点击显示 收起
Department of Infectious Diseases, Copenhagen Muscle Research Centre
Department of Clinical Immunology, Rigshospitalet, Copenhagen, Denmark
Background.
Immunological and virological consequences of low-level viremia in human immunodeficiency virus (HIV) type 1infected patients receiving highly active antiretroviral therapy (HAART) remain to be determined.
Methods.
For 24 months, 101 HAART-treated, HIV-1infected patients with HIV RNA levels 200 copies/mL were followed prospectively: HIV RNA level and CD4 and CD8 cell counts were investigated every 3 months, and proviral DNA and T cell subsets were investigated every 6 months.
Results.
During follow-up, 33 patients had HIV RNA levels 20 copies/mL at all visits (uVL patients), whereas 68 patients had HIV RNA levels >20 copies/mL at 1 visit (dVL patients) (median increase, 81 copies/mL [interquartile range, 37480 copies/mL]). dVL patients had higher concentrations of CD8 cells, activated and memory T cells, and proviral DNA, compared with uVL patients (P < .05). A higher HIV RNA level was independently associated with reduced CD4 gain (P < .001). A higher HIV RNA level also was associated with increases in activated CD8+CD38+ and CD8+HLA-DR+ cells (P < .05), and a higher level of activated CD8+CD38+ cells was independently associated with reduced CD4 gain (P < .05). A higher proviral DNA level was associated with increases in CD4+CD45RA-CD28- effector cells and reductions in naive CD4+CD45RA+CD62L+ and CD8+CD45RA+CD62L+ cells (P < .05). Higher levels of activated CD4+HLA-DR+ and early differentiated CD4+CD45RA-CD28+ cells predicted increased risk of subsequent detectable viremia in patients with undetectable HIV RNA (P < .05).
Conclusion.
These findings indicate that low-level viremia and proviral DNA are intimately associated with the immunological and virological equilibrium in patients receiving HAART.
Treatment with highly active antiretroviral therapy (HAART) leads to profound virological suppression, increases in CD4 cell count, and remarkable clinical improvement [1] in most HIV-1infected patients. A significant proportion of HIV-1infected patients receiving HAART experience virological rebound [27]. Although many patients with virological rebound continue to benefit clinically and to maintain CD4 cell counts above pretreatment levels for long periods [2], the virological, immunological, and clinical consequences of intermittent or persistent low-level viremia have been debated [2, 5, 715]. Some studies have reported a low risk of clinical progression [2] and persistent increases in CD4 cell count [5, 8] for patients with ongoing viral replication. Conversely, other studies have demonstrated that intermittent or persistent low-level viremia is associated with higher residual viremia [10], steady or increasing levels of HIV in the latent reservoir [11], selection of resistant HIV [12, 13], increased risk of virological failure [14], and higher levels of activated T cells [7, 15].
The present study was designed to investigate virological and immunological markers capable of predicting virological failure in HAART-treated, HIV-1infected patients followed prospectively after HIV RNA was found to be undetectable. The study was initiated in 1997, with the expectation that a substantial proportion of the patients would experience treatment failure. However, since most patients maintained good virological control during follow-up, the data obtained from this cohort were used to identify markers capable of predicting low-level viremia. Furthermore, data from this cohort were used to investigate associations between low-level viremia, proviral DNA, CD4 gain, and T cell subsets in patients receiving HAART.
PATIENTS AND METHODS
Study population.
The present study was done at the Department of Infectious Diseases, Rigshospitalet (Copenhagen, Denmark), as described elsewhere [16]. During the period September 1997August 1998, patients with reproducible plasma HIV RNA levels 200 copies/mL were included in the study. The patients were followed prospectively for 24 months: plasma HIV RNA level, CD4 cell count, and CD8 cell count were analyzed every 3 months, whereas T cell subsets and proviral DNA were analyzed every 6 months. Immunological and virological baseline data and history of antiretroviral treatment were extracted from patient files (table 1). The present study was based on results for 101 patients for whom flow-cytometric data at 2 time points were available; consequently, 2 patients from the cohort, described elsewhere [16], were excluded before data analysis. Written, informed consent was obtained from all subjects, and the study was approved by the local ethics committee for the Copenhagen and Frederiksberg Communities, Denmark (project 01-192/97).
Plasma HIV RNA and proviral DNA levels.
