Despite substantial reductions in mortality and morbidity among HIV-infected persons receiving highly active antiretroviral therapy (HAART), there continue to be individuals for whom therapy fails to suppress viral replication.1 To characterize and better understand those who do not respond to HAART, numerous predictors for virologic treatment failure have been proposed. Studies focusing on initial virologic response have found that younger age, African American race, poor adherence to treatment, missed visits, lower baseline CD4 counts, and higher baseline HIV RNA levels were associated with initial virologic treatment failure.2-9
Virologic failure (VF) is a concern not only when monitoring the initial response to therapy but also once viral suppression (VS) on therapy has been established. At approximately 2 years after initial suppression, cumulative rates of VF can range from 20% to 40%.4,8,10-12 In addition to many of the same predictors found for initial failure to suppress HIV RNA after starting HAART, factors associated with postsuppression VF include previous antiretroviral (ART) exposure, date of starting HAART, the duration of VS, presence of oral lesions, and change in treatment or prior treatment failure.6,7,10-15 A recently reported scoring algorithm to predict VF after suppression included a subset of the risk factors listed previously.14 Although the rates of VF predicted by the algorithm were very close to the observed rates, some settings may not have the capacity to identify all of these factors, and there are likely other predictive factors that are not yet identified.
One potential predictor that is readily available in most clinical settings, yet has not often been included in previous studies of VF, is the total CD8 cell count. The relationship between CD8 response and HIV outcomes is not completely understood but seems to be multifaceted. The CD8 response can be viewed in 3 different contexts: HIV-specific CD8 cells, the activation of CD8 subsets, and the total CD8 count. Associations with HIV outcomes vary by the type of CD8 response being considered.
Up to 20% of circulating CD8 cells in those with untreated chronic HIV infection are HIV specific.16 HIV-specific CD8 cells play an important role in the control of HIV viremia. Depletion of the CD8-specific response results in the loss of virologic control in animal models of HIV,17 and in humans, CD8 cells have been shown to suppress HIV replication through both cytolytic and noncytolytic mechanisms.18-22 A soluble factor involved in the noncytolytic antiviral response has been elaborated from CD8 cells from long-term survivors with asymptomatic chronic HIV infection,23 and better clinical outcomes have been noted with increased anti-HIV CD8 cell activity.24-29
However, a hyperdynamic or overstimulated CD8 immune response, reflected by the activation of CD8 subsets and elevated total CD8 counts, may accelerate immune dysfunction and certain disease processes. Current evidence suggests a correlation between CD4 loss and hyperdynamic CD8 response,16,30 although CD4 losses may also occur via alternate mechanisms.31 The expression of some CD8 subsets, particularly CD38+ CD8+, has been linked to HIV disease progression.32-35 Multiple studies have assessed the relationship between the activation of CD8 subsets and response to HAART.36-40 Motivated by the need to find markers for virological treatment failure in resource-limited settings lacking routine viral load capability, several studies have investigated the potential use of CD38 expression.38-40 Although significant correlations between increased CD38 expression and VF were observed, poor sensitivity and specificity of this marker precluded wider application.
Many studies have demonstrated the widespread elevation of total CD8 cell counts in untreated HIV infection,16,30,41 and some have reported a decrease in CD8 cells on HAART initiation.42 Observed increases in circulating CD8 cells in HIV-positive individuals have been attributed to declines in CD4 cells and subsequent CD8 compensation known as “blind T-cell homeostasis.”43 Some early studies conducted before HAART became available showed significant associations between elevated baseline total CD8 counts and progression to AIDS,44-46 whereas others failed to find such associations with AIDS47 or CD4 outcomes.48 In light of these conflicting results and as a greater understanding of the complexity of the CD8 response evolved, more studies focused on CD8 subsets. Thus, few studies have evaluated total CD8 counts in the context of virologic response to HAART.
If elevated total CD8 counts were a pathologic (hyperdynamic) as opposed to a benign (blind T-cell homeostasis) process, the elevated CD8 count could be a potential marker for adverse events, such as virologic treatment failure. This concept has not yet been adequately explored. In this study, we examine whether elevated total CD8 counts are predictive of virologic treatment failure. Because adherence has been shown to be an important factor in viral response, we focused on individuals who demonstrated initial VS and therefore may be more compliant with therapy.
