Update on HIV drug resistance

Daniel R. Kuritzkes, MD for

Resistance of HIV-1 to a growing number of drugs is a significant factor that limits the success of antiretroviral therapy.

In contrast to previous years, no single slide session was devoted to drug resistance at this year’s Conference on Retroviruses and Opportunistic Infections. However, several oral presentations at various sessions as well as a number of posters presented new or updated information on this topic.

The following review summarizes data regarding patterns of resistance to existing drugs, resistance to newer antiretroviral agents, the problem of primary drug resistance (spread of drug resistance through HIV-1 transmission), and drug resistance testing. Because of the large number of posters in this area, only those considered to be the most clinically significant or important will be covered.

Cross-resistance among nucleoside reverse transcriptase inhibitors (NRTIs)

When the NRTI’s first entered into clinical practice, resistance was thought to be relatively drug-specific. There were mutations that conferred zidovudine (ZDV; AZT) resistance, those that conferred resistance to didanosine (ddI) and zalcitabine (ddC), the unique pattern of resistance to lamivudine (3TC), and more recently mutations that confer resistance to abacavir (ABC). Resistance to d4T was difficult to pin down, either genotypically or phenotypically. Over the last one or two years, however, a growing appreciation for the extent of cross-resistance among the NRTI’s has emerged. In particular, it is now understood that high-level ZDV resistance can confer cross-resistance to most other NRTI’s.

Several presentations at this meeting provided additional evidence for cross-resistance between ZDV and d4T. Duan et al [1] studied inhibition of purified RT from a pair of ZDV sensitive and resistant clinical isolates of HIV-1. Enzyme from the ZDV-resistant isolate showed approximately 10-fold resistance to AZT-triphosphate and d4T-triphosphate (the activated forms of AZT and d4T), and also showed modest (3-fold) resistance to 3TC-triphosphate. These findings provide additional evidence for cross-resistance between ZDV and d4T and may provide a biochemical explanation for the lower efficacy of d4T in patients with prior ZDV treatment.

These results are supported by findings from NARVAL, a trial of genotypic and phenotypic resistance testing [2]. The association between presence of 3 or more ZDV resistance mutations with virologic response to d4T, ddI, ABC, or 3TC in the control (no resistance testing) arm was studied. Presence of these 3 or more ZDV resistance mutations at baseline was associated with a worse virologic response at week 12 to regimens that included d4T (P=0.0364) ddI (P=0.0213), and ABC (0.0676). These results confirm the effect of ZDV resistance mutations on response to other NRTI’s, even when they are part of a HAART regimen.

Similar results were obtained by Shulman et al [3] in an analysis of ACTG 302. In this study from the pre-HAART era, ZDV-experienced patients who subsequently received d4T monotherapy were unlikely to show a virologic response (0.3-log10 decrease in HIV-1 RNA from baseline) in the presence of ZDV resistance mutations (215Yor F, or a combination of mutations at codons 67, 70, and 219). Of note, a mutation at codon 70 alone (often the first ZDV resistance mutation to emerge) was associated with a greater likelihood of virologic response.

Different results were observed by Cohen et al in the VIRA3001 study, a randomised trial of phenotyping vs standard of care [4]. In that study, the frequency of ZDV resistance mutations among patients who had received only a single thymidine analogue prior to study entry as 61% for ZDV-treated patients and 69% for d4T-treated patients, respectively. Among the ZDV- or d4T-experienced patients who switched to a d4T- or ZDV-containing regimen, those who switched from ZDV to d4T (n=57) were more likely to have a treatment response at week 16 than were those who switched from d4T to ZDV (n=10) (60% vs 20%; P=0.036) in an intention-to-treat (missing=failure) analysis. These data should be interpreted with caution because of the small number of patients (10) who switched from d4T to ZDV, and should be viewed in the context of other studies presented here and at earlier meetings showing a reduced response to d4T in patients with ZDV-resistant virus.

Resistance to abacavir [Ziagen]

One problem in the interpretation of phenotypic drug resistance assays is defining the “break point” or cut-off between sensitive and resistant viruses. Up until recently, these definitions were based on assay variation, rather than on clinical data. Clinically relevant cut-offs for abacavir in the PhenoSense assay (ViroLogic) were determined by Lanier et al [5] using data from four clinical trials of abacavir (CAN 2003, 3001, 3002, and 3009). In each case, NRTI-experienced patients with detectable plasma HIV-1 RNA added ABC to a stable background regimen. Using several different statistical approaches, a significantly better response was observed when the patient’s virus was <4.5-fold resistant to ABC as compared to wild-type. As a result, a cut-off of 4.5-fold over wild-type has been established for ABC resistance in the PhenoSense assay.

