Materials and Methods
Study design and sample collection. This phase 2 study was an open-label, parallel-design, dose-ranging study of alb-IFN conducted in 56 patients with genotype 1 CHC who had not previously received IFNa therapy.10 Patients received 2 subcutaneous injections of alb-IFN administered 14 days apart in 5 dose cohorts (200 µg, 450 µg, 670 µg, 900 µg, and 1200 µg). Patients did not receive ribavirin in this study. Most patients were later enrolled into the phase 3 study of alb-IFN with ribavirin. Blood was collected on Days 0, 1, 4, 7, 14, 15, 18, 21, 28, 35, and 42. Levels of HCV RNA were assessed by real-time polymerase chain reaction (CE-marked COBAS® Ampliprep/COBAS® Taqman® HCV test, Roche Diagnostics, Basel, Switzerland) with a dynamic range of 10 IU/mL to 100 million IU/mL. Serum alb-IFN concentrations were measured at the same times using a validated enzyme-linked immunosorbent assay (limit of quantification 530 pg/mL). The analysis included data from 26 patients enrolled in the 2 higher dose groups (900 µg and 1200 µg), which are both being studied in phase 3 trials in combination with ribavirin.
Mathematical models. Pharmacokinetic model definition. Concentrations of alb-IFN were modeled using a standard 1-compartment pharmacokinetic model, where it is assumed that serum and liver drug concentrations are in equilibrium:
Bioavailable injected dose (FD) and volume of distribution (Vd) are coupled into the parameter , ka is the absorption rate from the injection site, and ke is the clearance rate from circulation. The alb-IFN t½ was calculated from the clearance rate:
Model of viral dynamics. Viral kinetics was analyzed by a model of viral dynamics described by the following differential equations5:
T(t) is the number of target cells produced at rate s and dying with death rate constant d. I(t) is the number of productively infected cells corresponding to constant de-novo infection rate ß and loss rate constant d (productively infected cell t½ is ln/d), coupled with the viral infection blocking factor ? (here assumed to be 0 based on previous results regarding the effect of IFNa on HCV). V(t) is the viral load resulting from viral production at constant rate p per cell per day and clearance rate c (free virion t½ is ln/c). The drug effect is modeled through the effectiveness in blocking production, which in this case is assumed to vary in time—e(t)—rather than being constant as in a previous study by Neumann et al.5
Model integrating pharmacokinetics and viral kinetics. The relationship between serum alb-IFN concentration and viral kinetics was assessed by modeling the varying effectiveness in blocking viral production—e(t)—as a function of alb-IFN concentration—e(IFN). This differs from previous studies of standard daily IFN injections, in which IFN levels and thus e were assumed to be approximately constant. As alb-IFN levels are not constant during treatment,6,13,14 a Hill function was assumed for the relationship between antiviral efficacy (e) and IFN concentrations—C(t)—during the treatment period: