 |
|
|
| |
New and emerging antiretrovirals, Comparative ART Trials, and when to begin HAART after an OI- now or later
|
| |
| |
CROI Boston 2008 report
David Margolis, MD, University of North Carolina
The steady advance of antivirals for the treatment of HIV infection continued at the 15th version of CROI (www.retroconference.org/2008). While the presentations did not offer the excitement of last years' announcement of the era of integrase inhibition, reports of several drugs, notably inhibitors of the CCR5 HIV entry co-receptor, were of significant interest. These presentations can be viewed at: http://app2.capitalreach.com/esp1204/servlet/tc?c=10164&cn=retro&e=8044&m=1&s=20376&&espmt=2&mp3file=8044&m4bfile=8044&br=80&audio=false
CCR5 Inhibitors
A phase 2 dose-finding study of the small molecule CCR5 antagonist vicriviroc (VCV) demonstrated that this agent is ready to move forward into full Phase 3 testing, to qualify for approval and licensure [abstr. 39LB]. Previous studies had raised questions about the dosing of VCV, and concerns that its use might be linked to the onset of lymphomas. Those past difficulties, likely due to flaws in study design and bad luck, seemed to be put to rest.
This study, optimistically entitled Victor-E1, compared VCV 20 or 30 mg once a day in combination with optimized background therapy (OBT) to OBT and placebo, over 48 weeks. Patients who had CCR5-tropic HIV detected at screening, had received NRTIs, NNRTIs, and PIs, and were failing ART with > 1000 copies/mL plasma HIV RNA were enrolled. OBT contained a ritonavir-boosted PI and any available ART except efavirenz.
116 subjects were randomized equally to VCV 20 mg, 30 mg or placebo QD, administered with a new OBT. Mean age was 45 years: 77% were male; 68% white, 17% black, 72% Latino. Mean CD4 count was 210 cells/mL, mean plasma HIV RNA was 4.5 log10 copies/mL, with about 30% having > 100,000 copies of HIV RNA/ml. 51% had an AIDS-defining event prior to study; 6 (5%) were HCV coinfected. About 25% of the patients used darunavir or enfuvirtide (T20) for the first time in the VCV arms, as compared to about 15% in the placebo arms. However, the number of active drugs by resistance assay in OBT was comparable between arms.
Virological responses were robust in both VCV groups. At 48 weeks, mean decline in HIV RNA was about -1.75 log10 in the VCV arms, and -0.80 in the placebo arm (p<0.002, active vs. PBO). Percent with HIV RNA <50 copies were slightly more than 50% in the VCV/OBT arms, and only 10% in the placebo/OBT group.
No toxicities or signals of safety issues were observed. Notably there was no significant liver toxicity, and only 1 case of lymphoma was observed. A trend towards greater effect in the 30 mg group was reported, especially in those with >100,000 copies/ml of HIV RNA, but not statistically significant. More patients receiving VCV were found to have CXCR4 or dual/mixed (X4/R5) virus emerge during the study.
A proof-of-activity study of a second R5 inhibitor, SCH532706 from Schering, was presented [Abstr. 38]. A 60 mg dose was given once a day with 100 mg of ritonavir for 10 days to 12 patients. A -1.6 log decline of viral load was seen. This response was variable, with 30% having a 2 log drop by day 15, and 17% having a < 1 log decline at day 10. Half-life is long (39.4 hours), suggesting once a day use. Most patients had mild side effects during dosing, and one had self-limited pericarditis after the study. Development for once-a-day use is planned. Resistance to the licensed R5 inhibitor maraviroc (MVC) sometimes develops due to the emergence of X4-using virus, as discussed below, but sometimes due to the development of resistance to MVC itself. Therefore, additional R5 inhibitor drugs in the development pipeline may be of use for overcoming R5 resistance, as well as other possible advantages such as improved side effect or safety profiles.
