Management of treatment experienced patients

Simon Collins, HIV i-Base

It is difficult to know the number of patients in the UK who are currently on a virologically failing treatment with resistance to three drug classes.

Low enrolment in the Optima trial comparing use of Mega-HAART >4 drugs to standard four-drug therapy, with or without a treatment interruption, has been reported to be due to a small pool of patients in this situation, perhaps due to the recent availability of lopinavir/r and tenofovir.

Nevertheless the numbers are likely to continue to grow before the next ‘best hope’ drugs for resistance virus become widely available. Tipranavir, a protease inhibitor from Boehringer with activity against current PI resistance, likely to be boosted by ritonavir, will not be available in expanded access until late 2003. The first fusion inhibitor, T-20 (enfuvirtide) from Roche/Trimeris, will not be affected by resistance to current agents but will also not become more widely available again until mid 2003 at the earliest. To be effective both these drugs need to be supported by other active drugs within a regimen.

Very encouraging results were presented in Barcelona on both these drugs for salvage therapy.

Although other long promised therapies are now in the early pipeline including integrase inhibitors and other entry inhibitors, the development process, although compressed for HIV, will never be fast enough for those dependent on access to these drugs. Also, access to these compounds within studies may be less likely to offer best treatment for those study participants unless other investigational agents are allowed.

Until these new drugs become available, it is essential to understand several principles in order to optimise individual patient care:

  1. accurately targeting active drugs to susceptible virus is the only mechanism for reducing and maintaining a reduced viral load;
  2. resistance is nearly always a sliding scale rather than an on-off process. Therefore some drugs may continue to provide benefit despite low level resistance and some resistance and may therefore be overcome given higher drug pressure.
    Both these point to the importance of individualising treatment and individualising dosing through the use of therapeutic drug monitoring.
  3. thirdly, drug resistance appears to result in less replicatively competent virus – and several studies reported on way to understand and exploit this resistance in a salvage setting.

Montaner overview favours Mega-HAART

Options in salvage therapy were supported in an important overview by Julio Montaner. [1] In practice, even with the use of new agents, he suggested that there are essentially only a few strategies:

  1. to change to a multiple drug combination – in the hope that a heterogeneous viral resistance population will be affected by multiple different drugs
  2. to use a period without treatment – to partially revert to wild type virus and the subsequent multiple drug regimen will have higher success.
    Results from Barcelona supported both approaches in different degrees, and a third option:
  3. to do nothing and monitor, (Deeks et al) has been shown to hold some patients who are using protease-based combinations clinically stable for up to two years. This is generally recognised as a limited measure because resistance is likely to continue to develop.

Without clear results from large studies, a highly individualised approach to each patient is essential. This takes time and involves a closer standard of care. Many additional common sense aspects of individualising treatment are only possible by using supportive technology such as therapeutic drug monitoring, resistance tests together with expert interpretation and adherence support.

Montaner also introduced the importance of the ‘virtual virus’ as the cumulative sum of worst resistance profile of an individual over time. This requires longitudinal resistance testing, timing of tests (ie when on-treatment) and a full treatment history. Even without baseline resistance tests, if a treatment history implies earlier resistance, then it is important to remember this when reading resistance test results from current treatment. A drug that shows sensitivity on a resistance test is not the same as a drug that is going to be active in any given situation. Resistance tests are only able to indicate resistance to drugs that a patient is using when the sample was taken. Even then, they only provide guidelines for subsequent treatment, and are dependent on sensitive drugs being available.

Montaner’s results using mega-HAART, supported by drug level monitoring with up to nine drugs (median 6, IQR 5-7) in a cohort of 248 multiply drug resistant patients, saw 69% on ITT analysis achieve viral load <400 copies/ml on at least two consecutive counts and 35-40% reach <50 copies/ml at week 48. [2] He also reported that this response is sustained out to 24 months in 80% of his patients who achieve undetectable viral load, and that aiming for this goal in a treatment experienced group is both realistic and sustainable. Even more optimistically he showed that this aggressive response to multiply drug resistant patients produced a Kaplan-Meier survival response similar to treatment naïve patients.

IQ and Virtual IQ

Numerous previous studies have shown that higher drug levels, especially of PIs, result in greater antiviral activity, and often sufficiently to overcome partial resistance and often shown in studies utilising the virtual IQ. Resistance is rarely an ‘on-off’ switch. The inhibitory quotient (IQ) is essentially the relationship of drug exposure to drug susceptibility of a particular pathogen.

Casado and colleagues from Ramon y Cajal Hospital, Madrid, found significant differences in response depending on whether IQ was > 1 in a prospective study of 51 PI-experienced patients using either indinavir/ritonavir or nelfinavir/saquinavir combinations. Virological response in patients who achieved IQ>1 rather than <1 was –1.68 vs –0.51 log (p=0.04) for nelfinavir; -1.1 vs -0.92 (p=0.03) for saquinavir, and -1.2 vs -0.83 log (p=0.09) for indinavir [3].

Many of the presentations at Barcelona again highlighted the importance of TDM to manage drug interactions used in salvage therapy.

Giga-HAART and drug levels

Two posters from the French Giga-HAART study looked at treatment interruption and drug levels together with multiple drug regimens. Although first presented in Athens last year, and notably not accepted for the Retrovirus conference in February, these results are sufficiently important to report again.

70 patients resistant to three classes of drugs were randomised to either change immediately to a new regimen containing upwards of eight drugs (immediate treatment arm) or to take an eight-week break from treatment before a similar regimen (deferred treatment arm). Giga-HAART regimens were constructed using 3-4 NRTIs + one NNRTI + hydroxyurea + ritonavir (400mg BID) + amprenavir (600mg BID) + another PI (either indinavir 400mg BID or saquinavir 600mg BID or nelfinavir 1250mg BID). This was an advanced population with median baseline viral load and CD4 of 5.3 logs and 27 cells/mm3 respectively. ARV treatment had been used for a mean 6.6 years with exposure to 11 ARV drugs.

