Three- or four-drug antiretroviral therapy is hard to take reliably and can be associated with short-term, and perhaps long-term side effects as we are learning with the metabolic toxicities reported in association with protease inhibitor use. Therefore, investigators are beginning to test a treatment strategy that features an aggressive initial combination followed by a scaled-back maintenance therapy. There is a scientific rationale to support this approach. Almost all of the virus that is measured in viral load assays is produced in productively infected cells -- activated, CD4+ lymphocytes. We know that HIV replicates rapidly in infected individuals with the viral life cycle lasting a little over one day, and the life-span of infected cells is approximately two days. Therefore, if viral replication can be totally inhibited it would not take a long time until all of the productively infected cells are eliminated from the body. What's left is a smaller population of chronically infected cells and latently infected cells that produce far less virus than that made by the productively infected cell population. Perhaps fewer drugs would be needed to stop viral replication in these cells.

A multi-center study in the Netherlands, the ADAM study, was designed to test this hypothesis. 62 treatment naïve patients with baseline viral loads > 1,000 copies/mL and CD4+ counts > 200 cells/mm3 were treated with an initial combination of d4T + 3TC + nelfinavir (750 mg tid) + saquinavir (600 mg tid) for 26 weeks. Those patients with viral loads < 50 copies/mL at weeks 25 and 26 were eligible to be randomized to receive one of three maintenance therapies: d4T + NFV, NFV + SQV, or continue on the original four-drug combination.

The patients' mean baseline viral load was 4.5 log (about 50,000) copies/mL and the mean baseline CD4+ count was 400 cells/mm3. At the present time, 31 patients have been randomized to a maintenance therapy and 25 have been followed for an additional 12 weeks. Of those patients who remained on the four drug induction regimen, 1/11 (9%) experienced a rebound in viral load to quantifiable levels, while 9/14 (64%) of those whose treatment was scaled back to a 2-drug combination had a rebound in viral load. The authors did not state which if either or the 2-drug therapies was more likely to fail.

The authors looked at several variables to determine if there were factors associated with viral load rebound among patients on 2-drug maintenance therapy. Baseline CD4+ counts and viral loads were not predictive of rebound, nor were changes in CD4+ or CD8+ counts on therapy. The most significant predictor of failure was the rate of clearance of HIV RNA after the initiation of therapy. 80% of those patients whose viral loads remained below quantifiable levels on maintenance therapy had a viral load < 50 copies after 4 weeks of induction therapy. However, it took 16 weeks for the viral loads to go below 50 copies/mL for 80% of patients whose viral loads rebounded on 2-drug maintenance.

This study was similar in approach to two others that were presented at the last Retrovirus Conference in Chicago. In ACTG 343, patients were given induction therapy with ZDV + 3TC + IDV for 24 weeks, and were randomized to ZDV + 3TC or IDV monotherapy, or continued on the three drug combination if viral loads were < 200 copies/mL at weeks 16, 20 and 24. Those patients assigned to the less intense maintenance therapies were much more likely to experience a rebound in their viral load above 200 copies/mL than those who continued on the three drugs. In the TRILEGE study, patients also had induction therapy with ZDV + 3TC + IDV. After only 16 weeks of induction, patients were randomized to continue on 3-drug therapy or a scaled back maintenance therapy if their viral load was < 50 copies/mL. Again patients on the less intense treatment were more likely to have viral load rebound.

ACTG 343 and TRILEGE differed than the ADAM study in several important aspects. In the former two studies induction therapy included three drugs rather than four. In the ADAM study patients were required to have viral loads < 50 copies/mL to be eligible for randomization to the maintenance phase compared to < 200 copies/mL in ACTG 343. It is now known that some of the patients who failed on maintenance therapy on ACTG 343 had quantifiable viral loads between 50 - 200 copies/mL. Despite a potentially more aggressive induction phase in the ADAM study compared to the others (although there really is no data comparing indinavir as the only PI in a 3-drug combination with NLV + SQV as a PI combo in a 4-drug regimen), it is clear that the induction-maintenance strategies followed by these three studies are not adequate to provide durable suppression of virus replication. Future induction-maintenance studies will include more aggressive induction therapy for longer periods of time, followed by 3-drug maintenance therapies that may be more easily tolerated than combinations including protease inhibitors.