A First Report in a Series on Resistance Testing; including a Geneva Report on Transmission of Protease Resistant Virus

This is a preliminary first report on genotypic and phenotypic resistance testing. This report includes the presentation by Doug Mayers, MD at the NATAP Post-Geneva Community Forum in July '98 at NYU Medical Center. It also includes additional information from Jules Levin. The draft of full proceedings from the July Post-Geneva Forum in July is finished and being finalized now for printing and distribution. It will also be posted on this web site. If you would like a printed hard copy please contact NATAP by our email address posted on the front page of this web site or by calling us at 212 219-0106.

Let's try first to define the two types of testing so you understand what they do. The types of tests are genotypic testing and phenotypic testing. Resistance testing can be a complicated topic. If you do not understand some of this report please feel free to call our offices for an explanation.

Phenotypic Resistance Test--What is phenotypic resistance testing? It was being used in laboratories for research purposes, but due to advances in technology it is being made available commercially for doctors and consumers. Using the phenotypic testing method, a sample of a person's blood is placed in a test tube along with a specific drug. The laboratory determines how much drug is needed to inhibit HIV by either 50% or 90%. When you are talking about the inhibitory concentration (IC) required to do this you'll hear the terms IC50 or IC90. IC50 means the drug can inhibit 50% of HIV production in the test tube and IC90 means 90% of HIV is inhibited. This is called phenotypic resistance testing. If you need more drug to suppress virus than is normally needed that means there is phenotypic resistance.

Genotypic Resistance Test-- A different way to look at resistance is to perform a genotypic test, where you look at the sequence of the genes in the virus itself. You are looking at the changes in the genes and these are called mutations; but, you have to look for mutations that have been proven to result in resistance for that drug, to make the findings useful for the patient. Unfortunately, we have not identified all the mutations that may be relevant. As we start combining multiple drugs different mutations may emerge. For example, failure of a NNRTI+ a NRTI may result in a different mutation than is known to be associated with NNRTI resistance. This subject is discussed more below.

Genotypic Resistance Testing

Strengths

• This testing method detects the major populations of circulating virsuses. What you see is actually growing the best. That essentially means that high level or moderate resistance is more easily detected, but low level resistance that may be present at a rate of about10% may not be detectable. How crucial is it not to be able to detect this low level resistance? It appears opinions are mixed about this
• It detects and reports to you the actual mutations detected. For example, if you have indinavir resistance it should report to you that they detected 82, 46, 84, 63 mutations. When using a phenotypic resistance test alone you will receive a report saying simply that you have low, medium or high level resistance to a specific drug; you won't receive a list of mutations unless you use genotypic testing. But, some doctors prefer phenotypic testing because the presence of mutations does not necessarily mean you have full resistance to a drug. Using both genotypic and pheno- typic testing can be preferable but expensive if your insurance won't cover it. So far, it appears doctors are able to obtain insurance coverage to pay for genotypic testing, but not phenotypic testing just yet. But work is being done to convince insurers to pay for the phenotypic testing.
• Rapid turnaround time. Currently, it can take 7-10 days to get results back from the lab. You can get back results within 2-3 days from some labs. In the future, it may be possible to get results back in 1-2 days
• It may detect non-adherence because if you fail a combination but no resistance mutations are reported it may be a sign the person is not taking the drug(s)