An analysis of 63 patients at weeks 12/20 showed a highly significant difference in response between patients who interrupted therapy compared to those who used continuous treatment and these results were sufficient to require the study to close early. [4]

Further analysis of the role of adequate drug levels in the Giga-HAART study (above or below the recommended Cmin) a change in viral load of +0.4 log was common in patients from either arm who were found not to be achieving optimal drug levels. Patients with an adequate Cmin in the immediate treatment arm achieved a viral load reduction of –0.4 logs. In the deferred treatment arm, patients with adequate Cmin levels achieved decreases of –2.0 and –2.6 logs depending on whether they had experienced a shift to wild-type virus during the period off treatment. Patients in the deferred arm with low Cmin levels only achieved reductions of –0.7 logs. [5] As with other studies, sub-optimal drug levels in this study were clearly associated with minimal comparable responses.

Viral load responses by drug level

Immediate arm Deferred arm
Adequate Cmin -0.4 -2.0 to -2.6*
Sub-optimal drug level +0.4 -0.7

Cmin levels used in Giga-HAART Study:

Efavirenz ³ 1100ng/ml; nevirapine ³ 4000ng/ml; ritonavir ³ 2100 ng/ml; amprenavir ³1000 ng/ml; lopinavir ³3000 ng/ml; saquinavir ³250 ng/ml; indinavir ³150 ng/ml; nelfinavir ³1000 ng/ml.

Sequential monotherapy in salvage?

While Barcelona provided practical support for most of the approaches to salvage therapy that have already been reported in HTB over the last two years, and which are already the basis of the i-Base ‘Guide to Second-line and Salvage Therapy’, there was very little that was new for salvage therapy.

A few days earlier, at the XIth resistance workshop in Seville, a paper from Andrew Phillips from the Royal Free Centre for HIV Medicine, London, suggested a basis for the most radical theoretical approach to salvage therapy discussed this summer. [6]

Phillips suggested that the reduced replicative capacity for most drug resistant virus – all except some NNRTI-associated mutations seem markedly less fit in the presence of drug compared to wild type – may be sufficient to protect CD4 depletion.

Using a previously described model allowing for 128 subpopulations of short-lived cells and resistant combinations of seven ARVs, he calculated that sequential daily (or weekly) monotherapy was as likely to produce sustained 3 log viral suppression over three years as a continuous seven-drug combination. According to the model, under some circumstances, dual or triple therapy may be more likely to work than monotherapy, and it would probably be best to start with double or triple combinations. The only advantages of monotherapy are reduced toxicity and cost.

Importantly, this model worked even in a ‘worst case’ scenario when all 128 resistant subpopulations are present. Although viral resistance would be present on each day of treatment, with effective replicative ratio >1, with whichever drug is being used, the sequential regimen would remain effective because this subpopulation does not have time to grow sufficiently during the short period during which the drug is being taken. The model allows there to be different amounts of each of the 128 different types of virus at any one time. It would also work if 100% of virus is resistant, so long as that virus is sufficiently impaired as a result of those resistance mutations.

This approach is not currently being followed in practice, and one caution is that it may require more classes of resistant drugs than are currently available. However, if the rationale is plausible, this is a study that would work best for people who have already developed strongest resistance, who arguably have few other options, and who would therefore be most likely to want to try this option as they have already developed extensive resistance.

Several other studies at the Seville meeting looked at defining and harnessing reduced viral fitness for clinical benefit and proof of concept case studies will be followed with interest.

See separate reports in this issue on the Seville meeting.


In the context of salvage therapy, where increasing the numbers of agents in a combination almost always increases the likelihood of a better response – largely due to a heterogeneity of resistance viral population during early failure – the increased number of drug-drug interactions, makes using TDM to manage treatment experienced patients essential.

In the UK this is facilitated by the validated laboratory service provided at Liverpool University, especially as costs of TDM are subsidised by the manufacturers of amprenavir (Agenerase), indinavirReferences (Crixivan), lopinavir/r (Kaletra), nelfinavir (Viracept) and saquinavir (Invirase, HGC; Fortovase, SGC).


  1. Montaner J – New challenges and perspectives in salvage therapy.XIV International AIDS Conference, Barcelona 7-12 July. Abstract WeOrB192.
  2. Montaner J et al – Multi-drug rescue therapy (MDRT) in HIV+ individuals. XIV International AIDS Conference, Abstract B10484.
  3. JL Casado, A Moreno, K Hertogs et al. Individualising salvage therapy: the inhibitory quotient as a predictor of virological response. Barcelona 7-12 July. Abstract TuPeB4488.
  4. C Katlama, S Dominguez, C Duvivier et al. Benefits of Treatment Interruption (TI) in Patients with Multiple Therapy Failures, CD4 cells <200 /mm3 and HIV RNA >50 000 cp/ml (GIGHAART ANRS 097). XIV International AIDS Conference, Barcelona 7-12 July. Abstract WePeB5887
  5. G Peytavin, M Legrand, C Duvivier et al – Relationship between trough plasma concentrations of HIV antiretroviral drugs and virological response and mutations reversion in gighaart trial (anrs 097). XIV International AIDS Conference, Barcelona 7-12 July. Abstract TuPeB4568
  6. Phillips A et al – Theoretical rationale for the use of single drug sequential antiretroviral therapy regimens. XIth International Drug Resistance Workshop, Seville. July 2-5th 2002. Abstract 73. Antiviral Therapy 2002; 7:S61

Links to other websites are current at date of posting but not maintained.