Share this article on:

Genotype–phenotype discordance: the evolution in our understanding HIV-1 drug resistance

Zolopa, Andrew R

Editorial Comments

From Stanford University School of Medicine, Stanford, California, USA.

See also p. 955

Requests for reprints to: Dr A. R. Zolopa, Stanford University School of Medicine, Stanford, California, USA.

Received: 8 November 2002; revised: 5 December 2002; accepted: 8 January 2003.

Although HIV resistance testing is now considered standard of care in resource-rich countries, both genotype and phenotype tests have limitations [1]. In this issue of AIDS, Parkin and colleagues from Virologic Inc. illustrate an important limitation of current genotype interpretation algorithms for antiretroviral agents such as lopinavir/ritonavir (LPV/r; Kaletra), which require multiple mutations for clinically significant resistance [2]. The algorithms for such drugs are complex both in terms of the number of mutations that contribute to resistance and in the various patterns of mutations that result in a virus with clinically significant resistance. Attempting to put this complexity into a simple set of rules is an ongoing challenge. Phenotype tests, however, appear somewhat less ‘sensitive’ than genotyping to what could be seen as evolving resistance. Faced with the limitations of resistance tests, many clinicians order both tests in an attempt to make optimal treatment decisions for their patients. Not infrequently, however, clinicians who obtain both a phenotype and genotype test find the results to be in disagreement, or ‘discordant', that is, one test result is interpreted as drug sensitive while the other is interpreted as drug resistant. Understanding the reasons for these discordant results should enable the clinician to make better use of these tests and thereby make better treatment decisions.

Some of the discordance between genotype and phenotype is simply a reflection of our incomplete and still evolving understanding of drug resistance. This discordance appears to be a consequence of incorrect interpretation of the test results, that is to say it is not the results that are in disagreement it is the interpretation. After all, a simple count of mutations or a single phenotype fold change (FC) in the 50% inhibitory concentration value used as a barrier to separate ‘sensitive’ from ‘resistant’ virus is somewhat arbitrarily drawn. These arbitrary separations lead to misclassification and apparent discordances. Apparent discordant results are illustrated in Parkin et al. [2] by those viruses classified by the original LPV/r scoring system as genotypically ‘sensitive’ (i.e., less than six mutations) but phenotypically ‘resistant’ (i.e., FC > 10) [3]. Parkin et al. developed a new genotypic interpretation algorithm that better predicts LPV/r phenotypic susceptibility. By looking closely at the results in which there was genotypic–phenotypic discordance (sensitive by genotyping but reduced susceptibility by phenotyping), the authors were able to improve our understanding of the LPV/r genotypic resistance profile and eliminate a significant amount of the discordance.

There are several examples of what appears to be genotype–phenotype discordance of actual test results [4]. In these examples, the phenotype appears sensitive (i.e., not different from wild type) while the genotype reveals mutations generally associated with drug resistance. As demonstrated by Parkin et al. [2], mixtures of drug-resistant mutants and wild-type virus may not always be detected by a shift in phenotype FC. It will depend on the specific mutation, the drug and the relative proportion of mutant/wild-type virus in the specimen. A second example involves transitional mutations. In cohorts of recently tested patients, investigators at the Centers for Disease Control and Prevention have shown that resistant viruses can back-mutate to intermediate or transitional forms that, by themselves, do not effect the phenotype and yet quickly re-emerge as resistant mutations with appropriate drug exposure [5]. Finally, antagonistic interactions between mutations in effect cancel each other's impact on the phenotype and result in a phenotype that appears ‘sensitive'. These interactions have been demonstrated to occur within the nucleoside reverse transcriptase inhibitor class and between the nucleoside and non-nucleoside reverse transcriptase inhibitor classes [6,7].

Whatever the source of discordance, the question for the clinician is how best to incorporate this kind of information into a treatment plan. Many practitioners rely on the interpretation that accompanies the results of a genotype or phenotype test. Under this circumstance, having more information about the patient's resistance profile will not necessarily lead to better outcomes if the results are discordant and the clinician is left confused. The clinician must be an informed and critical consumer of this technology and the interpretation that accompanies the results. If the test results are discordant, the clinician must ask why the results appear to be at odds and then base a treatment plan on this more informed view.

Parkin and colleagues have furthered our understanding of the ever increasing complexity of genotypic resistance patterns to LPV/r and the potential for at least some degree of cross-resistance with amprenavir [2]. For clinicians trying to chose between the various ‘boosted’ protease inhibitor antiretroviral regimens, a phenotype test will likely be most helpful in summarizing this genotypic complexity into a result that can be used to optimize the regimen. However, an accompanying genotype analysis can help to inform the phenotype result by identifying mixtures, transitional mutations and antagonistic mutational interactions. Clinicians should consider consulting an expert in HIV drug resistance if the results of a resistance test(s) are confusing or discordant. Ultimately, the discordant results need to be correlated with virological response to specific treatment regimens before we can fully realize the benefits of these evolving technologies.

Back to Top | Article Outline


1.Zolopa AR, D'Aquila TR. HIV resistance testing in clinical practice. In AIDS Therapy, 2nd edn. Edited by Dolin R, Masur H, Saag MS. New York: Elsevier; 2002:347–358.
2.Parkin NT, Chappey C, Petropoulos CJ. Improving lopinavir genotype algorithm through phenotype correlations: novel mutation patterns and amprenavir cross-resistance. AIDS 2003, 17:955–961.
3.Kempf DJ, Isaacson JD, King MS, Brun SC, Xu Y, Real K, et al. Identification of genotypic changes in human immunodeficiency virus protease that correlate with reduced susceptibility to the protease inhibitor lopinavir among viral isolates from protease inhibitor-experienced patients. J Virol 2001, 75:7462–7469.
4.Shafer RS. Genotypic testing for human immunodeficiency virus type 1 drug resistance. Clin Microbiol Rev 2002, 15:247–277.
5.Garcia-Lerma JG, Nidtha S, Blumoff K, Weinstock H, Heneine W. Increased ability for selection of zidovudine resistance in a distinct class of wild-type HIV-1 from drug-naive persons. Proc Natl Acad Sci USA 2001, 98:13907–13912.
6.Larder BA, Kemp SD, Harrigan PR. Potential mechanism for sustained antiretroviral efficacy of AZT–3TC combination therapy. Science 1995, 269:696–699.
7.Shulman N, Zolopa AR, Passaro D, Shafer RW, Huang W, Katzenstein D, et al. Phenotypic hypersusceptibility to non-nucleoside reverse transcriptase inhibitors in treatment-experienced HIV-infected patients: impact on virological response to efavirenz-based therapy. AIDS 2001, 15:1125–1132.

HIV; protease; genotype; phenotype; resistance; lopinavir; amprenavir

© 2003 Lippincott Williams & Wilkins, Inc.