Evaluating Drug Resistance Tests
Almost two years since genotype and phenotype testing have become generally, if not widely, available for the diagnosis and evaluation of HIV drug resistance, there is still much controversy about their accuracy as diagnostic tools and their usefulness as predictors of treatment success. In theory, the likelihood of treatment failure can be minimized by using resistance assays to determine ahead of time which drugs each person's particular strain of HIV is susceptible to, and choosing those drugs for that individual's treatment regimen.
Drug resistance could be called the Achilles' heel of the current strategies for the treatment of HIV infection. Resistance results from HIV's tendency to mutate its genetic material (RNA) in a way that impedes one or more of the antiviral drugs from inhibiting viral replication. The virus' genetic material is an instruction manual, telling the virus what to do and how it will look.
In the absence of antiviral treatment, HIV replicates freely in the blood stream, making billions of copies of itself each day. This virus is usually called "wild type" because it has not been exposed to any drugs. Since the virus replicates at such an extraordinarily rapid rate, errors frequently occur during the process of producing new copies. This results in HIV whose genetic material is not identical to the original and is in some way defective. These defective viruses are called mutants and occur at all times during the replication process. However, the mutations remain a relatively small portion of the total viral population, since wild type tends to be the prevalent and stronger strain of virus.
Some of the mutations that do occur make the virus so different from the wild type that one or more of the antiviral drugs are no longer effective, making those mutant versions of HIV resistant to a particular drug or class of drugs. However, not all mutations cause the virus to become resistant to all drugs. In some cases, the presence of a single mutation can confer resistance to a drug, while in other cases a set of mutations is required before a drug becomes ineffective. For example, HIV develops resistance to 3TC (Epivir®) due to the presence of a simple mutation (at position 184), while a complex series of mutations is required for the emergence of resistance to Crixivan® (indinavir) and other protease inhibitors. Each person's mutations are called resistance patterns.
If an antiviral combination is not completely suppressing viral replication, drug-resistant mutations will eventually emerge and become the dominant strain in an individual's viral population. The direct effect of this can be an increase in viral load, which is generally known as virologic failure, and subsequent drug or treatment failure.
Currently, there are two methods to determine whether a particular virus is sensitive or resistant to a drug: genotype and phenotype testing. Genotype testing looks for the mutations that make the virus resistant or potentially resistant while phenotype testing measures the potential sensitivity of a cultivated virus to a particular drug. The methods are very different.
Genotyping relies on the amplification of the gene sequence of the reverse transcriptase and protease enzymes of HIV. The net result of the genotype test is a map of the genetic sequence of the virus with its associated mutations. Because researchers have identified some of the mutations associated with certain resistance patterns, they can determine whether the virus being tested is resistant to a particular drug or drugs. The problem with this is that the presence (or absence) of a mutation or set of mutations known to confer resistance in a lab setting does not always correlate with treatment outcomes in a real-world setting. Likewise, drug failure in an HIV-positive person cannot always be explained by genotypic analysis.
Phenotype testing measures the potential sensitivity or resistance of HIV by growing virus in a test tube with genetic characteristics copied from blood samples submitted by individual patients. Cultures in which the virus is growing are then treated with various antiviral therapies to determine how much drug is needed to inhibit the virus. The results are compared to the amount of drug needed to inhibit laboratory standard or wild type virus.
If the virus cannot grow in the presence of a drug, then it is still sensitive to that drug. However, if the virus replicates freely or requires a larger than normal amount of a drug in order to be suppressed, then it is resistant or has a degree of resistance to that drug. The degree of resistance is often expressed in terms of the amount of drug required to suppress replication. Generally, anything above a 10-fold increase in the amount of drug required to inhibit the virus is considered highly resistant and likely to mean that the drugs are no longer capable of blocking the virus from replicating. A finding of a four- to ten-fold increase is considered intermediately resistant. Less than a four-fold increase is considered sensitive.
Because they directly test the patient's virus against the actual drugs used in treatment, phenotypic assays are sometimes considered the gold standard of resistance testing. The test is quite expensive, costing over $800, and takes 4-6 weeks. Genotyping is a simpler process and typically costs less, between $300 and $500, and takes less time. Both assays generally require a minimum viral load of 1,000 to 2,000 copies in order to be accurate.
Studies of Resistance Testing
While there has not been definitive proof that either test can predict whether a particular combination therapy will be successful, a number of recent reports have suggested that resistance assay results may reliably predict response to therapy. In a study presented at the Resistance Workshop in Italy last summer, indinavir failures whose HIV was sensitive to two or three of the drugs in their salvage combinations experienced a dramatic and sustained reduction in viral load through the 16 week study period. In comparison, patients with HIV sensitive to zero or one drug in their salvage regimen only experienced a transient drop in viral load in the first 4 weeks that rapidly returned to baseline.
Another study presented at the Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) in San Diego last fall demonstrated that patients with phenotypically or genotypically sensitive virus to ritonavir (Norvir®) with saquinavir (Invirase®) before therapy were 4- to 12-fold more likely to achieve a viral load below 500 copies/ml using this double protease inhibitor combination. However, patients whose tests showed sensitivity to both drugs were not always successfully treated. This correlates with information collected in databases that indicates that about 15% of HIV isolates without known protease-inhibitor mutations are still protease-inhibitor resistant.
At the International Congress on Drug Therapy and HIV Infection in Glasgow in November, preliminary results from the French VIRADAPT study were presented. This study looked at the value of genotyping in selecting salvage regimens. The 108 participants had previously experienced treatment failure and were divided into two groups. One group switched regimens based on physician judgment and treatment guidelines. The other group used genotypic testing to select a new regimen. Those in the group using genotyping were more likely to have a viral load below the limit of detection at both month three and month six.
