Rapid Suppression of HIV-RNA Is Associated with Improved Control of Immune Activation in Mozambican Adults Initiating Antiretroviral Therapy with Low CD4 Counts

Abstract

The rapidity of HIV-RNA suppression after initiation of combined antiretroviral therapy (cART) may impact immune reconstitution in developing countries, where patients initiate cART at low CD4 T cell counts. One hundred and thirty-five HIV-1 Mozambican adults initiating cART were prospectively followed over 16 months within a larger observational study. Plasma HIV-RNA, CD4 counts, and CD8 T cell activation were monitored at the pre-cART visit and at 4, 10, and 16 months during cART. Of the 89 patients with available HIV-RNA data at pre-cART and 4 and 10 months post-cART, 68% (60/89) suppressed HIV-RNA at 4 months and were defined as “early virological controllers”(EC). Twenty of the 29 remaining patients who did not control HIV-RNA at 4 months did so at 10 months and were classified as “late virological controllers”(LC). Nine (10%) patients did not control HIV-RNA at either time point. Both initiating an EFV-containing cART regimen and having pre-cART tuberculosis were significantly associated with early HIV-RNA suppression if locked into a multivariate model [EFV OR: 13.6 (95% CI 1.7; 108.1) p = 0.014) tuberculosis OR: 11.0 (95% CI 1.4; 87.9) p = 0.024]. EC demonstrated significantly lower median activated CD8 T cells at 4, 10, and 16 months post-cART than did LC. Approximately 63% (12/19) of LC experienced reappearance of detectable HIV-RNA at 6 months postcontrol as compared to 15% (2/60) of EC (p = 0.001). This study suggests that rapid suppression of HIV-RNA may lead to a lower rate of reappearance of HIV-RNA, which could impact CD8 T cell activation levels in patients initiating cART at low CD4 counts.

Introduction

C ombined antiretroviral therapy (cART) has been associated with dramatic reductions in morbidity and mortality in HIV-1-infected patients.^1,2 cART leads to a rapid reduction in plasma HIV-RNA levels and to an increase in peripheral CD4 T cell counts.^3
–5 However, the effective administration of cART faces numerous obstacles in resource-poor countries. Patients usually start at late phase AIDS with very low CD4 counts, which has been shown to lead to poorer clinical outcome.⁶ Many HIV-infected individuals are coinfected with tuberculosis and thus simultaneously initiate cART and tuberculosis therapy, along with cotrimoxazol prophylaxis, increasing the risk for drug interactions and adverse effects. Most resource-poor settings recommend the use of two nucleoside reverse transcriptase inhibitors (NRTIs) combined with a nonnucleoside reverse transcriptase inhibitor (NNRTI), which may be nevirapine (NVP) or efavirenz (EFV).⁷ However, most countries in sub-Saharan Africa use an NVP-based regimen for first line therapy because it is less expensive and available in generic fixed-dose combination regimens. Protease inhibitors are recommended as second line but are frequently unavailable. In addition, cART patients are managed in the absence of HIV viral load monitoring, which may lead to longer delays in detecting virological and immunological failure.⁸

It has been shown that initiating cART between 200 and 350 CD4/μl is associated with better immunological control and less mortality than initiating below 200 CD4 counts as is currently largely practiced in sub-Saharan Africa despite World Health Organization (WHO) recommendations in 2009.^6,9 Indeed, the duration of time spent at low CD4⁺ levels even after initiation of cART has been shown to be an important determinant of both AIDS and non-AIDS morbidity and mortality.^10,11 There is thus extensive literature on the benefits of initiating cART at higher CD4 levels and the controversies associated with this in resource-limited countries.^12
–14 However, there is very little information on the impact that time to HIV-RNA suppression may have on immune reconstitution and clinical outcome.

Efficient viral load suppression after cART initiation is characterized by a rapid decay in the first 10 days of treatment followed by a more gradual decline lasting up to several months.^15
–17 Delays in HIV-RNA suppression favor more frequent development of drug resistance.^18,19 However, it is unknown whether slower HIV-RNA suppression can also hamper immune recovery including the cART-associated decrease in generalized immune activation, particularly in patients initiating cART at low CD4 T cell counts. Chronic immune activation is the strongest correlate of HIV/AIDS disease progression and is associated with effective immune reconstitution.^20,21

The objective of this study was to determine predictors of early control of HIV-RNA after initiation of cART and to assess the impact of early viral load control on immunological and virological outcomes.

