The following sections look at how viral load results are used to predict the effectiveness of combination therapy. Further information of relevance can be found in Anti-HIV therapy: Choosing your treatment strategy.

In addition to its use to predict the prognosis of a currently untreated individual, viral load has become the most widely used marker for monitoring the effect of anti-HIV treatments. Studies have examined two questions:

• whether a single measurement of the viral load of people taking anti-HIV therapy can predict their risk of disease progression

• whether changes in viral load soon after taking an anti-HIV regimen can predict the longer-term effects of the drug(s) on disease progression and survival.

Studies of the first question have produced contradictory results. Researchers examined whether the initial viral load of participants in the American trial ACTG 116A predicted their prognosis. They found that the baseline viral load level was a good predictor for people who had never taken anti-HIV treatments. But for participants who had taken AZT before - all of whom had taken it for less than four months - a 'snapshot' measurement of their viral load on entering the study could no longer predict their clinical course. These researchers suggested that the predictive value of a single measurement of viral load may be lost if you have already taken anti-HIV treatments, even for a short time.

By contrast, another team of researchers have studied 620 people enrolled in two large American trials comparing AZT monotherapy with combination therapy with AZT plus ddC or 3TC. They wanted to know whether a participant's most recent viral load measurement predicted his/her risk of disease progression. Their results suggested that it did: people with higher viral load were at increased risk of progression. This team found that CD4 count measurements were also useful, and that the best way of predicting an individual's risk of developing AIDS was to monitor both tests.

Changes in viral load in response to treatment

There is now a large body of evidence that suggests changes in viral load in response to anti-HIV drugs can give an indication of the likely long-term effectiveness of the treatment. These data include virological sub-studies of three trials which showed clinical benefits to drug therapy (Delta, ACTG 175 and CPCRA 007), all of which compared AZT monotherapy versus the combination of AZT with either ddI or ddC.

In each of these studies the regimen which showed the greatest clinical benefits in long-term follow-up also had the greatest short-term effect on viral load. Among previously untreated people in Delta, AZT plus ddI was the most effective at delaying disease progression and death, and it was this regimen that showed the greatest reduction in viral load after eight weeks of treatment (Brun-Vezinet). A one log reduction in viral load after eight weeks of treatment was associated with a 50% reduced risk of developing AIDS over 96 weeks of follow-up, and a similar degree of risk reduction was seen in the study ACTG 241. For every one log reduction at eight weeks there was a 52% reduction in the risk of disease progression over 48 weeks. For further details of research linking viral load reduction to delayed disease progression, see the Key Research section below.

In presentations to the US Food and Drug Administration as part of a hearing to investigate the relationship between viral load changes and clinical outcome in recent HIV drug trials, the panel heard:

• The length of time that viral load remains suppressed below 5,000 copies correlates with the length of time for which disease progression can be delayed.

• Suppressing viral load beneath 400 copies resulted in a reduced risk of disease progression compared with lowering viral load to between 400 and 5,000 copies, but this in itself reduced the risk of disease progression compared to a lesser reduction to above 5,000 copies.

• Viral load at baseline (before commencing treatment) was a very strong predictor of the risk of disease progression, regardless of response to treatment.

Early monitoring

Monitoring viral load in the first few months after commencing an anti-HIV combination is used to assess the effectiveness of therapy and adherence to the regimen. However, as the studies below indicate, a very early viral load response to treatment is not necessarily a reliable guideline to a medium or long-term response.

Some experts are doubtful that very early results are a reliable indicator of response to treatment. John Mellors told the Seventh Conference on Retroviruses in early 2000 that it was very difficult to predict long-term response when looking at viral load measured eight weeks after starting treatment. Mellors cited a study of treatment-naive individuals who failed to reach undetectable viral load; only 3% of these individuals could be identified by an abnormally slow response during the first four weeks of treatment. Poor performance at week eight was not predictive of the subsequent viral response (Huang). However, viral response of highly treatment-experienced individuals during the first four weeks of treatment was associated with medium-term response.

