School of Pharmacy, Pharmacy Australia Centre of Excellence, University of Queensland, 20 Cornwall Street, Woolloongabba, Brisbane, QLD, 4102, Australia.
Princess Alexandra Hospital, Brisbane, QLD, Australia.
Clin Pharmacokinet. 2019 Mar;58(3):389-399. doi: 10.1007/s40262-018-0707-9.
The objective of this article is to investigate the influence of blood sampling times on tobramycin exposure estimation and clinical decisions and to determine the best sampling times for two estimation methods used for therapeutic drug monitoring.
Adult patients with cystic fibrosis, treated with once-daily intravenous tobramycin, were intensively sampled over one 24-h dosing interval to determine true exposure (AUC). The AUCs were then estimated using both log-linear regression and Bayesian forecasting methods for 21 different sampling time combinations. These were compared to true exposure using relative prediction errors. The differences in subsequent dose recommendations were calculated.
Twelve patients, with a median (range) age of 25 years (18-36) and weight of 66.5 kg (50.6-76.4) contributed 96 tobramycin concentrations. Five hundred and eighty-eight estimated AUCs were compared to 12 measured true AUC values. Median relative prediction errors ranged from - 34.7 to 45.5% for the log-linear regression method and from - 14.46 to 11.23% for the Bayesian forecasting method across the 21 sampling combinations. The most unbiased exposure estimation was provided from concentrations sampled at 100/640 min after the start of the infusion using log-linear regression and at 70/160 min using Bayesian forecasting. Subsequent dosing recommendations varied greatly depending on the estimation method and the sampling times used.
Sampling times markedly influence bias in AUC estimation, leading to greatly varied dose adjustments. The impact of blood sampling times on dosing decisions is reduced when using Bayesian forecasting.
本文旨在研究采血时间对妥布霉素暴露评估和临床决策的影响，并确定两种治疗药物监测中使用的估算方法的最佳采血时间。
对接受每日一次静脉注射妥布霉素治疗的成年囊性纤维化患者进行 24 小时密集采样，以确定真实暴露量（AUC）。然后使用对数线性回归和贝叶斯预测两种方法，对 21 种不同的采血时间组合进行 AUC 估算，并使用相对预测误差对真实暴露量进行比较。计算后续剂量建议的差异。
12 名患者（中位年龄 25 岁[18-36]，体重 66.5kg[50.6-76.4]）参与了研究，共采集了 96 份妥布霉素浓度。将 588 份估算的 AUC 与 12 份实测的真实 AUC 值进行比较。对数线性回归法的中位相对预测误差范围为-34.7%至 45.5%，贝叶斯预测法的中位相对预测误差范围为-14.46%至 11.23%，在 21 种采血组合中。最准确的暴露估计是通过输注开始后 100/640 分钟时采集的浓度（使用对数线性回归法）和 70/160 分钟时采集的浓度（使用贝叶斯预测法）获得的。后续剂量建议的差异因估算方法和使用的采血时间而异。
采血时间显著影响 AUC 估算的偏差，导致剂量调整差异很大。使用贝叶斯预测法可减少采血时间对剂量决策的影响。
