The Effect of Protease Inhibitor-Based Dual Antiretroviral Regimens on CD4/CD8 Ratio During the First Year of Therapy in ART-Naïve Patients With HIV-Infection

MI Figueroa; A Camiro-Zuñiga; PF Belaunzaran-Zamudio; J Sierra Madero; J Andrade Villanueva; JR Arribas; JR Lama; DM Cecchini; G Lopardo; B Crabtree-Ramírez; A Gun; P Patterson; VI Fink; OG Sued; P Cahn


HIV Medicine. 2021;22(4):254-261. 

In This Article


Study Design and Settings

The GARDEL (Global AntiRetroviral Design Encompassing Lopinavir/r and Lamivudine vs LPV/r based standard therapy; NCT01237444) and ANDES (Estudio Abierto para Sujetos HIV + Naive de Tratamiento Antirretroviral Basado en la Combinación de DRV/RTV/3TC, para Explorar Tolerabilidad, Eficacia y Seguridad; NCT02770508) studies were both phase 4 randomized, controlled, open-label, non-inferiority, multicentre clinical trials that compared the safety, tolerability, antiviral activity and emergence of HIV-1 resistance of boosted PI-based DT with standard TT regimens in ART-naïve HIV-1-infected patients.[7,11]

The GARDEL trial compared a DT regimen of lopinavir/ritonavir (LPV/r) 400/100 mg + 3TC 150 mg twice daily vs. a TT regimen of LPV/r 400/100 mg twice daily + 3TC or emtricitabine (FTC) + one other, investigator-selected nucleoside reverse transcriptase inhibitor (NRTI) in fixed-dose combination, including zidovudine (ZDV)/3TC, abacavir (ABC)/3TC, or tenofovir disoproxil fumarate (TDF)/3TC. The study enrolled participants from 19 centres in six countries: Argentina, Chile, Mexico, Peru, Spain and the USA. The GARDEL study concluded its follow-up period (96 weeks) for all the participating patients (TT, 206; DT, 216; total, 422).

The ANDES trial compared a DT regimen of DRV/r 800/100 mg + 3TC 300 mg once daily with a TT regimen of DRV/r 800/100 mg + 3TC/TDF 300/300 mg or FTC/TDF 200/300 mg once daily. The study was performed in Buenos Aires, Argentina in six different centers. The initial design established a first stage with 144 participants, followed by a second stage that would enroll 190 participants if the efficacy of the first stage participants in the dual arm reached at least a 75% of patients with viral suppression. ANDES study stage 1 concluded its follow-up period at the 48-week mark for all the participating patients (TT = 70, DT = 75, total = 145).

For both studies, randomization was stratified according to baseline viral load (< 100 000 vs. ≥ 100 000 copies/mL). Participants were randomly assigned (1:1) in blocks of four, by centre, by a central computer-generated randomization table.

Ethical Approval and Informed Consent

This study includes data from two studies (GARDEL and ANDES). Both studies were approved by the Comite de Bioética de Fundación Huésped FWA IORG0001557, IRB00002014 under the numbers GARDEL-FH10 and ANDES-FH15. Ethics committee approval was also obtained at all participating centres in accordance with the principles of the Declaration of Helsinki. Every patient gave written informed consent before undergoing any study procedure.


Patients were eligible for enrolment in both clinical trials if they were infected with HIV-1, ≥ 18 years, naive to ART, had a plasma HIV RNA measurement of at least 1000 copies/mL, did not have an active hepatitis B or hepatitis C infection, were in good general medical health, not pregnant and willing to use two contraceptive methods, had no abnormal laboratory results and did not misuse alcohol or other substance.