Study participants were enrolled in the US Military HIV Natural History Study (NHS), an ongoing multisite prospective cohort study of consenting HIV-infected Department of Defense beneficiaries.49,50 Since its inception in 1986, the NHS has enrolled >5000 participants. Medical history information and routine laboratory measures including CD4 cell counts, CD8 cell counts, and HIV RNA levels were collected at semiannual study visits. The central Infectious Diseases Institutional Review Board approved this study.
This study selected NHS participants who initiated HAART between 1996 and 2008 and demonstrated initial VS (defined below). Further inclusion criteria required at least 2 HIV RNA measures after VS and availability of a baseline CD8 cell count (Fig. 1). Comparisons between the 817 participants included and the 1559 participants excluded due to the lack of VS, insufficient follow-up data, or lack of CD8 count at baseline showed that the included group was more often white, initiated HAART in more recent calendar years (2000-2008 vs. 1996-1999), had a shorter time from HIV diagnosis to HAART initiation, and were generally healthier at the start of HAART (greater CD4 counts, lower HIV RNA levels). A lower percentage of those included used ART before HAART, and a lower percentage had any clinical AIDS events before HAART compared with the excluded participants. Classes of drugs used in the first HAART regimen also differed significantly by inclusion/exclusion group, reflecting the era in which HAART was initiated. In addition, the participants included in the analysis had significantly higher CD8 counts at the start of HAART compared with those who were excluded (difference in medians = 35 cells per cubic millimeter, P = 0.02).
VS was defined as both the 6- and 12-month post-HAART HIV RNA levels <400 copies per milliliter; this was the limit of detection for early assays, and we opted to standardize VS across time periods. Because the participants were not at risk for VF in this study until they demonstrated VS, baseline was defined as 12 months after HAART initiation. VF was defined as a confirmed HIV RNA measurement ≥400 copies per milliliter with the date of VF defined as the date of the first HIV RNA level ≥400 copies per milliliter after VS. CD8 counts ≥1200 cells per cubic millimeter, the upper limit of normal according to clinical laboratory reference values,51 were considered elevated. HAART was defined as previously described.9 ART refers to regimens not defined as HAART and typically indicated mono- or dual-therapy with nucleoside reverse transcriptase inhibitors. Variables defined at the time of starting HAART used the latest value occurring in a window of 6 months before HAART initiation. Baseline variables were those measured closest to the baseline date (±3 months).
CD4 cell counts, CD8 cell counts, and HIV RNA levels nearest to each 6-month anniversary (±3 months) from HAART initiation were selected for analysis. Follow-up data were censored at the last 6-month HIV RNA measurement before 3 consecutive missing HIV RNA values. Baseline group comparisons used chi-square tests for categorical variables and Wilcoxon 2-sample tests for continuous variables. Incidence rates of VF were calculated per 100 person-years (pyrs) of follow-up with exact Poisson 95% confidence intervals (CIs). Kaplan-Meier curves were used to estimate the cumulative proportion of the participants remaining free of VF during follow-up both overall and by era in which HAART was initiated. Cox proportional hazard models were used to estimate the associations between CD8 counts and risk of VF. Separate models were used to examine the effects of baseline and current CD8 counts. Baseline CD8 counts were categorized using quartiles at the start of HAART.
CD8 counts measured during analysis time were parameterized using 2 different time-updated CD8 covariates, each examined in separate models. The first time-updated covariate calculated the proportion of previous visits (starting at baseline) with elevated CD8 counts. Because CD8 counts were infrequently elevated, choices for the parameterization of this time-updated covariate were limited. Therefore, the proportion was categorized according to the upper quartile to compare those with >20% of prior visits with elevated CD8 counts to those with ≤20% of prior visits with elevated CD8 counts. The second time-updated CD8 covariate described change in CD8 counts from HAART initiation to each lagged visit and was categorized as an increase, decrease or stay the same, or pattern unknown due to missing data. All time-updated values were carried forward no further than 18 months, after which they were set to missing.