PF-232798, another CCR5 antagonist in development as a 2nd-generation drug by Pfizer (Poster 737), was shown to be active against MVC-resistant virus generated in culture. Structural modeling suggested that PF-232798 binds the CCR5 receptor in a slightly different way than MVC. PF-232798 was well tolerated in healthy human volunteers at single oral doses of up of 750 mg, and pharmacokinetics studies suggest that a once daily dose of 250 mg should be sufficient for antiviral effect.
Mike Westby from Pfizer then explored the reasons for failure of therapy in the MERIT study, a comparison of efavirenz (EFV) to MVC in combination with Combivir in ART-naive patients. The primary clinical results of this study, presented at IAS '07 in Sydney, showed that MVC was non-inferior to EFV at 48 weeks of therapy (65% and 69% suppressed to <50 copies/ml, respectively). 13.6% of EFV patients failed due to toxicities, whereas 11.9% of MVC patients failed due to loss of virological control.
A major concern, of course, in the use of R5 inhibitors is that they may result in "switching" in the viral swarm from viruses that use the CCR5 receptor to viruses that use the CXCR4 receptor for entry. There has been a longstanding chicken-and-the-egg debate about X4 virus. Although its detection as the predominant viral species in the peripheral blood is clearly associated with a higher risk of progression to clinical AIDS, it has never been clear that X4 virus is the DRIVER of accelerated pathogenesis, or a MARKER of a collapsing immune system.
As X4 viruses are associated with HIV disease progression, and found to predominate in about half of patients with clinical AIDS, there is great concern that a drug that allows for, or selects for, the emergence of X4 HIV may lead to worse outcomes. However, it is unclear whether the emergence of X4 virus is a clinically different event when it occurs during the progression of untreated HIV viremia than when it occurs in the face of R5 inhibitor selection in a patient failing R5 inhibitor therapy.
In the MERIT study, a screening receptor usage or tropism test was done using the phenotypic assay (Trofile), to prevent the entry of patients with circulating viruses that can use the X4 receptor from entering the study. Of course, MVC would be of no benefit to such patients, and theoretically might do harm by selecting for an expansion of the X4 population. Screening was done at a pre-entry visit, and on the day of entry.
Patients without dual/mixed virus (virus that is either a mixed population of X4 and R5-using virus or a virus population that can use either receptor) at screening entered the study. However, 3.5% patients (n = 25) were found to have D/M virus phenotype at study entry, only days after the initial screening viral phenotype was done. This is likely due to variability in the patient. For example, at screening 3% of the circulating virus population might have been D/M virus --- undetectable in the current Trofile assay --- but at on the day of entry the population might have shifted to 12% D/M virus and the patient scored as "having" D/M virus
The misclassification of these few patients is an expected phenomenon, as the current tropism assay cannot detect D/M virus reliably if its frequency in a population is <10%. There is evidence that using research assays very low levels of X4 virus can be found in patients that fail R5 therapy even before they receive an R5 inhibitor. This can be thought of just the same way we think of the pre-existence of the common M184V resistance mutation to 3TC and FTC, although the frequency of M184V is much higher. A major point of controversy was whether or not tropism assays should therefore be held to a higher standard, and be expected to detect populations of less than 5%, a feat currently unachieved by other standard resistance tests.
However, the detection of D/M virus makes MVC likely to fail, especially if it is a key part of the antiviral regimen. In the MERIT study, 11 of these patients with D/M HIV were randomized to efavirenz therapy, and 55% of these achieved <50 copies/ml at 48 weeks. Of the 14 patients with D/M virus randomized to MVC, only 1 achieved suppression at 48 weeks.
To understand why patients with R5 virus at baseline should fail maraviroc in the MERIT study, maraviroc plasma levels (obtained by periodic PK sampling) were obtained and studied. Of the 32 patients with R5 virus at screening and baseline who later failed therapy, 12 patients were found to have plasma levels of MVC that were below the level of quantification. In all cases this correlated with a rebound in viral load. In some patients this further correlated with documented interruptions or poor adherence documented in patient records. Such interruptions or poor adherence was documented in 5 of 6 patients whose viral load was <500 copies/ml at discontinuation, consistent with a period of poor adherence triggering a virological failure rule followed by re-adherence immediately prior to discontinuation from the study.
