Ultrasensitive Assessment of Residual HIV Viraemia in HAART-Treated Patients With Persistently Undetectable Plasma HIV-RNA: A Cross-Sectional Evaluation
Journal of Medical Virology 81:400-405 (2009)
Stefano Bonora,1* Emanuele Nicastri,2 Andrea Calcagno,1 Daniel Gonzalez de Requena,1
Gabriella D'Ettorre,3 Loredana Sarmati,4 Lucia Palmisano,5 Vincenzo Vullo,3
Giovanni Di Perri,1 and Massimo Andreoni4
1Department of Infectious Diseases of the University of Torino, Torino, Italy
2National Institute for Infectious Diseases ''Lazzaro Spallanzani'', Rome, Italy
3Department of Infectious and Tropical Diseases of the University of Rome ''La Sapienza'', Rome, Italy
4Department of Public Health and Cellular Biology, Univer
sity Tor Vergata, Rome, Italy
5Istituto Superiore di Sanita`, Rome, Italy
Grant sponsor: Abbott Laboratories.
Boehringer Ingelheim Pharmaceuticals provided funding for editorial assistance. Editorial assistance was provided by Euro RSCG Life UK.
Improvements in HIV-RNA assays have made accurate detection of as few as 2 copies/ml possible. This study objective was the evaluation of ultrasensitive HIV-RNA quantitation (beneath current threshold: 50 copies/ml) in patients receiving different antiretroviral regimens. A cross-sectional, ultrasensitive measurement of
HIV-RNA levels (detection limit: 2.5 HIV-RNA copies/ml) was performed in 154 HIV-1-infected patients receiving ARV therapy, all classed as full responders according to the 50 copies/ml cut-off. Patients were undergoing treatment with two nucleoside/nucleotide reverse transcriptase inhibitors (N/NtRTIs) plus nevirapine (NVP, n1/448), efavirenz (EFV, n1/457) or lopinavir/ritonavir (LPV/r, n1/449). Undetectable HIVRNA (<2.5 copies/ml) occurred in 29/48 (60.4%),
24/57 (42.1%) and 14/49 (28.6%) NVP, EFV and LPV/r recipients, respectively. Mean virologicalsuppression (<50 copies/ml) duration was 28.6 months (median1/422, SD1/417.8), and only in LPV/r recipients length of suppression was associated with significantly lower HIV-RNA levels (P1/40.015). Mean nadir CD4+ cell count of 270 cells/mm3 (median1/4240, SD1/4194.5) was significantly lower in the LPV/r arm (P<0.001). Nadir CD4+ level correlated with virological
suppression but had opposite trends between NVP (positive) and LPV/r (negative; two tailed P1/40.01). Logistic regression analysis showed
NVP was the only independent factor associated with virologic suppression. NVP has demonstrated a distinct virological advantage at subclinical viral loads, possibly due to its greater penetration in extra-vascular compartments, warranting further investigation in the context of persistent low-level viraemia in long-term HAART.
Circulating CD4+ T-lymphocyte cell count and plasma HIV-RNA levels are the major indicators of highly active antiretroviral therapy (HAART) efficacy. CD4+ cell count is regarded as a continuous variable but plasma HIV-RNA is a categorical variable, with an efficacy threshold of viral load of 50 copies/ml. This
threshold replaced the previous threshold of 400 copies/ ml because (i) increasingly potent HAART regimens developed over the last decade can achieve the more stringent target and, (ii) prolonged exposure to regimens capable of achieving viral load <400 but not <50 copies/ml carries a significant risk of eventual virological failure and drug-resistance development [Rong et al., 2007; DHHS, 2008].
Improved HIV-RNA assay technology has made the accurate detection of as few as 2 copies/ml possible, thus potentially making quantitation of low-level HIV-RNA below the 50 copies/ml threshold clinically important (from Jules: I don't see clinical significance to sustaining viral load below 2 copies as compared to below 50 copies, nor is there any evidence on this yet). [Piwowar-Manning et al., 2003; Palmisano et al., 2005]. Indeed, a recent study has demonstrated that persistent, low-level viraemia can remain stable for at least seven years [Palmer et al., 2008]. Accordingly, a cross-sectional evaluation of virologically suppressed (viral load <50 copies/ml), HIV-infected recipients of various HAART regimens is reported here.
