ABSTRACT
Enfuvirtide is the first fusion and entry inhibitor approved for use for the treatment of human immunodeficiency virus (HIV) type 1 infection and as such represents a novel class of agents. For the population of patients experienced with three antiretroviral classes, enfuvirtide provides an additional option for treatment. This prospective, open-label, 24-week, single-arm trial assessed the efficacy and safety of enfuvirtide (90 mg injected subcutaneously twice daily) in combination with darunavir-ritonavir (600/100 mg administered orally twice daily) in triple-antiretroviral-class-experienced adults failing their current regimen. The primary efficacy endpoint was the proportion of participants with plasma HIV RNA loads of <50 copies/ml. Other virological and immunological measures were also evaluated, as were the effects of the baseline viral coreceptor tropism and darunavir phenotype sensitivity scores on the outcomes. At week 24, 60.3%, 72.5%, and 84.0% of 131 participants achieved viral loads of <50 copies/ml and <400 copies/ml and a change from the baseline load of ≥1 log10 copies/ml, respectively. A baseline viral load of ≤5 log10 copies/ml was a significant predictor of achieving a viral load of <50 copies/ml at 24 weeks; however, neither background genotype sensitivity nor darunavir phenotype sensitivity was a significant predictor of the achievement of viral loads of <50 copies/ml. Although these findings are limited by the relatively small numbers of participants with darunavir susceptibility changes of ≥10-fold, they suggest that combining enfuvirtide and darunavir-ritonavir with an optimized background regimen in triple-class experienced participants naïve to these agents can result in positive virological and immunological responses regardless of most baseline parameters.
The introduction of highly active antiretroviral therapy has resulted in remarkable reductions in human immunodeficiency virus type 1 (HIV-1) infection-related morbidity and mortality (5). While an optimized regimen of at least two, but preferably three, active agents is critical to the success of antiretroviral therapy, the options for the selection of those active agents are limited for an increasing population of triple-antiretroviral-class-experienced patients harboring highly resistant viruses (10, 11). Among those demonstrating virological failure with antiretrovirals from the standard three classes of agents, 97% harbored viruses with at least one drug resistance mutation, with resistance to three or more agents within each class occurring in more than 40% of patients (11).
Enfuvirtide (ENF; Roche) is the first in a novel class of agents, known as fusion inhibitors, approved for the treatment of HIV infection. ENF, a 36-amino-acid synthetic peptide derived from the heptad repeat 2 (HR2) domain of HIV gp41, blocks the association of HR2 with HR1 (12) and prevents gp41 from undergoing a conformational change required for viral fusion and cell entry (6). Randomized, controlled trials have demonstrated virological and immunological benefits with the addition of ENF to an optimized background regimen (1, 2, 4, 9).
Darunavir (DRV; Tibotec Therapeutics), the latest protease inhibitor to have been developed, has also been shown to be highly effective in the triple-class-experienced patient population. A combined efficacy analysis of POWER trials 1 and 2 demonstrated that patients naïve to DRV that were switched from a failing regimen to DRV-ritonavir (DRV/r; 600/100 mg administered twice daily) with an optimized background regimen were significantly more likely to achieve undetectable levels of plasma HIV RNA and reductions of the load from the baseline load of ≥1 log10 copies/ml (1).
The Below the Level of Quantification (BLQ) Study (ClinicalTrials.gov registry number NCT00326963) evaluated the use of ENF in combination with DRV, obtained through an expanded-access program or commercially, and an optimized background regimen with triple-class-experienced HIV-infected patients. Virological and immunological determinants were assessed, as were the effects of the background regimen and DRV sensitivity on outcomes.
(This study was presented in part at the 47th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy, 17 to 20 September 2007, Chicago, IL.)
