Testing gains shown in antiretroviral resistance

Research addresses new recommendations

Resistance testing has proved useful in treating some HIV patients, but it’s expense and inconsistent value in predicting the future course of HIV disease have left clinicians with some doubts about when and how to best use it.

As published reports continue to suggest HIV drug resistance is increasing and may affect both an individual’s response to antiretroviral treatment as well as an HIV population’s clinical outcomes, the decision on whether to use a resistance test in any particular case grows more complicated.

"In some cases, the suggestion is that resistance, in fact, has an important clinical outcome impact in that people who have acquired viruses with more resistance seem not to respond as well to treatment," says Chuck Hicks, MD, associate professor of medicine at the Duke University Medical Center in Durham, NC.

Hicks and colleagues have studied HIV anti-retroviral resistance and the use of resistance testing in clinical treatment, addressing the 2002 resistance-testing guidelines.

The 2002 Antiretroviral Therapy Guidelines, published by the U.S. Department of Health and Human Services (DHHS), says viral-resistance testing helps doctors choose the most effective drugs.

It recommends resistance testing when a patient’s viral load is not controlled by new medications or when there is viral breakthrough during a regimen that has been working.

Resistance testing revisions

The latest guidelines, published by DHHS July 14, have a revised section on resistance testing. The new guidelines discuss in detail the limitations of genotyping and phenotyping assays, including the lack of uniform quality assurance for all available assays, the high cost, and insensitivity for minor viral species. The new guidelines state when therapy is about to be initiated in someone with acute HIV infection, resistance testing at baseline could optimize virologic response, although the strategy is untested.

The 2002 guidelines suggests that conducting resistance testing before treating a newly infected HIV patient is a reasonable clinical practice.

"What we wondered was: Are the kinds of resistance reported from San Francisco, San Diego, and New York City similar in our part of the country [ North Carolina]?" Hicks says. "And also, does this resistance impact how they respond to treatment?"

North Carolina investigators found that rates of antiretroviral resistance were quite uncommon in the population they studied, and the resistance that did exist seemed to have no impact on clinical outcomes under highly active antiretroviral therapy (HAART), Hicks says.

Out of a total of 31 patients with primary HIV infection, genotypic resistance testing showed there were two cases of resistance to nucleoside reverse transcriptase inhibitors (NRTIs) and one case of resistance to non-nucleoside reverse transcriptase inhibitors (NNRTIs). Phenotypic resistance testing indicated two cases of resistance to NRTIs, two cases of resistance to protease inhibitors, and three cases of resistance to NNRTIs.1

"In the most generous form of reporting resistance we found, there might be up to five patients who had some resistance to drugs they were prescribed for acute infection," Hicks says. "Four out of five patients achieved viral loads of less than 50 copies, so they had very good response to therapy; and one patient — who had the greatest amount of resistance — did not achieve less than 50 copies."

Investigators concluded that transmitted resistance among Southeastern U.S. patients is relatively infrequent and is an imperfect predictor of treatment response.1

Another difference between what the North Carolina study found and what has been published previously involved the time period of resistance.

"Other studies report increased rates of resistance over time; but in our study, the opposite was true," Hicks says. "Those diagnosed between 1998 and 1999 were more likely to have resistance than those diagnosed after 2000, so that’s a little bit different."

A possible explanation for the difference is that the study’s population was so small that it was just a quirk, he suggests.

"Another potential explanation could be that the population of people who were among gay white men who acquired it from other gay white men may be more likely to have received therapy and more likely to pass on resistant viruses," Hicks explains. "Our patients were not as likely to have acquired their infection from a population that has been treated a lot in the past, and so there perhaps was less potential to pass on resistance."

The issue of how relevant resistance testing is to one particular type of HIV population vs. others only is one of the challenges clinicians face when deciding whether to order a resistance test. Another difficulty involves the accuracy of the test itself.

Another study published this year investigated whether a lab generates the same real-time results in genotyping for resistance testing when duplicate clinical samples are examined.2

"Basically, I think genotyping is a fairly accurate assay, but there are instances where you will get discrepant data coming back if you had the opportunity to send replicate samples, and most people don’t," says Diana Huang, PhD, assistant professor at the Rush Medical College in Chicago.

Resistance testing typically is a brief look at data from one point in time, she says.

Therefore, Chicago investigators decided to submit replicate samples to genotyping labs to see how consistent the reporting was, Huang explains. "Generally, the labs were pretty good at being able to get similar results for both the replicate samples and the entire panel; but every now and then, you’d get discordant results," she points out.

