The optimal time to start antiretroviral therapy (ART) for patients infected with the HIV is not well established. There is currently considerable debate about whether ART should be initiated at a higher CD4 cell count threshold than the 350 cells/μl advocated in recent guidelines [1–11]. Because mortality and morbidity rates are low at higher CD4 cell counts, a randomized trial to address this question would require a large sample size and long follow-up [5,12–14]. CD4 cell counts are highly predictive of the short-term risk of HIV-related morbidity and mortality [15–18] and so there is considerable interest in evaluating long-term effects of ART on CD4 cell count. One major question is whether treatment can be deferred without irreversible immune system damage [19–22], as might be suggested if CD4 cell counts normalize irrespective of the pretreatment CD4 cell count. A second major question is whether the beneficial effects of ART in raising CD4 cell counts are maintained long-term or are time limited and CD4 cell counts eventually stabilize or decline in patients on ART.
There are conflicting reports about the pattern of CD4 cell counts after multiple years of ART. Studies [23–27] of North American and European adults have variously concluded that CD4 cell counts continue to rise after 4–7 or more years of treatment or that the CD4 cell count stabilizes, overall [28,29] or in a subset of patients [27,30,31]. Most analyses have restricted CD4 cell count assessment to times when plasma HIV-1 RNA (vRNA) was suppressed, for example, to less than 50 copies/ml. The characteristics of patients who remain virologically suppressed are likely different from those who do not, leading to possible selection biases. Hence, it is important to evaluate CD4 trajectories in complete populations initiating ART. Further biases may arise if dropout, often common, is ignored. For example, if more immunocompromised patients with lower CD4 cell counts drop out more frequently, an analysis that simply includes the patients in follow-up will overestimate the CD4 cell count trajectory in the entire population. Such issues may explain why different studies reach different conclusions regarding long-term CD4 cell count trajectories.
We evaluated the long-term CD4 cell count trajectory among patients who first initiated ART while participating in AIDS Clinical Trials Group (ACTG) study 384, a randomized clinical trial, which showed that efavirenz (EFV)–lamivudine (3TC)/zidovudine (ZDV) was a potent combination and helped to establish EFV-based regimens as initial treatment for HIV-infected patients [8,32,33]. Differences between the randomized arms were seen in vRNA suppression, but not in CD4 cell count increases over 3 years [32–34]. EFV–3TC/ZDV was the best initial regimen, having the lowest rate of protocol-defined virologic failure among the three-drug regimens; the four-drug regimens studied did not show benefit over this regimen [32,33].
We assessed three objectives regarding long-term CD4 cell count increases during ART, using data from ACTG 384 and subsequent follow-up in an observational study, ACTG A5001/AIDS Clinical Trials Group Longitudinal Linked Randomized Trials (ALLRT) . First, we evaluated the long-term CD4 cell count trajectory for the patients enrolled in ACTG 384 based on the treatment practices during the past few years, addressing the question ‘What would have happened to the CD4 cell count trajectory in this cohort had all patients been followed for 7 years’? During follow-up, however, treatment guidelines changed, notably following the finding that discontinuation of ART at higher CD4 cell counts leads to increased morbidity . Therefore, we also addressed the question ‘What would have happened to the CD4 cell count trajectory of these patients had all remained on ART throughout the 7 years’? Last, we addressed the question ‘What was the CD4 cell count trajectory among patients who remained in follow-up and were virologically suppressed for 7 years’? This focuses on a select subgroup of patients for whom treatment was successful over a long period, providing insight into results for the first two questions and an opportunity to compare our results with those of studies focusing on this subgroup. We also evaluate the long-term differences in CD4 cell count according to initial regimen in ACTG 384, and according to pretreatment CD4 cell count categories, because such differences inform the ongoing debate over the optimal CD4 cell count threshold at which ART should be initiated.
ACTG 384 [32,33] enrolled 898 patients in 1998–1999 at 58 sites in the United States. We excluded 82 participants enrolled in Italy, as they were ineligible for long-term follow-up in ALLRT. ACTG 384 ended in 2002. ART-naive patients were randomized to either EFV (a nonnucleoside reverse transcriptase inhibitor, NNRTI), nelfinavir (NFV, a protease inhibitor), or both EFV and NFV (NNRTI/protease inhibitor) (double-blind) and were further randomized to add two nucleoside reverse transcriptase inhibitors (NRTIs): either ZDV + 3TC or didanosine/stavudine (open-label). There were no enrollment restrictions on CD4 cell count; vRNA had to be at least 500 copies/ml.
AIDS Clinical Trials Group Longitudinal Linked Randomized Trials
Following ACTG 384, most (575, 64%) patients enrolled in ALLRT , a long-term observational study for patients enrolled in randomized treatment studies. ALLRT did not provide ART. At the time of analysis, all participants had potential follow-up of 7 years.
Evaluations in both studies continued regardless of changes in or discontinuation of ART, including CD4 cell count and vRNA measurements every 8 or 16 weeks.
Patients provided written informed consent for participation in ACTG 384 and ALLRT. All sites received institutional review board approval for both studies.
