Successful therapy following viral rebound has been problematic. Intensification of the existing regimen by adding 1 or 2 drugs generally has been avoided. However, successfully adding new drugs to an existing regimen would be advantageous since it would expose the patient to fewer antiretroviral agents in the overall treatment course. Recent evidence suggests that a significant proportion of patients who experience viral rebound while receiving a protease inhibitor (PI) actually have viral rebound with a PI-sensitive virus. Other studies have shown that treatment decisions based on resistance assays result in better virologic outcomes. This trial examines further the effect of resistance assay-directed intensification of a PI-containing antiretroviral regimen on viral load. [AS PER AMENDMENT 04/03/01: The antiretroviral regimen need not contain a protease inhibitor.]

Patients are stratified by baseline plasma HIV-1 RNA levels (5,000 copies/ml or less versus greater than 5,000 copies/ml). Patients undergo phenotypic drug resistance testing prior to study entry. Based on the phenotypic results, patients are [AS PER AMENDMENT 11/9/00: selectively] randomized equally to

1 of 3 [AS PER AMENDMENT 11/9/00: 1 of 2] intensification strategies while remaining on their current, initial [AS PER AMENDMENT 11/9/00: (background)] antiretroviral therapy (ART). [AS PER AMENDMENT 04/03/01: ART need not be initial.] A patient is excluded from randomization to an arm if that arm contains a drug to which the patient has phenotypic resistance. Arm A adds abacavir (ABC). Arm B adds amprenavir (APV) [AS PER AMENDMENT 11/9/00: and ritonavir (RTV)]. Arm C adds didanosine (ddI) plus hydroxyurea (HU). [AS PER AMENDMENT 11/9/00: Arm C has been discontinued.] A patient's HIV must be sensitive to at least 3 drugs. [AS PER AMENDMENT 11/9/00: Each patient must be taking at least 3 drugs to which his/her HIV isolate is sensitive, including ABC or APV and at least 2 other drugs that are part of the current, initial (background) ART. If phenotypic resistance testing at screening indicates resistance to a nucleoside reverse transcriptase inhibitor (NRTI) drug in the patient's current, initial (background) ART, then the local investigator may choose to discontinue that drug. However, the patient and local investigator may choose to continue the drug but it will not be considered an active drug per this protocol.] [AS PER AMENDMENT 04/03/01: ART need not be initial.] Patients have regular clinic visits for physical exams and blood tests, including CD4 and CD8 cell counts, plasma HIV-1 RNA assays, and tests for pharmacokinetic variability. In the event of viral rebound of 500 copies/ml or more at Week 12 or later, phenotypic/genotypic drug resistance is assayed. In addition, phenotypic drug resistance is assayed at the primary endpoint (Week 24) and at the end of treatment (Week 48) on all patients.