Current US guidelines define an expected response to treatment as a two log drop in viral load within four weeks. However, a failure to respond to HAART with a two log drop within four weeks doesn't necessarily indicate total treatment failure, according to a study of 335 individuals starting HAART at the Frankfurt UNI Clinic. Viral load was still detectable in the majority of patients at week 4. However by week eight, 70% of patients were responders, and by week 16, 80% were responders (Cozzi-Lepri).

The Frankfurt study suggested that viral load at week 8 is a more reliable guide to subsequent response. In those with viral load below 10,000 copies but still detectable at week eight, the chances of achieving undetectable viral load by week 16 diminished. Only 38% of those who remained above 10,000 copies at this point went on to reach undetectable viral load.

Baseline viral load is a key factor in determining the predictive value of early viral load results. Eleven out of 14 people with viral load between 100,000 copies and 1 million copies achieved undetectable viral load at week 16 despite failing to experience a 2 log drop in viral load within four weeks of starting treatment.

The viral load nadir

Several studies have suggested that one of the most important predictors of how long the benefits of an antiviral regimen will last is the nadir of the response, i.e. how low the viral load is driven by treatment.

A virological analysis of 1280 Delta participants showed that those who achieved viral load below 800 copies had a 3.1 fold lower risk of death compared with those who did not achieve viral load below 5,000 copies (Aboulker). However, the risk of disease progression was less profound between those who fell to between 800 and 5,000 copies (2.2 fold), and those who fell to between 5,000 and 10,000 (3.1 fold).

An analysis of 773 French patients found that people on anti-retroviral treatment who failed to maintain viral load below 5,000 copies were almost two and a half times more likely to become ill or die within 12 months than people who maintained viral load below 500 copies. Maintaining a viral load below 5,000 copies - but above 500 copies - didn't substantially increase the risk of becoming ill.

Every one log reduction in viral load after one month of treatment was associated with a 50% reduction in the risk of illness, according to researchers from the INSERM research centre in Bordeaux. On the other hand, a viral load reduction of less than 0.5 log regardless of baseline viral load by week four was strongly associated with a poor long-term response (Thiebaut).

A study of people receiving protease inhibitor therapy (either as monotherapy, in combination with nucleoside analogues or dual protease inhibitor combinations) found that the duration of viral load suppression (defined as the time from starting treatment until the point when viral load rebounded by at least 0.6 log from its lowest value) was predicted by the nadir that viral load reached on treatment (Kempf). Among people whose viral load was suppressed below 200, the average duration of response was 199 days. For those whose lowest viral load was between 200 and 1000, the duration of response was 128 days, and for people whose viral load failed to decline below 1000 the response lasted an average of 60 days.

Thus, if the viral load is not reduced below 200 copies, the risk of virological failure in the short-to-medium-term is quite high.

The lower the better

More recently, it has become clear that suppressing viral load below 50 or 25 copies may be necessary to block new rounds of viral replication and the eventual emergence of resistance.

Researchers examined virus from people with viral load consistently below 50 copies for at least two years, and found no evidence of viral evolution that would suggest ongoing replication of new generations of HIV. In comparison an individual with viral load just above 50 copies showed evidence of viral evolution that might eventually pave the way for the emergence of resistance (Wong). Furthermore, a study of 168 people with detectable virus below 400 copies found that 52 had viral rebound above 400 copies after 25 weeks follow-up (Pilcher 1999).

However, another study found no clear correlation between depth of suppression and viral evolution. Whilst one individual with viral load consistently below 20 copies showed evidence of viral evolution, two other patients with viral load in the range 20 to 200 copies for 14 and 23 months respectively showed no evidence of viral evolution (Martinez).

The INCAS study of AZT/ddI/nevirapine in people new to treatment also demonstrated that the nadir response predicted the length of time it would take for viral load to rebound (at least over the 52 weeks for which the trial ran). The risk of virologic failure (defined as viral load above 500 copies, or increase by half a log, or returning to its pre-treatment value), was almost ten times lower in those whose viral load nadir was below 20 copies (Raboud).