The discordant results could be attributed to a number of different causes, including the way a codon is read, which might result in an interpretation of a virus being wild type or containing a resistance mutation, Huang says.

Also, inherent within the sample itself are genomes that represent mixed populations of virus, and the person who is doing the genotyping assay will have to determine what is real and what isn’t, she adds.

"You can have situations where the performance of the assay itself may select out different populations to be predominant, and that could affect the identification of a resistance mutation in the sample," Huang says. "How the call is made may be based on proportionate populations in the sample."

Also, there could be technical and mechanical problems earlier in the assay that affect later steps of the assay, which generate the identification of resistance mutations, she adds.

Second samples

All of this suggests that physicians need to be careful when interpreting genotyping results and, occasionally, send in a second sample, Huang says. "The discrepancies occur, but they are not that common. I don’t want people to think it’s that widely prevalent, but occasionally, you will see this."

Although the latest HIV treatment guidelines suggest that antiretroviral resistance testing has some merit as part of baseline treatment of newly infected HIV patients, some recent research offers clues to its usefulness as a screening tool.

While physicians are divided on the issue of when to begin antiretroviral treatment, one outcome of this decision may be an increase or decrease in drug-resistant virus.

"It is in the interest of patients to have a genotypic profile done at the time of primary infection, because if you are infected with drug-resistant virus, it’s going to disappear with time," says Luc Perrin, MD, professor of clinical virology at Geneva University in Switzerland.

"In two to three years, the virus will be back to the wild type, but the one type of resistance is still there, so you won’t be able to detect it several years later," Perrin adds. "So you will treat patients and not take into account that in the reservoir, there is drug resistance and the response is going to be inefficient." The optimal time for the initial drug-resistance test is within the first six months, before antiretroviral treatment and at the time of primary HIV infection, Perrin says.

Perrin and other investigators from Switzerland found a 10.5% prevalence of drug resistance in recently infected Swiss patients, and they further discovered that about 30% of the recently infected patients were infected by other recently infected patients.3

"By making a calculation to find out the genetic profile of a potential transmitter, and by comparing the genetic profile of a potential transmitter to the genetic profile of acutely and recently infected HIV patients, you find out there’s a lot less Class 2 resistance than expected," Perrin says.

The study concluded that multidrug-resistant HIV-1 has a decreased transmission capacity, which is an explanation for the reduced spreading of variants in newly infected patients.3

However, the problem with that strategy is that the resistance tests are very expensive in the United States, where HIV patients do not have the advantage of a universal health care system.

So again, clinicians may find it impractical to routinely order resistance testing screening and will have to continue to make this decision on a per-patient, per-case basis.

"There are two main groups of people for whom resistance testing is potentially useful," Hicks says. "One is the group of people on therapy who are failing treatment and you’re trying to decide what to do next, and that’s someone you know is HIV-positive, who has been started on therapy and has not been as successful as you’d like."

The other group involves people who have never been on treatment, including those who are acutely infected or who have been infected for a year or two, he says. "Resistance testing could be helpful." And if resistance testing costs $5 per sample, then it would make sense to offer it routinely in the case of the newly or recently infected patients, Hicks adds.

"But it doesn’t cost $5, and the amount of money available to care for HIV patients is not infinite," he says. "So it would be hard to argue that it’s cost-effective to do resistance testing when there’s no acute infection."

Perhaps in areas where HIV drug resistance is known to be high, such as in San Francisco, a clinician could justify the cost of doing routine resistance testing of newly infected patients, Hicks suggests.

"By contrast, in North Carolina, it appears from an undersized sample that drug resistance is a much less common phenomenon. In what settings does the information, which is of excellent quality, really impact decision making that’s of benefit to the patient?" he asks.

"And as we study the question of what kind of patient will benefit from what kind of resistance testing, then cost-effectiveness arguments fade into the background," Hicks adds.


1. Hicks C, Eron J, Lennox J, et al. Antiretroviral resistance among patients with primary HIV infection in the South- eastern U.S. — impact on treatment outcome. Presented at the 10th Conference on Retroviruses and Opportunistic Infections. Boston; February 2003. Poster 502.

2. Huang D, Brambilla D, Ouma A, et al. Variability in quality assessment data from replicate genotyping of clinical samples. Presented at the 10th Conference on Retroviruses and Opportunistic Infections. Boston; February 2003. Poster 584.

3. Yerly S, Jost S, Telenti A, et al. Are drug-resistant variants transmitted with lower efficiency than wild-type? Presented at the 10th Conference on Retroviruses and Opportunistic Infections. Boston; February 2003. Abstract 145.