MATERIALS AND METHODS
Study design and participants.This prospective, open-label, 24-week, single-arm, clinical trial conducted in the United States and Australia was designed to assess the efficacy, safety, adherence, and tolerability of ENF with DRV/r. HIV-infected, triple-antiretroviral-class-experienced, ENF-naïve, DRV-naïve adults were enrolled between March and October 2006. The principal exclusion criteria were any known AIDS Clinical Trial Group grade 4 clinical or laboratory abnormality (asymptomatic grade 4 abnormalities were permitted at the discretion of the investigator), evidence of any active, untreated opportunistic infection, an intercurrent illness or drug toxicity, malignancy requiring chemotherapy or radiotherapy, and pregnancy or breast-feeding (or intention to become pregnant or breast-feed during the study). Participants with plasma HIV RNA loads of >2,000 copies/ml within 60 days of study initiation and receiving their current regimen received ENF (90 mg injected subcutaneously twice daily) in combination with DRV/r (600/100 mg administered orally twice daily) within an investigator-determined optimized background regimen. ENF was administered by a participant-selected injection device (either a 27-gauge, 1/2-in. needle and syringe, a 31-gauge, 8-mm needle and syringe, or a Biojector 2000 needle-free delivery system [Bioject Medical Technologies, Tualatin, OR]); the participants were permitted to switch devices at any time during the study. The study was conducted in accordance with the Declaration of Helsinki of 1975, as revised in 2000, and with institutional review board approval. Written consent was obtained from all participants.
Assessments.The following were evaluated at the baseline: the background regimen phenotype sensitivity (PhenoSense HIV), reported as n-fold change in the 50% inhibitory concentration (IC50); the background regimen genotype sensitivity, calculated by standard population-based genotype analysis (GeneSeq HIV); and coreceptor tropism (Trofile) (Monogram Biosciences, South San Francisco, CA). The genotype sensitivity score (GSS) was defined as the total number of drugs (excluding study drugs) in a participant's optimized background antiretroviral regimen to which their HIV isolate had genotypic sensitivity, as deduced from gene sequence and mutation analyses. The fold change in DRV susceptibility (a phenotypic HIV drug susceptibility assessment; Monogram Biosciences) was calculated according to the following formula: IC50 (patient)/IC50 (drug-sensitive reference virus), where IC50 is the drug concentration required to inhibit viral replication by 50%. The plasma HIV RNA load (number of copies/ml; Cobas Amplicor HIV-1 Monitor test, version 1.5; Roche Molecular Systems, Pleasanton, CA) and blood CD4+ T lymphocyte counts (number of cells/mm3) were measured at the baseline; at weeks 1, 4, 12, 16, and 24; and at 4 weeks after the end of treatment. At the same time points, reports of serious adverse events, deaths, serious AIDS-defining events, discontinuations (ENF or injection related), adverse events of special interest (any injection device-related adverse event other than the expected signs or symptoms of localized injection-site reactions), and localized injection-site reactions were collected.
End points and analyses.Efficacy analyses were conducted with the intent-to-treat population, which included participants who received at least one dose of trial medication and who had at least one postbaseline efficacy measurement. The safety analyses included all participants who received at least one dose of trial medication and one safety evaluation. The calculation of sample size assumes a 56% ENF response rate (the proportion of participants achieving HIV RNA loads of <50 copies/ml), as estimated from the results of previous trials (4, 7, 8, 9), and a predefined response rate for DRV of 43% (the proportion of participants with <50 HIV RNA copies/ml across all DRV change ranges; see http://www.fda.gov/cder/foi/label/2006/021976lbl.pdf). By using these criteria, a sample size of 120 participants was calculated to provide at least an 80% power to declare, with 95% confidence, a response rate greater than 43%. Sample size and power calculations were prepared by the Fisher exact test of nQuery Advisor (version 5.0).
The primary efficacy end point was the proportion of participants with plasma HIV RNA loads of <50 copies/ml at week 24. Additional 24-week efficacy end points included the proportion of participants with plasma HIV RNA loads of <400 copies/ml and a ≥1-log10-copy/ml decrease from the baseline, the mean change in the log10 number of plasma HIV RNA copies/ml from the baseline, the final mean CD4+ cell count, and the mean change in CD4+ counts from baseline.
For week 24 end points, the data were summarized by using proportions or means and standard deviations; two-sided 95% confidence intervals (CIs) were constructed. For the analysis of participants achieving plasma HIV RNA loads of <50, <400, and ≥1 log10 copies/ml, those with missing values were imputed as nonresponders; for the mean log10 change from the baseline and the immunological end points, the last observation carried forward method was used. The end points were next stratified by use of the baseline coreceptor (CCR5 versus CXCR4), and between-group differences were evaluated by the chi-square test for categorical variables and the t test for continuous variables. The end points also were stratified by the baseline DRV phenotype resistance status by using protocol-defined cutoffs (changes of <10-fold, ≥10- to 40-fold, and >40-fold), as used in previous trials (1, 7, 8), and by the fold change in tertile (tertile 1, 0.26 to 1.01; tertile 2, 1.07 to 5.23; tertile 3, 5.44 to 178.30); an additional post hoc analysis was performed by using changes in cutoffs of <3-fold, 3- to 7-fold, >7- to 10-fold, and >10-fold to explore the potential relevance of alternate cutoffs as predictors of outcomes. Overall comparisons were made by using either the chi-square test for categorical variables or analysis of variance for continuous variables. Finally, stepwise logistic regression was performed to evaluate the impact of the CD4+ cell count (≤100 [reference], >100), the baseline GSS (zero [reference], one, or greater than or equal to two), the log10 baseline viral load (≤5 [reference], >5), and the baseline DRV change (≤5 [reference], >5) on the primary end point dichotomized as a week 24 plasma HIV RNA load of < 50 copies/ml versus one of ≥50 copies/ml.
