GMHC Treatment Issues, Volume 10, Number 6/7 - June/July 1996
Theo Smart

According to its FDA-approved labeling, the Roche viral load test can be used to predict the risk of disease progression in people with HIV. The FDA also allowed mention on the label that the test has been used as an aid in assessing the activity of antiretroviral therapy, and in fact, that changes in viral load as measured by the test contributed to the approval of 3TC and the protease inhibitors. But the agency shied clear of endorsing the use of the test for patient monitoring by insisting on the following caveat: "the clinical significance of changes in HIV RNA measurements has not been fully established although several large studies that will more fully determine the role of comparative HIV RNA measurements in patient management are now in progress."

In contrast to the FDA, the IAS panel of experts believes that there are enough available data to use the test not only to determine the risk of disease progression, but to decide when to initiate therapy, whether a new treatment is having an effect and when to switch treatments because the current treatment is failing. Both the FDA and the IAS recommend that the viral load tests should be used in conjunction with other surrogate markers, in particular CD4 cell counts, which, the IAS panel said, remain, "the best predictor for the risk of developing an AIDS-related complication." High viral loads may predict rapid progression (see below) but it is the loss of CD4 cells that puts one in imminent danger of contracting PCP or another opportunistic infection.

One Virus, Three Ways to Count It

One of the FDA Blood Products Advisory Committee's chief stumbling blocks when reviewing Roche's application last March (see page twelve of the March Treatment Issues), was that many of the data supporting the use of viral load were generated using different assays (and blood samples processed several different ways).

The three available tests do indeed work differently:

* In Roche's PCR assay, plasma containing HIV RNA is exposed to an enzyme that converts the RNA into DNA. Repeated polymerase chain reactions (PCR) clone millions of copies of the DNA particles, which bind to primers (molecules designed to form a strong bond with the targeted genetic material) on a plate. The plates are then washed with a solution containing colored "tags" that stick to the DNA and are subsequently counted by a machine. The test measures from 400 copies to 800,000 copies of HIV RNA per milliliter of blood plasma.

* Chiron's "bDNA" test, in contrast, first captures the target HIV RNA with primers on a plate. The plate is washed with a solution containing "branched" DNA molecules (the bDNA), which bind to the HIV RNA. There is no cloning stage, rather the signal from the bound bDNA is amplified. Each bDNA molecule has multiple sites (the branches) for an alkaline phosphatase label capable of generating light in the presence of another reagent. The light emitted by the bDNA molecules is counted by a machine. The second generation bDNA assay can detect HIV RNA or DNA or other genetic targets within a range of 500 to 1,600,000 copies/ml.

* Lastly, Organon Tecknika's NASBA test (Nucleic Acid Sequence-Based Amplification) is similar to the Roche PCR assay and can detect between 400 and 5,000,000 copies of HIV RNA. In contrast to the other assays, heparin, an anticoagulant commonly added to blood specimens when extracting plasma, can be used in the initial processing of the blood sample before freezing. (Heparin added to samples tested with Roche's or Chiron's assay yield RNA counts around a third lower than they would if another anticoagulant had been used.)

The FDA was considering only the Roche viral load assay this spring, with Chiron's version now undergoing separate review. The IAS panel recommendations apply to all three viral load tests regardless of FDA status. While the tests are not standardized with each other, a number of studies have shown that they tally up very similar viral loads from the same plasma sample. Though the FDA's Blood Products Advisory Committee hesitated to extrapolate data generated by the bDNA or NASBA test to use of the Roche test, the IAS panel treated them as equivalent. "The IAS guidelines are guesswork by a bunch of people working with incomplete data, while the FDA was being more cautious," stated Robert Combs, M.D., a member of the IAS panel.

By necessity, the IAS group considered together data produced by all three tests. Yet there are slight variations between these tests, and, for reasons of consistency, the panel advised that people continue with the assay used to establish their initial value.

