JAIDS Journal of Acquired Immune Deficiency Syndromes:
Comparative Effectiveness of Initial Antiretroviral Therapy Regimens: ACTG 5095 and 5142 Clinical Trials Relative to ART-CC Cohort Study
Mugavero, Michael J. MD*; May, Margaret PhD†; Ribaudo, Heather J. PhD‡; Gulick, Roy M. MD, MPH§; Riddler, Sharon A. MD‖; Haubrich, Richard MD¶; Napravnik, Sonia PhD#; Abgrall, Sophie MD, PhD**; Phillips, Andrew PhD††; Harris, Ross MSc‡‡; Gill, M. John MD§§; de Wolf, Frank MD, PhD‖‖; Hogg, Robert PhD¶¶; Günthard, Huldrych F. MD##; Chêne, Geneviève MD, PhD***; D'Arminio Monforte, Antonella MD†††; Guest, Jodie L. PhD‡‡‡; Smith, Colette PhD††; Murillas, Javier MD§§§; Berenguer, Juan MD‖‖‖; Wyen, Christoph MD¶¶¶; Domingo, Pere MD###; Kitahata, Mari M. MD, MPH****; Sterne, Jonathan A. C. PhD†; Saag, Michael S. MD*; AIDS Clinical Trial Group DACS 241 Team; AIDS Clinical Trial Group Study 5095 Team; AIDS Clinical Trial Group Study 5142 team; The Antiretroviral Cohort Collaboration
*Division of Infectious Disease, Department of Medicine, University of Alabama, Birmingham, AL
†School of Social and Community Medicine, University of Bristol, Bristol, United Kingdom
‡Center for Biostatistics in AIDS Research, Harvard School of Public Health, Boston, MA
§Division of Infectious Diseases, Weill Medical College of Cornell University, New York, NY
‖Division of Infectious Diseases, Department of Medicine, University of Pittsburgh, PA
¶Division of Infectious Diseases University of California, San Diego, CA
#Division of Infectious Diseases, University of North Carolina at Chapel Hill, Chapel Hill, NC
**INSERM U943, Paris, F-75 013, France; UPMC Univ Paris 06, UMR_S 943, Paris, F-75013 France; AP-HP, Hôpital Avicenne, Service des maladies infectieuses et tropicales, Bobigny, F-93 000 France
††UCL Medical School, London, United Kingdom
‡‡Health Protection Agency, Colindale, London, United Kingdom
§§Division of Infectious Diseases, University of Calgary, Calgary, Canada
‖‖Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
¶¶Division of Epidemiology and Population Health, British Columbia Centre for Excellence in HIV/AIDS, Vancouver, Canada; and Faculty of Health Sciences, Simon Fraser University, Burnaby, Canada
##Division of Infectious Diseases and Hospital Epidemiology, University Hospital Zurich, University of Zurich, Zurich, Switzerland
***INSERM, U897, University Bordeaux Segalen, Bordeaux, France
†††Clinic of Infectious Diseases and Tropical Medicine, San Paolo Hospital, University of Milan, Milan, Italy
‡‡‡HIV Atlanta VA Cohort Study, Atlanta Veterans Affairs Medical Center, Decatur, GA
§§§Division of Infectious Diseases, Hospital Son Dureta, Palma de Mallorca, Spain
‖‖‖Infectious Diseases Unit, Hospital General Universitario Gregorio Marañón, Madrid, Spain
¶¶¶First Department of Internal Medicine, University of Köln, Köln, Germany
###Infectious Diseases Unit, Hospital de la Santa Creu i Sant Pau, Av Sant Antoni Maria Claret, Barcelona, Spain
****Department of Medicine, Division of Infectious Diseases, University of Washington, Washington, DC
The ACTG is supported by grant AI-68636 from the US National Institutes of Health, with additional support provided by grant K23MH082641 (MJM), U01AI068634 (HJR), and K24AI-51966 (RMG). The ART Cohort Collaboration is supported by the UK Medical Research Council grant G0700820. Sources of funding of individual cohorts include the Agence Nationale de Recherche contre le SIDA, the Institut National de la Santé et de la Recherche Médicale (INSERM), the French, Italian, Spanish and Swiss Ministries of Health, The Swiss HIV Cohort Study (SHCS), supported by the Swiss National Science Foundation (Grant No. 33CSC0- 08787) and by the SHCS research foundation, the Stichting HIV Monitoring, the European Commission, the British Columbia and Alberta Governments, the Michael Smith Foundation for Health Research, the Canadian Institutes of Health Research, the VHA Office of Research and Development and unrestricted grants from GlaxoSmithKline, Roche and Boehringer-Ingelheim. Supported in part by the “Spanish Network for AIDS Research (RIS; ISCIII-RETIC RD06/006).
