Phase 2 study of cobicistat versus ritonavir each with once-daily atazanavir and fixed-dose emtricitabine/tenofovir df in the initial treatment of HIV infection : AIDS

Secondary Logo

Journal Logo


Phase 2 study of cobicistat versus ritonavir each with once-daily atazanavir and fixed-dose emtricitabine/tenofovir df in the initial treatment of HIV infection

Elion, Richarda; Cohen, Calvinb; Gathe, Josephc; Shalit, Peterd; Hawkins, Trevore; Liu, Hui C.f; Mathias, Anita A.f; Chuck, Steven L.f; Kearney, Brian P.f; Warren, David R.f for the GS-US-216–0105 Study Team

Author Information
doi: 10.1097/QAD.0b013e32834b4d48
  • Free



Mechanism-based inhibition of cytochrome P4503A (CYP3A) enzymes by low doses (100–400 mg/day) of the HIV-1 protease inhibitor ritonavir (RTV) to pharmacologically enhance (‘boost’) plasma levels of other protease inhibitors that are metabolized by CYP3A enzymes was a serendipitous therapeutic advancement early in the use of protease inhibitors as components of highly active antiretroviral therapy (HAART) [1–5]. Pharmacologic boosting of protease inhibitors has several desirable outcomes: raises pharmacologic barrier to the development of virologic resistance; the protease inhibitor may be administered at lower dose and decreased pill burden; often decreases dosing frequency; and may decrease side-effects, thereby improving a HAART regimen [6]. RTV is used primarily to boost the eight approved protease inhibitors [7–11].

Cobicistat (COBI, formerly GS-9350) is a more specific inhibitor of CYP3A without antiretroviral activity. In-vitro studies demonstrate COBI, compared with RTV, is a weak inhibitor of CYP2D6 (half maximal inhibitory concentration, IC50 = 9.2 μmol/l), does not inhibit other CYP isoforms, and displays low liability for induction; thus, COBI exhibits a pharmacokinetic profile associated with fewer clinically significant drug–drug interactions. Other favorable characteristics of COBI include reduced perturbation of the normal adipocyte functions, such as lipid accumulation and/or response to insulin, which may offer the potential for fewer adverse biochemical effects relative to RTV [12,13].

COBI has high aqueous solubility allowing formulation as a tablet and coformulation with other drugs [12–16]. COBI is being developed specifically as a pharmacoenhancer for drugs metabolized by CYP3A to improve pharmacokinetic exposure to decrease total daily dose and/or less frequent dosing [12,14]. COBI and RTV boost plasma levels of atazanavir (ATV) and darunavir (DRV) comparably when co-administered to HIV-1-uninfected adults [15,16]. These results formed the rationale for initiating a phase 2 clinical trial in HIV-1-infected antiretroviral treatment-naive adults to test efficacy and safety of HAART consisting of COBI compared with RTV as pharmacoenhancers for ATV coadministered with fixed-dose emtricitabine/tenofovir df (FTC/TDF).


Study design

Inclusion criteria were as follows : HIV-1-infected adults (≥18 years), screening plasma HIV-1 RNA at least 5 000 copies/ml, CD4 cell count more than 50 cells/μl, no prior use of approved or experimental anti-HIV drugs and no nucleoside or non-nucleoside reverse transcriptase inhibitor, or primary protease inhibitor genotypic resistance mutations (International AIDS Society - U.S.A. guidelines), normal ECG, estimated glomerular filtration rate (eGFR, Cockcroft–Gault) at least 80 ml/min, aspartate amino transferase or alanine aminotransferase 2.5 times upper limit of normal or less, total bilirubin 1.5 mg/dl or less, and for women, a negative serum pregnancy test. Exclusion criteria were as follows: hepatitis B or C co-infection, new AIDS-defining condition within 30 days of screening, or vaccination within 90 days of study treatment dosing. The study was conducted in the United States, approved by Institutional Review Boards, and participants signed informed consent before screening.

Eligible participants were randomized 2 : 1 (stratified by screening HIV-1 RNA ≤ or > 100 000 copies/ml) to treatment with open-label ATV 300 mg and FTC/TDF 200/300 mg and either COBI 150 mg once-daily (QD) and RTV placebo QD (atazanavir/cobicistat group, ATV/co) or RTV 100 mg QD and COBI placebo QD (atazanavir/ritonavir group, ATV/r). Study assessments occurred at screening, baseline, weeks 2, 4, 8, 12, 16, and then every 8 weeks through week 48, including laboratory analyses (CD4 cell count, hematology, chemistry, and urinalysis; Covance Laboratories, Indianapolis, Indiana, USA), HIV-1 RNA (AMPLICOR HIV-1 Monitor assays; Roche Molecular Systems, Pleasanton, California, USA), and physical examinations. HIV-1 reverse transcriptase and protease genotype (GeneSeq Assay; Monogram Biosciences, South San Francisco, California, USA) was analyzed at screening.

