Another possible model of treatment involves giving a highly potent combination therapy regimen for the first six or twelve months and then subtracting several drugs once most of the virus population has been eliminated. This is sometimes described as a 'step-down' approach to treatment, or likened to the treatment of certain tumours or infections, in which an 'induction' period of aggressive therapy is followed by a longer term phase of less intensive 'maintenance' therapy.

The case for induction / maintenance

The concept of induction and maintenance therapy, or subtractive therapy, is attractive for several reasons:

• It involves less exposure to potentially harmful drugs.

• It preserves future treatment options.

• It minimises the risk of side-effects or resistance.

• There are fewer tablets to take, helping with compliance.

• It is less expensive.

Despite the discouraging preliminary results with step-down or maintenance therapy, several studies have shown some endorsement of the step-down principle.

One study accidentally demonstrated that the concept of induction and maintenance is sound. Participants in a study of AZT (zidovudine, Retrovir), ddI (didanosine, Videx / VidexEC) and nevirapine (Viramune) who stopped taking ddI after at least six weeks of viral suppression below 20 copies/ml showed no evidence of viral rebound or resistance to nevirapine after 18 months' follow-up, suggesting that once maximal suppression has been achieved, a reduction in the number of drugs need not permit a resurgence in viral activity (Hall 1998).

A non-randomised, open-label study of induction treatment with saquinavir/ritonavir plus 3TC (lamivudine, Epivir) and d4T (stavudine, Zerit) in people with high baseline viral load (ranging from 56,000 to over 1,000,000 copies/ml) found that 11 of 17 maintained viral suppression after switching to d4T, ddI and nevirapine after between 17 and 61 weeks on maintenance treatment (Michelmore 1998).

Similarly, early results from the maintenance phase of an uncontrolled Spanish study support further study of this treatment strategy. Twenty-eight people with viral load above 50,000 copies/ml began a first-line combination of twice daily d4T, 3TC, nelfinavir (Viracept), saquinavir (Invirase / Fortovase) and nevirapine. Those with viral loads below 50 copies/ml after nine months then de-intensified to d4T, 3TC and nevirapine. Although only 16 patients completed the full induction phase, 15 of these remained below 50 copies/ml three months after the switch (Soriano 2000).

The case against induction / maintenance

However, several larger studies have shown no advantage to particular induction and maintenance protocols. The Trilege study in France, ACTG 343 in the United States, the MIRO study in Europe and the ADAM study in Holland have all demonstrated a higher 'failure' rate in people randomised to 'step down' to one or two drugs rather than continue on three drugs.

Most disappointing have been the final results of a Swiss study following interim results which found that a simplified triple regimen could extend viral suppression. The study enrolled 161 people who had viral suppression below 50 copies/ml for at least six months. Participants were randomised either to continue their protease inhibitor-containing regimen, or to switch to abacavir (Ziagen), AZT and 3TC. After one year of follow-up, eleven people on the triple nucleoside reverse transcriptase inhibitor (NRTI) arm had viral rebound compared to five on the protease inhibitor therapy, suggesting the virological superiority of the protease inhibitor-based regimen. Those who experienced rebound were more likely to have prior NRTI experience, suggesting that a switch to a triple NRTI regimen should be avoided in those who took dual NRTI therapy prior to commencing triple therapy. Not surprisingly, lipid levels improved in those switching to the triple NRTI regimen, and there were a larger number of discontinuations in the protease inhibitor group (Opravil 2000).

In the Trilege study, participants were randomised to remain on AZT, 3TC and indinavir (Crixivan) or switch to either AZT and 3TC or to AZT and indinavir if their viral load had fallen below 500 copies/ml after three months of triple therapy. The rebound rate was almost four times greater amongst participants randomised to AZT and 3TC or to AZT and indinavir compared with those who remained on triple therapy. Some had rebound of wild type virus, suggesting that rebound did not occur due to resistance. Instead, researchers have attributed virological failure to problems with adherence and lack of antiviral potency. Rebound was defined as one viral load measurement above 500 copies/ml (Descamps 2000; Pialoux 1998).

In ACTG 343, participants were randomised to remain on triple therapy, or to switch to either AZT and 3TC or to indinavir monotherapy if their viral load was consistently below 200 copies/ml at weeks 16, 20 and 24. Failure in this study was also defined as rebound above the limit of detection. ACTG 343 also showed that switching to a less intensive maintenance regimen, this time at 24 weeks, was inferior to the continuation of triple therapy (Havlir 1998). Viral rebound amongst patients whose treatment was eased to indinavir alone occurred without the emergence of detectable genotypic resistance. All had wild-type virus, suggesting either poor compliance or poor absorption of the drug in this sub-group.