Initial models were stratified by era in which HAART was initiated (1996-1999 vs. 2000-2008), allowing a separate baseline hazard for each era while estimating a common CD8 effect. These eras were chosen a priori to reflect differences in HIV treatment and care in the first few years after HAART became available and thereafter. Additional models included an interaction between HAART start era and CD8 count, to estimate separate hazard ratios (HRs) for CD8 count by HAART start era. Adjustment variables considered for inclusion in multivariate models included baseline age, gender, race, ART before HAART, clinical AIDS events before baseline, years from HIV diagnosis to baseline, HIV RNA at HAART start, and CD4 cell count. In general, final multivariate models excluded covariates that were neither significant (P < 0.05) nor impacted the CD8 estimate by >20%. However, because of potential colinearity between CD8 and CD4 counts, models were constructed with and without CD4 counts included. Two sensitivity analyses were conducted: (1) time-updated values were carried no further than 12 months and (2) a percent increase greater than reported assay variability52 was modeled rather than an absolute increase in CD8 counts. SAS software, version 9.2 (SAS Institute) was used for all analyses.
Of the 817 eligible participants, 445 (54%) initiated HAART between 1996 and 1999, whereas 372 (46%) initiated HAART between 2000 and 2008. Baseline participant characteristics by era in which HAART was initiated and overall are shown in Table 1. Those starting HAART in the more recent era were significantly younger, less likely white, had a lower CD4 count and a higher HIV RNA level at the start of HAART, and a lower baseline CD8 count than those starting HAART in the 3 years after it became available in 1996. In addition, the more recent HAART era group had a significantly shorter time from HIV diagnosis to baseline, lower proportions of participants using ART before HAART, and lower proportions with any clinical AIDS events before baseline. Classes of drugs used in the initial HAART regimen also varied according to HAART start era.
CD8 counts decreased significantly from HAART initiation to baseline (1 year later) by a median of 61 cells per cubic millimeter (95% CI: 45 to 87 cells per cubic millimeter decrease). Greater decreases were observed among those initiating HAART in 2000-2008 (median decrease of 109 cells per cubic millimeter, 95% CI: 72 to 167 cells per cubic millimeter decrease) compared with those initiating HAART in 1996-1999 (median decrease of 32 cells per cubic millimeter, 95% CI: 1 to 55 cells per cubic millimeter decrease, P < 0.001).
Rates of Virologic Failure
The participants were followed for a median of 4 years [interquartile range (IQR): 2-7 years]. At 2 years after confirmed VS, Kaplan-Meier estimates of VF were 12% (95% CI: 10%, 15%) overall, 15% (95% CI: 12%, 18%) among those initiating HAART in 1996-1999 and 9% (95% CI 6%, 13%) among those initiating HAART in 2000-2008. During the entire follow-up period, a total of 216 participants had VF, giving a rate of 5.6 per 100 pyrs (95% CI: 4.9 to 6.4). VF rates varied by both era in which HAART was initiated and baseline CD8 count (Fig. 2). Rates were higher among those initiating HAART in 1996-1999 compared with those initiating HAART in 2000-2008 (unadjusted HR = 2.68, 95% CI: 1.90 to 3.78). Among those initiating HAART in 1996-1999, VF rates among categories of baseline CD8 count were similar. In contrast, VF rates were the highest for elevated baseline CD8 counts among those initiating HAART in 2000-2008.
Baseline CD8 Count and Risk of Virologic Failure
Corresponding to the patterns observed in VF rates, the relationship between baseline CD8 count and risk of VF varied by HAART era (Table 2). Among those initiating HAART in 1996-1999, baseline CD8 count was not significantly associated with the risk of VF, but among those initiating HAART in 2000-2008, the participants with an elevated (≥1200 cells per cubic millimeter) baseline CD8 count had a significantly greater risk of VF compared with those with a baseline CD8 count of 600 cells per cubic millimeter or lower (adjusted HR = 2.68, 95% CI: 1.13 to 6.35). Recent HAART era participants with baseline CD8 counts in the middle categories (601-849, 850-1199 cells per cubic millimeter) were not at a significantly greater risk of VF compared with those with a baseline CD8 count of ≤600 cells per cubic millimeter (Table 2).
Current CD8 Count and Risk of Virologic Failure
Of the 216 participants with VF, 67 (31%) had elevated CD8 counts at >20% of visits before the VF. In comparison, 121 (20%) of the 601 participants without a VF had elevated CD8 counts at >20% of visits before their censoring date. In accordance with this observation, Cox models including a time-updated CD8 covariate showed that participants with >20% of prior 6-month visits with elevated CD8 counts had a greater risk of VF than those with ≤20% of prior visits with elevated CD8 counts (adjusted HR = 1.53, 95% CI: 1.14 to 2.06, Table 3). This increased risk associated with elevated CD8 counts was greater among those initiating HAART in 2000-2008 (adjusted HR = 2.70, 95% CI: 1.42 to 5.13) than in those initiating HAART in 1996-1999 (adjusted HR = 1.34, 95% CI: 0.96 to 1.86), although this difference was marginally significant (P = 0.06 for interaction).