Materials and Methods

Study area and population

Patients

This study was conducted from April 2006 through November 2008 at the Centro de Investigação em Saúde de Manhiça (CISM)/Manhiça District Hospital (MDH) in Manhiça District, southern Mozambique. The present study is an ancillary study from a prospective surveillance cohort designed to assess the incidence, clinical spectrum, and risk factors of immune reconstitution inflammatory syndrome (IRIS) in this area.²²

The CISM has been conducting continuous demographic surveillance (DS) in the district since 1996, and vital events are regularly collected from 84,000 people living in the Manhiça District.²³ Patients attending the HIV AIDS voluntary counseling and testing (VCT) services at the MDH and meeting criteria for initiation of cART were invited to participate in the study if they were over 18 years old and living in the study area under DS.

Patients initiated cART according to the Mozambican National Antiretroviral Treatment Guidelines, which recommends starting with CD4 cells ≤200 cells/μl irrespective of the clinical stage, WHO stage III with CD4 cell count ≤350 cells/μl, or WHO stage IV irrespective of CD4 cell count. First line treatment consisted of two NRTIs (zidovudine, stavudine, or lamivudine) with an NNRTI (nevirapine or efavirenz).

The study was approved by the institutional review boards of the Mozambican National Bioethics Committee and the Hospital Clinic of Barcelona Ethics Review Committee and written informed consent was obtained from all study participants.

Study follow- up

Patients were followed for 16 months with scheduled visits according to the Mozambican Guidelines, which include a pretreatment assessment and regular assessments at week 2 and months 4, 10, and 16 after cART initiation. At each visit a clinical examination and monitoring HIV-1 viral load, CD4 and CD8 counting, and immune activation were performed.

Patients were required to claim antiretroviral medication monthly and this was used to assess adherence. Missed or late visits were recorded as insufficient adherence.

Laboratory procedures

HIV-1 viral load determinations

Plasma HIV-1 viral load levels were measured using a reverse transcriptase (RT) polymerase chain reaction (PCR) technique (Amplicor, Roche Monitor v1.5) from cryopreserved samples. The lower limit of detection of the assay was 400 copies/ml.

Immunology and hematological determinations

CD4 counting was performed after staining fresh whole blood samples with labeled antibodies: CD4, CD3, CD8, and CD45 in TruCount tubes (Becton Dickinson Biosciences, San Jose, CA). For analysis, CD4 counts were categorized as greater or less than 200 cells/μl.

To access the percentage of activated CD8 T lymphocytes, cell staining was performed with CD3, CD8, CD38, and HLA-DR antibodies. Activated T cells were defined as those CD8 T cells expressing both CD38 and HLA-DR surface markers.

Hematological analyses were performed using a Sysmex KX 21N hematological analyzer (Kobe, Japan).

Malaria and tuberculosis diagnosis

Malaria parasitemia was determined by thick blood smears stained with Giemsa and examined using optical microscopy. Positive slides were defined as those with at least one parasite per 100 high-power microscopy fields.

Tuberculosis cases included confirmed and suspected cases. Confirmed cases were sputum smear positive after staining by Ziehl-Nielsen and examination using optical microscopy or culture positive. Suspected cases that were not laboratory confirmed had a compatible chest x-ray or clinical response to antituberculosis drugs.

Statistical methods and definitions

Proportions for categorical variables were compared using Chi square or Fisher exact tests. Univariate and multivariate logistic regression was used to analyze the risk factors associated with early HIV virus control. Multivariate analyses were performed by a forward-stepwise procedure using p < 0.05 and p > 0.10 from the likelihood ratio test, as enter and remove criteria, respectively, using all variables showing a p-value <0.150 in the univariate logistic regression. Variable locking was used for models with EFV and tuberculosis and Akaike's information criteria (AIC) was used to assess goodness of fit.

