Abstract

Background: Integration of Phase 1 pharmacokinetic (PK) and non-clinical data provides for the potential to optimize antimicrobial dosing regimens for Phase 2/3 studies. A murine-thigh infection model identified the pharmacokinetic-pharmacodynamic (PK-PD) measure associated with efficacy (T>MIC) and the magnitude of T>MIC predictive of response (27, 35, and 43% for S. pneumoniae, S. aureus, and Gram-negative bacilli, respectively). MCS was applied to Phase 1 PK and nonclinical data to forecast doripenem (DOR) dosing strategies that maximize therapeutic benefit while minimizing drug exposure.

Methods: Using Phase 1 data from 24 subjects that received 1 of 4 regimens, 500 or 1000 mg given q12h or q8h, a population PK model was developed. Blood samples (n) were collected on days 1 (12), 4-6 (6), 7 (13), 8 (1), and 11 (1). Mean PK parameter estimates and a covariance matrix were used for a 5000 patient Monte Carlo Simulation (MCS) to evaluate PK-PD target attainment (based on free drug concentrations) for >100 different regimens (doses of 250, 500, 750, 1000, 2000, 3000 mg; intervals of q6h, q8h, q12h, q24h; infusion durations of 1-6 & 24 h). A range of doubling MIC dilutions from 0.25-16 mg/L was considered.

Results: Using a 2-compartment model with linear elimination and an additive plus proportional residual error model, CL (14.5 L/h), Vc (9.43 L), and t1/2 (0.5 h) were estimated. A subset of the results (below) demonstrate the impact of varying dose, interval, and infusion durations on PK-PD target attainment (%) at MICs of 2, 4, and 8 mg/L.

Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Chicago, Illinois, September 2003

By S. M. Bhavnani, J. P. Hammel, B.B. Cirincione, D. Thye, M. A. Wikler, P. G. Ambrose