HIV Therapy Strategies: initial, second line and salvage
     Reported for NATAP by Graeme Moyle, MD, Chelsea & Westminster Hospital, London, UK

Two presentations at the International Congress on Drug Therapy in HIV Infection dealt with the key tactical and strategic issues related to HIV therapy intervention. Professor Scott Hammer from the Columbia University in New York City began with a discussion of the pros and cons of early intervention with therapy:


Pathogenesis (immune damage early)  

Increased chance of success

Immune preservation.

Reduced transmission.

Long term clinical benefit.





Ability to undergo immune restoration

Quality of life

Absence of evidence of viral eradication


Professor Hammer, after outlining these points, said that particular issues around toxicity have currently swung the treatment pendulum in favour of a delayed time to intervention, but there is clearly a range with both CD4 count and viral load where there exists a quandary as to when to intervene. He suggested that most physicians would currently now defer therapy in individuals with a CD4 count above 500 cells/mm3 and that the need for treatment was clear in individuals below 200/mm3, as this has been established by several cohort studies to be the range where opportunistic events occur.

Therefore, the area of quandary lies between CD4 counts of 200 to 500/mm3 with a number of guidelines now recommending that intervention begin below 350 cells/mm3. Importantly, some evidence to be discussed at this meeting evaluates this quandary. An analysis by Alexandro Cozzi Lepri (Abstract PL3.5) presented at this meeting evaluated the mean CD4 count increase and time to virological rebound in individuals starting triple therapy regimens with CD4 counts of <200, 201-350 and >350. In the database study, which was conducted by the ICONA group in Italy, 1,329 previously naÔve individuals were included. Rises in CD4 count in the <200, 201-350 and >350/mm3 baseline CD4 strata after 24 months of therapy were a mean 296, 312 and 151 cells/mm3, respectively. Individuals with a CD4 count less than 200 at the initiation of therapy, however, had an increased relative risk of subsequent virological failure, compared with the group commencing with a CD4 count greater than 350.  (Relative hazard = 1.33, 95% CI: 1.02-1.73).

There was, however, no increased risk of virological failure relative between the 201-350 and the >350 group (relative hazard 1.01). This study therefore supported Professor Hammerís view that there is no particular advantage, as recommended in current guidelines to starting with a CD4 count greater that 350 relative to one between 350 and 200, however it does support the need to start therapy before the CD4 count falls below 200/mm3.

A similar analysis also presented at this meeting by Professor Andrew Phillips (Abstract PL3.4) which included a database of 2,742 European patients also found that there was an increased risk of viral rebound in individuals who initiated with a CD4 count below 200/mm3 and those who initiated with a viral load greater than 100,000 copies/ml. Using this data, Professor Hammer went on to discuss the quandary that exists in the range of viral loads. He suggested that some individuals may now be considering the CD4 count as being an exclusive basis for starting therapy.

However, reminding us of data from the MACS cohort published several years ago now by Mellors et al, these indicated that viral load has a strong predictive value of future risk of disease progression in persons with HIV, independent of CD4 count. Therefore, one should at least consider that the individuals in the highest stratum in this study (i.e. individuals with viral loads greater than 50,000 copies/ml) should be considered for starting therapy. He suggested that based on current guidelines in clinical practice, the majority of individuals with a viral load below 5,000 copies would be considered for deferment of treatment and, as supported by Andrew Phillipsí data that the majority of individuals above 100,000 should be considered as clearly requiring treatment.

Therefore, the quandary for viral load lay in the range of 5,000-100,000 copies per mil.  He provided some further data to support evidence for intervention at relatively lower viral loads as being more efficacious with a new analysis derived from the ACTG320 study which has been currently prepared for publication.  This study involved individuals who were zidovudine experienced but protease inhibitor and 3TC naÔve who added 3TC plus indinavir into their regimen. Looking at the proportion of patients who were less than 50 copies/ml by week 24 within that study in the group who started with a viral load of less than 5,000 copies, 59% were undetectable. In the range 5-50,000 copies 44% were undetectable and in the range 50,000 to 500,000 it was 24% and those starting therapy with greater than 500,000 copies 14% only achieved an undetectable viral load by this time point.

Whilst these data are supportive of the idea of commencing with a viral load less than 50,000, they ignore some data from an analysis of the DuPont 006 study, recently presented at the ICAAC meting, which looked at treatment naÔve patients commencing either Efavirenz or Indinavir in combination with AZT and 3TC. This analysis indicated that the time point of 24 weeks may be too soon to assess optimal virological response in individuals who commence with viral loads greater than 100,000 copies, hence the new ACTG 320 analysis may be underestimating the efficacy of this regimen at high viral loads by using a premature cut-off.

Professor Hammer went on to mention that when using both CD4 count and viral load, one needs to assess that there may be gender and ethnicity specific variation within CD4 counts and viral loads. Furthermore, physicians should not be considering making treatment decisions on the basis of single samples but should be looking at repeat sampling to assess the trajectory of viral load and CD4 count over time.