RESULTS
Of 142 participants enrolled in the study, 2 did not receive any study medication and 3 did not receive a safety evaluation, resulting in 137 participants in the safety population. Six participants received at least one dose of ENF but did not receive any efficacy assessments after the baseline; thus, 131 participants were included in the intent-to-treat population. The baseline characteristics of the intent-to-treat population are presented in Table 1. All participants received ENF and DRV during the study. In addition to these agents, 95.4% (125/131) of the participants received nucleoside reverse transcriptase inhibitors (NRTIs), 0.8% (1/131) received nonnucleoside reverse transcriptase inhibitors (NNRTIs), and 2.3% (3/131) received an NRTI plus an NNRTI. Two participants (2/131; 1.5%) received only a protease inhibitor (DRV) and ENF.
At week 24, 79/131 (60.3%; 95% CI, 0.515, 0.691) participants had plasma HIV RNA loads of <50 copies/ml and 95/131 (72.5%; 95% CI, 0.645, 0.805) had viral loads of <400 copies/ml. In addition, 110/131 (84.0%; lower 95% CI, 0.773) participants had a change from baseline of ≥1 log10 copies/ml. The mean change in the viral load from the baseline was −2.39 log10 copies/ml ± 1.21 (95% CI, −2.59, −2.18). The week 24 mean CD4+ count (cells/mm3) was 246.2 ± 166.6 (95% CI, 217.4, 275.0), for a mean change from the baseline of 84.0 ± 111.7 (95% CI, 64.6, 103.4).
Participants with CCR5 HIV coreceptor tropism by the Trofile assay had a significantly greater mean CD4+ cell count at baseline than those with dual/mixed HIV coreceptor tropism (218.2 ± 177.64 and 91.6 ± 106.29, respectively; P < 0.001). At 24 weeks, significantly more participants with CCR5-tropic virus (47/55, 85.5%) than participants with dual/mixed virus (32/48; 66.7%) achieved <400 copies/ml (P = 0.028). There were no other significant differences in outcomes on the basis of coreceptor tropism.
In the stepwise logistic regression analysis, participants with baseline plasma HIV viral loads of >5 log10 copies/ml were significantly less likely to achieve viral loads of <50 copies/ml than those with viral loads of ≤5 log10 copies/ml (n = 114; odds ratio, 0.32; 95% CI, 0.14, 0.74; P = 0.008). The baseline GSS (zero, one, or greater than or equal to two; GSS is the number of active agents, in addition to ENF and DRV) was not selected at the significance level of 0.25, indicating that there was no relationship between baseline antiretroviral drug susceptibility and the likelihood of achieving <50 copies/ml. The baseline CD4+ cell count (≤100 cells/mm3 or >100 cells/mm3) was also not a significant predictor of achieving <50 copies/ml (odds ratio, 1.943; 95% CI, 0.849, 4.447; P = 0.116).
Clinically relevant cutoffs for the n-fold change in DRV resistance were not established at the initiation of this trial; therefore, the change cutoffs used to assign DRV susceptibility were <10-fold, ≥10- to 40-fold, and >40-fold, as in other trials (1, 7, 8); change tertiles (as defined in Materials and Methods); and a post hoc analysis that used change cutoffs of <3-fold, 3- to 7-fold, >7- to 10-fold, and >10-fold, which include the cutoffs used by the U.S. Food and Drug Administration for DRV approval. There were no significant associations between baseline DRV resistance change (including the tertile analysis) and any baseline parameter or outcome variable (Table 2; tertile data not shown). In an additional post hoc analysis (data not shown) of DRV phenotypic sensitivity and GSS of the background antiretroviral regimen, 30/41 (73.2%) of participants with DRV changes of ≤40-fold and with a GSS of zero achieved viral loads of <50 copies/ml at 24 weeks, 37/41 (90.2%) had viral loads of <400 copies/ml, and 38/41 (92.7%) had reductions in HIV RNA loads of ≥1 log10 copies/ml. Participants with DRV changes of ≤40-fold and sensitivity to at least one other agent (GSS greater than or equal to one) did not have significant added benefits, with 34/50 (68%), 42/50 (84%), and 44/50 (88%) having viral loads of <50 copies/ml, viral loads of <400 copies/ml, and ≥1-log10 reductions in HIV RNA loads, respectively. Three of five participants with DRV changes of >40-fold at the baseline and without sensitivity to any other agents had viral loads of <50 copies/ml, while one of two participants with a DRV change of >40-fold at the baseline and sensitivity to at least one other agent had a viral load of <50 copies/ml. The results were comparable for the other end points.