M.J.M. has received consulting fees (advisory board) from Bristol-Myers Squibb, Gilead Sciences, and Merck Foundation; and grant support from Bristol-Myers Squibb, Pfizer, Inc, Tibotec Therapeutics, and Definicare, LLC; R.M.G. has received consulting fees from Bristol-Myers, Gilead, Merck, Tibotec, ViiV and Virostatics and research grant support (to Weill Cornell Medical College) from Merck, Pfizer, Tibotec, and ViiV; Sophie Abgrall has received support for conference attendance from pharmaceutical companies including Glaxo-SmithKline, Abbott, Tibotec, Gilead and Boehringer Ingelheim; M.J.G. has received consulting fees (advisory boards) from Abbott, Bristol-Myers Squibb, Viiv Healthcare, Janssen, Merck and Gilead and has had grant support from Gilead, ViiV Healthcare and Pfizer; H.F.G. has been an adviser and/or consultant for GlaxoSmithKline, Abbott, Novartis, Boehringer Ingelheim, Roche, Tibotec, and Bristol-Myers Squibb and has received unrestricted research and educational grants from Roche, Abbott, Bristol-Myers Squibb, GlaxoSmithKline, Tibotec, and Merck Sharp and Dohme (MSD) (all money went to institution); G.C. has received consulting fees (Scientific Committee) from Roche, and has had scientific responsibilities in projects receiving specific grant support from Gilead, Tibotec, Boehringer Ingelheim, GlaxoSmithKline, Roche, Pfizer, MSD, Bristol-Myers Squibb, Janssen, ViiV Healthcare, and managed through her Institution or a nonprofit society; C.S. has received consulting fees from Bristol-Myers Squibb, Gilead Sciences, Tibotec, Glaxo Smith Kline, Abbott and Roche and grant support from Bristol-Myers Squibb and Glaxo Smith Kline; C.W. has received consulting fees from Boehringer Ingelheim, fees for speaking engagments from Bristol-Myers Squibb, Gilead Sciences, ViiV Healthcare, MSD, Janssen-Cilag, Essex, Pfizer and Abbott; P.D. has received consulting fees (advisory board) from Bristol-Myers Squibb, Gilead Sciences, Boehringer Ingelheim, Merck, Abbott, Janssen & Cilag, ViiV healthcare and Theratechnologies and grant support from Abbott, Boehringer Ingelheim, Pfizer, Inc, and Tibotec Therapeutics; M.S.S. has received consulting fees from Ardea Biosciences, Avexa, Boehringer-Ingelheim, Bristol-Myers Squibb, Gilead Sciences, GlaxoSmithKline, Merck, Monogram Biosciences, Pain Therapeutics, Panacos, Pfizer, Progenics, Roche Laboratories, Tibotec, Tobira Therapeutics, and Vicro and research support from Achillion Pharmaceuticals, Avexa, Boehringer-Ingelheim, GlaxoSmithKline, Merck, Panacos, Pfizer, Progenics, Theratechnologies and Tibotec. All remaining authors have no conflicts of interest to declare.
The members of the ACTG 5095 and 5142 study team and their support are listed in Appendix 1.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.jaids.com).
Correspondence to: Michael J. Mugavero, MD, MHSc, CCB 142, 908, 20th Street South, Birmingham, AL 35294-2050 (e-mail: email@example.com).
Received April 27, 2011
Accepted July 28, 2011
Background: The generalizability of antiretroviral therapy (ART) clinical trial efficacy findings to routine care settings is not well studied. We compared the relative effectiveness of initial ART regimens estimated in AIDS Clinical Trial Group (ACTG) randomized controlled trials with that among patients receiving ART at Antiretroviral Therapy Cohort Collaboration (ART-CC) study sites.
Methods: Treatment-naive HIV-infected patients initiating identical ART regimens in ACTG trials (A5095 and A5142) and at 15 ART-CC cohort study sites were included. Virological failure (HIV-1 RNA >200 copies/mL) at 24 and 48 weeks, incident AIDS-defining events and mortality were measured according to study design (ART-CC cohort vs. ACTG trial) and stratified by third drug [abacavir (ABC), efavirenz (EFV), and lopinavir/r (LPV/r)]. We used logistic regression to estimate and compare odds ratios (OR) for virological failure between different regimens and study designs, and used Cox models to estimate and compare hazard ratios for AIDS and death.
Results: Compared with patients receiving ABC, those receiving EFV had roughly half the odds of 24-week virologic failure (>200 copies/mL) in both ACTG 5095 (OR = 0.53, 95% confidence interval: 0.36 to 0.79) and ART-CC (0.46, 0.37 to 0.57). Virologic superiority of EFV (vs. ABC) seemed comparable in ART-CC and ACTG 5095 (ratio of ORs 0.86, 95% confidence interval: 0.54 to 1.35). Odds ratios for 48-week virologic failure, comparing EFV with LPV/r, were also comparable in ACTG 5142 and ART-CC (ratio of ORs: 0.87, 0.45 to 1.69).
