![](images/null.gif) |
![](images/null.gif) |
![](images/null.gif) |
|
Report from the 13th International Workshop on HIV Drug Resistance
|
|
|
Written by David Margolis, M.D.
Dallas VA Medical Center & University of Texas-Southwestern
Although it took attendees from the left side of the Atlantic some time to find their way to this meeting in the volcanic isle of Tenerife in the Canary Islands, the content of the meeting made the trip worthwhile. It was interesting to again note that advances in the understanding of the molecular mechanisms of HIV drug resistance, a topic that may seem arcane and irrelevant to clinicians and patients, continues to drive new clinical applications and concepts to be explored in the treatment of people with HIV infection. Meeting abstracts will be available at www.intmedpress.com in several weeks.
TOPICS
(1) New Drugs and Resistance to them: integrase, entry inhibitors, TMC-125, new protease inhibitors
(2) Mechanisms of Drug Resistance and Implications for Therapy
How does HIV drug resistance occur: Evidence for frequent recombination
--Entry inhibitors
--NRTI resistance (K65R, tenofovir): Applications for Therapeutic Strategies
--Protease inhibitor resistance
(3) Clinical Implications of Resistance: K65R; tipranavir resistance; Non-clade B drug resistance
(4) How does HIV drug resistance occur: Evidence for frequent recombination
(5) Resistance rates to all 3 drug classes
(6) Preferred genotyping algorithm for Kaletra; Reyataz; d4T and ddI phenotypic cutoffs; Virtual Phenotype
New Drugs and Resistance to them
As for several years, the path to development of drugs that inhibit novel steps of the HIV life cycle, as well as new additions to the current reverse transcriptase (RT) and protease inhibitor (PI) stable includes an evaluation of drug resistance.
Integrase inhibitors
Resistance to several chemical classes of strand transfer integrase inhibitors (STIs) were discussed at the meeting. The most disappointing feature of this class of inhibitors is that despite periodic excitement generated by pre-clinical findings, we still do not have any of these drugs in people. The small and large companies that are still working on STIs deserve credit for not giving up yet.
The good news is that each of the chemical type of STIs reported seem to generate mutations in different parts of integrase. Although the chemical properties of some of these agents present significant obstacles to clinical drug development, these findings suggest that in the future resistance to one STI might not mean resistance to all STIs.
Daria Hazuda (abstr. 1) reported that resistance mutations to diketo acids, a class of STIs that are no longer leading candidates for clinical development, are mutually exclusive to mutations to napthyridine carboxamides, a new class of STIs now under development at Merck. Resistance mutations selected against strylquinolines (abstr. 2), another class of STIs under development in France, also mapped to a site that differs from DKA resistance. Viruses resistant to both napthyridine carboxamides and strylquinolines were reported to be highly replication impaired. No data on resistance to a 4rth class of STIs, carbazole derivatives, was presented (abstr. 3). Carbazoles appear potent but toxic, and further work will have to be done to improve their properties.
Entry Inhibitors
A small molecule in development at Bristol, BMS 488043, sticks into the pocket within HIV envelope gp120 that is directly contacted by the CD4 receptor, and exerts antiviral effect by blocking viral entry. Lin reported (abstr. 5) that resistance mutations to the compound were formed at multiple sites both within the binding pocket and at distant sites surrounding the pocket. As noted by Dr. Lin, this is reminiscent of the changes that HIV envelope makes to evade the body’s own neutralizing antibodies. As HIV is quite adept at shape-shifting its envelope, this suggested to this author that this drug faces significant challenges. But the proof will be in the testing.