Trilege and ACTG 343 have been criticised for flaws in their design. Many participants were AZT-experienced and hence effectively received 3TC or indinavir monotherapy. Furthermore, the level of suppression required before the switch has been criticised as insufficient. However, in the Trilege study, the failure of the two drug maintenance arms was also observed amongst a subgroup of people who had viral load below 50 copies/ml at randomisation to maintenance therapy. Another intriguing aspect of the ACTG 343 study was that people with lower increases in CD4 cell counts during induction had higher rates of sustained suppression of HIV.

In the MIRO study, 85 people were randomised to receive indinavir or ritonavir (Norvir)-boosted saquinavir, plus two NRTIs. After at least 16 weeks of treatment, people with viral loads below 400 copies/ml were randomised to either continue their initial regimen or to drop the protease inhibitor. Amongst the 40 patients who were underwent the second randomisation, those who dropped the protease inhibitor were significantly more likely to experience a viral load rebound.

In the ADAM study 43 patients who achieved undetectable viral load below 50 copies/ml on d4T, 3TC, nelfinavir and saquinavir were randomised either to continue with the existing regimen, to switch to two-drug maintenance with nelfinavir and either d4T or saquinavir. After ten weeks of maintenance therapy four of seven d4T recipients and five of seven saquinavir recipients had experienced viral load rebounds, compared to one of eleven of the quadruple therapy group.

After an interim analysis showed that maintenance treatment after 26 weeks produced an inferior response, the randomisation was postponed until week 50. Results from week 96 showed that this longer induction phase seemed to have no impact on response to the two-drug maintenance phase, although the study was limited by small numbers available for randomisation at week 50. Length of time to viral rebound among those on maintenance therapy was not affected by the longer induction period. The authors concluded that this was not surprising given the evidence that the failure of the step-down approach stemmed from the lack of antiviral potency of two-drug maintenance therapy in the presence of ongoing viral replication (Reijers 1998, 2001).

A study of 60 people who had viral loads below 50 copies/ml for over one year and CD4 cell counts above 200 cells/mm³ found that a switch to ddI and hydroxycarbamide (Hydrea) led to rapid viral rebound in most individuals. Only two of 30 people who switched had continuous viral suppression compared with the entire control group. The trial required that HAART be re-started when viral load rose above 5000 copies/ml. However, at 16 weeks, intent-to-treat analysis showed that only 54% of the ddI and hydroxycarbamide arm had viral loads below 5000 copies/ml. Furthermore, CD4 cell counts fell in 10% of the control group and 56% of the hydroxycarbamide group (Barreiro 2000).

Can it work?

Despite the increasingly gloomy picture surrounding induction and maintenance therapy for HIV, the failure of the subtractive approach in these studies does not necessarily mean that the fundamental concept is flawed. It has been suggested that subtractive therapy may stand a better chance of success if:

• A more aggressive initial regimen is used.

• Participants are not switched until they have been below the limit of detection for long period, in order to reduce viral reservoirs much further.

• Participants are required to achieve a greater degree of viral suppression before switching (e.g. below 50 copies/ml using sensitive assays).

• A more potent maintenance regimen, such as two protease inhibitors, or two drugs which can be given once a day such as efavirenz (Sustiva) and ddI, could be used.

• Participants are switched to a completely new regimen.

However, several of the above studies indicate that viral rebound occurs at higher rates among those on maintenance therapy due to a lack of antiviral potency. That is, the step-down regimens are often too weak to suppress HIV replication.

Researchers are now looking at single ritonavir (Norvir)-boosted protease inhibitors (lopinavir [Kaletra] and indinavir [Crixivan]) as maintenance agents, using no NRTIs. Part of the attraction of this strategy is that it avoids the long-term use of protease inhibitors and NRTIs, a combination which carries the highest risk of lipodystrophy. A Swiss / German study has reported no virologic rebound above 500 copies/ml in eleven patients followed for at least 48 weeks on ritonavir-boosted indinavir maintenance treatment (Hupfer 2003).

Updating the approach

Recent studies have tested the strategy of starting treatment with a four-drug multi-class regimen, and then stepping down to triple therapy after variable periods.

