International AIDS Society–USA Topics in HIV Medicine
Management of Antiretroviral Therapy
Timothy J. Wilkin, MD, C. Mhorag Hay, MD, Christine M. Hogan, MD, and Scott M. Hammer, MD
As in previous years, the 9th Conference Entry Inhibitors ants (Abstract 2). The drug is not cur- on Retroviruses and Opportunistic rently going forward in clinical develop- Infections provided a forum for a state- The chemokine receptors CCR5 and ment, but proof of principle appears to of-the-art update in antiretroviral thera- CXCR4 are coreceptors used by many have been established. py. Highlights included the status of strains of HIV-1 in addition to CD4 to new antiretroviral agents from both enter cells. Attempts are under way to CCR5 Receptor Blockers. Laughlin pre- existing and new drug classes; presenta- develop antiretroviral agents that inhib- sented the results for the first 12 HIV- tion of trials in antiretroviral-naive and it HIV-1 entry by blocking these requisite infected volunteers treated with SCH-C, antiretroviral-experienced persons; up- coreceptors. an orally bioavailable CCR5 receptor dates on strategic approaches to thera- antagonist with potent in vitro antiviral py, including when to start therapy, CXCR4 Receptor Blockers. AMD-3100 is a activity against a broad range of primary treatment interruptions, and immune- small-molecule CXCR4 receptor blocker HIV-1 isolates (Abstract 1). In this ongo- based therapies; mechanisms and evo- with potent in vitro anti-HIV activity. ing, sequential rising-dose trial, in which lution of viral drug resistance; clinical Results of an open-label dose-escala- there will be 12 subjects per group, HIV- applications of drug resistance testing; tion study to test the safety, pharma- infected volunteers receive 10 days of and therapeutic drug-level monitoring. cokinetics, and antiviral effect of AMD- SCH-C monotherapy with total daily 3100 were presented (Abstract 391-T). doses of 50, 100, and 200 mg. To date, 12 Forty HIV-infected volunteers with plas- HIV-1-infected adults, with CD4+ cell Antiretroviral Chemotherapy: ma HIV-1 RNA levels above 5000 counts above 250/µL, currently not New Investigational Agents copies/mL, on no or stable antiretroviral receiving antiretroviral therapy and with regimens, were enrolled and received 10 a non-SI phenotype, have received SCH- The advent of currently available days of continuous intravenous infusion C 25 mg orally every 12 hours for 10 days antiretroviral medications has substan- of AMD-3100 at a dose escalating from and have undergone intensive pharma- tially reduced HIV-1-related morbidity 2.5 µg/kg/hr up to 160 µg/kg/hr. Presence cokinetic and plasma HIV-1 RNA moni- and mortality. However, the increasing of the syncytium-inducing (SI) pheno- toring. emergence of drug-resistant HIV-1 vari- type (one method for determining pres- SCH-C was well tolerated, although ants as well as short-term and long-term ence of CXCR4-using variants) was not in electrocardiogram testing, subtle pro- toxic effects of these agents limit their part of the inclusion criteria, and only longation of the mean QTc interval was overall effectiveness. Several presenta- 28% of volunteers harbored SI variants. seen over 10 days of the lowest dose. Ten tions at the conference addressed new The study was discontinued early of 12 volunteers experienced at least a antiretroviral agents, including new because of adverse effects and lack of 0.5-log10 reduction in plasma HIV-1 RNA agents in existing classes with more antiviral effect. Adverse experiences, level from baseline during dosing, and 4 favorable resistance or toxicity profiles, particularly gastrointestinal in nature, subjects experienced a 1-log10 or greater as well as agents in new classes with were generally common. Two patients drop in plasma HIV-1 RNA level from non-overlapping resistance mecha- experienced unexpected adverse events baseline. Thus these preliminary proof- nisms compared with currently avail- of premature ventricle contractions, and of-principle data for SCH-C support fur- able classes. some patients in each dose group had ther exploration of the CCR5 receptor as atrial tachycardia. Three volunteers dis- a target for antiretroviral therapy. The Dr Wilkin is Post-Doctoral Research continued the study drug because of clinical future of SCH-C will depend in Fellow and Dr Hay is Instructor in Clinical adverse events. No volunteer harboring part on electrocardiographic findings at
Medicine in the Division of Infectious SI virus showed a 1-log10 or higher drop the higher doses. Diseases at Columbia University College in HIV-1 RNA, but 1 patient harboring SI One of the concerns of targeting the of Physicians and Surgeons. Dr Hogan is virus had a 0.9-log10 drop at day 10 of the CCR5 coreceptor is that doing so could Research Associate at the Aaron highest dose. Of note however, through select for X4-using strains of HIV-1 that Diamond AIDS Research Center, the use of an entry assay that can identify have been associated with more rapid Rockefeller University, and Instructor in HIV-1 isolates that are X4-, X5-, or dual- CD4+ T-cell count decline and clinical the Division of Infectious Diseases at tropic, investigators observed that 9 of progression. HIV-1 variants resistant to Columbia. Dr Hammer is Professor of 19 patients with dual (X4/R5) or mixed SCH-C were generated by in vitro pas- Medicine at Columbia and Chief of the (X4 and R5) virus at baseline exhibited a saging of HIV-1 strains in peripheral Division of Infectious Diseases at complete loss of X4 virus by day 11 of blood mononuclear cells (PBMCs) in the Columbia Presbyterian Medical Center in treatment. This result suggested the presence of increasing concentrations of New York, New York. ability of the drug to eliminate X4 vari- SCH-C (Abstract 397-T). Characterization
18 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
of the emerging variants with reduced located within either gp41 or the CD4 Tenofovir disoproxil fumarate (teno- sensitivity to SCH-C showed no switch to binding site of gp120. fovir) is a recently approved nucleotide other chemokine receptors including A cautionary note was sounded reverse transcriptase inhibitor for treat- CXCR4. However, it has not been deter- regarding the targeting of the very het- ing HIV infection. Louie and colleagues mined whether this switch occurs in erogeneous HIV-1 envelope protein: assessed the antiviral potency of this vivo. Similar findings in terms of contin- although 20 of 24 subtype-B isolates agent by obtaining frequent measure- ued CCR5 coreceptor use by resistant were inhibited by BMS-806 with an EC50 ments of HIV-1 RNA during the first 3 virus were reported from in vitro serial below 100 nM, 4 outliers had an EC50 as weeks of monotherapy with tenofovir in passaging experiments in the presence high as 10,000 nM. Future studies with 10 HIV-infected patients (Abstract 3). of SCH-D, another potent small- this promising compound will elucidate The decay rate of HIV-1 RNA was calcu- molecule antagonist of CCR5 in preclini- the mechanism of this diminished sus- lated and compared with the decay rates cal development (Abstract 396-T). ceptibility and the potential of this agent from similar previous experiments with PRO 140 is an anti-CCR5 monoclonal to inhibit a wide variety of HIV-1 strains. other agents. They found that tenofovir antibody that potently inhibits HIV-1 monotherapy led to a decay rate similar entry in vitro. Results of a study evaluat- Integrase Inhibitors to that seen in ritonavir monotherapy, ing the antiviral activity of PRO 140 in indicating that this agent has robust vivo in the hu-PBL-SCID mouse model of In vitro data for a new HIV-1 integrase activity against HIV-1. HIV-1 infection were presented (Abstract inhibitor in clinical development were presented by Fujiwara and colleagues 403-T). The mice were infected with the Nonnucleoside Reverse Transcriptase (Abstract 8). S-1360 is an orally available R5 isolate HIV-1 JR-CSF and then treated Inhibitors intraperitoneally with PRO 140 or a con- small-molecule inhibitor of HIV-1 inte- trol antibody. Both single-dose (1 mg) grase, the enzyme essential for the inte- TMC125 is a novel, second-generation and multiple-dose (0.1-1.0 mg every 3 gration of HIV-1 proviral DNA into host- investigational nonnucleoside reverse days for 3 weeks) PRO 140 reduced plas- cell chromosomes. The compound is an transcriptase inhibitor (NNRTI) with ma HIV-1 RNA to levels below detection integrase inhibitor with an EC50 of 28 to equally potent in vitro activity against in all treated animals, with the highest 74 ng/mL in vitro against clinical iso- wild-type HIV-1 and NNRTI-resistant plasma HIV-1 RNA level reduction being lates, and has demonstrated anti-HIV HIV-1 bearing the L100I, K103N, Y181C, activity in a mouse-MT4 in vivo assay. 1.8 log10 copies/mL. Dose-dependent dif- Y188L, or G190A/S mutations. Gazzard ferences were observed in the kinetics of HIV-1 variants resistant to S-1360 have (Abstract 4) presented results of a trial the antiviral response. Further studies of been isolated in vitro and the amino acid evaluating the antiviral activity of this compound in HIV-infected volun- substitutions responsible for drug resis- TMC125 in patients on failing NNRTI- teers are planned. tance have been shown to be in close containing regimens. Sixteen male vol- proximity to the integrase active site (eg, unteers with documented resistance to Novel Entry Inhibitors. Colonno de- T66I). This compound is currently being efavirenz (>10-fold decrease in suscepti- scribed the identification and characteri- investigated in a phase 1/2 study of HIV- bility by VirtualPhenotype and/or Anti- zation of a novel small-molecule in- infected volunteers. virogram [Tibotec-Virco, Mechelen, hibitor of HIV-1 entry, BMS-806, which Belgium]), in whom an efavirenz- or was identified using a cell-based screen Nucleoside and Nucleotide Reverse nevirapine-containing regimen was fail- and is believed to target the HIV-1 enve- Transcriptase Inhibitors ing, were enrolled. They received 900 mg lope protein (Abstracts 9 and 10). In vitro of TMC125 orally twice daily for 7 days in assays have demonstrated that the com- DPC 817 (Abstract 385-T) is a cytidine place of the failing NNRTI, and contin- pound competitively inhibits the binding nucleoside reverse transcriptase in- ued to take their nRTIs unchanged. One of gp120 to CD4 and that the radiola- hibitor (nRTI) with activity against many patient was withdrawn because of non- beled inhibitor binds selectively to puri- lamivudine- and zidovudine-resistant adherence. Median baseline plasma fied gp120 protein. The compound HIV-1 isolates. It is rapidly converted to HIV-1 RNA level was 10,753 copies/mL potently inhibits HIV-1 clinical isolates the active triphosphate form, which has and median baseline CD4+ cell count and laboratory strains (R5-, X4-, and an intracellular half-life of 13 to 17 was 389/µL. Median fold-change in EC50 dual-tropic variants) with a median hours. DPC 817 inhibits wild-type labo- to efavirenz at screening was 111 (range, effective concentration (EC50) of 62 nM in ratory and clinical isolates of HIV-1 with 16-659), and all patients had greater culture assays with subtype-B isolates, a mean 90% inhibitory concentration than 35-fold decreased sensitivity to and is inactive against HIV-2 and simian value of 855 nM, and shows less than 5- nevirapine. immunodeficiency virus (SIV). BMS-806 fold reduction in activity against recom- TMC125 was well tolerated in this has low cytotoxicity and is orally binant viruses containing as many as 10 study, with 11 patients reporting grade 1 bioavailable in rats, dogs, and monkeys. mutations, including M41L, M184V, adverse events, the most common of Preliminary animal toxicology studies D67N, L74V, K70R, T215Y, or K219Q. which were diarrhea and headache. No have revealed no safety concerns as of Thus DPC 817 may be effective in indi- lab abnormalities or electrocardiograph- yet. In vitro passaging of HIV-1 strains in viduals with resistance to other nRTIs ic changes were observed. The median the presence of the compound selected and is currently under investigation in a drop in plasma HIV-1 RNA level at day 8 for resistant strains, with mutations phase 1/2 clinical trial. was 0.89 log10 copies/mL (mean, 0.86
19 International AIDS Society–USA Topics in HIV Medicine
log10; range, 0.18-1.71 log10; intragroup failed in 61% and efavirenz had failed in ment discontinuations due to adverse comparison vs baseline, P<.001). Twelve 39%. In 94% of patients, baseline muta- effects in the atazanavir arms than in the patients (75%) experienced a decrease tions were consistent with virologic fail- ritonavir arm (9% vs 30%). The mean in plasma HIV-1 RNA level of at least 0.5 ure on an NNRTI regimen. The on-treat- decrease in plasma HIV-1 RNA levels at log10 copies/mL, and 7 patients (44%) ment response rate at week 8 (HIV-1 week 48 was 1.44 log10 copies/mL in the experienced a decrease greater than 1 RNA level <400 copies/mL), pooled for 400-mg atazanavir arm, 1.19 log10 log10 copies/mL. Thus TMC125 is well tol- the 100 mg and 200 mg doses, was 57% copies/mL in the 600-mg atazanavir arm, erated and has significant antiviral and varied according to the number of and 1.66 log10 copies/mL in the ritonavir potency against NNRTI-resistant HIV-1 new nRTIs used: 4 of 10 (40%) for arm. in vivo. patients starting no new nRTIs, 13 of 18 Subjects receiving the atazanavir- One question raised by this presen- (72%) for patients starting 1 new nRTI, containing regimen experienced de- tation is why the plasma HIV-1 RNA and 10 of 15 (66.7%) for patients start- creases in cholesterol, LDL, and triglyc- decline was less than that seen in ing 2 new nRTIs. Thus, data at 8 weeks eride levels with mean percent change NNRTI-naive individuals treated with suggest that DPC 083 has activity in from baseline of 1, −1, and −5, respec- TMC125 in a separate study (Abstract 5), patients in whom currently available tively, for the 400-mg dose of atazanavir; despite the in vitro activity profile of the NNRTIs are failing, but that the and −5, −7, and −27, respectively, for the drug against NNRTI-resistant isolates. response rate improves when DPC 083 600-mg dose of atazanavir. In contrast, The possibility of pharmacokinetic inter- is used in combination with at least 1 subjects in the ritonavir-containing arm actions with lingering efavirenz or nevi- new nRTI. However, almost 30% of experienced sustained increases in total rapine was given as one possible expla- patients enrolled were discontinued cholesterol, LDL, and triglyceride levels, nation, but this needs to be clarified in from the study, mostly because of pro- with mean percent change from baseline future studies. tocol violations. of +11, +23, and +93, respectively. A new In the study of TMC125 in 12 NNRTI- The 24-week results of a phase 2, protease inhibitor without adverse naive subjects (Abstract 5), the initial double-blind, dose-ranging study effects on lipids would be a welcome rate of decline of plasma HIV-1 RNA (Study DPC 083-201) of DPC 083 (50, addition to the antiretroviral armamen- observed during 1 week of monotherapy 100, and 200 mg once daily) versus tarium. with TMC125 900 mg twice daily was efavirenz in combination with fixed- compared retrospectively with that dose lamivudine/zidovudine in anti- Antiretroviral Chemotherapy in obtained during the previously per- retroviral-naive patients were also pre- formed 13-patient ERA study of a 5-drug sented (Abstract 7). As a result of this Antiretroviral-Naive Subjects combination of zidovudine, lamivudine, study, the 100-mg dose of DPC 083 was abacavir, indinavir, and nevirapine. The selected for a phase 3 study in NNRTI- Timing of Initial Antiretroviral median decline in plasma HIV-1 RNA naive patients. At this lower dose, DPC Therapy level was comparable between the 2 083 was associated with less dizziness groups:1.92 log10 copies/mL for TMC125- and the same frequency of rash as Optimal timing of antiretroviral therapy efavirenz. initiation for HIV-infected patients treated patients and 1.55 log10 copies/ mL for ERA patients (P=.40). The 2 remains unclear. Chaisson addressed this topic during the “Controversies in groups, however, did differ significantly Protease Inhibitors in terms of age (mean, 25 years for the Antiretroviral Therapy” symposium TMC125 group and 40 years for the ERA Atazanavir is a once-daily protease (Abstract S17). He discussed whether group), and there was a trend toward inhibitor in development that has been data exist to justify the following ratio- higher baseline CD4+ counts in the shown not to increase total cholesterol, nales for early treatment: TMC125 group (458 vs 360 cells/µL). The low-density lipoprotein (LDL), or authors concluded that TMC125 is a triglyceride levels in antiretroviral-naive 1. Will early treatment lead to eradica- highly potent NNRTI. individuals. Haas presented the results tion? No, not with current therapy. DPC 083 is another investigational of a randomized, active-controlled, NNRTI with activity against NNRTI-resis- blinded study that evaluated the safety, 2. Does early therapy lead to better tant isolates. Preliminary 8-week results tolerability, and efficacy of atazanavir virologic response? Probably not; in were presented of Study DPC 083-203, (400 or 600 mg qd)/saquinavir (1200 mg studies that have shown a worse virologic an ongoing study of the use of DPC 083 qd) and 2 nRTIs versus ritonavir (400 response when treatment is initiated with (either 100 mg or 200 mg once daily) and mg bid)/saquinavir (400 mg bid) and 2 CD4+ cell counts below 200/µL, the 2 nRTIs (selected based on treatment nRTIs (Abstract 42). Study subjects results may have been confounded by history and baseline genotype) in were volunteers in whom a prior regi- poorer adherence in the delayed initiators. patients on failing NNRTI-containing men was virologically failing. Efficacy regimens (Abstract 6). Fifty-one patients and lipid results at week 48 were pre- 3. Do patients who are treated later were enrolled, with a mean baseline sented for 85 subjects having a plasma experience more drug-related toxic- plasma HIV-1 RNA level of 3.85 log10 HIV-1 RNA level of 1000 to 100,000 ity? Perhaps in the short-term, but some copies/mL and a mean baseline CD4+ copies/mL and CD4+ cell counts of at long-term toxicities such as new-onset cell count of 473/µL. Nevirapine had least 100/µL. There were fewer treat- hyperglycemia are common to all.
20 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
4. Is immune function restored if treat- cohorts is problematic; and follow-up is Trials in Antiretroviral-Naive Subjects ment is begun late? Yes, although the limited—we do not as yet know whether The results of trials of initial antiretrovi- durability is unknown. 5 or 10 years of follow-up will show that ral therapy in treatment-naive patients either early or delayed initiation is more are selectively summarized in Table 1. Thus the remaining question is whether beneficial. Survival data have been early treatment improves clinical out- based on 2- or 3-year follow-up. AIDS Clinical Trials Group 388 Study. Fischl come. The answer to this question Palella presented data from the HIV presented the results of ACTG 388, a remains elusive in part because of the Outpatient Study (HOPS), which sug- phase 3 open-label randomized study difficulty in performing a randomized gested that patients who initiate comparing 2 4-drug antiretroviral regi- clinical trial to address the question. antiretroviral therapy at higher CD4+ mens with a 3-drug regimen in patients Chaisson discussed a previously pub- cell counts experience lower mortality with advanced HIV disease (CD4+ cell lished observational cohort study from (Abstract 13). They grouped patients count <200/µL or plasma HIV-1 RNA the Johns Hopkins HIV program that fol- observed at 8 US clinics participating in level >80,000 copies/mL) and no prior lowed 2 groups of patients who entered the HOPS from January 1996 to March antiretroviral or limited nRTI therapy the clinic after July 1, 1996 (Sterling et al, 2001 into 1 of 3 pretreatment CD4+ cell (Abstract 41). The trial compared the use AIDS, 2001). The first group initiated count strata: 501 to 750 cells/µL (n=126), of zidovudine and lamivudine with either antiretroviral therapy for at least 90 351 to 500 cells/µL (n=315), and 201 to indinavir (800 mg tid), efavirenz and indi- days, and the second group did not ini- 350 cells/µL (n=377). In a prospective navir (1000 mg tid), or nelfinavir (1250 tiate antiretroviral therapy. Patients who analysis, they looked at mortality rates mg bid) and indinavir (initially 1000 mg initiated antiretroviral therapy were in each preantiretroviral therapy CD4+ bid but increased to 1200 mg bid during more likely to be men, less likely to be count stratum, comparing patients who the study). Among the 517 enrolled, the African American or injection-drug initiated antiretroviral therapy while in mean baseline CD4+ cell count was users, and had lower baseline CD4+ cell that pretherapy stratum with patients 161/µL and the mean plasma HIV-1 RNA who delayed antiretroviral therapy until counts and higher baseline plasma HIV- level was 5.42 log10 copies/mL; median 1 RNA levels than the group that did not they reached a lower stratum. follow-up was 2.1 years. The primary end- initiate antiretroviral therapy. For patients with CD4+ counts of 201 point was time to virologic failure, which Among persons with CD4+ cell to 350 cells/µL, 20.8 deaths per 1000 per- occurred in 172 subjects: 52 in the indi- counts below 200/µL at entry, patients son-years of observation occurred in 325 navir arm, 39 in the efavirenz/indinavir who initiated antiretroviral therapy “initiators” and 70.6 deaths per 1000 per- arm, and 81 in the nelfinavir/indinavir experienced a lower rate of progression son-years of observation occurred in 52 arm. Virologic failure occurred at a lower to AIDS or death. Patients with CD4+ “delayers” (rate ratio [RR]=0.29, P<.001; rate in the efavirenz/indinavir arm cell counts from 201/µL to 350/µL or median years of follow-up, 3.0 for initia- (P=.04) and at a higher rate in the nelfi- higher experienced no change in clinical tors and 3.3 for delayers, P>.3). For navir/indinavir arm than in the indinavir- disease progression whether or not ther- patients with CD4+ counts between 351 only arm (P=.006). There was a trend apy was initiated. By multivariate analy- and 500 cells/µL, 10.7 deaths per 1000 toward less discontinuation in the sis of persons on antiretroviral therapy, person-years of observation occurred in efavirenz/indinavir arm. low CD4+ cell count (<200/µL) at the 229 initiators and 18.2 deaths per 1000 Grade 3 or 4 signs and symptoms time of initiation predicted progression person-years of observation occurred in occurred in 126 individuals in the study: to AIDS, but the plasma HIV-1 RNA level 86 delayers (RR=0.59, P>.3; 95% confi- 35 in the indinavir arm, 41 in the was not as predictive. dence interval [CI], 0.21, 1.65; median efavirenz/indinavir arm, and 50 in the Chaisson thus concluded that years of follow-up, 3.7 for initiators and nelfinavir/indinavir arm. Grade 3 or 4 lab- patients who initiate antiretroviral ther- 3.4 for delayers, P>.3). For patients with oratory abnormalities occurred in 178 apy at CD4+ cell counts above 350/µL CD4+ counts of 501 to 750 cells/µL, 7.5 subjects (57, indinavir; 58, efavirenz/indi- have not been shown to benefit clinical- deaths per 1000 person-years of obser- navir; and 63, nelfinavir/ indinavir). There ly and that the optimal time to initiate vation occurred in 54 initiators and 3.0 was a trend toward increased occurrence antiretroviral therapy for patients with per 1000 person-years occurred in 72 of adverse events in the nelfinavir/indi- CD4+ cell counts from 200/µL to 350/µL delayers (RR=2.25, P>.4), but only 3 navir arm as compared with the indinavir has not been determined. Patients with deaths occurred in this stratum, none arm (P=.07), but no significant difference CD4+ cell counts below 200/µL should apparently HIV-related (median years of was seen in occurrence of adverse events initiate treatment. He also concluded follow-up, 5.9 for initiators and 5.3 for between the efavirenz/indinavir arm and that a high plasma HIV-1 RNA level by delayers, P>.3). the indinavir arm (P=.97). Fischl and col- itself should not be an indication for The authors concluded that these leagues concluded that compared with treatment, but might suggest more fre- preliminary data suggest that initiation the 3-drug regimen, treatment with quent monitoring. He acknowledged the of antiretroviral therapy for patients with zidovudine, lamivudine, efavirenz, and following caveats to these conclusions: CD4+ counts of 201 to 350 cells/µL, and indinavir was comparably well tolerated observational cohorts are subject to possibly for patients with CD4+ counts and yielded a superior virologic measurable and unmeasurable con- of 351 to 500 cells/µL, is associated with response. Treatment with zidovudine, founding and cannot replace random- reduction in mortality in comparison lamivudine, nelfinavir, and indinavir, on ized controlled trials; variability between with those who delay therapy. the other hand, yielded an inferior viro-
21 International AIDS Society–USA Topics in HIV Medicine
Table 1. Trials in Antiretroviral-Naive Patients
Authors, Study (Abstract No.), N Follow-up Baseline Values Change in Values and Regimens HIV-1 RNA CD4+ HIV-1 RNA CD4+ (copies/mL) (cells/µL) (copies/mL) (cells/µL)
Fischl et al, ACTG 388 (41)
1. zidovudine/lamivudine/ 168 2.1 years 5.42 log10 161 86% <200 by Increase indinavir 800 mg tid (median) week 24
2. zidovudine/lamivudine/efavirenz/ 173 87% <200 by Increase indinavir 1000 mg tid week 24
3. zidovudine/lamivudine/nelfinavir 176 78% <200 by Increase 1250 mg bid/indinavir 1000 mg bid week 24 (later increased to 1200 mg bid)
Eron et al, M99-056 (409-W)
1. stavudine/lamivudine and 19 48 weeks 4.7 log10 264 74% <50 by ITT +235 lopinavir/ritonavir 800/200 mg qd at week 48
2. stavudine/lamivudine and 19 79% <50 by ITT +248 lopinavir/ritonavir 400/100 mg bid at week 48
Domula et al (408-W)
1. zidovudine/lamivudine/efavirenz 35 48 weeks 4.9 log10 275 37% <50 by ITT; +180 81% <50 by OT
2. lamivudine/stavudine/efavirenz 35 163 40% <50 by ITT; +245 82% <50 by OT
3. didanosine/stavudine/efavirenz 38 165 37% <50 by ITT; +281 74% <50 by OT
ACTG indicates AIDS Clinical Trials Group; ITT, intent-to-treat analysis; OT, on-treatment analysis.
