| |
Prevalence of Antiretroviral Drug Resistance Mutations in Chronically HIV--Infected, Treatment-Naive Patients: Implications for Routine Resistance Screening before Initiation of Antiretroviral Therapy
|
| |
| |
Clinical Infectious Diseases Feb 1, 2005;40:468-474
Richard M. Novak,1 Li Chen,2 Rodger D. MacArthur,3 John D. Baxter,4 Kathy Huppler Hullsiek,2 Grace Peng,2 Ying Xiang,2 Christopher Henely,5 Barry Schmetter,6 Jonathan Uy,1 Mary van den Berg-Wolf,7 Michael Kozal,8 and the Terry Beirn Community Programs for Clinical Research on AIDS 058 Study Team
1University of Illinois, Chicago; 2University of Minnesota, Minneapolis; 3Wayne State University, Detroit, Michigan; 4University of Medicine and Dentistry of New Jersey, Camden, and 5Southern New Jersey AIDS Clinical Trials, Belmar, New Jersey; 6Social and Scientific Systems, Silver Springs, Maryland; 7Temple University, Philadelphia, Pennsylvania; and 8Yale University, New Haven, Connecticut
This study analyzed retrospectively the prevalence of genotypic drug resistance in 490 HIV+ treatment naives. Average CD4 count was 269. Results suggest many patients have had HIV for 7-8 years. The study found 11% of study patients had genotypic drug resistance & 11% had >1 resistance mutation. 6 patients had the K103N mutation, the key mutation associated with resistance to NNRTIs nevirapine & efavirenz. They reported prevalence of drug resistance increased over time to 14% in 2001, last year of study. The authors conclude: These results demonstrate the persistence of drug resistance mutations in chronically HIV-infected patients and an increasing prevalence of resistance over time, and they support genotyping of virus at baseline for chronically HIV-infected patients.
The authors said:
This study provides one of the first descriptions of the prevalence of resistance to HIV drugs in chronically infected treatment-naive patients representing a geographically and racially diverse population in the United States. The results indicate that prevalence of mutations conferring antiretroviral resistance in chronically infected patients is common (10.8%) and that genotypic resistance testing of chronically infected patients before initiation of therapy may be appropriate. The population in this study had more advanced HIV disease and had a larger proportion of women than did previous studies and is representative of the ongoing epidemic of HIV infection in the United States and the type of HIV-infected patients who are currently presenting to care......A disturbing—but not unexpected—observation was the increasing trend in prevalence by year of enrollment. In 2001, the last full year of study enrollment, the prevalence of any mutation was 14.7%. This implies that the current prevalence may well be higher than that observed during the 3 years that this study accrued patients......The persistence of these mutations is of concern, because they increase the chance of secondary transmission of drug-resistant variants.....The most prevalent mutations were in the NRTI class. This is not surprising, given that the population had a mean CD4 cell count of 269 cells/mm3, suggesting that many of these patients may have been infected for 〜7--8 years. The NRTIs have been available longer than the other classes screened for in genotype resistance tests, and it would be expected that the greater length of exposure would lead to a higher prevalence of mutations.
15% had a history of injection drug use; 4% were positive for hepatitis B surface antigen, and 20% were coinfected with hepatitis C virus. Among the male subjects, 69% had sex with men....6 patients had the K103N NNRTI mutation...7 patients had the 108I NNRTI mutation....2 patients had the 181C NNRTI mutation....1 patient had the 84V PI mutation...2 patients had the 46L PI mutation & 1 patient had the 46I PI mutation..2 patients had the PI 30N mutation
Prevalence of antiretroviral resistance mutations. Overall, 57 (11.6%) of the 491 patients sampled had >1 resistance mutation, resulting in an estimated prevalence for the cohort of 10.8% (95% CI, 9.5%--12.1%). By drug class, the estimated prevalence of mutations conferring resistance was 7.8% for NRTIs, 3.0% for NNRTIs, and 0.7% for PIs. Additionally, samples for 0.7% of patients were resistant to >1 class. Of the 57 patients with resistance mutations, 33 (58%) had a susceptibility interpretation of intermediate or resistant to >1 drug. If the 118I mutation, which has been associated with lamivudine resistance but can also occur as a natural polymorphism, was removed from the definition for resistance mutations for NRTIs, as reported in a recent study by the CDC, then the estimated prevalence of mutations was 8.8% for any drug class and 5.6% for NRTIs.