Weaknesses

• Currently, there is a poor understanding of mutational interactions. For example, we are familar with the fact that in some cases 3TC resistance reverses AZT resistance; or, the presence of the 3TC resistance mutation 184 may improve adefovir efficacy. AZT mutations may be detected and reported using a genotypic test, but a phenotypic test may report only possible AZT resistance because the 3TC 184 mutation is present. There may be additional interactions we are not yet aware of, and so a test cannot take them into consideration although they may effect actual resistance.. Databases are being rapidly developed. A good database should improve use of testing. It's expected that these databases will allow predicting phenotypic resistance based upon the presence of specific mutations. An additional difficulty is that results can be reported back to doctors in a way that is very difficult to interpret, partially because they may not understand the importance of specific mutations, or because it is unclear how prevalent a specific mutation may be
• This testing method will miss new mutations. A mutation can occur that has never been seen before and therefore it is not known whether or not it contributes to resistance for that drug. However, in this case a phenotypic test can detect this resistance even if the genotypic test does not. For example, a 69 mutation that can emerge as a result of multi-drug NRTI resistance or may be associated with d4T resistance was recently discovered and first reported at the Resistance meeting. A genotypic test may not yet be able to detect this mutation
• Neither the genotypic nor the phenotypic testing can detect mutations that are present at a very low level
• The person has to be taking the drug to detect resistance. If a person stopped taking the drug a while ago, testing may be unable to detect resistance. But, once the person starts taking the drug again resistance can emerge quickly.

Phenotypic Test

Strengths

• They will tell you if they detect actual resistance to a drug rather than just seeing a list of mutations; you receive a report saying a person is either susceptible to a drug or has low, medium, or high level resistance to a drug; this makes the results more easily interpretable
• Until reliable databases have been constructed, this test may be more reliable in predicting resistance than the genotypic test.

Weaknesses

• The amount of resistance required for failure of a specific drug may not be known yet for all drugs. For example, a phenotypic test may report that a person has 4-7 fold resistance to a drug. That may be enough to fail one drug and may not be enough to fail a different drug
• It takes longer to get results back than it does for genotypic testing. It can take 9 to 14 days to get results back.
• It costs twice as much to perform a phenotypic test than a genotypic test. A genotypic test can cost about $400 and a phenotypic test can cost about $800
• As with genotypic testing, if a person has stopped taking the drug, resistance may not be detectable. This can deceive you into thinking the person is not resistant to a drug. But, resistance can quickly emerge after restarting a drug. Therefore, it is very important to consider a person's previous drug experience. If they had used and failed indinavir, but, they stopped using it 3 months ago. Resistance may not be detectable but might emerge after starting a new PI. This may be the cause of the weakness discussed in the next bullet point
• The test is better at reporting resistance than sensitivity to a drug. There is a good amount of data showing that if phenotypic resistance to a drug is reported to you, then you will not get a response to the drug. But, if it's reported to you that you are not resistant to a drug you may or may not respond to that drug. So, the test is better at reporting you will not respond to a drug than if you will respond to a drug

3 Recent Studies of Resistance Testing

These studies were reported at Geneva or the Resistance meeting.

Baseline Phenotypic Test and Response at Week 16. Dr Steven Deeks took 18 patients who failed indinavir and gave them a 4-drug regimen of abacavir+saquinavir (Fortovase)+nelfinavir+ either nevirapine or a NRTI. They defined sensitivity to a drug as having less than 2 fold resistance using the ViroLogic Phenotypic Test. If you had greater than two-fold resistance to a drug then you were defined as being resistant to it.

N	Drug	Response	<400 copies/ml	<50 copies/ml
7	0-1	none or transient	0/7	0/7
11	2-3	sustained for 16 weeks	8/10	4/9
0	4	-	-	-

As you can see in line 1 of the table, 7 of the 18 patients were sensitive to 0 or 1 drug in the regimen they received. That means they were resistant to 3 or 4 drugs. These 7 people had no response to the therapy they received. None of them had <400 copies/ml (undetectable viral load) after 16 weeks on therapy.

11 of the 18 patients were sensitive to 2 or 3 of the 4 drugs. They were resistant to 1 or 2 of the 4 drugs. 8 out of 10 (80%) people had <400 copies/ml (undetectable viral load) after 16 weeks of therapy. And, as you can see one of the 18 patients were sensitive to all 4 drugs.

The persons who were sensitive to at least 2 or 3 drugs had a median viral load reduction of 2.33 log at week 16. The individuals sensitive to 0 or 1 drug had a median viral load reduction of 1.2 log at week 16. The difference was statistically significant (p=0.007).

