Andrew Gewitz (1), Marc Weiner (2), Ron Keizer (3), William MacKenzie (4), Melissa Engle (2), William Whitworth (4), John Johnson (5), Pheona Nsubuga (6), Payam Nahid (1), Nhung Viet Nguyen (7), Charles Peloquin (8), Kelly Dooley (9), Susan Dorman (9), and Rada Savic (1)
(1) University of California, San Francisco, CA (2) UTHSC, San Antonio, TX (3) InsightRX, San Francisco, CA (4) Center for Disease Control, Atlanta, GA (5) Case Western Reserve University, Cleveland, OH (6) Uganda-Case Western Reserve University Research Collaboration, Kampala, Uganda (7) National Tuberculosis Program, Hanoi, Vietnam (8) University of Florida, Gainesville, FL (9) Johns Hopkins University, Baltimore, MD
Objectives: Treatment duration shortening is a major priority for current Tuberculosis (TB) drug development, yet recent clinical trials have failed to achieve this goal when testing shorter regimens in drug-sensitive TB patient populations. To identify putative sets of biomarkers that might facilitate stratification of patients by treatment duration, we developed integrated pharmacokinetic-pharmacodynamic (PK/PD) models of the outcome time-to-stable culture conversion as measured by Mycobacterial Growth Inhibitor Tube (MGIT) assay. Specifically, we wished to (i) characterize the relationship between MGIT assay dynamics and the above outcome in participants enrolled in either of two Phase 2 studies [1-2], and (ii) examine the effects of PK and clinical biomarkers on model parameters suggestive of possible shortening of treatment duration.
Methods: We developed longitudinal nonlinear mixed-effects PK/PD models using data from two randomized trials, the first a two-arm study comparing RFP and standard of-care (rifampin, RIF), and the second involving dose-ranging of RFP vs. RIF (nRIF =320, nRPT = 516, combined). Time-shifted logistic models were best described the dynamics of MGIT assay time-to-stabe conversion in both settings. Time-to-stable conversion was approximated using our model, and the effects of baseline biomarkers were assessed.
Results: Baseline smear grade and productive cough at entry were independently associated with time-to-stable conversion as derived from our model for RIF. For the model for RPT, baseline smear grade, cough grade at study entry, chest x-ray extent, and derived drug exposure (AUC) were independently associated with time-to-stable conversion, whereas dosage effects (both fixed and weight-based) were negligible. Further model simulations suggest, e.g., that for RIF, 95% of patients with high baseline smear grade and productive cough stably converge within 13 weeks versus within 9 weeks for 95% of patients with low smear grade and no cough at study entry.
Conclusions: Our results suggest that various baseline patient characteristics can be used to prospectively stratify patients by treatment duration in clinical TB studies, and highlight the general utility of PK/PD modeling to assess exposure-response relations and guide design of future clinical trials by identifying factors that may aid in predicting therapeutic response at earlier time points.
References:
[1] Dorman SE, Goldberg S, Stout JE, et al. Substitution of rifapentine for rifampin during intensive phase treatment of pulmonary tuberculosis: study 29 of the tuberculosis trials consortium. J Infect Dis 2012; 206: 1030–40.
[2] Dorman SE, Savic RM, Goldberg S, et al. Daily rifapentine for treatment of pulmonary tuberculosis. A randomized, dose-ranging trial. Am J Respir Crit Care Med 2015; 191: 333–43.
Reference: PAGE 25 (2016) Abstr 5851 [www.page-meeting.org/?abstract=5851]
Poster: Drug/Disease modeling - Infection