In a related poster, Melby et al presented data on the emergence of NRTI resistance mutations following long-term initial therapy with ZDV/3TC/ABC in CNA3005 (ZDV/3TC/ABC vs ZDV/3TC/IDV) [6]. To date, 43/262 patients (16%) of patients assigned to the triple-NRTI arm have experienced virologic failure (two consecutive plasma HIV-1 RNA levels >400 copies/mL). One or more NRTI resistance mutations were found in 34/40 samples available for genotyping. Thirty-two of these had an M184V mutation at the first genotype following virologic failure; two patients had 1-2 ZDV resistance mutations in addition to the M184V mutation.

These results demonstrate that as with PI-containing regimens, the most common mutation following failure of this ABC-containing triple-therapy regimen was the 184V mutation, which confers resistance to 3TC and low-level resistance to ABC. Patients who remained on the ZDV/3TC/ABC regimen after virologic failure had reasonably stable virus loads over a median follow-up of 80 weeks (median plasma HIV-1 RNA = 3.5 log10 copies/mL), but showed accumulation of additional resistance mutations in their virus samples (Y115F or ZDV resistance mutations). These results reinforce the concept that although virologic failure may initially be accompanied by limited evidence of drug resistance, more extensive resistance patterns emerge with continued administration of a failing regimen.

Nucleotide reverse transcriptase inhibitors

Tenofovir (TDF) is an investigational nucleotide analogue inhibitor of HIV-1 RT that is in phase 3 clinical trials and recently has become available through expanded access. Previous work has shown that most ZDV-resistant isolates remain susceptible to TDF, and that presence of the 3TC-resistance mutation (M184V) increases susceptibility to TDF.

Miller et al presented an analysis of TDF resistance data and clinical response from Gilead Study 902 [7]. In that study, addition of TDF to background failing therapy resulted in an average reduction in plasma HIV-1 RNA over 24 weeks of 0.58 log10 copies/mL; at 48 weeks patients receiving TDF showed a mean virus load reduction of 0.62 log10. In the current study, baseline phenotypic resistance testing of samples from 53 patients showed a mean reduction in TDF susceptibility of 1.9-fold compared to wild-type and 13.8-fold for ZDV (Antivirogram, Virco). Only four patients had >4-fold reduced susceptibility to TDF at baseline.

The HIV-1 RNA response to TDF was significantly associated with susceptibility to TDF (P=0.007) and ZDV (P=0.035), but not other NRTIs. Isolates from 14 patients obtained at week 48 showed >2.5-fold reductions in TDF susceptibility. A K65R mutation (which confers TDF resistance) emerged in 4 patients and was associated with reduced susceptibility to TDF. Although exact cut-offs for clinical response to TDF based on phenotypic testing have not yet been established, data from this study clearly show a relationship between TDF susceptibility and treatment response, and also suggest that ZDV resistance may play a role in determining response to TDF.


Daminopurine dioxalane (DAPD) is a prodrug of dioxalane guanosine (DXG), which is currently in phase I/II clinical trials. The drug has in vitro activity against ZDV- and 3TC-resistant isolates of HIV-1, including those that carry the multi-nucleoside resistance mutation at codon 151. Resistance is conferred to DAPD by the K65R mutation and by the insertion mutation at codon 69 that is also associated with multi-nucleoside resistance.

Studies presented by Feng et al [8] with purified RT from wild-type isolates of HIV-1 demonstrated that DXG-TP is a potent inhibitor, and that the enzyme incorporated the natural substrate dGTP 17 times more efficiently than DXG-TP. The overall efficiency of incorporation of DXG-TP was not affected by mutations that confer resistance to ZDV or 3TC. These data provide biochemical support to the rationale for use of DAPD against NRTI-resistance viruses. Preliminary data presented at ICAAC (Eron et al. 40th ICAAC, Toronto, 2000 [Abstract 690]) demonstrated short-term (15-days) activity of DAPD as a single agent when added to the regimen of patients failing antiretroviral therapy. However, the utility of DAPD in salvage therapy remains to be determined in randomised clinical trials.