Plasma HIV RNA was quantified by a standardized reverse-transcriptase polymerase chain reaction assay (Amplicor HIV-1 Monitor; Roche Diagnostic Systems). Some baseline samples were analyzed by a standard assay (lower limit of detection , 200 copies/mL), whereas all later samples were analyzed by an ultrasensitive assay (LLD, 20 copies/mL). Samples analyzed by the standard assay were reanalyzed by the ultrasensitive assay. Samples yielding a signal more than twice that of the mean from numerous negative control samples were recorded as detectable but nonquantifiable, whereas samples with weaker signals were recorded as undetectable. For statistical analysis, the samples were given values of 20 and 19 copies/mL, respectively, as described elsewhere [16]. Proviral HIV DNA copies per 106 peripheral blood mononuclear cells were quantified by a prototype assay (Amplicor HIV DNA assay; Roche Diagnostic Systems), in accordance with the manufacturer's recommendations, as described elsewhere [16].
Flow-cytometric analysis of T cells.
The total concentration of CD4 and CD8 T lymphocytes (×109 cells/L) were quantified with BD Tritest CD4/CD8/CD3 TruCount Tubes (Becton Dickinson), by use of an Epics XL-MCL flow cytometer (Beckman Coulter). T cell subsets were analyzed in fresh whole blood, at inclusion of patients in the study and every 6 months for 24 months, by direct cell-surface 4-color immunofluorescent staining using anti-CD45, -CD14, -CD3, -CD4, -CD8, -CD45RA, -CD62L, -CD28, and -CD38 and IgG1 (Beckman Coulter) and antiHLA-DR and IgG1 (DAKO). Labeled cells were analyzed by use of an Epics XL-MCL flow cytometer, and the subsequent computer analyses were done with WinList 4.0 (Verity Software House). Lymphocyte gates in forward/right-angle light scatter with <3% CD14+ monocytes were used in all analyses. The absolute cell concentration (×109 cells/L) was calculated by multiplying the proportion with the concentration of lymphocytes, obtained from hematological analysis. The following combinations of labeled antibodies were used (fluorescein isothiocyanate/phycoerythrin/R-phycoerythrin-cyanin-5.1/phycoerythrinTexas red): CD45/CD14/-/-, CD62L/CD45RA/CD4/CD3, CD62L/CD45RA/CD8/CD3, CD28/CD45RA/CD4/CD3, HLA-DR/-/CD4/CD3, and HLA-DR/CD38/CD8/CD3. The following cell definitions were used [1720]: naive CD4 or CD8 cells (CD45RA+CD62L+), memory CD4 or CD8 cells (CD45RA-CD45RO+), early differentiated CD4 cells (CD45RA-CD28+), late-differentiated CD4 cells (CD45RA+CD28-), effector CD4 cells (CD45RA-CD28-), and activated CD4 or CD8 cells (CD38+ and/or HLA-DR+).
Statistics.
Repeated-measures analyses for each investigated variable were done by use of a means model (the PROC MIXED model in the SAS software package), under the assumption of a first-order autoregressive covariance structure among the repeated measurements. Since the patients had been receiving HAART for various periods of time at inclusion, the timescale used was months from initiation of HAART. Group effects and time × group effects were included in the model, to investigate differences in the longitudinal changes between patients with and those without detectable HIV RNA during follow-up. Specific changes over time were analyzed by use of Bonferroni-adjusted, post hoc, paired t tests. Baseline variables for the 2 patient groups were compared by use of a 2-sample t test or by use of 2 and Fisher's exact tests. The goodness of fit of the mixed model was assessed by investigation of the residuals.
A random-effects model (the PROC MIXED model in the SAS software package) that assumed a variance-component covariance structure and random effects between subject levels [21] was used to investigate the statistical association between the level of an explanatory variable and the dependent variable, during follow-up. For each patient, data obtained at all 5 time points (including missing values) contributed to the analysis.
By means of this model, we first investigated whether the levels of HIV RNA or proviral DNA both at baseline and during follow-up were associated with changes in the proportions of the investigated T cell subsets (table 2). Results are presented as the mean relative change in the proportion of a T cell subset during follow-up that was associated with a 10-fold increase in HIV RNA level or a 2-fold increase in proviral DNA level.