Virologic failure required confirmation and was defined as a suboptimal virologic response (HIV-1 RNA < 1 log10 reduction from baseline and ≥50 copies/ml at the week 8 visit) or virologic rebound at any visit after achieving HIV-1 RNA of less than 50 copies/ml (an increase in HIV-1 RNA to ≥400 copies/ml or a > 1 log10 increase in HIV-1 RNA from nadir). Adverse events and laboratory abnormalities were graded for severity using toxicity grading scale adapted from the Division of AIDS Table for Grading the Severity of Adult and Pediatric Adverse Events (Clarification August 2009, Intensive pharmacokinetic assessments over 24 h occurred in a subset of participants at either week 2, 4, or 8, and trough samples were drawn 20–24-h postdose at weeks 8, 24, and 48.

Statistical methods

Primary efficacy endpoint was as follows: proportion of participants with HIV-1 RNA less than 50 copies/ml at week 24 using point estimates and 95% confidence interval for difference in response rates by normal approximation methods, stratified by baseline HIV-1 RNA level. Secondary endpoints were as follows: proportion of participants with HIV-1 RNA of less than 50 copies/ml at week 48, and CD4+ cell count at weeks 24 and 48. Safety and pharmacokinetic endpoints were summarized using descriptive statistics. Post-hoc intertreatment group statistical comparisons were performed for data presented in Table 1 that were not primary or secondary endpoints, excluding adverse events.

Table 1:
Baseline characteristics, efficacy, and safety results.


Eighty-five participants of 137 screened were randomized 2 : 1 to the ATV/co or ATV/r groups. Six participants randomized to ATV/co never received study treatment. Thus, 79 randomized participants constituted the intent-to-treat analysis efficacy and safety populations: ATV/co (n = 50) and ATV/r (n = 29). Table 1 displays baseline characteristics (demographic and disease), efficacy (primary and secondary endpoints), and safety (adverse events and laboratory abnormalities through week 48) for both treatment groups with post-hoc intertreatment group statistical comparisons. Baseline demographics and disease characteristics were similar between treatment groups. Seventy-six percent of ATV/co participants had HIV-1 RNA 100 000 copies/ml or less and 24% more than 100 000 copies/ml, and ATV/r participants had 62 and 38%, respectively.

The stratum-weighted difference in the response rate (ATV/co–ATV/r) was −1.9% (−18.4 to 14.7%) at week 24. The other efficacy endpoints were similar between the treatment arms. No participant experienced virologic failure.

Ten percent (eight of 79) participants discontinued study treatment prematurely. Five of the eight participants discontinued prior to day 28; five of 50 (10%) ATV/co and three of 29 (10%) ATV/r participants. Three of 79 (4%) participants discontinued study treatment prematurely due to adverse events related to study treatment; two of 50 (4%) ATV/co (moderate vomiting, day 3; severe generalized maculopapular rash, day 11) and one of 29 (3%) ATV/r participants (mild ocular icterus, day 21). Of the remaining five participants, one on each treatment was lost to follow-up, one ATV/co participant withdrew consent and another was discontinued at the investigator's discretion due to nonadherence to protocol, and one ATV/r participant had a protocol violation.

Through 48 weeks, 39 of 50 (78%) ATV/co and 24 of 29 (83%) ATV/r participants experienced at least one treatment-emergent adverse event (TEAE) and the majority was assessed as mild or moderate severity. Three serious adverse events occurred in three participants and were assessed as unrelated to study treatment: pneumonia (one on each treatment) and cellulitis (ATV/co). Study treatment-related TEAEs in at least 5% of participants on either treatment were reported in 18 of 50 (36%) ATV/co and 14 of 29 (48%) ATV/r participants; ocular icterus occurred most commonly six of 50 (12%) and four of 29 (14%) participants, respectively, followed by fatigue, diarrhea, nausea, and flatulence.