logic response. plasma HIV-1 RNA level of 4.6 log10 taking hydroxyurea (51.8%) than in those copies/mL. not taking hydroxyurea (26%; P<.001); CHARM Study. The CHARM study was a By intent-to-treat analysis at week adverse events also occurred more fre- phase 3, open-label, randomized, multi- 48, treatment failure had occurred in quently in subjects taking nevirapine center study to evaluate the efficacy and 60% of volunteers receiving nevirapine, (46.1%) than in those not taking nevirap- tolerability of adding nevirapine and/or compared with 62.3% of patients not ine (31.6%; P=.024). In the intent-to- hydroxyurea to a triple-nRTI regimen in receiving nevirapine (P=.826). The as- treat analysis, a slower increase in mean treatment-naive HIV-1-infected subjects treated failure rates were 21.1% and CD4+ cell count was seen in subjects (Abstract 410-W). Volunteers with plas- 27.1%, respectively (P=.445). By intent- receiving hydroxyurea than in subjects ma HIV-1 RNA levels above 5000 to-treat analysis at week 48, treatment not receiving hydroxyurea. Thus, neither copies/mL were randomized using a fac- failure had occurred in 69.3% of volun- the addition of hydroxyurea nor the torial design to add nevirapine and/or teers receiving hydroxyurea, compared addition of nevirapine to the triple-nRTI hydroxyurea to the combination of with 53.9% of patients not receiving regimen decreased the primary end- zidovudine, lamivudine, and abacavir. hydroxyurea (P=.017). The as-treated point of treatment failure, and addition The primary endpoint was treatment failure rates were 23.9% and 24.3%, of either hydroxyurea or nevirapine failure defined as plasma HIV-1 RNA respectively (P=.963). The odds ratio of increased the incidence of adverse level above 50 copies/mL at or after experiencing treatment failure at week events. week 24, or discontinuation of or change 48 was 1.71 (95% CI, 1.11, 2.62) for in randomized treatment. A total of 229 hydroxyurea use. Study M99-056. Eron presented the 48- volunteers were enrolled, with baseline Treatment-limiting adverse events week results of study M99-056 compar- CD4+ cell count of 269/µL and baseline occurred more frequently in subjects ing once-daily and twice-daily dosing of
22 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
missing data equaled failure, 74% and efavirenz in combination with any of the 79% of patients in the daily and twice- 3 different nRTI pairs is an effective ini- Comments daily groups, respectively, had plasma tial antiretroviral regimen. However, 33% HIV-1 RNA levels below 50 copies/mL at of volunteers terminated therapy early week 48. The mean CD4+ cell count because of adverse events, of which increases from baseline to week 48 were approximately two-thirds were therapy 238/µL and 248/µL for the daily and related, and these protease inhibitor- twice-daily groups, respectively. Adher- sparing regimens were not immune to The efavirenz/indinavir group had superior ence was similar between the treatment adverse effects on lipids. virologic response: 39 virologic failures vs groups as measured by medication 52 in indinavir arm (P=.04). The nelfi- event monitoring system (MEMS cap) Antiretroviral Chemotherapy in navir/indinavir group had inferior virologic analysis. The frequency of adverse response: 81 virologic failures vs 52 in events (nausea, diarrhea, lipid abnor- Antiretroviral-Experienced indinavir arm (P=.006). malities) was comparable across the 2 Subjects groups. Trials in Antiretroviral-Experienced Comparison of 3 nRTI Combinations With Patients Efavirenz. Domula presented the results of a study evaluating the safety and effi- The results of trials in antiretroviral ther- cacy of efavirenz in combination with 3 apy-experienced patients are summa- Adverse events and adherence were simi- different dual nRTI combinations in rized in Table 2. lar in the 2 groups. The qd group had lower lopinavir trough levels. antiretroviral-naive HIV-1-infected pa- tients (Abstract 408-W). Patients ACTG 398 Study. The results of ACTG 398 received efavirenz in combination with were given by Mellors in an oral presen- zidovudine/lamivudine (n=35), lamivu- tation that focused on the baseline pre- dine/stavudine (n=35), or didanosine/ dictors of virologic suppression (Ab- 36 patients (33%) discontinued therapy stavudine (n=38). Median baseline plas- stract 45). The stimulus for this trial was because of adverse events: 10, efavirenz- ma HIV-1 RNA level was 4.9 log10 to address whether the addition of a sec- related rash or central nervous system copies/mL and CD4+ cell count was ond protease inhibitor to a new 4-drug symptoms; 6, neuropathy; 3, nausea; 2, 174/µL. By intent-to-treat analysis at 48 salvage regimen would lead to improved leukopenia or anemia; 1, lipodystrophy; weeks, the percentage of patients with virologic suppression among patients in 1, asthenia; 2, emergence of resistance; plasma HIV-1 RNA level below 50 whom protease-inhibitor-based regi- and 11, reasons not related to therapy. copies/mL was 37% in the zidovudine/ mens were failing. Participants were eli- Overall cholesterol levels increased by lamivudine group, 40% in the lamivu- gible if they had a plasma HIV-1 RNA 28% from baseline. dine/stavudine group, and 37% in the level above 1000 copies/mL, prior expo- didanosine/stavudine group. (By on- sure to 1 to 3 protease inhibitors for treatment analysis, the percentages more than 16 weeks, and no prior expe- were 81%, 82%, and 74%, respectively.) rience with abacavir, adefovir, or ampre- coformulated lopinavir/ritonavir in com- There were no statistically significant navir. Fifty-six percent of participants bination with stavudine and didanosine differences in virologic response among were NNRTI-naive. All participants (Abstract 409-W). Thirty-eight antiretro- the 3 groups. received abacavir, adefovir, efavirenz, viral-naive patients with median base- Over 48 weeks, 36 patients (33%) dis- and amprenavir, and were randomized line plasma HIV-1 RNA level of 4.7 log10 continued therapy because of adverse to indinavir 1200 mg twice daily, copies/mL and CD4+ cell count of events, although in 11, these were not saquinavir soft-gel capsule 1600 mg 264/µL were randomized to receive related to therapy: 10 stopped because twice daily, nelfinavir 1250 mg twice lopinavir/ritonavir 800/200 mg once of efavirenz-related effects (rash and daily, or placebo. The median baseline daily or lopinavir/ritonavir 400/100 mg central nervous system effects), 6 HIV-1 RNA level was 4.7 log10 copies/mL twice daily in addition to the standard because of didanosine/stavudine-relat- and the median CD4+ cell count was doses of stavudine and lamivudine. ed neuropathy, 3 because of nausea, 2 202/µL. Virologic failure was defined as a Lopinavir area-under-the-curve and because of anemia or leukopenia, 1 confirmed rise of HIV-1 RNA level above maximum concentration (Cmax) values because of lipodystrophy, 1 because of baseline, a less than 0.5-log10 decline in were similar for the daily and twice-daily asthenia, and 2 because of development HIV-1 RNA level by week 8, a confirmed regimens, but overall median lopinavir of resistance. Overall, cholesterol levels 1.0-log10 rise above HIV-1 RNA nadir, 2 trough concentration/IC50 was lower in increased by 28% from baseline to week consecutive HIV-1 RNA levels of at least the daily than the twice-daily regimen 48 (P<.0001), with a preponderance of 200 copies/mL after 2 consecutive val- (40 vs 84). Pharmacokinetic data from cholesterol elevations in the didano- ues below this limit, or a confirmed HIV- this study were also given in an oral pre- sine/stavudine arm, and triglyceride lev- 1 RNA level of at least 200 copies/mL at sentation by Bertz (Abstract 126). els increased by 38% from baseline week 24 or 48. The virologic failure rate By intent-to-treat analysis in which (P=.004). The authors conclude that in the 3 dual protease inhibitor arms
23 International AIDS Society–USA Topics in HIV Medicine
Table 2. Trials in Treatment-Experienced Patients
Authors, Study (Abstract No.), N Follow-up Baseline Values Change in Values and Regimen (s) (weeks) HIV-1 RNA CD4+ HIV-1 RNA CD4+ (copies/mL) (cells/µL) (copies/mL) (cells/µL)
Haas et al, AI424-009 (42)
1. atazanavir 400 mg/saquinavir 34 48 3-5 log10 >100 -1.44 log10 +109 sgc 1200 mg qd/2 nRTIs (mean change)
2. atazanavir 600 mg/saquinavir 28 -1.19 log10 +55 sgc 1200 mg qd/2 nRTIs
3. ritonavir 400 mg/saquinavir 23 -1.66 log10 +149 sgc 400 bid/2 nRTIs
Squires et al, Study 907 (413 W)
1. tenofovir 300 qd plus stable 368 24 (48) 3.4 log10 427 24 weeks: +12.5 background antiretrovirals -0.61 log10 (48 weeks:
-0.56 log10)
2. placebo plus stable background 182 24 (on 24 weeks: -10.8
antiretrovirals (for 24 weeks prior to placebo) -0.03 log10 cross-over)
3. group 2 after cross-over to 24 (on 24-48 weeks: tenofovir tenofovir) -0.7 log10
Mellors et al, ACTG 398 (45) 1. amprenavir/abacavir/ 322 48 4.7 log10 212 23% <200 N/A efavirenz/adefovir/saquinavir sgc 1600 mg or indinavir 1200 mg or nelfinavir 1250 mg bid
2. amprenavir/abacavir/ 157 18% <200 efavirenz/adefovir/protease inhibitor (P=.17) placebo
ACTG indicates AIDS Clinical Trials Group; NNRTI, nonnucleoside reverse transcriptase inhibitor; nRTI, nucleoside reverse transcriptase inhibitor; sgc, soft-gel capsule; tenofovir, tenofovir disoproxil fumarate. combined was not significantly different notypic susceptibility to efavirenz was and specific adherence to efavirenz. than the placebo arm at 48 weeks (77% the most important predictor of virolog- NNRTI-naive patients with a 0.4-fold or vs 82%, respectively; P=.17). ic failure. Other important predictors less change in baseline phenotypic
Mellors presented an extensive anal- were oral efavirenz clearance, NNRTI resistance to efavirenz (ie, IC50 60% ysis of the predictors of virologic failure. experience, medication adherence as lower than laboratory strain), also In a multivariate analysis, baseline phe- measured by MEMS cap monitoring, known as “hypersusceptibility,” had less
24 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
(Abstract 413-W). Participants enrolled stopping therapy were to decrease cost, in this study were antiretroviral-experi- avoid adverse effects, and boost HIV- Comments enced and were on a stable regimen with specific immunity. The 3 situations in a plasma HIV-1 RNA level of 400 to which STI is currently being studied are 10,000 copies/mL. The mean baseline (1) patients who began potent antiretro-
