Salvage Therapy

While many HIV-infected persons treated with potent combination antiretroviral drug regimens (so-called highly active antiretroviral therapy or HAART) will achieve prolonged suppression of HIV replication to undetectable levels, many others will not. Failure to obtain durable benefit from HAART is often the result of the development in treated patients of HIV variants that are resistant to the specific antiretroviral drugs used. However, there is substantial cross-resistance observed between different antiretroviral drugs within the same drug class (e.g., if a HIV variant is resistant to one protease inhibitor, it may also be resistant to other protease inhibitors). Deciding how to manage antiretroviral therapy in a patient that is no longer responding to their initial HAART regimen is a major challenge in current HIV therapy. With each successful failure, the chances of obtaining durable suppression of HIV replication in treated persons diminish substantially.

The selection of second (or subsequent) line treatment strategies is commonly referred to as "salvage therapy." Salvage therapy is an extremely complex subject, where concepts and strategies are evolving rapidly. Scenarios of salvage therapy to be considered include first treatment failure, second treatment failure, and multiple treatment failure following prior multi-drug treatment experiences. The goal of all anti-HIV treatment strategies is to suppress the viral load to undetectable levels, as this is the only way to circumvent the inherent tendency of HIV to develop drug resistance. However, with successive failures of treatment regimens in individual patients, suppression of viral load to undetectable levels is often not achievable and the goal of antiviral therapy shifts to target suppression of virus replication as much as possible for as long as possible. In this way, meaningful delay of progression to life-threatening immunodeficiency may be achieved, even if virus replication continues at a detectable level. However, incomplete suppression of HIV replication makes the ultimate development of high-level drug resistance inevitable, which sets the stage for eventual progression to AIDS and death.

"The approach to salvage therapy is not monolithic. There is a spectrum of salvage therapy, with different treatment options and different treatment goals," explained Patrick Yeni, MD, of Hospital Bichat, Paris, France.

To make appropriate decisions, the clinician must take into account the circumstances of the particular scenario to select appropriate drugs and sequence them correctly. Predictors of antiretroviral treatment (ART) failure include advanced HIV disease at baseline (characterized by increased plasma viral load and decreased CD4+ cell counts), prior antiviral therapy experience, viral resistance, and low drug concentrations in plasma; the latter can be due to poor adherence, interpatient variability in drug metabolism, or drug/drug interactions.

In the setting of initial treatment failure, antiviral drug resistance may not always be the problem. This was illustrated by the recent TRILEGE study, which showed that the major cause of first failure was poor adherence to therapy-not drug resistance. If drug resistance is not present, the clinician should identify and act on the factors that led to failure, whether it is poor adherence, interpatient variability, or drug/drug interactions. The clinician should also consider intensification of the regimen by adding new agents to an established treatment regimen. However, simply adding one or two new drugs to a regimen that is failing due to the presence of drug-resistant HIV variants is unlikely to durably suppress HIV replication and may lead to the rapid development of resistance to the newly-added antiviral drugs as well.

"At the time of first failure, it is extremely important to identify the reason for low drug concentration. Different actions are required for each reason," he emphasized. With each successive treatment failure, drug resistance is increasingly likely to be responsible. In the setting of second failure, Dr. Yeni advised that, if possible, all drugs in the regimen should be changed, and a new class of drug should be introduced. Treatment history should be considered in selecting an alternative regimen that is not cross resistant to the failing therapy. Three or more active drugs should be selected, and sequencing these drugs correctly is very important, emphasized Dr. Yeni. For example, data from studies using two nucleoside reverse transcriptase inhibitors (NRTIs) suggest that d4T followed by zidovudine (AZT) results in cross-resistance, but the development of cross-resistance is less likely to occur when the sequence of these two drugs is reversed, with AZT followed by d4T. Protease inhibitors (PIs) are considered necessary for most salvage situations, including those that involve PI failure. For example, recent evidence suggests that the double PI regimen of ritonavir/saquinavir can rescue nelfinavir (NFV) failures.

There are fewer data on the combination of NFV/SQV as salvage therapy. It appears that for indinavir (IND) failures, the combination of RTV/SQV is preferable to NFV/SQV. There are new PIs (such as ABT-238) currently in clinical development. These new agents may not display cross-resistance to available PIs and would be very useful for failures that occur on current PI-containing regimens. In addition, non-nucleoside reverse transcriptase inhibitors (NNRTIs), in combination with drugs from other classes, have an important role in this setting. Many (but not all) HIV treatment researchers believe that patients who fail on multi-drug antiretroviral therapy should undergo resistance testing to guide the selection of drugs that may be successfully employed in subsequent treatment regimens. Current salvage regimens often will include three to five drugs. When the goal of decreasing the viral load to undetectable levels is not possible, maximal achievable suppression of viral load is sought. Dr. Yeni described situations in which an even larger "cocktail" of drugs-from six to nine agents-may be required to accomplish adequate suppression of HIV infection in heavily pretreated patients. Studies of such extensive cocktails in patients with significant prior experience with antiviral drugs, show that even though multi-component HAART salvage regimens may sometimes result in transiently undetectable viral loads, these responses are often not durable. In addition, such complicated regimens involving numerous drugs are associated with problems of adherence, tolerance, and drug/drug interactions.

