Efficacy, safety and predictive factors of virological success of a boosted amprenavir-based salvage regimen in heavily antiretroviral-experienced HIV-1-infected patients

Abstract

Background. Amprenavir (APV) has been shown to be effective in naive or treatment-experienced HIV-1 infected patients. However, the safety and efficacy of the APV 600 mg/ritonavir (RTV) 100 mg twice a day (bid) combination, the usually recommended dosage for boosted APV, have been less well studied. We assessed the predictive factors associated with virological success of APV/RTV-based regimens. Methods. Patients in the PharmAdapt study receiving an APV/RTV-containing regimen were included in the study. The predictivity of covariates on virological response at 4 months was analysed according to the data analysis plan. We processed logistic regression using bootstrapping to allow several covariates in the models. Results. Forty patients received an APV/RTV-containing regimen, 38 of whom were male (95%). Risk factors were heterosexual contacts (four patients; 10%), homosexual contacts (31 patients; 78%), and intravenous drug use (four patients; 10%). Twenty-seven per cent of patients were Centers for Disease Control and Prevention Classification System (CDC) stage A, 38% were stage B and 35% were stage C. The median baseline CD4 count was 313 cells/μL [interquartile range (IQR) 211, 414], and the median baseline viral load was 4.4 log10 HIV-1 RNA copies/mL (IQR 3.7, 4.9). Patients were exposed to a median number of 7.5 (IQR 6, 9) drugs for a median number of 3.8 (IQR 3.3, 4.3) years. The baseline number of resistance mutations was 4 [IQR 3, 5 for nucleoside reverse transcriptase inhibitors (NRTIs), 1 (IQR 0, 2)] for non-nucleoside reverse transcriptase inhibitors (NNRTIs) and 6 [IQR 5, 8 for protease inhibitors (PIs)]. At month 4, median viral load decreased to 1.2 log10 copies/mL (IQR 0.3, 1.6); 50% of patients had a viral load < 200 copies/mL by intention-to-treat analysis. The number of APV resistance mutations was associated with viral load changes. Median APV concentration was 1750 ng/mL (IQR 1130, 2520). At month 4, using several cut-offs, neither APV concentration nor the genotypic inhibitory quotient was predictive of viral load changes. Baseline viral load and the number of protease mutations were associated with outcome. Conclusions. Efficacious APV concentrations need to be determined for antiretroviral-experienced patients. Baseline viral load and the number of mutations on the protease coding region (PRO) were associated with the virological outcome of APV/RTV-based regimens. © 2004 British HIV Association.