CD4 cell count measurements
We focused on annual CD4 cell counts, enabling us to develop a model and hence adjust for loss-to-follow-up in the corresponding 1-year periods. Because ALLRT followed participants every 16 weeks, a patient's annual CD4 cell count was defined as the mean of up to three 16-week values closest to the actual year since start of ART. Baseline CD4 cell count was the mean of the two pretreatment CD4 cell counts in ACTG 384.
Virological suppression was defined as no confirmed vRNA above 400 copies/ml, based on measurements every 16 weeks.
End of follow-up
End of follow-up was defined as the year of death or, if not known to be deceased, the last year a CD4 cell count was available.
We estimated the CD4 cell count trajectory, as represented by the median CD4 cell count for the 898 patients originally enrolled in the 7 successive years since starting ART, for each of the three scenarios described in the ‘Introduction’ section. For the analysis had everyone remained on treatment, we censored patients if they discontinued ART for more than 60 days. In estimating the median CD4 cell count, if a patient died, we ranked the patient lower than any surviving patient; this avoids the selection biases arising if the CD4 cell count trajectory is only evaluated among surviving patients. The median CD4 cell count then has the interpretation that 50% of the original population is alive with a CD4 cell count above that median.
To address informative censoring due to selective follow-up, that is, individuals with certain characteristics may be more likely to remain in follow-up (or on ART), we used inverse probability of censoring weighting (IPCW) [37,38]. To illustrate the concept behind IPCW, suppose that a study population includes a subset of 10 identical patients of whom five are lost to follow-up. IPCW then assigns the five with observed data double the weight in the analysis. In general, IPCW assigns weights to patients on follow-up: the inverse of their estimated probability of being on follow-up based on their patient characteristics. This approach has been proven to lead to estimators with desirable statistical properties . This analytical approach estimates outcomes in the entire study population, in contrast to describing a subgroup selected over time. In practice, the (unobserved) outcomes of patients who were lost-to-follow-up are represented by increasing the weight given to the outcomes of ‘similar’ patients who remained in follow-up. The weights are obtained from a logistic regression model used to describe the probability of remaining in follow-up at a given time in terms of covariate values observed in the past (so these covariates are the basis for identifying ‘similar’ patients). The main assumption behind this method is that there are no (possibly unmeasured) factors that both affect the subsequent CD4 cell count trajectory and independently predict the probability of remaining in follow-up, which are not included in the logistic regression model.
The logistic regression model used to estimate the weights, by estimating the probabilities of remaining in follow-up, used separate models during ACTG 384, the transition from ACTG 384 to ALLRT, and during ALLRT, as factors influencing follow-up might differ across these three situations. The covariates included were those that we considered might be confounding variables , that is, predict loss to follow-up and also long-term CD4 cell count outcome: initial treatment, pre-ART CD4 cell count and log10 vRNA, age, sex, race/ethnicity (white non-Hispanic/other), injection drug use (ever/never), and time-varying variables, that is, recent CD4 cell count, change in CD4 cell count over the previous year, vRNA (<400 versus ≥400 copies/ml), and year since starting ART. In sensitivity analyses, we repeated the IPCW analysis using a single censoring model and a reduced single model obtained from stepwise variable selection; results (not shown) changed minimally. Standard errors were obtained by bootstrap . For comparison, we also present results from a nonweighted (naive) analysis, which could be biased, for reasons discussed above. SAS 9.1.3 (SAS Institute Inc., Cary, North Carolina, USA) was used for all analyses.
Selected baseline (pre-ART) characteristics of the 898 ACTG 384 patients enrolled in the United States are 82% men; 9% ever used intravenous drugs; 42% white non-Hispanic; median age 36 years [interquartile range (IQR) 30–43 years]; median vRNA (log10 copies/ml) 4.95 (IQR 4.33–5.53); CD4 cell count (cells/μl) of more than 500 (17%), 351–500 (20%), 201–350 (24%), and 200 or less (39%). These baseline characteristics did not differ significantly between baseline CD4 cell count categories except that there was a significant trend for higher vRNA with lower CD4 cell counts. During follow-up in ACTG 384, 18 (2.0%) patients died and 683 (76%) patients contributed CD4 cell data until the end of that study; 575 (64% of 898) continued observational follow-up in ALLRT, of whom 14 (2.4% of 575) died during ALLRT follow-up; and 356 (62% of 575) patients contributed CD4 cell data through 7 years after starting ART. This resulted in 543 patients (60%) who were in follow-up in year 4 or had died previously, and 388 (43%) in year 7. For the analysis had everyone remained on treatment, 430 patients (48%) were in follow-up and on ART without any interruptions of more than 60 days in year 4, and 244 (27%) in year 7.
We found that older age, white non-Hispanic racial/ethnicity group, female sex, noninjection drug user, higher pretreatment viral load, lower current CD4 cell count, higher CD4 cell count increase in the last year, and vRNA less than 400 copies/ml were significantly associated with increased probability of staying in follow-up in at least one of the models for retention. Initial treatment and year of follow-up were also significant.