For those with VF, CD8 counts increased from the start of HAART to the visit immediately preceding the VF by a median of 51 cells per cubic millimeter (IQR: −237 to +241 cells per cubic millimeter), whereas for those without a VF, CD8 counts decreased from the start of HAART to the visit before censoring by a median of 108 cells per cubic millimeter (IQR: −382 to +102 cells per cubic millimeter). In Cox models, increases in CD8 count from the start of HAART were significantly associated with VF when compared with CD8 counts that decreased or remained the same (adjusted HR = 1.59, 95% CI: 1.19 to 2.13, Table 3). This comparison did not vary significantly by HAART start era (P = 0.68 for interaction). In sensitivity analysis, those with a >4% increase in CD8 counts also had significantly greater risk of VF compared with those with ≤4% increases (adjusted HR = 1.71, 95% CI: 1.28 to 2.28). All adjusted CD8 model estimates were similar when CD4 was not included in the multivariate model and when time-updated values were carried forward for only 12 months (data not shown).
In a cohort of HIV-infected military beneficiaries who demonstrated initial VS after the initiation of HAART, we found that elevated total CD8 counts were associated with a greater risk of future VF. This association persisted using 3 separate evaluations of CD8 counts: one measured at the time of confirmed initial VS (defined as 12 months post-HAART in this study), and 2 different metrics for CD8 measured after initial VS: proportion of prior visits with elevation and change in CD8 counts from HAART initiation.
Few studies have explored the total CD8 count as a potential predictor for virologic treatment failure. One study found no significant association between total CD8 count and categories of suppressed viral replication versus continuing viral replication.40 However, this study was cross-sectional and thus not designed to capture cumulative effects of total CD8 counts. In our study, we used time-updated covariates to relate the cumulative history of elevated CD8 counts (from the time of initial suppression) to the probability of future VF. We found that baseline and serial elevation in the total CD8 count and an increase in CD8 counts from HAART initiation to the current visit was associated with an increased risk of VF. Possible explanations for these associations include the adverse effects of a hyperdynamic or overstimulated immune response, which may be reflected in the total CD8 count. Alternatively, elevated CD8 cells may have resulted from low-level viremia, which is associated with increased risk of VF.53 In addition, coinfections may have been responsible for both elevated CD8 counts and increased risk of failure, although this is temporally less plausible.
For both CD8 count measured at the time of initial VS and CD8 count measured during follow-up (when evaluating the proportion of CD8 counts that were elevated), we found stronger associations with VF when HAART was initiated in the more recent era of 2000-2008, rather than that in 1996-1999. This retrospective study was not designed to address the basis for this differential effect, but some explanations are consistent with our data. First, the effect may relate to the duration of HIV infection at the time of HAART initiation. The 2000-2008 HAART group had been infected with HIV for a much shorter time compared with those initiating HAART in 1996-1999. A recent study showed a sharper decline in CD8 counts after the initiation of HAART among those with early versus chronic HIV infection.54 In the first year after starting HAART, we also observed a steeper CD8 count decline for the recent HAART era group with early HAART initiation compared with the 1996-1999 group that had been infected longer before starting HAART. Immune system recovery may be greater for the treatment initiated earlier in the course of infection; poor response among those initiating treatment later in the course of HIV infection may obscure a relationship between elevated CD8 counts and VF. A second possible explanation relates to toxicity of HAART regimens prescribed in the 1990s. Higher rates of VF among those initiating HAART in 1996-1999 may more often be linked to toxicity-related discontinuation of therapy, rather than to intermittent noncompliance. In this scenario, low-level viremia leading to elevated CD8 counts and VF may be less plausible.
