Therapeutic interventions in HIV infection – a critical view


correspondence Janet Darbyshire, Medical Research Council, Clinical Trials Unit, 222 Euston Road, London NW1 2DA, UK. E-mail:


Summary The introduction of highly effective combination regimens of antiretroviral drugs has led in recent years to substantial improvements in morbidity and mortality. As yet immune-based therapies have had little if any impact. However it is clear that eradication of HIV is not achievable with existing anti-HIV drugs and in spite of the major advances there remain many challenges in the clinical management of HIV-infected individuals. These benefits are unlikely to be extended to resource poor countries in the foreseeable future. The barriers are primarily but not only the costs. In resource-rich countries there are concerns about long term toxicities and many people have already exhausted all of the current therapeutic options. There is an urgent need for new drugs, ideally attacking new targets or with no cross resistance to existing drugs, and which are well tolerated and safe, easy to take and cheap. Many important questions still remain unanswered, in particular when to start antiretroviral therapy.


In HIV-infected individuals the primary target of therapy is the human immunodeficiency virus (HIV-1), but most of the clinical manifestations are related to the effect of HIV on the immune system, which leads to progressive immunodeficiency. Consequently the other potential therapeutic interventions are aimed at the immune system or the prevention and treatment of the opportunistic infections (OIs) and tumours which define AIDS and are the main causes of mortality and morbidity. Improvements in morbidity and mortality were seen before the development of effective antiretroviral drugs, due to better prevention, diagnosis and treatment of AIDS-defining events. Recently, the introduction of highly effective combination regimens of antiretroviral drugs has led to substantial improvements in morbidity and mortality but as yet immune-based therapies have had little if any impact. The most promising immune-based therapy, which is currently in large scale clinical trials, is recombinant interleukin 2 (rIL-2). This has been shown in a number of trials to result in substantial and sustained increases in CD4 cell counts ( Emery et al. 1998 ). Current trials will determine whether these increases in CD4 cell count lead to improvements in mortality and morbidity concurrently with combination antiretroviral therapy (ART). Another approach, which is currently under investigation, is the use of therapeutic vaccines. Inactivated HIV has been shown to have promising effects on the immune system but as yet there are no data showing clinical benefit. It is clear that eradication of HIV is not achievable with existing anti-HIV drugs and therefore the use of immune-based therapies either to increase the durability of viral suppression or to eradicate HIV from latent reservoirs is being explored by many groups.

Antiretroviral drugs

Several classes of anti-HIV drugs have been developed which target different stages of the viral replicative cycle. Some of these have proved to be ineffective and have not been developed further and there are currently only three classes of drugs in routine clinical use with others at different stages of development. The first drugs which were developed, the nucleoside analogue reverse transcriptase inhibitors (NRTIs), act as inhibitors of and substrates for the HIV reverse transcriptase (RT) enzyme, preventing the formation of proviral DNA. There are now six NRTIs licensed and several more in different stages of evaluation. The second class of drugs also attacks the RT but by different mechanisms and they are usually referred to as non-nucleoside RT inhibitors (NNRTIs). The third class of drugs acts by inhibiting the HIV protease enzyme (PIs). There are now 12 drugs from these three classes licensed and in routine clinical practice in Europe ( Table 1); most have to be given at least twice a day, many require several tablets or capsules to be taken at each dose and most are associated with both major and minor toxicity. Moreover, there is considerable cross-resistance between drugs of the same class, particularly the NNRTIs and PIs, although not between classes. Even with combinations of three or four drugs, which can achieve substantial and sustained viral suppression if taken appropriately, there is no evidence that eradication or cure is an achievable target with current therapies nor that treatment can be stopped without viral rebound ( British HIV Association (BHIVA) 2000).

Table 1.  Antiretroviral drugs in routine clinical use Thumbnail image of

Therapeutic failure

Therapeutic failure is becoming an increasingly common problem and may be due to one or more reasons. Problems of adherence to therapy are probably the most common reason for failure, but the development of viral resistance and pharmacological issues such as variations in bioavailability or interactions between antiretroviral drugs or with other drugs may play a role. The search continues both for new drugs with different modes of action and for new drugs from existing classes but with better bioavailability (and therefore a smaller pill burden), or a longer half life (permitting once daily dosage) or with increased potency, less toxicity or a different resistance profile. The introduction of combined preparations may also aid adherence and reduce the risk of failure, as in other diseases such as tuberculosis.

