 |
|
|
| |
Selected review of HIV resistance presented at the 41st ICAAC
Andrew R. Zolopa, MD
Stanford University
|
| |
| |
Topics:
--HIV Drug Resistance in treatment experienced and newly infected
--Interpretation of genotypic test results: ENVA-3 Study is a survey of
reported results from 132 labs
--Impact of the M184V mutation (3TC resistance) on phenotypic resistance to
NRTIs
--Response to SQV/RTV when PI resistance is present (82A and 90M mutations)
--Phenotypic resistance test (Phenosense assay) adapted to test for non-clade
B HIV
--Tenofovir resistance report
--Amprenavir Pharmacokinetics: interaction of new "908" amprenavir prodrug
with efavirenz in once & twice daily regimens when using RTV for boosting
The Current Epidemiology of HIV-1 drug resistance
In a late breaker session on Tuesday December 18, Dr. Doug Richman from UCSD, presented data on drug resistance from a population-based sample of HIV-infected adults in care in the US-based clinics during the late 1990's. The presentation received wide coverage by the media and as is usually the case some of these reports have, unfortunately, misinterpreted some of the study findings.
The HCSUS (HIV Cost and Services Utilization Study) is a well-described population-based weighted sample of HIV-infected adults who were receiving care in 1996 in the US. The sample has been carefully constructed so as to be representative of all HIV-infected adults in care in the late 1990's. In this study, the investigators took a random sample of plasma specimens from 1,906 individuals collected between 1998 and 1999. From the samples in which the viral load was detectable (ie >500 copies) the phenotypes were determined using ViroLogic's Phenosense test.
Over 1/3 (36.5%) of the sample had viral loads below 500 copies and were not phenotyped and assumed NOT to have clinically significant drug resistance. In the remaining 63.5% of samples, the investigators found that 78% had at least 1 drug in which the phenotype showed resistance. Drug resistance to the nucleoside class was most common with 75% showing resistance to at least one drug from this class with resistance to 3TC being most prevalent. Resistance to PI's was seen in 42% of samples and NNRTI resistance was seen in 31%. Over half (51%) of the sample showed resistance to 2 or more classes for antiretrovirals and 14% had triple class resistance.
Factors associated with resistance were investigated. Currently being on antiretrovirals, having more advanced disease and lowest CD4 count were all associated with more drug resistance. In addition, several demographic features appeared to be associated with resistance including, male sex, men who have sex with men, white race and living outside of the midwest region. It is likely that all of these demographic variables are related to better health care and earlier access to antiretroviral therapy that was likely to have been introduced in a suboptimal manner as pointed out by Dr. Richman. It is important to emphasize that these factors should not be seen as "causing" resistance. Furthermore, these results can not be extrapolated to other populations, for example, newly infected populations. Finally, of interest, the investigators found that individuals being care for at centers with large HIV practices tended to have LESS resistance - implying that expert care found at large centers may be better able to treat effectively and avoid drug resistance.
In another study, investigators from Consolidated labs (Van Nays, CA) showed that drug resistance mutations were commonly seen in a small group (n=38) of clinical isolates from patients on ART (which was not adequately characterized in the presentation) with low level viremia (viral loads 100-<400 copies). The implications of this study are not clear as it can not be determined how much of the resistance measured is from failed prior therapy and how much is evolving on current therapy [I-1750, Lee]
Drug resistance and treatment response in treatment naive groups
In an ongoing cohort analysis, investigatorss from Boston presented the results of drug resistance in a group of 88 HIV-1 infected treatment naive subjects. Drug-resistance was seen in 16 (18%) with resistance to nucleosides most commonly seen. Twelve subjects with resistance mutations at baseline started on ART and all achieved viral loads of <50 copies by 48 weeks including 3 subjects in whom resistance was present to at least one of the drugs in their initial ARV regimen. In this small subgroup, time to undetectable viral load appeared delayed but all achieved <50 copies by week 48.