In the Forte study, 62 treatment-naive patients were randomised to receive two NRTIs, one non-nucleoside reverse transcriptase inhibitor (NNRTI) and one protease inhibitor for 24 to 32 weeks, until viral load reached below 50 copies/ml. This was followed by standard therapy of two NRTIs and one NNRTI. They were compared with 60 treatment-naive patients randomised to receive standard dual NRTI, single NNRTI HAART.

Adherence after 48 weeks was better in the standard therapy arm (89%) compared with 83% in the induction and maintenance arm, in which 19 patients (31%) stopped more than one drug before 24 weeks, including 15 for toxicity. In intent-to-treat analyses, however, more patients on the standard therapy arm had virological failure at 24 and 32 weeks. After 48 weeks, the mean fall in viral load was 0.86 log₁₀ greater in the induction/maintenance arm, more patients had viral loads below 50 copies/ml, and the median increase in the CD4 cell count was slightly greater. After an average of 81 weeks follow-up, there were no significant differences in the number of patients with serious adverse events, grade 3 and 4 adverse events, adverse events leading to stopping an HIV drug or progression to AIDS or death between the two arms (Williams 2004).

In the TIME study, patients were treated with AZT, 3TC, abacavir and efavirenz for 48 weeks, before randomisation to continue with the four-drug regimen or to remove efavirenz. Despite low tolerability of the initial four-drug combination, which may have been due to the advanced disease of the patient cohort, intent-to-treat analysis revealed that the two approaches were equivalent. However, this may have been due to a high failure rate in the three-drug arm, balanced by the low tolerability of the four-drug arm (Johnson 2003). A similar approach was tested in the ESS40013 study, with no significant differences between the two arms of the study 48 weeks after withdrawal of efavirenz from the maintenance arm in terms of the proportions of patients with viral loads below 50 copies/ml and time to treatment failure. However, the patients who stopped taking efavirenz reported side-effects less often and improvements in blood fat levels, as well as higher rates of adherence (Markowitz 2005).

Stepping down to fewer than three drugs

Step-down approaches do not always require the use of four drugs at the outset. Several recent studies have demonstrated successful responses when patients go from three drugs to two drugs or even to one drug.

The OK study is the most radical approach to induction and maintenance yet tested. In this study, 42 patients with viral loads below 50 copies/ml after at least four weeks taking Kaletra plus two NRTIs were randomised to receive Kaletra 400 / 100mg twice daily alone, or to continue with the triple drug regimen.

Patients in both groups maintained viral suppression and CD4 cell counts over 24 weeks. Although three of the patients in the Kaletra-only arm experienced virological failure, genotypic analysis revealed that this was not due to the development of PI resistance mutations. Patients who stopped taking NRTIs had significant improvements in red blood cell counts but there was no significant difference in lipid levels (Arribas 2004).

Similar findings have been reported from a pilot study of patients switching from an NNRTI-based regimen to Kaletra monotherapy. At baseline, all 18 patients had undetectable viral loads and were taking one NNRTI and two NRTIs. Two weeks after replacing the NNRTI with twice-daily Kaletra at a dose of 400 / 100mg the NRTIs were stopped, leaving only Kaletra. Twenty-four weeks later, 14 (78%) of the patients remained in the study, 13 (72%) of whom had maintained their viral load below 75 copies/ml.

Of the remaining four patients, three left the study due to diarrhoea, while the fourth developed the protease inhibitor resistance mutation M36I and withdrew. Five of the patients remaining within the trial for 24 weeks exhibited a rise in blood lipids that required the addition or increased dosing of lipid-lowering agents. This study is continuing for another 48 weeks (Pierrone 2004).

Another small study of Kaletra monotherapy as maintenance treatment added the drug to an existing HAART regimen in patients who had undetectable viral load for at least nine months. After two weeks lopinavir levels were measured and all other drugs were discontinued if trough lopinavir levels above 3µg/ml were recorded. Preliminary results showed that all 16 patients had viral load below 400 copies/ml at week 16, and all but one had viral load below 75 copies/ml at this point (Ruane 2004).

Simplification regimens

Simplification regimens are designed to offer equal antiviral potency, but fewer adverse events, lower pill burden and exposure to fewer antiretroviral drug classes, and so differ from the original concept of induction and maintenance.

This line of research has been pursued since it became clear that induction and maintenance strategies were likely to be unsuccessful.

Simplification has usually consisted of switching from PI-containing regimens to NNRTI or abacavir-containing regimens, and these studies are discussed in the following sections:

• Abacavir - key research in Drugs used by people with HIV: Nucleoside reverse transcriptase inhibitors.