plasma HIV-1 RNA level was 3.36 log10 viral therapy prior to seroconversion copies/mL and mean CD4+ cell count and are chronically virologically sup- was 427/µL. These participants were pressed, (2) patients who began potent Atazanavir arms had fewer discontinua- highly treatment-experienced with a antiretroviral therapy during the chronic tions due to adverse effects (10% vs 30%) mean of 5.4 years of prior antiretroviral phase of infection (ie, established infec- and had a more favorable lipid profile. use. They were randomized 2:1 to the tion) and who are virologically sup- addition of tenofovir or placebo to the pressed, and (3) prior to a salvage regi- patient’s own stable background men in patients with highly drug-resis- antiretroviral regimen. After 24 weeks, tant virus, to induce resensitization of all participants received open-label virus. He stressed that large studies are tenofovir. The primary outcome was needed to compare STI to simply dis- time-weighted average change from continuing therapy and to weigh the baseline log plasma HIV-1 RNA at 24 risks and benefits of STI. Abstracts from Outcome was a time-weighted average 10 weeks (DAVG ). The DAVG was −.61 in the conference examining STI in each of decline of log plasma HIV-1 RNA level. 24 24 10 the tenofovir group and −.03 in the these situations will be discussed in the Tenofovir had a significantly greater following sections. decline in plasma HIV-1 RNA levels than placebo group (P<.0001). The partici- pants crossing over to tenofovir at 24 placebo (P<.0001). The subjects were Primary Infection. Miró and colleagues weeks showed a similar subsequent highly antiretroviral-experienced: 48% had (Abstract 529-M) reported preliminary decline in log plasma HIV-1 RNA level baseline NNRTI mutations, 58% had base- 10 results from an ongoing study of 12 (DAVG , −.71). Those originally random- line protease inhibitor mutations, and 24 patients who began highly active 94% had nRTI mutations. The mean ized to tenofovir maintained a .56-log 10 antiretroviral therapy (HAART) during length of antiretroviral use was 5.4 years. decline at 48 weeks. primary HIV infection at least 1 year Grade 3 or 4 abnormalities were <2% in A detailed discussion of the virologic prior to enrollment and who had sus- all groups. Serum creatinine did not rise response as predicted by baseline muta- above 2.0 mg/dL in any patients. A change tained virologic suppression. These par- tions in the reverse transcriptase gene is ticipants will undergo 4 cycles of STI: 2 of >.5 mg/dL was seen in 12 (3%) partici- given in the “Viral Resistance” section in pants receiving tenofovir for 48 weeks months off followed by 2 to 4 months on this review. There were a low number of therapy. Viral suppression to below 5000 compared with 2 (1%) receiving placebo adverse events that did not vary signifi- for 24 weeks. No one discontinued study plasma HIV-1 RNA copies/mL was cantly by treatment group. Specifically, drug because of change in creatinine or achieved in 4 of 12 patients after the no participants developed a creatinine hypophosphatemia. third STI. A strong CD8+ cytotoxic T level of more than 2.0 mg/dL or discon- lymphocyte (CTL) response developed tinued the drug because of creatinine in 7 of 9 participants after the third STI. elevations or hypophosphatemia. Simi- Yu and colleagues described the CTL Of the participants, 44% were NNRTI- lar conclusions were drawn in a safety response in a single patient treated with experienced. Virologic suppression was analysis combining tenofovir DF 907 HAART during acute HIV infection strongly predicted by efavirenz hypersus- study data with data from previous teno- (Abstract 537-M). They were able to doc- ceptibility (<.4-fold resistance) having a fovir clinical trials (Abstract 416-W). ument a broadening CTL response in 0.27 odds of virologic failure (P<.001). Similar virologic response and safety this patient during 2 subsequent STIs There was a high dropout rate due to tox- profiles were reported in the expanded using 505 overlapping peptides span- icities (30% at 24 weeks and 42% at 48 access program of 7317 subjects ning the entire expressed HIV-1 weeks). (Abstract 415-W). sequence. Only 2 epitopes were target- ed during acute infection compared with (table continued on next page) Strategies for Antiretroviral 25 at the end of the second STI. Therapy However, as of the end of the second STI, virologic control in this patient was Structured Treatment Interruptions not maintained off HAART. virologic failure at 48 weeks than NNRTI- Hirschel gave an overview of the contro- Chronic (Established) Infection. Several naive patients with a more than 0.4- to versies associated with structured treat- presentations examined the outcomes 2.5-fold change (43% vs 74%, P<.01). ment interruptions (STIs) during the associated with treatment interruptions “Controversies in Antiretroviral Therapy” in the setting of established infection. Study 907. Squires presented the results symposium (Abstract S18). The main The largest of these studies came from of Study 907 in a poster presentation reasons he outlined for considering the EuroSIDA group (Abstract 48) in
25 International AIDS Society–USA Topics in HIV Medicine
Table 2. Trials in Treatment-Experienced Patients (continued from page 24)
Authors, Study (Abstract No.), N Follow-up Baseline Values Change in Values and Regimen (s) (weeks) HIV-1 RNA CD4+ HIV-1 RNA CD4+ (copies/mL) (cells/µL) (copies/mL) (cells/µL)
Lalezari et al, T20-206 (418-W)
1. Pentafuside (50, 75, or 52 48 4.28 log10 279 54.9% <400 (penta- +132 100 mg) bid/abacavir/efavirenz/ fuside arms grouped); amprenavir/ritonavir bid 47.1% <50
2. abacavir/efavirenz/ 19 4.25 log10 201 36.8% <400; +90 amprenavir/ritonavir bid 36.8% <50
Boyd et al, HIV-NAT 005 (422-W)
1. zidovudine/lamivudine/indinavir 50 76 4.0 log10 168 1.8-log10 decline +101 800 mg/ritonavir 100 mg bid (66% <50)
2. zidovudine/lamivudine/ 54 1.6-log10 decline +128 indinavir tid (69% <50)
Baldini et al (423-W) Single arm: lopinavir/ritonavir 400 22 24 4.8 log10 177 1.18-log10 decline +88 mg/100 mg bid/amprenavir 600 mg at week 8; bid/and 2 nRTIs 1.13-log10 decline at week 24
Albrecht et al, ACTG 364 (425-W) 1. nelfinavir/placebo/2 nRTIs 195 144 3.9 log10 350 48% <50 +171
2. efavirenz/placebo/2 nRTIs 58% <50
3. nelfinavir/efavirenz/2 nRTIs 71% <50
Lafeuillade et al, HYDILE (424-W)
1. didanosine/stavudine/abacavir/ 24 48 4.2 log10 386 21% <50 +118 efavirenz
2. didanosine/stavudine/ 22 3.8 log10 55% <50 -27 abacavir/efavirenz/hydroxyurea
3. didanosine/stavudine/ 23 3.9 log10 48% <50 +78 abacavir/efavirenz/hydroxyurea/ interleukin-2
ACTG indicates AIDS Clinical Trials Group; NNRTI, nonnucleoside reverse transcriptase inhibitor; nRTI, nucleoside reverse transcriptase inhibitor; sgc, soft-gel capsule; tenofovir, tenofovir disoproxil fumarate. which 565 of 3610 assessable patients CD4+ cell count at time of interruption ing event or death was 6 times higher interrupted HAART. The median nadir was 242/µL. Participants had to be fol- among those stopping antiretroviral CD4+ cell count prior to beginning lowed up for at least 3 months after therapy than among those who did not. HAART for those subsequently interrupt- stopping to be included in this analysis. When controlling for the last-known ing therapy was 130/µL and the median The risk of developing a new AIDS-defin- CD4+ cell count and HIV-1 RNA level,
26 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
disease progression. The magnitude of median CD4+ cell count of 740/µL. All this risk was strongly correlated with participants underwent the same proto- Comments CD4+ cell count. The rate of AIDS-defin- col of STI: 2 weeks off therapy followed ing events or death was .48 per person- by 8 weeks on therapy for 4 cycles, fol- years of observation in those interrupt- lowed by discontinuation of all ing therapy at a CD4+ cell count below antiretroviral medications. The primary 200/µL, compared with .03 per person- outcome was proportion with plasma years of observation in those interrupt- HIV-1 RNA level below 5000 copies/mL The most common adverse effect was ing therapy at a CD4+ cell count above (“responders”) at week 52 (12 weeks off local injection site reaction (in 69%), lead- 200/µL. As this is an observational therapy) and week 96 (56 weeks off ther- ing to discontinuation in 3 participants cohort study, caution should be exer- apy). The proportion of responders was (8%). There were no other adverse effects cised in interpreting the implications of 18% at week 52 and 11% at week 96. clearly related to pentafuside. This study the findings for STIs in the setting of Response was predicted by a low pre- was not powered to show efficacy. chronic HIV infection. HAART plasma HIV-1 RNA level and lack A note of caution was also made by of rebound during STI and was not cor- investigators studying the dynamics of related to HIV-specific immunity as HIV in the cerebrospinal fluid (CSF) after measured. This led the investigators to All participants had at least 3 months of treatment interruption (Abstract 49). conclude that this intervention would prior zidovudine use. Although not statisti- Ellis and colleagues demonstrated a rarely be sufficient for the long-term cally significant, nephrolithiasis, hyper- rapid rebound of HIV in the CSF among management of patients and that the bilirubinemia, and hyperlipidemia were more common in the bid dose group. patients interrupting therapy, which “autovaccination” hypothesis was prob- Drug interruptions or dose reduction was lagged only several days behind the lematic. The largest CD4+ cell drop more common in the bid arm than in the plasma HIV-1 RNA level. Four of 12 occurred during the first 12 weeks off tid arm (48% vs 30%, P=.05). patients tested also developed a brisk therapy followed by a slower decline pleocytosis (white blood cell count thereafter. One person required salvage >10/µL), which followed viral rebound. therapy because of the emergence of The neurocognitive consequences of virologic resistance, and 2 patients Pharmacokinetic data showed lopinavir this viral rebound are a planned focus of developed symptoms consistent with levels were in the expected range but future investigations. the acute retroviral syndrome. amprenavir levels were reduced by about Another theoretical concern about Plana and colleagues evaluated STI 30%. treatment interruption is that viral with and without hydroxyurea (Abstract rebound can expand the size of the viral 535-M). Twenty patients who began reservoir in latently infected T-cells that HAART in chronic infection with a base- The efavirenz and efavirenz/nelfinavir arms decay slowly on HAART. Blankson and line plasma HIV-1 RNA level above 5000 were superior to the nelfinavir-only arm colleagues (Abstract 491-M) closely fol- copies/mL and CD4+ cell count above (P=.028 and P<.001, respectively). The lowed 5 individuals during STI to char- 500/µL who had achieved virologic sup- nRTIs were open-label and based on prior acterize the latent reservoir in CD4+ T pression for more than 32 weeks were regimens. cells. They found that the frequency of enrolled in this study. The antiretroviral latently infected CD4+ T cells rapidly regimens were changed to stavudine/ increased to levels similar to those in didanosine/nelfinavir with or without All patients were naive to NNRTIs and aba- primary HIV infection, but also decayed hydroxyurea for 24 weeks. The patients cavir, but nearly all had received stavudine rapidly after reinstitution of HAART, then underwent 5 treatment interrup- and didanosine. Plasma HIV-1 RNA level in leading the investigators to conclude tions separated by 2 months using the the hydroxyurea arms was more likely to that the size of the latent reservoir was same HAART regimen. Hydroxyurea was be suppressed (P=.008), but toxicities were not dramatically increased during STI. continued after the third interruption more common in these groups, including The Swiss-Spanish Intermittent without subsequent discontinuation. 2 cases of lactic acidosis, 2 cases of Treatment Trial is a prospective trial of They found that 8 of 9 hydroxyurea peripheral neuropathy, and 1 case of treatment interruptions in patients recipients maintained a plasma HIV-1 grade 3 hepatitis. beginning HAART in the chronic phase RNA level below 5000 copies/mL com- of infection. Hirschel presented data on pared with 4 of 10 without hydroxyurea, behalf of the study team in both the pre- (P=.02). The HIV-specific CTL did not viously mentioned symposium on con- differ between arms. It is not possible to troversies in the management of HIV distinguish whether STI had an effect in and a poster session (Abstracts 528-M, addition to the antiretroviral activity of the risk was still 2.4 times higher in S18). This is the largest study to date hydroxyurea in this study. those interrupting therapy (P=.0001). examining STI in this situation. They This implies that antiretroviral regimens studied 133 individuals, 96% of whom Prior to Salvage Therapy. Ruiz and col- having poor virologic and immunologic began antiretroviral therapy during the leagues evaluated the utility of an STI response are associated with a slower chronic phase of their infection with a prior to a salvage regimen in 46 patients
