| |
Single HIV Genotypic Resistance Test Can Underestimate Degree of Drug Resistance
|
| |
| |
"HIV-1 Drug Resistance: Degree of Underestimation by a Cross-Sectional versus a Longitudinal Testing Approach"
The Journal of Infectious Diseases April 2005;191:1325-1330
P. Richard Harrigan,1,2 Brian Wynhoven,1 Zabrina L. Brumme,1,2 Chanson J. Brumme,1 Beheroze Sattha,1 Jennifer C. Major,1 Rafael de la Rosa,3 and Julio S. G. Montaner1,2
1British Columbia Centre for Excellence in HIV/AIDS, St. Paul's Hospital, and 2Department of Medicine, University of British Columbia, Vancouver, British Columbia, Canada; 3AIDS and Viral Hepatitis Study Group, Department of Internal Medicine, Virgen del Rocio University Hospital, Seville, Spain
"....cross-sectional genotypic analyses may seriously underestimate the prevalence of antiretroviral drug resistance in treatment-experienced populations...the degree of underestimation reported here should be taken as a minimum estimate...treatment history should be used in combination with resistance testing in guidance for selecting therapy options...."
This study found that- Use of only 1 genotype test underestimated resistance detected compared to looking at 3 or more genotypic resistance tests:
--The 10 most common resistance mutations present in the historical genotype but absent in the most recent genotype, according to the prevalence potentially underestimated when only the most recent genotype was considered, were, in the HIV-1 RT gene, M184V (33.1%), M41L (11.9%), T215Y (11.4%), D67N (10.5%), K70R (10.1%), K103N (9.4%), Y181C (7.4%), and L210W (7.1%), and, in the HIV-1 PR gene, L90M (6.1%) and M46I (5.7%).
--243 (14.0%) subjects harbored the L90M mutation at some point in their genotypic history, compared with only 151 (9.6%) subjects who had this mutation during the most recent genotypic test performed while they were prescribed antiretroviral drugs and 138 (8.0%) subjects who had this mutation in their most recent genotype regardless of therapy type
--Most notably, the M184V mutation, which is associated with resistance to lamivudine, was prevalent in 999 (57.6%) of the historical genotypes, 559 (35.6%) of the most recent genotypes tested while the subjects were prescribed antiretroviral drugs, and 425 (24.5%) of the most recent genotypes regardless of therapy type...The differences in inferred antiretroviral drug resistance found when historical and cross-sectional approaches are used appear to be largest for NRTIs
--The 10 most common resistance mutations present in the historical genotype but absent in the most recent genotype, according to the prevalence potentially underestimated when only the most recent genotype was considered, were, in the HIV-1 RT gene, M184V (33.1%), M41L (11.9%), T215Y (11.4%), D67N (10.5%), K70R (10.1%), K103N (9.4%), Y181C (7.4%), and L210W (7.1%), and, in the HIV-1 PR gene, L90M (6.1%) and M46I (5.7%).
-- mutations often described as "revertants" at RT codon 215 C/D/S (or combinations of these with other amino acid variants) were observed in 1.9%, 2.7%, and 7.0% of the historical genotypes, respectively, and in 0.5%, 1.8%, and 2.0% of the most recent genotypes. The RT mutation 74I was observed in 1.9% of the historical genotypes and in only 1.0% of the most recent genotypes.
ABSTRACT
Genotyping of human immunodeficiency virus type 1 (HIV-1) for antiretroviral drug resistance is routinely used both in clinical practice, to guide the selection of options for an individual's antiretroviral therapy, and in epidemiological studies, to estimate levels of antiretroviral drug resistance in a patient population. However, reliance on results of a single test can result in an underestimation of antiretroviral drug resistance. In the present study, we quantified the prevalence of resistance-associated mutations found in recent genotypic tests of 1734 HIV-1-infected, treatment-experienced subjects who had at least 3 genotypic tests (n = 11,404 genotypic tests total; median, 5 tests/subject) and compared it with that of resistance-associated mutations ever detected in these subjects between 1996 and 2004. Single-point analyses underestimated antiretroviral drug resistance, particularly for nucleoside analogues, in both individuals and patient populations. For example, the prevalence of resistance-associated mutation M184V/I was 25.5% in the most recent genotypes and 58.8% in available historical genotypes. Our results suggest that analysis of a combined historical genotype rather than of a cross-sectional genotype may lead to more accurate estimates of antiretroviral drug resistance in individual patients and in patient populations.
