|
Thomas Merigan, M.D., is Professor of Medicine and Director of the Center for AIDS Research at Stanford, and a leading expert in HIV treatment research and viral drug resistance. Dr. Merigan was also a co-principal investigator of the GART study,(1) reported last month at the Retroviruses conference, which showed that HIV resistance testing -- plus expert assistance in interpreting the test -- could make antiretroviral therapy more effective. AIDS Treatment News interviewed Dr. Merigan on March 23, 1999.
Background
When treating bacterial infections with antibiotics, physicians have long had tests available to tell them which drugs should be avoided for a particular patient, because the bacteria had become resistant to that drug. It is much harder to develop such tests for viruses, but today two different kinds of tests for HIV resistance to antiretrovirals are in experimental use. These tests are expensive at this time, and not yet approved by the FDA, but physicians can use them. Recently two clinical trials -- GART,(1) which was mentioned above, and the VIRADAPT study(2) in France -- have provided clear evidence that viral resistance testing, when properly used, can improve the effectiveness of HIV treatment in certain cases.
The two kinds of tests are called genotypic and phenotypic viral resistance testing. Genotypic tests look for mutations in the virus which are known to be associated with viral resistance to particular drugs. Phenotypic tests grow the virus in the laboratory to see how well the drugs work to stop it; these tests are time-consuming and expensive, in part because virus from patients generally does not grow well in the laboratory, so genes from the patient's virus must be transplanted into a laboratory-adapted virus before the drug tests can be run. In HIV infection, a single variant of the virus is usually transmitted, but in the patient, this HIV evolves into millions of different genetic variations. The resistance tests currently available cannot detect minor populations. If a patient is being treated with an antiretroviral and resistance develops, and then the patient stops the drug, the resistant virus can over time become a minor population temporarily (since non-resistant virus is generally more fit and reproduces more rapidly, as viruses usually must pay a price for drug resistance). The resistant virus would then be missed by the test -- but can come back very quickly if the drug is started again. Therefore, a patient whose treatment is failing, and who is being tested to determine which of the drugs they are taking have become ineffective for him or her, should best be on the regimen when their virus is tested for resistance.
Interview with Dr. Merigan
AIDS Treatment News: What do people need to know now about viral resistance testing?
Merigan: There are three settings where HIV resistance testing might be used clinically:
(1) In case of exposure such as needlestick or possible maternal transmission, the virus from the contact may be tested, to make sure that this virus is not already resistant to the drugs being used in hope of preventing infection. Of course treatment will often need to start empirically before the results of the resistance test are available.
(2) Resistance testing might also be used before starting a person's first antiretroviral treatment -- especially if they are likely to have been infected with a virus already resistant to antiretrovirals.
(3) The third situation is when an antiretroviral drug regimen has clearly failed, and the doctor believes that the drugs are all being taken and the most likely cause of the drug failure is antiretroviral resistance.
So resistance testing can be used either to realign existing therapy, or to help pick the best initial therapy for the patient.
ATN: And when trying to analyze which drugs in a regimen are not working, it is important to still be on the regimen when the sample is taken for resistance testing, to prevent "wild type" virus from growing back and hiding the resistant virus from the test?
Merigan: That's true. But also, the regimen will usually be partially active for the patient -- another reason the doctor may not want to stop that treatment until a new regimen has been chosen to replace it.
Genotypic Vs. Phenotypic Resistance Tests
ATN: How should our readers understand the difference between genotypic and phenotypic testing? The general impression in the community is that the phenotypic test is best, or at least easiest to interpret, but also that this test is more expensive and takes a longer time before the results are available.
Merigan: It is important to know that any interpretation of a genotypic result must be based on phenotypic testing of that drug done sometime in the past [not with the same patient, usually]. Every time I interpret a genotypic assay [test], and say that this patient's virus is resistant to drugs X, Y, and Z, it means that somebody had already taken that kind of virus, with those mutations, and tested that virus in a phenotypic assay. So the genotypic test is basically a shortcut, based on knowing that the given amino acid change in the reverse transcriptase gene, or in the protease gene, results in phenotypic resistance to that drug.