Interpretation and Application of Results-- So, what does this mean? These study results are potentially very important. You may be able to select a salvage regimen that will work for you by using phenotypic resistance testing. If you are not resistant to 2 or 3 drugs in a regimen it is more likely to work. You may have heard the results of this study reported wrong to you. Some have said that a 4-drug salvage regimen consisting of nevirapine+saquinavir+nelfinavir+abacavir was successful at suppressing viral load for individuals who had indinavir resistance. That is not what this study says. This study says that if a person was sensitive, using the phenotypic resistance test, to 2 or more drugs then they were likely (8/10) to reduce their viral load to undetectable (<400 copies/ml). If a person took all 4 of these drugs but was resistant 3 or 4 of them they did not lower their viral load <400 copies/ml. The key is not which drugs you use but if you are sensitive to at least 2 or 3 of the drugs. You could have used any 4 drugs in this study.

Response to Ritonavir+Saquinavir in PI experienced Patients Using Genotypic Test. Dr Andrew Zolopa took 51 patients who had failed at least 1, 2 or 3 previous PIs. He looked at viral load responses at 4 and 12 weeks. He found the following baseline information about the patient was predictive of the response: stage of HIV disease (advanced or moderate), CD4 count, viral load, HIV drug history (number and duration of prior antiviral treatment). Individuals with higher CD4s and lower viral loads were more likely to respond to the treatment.

The key finding from this study is that independent of all the predictors mentioned above, Zolopa found that simply the number of PI mutations (30, 46, 48, 54, 82, 84, 90) that were found upon performing a genotypic test significantly predicted (p<0.001) the response to a ritonavir+saquinavir regimen.

So, in the first study phenotype testing predicted response, and in the second study genotypic testing predicted response.

Using Both Genotypic and Phenotypic Resistance Tests to Predict Response to Ritonavir+Saquinavir. Richard Harrigan reported at the Resistance meeting just before Geneva on 84 patients who had significant exposure to Pis and received a ritonavir+saquinavir salvage regimen. Drug sensitivity in this study was defined as <4-fold resistance using the Virco phenotypic resistance test. Harrigan found that CD4 count and viral load were factors in predicting the outcome. As expected, individuals with higher CD4 counts and lower viral loads were more likely to respond better to the salvage regimen. Most significantly--

If a person had AIDS, they had a 4-fold greater risk of failure

Phenotypic resistance predicted a 5.6 fold greater risk of failure

Genotypic resistance predicted a 3.5 fold greater risk of failure

Individuals with phenotypic resistance at baseline did not respond well to RTV+SQV, but the data suggests that for those who appeared to be phenotypically sensitive at baseline some of them appeared not to be responsive to therapy with a good viral load reduction. This reflects what we've already said--that phenotypic tests are better at detecting resistance than sensitivity

Baseline Nelfinavir Genotypic/Phenotypic Resistance in Expanded Access Program. Amy Patick reported at the Resistance meeting on 65 heavily pre-treated patients who received nelfinavir in the Agouron expanded access program. This data is from 24 weeks follow-up. They evaluated mutations 48V, 82A/F/T, 84V, and 90M because these are the mutations associated with nelfinavir cross-resistance.

Looking at genotypic mutations--

# of Genotypic mutations	0	1	2	3
Responders	12/14	4/5	2/8	1/4
p=0.02

12 out of 14 individuals with 0 mutations got a good response; 4 out 5 individuals with 1 mutation got a good response; while, individuals with 2 or more mutations did not respond well.

Looking at phenotypic resistance--

Phenotypic Resistance	<4-fold	4-10 fold	>10-fold
Responders	14/19	1/1	1/5
p=0.04

14 out of 19 individuals with <4-fold phenotypic resistance responded well to nelfinavir. So, both genotypic and phenotypic testing predicted success or failure. All these studies are retrospective studies. This means they looked back at the individuals blood samples and tested them for resistance after they had started the salvage regimens. Researchers agree that prospective studies are necessary to confirm these results. But, there is an increasing body of data suggesting that you may be able to predict the short-term viral load response a person will have to a drug using either genotypic or phenotypic test results. Perhaps the best predictor is --how many drugs in a new regimen is the person sensitive to?