So overall: 360 patients randomized to receive MVC in MERIT
235 were < 50 copies/ml at 48 weeks
43 failed due to lack of virological efficacy (12%)
Of 14 with D/M virus at entry 13 failed (93% with D/M)
346 had R5 virus at screening & entry 32 failed (9% with R5)
Of those with R5 at screening & entry:
10 failures with X4 virus (31% of failures, 3% of total)
22 failures with R5 or untyped virus (69% of failures, 6% of total)
50% with R5/untyped virus failed with low MVC levels, but half did not
Therefore when MVC failed, about a third of the time it seemed due to X4-using HIV, a third of the time it seemed due to inadequate adherence, and a third of the time MVC failed despite apparent adherence and the predominance of R5 virus. There is yet no evidence that failing with X4 virus leads to a worse clinical outcome, but careful observation of these study cohorts is ongoing. In most cases when MVC is stopped, R5 virus predominance appears to return.
However, when MVC works it seems to have durable effect. 48-week data from the MOTIVATE studies (poster 792), initially presented at ICAAC, EACS, and CROI 2007, was shown. In these patients with 3-class resistance treated with optimized background therapy and once-daily or twice-daily MVC or placebo, MVC arms were superior with about 45% achieving <50 copies at 48 weeks, and the proportion of patients who had achieved virological success by week 24 appeared to remain stable through week 48.
Many investigators have noticed that R5 inhibitors may have a unique effect on CD4 count. Wilkin, Ribaudo, and Gulick (abstr. 800) conducted a meta-analysis of recent phase II and III trials in treatment-experienced subjects evaluating newer non-nucleosides (etravirine), protease inhibitors (daruanvir, tipranavir), integrase inhibitors (raltegravir), fusion inhibitors (enfuvirtide), and CCR5 inhibitors (maraviroc, vicriviroc), each in combination with an optimized background regimen.
They found that the use of a CCR5 inhibitor was associated with an additional gain of +32 cells/_L at 24 weeks compared to other regimens without a CCR5 inhibitor. In this model, CD4 cell increase was not associated with the baseline CD4 cell count, gender, or median age, and as might be expected was associated with virologic response at week 24. However, it has yet to be demonstrated that any additional increase of CD4 cells that might be obtained in regimens that contain an R5 inhibitor is durable, and that this increase is correlated with a functional increase in immune sufficiency (i.e. decreased clinical events).
Integrase Inhibitors
There was surprisingly sparse information on raltegravir (RAL) or other integrase inhibitors in development at the 15th CROI. Interesting and useful data on the 48-week responses to salvage therapy in the BENCHMRK studies was consigned to posters (788, 789). The data on the two posters was virtually identical, and superimposable over the response data previously shown over 24 week. In these studies, once patients achieved viral suppression at <50 copies/ml, failure and breakthrough viremia did not appear to occur at appreciable frequency. The 60-65% response rate achieved at weeks 16 and 24 appeared to be stable to 48 weeks, with nearly the full cohort of patients accounted for out to 48 weeks (462 patients on RAL in the studies at entry, 459 reported at 48 weeks).
As in the initial presentations at CROI 2007, 45-50% of patients with no active agents in background therapy (OBT) by genotypic score or by phenotypic score maintained <50 copies/ml at 48 weeks, only a little worse than 50-65% of patients with 1 active agent in OBT.
These results give reason to reconsider the view of raltegravir as a drug with a "low genetic barrier." Clearly, a single mutation results in significant viral resistance to RAL, however RAL appears to behave differently than an NNRTI. It is difficult to imagine that one would see success similar to that seen in BENCHMRK if one had treated patients with PI- and NRTI-resistance with OBT + efavirenz. Certainly, RAL should be used with at least one, if not two, fully active antiviral agents. But if that is not possible, there is still perhaps a nearly 50-50 chance of success.