Next, we investigated whether the levels of specific T cell subsets both at baseline and during follow-up were associated with changes in CD4 cell count (i.e., CD4 gain) (table 3). Results are presented as the mean relative change in CD4 gain during follow-up that was associated with a 2-fold increase in the proportion of a T cell subset. To adjust for baseline variables with a known or potential effect on immune reconstitution, we investigated whether CD4 gain was affected by various baseline variables (age, AIDS diagnosis, pre-HAART CD4 and CD8 cell counts, pre-HAART HIV RNA level, preinclusion CD4 and CD8 gains per month, preinclusion decrease in HIV RNA level, treatment status, number of years HIV antibody positive at inclusion, or months of HAART before inclusion). A multivariate random-effects model, adjusted for significant univariate baseline variables, was subsequently fitted (table 3). All results from the random-effects model are presented with 95% confidence intervals (CIs). Goodness of fit was assessed by the plotting of residuals against predicted values, simultaneously for all data as well as separately for each subject.
A generalized linear model (the PROC GENMOD model in the SAS software package, repeated covariance structure) with a binomial response variable and a logit link function [22] was used to investigate whether a higher level of an explanatory variable at a given time point could predict the risk of subsequent detectable viremia (HIV RNA level >20 copies/mL [yes or no]) at the subsequent 3-month visit. To predict detectable viremia in patients with currently undetectable HIV RNA, the model only included data from time points when HIV RNA was undetectable. Results are presented as the odds ratio (OR), with 95% CI, for detectable viremia at the subsequent 3-month visit, as predicted by a 2-fold increase in the proportion of a T cell subset. Goodness of fit was assessed by the plotting of grouped explanatory variables against the logit function and by inclusion of continuous and grouped explanatory variables in the same model.
Data are presented as medians with ranges or interquartile ranges (IQRs). P < .05 was considered to be significant. Statistical calculations were performed by use of SAS 8.2 (SAS Institute).
RESULTS
The patients were stratified according to HIV RNA level during the study period: 33 patients had HIV RNA levels 20 copies/mL at all visits (uVL patients), whereas 68 patients had HIV RNA levels >20 copies/mL at 1 visit (dVL patients) (table 1). Only dVL patients had HIV RNA levels of 21200 copies/mL at inclusion (table 1). During follow-up, HIV RNA and proviral DNA levels were higher in the dVL patients (group-effect P < .001 and P = .003, respectively), and HIV RNA levels increased in the dVL patients (P < .001), whereas proviral DNA levels remained unchanged in both groups (data not shown). The increase in HIV RNA levels at time points when virological rebound occurred in the dVL patients was modest, with a median increase of 81 HIV RNA copies/mL (IQR, 37480 HIV RNA copies/mL). Only 4 patients had 2 consecutive episodes with >10,000 HIV RNA copies/mL. In comparisons of baseline characteristics between the uVL and the dVL patients, pre-HAART CD8 cell count was higher among the dVL patients. No other baseline characteristics differed between the 2 groups (table 1).
Antiretroviral treatment.
The patients had received HAART for a median period of 14 months (range, 432 months) at inclusion (table 1). Most patients had received protease inhibitor (PI)based therapy with 3 or 4 drugs, and some patients also had received a nonnucleoside reverse-transcriptase inhibitor (NNRTI) (table 1). One patient had received 2 nucleoside reverse-transcriptase inhibitors (NRTIs) at inclusion, but, since similar results were obtained whether the statistical analyses excluded or included this patient, results are shown for all patients. No patients discontinued HAART during follow-up, but 37 patients had their antiretroviral treatment modified, either to another NRTI or PI combination or to include an NNRTI (table 1). For the majority of patients, the modification of therapy was due to adverse effects, rather than to treatment failure. After modification of HAART, the patients were still included in the statistical analyses.
Initial antiretroviral therapy before initiation of HAART consisted of 1, 2, 3, or 4 drugs in 38, 24, 34, and 5 patients, as described elsewhere [16]. The study groups had comparable treatment history and treatment regimens during follow-up (table 1).
Longitudinal changes in T cell subsets in uVL and dVL patients.
