AIDS Treatment Update, Issue 59, November 1997
Keith Alcorn

In AIDS Treatment Update issue 56/57, we questioned British doctors about the wisdom of establishing `undetectable' viral load as the main aim of treatment. Whilst they agreed that on theoretical grounds this was the ideal response to therapy, they saw many problems with this target in practice.

This month we look at `undetectable' viral load as the goal of treatment in more detail.

**What is `undetectable' viral load?

Viral load tests count the number of HIV particles in a sample such as blood plasma by looking for the virus' genetic material, RNA. The test result is described as the number of `copies' of HIV RNA per millilitre (copies/ml), and indicates the rate at which new HIV particles are being produced from infected cells.

There are several different tests currently in use, but each has a cut-off point below which it cannot reliably detect HIV. This point is called `the limit of detection', and its level varies from one test to another. The most commonly used viral load test, the Roche HIV-1 Amplicor test, has a lower limit of 400 copies/ml, which means that any sample with between zero and 399 copies/ml is described as `undetectable'. The Chiron Quantiplex test, which uses a different method to assess viral load, has a slightly higher detection limit of 500 copies/ml.

However, just because HIV cannot be detected, this does not mean it is not there. More sensitive tests which can measure viral load accurately down to the level of 20 copies/ml have now been developed. Studies show that many individuals whose viral load is `undetectable' using tests with a lower limit of 400 or 500 copies do have detectable viral load if tested with a more sensitive assay. For example, in a study of 72 people whose viral load was `undetectable' using a test with a lower limit of 500, just over two-thirds were found to have detectable viral load when using a more sensitive test with a lower limit of 50 copies/ml (abstract 51 at the International Workshop on HIV Drug Resistance, Treatment Strategies and Eradication, Florida, June 1997).

Furthermore, clinics only routinely measure viral load in blood samples, whereas the vast majority of the HIV in the body is in the lymphoid tissue. Several research studies have shown that viral load in these tissues is minimally affected by dual therapy with nucleoside analogue reverse transcriptase inhibitors (NRTIs) such as AZT plus ddI or 3TC, but usually falls dramatically during more potent triple therapy. Nevertheless, HIV replication may remain detectable in the lymphoid tissue long after viral load in blood has fallen below the limit of detection. Any interruptions in anti-HIV therapy or rebounds in viral load may lead to a rapid `re-seeding' of the lymphoid tissue with HIV, resetting lymphoid viral load to its pre-treatment levels.

Finally, standard viral load tests only count the number of HIV particles currently being produced from infected cells, and not the number of HIV-infected cells themselves (i.e. cells that contain HIV's genetic material integrated into the human DNA). These HIV-infected cells can persist long after viral load test results fall below the limit of detection. Although the viral DNA in many of these cells is thought to be defective, recent studies have shown that infectious HIV can be grown from some of them.

Suppressing viral load to below the limit of detection is thought to be a good thing because it is associated with two desirable outcomes in the short-term:

- a very low risk of disease progression

- a very low risk of developing resistance to your current drug regimen.

**Preventing disease progression

Several studies have compared the risks of disease progression among people at various low levels of viral load, to see whether it is necessary to reduce viral load to very low levels to avoid disease progression in the short-term.

A recent meeting at the US Food and Drug Administration (FDA) looking at HIV viral load heard an analysis of the results of a number of studies conducted by Glaxo Wellcome and the US AIDS Clinical Trials Group. It found that the lower the viral load after 24 weeks of anti-HIV therapy, the lower the risk of developing an AIDS-related illness during the first year of treatment. The maximum benefit was seen if viral load at 24 weeks was below 5000; however, because relatively few people reached this level it was difficult to assess any differences between viral load levels lower than this. Most of the people in these studies were receiving combinations of NRTIs alone.

An analysis of 1280 participants in the Delta trial comparing AZT, AZT/ddI and AZT/ddC explored the link between the lowest point that viral load reached during the first 16 weeks of treatment, and the subsequent risk of death (Florida, abstract 59). It showed that people whose viral load never dropped below 10,000 copies were 6.3 times more likely to die than those who went below 800, and twice as likely to die as those whose viral load fell to between 800 and 5000.