UK-427,857 is a CCR5 chemokine receptor blocker soon advanced stage clinical trials in HIV infection. CCR5 is the major 2nd receptor that HIV uses in addition to the CD4 receptor to enter cells. As patients have not been treated with UK-427,857 for more than a few days, resistant virus has not yet been observed in the clinic. To model what is to be expected, Westby (abstr. 6) presented the results of laboratory studies in which viral isolates from patients where grown in the lab under conditions that allowed resistance to emerge. Resistance was difficult to generate in the lab, but when it did occur two types of resistance were found. One type of virus had evolved the ability to use the alternate 2nd receptor CXCR4. Other resistant strains had evolved to access CCR5 despite the presence of the inhibitor, either by sticking to a slightly different part of the receptor or sticking more tightly. All of these resistant viruses had mutations that mapped to area of the gp120 envelope molecule that binds CCR5, so that current clinical resistance testing techniques should be able to be adapted to detect R5 inhibitor resistance.
Miller and colleagues from Merck (abstr. 9) described the isolation and development of a synthetic antibody-like molecule that binds the HIV envelope region in a similar fashion as the entry inhibitor enfuvirtide (T20, Fuzeon). While HIV will likely be able to mutate, just as resistance to enfuvirtide can develop, artificial antibodies to this important therapeutic target had never been developed before.
The usual suspects
Information was presented about resistance to TMC-125, a new non-nucleoside RT inhibitor also entering advanced stages of clinical trials in HIV infection (abstr. 16). Previously reported pilot studies showed that patients with K103N and other NNRTI mutations displayed a clinical response to TMC-125, but due to the initial nature of these studies drug was not given for more than 2 weeks. As with UK-427,857, laboratory studies of viruses with various levels of NNRTI resistance grown in the presence of various concentrations of TMC-125 showed that the evolution of mutations V179F, Y181C, L214F, and M230L conferred high level resistance to TMC-125. TMC-125 might prove to have better potency or tolerability than current NNRTIs, but TMC-125 has also generated excitement in the hope that it might have durable activity against NNRTI resistant HIV. Although one can hope for the best in clinical trials, the presence of the common NNRTI mutation Y181C in a lab-grown resistant virus is of concern.
Patterns of resistance to new protease inhibitors under development by Glaxo (abstr. 11 and 12) and Pfizer (abstr. 13) were presented. Further study will be needed to show if these drugs will offer a real clinical advantage over current PIs. A new class of protease inhibitors in early development by Sequoia Pharmaceuticals (abstr. 14) appeared to retain significant potency against 3 mutant clones with broad PI-resistance in laboratory testing.
Mechanisms of Drug Resistance and Implications for Therapy
How does HIV drug resistance occur: Evidence for frequent recombination
In tandem presentations, the NCI’s Frederick Drug Resistance Program group presented strong evidence that robust rates of recombination drive the evolution capabilities of HIV. Hu (abst. 44) presented the results of an elegant laboratory experiment using 2 viral clones tagged with different marker genes encoding proteins detectable in the flow cytometer. His studies showed that a recombination rate of 11% could be achieved during a single replication cycle. Maldarelli (abstr. 45) then followed, reporting studies based on single genome sequencing from recently infected patients. Linkage analysis yielded an estimate of 2.4 recombination events per viral genome infection cycle. An alternative analysis, linkage disequilibrium, yielded a consistent estimate of 0.5 to 8 events. The frequency of recombination suggested that up to 20% of the productively infected cells were infected by more than one virus. This is most likely to happen by cell-to-cell infection, or infection over short distances (e.g. within a lymph node). This is consistent with the hypotheses of others that HIV infection in maintained by “infectious centers” within tissue, and by passage of virus from presenting cells to T cells a the “infectious synapse.” These findings inject new and important information into a long-standing debate: does resistance comes from evolution of a single virus (eg M184V virus adds a TAM, etc) or does recombination of viruses carrying different sets of mutations occur? and how much resistance comes from evolution of a single virus vs how much if any can come from Recombination. In theory recombination of two mutant viruses could evolve more quickly, as for example 2 different double mutants recombine, quickly generating a quaduple mutant.