Patients were excluded if they were pregnant or breastfeeding, had a confirmed HIV-2 infection, moderate or severe hepatic impairment, any clinically significant active disease, and any active AIDS-associated opportunistic disease within 30 days after screening that, in the investigator's opinion, might have risked the patient's safety, the results of the study or protocol adherence. Patients with evidence of viral resistance to LPV/r, 3TC, FTC or other NRTIs at screening based on the 2009 International Antiviral Society USA (IAS-USA) resistance list were not eligible for the GARDEL study.[7] Likewise, patients who had evidence of resistance to DRV/r, TDF, 3TC or FTC at a screening visit on the basis of the 2013 International Antiviral Society USA (IAS-USA) resistance list were deemed ineligible to participate in the ANDES study.[18] Finally, patients were excluded if they had non-compliance with study procedures, consent withdrawal, demonstration of non-adherence or were lost to follow-up.

The enrolment period for the GARDEL study was from 10 December 2010 to 15 May 2012, while the ANDES participants were enrolled between 1 November 2015 and 30 November 2016.


Patients were followed up at baseline, 4, 8 and 12 weeks and every 12 weeks thereafter. HIV viral loads and CD4 and CD8 counts were measured at each visit. In both trials, patients were censored at viral failure, opportunistic infection and any disease requiring hospitalization or pregnancy. Censorship for whatever reason was similar across treatment groups in both studies so we did not adjust the analysis for differentiated attrition. Viral failure was defined as two consecutive quantitative viral loads ≥ 7 days and < 30 days apart, > 400 copies/mL at week 24 or thereafter or ≥ 50 copies/mL at week 48).

In the GARDEL study, all laboratory tests were processed at a designated local laboratory for each study site. Plasma HIV-RNA concentrations were measured using either the Abbott RealTime HIV-1 assay (Abbott Molecular, Des Plaines, IL, USA), Roche Amplicor assay (Roche Molecular Systems, Branchburg, NJ, USA), or COBAS-TaqMan Assay (HIV-1 Test, version 2.0; Indianapolis, IN, USA), according to availability at each site. Each site used the same assay across the whole duration of the study. In the ANDES study, all samples were processed in the Fundación Huésped laboratory using the Abbott RealTime HIV-1 assay for plasma HIV-RNA concentrations.


We compared patients' CD4/CD8 ratios between DT and TT arms at baseline and at 4, 12, 24, 36 and 48 weeks. We also compared the data from both arms amongst the patients from the GARDEL study at 96 weeks of follow-up.

The main outcomes were median CD4/CD8 ratio and proportion of patients who achieved a CD4/CD8 ratio > 1 measured at 48 weeks' follow-up. The cut-off of a CD4/CD8 ratio > 1 was chosen due to the fact that inverted CD4/CD8 ratios (< 1) represent an immune risk phenotype that is associated with altered immune function, immune senescence and chronic inflammation in both HIV-infected and uninfected populations.[19] In addition, we conducted a sensitivity analysis by calculating the percentage of patients who achieved a CD4/CD8 ratio > 0.5. We also compared absolute CD4 and CD8 counts at each point in time. We performed sub-analysis in groups of people with risk factors for a worse immune reconstitution: ≥ 50 years of age, baseline CD4 count ≤ 200 cells/μL, baseline CD4 count < 350 cells/μL, viral load > 100 000 copies/mL, people with baseline CD4 count ≤ 200 cells/μL and viral load > 100 000 copies/mL, and LPV/r based therapy.

Statistical Analysis

We used mean and standard deviation, or median and interquartile range (IQR), as appropriate, to summarize continuous variables. We used frequencies and simple proportions to summarize categorical variables. We compared CD4/CD8 ratio between treatment arms using the Mann–Whitney U-test and compared the proportion of patients with a CD4/CD8 ratio ≥ 1 with the χ 2 test. In a secondary analysis, we used multi-level models to estimate and compare changes in CD4/CD8 ratios over time. We carried out a multivariate logistic regression model, including age, gender, body mass index (BMI), race, CD4 count, HIV-RNA, and treatment group to identify baseline factors associated with the probability of achieving a CD4/CD8 ratio > 1. We performed all analyses using STATA v.14.0 (StataCorp, College Station, TX, USA).