Mayers discussed data he collected from patients treated with HAART in the military. These were patients failing to respond to HAART regimens. All were on a HAART regimen at the time of failure. On average these patients had a history of supposedly taking 5 NRTIs and 2 PIs. 22% of the patients tested had no resistance mutations. 24% were protease sensitive with some NRTI resistance. About 54% were NRTI and protease resistant. They found that many of these patients (22% without resistance) just could not take the regimens. The patients were unable to tell their doctors that they couldn't take the medications. Knowing which of these 3 categories a patient may be in is important to patient management. If a person does not have resistance but is not responding because they are not taking their medications, they might respond well to treatment if you could design a regimen that would be easier for them to take.

Mayers listed reasons why people can fail a regimen without resistance being detectable--

• Drug potency-- a drug may be so weak that a person could fail it without resistance developing
• Absorption-- the patient may not be absorbing a drug
• Metabolism--the person's metabolism may clear the drug too quickly
• Compliance--as stated above, the person may not be taking the drug
• Viral Virulence

In summary Mayers said--

• Drug resistance testing offers a technology to determine frequently or moderately occurring mutations. But is more limted in its ability to identify mutations that occur at low levels for people on aniretroviral therapy
• Can distinguish between drug resistant virus requiring salvage therapy from problems with adherence requiring a patient-friendly drug regimen
• After a period of time following the stopping of taking a drug, resistance may no longer be detectable. You don't want to test for resistance after a person has been off a drug for a period of time that might be as short as 3-4 weeks. Resistance can emerge again quickly after starting treatment. The best way to get useful information is to test for resistance while the person is still taking the failing regimen
• It is important to consider a person's drug history in the context of reviewing the results of resistance testing because of the reasons reviewed above
• the usefulness of using drug resistance mutations to select salvage regimens needs to be further validated in studies, many of which are now in progress
• At this point Mayers thinks that results from genotypic tests or phenotypic tests will furnish equivalent information, so either could be used. I'm not sure I agree with that. I feel more confident in phenotypic test predicting resistance than genotypic test results. But, I think the best way to try and select a regimen is to use both tests, although it can be expensive without insurance coverage.

Transmission of Drug Resistant Virus

AZT was first started to be used in1987, From 1988 to 1991 they discovered a 1.4% rate of transmission of AZT resistance from one person to another. By 1992 that rate increased to 7.5%. In 1993-1994 that rate increased to 10.4%. It's now holding at 7.5 to 10% in all patient populations that are being surveyed now. That means, 10% of people who are now becoming infected with HIV have AZT resistance before ever taking the drug. At Geneva, it was reported that--

• 7.5% of newly HIV infected individuals have AZT resistance
• 8% are resistant to 3TC
• NNRTIs are too new so there is no data yet
• 3 to 5% are protease inhibitor resistant

Individuals infected with virus resistant to Pis may have severely limited or no good treatment options. That is why it is crucial to use safer sex techniques. For individuals who take treatments, it is crucial that they realize that unsafe sex can leave a person not only infected with HIV but it may leave that person with no treatment options. If a person is treatment naive, it may be advisable to use resistance testing to check if they have pre-existing resistance that they may have contracted through transmission.

Dr Nick Hellman works for ViroLogic, the company that makes a phenotypic resistance test. He discussed at this forum a person who was infected with HIV that was resistant to delavridine and nevirapine. He described a second person who they performed a phenotypic resistance test on 23 days after the person was infected with HIV. There was evidence of decreased susceptibility to AZT, 3TC, saquinavir, nelfinavir, ritonavir and indinavir. Additional reports were presented of individuals infected through transmission with PI resistant virus.