Hyperbilirubinemia (grade 1 or greater increase in total bilirubin) occurred in 96% ATV/co and 100% ATV/r participants. Sixty-three percent of ATV/co and 45% of ATV/r participants developed at least grade 3 hyperbilirubinemia. Mean indirect bilirubin (mg/dl) was numerically higher among the ATV/co versus ATV/r participants, but was statistically significant only at week 2 (3.2 and 2.1, respectively) (P = 0.05). Intertreatment group statistical comparison of mean change from baseline in fasting lipids revealed no statistically significant difference. Mean percentage change from baseline in eGFR was significantly decreased only at week 2, −8 (ATV/co group) versus −3% (ATV/r group) (P = 0.02), but not at week 24 or 48.

Atazanavir exposure [maximum concentration, Cmax and area under the plasma concentration-time curve for a dosing interval, AUCtau (mean % coefficient of variation)] was estimated from the intensive pharmacokinetic samples: ATV/co participants (n = 8) Cmax, 3880 ng/ml (36%) and AUCtau 41 300 ng h/ml (33%); and ATV/r participants (n = 10) Cmax 4390 ng/ml (47%) and AUCtau 49 900 ng hr/ml (47%). Figure 1 illustrates the values for geometric mean ATV trough at week 8 for both treatment groups and the table inset shows the ATV concentration before dosing, Ctrough (geometric mean) at weeks 8, 24, and 48 for both treatments as well as their respective geometric mean ratio (90% confidence interval).

Fig. 1:
Atazanavir trough plasma concentrations (ng/ml) at week 8 following administration of ATV/co with TVD and ATV/r with TVD.Data presented as geometric mean (±95% confidence interval). GLS mean, geometric least squares mean; only weeks 8, 24 and 48 trough concentration data with trough sampling time within range 20–24 h after observed dosing time are summarized. Two subjects each had one PK sample (at weeks 8 and 24) not evaluable and thus were excluded. All subjects had evaluable samples at week 48. Plasma concentrations below limit of quantitation (BLQ) were excluded from computing geometric means.


The results of this phase 2 study demonstrated that HIV-1-infected treatment-naive participants using COBI or RTV as pharmacoenhancers for ATV coadministered with FTC/TDF achieved and maintained similar rates of virologic suppression (82 and 86%, respectively) and CD4 cell count increase (230 and 206 cells/μl, respectively) through week 48. These rates of virologic suppression are comparable to 48-week results from three randomized clinical trials, wherein ATV/r and FTC/TDF were compared with another HAART: fosamprenavir/r and FTC/TDF once daily, 83% [17]; lopinavir/r twice daily with FTC/TDF once daily, 78% [18]; and efavirenz and FTC/TDF or ATV/r with abacavir/lamivudine, both once daily, 84% [19].

Statistical analysis of rates of AEs occurring between treatment groups is generally not preplanned for studies with 2 : 1 randomization of a relatively small number of participants. Both treatment groups appeared to have similar numbers of participants with treatment-related adverse events except for nausea, which occurred in five participants receiving ATV/co versus one ATV/r participant and did not cause study drug interruption or discontinuation. One ATV/co participant reported an episode of vomiting on day 3, which was assessed as related to study treatment. The participant discontinued study treatment and did not attend any postbaseline study visits. The AE profile for COBI when given with ATV and FTC/TDF will be characterized more accurately with the accrual of additional safety data from the ongoing phase 3 study with the same treatment groups.

Elevation of indirect bilirubin was the most common laboratory abnormality, a side-effect of atazanavir mediated through competitive inhibition of UGT1A1 enzyme and influenced by genetic factors. In one study, plasma levels of ATV and the risk of hyperbilirubinemia were predicted by the 3435C→T polymorphism at the multidrug resistance gene 1: significantly higher plasma atazanavir plasma levels in participants with CC, compared with CT and TT genotypes, and bilirubin levels correlated with atazanavir concentrations [20]. In our study, the incidence of hyperbilirubinemia of at least grade 3 hyperbilirubinemia was similar between treatment groups. Mean indirect bilirubin levels at most postbaseline time points were slightly higher in participants receiving ATV/co compared with ATV/r; however, there was not a disproportionate number of participants manifesting clinical sequelae of elevated indirect bilirubin, namely ocular icterus and jaundice. The slightly higher mean indirect bilirubin in the ATV/co group might be related to differences between COBI and RTV in affinity for off-target metabolic enzymes. For example, RTV is a better pregnane X receptor activator than COBI, and so might increase UGT1A1 levels and, thus, decrease indirect bilirubin by increased metabolism independent of ATV plasma level.

