Success and failure of treatment can be evaluated using virological, immunological or clinical criteria. Although they are often associated with each other, they should be judged separately.
The earliest indicator is virological success or failure. This mainly means decrease, absence of decrease, or increase in viral load. This is followed, often a little later, by immunological treatment success, measured through the CD4 cells, or immu-nological treatment failure. Clinical treatment failure, if it occurs, usually only becomes apparent much later - first the lab values deteriorate, then the patient! Although the incidence of opportunistic infections after only three months on HAART is approximately halved, the clinical treatment success is not noticed by asymptomatic patients (Ledergerber 1999).
Virological treatment success is generally a decrease in viral load to below the level of detection of 50 copies/ml. This is based on the experience that, the more rapid and greater the decrease in viral load, the longer the therapeutic effect (Kempf 1998, Powderly 1999).
In as early as the INCAS Trial, the relative risk of treatment failure (defined here as an increase to above 5,000 copies/ml) in patients who had reached a viral load below 20 copies/ml was 20 times lower than in those who had never reached a level under 400 copies/ml (Raboud 1998). It is not completely clear whether the data obtained from the early HAART era are still valid.
On HAART, viral load declines in two phases. An initial, very rapid decrease in the first few weeks is followed by a slower phase, in which plasma viremia declines only slowly. A decay to below the level of detection should be reached after 34 months; in cases of very high baseline viral load, it may even take longer. However, a viral load above the level of detection after 6 months of treatment is almost always seen as failure. The same is true if a rebound in viral load is confirmed, in which - following fast confirmation - it should be considered what can be improved in the therapy (resorption, resistance, compliance?).
Virological treatment failure can be recognized quite early - therefore, initial monitoring even after four weeks is useful not only to the patient for psychological reasons ("less viruses, more helper cells"). If the viral load after 4 weeks is not at least substantially less than 5,000 copies/ml, failure of therapy is likely later (Mag-giolo 2000). Of those patients, in whom the viral load is not below 500 copies/ml after 8 weeks or at least one log lower than baseline, only 9 % will reach less than 500 copies/ml at 24 weeks (Demeter 2001). According to a new prospective study, the response at 48 weeks can be predicted as early as 7 days (Haubrich 2007) - a control so early on is not part of the routine management.
The cut-off point of 50 copies/ml as the success criterion is arbitrary. It is based on the currently available assays for measurement of viral load. Whether 60 copies/ml are worse than 30 copies/ml indicating successful treatment is as yet not proven. At these low levels, methodological inaccuracies must also be considered. A single viral load rebound ("blip") to low levels is often irrelevant (see below). However, to distinguish them, blips are from lower, but still measurable viral loads (50400 copies/ml). Resistances are frequently detectable in these patients - in one study this was so in 43 % of cases (Nettles 2004).
A viral load "below the level of detection" of 50 copies/ml means just that - no more, no less. Even so, numerous studies indicate that replication and therefore development of resistance can continue with an undetectable virus load. 50 viral copies/ml indicate that 5 liters of blood contain 250,000 viruses; in addition, even more actively replicating viruses are present in the lymphatic organs. Theoretically, a measurable viremia, even at very low levels, may possibly translate to a higher risk of resistance in the long-term. Perhaps there is indeed a relevant difference between 100 and 10 copies/ml with regard to the risk of developing resistance. But we just don't know yet.
The most important risk factors for virological treatment failure are antiretroviral treatment (pre-existing resistances) and poor compliance (review: Deeks 2000). It has still not been proven whether the CD4-cell count at the start of therapy plays an important role. In several cohorts, no association has been found (Cozzi Lepri 2001, Phillips 2001, Le Moing 2002). See also the discussion in the chapter "When to start HAART".
Many other factors that influence the success of therapy probably remain unknown. Pharmacogenetics is a new field, which, although still in the early phases, is starting to gain importance. It investigates individual genetic factors which influence the success of therapy. Until now, most factors that have been uncovered have predicted intolerance or allergies to, for example, abacavir or nevirapine (see relevant section). But one day, tests will be available that will help antiretroviral therapy to be individualized and the success of therapy to be improved. These will range from individual dosing to tests, which predict results - eg. the CCR5 antagonists (Review: Haas 2006).
