(W0930-10-66) Leveraging Nausea Exposure-Response Relationship to Optimize Titration and Dose Selection Decisions for Cotadutide

Purpose: Cotadutide is a balanced dual glucagon-like peptide-1 (GLP-1)/glucagon receptor agonist. Gastrointestinal (GI) adverse events (AEs) of nausea/vomiting are known to be associated with GLP-1 receptor agonism and may result in significant patient dropout in clinical studies. The GI related AEs can be mitigated by slow up-titration from a low initial dose to the therapeutic range, allowing for tolerance development. In this study, we developed a PK-PD model to characterize the relationship between dose/exposure and the nausea incidence using cotadutide clinical data to facilitate dose selection decisions and titration optimization strategies.

Methods: Data from four clinical studies of cotadutide, comprising 675 subjects and 86,462 daily nausea records, were used in the study. A descriptive analysis was performed on pooled nausea data of the studies, administered subcutaneously once daily from 50 to 600 μg with weekly, biweekly, or monthly uptitration schemes. The dose/exposure-nausea analyses were implemented using a two-part discrete time Markov model approach, and the severity of nausea was modeled as three different categorical states (non-nausea, mild, moderate/severe) within the model. The onset and duration of an event was governed by the transitions between states. The model was used to simulate the nausea rate for weekly, biweekly, and monthly titration schemes with 50, 100, 200, 300, 400, 500 and 600 μg doses as designed in clinical studies.

Results: The descriptive analysis revealed a large variability in the nausea incidence rate, ranging from 17% to 75% at the highest dose levels across the studies. The posterior predictive check (PPC) in Figure 1 indicated that the two-part Markov model described the relationship between exposure and the incidence rate of nausea adequately. This model was subsequently used to simulate the nausea rates for exploring different titration schemes. The simulations found that a weekly titration can contain the highest nausea rate to a ~32%, whereas a monthly titration results in a ~30% nausea rate (Figure 2).

Conclusion: The model-based analyses successfully delineated the relationship between dose escalation regimens and the incidence of nausea for cotadutide, underscoring the importance of fine-tuning the titration schedule to achieve an optimal risk-benefit balance and getting to the therapeutic dose quickly – i.e., weighing potential dropout rates against the total length of titration. There is no significant improvement in nausea incidence as comparing the monthly with weekly titration schemes, however, the weekly titration shortens the titration interval significantly and thus appears to be optimal with manageable nausea rates clinically. Additionally, this model can be utilized to optimize dose and titration schemes for other therapeutics in clinical trials with the aim of minimizing time to reach the therapeutic doses with acceptable tolerability in patients.

Acknowledgements:

Funding: This study was funded by AstraZeneca. Competing interests: U.S., R.D., C.J., H.B., K.A., Y.H. are employees of AstraZeneca and own stocks in AstraZeneca.

Posterior predictive check of different models for total number of transitions between different levels of nausea for 100 simulated datasets. The vertical line indicates observed number of transitions in the original dataset.
Simulated incidence rate of categorized nausea severity by different titration schemes using Two-part model. Simulated probabilities are presented as means (lines) with 90% confidence interval (CI) (shades) from 100 simulated clinical studies.