Evolving information about resistance and resistance testing that might affect clinical practice would include newly identified mutations and clinically relevant phenotypic susceptibility cut-offs for approved drugs, prospective virologic outcome data involving resistance testing, and comparative data pertaining to different assays and test interpretations systems. Although there was not a lot of such information presented at the 9th Conference on Retroviruses and Opportunistic Infections, there were a few tidbits.

Of course, limitations in our ability to treat drug-resistant HIV infection is likely due more to innate pharmacologic shortcomings and cross-resistance patterns of available drugs than to a clinician's inability to identify and understand resistance using any particular assay or interpretation technology. We therefore introduce this report on HIV resistance test information with an update on NRTI cross resistance.

NRTI cross-resistance
NRTI cross-resistance (and our growing appreciation of this phenomenon) continues to be among the most daunting limitations of our ability to construct fully suppressive antiretroviral (ARV) regimens for patients with drug-resistant HIV. An analysis of the relationship between NRTI genotypes and phenotypic susceptibility from 4,174 samples showed a linear decrease in susceptibility across all NRTIs as the number of nucleoside analog mutations (NAMs) increased from 0 to 6 [1]. This presentation also clearly demonstrated the effect of the M184V mutation on susceptibility across the range of NAMs. The M184V mutation causes further reductions in susceptibility (when compared to genotypes without M184V) at each NAM level for 3TC, ddI, ddC and ABC, while it causes an increase in susceptibility to AZT, tenofovir (TDF) and to a lesser extent d4T.

While TDF therapy does not select for resistance mutations with significant cross-resistance other than that at codon 65, TDF susceptibility is dramatically reduced in the presence of 3 or more thymidine analog mutations (TAMs) when these include mutations M41L and/or L210W [2]. Thus, while TDF offers an additional NRTI option for some highly NRTI-experienced patients, many of these patients are unlikely to experience more than a modest virologic benefit.

Assay comparisons
Although only one kit-based assay is currently cleared by FDA for marketing (TRUGENE^TM, Visible Genetics), application for Applied Biosystems' HIV Genotyping System (ViroSeq^TM) has been submitted to the FDA and is likely to receive clearance soon. A comparison of genotypes for 34 non-subtype B HIV-1 isolates (subtypes A-H) run with both kits showed only minor differences in sequence results [3]. However, the TRUGENE system was unable to genotype 2 of 34 isolates (1 subtype C and 1 subtype G sample), suggesting there may be some limitations in the ability to amplify the RT and protease gene regions in some non-B subtypes. The authors point out, though, that clinical specimens have failed in both assays with a success rate of approximately 90% for both assays.

There are also several published comparisons addressing the relative performance of these two kits in B or non-B subtypes, which describe varying degrees of sequence concordance, and ability to detect insertion mutations, between the two kits [4-6]. A study by Erali et al also addresses the relative technical issues and training required to perform each assay. Future studies should address the relative capital and per-assay costs to establish and run these two systems in order to assist laboratories in deciding which system to implement as they evaluate performance characteristics and data. The ease of use of the report, and the reliability of the interpretation tools associated with each kit, will also be critical factors in deciding which system and kit to use.

Interpretation tools
Rules-based algorithms derived from the literature are the major tool used to interpret genotypic resistance test results. Each kit-based assay manufacturer and proprietary non-kit based laboratory assay uses its own algorithm, when an interpretation is provided. (Please see "Interpreting Genotypic Resistance Tests" by Robert Shafer and Jonathan Schapiro for a thorough review of interpretation systems and comparative approaches to identifying the best rules for the interpretation of genotypic resistance test results.)

The Virtual Phenotype^TM (Tibotec-Virco) uses a database of related genotypes and phenotypes to predict the phenotypic susceptibility of an isolate based on a given genotype. The first prospective, controlled, virologic outcomes studies addressing the clinical utility of the Virtual Phenotype^TM compared to a real phenotype was presented at the 9th CROI [7]. In this randomized study, changes in ARV were informed by either an actual phenotype or Virtual Phenotype^TM for 201 highly ARV-experienced patients. Endpoints included the percent of patients with HIV-1 RNA <400 and the percent with an HIV-1 RNA reduction > 0.5 log₁₀ copies/mL at weeks 4, 16, 32 and 48. The groups were well matched at baseline for viral load, CD4 cell count, and ARV history. At 48 weeks, only approximately 15% of subjects in either arm had achieved an HIV-1 RNA reduction to < 400 copies/mL in intent-to-treat analysis. There was no significant difference in mean reductions in viral load or the proportion of subjects with > 0.5 log₁₀ copies/mL reduction (approximately 50%). There was an intriguing difference in CD4 cell count increase in favor of the Virtual Phenotype^TM arm (41.6 vs 94.4 cells; P = 0.16).

The major limitation of this study is that it compared the Virtual Phenotype^TM to the real phenotype, rather than comparing the Virtual Phenotype^TM to one or more of the common rules-based algorithm approaches to interpreting a genotype (or to no test). Previous prospective studies of phenotypic assays have not consistently shown that patients with access to these assays had better virologic responses than patients treated without a resistance test [8-13]. Thus it is still not clear if the Virtual Phenotype^TM provides information equivalent to a genotype interpreted with a rules-based algorithm when applied to selection of new ARV therapy.

Prospective virologic outcomes studies
There were no new prospective clinical utility studies presented, although one group presented a meta-analysis of six published or reported prospective studies (4 genotype vs SOC, 1 phenotype vs SOC, and 1 genotype vs phenotype vs SOC) [13]. At 3 months, 42.6% with a genotype versus 33.2% without had undetectable HIV-1 RNA (OR 1.7; 95% CI, 1.3-2.2). At 6 months (in a sub-set of the studies), 38.8% versus 28.7% had undetectable HIV-1 RNA (OR 1.6; 95% CI, 1.2-2.2). However, there was no difference in viral load suppression between those with and without a phenotype, 37.5% versus 33.8% (OR 1.1; 95% CI 0.8-1.6).

Phenotypic cut-offs and graded response
No new phenotypic cut-offs were presented, although tenofovir data suggesting graded reductions in susceptibility between 1.8- and 3.8 fold (for the Tibotec-Virco assay) were discussed [2].

One of the interesting challenges to phenotypic resistance test providers is to define the levels at which there is minimal, moderate and total reduction in susceptibility, so a risk-benefit assessment regarding the use of a certain drug can be made if partial susceptibility is reported. Two presentations began to address this issue. In one, regression analyses of IC50 values for baseline samples from 142 patients starting a salvage regimen in a clinical trial were used to investigate the relationship between IC50 and virologic response at 16 weeks [14]. Although individual dichotomous scores of resistant and sensitive combined into a phenotypic sensitivity score were not associated with week 16 virologic response, continuous combined phenotypic sensitivity scores that were also weighted based on expected relative drug potency were associated with week 16 virologic response (P = 0.05). Although the concept or using a continuous scale to assess susceptibility appears promising from studies such as this, determining scales for individual drugs that can be used in clinical decision making, rather than retrospective scores of combinations of susceptibilities, remains a challenge.

It would be a great service if clinicians could request an updated report utilizing the most current clinically relevant cut-offs when reviewing older phenotypic test results in order to make treatment decisions.

A study of phenotypic indinavir susceptibility testing in subjects receiving indinavir and ritonavir was undertaken to determine the degree of reduced indinavir susceptibility that could be overcome with ritonavir boosting [15]. Virologic activity was seen in patients on a variety of combination regimens when the indinavir IC50 fold change was >2.5 but <25 if they also received an NNRTI. The authors concluded that the best responses were seen in those with an indinavir IC50 change of <2.5-fold, but that there was no clear cut-off for defining partial versus no response. It is unclear if the virologic responses in the patients with indinavir IC50 changes of >2.5-fold were entirely due to the NNRTI component, and this study offers only a small step in the progress towards defining a continuum of phenotypic susceptibility for indinavir.

