance mutations including the following: D67N, L210F, K101E, V106M, V179I/D, G190A/S/E, and G48V (which were more common in CRF01_AE virus) and M41L, T215Y, and V82A (which were less common in CRF01_AE virus). CONCLUSIONS: The pattern of treatment-related mutations in CRF01_AE virus differs from that in subtype B virus at a number of positions determining drug resistance. Understanding these differences is important for interpreting results of resistance tests and determining treatment choices.