• Efavirenz - key research in Drugs used by people with HIV: Non-nucleoside reverse transcriptase inhibitors.

• Nevirapine - key research in Drugs used by people with HIV: Non-nucleoside reverse transcriptase inhibitors.

• Options for simplifying therapy in Anti-HIV therapy: Choosing a combination, specifically with reference to simplification to once daily regimens.

• Anti-HIV therapy: Body fat and metabolic changes whilst on treatment.

References

Arribas JR et al. Simplification to Lopinavir/r single-drug HAART: 24 weeks results of a randomized, controlled, open label, pilot clinical trial (OK Study). Fifteenth International AIDS Conference, Bangkok, abstract TuPeB4486, 2004.

Barreiro P et al. Hydroxyurea plus didanosine as maintenance therapy after 1 year on highly active antiretroviral therapy. AIDS 14: 207-208, 2000.

Clumeck N et al. MIRO: a randomised trial of three maintenance regimens given after four months of triple or quadruple induction therapy. Fourth Congress on Drug Therapy in HIV Infection, Glasgow, abstract OP2.1, 1998.

Cooper D et al. Therapeutic strategies for HIV infection - time to think hard. N Engl J Med 339: 1319-1321, 1998.

Descamps D et al. Mechanisms of virologic failure in previously untreated HIV-infected patients from a trial of induction-maintenance therapy. JAMA 283: 205-211, 2000.

Hall D et al. Induction-maintenance antiretroviral therapy: proof of concept. AIDS 12: F41-F44, 1998.

Havlir DV et al. Maintenance antiretroviral therapies in HIV-infected subjects with undetectable plasma HIV RNA after triple-drug therapy. N Engl J Med 339: 1261-1268, 1998.

Hupfer M et al. Pilot study: ritonavir boosted indinavir treatment as a simplified maintenance monotherapy for HIV infection. Second International AIDS Society Conference on HIV Pathogenesis and Treatment, Paris, abstract 589, 2003.

Johnson M et al. Induction therapy with Trizivir (zidovudine/lamivudine/abacavir) (TZV) plus efavirenz (EFV): TIME study (AZL30004) results at 24 weeks. Ninth European AIDS Conference, Warsaw, abstract F1/4, 2003.

Markowitz M et al. Induction with abacavir/lamivudine/zidovudine plus efavirenz for 48 weeks followed by 48-week maintenance with abacavir/lamivudine/zidovudine alone in antiretroviral-naive HIV-1-infected patients. J Acquir Immune Defic Syndr 39: 257-264, 2005.

Michelmore HM et al. Successful induction/maintenance antiretroviral therapy in treatment-naive HIV-infected subjects - a pilot. Fourth Congress on Drug Therapy in HIV Infection, Glasgow, abstract P135, 1998.

Opravil M et al. A randomized trial of simplified maintenance therapy with abacavir, lamivudine, and zidovudine in human immunodeficiency virus infection. J Infect Dis 185: 1251-1260, 2002.

Pierone G et al. Simplification to lopinavir/ritonavir monotherapy from NNRTI-based HAART in HIV-infected patients with complete viral suppression. XV International AIDS Conference, Bangkok, abstract TuPeB4595, 2004.

Pialoux G et al. A randomised trial of three maintenance regimens given after three months of induction therapy with zidovudine, lamivudine, and indinavir in previously untreated HIV-1-infected patients. N Engl J Med 339: 1269-1276, 1998.

Reijers MHE et al. Maintenance therapy after quadruple induction therapy in HIV-1 infected individuals: Amsterdam Duration of Antiretroviral Medication (ADAM) study. Lancet 352: 185-190, 1998.

Reijers MHE et al. The ADAM study continued: maintenance therapy after 50 weeks of induction therapy - research letter. AIDS 15: 129-130, 2001.

Ruane P et al. Maintenance therapy using lopinavir/ritonavir alone with well controlled HIV infection. Fifteenth International AIDS Conference, Bangkok, abstract TuPeB4577, 2004.

Soriano V et al. Induction-maintenance (5 --> 3 drugs) in HIV-infected patients with high viral load. AIDS Patient Care STDS 14: 573-574, 2000.

Williams I et al. A Virological Benefit from an Induction/Maintenance Strategy Compared with a Standard 3-drug Regimen in Antiretroviral Naﶥ Patients: the FORTE Trial. Eleventh Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 564, 2004.