Conclusions: Between ART regimen virologic efficacy of third drugs ABC, EFV, and LPV/r observed in the ACTG 5095 and 5142 trials seem generalizable to the routine care setting of ART-CC clinical cohorts.
Randomized controlled clinical trials are the cornerstone of evidence-based medicine and are essential to inform HIV antiretroviral treatment (ART) guidelines and clinical practice decisions.1–3 For over 2 decades, the Adult AIDS Clinical Trial Group (ACTG) has been a leading organization in the conduct of clinical trials, including those comparing the efficacy of initial ART regimens (https://actgnetwork.org/). Because of the potential for selection bias imposed by trial eligibility criteria and volunteer bias for participation in clinical trials, there is always uncertainty whether clinical trial findings will be generalizable to the broader patient population treated through routine clinical care outside the context of a study.4,5 Regardless, randomized controlled trials remain the optimal means to compare the efficacy between treatment strategies and the only study methodology able to directly assess causality.
Observational cohort studies offer a complementary research design that allows for a comparison of the effectiveness of different treatment strategies in routine care settings. Analyses of HIV cohort studies have similarly made important contributions to treatment strategies, particularly in assessing effects on clinical events and mortality, which often cannot be adequately evaluated in clinical trials because of limited duration of RCTs and low event frequency. The Antiretroviral Therapy Cohort Collaboration (ART-CC) has been a leading, international multisite HIV cohort study for over a decade (http://www.art-cohort-collaboration.org).6,7 A notable limitation inherent to cohort studies is the potential for confounding by indication and unmeasured confounding, which may impact outcome interpretation and reliability of study findings.4,5,8,9 Regardless, well-conducted observational cohort studies play an important role in HIV treatment decisions as they are inclusive and reflective of treatment responses and outcomes of the broader population of HIV-infected persons than typically studied through clinical trials and generally can provide longer follow-up.
In recent years, considerable emphasis has been placed on the importance of comparative effectiveness research (CER) to allow for informed treatment decisions to improve individual and population level health.10 Inherent to the definition of CER is the direct comparison of alternative treatments in patients typical of those treated in day-to-day clinical care. Among the priority areas in the CER agenda is the development and evaluation of methodologies of clinical research that address limitations of existing study designs to generate novel data elements to augment the traditional evidence base thereby fostering more informed treatment decisions.10 Here, we compare patient-level virologic and clinical effectiveness of a number of initial ART regimens estimated in ACTG clinical trials with that estimated in patients treated in routine clinical care and enrolled in cohorts participating in the ART-CC. To our knowledge, this is the first large-scale regimen-level comparison of contemporary initial ART regimens among patients receiving treatment through clinical trials versus routine care. These analyses address the pivotal questions of whether ART efficacy findings observed in clinical trials translate to routine care settings, and whether differential ART regimen-level effects are observed across study designs and treatment settings.
Setting and Participants
AIDS Clinical Trial Group (ACTG) Study 5095
ACTG 5095 is a randomized double-blind study that compared 3 ART regimens for the initial treatment of subjects infected with HIV-1 and has been described in detail.11,12 We used data from the 2 arms of the trial that compared 3 drug regimens: zidovudine (AZT), lamivudine (3TC), and abacavir (ABC) with AZT + 3TC plus efavirenz (EFV).
Eligible patients were HIV-1–infected adults who had received no previous ART and who had a plasma HIV-1 RNA level of at least 400 copies per milliliter and acceptable safety laboratory results across a range of different measures. Patients were excluded if they had received immunomodulator or investigational therapy or vaccines within the previous 30 days, if they weighed less than 40 kg or if they were pregnant or breastfeeding. Subjects enrolled in the study between March 2001 and November 2002.
ACTG Study 5142
ACTG 5142 is a randomized open-label trial that compared 3 ART regimens for the initial treatment of subjects infected with HIV-1 and has been described in detail.13 We used data from the 2 arms of the trial that compared EFV with lopinavir-boosted with ritonavir (LPV/r) each together with 2 nucleoside reverse transcriptase inhibitors (NRTIs) [3TC with either AZT or stavudine (D4T) or tenofovir (TDF)].
The study population consisted of HIV-1–infected patients at least 13 years of age who had not received previous ART. All patients had a plasma HIV-1 RNA level of at least 2000 copies per milliliter with any CD4 cell count and acceptable laboratory results across a range of different measures. Patients were enrolled from January 2003 to May 2004.