Factors associated with resistance mutations. In univariate analyses, there was a trend for increasing prevalence of mutations from 1999 to 2001 (P = .08). In a multiple logistic regression model, there was a 40% increase per year in prevalence of mutations by later year of enrollment (adjusted OR, 1.4; 95% CI, 1.0--2.1; P = .05). In addition, non-Hispanic white subjects were more likely than African American subjects to have resistance (16.7% vs. 9.1%; adjusted OR, 2.1; 95% CI, 1.1--4.1; P = .03). The percentage of Hispanic patients with mutations was lower than the percentage of non-Hispanic white patients with mutations, but the difference was not significant (11.6% vs. 16.7%; OR, 0.6; 95% CI, 0.3--1.5; P = .29). The prevalence of resistance mutations did not differ by age, sex, injection drug use, CD4 cell count, HIV RNA level, previous diagnosis of AIDS-defining condition, or geographic distribution. Because there were too few patients (n = 16) with an ethnicity other than non-Hispanic white, African American, or Hispanic, those patients were removed from the multiple logistic regression analysis.
Specific antiretroviral resistance mutations. Note that, although no 184V mutations were observed, 17 patients had a 118I mutation, the most frequently observed amino acid change in this sample. Mutations at the 215 position of reverse transcriptase (RT), which have been associated with resistance to thymidine analogues, were also frequently seen (14 samples), as well as were revertants from 215Y, the most common 215 mutation associated with thymidine analogue resistance. Variants in the 69 position were also common, found in 12 samples. In addition, other thymidine analogue mutations observed were 41L, 44D, 70R, 210W, and 219Q. The K65R and I74V mutations were not observed. One patient had a combination of 151M, 70R, 77L, and 116Y mutations, associated with resistance to multiple drugs (data not shown). Other thymidine analogue mutations observed together in patients with multiple mutations were 70R, 69D, and 219Q in 1 patient and 41L and 210W in 2 patients. In the protease gene, multiple polymorphisms were observed, known to be naturally occurring, and are not reported here. However, 30N and 88D were observed together. The NNRTI mutations most often seen were 103N and 108I.
AUTHOR DISCUSSION: Among the detected NRTI mutations, the thymidine analogue mutations were the most common. In addition, RT codon 215 changes known as revertants (e.g., C/S/D) were commonly observed and are indicative of a prior 215Y mutation. These codon 215 substitutions have been associated with virological failure during treatment with HAART regimens that contained stavudine or zidovudine.
In general, NNRTI and PI mutations were seen less frequently, possibly reflecting their later introduction into the antiretroviral armamentarium. However, 6 patients had the RT codon 103N mutation, which confers resistance to the entire class of NNRTIs and has been recently shown to persist for years as the dominant viral variant in patients acutely infected with a drug-resistant K103N strain. Seven patients were found to have 108I mutations. Although this mutation is not associated with reduction in activity of NNRTIs, it is indicative of NNRTI selection pressure and therefore may serve as a marker for other more important, archived mutations. This principle applies to all detectable mutations; thus, the phenotypic interpretation provided with a genotype resistance test might underestimate the prevalence of resistance. The presence of minor NRTI and NNRTI mutations should raise suspicion of other archived mutations in the same class. In this respect, genotype testing will likely be more useful than phenotype testing at baseline.
A major strength of this study design is that it uses samples obtained from a large, clinical end point study. The resistance data will ultimately be applied to the clinical end points of the larger study, once completed and unblinded, to determine whether baseline resistance mutations have an impact on clinical outcomes.
A weakness of this study design was a lack of access to detuned assay (a modified HIV antibody assay used to indicate recent seroconversion) data to support the contention that the patients in this study were chronically infected. However, no difference in prevalence of mutations by CD4 cell count was observed in our study, which makes the impact of recent infections, if present, in this group less relevant. Additionally, in the CDC study by Weinstock et al, there was no difference in prevalence of resistance mutations by duration of infection on the basis of detuned assay testing. Last, the decision to perform resistance testing at baseline will be made by clinicians without the benefit of detuned assay results, and therefore, their judgment regarding resistance testing will more likely be guided by clinical data, as in our study.