This study had several limitations. The participants in this analysis were a select group that showed initial response to HAART by having <400 copies per milliliter of HIV RNA at both 6 and 12 months after initiation. This group differed significantly from the excluded participants in many respects, including CD8 counts at HAART start. However, the purpose of these selection criteria was to identify and restrict the analysis to the participants who were more likely adherent to treatment; data regarding adherence was not available for the entire study period, so selection was used to attempt to control for the important role of this variable in virologic treatment failure. A second limitation is that our categorization of the proportion of previous visits with elevated CD8 counts was fairly arbitrary. However, additional evidence for the effect of elevated CD8 levels on virologic treatment failure was found using baseline CD8 counts and using the change in CD8 counts from HAART start to each visit. Confirmation of these findings in a larger cohort is advisable.
The rate of VF observed in this study (5.6 per 100 pyrs overall or 12% at 2 years) was lower than what has been reported in earlier years.4,8,10-12 However, rates of treatment failure in general have been decreasing in recent years,55-57 and a more current estimate of VF58 was closer to what we have observed in our cohort, which includes the participants initiating HAART as late as 2008. Despite this decline in failure rates, VF remains an important concern when monitoring HIV-infected patients on lifelong treatment. Thus, clinically relevant tools that are readily available to predict treatment failure are needed. In this study, we have highlighted the potential of the total CD8 count as one such tool. Although the CD8 count alone may not have adequate sensitivity to predict VF, its use in combination with the battery of existing predictors for VF could make a significant contribution to prediction algorithms. Because some of the other suggested predictors involved characteristics that are not common in our study (prior VF, suboptimal adherence, or suppression <12 months), an elevated CD8 count may be one of the few indicators of future VF among virally suppressed individuals who may not otherwise be viewed as high risk for failure. By identifying those patients at increased risk of VF, targeted efforts to confirm treatment adherence or increase the frequency of monitoring could be implemented with the goal of preventing VF among those who are currently maintaining VS.
Braden Hale conceived the idea for this analysis. Jason Okulicz, Amy Weintrob, Brian Agan, Nancy Crum-Cianflone, Anuradha Ganesan, Tomas Ferguson, and Braden Hale helped implement the study, collect data, and oversee the individual participating sites. Elizabeth Krantz performed the statistical analyses and drafted the manuscript. Katherine Huppler Hullsiek provided valuable input regarding the statistical analyses. All the authors provided critical reads that helped shape the manuscript.
Additional members of the IDCRP HIV/STI Working Group include Susan Banks, Mary Bavaro, Helen Chun, Cathy Decker, Lynn Eberly, Connor Eggleston, Susan Fraser, Heather Hairston, Josh Hartzell, Arthur Johnson, Michael Landrum, Alan Lifson, Michelle Linfesty, Grace Macalino, Jason Maguire, Scott Merritt, Robert O'Connell, Sheila Peel, Michael Polis, John Powers, Timothy Whitman, Glenn Wortmann, and Michael Zapor.
1. Collaboration of Observational HIV Epidemiological Research Europe (COHERE) study group. Response to combination antiretroviral therapy: variation by age. AIDS
2. Deeks SG, Hecht FM, Swanson M, et al. HIV RNA and CD4 cell count response to protease inhibitor therapy in an urban AIDS clinic: response to both initial and salvage therapy. AIDS
3. Hartzell JD, Spooner K, Howard R, et al. Race and mental health diagnosis are risk factors for highly active antiretroviral therapy in a military cohort despite equal access to care. J Acquir Immune Defic Syndr
4. Lederberger 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. Lancet
5. Lucas GM, Chaisson RE, Moore RD. Highly active antiretroviral therapy in a large urban clinic: risk factors for virologic failure and adverse drug reactions. Ann Intern Med
6. Mocroft A, Gill MJ, Davidson W, et al. Predictors of a viral response and subsequent virological treatment failure in patients starting a protease inhibitor. AIDS