Lopinavir [Kaletra]

An important issue in determining strategies for antiretroviral therapy is the pattern of resistance that can be expected upon failure of a particular regimen. Although a considerable amount of data have accumulated regarding resistance to lopinavir (ABT-378) in isolates from patients failing other protease inhibitors, the small number of patients who fail first-line therapy with lopinavir/ritonavir has limited knowledge regarding expected patterns of resistance and cross-resistance in viruses from such patients. Two posters attempted to address this question.

Brun et al analysed cross-resistance to protease inhibitors among 56 viruses that were highly resistant to lopinavir in an attempt to estimate what patterns of protease inhibitor (PI) resistance might be expected from patients failing lopinavir/ritonavir regimens [9]. Five of these isolates came from PI-experienced patients who had viral rebound when receiving salvage therapy with lopinavir/ritonavir. (To date, no lopinavir-resistant viruses have been recovered from more than 470 treatment-naive patients who have received this combination for more than 48 weeks.) The fold-change in susceptibility to lopinavir was highly correlated with susceptibility to indinavir and ritonavir, but correlated poorly with amprenavir and saquinavir resistance. For example, viruses with a median 44-fold resistance to lopinavir were only 6-fold resistant to amprenavir. Similarly, viruses from PI-experienced patients failing lopinavir-based salvage therapy had 9- to 99-fold resistance to lopinavir but were <8.5-fold resistant to amprenavir; three of these isolates also remained saquinavir susceptible.

These data suggest that patients who experience treatment failure on lopinavir might still have a virologic response if treated subsequently with amprenavir- or saquinavir-based regimens. Clinical data from a single patient with a virus that developed 25-fold resistance to lopinavir but retained wild-type sensitivity to amprenavir, and had a good treatment response to that drug with plasma HIV-1 RNA levels that dropped to <400 copies/mL at week 12. Although these data are encouraging, it is important to note that the patients in this study had all been treated with other protease inhibitors prior to the use of lopinavir. Different patterns of cross-resistance might be observed depending on the specific mutations that emerge during treatment with lopinavir in previously PI-naive patients.

In a related study, Bernstein et al examined resistance patterns in HIV-1 isolates from treatment-naive patients failing antiretroviral therapy in the Abbott M98-863 study (lopinavir vs nelfinavir) [10]. Genotypic data are available from 31 patients who failed lopinavir/ritonavir and from 65 patients who failed nelfinavir. Whereas 13/31 (42%) isolates from lopinavir/ritonavir failures showed 3TC resistance, none had accumulated resistance mutations in the protease gene. By contrast, isolates from 21/65 (32%) patients failing nelfinavir had genotypic resistance of nelfinavir resistance, and 86% had resistance to 3TC. These findings are consistent with the prediction that if patients are monitored closely it is unlikely that lopinavir-resistant viruses will be found in patients failing first-line therapy with lopinavir/ritonavir.

Although the data remain somewhat limited, they provide some reassurance regarding the use of lopinavir/ritonavir as an initial protease inhibitor-containing regimen in patients for whom such therapy is appropriate.

Primary  drug resistance

The transmission of drug-resistant HIV-1 was first documented in 1993. More recently, a several well-documented cases of transmission of multi-drug resistant HIV-1 have been reported. Since then, there has been growing concern about the potential for widespread transmission of drug-resistant virus. Several presentations and posters at the 8th CROI in Chicago documented the prevalence of drug resistance among patients with newly acquired HIV-1 infection from around the world.

Weinstock et al [11] presented data on the prevalence of drug resistance mutations in HIV-1 samples from 437 treatment-naive patients from 10 U.S. cities collected during 1997-99. Approximately 10% of these patients were recently seroconverters; the remainder had chronic established HIV-1 infection. Overall, 45 (10%) had mutations associated with drug resistance: 8.5% for NRTI’s, 2.5% for nNRTI’s, and 0.7% for PI’s. Only 1% had evidence of multi-drug resistance. The prevalence of resistance mutations was similar in samples from patients with recent or established infection, arguing against a substantial increase over time in this population.

A contrasting view was presented by Little et al [12]. Patients are identified as having acute infection or as recent seroconverters from eight cities in the U.S. and Canada. Drug resistance was determined by the PhenoSense assay (ViroLogic, South San Francisco) and by genotyping (using ABI sequencing). Samples were characterized as having >2.5-fold or >10-fold resistance to drugs in each class of antiretroviral agent.