Entry Inhibitors
As it is hoped that multiple entry inhibitor drugs will be available in the clinic in the next several years, Chris Petropoulous’ talk about the mechanisms of resistance to this class of antivirals was timely and informative (Abstr. 19). Four mechanisms can be easily imagined and were discussed by Petropoulous:
- Competitive resistance: the virus evolves to stick to entry receptors targeted by the drug harder than the drug can
- Non-competitive resistance: the virus evolves to use the receptor despite the fact that the drug is there. This could be accomplished either by using a slightly different surface of the receptor or by using the receptor-drug complex for entry.
- Switching: evolving to use the CXCR4 or CCR5 receptor, or both, if one is blocked.
- Escape: evolving to avoid the use of the CD4 entry pathway altogether and enter via the use of other cellular receptors. This has been described in the laboratory and for other retroviruses, but it has not been described clinically (at least so far) with HIV-1.
These concepts reinforce the idea that we must use entry inhibitors in the context of other antiretroviral therapy and, as has been suggested in the past, that combination entry inhibitor therapy may be highly potent and synergistic. Further, phenotypic and genotypic assays to detect and measure these types of resistance are already available. However, as the biology of cell entry is more complex than that of reverse transcription or protease function, the clinical use of such resistance assays will require a great deal of validation and testing.
NRTI Resistance
Parniak (Abstr. 26) and Meyer (Abstr. 31) both presented the novel concept to trying to use the mechanism of AZT resistance against AZT-resistant virus. Perhaps this is a good idea, as it appears to have occurred to two smart people at around the same time. Before diving into this, a quick review of mechanisms of NRTI resistance is in order. When added to the HIV DNA chain NRTIs block HIV replication by terminating the further copying of the HIV RNA genome into DNA.
The simplest example of resistance is that acquired by mutation of the methionine (M) amino acid at position 184 of RT (for example methionine-184-to-valine, or M184V). This mutation directly inhibits binding of 3TC or FTC and its incorporation into the replicating HIV DNA chain.
Thymidine analog mutations (or TAMS), so named because they are selected for by the thymidine analogs AZT and D4T, help the virus escape NRTI drugs by excising them from the growing HIV DNA chain, and allowing HIV replication to continue. Examples of these are M41L, L210W, T215Y or F, D67N, K70R, and K219E or Q.
The K65R mutation is unique and complex in that it inhibits binding and incorporation of all NRTIs, but also decreases excision of NRTIs. The extent of these two counteracting effects differs for each NRTI and determines whether K65R (alone) will increase or decrease sensitivity of the virus to a particular NRTI.
For AZT, even though there is decreased binding/incorporation by K65R, greatly decreased excision of AZT results in overall increased susceptibility of the K65R virus to AZT. For abacavir, despite decreased excision by K65R, more strongly decreased binding/incorporation results in somewhat reduced susceptibility to K65R virus to ABC. For the other NRTI, particularly tenofovir, decreased binding/incorporation appears to be responsible for the decreased susceptibility. Hence TDF and ABC, as single drugs, select strongly for the K65R mutation.
A phosphate donor is required to provide the energy for TAMs to excise NRTI molecules. Parniak reported that a bisphosphonate (a relative of foscarnet) could inhibit excision and increase the sensitivity of both wild type and TAM-containing HIV to AZT. Meyer reported a similar effect for dideoxynucleoside tetraphosphates (ddN-P-P-P-P-ddN). Although there are likely to be significant hurdles before these molecules are “drugable,” and cell toxicity induced by such compounds is an obvious possibility, this is a new idea that deserves more work.
Applications for Therapeutic Strategies
The antiviral strategy of inhibiting drug excision by TAM-containing RT is actually an old one. The well-known 3TC/FTC mutation M184V decreases the ability of RT to excise nucleotides, and is one of the reasons given for the clinical antiviral synergy of AZT or D4T + 3TC or FTC combinations. At this meeting Lennerstrand (abstr. 22) showed that in vitro M184V inhibited the excision of tenofovir as well, in concert with the increased clinical response to tenofovir of patients whose genotypes show M184V in isolation.