ART-CC Cohort Collaboration
ART-CC is an international collaboration between the investigators of cohort studies of HIV-1–infected patients from Europe and North America that was established in 2000 to estimate prognosis of HIV-1–infected, treatment-naive patients initiating combination ART. The collaboration has been described in detail elsewhere.6,7 Prospective cohort studies were eligible for participation if they had enrolled at least 100 HIV-1–infected patients aged 16 years or older who had not previously received ART, started ART with a combination of at least 3 antiretroviral drugs after 1996, and been followed for a median duration of at least 1 year after ART initiation. The dataset analyzed here was assembled during 2009 and included data from 15 cohorts: the AIDS Therapy Evaluation Project, Netherlands (ATHENA),14 the Agence Nationale de la Recherche sur le SIDA et les hépatites virales CO3 Aquitaine Cohort,15 the Agence Nationale de la Recherche sur le SIDA et les hépatites CO4 French Hospital Database on HIV,16 the Italian Cohort of Antiretroviral-Naive Patients,17 the Köln/Bonn Cohort, Germany,18 the Proyecto para la Informatización del Seguimiento Clínico-epidemiológico de la Infección por HIV y SIDA Cohort,19 Cohorte de la Red de Investigación en Sida (CoRIS Cohort),20 VIH-Aplicación de Control Hospitalario Cohort, Spain, the Royal Free Hospital Cohort, United Kingdom,21 the British Columbia Center for Excellence in HIV,22 the South Alberta Cohort,23 Canada, the Swiss HIV Cohort Study,24 the 1917 Clinic Cohort from the University of Alabama,25 the HIV Atlanta VA Cohort Study,26 and the University of Washington HIV Cohort, United States. At all sites, institutional review boards approved the collection of data. All cohorts provided anonymized data on a predefined set of demographic, laboratory, and clinical variables, which were then pooled and analyzed centrally.
All patients enrolled in the specified arms of ACTG 5095 and in ACTG 5142 and who had measurements of CD4 count and HIV-1 RNA at start of ART were included. Patients in ART-CC included in this study initiated ART after January 1, 2000, and had at least 1 year of follow-up before end of study. Comparator patients in ART-CC started on the same regimens as in the corresponding trial: for ACTG 5095 EFV or ABC together with AZT + 3TC, and for ACTG 5142 EFV or LPV/r together with 3TC plus choice of AZT or D4T or TDF. Our analysis included all subjects meeting our eligibility criteria from each study. Formal matching according to ART regimen constituents was not done, as we did not conduct a case–control study. Potential imbalances across study designs with respect to regimen constituents were controlled for in multivariate analyses.
In all analyses, we used an intent-to-continue-treatment approach, and thus ignored changes to treatment regimen, including treatment interruptions and terminations. In ART-CC, baseline measurements of CD4 count and viral load were the nearest to date of starting ART within 3 months before start date. Measurements at 24 and 48 weeks were the nearest within a window of ±7 weeks. Primary analyses included patients with available HIV RNA measures, and patients with missing outcome data were excluded.
Comparison of ACTG 5095 and ART-CC
We defined virologic failure as a single HIV-1 RNA level >200 copies per milliliter at 24 weeks. Because of early termination of the ABC arm in A5095,12 24-week virologic failure was selected over 48 weeks as the primary outcome measure. We used logistic regression to estimate crude and adjusted odds ratios (ORs) for virological failure comparing EFV versus ABC as a third drug in both ACTG 5095 “trial” and ART-CC “cohort”. Models were adjusted for year of starting ART, age, sex, assumed transmission via injection drug use, prior AIDS diagnosis, CD4 count, and HIV RNA at start of ART. We estimated ratios of OR (OR in cohort\OR in trial) to compare relative effects of ART regimens in the cohort and trial settings; formal criteria for these comparisons are not established. We also estimated the OR for virological failure comparing trial with cohort, separately in patients on EFV and on ABC.
We used Cox proportional hazards models to estimate crude and adjusted hazard ratios (HRs) comparing the effect of ART regimens on the clinical endpoint of time to AIDS or death, in both the trial and the cohort. We estimated ratios of HR (HR in cohort\HR in trial) to compare the relative effects of ART regimens in the cohort and trial settings. We estimated HR for AIDS or death comparing the cohort with trial, within strata defined by drug regimens. As there were few deaths in the trial, mortality HRs were only estimated using cohort data (analysis stratified by constituent cohorts).
Comparison of ACTG 5142 and ART-CC
Analyses described above were repeated to compare EFV with LPV/r as a third drug using data from ACTG 5142 and ART-CC. The primary end point for these analyses was 48-week virological failure, defined as a single HIV-1 RNA level >200 copies per milliliter. We also investigated progression to AIDS and death.
Analyses were repeated using a higher threshold for virological failure (HIV-1 RNA > 400 copies/mL), which allowed the inclusion of additional patients from ART-CC who had viral load measured using a less sensitive assay. We repeated all analyses stratifying by baseline HIV-1 RNA (<100,000, ≥ 100,000 copies/mL) and CD4 count (<200, ≥200 cells/mL). Finally, we conducted sensitivity analyses using a 3-month window around the 24-week and 48-week endpoints and also conducted analyses carrying the last viral load value forward to evaluate the potential impact of missing viral load measurements using a ±7-week window.
Role of the Funding Source
This study was supported by the UK Medical Research Council and the US National Institutes of Health, neither of which played a role in the study's design, conduct, and reporting.
Patient characteristics are presented by third drug (ABC, EFV, and LPV/r) and study design (Table 1); ACTG 5095 (n = 753) and identical regimens in ART-CC (n = 4610), and ACTG 5142 (n = 498) and identical regimens in ART-CC (n = 8212). In general, the proportion of female patients was slightly higher for ART-CC than for ACTG, whereas median age at ART initiation was similar (roughly 38 years) across arms and study designs. Median CD4 counts at ART initiation were approximately 200 cells per milliliter in all ACTG treatment arms and among those receiving EFV in ART-CC, with notable differences among ABC (median 250 cells/mL) and LPV/r (median 146 cells/mL) treated patients in the ART-CC. Virologic failure (>200 copies/mL) was higher among ABC-treated patients in both ACTG 5095 and ART-CC relative to EFV and LPV/r–treated patients in A5095, A5142, and ART-CC. AIDS-defining events and deaths were relatively infrequent at 48 weeks after ART initiation and notably higher among ART-CC patients.
Adjusted estimates of virologic effectiveness stratified by third drug (Table 2) showed that patients receiving EFV had roughly half the odds of 24-week virologic failure (>200 copies/mL) compared with ABC-treated patients for both ACTG 5095 (OR = 0.53, 95% confidence interval (CI) = 0.36 to 0.79) and ART-CC (0.46, 0.37 to 0.57). The ratio of ORs (0.86, 0.54 to 1.35), (cohort study effectiveness/clinical trial efficacy, or effectiveness/efficacy ratio), indicates that the virologic superiority of EFV versus ABC was comparable in ART-CC and ACTG 5095 (Fig. 1). Adjusted estimates of 48-week virologic effectiveness showed that the odds of failure were similar for EFV compared with LPV/r in ACTG 5142 (0.97, 0.51 to 1.85), but somewhat lower for EFV compared with LPV/r in ART-CC (0.84, 0.71 to 1.00). However, the ratio of ORs (0.87, 0.45 to 1.69) did not provide evidence that cohort effectiveness differed from trial efficacy (Fig. 1). Adjusted analyses comparing study design (ACTG 5095 versus ART-CC and A5142 versus ART-CC, Table 2), found comparable odds of virologic failure in trials compared with the cohort study across all treatment regimens.
When comparing 48-week AIDS-defining events or deaths stratified by third drug (Table 3), there was clear evidence of confounding by indication in estimates based on cohort data, suggested by sizeable shifts in parameter estimates in adjusted analyses. Consistent with the patterns of prognostic factors in the cohort data presented in Table 1, the effect of EFV compared with ABC was more beneficial after adjustment, whereas the effect of EFV compared with LPV was attenuated toward 1. Estimates from trials were imprecise because of the small number of events. In adjusted analyses, patients receiving EFV had lower rates of AIDS and death compared with those receiving ABC in both ACTG 5095 (0.60, 0.26 to 1.41) and ART-CC (0.73, 0.54 to 0.99). Rates of AIDS and death seemed similar in patients treated with EFV and LPV/r in both ACTG 5142 (0.96, 0.40 to 2.30) and ART-CC (0.88, 0.73 to 1.06). Effectiveness estimated in ART-CC seemed similar to efficacy estimated in ACTG trials for each regimen comparison (ratios of ORs EFV versus ABC; 1.21, 0.50 to 2.95 and EFV versus LPV/r; 0.92, 0.39 to 2.17).
Adjusted analyses of all-cause 48-week mortality by ART regimen were restricted to patients in ART-CC (because there were insufficient deaths in both ACTG 5095 and ACTG 5142). There was little evidence of between-regimen differences in mortality rates (EFV vs. ABC, 0.81, 0.48 to 1.38; EFV versus LPV/r 0.85, 0.63 to 1.15), although CIs were wide.
Findings from sensitivity analyses using a virologic failure threshold of 400 copies per milliliter were largely in accordance to those from primary analyses (see Appendix, Supplemental Digital Content 1, http://links.lww.com/QAI/A219). Analyses of virological failure stratified by viral load and CD4 count at ART initiation showed that the benefit of EFV over ABC is greater in patients with viral load ≥100,000 (vs. <100,000) copies/mL and in those with CD4 count <200 (vs. ≥200) cells per milliliter (see Appendix, Supplemental Digital Content 2, http://links.lww.com/QAI/A220). Finally, at viral load ≥100,000 (vs. <100,000) copies per milliliter, more disparate findings in comparisons of EFV versus LPV/r stratified by study design and cohort versus trial stratified by regimen are observed relative to primary analyses, albeit with wide CIs (see Appendix, Supplemental Digital Content 2, http://links.lww.com/QAI/A220).