In another related anecdote of interest, Harris and colleagues reported their experience in patients switched from enfuvirtide to raltegravir while on a virologically suppressive regimen (Abstr. 99). This clinical maneuver may appear attractive to patients having difficulty with continued administration of enfuvirtide. 29 patients had received ENF for 7 to 75 months (median 27.5), and had viral load of <50 copies/mL for 1 to 72 months (median 24) before starting RAL. Viral load remained <50 copies/mL in all patients at all time-points measured: month 1 (n = 21), month 2 (n = 13), month 3 (n = 9), and month 4 (n = 7). At least thus far, this preliminary, limited data suggests that such a switch may be safe.
Reverse Transcriptase Inhibitors
Following the 24-week results of the DUET studies, the 48-week interim analysis of these 96-week studies was presented (posters 790, 791). DUET-1 and -2 are identical, ongoing 96-week randomized double blind Phase III trials performed in different parts of the world evaluating the efficacy and safety of the new NNRTI etravirine (ETR) 200mg vs placebo, both given twice a day. All patients received a background regimen of darunavir with low-dose ritonavir (DRV/r), investigator-selected NRTI(s) and optional enfuvirtide. Patients had documented NNRTI resistance and 3 or more primary PI mutations.
In both studies about 60% of patients achieved the efficacy endpoint of HIV RNA <50 copies/ml by 24 weeks and maintained this response at the week 48 interim analysis. This compared with such a response in 40% of the placebo recipients, who had the benefit of DRV/r. As in BENCHMRK, responses improved when background therapy contained more active agents, but even with no other active agents and high levels of DRV resistance, about a third of the patients achieved and maintained suppression. Again, ETR appears to have a genetic barrier to resistance that is higher than first-generation NNRTIs. Pre-clinical data on four other novel NNRTIs with good antiviral activity despite the presence of NNRTI resistance mutations were also presented (posters 729, 730, 731).
Apricitabine (ATC) is a new cytidine analogue NRTI with activity against viruses with M184V (3TC, FTC resistance) and other NRTI resistance mutations (poster 793). In a phase IIb study, 600 mg or 800 mg ATC bid was compared to lamivudine (3TC) in patients failing therapy and known to have the M184V resistance mutation. During the first 3 weeks of the study, patients continued failing therapy and added one of the doses to ATC or 3TC in a blinded fashion. Background therapy was optimized at day 21. At baseline, 52% had ≥3 thymidine analogue mutations (TAM) and 76% had ≥1 NNRTI mutation.
ATC showed a -0.71 to -0.9 log reduction in viral load at day 21 (the primary endpoint) compared to 3TC (-0.03 log, p ≦0.05; ANCOVA). After therapy was optimized, all three arms achieved roughly equivalent declines of viral load by week 24, of ca. 1.6 logs. Further development of ATC may provide another NRTI option for therapy.
Amdoxovir (DAPD) another NRTI in development for several years, was known to be synergistic with zidovudine (ZDV) in culture. The anti-HIV activity of DAPD with and without 2 doses of ZDV was tested (poster 794).
24 subjects not receiving ART were randomized to DAPD 500 mg twice daily, DAPD and/or ZDV 200 mg or 300 mg twice daily, or placebo for 10 days. Viral load was determined daily. On day 10, mean VL log10 change was +0.10 with placebo, -0.65 with ZDV 200, -0.45 with ZDV 300, -1.07±0.80 with DAPD, -1.97±0.16 with DAPD/ZDV 200, and -1.67±0.21 with DAPD/ZDV 300. DAPD/ZDV 200 was significantly more potent than DAPD (p <0.04), suggesting synergy, and there was markedly decreased viral load variability with DAPD/AZT compared with DAPD alone. Viral load decline was significantly improved with DAPD/ZDV 200 and 300 mg compared with ZDV monotherapy (p ≦0.0001). The investigators concluded that co-formulated DAPD/ZDV warrants further development for second-line therapy of HIV infections.
