Background: Controversy exists about the impact of a treatment interruption on response to a salvage regimen. The objective of this analysis was to explore factors associated with short-term virologic response to LPVr in experienced patients (pts) who were currently on or off therapy when a LPVr regimen was begun.

Methods: This is an analysis from CCTG 578, an on-going, randomized, 3x2 factorial study of 3 adherence interventions crossed with therapeutic drug monitoring. LPVr levels drawn pre-, 2- and 4-hrs post-witnessed dose at wk 2, and randomly at wks 4 and 6, as well as phenotype (ViroLogic) and HIV RNA (week 1, 2, 4, and 6) were assessed. Treatment experienced pts were either on drug (ON) or had interrupted therapy (OFF) for ≥ 4 months when the LPVr regimen was initiated.

Results: The log₁₀ baseline HIV RNA was higher for patients OFF (5.2, n = 14) than ON (4.5, n = 16; p = 0.003). Mean baseline fold change in IC₅₀ to LPV was lower for pts OFF (0.7, range 0.3–1.6) versus ON (2.3, 0.5–9.2). The LPV C₁₂/IC₅₀ was greater for pts OFF (26) than ON (17). Duration of prior treatment and number of agents in the LPVr regimen were similar. HIV RNA reduction at wks 1–6 was significantly greater (p < 0.02) for patients OFF than ON (2.1 vs 1.2 at wk 6). This difference was not explained by introduction of baseline HIV RNA, LPV concentration or C₁₂/IC₅₀ into multivariate models. However, introduction of an interaction term (base RNA * group (ON vs OFF)) abrogated statistical significance of ON status. The mean viral load achieved at wk 6 was similar in both groups (3.2 ON vs 3.0 OFF).

Conclusions: Viral load reduction with LPVr regimens in experienced pts appeared to be greater after a treatment interruption, but this difference was due to an interaction with baseline viral load. These data do not support use of interruptions to improve salvage responses.