Figure 1 shows the nonweighted and IPCW-estimated intent-to-treat (i.e. ignoring ART discontinuations) median CD4 cell count over time, for patients initially on the standard of care regimen EFV–3TC/ZDV versus the other three-drug regimens versus the four-drug regimens. These IPCW-estimated medians show the CD4 cell count trajectory over time in the study population given the treatment practices between 1998 and 2007, had all patients actually remained in follow-up for 7 years (or until death). Although weights ranged from 1.02 to 15, there were only small differences between results from the nonweighted and IPCW analyses. Thus, although there were strong predictors of loss to follow-up, these did not markedly confound the evaluation of the long-term CD4 cell count trajectory. There was no significant difference in long-term median CD4 cell count trajectory between patients initially assigned to EFV–3TC/ZDV versus other initial regimens. In the overall study population, combining across initial treatments, the median CD4 cell count increased from 270 cells/μl prior to starting treatment to 433, 518, and 556 cells/μl at 1, 2, and 3 years and changed minimally thereafter to a median of 532 cells/μl at 7 years. Table 1 also shows the estimated 25th and 75th percentiles for CD4 cell counts in the study population at 3 and 7 years. These also changed very little after 3 years, indicating stability in CD4 cell counts between 3 and 7 years in the study population.
Figure 2(a) shows the nonweighted and IPCW-estimated median CD4 cell counts over time by pretreatment CD4 cell category, showing again only small differences between the two analyses. The absolute changes in median CD4 cell count were similar for 3 years irrespective of pretreatment CD4 cell count category; thereafter, the trajectories differed, with an increase in the median CD4 cell count (from 369 to 453 cells/μl) for patients with pretreatment CD4 cell count of 200 cells/μl or less, and declines for patients in the pretreatment CD4 cell count categories of 201–350, 351–500, and more than 500 cells/μl (Table 1). The IPCW-estimated proportion of patients with a CD4 cell count of 350 cells/μl or less at (or dying before) 7 years, by pretreatment CD4 count category, was 33% for pretreatment CD4 cell count of 200 cells/μl or less, 24% for CD4 cell count of 201–350 cells/μl, 16% for CD4 cell count of 351–500 cells/μl, and 8% for CD4 cell count of more than 500 cells/μl (overall 23%). Table 1 also includes these percentages for thresholds of 200 and 500 cells/μl.
Some patients discontinued ART for more than 60 days, and the proportion discontinuing increased with higher pretreatment CD4 cell counts, possibly explaining the decline in median CD4 cell counts in the intent-to-treat analysis for patients in the higher pretreatment CD4 cell count categories. This is addressed in the ‘on-treatment’ analysis, in which it is assumed that patients who discontinued ART might have remained on ART had they known of the potential adverse consequences of discontinuing ART identified in the Strategies for Management of Antiretroviral Therapy (SMART) study . In the models for the on-treatment analysis, older age and vRNA below 400 copies/ml, but not CD4 cell count, significantly increased the probability of staying on treatment, given a patient remained in follow-up.
Figure 2(b) shows the nonweighted and IPCW-estimated median CD4 cell counts over time, had all patients remained on ART throughout the 7 years without interruptions longer than 60 days. As for the intent-to-treat analysis, the differences between results from the two analyses were generally small except for the highest pretreatment CD4 category (>500 cells/μl), for which the difference increased after year 4. In all categories of pretreatment CD4 cell count, the median CD4 cell count either stabilized or continued to rise for 7 years, although more slowly after the first 3 years of ART. Table 2 gives the estimated medians and 25th and 75th percentiles of the distribution of CD4 cell counts during years 3 and 7, had patients remained on ART. As with the intent-to-treat analysis, there were minimal differences in median CD4 cell count according to type of initial ART. The medians and 25th and 75th percentiles typically showed modest increases between years 3 and 7, suggesting a general upward shift in the distribution of CD4 cell counts in the study population, both overall and by pretreatment CD4 cell count category. Had all patients remained on ART, the IPCW-estimated proportion of patients with a CD4 cell count below 350 cells/μl at (or dying before) 7 years after starting ART was 25% for pretreatment CD4 cell count of 200 cells/μl or less, 9% for CD4 cell count of 201–350 cells/μl, 3% for CD4 cell count of 351–500 cells/μl, and 2% for CD4 cell count of more than 500 cells/μl (overall 14%). Table 2 also includes these percentages for thresholds of 200 and 500 cells/μl.
Figure 3 shows the median CD4 cell counts over time in the patients who were followed for 7 years by pretreatment CD4 cell count category, comparing those who did versus did not maintain virologic suppression throughout. In those patients who remained virologically suppressed for 7 years, the median CD4 cell count continued to rise over 7 years in the lower pretreatment CD4 cell count categories, and stabilized in the higher ones. The median CD4 cell count declined in later years among patients who were not virologically suppressed throughout 7 years.
Our study found substantial improvements in median CD4 cell count during 7 years after HIV-infected patients first started ART in the randomized clinical trial ACTG 384, which established the combination of EFV–3TC/ZDV as a standard of care for initial treatment of HIV infection. The largest increases occurred during the first 1–2 years of follow-up and, in the study population as a whole, showed little change between 3 and 7 years. When changes in median CD4 cell count were evaluated by pretreatment CD4 cell count, patients in the lowest category considered, 200 cells/μl or less, experienced a modest gain between 3 and 7 years, whereas patients in the highest category, more than 500 cells/μl, experienced a decline.