The K65R mutation is known to decrease excision capacity as well. Mark Wainberg, a grandfather of 3TC resistance who proudly announced that he is recently an actual grandfather, showed in work from his lab (abstr. 27) that the L74V mutation, selected in the clinic by abacavir or DDI, reduced by 50% the ability of RT to excise. The effect of L74V was similar to that of K65R, and less that M184V. In the past the non-nucleoside mutations 100I and 181C have also been shown to decrease excision.
In general, most would agree that we should select combination antiretroviral therapy based on criteria that make a given therapeutic combination most likely to be successful for an individual patient. However, planning for fallback strategies in the case of clinical failure plays an important role in the decision process of many clinicians and patients.
Understanding factors that inhibit the excision of RT inhibitor drug molecules by drug-resistant HIV RT leads back to the idea that antiviral drug combinations should select for mutually antagonistic drug resistant mutations. This is not a new concept, but lent considerable scientific rationale by several presentations in Tenerife. Faruki and colleagues (abstr. 80) presented prevalence data on NRTI drug resistance mutations from the large LabCorp database. Sequence data from over 80,000 clinical samples sent to LabCorp since 1999 for clinical resistance testing were analyzed. Over this period of time, the prevalence of TAM mutations decreased, and the prevalence of K65R and Y115F increased. This is likely due to the decreasing use of AZT and d4T, and the increasing use of abacavir and then tenofovir during this period of time.
Given these changes in “viral demographics” it is interesting to note that the detection of viruses encoding both the K65R mutation and multiple TAMs (defined in this study as the TAM triplets of 41L/210W/512Y or F, or 67N/70R/219Q or E) was very rare. In samples sent to LabCorp for clinical testing from 1999-2001, 0-0.03% encoded this combination, while in 2002-2003 only 0.02% to 0.05% encoded both TAMs and K65R. This was despite the increase in detection of K65R and the increased use of drugs that select for this mutation. This observation does not prove that the use of TAM-selecting and K65R-selecting drug together will slow the development of both mutations, but it suggests that the virus prefers to develop resistance along one pathway or the other. Perhaps forcing the virus to take two pathways will be clinically beneficial. This will require study in clinical trials before it is accepted as dogma.
Protease Inhibitor Resistance--
(potential novel way in which protease inhibitor resistance may develop; resistance: PI resistance in the gag region of the protease)
Protease inhibitor drugs fit into a cleft in the jaws of the active site of the HIV protease enzyme, jamming up the works and preventing protease from processing HIV virion protein precursors into a finished virus. However, HIV appears to have other means of evading the effect of protease inhibitor drugs.
Van Maarseeven presented work of the Boucher lab in Utrecht (abstr. 36) in which viruses were grown in the presence of high concentrations of an experimental Roche protease inhibitor. The drug used in selection was reported to have a high genetic barrier, presumably meaning that multiple active site mutations are be required to give drug resistance, and could not serially evolve during culture with high levels of drug. This would force the virus to find another, shorter pathway to PI resistance. After more than a year of selection, viruses with 6- to 8-fold resistance to the Roche drug, and 2- to 5-fold resistance to all PIs were isolated. This virus had mutations in the gag region of the HIV genome, that domain that encodes inner viral proteins. Although the exact mechanism of resistance of these viruses is not yet defined, cleavage of the gag pre-protein of this drug-resistant virus into individual viral proteins appeared to proceed more quickly than that of wild-type, drug-sensitive virus.
This suggested that the virus has developed PI resistance by evolving changes in the gag region that allowed protease to bind more tightly and cleave despite the presence of drug, or by other mechanisms such as producing higher levels of protease enzyme. To fully understand protease resistance, it may therefore be important to study the viral genome outside of the protease enzyme itself. Some drug companies have already begun to study the gag region as part of the pre-clinical evaluation of protease inhibitors. Some meeting attendees speculated that this mechanism of resistance might explain clinical PI resistance in the absence of detected resistance mutations within the protease gene, as has been described for Kaletra. Baxter and colleagues (abstr. 42) described the detection of gag region mutations in 40% of 148 patients failing PI therapy in CPCRA 046, the GART study.