It is important to help people understand that the genotypic information, when interpreted properly, is also giving you something equivalent to a phenotypic assay. An exception is that if you are encountering a new kind of virus for the first time, then you would have to do the phenotypic assay to know what you're up against. Our laboratory has been doing this kind of work for several years now; we have discovered the multi-nucleoside and the multi-drug resistance pathways. We look carefully at new patterns of amino acid changes if they develop during antiretroviral treatment -- especially if we can associate them with rising viral load.
You need to educate the public that no genotypic result stands alone; it depends on phenotypic verification being done at some time. That is why we can make a projection of resistance with a genotypic result. On the other hand, doing a phenotypic assay on every person's virus is like going back to first principles -- proving against each drug tested that you have resistance or sensitivity.
There are complications; for example, combinations of drugs can have actions in the body that are not seen in laboratory tests. You need to know the pharmacology of the drugs in patients to know that a given level is associated with resistance or sensitivity. Even if you measure resistance directly for each patient, a given number does not necessarily imply that the drug will fail clinically; that may depend, for example, on whether or not drug levels in the blood go below certain thresholds. And in the future, there may be drugs that are so active that they are effective even against virus which might have been judged resistant by other standards.
ATN: If a company is offering a test for viral resistance to particular drugs, wouldn't they be able to provide adequate guidance on how to interpret the results?
Merigan: They try to, but our understanding can change.
In drug development, you have to look at the pharmacology, as well as resistance patterns from genotypic or phenotypic tests. And sometimes companies do not do all of the necessary pharmacology, especially during the first few years in the drug's development.
In AIDS we are pushing the envelope in developing these drugs. And sometimes we would need data that has not yet been obtained, to come to firm conclusions.
ATN: How much of a problem do doctors have at this time in deciding which genotypic or phenotypic lab to use?
Merigan: The only phenotypic group now doing work on patient-submitted specimens is VIRCO in Belgium; they are working through LabCorp in the U.S. But there are many laboratories doing genotypic testing.
Because there have never been head-to-head trials of the two assays, comparing genotypic and phenotypic testing directly against each other, we do no know which niche applications might be better for one test or the other. It is likely that both of them give almost identical information for many, many uses -- that would be my guess. There may be specific cases where one would have advantages over the other, but this would have to be proven; it cannot just be hypothesized. We need that head-to-head test.
Our group is beginning work on designing one now. But to do the trial, we will need two test manufacturers who are willing to go head to head. That's like finding two drug manufacturers who are willing to put their drugs in trial against each other -- hard to do. We are working on it.
ATN: Especially with genotypic testing, it would help greatly if physicians could get expert assistance in interpreting the results. The GART(1) study showed a big difference between physicians who followed the advice of expert virologists and those who did not -- and these physicians were all HIV experts. Where are we now and in the near future on feasible ways to make such help widely available?
Merigan: Many companies doing these tests are already handing out tables showing the relationship between amino acid changes and drug resistance. Also, some of them have experts for doctors to call to discuss individual cases.
For the future, we should have expert meetings once or twice a year, which publish guidelines that summarize the best advice. But that best advice will always have to be tempered by the individual patients' side effects, drug history, and of course the possibility of drug interactions. No one can completely codify these; they need to be translated by a thoughtful physician who knows the full story of the drugs being used, the viral-resistance results, and the patient.
ATN: What about artificial-intelligence approaches for guiding physicians in choices of therapy?
Merigan: People have various software approaches. But none of them are either licensed or widely distributed. And there are fears about legal responsibility if the system fails to work in some case; what if you left out a fact that made it predictable that it would not work? I am not sure how we are going to handle this issue. It is easy to look at what the genotype means from a phenotypic standpoint; but once you start bringing in all the other factors, the decision approaches the usual complexities of the practice of medicine, where the most important thing is for the patient and the doctor to have a good understanding of what they are up against, and share the issues. And it means you have to take a risk; there will always be some unknowns.
ATN: Even with consensus medications, most doctors even in HIV practices will not have time to keep up with the intricacies of all the different mutations.
Merigan: That is why at least converting the mutations into generally accepted phenotypic results is a logical thing to do; this is happening already, and will improve with time.
Reimbursement for Viral Resistance Tests
ATN: What about reimbursement for these tests by private or public insurance?