In conclusion, the prevalence of primary HIV antiretroviral resistance in a chronically infected, geographically, racially, and sexually diverse group of patients with moderately advanced HIV disease is 〜9%--11% and appears to be increasing over time. These data support routine genotypic resistance testing of all treatment-naive patients before initiation of antiretroviral therapy. It is unknown whether this resistance will translate into slower treatment response rates and higher rates of treatment failure among chronically infected patients, but the increased number of treatment failures observed among newly infected patients undergoing treatment may portend a similar outcome for persons with more advanced disease. The FIRST study is a clinical end point trial comparing 3 treatment strategies for long-term outcomes, and this study, once completed, will be able to report the effect of primary resistance on clinical outcomes. Clinical management of treatment-naive, chronically HIV-infected patients in the United States will need to address the growing problem of primary HIV drug resistance.
ABSTRACT
Background. The prevalence of drug resistance among persons with newly acquired human immunodeficiency virus (HIV) infection is well documented. However, it is unclear to what extent these mutations persist in chronically infected, treatment-naive patients.
Methods. Prevalence of and factors associated with genotypic drug resistance were analyzed retrospectively in a subset of 491 chronically HIV-infected, antiretroviral-naive patients enrolled at 25 cities in the Terry Beirn Community Programs for Clinical Research on Acquired Immune Deficiency Syndrome (AIDS) Flexible Initial Retrovirus Suppressive Therapies trial during 1999--2001. Resistance was defined on the basis of the International AIDS Society 2003 definition, as well as the presence of additional mutations at codons 215 (C/D/E/S) and 69 (A/N/S) in the pol gene. Prevalence of mutations was estimated by use of techniques for stratified random samples. Logistic regression models were used to determine factors associated with resistance.
Results. Among the 491 chronically HIV-infected patients (mean CD4 cell count, 269 cells/mm3; 31% of patients had a prior AIDS diagnosis), 57 (11.6%) had >1 resistance mutation, resulting in an estimated prevalence for the cohort of 10.8% (95% confidence interval [CI], 9.5%--12.1%). The prevalence was 8.8% if the 118I mutation was excluded.
By drug class, the estimated prevalence of mutations conferring resistance to nucleoside reverse-transcriptase inhibitors was 7.8%, and the prevalence was 3.0% for nonnucleoside reverse-transcriptase inhibitors and 0.7% for protease inhibitors.
In a multiple logistic regression analysis, non-Hispanic white subjects were twice as likely than African American subjects to have resistance (odds ratio [OR], 2.1; 95% CI, 1.1--4.1; P = .03), and there was a 40% increase per year in prevalence of mutations by later year of enrollment (OR, 1.4; 95% CI, 1.0--2.1; P = .05).
Conclusions. These results demonstrate the persistence of drug resistance mutations in chronically HIV-infected patients and an increasing prevalence of resistance over time, and they support genotyping of virus at baseline for chronically HIV-infected patients.
RESULTS
Patient characteristics. Among the 504 patients selected, 13 had used NRTIs and were excluded from the analysis. The characteristics of the remaining 491 patients are described in table 1. The average age was 38 years. Most of the patients were African American (52%); 16% were Hispanic. Nineteen percent of subjects were women. Overall, 15% had a history of injection drug use; 4% were positive for hepatitis B surface antigen, and 20% were coinfected with hepatitis C virus. Among the male subjects, 69% had sex with men. The mean CD4 cell count was 269 cells/mm3, and the median was 266 cells/mm3 (interquartile range, 76--395 cells/mm3). The mean log10 HIV RNA level was 5.0 copies/mL, and the median was 5.0 copies/mL (interquartile range, 4.4--5.5 copies/mL). Approximately 69% of patients had a CD4 cell count below 350 cells/mm3, and 31% of patients previously had an AIDS-defining event. In general, the characteristics of patients with mutations were not significantly different from those of patients without mutations. However, 42% of patients with mutations and 28% of those without mutations were non-Hispanic white subjects (P = .02).