Owing to the low number of subjects included in the period outside 1236 months after initiation of HAART (figure 1), the longitudinal changes in T cell subsets were only investigated during this period (figure 2C2H). The HAART-induced changes in CD4 and CD8 cell counts were equal in the uVL and the dVL patients (figure 2A and 2B), although CD8 cell count was higher in the dVL patients (figure 2B). The concentrations of naive CD4 cells, naive CD8 cells, and activated CD4 cells increased equally in the 2 groups (figure 2C, 2D, and 2G), although the concentration of activated CD4 cells was higher in the dVL patients (figure 2G). The concentrations of memory CD4 cells, memory CD8 cells, and activated CD8 cells did not change during follow-up, but they were higher in the dVL patients (figure 2E, 2F, and 2H).
A higher proviral DNA level was associated with reductions in the proportions of naive CD4 and CD8 cells (table 2). Conversely, a higher proviral DNA level was associated with an increase in the proportion of effector CD4 cells (table 2). Similar results were found for concentrations of the investigated T cell subsets (data not shown).
Association between HIV RNA level or T cell subsets and CD4 gain.
In accordance with previous findings [23, 24], previous AIDS diagnosis, pre-HAART CD4 and CD8 cell counts, and preinclusion CD4 gain were found to affect CD4 gain during follow-up, in the univariate analysis (data not shown). Consequently, these variables were included in the multivariate model (table 3).
A 10-fold increase in HIV RNA level during follow-up was independently associated with a 31% reduction in CD4 gain, whereas proviral DNA level had no effect on CD4 gain (table 3). Furthermore, a 2-fold increase in the percentage of total CD8 cells was independently associated with a 10% reduction in CD4 gain (relative change, 0.90; 95% CI, 0.830.98; P < .001).
Higher proportions of naive CD4 and CD8 cells were independently associated with increases in CD4 gain, whereas a higher proportion of activated CD8 cells was independently associated with a reduction in CD4 gain (table 3). Higher proportions of memory CD4 cells, early differentiated CD4 cells, and activated CD8 cells were associated with reductions in CD4 gain, in the univariate analysis (table 3).
Risk of subsequent detectable viremia predicted by increases in specific T cell subsets.
We investigated whether a higher proportion of a specific T cell subset at a given time point could predict the risk of detectable viremia at the subsequent 3-month visit. Since none of the investigated baseline variables (mentioned in the "Statistics" subsection of Patients and Methods) could predict subsequent detectable viremia, only a univariate model was fitted.
A 2-fold increase in the proportion of activated CD4 cells predicted a 39% increased risk of subsequent detectable viremia (table 4). Furthermore, a higher proportion of early differentiated CD4 cells and late-differentiated CD4 cells predicted increased and reduced risk of subsequent detectable viremia, respectively (table 4). Similar results were found when concentrations of T cells were used to predict subsequent detectable viremia (data not shown).
DISCUSSION
The present study investigated the associations between low-level viremia, proviral DNA, CD4 gain, and T cell subsets in HIV-1infected patients with undetectable HIV RNA after initiation of HAART. The following were new findings: (1) even low-level viremia was associated with a reduction in CD4 gain; (2) proviral DNA level was associated with CD4 T cell activation; (3) CD4 T cell activation predicted subsequent viremia; and (4) early and late-differentiated CD4 cells had opposite effects on the risk of subsequent viremia.
First, in the present study, a higher HIV RNA level was independently associated with a reduction in CD4 gain, as well as with increases in the proportions of memory and activated CD8 cells. Since a large majority of the patients maintained relatively good virological control during follow-up (median increase in HIV RNA level at time points with virological rebound, 81 copies/mL [IQR, 37480 copies/mL]), these findings can most likely be attributed to low-level viremia.
It recently has been demonstrated that viral replication might continue in patients with undetectable HIV RNA [25, 26] and that intensification of HAART for these patients leads to significant reductions in viral replication [27, 28] and T cell activation [27]. In contrast to the results of the present study, Havlir et al. [27] found no association between low-level viremia and CD4 gain in patients receiving HAART. However, the divergent findings with regard to CD4 gain could be explained by differences in the magnitude of the low-level viremia (23 HIV RNA copies/mL in [27]) or by the low number of subjects (n = 5) included in the study by Havlir et al. [27]. Although the present study is the first, to our knowledge, to demonstrate a negative association between even low-level viremia and immune reconstitution in patients receiving HAART, this potentially important result should be confirmed in later studies.