In order to maximize the utility of currently available phenotypic resistance tests and the evolving science of developing clinically relevant cut-offs, it would be a great service if clinicians could request an updated report utilizing the most current clinically relevant cut-offs when reviewing older phenotypic test results in order to make treatment decisions.

Conclusions
One of the most interesting discussions I had at this conference occurred at a table in the back of the poster sessions where a clinical research colleague and I "came out" to each other about our cynicism concerning resistance testing, and the isolation we sometimes feel in our clinical environments regarding this perspective.

Unlike my colleague, I do use genotypic resistance tests frequently, but I do so with great caution. First, I document what mutations I expect to see, then which ones I am pretty sure I know are there even if I don't see them (due to remote drug selection pressure or insensitivity of the test). Next, I record the questions I hope the test can answer and how certain I will be of those answers (e.g., how many and which TAMs or PI mutations, and how confident will I be of the results based on treatment history). Finally, I record and discuss with my patients the drugs I think I will probably recommend, and modify those recommendations based on patient preference and other medical and psychosocial factors.

In my own practice, I find the tests often confirm what I suspected, and the drugs I choose seldom vary far from what I predicted. Of course, there are surprises, and these most often make me wonder about a patient's report regarding adherence. The test often answers my question(s), but cannot provide reassurance when I knew it could not answer with much certainty (e.g., I only see one TAM, but the patient has not been on a thymidine analog for many months or years so I am not confident about archived resistance mutations).

Perhaps the resistance patterns I have learned in order to interpret these tests have helped me learn how to predict resistance fairly well? Yet I find comfort in this confirmation, and several studies have suggested that the tests add to clinical factors and predictions [16,17]. But did I really need the test? My colleague would respond that I do not, and prospective virologic outcomes data available to date (which are very short term) currently support both our views and practices nearly equally well.