However, some patients discontinued ART during follow-up, particularly those with higher pretreatment CD4 cell counts, a practice that is no longer recommended because of the SMART study  results. We, therefore, conducted a novel statistical analysis, IPCW, which estimated CD4 cell count trajectories, had all patients remained on ART. This showed that the median CD4 cell count continued to show a modest increase between 3 and 7 years after starting ART, both in the overall study population and across pretreatment CD4 cell count categories. Of note, this pattern was also seen when considering the 25th and 75th percentiles of the distribution of CD4 cell counts, indicating ongoing peripheral CD4 cell count reconstitution with continued ART in the broad population of patients.
Patients who began ART at lower CD4 cell counts continued to have lower CD4 cell counts than those who began ART at higher CD4 cell counts. We found this not only in the subset of patients for whom virologic suppression was maintained throughout the 7 years of follow-up, as found in other studies [25–27,29,31], but also in our analyses, which evaluated the entire study population who started ART. We estimated that 33% of patients who originally started ART when their CD4 cell counts were 200 cells/μl or less would have had CD4 cell counts 350 cells/μl or less at (or died before) 7 years in the intent-to-treat analysis; the IPCW model suggested that had all patients remained on ART for 7 years, 25% would still have CD4 cell counts below 350 cells/μl after 7 years. In contrast, among patients with pretreatment CD4 cell counts in the 351–500 and more than 500 cells/μl ranges, the latter estimates were 3 and 2%, respectively.
Thus, initiation of ART at higher CD4 cell counts than typically recommended in recent treatment guidelines appears to be associated with maintenance of CD4 cell counts at levels above 350 cells/μl (when the risk of HIV-associated morbidity and mortality is low) for almost all patients for at least 7 years. Postponing ART until the CD4 cell count drops below 350 cells/μl might be too conservative, as patients who are in the pretreatment CD4 cell count category of 201–350 cells/μl have a median 7-year CD4 cell count similar to patients with pretreatment CD4 cell count of 200 cells/μl or less.
In the subgroup of patients who maintained vRNA below 400 copies/ml for 7 years, we observed ongoing improvements in median CD4 cell counts among patients who started ART at CD4 cell counts less than 350 cells/μl. Although this is a selected subgroup, this finding and similar findings in other studies [25,29,31] provide support for treatment management practices that promote sustained virologic suppression. Additionally, individuals without sustained virologic suppression are at risk for acquiring drug resistance mutations. However, it is difficult to distinguish between virologic failure due to drug resistance and virologic failure due to nonadherence or toxicity/intolerance, as it is often due to a combination of these factors.
We found no long-term differences in CD4 cell counts achieved according to initial category of randomized treatment between EFV–3TC/ZDV and the other three and four-drug regimens, a finding made possible by the ALLRT study wherein participants are followed long-term after their participation in randomized ACTG clinical trials and through changes in ART regimens. Like ACTG 384, two other studies [39,40] found no difference between randomized initial ART in short-term CD4 cell count increases (2–3 years). The lack of long-term differences found in our study likely reflects the availability of multiple potent antiretroviral regimens, so that even after 7 years, most patients still have treatment options. In addition, of the six regimens studied in ACTG 384, only the EFV–3TC/ZDV regimen has remained a preferred regimen, and most patients discontinued use of the other regimens; of those randomized to EFV–3TC/ZDV and alive and in follow-up after 7 years, 54% were on their initial regimen as compared with 11% for the other three-drug regimens and 4% for the four-drug regimens.
A notable strength of our study is the well characterized patient population and the standardized prospective collection of data. The main limitation is that some patients in ACTG 384 did not enroll in the ALLRT study and some patients were lost to follow-up. In addition, some patients discontinued treatment during follow-up, particularly at higher CD4 cell counts, consistent with treatment management practices during the study period. We attempted to address these limitations by using the IPCW method. This method adjusts for factors that are included in the model used to derive weights related to the probability of remaining in follow-up (or on ART), which might also affect a patient's future CD4 cell count trajectory. It is possible, however, that there are other (unmeasured) factors that influenced patient decisions to stay in follow-up or stay on ART that might also have been associated with CD4 cell count outcome. Omitting such factors may lead to bias in our results. For example, among patients with similar covariate histories, if patients who discontinued ART would have had lower CD4 cell counts, had they continued on ART than patients who did actually continue on ART, then the IPCW method might overestimate the subsequent CD4 trajectory that could be achieved in the whole population, had everyone stayed on ART.
This study has other limitations. First, recommended ART regimens have changed during recent years, so CD4 cell count outcomes might be different for patients who are starting ART now. As current regimens tend to be better tolerated and possibly more efficacious, in part because of simpler dosing (e.g. once daily) leading to improved adherence, patients starting ART now might have better CD4 cell count outcomes than in our study. Second, we studied a group of patients who enrolled in a clinical trial, most of whom subsequently participated in a long-term observational study. Outcomes might be different in a more general patient population, although the demographic characteristics of our patients were quite diverse. Third, it is possible that the better outcomes among patients starting ART at higher CD4 cell counts are attributable to differences in other patient characteristics. For example, it is possible that starting ART at higher CD4 cell counts is associated with health-seeking behavior, so patients starting ART at higher CD4 cell counts might be expected to fare better for this reason.