Treatment interruption prior to starting a new drug regimen has been studied in the salvage therapy setting, but results have not been particularly promising. One study showed that two thirds of patients had plasma virus populations that appeared to have reverted to wild-type following treatment interruption. However, the viral level rose to an even higher level than before upon reinstitution of treatment coincident with the rapid outgrowth of pre-existing (but undetected) drug-resistant variants. Both multi-drug (HAART) regimens and treatment interruption strategies require further clinical trials before they can be recommended in the setting of salvage therapy, stated Dr. Yeni.

Salvage Therapy: New Drugs

"The operative words regarding new drug development are 'cautious optimism.' There are 14 FDA-approved drugs for the treatment of AIDS, and literally dozens in various stages of clinical development. A major effort is being made to develop drugs within existing classes, and efforts are also focused on developing therapies directed at novel targets," stated Joseph Eron, MD, of the University of North Carolina, Chapel Hill.

New Protease Inhibitors

Three new PIs furthest along in the stages of clinical development are:

* ABT-378/r. ABT-378/r, currently in Phase III studies, was designed by Abbott to decrease the interaction of this drug with the codon 82 mutation of the HIV protease that is associated with resistance to ritonavir (RTV). ABT-378/r combines ABT-378 with RTV and is 10 times more potent than RTV alone. Studies in treatment-naпve patients suggest that the drug has activity against viruses that have been exposed to other PIs. Studies in PI-exposed, treatment-experienced patients with several failures might have different results.

* Tipranavir. In vitro studies of tipranavir have shown promising activity against highly resistant PI isolates, but a new formulation of the drug was developed to improve its bioavailability. Tipranavir is currently in Phase II studies. "The new formulation is not out of the woods yet, because it has strong protein binding," said Dr. Eron.

* BMS-232632. BMS-232632, currently in Phase II studies, demonstrates potency in vitro, with modest protein binding. This agent appears to have activity against some strains of HIV that are resistant to currently available PIs.

* PIs in preliminary stages of development include Merck 944A, Agouron 1776, Parke Davis 178390, and several from DuPont. Studies of Merck 944A have recently been halted due to kidney toxicity in animal studies.

Reverse Transcriptase Inhibitors

* New NRTIs: Amdoxovir (DAPD) (the most promising one in development) and deoxythyocytidine (doTC).

* New nucleotide analog: tenofovir, a drug currently in Phase III trials, that can be taken once a day.

* New NNRTIs include AG 1579 (capravirine), DuPont compounds that are efavirenz (EFZ) derivatives, and calonolide A.

Agents Directed at Novel Targets

Agents directed at novel targets include:

* chemokine receptor inhibitors that block the virus from binding to an essential co-receptor in the first step of the HIV replication cycle (CCR5 inhibitors and CXCR4 inhibitors).

* integrase inhibitors that block interpretation of HIV into the host cell chromosome (a Merck compound is in pre-clinical development).

* fusion inhibitors that block fusion of the virus membrane with the reagent cell membrane-an essential early step of the HIV replication cycle. (Of these, T-20 is the most well studied; T-1249, being developed by the same company, is in Phase I/II trials.)

The optimal way to study new drugs is being debated. Some questions include: Should new agents be studied in the setting of first treatment failure or as late salvage therapy? Will they be useful in treating individuals with viruses resistant to NNRTI, NRTI, or PI?

Dr. Eron said that not all of these new agents would pass the hurdles for FDA approval. In his opinion, among the drugs currently being studied, "ABT-378, T- 20, and CCR5 inhibitors are the ones most likely to become available."

Drug Resistance and Treatment Failure

According to a number of experts who spoke at the meeting, resistance testing has come of age. Commercially available genotype and phenotype tests are being widely used at AIDS clinics around the country. Many AIDS experts believe that resistance testing should be the standard of care in cases of treatment failure. However, third-party coverage for resistance testing may be problematic in certain states and not all commercial labs have the ability to perform these complex assays in a high-quality, reproducible manner. None of the available tests for HIV drug resistance have been approved by the FDA.