Analysis of average monthly percent decline in activated CD8 T cells was calculated using the difference in percent activated cells between fixed time points divided by the number of months between time points. Linear regression of the monthly percent decline of activated CD8 T cells was performed for two time periods: prior to HIV-RNA control using the pre-cART timepoint as baseline and during the period of 6 months following HIV-RNA control using the HIV-RNA control time point as baseline (i.e., early virological controllers at 4 months and late virological controllers at 10 months).

Data entry, validation, and cleaning were performed using Microsoft FOXPRO version 5.0 for Windows (Microsoft Corp., Seattle, WA) and statistical analysis was performed using STATA version 9.0 (StataCorp., College Station, TX).

Anemia was defined as hematocrit below 30% packed cell volume.

Virological response was defined in relation to the scheduled visits recommended by the Mozambican guidelines. Thus, early virological controllers were defined as those patients with HIV-1 viral load <400 copies/ml at month 4 after starting cART; late virological controllers were those with HIV-1 viral load >400 copies/ml at month 4 and <400 copies/ml at month 10.

CDC AIDS status was categorized in a dichotomous variable comparing non-AIDS status including stages with CD4 >200 and/or symptoms not indicative of AIDS = A1, A2, B1, B2 versus AIDS-defining clinical condition and/or CD4 <200 = A3, B3, C1, C2, C3.

Results

Study patient characteristics and time to suppression of HIV-RNA load

Between April 2006 and November 2008, 135 cART-naive HIV-positive adults initiated cART, of whom 129 had baseline laboratory measures. These patients were followed up for 16 months in the overall prospective IRIS study. This study assessed a subset of 89 individuals with viral load and CD4 measures available at the pre-cART, 4-month, and 10-month visit. The study profile is illustrated in Fig. 1. Deaths between pre-cART and 4-month visit were due to opportunistic infections common in advanced AIDS patients. Lost to follow-up was due to migrations.

FIG. 1.

Summary of patient recruitment and classification into early and late viral load (VL) controllers. cART, combined antiretroviral therapy; VL, viral load; LTFU, lost to follow up.

To classify patients into early or late HIV viral load controllers, HIV-RNA was assessed at 4 and 10 months. At 4 months post-cART, 60/89 (68%) patients had undetectable HIV-1 viral load whereas 29/89 (33%) patients had detectable HIV-RNA. At 10 months post-cART, 20/29 (69%) of those who had detectable HIV-RNA at 4 months were able to reduce HIV-1 viral load to below 400 copies/ml. Thus, those patients suppressing viral load at 4 months with viral load data available at 10 months were classified as “early virological controllers (EC)” (n = 60) and those who did not control HIV-RNA at 4 months and did so at 10 months were classified as “late virological controllers (LC)” (n = 20) (Fig. 1). Those excluded from the analysis did not suppress HIV-RNA at either 4 or 10 months (n = 9). The pre-cART characteristics of the population studied are shown in Table 1 and did not differ significantly from the overall population.

Table 1.

Pre-cART Characteristics of the Study Population

Patient characteristics	n/N (%) N = 89
Gender (female)	49/89 (55)
Immunovirology
HIV RNA (>5 log 10 copies/ml)	85/89 (96)
CD4 counts < 200 cells/μl	67/89 (75)
Clinical data
WHO status 3/4	64/89 (72)
AIDS-defining CDC status^a	83/89 (93)
Anemia	21/89 (24)
Tuberculosis	27/89 (30)
Pf^b malaria	5/89 (6)
EFV-containing regimen	31/89 (35)

CDC status categorized as defined in Materials and Methods.

Pf, plasmodium falciparum.

Factors associated with early and late response to cART

Assessment by univariate logistic regression of associations between early or late HIV-RNA control and pre-cART characteristics showed that being an early HIV controller was associated with having an HIV-RNA >5 log₁₀ copies/ml, initiating cART with an EFV-containing regimen, and having baseline tuberculosis (Table 2). As EFV is given to tuberculosis patients in place of NVP, 26/31 (83.9%) of the patients receiving EFV had pre-cART tuberculosis. Other baseline characteristics assessed such as the proportion of individuals with CD4 <200 cells/μl and CDC stage showed no significant association with early or late response to cART.

Table 2.