A total of 408 patients have been identified by this group. The prevalence of high-level (>10-fold) resistance to nNRTI’s and PI’s in samples from newly infected individuals increased sharply during 1999-2000 compared to 1996-98. The proportion of isolates with high-level resistance to the nNRTI’s increased from 1% to 7% (P=0.001) and for PI’s from 2% to 6% (P=0.05). Eight percent of newly infected patients identified after 1998 carried viruses that showed >10-fold resistance to at least one drug, and 4% had resistance to two or more drugs.

The time to virologic suppression (plasma HIV-1 RNA <500 copies/mL) was significantly longer for newly infected patients with viruses that had >10-fold resistance to one or more drugs as compared to those that had sensitive viruses (P=0.05). Among patients who achieved complete viral suppression on an initial regimen instituted during primary HIV infection, there was also a trend toward earlier failure of viral suppression for those with viruses that had >2.5-fold reduction in susceptibility to one or more drugs.

An increasing prevalence of transmitted drug-resistant virus was also noted by Simon et al [13], who reported on 61 newly infected patients identified during 1999-2000 in New York City and Montreal. Overall prevalence of primary drug resistance mutations in samples from this cohort was 26%, which represents an increase from 16% for the years 1995-98. These numbers may not be representative of the country as a whole, however, given the relatively small numbers and highly selected patient population. Nevertheless, these data show that rates of transmission of resistant virus can vary considerably depending on the group and geographic area under study.

A question that is often asked is how stable drug-resistant variants are after transmission, in the absence of antiretroviral therapy in the newly infected patient. Two abstracts addressed this issue. Garcia-Lerma et al from the CDC [14] identified 10 newly infected patients with unusual mutations at codon 215. A change from threonine (T) to tyrosine (Y) or phenylalanine (F) at this position is associated with ZDV resistance. This resistance mutation requires a change at two nucleotides within the 215 codon (eg, ACC to TAC). Novel variants reported in this poster included 215D, 215C, 215S, and 215E, which are thought to represent intermediates of a 215Y revertant (ie, only one of the two nucleotides at codon reverted back to wild-type).

Fitness assays showed that these partial revertants were more fit than the ZDV-resistant 215Y variant. Although these novel mutations do not by themselves confer resistance to ZDV, they should be considered as markers of previous selection by ZDV, and set the stage for more rapid emergence of ZDV resistance if treatment with ZDV is initiated in patients carrying such viruses.

Daar et al [15] reported on the case of a patient who presented with acute HIV-1 infection and was found to have a multidrug-resistant virus that carried both PR and RT resistance mutations, including the insertion mutation at RT codon 69. Virus load, which initially was 5.48 log10 copies/mL, fell to approximately 3.0 log (1,000 copies/mL) in the absence of treatment. More susceptible virus evolved over a 4-5 month period. At 5 months, a 10-fold increase in virus load above the previous set point of 1000 copies/mL corresponded to emergence of a virus that carried no primary drug resistance mutations.

These data demonstrate that although highly resistant viruses can be transmitted, their reduced replicative capacity may be associated with relatively low virus loads following primary infection. However, emergence of wild-type virus in the absence of drug treatment may subsequently be accompanied by a rise in virus load, most likely as a consequence of superior fitness of the wild-type virus.

Two other reports tracked the spread of drug-resistant HIV-1 in Europe. Yerly et al [16] reported on rates of drug resistance among patients with primary HIV infection in Switzerland. In that country, transmission of drug-resistant virus appears to have decreased over the last three years (1998-2000) as compared to the period 1996-97. The investigators attributed this reduction to the high rates of viral suppression achieved by the widespread availability and use of potent antiretroviral therapy.

Whereas only 10% of patients followed in the Swiss HIV Cohort study had undetectable levels plasma HIV-1 RNA in 1996, that figure increased to 53% in 1999. Yerly et al also noticed significant clustering of cases of primary HIV infection, which is a worrisome finding. Through nucleic acid sequence analysis, they were able to show epidemiologic links among several cases of primary HIV infection. In many instances, transmission occurred even before the index case developed symptoms. This situation is analogous to the spread of other viral infections (e.g., hepatitis A), in which the patient is most infectious prior to development of recognizable symptoms that lead to diagnosis. Efforts at contact tracing and counselling to halt the spread of HIV will be made more difficult if a substantial number of individuals spread the virus during this early “window” period.