27 International AIDS Society–USA Topics in HIV Medicine
in whom a third HAART regimen was Hydroxyurea combined with didano- helper response was increased among failing and who had resistance muta- sine was studied as an attempt to simpli- the long-term nonprogressors and nega- tions in all 3 drug classes (Abstract 421- fy the antiretroviral therapy regimens of tively correlated with HIV-1 RNA level W). Patients eventually received a patients with virologic suppression on and viral DNA. In addition, the antibody “mega-HAART” regimen consisting of HAART for more than 1 year (Abstract response was also negatively correlated didanosine, abacavir, lamivudine, saq- 533-M). Patients were randomized to with HIV-1 RNA level, in particular the uinavir soft-gel capsule, and lopinavir/ continuing HAART or changing to IgG2 response. The combination of IgG2 ritonavir. They were randomized to an hydroxyurea/didanosine. In an intent-to- response and T-helper response as mea- immediate switch to this regimen or to treat analysis, 41 (38%) of 107 patients in sured by interferon-gamma ELISPOT starting after discontinuing their previ- the continuing HAART arm maintained was the best predictor of the mainte- ous antiretroviral medications for 3 virologic suppression at 48 weeks com- nance of the long-term nonprogressor months. There was a high prevalence of pared with 45 (39%) of 116 patients in the status in a multivariate model. genotypic mutations that subsequently hydroxyurea arm. Success on the hydrox- The immunologic goals of a thera- faded in those randomized to treatment yurea/didanosine regimen was correlated peutic vaccine, as outlined by Autran, interruption. The plasma HIV-1 RNA with lower baseline plasma HIV-1 RNA are to produce a strong and durable T- level increased by 0.9 log10 copies/mL level, lack of didanosine experience, helper response and a more intense and and the CD4+ cell count decreased by higher nadir CD4+ cell count, and shorter broader CTL response. This vaccine 131/µL in those off therapy. In 47% of time on HAART. Patients continuing on would be given during HAART prior to a patients receiving the STI, plasma HIV-1 HAART had an increased number of treatment interruption. The clinical goal RNA level declined below 80 copies/mL adverse reactions compared with pa- would be to lower the viral set point and compared with 36% of those not receiv- tients in the hydroxyurea arm. prolong the time off antiretroviral thera- ing the STI (P=not significant). This py. Several trials constructed in this study had limited power to detect subtle Therapeutic Vaccines manner are currently under way using treatment effects. No adverse events viral vectors (canarypox and replication- were noticed, but participants started Autran reviewed immune reconstitution defective adenovirus) that express por- with a high median CD4+ cell count, and the state of therapeutic vaccine trials tions of the HIV genome, some in com- 392/µL and 322/µL in the STI and imme- (Abstract L3). Patients in the chronic bination with interleukin-2 (IL-2) or a diate switch arms, respectively. phase of HIV infection commonly have DNA prime. Although no results are HIV-specific T-cell responses as deter- available, these vaccines were able to
Switch Studies and Simplification mined by more sensitive assays such as induce TH1 responses, and the measure- the interferon-gamma enzyme-linked ment of the CD8+ cell response is in NEFA is a multicenter, randomized, immunospot (ELISPOT). However, the progress. She ended with the promise of open-label trial examining the virologic strength of these responses is dimin- new trials involving improved vaccines response and adverse effect profiles of ished compared with those in individuals in the context of a large international changing either a single or dual protease treated during primary infection. The effort. inhibitor to one of 3 drugs: nevirapine, HIV-specific CD8+ T cells decrease pro- Several poster presentations dis- efavirenz, or abacavir (Abstract LB17). portionately with the decline in plasma cussed therapeutic vaccines in both ani- Four hundred sixty participants were HIV-1 RNA level associated with HAART. mal models and human subjects enrolled and the primary outcome was This led researchers to explore reconsti- (Abstract 312-W). DermaVir (George- proportion with undetectable plasma tution of HIV-specific T-cell responses town University Research Institute for HIV-1 RNA after 12 months. The intent- using HIV itself through STI. Preexisting Genetic and Human Therapy, Washing- to-treat analysis showed 78%, 74%, and immunodominant HIV-specific CD8 T ton, DC) is a novel topical immunization 77% of individuals had plasma HIV-1 cells have been found to rebound with composed of plasmid DNA combined RNA levels below 200 copies/mL at 12 STI, but this rebound is directly related to with polyethylenimine-mannose (PEIm). months in the nevirapine, efavirenz, and plasma HIV-1 RNA level, and control of The PEIm allows the plasmid to enter abacavir arms, respectively (P=.7). The viremia in the absence of HAART has not the dendritic cells and avoid degrada- on-treatment analysis showed 94%, 94%, been predictably or routinely achieved. tion by the endosome. The dendritic and 87% had plasma HIV-1 RNA levels Autran then discussed the work of her cells then present the DNA-based anti- below 200 copies/mL in the 3 arms laboratory and others in the search for gens to naive T cells. The investigators respectively (P=.06). More treatment dis- immune, viral, and genetic correlates of studied 10 SIV-infected macaques with continuations due to adverse effects protection against HIV by studying long- late-stage AIDS that were randomized to were seen in the efavirenz (17%) and term nonprogressors. She was able to HAART alone or STI/HAART for 6 cycles nevirapine (16%) arms than in the aba- rule out any viral factors being responsi- followed by DermaVir (using SIV DNA) cavir arm (7%), but abacavir was associ- ble in the patients studied. There did and 4 additional cycles of STI/HAART. All ated with a higher rate of virologic fail- seem to be genetic factors such as CCR5- of the continuous HAART macaques ure. Twenty-three of 28 participants from ∆32 heterozygosity and various human died, but only 1 STI macaque died. The all 3 arms who developed virologic fail- leukocyte antigen alleles that were asso- macaques had consistent viral rebounds ure had received single or dual nRTI ciated with virologic control. Both the during the first 6 cycles, but these therapy prior to beginning HAART. intensity and breadth of the CD4+ T- rebounds declined sharply after
28 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
DermaVir and the median viral load after the use of dual or triple antiretroviral sion). Data were presented on the first the last STI was below 200 copies/mL. therapy. 43 subjects (Abstract 527-M). At 48 Virologic control was associated with a weeks, the mean CD4+ cell count was strong SIV-specific T-cell response. The Immunotherapy higher in the IL-2 group (1669/µL vs authors note that the pattern of progres- 686/µL, P<.001), with 87% having an HIV- sive virologic control in association with Levy and colleagues presented the final 1 RNA level below 50 copies/mL com- DermaVir was similar to macaques treat- analysis of the ANRS 079 trial, which pared with 70% of controls (P=not sig- ed early after infection with STI/HAART, studied the efficacy of IL-2 therapy in nificant). Three of 5 participants tested and was remarkable considering the HIV infection (Abstract 514-M). Patients in the IL-2 arm developed CTL respons- macaques' advanced stage of disease naive to antiretrovirals or protease es to new epitopes compared with 1 of 7 when starting this treatment. inhibitors were randomized to start controls. By protocol, all subjects were Tryniszewska and colleagues studied HAART (stavudine, lamivudine, and to remain on HAART, but the 1 respond- 2 highly attenuated poxvirus vectors indinavir) or HAART plus 10 cycles of IL- ing participant in the HAART-only arm expressing different SIV proteins 2 given over 74 weeks. The median and 1 of the 3 responding patients in the (NYVAC-SIV-gpe and NYVAC-SIV-rtn) with length of follow-up of the 118 patients IL-2 arm had a treatment interruption and without IL-2 in SIV251-infected rhe- randomized was 35 months. The study that can increase the CTL response. The sus macaques receiving HAART com- had some protocol deviations, with par- authors found the 2 additional CTL pared with controls for 8 months ticipants in both arms receiving cycles of responses in the IL-2 arm encouraging (Abstract 313-W). They were able to IL-2 after week 74. The IL-2 arm had a but say that they may have been due to demonstrate CD8+ T-cell response to significantly higher and faster CD4+ cell undetected “blips” associated with IL-2 the immunodominant Gag epitope in count rise through week 74 than the injections. These data are too prelimi- vaccinated macaques. The vaccinated HAART-only arm (835/µL vs 262/µL, nary to draw any conclusions at this monkeys had a significantly lower viral P<.0001). The CD4+ cell count rise time. rebound after discontinuing HAART. The remained significantly higher in the IL-2 Lu and colleagues reviewed the gene authors argue that this provides proof- arm than in the HAART-only arm at the therapy studies of the National of-concept for therapeutic vaccines in last analysis (604/µL vs 365/µL, Institutes of Health/National Institute of the model that best predicts HIV infec- P=.0002). The proportion of participants Allergy and Infectious Diseases involv- tion. with a plasma HIV-1 RNA level below 50 ing serodiscordant monozygotic twins There were several studies using copies/mL was similar in both groups (Abstract 525-M). The 2 studies involved REMUNE (Immune Response Corpora- (76% vs 78%). syngeneic lymphocytes virally trans- tion, Carlsbad, Calif), an envelope- In a substudy of ANRS 079, investiga- duced with the chimeric receptor gene depleted, inactivated Zairian isolate tors examined the change in HIV provi- (CD4/CD3-zeta) or the neomycin phos- given with incomplete Freund's adju- ral DNA in PBMCs over time (Abstract photransferase gene. Genetically modi- vant. Bucy and colleagues studied 515-M). The decay in log DNA/106 fied CD4+ T cells persisted for a mean of REMUNE compared with the adjuvant in PBMCs was similar in the HAART-only 3.3 and 5.2 years, respectively, and this 28 human subjects who underwent and IL-2 arms, suggesting that IL-2 did persistence was enhanced by IL-2 treatment interruption (Abstract 314-W). not enhance the effect of HAART on the administration. They argue that these They found a non-significant decrease in reservoir of proviral DNA in PBMCs nor results provide a basis for pursuing the peak and post-peak viral loads off did it expand the pool of latently infect- genetic strategies for the treatment of therapy. Robbins and colleagues were ed T cells. HIV infection. able to demonstrate an augmentation of The ESPRIT trial is an open-label, Pomerantz and colleagues presented HIV-specific helper T-cell responses in 5 