Joel E. Gallant, M.D., M.P.H.
The Hopkins HIV Report - November 1998

The Target Cell Availability Hypothesis Starting Therapy: When to Start and How Hard to Hit Efavirenz Adefovir

Twice-Daily Protease Inhibitor Resistance and Salvage Therapy Conclusion

Because much of the latest data on antiretroviral therapy was presented in Geneva at the 12^th World AIDS Conference just a few months ago, the news from San Diego seemed sparse. Nevertheless, there was a number of important highlights, both in updates of previously reported data and in reports of new studies.

The Target Cell Availability Hypothesis

Perhaps the most provocative subject discussed at ICAAC came not from reports presented at the conference, but from data presented at the recent resistance meeting in Lago Maggiore, which was frequently referred to by speakers in San Diego. In ACTG 343, one of three trials demonstrating the failure of the induction-maintenance strategy, patients who achieved undetectable viral loads after six months of an "induction" regimen of AZT/3TC/indinavir (IDV) were randomized to continue the three-drug regimen or to be switched to a "maintenance" regimen of either AZT/3TC or IDV alone. Interestingly, failure of the three-drug regimen was associated with the appearance of genotypic and phenotypic 3TC resistance but not with resistance to either AZT or indinavir. Even more surprising, no indinavir resistance was observed in those who experienced early failure of indinavir monotherapy, regardless of viral load. Furthermore, in those randomized to receive maintenance therapy with indinavir, there was a 40% increase in the risk of failure for every 100 cell increase in the CD4 count. One admittedly controversial explanation for these puzzling observations has been referred to as the "target cell availability hypothesis." Proponents of this hypothesis argue that HIV replication requires activated CD4 cells, target cells that increase in number with highly active antiretroviral therapy (HAART). Thus, HAART creates a "permissive environment" for viral replication. Because wild-type virus is more fit than resistant virus, it may out-compete resistant virus, especially when the factory for viral replication (activated CD4 cells) is large. Thus, initial rebound may often occur with wild type rather than resistant virus.

One implication of this hypothesis is that failure without evidence of resistance does not mean that the patient has been non-adherent. Another is that a strategy of early intensification may make sense. If a patient with low-level, detectable virus on a HAART regimen has no drug resistance or resistance to a single nucleoside analog, then changing to an entirely new combination may be unnecessary. Instead, the addition of additional agents might push the viral load below the limit of detection and prevent the development of resistance without forcing a switch from a partially effective combination. There is growing enthusiasm for intensification, based in part on clinical trial data and in part on our ability to detect virus at very low levels using ultrasensitive assays. The most appropriate candidates for intensification might be those who fail to achieve undetectable viral loads by the ultrasensitive assay after 20-24 weeks or those who experience confirmed rebound after having reached undetectable levels. Most intensification involves the addition of nucleoside analogs or hydroxyurea to a pre-existing HAART regimen. Intensification of a non-HAART regimen with protease inhibitors or NNRTIs may be a much riskier strategy because of the likelihood of resistance to NRTIs in an incompletely suppressive regimen.

Resistance testing may play an important role in deciding whether a drug regimen should be intensified or switched completely. However, currently available assays cannot be performed or may be less accurate when the viral load is very low, which is exactly when such information would be expected to be most useful.

Back to top

The target cell availability hypothesis, discussed above, would imply that early therapy, initiated when target cells are plentiful, would increase the likelihood of drug failure. Many studies looking at adherence and antiretroviral therapy in the "real world" also support the deferral of therapy for chronically infected individuals. On the other hand, in the opening session of the conference, Bruce Walker once again discussed evidence that therapy initiated during or shortly after the acute retroviral syndrome may preserve the CTL and CD4 responses to HIV itself, thereby turning patients into "long-term non-progressors." Another study supporting therapy during acute infection was presented by Perrin [S-102], who reported results from a community-based trial in which 56 patients were treated with HAART. Eight of the ten patients treated within six months of seroconversion achieved a profound and durable response, with viral loads sustained at <5 c/ml. In contrast, only about half of those with chronic infection and CD4 >500 cells/mm3 and virtually none of those with CD4 counts <500 achieved that goal, and in those who did the response was not sustained. Those treated early may have accumulated less proviral DNA in resting CD4 cells, the latent reservoirs of chronic HIV infection, and thus were able to achieve better suppression of viral replication.