Chaix et al reported the prevalence of drug-resistant HIV-1 in France [17]. 108 patients diagnosed with primary HIV-1 infection during 1999 were identified from three cohort studies. The frequency of drug resistance (identified by genotyping) was 6.5% having resistance to NRTI’s, 3.7% to the nNRTI’s, and 2.8% for the PI’s. Samples from two patients showed evidence of multi-drug resistance. Taken together, these results provide strong support for the need to perform drug resistance testing in patients with primary HIV-1 infection prior to initiating antiretroviral therapy.

Resistance testing and effect of drug resistance on response to treatment

Resistance testing is rapidly becoming an essential laboratory tool for help in selecting optimal antiretroviral regimens in HIV-1 infected patients. Several posters at the 8th CROI addressed important issues related to resistance testing and the ability of resistance testing to predict treatment success or failure. A critical question with regard to phenotypic resistance assays is the way in which “sensitive” and “resistant” viruses are defined. Ideally, such definitions would depend on showing that a given drug lost the ability to suppress virus replication at a particular IC50 or fold-resistance. Such data are available for only a handful of drugs, however (see discussion of cut-offs for abacavir, above). In the absence of such data, resistance has usually been defined by default as any level of susceptibility that exceeds the interassay variation (eg, 2.5- or 4-fold). This definition ignores the natural variation in susceptibility of wild-type viruses, however.

To address this shortcoming, Harrigan et al [18] analysed susceptibility data from 1000 isolates obtained from treatment-naive patients around the world using the Antivirogram assay (Virco). They found that 97.5% of isolates had shifts of less than 2.5- to 4-fold for the PI’s, less than 3.5-4.5-fold for the NRTI’s, but up to 5-10-fold for the nNRTI’s. Based on these data, new cut-offs have been defined for the Antivirogram assay.

Virtual phenotype

Interpretation of genotypic data is often difficult and confusing for patients and clinicians who are not experts in HIV-1 drug resistance. One approach to interpreting the genotype is the so-called “Virtual Phenotype”. The Virtual Phenotype uses computer-based artificial intelligence software to predict the likely phenotype for a given genotype by “matching” the genotype of a patient’s virus to other viruses in a database of >18,000 samples with paired genotypes and phenotypes. The phenotype of all the matching viruses is averaged, and reported as the Virtual Phenotype.

Previous work has shown that in most cases the virtual (predicted) phenotype is reasonably well correlated with the actual (measured) phenotype of a given virus. To validate the Virtual Phenotype against clinical response data, Graham et al [19] analysed baseline samples from 191 patients enrolled in VIRA 3001, a randomised trial of phenotyping vs standard of care. The Virtual Phenotype was a significant independent predictor of treatment outcome in several statistical models, and in some cases appeared to be more predictive than genotype.

These results provide additional support to the validity of the Virtual Phenotype as a tool for interpreting viral genotypes. However, the reliability of the Virtual Phenotype depends on a number of factors, including the specific mutations that are put into the search for a match, the number of matches found, and the distribution of drug susceptibility among the matches. Further work is needed to validate the Virtual Phenotype against other forms of resistance testing in clinical trials.

Inhibitory quotient

Another potential tool for the interpretation of resistance data is the “Inhibitory Quotient”. The IQ attempts to relate the measured trough level of a drug to the IC50 of the patient’s isolate for that drug (corrected for binding to plasma proteins). A high IQ means that the trough plasma concentration significantly exceeds the amount of drug needed to inhibit the virus in question; a low IQ suggests inadequate drug levels or a highly resistant virus.

Kempf et al [20] estimated IQ’s for patients receiving RTV-boosted IDV therapy for treatment of IDV-resistant virus. Phenotypes were predicted from genotypic data using the Virtual Phenotype. Response rates were significantly greater in patients with HIV-1 that was <6-fold resistant to IDV by Virtual Phenotype as compared to patients with virus that was >6-fold resistant to IDV (P<0.05). Using the Virtual Phenotype and measured trough concentrations of IDV to calculate a “virtual” IQ, the investigators found that response rates were significantly higher among patients with an IQ>2 as compared to those with an IQ<2 (P<0.003). These results suggest that combining phenotypic data with drug levels might be particularly useful in predicting treatment response. However, adjusting drug doses on the basis of the IQ in an attempt to overcome drug resistance may not be advisable, since the safety of very high drug levels that might be required in certain cases has not been evaluated.