Vaccinations (against the flu, hepatitis, tetanus and pneumococcus) have been shown to cause a significant rise in viral load that can persist for weeks after the immunization. Intercurrent herpes outbreaks and opportunistic infections can boost viral loads even more. The IAS panel recommends that viral load should not be measured until one month after immunizations or acute illnesses, but some studies suggest that vaccination induced increases in viral load can persist for a longer period.

Each company's tests are subject to a certain amount of technical variability which, according to one study, may increase substantially for measurements at the lower limit of detection (see Treatment Issues, September 1995, pages 12- 14). There also may be a degree of biologic variability unrelated to major infections or vaccinations. Together, these variations seldom exceed 0.5 log (a three-fold difference). To be on the safe side, Dr. Coombs, who is one of the leading experts on viral load, has said, "You always need two baselines [at least one confirmatory test] for making clinical decisions."

Prognosis

The FDA and the IAS panel agreed that data from several cohorts show viral load to be a good prognostic indicator. Treatment Issues has covered this application of viral load extensively in the past (see November, 1994, pages 4-6, 10- 11; December 1995, pages 1, 3-4; February 1996, pages 1, 12- 16) In summary, the NIH, the Aaron Diamond Institute, researchers studying disease course in the MACS cohort and others have found that viral load several months after seroconversion, and at any timepoint thereafter, can be used to determine whether the individual will have a rapid, moderate or slow rate of disease progression.

There is a range of opinion on what precise figures to use for predicting disease outcome. The FDA cited data from trials ACTG 116A and 116B/117 (AZT vs. ddI studies), where the frequency of disease progression within five years exceeded 60 percent in patients with viral loads over 250,000 copies/ml. Progression was rare in those with fewer than 11,000 copies/ml unless they also had very low CD4 cells counts. This observation underscores the need to consider viral load in conjunction with CD4 count. The disadvantage of the two ACTG trials is they say little about the use of viral load for prognosis beyond five years.

The IAS guidelines also refer to recently published data (using the bDNA assay) from 180 patients from the MACS cohort with a follow-up period of up to 11.2 years. In his analysis, John Mellors, M.D., determined that time to progression within ten years in patients with more than 500 CD4 cells was greater than 70 percent if their viral load was above 10,200 copies/ml.2 But Dr. Mellors retrospectively tested frozen samples of heparinized blood. As noted above, heparinization reduces measurable viral load by more than one third, and the effect of long-term storage on further diminishing observed viral load is unclear.

Monitoring the Effect of Treatment

Pivotal studies testing AZT/3TC, ritonavir, and indinavir and NIH-sponsored studies ACTG 229 (testing saquinavir/ddC, saquinavir/AZT, and all three together), ACTG 175 (AZT, ddI, AZT/ddI and AZT/ddC), ACTG 116A, 116B/117 and VA 298 (early versus deferred AZT) all found that viral load tests indicate when a drug is having an antiviral effect. Even when the relatively impotent AZT is used, there is a very rapid drop in viral load within the first several days. This has been seen with every approved antiviral tested. Furthermore, rebounds in viral load back to baseline levels are correlated with loss of CD4 cells and clinical deterioration in a number of studies. Although mentioned in the Roche test's labeling, it is hard to understand why the FDA could not explicitly approve the use of the test to demonstrate whether a new drug is working. The proposition that antiretroviral therapies could have their effect by any mechanism other than anti-HIV activity seems straight out of Lewis Carroll.

If an antiretroviral drug does not reduce the level of HIV, why take it? The FDA has implied that it does not really intend to constrain frequent monitoring of viral load while on therapy. In the FDA's internal briefing paper ("Talk Paper") on the Roche test, the agency employs a broad definition of prognosis, noting that in ACTG 116B/117, a five-fold increase in viral load eight weeks after beginning therapy was "prognostic of progression." If this ACTG finding comes under the "prognosis" rubric, one might then be justified in checking viral load every other month while on therapy to make certain that it has not gone up.