Univariate Analysis of Baseline Pre-cART Characteristics Associated with Suppression of HIV-RNA at 4 Months (Early Virological Control) after cART Initiation (n = 80)

Baseline pre-cART characteristics	Early virological controllers ^a n/N (%)	Late virological controllers ^a n/N (%)	OR	95% CI	p
Gender
Male	25/60 (42)	10/20 (50)	1.0	—	—
Female	35/60 (58)	10/20 (50)	1.4	5.1–3.8	0.52
HIV-RNA
<5 log 10 copies/ml	1/60 (2)	3/20 (15)	1.0	—	—
>5 log 10 copies/ml	59/60 (98)	17/20 (85)	10.4	1.0–106.6	0.05
CD4 counts
>200 cells/μl	13/60 (22)	8/20 (40)	1.0
<200 cells/μl	47/60 (78)	12/20 (60)	0.4	0.14–1.2	0.11
CD8 activation
<66%	31/60 (52)	14/20 (70)	1.0
>66%	29/60 (48)	6/20 (30)	2.2	0.74–6.4	0.16
CDC status^b
Non-AIDS	2/60 (3)	3/20 (15)	1.0
AIDS-defining	58/60 (97)	17/20 (85)	5.1	0.79–33.2	0.09
Pf malaria
Negative	58/60 (97)	17/20 (85)	1.0
Positive	2/60 (3)	3/20 (15)	0.20	0.03–1.3	0.09
Tuberculosis
Negative	38/60 (63)	19/20 (95)	1.0
Positive	22/60 (37)	1/20 (5)	11.0	1.38–87.9	0.02
EFV-containing regimen
No EFV	35/60 (58)	19/20 (95)	1.0
EFV	25/60 (42)	1/20 (5)	13.5	1.7–108.1	0.01

Early and late virological controllers defined in Materials and Methods. The n = 9 patients who did not control HIV-RNA are not included in the analysis.

CDC status categorized as defined in Materials and Methods.

For multivariate analysis, because of the collinearity between tuberculosis and initiation with an EFV-containing regimen, we fitted two forward stepwise models locking either EFV or tuberculosis and adding the other variables as described in Materials and Methods. Both initiating an EFV-containing cART regimen and having pre-cART tuberculosis were significantly associated with early HIV-RNA suppression if locked into the model [EFV OR: 13.6 (95% CI 1.7; 108.1, p = 0.014) tuberculosis OR: 11.0 (95% CI 1.4; 87.9) p = 0.024]. These models used a stepwise procedure including all variables showing a p-value <0.150 in univariate analysis as described in Materials and Methods. Using Akaike's information criteria (AIC) to assess goodness of fit, the model locking EFV had a slightly better goodness of fit (AIC 81.6) versus the model locking pre-cART tuberculosis (AIC 83.4).

Association between rapidity of HIV-RNA control and immunological and virological outcomes

Parameters of T cell activation and HIV-RNA suppression were compared between patients with early and late virological control. The median percentage of activated CD8 T cells was similar at the pre-cART visit between early and late virological controllers [65.8% (IQR: 57.5–74.5) and 61.5% (IQR: 59.5–70.7) respectively, p = 0.60] (Table 3). However, at each time point thereafter, early virological controllers had a significantly lower median percentage of activated CD8 T cells as compared to late controllers (Table 3).

Table 3.

Dynamics of Activated CD8 T Cells and CD4 Counts over 16 Months after Initiation of cART According to Early or Late Suppression of HIV-RNA

Visits	Early virological controllers (n = 60) Median (IQR)	Late virological controllers (n = 20) Median (IQR)	p^a
CD8 T cell activation (%)
Pre-cART	65.8 (57.5–74.5)	61.5 (59.5–70.7)	0.602
4 months	42.7 (33.7–51.9)	55.7 (41.2–66.3)	0.034
10 months	30.2 (23.2–41.8)	39.6 (34.7–59.3)	0.004
16 months^b	25.1 (18.8–33.1)	34.7 (21.5–68.5)	0.040
CD4 counts (cells/μl)
Pre-cART	134 (80–190)	164 (84–226)	0.211
4 months	272 (216–407)	294 (177–357)	0.960
10 months	303 (205–386)	285 (214–401)	0.781
16 months^b	318 (223–417)	254 (137–338)	0.192

Based on Kruskal-Wallis test comparing medians.