Clearly, however, early treatment does not lead to eradication. In an Australian study 13 patients were treated during seroconversion [Zaunders J, et al., I-185]. Although all responded with suppression of viral load to <50 c/ml and normalization of CD4 counts, proviral DNA persisted at levels typically seen in chronically infected individuals. Even with early therapy, the increase in naive CD4 cells was small, and the number of activated CD8 cells remained higher than normal, suggesting ongoing immune activation as a result of low-level viral replication.

Unfortunately, few individuals are diagnosed during the acute retroviral syndrome, and it is not known how quickly therapy must be initiated before it is too late to preserve or restore anti-HIV immunity. Thus, for most patients the decision of whether to initiate therapy is one determined by weighing the risk of progression against the risks of therapy, which include toxicity, resistance, and premature elimination of therapeutic options.

There¹s HAART and then there¹s HAART. In a study of a four-drug regimen of AZT, 3TC, indinavir, and nevirapine, the decline in viral load was more rapid than has been observed in studies of three-drug HAART regimens [Polis M, et al., Sun LB-3]. Whether this more rapid decay translates into lower viral nadirs or greater durability is not known. If it does, however, it might revive enthusiasm for the now discredited induction-maintenance strategy, discussed above. However, instead of inducing patients with standard HAART combinations and then "deintensifying" with suboptimal regimens, there might be a stronger rationale for starting with highly aggressive, multi-drug cocktails, and "backing off" to a three-drug HAART combination.

Back to top

One of the most important clinical trials presented in Geneva was DMP-006, a study comparing efavirenz (EFV) plus AZT/3TC, indinavir plus AZT/3TC, and EFV plus IDV. At that time only 24 week data were available; since then, efavirenz has been approved by the FDA. At ICAAC 36-week data were presented which continue to demonstrate the same impressive results [Morales-Ramirez J, et al., I-103]. The AZT/3TC/EFV arm remains as effective as AZT/3TC/IDV in an as-treated analysis. In two forms of intent-to-treat analysis, patients randomized to the AZT/3TC/EFV arm had a better virologic response than those taking AZT/3TC/IDV. In the last observation carried forward (LOCF) analysis, 85% of those in the EFV arm achieved a viral load <400 c/ml compared to 65% of those in the IDV arm (66% vs. 50% RNA <50 c/ml). By the more conservative non-completer = failure (NC=F) analysis, the results were 72% vs. 53% for a viral load <400 c/ml and 64% vs. 44% for viral load <50 c/ml. The differences seen in the two intention-to-treat analyses may be driven largely by the unusually high rate of discontinuation in the AZT/3TC/IDV arm (41%). Nevertheless, the results strongly suggest that the combination of efavirenz plus two NRTIs is a true HAART regimen that may provide an effective alternative for initial therapy and allow the deferral of protease inhibitor use. Longer-term data are needed, but the prospects for durability are promising given the high proportion of patients whose viral loads were undetectable by the ultrasensitive assay.

The combination of efavirenz with protease inhibitors was studied not only in DMP-006 but also in DMP-003 (EFV/IDV) and DMP-024 (EFV/nelfinavir). In the 003 trial, 75% of patients treated with EFV/IDV achieved a viral load <500 c/ml, and 74% were <50 c/ml at week 84 by intent-to-treat, LOCF analysis [Havlir D, et al., I-104]. In study 024, 53% had RNA <50 c/ml at week 24 by intent-to-treat, NC=F analysis. However, treatment-experienced patients had a poorer response than treatment-naive patients (41% vs. 64% <50 c/ml), despite the fact that none of the experienced patients had taken PIs or NNRTIs. The reason for this difference is not understood. While these PI/NNRTI combinations demonstrate excellent antiviral efficacy, most physicians are reluctant to prescribe them for naive patients. Patients who become resistant to such regimens may be left with few options for future therapy given the extensive cross-resistance within both classes of drugs. Such combinations may be useful in salvage regimens, however, especially for patients with extensive exposure to NRTIs.