Merigan: Policies for reimbursement seem to come rapidly in HIV. I don't believe there will be major gaps. Studies like GART(1) and the French VIRADAPT(2) have been presented at conferences; when these are fully published it will be easier to get reimbursement. Both third-party carriers and government agencies are already paying for these tests for some patients.
ATN: Could you summarize the GART study?
Merigan: It was a randomized trial to examine the usefulness of genotypic testing, plus expert advice on the results, when the volunteers needed to change their treatment regimen. Everyone in the trial was tested for HIV resistance mutations, but for 16 weeks, volunteers were randomly assigned to have their doctors receive the resistance-test results or not; at the end of the 16-week study period, everyone was given the test results.
ATN: I understand you were one of the three virologists who provided expert advice to the physicians.
Merigan: The others were John Baxter at Rutgers, and Doug Mayers from Henry Ford Hospital in Detroit.
ATN: And on the whole, there was about a half-log [3-fold] greater reduction in viral load when physicians received the results -- and that approached a one-log [10-fold] improvement at those institutions where doctors tended to follow the expert advice?
Merigan: That is correct. Remember that we only did this study for 16 weeks [since patients would be reluctant to volunteer for a long study where test results which might help them were being withheld].
ATN: And the advice was to use an average of about four drugs; without the test results or the advice, the physicians used an average of about three?
Merigan: The principle advantage of GART availability was that it allows selection of the best four drugs -- not just four drugs, but ones which were likely to work.
Viral Resistance and Developing Better Drugs
ATN: You mentioned that your laboratory is working toward the design of new protease inhibitors which specifically target viruses resistant to the current drugs. Explain how this approach could help everyone, even patients starting therapy, not only those already resistant to the protease inhibitors currently in use.
Merigan: Ten years ago, when everybody knew that viral protease was an important target, they all made the same three-dimensional structure of the protease molecule, and tried to fit compounds into it. As a result, there is such a close relationship between all the HIV protease inhibitors that have been developed, that it is not surprising that we are out of options. If we could inhibit the virus by targeting viral protease which has become drug resistant, maybe we could use standard protease inhibitors more rationally, and make them last much longer.
New drugs developed this way could be very good candidates for combination treatment; they might be as different from conventional protease inhibitors as non-nucleosides [efavirenz, delavirdine, nevirapine] are from nucleoside analog drugs [AZT, ddI, d4T, abacavir, etc.]. Why not let what the virus has done drive us in new directions? Especially since we are not moving forward very well with targeting the other genes of HIV.
ATN: Two weeks ago in San Francisco you mentioned that you had to go to an X-ray crystallographer in an academic institution, rather than to a pharmaceutical company, to pursue this research approach. Why have the companies been slow to move?
Merigan: It is a departure for the pharmaceutical companies, because they would rather focus on first-line therapy. We are starting with the most troublesome virus, and trying to extend it backwards some day to first-line therapy as well. But the first question is, can we come up with antiretrovirals which are tailored for third-line use, for heavily treated patients who are otherwise out of options?
ATN: So this is an important research front that might lead to better treatment for everyone?
Merigan: It has that broader potential. And when you do science, if it's done very well, it surprises you. It is not too much to hope for a pleasant surprise.
ATN: Where do you think viral resistance testing will fit in clinical care in the future?
Merigan: I believe these tests are a new phase, like viral load, in the rational treatment of patients with HIV. There are unknowns, but to me it seems clear that we should push ahead, and that patients need the advantage of this approach.
References
1. Baxter J.D., Mayers D.L., Wentworth D.N., Neaton J.D., and Merigan T.C. A pilot study of the short-term effects of antiretroviral management based on plasma genotypic antiretroviral resistance testing (GART) in patients failing antiretroviral therapy. 6th Conference on Retroviruses and Opportunistic Infections, Chicago, January 31 - February 4, 1999 [abstract LB8].
2. Durant J, Clevenbergh P, Halfon P, and others. Can HIV genotype determination be useful for individualized adaptions of antiretroviral therapy: The VIRADAPT French study [abstract OP 7.1]. Abstracts of the Fourth International Congress on Drug Therapy in HIV Infection, Glasgow, Scotland. AIDS. 1998; volume 12, supplement 4, S16.
| 