At the recent Sixth Conference on Retroviruses and Opportunistic Infections held in Chicago, a similar study on genotype testing was presented. This study enrolled 153 volunteers with evidence of viral load rebound on a triple combination regimen. The participants were randomized either to receive a genotype with expert guidance on its interpretation (GART), or the current standard of care in which the clinician chose the next regimen based on the patient's treatment history. Overall, there was a significant difference in the viral loads in the two groups, with a lower viral load in the group using GART. In both arms, the viral load decline was transient and viral load rebound was noted by week 12. Nonetheless, these results suggest that there is likely to be a role for genotypic testing in the selection of antiviral medications for patients experiencing treatment failure.
The Role of Resistance Testing in Clinical Practice
Due to the growing body of data on resistance patterns in the context of scientific and clinical research, there is a better understanding of how the presence of HIV mutations in an individual impacts his or her treatment options. Genotypic and phenotypic analysis could be used as diagnostic tools to predict the success of initial therapy, to evaluate and perhaps measure the onset of resistant patterns, to determine salvage therapies and to follow the spread of multi-drug resistant virus.
Supporters of the use of resistance testing in treatment-naïve patients maintain that prior knowledge of pre-existing resistance patterns would enable the selection of an initial therapy with the greatest chance of success. The problem with this approach is that most people who are treatment naïve have wild type virus, making these tests an unnecessary expense. However, there is evidence that strains of drug-resistant virus are being transmitted. It has been reported that AZT (Retrovir®)-resistant virus is present in 10% of new infections. There was also a report of a virus resistant to all commercially-available HIV drugs being transmitted in San Francisco.
Genotypic and phenotypic analysis could be used to track the spread of drug-resistant HIV. A presentation in Chicago focused on the resistance profiles of HIV isolated from 114 treatment-naïve people with documented infection within the last three years. Using genotype and phenotype testing, the researchers found that the prevalence of resistance to two different classes of drugs was 3.2% by genotype and 2.2% by phenotype, and the prevalence of resistance to all three classes of drugs was 2.1% by genotype and 3.3% by phenotype. Although the investigators suggested that based on these data, "resistance testing prior to starting therapy may be useful to optimize initial therapy," it would be hard to recommend general resistance testing based on such low levels of resistance in treatment-naïve patients. These data could be used to monitor the spread of drug-resistant HIV for both epidemiological and prevention purposes and, in some cases, to recommend a more widespread use of resistance testing.
The value of genotype and phenotype testing is more evident in the diagnosis of resistance in treatment-experienced patients and in the development of strategies for treatment failure. Up to now, the most reliable method of determining drug resistance and treatment and virologic failure is testing viral load. An increase in viral load from below the level of detection may be an indicator of the onset of viral resistance and drug failure. One possible use for the phenotype and genotype assays is to catch the development of resistance early so that a change in treatment can be made before a complex pattern of resistance evolves. In fact, there is some evidence to suggest that a virus that shows early evidence of resistance to one protease inhibitor may still be sensitive to other PIs if a switch is made quickly enough. This would prevent the development of cross-resistance within a class. If this is proven to be true, then genotype and phenotype testing will have an important role in monitoring the efficacy of treatment.
The use of genotype and phenotype testing in designing salvage-therapy regimens is clearly the most compelling at this point. Both tests have shown some promise in improving the odds for success in salvage therapy, although the clinical benefit has yet to be definitively demonstrated. A presentation at the Retrovirus conference showed that evaluating failure with genotype and phenotype testing can help make an informed decision without sacrificing treatment options. In this case, the researchers looked at patients failing therapies that included indinavir and two nucleoside analogs. They concluded that the patients failing these regimens had virus more likely to present resistance to the nucleoside analogs than to the protease inhibitor, therefore allowing the change of the ineffective portion of the regimen only. Without resistance testing, the protease inhibitor might have been replaced as well, even though it was still effective.
While both methods need to be further fine-tuned and researched, the ability to use resistance assays to pinpoint which drugs in a failing combination are not working would allow only the problem drugs to be replaced and retain the benefit of any drugs that are still working. This would greatly increase treatment options.
Another example of the use of this new technology was presented at the Retrovirus conference. Baseline resistance testing was used to evaluate virologic response to a salvage protocol using Mega-HAART (more than six antiviral drugs). This study looked at 37 patients failing therapy, with phenotyping available for 24. The results of treating with Mega-HAART were given in terms of response (achieving and sustaining a viral load below 500 copies) and failure (never achieving a viral load below 500 copies).
The correlation between baseline resistance and response or failure was evident. While two out of the ten responders had virus resistant to some protease inhibitors, four out of six failures had viral isolates resistant to all protease inhibitors. Of the ten responders, seven had virus sensitive to three or four protease inhibitors, while only one of the six failures had virus sensitive to three or four protease inhibitors. The aim of this study was to measure the Mega-HAART response by baseline resistance and not the predictive value of genotype testing, but it is clear that such information is relevant in the management of heavily-pretreated patients.
While genotypic analysis points to specific mutation patterns that can be associated with resistance to a drug or combination of drugs, phenotypic analysis can help gauge the degree of resistance. Several studies have found a correlation between phenotypic and genotypic resistance. Perhaps an approach using both methods could provide the best results for predicting success or failure of salvage regimens, especially in heavily-pretreated patients.
This article was provided by AIDS Community Research Initiative of America. It is a part of the publication CRIA Update. Visit ACRIA's website to find out more about their activities, publications and services.