At 16 months post-cART the medians were based on n = 54 early controllers and n = 19 late controllers.

Over the time preceding control of HIV-RNA, early virological controllers had a larger monthly decrease in the median percentage of activated CD8 T cells as compared to late virological controllers [5.28%, IQR (2.52–7.56) vs. 2.29%, IQR (0.90–2.78), p = 0.0001; Table 4]. Early virological controllers decreased the percentage of activated CD8 T cells by an average of 3.17% per month more than did late virological controllers after adjusting for pre-cART HIV-RNA (Table 4). Over the 6 months following control of HIV-RNA, early controllers also decreased the percentage of activated CD8 T cells more rapidly than did late controllers (Table 4). However, after adjustment by HIV-RNA load at 6 months postcontrol, there was no significant difference in monthly decline of CD8 activated T cells between early and late virological controllers (Table 4). Indeed, HIV viral load at 6 months post HIV-RNA control was associated with a –0.62% decrease in the monthly percent decline of activated CD8 T cells per log HIV-RNA after adjustment for early vs. late control [(95% CI –1.01; –0.22), p = 0.002].

Table 4.

Median Monthly Percent Decline in Activated CD8 T Cells According to Early or Late Suppression of HIV-RNA ^a

	Early virological controllers Median % (IQR)	Late virological controllers Median % (IQR)	p^b	Coefficient: adjusted difference	95% CI	p^c
Prior to HIV-RNA control	5.28 (2.52–7.56)	2.29 (0.90–2.78)	0.0001	3.17	1.18; 5.16	0.002
Following HIV-RNA control	1.99 (−0.58–3.59)	0.29 (−1.85–3.32)	0.08	0.26	−1.32, 1.84	0.742

Analysis is presented for monthly percent decrease prior to HIV-RNA control, and for monthly percent decrease during the 6 months following HIV-RNA control as described in Materials and Methods. Multivariate linear regression compares early vs. late controllers after adjusting for viral load.

Kruskal-Wallis.

Multivariate linear regression adjusted by viral load at pre-cART for analysis prior to HIV-RNA control and adjusted by viral load at 6 months postcontrol for analysis posterior to HIV-RNA control.

Median CD4 counts were not significantly different between early and late virological controllers at all time points (Table 3). However, the median gain in CD4 count between pre-cART to 16 months post-cART was significantly greater in early as compared to late virological controllers [median CD4 gain in early controllers: 177 cells/μl (IQR 114–248) vs. median gain in late controllers 98 cells/μl (IQR –22–178), p = 0.012 by Kruskal–Wallis].

Early and late virological controllers differed in their ability to maintain suppression of HIV-RNA 6 months following achievement of suppression. Approximately 15% (2/60) of early HIV controllers had a reappearance of detectable HIV-RNA at 6 months postcontrol as compared to 63% (12/19) of late HIV controllers (p = 0.001).

Discussion

This observational study in a rural area of southern Mozambique suggests that those patients with rapid virological control have less reappearance of HIV-RNA and thus may have lower levels of CD8 T cell activation at various times post-cART initiation as compared to those patients with later HIV-RNA suppression.

In this study, approximately 30% patients initiated treatment with an EFV-based cART, of whom approximately 84% had tuberculosis. Most countries in sub-Saharan Africa, including Mozambique, use an NVP-based regimen for first-line therapy and EFV is given in tuberculosis patients because its interaction with rifampicin is weaker than that of NVP.^24,25 In this study, EFV and tuberculosis were both associated with more rapid reduction of HIV-RNA (OR: 13.6 and OR: 11.0, respectively). Because of the great overlap between EFV and tuberculosis, it is difficult to attribute early virological control to either tuberculosis or an EFV-containing regimen. Since tuberculosis is a criterion for AIDS, and itself immunosuppressive and capable of stimulating HIV replication, tuberculosis may mask an otherwise less advanced immunosuppressed state. Tuberculosis patients were treated with standard four-drug therapy and thus a rapid improvement could unmask a less advanced AIDS stage, which could explain a more rapid virological control. On the other hand, there is controversy as to whether EFV is a more potent antiretroviral than NVP.^26