HIV-infected patients, aged >18 years, who achieved virological suppression (two consecutive HIV-RNA measurements <50 copies/ml in the course of regular clinical follow up every 3 months) on HAART were eligible for study enrolment. Patients meeting these criteria were screened at the Clinical Departments of
Infectious Diseases of the University of Torino and of Rome ''La Sapienza'', and of the National Institute for Infectious Diseases ''Lazzaro Spallanzani'', Rome, Italy.
Eligible patients were screened locally for further confirmation of their virological-suppression status; two plasma aliquots were stored at -80 degrees C. Upon confirmation of HIV-RNA <50 copies/ml, these aliquots were subjected to ultrasensitive HIV-RNA assay at a single centre.
Age, gender, nadir CD4+ cell count/mm3 and percentage, Current ARV regimen,number of previous regimens, total duration of HAART exposure, duration of viral load suppression under current treatment, CD4+ cell count/
mm3 and percentage were determined at the time of study entry. Viral load suppression durationwas defined as the time (months) with viral load <50 copies/ml under current treatment; in cases of switching from a previous regimen, with uninterrupted suppression (i.e., no virological failure or ''blips''), the suppression duration was that under the current regimen.
The primary endpoint was verification of whether any of the individual parameters recorded (nadir CD4+ cell count [absolute value and percentage], CD4+ cell
increase [absolute value and percentage] and treatment duration) were associated with the magnitude of viral suppression. The secondary endpoint was to see whether the magnitude of viral suppression was associated with
the intake of any specific drug or drug combination. Participation in the trial, between March 2005 and January 2007, was voluntary and written, informed
consent to perform ultrasensitive HIV-RNA assays of plasma aliquots from blood samples taken during their usual monitoring was obtained from all subjects. The
protocol, informed consent and subject information form were reviewed and approved by the local Institutional Review Board.
Plasma, obtained from blood collected in ethylenediamine tetra-acetate, was frozen (-80 degrees C) within 6 hr of collection and stored at this temperature until assayed at the Istituto Superiore di Sanita, Rome, Italy. Residual viraemia was quantified by ultrasensitive assay (detection limit: 2.5 copies/ml), using a modified Amplicor HIV-1-1 Monitor test version 1.5 (Roche Molecular Systems, Branchburg, NJ) [Palmisano et al., 2005]. Assay modifications included pelleting virus from 2 ml of plasma at 23,600 g at 488C for 2 hr, adding half the normal volume of quantification standard and resuspending the RNA pellet in 50 ml diluent. The entire re-suspension volume was assayed by reversetranscription
polymerase chain reaction amplification.
Differences between patient groups were assessed by w2 and Mann-Whitney U tests for categorical and continuous variables, respectively. Correlations were
performed by Pearson's and Spearman's analyses for continuous variables (ultrasensitive RNA and other factors). To evaluate the impact of each treatment on the suppression level, t-test and odds ratio (OR) were calculated considering viral load >2.5 copies as unsuppressed. Amultiple logistic regression model was used to obtain adjusted analyses for all considered factors. All analyses were performed using SPSS for Windows 13.0 (SPSS, Chicago, IL).
A total of 166 HIV-infected patients were screened at enrolment. Mean time from enrolment to ultrasensitive evaluation was 3 months. Twelve patients (7.2%), initially found to have plasma HIV-RNA values <50 copies/ml, based on their last two evaluations, were excluded from the analysis due to viral load >50 copies/ ml, while 154 patients were eligible for analysis (viral load <50 copies/ml) (Table I). Excluded patients were administered with EFV, NVP, and LPV/r in 50%, 25%,
and 25%of cases, respectively. Of these eligible patients, 97/154 (63%) were receiving first-line therapy; of the remaining patients, 33/57 (57.9%) had changed from a previous virologically suppressive regimen due mainly to convenience, and 24/57 (42.1%) had experienced virological failure and changed to a new, virologically suppressive regimen. All patients were receiving conventional triple-drug regimens comprising two nucleoside/nucleotide reverse transcriptase inhibitors (N/NtRTIs) plus either a non-nucleoside reverse transcriptase inhibitor (NNRTI; nevirapine [NVP, n=48] or efavirenz [EFV, n=57]) or the protease inhibitor (PI) combination lopinavir/ritonavir (LPV/r, n=49). All but 12 patients (92.2%) were receiving a cytidine analogue (lamivudine or emtricitabine), 91 patients (59%) were receiving zidovudine, 48 patients (31.2%) tenofovir, 12 (7.8%) didanosine, 10 (6.5%) abacavir and 5 (3.2%) stavudine.