Professor Hammer then went on to discuss the options for first therapy, the basis of this choice being:

  • Potency.

  • Side effect, their severity and frequency.

  • Patientsí commitment to therapy as an important determinant of future adherence.

  • Impact on quality of life of therapy.

  • Impact of therapy choice on future treatment options.

  • The presence of co-infections, eg Hepatitis B or C or TB.

  • The need for concomitant medications which may have potential drug interactions.

  • Whether the individual may have acquired a drug-resistant HIV strain.

  • Issues around cost and access to therapy.

The discussed options included:

  • two nucleoside analogues in combination with a protease inhibitor, which may be boosted with a ëbabyí dose of ritonavir,

  • two nucleoside analogues in combination with a non-nucleoside reverse transcriptase inhibitor (NNRTI),

  • two nucleoside analogues in combination with abacavir,

  • all three drug classes concomitantly or

  • a nucleoside sparing regimen of protease inhibitor with an NNRTI

He said that while the overall impression of these regimens was one of relative similarity between different combinations, some studies have suggested that differences may exists between therapies. The difficulty with interpretation of some of these studies being that the administration characteristics may vary between the components of these regimens, for example, three times a day versus twice daily dosing or food restrictive versus unrestricted dosing. This may mean the studies are really testing adherence rather than potency. In recent times, randomised clinical trials have demonstrated superior virological responses at week 48 to Lopinavir/ritonavir over nelfinavir (ABT378 study, M98, 863, Abstract PL65 at this meeting), nevirapine over nelfinavir (The Combine Study), Efavirenz over Indinavir (006 Study), Indinavir over Abacavir (CNA3005 Study) and, out to 24 weeks, similar response between Indinavir and Abacavir (CNA30014 Study).  It is important to note, that when studies are repeated to confirm the results of a first study, such as with the comparisons of abacavir and indinavir, different results may be observed. Physicians need to take into account the difficulties of interpreting the results derived from these comparative studies to draw pragmatic conclusions for individual patients. Professor Hammer underlined that perhaps the group where the potency of regimens is best assessed is in individuals with high viral load, e. g. >100,000 copies/ml and with follow-up of 48 weeks or greater.

With regard to switching of therapy, Professor Hammer commented that studies looking at Efavirenz, Nevirapine or Abacavir as replacements for Protease Inhibitors had demonstrated maintenance of virological control in the vast majority of individuals who were on their first ever treatment regimen. However, some loss of virological control may occur, with at least some of the drugs, in a proportion of patients who had prior nucleoside analogue experience before switching. This has been best demonstrated in the Abacavir Swiss Cohort Study where patients with archived zidovudine mutations from previous therapy had an increased risk of virological rebound when switched to an Abacavir-based triple nucleoside regimen. 

Professor Hammer also mentioned the potential to investigate protease inhibitor plus non-nucleoside combinations for the management of nucleoside analogue toxicity, but said that, thus far, no data had been reported using this approach in a substitution circumstance.

He underlined that it is important when considering first choice therapy to consider what the second choice therapy will be and underlined that second choices of therapy should always be guided by resistance tests, given the appreciation of the limitations that exist in interpretation of resistance test results.  Listing off his first line therapy regimens, he indicated that, for example, if an individual fails on two nucleoside analogues and a protease inhibitor, if 3TC is present in the regimen, this is likely to be the only drug to which resistance is present, whereas if an individual commences on a non-nucleoside based regimen, virological resistance to 3TC (if present) and the NNRTI is also likely to be present at the time of rebound. Similarly with Abacavir-based regimen, 3TC failure has been most commonly the first reported rebound. With triple class regimens, if one includes 3TC and an NNRTI resistance to both of the agents may be present at the time of rebound.  In the few studies that have evaluated protease inhibitor plus non-nucleoside patients, it tends to be non-nucleoside resistance that is present at the time of rebound.  Professor Hammer made no further comment about the issue regarding 3TC failure, although previous guidelines from the USA have suggested that 3TC should only be used in regimens that the physician assumes will be fully suppressive.  The contribution of this drug to the failure of first line regimens raises the issue of whether this drug should really be included in first line therapy choice or it may be best included in intensified second line therapy where 3TC use has also been demonstrated to be beneficial.

Professor Hammer went on to comment that several strategic studies including the Initio Study in Europe and Australia and North America, the ACTG384 and ACTGA5095studies will be going on to investigate strategic therapy approaches. In these studies, specific initial regimens are then followed by specific second line regimens to look at the issues of sequencing. As a special note, he commented on the recent new availability of lopinavir/ritonavir combination capsules and discussed that, as this drug may be useful both as an initial PI regimen, where it has indicated superiority to nelfinavir, but also has demonstrated potent activity in the circumstances of protease inhibitor pre-treated patients. Therefore, more data are needed to clarify whether this drug should be used as a first choice drug from the protease class or whether it should be saved for salvage after prior protease inhibitor use. 