Among patients initiating antiretroviral therapy for HIV infection, differences in virologic and clinical effectiveness between ART regimens compared in the ACTG 5095 and ACTG 5142 were similar to observed differences when comparing identical regimens administered in routine care in the ART Cohort Collaboration. The virologic superiority of EFV over ABC observed in ACTG 5095 was similar to that seen in the ART-CC, suggesting the findings of this trial translated well to a routine care setting. The comparable efficacy of EFV and LPV/r as a third drug paired with 2 NRTIs in ACTG 5142 was mirrored by the similar virological and clinical effectiveness of these regimens in ART-CC. To our knowledge, this is the first large-scale regimen level evaluation of the comparative effectiveness of initial ART when administered in a clinical trial and routine care settings. Our finding that ART regimen differences in the ACTG clinical trials correlated with the routine care setting of ART-CC suggests the generalizability of the results and provides an important link between clinical trials and their applicability to a broader patient population.
Although comparisons of clinical trial efficacy with routine care effectiveness are well documented for other medical conditions,27–30 evaluation of ART regimens for the treatment of HIV infection in clinical trials versus routine clinical care has not been widely studied.31,32 A recent study from a single academic US HIV clinic found similar rates of virologic suppression among patients receiving ART through clinical trials and routine care, but could only assess treatment strategy (trial vs. routine care) as the sample was too small for regimen level comparisons.32 Because care providers must choose among numerous initial ART regimen options, the ability of the current study to move beyond treatment strategy to the regimen level should provide valuable information to inform HIV treatment decisions in accordance with the goals of comparative effectiveness research.10 In the current study, the clinical trial efficacy of evaluated regimens (EFV, ABC, and LPV/r paired with 2 NRTIs) was mirrored by the effectiveness in routine care settings.
In the evaluation of virologic effectiveness, the definition of virologic failure and analytic approach to assess failure in the current study differed from the primary analyses in the original ACTG 5095 and 5142 studies.11–13 This is particularly noteworthy with regards to the comparison of virologic effectiveness between EFV vs. LPV/r. In the original ACTG 5142 study, EFV was found to have superior virologic efficacy relative to LPV/r using survival methods and with assessment of virologic failure starting at 32 weeks after ART start.13 Of note, as seen with the current analyses, similar rates of cross-sectional 48-week virologic failure were observed between the EFV and LPV/r groups in original analyses of the ACTG 5142 study, although at 96 weeks patients receiving EFV were significantly more likely to achieve a viral load <200 copies per milliliter (93% vs. 86%, P = 0.04).13 The analytic approach employed for the current study, cross-sectional evaluation of unconfirmed virologic failure (>200 copies/mL) at 48 weeks after ART start, was selected based on availability of plasma HIV RNA measures in the ART-CC. In contrast to the ACTG studies, plasma HIV RNA measures are obtained considerably less frequently in routine care, making the definition and analytic methods employed in clinical trials impractical for cohort data based on the limited availability of outcome measures.
Several strengths of our study are noteworthy. This study represents an initial collaboration between the Adult ACTG and ART-CC to allow for innovative approaches to evaluate ART regimen performance at the patient level. The evaluation of clinical events in additional to virologic failure adds contextual richness to the between ART regimen comparisons and provides important information for patients and providers. Beyond the evaluation of between ART regimen differences in surrogate HIV biomarkers across study design, evaluation of clinical events has important implications for patient health. The conduct of sensitivity analyses, the results of which are largely in accordance with findings from primary analyses (see web appendices), provides additional confidence in the interpretation of study findings.
Our study has limitations. As with all observational studies, there is potential for unmeasured confounding that may bias estimates from cohort data.4,5 The imbalances in prognostic factors between cohort patients who received different regimens were reflected in differences between the crude and adjusted hazard ratios for clinical events presented in Table 3. However, we cannot exclude the possibility that further unmeasured prognostic factors were used by physicians choosing between different ART regimens in routine care settings. Many of the NRTIs evaluated in the current study are not among those recommended in updated HIV treatment guidelines.1–3 Future analyses should assess more modern ART regimens. ART-CC includes patients in Europe, Canada, and the United States, whereas with the exception of a handful of individuals in A5142 enrolled in South Africa, ACTG sites in these studies are restricted to the United States and Puerto Rico. Analyses were conducted according to intent-to-continue-treatment principles and ignoring missing outcome data as done for the original ACTG clinical trials such that the impact of treatment changes, terminations, and missing data were not evaluated in the current analyses. On-going studies in the ART-CC are evaluating ART interruptions, terminations, and switches. Because patient follow-up in ACTG studies continues beyond initial treatment change or failure, future analyses may be possible to evaluate these factors. At 24 weeks, missing viral load data was observed in 7% of ACTG participants and 31% of ART-CC patients, which may impact outcomes interpretation. However, missing viral load data was observed in 7% of ACTG participants and 13% of ART-CC patients in sensitivity analyses using a 3-month window around the 24-week endpoint. Sensitivity analyses using this wider viral load measurement window and those using a last viral load carried forward approach yielded findings consistent with primary analyses (see Appendix, Supplemental Digital Content 3, http://links.lww.com/QAI/A221). CIs were fairly wide, particularly for the ratio of odds and hazards ratios, limiting the precision of parameter estimates. Although caution must be exercised when comparing findings across studies and designs, we suggest our analyses have an important role and are particularly germane in light of considerable recent emphasis on comparative effectiveness research and methodology.10
In conclusion, our study found ART regimen virologic and clinical efficacy for ABC, EFV, and LPV/r in combination with 2 NRTIs observed in ACTG 5095 and 5142 clinical trials were mirrored when these regimens were administered in the routine care setting at ART-CC clinical sites. The generalizability of findings for these trials to routine care settings suggests ART regimen performance in clinical trials likely translates to routine care settings. Although additional studies are needed to confirm our findings and evaluate other ART regimens, we believe our study provides pivotal new evidence demonstrating the comparative effectiveness of ARTs evaluated in clinical trials and a large clinical cohort study; this research will help inform ART treatment decisions for HIV-infected patients in routine clinical care.