randomized, controlled trial assessing results from the Residual HIV-1 Disease of 5 REMUNE-vaccinated individuals HAART versus HAART plus IL-2 in terms Eradication Trial, an attempt to eradi- compared with 0 of 4 controls (Abstract of clinical outcome (Abstract 517-M). cate the latent reservoir of HIV in 3 315-W). All patients remained on thera- This ambitious effort will study 4000 patients (Abstract 405-T). These patients py and did not undergo a treatment participants for 5 years. Labriola and were on potent antiretroviral therapy interruption. colleagues presented preliminary data with HIV-1 RNA levels below 50 Fernandez-Cruz and colleagues pre- on the predictors of CD4+ cell count rise copies/mL for 1 year prior to enrollment. sented data from the STIR-2102 trial that associated with IL-2. The CD4+ cell Their regimens were intensified for at studied time to virologic failure for count response among patients receiv- least 1 month with hydroxyurea and patients receiving zidovudine/didano- ing IL-2 was greater among those with a didanosine. After receiving anti-CD3 sine or stavudine/lamivudine/indinavir higher nadir CD4+ cell count and was antibodies and IL-2, they continued the with and without REMUNE every 3 inversely related to length of time on same antiretroviral regimen for an addi- months (Abstract 318-W). They found antiretroviral therapy. tional 5 to 6 months. Tonsillar biopsies that REMUNE was associated with more Hecht and colleagues presented pre- showed no HIV-1 RNA by in situ durable virologic suppression (plasma liminary results of a randomized study hybridization in either follicular dendrit- HIV-1 RNA <5000 copies/mL) than comparing HAART alone and HAART ic cells or lymphocytes. Unfortunately, antiretroviral therapy alone. They did with IL-2 in the treatment of early HIV the plasma HIV-1 RNA level eventually not stratify time to virologic failure by infection (<12 months post-seroconver- rebounded in all 3 participants.
29 International AIDS Society–USA Topics in HIV Medicine
Blips medications. Sexual behavior and HIV tive was to determine if emergence of serostatus were assessed at baseline the M184V mutation during virologic Blips in viral load are defined as single and every 6 months for 2 years. suppression was predictive of virologic measurements of HIV-1 RNA level above The investigators found that high-risk failure. Of 26 patients having M184V at 50 copies/mL with subsequent measure- sexual behavior actually decreased after 24 weeks, 12 (46%) eventually had treat- ments below 50 copies/mL. Previous beginning the study. Seventy-three par- ment failure compared with 28 of 93 reports indicate that these blips are not ticipants began PEP 110 times, and 101 subjects (30%) without M184V. These associated with subsequent virologic (91.8%) of the sexual acts that preceded differences were not statistically signifi- failure. Havlir and colleagues presented PEP initiation met criteria for high-risk cant and the authors concluded that data on blips occurring during salvage behavior, prompting investigators to emergence of M184V during virologic therapy of heavily treatment-experi- prescribe a 28-day course of zidovu- suppression was not predictive of treat- enced patients (Abstract 93). These data dine/lamivudine. Eleven seroconver- ment failure. came from participants enrolled in sions occurred during the study, 10 ACTG 398—a study of protease among participants who did not start Resistance to Zidovudine and inhibitor- and NNRTI-experienced pa- PEP after high-risk encounters and 1 Stavudine tients randomized to efavirenz, abacavir, who did. The viral genotype from this adefovir, and amprenavir, with or with- person showed the M184V mutation. Mellors’ overview of nRTI resistance out a second protease inhibitor—and (Abstract L6) also described the current- ACTG 359, in which indinavir- and nRTI- ly accepted mechanism of zidovudine experienced patients were randomized Viral Resistance resistance. Mutations conferring resis- to nelfinavir/saquinavir or ritonavir/ tance to zidovudine do not appear to saquinavir with either delavirdine, ade- A number of presentations focused on affect incorporation of zidovudine into fovir, or both. They found that these resistance to antiretroviral therapies and the growing DNA chain in the manner blips did not predict subsequent viro- the mechanisms thought to underlie seen with lamivudine resistance. logic failure (RR, 1.13; 95% CI, .42-3.05). resistance. Phosphorolysis, a process by which ter- Di Mascio and colleagues examined minal dNTPs are removed from the DNA whether the frequency of blips in viral Resistance to Lamivudine chain by a phosphate donor such as load varied with time. They found that adenosine triphosphate (ATP), is 77.5% of 123 patients on their first In an overview of the major resistance thought to play a significant role in HAART regimen exhibited blips at a mechanism for lamivudine (and other zidovudine resistance. It is believed that mean frequency of .09 per sample with a “L-nRTIs,” including emtricitabine, L- phosphorolysis, which is more efficient mean amplitude of 165 plasma HIV-1 d4FC, and dOTC), Mellors (Abstract L6) in zidovudine-resistant viruses, removes RNA copies/mL. They concluded that described the mechanism of action of bound zidovudine from the DNA chain, viral blips appeared randomly, and that the M184V mutation, which confers 50- freeing the 3’-OH group and allowing for neither blip frequency nor amplitude to 100-fold resistance to these drugs. DNA synthesis to continue. Because increased with time. Studies have shown that M184V does zidovudine has a relatively long side not affect lamivudine's binding to the chain, it is not efficiently translocated Postexposure Prophylaxis active site of reverse transcriptase but away from the nucleotide binding posi- does markedly decrease its incorpora- tion (N) to the polymerization position A concern that has been expressed tion into the growing DNA chain. This is (P) that protects nucleotides from phos- about offering postexposure prophylaxis thought to occur because of steric hin- phorolysis. Mutations such as T215Y (PEP) after high-risk sexual encounters drance. In the 3-dimensional view, the increase ATP binding to the template, is that sexual behavior may be disinhib- ring structure of L-nRTIs points up resulting in more efficient phosphoroly- ited, leading to an increased frequency whereas the ring structures of endoge- sis. of high-risk behavior. To address this nous deoxynucleotide triphosphates The importance of nRTI-associated question, Schechter and colleagues (dNTPs) point down. When the methion- mutations (NAMs) in cross-nRTI resis- prospectively studied 202 HIV-seronega- ine at position 184 mutates to a valine, tance was stressed by a number of pre- tive homosexual men in a trial of the steric hindrance between the relatively senters. Mellors noted that increasing acceptability and behavioral impact of bulky valine molecule and the lamivu- numbers of NAMs correlated with PEP (Abstract 15). All participants dine ring occurs, malpositioning lamivu- increased cross-resistance among zido- received counseling on high-risk sexual dine and preventing its incorporation vudine, stavudine, and abacavir (Ab- behavior at baseline and were provided into the growing DNA chain. stract L6). Ross and colleagues (Abstract with a 4-day supply of zidovudine/ Eron and colleagues (Abstract 570-T) 568-T) noted that NAMs, specifically lamivudine. During the 2-year follow-up evaluated 119 subjects with plasma HIV- M41L, K70R, L219W, T215Y/F, and period, participants began antiretrovi- 1 RNA levels below 120 copies/mL who K219Q/E, were observed more frequent- rals after any sexual encounter they were enrolled in the NZTA4002 study, an ly in patients treated with the combina- deemed to be high risk. They were open-label trial of zidovudine/lamivu- tion of stavudine and didanosine than in instructed to present to the study site dine/abacavir/amprenavir versus zidovu- patients treated with stavudine and for evaluation within 4 days of beginning dine/lamivudine/nelfinavir. Their objec- lamivudine (41% vs 31%), suggesting
30 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
that the choice of dual-nRTI therapy may decrease in susceptibility correlates ed with hypersusceptibility to ampre- influence the development of NAMs. with a diminished response to tenofovir. navir and D30N confers hypersuscepti- Garcia-Lerma and colleagues (Ab- Tuske and colleagues (Abstract 44) bility to ritonavir. stract 571-T) have previously shown that solved the crystal structure of tenofovir Isaacson and colleagues used data viruses with T215D/C, a naturally occur- complexed with reverse transcriptase from the lopinavir/ritonavir expanded ring polymorphism that may be due to and suggest that the acyclic nature of access database comprising 792 revertant T215Y mutants, are more likely tenofovir allows it to avoid steric hin- antiretroviral-experienced patients who than wild-type viruses to select for drance, as occurs with lamivudine and started on combination regimens T215Y in the presence of zidovudine. In M184V. Additionally, the flexible nature including lopinavir/ritonavir to deter- this presentation they showed that 4 of of tenofovir allows it to assume multiple mine the genotypic predictors of 6 viruses with T215D/C selected for conformations at the reverse transcrip- lopinavir/ritonavir failure (Abstract 559- T215Y in the presence of stavudine in tase active site, making it a poor sub- T). They determined that the presence vitro, compared with 0 of 3 wild-type strate for excision by phosphorolysis as of mutations at positions 10, 20, 33, 36, viruses. They suggest that patients is seen with zidovudine. They concluded 54, and 82 in the presence of multiple infected with T215D/C viruses and treat- that the structure of tenofovir could other mutations was statistically signifi- ed with zidovudine or stavudine may be account for its slow development of cantly associated with virologic failure. at increased risk of quickly developing resistance. Mutations at 24, 47, 48, and 84 also T215Y mutants. seemed to influence virologic failure, Whitcomb and colleagues (Abstract NNRTI Hypersusceptibility but the association was not statistically 569-T) looked at nRTI susceptibilities Several sessions touched on NNRTI significant. They suggested that a from 2500 samples with matched phe- hypersusceptibility, which has been weighted algorithm based on these notypic and genotypic measurements described in the presence of nRTI muta- findings may be a better predictor of and saw that increasing numbers of tions. Swanstrom and colleagues looked response to lopinavir/ritonavir than the NAMs correlated with decreased sus- at the development of resistance to current mutation score. Conversely, ceptibilities to all nRTIs. M184V modu- delavirdine in the presence of nRTI- Loutfy and colleagues looked at 167 lates this cross-resistance, decreasing associated mutations conferring NNRTI patients on lopinavir/ritonavir therapy susceptibility to didanosine, zalcitabine, hypersusceptibility (Abstract 567-T). In and found that the current mutation abacavir, and lamivudine and increasing the subset of patients who had NNRTI score had a linear relationship to viro- susceptibility to zidovudine, stavudine, hypersusceptibility at baseline and who logic failure (Abstract 560-T). They fur- and adefovir. developed delavirdine resistance ther found that the presence of L90M (K103N), nRTI-associated mutations was highly predictive of failure. Resistance to Tenofovir conferring NNRTI hypersusceptibility