Srishti Nagaraj 1,2, Daniël M. Jonker 1, Trine M. Lund 2, Mats O. Karlsson 3
1 Ferring Pharmaceuticals A/S (Kastrup, Denmark), 2 Department of Drug Design and Pharmacology, University of Copenhagen (Copenhagen, Denmark), 3 Department of Pharmacy, Uppsala University (Uppsala, Sweden)
Introduction: In vitro fertilisation (IVF) involves stimulating the development of multiple ovarian follicles using follicle-stimulating hormone (FSH) to achieve pregnancy. Once enough follicles develop to a sufficient size, ovulation is triggered, oocytes are collected and fertilised. This constitutes one cycle of controlled ovarian stimulation (COS).
Follitropin delta (REKOVELLE) is a recombinant FSH (rFSH) developed by Ferring Pharmaceuticals for use in IVF. A dose-response model describing the number of oocytes retrieved in the first treatment cycle for follitropin delta has been published [1]. This model includes baseline anti-Müllerian hormone (AMH) levels as a covariate on Emax and was used to design an individualised dosing algorithm for follitropin delta administration in the first COS cycle. It demonstrated that more oocytes were retrieved in patients with higher AMH levels, who therefore required a lower dose of rFSH.
Dose selection for a cycle balances the need to retrieve enough oocytes against the risk of ovarian hyperstimulation syndrome, and since many women require multiple cycles of IVF, it is of interest how previous responses can guide the selection of future doses. Furthermore, to optimise pregnancy rates over multiple cycles, it is relevant to understand factors affecting patient enrolment in subsequent cycles.
Objectives:
i) Update the existing dose-response model for follitropin delta, using data from clinical trials where subjects completed up to three cycles of IVF.
ii) Identify predictors of the probability of patient enrolment in subsequent cycles to allow adequate simulation-based predictions.
Methods: Data from women undergoing IVF treatment with follitropin delta from three trials were utilised – a phase 2 dose-finding trial [2], including 220 subjects randomised to five different dose levels of follitropin delta, and two phase 3 trials [3,4]. These phase 3 trials included 665 women undergoing multiple treatment cycles with follitropin delta, of whom 252 and 95 completed two and three treatment cycles, respectively. In the second and third cycles, the dose of rFSH was increased, decreased, or maintained based on the number of oocytes retrieved in the previous cycle. The inclusion of multiple COS cycles and a wider range of AMH levels enabled the current update of the existing dose-response model with focus on interindividual variability aspects.
Since not all subjects continued to the next COS cycle, a random drop-out mechanism, i.e., dependent on observed data, was investigated [5]. A logistic regression model estimating the probability of patient enrolment to the next cycle was developed, and its effect on simulation-based predictions was explored. All modelling and simulations were performed using R, PsN and NONMEM®.
Results: An interindividual random effect for Emax was added to the dose-response model. The function NAMH/(β+NAMH), where NAMH is normalised AMH and β is the covariate coefficient, described the influence of AMH on Emax and improved on the previous covariate model. The final estimates for the dose-response model were Emax = 25.3 oocytes (RSE 17%), ED50 = 6 μg (RSE 13%), Hill factor = 1.97 (RSE 51%), β = 0.828 pmol/L (RSE 12%), and overdispersion of negative binomial model = 0.0615 (RSE 35%). The model-estimated variability between individuals (%CV) on Emax was approximately 32%.
In the patient enrolment model, baseline AMH levels and the number of oocytes retrieved in the previous cycle were important factors affecting the probability of patient enrolment into the following cycle. The model estimated that, for typical women with an AMH level of 16.5 pmol/L and 9 oocytes retrieved previously, 39% were enrolled in the next cycle. The probability of a patient enrolling in a subsequent IVF treatment cycle decreases with a higher number of oocytes retrieved and higher AMH levels.
Conclusions: The study implemented interindividual variability and an improved AMH covariate function for the prediction of oocyte retrieval. A logistic regression model predicting patient enrolment probability improved simulations. The dose-response model is well suited to simulate and evaluate recommendations for dose adjustments in repeated cycles.
References:
[1] Seifer DB, Tal R (eds). Anti-Müllerian Hormone: Biology, Role in Ovarian Function and Clinical Significance. Nova Science Publishers, 2016.
[2] Arce JC et al. Fertil Steril. 2014;102(6):1633–1640.e5.
[3] Nyboe Andersen A et al. Fertil Steril. 2017;107(2):387–396.e4.
[4] Bosch E et al. Reprod Biomed Online. 2019;38(2):195–205.
[5] Friberg L et al. Clin Pharmacol Ther. 2009;86(1):84–91.
References:
References:
[1] Seifer DB, Tal R (eds). Anti-Müllerian Hormone: Biology, Role in Ovarian Function and Clinical Significance. Nova Science Publishers, 2016.
[2] Arce JC et al. Fertil Steril. 2014;102(6):1633–1640.e5.
[3] Nyboe Andersen A et al. Fertil Steril. 2017;107(2):387–396.e4.
[4] Bosch E et al. Reprod Biomed Online. 2019;38(2):195–205.
[5] Friberg L et al. Clin Pharmacol Ther. 2009;86(1):84–91.
Reference: PAGE 34 (2026) Abstr 11947 [www.page-meeting.org/?abstract=11947]
Poster: Drug/Disease Modelling - Endocrine