To the FDA's credit, treatment should improve one's physical condition, not merely ameliorate a laboratory marker like viral load. Although therapeutically induced reductions in viral load are generally associated with increases in CD4 cell counts and, in some studies, with a reduction in opportunistic conditions and death, the data do not show that the greatest reductions in viral load consistently produce the greatest physical or CD4 cell improvements. For example, in ACTG 229, AZT/ddC effected the largest reduction of viral load, but AZT/saquinavir produced a better rise in CD4 cell counts. One factor could be the toxicity of the nucleoside analog ddC compared to the protease inhibitor saquinavir -- one-third of the volunteers on AZT/ddC had reduced CD4 counts even as their viral load went down, compared to one-tenth of those in the AZT/saquinavir arm. Dr. Coombs commented, "There might be other ways a protease inhibitor improves CD4 count. Viral load didn't explain all the treatment effect of the therapy."

The rate of CD4 decline after a rebound in viral load also appears related to the particular drugs being administered and not merely to the increase in viral load itself. Preservation of CD4 cell count gains have been reported in protease inhibitor studies even as viral load has returned to baseline.

In ACTG 175, the AZT/ddI and AZT/ddC arms reduced viral burden more than the ddI arm, but the volunteers on ddI alone did at least as well physically as those on the more potent combinations. At the advisory committee meeting in March, representatives of the FDA noted that such discrepancies made it difficult to establish algorithms to aid physicians employing viral load to manage individual patients. Such algorithms should be established by a number of ongoing or planned studies that are designed to show, regardless of the therapy used, that it is the reduction in viral load that slows disease progression. But it is unlikely that equally potent regimens will always have the same clinical effect since their differing toxicity can impair responses.

The guidelines from IAS panel (which do not mention many of the conflicting data) recommend a general algorithm based nearly as much upon theory as clinical data. The IAS panel believes that the goal of therapy should be to reduce viral load levels to below the limit of detection, or at least below 5,000 copies/ml. This desirable goal may not be achievable in all patients. There is a danger that while attempting to reach the 5,000 copy level, physicians will rapidly shift through their entire therapeutic arsenal and create HIV resistant to all available agents. Less dramatic reductions in viral burden may stabilize a patient and stretch out the available therapies so that they can be relied on when symptoms reflecting serious immune deficiency appear. But there are no data at present to support either the "hard and early" or the "soft and early" approach.

The odds are that there is less of a chance for HIV to become resistant to a drug when there is less of it replicating. Even reducing viral load to undetectable levels may not be enough to prevent eventual treatment failure, though. At the lower limit of detection for the present viral load tests (around 500 HIV RNA copies/ml), the virus may still be replicating in the blood, not to mention viral reservoirs in the lymph nodes, where HIV concentrations are much higher than in the blood, or in other organs including the brain that drugs may not reach well. As long as the virus can reproduce, there is a chance of drug resistance emerging.

Initiating Therapy

The most controversial aspect of the IAS viral load recommendations is the use of the test to decide when to initiate therapy. The panel concluded that practitioners should consider treatment of individuals with viral loads greater than 5,000 to 10,000 copies/ml and definitely should treat those with viral loads in excess of 30,000 to 50,000 because such individuals are supposed to be in immediate danger of progression.

The viral load ranges given indicate that the panel members could not settle on a single absolute value that predicted progression. But long-term prognosis is not the only factor to consider when initiating treatment -- especially when there is a potential to exhaust the available therapies and no clear evidence that treating low levels of viral load early in the course of disease has any effect on progression.

There is a wide range of opinion among clinicians in our survey regarding when to treat on the basis of viral load. Most respondents still rely on a combination of symptoms and lab tests to initiate treatment. They usually believe therapy should drive viral load down below at least 10,000 copies/ml to achieve disease stability.