References

J. M. Whitcomb, E. Paxinos, W. Huang, M. Maranta, K. Limoli, C. Chappey, N. T. Parkin, N. S. Hellmann, C. J. Petropoulos. The Presence of Nucleoside Analogue Mutations (NAMs) Highly Correlated with Reduced Susceptibility to all NRTIs. 9th Conference on Retroviruses and Opportunistic Infections. 24-28-4 Feb 2002, Seattle, WA. Abstract 569.
M. D. Miller, N. A. Margot, and B. Lu. Effect of Baseline Nucleoside-Associated Resistance on Response to Tenofovir DF (TDF) Therapy: Integrated Analyses of Studies 902 and 907. 9th Conference on Retroviruses and Opportunistic Infections. 24-28-4 Feb 2002, Seattle, WA. Abstract 43.
L. Jagodzinski, J. Cooley, S. Kelly, N. Michael.Performance of the TRUGENE HIV-1 Genotyping Kit and the Applied Biosystems HIV-1 Genotyping System in Sequence-Based Analysis of Non-B HIV-1 Subtypes. 9th Conference on Retroviruses and Opportunistic Infections. 24-28-4 Feb 2002, Seattle, WA. Abstract 594.
Fontaine E, Riva C, Peeters M, et al. Evaluation of two commercial kits for the detection of genotypic drug resistance on a panel of HIV type 1 subtypes A through J. J Acquir Immune Defic Syndr. 2001; 28:254-8.
Erali M, Page S, Reimer LG, Hillyard DR. Human immunodeficiency virus type 1 drug resistance testing: a comparison of three sequence-based methods. J Clin Microbiol. 2001; 39:2157-65.
Koch N, Tamalet C, Tivoli N, Fantini J, Yahi N. Comparison of two commercial assays for the detection of insertion mutations of HIV-1 reverse transcriptase. J Clin Virol. 2001; 21:153-62.
M. J. Perez-Elias, I. Garcia-Arata, V. Muñoz, I. Santos, J. Sanz, V. Abraira, A. Moreno, J. R. Arribas, J. González, A. Antela, F. Dronda, M. Pumares, P. Martí-Belda, S. Moreno for the Realvirfen Study Group. A Randomized, Prospective Study of Phenotype (P) versus Virtual Phenotype (VirtualP) Testing for Patients Failing Antiretroviral Therapy (ART). 9th Conference on Retroviruses and Opportunistic Infections. 24-28-4 Feb 2002, Seattle, WA. Abstract 586.
Meynard JL, Vray M, Morand-Joubert L, Race E, Descamps D, Peytavin G, Matheron S, Lamotte C, Guiramand S, Costagliola D, Brun-Vezinet F, Clavel F, Girard PM. Phenotypic or genotypic resistance testing for choosing antiretroviral therapy after treatment failure: a randomized trial. AIDS. 2002;16(5):727-36.
Baxter JD, Mayers DL, Wentworth DN, Neaton JD, Hoover ML, Winters MA, Mannheimer SB, Thompson MA, Abrams DI, Brizz BJ, Ioannidis JP, Merigan TC. A randomized study of antiretroviral management based on plasma genotypic antiretroviral resistance testing in patients failing therapy. CPCRA 046 Study Team for the Terry Beirn Community Programs for Clinical Research on AIDS. AIDS. 2000;16:F83-93.
Clevenbergh P, Durant J, Halfon P, del Giudice P, Mondain V, Montagne N, Schapiro JM, Boucher CA, Dellamonica P. Persisting long-term benefit of genotype-guided treatment for HIV-infected patients failing HAART. The Viradapt Study: week 48 follow-up. Antivir Ther. 2000;5:65-70.
Durant J, Clevenbergh P, Halfon P, Delgiudice P, Porsin S, Simonet P, Montagne N, Boucher CA, Schapiro JM, Dellamonica P. Drug-resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. Lancet. 1999;353:2195-9.
Haubrich R, Keiser P, Kemper C, Witt M, Leedom J, Forthold D, Hellmann N. CCTG 575: A randomized, prospective study of phenotype testing (Pheno) verus standard of care (SOC) for patients failing antiretroviral therapy. The 1st. IAS Conference on HIV Pathogenesis and Treatment, Buenos Aires, Argentina, 8-11 July 2001. Abstract 127.
D. Torre, R. Tambini. Meta-Analysis of Antiretroviral Drug Resistance Testing in HIV-1 Infection. 9th Conference on Retroviruses and Opportunistic Infections. 24-28-4 Feb 2002, Seattle, WA. Abstract 584.
R. Swanstrom, H. Cheng, N. S. Hellmann, D. Katzenstein, R. J. Bosch, S. A. Fiscus, R. Haubrich, R. Gulick. Alternative Phenotypic Sensitivity Scores and Outcome Measures Increase Detection of the Impact of Drug Resistance on Response to Salvage Therapy -- Results from ACTG 359. 9th Conference on Retroviruses and Opportunistic Infections. 24-28-4 Feb 2002, Seattle, WA. Abstract 592.
H. Rice, A. Zolopa, M. Coram, U. Murlidharan, N. Shulman, C. Vaamonde, J. H. Condra, N. S. Hellmann, H. King4, and M. Bates. Correlation of Phenotypic Resistance and Virologic Response to Indinavir/Ritonavir Boosted Regimens. 9th Conference on Retroviruses and Opportunistic Infections. 24-28-4 Feb 2002, Seattle, WA. Abstract 558.
Zolopa AR, Shafer RW, Warford A, et al. HIV-1 genotypic resistance patterns predict response to saquinavir- ritonavir therapy in patients in whom previous protease inhibitor therapy had failed. Ann Intern Med. 1999; 131:813-21.
Saag MS. HIV resistance testing in clinical practice: a Qaly-fied success. Ann Intern Med. 2001; 134:475-7.