Genotypic resistance

Updates on two recently completed studies of genotypic resistance testing were presented at the meeting. Tural et al presented follow-up data on the Havana study, which compared the utility of genotypic resistance testing, expert advice, or both as compared to standard of care in selecting salvage regimens for patients experiencing failure of antiretroviral therapy.

Genotyping and expert advice each resulted in significantly better virologic response. Response rates (% of patients with virus load <400 copies/mL at week 24) were 57.5% for the genotype arm vs 42.4% for the control arm (P=0.01), and 59.1% for the expert advice arm vs 41.1% for the no-advice arm (P=0.003). The best response rates were observed in patients who received both genotyping and expert advice as compared to patients who received neither genotyping nor expert advice (69.2% vs 36.4%, respectively; P=0.001). These results suggest that although expert advice is helpful, the availability of genotypic resistance assays leads to further improvements in virologic outcome of salvage therapy. In contrast to some other studies, the Havana study found that genotyping made the biggest difference in patients who had failed 3 or more regimens.

De Luca et al [21] presented final results of the Argenta study, which had been presented in part at the Durban meeting. In this study, 174 patients were randomised to genotype or standard of care. Twenty-five percent had failed >3 HAART regimens and 41% were triple-class experienced. Although a significant difference in the percent of patients with virus loads <500 copies/mL was observed at 12 weeks favouring the genotyping arm (27% vs 12%, P=0.02 for genotype and standard of care, respectively) the difference between arms was not statistically significant at week 24 (21% vs 17%, respectively).

Failure to observe a larger difference between the groups may be a result of the imbalance between arms with regard to a number of important parameters at study entry, including a larger proportion of patients with >2 primary resistance mutations and greater proportion of nNRTI-experienced patients in the genotyping arm. By contrast, duration of prior nNRTI therapy was longer in the standard of care arm. A beneficial effect of genotyping was more easily demonstrated when the analysis focused on adherent patients failing their third or earlier HAART regimen. These results are consistent with those of Viradapt and the NARVAL study, which suggest that treatment adherence and number of prior treatment regimens play an important role in determining the usefulness of drug resistance testing.

Two presentations on resistance testing reported the seemingly paradoxical observation that in some cases presence of resistance mutations appeared associated with an increased chance of treatment success. Genotypic analysis of isolates from patients participating in the NOVAVIR study (ANRS 073) provides additional evidence of cross-resistance between ZDV and d4T [22]. In that study, ZDV-, ddI-, or ddC-experienced patients who were 3TC- and PI-naive were randomised to ZDV/3TC/IDV or d4T/3TC/IDV. Similar virologic responses were observed in the two arms. A majority of patients had evidence of one or more ZDV resistance mutations at entry.

Curiously, presence of ZDV resistance mutations was associated with a significantly lower risk of virologic failure in both the ZDV- and d4T-containing arms. Similar results were reported by Tasker et al [23] who found greater decreases in plasma HIV-1 RNA and greater increases in CD4 cell count at six months among patients carrying drug-resistant virus. These results suggest the possibility that patients with wild-type virus were poorly adherent to therapy while on NRTI therapy, hence the absence of resistance mutations at baseline.


What were the most important take-home messages of the conference with regard to drug resistance? First, that cross-resistance between the NRTIs is more widespread than previously recognized, but that newer drugs in this class (as well as the nucleotide RT inhibitors) may have an important role in salvage therapy. Second, although transmission of drug resistant virus remains low in some cities, rates of primary resistance to the newer drugs (nNRTI’s and PI’s) increased sharply over the last two years in other cities, and might be associated with reduced effectiveness of treatment.

These data provide a rationale for performing drug resistance testing prior to initiating therapy in settings with high prevalence of drug resistant virus among newly infected individuals. Third, more data are needed in order to better define the cut-offs for sensitive and resistant virus in phenotypic assays, but improvements are being made. Fourth, the Virtual Phenotype appears to be a useful tool for interpreting genotypic data, although prospective clinical validation is needed. Finally, resistance testing appears to be useful in selecting salvage regimens in most studies, but further refinements are needed.