Drug resistance in a cohort of chronically infected but ARV naive individuals from Spain appeared to remain at low levels but increasing slightly over the four years of study (1997-2001).
Overall, from these presentations it appears that HIV-1 drug resistance is very common in treated populations and is increasingly common finding in cohorts of newly infected cohorts. The implications of these trends are indeed worrisome and require concerted efforts at HIV prevention and development of new treatments.
Interpretation of Genotypes: Are we making progress?
Dr. Schuurman presented results from the ENVA-3 study - a world wide evaluation of genotyping proficiency. A panel of 5 isolates with different patterns of drug-resistance mutations were sent to 132 centers for a blinded analysis. The laboratory proficiency was presented at this past summer's resistance workshop and showed that there was about the same level of accuracy as seen in the earlier ENVA 2 study, that is that about half the labs are very accurate and the other half are variably accurate. In addition to studying the laboratory proficiency the investigators asked participating lab directors to "interpret" the genotype result. Centers were asked to rate the expected drug activity for the determined genotypes. There was a variety of different interpretation methods used including "homebrew" algorithms (28%) VGI's TrueGene software (24%), ABI HIV-GT (13%) etc. Dr. Schuurman showed that there was wide variation in the interpretation of the genotypes. For example, in an isolate with reverse transcriptase mutations at 41, 184 and 215; 32% of centers scored 3TC as an active drug while the rest scored it inactive. The same variation could be seen with PI's, for example in a isolate with multiple PI-related mutations including mutations at codon 82, 1/3 of centers called nelfinavir active, 1/3 called nelfinavir partially active and 1/3 called nelfinavir inactive. Other recent studies have shown similar variation across different interpretation algorithms.
The limitation of this particular analysis is the exact meaning of "active" drug. It is likely that different lab directors interpreted "activity" differently, confusing phenotypic sensitivity, clinical activity and willingness to use a particular drug in a treatment regimen. Despite these findings it appears that there is a consensus evolving in the area of genotypic interpretation and we now have a FDA approved algorithm (TruGene report) which is likely to lead to greater standardization in this rapidly evolving field.
Another area of concern in the interpretation of resistance testing is how representative are the results of these plasma-based tests? Do they accurately reflect viral populations of other tissue or cellular compartments? One of the most important reservoirs of viral replication is the lymphatic system. Investigators from UCLA in collaboration with Virco demonstrated that the genotypes and phenotypes from plasma specimens were highly concordant with that of paired gut-associated lymphatic tissue (GALT) specimens. In a few cases, major drug -resistance associated mutations were present in the plasma but appeared only later in the GALT. The GALT was collected through colonscopic biospy of GI mucosa of 12 patients on stable antiretroviral therapy [I-1749] How concordant the results would be in other more heterogenous groups of patients remains to be defined but on the basis of this study and others it appears that the plasma is reasonably reflective of the dominate quazispecies in the lymphatics.
Editorial Note: Interpretations by labs of resistance test results may not be reliable.
Impact of the M184V mutation on nucleoside analogs (nRTI's) phenotypes
The M184V mutation associated with high level resistance to 3TC has also been associated with attenuation of AZT resistance in isolates with 1 to 2 AZT-associated resistance mutations (AKA TAM's - thymidine analog associated mutations or NEM's -nucleoside excision mutations). Shulman and colleagues demonstrated the impact of the 184V mutation on D4T, AZT, ABC and DDI phenotypes from samples taken from the ACTG 364 study. In this study, patients were previously treated with mono and dual nRTI therapy for several years received two nRTI's with efavirenz and or nelfinavir. Using baseline samples from 130 of the original 196 subjects enrolled, the investigators evaluated the impact of the 184V on baseline susceptibility of nRTI's. As expected, the median IC50 fold change for AZT was 95 in the group without 184V and only 9.3 in the group with the 184V despite similar distribution of TAM's in the two groups. Similarly, the median fold change for D4T although a much smaller dynamic range was also attenuated by the 184V, 2.7 versus 2.0 (p<0.01). Conversely, the 184V was associated with significant increases in IC50's of DDI (increased from 1.4 to 1.8) and abacavir (increased from 3.3 to 6.9). Unfortunately, the investigators could not demonstrate any improvement in virologic response related to the presence of the 184V mutation in subjects who received D4T containing regimens.