Developments in antiretroviral therapy (ART)

Since the first drug, zidovudine (AZT, ZDV) became available in the late 1980s there have been a number of key developments in therapy. The first trials of AZT showed a substantial short-term reduction in mortality in people with AIDS but subsequent trials in earlier stages of infection, particularly asymptomatic individuals, showed no evidence of long-term benefit ( HIV Trialists' Collaborative Group 1999).

Similarly, monotherapy with drugs from the other two classes showed short-term effects on laboratory markers (CD4 cell counts and viral load, measured by the levels of HIV RNA in the plasma using molecular techniques) but these were not sustained. Combinations of two NRTIs – AZT plus didanosine (ddI) or zalcitabine (ddC) or lamivudine (3TC) – were shown in several trials to result in substantial reductions in disease progression and mortality compared to AZT monotherapy, particularly in people who had not taken prior AZT ( HIV Trialists' Collaborative Group 1999) and were introduced into routine practice in the mid-1990s. The next step forward was the demonstration that the addition of a PI to two NRTIs led to substantial and sustained suppression of viral load and increase in CD4 cell count and to further reduction in disease progression ( Hammer et al. 1997 ). Subsequently triple combinations, in which the PI is replaced by an NNRTI or, more recently, a third NRTI, have been shown to have similar effects on laboratory markers ( Montaner et al. 1998 ; Staszewski et al. 1999 ).

The term ‘highly active antiretroviral therapy’ or ‘HAART’ has been widely used to refer to triple combinations including either a PI or an NNRTI which are now recommended as treatment of choice in most national and international guidelines ( British HIV Association (BHIVA) 2000; Panel on Clinical Practices for the Treatment of HIV Infection 2000; Carpenter et al. 2000 ) and have been widely adopted since the late 1990s. Several observational studies have shown marked reductions in deaths from AIDS and hospital admissions for OIs which are temporally associated with the introduction of HAART in the US and Europe although there is some evidence that these declines may now be levelling off ( Mocoft et al. 1998 ; Palella et al. 1998 ).

Time of initiation of ART

In spite of major advances in the therapy of HIV infection in the last 10 years, there remain several key questions concerning the optimal approach to therapy. First, there are no data which define the best time to start ART. Those who advocate starting early, even at the time of primary infection if this can be identified, argue that early intervention provides the best chance of immune reconstitution, of sustained suppression of viral replication and, potentially, of eradication. However, early optimism that eradication might be feasible was based on the profound viral suppression achieved by potent regimens when started close to primary infection. Subsequent studies have identified long-lived latently infected cells and current estimates suggest that maximally suppressive therapy may need to be continued for many years if HIV release from these cells cannot be controlled by the immune system ( Finzi et al. 1997 ; Wong et al. 1997 ; Ho 1998). The arguments against early intervention relate primarily to the difficulties of maintaining adherence to complex schedules, particularly in the presence of major or minor toxicity, and to concerns that the development of therapeutic failure with the emergence of viral resistance may rapidly exhaust the limited therapeutic options. In view of the long clinical course of HIV infection – even without ART the median survival is around 10 years and individuals may remain asymptomatic for most of that time – the initiation of therapy in the asymptomatic phase with all of the associated difficulties may be both undesirable and unsuccessful. As in many other chronic diseases, particularly where drugs with unpleasant side-effects are given to people without symptoms, poor adherence is the major threat to successful ART. Although there have been no randomised trials on when to start therapy since the trials of AZT monotherapy, a recent analysis of data from observational cohorts suggested that the virological response was similar whether the CD4 count was 200–349 or 350 and above and that increases in CD4 cell count occurred even if the baseline value was very low ( Phillips 1999). Further, of those who started therapy, more than a third had stopped at least one drug within a year, largely for toxicity or adherence problems. Reflecting the lack of data on which to base recommendations about when to start, national and international guidelines vary considerably. For example the UK guidelines are more conservative than those of the International AIDS Society ( British HIV Association (BHIVA) 2000; Carpenter et al. 2000 ).