Comparative ART Clinical studies
Truvada (tenofovir/FTC) is the most prescribed dual NRTI combination, but has not been directly compared to Epzicom (abacavir/3TC). Smith and collaborators presented a GSK-sponsored Phase IV, randomized, double blind, 96-week study conducted at 78 sites in the US and Puerto Rico to directly compare the efficacy of Epzicom to Truvada in 688 ART-naive subjects (Abstr. 774). HIV-1-infected, ARV-naive subjects with plasma HIV-1 RNA ≥1000 copies/mL (stratified < or ≥100,000 copies/mL) and any CD4+ count received either blinded ABC/3TC or TDF/FTC with open-label lopinavir/ritonavir (LPV/r) soft gel capsule once daily.
Protocol-defined virologic failure was defined as failure to achieve viral load <200 copies/mL by week 24 or confirmed rebound to ≥200 copies/mL. Primary efficacy endpoint was the proportion of subjects with viral load <50 copies/mL at week 48 (intent to treat, M = F, switch included analysis). For the difference in responses between arms, 95%CI was calculated to show non-inferiority using a 12% margin. Within-class switches were allowed for toxicity. HLA-B*5701 screening was not performed.
ABC/3TC was non-inferior to TDF/FTC when combined with once-daily LPV/r (68% vs. 67% achieved <50 copies at 48 weeks). The frequency of virological failure was indistinguishable between the two arms. Median CD4+ increase was numerically greater in the ABC/3TC arm at week 48. Rates of drug-related grade-2 to -4 adverse events were comparable between arms. In this study, treatment-emergent NRTI mutations were (numerically) more frequent in the Truvada arm. Both once-daily treatment regimens proved efficacious in this ART-naive population through 48 weeks.
On the protease inhibitor side of the therapy equation, BMS sponsored a head-to-head study to compare the efficacy of ritonavir-boosted Reyataz (ATV/r) to lopinavir/ritonavir (Kaletra, KAL). 48-week results were presented from a randomized, open-label, multicenter, ongoing 96-week study to assess non-inferiority (10% margin) of ATV/r 300 mg/100 mg once-daily vs KAL twice-daily, both in combination with fixed-dose tenofovir (TDF) 300 mg/emtricitabine (FTC) 200 mg once-daily, in treatment-naive patients (Abstr. 36). The primary endpoint was the proportion of patients with HIV RNA<50 copies/mL at week 48. 883 patients we randomized, and 878 were treated. Median CD4 cells were 205 cells/mm3; median plasma HIV RNA was 4.98 log10 copies/mL.
At week 48, mean CD4 increases from baseline for ATV/r and LPV/r were 203 and 219 cells/mm3, respectively. Fewer patients on ATV/r (2%) than LPV/r (8%) initiated lipid-lowering therapy. Patients on ATV/r had a lower incidence of grade 2-4 treatment-related diarrhea (2% vs 11%) and nausea (4% vs 8%) than LPV/r. Grade 3-4 ALT/AST elevations were low (≦2%) on both arms. Discontinuations prior to week 48 were: ATV/r, 9%; LPV/r, 13%. Adverse event-related discontinuations were 2% and 3% on ATV/r and LPV/r, respectively; 3 patients (<1%) discontinued ATV/r due to jaundice/hyperbilirubinemia. During this study KAL capsules were used, and so discontinuations that blunted the efficacy of KAL therapy may be decreased if KAL tablets are used.
