Abstract 957 Poster Session IV, Tuesday, 5/4 (poster 101)

There is limited information on the pharmacodynamic relationship of antiretroviral agents to HIV-1 viral load response in children receiving HAART. Intensive NFV pharmacokinetics (PK) was performed 4 weeks after randomization to 3 of 4 protocol treatment arms containing NFV in antiretroviral experienced children aged ≥4 months to 17 years enrolled in PACTG 377. Association of the NFV PK with HIV-1 RNA response at weeks 4, 8, and 12 was evaluated in these children. NFV concentrations were determined using a validated HPLC assay and PK parameters estimated using noncompartmental methods. Suppression of HIV replication was monitored by quantitative HIV-1 RNA PCR measurement (Roche Amplicor). Responses were assessed in 38 children receiving NFV with PK data available, including 7 receiving NFV BID. NFV AUC or Cmin (typically obtained 10-15 hours following TID dosing) were correlated with % of children with virologic success, undetectable HIV-1 RNA (<400 copies) at week 12 or a change in RNA from baseline to week 12.

Correlation of baseline characteristics with viral response at week 12 in the 96 children (including the 38 with PK data available) assigned to a NFV-containing treatment arm revealed that baseline viral load was strongly associated, p<0.001, with virologic success. Univariate analysis of PK parameters in the 38 children with PK indicated an association of very low (≤150) NFV Cmin at week 4 with proportion with undetectable RNA at week 8 (13% vs. 62%), p=0.019 and week 12 (13% vs. 73%), p=0.003. Correlation of NFV AUC with viral response did not achieve statistical significance at these timepoints. Multivariate analysis in this group of children confirmed that baseline viral load in a logistic model correlated with viral response, p=0.028, whereas NFV Cmin did not, p=0.39. Additional analysis indicated that a higher proportion, 71% (12/17), of small children (<25 kg) achieved lower Cmin (≤ median of the entire Pk group) than children ≥25 kg, 33% (7/21), but was of borderline statistical significance (p=0.049 by Fisher's exact test) in this preliminary analysis.

We conclude that high baseline viral load was a strong predictor of viral response and that very low NFV Cmin values at week 4 were associated with a low rate of undetectable RNA at week 12. This association of a lower rate of virologic success with very low Cmin may reflect problems with adherence or prolonged overnight dosing intervals in children on TID dosing with resulting low trough levels may suggest closer adherence to TID dosing or support BID dosing to improve response. This evaluation is not definitive but suggests that the NFV dose may be low for children <25 kg, although viral response did not vary with weight. Analysis of children subsequently enrolled on this protocol is ongoing. Further correlation of viral response to measures of NFV and other antiretroviral agent exposure is needed.