Cleo Demeester1, Pieter Annaert1,2, Patrick Augustijns1, Jan Schlender3, André Dallmann4,6, Thomas Wendl5,6
1Drug Delivery and Disposition, KU Leuven, 2BioNotus, 3Novartis Institute of Biomedical Research, 4Bayer HealthCare SAS on behalf of: Research & Development, Pharmaceuticals, Model-Informed Drug Development, Bayer AG, Leverkusen, Germany, 5Bayer AG, Research & Development, Pharmaceuticals, Model-Informed Drug Development, 6These authors share last authorship
Introduction: While older adults receive the vast majority of drug prescriptions, current formulations are based on clinical studies in younger adults. Due to multi-morbidity, frailty, or polypharmacy, older adults are usually not enrolled in clinical trials and therefore under-represented in these studies [1], [2]. A useful alternative for investigating pharmacokinetics (PK) in older people is physiologically based pharmacokinetic (PBPK) modelling and simulation. Previously, a huge effort has been made to inform the anthropometric and physiological changes of older adults and are incorporated into the open-source PBPK software PK-Sim® [3]. However, PBPK models presently do not account for age-related changes in plasma proteins [4], haematocrit, cytochrome P450 (CYP) 3A4 activity [5], and absorption-related parameters for the older population [6], [7]. Objectives: This study investigates the age-related changes in plasma protein concentrations, haematocrit values, and CYP3A4 activity to define senescence functions and test whether the implementation of these changes in a novel PBPK framework enhances the predictive performance of PBPK models for older adults. Methods: The plasma protein scaling factors for males and females over age were established by incorporating the plasma protein concentrations over age. Human serum albumin concentrations and haematocrit values were obtained from the National Health and Nutrition Examination Survey (NHANES) [8], while for alpha-1-acid glycoprotein (AAG) concentrations were sourced from a systematic screening of the literature. Linear regression analysis was applied to determine the relationship between age and protein concentrations. The scaling factors for plasma proteins were validated by predicting known fraction unbound (fu) values using the single protein binding model [9]. Additionally, a linear ontogeny function for CYP3A4 was developed by calculating the intrinsic unbound clearance from published plasma clearance data for the probe compounds midazolam and alfentanil, using the well-stirred model [10]. This model accounted for age-dependent changes in liver blood flow, renal clearance, the blood/plasma concentration ratio, and fu. The senescence functions were then integrated into existing PBPK models for midazolam, alfentanil, and nimodipine, which had been previously validated in younger populations [11], to simulate pharmacokinetics in older adults using PK-Sim®. Results: Data from 20 different publications were analysed to study the AAG concentrations and showed a slight but significant increase in AAG concentrations with age for males (y=0.583+0.00318x, p=0.007, r2=0.08). In contrast, albumin concentrations significantly decreased with age in both sexes, with a more pronounced decline in males (y=44.7-0.0657x, p<0.001, r2=0.14) compared to females (y=40.1-0.00859x, p=0.014, r2<0.01). The plasma protein concentrations were able to predict the fu within a 2-fold error in the older population; however, the fold change in fu was not adequately predicted. Furthermore, a decrease in haematocrit levels over age was observed in males, whereas a slight increase in haematocrit was seen in females. Moreover, the intrinsic unbound clearance of midazolam and alfentanil exhibited a significant linear decrease over age (p=3.3E-4, r2=0.11). Also, a larger variability was seen for the older adults as the coefficient of variation was 81.7% for the older adults (65+) and 47.3% for the younger adults (20 – 40 years old). The senescence functions for the plasma proteins, haematocrit levels, and the CYP3A4 ontogeny were implemented in the PBPK models. Overall, the original PBPK models without any adjustments aligned more closely with the observed data with a geometric mean fold error (GMFE) of 1.33 compared to a GMFE of 1.44 for the adapted models. An exception was observed for the intravenous administration of nimodipine, where the model adaptation demonstrated a better prediction. Conclusion: Interestingly, despite midazolam and alfentanil being part of the training set for the CYP3A4 ontogeny, the original PBPK models without the age-related changes described above outperformed the novel models, raising questions about the underlying cause of this discrepancy. Further investigation is required to understand the root cause of the observed differences. Acknowledgements: This research is supported by a Horizon 2020 research and innovation program; Marie Sklodowska-Curie grant (No. 956146).