Schwartz and colleagues (Abstract were retained. Additionally, the pres- 562-T) followed 41 patients on the Miller and colleagues reported on the ence of NNRTI hypersusceptibility did investigational drug tipranavir for 1 year effect of baseline nRTI mutations in not improve treatment outcome in this looking for reduced susceptibility. response to tenofovir therapy (Abstract study although it was associated with Commonly seen protease inhibitor 43). Based on pooled data from 2 stud- 0.54-log10 copies/mL lower baseline cross-resistance mutations including ies, they noted that the type and number plasma HIV-1 RNA level. Mellors and 46, 82, 84, and 90 did not confer of NAMs at baseline had significant colleagues reported on the positive tipranavir resistance. The V82T muta- effects on tenofovir response. Patients effect of NNRTI hypersusceptibility on tion in combination with an L33I/F/V with no NAMs at baseline had a 0.8-log10 the virologic response to an efavirenz- was associated with decreased suscepti- copies/mL drop in HIV-1 RNA on teno- containing salvage regimen in ACTG 398 bility in 4 of 6 patients. fovir, and those with 1 to 2 NAMs had a (Abstract 45; please see the “Trials in A molecular mechanism for ampre- 0.7-log10 copies/mL drop. Patients with 3 Treatment-Experienced Patients” sec- navir resistance was suggested by Xu or more NAMs including either M41L, tion). and colleagues (Abstract 563-T). L210W, or T215Y/F had only a 0.2-log10 Amprenavir resistance is primarily copies/mL decline in HIV-1 RNA on Protease Inhibitor Resistance mediated through I50V, although multi- tenofovir. Patients with 3 or more NAMs Obry and colleagues reported on hyper- ple other mutations are required for but without M41L, L210W, or T215Y/F susceptibility to protease inhibitors high-level resistance. It is known that had a 0.7-log10 copies/mL decline. using resistance phenotypes obtained the presence of I50V decreases the affin- Furthermore, the presence of L210W was from participants in the ANRS 088 ity of protease for amprenavir, but the diagnostic of 3 or more NAMs and was (NARVAL) trial of the effect of resistance molecular correlates of this finding are the strongest single marker of lack of testing on outcome (Abstract 557-T). not known. The authors suggest that the response to tenofovir. Resistance to They reported that mutations at codons loss of a methyl group when valine (V) tenofovir now appears to be mediated 30 and 88 in protease are associated replaces isoleucine (I) results in loss of by K65R (a rarely occurring mutation with significant cross-protease inhibitor hydrophobic contacts, creation of with tenofovir therapy), the T69S inser- resistance, but they confer hypersuscep- space, and loss of the close contact of tion, and multiple NAMs (>3.8 by statis- tibility to amprenavir and ritonavir. They protease with amprenavir needed for tical analysis). Phenotypically, a 4-fold further noted that N88S/T was associat- the drug to inhibit enzymatic activity.
31 International AIDS Society–USA Topics in HIV Medicine
Other Agents Grossman and colleagues compared 41 relation to viral fitness (Abstract 577-T). subtype-C patients and 67 subtype-B They determined that, although sub- True diketoacid integrase inhibitors patients in whom antiretroviral therapy stantial variability in fitness exists, in have now been identified and resistance had failed (Abstract 565-T). They found a general, viral fitness decreases as the has been induced in vitro by numerous mutation pattern that was generally number of NNRTI mutations increases. passages of virus in the presence of similar in both groups. The M184V lamivudine-associated these agents. Witvrouw and colleagues Additional information regarding mutation has been shown to result in (Abstract 573-T) identified new inte- treatment and resistance of non-B sub- decreased viral fitness. Wei and col- grase inhibitor resistance mutations types should be forthcoming as leagues analyzed a number of M184V selected for by L-708,906, a novel inves- antiretroviral treatment programs in mutant viruses to determine the bio- tigational integrase inhibitor, in vitro. Africa and Asia expand. Non-B subtypes chemical explanation for this finding As had been shown previously by continue to spread through Europe, as (Abstract 578-T). They presented data Hazuda and colleagues, the T66I muta- evidenced by J/A recombinants in newly suggesting that reverse transcriptase tion appeared early. With further pas- seroconverting injection drug users in containing the M184V mutation does sages, L74M and S230R mutations were Lausanne, Switzerland (Abstract 17), not recognize the initiation complex for also detected. This virus was highly and by the diverse subtype-G recombi- synthesis of viral DNA as well as wild- resistant to L-708,906 but retained full nants spreading through Southern type enzyme, and they postulated that susceptibility to entry inhibitors, nRTIs, Europe (Abstract 759-W). this may account for decreased viral fit- and NNRTIs, and protease inhibitors. ness in these viruses. Resistance and Fitness Resistance in Non-B Subtypes Resistance Testing There has been considerable interest in Several abstracts addressed potential the issue of viral fitness in patients with The utility of genotypic or phenotypic differences in treatment and resistance multiply resistant viruses. Barbour and testing in managing HIV infection has of non-B subtypes. Agwale and col- colleagues evaluated specimens from 20 been controversial. Several studies leagues phenotypically analyzed 14 sub- patients who were experiencing virolog- addressing the issue were presented. type-G and 4 subtype-A/G recombinants ic failure on protease inhibitor-based Torre and colleagues presented the from patients in Nigeria prior to starting regimens but who had plasma HIV-1 results of a meta-analysis of 6 random- antiretroviral therapy and found that all RNA levels lower than their baseline lev- ized controlled trials designed to deter- but 1 exhibited full susceptibility to all els (Abstract 575-T). Early virologic fail- mine the impact of resistance-guided antiretroviral medications, with the ure (ie, <6 months) in these patients antiretroviral therapy on virologic out- remaining sample having a 3.4-fold was associated with reduced viral come (Abstract 584-T). At 6 months, change to nelfinavir and ritonavir replicative capacity. The degree of reduc- 38.8% of patients who received genotyp- (Abstract 461-W). Portugal has an tion in replicative capacity was directly ic resistance testing had undetectable increasing prevalence of subtype-G associated with the decrease in plasma plasma HIV-1 RNA levels versus 28.7% of infections, leading Gomes and col- HIV-1 RNA level relative to pretherapy patients receiving standard of care. leagues to review the patterns of resis- level. However, with long-term failure, Expert interpretation offered in conjunc- tance for individuals in whom a nelfi- plasma HIV-1 RNA levels did increase tion with genotypic testing increased navir-based antiretroviral regimen is and this was associated directly with the rate of viral suppression (50.7% vs failing (Abstract 46). They found that a protease inhibitor resistance. 35.8%). Phenotypic testing did not seem nelfinavir-based regimen was 3 times Weber and colleagues evaluated to offer an advantage over standard of more likely to fail in subtype-G patients samples from 4 patients participating in care in this small meta-analysis. than in subtype-B patients. Moreover, ACTG 315 (zidovudine/lamivudine/riton- Perez-Elias and colleagues present- none of the 10 subtype-G patients in avir recipients) to determine if the base- ed the results of a prospective study of whom a nelfinavir-based regimen had line genetic background of the virus cor- 276 patients in whom antiretroviral ther- failed developed the D30N mutation related with viral fitness (Abstract 576- apy was failing who were randomized to compared with 6 of 11 patients with T). Two patients retained wild-type virus receive either actual phenotype or virtu- subtype-B (P<.01). The subtype-G and 2 patients developed resistant virus. al phenotype testing (Realvirfen study, patients exhibited either the L90M The resistant viruses had lower replica- Abstract 586-T). In a multivariate linear pathway or other mutations on genotyp- tive capacities than wild-type virus. One regression analysis, patients in the virtu- ic analysis. of the 2 patients had baseline natural al phenotype arm had a significantly Pillay and colleagues reviewed 113 compensatory mutations, and this greater decrease in plasma HIV-1 RNA pediatric patients from the Penta 5 trial, patient’s virus had an increase in viral level than those in the actual phenotype 67 (59%) of whom had non-B subtypes fitness over time. The authors suggest group (P=.01). Mazzotta and colleagues (Abstract 813-W). They did not find a dif- that the presence of background com- presented the results of the GenPherex ference in treatment outcomes or resis- pensatory mutations may allow for more trial comparing virtual and real pheno- tance patterns with respect to subtype, rapid recovery of replicative capacity. types in 173 patients in whom antiretro- but the number of each of the 7 non-B Soderberg and colleagues looked at viral therapy was failing (Abstract 589-T). subtypes or recombinants was small. the accumulation of NNRTI mutations in Virologic outcome did not differ
32 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
between the 2 groups and CD4+ cell receive TDM or standard of care. Those and colleagues suggested that the effect counts rose in both groups, with a trend in the TDM arm had significantly lower of plasma protein binding to the pro- toward a greater increase in the virtual toxicity than those in the standard-of- tease inhibitors may be overestimated phenotype group. The authors conclude care arm. and may not need to be included in IQ that virtual phenotype is as reliable as Castagna and colleagues looked at calculations (Abstract 449-W). They standard phenotype. the predictive value of the normalized reached this conclusion by looking at Workman and colleagues presented inhibitory quotient (NIQ) in determining the effect of plasma protein binding on the results of CREST, a multicenter ran- response to lopinavir in combination physiologic doses of saquinavir, riton- domized comparison of the effect of with ritonavir (Abstract 128). An NIQ was avir, nelfinavir, amprenavir, and the genotype and virtual phenotype testing defined as the ratio of the patient’s investigational protease inhibitor TMC on antiretroviral therapy-prescribing trough drug concentration to the fold 114. There was a less than 5-fold patterns (Abstract 587-T). The study change in IC50 of the patient's virus decrease in potency in the presence of included 330 patients in whom (inhibitory quotient [IQ]) divided by the the major protease inhibitor binding antiretroviral therapy was failing. No dif- ratio of the median population trough protein alpha-1-acid glycoprotein. ference was seen between the genotype drug concentration to the population TDM has been thought to be difficult and virtual phenotype groups in the cutoff fold change in IC50. This NIQ obvi- to implement for nRTIs because it is the numbers of changes made to planned ates the need for correcting for protein concentration of the intracellular drug regimens after resistance-testing results binding. In the study, 52 antiretroviral triphosphates that is important in their were available. It was noted that the vir- therapy-experienced patients were given efficacy. Becher and colleagues (Abstract