See Suppressing viral load as low as possible? in Anti-HIV therapy: Choosing your treatment strategy

The speed of viral load decline

The rate at which a new drug combination clears virus from the blood has also been proposed as a potential predictor of the durability of the regimen. The assumption is that the speed of the fall in viral load (also called the 'decay rate') is a function of the potency of the regimen, and that a rapid fall in viral load is most likely to delay the emergence of resistance.

The ADAM study was one of the first to report on this relationship. It evaluated an induction-maintenance strategy and found that those participants who had the most rapid fall in viral load during four-drug induction were most likely to maintain viral suppression when de-intensifying to two drugs (Hoetelmans).

More data on this subject can be found in Suppressing viral load as fast as possible? in Anti-HIV therapy: Choosing your treatment strategy.

Gender differences in the speed of response to HAART have also been observed. A review of 378 patients at the Royal Free Hospital in London found that women were significantly more likely to have undetectable viral load (below 500 copies) within 16 weeks of starting treatment after controlling for baseline viral load (Moore)

Future measurements of treatment efficacy

Currently, researchers describe the effects of drugs in terms of the proportion of treated people whose viral load falls below the detectable limits of the assay. The significance of this depends on the viral load level before treatment: if people start with a low viral load, it does not have to fall very much to become `undetectable'. Also, the lower threshold of different tests does vary, so the same test must be used for monitoring viral load over time.

Experimental studies have begun to look at the slope of decline of viral load during the first week of treatment in order to assess new drugs given as monotherapy. One study found that the decline in viral load at day 6 of treatment predicted the likelihood either of viral load below 50 copies/ml or a -1.5log reduction at weeks 4, 8 or 12; a reduction of less than 0.75 log by this point predicted non-response (Polis 2001).

Researchers have also begun to look at levels of virus in lymphoid tissue such as the tonsils and also at levels of proviral DNA as an indicator of the proportion of CD4 cells which remain infected in plasma and lymphoid tissue. These measurements are relevant in situations where eradication of HIV is being aimed for, and when HIV RNA is undetectable in plasma. They are discussed in more detail in Eradicating HIV? in Anti-HIV therapy: Choosing your treatment strategy.

A team of Swiss researchers has recently reported on the use of a modified viral load test to measure HIV transcription in cells, rather than plasma viral load. Of 16 patients followed for 48 to 90 weeks, 10 had plasma viral load below 50 copies but still had detectable cell-associated HIV transcription (Fischer). This finding was supported by a team from the Aaron Diamond Centre in New York (Lewin).

The `LTR circle' PCR test is a further experimental technique which has been applied in the study of people whose plasma viral load is undetectable using the most sensitive of tests. This test (only used in research settings at present), detects the double loops of long terminal repeats (LTRs) which occur when the HIV DNA formed after reverse transcription either inserts itself into the host cell's DNA or doubles back on itself to form a loop. (Long terminal repeat sequences are found at either end of the HIV genome and contain binding sites for cellular proteins). The appearance of double LTR circles confirms ongoing low-level HIV replication.

Using this technique, a group from Hammersmith Hospital in London found that 76% of individuals with viral load below 50 copies had evidence of ongoing viral replication, despite having no detectable HIV according to the most sensitive tests currently in use (Sharkey).

Key research

Thiebaut reported on 744 individuals commencing a new regimen with 4-12 months follow-uop data available. Forty-seven per cent were receiving triple therapy. Thirty-nine per cent of the cohort were treatment-naﶥ, but the rest had taken at least one dual nucleoside analogue combination before the follow-up period began in mid-1996. Prior treatment experience did not significantly increase the risk of a poor response or subsequent illness. Baseline viral load was around 25,000 to 40,000 copies/mL, and 45% of patients achieved viral load below 500 copies, with 26% falling below 5,000 but not reaching 500 copies, and 29% failing to achieve viral load below 5,000 copies. During the subsequent 12 month follow-up period, poor responders were significantly more likely to experience disease progression than intermediate or good responders. Eighteen percent of poor responders developed an AIDS-defining illness compared to 7% of intermediate responders and 8% of good responders. Following multivariate analysis, poor virological response remained significantly associated with a poorer outcome.

