| |
Low-Level HIV Viremia May Cause Failure
|
| |
| |
"Genotypic Resistance in HIV-1--Infected Patients with Persistently Detectable Low-Level Viremia while Receiving Highly Active Antiretroviral Therapy"
Clinical Infectious Diseases October 1, 2004;39:1030-1037
Richard E. Nettles,1 Tara L. Kieffer,1 Rachel P. Simmons,1 Joseph Cofrancesco, Jr.,1 Richard D. Moore,1 Joel E. Gallant,1 Deborah Persaud,2 and Robert F. Siliciano1,3
Departments of 1Medicine and 2Pediatrics, Johns Hopkins University School of Medicine, and 3Howard Hughes Medical Institute, Baltimore, Maryland
AUTHOR COMMENTS:
Once a nadir pVL of <50 copies/mL is achieved, 25%--50% of patients subsequently experience transient episodes of detectable, low-level viremia. In most studies, these episodes have not correlated with virologic failure. Among patients with previously undetectable viremia receiving HAART, 〜4% per year will subsequently experience persistent, detectable, low-level viremia that often leads to overt virologic failure if the HAART regimen is not altered
For the 43% of patients with isolates that were resistant to at least 1 antiretroviral agent in the HAART regimen received during the study, ongoing, persistent, low-level viremia was particularly concerning, because sequential accumulation of additional mutations conferring resistance to the remaining antiretroviral drugs in the regimen may lead to virologic failure and may limit future treatment options. Early intensification or modification of HAART regimens for these patients may be warranted
As in the 2 previously published reports on genotypic resistance during low-level viremia, RT mutations—specifically, M184V and M41L/T215Y—were particularly common in isolates from patients with detectable, low-level viremia in this study. It is of concern that isolates from our population also showed a high frequency of the K65R mutation during detectable, low-level viremia
patients who did not have isolates with resistance mutations were able to maintain a median pVL of 89 copies/mL for a median duration of >14 months without detection of isolates that developed resistance. It appears that a persistent pVL slightly >50 copies/mL may not always lead to the evolution of detectable drug resistance mutations
ABSTRACT
Background. Technical limitations in the sensitivity of commercial genotyping methods may prevent clinicians from determining whether drug-resistant human immunodeficiency virus type 1 (HIV-1) is present in patients with low-level viremia. We performed ultrasensitive HIV-1 genotyping for patients with persistent plasma virus loads of 50--400 copies/mL to better define the prevalence of drug resistance and the most common resistance mutations during persistently detectable low-level viremia.
Methods. Genotyping of HIV-1 was performed with an ultrasensitive clonal genotyping method.
Results. We studied 21 patients who had persistent, detectable, low-level viremia for a median of 11 months. Nine (43%) of 21 patients had HIV-1 isolates with significant resistance mutations. The most common mutations were M184V, K65R, and M41L/T215Y.
Conclusions. The finding that clinically significant resistance mutations were present in some but not all patients with persistent viremia (range, 50--400 copies/mL) highlights the need to improve the sensitivity of current clinical assays for detection of drug resistance.
INTRODUCTION
For the >40 million people infected with HIV-1, the best current hope for delaying disease progression and prolonging survival lies in treatment with HAART, which consists of combinations of ⩾3 drugs that inhibit HIV-1 reverse transcriptase (RT), protease, or the virus itself. HAART can reduce the plasma virus load (pVL) to below the limit of detection of 50 copies/mL of ultrasensitive clinical assays. This reduction in viremia allows a considerable degree of immune reconstitution in the majority of patients.
Although the benefits of HAART in reducing morbidity and mortality are clear, many questions remain concerning its optimal use. Overt HAART failure (characterized by pVLs that are persistently >400 copies/mL) can be caused by poor adherence to therapy, inadequate antiretroviral drug bioavailability or potency, or antiretroviral drug resistance. The nadir pVL during treatment is predictive of the duration of suppression of viremia. Durable suppression generally requires that the regimen reduce HIV-1 RNA viremia to <50 copies/mL by 4--6 months after initiation of therapy. Multiple clinical trials have shown that the long-term durability of response is improved for patients who achieve and maintain a nadir pVL of <50 copies/mL, compared with those with nadir pVLs of 50--400 copies/mL. Thus, suppression to <50 copies/mL is the goal of current therapy.
Once a nadir pVL of <50 copies/mL is achieved, 25%--50% of patients subsequently experience transient episodes of detectable, low-level viremia. In most studies, these episodes have not correlated with virologic failure. However, among patients with previously undetectable viremia receiving HAART, 〜4% per year will subsequently experience persistent, detectable, low-level viremia that often leads to overt virologic failure if the HAART regimen is not altered.