tual phenotype arm had significantly lopinavir/ritonavir in combination with 452-W) developed a direct liquid chro- less resistance reported than did the other antiretroviral therapy chosen by matography/tandem mass spectrometry genotype arm; however, the clinical sig- their health care providers. Although assay to measure the active intracellular nificance of this finding is not known. NIQ was not predictive of initial triphosphorylated anabolites of stavu- This, however, is to be expected, response, it was highly predictive of dine and didanosine in peripheral blood because for a number of drugs, multiple response from week 12 onward, and at mononuclear cells. They were able to mutations are needed before a signifi- 48 weeks was more predictive than resis- quantify median intracellular concentra- cant change in susceptibility can be tance testing. Also, although 100% of tion (ICC) in 22 patients with plasma detected. patients with the highest NIQs had plas- HIV-1 RNA levels below detection limits ma HIV-1 RNA levels below detection for both stavudine and didanosine and Pharmacology limits, only 15% of those with the lowest suggest that further studies of ICCs of NIQ responded to therapy. nRTIs may be useful. The effect of intracellular triphos- Therapeutic Drug Monitoring Fletcher presented the results of the predictive value of IQs in patients on phate anabolite concentrations was also The clinical utility of therapeutic drug saquinavir plus either ritonavir or nelfi- studied by Hoggard and colleagues monitoring (TDM) is currently being navir as part of ACTG 359 (Abstract 129). (Abstract 455-W). They looked at levels debated. In a plenary talk, Back explored They found that IQ predicted response of intracellular lamivudine triphosphate the pros and cons of TDM and reviewed at weeks 4, 8, and 12 but not at week 16; (lamivudine-TP) and the ratio of lamivu- a number of trials (Abstract S20). In and had decreasing correlation with out- dine-TP to endogenous deoxycytidine favor of TDM are the well-defined rela- come after week 4. triphosphate (dCTP) in patients tionship between drug exposure and Phillips and colleagues reported on a responding to therapy and those in efficacy and toxicity, the large interpa- study of virtual inhibitory quotients whom therapy had failed. They deter- tient variability in drug levels, and the (VIQs) in antiretroviral therapy-experi- mined that ICCs of lamivudine-TP and narrow therapeutic window. Against the enced patients taking amprenavir and lamivudine-TP/dCTP were lower in non- use of TDM are high intrapatient vari- lopinavir/ritonavir in combination responders than in those responding to ability, unclear concepts of the role of (Abstract 130). A VIQ was defined as the therapy despite no differences in levels protein binding, logistics (more blood patient's trough drug concentration of endogenous dCTP between the draws, scheduling, etc), lack of expert divided by the virtual phenotype and the groups. interpretation, and the lack of random- protein-adjusted IC50. The VIQ of ampre- ized controlled trials showing efficacy. navir was highly predictive of suppres- Pharmacokinetics of Antiviral Drugs Only a few randomized controlled trials sion and there was a trend toward this of TDM have been carried out to date. with lopinavir. Eighty percent of patients Lopinavir/ritonavir Daily Versus Twice The ATHENA trial looked at the role of with a lopinavir VIQ greater than 15 and Daily. A great deal of interest is current- TDM in 92 antiretroviral therapy-naive an amprenavir VIQ greater than 1.3 had ly being shown in once-daily prepara- patients starting on a nelfinavir-based plasma HIV-1 RNA levels less than 50 tions of antiviral agents. M99-056, a regimen. At 12 months, those patients copies/mL. study of once-daily versus standard receiving TDM had a significantly lower The protein binding of antiretroviral twice-daily dosing of lopinavir/ritonavir plasma HIV-1 RNA level than those with- drugs, especially the protease inhib- was presented by Bertz and colleagues out TDM. In a separate arm, 55 patients itors, is thought to limit the usefulness (Abstract 126). Lopinavir/ritonavir was starting on indinavir were randomized to of IQs. A presentation by de Béthune given as a single 800 mg/200 mg dose
33 International AIDS Society–USA Topics in HIV Medicine
daily or as 400 mg/100 mg twice daily in interactions among antiretroviral tered with atazanavir without dose mod- a prospective trial. At 38 weeks, no sig- agents. Wire and colleagues studied the ification of atazanavir but suggested nificant difference was seen between the pharmacokinetics of amprenavir, given that rifabutin dose modification may be groups in the numbers of patients who as the prodrug GW433908, when given in necessary. Moyle and colleagues looked achieved plasma HIV-1 RNA levels combination with ritonavir (100 or 200 at the effect of pravastatin on protease below detection limits. However, it was mg) and efavirenz (Abstract 431-W). In inhibitors in patients suppressed on noted that the median IQ was lower in the 26 patients completing pharmacoki- protease inhibitor therapy (Abstract patients receiving the once-a-day dose netic analysis, the addition of efavirenz 446-W). No impact of pravastatin was compared with those on standard dos- to amprenavir and ritonavir had no seen on protease inhibitor trough con- ing (40 vs 84) and that the 24-hour effect on plasma concentrations of centrations and the authors concluded trough concentration was lower in amprenavir. Additionally, giving 200 mg that the efficiency and lack of interac- patients receiving once-daily dosing. of ritonavir instead of 100 mg did not tions with this agent make it a good These findings suggested that once- significantly increase plasma ampre- choice for therapy in patients with daily dosing may have a narrower thera- navir levels. hypercholesterolemia on antiretroviral peutic window than standard dosing, Solas and colleagues (Abstract 440- therapy. Cardiello and colleagues stud- and the conclusion was reached that W) looked at the effect of coadministra- ied the effect of itraconazole on twice-daily dosing is still the optimal tion of lopinavir/ritonavir with ampre- saquinavir soft-gel formulation way to dose lopinavir/ritonavir. navir. Plasma trough concentrations of (Abstract 447-W). They determined that amprenavir and lopinavir were mea- lower doses of saquinavir (800 or 1200 Stavudine Extended Release. Kaul and sured in 46 patients on 2 different doses mg vs standard 1400 mg bid) resulted in colleagues studied the pharmacokinet- of amprenavir (600 mg vs 750 mg bid) in adequate pharmacokinetic values when ics of once-daily versus twice-daily combination with lopinavir/ritonavir combined with 100 mg of itraconazole. stavudine in 2 studies. The first was a (400 mg/100 mg). Decreases of 51% in multiple-dose study (Abstract 429-W). the 600-mg dose group and 33% in the Compartmental Penetration. The distri- Fifteen HIV-infected volunteers received 750-mg dose group were noted for bution of antiretroviral agents into the either stavudine extended release 100 patients taking amprenavir in combina- cerebrospinal fluid and genital tract was mg daily or the standard 40 mg twice tion with lopinavir/ritonavir when com- the subject of several presentations. daily dosing over a 9-day period. The pared with those taking amprenavir/ Haas and colleagues looked at the effect once-daily dosing formulation achieved ritonavir. It was noted that 85% of these of ritonavir boosting on indinavir levels higher maximal concentrations and 24- patients still had median minimum con- in the cerebrospinal fluid (Abstract 437- hour area-under-the-curve values than centration values up to 3-fold higher W) and found that low-dose ritonavir standard dosing. No accumulation of than standard, non-boosted amprenavir increased indinavir levels 3-fold primar- drug was seen, and interpatient and dosing. Lopinavir levels were not affect- ily by increasing plasma levels. intrapatient variability was low. It was ed by coadministration with amprenavir. Sankatsing and colleagues looked at believed that the pharmacokinetic pro- It was suggested that TDM may play a seminal plasma levels of lopinavir in file supported once-daily dosing. role in amprenavir/lopinavir/ritonavir HIV-infected men on lopinavir-contain- The second study was a 48-week combination regimens. ing regimens for more than 4 weeks phase 2 trial of stavudine extended Preston and colleagues evaluated (Abstract 439-W) and found that seminal release versus standard stavudine in the pharmacokinetics of atazanavir in plasma levels were significantly lower combination with lamivudine and combination with efavirenz (Abstract than plasma levels. They concluded that efavirenz (Abstract 430-W). Intensive 443-W). Coadministration of these lopinavir has poor penetration into the pharmacokinetic studies on a subset of agents resulted in atazanavir levels that seminal plasma and are currently inves- patients were performed during the first were significantly lower than atazanavir tigating whether this will lead to subop- 2 weeks of the trial. For both prepara- alone, leading the authors to suggest timal viral suppression and resistance in tions, pharmacokinetic parameters were that atazanavir dosing should be adjust- this compartment. similar on days 1 and 14. The geometric ed if used in combination with efavirenz. Conclusion mean maximum concentration of the O'Mara and colleagues looked at the once-daily preparation was approxi- effect of adding low-dose ritonavir to The 9th Conference on Retroviruses and mately 50% lower than the standard atazanavir plus efavirenz (Abstract 444- Opportunistic Infections once again preparation. However, the geometric W). Adding 200 mg of ritonavir to the revealed itself to be a premier scientific minimum concentration was 5.5 times regimen increased atazanavir levels 3- meeting devoted to bringing together higher for the once-daily preparation. fold over atazanavir alone. advances in basic and clinical HIV These authors again concluded that the Other studies looked at the effect of research in 1 forum. Antiretroviral thera- pharmacokinetic studies supported the other agents on antiretroviral pharma- py remained a dominant component of use of once-daily stavudine prepara- cokinetics. Agarwala and colleagues the meeting and although no major tions. evaluated the effect of rifabutin on breakthroughs were evident, substantial atazanavir with or without ritonavir advances were reported with respect to Drug-Drug Interactions. Several presen- boosting (Abstract 445-W). They con- new antiretroviral agents, strategies of tations focused on the pharmacokinetic cluded that rifabutin may be coadminis- therapy, and the applications of drug
34 Conference Highlights - Antiretroviral Therapy Volume 10 Issue 1 March/April 2002
resistance testing and therapeutic drug Financial Disclosure: Drs Wilkin, Hay, and Hogan have level monitoring. In parallel, the many Additional Suggested Reading no affiliations with commercial organizations that may unanswered questions in the field that have interests related to the content of this article. Dr Sterling TR, Chaisson RE, Moore RD. HIV-1 Hammer has served as a site investigator for or consul- were highlighted will no doubt form the RNA, CD4 T-lymphocytes, and clinical response tant to Boehringer Ingelheim, Bristol-Myers Squibb, focus of many of next year's presenta- to highly active antiretroviral therapy. AIDS. GlaxoSmithKline, Pfizer, Roche-Trimeris, Shionogi, Shire tions. 2001;15: 2251-2257. Biochem, Tibotec-Virco, and Triangle.
35