INTRODUCTION
Since the introduction of antiretroviral therapy for persons with HIV infection 17 years ago, the rates of mortality and morbidity related to HIV disease have decreased. However, an undesired consequence of antiretroviral therapy has been the selection of virus variants with mutations that lead to decreased drug susceptibility. Persons infected with drug-resistant HIV have decreased treatment options and can have poor therapeutic responses. Recent data suggest that new infection with a virus strain already resistant to antiretroviral drugs has a negative impact on initial treatment response and also shortens the time to first virological failure. The transmission of HIV with mutations that confer antiretroviral resistance during the acute phase of HIV infection has been documented in numerous studies from many geographic areas where patients have ongoing access to antiretroviral medicines. Most of these studies have focused on those with recent seroconversion to HIV. Little is known about how long resistance mutations acquired during primary infection persist in the plasma of patients with established infection who have not yet been exposed to antiretroviral drugs. A recent report that described 11 patients with transmitted resistant virus indicated that, after transmission of a virus bearing a K103N mutation, the average time to reversion of this mutation back to wild-type was 196 days, and no reversions were seen for protease inhibitor (PI) mutations out to 342 days. Similarly, a study of 6 patients with primary resistance demonstrated no change in genotypic or phenotypic resistance to nonnucleoside reverse-transcriptase inhibitors (NNRTIs) or PIs over a median of 12 months of follow-up and partial loss of resistance to nucleoside reverse-transcriptase inhibitors (NRTIs) in 3 persons. Given these observations, a critical question in the field of HIV study is whether genotypic resistance testing is indicated for chronically infected, antiretroviral-naive patients preparing to initiate therapy. The current Department of Health and Human Services guidelines for antiretroviral therapy do not generally recommend resistance testing for chronically infected, treatment-naive patients, primarily because of a lack of information concerning the prevalence of drug resistance in this population. They cite a prevalence of 5% as a possible level at which routine testing may be cost-effective. The guidelines do recommend that genotypic resistance testing be considered for patients with acute infection. The International AIDS Society--USA recommends resistance testing for persons with established infections up to 2 years from the time of infection and possibly longer but, again, cite the lack of data to support testing in cases of chronic infection. Most patients with new diagnoses who seek clinical care fit the definition of chronic HIV infection rather than that of acute infection, and the prevalence of resistance mutations among these patients needs elucidation.
The Terry Beirn Community Program for Clinical Research on AIDS (CPCRA), a multicenter network, began a treatment strategy trial for HIV-infected persons naive to treatment in 1999. The Flexible Initial Retrovirus Suppressive Therapies (FIRST) trial (CPCRA 058) enrolled a cohort of 1397 demographically diverse treatment-naive subjects from 18 sites encompassing 25 US communities. A subset of baseline samples from the FIRST cohort was randomly selected and analyzed to determine the prevalence of primary drug resistance in this large, demographically diverse cohort of chronically infected patients, recruited over a 3-year period. The intent of this project was to inform the HIV clinical community regarding the necessity of routine resistance screening for such patients before initiation of antiretroviral therapy.
METHODS
Study design. The CPCRA FIRST study is an ongoing, prospective, randomized trial that compares 3 treatment strategies for the initial treatment of antiretroviral-naive patients with HIV infection: 1--2 PIs plus NRTI, NNRTI plus NRTI, or PI plus an NNRTI plus 1 or 2 NRTIs. Treatment strategies were defined by classes of drugs, not by specific antiretrovirals. For each of the 3 strategy arms and before randomization, patients and their clinicians were given the option of either preselecting the particular drugs to be used in a class or indicating that they would allow additional randomizations to study-specific drugs within each class. Full details of the study design have been published elsewhere. The FIRST study reached its target enrollment and closed to accrual on 11 January 2002, with 1397 patients enrolled. The FIRST study was approved by the institutional review boards at the institutions where the study is conducted, and written informed consent was obtained from all study participants.
Study population. HIV-infected persons who were >13 years of age, and who, in the cases of women with childbearing potential, were willing to use a barrier method of birth control throughout the course of the study, were eligible to participate. Patients were excluded for the following reasons: they were pregnant or breast-feeding or they had any conditions that precluded successful participation (in the judgment of the investigator), any previous PI or NNRTI use, a cumulative total of >4 weeks of NRTI use, or use of lamivudine for >1 week.
Data collection. Before randomization, a baseline evaluation was done that included medical and treatment history, physical examination, and determination of hepatitis B and C serostatus and previous diagnosis of AIDS-defining conditions, as well as obtainment of specimens for laboratory measures, such as CD4 cell count, HIV RNA level, and aspartate aminotransferase and alanine aminotransferase levels. A modified version of the previously published Centers for Disease Control and Prevention (CDC) criteria for AIDS-defining conditions was used. Genotypic resistance tests were performed on stored specimens at a CPCRA-designated central laboratory, with plasma-derived viral RNA, by use of the TRUGENE HIV-1 genotyping kit and OpenGene DNA sequencing system. The susceptibility to each drug was reported as susceptible, intermediate, or resistant, according to the CPCRA genotypic interpretive algorithm.
|
|
| |
| |
|
|
|