In summary, our study suggests that the median CD4 cell count in this population of patients who initiated NNRTI and protease inhibitor-based regimens increased for 3 years and then stabilized through 7 years of follow-up. However, when adjustment was made for discontinuation of ART, the median CD4 cell count continued to rise in later years. This trend was seen across all categories of pretreatment CD4 cell counts. However, even after 7 years, a significant proportion of patients who initiated ART with CD4 cell counts below 200 cells/μl still had counts below 350 cells/μl. The inability to normalize CD4 cell counts among many patients starting ART at low CD4 cell counts, even after 7 years of treatment, provides additional support to consider initiation of therapy at higher CD4 cell counts, consistent with recent observational studies that suggested benefits in terms of HIV-related morbidity and mortality [9,41,42] and the newly updated treatment guidelines .
We are indebted to the patients who volunteered for ACTG 384 and subsequently to ALLRT, the ACTG sites, and the ACTG 384 and ALLRT study teams. We also want to thank Joe Eron for productive discussions and Andrea Rotnitzky for her valuable insights into inverse probability weighting. This study was supported in part by the ACTG funded by the National Institute of Allergy and Infectious Diseases (NIAID; AI 38858, AI 68636, AI 069434, AI 069472, and AI 062435), the Statistical and Data Management Center (AI 38855 and AI 68634), and the National Institute of Health (NIAID R01 AI 51164, AI 024643, and NIH-R01-GM48704).
R.S. and G.R. were involved in design and conduct of ACTG 384. R.B., C.B., and A.C. were involved in the design and conduct of the ALLRT study. J.L., R.B., and M.H. carried out the analyses. J.L. drafted a first version of the manuscript. All authors contributed to the final manuscript.
ACTG 384: http://clinicaltrials.gov/ct2/show/NCT00000919.
ACTG A5001/ALLRT: http://clinicaltrials.gov/ct2/show/NCT00001137.
This work was also supported by the NIAID grant numbers AI069474, AI27664, and AI69432. We would like to thank the ACTG 384 and ALLRT participants and acknowledge the following persons and institutions who participated in the conduct of this study: Massachusetts General Hospital, Amy Sbrolla, BSN, RN and Nicole Burgett-Yandow, BSN, RN; Beth Israel Deaconess Medical Center (BIDMC) CRS, Mary Albrecht, MD and Neah Kim, MSN, FNP-C, CTU grant number AI069472, CFAR grant number, P30 AI060354 A0104; Boston Medical Center ACTG CRS, Paul R. Skolnik, MD; Betsy Adams, RN, CTU grant number 5U01AI069472, GCRC grant number M01- RR00533; Brigham and Women's Hospital, Paul Sax, MD and Joanne Delaney RN, CTU grant number AI069472; Johns Hopkins University, Denice Jones and Ilene Wiggins, RN, CTU grant number AI-69465, GCRC grant number RR-00052; NYU/NYC HHC at Bellevue, Janet Forcht, RN and Richardson St. Louis, CTU grant numbers AI27665 and AI69532, GCRC Grant RR00096; Mount Sinai Medical Center; Stanford University, Sandra Valle, PA-C & Jane Norris, PA-C, CTU grant number UOI-A1069556; San Mateo County AIDS Program; Santa Clara Valley Medical Center; Willow Clinic; UCLA School of Medicine, Judith Currier, MD, MSc and Eric Daar, MD, CTU grant number AI069424; Harbor-UCLA Medical Center; University of California, San Diego Antiviral Res, Susan Cahill, RN and Linda Meixner, RN, CTU grant number AI069432; San Francisco General Hospital, C. Bradley Hare, MD and Diane Havlir, MD, CTU grant number AI69502; Marin County Department of Health; University of Miami, Hector H. Bolivar, MD and Margaret A. Fischl, MD, CTU grant numbers AI27675 and AI69477; University of Pittsburgh, Deborah McMahon, MD and Barbara Rutecki, MSN, MPH, CTU grant number AI69494; Georgetown University, Princy Kumar, MD and Karyn Hawkins, RN; University of Rochester Medical Center, Jane Reid, RNc, MS, ANP and Mary Adams, RN, MPH, CTU grant numbers AI27658 and AI69511, GCRC grant number RR00044; SUNY-Buffalo, Erie County Med Ctr, Gene Morse PharmD, CTU: AI069511-02, CRC: 5-MO1 RR00044; McCree McCuller Wellness Center, Nyef El-Daher MD, CTU: AI069511-02, CRC: 5-MO1 RR00044; AIDS Community Health Center, Christine Hurley RN and Roberto Corales DO, CTU: AI069511-02, CRC: 5-MO1 RR00044; University of Southern California, Fred R. Sattler, MD and Frances Marie Canchola, RN, CTU grant numbers AI27673 and AI69428; University of Washington (Seattle), Sheryl S. Storey, PA-C and Shelia Dunaway, MD, CTU grant number AI069434, CTU grant number AI27664; University of Minnesota, Henry