Scholarlink

Background <p>We aimed to design an optimal accompanying diagnosis, defined as a model-informed limited sampling strategy (LSS) that uses validated population pharmacokinetic (popPK) models to predict moxifloxacin exposures with minimal blood samples, facilitating therapeutic drug (TDM) while reducing patient burden, through external validation of published popPK models for moxifloxacin in patients with tuberculosis, for developing a limited sampling strategy to optimize drug monitoring.</p> Methods <p>A systematic literature review identified seven popPK models for moxifloxacin. These models were externally validated using a dataset of 1,042 moxifloxacin concentrations from 113 multidrug-resistant tuberculosis (MDR-TB) patients. The model performance was evaluated using various predictive performance metrics, including the median prediction error, mean absolute error, and root mean square error. A Bayesian forecasting analysis was conducted to determine the optimal sampling points for accompanying diagnosis.</p> Results <p>Most models showed acceptable predictive performance within ± 30% prediction error (F30). The model proposed by Pranger et al. demonstrated the best overall performance for external validation. Analysis of the accumulated data revealed that sampling at even-numbered weeks (2, 4, 6, 8, and 10) provided better predictive power than sampling at odd-numbered weeks. Data from the first week was insufficient for an accurate prediction. The model performance stabilized between two and eight weeks of accumulated data.</p> Conclusion <p>This study suggests that popPK modeling with a limited sampling strategy can effectively predict moxifloxacin concentrations in patients with tuberculosis. Optimal accompanying diagnosis can be achieved with data collected within 1 month, preferably at even-numbered weeks, thereby reducing the burden of frequent blood sampling. These findings have implications for improving therapeutic drug monitoring in the treatment of multidrug-resistant tuberculosis, potentially improving patient compliance and treatment outcomes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimal accompanying diagnosis design through external validation of population pharmacokinetic models for moxifloxacin in tuberculosis patients

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月24日

查看全文
  • Hyunjung Lee,
  • Woojin Jung,
  • Min Jung Chang,
  • Soyoung Lee,
  • Jung-woo Chae,
  • Radojka M. Savic,
  • Hwi-yeol Yun

摘要

Background

We aimed to design an optimal accompanying diagnosis, defined as a model-informed limited sampling strategy (LSS) that uses validated population pharmacokinetic (popPK) models to predict moxifloxacin exposures with minimal blood samples, facilitating therapeutic drug (TDM) while reducing patient burden, through external validation of published popPK models for moxifloxacin in patients with tuberculosis, for developing a limited sampling strategy to optimize drug monitoring.

Methods

A systematic literature review identified seven popPK models for moxifloxacin. These models were externally validated using a dataset of 1,042 moxifloxacin concentrations from 113 multidrug-resistant tuberculosis (MDR-TB) patients. The model performance was evaluated using various predictive performance metrics, including the median prediction error, mean absolute error, and root mean square error. A Bayesian forecasting analysis was conducted to determine the optimal sampling points for accompanying diagnosis.

Results

Most models showed acceptable predictive performance within ± 30% prediction error (F30). The model proposed by Pranger et al. demonstrated the best overall performance for external validation. Analysis of the accumulated data revealed that sampling at even-numbered weeks (2, 4, 6, 8, and 10) provided better predictive power than sampling at odd-numbered weeks. Data from the first week was insufficient for an accurate prediction. The model performance stabilized between two and eight weeks of accumulated data.

Conclusion

This study suggests that popPK modeling with a limited sampling strategy can effectively predict moxifloxacin concentrations in patients with tuberculosis. Optimal accompanying diagnosis can be achieved with data collected within 1 month, preferably at even-numbered weeks, thereby reducing the burden of frequent blood sampling. These findings have implications for improving therapeutic drug monitoring in the treatment of multidrug-resistant tuberculosis, potentially improving patient compliance and treatment outcomes.