Protease mutations 82A and 90M, Saquinavir phenotype and response to SQV/RTV
Recent reports have described saquinavir phenotypes with no or only low level reduced susceptibility despite the presence of major protease inhibitor-related resistance mutations. For example, Parkin and colleagues at ViroLogic have shown that over a third of viral isolates with major mutations remain susceptible to saquinavir. [Parkin Antiviral therapy 2000:5 (Supplement 3): 50] The question remains as to whether patients with major protease inhibitor mutations but no more than low level reduced susceptibility to SQV will respond to a boosted SQV regimen. This has potentially important implications for protease experienced patients with this "discordant" pattern.
Zolopa and colleagues at Stanford evaluated the virologic response to Saquinavir -Ritonavir (400/400mg) ARV regimens in patients who had failed prior PI-regimens and who had major PI-resistance mutations. In a cohort of 54 previously reported patients the investigators identified 11 patients who had the combination of protease mutations 82A and 90M. Despite these major mutations, only 2 of the 11 had a SQV IC50 fold change above 10. These two patients both had the 48V mutation in addition to the 82 and 90 mutations. The other 9 patients had SQV FC that ranged from 0.5 to 9. In contrast to the saquinavir phenotype most other PI's showed high level resistance with the exception of amprenavir. All patients received SQV/RTV and recycled nucleoside analogs. No patient received a new class of antiretroviral agent. The over all response to the boosted SQV regimen was disappointing, median viral load change at week 4 was -0.38 log decline and all patients were near baseline or above by week 24.
The investigators also preformed follow up genotypes and phenotypes expecting to see rapid evolution of highly resistant viruses with additional mutations like the 48V. This however, was not the case there was little genotypic evolution (5 patients isolates added a G73S -a poorly defined accessory mutation) and little change in the SQV phenotype. It is not clear why there was so little response to the SQV/RTV and the investigators recommend caution in the interpretation of SQV phenotypes in the setting of protease experienced patients with multiple PI-resistance mutations.
Improved phenotype for non-clade B HIV-1
Non-clade B HIV-1 accounts for the vast majority of HIV -1 infection worldwide. It is important that the technologies used to measure viral load and viral resistance be adapted to function accurately for these subtypes of the virus. Investigators from ViroLogic have adapted their PhenoSense tm assay and demonstrated that the enhanced version performs well with non-clade B subtypes with an improved sensitivity (down to viral loads of 125 copies/cc). They were also able to adapt the assay to measure replication capacities of these subtypes which may play a larger role in the monitoring of antiretroviral therapy in the future as interest in measuring viral "fitness" increases amongst treaters. [I-1748]
Tenofovir DF the first nucleotide analog for HIV-1
As most readers of NATAP reports already know, tenofovir DF was approved by the FDA for the treatment of HIV-1 infection last September. The drug is a welcome addition to the antiretroviral armamentarium for a number of reasons. I will focus on the resistance data that was presented at ICAAC conference. The so-called biological cutoff for TDF was determined to be 3.0 using isolates from treatment naive subjects. This means that 97.5% of subjects with wildtype virus will have a pheotype (Virco) with a fold change of 3.0 or less. The group from Virco demonstrated that in 5,000 isolates from treatment experience patients (but TDF naive) 88% had TDF IC50 fold change of <3.0 while only 4% had fold change >5.0. [1756] Moreover, the investigators from Gilead Sciences have shown that patients with a fold change of up to 4.0 appear to respond to TDF when added onto stable background therapy. Therefore, although not conclusive it appears that the clinical cutoff for tenofovir will be 4.0.