Clinical Implications of Resistance
K65R musings
Valer et al. (abstr. 40) reported on the prevalence of K65R. In a cohort of 1846 patients from Madrid they found that the prevalence of K65R increased from 0-0.6% in 1999-2001, to 3% in 2002, 7.3% in 2003, and 11.5% in the first few months of 2004. The rising incidence of the K65R mutation may reflect the broad and growing use of tenofovir, and of other drugs such as abacavir and didanosine that can select this mutation.
In a cohort from Rome reported by Antinori (abstr. 152), 1392 genotypes were done on 771 patients who were never treated with tenofovir between 1999 and 2003. In contrast to the Spanish cohort, they found the incidence of K65R to be stable (overall prevalence 1.6% of these genotypes performed for clinical management), but K65R more likely to be found in patients with a history of clinical AIDS, or current use of abacavir or efavirenz. The presence of M184V or TAMs decreased the likelihood of K65R being found, while the mutlinucleoside resistance mutation Q151M or the NNRTI resistance mutation L100I increased the likelihood of K65R. Most of these observations would be expected, but the association with NNRTI use and NNRTI resistance should prompt further examination.
On the other hand evidence from the Gilead 903 study presented by Miller (abstr. 157), a carefully performed comparison of tenofovir to D4T as initial therapy with 3TC and efavirenz, suggests that the emergence of K65R in initial therapy should be infrequent. Of the 600 patients in the study, only 15.7% and 16.3% suffered virologic failure after 2 years of therapy in the tenofovir and D4T arms, respectively. More than a third of these patients did not have new (or any detectable) NRTI resistance mutations. However, 8 failing patients of the 299 treated with tenofovir displayed K65R, whereas only 2 of the 301 patients treated with D4T displayed K65R. Of these eight, 5 also had efavirenz and M184V/I (3TC) resistance, whereas only 1 of the D4T-treated patients displayed this multiple resistance pattern. This might be viewed as a glass that is much more than half-full, as 98.3% of the tenofovir-treated patients did not display multiple resistance mutations.
Unexpectedly bad results:
As reported at the Paris IAS meeting last summer, and the ICAAC meeting last fall, several triple nucleoside regimens can lead to rapid non-response. A resistance analysis of two of these studies, Tonus and ESS30009 was presented. Descamps (abstr. 155) found that non-response to abacavir, tenofovir, and 3TC once a day was associated with M184V or M184I in most of 21 patients by week 4. A few patients had K65R alone or with 184V by week 4. By week 12, 18 of 21 patients had M184V and 13 of these also had K65R. However, clonal analysis of selected patients at week 4 detected K65R separately, but not on the same viral genome as the M184V until week 12.
Similarly, Ross (abstr. 159) found M184I and M184V most often in individual clones from the viral swarm at week 2 of therapy in patients who failed to respond to abacavir, tenofovir, and 3TC in ESS30009. K65R alone was found in a few clones. By week 4 and 12, the numbers of resistant clones encoding M184I/V or K65R increased, and a few clones encoding both M184I/V and K65R appeared. Both studies seemed to show that resistance to 3TC and abacavir/tenofovir evolved on separate viruses. It is not yet clear whether further resistance then evolved from a single-mutant viral genome, or whether 2 resistance genomes recombined to generate the M184V + K65R virus.
What is tipranavir resistance
In a vast improvement over studies of several years ago, Boerhinger’s 1182.51 study (abstr. 147) provided a clear sense of resistance and clinical response to tipranavir. Patients in this study had resistance to NRTIs, NNRTIs, and PIs, viral load of >1000 copies/ml, mutations at 3 or more of PR codons 33, 82, 84, and 90. Patients received optimized background (other than PIs) therapy and ritonavir-boosted tipranavir, lopinavir (LPV), amprenavir (APV), or saquinavir (SQV) for 2 weeks. After 2 weeks, tipranavir was added the therapy of those on LPV, APV, or SQV.