  1. Duan CY et al. Biochemical evidence of cross-resistant to stavudine (d4T) triphosphate in purified HIV-1 reverse transcriptase (RT) derived from a zidovudine (AZT)-resistant isolate. 8th CROI Abstract 442.
  2. Costagliola D et al. Presence of thymidine-associated mutations and response to d4T, abacavir, and ddI in the control arm of the NARVAL ANRS 088 trial. 8th CROI Abstract 450.
  3. vShulman N et al. Genotypic predictors of virologic response to stavudine after zidovudine monotherapy (ACTG 302). 8th CROI Abstract 437.
  4. Cohen C et al. Virologic suppression from different thymidine analogue (TA)-containing HAART regimen sequencing strategies:VIRA3001. 8th CROI Abstract 444.
  5. Lanier ER et al. Determination of a clinically relevant phenotypic resistance “cutoff” for abacavir using the PhenoSense assay. 8th CROI Abstract 254.
  6. Melby T et al. Time to Appearance of NRTI-Associated Mutations and Response to Subsequent Therapy for Patients on Failing ABC/COM. 8th CROI. Abstract 448.
  7. Miller MD et al. Baseline and week 48 final phenotypic analysis of HIV-1 from patients adding tenofovir disoproxil fumarate (TDF) therapy to background ART. 8th CROI Abstract 441.
  8. Feng J et al. Mechanistic studies of dioxolane guanosine 5′-triphosphate:implications for efficacy, lack of cross-resistance and selectivity of DAPD. 8th CROI Abstract 306.
  9. Brun S et al. Patterns of protease inhibitor cross-resistance in viral isolates with reduced susceptibility to ABT-378. 8th CROI Abstract 452.
  10. Bernstein B et al. Absence of resistance to Kaletra (ABT-378/r) observed through 48 weeks of therapy in antiretroviral naive subjects. 8th CROI Abstract 453.
  11. Weinstock H et al. Prevalence of mutations associated with decreased antiretroviral drug susceptibility among recently and chronically HIV-1-infected persons in 10 US cities, 1997-99. 8th CROI Abstract 265.
  12. Little SJ et al. Antiretroviral drug susceptibility and response to initial therapy among recently HIV-infected subjects in North America. 8th CROI Abstract 756.
  13. Simon V et al. Prevalence of drug-resistant HIV-1 variants in newly infected individuals during 1999-2000. 8th CROI Abstract 423.
  14. Garc’a-Lerma G et al. Unusual Mutations at Codon 215 of HIV-1 Reverse Transcriptase in Treatment- Naive, HIV-1-Infected Persons: Prevalence, Drug Susceptibility, and Replicative Fitness. 8th CROI Abstract 426
  15. Daar E et al. Viral evolution in an untreated patient who acquired multi-drug-resistant HIV during primary infection. 8th CROI Abstract 427.
  16. Yerly S et al. HIV drug resistance and molecular epidemiology in patients with primary HIV infection. 8th CROI Abstract 754.
  17. Chaix ML et al. Prevalence of genotypic drug resistance among French patients infected during the year 1999. 8th CROI Abstract 755.
  18. Harrigan PR et al. Worldwide Variation in Antiretroviral Phenotypic Susceptibility in Untreated Individuals. 8th CROI Abstract 455
  19. Graham N, Peeters M, Vergiest W, Harrigan R, Larder B. The Virtual Phenotype is an independent predictor of clinical response. 8th CROI Abstract 524.
  20. Kempf D et al. Response to ritonavir (RTV) intensification in indinavir (IDV) recipients is highly correlated with virtual inhibitory quotient. 8th CROI Abstract 523.
  21. De Luca A et al. A prospective, randomised study on the usefulnes of genotypic resistance testing and the assessment of patient-reported adherence in unselected patients failing potent HIV therapy (ARGENTA): final 6-month results. 8th CROI Abstract 433.
  22. Decamps D et al. Genotypic resistance to zidovudine (ZDV) and relationship to subsequent virological response in NOVAVIR ANRS 073 trial. 8th CROI Abstract 438.
  23. Tasker SA et al. Clinical impact of baseline genotypic resistance. 8th CROI Abstract 436.

Source: HIVand, Copyright 2001 by HIV and Hepatitis Treatment Advocates and HIV and All Rights Reserved.

Links to other websites are current at date of posting but not maintained.