Individuals with sustained viral loads between 2.7-3.7 log10 HIV RNA copies/ml during the 4-12 months treatment had equivalent rates of progression to good responders during the second year of the study. The authors indicated that a change of treatment could be deferred in this group depending on individual treatment history and available treatment options.

Virological response during the first month of treatment was crucial in predicting disease progression after 12 months, whereas baseline viral load was not. A viral load reduction of less than half a log at week 4 was significantly associated with a poor responder status after 12 months treatment

Katzenstein (ACTG 175) examined the size of the reduction in viral load after eight weeks of treatment predicted future disease progression. A 1 log decrease at week 8 sustained to 56 weeks was associated with an 87% reduced risk of death and an 80% reduced risk of developing any further AIDS-defining illness over the three years of the study. Every subsequent 1 log decrease was associated with a further 66% reduced risk of developing any AIDS defining illness

Mayers (CPCRA 007)reported that AZT plus ddI had the greatest effect on viral load and the greatest clinical benefits; a one-log (10-fold) decrease in viral load after six months of treatment was associated with a 60% reduced risk of disease progression or death.

Yerly found that the response of viral load to ddI after one month of treatment was predictive of survival.

OBrien (VA298) reported that changes in CD4 cell count and viral load strongly predicted the clinical benefits of AZT; the researchers calculated that a 75% decrease in viral load along with a 10% increase in CD4 count accounted for 79% of the benefit of treatment in terms of reduced disease progression.

Staszewski conducted a meta-analysis of 1448 people who received AZT/3TC in six trials showed a 2% risk of developing AIDS over the course of a year's follow-up if viral load was suppressed below 5,000 copies.

Romeu analysed 444 patients receiving a variety of anti-HIV drugs. People whose viral load remained below 35,000 copies by the Roche RT-PCR assay over three years of follow-up had a 6.9% risk of developing AIDS or dying after three years. Those whose viral load went above 35,000 copies whilst never exceeding 200,000 copies over three years of follow-up had a 33% risk of developing AIDS over three years. Those who went above 200,000 copies at any point had a 70% risk of developing AIDS after three years. The risk of death was 23 times greater in those who could not maintain viral load below 500,000 copies compared with those who never went above 150,000 copies.

References

Concorde Coordinating Committee. MRC/ANRS randomised double-blind controlled trial of immediate and deferred zidovudine in symptom-free HIV infection. Lancet 343: 871-881, 1994.

Cozzi-Lepri A et al. The virological response to highly active antiretroviral therapy over the first 24 weeks of therapy according to the pre-therapy viral load and the weeks 4-8 viral load. AIDS 15(1): 47-54, 2001.

Delta Coordinating Committee and Delta Virology Committee. HIV-1 RNA response to antiretroviral treatment in 1280 participants in the Delta Trial: an extended virology study. AIDS 13(1): 57-65, 1999.

Fischer M et al. HIV-1 RNA transcription in blood cells of patients with undetectable plasma viral loads. Sixth Conference on Retroviruses and Opportunistic Infections, Chicago, abstract 154, 1999.

Hoetelmans RMW et al. The effect of plasma drug concentrations on HIV-1 clearance rate in anti-retroviral naive patients during a quadruple drug regimen. AIDS 12:8, F111-5, 1998.

Huang W et al. Patterns of plasma human immunodeficiency virus type 1 RNA response to antiretroviral therapy. J Infect Dis 183(10): 1455-1465, 2001.