Virologic failure associated with detectable, low-level viremia appears to be linked to the early development of drug resistance mutations. Gunthard et al. found evidence for evolving genotypic resistance in patients with a consistent pVL of 20--400 copies/mL. Parkin et al. retrospectively evaluated HIV-1 genotypes in a longitudinal study involving 16 patients enrolled in a 4-drug salvage therapy trial and demonstrated that drug resistance could be detected at virus loads of <1000 copies/mL and that failure of the salvage regimen was typically associated with early evolution of drug resistance. Aleman et al. retrospectively examined HIV-1 genotypes in 14 patients receiving HAART who had a nadir pVL of <500 copies/mL, but who thereafter experienced persistent, low-grade viremia (<1000 copies/mL) or a slow increase to <4000 copies/mL, and showed that treatment failure, even during low-level viremia, can lead to the emergence of new drug resistance mutations.
Although there is general agreement that patients who have repeated pVL measurements of 50--400 copies/mL are at risk for overt HAART failure, it remains unclear whether resistant viruses are present in such patients and whether genotypic or phenotypic analysis of resistance would improve the clinical management of such patients. Currently, clinical assays for resistance often fail to yield results when the pVL is <1000 copies/mL. In previous studies, we have shown that it is possible to obtain genotypic information even when the pVL is <50 copies/mL. In the present study, we used this sensitive genotyping method to evaluate the extent of resistance in patients receiving HAART who had successive pVLs of 50--400 copies/mL.
Author Discussion
In this cohort of patients with persistent, detectable, low-level viremia of 50--400 copies/mL, nearly one-half of the participants had HIV-1 isolates with genotypic evidence of resistance to drugs included in the antiretroviral regimen received during the study. For isolates in certain patients, the resistance was both diverse (i.e., there was resistance to all classes of antiretroviral drugs) and significant (i.e., there was resistance to a median of 3 of 5 antiretroviral drugs received by the patients during the study). In certain cases, the resistance mutations appeared to have been selected for by the study regimen and may have arisen during the period of low-level viremia. In such cases, we cannot exclude transmission or superinfection with resistant virus. All of the patients in whom resistance mutations were detected had mutations that could compromise that efficacy of the current regimen. Thus, the availability of genotypic information at this level of viremia has the clear potential to guide the choice of an alternative regimen before overt failure occurs. For the 43% of patients with isolates that were resistant to at least 1 antiretroviral agent in the HAART regimen received during the study, ongoing, persistent, low-level viremia was particularly concerning, because sequential accumulation of additional mutations conferring resistance to the remaining antiretroviral drugs in the regimen may lead to virologic failure and may limit future treatment options. Early intensification or modification of HAART regimens for these patients may be warranted.
The prevalence of HIV-1 mutations conferring significant drug resistance in patients who had successive samples with detectable, low-level pVL supports the need to improve the sensitivity of clinical assays for detection of resistance, so that mutations can be detected in isolates from patients with this level of viremia. The need for improved methods is highlighted by the fact that the presence of low-level viremia alone is an insufficient justification for a change in therapy. In our study, 57% of the patients with detectable, low-level viremia did not have evidence of genotypic resistance. For these patients, the current HAART regimen should have been adequate to produce and maintain an undetectable pVL, and modification of the current regimen might have led to unnecessary exposure to additional antiretroviral agents. However, this assertion must be tempered by the fact that, despite our attempts to analyze multiple viral clones in each patient, it is possible that there were resistance mutations present in a minority of the virus species that we failed to detect. In general, more clones were obtained from patients who had isolates with demonstrable resistance than from those who had isolates without resistance. Among individual patients, there was heterogeneity in resistance patterns of individual clones. Therefore, it is possible that the analysis of additional clones from patients with isolates in which resistance was not detected would have led to the discovery of minority variants with resistance mutations. Thus, the study may have underestimated the amount of resistance. Indeed, this is a problem with all forms of resistance testing. In isolates in which resistance was found, a variety of different mutations were detected. Thus, improvements in the sensitivity of clinical genotyping methods down to the level that was achieved here would enable health care professionals to determine how to modify the HAART regimen to prevent the accumulation of additional mutations.
As in the 2 previously published reports on genotypic resistance during low-level viremia, RT mutations—specifically, M184V and M41L/T215Y—were particularly common in isolates from patients with detectable, low-level viremia. It is of concern that isolates from our population also showed a high frequency of the K65R mutation during detectable, low-level viremia. The increasing frequency of K65R may be associated with the use of tenofovir DF, abacavir, and didanosine in combinations that do not include thymidine analogs, which are thought to limit the acquisition of this mutation.