For the time being, the good news is: morbidity and mortality may be lowered significantly even if the virological success is not complete, i.e., the viral load is not decreased below the level of detection (Mezzaroma 1999, Deeks 2000, Grabar 2000). This is important in patients who have a limited number of treatment options. In such cases, it can sometimes be more sensible to temporarily abandon viral load as a measure of success (see also chapter on "Salvage Therapy"); and to make the stabilization of the CD4 cells the top priority. Patients often remain immu-nologically stable for relatively long periods of time, even with insufficient viral suppression. A cohort study has shown that CD4 cells do not drop as long as the viral load remains below 10,000 copies/ml or at least 1.5 logs below the individual set point (Lederberger 2004).
However, in comparison to a few years previously, more has become possible through new drugs and classes of drugs. In the age of T-20, tipranavir, darunavir, etravirine, maraviroc and raltegravir a repeat attempt should be made to reduce the viral load to below the level of detection, even in intensively pre-treated patients.
How long does virological treatment success last?
Little is known about how long treatments remain effective. The rumor that treatment success is limited to only a few years is still widespread. It originates from the early years of HAART. However, many patients at the time were still inadequately treated or had been pretreated with mono- or dual therapy, and had thus developed extensive resistance. In such patients, the effect of treatment was often limited, as even a single point mutation was often enough to topple a whole regimen. Today, especially in therapy-naive patients without pre-existing mutations, the risk of treatment failure is much less.
After ten years of HAART, a surprisingly high number of patients still have viral loads below the level of detection. This is particularly true for patients who were adequately treated from the start, as judged by today's standards (starting with triple therapy and/or rapid switching of several drugs). One of the few trials with a longer follow-up period studied 336 antiretroviral-naive patients who had reached a viral load below 50 copies/ml within 24 weeks (Phillips 2001). After 3.3 years, the risk of viral rebound seemed at first glance to be relatively high at 25.3 %>. More detailed analysis showed that a large proportion of the patients experiencing viral rebound had actually interrupted HAART. True virological failure was only seen in 14 patients, which corresponds to a risk of 5.2 % after 3.3 years. Most importantly, the risk of virological failure decreased significantly with time. In the Phase II M97-720 Study, in which 100 patients were originally treated with d4T+3TC+lopinavir/r, 62 % still had less than 50 copies/ml after six years in the ITT analysis, compared to 98 % in the On-Treatment-Analysis (Gulick 2004). Real virological failure was very rare. In the Merck 035 subanalysis, patients on AZT+3TC+indinavir were also followed for six years. In the last ITT analysis, 58 % were still below the level of detection, despite the fact that these patients had been pre-treated with nucleoside analogs (Gulick 2003).
These studies clearly show that, providing treatment is not interrupted, viral load may remain below the level of detection for many years, perhaps even decades. Resistances are by no means unavoidable. This has been confirmed by cohort studies, in which virological treatment failure has become significantly less in the last few years (Lohase 2005, Lampe 2006). HAART is getting better and better. In 1996, only 58 % of patients had a viral load of less than 500 copies/ml; in 2003, it was 83 % (May 2006).
"Blips" - Do they mean virological failure?
Blips are thought to be transient and, almost always, small increases in viral load, so long as the viral load before and after the blip was below the borderline value of 50 copies/ml. At least three measurements of viral load are therefore required to be able to identify a blip. Blips are a frequent phenomenon of HIV patients on HAART and are observed in 20-40 % of patients on HAART (Sungkanuparph 2005). Blips often worry both patients and clinicians: do they signal treatment failure?
Although studies indicate that this is not the case in the medium term (Havlir 2001, Moore 2002, Sklar 2002, Mira 2003, Sungkanuparph 2005), the causes of blips have, to a large extent, not been investigated. There has been no association found with compliance. (Di Mascio 2003, Miller 2004). It is possible that blips are the result of immunological mechanisms. The earlier patients are treated in the course of infection, i.e., the higher the CD4 cell count at the start of therapy, the more seldom blips seem to occur (Di Mascio 2003+2004, Sungkanuparph 2005). There does not appear to be any association with particular antiretroviral combinations - in a large cohort (Sungkanuparph 2005), the frequency of blips on NNRTIs was 34 versus 33 % on PIs, and even the size of the blips were equivalent (median 140 and 144 copies/ml respectively). In both groups, the risk of virological failure at 2 years was approximately 8 %. One important observation was that the blips did not increase the risk of treatment failure, even with NNRTIs (Martinez 2005).