"Although genotype and phenotype drug resistance testing is available and evidence to date suggests such testing is reliable, more information is needed to determine how best to use results of these tests to manage patients," explained Lisa Demeter, MD, an Associate Professor of Medicine at the University of Rochester School of Medicine, Rochester, New York. Resistance detected at baseline does not fully explain treatment failure, and it may be that plasma drug concentration levels are contributing to failure.

A review of existing drug resistance studies was conducted by the Resistance Collaborative Group (RCG), a composite of representatives from industry, academia, and the FDA. The RCG used a uniform definition of virological failure (plasma RNA >400 ml at week 24) in their analysis. This review demonstrated consistent benefit for both genotypic and phenotypic resistance testing when analyzed according to uniform criteria. These results led many experts to recommend that resistance testing should be the standard of care when designing salvage regimens. Genotypic and phenotypic testing has relative advantages and disadvantages. Genotypic assays (based on nucleotide sequence analysis) are more widely available and cost less than phenotypic assays (based on inhibition of HIV replication in tissue culture), and they may be more useful than phenotypic testing for early treatment failure. Current phenotypic assays may be particularly useful in extensively-treated patients. Phenotypic testing relies on recombinant viruses and is expensive, which may limit clinical utility.

Dr. Demeter suggested a number of areas for future study. These include: the relative roles of genotypic and phenotypic testing in early and late salvage therapy; to determine the resistance cutoff and predict success of an entire regimen rather than a single component; therapeutic drug monitoring; pathogenesis of drug resistance; and development of new assays with increased sensitivity at low viral loads, increased sensitivity for combinations of drugs, and improved detection of genetic variants.

CRITICAL ISSUES IN ANTIRETROVIRAL THERAPY

Although there are potent agents for the treatment of AIDS, there are several important issues to address, including the implications of viral reservoirs, the role of strategic treatment interruption, and the implications of the discordant immune response in patients with virological failure (i.e., stable CD4+ cell counts in the presence of virological rebound), explained Daniel Kuritzkes, MD, of the University of Colorado, Denver.

Viral Reservoirs

There is now compelling evidence that low-level viral replication continues even in patients with undetectable viral levels, and that patients with HIV infection have "viral reservoirs."

Dr. Kuritzkes stated, "Just because you can't measure something, doesn't mean it isn't there." Research has shown that viral reservoirs are established early in the course of HIV infection and are slow to decay. The timing of decay and the mechanism for maintaining such viral reservoirs while patients are taking potent antiretroviral drugs remain undetermined. Studies on the rate of decay of persistent reservoirs of HIV infection have been equivocal, and it appears that the rate may be tied to the degree and durability of viral suppression achieved in the treated patients studied.

Studies of techniques to purge viral reservoirs using interleukin-2 have been disappointing thus far. More potent regimens and more aggressive treatment may be needed to eliminate viral reservoirs. Specific drugs targeted to definitive types of viral reservoirs may prove useful in future studies.

Treatment Interruptions

Treatment interruption has been studied in the setting of salvage therapy and as a strategy for stimulating immune function. A study presented at the 1999 meeting of the ICAAC by V. Miller et al., showed that treatment interruption achieved reversion to wild-type virus in two thirds of patients in the salvage therapy setting; however, this was accompanied by a significant decline in CD4+ cell count, and it took almost a year for these counts to recover.

"It may be that treatment interruption initiates a continuous cycle between drug-sensitive and wild-type drug-resistant virus. This leaves the question open whether treatment interruption is a successful long-term strategy or just a temporizing measure (in the salvage setting). We may just be shuffling the chairs on the Titanic with this approach," said Dr. Kuritzkes.

Treatment interruption to stimulate immune function is based on the rationale that antigenic stimulation is required to maintain HIV-specific immunity. "It is a paradox that virus is needed to stimulate the immune response. In the majority of cases, viral rebound is almost always accompanied by evidence of immune activation," he told listeners.

Despite the theoretical benefit of treatment interruption, studies thus far have shown no long-term benefit in immune function. Also, undetected plasma HIV-RNA has been seen in individuals stopping therapy without treatment interruption, and virus-specific immunity may be short-lived. It remains to be determined what the consequences of treatment interruption are on viral load and on CD4+ cell count. It is not known whether enhanced immunity would persist beyond control of viral replication. It also remains unknown whether the immune response helps maintain virus replication and whether treatment interruption, an endogenous strategy, is better than exogenous immune stimulation with a vaccine.

CD4+ Response

In patients taking potent antiretroviral drugs, increased CD4+ cell counts correlate with the magnitude of viral suppression. CD4+ cell count does decline in patients with partial viral rebound, but this is a slow process. Cessation of antiretroviral therapy leads to increased viral rebound and a decline in CD4+ cell count. In some patients, however, CD4+ counts remain stable in the presence of detectable replicating virus, which Dr. Kuritzkes called "a discordant CD4+ response."

Questions about the discordant CD4+ response include: Is the effect of continued therapy mediated by persistent inhibition of virus replication per se or by maintaining selective pressure for replication-impaired variants? Is this response limited to protease inhibitors (PIs), or is it seen with NNRTIs and triple NRTI regimens? What determines the relationship between virus replication and CD4+ cell loss? "This last question is a fundamental question we have been asking since the beginning of the AIDS epidemic," concluded Dr. Kuritzkes.

Metabolic Complications of Antiretroviral Therapy

The life-extending effects of antiretroviral therapy were greeted with joy by HIV-infected persons and by physicians. However, experience with these drugs has shown that they have a number of worrisome side effects. Well known side effects include body fat redistribution syndromes (i.e., lipodystrophy), undesirable lipid changes, and mitochondrial damage. Two studies presented at the CROI meeting showed that antiretroviral therapy causes a substantial decrease in bone mineral density (BMD), raising the potential for osteoporosis and subsequent bone fracture. In one study, 21% of 122 patients taking antiretroviral therapy could be classified as having osteoporosis, while 50% had evidence of bone thinning, or osteopenia. A second study found evidence of osteopenia and/or osteoporosis in approximately 30% of 80 HIV-infected men with lipodystrophy syndrome who were taking a PI-based regimen. Neither study could confirm a specific association between PI therapy and osteoporosis.

David Cooper's group in Sydney, Australia, first reported the phenomenon of body fat composition changes seen in patients on highly active antiretroviral therapy (HAART). These changes include the "buffalo hump," or fat pads around the back of the neck, and central adiposis, characterized by fat accumulation in the abdomen and visceral organs and thinning of the trunk, face, and limbs. The fat redistribution syndrome has varying degrees of expression in different patients. At first it was thought that lipodystrophy was a side effect of PI therapy, but it has now become evident that this can be a side effect of NRTIs as well (or perhaps all effective HAART regimens). "We don't know if these findings will change the way we treat patients. We don't have a certainty about how common these side effects are, but some degree of metabolic changes have been reported in 50% to 60% of patients who have taken HAART for 5 to 6 years," explained William Powderly, MD, of Washington University, St. Louis, Missouri.

The etiology of the fat redistribution syndrome is not clear. A number of hypotheses have been proposed, including an enzyme-blocking effect of PIs, the HIV disease process itself, or some combination of these two explanations.

"We don't have clear-cut proof of what causes the fat redistribution syndrome. But these changes were not seen in the pre-HAART era, and most experts believe that the drugs are the cause. The role of PIs is not clear," said David A.Cooper, MD, of St. Vincent's Hospital, National Center in HIV Epidemiology and Clinical Research, Sydney, Australia.

New evidence suggests that both PIs and NRTIs are associated with body fat redistribution and lipid changes, even though these agents have different mechanisms of action. The toxicities of both classes of drug appear to be primarily metabolic. It may be possible to dissect out the causes by studying the intersection of toxicities, noted Dr. Cooper.

The pathogenesis of lipodystrophy appears to be complex. The prevalence and severity of body fat composition changes varies considerably among patients. Epidemiologic data suggest that specific drugs may be more likely to cause fat redistribution than others. In addition, studies suggest that treatment duration and patient age are also important factors associated with the development of this syndrome. People on the same HAART regimen can have opposite effects, and they can have mixed syndromes. Buffalo hump and central adiposis are now considered to represent different biological processes.

An important concern is that the changes in body fat may also contribute to the development of insulin resistance, a precursor of diabetes. This, in combination with the adverse changes on lipids, may potentially lead to an increased risk of heart disease in persons treated with HAART. Studies are now being conducted to determine whether patients with HIV who develop risk factors for heart disease should be treated in a prophylactic manner with drugs like the HMG-COA reductase inhibitors, known as statins.

Experts at the meeting agreed that the benefits of HAART far outweigh the risks for persons with AIDS. Several of the speakers said that specific drugs should not be singled out as the culprits for metabolic changes.

"The story is evolving. Older patients seem to be most affected by metabolic abnormalities, but we do know that lipodystrophy is not a normal consequence of aging. Duration of exposure to drug treatment appears to be the major risk factor. We urge caution in directly linking specific drugs to changes in body fat composition. Until randomized controlled trials are done, the best data we have suggest that duration of exposure is the biggest risk factor," stated Karl Grunfeld, MD, of the University of California at San Francisco.

Both retrospective and prospective studies are needed to tease out the effects of duration of therapy and of specific therapies. Most studies that have been conducted to date are cross-sectional, providing evidence at one point in time; longitudinal studies could help shed light on the evolution of metabolic changes in patients on HAART.