Of the 137 participants in the safety population, 4 (2.9%) discontinued the study for safety reasons (two adverse events or intercurrent illnesses, one injection-site reaction, and one death). The serious adverse events reported were one case of cellulitis (0.7%) possibly related to ENF (the only adverse event of special interest, as defined in Materials and Methods) and one neutrophil count decrease (0.7%) with a remote possibility of a relationship to ENF. One case of rash (0.7%) was reported as being possibly related to ENF or possibly related to DRV. Other serious adverse events included pneumonia (n = 4; 2.9%); cellulitis (n = 2; 1.5%); renal failure (n = 2; 1.5%); and one (0.7%) case each of anemia, asthma, dyspnea, erythema nodosum, metabolic acidosis, steroid myopathy, sinusitis, and upper gastrointestinal hemorrhage. One (0.7%) death due to sepsis was not considered treatment related. The ranges in the percentages of participants experiencing grade ≥2 of certain injection-site reactions as measured at weeks 1, 12, and 24 and the mean numbers of reactions/participant were as follows: ongoing pain/discomfort (9 to 13%, 1.8 to 2.8), erythema (16 to 32%, 1.6 to 2.1), induration (32 to 44%, 2.2 to 3.7), pruritis (0 to 3%, 0 to 2.5), nodules and cysts (5 to 14%, 1.2 to 3.9), and ecchymosis (13 to 15%, 2.3 to 2.5). The numbers and percentages of participants experiencing any injection-site reaction of grade ≥2 at weeks 1, 12, and 24 were 58/131 (44.3%), 63/117 (53.8%), and 51/110 (46.4%), respectively. Overall, 98/137 (71.5%) participants had an injection-site reaction of grade ≥2 at any time during the study. The number and percentage of participants receiving ≥85% of their expected doses was 122/131 and 93.1%, respectively.
DISCUSSION
The use of combination antiretroviral therapy with two or more active agents is critical to the achievement and maintenance of plasma HIV RNA levels below the limits of detection (<50 copies/ml), which is the therapeutic goal even for highly treatment-experienced patients (3). In this population, ENF is an important treatment option, in that many patients have not had prior exposure to it; therefore, robust activity is anticipated and has been documented in trials involving patients initiating ENF therapy for the first time (http://www.fda.gov/ohrms/dockets/ac/07/briefing/2007-4314b1-01-Merck.pdf) (1, 2, 4, 9). DRV has also been shown to be highly effective in this patient population (1). In the present study, ENF in combination with DRV/r plus an optimized background regimen in triple-antiretroviral-class-experienced participants naïve to these two agents was well tolerated and produced excellent outcomes at 24 weeks. While the baseline DRV change was not a significant predictor of response (with approximately 60% of all participants achieving viral loads of <50 copies/ml), the small numbers of participants with DRV changes of >10-fold should be considered when these findings are interpreted.
The combination of ENF and DRV/r with an optimized background regimen has also been evaluated in the POWER trials (1a). Of the participants with one DRV resistance mutation or less, 64% of those receiving ENF de novo (n = 52) versus 55% of those not receiving ENF (n = 112) had undetectable viral loads (P = 0.33); the values were adjusted to account for the fact that participants receiving ENF tended to have more severe disease and had received fewer active background agents than those not receiving ENF. Of the participants with two DRV resistance mutations, 62% of those receiving ENF (n = 34) and 40% of those not receiving ENF (n = 78) achieved undetectable viral loads; among the participants with three or more DRV resistance mutations, 43% of those receiving ENF (n = 35) and 14% of those not receiving ENF (n = 58) achieved undetectable viral loads.
