565    

Background: Prior attempts to treat patients with "induction-maintenance" therapy were abandoned due to unacceptable rates of virologic failure. Subsequent studies show that some subjects sustain viral suppression for years with a boosted PI-only regimen. We reasoned that HIV DNA and more sensitive measures of HIV RNA (VL) might predict a successful outcome with less potent regimens and evaluated these parameters in the induction-maintenance trial ACTG 343.

Methods:  Viral load was measured at week 24 prior to starting randomized maintenance therapy using a modification (nominal LOD 2.5 copies/mL) of the Roche Amplicor. HIV DNA was measured from stored baseline and week 24 PBMC using the Roche DNA assay. HIV DNA and viral load at baseline and week 24, CD4 count at baseline, treatment arm, and baseline resistance mutations in HIV RNA were examined as predictors of virologic failure during maintenance. Times to failure were estimated using the K-M method and were compared using the log-rank test.

Results:  Subjects who achieved VL <200 copies/mL after 24 weeks of ZDV/IDV/3TC (induction) followed by IDV only (N= 52) or ZDV/3TC (N= 54) were evaluated. Baseline median values were viral load, 13,000 copies/mL; CD4, 455 copies/mm³. A total of 33 subjects experienced failure (confirmed viral load >200 copies/mL) before restarting triple therapy. Patients who did not fail were followed for a median of 14 weeks on maintenance therapy. At week 24, 44/106 (42%) of patients had undetectable VL (LOD 2.5 copies/mL) and HIV DNA < 100 copies/µg. In an analysis of the time to failure, stratified by prior ZDV experience, HIV DNA at week 24 (p = 0.030) and a resistance mutation at codon 215 (p = 0.023) were independently associated with failure. Continued virologic suppression was observed in 35/44 (80%) of subjects with HIV DNA< 100 and undetectable viral load at week 24. In addition, all 16 subjects that had viral load and DNA measurements at baseline and at week 24 that were below the median maintained virologic suppression throughout maintenance therapy.

Conclusions:  Low HIV DNA levels and the absence of resistance at codon 215 were the strongest predictors of viral suppression during maintenance therapy. All subjects with a combination of low baseline and wk 24 VL and DNA maintained continued viral suppression during maintenance. Although follow-up is limited because the study was stopped prematurely, these findings provide insight for future studies evaluating efficacy of less intensive therapy into parameters that may identify patients most likely to succeed.

Keywords: treatment strategy; maintenance; HIV DNA