Administration mode matters for 5‐fluorouracil therapy: Physiologically based pharmacokinetic evidence for avoidance of myelotoxicity by continuous infusion but not intravenous bolus

• Published in: British journal of clinical pharmacology Vol. 91; no. 4; pp. 1031 - 1040
• Main Authors: Chao, Chih‐Jia, Gardner, Iain, Lin, Chun‐Jung, Yeh, Kun‐Huei, Lu, Wan‐Chen, Abduljalil, Khaled, Ho, Yunn‐Fang
• Format: Journal Article
• Language: English
• Published: England 01.04.2025
• Subjects: 5‐fluorouracil; Antimetabolites, Antineoplastic - administration & dosage; Antimetabolites, Antineoplastic - adverse effects; Antimetabolites, Antineoplastic - pharmacokinetics; Bone Marrow - drug effects; Bone Marrow - metabolism; Dihydrouracil Dehydrogenase (NADP) - genetics; dose optimization; Dose-Response Relationship, Drug; Female; Fluorouracil - administration & dosage; Fluorouracil - adverse effects; Fluorouracil - pharmacokinetics; Humans; Infusions, Intravenous; Male; Middle Aged; Models, Biological; myelotoxicity; physiologically based pharmacokinetics; myelotoxicity; 5‐fluorouracil; dose optimization; physiologically based pharmacokinetics
• ISSN: 0306-5251; 1365-2125; 1365-2125
• DOI: 10.1111/bcp.16061

Aims Pre‐emptive prediction to avoid myelosuppression and harmful sequelae is difficult given the complex interplay among patients, drugs and treatment protocols. This study aimed to model plasma and bone marrow concentrations and the likelihood of myelotoxicity following administration of 5‐fluorouracil (5‐FU) by diverse intravenous (IV) bolus or continuous infusion (cIF) regimens. Methods Using physicochemical, in vitro and clinical data obtained from the literature consisting of various regimens and patient cohorts, a 5‐FU physiologically based pharmacokinetic (PBPK) model was developed. The predicted and observed PK values were compared to assess model performance prior to examining myelotoxicity potential of IV bolus vs. cIF and DPYD wild type vs. genetic variant. Results The established model was verified by utilizing 5‐FU concentration–time profiles of adequate heterogeneity contributed by 36 regimens from 15 studies. The study provided corroborative evidence to explain why cIF (vs. IV bolus) had lower myelotoxicity risk despite much higher total doses. The PBPK model was used to estimate the optimal dosage in patients heterozygous for the DPYD c.1905 + 1G > A allele and suggested that a dose reduction of at least 25% was needed (compared to the dose in wild‐type subjects). Conclusion A verified PBPK model was used to explain the lower myelotoxicity risk of cIF vs. IV bolus administration of 5‐FU and to estimate the dose reduction needed in carriers of a DPYD variant. With appropriate data, expertise and resources, PBPK models have many potential uses in precision medicine application of oncology drugs.