We are grateful to all patients, doctors, nurses, pharmacists, laboratories, and other study personnel who were involved in the participating studies.
2. Gazzard BG, Anderson J, Babiker A, et al. British HIV association guidelines for the treatment of HIV-1-infected adults with antiretroviral therapy 2008. HIV Med. 2008;9:563–608
3. Thompson MA, Aberg JA, Cahn P, et al. Antiretroviral treatment of adult HIV infection: 2010 recommendations of the international AIDS Society-USA panel. JAMA. 2010;304:321–333
4. D'Agostino RB Jr, D'Agostino RB Sr. Estimating treatment effects using observational data. JAMA. 2007;297:314–316
5. Pocock SJ, Elbourne DR. Randomized trials or observational tribulations? N Engl J Med. 2000;342:1907–1909
6. Chene G, Sterne JA, May M, et al. Prognostic importance of initial response in HIV-1 infected patients starting potent antiretroviral therapy: analysis of prospective studies. Lancet. 2003;362:679–686
7. Egger M, May M, Chene G, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet. 2002;360:119–129
8. Walker AM. Confounding by indication. Epidemiology. 1996;7:335–336
9. Wood E, Hogg RS, Heath KV, et al. Provider bias in the selection of non-nucleoside reverse transcriptase inhibitor and protease inhibitor-based highly active antiretroviral therapy and HIV treatment outcomes in observational studies. AIDS. 2003;17:2629–2634
10. Sox HC, Greenfield S. Comparative effectiveness research: a report from the Institute of Medicine. Ann Intern Med. 2009;151:203–205
11. Gulick RM, Ribaudo HJ, Shikuma CM, et al. Three- vs four-drug antiretroviral regimens for the initial treatment of HIV-1 infection: a randomized controlled trial. JAMA. 2006;296:769–781
12. Gulick RM, Ribaudo HJ, Shikuma CM, et al. Triple-nucleoside regimens versus efavirenz-containing regimens for the initial treatment of HIV-1 infection. N Engl J Med. 2004;350:1850–1861
13. Riddler SA, Haubrich R, DiRienzo AG, et al. Class-sparing regimens for initial treatment of HIV-1 infection. N Engl J Med. 2008;358:2095–2106
14. van Sighem A, Zhang S, Reiss P, et al. Immunologic, virologic, and clinical consequences of episodes of transient viremia during suppressive combination antiretroviral therapy. J Acquir Immune Defic Syndr. 2008;48:104–108
15. Bonnet F, Thiebaut R, Chene G, et al. Determinants of clinical progression in antiretroviral-naive HIV-infected patients starting highly active antiretroviral therapy. Aquitaine Cohort, France, 1996–2002. HIV Med. 2005;6:198–205
16. Kousignian I, Abgrall S, Grabar S, et al. Maintaining antiretroviral therapy reduces the risk of AIDS-defining events in patients with uncontrolled viral replication and profound immunodeficiency. Clin Infect Dis. 2008;46:296–304
17. d'Arminio Monforte A, Lepri AC, Rezza G, et al. Insights into the reasons for discontinuation of the first highly active antiretroviral therapy (HAART) regimen in a cohort of antiretroviral naive patients. I.CO.N.A. Study Group. Italian Cohort of Antiretroviral-Naive Patients. AIDS. 2000;14:499–507
18. Fatkenheuer G, Theisen A, Rockstroh J, et al. Virological treatment failure of protease inhibitor therapy in an unselected cohort of HIV-infected patients. AIDS. 1997;11:F113–F116
19. Jaen A, Esteve A, Miro JM, et al. Determinants of HIV progression and assessment of the optimal time to initiate highly active antiretroviral therapy: PISCIS Cohort (Spain). J Acquir Immune Defic Syndr. 2008;47:212–220
20. Caro-Murillo AM, Castilla J, Perez-Hoyos S, et al. Spanish cohort of naive HIV-infected patients (CoRIS): rationale, organization and initial results. [in Spanish] Enferm Infecc Microbiol Clin. 2007;25:23–31
21. Booth CL, Garcia-Diaz AM, Youle MS, et al. Prevalence and predictors of antiretroviral drug resistance in newly diagnosed HIV-1 infection. J Antimicrob Chemother. 2007;59:517–524