The most important side effects of efavirenz are neurologic, and include dizziness, vivid dreams or nightmares, and a sense of "disconnectedness." These side effects usually resolve within one to two weeks, and can be partially alleviated by taking the drug once daily before bedtime.

Back to top

Interim 20-week data were presented from a trial comparing a variety of adefovir (ADV) and IDV-containing regimens compared with AZT/3TC/IDV [Myers RA, et al., I-108]. Patients were randomized to receive either the control regimen using AZT/3TC/IDV or ADV/IDV plus either AZT, 3TC, d4T, or AZT/3TC. All of the three-drug arms performed well. The study was not powered for equivalence, but there were no obvious differences between groups in virologic or immunologic response. By intent-to-treat analysis, results were less impressive for the four-drug regimen because of poor tolerability and a high rate of drop-out. The three-drug regimens were well-tolerated; however, it should be noted that the proximal renal tubular dysfunction associated with adefovir typically occurs after 20 weeks.

Forty-eight week data were also presented from a study in which patients who had detectable virus (>2500 c/ml) on a variety of standard regimens added either adefovir or placebo to their current regimen [Kahn J, et al., I-108]. After 24 weeks, all patients received adefovir. Approximately 40% made further changes to their regimens by 48 weeks. At 24 weeks those taking adefovir had a mean reduction in viral load of 0.4 logs, which was sustained throughout the 48-week study period. The degree of benefit appeared to be associated with reverse transcriptase genotypic resistance patterns. Those with 3TC resistance (M184V) had the best response, with an average viral load reduction of 0.94 log, while those with high-level AZT resistance without M184V had no virologic response. Those with high-level AZT resistance who had the M184V mutation had a viral load response of -0.51 log [Cherrington JM, et al., I-84]. These data would suggest that adefovir should usually be co-administered with 3TC to improve efficacy.

In this study, 40% of patients, all of whom were taking 120 mg daily, developed proximal renal tubular dysfunction by week 48. The onset occurred after 24 weeks in all patients. This syndrome is characterized by azotemia, hypophosphatemia, proteinuria, glycosuria, and reduction in serum bicarbonate. Gradual resolution was observed following discontinuation of adefovir. Adefovir is now being used at a dose of 60 mg, which appears to result in a lower rate of nephrotoxicity. More data are needed on the efficacy and toxicity of this lower dose of the drug.

Back to top

In an analysis of combined data from the INCAS and AVANTI trials, adherence was an even more important predictor of long-term virologic success than viral load nadir [Sun LB-10]. Thus, the push for simpler twice-daily regimens is strong, and once-daily regimens may be just around the corner. In a poster presentation 48-week data from the bid vs. tid nelfinavir trial were presented [Johnson, et al., I-216]. The data continue to support the use of nelfinavir at a dose of 1250 mg bid. Approximately 80% of patients in both arms had viral loads <400 c/ml, and 80% were below 50 copies (56% by intent-to-treat analysis).

Although twice-daily indinavir trials were recently halted because of decreased efficacy with bid dosing, co-administration of indinavir with ritonavir may be an effective way to turn indinavir into a bid drug. Both drugs are given at a dose of 400 mg bid and may be taken with food. In a small, non-comparative trial of ritonavir/indinavir plus NRTIs in naive patients, all 12 who reached 16 weeks had viral loads <400 c/ml, and the majority had viral loads <80 c/ml [Rockstroh, et al., I-213].

