Long-acting formulation drug-drug interactions: evaluating the importance of timing using a subcutaneous levonorgestrel PBPK example
Brian Cicali (1), Rodrigo Cristofoletti (1), Michelle Pressly (1), Soyoung Kim (1), Thomas Wendl (2), Valvanera Vozmediano (1), Stephan Schmidt (1)
(1) Center for Pharmacometrics and Systems Pharmacology, University of Florida, USA, (2) Bayer AG, Leverkusen, Germany
Introduction: Long-acting drug formulations, such as injectables or subcutaneous implants, can provide significant benefits for drugs being taken for long periods, compared to their traditionally administered counterparts, e.g. oral doses. However, given the length of the dose (months to years), questions surrounding clinical study development, e.g. when efficacy and safety-based drug-drug interaction (DDI) evaluations should be conducted, are difficult to address. Levonorgestrel (LNG) is a progestin used in many contraceptive formulations including subcutaneous implants, which are inserted and remain under the skin for up to five years. Due to this length of application and intended use of the product, LNG implants are considered desirable for medication adherence benefits compared to oral administrations, however there is concern that these implants are more susceptible to contraceptive failure via CYP3A4 drug-drug interactions (DDI) due to lower systemic concentrations relative to oral LNG formulations. Whether there is an increased risk of contraceptive failure for LNG implants in DDI scenarios, and at what time points post-insertion are critical for such risk evaluation, is unknown. Therefore, we aimed to implement a subcutaneous release model for LNG implants into a physiologically-based pharmacokinetic (PBPK) modeling framework for DDI investigation at different timepoints post-insertion.
Objectives:
1) Utilize a PBPK modeling framework for subcutaneous LNG administration to predict systemic exposure levels over a 5-year period.
2) Evaluate the impact of CYP3A4 DDIs on LNG systemic exposure at various timepoints throughout the 5 years of drug implantation.
Methods: Utilizing a previously developed LNG PBPK model1, administration of 150 mg of LNG into the subcutaneous interstitial space was implemented using a modified biphasic release function and incorporation of formulation-specific components (decay rate and fraction available). Formulation-specific parameters were optimized using Jadelle® clinical trial data (n=321) and then verified using 2 additional clinical trials (n=216). Once verified, the model was simulated using a healthy women population (n=100) to represent DDI scenarios of CYP3A4 induction or inhibition via rifampin (600mg QD) or itraconazole (200mg QD) at steady state, respectively. These DDI simulations were conducted at time=0 (day of implantation) and replicated in 3-month intervals until time=5 years (implant removal). Predictions for LNG exposure were compiled and compared to simulation predictions of LNG exposure without DDI.
Results: Model simulations of subcutaneous LNG administration indicated a mean plasma concentration range of 0.35 and 0.19 ng/mL, with the lowest concentrations observed at late time points (year 5). DDI simulations with rifampin administered at the same time as the implant application resulted in a delayed Tmax (640 vs 280 hours) but similar mean concentration as the non-DDI scenario. However, as the DDI scenarios are delayed from the initial LNG application, decreases in mean LNG concentration becomes apparent; mean = 0.19 ng/mL when DDI at 3 months (vs 0.35 ng/mL) and mean = 0.14 ng/mL when DDI at 3 years (vs 0.25 ng/mL). Inhibition DDIs followed a similar trend for increased concentrations, however these levels remained well below the safety threshold at all timepoints. LNG concentrations returned to normal after ~30 days from the end of the DDI.
Conclusion: Using subcutaneous LNG implants as an example, our findings show that the time of DDI evaluation is important when assessing the safety and efficacy of long-acting drug formulation products. In the case of LNG, the consequence of inducer DDIs seems to be significant, especially as time post-implant increases, assuming implant and oral LNG follow the same exposure-response relationship. Future applications of this work will focus on translating these PK alterations to pharmacodynamic effects and real-world outcomes to help the support a developmental framework for optimizing long-acting formulation clinical trials.
References:
[1] Cicali, B, et al. CPT:PSP (2021) 10, 48-58