–29 Cohort studies have suggested that EFV is more effective in reducing HIV-RNA than NVP, but a multicenter randomized clinical trial and other observational studies have shown equivalent efficacy of the two NNRTIs in HIV-RNA reduction and CD4 increase³⁰ and there is limited and conflicting data from resource-poor populations regarding the direct comparison of effectiveness of EFV and NVP in improving immunological and clinical outcomes.^30

–33

The results from this study suggest that earlier HIV-RNA control leads to more rapid decrease of activated CD8 T cells independently of pre-cART viral load. The ability to reduce CD8 T cell activation is associated with better immune reconstitution and thus is likely to impact the patients' ability to limit opportunistic infections as CD8 T cell activation is the strongest predictor of AIDS progression.^20,21,34 This is particularly relevant in the sub-Saharan African context where patients initiate cART at late HIV/AIDS stages with very low CD4 counts. CD8 T cell activation was found to be lower in early virological controllers at all time points after initiation of cART. However, the results suggest that once HIV-RNA suppression was achieved, the average monthly percent decrease of CD8 T cell activation was not significantly different between early and late virological controllers, but rather was more strongly associated with viral load reappearance 6 months postcontrol. Indeed, 63% of late virological controllers had HIV RNA reappearance at 6 months postcontrol as compared to 15% of early virological controllers. Early virological control may thus lead to more durable HIV-RNA suppression. A limitation of this study is that time points were fixed according to national guidelines and thus there are no intermediate measures of CD8 T cell activation. The actual monthly percent decline of activated CD8 T cells between time points may vary more than reflected in this work where it was assumed to be constant.

The HIV-RNA reappearance at 6 months postcontrol could be attributed to a viral load rebound or a viral blip. In both instances, the reappearance of HIV-RNA could be due to either poor adherence or development of resistance, not assessed in our study. Compliance appeared similar between the two groups, although our assessment of adherence included only records of monthly medication pick-ups at the pharmacy. Late virological controllers could be in a state of suboptimal treatment, which would allow HIV to replicate for a longer lapse of time than in early virological controllers, thus increasing the opportunity for the development of resistance.¹⁹ Given that NNRTIs have a low genetic barrier for resistance, generation of resistance mutations to cART is not uncommon in this context. Indeed, a limitation of this study was the unavailability of HIV drug resistance testing for those patients with late HIV-RNA control and viral rebound or blips.

The question of cART monitoring and the minimum frequency and timing of HIV viral load measurement in the follow-up of cART patients in resource-poor countries is a subject of intense debate.^35,36 The main obstacles are the elevated cost, the human resources, and technological requirements. Recent evidence from the DART (Development of AntiRetroviral Therapy in Africa) study has suggested that during the first year of cART, clinically driven use of CD4 testing is as effective as regularly scheduled laboratory analyses for patient survival.³⁷ However, evidence has suggested that clinical or CD4 monitoring is less effective than HIV-RNA for detecting treatment failure and accumulation of resistance mutations.^35,36,38,39 Evidence from our current study may suggest that the application of a single viral load measurement at 4 months could lead to improved monitoring of patients initiating cART at low CD4 counts. Furthermore, if those patients who do not control viral load at 4 months (late controllers) are less likely to maintain suppression once achieved, this group could be a target for which repeat HIV-RNA testing would be advisable.

Many research efforts have been dedicated to determining the time at which to initiate cART. However, the rapidity in achieving HIV-RNA suppression may be equally important in improving virological outcomes. This study emphasizes the importance of achieving rapid suppression of HIV-RNA after initiation of cART in a sub-Saharan African population initiating cART.

Footnotes

Acknowledgments

The authors are grateful to the study participants and for the continued dedication of the VCT, HIV day hospital staff, and data management staff at the Centro de Investigação em Saúde de Manhiça. The authors are particularly grateful to Dr. Deolinda Nhassone, MDH director for her logistic support, to John Aponte, for his helpful inputs on statistical analysis, and to Elsa Banze, Recifal Iaço, Atanasio Xirindza, and Nelito Ernesto José for their participation in sample processing and patient logistics. Financial support was received from the Fundació “la Caixa” and the Agència Catalana de Cooperació al Desenvolupament (ACCD). The “Centro de Investigação em Saúde de Manhiça” receives core funding from the Spanish Agency for International Cooperation and Development (AECID). The VCT center is supported by the Agència Catalana de Cooperació al Desenvolupament (ACCD). D.N. was supported by a grant from the Spanish Ministry of Education and Science (Ramon y Cajal).