The genotypic patterns associated with resistance to TDF include the 65R mutation, which is the signature mutation for this drug and also contributes to low level resistance to DDI, DDC, ABC and 3TC. Fortunately this mutation is relatively infrequent (<3-5%) in cohorts of nucleoside experienced patients. It also appears not to develop rapidly in patients treated with TDF in whom multiple nucleoside analog- related mutations already exist. It has not yet been determined how rapid or prevalent the 65R mutation will be in patients treated with TDF who have no other nRTI-associated mutations but data from studies treating naive patients is forthcoming.
In addition to the 65R, the 69 insert multi-nucleoside resistance mutation also causes high level resistance to TDF but the other multi-nucleoside mutation (Q151M complex) does not cause resistance to TDF. Finally, it appears that the classical AZT-related mutations or TAM's can cause at least partial resistance to TDF, however, it appears to require the presence of the 210W mutation along with 41L and 215Y. In subjects treated with TDF as a single add on to stable background therapy those subjects with >=3 TAM's inclusive of 210W or 41L had a diminished response to TDF intensification.
To read a detailed NATAP report on PMPA abstracts at ICAAC (100-week safety, antiviral effectiveness) see:
www.natap.org/2001/ICAAC/day4.htm
Amprenavir Pharmacokinetics
Amprenavir (APV) is frequently used in combination with other PI's as part of salvage regimens or regimens for PI-experienced patients. The correct dosage of APV in combinations with LPV/r for example remains uncertain. There were several presentations that evaluated the PK profiles in patients on boosted APV regimens.
Investigators from France demonstrated that the steady state PK profile of amprenavir is significantly increased with the addition of low dose ritonavir. Garraffo and colleagues showed that both plasma levels and intracellular concentrations in PBMC's were improved with two different dosing schedules of RTV boosted APV, APV 600mg with RTV 100mg BID and APV 1200 with RTV 200 as a once daily dose. [A 489]
In another French APV PK study, Meynard and colleagues evaluated the PK profile of APV and LPV/r in patients taking these two PI's together in different BID dosages (LPV/r 400/100 with either APV 600mg or 750mg) compared to patients taking APV (600mg) with RTV (100mg) boosting. The investigators showed that the Cmin for APV was less than 50% in the groups receiving it with LPV/r compared to the group who received APV 600 with RTV100mg boosting. Furthermore, the Cmin for LPV was decreased in the group receiving LPV/r with APV compared to the group on LPV/r without APV. The implications of these observations are not clear but point out the need for careful PK studies which relate virologic response to PK profile are required to determine the optimal dosing of multi-PI regimens.
Editorial note: Amprenavir "908" prodrug once & twice a day with Efavirenz. GSK researchers reported (abstract 1737) at ICAAC on the effect of adding efavirenz to the newly formulated amprenavir prodrug called "908" when used in combination with RTV to boost 908 levels. Clinical studies are exploring "908" as a once and twice daily PI regimen for treatment-naive and for patients with PI-resistance. In a once daily regimen 1400 mg 908 + 200 mg RTV is being studied. In this study, after 14 days of dual PI therapy (908+RTV once daily), EFV 600 mg QD was added for an additional 14 days. We know from previous studies that EFV can reduce APV Cmin by 40%. Similar results were seen in this study. Results from this study show using only 200 mg of RTV with EFV+908 in a QD regimen reduced APV trough by about 35%. But, when 300 mg of RTV was added 908 trough level was the same as without adding EFV. For patients with PI or APV resistance using 300 mg of RTV appears preferable when combining EFV with 908 in a QD regimen. In a BID regimen, 908 700mg + RTV 100mg BID is the currently recommended dose, with or without EFV. It appears that an extra 100mg RTV is only needed in the QD regimen.
|
|
| |
|
|
|
|
|