The task was then left to Professor Sven Danner to discuss choices in salvage therapy. First, Professor Danner discussed how one may need to define failure, pointing out that the choice of definition is likely to change with individual circumstance. In some cases, one may describe failure either by a virological definition (say rise above 50 or 400 copies/ml), but in others by a fall of the CD4 count below a specific threshold (say 200 or 50 cells/mm3). The clinical trials definitions during late 80ís and early 90ís of treatment failure was a clinical one ñ the development of an AIDS-defining opportunistic infection or tumour ñ however this is the aspact of failure we must wotk hardest to avoid. To put it another way, failure may be based on oneís definition of success, failure being the absence of achievement of oneís success definition. Professor Danner, reiterated some of the points made by Professor Hammer in this regard; the ideas of lower viral load and higher CD4 count patients being associated with longer responses to therapy, more complete suppression of viral load. However, it is unfortunate that the nature of presentation of HIV diseases means many individuals present with characteristics which are unfavourable prognostically in this regard. Additionally, many patients have benefited from advice in the late 80ís and early 90ís which, whilst the best available advice at the time, has had an impact on their response to currently available therapies.

In this regard he moved on to discus the issue of discordant responses, that is to say individuals who have an incomplete virological response to therapy but maintain a rise in CD4 count. This phenomenon was first described during the AZT monotherapy era where complete reduction of viral load was rare but sustained elevations of CD4 counts over a year or more were observed.  The sustaining of CD4 count rises in people on triple therapy as described by Stephen Deeks at the San Francisco General is, on average, for about three years after a viral rebound on a protease inhibitor -based regimen.  Professor Danner describes some recent work published by Grabar (in the Annals of Internal Medicine 2000, vol 133: 401-419), which defined responses in a large cohort of patients as being either a CD4 count rise of greater than 50 cells, a viral load fall of greater than 1 log or both. Individuals commencing therapy who experienced both these response had few clinic events over the course of follow-up, whereas individuals who have neither of these responses have the poorest prognosis and were most likely to experience clinical disease progression. Interestingly, the individuals who experience an immune response to therapy did better than those who were virological respondents to therapy, with regard to risk of future disease progression.

Professor Danner then discussed the issue of insufficient virological response, which he defined as an individual commencing on therapy who, by a particular time point, (e.g. week 16 or week 24 into therapy) had not achieved an appropriate therapy goal, (e.g. less than 400 copies per ml at week 16 or less than 50 copies/ ml at week 24). He suggested that such patients may be suitable for treatment intensification, that is to say the addition of at least one new drug into that treatment regimen. He cited data from the Prometheus Study in which patients who had incompletely responded to a dual protease regimen of ritonavir and saquinavir added D4T at week 16 or week 17 and subsequently achieved similar virological response rates at week 48 to those patients who commenced all three drugs simultaneously. He underlined that whilst intensification may be appropriate in the circumstances of a fall in viral load where it appears that the regimen may be insufficiently potent, it is different from the circumstance where viral load is initially suppressed and then subsequently rebounds, where resistance to drugs may be an important aspect of the causation of rebound.

He underlined that, in these circumstances, there is a need to use resistance tests to guide the choice of future treatment options at that, indeed, the number of drugs may need to be increased with each new regimen to increase the chances of adequate antiviral immunological response. He also discussed the issue of toxicity related to therapy pointing to data from the Athena cohort in the Netherlands, where 30-40% of patients had an interruption or change of therapy over three years of follow up, secondary to toxicity. In assessing an individual, experiencing rebound of viral load on therapy, he underlined that, beyond issues around drug potency and viral resistance, physicians had available tools to assess whether adequate drug exposures were being achieved. Citing data from several studies, including the Viradapt study and Prometheus study where therapeutic drug monitoring had been used to correlate response with drug exposure. There is also the need to assess with patients whether they are taking their therapy reliably, not simply in terms of with the appropriate dosing intervals, but whether the medication is being taken with regard to food and fluid requirements. 

In a broad summary of the management of treatment failure, he suggested that we need to be looking at using new drugs which have non-overlapping resistance profiles, at the present time that often involves choosing drugs from within class and that, as Professor Hammer had underlined, the availability of drugs from the same class for future use will depend on what drugs had been chosen in the first line and other lines of the regimen prior to the this time point.

He also discussed the potential to boosted drug exposures, such as using baby dose ritonavir, with protease inhibitors to boost the ratio of minimum drug exposure to the inhibitory concentration required for that virus, sometimes called the ëinhibitory quotientí and that to achieve success in this regard, one may even consider using no just a single protease inhibitor boosted by ritonavir but potentially two protease inhibitors boosted by ritonavir. He also looked forward to the opportunity of using new classes of drugs, such as when fusion inhibitors become available. When he assesses an individual who has experienced virological failure he goes through a checklist of adherence, drug levels and resistance before considering change based on whether the individual has appropriate treatment options available for them and whether they are having the persistence of a discordant immunological response which may provide him with the opportunity to delay the decision to change therapy until better treatment options have accumulated.