The concentration of total CD4 cells and naive CD4 and CD8 cells increased equally in uVL and dVL patients 1236 months after initiation of HAART. Since most of the dVL patients had intermittent viremia, the divergent findings in the present study could indicate that CD4 gain was reduced at time points when viremia was detectable but not at time points when HIV RNA was undetectable (the random-effects model estimated the association between a higher HIV RNA level [i.e., detectable] and CD4 gain, at individual time points). This notionthat is, time points when even low-level viremia was detected may have been associated with altered immune functionis supported by a recent study demonstrating varying immune function at time points when viremia was or was not detectable, in patients with intermittent low-level viremia [7].
In the present study, the changes in the concentration of total CD8 cells, activated CD8 cells, memory CD4 cells, and memory CD8 cells and in proviral DNA level were comparable in uVL and dVL patients. However, these variables were higher in the dVL patients, indicating that low-level viremia may be associated with T cell activation and a higher proviral DNA level. The present study also found a negative association between T cell activation and CD4 gain. This finding is in accordance with a recent study, by Hunt et al. [29], demonstrating that higher levels of activated CD4 and CD8 cells in HAART-treated patients are associated with reduced early and reduced late CD4 gains, respectively. The negative effect of T cell activation on the progression of HIV disease and on CD4 gain in untreated and HAART-treated patients, respectively, most likely reflects that T cell depletion in HIV infection is driven mainly by immune activation [3034]. Overall, the findings in the present study indicate that periods of even low-level viremia in patients receiving HAART are negatively associated with immune reconstitution, either directly or through induction of immune activation.
Second, a higher proviral DNA level was associated with an increase in the proportion of effector CD4 cells and with reductions in the proportions of naive CD4 and CD8 cells. Whether this finding can be attributed to replication-competent or replication-defective genomes or to both, through production of viral proteins, is not known, but this finding indicates that the proviral DNA level interferes with immune reconstitution in patients receiving HAART. Further studies are needed to confirm this finding and to understand the potential biological effect of proviral DNA in patients receiving HAART.
Third, the present study investigated whether a higher proportion of specific T cell subsets could predict the risk of detectable viremia, at the subsequent 3-month visit, in patients with undetectable HIV RNA. Higher proportions of activated CD4 cells predicted an increased risk of subsequent detectable viremia. The finding that the proportion of activated CD8 cells could not predict subsequent detectable viremia may seem notable but most likely reflects that activation of HIV-infected CD4 cells leads to HIV replication. Although the CD4 T cell activation detected in the present study may be attributed to both low-level viral replication and concurrent infections, this finding suggests that CD4 T cell activation favors viral replication even during HAART.
Fourth, we found that higher proportions of early or late-differentiated CD4 cells predicted an increased or a reduced risk, respectively, of subsequent detectable viremia. Different functions of early and late-differentiated CD4 cells may explain this finding. In brief, early differentiated CD4 cells have a high proliferative capacity but no cytotoxic potential, whereas late-differentiated CD4 cells have a reduced proliferative capacity but a cytotoxic potential [19]. Thus, cytotoxic late-differentiated CD4 cells possibly exhibit cytotoxicity against HIV-infected cells. However, the opposite predictive value of early and late-differentiated CD4 cells also could reflect a high or a low degree of HIV infection, respectively, for these cells.
When interpreting the findings in the present study, some limitations should be considered. (1) The study was designed to investigate virological and immunological markers capable of predicting virological failure in HAART-treated patients. However, since most patients maintained good virological control during follow-up, the data from this cohort were used to identify markers capable of predicting low-level viremia. Thus, the results from the present study should be confirmed by other studies. (2) The patients were stratified according to HIV RNA level during follow-up (detectable viremia vs. undetectable HIV RNA), without distinction between patients with intermittent and those with persistent low-level viremia. However, the present study did investigate whether the longitudinal changes in the investigated T cell subsets and proviral DNA differed between patients with intermittent (detectable viremia at 50% of the time points [n = 47]) and those with persistent (detectable viremia at >50% of the time points [n = 21]) low-level viremia, but comparable results were found for these groups (data not shown). (3) The random-effects model used in the present study had some limitations. Since the investigated virological and immunological variables were measured concurrently, cause-and-effect inferences simply cannot be made. Thus, the finding that, for example, a higher level of T cell activation was associated with a reduction in CD4 gain also could reflect that persistent immunodeficiency with a low CD4 cell count results in persistent T cell activation, owing to the presence of concurrent infections. Furthermore, since the patients had been receiving HAART for varying times at inclusion, the estimated changes during follow-up represent an mean of early (larger) and late (smaller) changes in the investigated parameters. Finally, although several markers were investigated, P values were not adjusted for multiple comparisons, which further strengthens the need for confirmation of the results in studies of independent cohorts.
