A univariate and a multivariate linear regression model were used to investigate factors RGFP966 correlated with ATV plasma concentration.
Receiver operating characteristic (ROC) curves were used to identify a plasma ATV concentration cut-off predictive of virological response and toxicity. Predictors of virological outcome were investigated using a logistic regression model: variables significantly associated with virological response in univariate analysis were subsequently evaluated in a multivariate model. Spearman correlation coefficients (r) were calculated to evaluate the correlation between ATV concentration and bilirubin levels. All analyses were performed using the spss version 13.0 software package (SPSS Inc., Chicago, IL, USA). A total of 115 plasma samples from 86 patients fulfilling the selection criteria were analysed. Baseline features of the studied population, also split according to the use of ritonavir boosting, are shown in Table 1. Groups of patients
taking boosted or unboosted ATV differed for time since ATV initiation [median 11 months (interquartile range (IQR) 5–20 months) vs. 7 months (IQR 2–12 months), respectively; P=0.041] and concomitant tenofovir use (87.9 vs. 25%, respectively; P<0.001). A MK-1775 mw genotypic resistance test was available in 49 patients (57%); the median interval between the genotypic resistance test and drug measurement was 30 months (IQR 16–42 months). ATV plasma concentrations showed high inter-individual variability both globally and in ritonavir-boosted or unboosted regimens (CV 83.1, 71.4 and 86.5%, respectively) (Fig. 1). Overall, the median ATV plasma concentration was 1.50 mg/L
(IQR 0.70–2.30 mg/L); it was higher in samples obtained from patients taking boosted regimens [1.91 mg/L (IQR 1.20–2.81 mg/L) vs. 1.00 mg/L (IQR 0.22–1.34 mg/L) for unboosted regimens; P<0.001] and not concomitantly receiving acid-reducing agents [1.64 mg/L (IQR 0.87–2.46 mg/L) vs. 0.28 mg/L (IQR 0.16–1.00 mg/L) in those receiving acid-reducing L-gulonolactone oxidase agents; P=0.007]. There were no significant differences in median ATV concentration between patients whose prescribed combination regimens contained tenofovir and those whose regimens did not [1.80 mg/L (IQR 0.90–2.57) for regimens containing tenofovir vs. 1.24 mg/L (IQR 0.38–2.00) for regimens not containing tenofovir; P=0.065]. However, when we considered only the subgroup of patients receiving unboosted ATV, median ATV concentration was lower when tenofovir was coadministered [0.22 mg/L (IQR 0.04–0.80 mg/L) vs. 1.07 mg/L (0.38–1.55 mg/L) when tenofovir was not coadministered; P=0.024]. When patients with ritonavir boosting were considered, no significant difference in ATV concentration was observed between those concomitantly receiving tenofovir and those not receiving tenofovir [median concentration 1.86 mg/L (IQR 1.20–2.81 mg/L) for those receiving tenofovir vs. 2.18 mg/L (IQR 0.59–3.19 mg/L) for those not receiving tenofovir; P=0.748].