Early intervention: David Ho, M.D., has written that the establishment of a viral load set-point after the initial acute phase of HIV infection provides a rationale for early intervention -- that if you can lower the viral load during this critical period, you may be able to effect a long-term delay in disease progression.3 One six-month study of AZT in early patients with primary HIV infection found that those who received treatment were less likely to develop "soft" clinical endpoints, such as thrush, than those who received no treatment.4 This is not resounding proof that treatment at this stage of disease will have any lasting impact, but Dr. Ho is conducting several studies with more potent antiviral regimens that should provide clearer evidence of benefit.

Other Factors in Treatment Decisions

The FDA approval of the Roche viral load test is a watershed event that will increase access to the tests in the U.S. But again, viral load is not the only factor to take into consideration when making treatment decisions. Some strains of the virus may be more dangerous than others, and there are suggestions that this may be the case for AZT-resistant strains. Conversely, if drug resistance one day breeds a less pathogenic virus, it might be missed if people are quickly shuffled to a different therapy just because their viral load is once again on the rise.

Dr. Coombs noted, "I think there is a host range in which patients contain the virus. There is no evidence of clinical benefit to driving the virus population very much lower." Everyone's immune system is different, though. Some people may be able to sustain higher viral loads without progressing or, alternatively, progress at much lower viral loads.

In ACTG 116A and 116B/117, disease progression continued in a number of advanced patients with low viral loads and in patients whose viral load was reduced to low levels in response to therapy. If viral load and CD4 cell counts both decline on therapy, is that treatment adequate? Clearly no. Finally, starting therapy in asymptomatic patients with low viral loads who are not prepared to comply with arduous therapy may be irresponsible or downright dangerous since there is always the potential of exhausting therapeutic options by breeding drug-resistant HIV. Given the dangers involved, initiating and switching therapy are matters that must be carefully weighed using all available pertinent information. Viral load is but one of a constellation of tools to help people make those choices, but no single surrogate marker can dictate all treatment decisions.

1 Saag MS. et al. Nature Medicine. June 1996; 2(6):625- 629.

2 Mellors JW et al. Science. May 24, 1996; 272(5265):1167- 1170.

3 Ho D. The New England Journal of Medicine. August 17, 1995; 333(7):450-1.

4 Kinloch De Loæs S et al. The New England Journal of Medicine. August 17, 1995; 333(7):408-13.

Table 1. Summary of the IAS Interim Recommendations

Parameter	Recommendations
Plasma HIV RNA level	More than 5,000-10,000 copies/ml and
that suggests initiation	a CD4 cell count/clinical status
of treatment.	suggestive of progression; >30,000-
50,0	00 regardless of laboratory/
clin	ical status.
Target HIV RNA level after	Undetectable; <5,000 copies/ml
initiation of treatment	is acceptable.
Minimal decrease in HIV RNA	>0.5 log decrease.
indicative of antiviral
activity
Change in HIV RNA that	Return to (or within 0.3 to 0.5 log
suggests drug treatment	of) pretreatment value.
failure
Suggested frequency of	* At baseline: two measurements,
HIV RNA measurement	2 to 4 weeks apart.
* Ev	ery 3 to 4 months or in
co	njunction with CD4 counts.
* Sh	orter intervals as critical
deci	sion points are neared.
* 3	to 4 weeks after
in	itiating/changing

960701
GM100707

Copyright © 1996 - Treatment Issues. Reproduced with permission. Treatment Issues is published twelve times yearly by GMHC, Inc. All rights reserved. Noncommercial reproduction is encouraged. Subscription lists are kept confidential. GMHC Treatment Issues, The Tisch Building, 119 West 24th Street, New York, NY 10011  fredg@gmhc.org  http://www.gmhc.org

AEGiS is made possible through unrestricted grants from Boehringer Ingelheim, Elton John AIDS Foundation, iMetrikus, Inc., John M. Lloyd Foundation, the National Library of Medicine, and donations from users like you. Always watch for outdated information. This article first appeared in 2003. This material is designed to support, not replace, the relationship that exists between you and your doctor.

AEGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.

Copyright ©1980, 2003. AEGiS. All materials appearing on AEGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of AEGiS, or the party credited as the provider of the content. .