At the beginning of 2005, the study team led by Bob Siciliano set out to determine the meaning of "blips". In a labor-intensive study, 10 stalwart patients who had had a viral load of less than 50 copies/ml for at least six months, had blood samples taken every 2-3 days (!) over a period of 3-4 months (Nettles 2005). The obvious result: the more you look, the more you will find. During the observation time, at least one transient increase in the viral load was measurable above 50 copies/ml, in nine of the ten patients. Each blip was moderate, with a median value of 79 copies/ml, ranging from 51 to 201 copies/ml. The blips could not be associated with either specific clinical data, low plasma levels, or resistances. This observation led the authors to believe, that blips (with low, measurable values) mainly represent biological or statistical exceptions, and are not associatedwith treatment failure. In an estimated steady state level of viral load at around 20 copies/ml, the values are distributed randomly. However, 96 % of the randomly distributed measurements were less than 200 copies/ml.
It should be noted that other factors may also be responsible for intermittent vire-mia. In one large, retrospective analysis, 26 % were caused by intercurrent infections (Easterbrook 2002). For example, syphilis can cause a significant increase in viral load and reduction of CD4 cells (Buchacz 2004). Viral load can also increase temporarily after immunizations (Kolber 2002).
Summary: Based on available data, HAART should not be changed, even after repetitive blips. However, caution should be executed for higher blips (> 200500 copies/ml). Blips are distinguishable from low, repetitive, measurable plasma viremias, in that the risk of resistance is actually increased (Gunthard 1998, Nettlers 2004). Although there does not seem to be a relationship to compliance or drug levels, blips should raise the opportunity to talk to the patient about the subject of adherence. Compliance cannot be discussed often enough. Does the patient take his or her drugs regularly, or are doses occasionally missed? Are the dosing directions (on an empty stomach or with a meal) followed correctly?
Everything should be considered before changing a therapy. Each new therapy can cause new problems. Therefore, every suspected increase in the viral load should be controlled within a short interval, before treatment is prematurely changed.
Immunological treatment success is an increase in the CD4 cells - it is not defined more precisely. Depending on the study, increases by 50, 100 or 200 CD4 cells/^l or increases to above 200 or 500 CD4 cells/^l are defined as success. Failure is usually described as the absence of an increase or as a decrease in the CD4 count in patients receiving HAART.
It is difficult to individually predict the immunological success of therapy for patients on HAART, as it varies significantly from one person to another. As with the decrease in viral load, the increase in CD4 count also occurs in two phases. After a first, usually rapid increase over the first three to four months, further increases are considerably less pronounced. In a prospective study involving some 1,000 patients, the CD4-cell count increased during the first three months by a median of 21.2 CD4 cells/^l per month; in the following months the increase was only 5.5 CD4 cells/^l (Le Moing 2002). It is still under debate whether the immune system is restored continuously after a long period of viral suppression or whether a plateau is reached after three to four years (Smith 2004, Viard 2004). In our experience, both occurs: patients with CD4 cells that still increase only slowly after 5 or 6 years, and patients whose CD4 cells remain at relatively low level after just a relatively short period of time. The immunological success is not predictable in individual cases.
However, the lower the CD4 cell count at baseline, the less likely they are to normalize completely (Valdez 2002, Kaufmann 2003+2005). The immune system often does not recover completely. In the Swiss Cohort, only 39 % of 2,235 patients who had begun HAART in 1996-97 reached a CD4-cell count above 500/^l (Kaufmann 2003, see also below). The introduction of treatment within the first 3-6 months possibly provides certain clues as to how well the immune system will be restored (Kaufmann 2005).
Immunological treatment success is not necessarily linked to maximal viral suppression; even partial suppression can result in improved CD4-cell count (Kaufmann 1998, Mezzaroma 1999, Ledergerber 2004). The initial level of viral load is also not significant; what seems to be decisive is that the viral load remains lower than before treatment (Deeks 2002, Ledergerber 2004). In view of the many factors that occur, which are able to influence the success of therapy as well as the individ ual regeneration capacity (independent of HAART), it no longer makes sense to depend on the CD4-cell count as the deciding criterion for the success of HAART. Virological success is more appropriate for judging the efficacy of specific regimens.
A discordant response occurs when the therapeutic goals - clinical, immunological and virological - cannot all be achieved (see Table 4.1 for frequencies). So, the treatment may be virologically successful, with no visible immunological response; with the CD4 cells remaining on the low side despite undetectable viral load (Piketty 1998, Renaud 1999, Gabrar 2000, Piketty 2001). In contrast, a HAART regimen can bring about a considerable increase in CD4 cells, even when the viral load is detectable. This is sometimes seen in children (see Pediatric chapter).