Patients were grouped for analysis according to the third drug in their regimens (NVP, EFV, or LPV/r). The N/NtRTI combination distributions across these three groups were almost identical. Overall, the viral load in the three groups was as follows: NVP mean=3.5 copies/ml (median <2.5 copies/ml, 95% CI: <2.5-5.42);
EFV mean=7.46 copies/ml (median=4.30 copies/ml, 95% CI: 4.8-10.1); LPV/r mean=11.40 copies/ml (median=10 copies/ml, 95% CI: 8.26-14.56). HIV RNA
ultrasensitive assay showed that a total of 67/154 patients (43.5%) had undetectable plasma viral load (<2.5 copies/ml) while the remaining 87 had detectable viral load (2.6-45.5 copies/ml) (Fig. 1). The greatest virological suppression occurred in NVP recipients (Fig. 1), of whom 29/48 (60.4%) had undetectable plasma viral load. EFV recipients had undetectable plasma viral
load in 24/57 (42.1%) cases and LPV/r recipients had undetectable plasma in viral load 14/49 (28.6%) cases. The superior virological efficacy of NVP versus EFV achieved statistical significance when viral load was regarded as a continuous variable (P<0.01, 95% CI for difference in means: 0.71-7.21) and approached statistical significance for viral load as a categorical variable (HIV-RNA2.5;OR1/42.09,95%CI: 0.97-4.43). The superiority of NVP versus LPV/r was statistically significant for viral load as both continuous (P<0.001, 95% CI for the difference in means: 4.23-11.59) and categorical variables (OR=3.81, 95% CI: 1.61-8.75). Virological efficacy was non-significantly different between EFV and LPV/r for both viral load assumptions. No significant impact on the probability of having HIVRNA <2.5 copies/ml was associated with type of NRTIs used in the backbone (Table II).
Mean and median treatment durations were 3.37 (SD1/42.07) and 2.8 years, respectively. LPV/r recipients had the longest mean time under antiretroviral
treatment (4.04 years; median=3.3, SD=2.4), followed by NVP (3.33 years; median=2.8, SD=2.04) and EFV recipients (2.8 years, median 2.4, SD=1.55).
HAART-exposure duration was significantly different only between LPV/r and EFV (P=0.0015, 95% CI: 0.49-1.98). HAART exposure was not significantly
associated with virological outcome (Table II). The 154 patients had a mean virological suppression duration of 28.6 months (median 22, SD=17.8) and comparable mean values occurred in the defined treatment groups (EFV=27.9, SD=15.7 months; NVP=31.1, SD=19.3 months; LPV/r=27.0, SD=18.6 months). Overall suppression duration was significantly (P=0.03) associated with level of residual plasma viral load below 50 copies/ml (Table II). However, a significant association occurred only for LPV/r among the defined treatment groups (P=0.015).
A total of 97 (63%) patients were on first-line antiretroviral treatment; of the remaining 57, 57.4% switched from a previous efficacious regimen while 42.6% changed with unsuppressed viral load. No significant impact on the probability of having an undetectable viral load (HIV RNA <2.5 copies/ml) was associated with the proportion of naive patients among the different treatment groups (EFV, 42/57 [73.7%]; NVP, 30/48 [62.5%]; and LPV/r, 25/49 [51.0%]; p for the difference among the groups=0.076). Among naïve patients, proportions of subjects with HIV-RNA <2.5 copies/mol were 71% (22/30), 48.8% (20/42), and 44% (11/25) for NVP, EFV, and LPV/r, respectively.
No significant association was found between the magnitude of residual plasma viral load <50 copies/ml and increased CD4+ cell count on HAART (Table II).