A major barrier to successful viral suppression in HIV type 1 (HIV-1)-infected individuals is the emergence of virus resistant to antiretroviral drugs [1-3]. Genotypic or phenotypic tests of HIV-1 for antiretroviral drug resistance can help predict responses to antiretroviral therapy and are now recommended for routine management of HIV-1 infections. However, because these tests detect only the most prominent HIV-1 genetic variant(s) circulating at the time, the number of resistance-associated mutations present may potentially be underestimated when results from a single antiretroviral drug resistance test are used in the determination. Interruptions or changes in antiretroviral therapy that alter the selective pressure of antiretroviral drug exposure may result in a reversion to predominantly wild-type virus or in the development of alternate resistance-associated mutations [4-7]. However, under these conditions, drug-resistant variants are known to persist as undetected minority species and may reemerge rapidly when the selective pressure of drug exposure resumes [4, 8].
Although a determination based on the results of a single antiretroviral drug resistance test may lead to an underestimation in the levels of resistance, current clinical practice relies heavily on the use of a single genotypic antiretroviral drug resistance test to direct changes in therapy and/or to screen individuals for participation in clinical trials. In addition, epidemiological studies designed to estimate the prevalence of resistance-associated mutations in treatment-experienced individuals may also underestimate the extent of resistance if only cross-sectional data are used. The magnitude of this underestimation is not known.
To quantify the extent to which antiretroviral drug resistance may potentially be underestimated on both an individual and a population basis when results from a single genotypic test are used in the determination, we compared the prevalences of major resistance-associated mutations in the most recent HIV-1 genotype with the prevalences of those in a "historical" genotype that combined all resistance-associated mutations ever observed in an individual's available genotypic history-a "virtual virus." In addition, we attempted to identify specific resistance-associated mutations whose prevalences are most likely to be underestimated when a cross-sectional approach is used.
AUTHOR DISCUSSION
Reliance on results of a genotypic test performed on a single sample of HIV-1 RNA may lead to an underestimation of the level of resistance to antiretroviral drugs in an individual as well as in a patient population, because these tests detect only the most prominent HIV-1 variant(s). Undetected drug-resistant minority species are known to persist when individuals discontinue or change antiretroviral therapies, and the rapid reappearance of these mutations on reinitiation of therapy has been extensively documented [4, 5, 8, 14]. For this reason, current guidelines for antiretroviral drug resistance testing suggest that genotypes should be used in combination with treatment history to guide the selection of therapy options [4, 7]. Our data show that there is a considerable difference between resistance-associated mutations found in current genotypes and those found in historical genotypes, confirming that background treatment information is vital for full characterization of an individual's HIV-1 antiretroviral drug resistance profile [4, 7]. More specifically, our results indicate that treatment history-related factors, including the number of genotypic tests performed, the number of antiretroviral drugs prescribed over the subject's history, the time elapsed between genotypic tests, as well as the proportion of time prescribed antiretroviral drugs, significantly influenced whether a discrepancy was observed when the mutational profiles of the most recent genotypes were compared with those of historical genotypes. From an epidemiological viewpoint, our results indicate that cross-sectional genotypic analyses may seriously underestimate the prevalence of antiretroviral drug resistance in treatment-experienced populations-an observation that is consistent with findings of previous studies [15].
The differences in inferred antiretroviral drug resistance found when historical and cross-sectional approaches are used affect all drug classes but appear to be largest for NRTIs. Most commonly, prevalence of the M184V mutation appears to be underestimated, and this results in the greatest impact on the perceived level of resistance to lamivudine. However, even when the impact of the M184V mutation is excluded, the observation that resistance to NRTIs as a class may be underestimated when a cross-sectional analysis is used is somewhat surprising, because nucleoside analogue components tend to be retained during most changes in antiretroviral therapy. Of interest, in some subjects, some genotypes appeared to harbor no resistance-associated mutations even when the subject was prescribed antiretroviral drugs. A history of having multiple wild-type genotypes, observed in `28% of the study population, likely represents lack of exposure of virus to selective pressure from antiretroviral drugs, because of factors including adherence to the drug regimen and pharmacokinetics.