At 48 weeks 78% of patients receiving ATV/r and 76% receiving KAL had suppressed to < 50 copies/ml. The 95% CI for the difference between ATV/r and KAL was +7% (favoring ATV/r) to -3.8%; thus ATV/r was non-inferior. The study was performed in Paris, Mexico, Peru, Thailand, Argentina, Brooklyn, and the Western Cape, South Africa. In this study more patients with low CD4 counts responded in the ATV/r arm than the KAL arm; this finding is not statistically robust enough to claim superiority for ATV/r in patients with low CD4 counts, but does suggest that ATV/r can be successful in "sicker" patients.
After an OI: to start ART now or later?
A Randomized Strategy Trial, ACTG 5164 (abstr. 142)
In a clinically important study, the ACTG attempted to measure the benefit, or lack thereof, of the initiation of ART in the setting of acute opportunistic infections. Most practitioners wait until patients have left the hospital and recovered from their OI, before initiating ART in an outpatient setting. In some cases, this may delay the initiation of ART for several months.
ACTG 5164 was a randomized strategy trial of immediate ART given within 14 days of starting acute opportunistic infections treatment vs deferred ART given after acute opportunistic infections treatment is completed (at least 4 weeks after randomization). Randomization was stratified by opportunistic infections and CD4 (< or ≥50 cells/mm3). Patients with tuberculosis were excluded. The study provided drugs including lopinavir/ritonavir (LPV/r), stavudine (d4T), and tenofovir/emtricitabine (TDF/FTC), but clinicians were free to use any standard ART. The primary week 48 endpoint was an assignment to one of 3 outcomes: a) death or AIDS progression; b) no progression and HIV viral load ≥50; or c) no progression and viral load <50 copies/mL.
141 patients were randomized to each arm (immediate vs. deferred ART). Subjects were 85% men, 37% black, 36% Hispanic, and 23% white with a median age of 38 years. Entry median CD4 count was 29 cells/mm3, and log10 viral load was 5.07 copies/mL. Entry opportunistic infections included Pneumocystic carinii pneumonia (PCP) 63%, cryptococcal meningitis 13%, pneumonia 10%. 30% had more than one OI identified at entry.
Immediate and deferred arms started ART at a median of 12 (range 9-13 days) and 45 (range 41-55) days, respectively, after treatment for the opportunistic infection had started. Note that even the "deferred treatment" arm started ART in less than 2 months. There were no significant differences in initial ART regimens.
Each arm had 18 subjects for whom "no endpoint information" was coded as endpoint, a lost to follow-up rate of 13%. There was no statistically significant difference in the primary endpoint: 14% vs 24% for death or AIDS progression; 38% vs 31% for no progression and HIV viral load ≥50; and 48% vs 45% for no progression and viral load <50 copies/mL in immediate vs deferred arms, respectively. Both arms achieved similar CD4 and viral load by week 24.
Importantly, the immediate arm had: fewer deaths/AIDS progressions (20 vs. 34, p =0.035), longer time to death/AIDS progression (stratified HR = 0.53, p = 0.02), and shorter time to achieving an increase in CD4 counts to >50 and >100 (median 8.1 vs 3.9 weeks and 11.8 vs 4.2 weeks), respectively. There was a trend of earlier ART changes in the immediate arm (p = 0.15), but no significant differences in grade 3 or 4 adverse events, adherence, hospitalizations, or immune reconstitution inflammatory syndrome (8 immediate vs 12 deferred).
Although there was no significant difference between immediate and deferred ART in the primary endpoint that includes both clinical and virological response, immediate ART reduced death/AIDS progression over 48 weeks. The authors concluded that, absent contraindications, consideration should be given to early use of ART in HIV-infected patients presenting with an acute opportunistic infections.
It should be remembered that this was a clinical trial, using the resources of study coordinators and other adjunct staff available to study sites. Further, sites were highly incentivized to maintain follow-up in study patients, as the ACTG penalizes sites for lost-to-follow-up. On the other hand, there was a trend towards benefit of early ART, even though deferred ART was not very delayed. In our practice, this has stimulated a reconsideration of immediate ART in the setting of an OI.
|
| |
|
|
|
|
|