Charles Farthing presented interim results of a trial evaluating bid soft-gel saquinavir (Fortovase) as well as bid saquinavir/nelfinavir (SQV/NFV) [Farthing C, et al., I-105]. Data were presented from 240 patients randomized to receive SQV 1600 mg bid plus two NRTIs, SQV 1200 mg tid plus two NRTIs, or SQV/NFV (1600/1250 mg bid) plus one NRTI. Although these are preliminary data, there were no obvious differences in virologic response at 24 weeks. By intent-to-treat analysis, approximately 60% of patients had viral loads <400 c/ml, and 40% had <50 c/ml. The pill burden in these regimens is relatively high, however, especially in the dual-PI arm, which had the highest drop-out rate.

Back to top

In his lecture on salvage therapy, John Mellors reminded us that not all failure is virologic, and not all virologic failure is failure. Patients who are experiencing virologic failure (detectable or rebounding viral loads) may be enjoying excellent clinical and immunologic responses, at least in the short term. However, he emphasized that for those who have further treatment options, salvage therapy should be initiated early despite a lack of clinical progression or decline in CD4 count. This is especially true in the case of protease inhibitors, since continued use may lead to the accumulation of additional resistance mutations and decrease the likelihood of successful salvage therapy. In contrast, for those with extensive drug experience, it is often better to continue a failing regimen and wait for new drugs than to make repeated attempts at salvage using suboptimal regimens.

The choice of salvage agents is based on the resistance profile of the patient¹s virus, which is currently determined by the treatment history. Genotypic or phenotypic resistance testing can also be useful, provided the clinician is aware of the limitations of these tests, which predict failure more accurately than success and are better able to detect resistance to drugs the patient is currently taking than to drugs taken in the past. A study presented by Richard Harrigan emphasized this point [I-78]. In patients taking ritonavir/saquinavir after failure of other PIs, resistance to ritonavir and especially to saquinavir at the time of switch was highly correlated with failure of the salvage regimen. None of the patients who were resistant to both drugs achieved an undetectable viral load. However, successful salvage therapy was achieved in only half of the patients whose genotypic and phenotypic analyses indicated sensitivity to these agents.

Although cross-resistance among NRTIs may be predicted by the clinical history or resistance testing, there are situations in which the prediction of resistance is more complex. For example, decreased susceptibility to abacavir is seen in some patients with resistance to AZT and 3TC. When resistance to a third NRTI, such as d4T, has occurred, abacavir resistance is almost certain. However, the genotypic correlates of d4T resistance are not well understood, making it difficult to use genotypic analysis to predict abacavir susceptibility in some cases. Adefovir has diminished activity against virus with high-level resistance to AZT and has improved activity in the face of 3TC resistance. Among the NNRTIs, complete cross-resistance is the rule; however, genotype testing may identify a subset of patients with nevirapine resistance who do not have the K103N mutation and who may respond to efavirenz. Cross-resistance is widespread but less predictable among the protease inhibitors, and although nelfinavir and amprenavir have unique primary mutations, continued exposure would be expected to lead to additional accessory mutations for which there is greater overlap among the other PIs. Mellors concluded that for protease inhibitors, early switching and the judicious use of resistance testing may help to achieve more effective salvage regimens.

The controversial hypothesis that AZT interferes with subsequent phosphorylation of d4T has generated a great deal of heat at previous conferences. Boucher presented data supporting another potential explanation for the diminished effectiveness of other NRTIs after exposure to AZT [Keulen W, et al., I-111]. While AZT resistance mutations may lead to a decrease in viral fitness, subsequent compensatory mutations may actually increase viral fitness and lead to a diminished response to other nucleoside analogs. This may explain why AZT mutations, unlike the 3TC mutation which decreases viral fitness, often persist long after AZT has been discontinued [Bocket L, I-116].

Disturbing data on multi-drug resistance were presented in a study of 892 patients who had developed resistance to NRTI therapy [Larder BA, et al., Sun LB-4]. In this study, the usual NRTI resistance mutations were seen. Moreover, approximately 10% had the insertion mutation at codon 69, which is associated with resistance to all NRTIs. The other multi-drug resistance mutation, found at codon 151, was observed in 5%. While still uncommon, these mutations may account for some of the drug failure observed in patients whose treatment history would not suggest cross-resistance and may be another argument for the use of resistance testing assays that are capable of detecting insertion mutations.