It is important to note that our results may not be representative of resistance-associated mutation patterns in all HIV-1-infected individuals prescribed antiretroviral drugs, because the subjects in the present study generally represent individuals in whom therapy was failing and for whom multiple antiretroviral drug resistance tests were performed. The study population represents `29% of treated HIV-1-infected individuals in British Columbia during the study period; for the remaining 71% of such individuals, longitudinal genotypic data were not available. Furthermore, although plasma samples were collected at `3-month intervals, genotypic testing was not performed on all collected samples, so the degree of underestimation reported here should be taken as a minimum estimate. Despite these limitations, our data demonstrate that consideration of cross-sectional resistance data alone is likely to underestimate the effects of previous treatments-in some cases by a substantial amount. Our results strongly suggest that the use of combined historical genotypic data rather than cross-sectional genotypic data only may lead to more accurate estimates of antiretroviral drug resistance in treatment-experienced populations. In addition, consideration of a subject's history of HIV-1 mutation may also be of relevance in both clinical trials, for which antiretroviral drug resistance tests are part of the inclusion criteria, and in routine management of HIV-1 infections, in which genotypic data are used to guide the selection of therapy options. Historical genotypic data may add valuable information in the same way that longitudinal CD4 cell counts or viral load data may be more useful than results from a single test. As genotypic resistance testing becomes more widely available, it will become increasingly practical to adopt a comprehensive interpretation of antiretroviral drug resistance in both clinical and epidemiological studies.
RESULTS
Study population. The study population was composed of all HIV-1-infected individuals in British Columbia who had undergone >3 genotypic antiretroviral drug resistance tests during the study period (May 1996-July 2004) (n = 1734). The study group represents `29% of the 6014 HIV-1-infected British Columbians who were prescribed antiretroviral drugs at some point during the study period. Of the remaining 4280 individuals not included in the study group, 2458, 1215, and 607 had undergone 0, 1, or 2 genotypic tests during the study period, respectively.
A total of 11,404 genotypic tests were performed on 1734 subjects (median, 5 tests/subject; interquartile range [IQR], 4-8 tests/subject). The median time that elapsed between the earliest genotypic test and the most recent genotypic test was 34.1 months (IQR, 20.7-56.8 months). Subjects were prescribed a median of 5 different antiretroviral drugs (IQR, 3-8 antiretroviral drugs) between their earliest genotypic test and their most recent genotypic test. At the time of their first genotypic test, 47.6% of subjects were not prescribed antiretroviral drugs, 24.4% were prescribed monotherapy or dual therapy, and 28.0% were prescribed >3 antiretroviral drugs. At the time of the most recent genotypic test, 39.6% of subjects were not prescribed antiretroviral drugs, 5.7% were prescribed monotherapy or dual therapy, and 54.7% were prescribed >3 antiretroviral drugs. Of the 1734 subjects, 1569 (90.5%) changed antiretroviral therapy (defined as adding and/or dropping >1 drug from the regimen) during the study period; the median number of therapy changes per study subject was 4 (IQR, 2-7 therapy changes). The median pVLs were 4.6 log10 HIV-1 RNA copies/mL (IQR, 3.8-5.1 log10 HIV-1 RNA copies/mL) at the time of the earliest genotypic test and 4.3 log10 HIV-1 RNA copies/mL (IQR, 3.4-5.0 log10 HIV-1 RNA copies/mL) at the time of the most recent genotypic test.
Prevalences of resistance-associated mutations in most recent genotypes versus historical genotypes. We compared the prevalences of resistance-associated mutations observed in the most recent genotype regardless of therapy type (n = 1734), the prevalences of those observed during the most recent genotypic test performed while the subject was prescribed antiretroviral drugs (if a genotypic test was performed; n = 1571), and the prevalences of those in the historical genotype (n = 1734). Mutations associated with resistance to PI were somewhat more common in the historical genotypes than in the most recent genotypes; for example, 243 (14.0%) subjects harbored the L90M mutation at some point in their genotypic history, compared with only 151 (9.6%) subjects who had this mutation during the most recent genotypic test performed while they were prescribed antiretroviral drugs and 138 (8.0%) subjects who had this mutation in their most recent genotype regardless of therapy type. Figure 1B and 1C shows similar comparisons for mutations associated with resistance to NRTIs and NNRTIs; again, there was a substantial difference between the most recent genotypes and the historical genotypes in the prevalences of mutations at some codons. Most notably, the M184V mutation, which is associated with resistance to lamivudine, was prevalent in 999 (57.6%) of the historical genotypes, 559 (35.6%) of the most recent genotypes tested while the subjects were prescribed antiretroviral drugs, and 425 (24.5%) of the most recent genotypes regardless of therapy type. The 10 most common resistance mutations present in the historical genotype but absent in the most recent genotype, according to the prevalence potentially underestimated when only the most recent genotype was considered, were, in the HIV-1 RT gene, M184V (33.1%), M41L (11.9%), T215Y (11.4%), D67N (10.5%), K70R (10.1%), K103N (9.4%), Y181C (7.4%), and L210W (7.1%), and, in the HIV-1 PR gene, L90M (6.1%) and M46I (5.7%).