22. Alfonso V, Geller J, Bermbach N, et al. Becoming a “treatment success”: what helps and what hinders patients from achieving and sustaining undetectable viral loads. AIDS Patient Care STDS. 2006;20:326–334
23. Mocroft A, Gill MJ, Davidson W, et al. Predictors of a viral response and subsequent virological treatment failure in patients with HIV starting a protease inhibitor. AIDS. 1998;12:2161–2167
24. Schoeni-Affolter F, Ledergerber B, Rickenbach M, et al. Cohort profile: the Swiss HIV cohort study. Int J Epidemiol. 2010;39:1179–1189
25. Chen RY, Accortt NA, Westfall AO, et al. Distribution of health care expenditures for HIV-infected patients. Clin Infect Dis. 2006;42:1003–1010
26. Anderson KB, Guest JL, Rimland D. Hepatitis C virus coinfection increases mortality in HIV-infected patients in the highly active antiretroviral therapy era: data from the HIV Atlanta VA Cohort Study. Clin Infect Dis. 2004;39:1507–1513
27. Bahit MC, Cannon CP, Antman EM, et al. Direct comparison of characteristics, treatment, and outcomes of patients enrolled versus patients not enrolled in a clinical trial at centers participating in the TIMI 9 Trial and TIMI 9 Registry. Am Heart J. 2003;145:109–117
28. Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. N Engl J Med. 2000;342:1878–1886
29. Hallfors D, Cho H, Sanchez V, et al. Efficacy vs effectiveness trial results of an indicated “model” substance abuse program: implications for public health. Am J Public Health. 2006;96:2254–2259
30. Steg PG, Lopez-Sendon J, Lopez de Sa E, et al. External validity of clinical trials in acute myocardial infarction. Arch Intern Med. 2007;167:68–73
31. Hansen AB, Gerstoft J, Kirk O, et al. Unmeasured confounding caused slightly better response to HAART within than outside a randomized controlled trial. J Clin Epidemiol. 2008;61:87–94
32. Routman JS, Willig JH, Westfall AO, et al. Comparative efficacy versus effectiveness of initial antiretroviral therapy in clinical trials versus routine care. Clin Infect Dis. 2010;50:574–584
ACTG 5095 and 5142 Study Team ACTG 5095 Study Acknowledgments and Support
We gratefully acknowledge the efforts of the ACTG 5095 protocol team, the participating AIDS Clinical Trial Units of the ACTG, and the patient volunteers. ACTG 5095 was supported by National Institute of Allergy and Infectious Diseases (NIAID) grants AI38858 and AI068636). Bristol-Myers Squibb, Boehringer-Ingleheim, and GlaxoSmithKline provided study drugs. Bristol-Myers Squibb and GlaxoSmithKline also provided funding for plasma HIV-1 RNA assays and resistance testing.
ACTG 5095 Study Team
Roy M. Gulick (Chair), Cecilia M. Shikuma (Co-chair), Heather Ribaudo, Christina Lalama, Karin K. Klingman, Barbara Bastow, Anne Kmack, William A. Meyer, Daniel R. Kutitzkes, Edward P. Acosta, Valery Hughes, Kathleen E. Squires, Bruce R. Shackman, Jeffrey T. Schouten, Vincent Parrillo, Ana I. Martinez, Richard Fallis, Stephen P. Storfer, Michael Giordano, Marita McDonough, James Rooney, Lynn Rugh, Kirk Ryan, Jerry Tolson, Amy S. van Kempen, Carol Schnizlein Bick, Nancy Webb.
ACTG 5142 Study Acknowledgements and Support
We gratefully acknowledge the efforts of the ACTG 5142 protocol team, the participating AIDS Clinical Trial Units of the ACTG, and the patient volunteers. ACTG 5142 was supported by National Institute of Allergy and Infectious Diseases (NIAID) grant AI68636).
ACTG 5142 Study Team
Sharon A. Riddler (Co-chair), Richard Haubrich (Co-chair), A. Gregory DiRienzo, Lynne Peeples, William G. Powderly, Karin L. Klingman, Kevin W. Garren, Tania George, James F. Rooney, Barbara Brizz, Umesh G. Lalloo, Robert L. Murphy, Susan Swindells, Diane Havlir, John W. Mellors.
antiretroviral therapy; AIDS; comparative effectiveness; HIV; viral load
Supplemental Digital Content
© 2011 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.