Additional study of the etiology and consequences of detectable, low-level viremia in the absence of genotypic resistance is necessary. In our study, patients who did not have isolates with resistance mutations were able to maintain a median pVL of 89 copies/mL for a median duration of >14 months without detection of isolates that developed resistance. It appears that a persistent pVL slightly >50 copies/mL may not always lead to the evolution of detectable drug resistance mutations, even over this extended period. This study was not designed to differentiate between mutations selected for by previous HAART regimens and those received during the study. Therefore, we could not precisely define the pVL threshold above which resistance mutations may accumulate. For patients without genotypic evidence of drug resistance, drug potency, therapy adherence, and bioavailability may need to be maximized. Improving the technology of genotype resistance testing available to clinicians will be vital in differentiating between patients with and patients without resistant virus during low-level viremia.
We also examined clinical correlates of resistance in isolates from patients with persistent, detectable, low-level viremia. Previous nonsuppressive HAART, exposure to antiretroviral drugs before receipt of the HAART regimen administered during the study, and duration of exposure to antiretroviral drugs, particularly protease inhibitors, are associated with resistance. All of these findings fit with the current understanding of the natural history of the development of antiretroviral resistance.
It is surprising that previous non-HAART antiretroviral exposure did not predict resistance, despite the fact that some patients received previous NRTI monotherapy. This unexpected finding may be partially explained by the difficulty in retrospectively determining the degree of viral suppression achieved by the exposure to non-HAART antiretroviral therapy. pVL measurements were unavailable when monotherapy was being prescribed, and it is difficult to retrospectively access patient adherence. It is important to note that neither changes in pVLs nor changes in CD4 cell counts distinguished patients with from without isolates with genotypic resistance during persistent, low-level viremia.
We found an unexpected association between African American race and resistance. The association between African American race and longer duration of PI use, exposure to antiretroviral drugs before current HAART regimen, and previous receipt of nonsuppressive HAART likely explain the association found between race and resistance. Unfortunately, the small sample size limited our ability to use a multivariate model with all of these variables to examine the association between race and resistance more thoroughly.
Three of the patients in this cohort had isolates with genotypic evidence of resistance to all of the drugs in the HAART regimen received during the study, yet their pVL remained suppressed. The mechanism of this phenomenon is poorly understood but could be critical to designing salvage HAART regimens. Additionally, improving the sensitivity of phenotypic assays and measurements of replicative capacity will assist in understanding which antiretroviral drugs remain efficacious in these situations.
In conclusion, because persistent, detectable, low-level viremia has been associated with the development of resistance and the failure of HAART regimens, and because of the technological limitations of commercial genotyping laboratories, clinicians currently have to make educated guesses as to the presence of resistance in HIV-1 isolates from patients with low-level viremia. This study used novel ultrasensitive genotype assays at the clonal level to confirm that detectable, low-level viremia is frequently, but not always, associated with resistance. We have found that previous nonsuppressive HAART, exposure to antiretroviral drugs before the current HAART regimen, and longer duration of protease inhibitor exposure were associated with the presence of resistance. The findings of this study point to the importance of improving genotype technology and demonstrate multiple clinical applications for such ultrasensitive genotypes when available.
Patient Population
Patients were identified in collaboration with health care professionals at the Moore Clinic, an HIV-specialty clinic at the Johns Hopkins Hospital (Baltimore, MD). Informed consent was obtained from all study participants. Patients were selected on the basis of the following 4 criteria: (1) documented HIV-1 infection, with a pVL of >1000 copies/mL at some point during infection (to exclude patients who naturally control viremia); (2) treatment with a stable HAART regimen for at least 3 months before genotyping, during which intensification of therapy with drugs in the same class as those the patient was already receiving (e.g., addition of abacavir to a regimen of zidovudine, lamivudine, and efavirenz) and substitution of drugs in the same class (e.g., substitution of lopinavir/ritonavir for saquinavir and ritonavir) were permitted; (3) viremia of 50--400 copies/mL for at least 3 months, with a minimum of 2 consecutive pVLs in this range (patients could have pVLs of <50 copies/mL or >400 copies/mL during the period studied, but the final pVL during routine clinical follow-up must have returned to 50--400 copies/mL before genotyping); and (4) for patients recently initiating a new HAART regimen, the pVL must have rebounded from the nadir level (to exclude patients who were slowly reaching a pVL of <50 copies/mL after the initiation of HAART).