If the prevalences of resistance-associated mutations in multiple resistance categories were estimated on the basis of the results of the most recent genotypic test only, 4.5% of the study group would harbor resistance-associated mutations to all 4 drug categories (lamivudine, other NRTIs, NNRTIs, and PIs), compared with 12.2% of the study group if all available historical genotypes were also considered (table 1). Similarly, results from the most recent genotypic test indicated that 53.7% of genotypes had no detectable resistance-associated mutations, compared with 28.3% of genotypes if historical data were included (table 1). When analyzed by drug category, resistance to lamivudine (conferred by the M184V and/or M184I mutations) was most likely to be observed in a subject's genotypic history but not in the most recent genotypic test (a difference in prevalence of 33.3%; table 1).
Although they are not considered to be "major" resistance-associated mutations, mutations at certain other codons not indicated in figure 1 are also of interest. In particular, the mutations often described as "revertants" at RT codon 215 C/D/S (or combinations of these with other amino acid variants) were observed in 1.9%, 2.7%, and 7.0% of the historical genotypes, respectively, and in 0.5%, 1.8%, and 2.0% of the most recent genotypes. The RT mutation 74I was observed in 1.9% of the historical genotypes and in only 1.0% of the most recent genotypes.
Identification of factors associated with differences between the most recent genotypes and the historical genotypes. A total of 808 (46.6%) subjects displayed no change in their mutational profiles when their most recent genotypes and their historical genotypes were compared. (Of interest, the majority of this group [490/808; 60.6%] was composed of subjects who consistently displayed no major resistance-associated mutations in any genotypic test performed during the study period.) The remainder of the study subjects (926/1734; 53.4%) displayed at least 1 major mutational difference when their most recent genotypes and their historical genotypes were compared (median, 2 mutational differences; range, 1-13 mutational differences).
To identify factors that influence the magnitude of discrepancy between the most recent genotype and the historical genotype, we compared clinical and treatment history variables in these 2 groups. We observed a significant difference with respect to the median number of genotypic tests performed during the study period (4 tests [IQR, 3-6 tests] in subjects with no differences in their mutational profile when their most recent genotypes and their historical genotypes were compared vs. 6 tests [IQR, 4-10 tests] in subjects with at least 1 mutational difference; OR, 1.2; 95% CI, 1.2-1.3 per additional test performed; P < .0001). We observed a significant difference in the median number of antiretroviral drugs prescribed (4 drugs [IQR, 3-5 drugs] vs. 7 drugs [IQR, 5-9 drugs]; OR, 1.5; 95% CI, 1.4-1.5 per additional drug prescribed; P < .0001). The median time that elapsed between the earliest genotypic test and the most recent genotypic test was also significantly different (27 months [IQR, 15-42 months] vs. 46 months [IQR, 28-68 months]; OR, 1.5; 95% CI, 1.4-1.6 per additional 12-month period; P < .0001). Finally, the proportion of time that antiretroviral drugs were prescribed during the study period was also significantly different (61% of the study period [IQR, 25%-88% of the study period] vs. 78% of the study period [IQR, 55%-93% of the study period]; OR, 1.2; 95% CI, 1.4-1.6 per additional 10% increment; P < .0001).
We observed no significant difference in the 2 groups with respect to pVL at most recent genotypic test (P = .5; data not shown) or with respect to the time elapsed between discontinuation of antiretroviral therapy and the most recent genotypic test (median, 0 months; P = .6). Because the parameters we investigated were interrelated, multivariate analyses were not performed.
PATIENTS AND METHODS
Study population. The British Columbia Centre for Excellence in HIV/AIDS (BC CfE) in Vancouver, British Columbia, Canada, provides antiretroviral drugs and treatment for HIV free of charge to all HIV-infected residents of British Columbia via the British Columbia HIV/AIDS Drug Treatment Program. Briefly, the treatment guidelines of this program recommend that testing for plasma viral load (pVL) and CD4 cell counts be performed at baseline, at 4 weeks after initiation of or alteration in antiretroviral therapy, and at `3-month intervals thereafter [9]. Genotypic antiretroviral drug resistance testing is available at the discretion of the treating physician and may be performed retrospectively on plasma samples stored since 1996. Patient demographic information and clinical, virological, and immunological data are maintained in a centralized database. Ethical approval for this study was obtained from the institutional ethics board of Providence Health Care/University of British Columbia.
