| |
Brief Report: Limited Evolution of HIV Antiretroviral Drug
Resistance-Associated Mutations During the Performance of
Drug Resistance Testing
Sean Emery
|
| |
| |
Summary:
We investigated the evolution of HIV reverse transcriptase (RT)- and protease-associated antiretroviral (ARV) drug resistance mutations during the time taken to perform genotypic drug resistance testing. Thirty treatment-experienced patients who were adherent to therapy and who underwent genotypic drug resistance testing provided blood samples at randomization and when reviewing the test results (baseline). Patients remained on their existing therapy between randomization and baseline. The predominant HIV strains in 10 patients (33%) either lost and/or gained primary RT inhibitor (RTI) or protease inhibitor (PI) associated resistance mutations during the testing period. Of the 9 patients with RT mutations, 2 lost, 5 gained, and 2 both lost and gained RTI resistance mutations. One patient gained a
significant PI-associated resistance mutation on an existing PI-resistant background. The evolution that occurred in the RT may have altered the effectiveness of subsequent ARV therapy in some patients (5 of 7 patients). Neither viral load at randomization, ARV drug class used at randomization, time between collection of blood samples, duration of current therapy, nor number of ARV drugs used influenced gain or loss of resistance mutations. There was a significant association between duration of previous ARV therapy and gain of RTI-associated resistance mutations (p = .02), however. In general, our results suggest that patients should continue current therapy until test results are available. A few patients would be expected to gain ARV drug-associated resistance mutations during this time, however.
...in the majority of patients either did not change between randomization and baseline or changed only in sequences not thought to be associated with drug resistance. Nevertheless, of the 10 patients identified who had changes at codons in the RT or protease associated with ARV drug resistance, 7 (23%) acquired mutations that could generate or increase preexisting resistance to RTIs and 1 (3%) gained an additional mutation associated with genotypic PI resistance. The generation of thymidine-associated mutations in the remaining patients (5) may have contributed to reduced susceptibility to several NRTIs, however, including abacavir and stavudine... evolution of virus to become resistant or more resistant (through the accumulation of mutations) during the testing period occurred in one quarter of the patients we investigated. This is a much higher incidence than that reported in a similar study, where only 2 of 158 patients (1.2%) acquired mutations likely to confer resistance to RTIs or
PIs. Investigations involving larger numbers of patients may resolve this..... it is likely that a minority of patients will gain resistance mutations during this time and that these may have an impact on subsequent therapy
JAIDS Journal of Acquired Immune Deficiency Syndromes 2003; 32(1):57-61
*Chris Birch; *Tracey Middleton; Gillian Hales; David Cooper; Matthew Law; Suzanne Crowe; Jennifer Hoy; Sean Emery.
*Victorian Infectious Diseases Reference Laboratory, North Melbourne; National Centre for HIV Epidemiology and Clinical Research, University
of New South Wales, Sydney; Burnet Centre, Melbourne; and Alfred Hospital and Department of Medicine, Monash University, Melbourne,
Australia
Antiretroviral (ARV) drug resistance testing has become the standard of care in many developed countries because of the potential for the test results to have an impact on subsequent virologic responses to new therapies. Several studies examining the utility of resistance testing have shown that a sustained reduction in HIV viral load can be achieved by modifying therapy in response to the results. Although it has been recommended that blood used for resistance testing be drawn before the current drug regimen is stopped, no guidelines are available to suggest whether treatment should be continued or discontinued during the time testing is being done.
Using conventional assays, many resistance mutations rapidly become undetectable after ceasing therapy. Minor populations expressing important resistance mutations can be detected if sensitive assays are used, however, and phenotypic zidovudine resistance persists in some patients who ceased zidovudine therapy at least 3 months previously. There are few if any studies examining whether resistance is generated or continuing evolution of drug
resistance mutations occurs in HIV strains of patients who remain on existing therapy while resistance testing is performed. This is an important issue, because continuation of current therapy, although possibly resulting in the
generation of virus with increasing resistance or cross-resistance, may also be associated with the preservation of virus with reduced replicative fitness. Conversely, cessation of therapy immediately after blood is taken for testing may select for viruses of lesser resistance but greater replicative fitness.
The CREST study (Can Resistance Testing Enhance Subsequent Therapy) is an ongoing investigation comparing the utility of genotyping versus virtual phenotyping in individuals failing ARV therapy. We report here the results of a CREST substudy designed to address the issue of whether evolution of drug resistance mutations occurs in the reverse transcriptase (RT) and protease genes during the time taken for drug resistance test results to become available.
A rules-based algorithm was used by all participating laboratories and clinicians to interpret amino acid sequences for each of the drug classes (nucleoside RT inhibitors [NRTIs], non-NRTIs [NNRTIs], and protease inhibitors [PIs]). The algorithm is available at http://www. nrl.gov.au. The identification of mutations capable of conferring ARV drug resistance was based on published data.
Genotyping of the predominant HIV species in plasma samples was performed using in-house methods (12), the TruGene HIV-1 Assay (Visible Genetics, Toronto, Ontario, Canada), or the Viroseq HIV Genotyping Kit (Applied
Biosystems, Foster City, CA, U.S.A.). A written genotyping report was
provided to the requesting doctor within 28 days of collection of blood in all cases.