In summary, the present study found that even low-level viremia was associated with immunological and virological consequences in patients receiving HAART. Although low-level viremia did not result in obvious treatment failure, it was associated with reductions in CD4 gain and with T cell activation. Furthermore, CD4 T cell activation predicted subsequent increases in HIV replication even during HAART. Finally, a higher proviral DNA level was associated with CD4 T cell activation and reductions in naive CD4 and CD8 cells, indicating an immunological effect of proviral DNA even in patients with sustained virological suppression. It is not known to what degree persistent immune activation and immunological dysregulation accounts for the emergence of, for example, metabolic abnormalities and putative HIV-related dementia in patients receiving HAART. However, it is tempting to speculate whether further reductions in HIV replication and T cell activation through intensification of HAART may have a positive effect on immune reconstitution and certain long-term complications.
Acknowledgments
We thank the patients for their participation in the present study and Thomas Scheike (Department of Biostatistics, University of Copenhagen, Copenhagen) for his invaluable help with the statistical models. We acknowledge Gitte Grauert for help with the flow-cytometric analyses.
References
1. Hammer SM, Squires KE, Hughes MD, et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. AIDS Clinical Trials Group 320 Study Team. N Engl J Med 1997; 337:72533. First citation in article
2. Ledergerber B, Egger M, Opravil M, et al. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. Swiss HIV Cohort Study. Lancet 1999; 353:8638. First citation in article
3. Paredes R, Mocroft A, Kirk O, et al. Predictors of virological success and ensuing failure in HIV-positive patients starting highly active antiretroviral therapy in Europe: results from the EuroSIDA study. Arch Intern Med 2000; 160:112332. First citation in article
4. Mocroft A, Gill MJ, Davidson W, Phillips AN. Predictors of a viral response and subsequent virological treatment failure in patients with HIV starting a protease inhibitor. AIDS 1998; 12:21617. First citation in article
5. Deeks SG, Barbour JD, Martin JN, Swanson MS, Grant RM. Sustained CD4+ T cell response after virologic failure of protease inhibitorbased regimens in patients with human immunodeficiency virus infection. J Infect Dis 2000; 181:94653. First citation in article
6. Mocroft A, Ruiz L, Reiss P, et al. Virological rebound after suppression on highly active antiretroviral therapy. AIDS 2003; 17:174151. First citation in article
7. Karlsson AC, Younger SR, Martin JN, et al. Immunologic and virologic evolution during periods of intermittent and persistent low-level viremia. AIDS 2004; 18:9819. First citation in article
8. Kaufmann D, Pantaleo G, Sudre P, Telenti A. CD4-cell count in HIV-1infected individuals remaining viraemic with highly active antiretroviral therapy (HAART). Swiss HIV Cohort Study. Lancet 1998; 351:7234. First citation in article
9. Goetz MB. Discordance between virological, immunologic, and clinical outcomes of therapy with protease inhibitors among human immunodeficiency virusinfected patients. Clin Infect Dis 1999; 29:14314. First citation in article
10. Havlir DV, Bassett R, Levitan D, et al. Prevalence and predictive value of intermittent viremia with combination HIV therapy. JAMA 2001; 286:1719. First citation in article
11. Ramratnam B, Mittler JE, Zhang L, et al. The decay of the latent reservoir of replication-competent HIV-1 is inversely correlated with the extent of residual viral replication during prolonged anti-retroviral therapy. Nat Med 2000; 6:825. First citation in article
12. Martinez-Picado J, DePasquale MP, Kartsonis N, et al. Antiretroviral resistance during successful therapy of HIV type 1 infection. Proc Natl Acad Sci USA 2000; 97:1094853. First citation in article
13. Cohen Stuart JW, Wensing AM, Kovacs C, et al. Transient relapses ("blips") of plasma HIV RNA levels during HAART are associated with drug resistance. J Acquir Immune Defic Syndr 2001; 28:10513. First citation in article