Initial therapy regimens

There are limited data from randomised trials on what regimen to start with although these are largely relatively short-term and based on laboratory markers. The Atlantic study provides a direct comparison of the three types of triple combination, namely two NRTIs plus either a PI, an NNRTI or a third NRTI ( Murphy et al. 1999 ). However, the participants had relatively early disease with a median CD4 count of over 400 cells/mm3. At 48 weeks, the proportion with a viral load, measured by plasma HIV RNA, below the limits of detection of the standard assay (500 copies per ml), was similar in the three groups but there was a suggestion that the proportion below the detection limit of the more sensitive assay (50 copies/ml) was similar in the PI and NNRT1 groups but lower in the triple NRTI group. There may be differences between drugs of the same class in terms of potency as well as in toxicity, which means that caution is required when extrapolating from the results of one triple combination to others based on different drugs. Treatment guidelines are more consistent on what to start with, reflecting the availability of some data on at least the PI- and NNRTI-containing triple regimens, but also highlighting the limited data on combinations of three NRTIs and also of four drug combinations including all three classes ( British HIV Association (BHIVA) 2000; Panel on Clinical Practices for the Treatment of HIV Infection 2000; Carpenter et al. 2000 ). Several large trials are currently underway which address this key question.

Therapeutic failure and change of therapy: when to change

In the absence of a completely effective ART regimen, the second major therapeutic issue is when to change treatment and what to change to. Therapeutic failure undoubtedly occurs with all of the currently available regimens and is most often due to problems of adherence or pharmacological issues such as variations in bioavailability or interactions between ART drugs and with other drugs. In the absence of evidence from randomised trials, the decision to change may be based on the development of virological, immunological, or clinical failure alone or in combination. Virological failure is usually defined as a failure of the viral load to drop below a certain level or as a rebound in viral load after successful suppression. It frequently predates immunological failure, defined by CD4 cell count changes, or clinical failure. The definition of virological failure may be influenced by several factors such as the sensitivity of the HIV RNA assay, the clinical status of the individual, in particular whether or not they have had prior ART, and the viral load at initiation of treatment. BHIVA guidelines currently define first virological failure (i.e. in those previously ART naive) as a plasma HIV RNA (confirmed on two occasions at least two weeks apart) > 50 copies/mL at 24 weeks or > 50 copies/mL at any time if previously below 50, after an assessment of adherence and pharmacokinetics ( British HIV Association (BHIVA) 2000).

What to change to

The choice of regimen to change to has, to date, largely been made on the basis of prior drug history. Resistance assays are now becoming increasingly widely available. These are either genotypic assays, which identify specific changes in the HIV genome which are likely to be associated with resistance, or phenotypic assays, which directly measure the ability of the HIV to replicate in the presence of the drug in a range of concentrations. There are a number of issues which need to be addressed both technically – in relation to variability between the assay techniques, their ability to detect minority species, their reliability and the link between the result and clinical outcome – and practically – in terms of resources, both financial and personal, and time. The potential advantage of using a resistance assay is to aid the selection of more potent regimens, based on drugs to which the virus is still sensitive. The potential disadvantages are that therapy may be changed unnecessarily and the individual may therefore rapidly run out of options, which at present are limited. The randomised evidence to date on the use of genotypic assays in highly ART-experienced individuals suggests at least a short-term benefit in terms of viral load but data on the longer-term risks and benefits are needed to fully evaluate their value and optimal role in management ( Baxter et al. 1999 ; Durant et al. 1999 ).

Drug-sparing approaches

Although the current triple ART regimens are a major therapeutic advance, the considerable difficulties experienced by people in adhering to complex schedules for long periods of time has led to several approaches to develop drug sparing regimens. Maximizing the potency of available drugs, whether by the use of potentiators such as hydroxyurea, or by therapeutic drug monitoring are both being explored ( Barry et al. 1998 ; Rutschmann et al. 1998 ; ACTG 1999). Structured treatment interruptions or ‘drug holidays’ are currently being investigated in several scenarios. First, there is anecdotal evidence that if therapy is temporarily interrupted in individuals treated near the time of primary infection and whilst their viral load is below the limit of detection, the level of viral rebound may be reduced in successive cycles of therapy followed by further interruptions ( Garcia et al. 2000 ). Secondly there is some evidence that, in individuals who have failed after several courses of therapy and exposure to all three classes, a period without drugs before starting a multidrug regimen or ‘mega HAART’, may lead to the reappearance of wild type virus and potentially improve the effect of the mega-HAART ( Miller et al. 1999 ). The clinical status and quality of life of the individuals may also improve at least in the short term. Thirdly, evidence that CD4 increases are sustained even if viral suppression is not maintained, has raised the possibility that a “drug holiday” might be used to give people time off therapy without developing clinical disease progression, restarting when the CD4 cell count begins to decline ( Deeks et al. 2000 ). All these approaches remain as yet unproven and are not recommended outside randomised trials.