Author Disclosure Statement

No competing financial interests exist.

References

Demeter

, Hughes

, Coombs

et al. Predictors of virologic and clinical outcomes in HIV-1-infected patients receiving concurrent treatment with indinavir, zidovudine, and lamivudine. AIDS Clinical Trials Group Protocol 320. Ann Intern Med, 2001; 135:954–964.

Palella

Jr , Delaney

, Moorman

et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med, 1998; 338:853–860.

Gulick

, Mellors

, Havlir

et al. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy. N Engl J Med, 1997; 337:734–739.

Hammer

, Squires

, Hughes

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; 337:725–733.

Montaner

, Reiss

, Cooper

et al. A randomized, double-blind trial comparing combinations of nevirapine, didanosine, and zidovudine for HIV-infected patients: The INCAS Trial. Italy, The Netherlands, Canada and Australia Study. JAMA, 1998; 279:930–937.

Sterne

, May

, Costagliola

et al. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: A collaborative analysis of 18 HIV cohort studies. Lancet, 2009; 373:1352–1363.

World Health Organization. Scaling up antiretroviral therapy in resource-limited settings: treatment guidelines for a public health approach. World Health Organization: Geneva, 2005.

Stevens

, Scott

, Crowe

. Quantifying HIV for monitoring antiretroviral therapy in resource-poor settings. J Infect Dis, 2010; 201,Suppl 1:S16–26.

World Health Organization: WHO. Rapid advice: Antiretroviral therapy for HIV infection in adults and adolescents. Geneva, 2009.

10.

El-Sadr

, Lundgren

, Neaton

et al. CD4+ count-guided interruption of antiretroviral treatment. N Engl J Med, 2006; 355:2283–2296.

11.

Smit

, Geskus

, Walker

et al. Effective therapy has altered the spectrum of cause-specific mortality following HIV seroconversion. AIDS, 2006; 20:741–749.

12.

Bartlett

, Shao

. Successes, challenges, and limitations of current antiretroviral therapy in low-income and middle-income countries. Lancet Infect Dis, 2009; 9:637–649.

13.

Munderi

. When to start antiretroviral therapy in adults in low- and middle-income countries: Science and practice. Curr Opin HIV AIDS, 2010; 5:6–11.

14.

Wood

, Hogg

, Harrigan

, Montaner

. When to initiate antiretroviral therapy in HIV-1-infected adults: A review for clinicians and patients. Lancet Infect Dis, 2005; 5:407–414.

15.

, Neumann

, Perelson

, Chen

, Leonard

, Markowitz

. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature, 1995; 373:123–126.

16.

Wei

, Ghosh

, Taylor

et al. Viral dynamics in human immunodeficiency virus type 1 infection. Nature, 1995; 373:117–122.

17.

Perelson

, Essunger

, Cao

et al. Decay characteristics of HIV-1-infected compartments during combination therapy. Nature, 1997; 387:188–191.

18.

Metzner

, Allers

, Rauch

, Harrer

. Rapid selection of drug-resistant HIV-1 during the first months of suppressive ART in treatment-naive patients. AIDS, 2007; 21:703–711.

19.

Townsend

, Troya

, Maida

et al. First HAART in HIV-infected patients with high viral load: Value of HIV RNA levels at 12 weeks to predict virologic outcome. J Int Assoc Physicians AIDS Care (Chic Ill), 2009; 8:314–317.

20.

Giorgi

, Lyles

, Matud

et al. Predictive value of immunologic and virologic markers after long or short duration of HIV-1 infection. J Acquir Immune Defic Syndr, 2002; 29:346–355.

21.

Hazenberg

, Otto

, van Benthem

et al. Persistent immune activation in HIV-1 infection is associated with progression to AIDS. AIDS, 2003; 17:1881–1888.

22.