At two weeks, there was no significant viral load decline in the “old” PI arms, and a 1.15 log decline in HIV RNA in the tipranavir arm. 4 weeks after adding tipranavir, the “old” PI arms had displayed a 1.2 log decline. Response in all arms (including the tipranavir arm) was transient and lost by week 8. Preliminarily, double boosting of another PI with tipranavir is no longer recommended in most patients due to cumulative intolerability (potential for PI drug-drug interactions), and little additional perceived benefit compared with tipranavir alone.
A 1.5 log drop was seen in patients treated with tipranavir if less than 10 PR mutations were seen at baseline, and a 1.0 log drop if less than 15 PR mutations were seen. Response was worst to tipranavir for subjects with mutations at PR codons 33, 82, and 84. It was worst to SQV if 33, 82, and 90 were mutated, and worst to LPV or APV if 82, 84, and 90 were mutated. Obviously for patients with very advanced resistance, the addition of tipranavir alone will provide only temporary benefit, but this study well defines the drug’s potential benefits.
Non-clade B drug resistance
With increasing world travel and increasing HIV therapy outside of Europe and the US, drug resistance in the non-clade B HIV strains that predominate elsewhere received needed attention in Tenerife. Although the hopeful might feel that there is yet no evidence of significant differences in drug resistance in non-B virus, it seems likely that caution is warranted pending further study.
Eshleman (abstr. 50) showed that NNRTI resistance mutations in the plasma of women given single-dose nevirapine prenatal prophylaxis appeared with similar kinetics, but decayed to undetectable levels faster in women infected with clade A than clade D HIV. A large European database that included more than 600 non-B viruses reported no major differences between subtype B and subtypes A, C, D, F, G, J, CRF01-AE, and CRF02-AG in the number of genetic changes required to evolve to known major RT or PR drug resistance mutations. More than 200 type C sequences were studies, but only 20-90 sequences from each of the other subtypes.
However Turner (abstr. 88) did find that in subtype C the V106M resistance mutation could be facilitated by a single base change, rather than the 2 changes needed in other subtypes. Eshleman (abstr. 102) also found that a domain of the gag gene was frequently duplicated in type C virus, as compared to other subtypes, opening up the possibility of differences in drug susceptibility, or ability to recombine.
Further, Chaix (abstr. 160) reported an apparent difference in the likelihood of acquisition of nevirapine resistance after single-dose nevirapine prenatal prophylaxis between viral subtypes. In was unclear, however, if this was a direct effect, or a marker for differences in persistence of nevirapine concentrations in the small group of 74 women studied.
Resistance to all 3 drug classes
A few years ago some HIV treaters feared that the emergence of resistance to antiretrovirals meant that we would soon be cast back to the dark days of the early epidemic, without available effective therapies. Certainly this is still a feared outcome, and one that the Resistance Workshop was founded in an effort to avoid. But the good news is that, while new and better drugs and strategies to treat drug-resistant HIV are needed, we have all done a reasonably good job with the tools at hand so far.
In a large UK data base, Pillay (abstr. 77) that the prevalence of triple class resistance was considerable in patients in whom resistance tests were being done (ca. 15%), and that the overall number of patients with triple class resistance was increasing somewhat. However the prevalence of patients with triple class resistance in the HIV-infected population was relatively stable. This finding was echoed by Brun-Vezinet and colleagues in France (abstr. 78). In 17 centers participating in ARNS genotyping, roughly 2.5% of the genotypes performed showed triple class resistance between 2001 and 2002. The Virco database (abstr. 79), largely from the US and Europe, found that genotypes submitted with no resistance had increased from 2001 to 2003 (up to 40%), NNRTI resistance has increased from 50 to 66%, but NRTI and PI resistance had decreased (90 to 78%, and 68 to 39%, respectively). The prevalence of triple class resistance was also found to be stable at somewhat less than 15%. This was similar to the 18% incidence of 3-class resistance reported in a CDC study (abstr. 101)
Grant from UCSF, working in was is considered a hotbed of drug resistant HIV, reported that the transmission of drug-resistant HIV detected in primary HIV infection was decreasing in S.F. (abstr. 93). The incidence of any drug resistance mutations was 18-27% in 2001-2002, and of 2-class or 3-class resistance ca. 8% and 1.5% in that time. By 2003, any resistance was only found in 9.8% of primary infection, and of 2-class or 3-class resistance in 2 and 0%. Drug-resistant HIV seemed to be mostly transmitted in clusters of recently infected individuals, highlighting the difficulty that these unfit viruses may face in being transmitted.