RESULTS
The duration of prior therapy was on average 4-8 years for the patients. The duration of current therapy was on average 1-2 years. The number of drugs previously used ranged from about 3-12. Thirty patients (28 male and 2 female) enrolled in the substudy, and each participant provided samples that gave evaluable sequences at both randomization and baseline. The mean age of the group was 45.9 years. All were receiving at least one NRTI; 2 received NRTI therapy only, 18 received PI-based therapy, 8 received NNRTI-based therapy, and 2 received treatment with drugs from each class (NRTI, NNRTI, and PI). Randomization HIV resistance genotype results were provided to the requesting doctor within 28 days for all 30 patients. The mean number of days between collection of the randomization and baseline blood samples was 37.3 days (range: 22-135 days).
Nine of the 30 patients (30%) either lost and/or gained mutations considered to be associated with genotypic resistance to RTIs during the testing period (Table 2). Each had evidence of preexisting genotypic RT resistance at that time. Two of these patients lost RTI resistance mutations (Patients 24 and 28), 5 gained RTI mutations (Patients 5, 6, 14, 16, and 18), and 2 both lost and gained RTI mutations (Patients 2 and 7). HIV from 21 patients had either no RT sequence changes or changes not considered to contribute to RTI resistance.
One patient had changes in protease sequence associated with genotypic resistance during the testing period. This involved the selection in Patient 3 of M46I on a background already containing the L90M mutation associated with saquinavir and nelfinavir resistance (results not shown). A total of 29 patients had either no protease sequence changes or changes (polymorphisms) not considered to contribute to resistance to PIs.
The 5 patients whose predominant HIV strains gained RT mutations during testing had a mean viral load at randomization of 25,118 copies/mL (range: 1100 to >100,000 copies/mL) compared with a mean viral load for the 2
patients who lost significant RT mutations of 8912 copies/mL (range: 3162-25,118 copies/mL). The mean viral load for patients who both lost and gained mutations was 10,000 copies/mL. Patients whose predominant strains had no codon changes detected (n = 11) had a mean viral load of 24,547 copies/mL (range: 1400 to >100,000 copies/mL). The differences between these (log10-transformed) viral load values were not statistically significant using the nonparametric rank sum test (p = .667, p = .451, and p = .914 for mutations gained, lost, and lost and gained, respectively).
The mean duration of all previous therapy in the 5 patients who gained resistance mutations during the testing period was 8.7 years compared with 5.0 years for those who lost mutations, 5.5 years for those who both lost and gained mutations, and 5.0 years for those having no change in sequence. This apparent association between duration of all previous therapy and the likelihood that resistance mutations were gained was statistically significant (p = .02). Similar analysis of time between collection of randomization and baseline blood samples, duration of current ARV drug therapy, ARV drug class at randomization, and number of ARV drugs ever used was unable to detect any association with gain, loss, or both loss and gain of RTI-associated mutations detected during the testing period (p > .06 in all cases).
Discussion by authors
Our results suggest that most patients failing therapy with resistant virus at baseline do not subsequently accumulate or select for mutations likely to generate or increase existing ARV drug resistance if they remain on that therapy while drug resistance testing is performed. In the cohort studied, the RT and protease sequences of the predominant HIV strain in the majority of patients either did not change between randomization and baseline or changed only in sequences not thought to be associated with drug resistance. Nevertheless, of the 10 patients identified who had changes at codons in the RT or protease associated with ARV drug resistance, 7 (23%) acquired mutations that could generate or increase preexisting resistance to RTIs and 1 (3%) gained an additional mutation associated with genotypic PI resistance. The generation of additional RTI-associated mutations in 2 of the 7 patients in whom this occurred (Patients 5 and 18) is unlikely to have influenced subsequent choice of drug therapy according to the genotype interpretive algorithm used for the study. The generation of thymidine-associated mutations in the remaining patients may have contributed to reduced susceptibility to several NRTIs, however, including abacavir and stavudine.
The HIV strains from 4 patients (Patients 2, 7, 24, and 28) lost mutations that have been associated with resistance to RTIs. In the absence of noncompliance, the most likely reason for this was the disappearance of mutations selected as a result of a previous therapy, perhaps associated with a reduction in replicative fitness. This coincidental disappearance of
mutations is likely to have occurred in 2 of these patients, both of whom lost mutations not directly associated with the therapy they were receiving at randomization (K103N in Patient 2, who was not receiving NNRTI therapy, and E44A in Patient 7, who was not receiving treatment with lamivudine, although this mutation has recently been associated with resistance to other NRTIs). The loss of K65R in Patient 28, who had no history of treatment with tenofovir, didanosine, zalcitabine, or abacavir, may have been associated with the rare and transient appearance of a mutation normally expected to interfere with enzyme binding by the natural deoxynucleotide. In patients with resistance to PIs at baseline, any further mutations lost or gained during the testing period were not associated with known resistance mutations, with the exception of the selection of M46I in Patient 3, which may have predisposed to cross-resistance to all PIs.
Overall, evolution of virus to become resistant or more resistant (through the accumulation of mutations) during the testing period occurred in one quarter of the patients we investigated. This is a much higher incidence than that reported in a similar study, where only 2 of 158 patients (1.2%) acquired mutations likely to confer resistance to RTIs or PIs. Investigations involving larger numbers of patients may resolve this. In the current study, only the duration of all previous ARV therapy was significantly associated with the likelihood of resistance mutations being gained. Our results show that in patients failing therapy, it is more unlikely than likely that significant mutations will be generated if the time for testing and review is kept to within approximately 5 weeks. Nevertheless, it is likely that a minority of patients will gain resistance mutations during this time and that these may have an impact on subsequent therapy.
|
|
| |
| |
|
|
|