Table 4.1: Treatment success in prospective cohort studies
Response to HAART
Piketty 2001 n = 150
Grabar 2000 n = 2,236
Moore 2005 n = 1,527
Virological and immunological 60 %
Discordant: only immunological 19 %
Discordant: only virological 9 %
No treatment response 12 %
Definition of immunological success: increase in CD4 cells > 100/|jl after 30 months (Piketty 2001) or > 50/jl after 6 months (Grabar 2000) or at least once > 50/jl (Moore 2005)
Definition of virological success: continually > 1 log less than baseline or < 500 copies/ ml (Piketty 2001) or < 1,000 (Grabar 2000) or < 500 copies/ml (Moore 2005)
The reasons for the poor immunological treatment success despite good viral suppression are heterogeneous (Review: Aiuti 2006).
Low CD4 cells at baseline as well as low viral load before the start of therapy are just two of many factors (Florence 2003, Kaufmann 2005, Moore 2005, Wolbers 2007). Age also plays an important role: in older patients, immunological response is often only moderate in comparison to virological response. Several studies demonstrated that the probability of not achieving a rise in the CD4-cell count increases with patient age and with progressive decrease in thymus size as detected by computed tomography (Goetz 2001, Marimoutou 2001, Piketty 2001, Teixera 2001, Viard 2001, Wolbers 2007). Patients who are intravenous drug users also have relatively poor increases in CD4 cells (Dragstedt 2004). In the Swiss cohort, the CD4 cells increased more in women than in men (Wolbers 2007). Other causes for a lack of immunological response may be immuno- or myelosup-pressive concomitant therapies. We have seen patients, who have had a suppressed viral load below 50 CD4 cells/^l for years, who only experience significant immu-nological reconstitution when gancyclovir, cotrimoxazole or azathioprine are withdrawn. Concurrent illnesses such as various autoimmune diseases (Crohn's disease, Lupus erythmatosis) or liver cirrhosis can also have negative effects on the immune response.
There is also evidence that certain antiretroviral drugs have negative effects on immune reconstitution. On a combination of TDF+ddI (plus nevirapine) significant decreases in CD4 cells occur (Negredo 2004). The reason may lie in an unfavorable interaction between ddl and tenofovir. In two other studies, the CD4 cells increased significantly more on ABC+3TC, as well as on TDF+FTC, than on AZT+3TC, despite comparable virological success. However, it remains to be shown whether this is really a problem of myelotoxicity of AZT or a pure coincidence (DeJesus 2004, Pozniak 2006).
Practical considerations in dealing with viral load and CD4 cells
■ Viral load is the most important parameter in treatment monitoring.
■ If possible, use only one type of assay (in the same lab) - bear in mind that there is considerable methodological variability (up to half a log)!
■ Virological success should be monitored one month after initiation or modification of HAART.
■ Viral load should be below 50 copies/ml after 3-4 months (with high initial viral load, after 6 months at the latest) - if it is not, look for a cause!
■ The greater the decrease in viral load, the more durable the response to treatment.
■ Transient, low-level increases in viral load (blips) are usually insignificant -but VL should be monitored at short intervals (e.g. 2-4 weeks after such blips).
■ The older the patient, the greater the risk of a discordant response (well-suppressed viral load with no significant increase in CD4 count).
■ In contrast to viral load, increase in CD4 cells, i.e. immunological success, is difficult to influence.
■ CD4 cells are probably better predictors of the individual risk of AIDS.
■ Once CD4 count is good, it requires less frequent monitoring. Remember that with higher CD4 counts, values may vary considerably from one measurement to the next (which may mislead the patient to either a false sense of euphoria or unnecessary concern).
Clinical success is almost always evaluated on the reduction of clinical endpoints (AIDS-defining illnesses, death), although the improvement on HAART in a patient with considerable constitutional symptoms should also be seen as clinical success. Clinical treatment success is not always easy to measure and is dependent on both virological and immunological treatment success.
Table 4.2: Morbidity and Mortality, based on virological or immunological treatment success. Definition see Table 4.1. 95 % confidence intervals are shown in parentheses
Grabar 2000 Piketty 2001 Moore 2005
CD4 cells at baseline* Treatment success: Complete = Reference Only immunological, RR Only virological, RR None, RR
Was this article helpful?
All Natural Immune Boosters Proven To Fight Infection, Disease And More. Discover A Natural, Safe Effective Way To Boost Your Immune System Using Ingredients From Your Kitchen Cupboard. The only common sense, no holds barred guide to hit the market today no gimmicks, no pills, just old fashioned common sense remedies to cure colds, influenza, viral infections and more.