14. Greub G, Cozzi-Lepri A, Ledergerber B, et al. Intermittent and sustained low-level HIV viral rebound in patients receiving potent antiretroviral therapy. AIDS 2002; 16:19679. First citation in article
15. Wellons MF, Ottinger JS, Weinhold KJ, et al. Immunologic profile of human immunodeficiency virusinfected patients during viral remission and relapse on antiretroviral therapy. J Infect Dis 2001; 183:15225. First citation in article
16. Katzenstein TL, Ullum H, Roge BT, et al. Virological and immunological profiles among patients with undetectable viral load followed prospectively for 24 months. HIV Med 2003; 4:5361. First citation in article
17. De Rosa SC, Herzenberg LA, Herzenberg LA, Roederer M. 11-color, 13-parameter flow cytometry: identification of human naive T cells by phenotype, function, and T-cell receptor diversity. Nat Med 2001; 7:2458. First citation in article
18. Appay V, Zaunders JJ, Papagno L, et al. Characterization of CD4+ CTLs ex vivo. J Immunol 2002; 168:59548. First citation in article
19. Appay V, Rowland-Jones SL. Premature aging of the immune system: the cause of AIDS Trends Immunol 2002; 23:5805. First citation in article
20. Bisset LR, Lung TL, Kaelin M, Ludwig E, Dubs RW. Reference values for peripheral blood lymphocyte phenotypes applicable to the healthy adult population in Switzerland. Eur J Haematol 2004; 72:20312. First citation in article
21. Littell RC, Henry PR, Ammerman CB. Statistical analysis of repeated measures data using SAS procedures. J Anim Sci 1998; 76:121631. First citation in article
22. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986; 73:1322. First citation in article
23. Egger M, May M, Chene G, et al. Prognosis of HIV-1infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet 2002; 360:11929. First citation in article
24. Chene G, Sterne JA, May M, et al. Prognostic importance of initial response in HIV-1 infected patients starting potent antiretroviral therapy: analysis of prospective studies. Lancet 2003; 362:67986. First citation in article
25. Furtado MR, Callaway DS, Phair JP, et al. Persistence of HIV-1 transcription in peripheral-blood mononuclear cells in patients receiving potent antiretroviral therapy. N Engl J Med 1999; 340:161422. First citation in article
26. Sharkey ME, Teo I, Greenough T, et al. Persistence of episomal HIV-1 infection intermediates in patients on highly active anti-retroviral therapy. Nat Med 2000; 6:7681. First citation in article
27. Havlir DV, Strain MC, Clerici M, et al. Productive infection maintains a dynamic steady state of residual viremia in human immunodeficiency virus type 1infected persons treated with suppressive antiretroviral therapy for five years. J Virol 2003; 77:112129. First citation in article
28. Ramratnam B, Ribeiro R, He T, et al. Intensification of antiretroviral therapy accelerates the decay of the HIV-1 latent reservoir and decreases, but does not eliminate, ongoing virus replication. J Acquir Immune Defic Syndr 2004; 35:337. First citation in article
29. Hunt PW, Martin JN, Sinclair E, et al. T cell activation is associated with lower CD4+ T cell gains in human immunodeficiency virusinfected patients with sustained viral suppression during antiretroviral therapy. J Infect Dis 2003; 187:153443. First citation in article
30. Grossman Z, Feinberg MB, Paul WE. Multiple modes of cellular activation and virus transmission in HIV infection: a role for chronically and latently infected cells in sustaining viral replication. Proc Natl Acad Sci USA 1998; 95:63149. First citation in article
31. Hazenberg MD, Hamann D, Schuitemaker H, Miedema F. T cell depletion in HIV-1 infection: how CD4+ T cells go out of stock. Nat Immunol 2000; 1:2859. First citation in article
32. Grossman Z, Meier-Schellersheim M, Sousa AE, Victorino RM, Paul WE. CD4+ T-cell depletion in HIV infection: are we closer to understanding the cause Nat Med 2002; 8:31923. First citation in article
33. Ribeiro RM, Mohri H, Ho DD, Perelson AS. In vivo dynamics of T cell activation, proliferation, and death in HIV-1 infection: why are CD4+ but not CD8+ T cells depleted Proc Natl Acad Sci USA 2002; 99:155727. First citation in article
34. Deeks SG, Walker BD. The immune response to AIDS virus infection: good, bad, or both J Clin Invest 2004; 113:80810. First citation in article