7. 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
8. Phillips AN, Staszewski S, Weber R, et al. HIV viral load response to antiretroviral therapy according to the baseline CD4 cell count and viral load. JAMA
9. Weintrob AC, Grandits GA, Agan BK, et al. Virologic response differences between African Americans and European Americans initiating highly active antiretroviral therapy with equal access to care. J Acquir Immune Defic Syndr
10. Le Moing V, Chene G, Carrieri MP, et al. Predictors of virological rebound in HIV-1 infected patients initiating a protease inhibitor-containing regimen. AIDS
11. Mocroft A, Ruiz L, Reiss P, et al. Virological rebound after suppression on highly active antiretroviral therapy. AIDS
12. Phillips AN, Staszewski S, Lampe F, et al. Human immunodeficiency virus rebound after suppression to <400 copies/mL during initial highly active antiretroviral therapy regimens, according to prior nucleoside experience and duration of suppression. Infect Dis
13. Robbins GK, Daniels B, Zheng H, et al. Predictors of antiretroviral treatment failure in an urban HIV clinic. J Acquir Immune Defic Syndr
14. Robbins GK, Johnson KL, Chang Y, et al. Predicting virologic failure in an HIV clinic. Clin Infect Dis
15. Ramirez-Amador V, Ponce-de-Leon S, Anaya-Saavedra G, et al. Oral lesions as clinical markers of highly active antiretroviral therapy failure: a nested case-control study in Mexico City. Clin Infect Dis
16. Appay V, Sauce D. Immune activation and inflammation in HIV-1 infection: causes and consequences. J Pathol
17. Silvestri G. Naturally SIV-infected sooty mangabeys: are we closer to understanding why they do not develop AIDS? J Med Primatol
18. Saksena NK, Wu JQ, Potter SJ, et al. Human immunodeficiency virus interactions with CD8+ T Lymphocytes. Curr HIV Res
19. Walker CM, Erickson AL, Hsueh FC, et al. Inhibition of human immunodeficiency virus replication in acutely infected CD4+ cells by CD8+ cells involves a noncytotoxic mechanism. J Virol
20. Levy JA, Mackewicz CE, Barker E. Controlling HIV pathogenesis: the role of the noncytotoxic anti-HIV response of CD8+ T cells. Immunol Today
21. Levy JA. The importance of the innate immune system in controlling HIV infection and disease. Trends Immunol
22. Levy JA. The search for the CD8+ cell anti-HIV factor (CAF). Trends Immunol
23. Ferbas J, Kaplan AH, Hausner MA, et al. Virus burden in long-term survivors of Human Immunodeficiency (HIV) infection is a determinant of anti-HIV CD8+ lymphocyte activity. Infect Dis
24. Gomez AM, Smaill FM, Rosenthal KL. Inhibition of HIV replication by CD8+ T cells correlates with CD4 counts and clinical stage of disease. Clin Exp Immunol
25. Mackewicz CE, Ortega HW, Levy JA. CD8+ cell anti-HIV activity correlates with the clinical state of the individual. J Clin Invest
26. Landay AL, Mackewicz CE, Levy JA. An activated CD8+ T cell phenotype correlates with anti-HIV activity and asymptomatic clinical status. Clin Immunol Immunopathol
27. Barker E, Mackewicz CE, Reyes-Teran G, et al. Virological and immunological features of long-term human immunodeficiency virus-infected individuals who have remained asymptomatic compared with those who have progressed to acquired immunodeficiency syndrome. Blood
28. Castelli JC, Deeks SG, Shiboski S, et al. Relationship of CD8+ T cell noncytotoxic anti-HIV response to CD4+ T cell number in untreated asymptomatic HIV-infected individuals. Blood
29. Killian MS, Johnson C, Teque F, et al. Natural suppression of human immunodeficiency virus type 1 replication is mediated by transitional memory CD8+ T cells. J Virol
30. Hazenberg MD, Otto SA, van Benthem BHB, et al. Persistent immune activation in HIV-1 infection is associated with progression to AIDS. AIDS
31. Doitsh G, Cavrois M, Lassen KG, et al. Abortive HIV infection mediates CD4 T cell depletion and inflammation in human lymphoid tissue. Cell
32. Bofill M, Mocroft A, Lipman M, et al. Increased numbers of primed activated CD8+ CD38+ CD45RO+ T cells predict the decline of CD4+
T cells in HIV-1 infected patients. AIDS
33. Giorgi JV, Liu Z, Hultin LE, et al. Elevated levels of CD38+CD8+ T cells in HIV infection add to the prognostic value of low CD4+ T cell levels: results of 6 years of follow-up. J Acquir Immune Defic Syndr