The lowest mean and median nadir CD4+ cell counts were recorded in LPV/r recipients (136 and 126 cells/mm3, respectively; SD1/4105), followed by NVP (306 and 299 cells/mm3, respectively; SD1/4167) and EFV (353 and 333 cells/mm3, respectively; SD=215) recipients. Mean nadir CD4+ cell count in LPV/r recipients was significantly lower compared with NVP (P<0.001, 95%
CI for difference in means: -225 to -114) and EFV recipients (P<0.001, 95% CI: 279-153) but NVP and EFV recipients were non-significantly different. In NVP recipients, mean nadir CD4+ cell count and virological outcome (likelihood of achieving viral load <2.5 copies/ml) were significantly associated. NVP
recipients achieving viral load <2.5 copies/ml had a mean nadir CD4+ cell count of 350 cells/mm3 (SD1/4166), while those with viral load >2.5 copies/ml
had a mean count of 238 cells/mm3 (SD=148; P=0.014; 95% CI: 22-202). The opposite occurred in LPV/r recipients; those with viral load <2.5 copies/ml had a
mean nadir CD4+ cell count of 86 cells/mm3 (SD=59), while patients with viral load >2.5 copies/ml had a mean count of 156 cells/mm3 (SD=113; P=0.007; 95% CI: 21-118).
Fig. 1. UltrasensitiveHIV-RNA measurement according to different treatment. *Box plots of ultrasensitive HIV-RNA measurements (detection limit: 2.5 copies/ml) according to third drug of current HAART regimen (NVP, nevirapine; EFV, efavirenz; LPV/r, lopinavir/ritonavir). Black bars indicate median values, boxes cover the interquartil range (25th through 75th percentiles) and tails show the 5th and 95th percentiles, while circles and stars represent respectively outlyer (between 1.5 and 3 times the interquartile range) and extreme values (more than three times the interquartile range). The superior virological performance of NVP was statistically significant versus EFV, with HIV-RNA levels taken as continuous variables (P<0.01, 95% CI for difference in means: 0.71-7.21) and approached statistical significance when categoric variables were instead taken (HIV-RNA 2.5 copies/ml; Odds ratio [OR] 2.09, 95% CI: 0.97-4.43), while superiority of NVP versus LPV/r was statistically significant for both
continuous (P<0.001, 95% CI for difference in means: 4.23-11.59) and categorical variables (OR 3.81, 95% CI: 1.61-8.75). The difference between EFV and LPV/r did not reach statistical significance for either variable analysis.
In ARV therapy, the ability to evaluate treatment efficacy and estimate its performance over time is based on measurements of immunological recovery magnitude and viral suppression [DHHS, 2008]. The present study evaluated the virological performance of a series of patients, who were classifiable as full virological responders (based on the current efficacy threshold: 50 HIV-RNA copies/ml), using an assay capable of detecting 2.5 copies/ml. HAART responders were analyzed under different regimens and in terms of HAART
exposure duration and nadir CD4+ cell count.
The study design has the fundamental bias of recruiting only virological responders to HAART and therefore cannot serve as a conventional comparison among drugs or regimens. However, lower threshold HIV-RNA values are likely to be used increasingly for the virological monitoring of ARV therapy in the near
future; efforts toward attaining a better knowledge of virus levels occurring at a viral load below the current threshold for virological efficacy are warranted.
Initially, investigations into whether certain parameters, such as individual immunological profile at the time of HAART initiation and duration of successful HAART exposure, had any effect on the magnitude of the residual viral burden measurable below the established efficacy threshold of 50 HIV-RNA copies/
ml, were made. Observations in this study indicate that nadir CD4+ cell count exerts a degree of influence on the magnitude of virological suppression in both NVP and LPV/r recipients. However, while for the former, it is plausible that a higher nadir CD4+ cell count has a positive influence on the overall treatment outcome [Powderly et al., 1999], the same does not apply to the findings recorded for patients taking LPV/r, which are the opposite of those observed for NVP. It should be noted that the cohort studied here was somewhat heterogeneous, reflecting the ''real-world'' use of NVP, LPV/r, and EFV in clinical practice. LPV/r, particularly in Italy, is perceived as a safer drug for HIV patients
with a more progressed disease state. This view tends to be confirmed by the significantly different nadir CD4+ cell count of LPV/r recipients (mean=136.5 cells/mm3) compared with that recorded in NVP (mean=306.5 cells/mm3) and EFV (mean=353.6 cells/mm3) recipients. A possible interpretation might be differences in treatment adherence rates between more advanced naïve patients (who started HAART with greater evidence of AIDS-associated disease) and HIV-infected patients with less deteriorated immune and clinical conditions, who may perceive a less critical need for adherence to HAART [Horne et al., 2007]. Short interruptions of drug intake might not necessarily result in detectable viral rebounds with conventional HIV-RNA assay. However, rebounds currently regarded as being of lesser clinical relevance may become apparent following an ultrasensitive HIV-RNA assay, despite patients remaining classifiable as full responders on the basis of the 50 copies/ml cut-off. LPV/r would seem the most vulnerable to poor adherence of the three antiretrovirals studied. This may be due to it a shorter elimination half-life [Klein et al., 2007], thus making it more likely that a greater viral rebound beneath the current viral load threshold would
occur in the event of suboptimal adherence to LPV/r as compared with suboptimal adherence to NNRTIs. Further study is required in order to test this hypothesis.