[1] Q. Liu et al., “Roadmap to 2030 for Drug Evaluation in Older Adults.,” Clin. Pharmacol. Ther., vol. 112, no. 2, pp. 210–223, Aug. 2022, doi: 10.1002/cpt.2452. [2] R. Ruiter, J. Burggraaf, and R. Rissmann, “Under-representation of elderly in clinical trials: An analysis of the initial approval documents in the Food and Drug Administration database,” Br. J. Clin. Pharmacol., vol. 85, no. 4, pp. 838–844, 2019, doi: 10.1111/bcp.13876. [3] J. F. Schlender et al., “Development of a Whole-Body Physiologically Based Pharmacokinetic Approach to Assess the Pharmacokinetics of Drugs in Elderly Individuals,” Clin. Pharmacokinet., vol. 55, no. 12, pp. 1573–1589, 2016, doi: 10.1007/s40262-016-0422-3. [4] J. Al-Qassabi, S. P. F. Tan, P. Phonboon, A. Galetin, A. Rostami-Hodjegan, and D. Scotcher, “Facing the Facts of Altered Plasma Protein Binding: Do Current Models Correctly Predict Changes in Fraction Unbound in Special Populations?,” J. Pharm. Sci., vol. 113, no. 6, pp. 1664–1673, 2024, doi: https://doi.org/10.1016/j.xphs.2024.02.024. [5] J. George, K. Byth, and G. C. Farrell, “Age but not gender selectively affects expression of individual cytochrome P450 proteins in human liver.,” Biochem. Pharmacol., vol. 50, no. 5, pp. 727–730, Aug. 1995, doi: 10.1016/0006-2952(95)00192-3. [6] C. Demeester et al., “Physiologically based pharmacokinetic (PBPK) modelling of oral drug absorption in older adults – an AGePOP review,” Eur. J. Pharm. Sci., vol. 188, p. 106496, 2023, doi: https://doi.org/10.1016/j.ejps.2023.106496. [7] C. Demeester et al., “A quantification of gastric and duodenal fluid volumes in older adults using MRI.,” Int. J. Pharm., vol. 666, p. 124831, Dec. 2024, doi: 10.1016/j.ijpharm.2024.124831. [8] Centers for Disease Control and Prevention (CDC). National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey Questionnaire (or Examination Protocol, or Laboratory Protocol). Hyattsville, MD: U.S. Department of Health an. [Online]. Available: https://wwwn.cdc.gov/nchs/nhanes/search/datapage.aspx?Component=Laboratory. [9] P. J. McNamara and D. Meiman, “Predicting Drug Binding to Human Serum Albumin and Alpha One Acid Glycoprotein in Diseased and Age Patient Populations.,” J. Pharm. Sci., vol. 108, no. 8, pp. 2737–2747, Aug. 2019, doi: 10.1016/j.xphs.2019.03.018. [10] F. Salem, T. N. Johnson, K. Abduljalil, G. T. Tucker, and A. Rostami-Hodjegan, “A re-evaluation and validation of ontogeny functions for cytochrome P450 1A2 and 3A4 based on in vivo data.,” Clin. Pharmacokinet., vol. 53, no. 7, pp. 625–636, Jul. 2014, doi: 10.1007/s40262-014-0140-7. [11] “OSP-PBPK-Model-Library,” Open Systems Pharmacology. [Online]. Available: https://github.com/Open-Systems-Pharmacology/OSP-PBPK-Model-Library.
Reference: PAGE 33 (2025) Abstr 11402 [www.page-meeting.org/?abstract=11402]
Poster: Drug/Disease Modelling - Absorption & PBPK