34. Liu Z, Cumberland WG, Hultin LE, et al. Elevated CD38 antigen expression on CD8+ T cells is a stronger marker for the risk of chronic HIV disease progression to AIDS and death in the multicenter AIDS Cohort Study than CD4+ cell count, soluble immune activation markers, or combinations if HLA-DR and CD38 expression. J Acquir Immune Defic Syndr Hum Retrovirol
35. Perfetto SP, Malone JD, Hawkes C, et al. CD38 expression on cryopreserved CD8+ T cells predicts disease progression. Cytometry
36. Hunt PW, Martin JN, Sinclair E, et al. T cell activation is associated with lower CD4+ T cell gains in human immunodeficiency virus-infected patients with sustained viral suppression during antiretroviral therapy. Infect Dis
37. Benito JM, Lopez M, Lozano S, et al. CD4+ T cell recovery beyond the first year of complete suppression of viral replication during highly active antiretroviral therapy is not influenced by CD8+ T cell activation. Infect Dis
38. Ondoa P, Koblavi-Deme S, Borget MY, et al. Assessment of CD8+ T cell immune activation markers to monitor response to antiretroviral therapy among HIV-1 infected patients in Cote d'Ivoire. Clin Exp Immunol
39. Onlamoon N, Tabprasit S, Suwanagool S, et al. Studies on the potential use of CD38 expression as a marker for the efficacy of anti-retroviral therapy in HIV-1 infected patients in Thailand. Virology
40. Steel A, John L, Shamji MH, et al. CD38 expression on CD8 T cells has a weak association with CD4 T-cell recovery and is a poor marker of viral replication in HIV-1 infected patients on antiretroviral therapy. HIV Med
41. Appay V, Almeida JR, Sauce D, et al. Accelerated immune senescence and HIV-1 infection. Exp Gerontol
42. Hazenberg MD, Stuart JWTC, Otto SA, et al. T-cell division in human immunodeficiency virus (HIV)-1 infection is mainly due to immune activation: a longitudinal analysis in patients before and during highly active antiretroviral therapy (HAART). Blood
43. Adleman LM, Wofsy D. T-cell homeostasis: implications in HIV infection. J Acquir Immune Defic Syndr
44. Anderson RE, Shiboski SC, Royce R, et al. CD8+ T lymphocytes and progression to AIDS in HIV-infected men: some observations. AIDS
45. Phillips AN, Sabin CA, Elford J, et al. CD8 lymphocytes and serum immunoglobulin A levels early in HIV infection as predictors of CD4 lymphocyte depletion during 8 years of follow-up. AIDS
46. Schechter MT, Craib KJP, Le TN, et al. Susceptibility to AIDS progression appears early in HIV infection. AIDS
47. Fahey JL, Taylor JMG, Detels R, et al. The prognostic value of cellular and serologic markers in infection with human immunodeficiency virus type 1. N Engl J Med
48. Chevret S, Roquin H, Ganne P, et al. Prognostic value of an elevated CD8 lymphocyte count in HIV infection. Results of a prospective study of 152 asymptomatic HIV-positive individuals. AIDS
49. Weintrob AC, Fieberg AM, Agan BK, et al. Increasing age at HIV seroconversion from 18 to 40 years is associated with favorable virologic and immunologic responses to HAART. J Acquir Immune Defic Syndr
50. Marconi VC, Grandits GA, Weintrob AC, et al. Outcomes of highly active antiretroviral therapy in the context of universal access to healthcare: the U.S. Military HIV Natural History Study. AIDS Res Ther
51. Wolfe WH, Miner JC, Michalek JE. Immunological parameters in current and former US Air Force personnel. Vaccine
52. Backteman K, Ernerudh J. Biological and methodological variation of lymphocyte subsets in blood of human adults. J Immunol Meth
53. Aldous JL, Haubrich RH. Defining treatment failure in resource-rich settings. Curr Opin HIV AIDS
54. Moir S, Buckner CM, Ho J, et al. B cells in early and chronic HIV infection: evidence for preservation of immune function associated with early initiation of antiretroviral therapy. Blood
. 2010; pre-published online.
55. Hull MW, Lima VD, Hogg RS, et al. Epidemiology of treatment failure: a focus on recent trends. Curr Opin HIV AIDS
56. Vo TTN, Ledergerber B, Keiser O, et al. Durability and outcome of initial antiretroviral treatments received during 2000-2005 by patients in the Swiss HIV Cohort study. Infect Dis
57. Reekie J, Mocroft A, Sambatakou H, et al. Does less frequent routine monitoring of patients on a stable, fully suppressed cART regimen lead to an increased risk of treatment failure? AIDS
58. Smith CJ, Phillips AN, Hill T, et al. The rate of viral rebound after attainment of an HIV load <50 copies/mL according to specific antiretroviral drugs in use: results from a multicenter cohort study. Infect Dis