Education to diminish the transmission of drug-resistant (or other) HIV among injecting drug users might pose challenges, as suggested by surveys performed by Kozal and colleagues in New Haven (abstr. 89). These investigators found that the vast majority of active IDUs with drug-resistant HIV or wild-type HIV understood that sharing needles with others could transmit resistant virus, yet still engaged in risky behavior.
Harrigan (abstr, 84) reported the experience of the British Columbia HIV Treatment program, finding that triple class resistance was not a significant direct cause of death in their patient population between 1997 and 2001. Of 554 non-accidental deaths surveyed, fewer than 5% (ca. 26 patients) of the patients who died had triple class resistance. Other factors such as illicit drug use or recent clinical AIDS were associated with death.
However, this was not entirely the case in Italy. Zaccarelli (abstr. 139) found that the 36 month Kaplan-Meier risk of death was 9% in subjects without multi-resistance to a drug class, 11% with multi-resistance to one drug class, 15% with multi-resistance to two drug classes, and 30% with multi-resistance to all 3 drug classes.
Brief new points on resistance testing: preferred genotyping algorithm for Kaletra; Reyataz; d4T and ddI phenotypic cutoffs; Virtual Phenotype
Norton (abstr. 117) reported that a genotyping algorithm specifically designed for Kaletra outperformed several standard algorithms designed to predict PI sensitivity in general.
Shulman (abstr. 118) found that d4T resistance could be predicted by Virco virtual phenotype cutoff value of 1.0. She found that a 1-fold increase in d4T virtual phenotype score had 100% sensitivity: all the subjects who did not respond to d4T in a clinical trial (ACTG 302) had a VPT >1.0. The test had less specificity (74%), in that 7 of 22 non-responders to d4T were falsely predicted to respond to d4T by a VPT of less than 1.0. Similarly, using isolates from an old ACTG study of ddI monotherapy (ACTG 175, abstr. 119), Shulman’s group suggested that the fold-change cutoff for ddI of 1.7 may be too high, whereas 13 or 13 patients with an FC of < 1.0 responded to ddI, 6 of 11 with FCs of > 1.0 failed to respond to ddI.
Bacheler (abstr. 138) presented improved estimations of phenotypic clinical cutoffs for VirtualPhenotype using a meta-analyses of 11 clinical trials and 2 cohorts. Models of virological response as a function of baseline fold change in IC50 were constructed using linear and logistic regression, and phenotypic susceptibility scores of the background regimens, baseline viral load, and fold change were included in each model. Fold changes for VPT that predicted a 20% or 80% reduction in the maximum virologic response (RR) at 8 weeks of therapy were: A report of the response of 25 PI-experienced patients switched to 300 mg qd of atazanavir and 100 mg a day of ritonavir in the Swiss cohort was presented by Yerly (abstr. 149). Subjects were switched for failure (18), adherence (4), or intolerance (3). Despite mutations associated with ATZ resistance in 12 of these patients, virologic response was seen when background therapy was optimized and ATZ/r given in 6 of these 12. There was little or no ATZ/r response in subjects with 5 or more PR mutations at 10, 20, 33, 46, 47, 54, 82, 84, 90, mutations seen to give cross-resistance to many PIs.
|
|
|
![](images/null.gif) |
![](images/null.gif) |
|
|