Cyclical therapy has only been explored with monotherapy regimens to date and has been disappointing ( The MRC Quattro Steering Committee 1999). It may be worth further investigation with combination regimens. Induction/ maintenance therapy (starting with an intensive phase of three or more drugs and reducing to a maintenance phase of fewer drugs) is an attractive concept. It was unsuccessful in trials which explored dropping from three or four to only two drugs, but may be more successful when dropping from four to three ( Havlir et al. 1998 ; Pialoux et al. 1998 ).

Treatment of children

The general principles of therapy developed for adults apply equally to children but there are a number of specific issues and problems in the clinical management of HIV-infected children. The number of drugs available is limited by difficulties and delays in the development of suitable formulations for children, particularly the PIs. The dosage for children, particularly the very young, may be much higher on a mg/kg basis than for adults and pharmacokinetic studies are a key part of drug development for children. There is no evidence that toxicity is more severe or more common in children but minor side-effects or factors such as the taste or volume of syrup may seriously affect the tolerability of the drug and hence adherence with the regimens. The use of laboratory markers to assess prognosis and the effects of treatment is more difficult in children. Young children have much higher CD4 counts than adults and these decrease with age as well as with disease progression in HIV-infected children. Viral load is very high in vertically infected infants and declines with age in those who do not progress rapidly to AIDS in the first year of life. Although it is more difficult to achieve complete viral suppression in children, they may respond very well to combination therapy.

Antiretroviral therapy in resource-poor countries

The introduction of potent antiretroviral therapy regimens in the US and Europe has led to improvements in morbidity and mortality which are unlikely to be extended to resource-poor countries in the foreseeable future. The barriers to ART are primarily, but not only, cost. These include not only the cost of the drugs but also the costs of monitoring response to therapy and toxicity and of supervision of therapy to achieve adequate adherence. The difficulties of ensuring adherence are compounded by the potentially serious toxicity associated with all antiretroviral drugs and the minor toxicity, which may reduce tolerability. There are major concerns that inadequate regimens and poor adherence will lead rapidly to failure of ART regimens with the emergence of resistant strains which, if widely transmitted, may further threaten the success of ART. One of the few areas where ART may be feasible in resource-poor countries is in the prevention of perinatal transmission. In the first trial, AZT monotherapy given from 14 weeks, to the mother intravenously during labour and orally to the baby for six weeks substantially reduced transmission ( Connor et al. 1994 ). Subsequent studies showed that shorter periods of AZT monotherapy or a combination of AZT plus 3TC also had a marked effect on transmission ( Saba 1999). Perhaps the most exciting recent development is the evidence that two doses of nevirapine, one to the mother in labour and one to the baby, halved the transmission rate ( Guay et al. 1999 ). This may be a practicable and affordable intervention although there are issues about the longer-term effects particularly in a breastfeeding population, as transmission can occur via breast milk, and concerns that misuse of nevirapine as monotherapy may lead to the rapid development of resistance.

Future concerns and challenges

There are a number of concerns for the future. First, if therapeutic failure with the development of resistance becomes widespread and leads to the transmission of resistant strains, this may jeopardise the success of currently potent regimens. This is particularly worrying where there is cross-resistance between drugs of the same class. Secondly, if drugs, which have to be taken for many years if not for life, are increasingly licensed on the basis of relatively short-term data, the longer-term risks of toxicity may not be detected. It has recently become apparent that severe disturbances of lipid and glucose metabolism and disorders of fat distribution are common in people taking triple therapy, particularly containing PIs ( Carr et al. 1998 ; Miller et al. 1998 ; Lo et al. 1998 ). These disorders may theoretically increase the risks of cardiovascular disease. However, metabolic disorders have been shown to be associated with HIV infection in the absence of therapy, with PI- and non PI-containing and with other ART regimens and with the total duration of therapy. The reported prevalence is variable depending on the definition and diagnostic methods. There is some evidence that the symptoms and signs may resolve if treatment is changed or stopped. Furthermore, the development of lactic acidosis is increasingly being recognized in individuals who have taken long-term NRTI therapy ( Brinkman et al. 1998 ; Andrew et al. 2000 ).

In spite of the major advances in the last few years, there remain many challenges in the clinical management of HIV-infected individuals even in resource-rich countries. Many people have already exhausted all of the current therapeutic options and there is an urgent need for new drugs ideally attacking different targets or with no cross-resistance to existing drugs and which are well-tolerated and safe. The potential benefits of immune-based therapy are yet to be realized. The most important question, namely when to start therapy, remains unanswered. There is an urgent need to develop ART regimens which ultimately could be used in resource-poor countries in parallel with the search for even more potent regimens which can achieve prolonged viral suppression and delay disease progression and death.