Hughes MD et al. Monitoring plasma HIV-1 RNA levels in addition to CD4 lymphocyte count improves assessment of antiretroviral therapeutic response. Annals of Internal Medicine 126: 929-938, 1997.

Jacobson MA et al. Surrogate markers for survival in patients with AIDS and AIDS related complex treated with zidovudine. British Medical Journal 302: 73-78, 1991.

Johnson V et al. HIV drug resistance and virological failure during combination therapy with nevirapine, zidovudine and didanosine. International Workshop on Drug Resistance, Treatment Strategies and Eradication, Florida, abstract 58, 1997.

Katzenstein DA et al. The relation of virologic and immunologic markers to clinical outcomes after nucleoside therapy in HIV-infected adults with 200 to 500 CD4 cells per cubic millimeter. New England Journal of Medicine 335: 1091-1098, 1996.

Kempf D et al. The duration of viral suppression during protease inhibitor therapy for HIV-1 infection is predicted by plasma HIV-1 RNA at the nadir. AIDS 12(5): F09-F14, 1998.

Lewin S et al. Use of real-time PCR and molecular beacons to detect virus replication in human immunodeficiency virus type 1-infected individuals on prolonged effective antiretroviral therapy. J Virol 73(7): 6099-6103, 1999.

Machado SG et al. On the use of laboratory markers as surrogates for clinical endpoints in the evaluation of treatment for HIV infection. J AIDS 3: 1065-1073, 1990.

Martinez MA et al. Human immunodeficiency virus type 1 genetic evolution in patients with prolonged suppression of plasma viremia. Virology 10; 256(2): 180-187, 1999.

Mayers D et al. Viral burden measures in CPCRA 007. Eleventh International Conference on AIDS, Vancouver, abstract Th.B.911, 1996.

Moore AL et al. Gender differences in virologic response to treatment in an HIV-positive population: a cohort study. JAIDS 26: 159-163, 2001.

O'Brien WA et al. Changes in plasma HIV-1 RNA and CD4+ lymphocyte counts and the risk of progression to AIDS. New England Journal of Medicine 334(7): 426-430, 1996.

O'Brien WA et al. Changes in plasma HIV RNA levels and CD4 lymphocyte counts predict both response to anti-retroviral therapy and therapeutic failure. Annals of Internal Medicine 126: 939-945, 1997.

Pilcher CD et al. Detectable HIV-1RNA at levels below quantifiable limits by Amplicor HIV-1 monitor is associated with virologic relapse on antiretroviral therapy. AIDS 13(11):1337-1342, 1999.

Polis MA et al. Correlation between reduction in plasma HIV-1 RNA concentration 1 week after start of antiretroviral treatment and longer-term efficacy. Lancet 358(9295): 1760-1765, 2001.

Raboud JM et al. Suppression of plasma viral load below 20 copies/ml is required to achieve a long-term response to therapy. AIDS 12: 1619-1624, 1998.

Romeu J et al. Effect of maintaining sequential HIV-1 viral load values below a certain threshold on progression and mortality. 37th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, abstract I-137, 1997.

Saag MS et al. HIV viral load markers in clinical practice. Nature Medicine 2(6): 625-629, 1996.

Sharkey M et al. Persistence of episomal HIV-1 infection intermediates in patients on highly active anti-retroviral therapy. Nature Medicine 6:1, 76-81, 2000.

Staszewski S et al. Progression to AIDS very rare when HIV-1 RNA below 5,000 copies. 37th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, abstract I-129, 1997.

Thiebaut R et al. Clinical progression of HIV-1 infection according to viral response during the first year of antiretroviral treatment. AIDS 14(8): 971-978, 2000.

Wong J et al. HIV recovery from PBMC of patients with undetectable plasma viremia is not due to residual viral replication. Second International Workshop on Drug Resistance and Treatment Strategies, Lake Maggiore, abstract 148, 1998.

Yerly S et al. Response of HIV RNA to didanosine as a predictive marker of survival. AIDS 9: 159-163, 1995.