Abstract:

Objective: To establish a human physiologically based pharmacokinetic (PBPK) model of pretomanid, and conduct external validation of its predictive performance under different dosage regimens based on clinical data, so as to provide evidence for individualized precision pharmacotherapy for drug-resistant pulmonary tuberculosis. Methods: The PBPK model of pretomanid was developed based on plasma concentration data from 6 patients with drug-resistant pulmonary tuberculosis. WinNonlin and NLME software were applied for model construction, fitting and simulation, and pharmacokinetic parameters were calculated via non-compartmental analysis. For external validation, we integrated published data from healthy subjects and phase Ⅰ clinical studies (a total of 221 observed plasma concentration points), covering single oral doses (50-1000 mg) and a 200 mg multiple-dose regimen. Sensitivity analysis was performed on key parameters including plasma unbound fraction, effective intestinal permeability coefficient and hepatic intrinsic clearance. Results: The established pretomanid PBPK model could satisfactorily predict clinical plasma concentration-time profiles, with good consistency between predicted and observed values of major PK parameters. Validated with 221 external clinical plasma concentration data points, 92.20% of the observed values fell within the 90% confidence interval of model predictions, and the mean fold errors of main pharmacokinetic parameters all ranged from 0.5 to 2.0, indicating favorable predictive performance of the model. Multiple-dose simulation showed that effective steady-state trough concentration could be achieved after approximately 3 days of continuous administration; steady-state peak concentration was reached on day 7, with predicted exposure levels within the safe range. Sensitivity analysis revealed that plasma unbound fraction, effective intestinal permeability coefficient, hepatic portal blood flow and hepatic intrinsic clearance all affected drug exposure; among them, effective intestinal permeability coefficient exerted a significant effect on time to peak concentration, while hepatic intrinsic clearance had a substantial impact on drug exposure levels. Conclusion: The pretomanid PBPK model established in this study has satisfactory external predictive performance and can accurately simulate in vivo exposure characteristics under different dosage regimens. For pretomanid administered at 200 mg once daily, steady-state effective exposure can be achieved after approximately 3 days of treatment. It is recommended to monitor trough concentration (approximately 770 ng/ml) on the morning of day 4, and peak concentration (safety threshold of approximately 5000 ng/ml) at 4-6 hours after dosing on day 7, to guide efficacy and safety evaluation. The model can support dose optimization, selection of monitoring time points and individualized medication guidance for special populations in pretomanid-based treatment of drug-resistant pulmonary tuberculosis, and holds important reference value for the rational clinical application of the B-Pa-L regimen.

Key words: Tuberculosis, pulmonary, Drug resistance, Antitubercular agents, Pharmacokinetics, Pretomanid