The effects of DRV resistance on the responses to ENF were also evaluated in the DUET trials (7, 8), in which participants received DRV/r as a component of an optimized background regimen and were randomized to receive etravirine, a new NNRTI, or placebo. A subset of participants in each group received ENF. Of those receiving ENF in the placebo group, which represents the most similar group for comparison to our study group, 70% to 73% of participants with a DRV change of <10-fold (representing a majority of the participants in the trials) achieved viral loads of <50 copies/ml at 24 weeks, consistent with our findings.
The present study indicates that the use of ENF with an optimized background regimen may be a valuable option for triple-class-experienced patients who are failing their current regimens and has a high likelihood of achieving virological and immunological treatment goals. While DRV was a newly available agent at the time that this study was initiated, the results showed that ENF had additive effects when it was used with DRV. However, newly available potent oral agents may also be reasonably expected to offer benefits to this population. Clinicians must critically evaluate the resistance profiles when selecting an optimal regimen or any particular agent. Although the findings of this study were limited by the relatively small number of participants with DRV changes of 10-fold or greater, they suggest that a regimen that includes ENF may provide benefits even when sensitivity to DRV is less than optimal.
ACKNOWLEDGMENTS
This study was sponsored by Roche.
The following authors have relevant financial relationships to disclose: E. DeJesus received research support from or is a consultant/advisory board member for Roche, Tibotec; M. Greenberg was an employee of Trimeris during the conduct of the study; M. Gottlieb received research support from and is a consultant/advisory board member for Roche; C. Guittari is an employee of Roche; J. Gathe and A. Zolopa have nothing to disclose.
We thank the study participants and the investigators, as well as James Thommes for his invaluable assistance in the trial. Statistical support was provided by Pi-Yeong Chi. Editorial support was provided by Linda Whetter of Zola Associates.
The participating investigators and their institutions are as follows: Robert Bolan, The Los Angeles Gay and Lesbian Community Services Center, Inc., Los Angeles, CA; Marcus Conant, San Francisco, CA; Gordon Crofoot, Houston, TX; Edwin DeJesus, Orlando Immunology Center, Orlando, FL; Jay Dobkin, Columbia University Medical Center, New York, NY; Robin Dretler, Infectious Disease Specialists of Atlanta, PC, Decatur, GA; Richard Elion, Clinical Alliance for Research & Education—Infectious Disease, LLC, Washington, DC; Charles Farthing, AIDS Healthcare Foundation, Beverly Hills, CA; Joseph Gathe, Therapeutic Concepts, Houston, TX; Michael Gottlieb, Kenmar Research Institute, LLC, Los Angeles, CA; Stephen Green, Hampton Roads Medical Specialists, Hampton, VA; Harold Katner, Mercer University School of Medicine, Macon, GA; Marah Lee, Lifeway, Inc., Ft. Lauderdale, FL; Ralph Liporace, Albany Medical College, Albany, NY; Christopher Lucasti, South Jersey Infectious Disease, Somers Point, NJ; David McDonough, Vista Medical Partners, Beverly Hills, CA; Patrick McLeroth, Chase Brexton Health Services, Inc., Baltimore, MD; Patrick McNamara, Houston, TX; Donna Mildvan, Beth Israel Medical Center, New York, NY; Karam Mounzer, Philadelphia FIGHT, Philadelphia, PA; Robert Myers, Body Positive Inc., Phoenix, AZ; David Prelutsky, Southampton Healthcare Inc., St. Louis, MO; Moti Ramgopal, Associates in Infectious Diseases, Port St. Lucie, FL; James Sampson, The Research and Education Group, Portland, OR; Jihad Slim, Saint Michael's Medical Center, Newark, NJ; Louis Sloan, North Texas Infectious Diseases Consultants, PA, Dallas, TX; David Wheeler, Clinical Alliance for Research & Education—Infectious Disease, LLC, Annandale, VA; and Andrew Zolopa, Stanford University Medical Center, Stanford, CA.
FOOTNOTES
- Received 8 April 2008.
- Returned for modification 4 May 2008.
- Accepted 13 September 2008.
- ↵*Corresponding author. Mailing address: Orlando Immunology Center, 1701 N. Mills Avenue, Orlando, FL 32803. Phone: (407) 647-3960. Fax: (407) 367-0856. E-mail: edejesus{at}oicorlando.com
↵▿ Published ahead of print on 22 September 2008.
REFERENCES
- American Society for Microbiology