These studies looked at a `snapshot' of the viral load reduction at a certain point in time to see how well it predicted the risk of disease progression in the short-term, regardless of whether the viral load subsequently stayed at that level or increased. In the real world, however, people with HIV and their doctors usually monitor the viral load over time, and change treatment as necessary to try to keep it to a level with which they are comfortable (which may vary from person to person). Some recent studies have tried to assess what level of continuing viral load suppression is required if the risk of disease progression is to be kept to a minimum.

A German study of viral load and disease progression in 1593 people taking various anti-HIV therapies, presented at the recent 37th Interscience Congress on Antiviral Agents & Chemotherapy (ICAAC) in Toronto (abstract I-129), suggested that progression to AIDS is very rare when viral load is detectable, but below 5000. Out of 323 people with viral load below 5000, only 7 (2%) experienced an AIDS-defining illness. Conversely, of 195 people who did develop an AIDS-defining illness, 188 (96%) had viral load greater than 5000 despite anti-HIV therapy.

Similarly, a Spanish study of 444 people reported very low rates of disease progression so long as viral load was kept below 35,000 in sequential tests (ICAAC, abstract I-137). Among people whose viral load remained below 35,000, the probability of progression was only 1.2% after one year, 5.3% after two years, and 6.9% after 3 years. The one-year risk of disease progression was over 8 times higher for people whose viral load was not kept below 35,000.

**Prolonging the benefits

These studies suggest that it not be necessary to achieve `undetectable' viral load in order to keep your short-term risk of disease progression very low. You are very unlikely to experience new HIV-related conditions for as long as you can sustain your viral load at a very low but detectable level.

However, this raises a further question: what is the best way to maintain the benefits of a treatment regimen for as long as possible? On this point, recent research has found that the length of time it takes for resistance to a specific regimen to emerge, and/or for viral load to start to increase, is linked to how low your viral load is reduced. In other words, you may be most likely to achieve long-term benefits from an anti-HIV regimen if it suppresses your viral load to below detection.

One study looked at the duration of viral load suppression (defined as the time from starting treatment until the time when viral load rebounded by at least 0.6 log from its lowest point or `nadir') among people receiving combination therapy that included a protease inhibitor. It found that the people who achieved the lowest levels of viral load also achieved the longest duration of viral suppression (Florida, abstract 62). Among people whose lowest viral load was below 200, the average duration of suppression was 199 days. For those whose lowest viral load was between 200 and 1,000 the duration of suppression was 128 days, and for those whose viral load failed to decline below 1,000 the suppression lasted an average of only 60 days. This study suggests that if viral load is not reduced below 200 during protease inhibitor therapy, the chance that viral load will rebound in the short- to medium-term is quite high. This is why some doctors would recommend adding more drugs if viral load falls to decrease below 200 within a few months after starting a protease inhibitor-containing regimen.

The INCAS study of AZT/ddI/nevirapine demonstrated a similar link between the lowest level of viral load suppression achieved, and how long viral load remained suppressed to within 0.5 log of that lowest level. The average duration of suppression was 154 days for people whose viral load fell to below 20, compared with only 7 days for people whose viral load never fell below 20. The findings were presented at last month's Sixth European Conference on Clinical Aspects and Treatment of HIV Infection in Hamburg (abstract 105).

Why should there be such a link between the viral load nadir and the duration of suppression? The probable explanation is that the lower the viral load, the less likely it is that drug-resistant HIV strains will emerge, thus prolonging the effectiveness of the current regimen.

