Alan Faraj 1, Helene Hausner 2, Katarina Cepo 3
1 Pharmacometrics, Novo Nordisk A/S (Uppsala, Sweden), 2 Clinical Pharmacology, Novo Nordisk A/S (Copenhagen, Denmark), 3 Medical Science, Novo Nordisk A/S (Copenhagen, Denmark)
Objectives: Sickle cell disease (SCD) is driven by polymerization of mutated hemoglobin in red blood cells [1]. Fetal hemoglobin (HbF) reduces the effects of abnormal polymerization but is silenced during infancy and onwards by the epigenetic regulator DNA methyltransferase 1 (DNMT1 [2]). NDec is a fixed oral combination modified release product of decitabine (Dec) and tetrahydrouridine (THU). Dec incorporates into DNA and binds irreversibly to DNMT1, hereby inactivating DNMT1 and inducing HbF production [2] while THU inhibits the rapid degradation of Dec and decreases exposure to unfavorable CDA-mediated uridine metabolites [3]. This work aimed to characterize the population pharmacokinetics (PK) of Dec and THU and study the expected exposure levels of Dec in adult and adolescent patients following a proposed weight band-tiered dosing strategy [4] by using clinical trial simulations.
Methods: PK data from two phase-1 studies (NCT06291285 and NCT04086238) in which NDec was administered orally (15 mg Dec and 750 mg THU) with breakfast or a high-fat meal, respectively, were used for model building. Samples below the limit of quantification were modelled using the M3 method [5] for Dec (>30%) and ignored for THU (4.7%). The starting models were two compartment disposition models with linear elimination. Lag-time, zero-order followed by first order and transit models were tested to model the absorption for both drugs. As no intravenous data was available, apparent parameters were estimated. Inter-individual variability (IIV) was tested for relevant parameters. Age, bodyweight, race, ethnicity, sex and meal type were tested as covariates for both drugs using stepwise-covariate model building [6] including forward inclusion (p=0.01) and backward exclusion (p=0.001). Bodyweight was included as a covariate a priori through allometric scaling using fixed exponents of 0.75 and 1 for clearance and volume of distribution terms, respectively. Models were built using NONMEM v7.5.1 and evaluated using visual predictive checks, parameters plausibility and uncertainty. Based on the developed Dec model, clinical trial simulations were performed using a weight band-tiered dosing strategy that previously has been proposed [4] and compared to a flat-dosing approach. Plausible bodyweight distributions were sampled using bodyweight data from adult and adolescent sickle cell patients (40-125 kg). Based on the simulated of concentrations over time, maximum concentration (Cmax) and the area under the curve over 24 hours (AUC0-24h) were derived and compared between adults and adolescents.
Results: The final Dec population PK model was a two-compartment model with transit absorption, linear elimination and IIV on mean transit time, bioavailability and volume of distribution (CV% = 28, 40, 55, respectively) and a combined additive and proportional residual error model. Statistically significant (p=0.01) lower mean transit time, bioavailability and absorption rate was found when dosing with a heavy meal compared to dosing with a breakfast meal. For THU, the PK was also described using a transit absorption model with linear elimination. IIV was supported on bioavailability and mean transit time (CV% = 63, 78, respectively) with a full covariance block and the residual error model was a combined additive and proportional model. No statistically significant covariates were found for the PK of THU. Both developed models described the PK data well. The simulated Cmax of Dec after a breakfast meal (median and 90% prediction interval of median [ng/ml]) was found to be 38 (36-41) and 37 (29-46) for adults and adolescents, respectively, following a weight-band-tiered dosing approach and 44 (42-47) and 53 (43-68), respectively, for a flat-dosing approach (15 mg). The corresponding numbers for AUC0-24h (h*ng/ml) were 88 (84-92) and 83 (71-100) for a weight-banded dosing approach and 103 (99-108) and 122 (103-143) for a flat-dosing approach. The same trend of higher discrepancy between adults and adolescents when applying flat-dosing approach versus the weight banded approach was seen when simulating PK after a heavy meal.
Conclusions: The clinical trial simulations of previously assessed THU and Dec PK data suggested that exposures to both components in adolescents are expected to be similar to those reported in adults when following the suggested weight-banded dosing approach. Further development of population PK and PD models using data collected from treated adult patients will help predict clinical response in adolescents and tailor dosing for optimal disease control.
Acknowledgements: The authors would like to thank Walter Mibei for significant contributions informing the conduct of this work.
References:
[1] Kato GJ et al. Nat Rev Dis Primers 2018;4;18010; 2
[2] Molokie R et al. PLoS Med 2017;14;e1002382; 3
[3] Hausner H et al. ASH 2025; 4737
[4] Andemariam B et al. Blood 2022; 140 (Supplement1):5420-5421
[5] Bergstrand M et al. AAPS 2009 May 19;11(2):371-380
[6] Khandelwal A et al. AAPS 2011 Sep;13(3):464-72
Reference: PAGE 34 (2026) Abstr 12217 [www.page-meeting.org/?abstract=12217]
Poster: Drug/Disease Modelling - Other Topics