Letang

, Almeida

, Miro

et al. Predictors of immune reconstitution inflammatory syndrome-associated with Kaposi sarcoma in Mozambique: A prospective study. J Acquir Immune Defic Syndr, 2010; 53:589–597.

23.

Nhacolo

, Nhalungo

, Sacoor

, Aponte

, Thompson

, Alonso

. Levels and trends of demographic indices in southern rural Mozambique: Evidence from demographic surveillance in Manhica district. BMC Public Health, 2006; 6:291.

24.

Manosuthi

, Sungkanuparph

, Tantanathip

et al. A randomized trial comparing plasma drug concentrations and efficacies between 2 nonnucleoside reverse-transcriptase inhibitor-based regimens in HIV-infected patients receiving rifampicin: The N2R Study. Clin Infect Dis, 2009; 48:1752–1759.

25.

Manosuthi

, Sungkanuparph

, Thakkinstian

et al. Efavirenz levels and 24-week efficacy in HIV-infected patients with tuberculosis receiving highly active antiretroviral therapy and rifampicin. AIDS, 2005; 19:1481–1486.

26.

Keiser

, Nassar

, White

, Koen

, Moreno

. Comparison of nevirapine- and efavirenz-containing antiretroviral regimens in antiretroviral-naive patients: A cohort study. HIV Clin Trials, 2002; 3:296–303.

27.

Phillips

, Ledergerber

, Horban

et al. Rate of viral rebound according to specific drugs in the regimen in 2120 patients with HIV suppression. AIDS, 2004; 18:1795–1804.

28.

Ribaudo

, Kuritzkes

, Lalama

et al. Efavirenz-based regimens in treatment-naive patients with a range of pretreatment HIV-1 RNA levels and CD4 cell counts. J Infect Dis, 2008; 197:1006–1010.

29.

Cozzi-Lepri

, Phillips

, d'Arminio Monforte

et al. Virologic and immunologic response to regimens containing nevirapine or efavirenz in combination with 2 nucleoside analogues in the Italian Cohort Naive Antiretrovirals (I.Co.N.A.) study. J Infect Dis, 2002; 185:1062–1069.

30.

van Leth

, Phanuphak

, Ruxrungtham

et al. Comparison of first-line antiretroviral therapy with regimens including nevirapine, efavirenz, or both drugs, plus stavudine and lamivudine: A randomised open-label trial, the 2NN Study. Lancet, 2004; 363:1253–1263.

31.

Keiser

, Orrell

, Egger

et al. Public-health and individual approaches to antiretroviral therapy: Township South Africa and Switzerland compared. PLoS Med, 2008; 5:e148.

32.

Matthews

, Sabin

, Mandalia

et al. Virological suppression at 6 months is related to choice of initial regimen in antiretroviral-naive patients: A cohort study. AIDS, 2002; 16:53–61.

33.

Nachega

, Hislop

, Dowdy

et al. Efavirenz versus nevirapine-based initial treatment of HIV infection: Clinical and virological outcomes in Southern African adults. AIDS, 2008; 22:2117–2125.

34.

Sodora

, Silvestri

. Immune activation and AIDS pathogenesis. AIDS, 2008; 22:439–446.

35.

Calmy

, Ford

, Hirschel

et al.

HIV viral load monitoring in resource-limited regions: Optional or necessary?

Clin Infect Dis, 2007; 44:128–134.

36.

Lawn

, Bekker

, Calmy

, Wood

. Monitoring of antiretroviral therapy in low-resource settings. Lancet, 2008; 372:287–288author reply 289.

37.

Mugyenyi

, Walker

, Hakim

et al. Routine versus clinically driven laboratory monitoring of HIV antiretroviral therapy in Africa (DART): A randomised non-inferiority trial. Lancet, 2010; 375:123–131.

38.

Mee

, Fielding

, Charalambous

, Churchyard

, Grant

. Evaluation of the WHO criteria for antiretroviral treatment failure among adults in South Africa. AIDS, 2008; 22:1971–1977.

39.

Moore

, Awor

, Downing

et al. CD4+ T cell count monitoring does not accurately identify HIV-infected adults with virologic failure receiving antiretroviral therapy. J Acquir Immune Defic Syndr, 2008; 49:477–484.