Other critical individual variables, such as the possible influence of treatment duration on the magnitude of viral load between 0 and 50 copies/ml may have played a role. However, no such association was found, either by inter- or intra-group analyses. Although a significantly longer duration on HAART was observed in LPV/r recipients compared with EFV or NVP recipients, this seems unlikely to have had any influence on virological outcome (LPV/r virological performance was inferior to that of the other two groups). Overall, a positive association was found between the duration of suppression to <50 copies/ml and the viral load measured between 2.5 and 50 copies/ml but this was confirmed only for LPV/r recipients by intra-group analysis.
Further investigations into whether the magnitude of viral inhibition was associated with any specific ARV were conducted. In the case series investigated here, the best performance in absolute terms was achieved by NVP-based regimens; to some extent this was an unexpected finding given that NVP is not among the first-line options in HIV treatment guidelines [DHHS, 2008]. Although NVP recipients in this series undoubtedly had some initial advantage over LPV/r recipients in terms of a higher nadir CD4+ cell count (this turned out to be a paradoxical advantage for the latter drug, possibly for adherence reasons), comparable nadir figures were recorded for EFV. This unpredicted suppressive characteristic associated with NVP remained significant; 45/48 NVP (94%) recipients achieved viral load <10 copies/ml, whereas the same characteristic was observed in 41/57 (72%) EFV and 24/49 (49%) LPV/r recipients, respectively.
Different explanations may be given for the superior virological efficacy of NVP. Firstly, few randomized clinical trials have demonstrated non- inferiority between NVP- and EFV-based HAART, either as an initial regimen in treatment-naïve patients [van Leth et al., 2004] or as a switching option in treatment-experienced
patients [Martinez et al., 2003]. In particular, recent data from the NEFA trial at 3-year followup show that, in 2 out of 3 intention-to-treat analyses, NVP had a significantly better virological performance compared with EFV and abacavir (ABC) as simplification strategy from a PI-based HAART [Martinez and
Gatell, 2007]. Secondly, NVP data from this study might be explained by the greater tissue penetration of NVP compared with PIs and EFV. NVP significantly penetrates certain body compartments, such as the central nervous system, genital tracts and breast milk [Taylor et al., 2000; Reddy et al., 2003; Shapiro et al., 2005]. For instance, NVP showed a better penetration in central nervous system as compared to EFV [Letendre et al., 2008], and a better penetration in semen as compared to both EFV and LPV/r [Chan and Ray, 2008]. This property, which is quantitatively remarkable as compared with other antiretrovirals, might correspond to the ability of the drug to exert a significant
suppressive action even in those ''PK sanctuaries'' (above-cited body compartments), where other antiretrovirals have poor penetration and are not regarded as being capable of any marked inhibition of locally produced virions. We might suppose that some of the virions remaining measurable by ultrasensitive assay in the <2.5-50 copies/ml interval are produced in certain
body compartments, which NVP is able to penetrate and, hence, exert a stronger inhibitory effect compared with either EFV or LPV/r. This perspective is lent credence by a recent study, which has reported that low level, persistent viraemia under suppressive antiretroviral therapy (LPV/r plus stavudine and lamivudine) appears to arise from at least two cell compartments; one in which viral replication declines with time and another in which replication remains stable for at least seven years [Palmer et al., 2008].
Among patients with HIV infection who achieved virological suppression on HAART, those receiving NVP-based regimens were able to exert a more profound inhibition of viral replication (observed below the currently adopted threshold of virological suppression) than those receiving EFV- or LPV/r-based regimens. The clinical importance of this finding deserves further study in the context of persistent low-level viraemia under long-term HAART and the impact of antiretroviral penetration into long-lived reservoirs of virus.