Nevertheless, we should be cautious about drawing the conclusion that viral load must always be suppressed below 20 to achieve a durable response. Although this was true in the INCAS study of nevirapine-containing combinations, it does not necessarily follow that the same will apply to other regimens. Nevirapine may be an unusual case because it is very easy for nevirapine-resistant HIV strains to emerge if a significant amount of viral replication is still taking place. It is plausible that with other drugs such as d4T, ddI and ddC (to which resistance is relatively rare and takes a relatively long time to develop), it may be possible to suppress viral load to low but still detectable levels and to maintain them for a long time. The same may apply be true of drugs which require HIV to develop a number of different mutations over time in order to gain a significant degree of resistance.

**Can `undetectable' be maintained?

Several studies presented at the Hamburg conference highlighted how difficult it is to achieve and maintain `undetectable' viral load using dual combinations of NRTIs, even for people whose viral load is relatively low when starting therapy.

A study from Vancouver showed that 67% of people who started dual NRTI therapy when their viral load was below 20,000 copies failed to suppress their viral load to below 500 and sustain it at that level (abstract 111).

Similarly, only 32% of people with viral load between 5,000 and 20,000 sustained a viral load below 400 for 48 weeks on AZT/3TC in six different studies analysed by Glaxo Wellcome (abstract 112). However, among people whose viral load was below 5,000 copies when they started treatment, 72% did achieve sustained viral load suppression over the 48 weeks.

Several studies have shown that the chance of achieving `undetectable' viral load is normally much greater with triple therapy regimens. However, even combinations that include a protease inhibitor may fail to achieve and maintain viral load suppression to below the limit of detection in a substantial proportion of cases, especially among people who have already had some anti-HIV drug treatment.

Doctors at San Francisco General Hospital recently reported that only 47% of people who had received at least six month's therapy with ritonavir or indinavir had viral load below 500 in their two most recent tests. People who started the protease inhibitor without changing any of their other drugs, as well as those with low CD4 counts or high viral loads, or documented problems adhering to therapy, were the least likely to experience and sustain `undetectable' viral load (35th Infectious Diseases Society of America meeting, San Francisco, September 1997, abstract LB-2).

Data from Professor Clive Loveday of the Royal Free Hospital in London also indicate just how difficult it can be to achieve `undetectable' viral load. Of the first 50 previously-untreated people who started on triple therapy at that clinic, after three months only 8 (16%) had achieved viral load below 400. Even during a further six months of triple therapy treatment, no more than 12 (26%) achieved `undetectable' viral load.

Although these results have not yet been analysed to distinguish between drug regimens, Professor Loveday says he believes that major problems with compliance to current regimens are likely to be a significant factor in the low attainment of `undetectable' viral load.

Professor Loveday also notes that the viruses which emerge after a rebound in viral load may be different from those which existed before treatment. They may be less fit and may replicate more slowly. They may also be less infectious, and less effective at killing CD4 cells.

**Practical implications

What does all this mean if you are thinking about starting treatment, or you are already on treatment and thinking about changing because you don't have `undetectable' viral load at the moment?

First, most authorities agree that in terms of delaying resistance and prolonging the benefit from a specific regimen, `undetectable' viral load is clearly better than viral load above 10,000 or even 5,000 copies. If you want to aim for this you should tailor your drug treatment accordingly. This means that unless your viral load is below 10,000 copies - a level at which few British doctors would recommend treatment in any case - you should consider a combination of (at least) three anti-HIV drugs, since you are unlikely to achieve `undetectable' viral load with a combination of two NRTI drugs alone.

If you adopt this strategy and succeed in achieving `undetectable' viral load, you may well decide to change treatment as soon as possible if your viral load increases to detectable levels again. The standard advice is to change to at least two new drugs, and to choose ones to which your virus is still likely to be susceptible even if it has developed resistance to your current regimen.

On the other hand, you may decide simply to try to keep your viral load in the range where disease progression is highly unlikely, which seems to be below 5000. If this is your strategy, you might still choose dual NRTI therapy with drugs to which resistance is slow to develop, such as d4T plus ddI, in order to preserve as many options as possible for the future.

This approach will reduce your risk of disease progression in the short- to medium-term, although perhaps not as much as if you use three drugs including a protease inhibitor. The risk with this strategy is that if your viral load remains detectable, you may well be at a greater risk of developing resistance to your chosen drugs, which may also reduce your treatment options in the future.