H. Balfour, Jr and Christine Fietzer, CTU grant number AI27661; Hennepin County Medical Ctr, Keith Henry, MD and Bette Bordenave, RN; University of Iowa Hospitals and Clinics; University of Nebraska Med. Ctr., Susan Swindells, MBBS and Frances Van Meter, APRN, CTU grant number AI27661; Duke University Medical Center, Gary M. Cox, MD and Martha Silberman, RN, CTU grant number 1U01-AI069484; Washington University (St. Louis), Mark Rodriguez RN BSN and Ge-Youl Kim, RN, BSN, CTU grant UO1 AI069495; St. Louis Connect Care; Ohio State University, Michael F. Para, MD and Diane Gochnour, RN, CTU grant number AI069474; University of Cincinnati, Judith Feinberg, MD and Jenifer Baer, RN, CTU grant number AI-069513; Case Western Reserve University, Benigno Rodriguez, MD, MSc and Barbara Philpotts, RN, BSN, CTU grant numbers AI25879 and AI69501; MetroHealth CRS; Cleveland Clinic; Indiana University, Mitchell Goldman, MD and Beth Zwickl, NP, CTU grant number AI25859-19, GCRC grant number RR000750; Methodist Hospital of Indiana; Wishard Memorial Hospital; Northwestern University, Robert L. Murphy, MD and Baiba Berzins, MPH, CTU grant number AI69471; Rush University Medical Center in Chicago, Beverly E. Sha, MD and Janice Fritsche, MS, APRN, CTU grant number U01 AI069471; Cook County Hospital Core Center, Oluwatoyin Adeyemi, MD and Joanne Despotes, RN, MPH; Beth Israel Medical Center, Donna Mildvan, MD and Gwendolyn Costantini, FNP, CTU grant number AI46370; The Miriam Hospital, Karen T. Tashima, MD, Pamela Poethke, RN, BSN, and Katherine Wright and Kim Raposa, CTU grant numbers AI46381 and AI69472; University of North Carolina, David Ragan, RN and Joseph J. Eron Jr, MD, CTU grant number U01 AI069423, CFAR grant number P30AI050410(-11), GCRC grant number M01 RR000046-48; Moses H. Cone Hospital, Kim Epperson, RN and Timothy Lane, MD; Carolinas Medical Center; Wake County Human Services, David Currin, RN and Kristine Patterson, MD; Vanderbilt University, Michael Morgan, FNP, Brenda Jackson, RN, Vicki Bailey, RN and Janet Nicotera, RN, BSN, CTU grant number AI069439; University of Texas, Southwestern Medical Center, Philip Keiser, MD and Tianna Petersen, MS, CTU grant number AI46376; University of California, Davis Medical Center, Melissa Schreiber, PA and Abby Olusanya, NP, CTU grant number AI38858-09S1; UC Davis Medical Center; University of Maryland, Institute of Human Virology, Charles Davis, MD and Onyinye Erondu, RN, CTU grant number AI69447; University of Hawaii, Nancy Hanks, RN and Lorna Nagamine, RN, CTU grant number AI34853; Puerto Rico-AIDS Clinical Trial Unit (PR-ACTU), Jorge L Santana, MD and Santiago Marrero, MD, CTU grant number Al69415; University of Alabama at Birmingham, Michael Saag, MD and Kerry Upton, RN, CTU grant number AI069452, GCRC grant number M01-RR00032; Emory University, The Ponce de Leon Center, Jeffrey Lennox, MD and Carlos del Rio, MD, CTU grant number AI69418, CFAR grant number AI50409; University of Colorado Health Sciences Center, Denver, Beverly Putnam, MSN and Cathi Basler, MSN, CTU grant number AI069450, GCRC grant number RR00051, CFAR grant number AI054907; University of Pennsylvania, Philadelphia, Harvey Friedman, MD and Rosemarie Kappes, MPH, UO1-AI 032783-14 and AI69467, CFAR grant number 5-P30-AI-045008-09; Presbyterian Medical Center; University of Texas, Galveston, William A. O'Brien, MD and Gerianne Casey, RN, CTU grant number AI32782; Aaron Diamond AIDS Research Center; Columbia University, Jolene Noel Connor, RN, RNC and Madeline Torres, RN, RNC, CTU grant number AI069470, GCRC grant number RR024156; Weill Medical College, Valery Hughes, FNP and Todd Stroberg, RN, GCRC grant number RR00047; Cornell University CRS; Emory University Comprehensive Hemophilia Program, James Paul Steinberg, MD; Tulane University, Cindy Leissinger, MD.
M.D.H. is a paid member of Data and Safety Monitoring Boards for Boehringer-Ingelheim, Medicines Development Ltd., Pfizer, and Tibotec. These companies are all manufacturers or developers of ART or other therapy for HIV infection. In the last year and currently, C.A.B. has served on advisory boards for GlaxoSmithKline and Merck, and served on a Data and Safety Monitoring Board for Achillion, A.C.C. receives research support from Merck & Co. and Schering-Plough, has served on advisory boards for GlaxoSmithKline and Pfizer, is a member of a Data and Safety Monitoring Board for Merck & Co., and owns stock in Abbott Laboratories and Bristol Myers Squibb. In the last year, G.K.R. was a consultant for Johnson and Johnson.