Mean eGFR decreased from baseline in both treatment groups, was statistically significantly greater only at week 2, reached a nadir by week 24, and did not progress further through week 48. Three ATV/co participants had confirmed grade 1 elevation of serum creatinine. No participant receiving either treatment developed persistent serum electrolyte or urinalysis abnormalities, or discontinued or interrupted study treatment due to change in serum creatinine or eGFR [21,22]. A pharmacokinetic/pharmacodynamic study in healthy HIV-uninfected adults using iohexol [23], which exclusively undergoes glomerular filtration and is not secreted or reabsorbed, has shown that COBI can cause a mild increase in serum creatinine (mild decrease in eGFR), but does not affect iohexol-measured (actual) GFR, suggesting that the observed increase in serum creatinine may be due to an effect of COBI on the tubular secretion and not glomerular filtration of serum creatinine [21]. Additional in-vitro and clinical studies of COBI in both HIV-infected and HIV-uninfected participants are in progress and may provide additional insight into the biological mechanism responsible for the mild increase in serum creatinine seen with COBI.

Pharmacokinetic analyses demonstrated that COBI and RTV boost ATV exposures appear to be comparable including adequate ATV trough concentrations relative to its protein-binding adjusted concentration that leads to 90% maximal concentration, EC90, of 14 ng/ml when boosted with COBI or RTV. Greater than 90% of quantifiable values for ATV Ctrough at these visits were above the Department of Health and Human Services recommended target of 150 ng/ml for both treatments. These ATV exposures are consistent with historical pharmacokinetic data from a phase 1 study when ATV was boosted with COBI and RTV [15], and lends additional support for COBI as an appropriate pharmacoenhancer for the once-daily HAART regimen of ATV and FTC/TDF.

The results of this phase 2 study helped design and initiate a 96-week phase 3 randomized, partially double-blind study to evaluate the efficacy and safety of ATV/co versus ATV/r each administered with FTC/TDF in HIV-1-infected antiretroviral treatment-naive adults that is ongoing.


The authors thank the site investigators, study coordinators and staff, participants, and study management team for their careful attention to the study protocol.

R.E. was a study site investigator and contributed to writing and editing the manuscript.

C.C. was a study site investigator and contributed to writing and editing the manuscript.

J.G. was a study site investigator and contributed to writing and editing the manuscript.

P.S. was a study site investigator and contributed to writing and editing the manuscript.

T.H. was a study site investigator and contributed to writing and editing the manuscript.

H.C.L. was the study biostatistician and contributed to writing and editing the manuscript.

A.A.M. was the study pharmcokineticist and contributed to writing and editing the manuscript.

S.L.C. was a study medical director and contributed to writing and editing the manuscript.

B.P.K. was study pharmacologist and contributed to writing and editing the manuscript.

D.R.W. was a study medical director and contributed to writing and editing the manuscript.

Conflicts of interest

None declared.