With either strategy, it seems clear that if you start treatment when your viral load is relatively low, you have a greater chance of suppressing it to very low or `undetectable' levels. If you delay starting treatment until your viral load is higher, you may find you need to take more drugs to reduce it to the extent that you want.

Likewise, if you have already taken several NRTIs and your viral load is currently high, you may need very aggressive therapy (such as combinations that include more than one protease inhibitor) to suppress your viral load to below the limit of detection. Several studies have highlighted poor responses among people who have taken several NRTIs (and thus might be expected to have developed some degree of resistance to these drugs) and then start a triple combination containing only a single protease inhibitor with NRTIs.

**Key conclusions

Different viral load tests have different lower limits of detection

Viral load that is `undetectable' with one test may be detectable with a more sensitive test

Disease progression is very unlikely while your viral load is below 5,000 during treatment

Suppressing viral load to below detection may minimise the risk of developing resistance to your current drugs

**Web links

Reports on the Florida resistance workshop and the ICAAC conference are available on the Internet at http://www.healthcg.com

The abstracts of research presented at the Hamburg conference are available on the Internet at http://www.euro-aids97.com

**BOX: The state of the eradication hypothesis

Over the last 18 months there has been much discussion of the possibility of eradicating HIV from the body completely after three to four years of highly active anti-retroviral therapy (see AIDS Treatment Update issue 44). This approach is based on calculations by Dr David Ho and studies monitoring the rates of viral load reductions among previously untreated people who started triple or quadruple anti-HIV therapy very soon after they became infected.

However, the time scale calculated by Ho has been questioned by many researchers, and recent evidence suggests that HIV could persist in a latent form in some cells for more than ten years, if not for life.

HIV's genetic material, in a form that appears to be capable of producing new, infectious HIV particles, can be found integrated into the DNA of infected but resting (i.e. dormant) memory CD4 cells after as much as 30 months of triple or quadruple therapy. Dr Ho had previously been unable to stimulate this HIV proviral DNA to produce new HIV copies, and he argued that it represented defective HIV genes that he called `dead archival material'. However, in a recent study researchers at Johns Hopkins University have been able to stimulate HIV production from this dormant genetic material. This research suggests that a tiny reservoir of genetic material could persist despite many years of therapy, providing the basis for a resurgence in HIV levels if treatment was stopped.

In Hamburg last month, Dr Ho himself reported that patients treated with triple therapy for up to 28 months at his clinic still showed evidence of extremely low levels of virus replication. He suggested that infected cells might still be available to support a renewed burst of viral replication if therapy was stopped, although it was unlikely that they would all begin virus production at the same time.

Other researchers continue to raise doubts about the ability of drugs to clear HIV from brain tissue, even though a range of drugs, including a combination of ritonavir and saquinavir, have been shown to clear HIV from the cerebrospinal fluid (CSF) that bathes the spinal cord and brain. The brain tissue may be the last `sanctuary site' for HIV in the body, and at a Hamburg satellite symposium sponsored by Glaxo-Wellcome, Professor Peter Portegies of Amsterdam University argued that AZT was the only drug which has been shown to reduce HIV levels in brain tissue as well as in the CSF.

What is unclear is whether anti-retroviral therapy will need to persist until every last infected cell can be eradicated from the body - if this is ever possible - or whether at some point the immune system may recover enough to control these tiny remnants of infection on its own.

971101
ATU5901

Copyright © 1997 - AIDS Treatment Update. Permission granted for noncommercial reproduction, provided that our address and phone number are included if more than short quotations are used. Subscription lists are kept confidential. NAM Publications 16a Clapham Common Southside, London, England SW4 7AB; TEL: 01-71-627-3200 (from outside the UK: +44-171-627-3200); FAX: 01-71-627-3101 (from outside the UK: +44=171-627-3101)  info@nam.org.uk  http://www.nam.org.uk