Predictive value of pharmacokinetics-adjusted phenotypic susceptibility on response to ritonavir-enhanced protease inhibitors (PIs) in human immunodeficiency virus-infected subjects failing prior PI therapy

Abstract

The activities of protease inhibitors in vivo may depend on plasma concentrations and viral susceptibility. This nonrandomized, open-label study evaluated the relationship of the inhibitory quotient (IQ [the ratio of drug exposure to viral phenotypic susceptibility]) to the human immunodeficiency virus type 1 (HIV-1) viral load (VL) change for ritonavir-enhanced protease inhibitors (PIs). Subjects on PI-based regimens replaced their PIs with ritonavir-enhanced indinavir (IDV/r) 800/200 mg, fosamprenavir (FPV/r) 700/100 mg, or lopinavir (LPV/r) 400/200 mg twice daily. Pharmacokinetics were assessed at day 14; follow-up lasted 24 weeks. Associations between IQ and VL changes were examined. Fifty-three subjects enrolled, 12 on IDV/r, 33 on FPV/r, and 8 on LPV/r. Median changes (n-fold) (FC) of 50% inhibitory concentrations (IC(50)s) to the study PI were high. Median 2-week VL changes were -0.7, -0.1, and -1.0 log(10) for IDV/r, FPV/r, and LPV/r. With FPV/r, correlations between the IQ and the 2-week change in VL were significant (Spearman's r range, -0.39 to -0.50; P < or = 0.029). The strongest correlation with response to FPV/r was the IC(50) FC (r = 0.57; P = 0.001), which improved when only adherent subjects were included (r = 0.68; P = 0.001). In multivariable analyses of the FPV/r arm that included FC, one measure of the drug concentration, corresponding IQ, baseline VL, and CD4, the FC to FPV was the only significant predictor of VL decline (P < 0.001). In exploratory analyses of all arms, the area under the concentration-time curve IQ was correlated with the week 2 VL change (r = -0.72; P < 0.001). In conclusion, in PI-experienced subjects with highly resistant HIV-1, short-term VL responses to RTV-enhanced FPV/r correlated best with baseline susceptibility. The IQ improved correlation in analyses of all arms where a greater range of virologic responses was observed.

FIG. 1.

(a) Plot of log RNA change versus IC₅₀ FC (on a log_e scale) and estimated log RNA change as a function of IC₅₀ FC (linear regression model). (b) Plot of log RNA change versus IC₅₀ FC (on a log_e scale) and estimated log RNA change as a function of IC₅₀ FC for subjects with at least 80% adherence (AH) (linear regression model).

FIG. 2.

(a) Plot of log RNA change versus the reciprocal of AUC_IQ_FC (on a log_e scale) and estimated log RNA change as a function of the reciprocal of AUC_IQ_FC (linear regression model). AUC_IQ_FC = AUC/[mean wild-type IC₅₀ of reference strains × IC₅₀ FC]. (b) Plot of log RNA change versus the reciprocal of C_{12_}IQ_FC (on a log_e scale) and estimated log RNA change as a function of the reciprocal of C₁₂ _IQ_FC (linear regression model).