1. Kempf DJ, Marsh KC, Kumar G, Rodrigues AD, Denissen JF, McDonald E, et al. Pharmacokinetic enhancement of inhibitors of the human immunodeficiency virus protease by coadministration with ritonavir. Antimicrob Agents Chemother 1997; 41:654–660.
2. Merry C, Barry MG, Mulcahy F, Ryan M, Heavey J, Tjia JF, et al. Saquinavir pharmacokinetics alone and in combination with ritonavir in HIV-infected patients. AIDS 1997; 11:F29–F33.
3. van Heeswijk RPG, Veldkamp AI, Hoetelmans RMW, Mulder JW, Schreij G, Hsu A, et al. The steady-state plasma pharmacokinetics of indinavir alone and in combination with a low dose of ritonavir in twice daily dosing regimens in HIV-1-infected individuals. AIDS 1999; 13:F95–F99.
4. Lal R, Hsu A, Chen P, et al.Single dose pharmacokinetics of ABT-378 in combination with ritonavir [abstract I-194.]. In: 37th Interscience Conference on Antimicrobial Agents and Chemotherapy; 28 September to 1 October 1997; Toronto, Ontario, Canada.
5. Lal R, Hsu A, Granneman R, El-Shourbagy T, Johnson M, Lam W, et al.Multiple dose safety, tolerability and pharmacokinetics of ABT-378 in combination with ritonavir [abstract 647]. In: 5th Conference on Retroviruses and Opportunistic Infections; 1–5 February 1998; Chicago, Illinois, USA.
6. Hsu A, Granneman GR, Bertz RJ. Ritonavir: clinical pharmacokinetics and interactions with other anti-HIV agents. Clin Pharmacokinet 1998; 35:275–291.
7. Invirase U.S. label information,,021785s010lbl.pdf.
8. Lexiva U.S. label information,,022116s010lbl.pdf.
9. Reyataz U.S. label information,
10. Aptivus U.S. label information,
11. Prezista U.S. label information,
12. Xu L, Liu H, Murray BP, Callebaut C, Lee MS, Hong A, et al. Cobicistat (GS-9350): a potent and selective inhibitor of human CYP3A as a novel pharmacoenhancer. ACS Med Chem Lett 2010; 1:209–213.
13. Callebaut C, Tsai L, Stray K, Liu H, Xu L, Desai M, et al. Biological profiling of GS-9350, a novel pharmacoenhancer that lacks anti-HIV activity and exhibits low potential for metabolic adverse effects in vitro [abstract]. Antivir Res 2010; 86:A31.
14. Mathias AA, German P, Murray BP, Wei L, Jain A, West S, et al. Pharmacokinetics and pharmacodynamics of GS-9350: a novel pharmacokinetic enhancer without anti-HIV activity. Clin Pharmacol Ther 2010; 87:322–329.
15. Ramanathan S, Warren D, Wei L, Kearney BP. Pharmacokinetic boosting of atazanavir with the pharmacoenhancer GS-9350 versus ritonavir [abstract 614]. In: Interscience Conference on Antimicrobial Agents and Chemotherapy; 2009; San Francisco, California, USA.
16. Mathias A, Liu HC, Warren D, Sekar V, Kearney BP. Relative bioavailability and pharmacokinetics of darunavir when boosted with the pharmacoenhancer GS-9350 versus ritonavir [abstract 28]. In: International Workshop on Clinical Pharmacology of HIV Therapy; 2010; Sorrento, Italy.
17. Smith KY, Weinberg WG, Dejesus E, Fischl MA, Liao Q, Ross LL, et al. Fosamprenavir or atazanavir once daily boosted with ritonavir 100 mg, plus tenofovir/emtricitabine, for the initial treatment of HIV infection: 48-week results of ALERT. AIDS Res Ther 2008; 5:5.
18. Molina JM, Andrade-Villanueva J, Echevarria J, Chetchotisakd P, Corral J, David N, et al. for the CASTLE Study TeamOnce-daily atazanavir/ritonavir versus twice-daily lopinavir/ritonavir, each in combination with tenofovir and emtricitabine, for management of antiretroviral-naive HIV-1-infected patients: 48 week efficacy and safety results of the CASTLE study. Lancet 2008; 372:646–655.
19. Daar ES. ACTG 5202: ATV/r or EFV in combination with ABC/3TC or TDF/FTC in antiretroviral naïve patients [abstract 59LB]. In: 17th Conference on Retroviruses and Opportunistic Infections (CROI); 16–19 February 2010; San Francisco, California, USA.
20. Rodríguez-Novoa R, Barreiro P, Rendon A, Barrios A, Corral A, Jimenez-Nacher I, et al. Plasma levels of atazanavir and the risk of hyperbilirubinemia are predicted by the 3435CrT polymorphism at the multidrug resistance gene 1. Clin Infect Dis 2006; 42:291–295.
21. Cohen C, Shamblaw D, Ruane P, Elion R, DeJesus E, Liu HC, et al.The single-tablet, fixed-dose regimen of elvitegravir/GS-9350/emtricitabine/tenofovir DF (Quad) achieves a high rate of virologic suppression and GS-9350 is an effective booster [abstract 58LB]. In: 17th Conference on Retroviruses and Opportunistic Infections (CROI) ; 16–19 February 2010; San Francisco, California, USA.
22. Elion R, Gathe J, Rashbaum B, Shalit P, Hawkins T, Liu C, et al.The single-tablet regimen elvitegravir/cobicistat/emtricitabine/tenofovir disoproxil fumarate (EVG/COBI/FTC/TDF; ‘Quad’) maintains a high rate of virologic suppression, and cobicistat (COBI) is an effective pharmacoenhancer through 48 weeks [abstract H-938B]. In: 50th ICAAC; 2010; Boston, Massachusetts.
23. Krutzén E, Bäck S-E, Nilsson-Ehle I, Nilsson-Ehle P. Plasma clearance of a new contrast agent, iohexol: a method for the assessment of glomerular filtration rate. J Lab Clin Med 1984; 104:955–961.

booster; cobicistat; GS-9350; HIV treatment-naive

© 2011 Lippincott Williams & Wilkins, Inc.