At ICAAC we learned of yet another study demonstrating dismal results of ritonavir/saquinavir salvage therapy following indinavir failure [Lallemand, et al., I-194]. Only three of 43 patients achieved undetectable viral loads (<500 c/ml), and 37 of the 43 failed to achieve more than a 1.0 log decrease. However, these patients had extensive experience with NRTIs and were not started on salvage therapy unless their viral load had returned to baseline or they had shown no significant response to indinavir. The majority of those tested had multiple protease inhibitor mutations. Not surprisingly, the three who achieved sustained viral suppression were the only three who had no resistance mutations at the time of switch. Other studies have suggested that RTV/SQV salvage therapy is more likely to be effective if initiated soon after indinavir failure.

One of the controversies in current discussions of salvage therapy is what is sometimes termed "mega-HAART," a combination of six, seven, eight, or even nine-drug regimens, which necessarily include a number of recycled agents. The rationale for this approach is that while patients with extensive drug resistance will have virus resistant to numerous one, two, and three-drug combinations, it is unlikely that they will have virus resistant to all of the drugs in such a complex regimen. Previous studies examining this strategy, such as those by Staszewski and Workman, suggested that this approach may lead to significant reduction in viral load, at least in the short term. However, toxicity is high and many patients are forced to discontinue all or part of these complex regimens due to intolerability. At ICAAC, Julio Montaner presented data on 83 patients who were treated with salvage regimens containing up to nine drugs [I-201]. Fifty-eight percent were NNRTI-naive, but all had extensive exposure to NRTIs and PIs. By intent-to-treat analysis, only a quarter of the patients sustained viral suppression (RNA <400 c/ml) to their third follow-up visit. One third had to modify their regimens due to adverse events, and nine stopped therapy altogether. There may be a role for such an approach in highly motivated and adherent individuals. Nevertheless, the cost, toxicity, and relatively poor efficacy of mega-HAART make it unlikely that this approach will be the answer to the problem of multi-drug resistance for the majority of patients.

Studies presented in Geneva and San Diego have shown that drug-resistant strains of HIV can be transmitted to HIV-negative individuals. However, there is also a heated debate in the lay press regarding the issue of "reinfection" with HIV--whether HIV-infected individuals can be reinfected with different strains, including drug resistant strains. An animal study presented at ICAAC brought us closer to answering that question [Fultz PN, et al., I-33]. Infected chimpanzees were exposed to a different strain of HIV that could be distinguished from the original strain by PCR. Viral load was not affected, and PCR testing demonstrated that the original strain remained dominant. However, using specific PCR testing, it was clear that the new strain had established itself. While it is not clear whether this second strain could become predominant under selective pressure, this study certainly should give pause to the proponents of so-called "bare-back sex."

Back to top

The pace of scientific discovery and drug development in the field of HIV is so rapid that it's now necessary to have several conferences per year. However, following so closely on the tails of both the 12^th World AIDS Conference and the annual Retrovirus conference, this year's ICAAC held little promise of earth-shaking presentations in the field of HIV. While not earth shaking, the presentations were valuable for their reports of incremental changes, many of which were quite significant. Incremental progress is, after all, still progress. And it's not over yet for this year -- both IDSA (Denver) and the European HIV conference (Glasgow) take place in November, just three months before the retrovirus conference in Chicago. Stay tuned.

981101
JH981101

©1998. The Johns Hopkins University AIDS Service, Division of Infectious Diseases. Permission to use and reproduce portions of this newsletter is hereby granted provided that author and publication are fully credited and both copyright and permission notice appear with reprinted material. Inquiries may be directed to Sharon McAvinue, Managing Editor. Website: Johns Hopkins AIDS Service.

The original of this article can be found at http://hopkins-aids.edu/publications/report/nov98_1.html