RESULTS
By means of a sensitive clonal genotyping method, we analyzed drug resistance mutations in HIV-1 isolates from 21 patients receiving HAART who had persistent, detectable, low-level viremia for a median duration of 11 months (range, 3--45 months) (table 1). A mean of 5 samples with a pVL of 50--400 copies/mL (range, 3--9 measurements) were collected from each patient during the study period. Genotyping revealed that 9 (43%) of 21 patients had isolates with significant genotypic resistance to at least 1 antiretroviral drug in the HAART regimen they were currently receiving. Eight (38%) of the 21 patients receiving nucleoside/nucleotide RT inhibitors (NRTIs) had NRTI-resistant isolates, 2 (12.5%) of 16 patients receiving a nonnucleoside RT inhibitor (NNRTI) had NNRTI-resistant isolates, and 2 (29%) of 7 patients receiving a protease inhibitor (PI) had PI-resistant isolates. Two (10%) of the 21 patients had isolates with resistance to 2 classes of antiretroviral drugs, and 1 (5%) of 21 had isolates with resistance to 3 classes. Two patients (10%) had isolates with genotypic resistance mutations that conferred resistance to all of the medications in their HAART regimen. All resistant isolates had some RT resistance; 8 of 9 patients had NRTI-resistant isolates, and 4 of 9 had NNRTI-resistant isolates. The most common RT mutations were M184V, K65R, and M41L/T215Y; each of these were present in one-third of the patients with resistant isolates. Patients with resistant HIV-1 were receiving a median of 5 antiretroviral drugs during the study, and, for these patients, there was genotypic evidence for resistance to a median of 3 of the antiretroviral drugs they were receiving.
Eight of the 9 patients who had isolates with resistance mutations had previous exposure to antiretroviral drugs before initiating the current regiment. Thus, it was possible that the resistance mutations detected were selected by previous therapy, rather than by therapy received during the study. However, as shown in table 1, there were 3 cases in which the mutations detected could be clearly attributed to selection by the regimen received during the study. For example, in an isolate from patient 114, who had no history of virologic failure while receiving NRTIs, the K65R mutation was detected. This mutation confers intermediate resistance to lamivudine, abacavir, and tenofovir DF, all of which were administered to the patient during the study period. Barring transmission or superinfection with resistant virus, our results are consistent with the selection of resistance mutations during detectable, low-level viremia.
The potential usefulness of genotyping during low-level viremia is further highlighted by the fact that the majority of patients did not have isolates in which significant resistance was detected. Although the failure to detect resistance does not rule out the presence of resistance (in particular, the presence of resistance mutations stored in cellular reservoirs), the analysis of multiple plasma clones is likely to reveal the dominant species present at the time of sampling. the protease and RT regions in a median of 4 and 2 clones, respectively, were analyzed per patient. In the majority of patients studied, no resistance was detected, despite recurrent, detectable, low-level viremia. The absence of mutations in isolates from these patients was not the result of decrease adherence at the time of genotypic sampling. In fact, on the day of sampling, most patients had a pVL of <50 copies/mL. These results suggest that, in some patients receiving HAART, low-level viremia of 50--400 copies/mL was not associated with persistent, dominant, drug resistance mutations.
Patients with and patients without HIV-1 isolates with significant resistance were similar with respect to most demographic characteristics, including sex, age, and mode of HIV-1 transmission. However, African Americans were more likely to have resistance than were white non-Hispanics (P = .005). African American race was also associated with shorter duration of low-level viremia (P = .03), longer duration of PI use (P = .04), exposure to antiretroviral drugs before the current HAART regimen (P = .02), and receipt of previous nonsuppressive HAART (i.e., pVL was >50 copies/mL) (P = .05) (data not shown).
The characteristics of the pVL (duration, magnitude, and peak) and the CD4 cell count (baseline count, nadir count, and change during low-level viremia) did not differ between patients with isolates with significant resistance and those without such isolates.
Characteristics of the current treatment regimen were similar between patients with and patients without resistance (table 4), but treatment history differed between the 2 groups. Patients with resistant isolates were more likely to have received previous nonsuppressive HAART, at least 2 consecutive pVLs of >400 copies/mL, and a longer median duration of PI exposure. Patients who were naive to antiretroviral drugs before receipt of the HAART regimen during the study were less likely to have resistant isolates (P = .06). Previous non-HAART antiretroviral exposure (e.g., zidovudine monotherapy) and previous nonsuppressive HAART with at least 2 consecutive pVLs of >50 copies/mL were not associated with an increased likelihood of having isolates with resistance to current HAART regimens.
|
|
| |
| |
|
|
|