In this retrospective study, we examined the BC CfE database for all HIV-1-infected individuals who had undergone >3 genotypic antiretroviral drug resistance tests between May 1996 and July 2004 (n = 1734 subjects; 11,404 genotypic antiretroviral drug resistance tests in total). In 1050 (60.6%) cases, subjects were prescribed antiretroviral drugs at the time of the most recent genotypic test.
REFERENCES
1. DeGruttola V, Dix L, D'Aquila R, et al. The relation between baseline HIV drug resistance and response to antiretroviral therapy: re-analysis of retrospective and prospective studies using a standardized data analysis plan. Antivir Ther 2000; 5:41-8. First citation in article | PubMed
2. Durant J, Clevenbergh P, Halfon P, et al. Drug-resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. Lancet 1999; 353:2195-9. First citation in article | PubMed | CrossRef
3. Baxter JD, Mayers DL, Wentworth DN, et al. A randomized study of antiretroviral management based on plasma genotypic antiretroviral resistance testing in patients failing therapy. AIDS 2000; 14:F83-93. First citation in article | PubMed | CrossRef
4. Hirsch MS, Brun-Vezinet F, Clotet B, et al. Antiretroviral drug resistance testing in adults infected with human immunodeficiency virus type 1: 2003 recommendations of an International AIDS Society-USA Panel. Clin Infect Dis 2003; 37:113-28. First citation in article | Full Text | PubMed
5. Hance AJ, Lemiale V, Izopet J, et al. Changes in human immunodeficiency virus type 1 populations after treatment interruption in patients failing antiretroviral therapy. J Virol 2001; 75:6410-7. First citation in article | PubMed | CrossRef
6. Devereux HL, Youle M, Johnson MA, Loveday C. Rapid decline in detectability of HIV-1 drug resistance mutations after stopping therapy. AIDS 1999; 13:F123-7. First citation in article | PubMed | CrossRef
7. Verhofstede C, Wanzeele FV, Van Der Gucht B, De Cabooter N, Plum J. Interruption of reverse transcriptase inhibitors or a switch from reverse transcriptase to protease inhibitors resulted in a fast reappearance of virus strains with a reverse transcriptase inhibitor-sensitive genotype. AIDS 1999; 13:2541-6. First citation in article | PubMed | CrossRef
8. Izopet J, Souyris C, Hance A, et al. Evolution of human immunodeficiency virus type 1 populations after resumption of therapy following treatment interruption and shift in resistance genotype. J Infect Dis 2002; 185:1506-10. First citation in article | Full Text | PubMed
9. Hogg RS, Yip B, Chan KJ, et al. Rates of disease progression by baseline CD4 cell count and viral load after initiating triple-drug therapy. JAMA 2001; 286:2568-77. First citation in article | PubMed | CrossRef
10. Alexander CS, Dong W, Chan K, et al. HIV protease and reverse transcriptase variation and therapy outcome in antiretroviral-nave individuals from a large North American cohort. AIDS 2001; 15:601-7. First citation in article | PubMed | CrossRef
11. Galli R, Wynhoven B, Harrigan R. Quality management of sequence-based methods for measuring HIV-1 drug resistance. Am Clin Lab 2002; 21:13-5. First citation in article | PubMed
12. Johnson VA, Brun-Vezinet F, Clotet B, et al. International AIDS Society-USA Drug Resistance Mutations Group. Drug resistance mutations in HIV-1. Top HIV Med 2003; 11:215-21. First citation in article | PubMed
13. Hogg RS, Heath K, Bangsberg D, et al. Intermittent use of triple-combination therapy is predictive of mortality at baseline and after 1 year of follow-up. AIDS 2002; 16:1051-8. First citation in article | PubMed | CrossRef
14. Devereux HL, Loveday C, Youle M, Sabin CA, Burke A, Johnson MA. Reduction in human immunodeficiency virus type 1 mutations associated with drug resistance after initiating new therapeutic regimens in pretreated patients. J Infect Dis 2000; 181:1804-7. First citation in article | Full Text | PubMed
15. Pillay D, Green H, Gazzard B, et al. Estimating resistance in drug experienced patients in the UK. Antivir Ther 2004; 9:S88. First citation in article
|
|
| |
| |
|
|
|
