(T-087) Exposure-Response Analysis of Bleeding Events Following Weekly SC Flat Dose Marstacimab Administration in Hemophilia Participants Using Regression Modeling and Repeated-Time-To-Event (RTTE) Analyses
Objectives: Hemophilia is an X-linked recessive genetic disorder which is characterized by a reduced ability to form clots in response to bleeds. Marstacimab, an inhibitor of the Tissue Factor Pathway Inhibitor (TFPI, a protein in the extrinsic coagulation pathway) was recently approved for routine prophylaxis to prevent or reduce the frequency of bleeding episodes in adult and pediatric participants 12 years of age and older with severe Hemophilia A or B. The recommended dosing regimen is a once-weekly (QW), subcutaneous (SC), flat (weight-independent) dose of 150 mg with a loading dose of 300 mg SC. The analyses presented here were performed to characterize the exposure-response relationship for marstacimab in hemophilia participants.
Methods: A Poisson regression analysis was used to describe the relationship between marstacimab exposure metrics (steady state Cavg, Cmax and Cmin) and number of bleeds for hemophilia subjects in Phase 2 and Phase 3 studies. Different transformations of all the exposure metrics (log, square root, nominal value) as well as various forms of the Poisson regression model (Poisson, Quasi-Poisson, GEE Poisson) were explored to identify the model that best describes the relationship between marstacimab exposure and the clinical endpoint of interest i.e. Annualized Bleeding Rate (ABR). In addition, a repeated time to event (RTTE) model was used to link pharmacokinetics to spontaneous bleeding events in Hemophilia participants. An RTTE model is a survival model which can consider recurrent events from the same individual (temporal aspect) and help explain bleed data better.
Results: For the Poisson regression analysis, the final model was a Quasi-Poisson regression model with log PK as the independent variable and ABR as the dependent variable. A dispersion parameter far above 1 implied that variance in the data was much higher than the average and justified the application of Quasi-Poisson model instead of a Poisson model. As expected, a negative relationship is seen between marstacimab exposure and Annualized Bleeding Rate (ABR) across all three exposure metrics examined. Over the range of therapeutic concentrations seen across dose groups and age groups (13000 - 68000 ng/mL), the model-predicted ABRs matched well with the observed values. For the RTTE analysis, a constant baseline hazard model (base model) and a Gompertz model of hazard function (final model) were studied. The final model also incorporated the effect of drug concentration (obtained from the population PK model) via the hazard parameter, which resulted in a significant decrease in the overall objective function.
Conclusions: Bleed event data from Phase 2 and Phase 3 studies in hemophilia participants were analyzed using 2 different methodologies i.e. Poisson regression and RTTE analysis, to understand the exposure-response relationship for marstacimab. The RTTE model provides temporal characterization of drug effect, more easily allowing for time-variable hazard function and handling of right-censoring. An inverse relationship between bleed events and drug concentration, with predicted ABRs within observed ranges, was demonstrated with both methodologies.
Citations: N/A
Keywords: Repeated Time to Event (RTTE), Poisson Regression, PKPD Modeling