1. Hammer SM, Eron JJ, Reiss P, Schooley RT, Thompson MA, Walmsley S, et al
. Antiretroviral treatment of adult HIV infection: 2008 recommendations of the International AIDS Society-USA panel. JAMA 2008; 300:555–570.
2. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents
. Washington, DC, USA: Department of Health and Human Services; 2008. pp. 1–139. http://aidsinfo.nih.gov/contentfiles/AdultandAdolescentGL001226.pdf
. [Accessed 9 December 2009]
3. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents
. Department of Health and Human Services; 2009. pp. 1–161. Available at http://aidsinfo.nih.gov/contentfiles/AdultandAdolescentGL.pdf
. [Accessed 9 December 2009]
4. Wilkin TJ, Gulick RM. When to start antiretroviral therapy? Clin Infect Dis 2008; 47:1580–1586.
5. Phillips AN, Emery S. Predicting the potential benefits of early initiation of ART: time to do a trial to find out. Curr Opin HIV AIDS 2009; 4:165–166.
6. Phillips AN, Gazzard BG, Clumeck N, Losso MH, Lundgren JD. When should antiretroviral therapy for HIV be started? BMJ 2007; 334:76–78.
7. Wood R, Lawn SD. Should the CD4 threshold for starting ART be raised? Lancet 2009; 373:1314–1316.
8. Robbins GK, Spritzler JG, Chan ES, Asmuth DM, Rajesh TG, Rodriguez BA, et al
. Incomplete reconstitution of T cell subsets on combination antiretroviral therapy in the AIDS Clinical Trials Group protocol 384. Clin Infect Dis 2009; 48:350–361.
9. Kitahata MM, Gange SJ, Abraham AG, Merriman B, Saag MS, Justice AC, et al
. Effect of early versus deferred antiretroviral therapy for HIV on survival. N Engl J Med 2009; 360:1815–1826.
10. Sabin CA, Phillips AN. Should HIV therapy be started at a CD4 cell count above 350 cells/μl in asymptomatic HIV-1-infected patients? Curr Opin Infect Dis 2009; 22:191–197.
11. WHO treatment guidelines (with addendum). Antiretroviral therapy for HIV infection in adults and adolescents
12. Antiretroviral Therapy Cohort Collaboration. Rates of disease progression according to initial highly active antiretroviral therapy regimen: a collaborative analysis of 12 prospective cohort studies
. J Infect Dis
13. Hughes MD, Ribaudo HR. The search for data on when to start treatment for HIV infection. J Infect Dis 2008; 197:1084–1086.
14. Strategies for Management of Antiretroviral Therapy (SMART) Study Group. Major clinical outcomes in antiretroviral therapy (ART)-naive participants and in those not receiving ART at baseline in the SMART study
. J Infect Dis
15. Phillips AN, Lundgren JD. The CD4 lymphocyte count and risk of clinical progression. Curr Opin HIV AIDS 2006; 1:43–49.
16. Guiguet M, Porter K, Phillips A, Costagliola D, Babiker A. Clinical progression rates by CD4 cell category before and after the initiation of combination antiretroviral therapy (cART). Open AIDS J 2008; 2:3–9.
17. Monforte A, Abrams D, Pradier C, Weber R, Reiss P, Bonnet F, et al
. HIV-induced immunodeficiency and mortality from AIDS-defining and non-AIDS-defining malignancies. AIDS 2008; 22:2143–2153.
18. Baker JV, Peng G, Rapkin J, Abrams DI, Silverberg MJ, MacArthur RD, et al
count and risk of non-AIDS diseases following initial treatment for HIV infection. AIDS 2008; 22:841–848.
19. Lange CG, Lederman MM, Medvik K, Asaad R, Wild M, Kalayjian, et al
. Nadir CD4+T-cell count and numbers of CD28+CD4+T-cells predict functional responses to immunizations in chronic HIV-1 infection
20. Guadalupe M, Reay E, Sankaran S, Prindiville T, Flamm J, McNeil A, et al
. Severe CD4+
T-cell depletion in gut lymphoid tissue during primary human immunodeficiency virus type 1 infection and substantial delay in restoration following highly active antiretroviral therapy. J Virol 2003; 77:11708–11717.
21. Schacker T. The role of secondary lymphatic tissue in immune deficiency of HIV infection. AIDS 2008; 22:S13–S18.
22. Estes J, Baker JV, Brenchley JM, Khoruts A, Barthold JL, Bantle A, et al
. Collagen deposition limits immune reconstitution in the gut. J Infect Dis 2008; 198:456–464.
23. Hunt PW, Deeks SG, Rodriguez B, Valdez H, Shade SB, Abrams DI, et al
. Continued CD4 cell count increases in HIV-infected adults experiencing 4 years of viral suppression on antiretroviral therapy. AIDS 2003; 17:1907–1915.
24. Smith K, Aga E, Bosch RJ, Valdez H, Connick E, Landay A, et al
. Long-term changes in circulating CD4 T lymphocytes in virologically suppressed patients after 6 years of highly active antiretroviral therapy. AIDS 2004; 18:1953–1956.
25. Landay A, da Silva BA, King MS, Albrecht M, Benson C, Eron J, et al
. Evidence of ongoing immune reconstitution in subjects with sustained viral suppression following 6 years of lopinavir–ritonavir treatment. Clin Infect Dis 2007; 44:749–754.
26. Mocroft A, Phillips AN, Gatell J, Ledergerber B, Fisher M, Clumeck N, et al
. Normalisation of CD4 counts in patients with HIV-1 infection and maximum virological suppression who are taking combination antiretroviral therapy: an observational cohort study. Lancet 2007; 370:407–413.
27. Kelley CF, Kitchen CM, Hunt PW, Rodriguez B, Hecht FM, Kitahata M, et al
. Incomplete peripheral CD4+
cell count restoration in HIV-infected patients receiving long-term antiretroviral treatment. Clin Infect Dis 2009; 48:787–794.
28. García F, de Lazzari E, Plana M, Castro P, Mestre G, Nomdedeu M, et al
. Long-term CD4+
T-cell response to highly active antiretroviral therapy according to baseline CD4+
T-cell count. J Acquir Immune Defic Syndr 2004; 36:702–713.
29. Moore RD, Keruly JC. CD4+
cell count 6 years after commencement of highly active antiretroviral therapy in persons with sustained virologic suppression. Clin Infect Dis 2007; 44:441–446.
30. Kaufmann GR, Furrer H, Ledergerber B, Perrin L, Opravil M, Vernazza P, et al
. Characteristics, determinants, and clinical relevance of CD4 T cell recovery to <500 cells/μl in HIV type 1-infected individuals receiving potent antiretroviral therapy. Clin Infect Dis 2005; 41:361–372.
31. Gras L, Kesselring AM, Griffin JT, van Sighem AI, Fraser C, Ghani AC, et al
. CD4 cell counts of 800 cells/mm3
or greater after 7 years of highly active antiretroviral therapy are feasible in most patients starting with 350 cells/mm3
or greater. JAIDS 2007; 45:183–192.
32. Robbins GK, De Gruttola V, Shafer RW, Smeaton LM, Snyder SW, Pettinelli C, et al
. Comparison of sequential three-drug regimens as initial therapy for HIV-1 infection. N Engl J Med 2003; 349:2293–2303.
33. Shafer RW, Smeaton LM, Robbins GK, De Gruttola V, Snyder SW, D'Aquila RT, et al
. Comparison of four-drug regimens and pairs of sequential three-drug regimens as initial therapy for HIV-1 infection. N Engl J Med 2003; 349:2304–2315.
34. Gandhi RT, Spritzler J, Chan E, Asmuth DM, Rodriguez B, Merigan TC, et al
. Effect of baseline- and treatment-related factors on immunologic recovery after initiation of antiretroviral therapy in HIV-1-positive subjects: results from ACTG 384. J Acquir Immune Defic Syndr 2006; 42:426–434.
35. Smurzynski M, Collier AC, Koletar SL, Bosch RJ, Wu K, Bastow B, et al
. AIDS clinical trials group longitudinal linked randomized trials (ALLRT): rationale, design, and baseline characteristics. HIV Clin Trials 2008; 9:269–282.
36. Strategies for Management of Antiretroviral Therapy (SMART) Study Group, El-Sadr WM, Lundgren JD, Neaton JD, Gordin F, Abrams D, Arduino RC, et al
. CD4+count-guided interruption of antiretroviral treatment
. N Engl J Med
37. Robins JM, Rotnitzky A, Zhao LP. Analysis of semiparametric regression models for repeated outcomes in the presence of missing data. J Am Stat Assoc 1995; 90:106–121.
38. Cole RC, Hernan MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol 2008; 186:656–664.
39. Trial International Co-ordinating Committee INITIO, Yeni P, Cooper DA, Aboulker JP, Babiker AG, Carey D, Darbyshire JH, et al
. Virological and immunological outcomes at 3 years after starting antiretroviral therapy with regimens containing nonnucleoside reverse transcriptase inhibitor, protease inhibitor, or both in INITIO: open-label randomised trial. Lancet 2006; 368:287–298.
40. MacArthur RD, Novak RM, Peng G, Chen L, Xiang Y, Hullsiek KH, et al
. A comparison of three highly active antiretroviral treatment strategies consisting of nonnucleoside reverse transcriptase inhibitors, protease inhibitors, or both in the presence of nucleoside reverse transcriptase inhibitors as initial therapy (CPCRA 058 FIRST Study): a long-term randomized trial. Lancet 2006; 368:2125–2135.
41. When To Start Consortium, Sterne JA, May M, Costagliola D, de Wolf F, Phillips AN, Harris R, et al
. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies
42. HIV-causal Collaboration (2010). The effect of combined antiretroviral therapy on the overall mortality of HIV-infected individuals