Kazutaka Oda1, Takashi Kasada, Masahiro Yoshikawa, Mitsunori Tanoue, Takuro Yamashita, Yusuke Takeshita. 1. *Department of Pharmacy, Social Insurance Medical Association, Omuta Tenryo Hospital, Omuta-shi, Fukuoka, Japan; †Department of Pharmacy, Omuta City Hospital, Omuta-shi, Fukuoka, Japan; and ‡Department of Circulation, Social Insurance Medical Association, Omuta Tenryo Hospital, Omuta-shi, Fukuoka, Japan.
Abstract
BACKGROUND: The serum level of teicoplanin (TEIC) is immediately elevated following administration of the recommended dose. In this study, the predictability of the serum trough for TEIC was investigated at day 2 or 3 (C(2-3)), and the authors performed a simulation based on the Bayesian method using C(2-3) in Japanese patients. METHODS: Patients whose the serum trough level was measured within 48 hours (C(2-3)) and at steady state (Css) were eligible for the study. C(2-3) was compared with the predicted level based on the population mean method, and Css was compared with the predicted Css based on both the Bayesian method using C(2-3) and the population mean method. Bias and prediction accuracy were evaluated by the mean prediction error and the mean absolute prediction error (MAE), respectively. RESULTS: The observed and predicted C(2-3) values were 13.2 ± 4.2 μg/mL and 10.4 ± 2.1 μg/mL, respectively. The observed Css was 17.1 ± 3.7 μg/mL, and the predicted Css values based on the Bayesian method and the population mean method were 16.8 ± 2.4 μg/mL and 15.3 ± 2.1 μg/mL, respectively. The mean prediction error and MAE for Css based on the population mean method were -1.87 μg/mL (not significant) and 3.45 μg/mL, respectively, and those based on the Bayesian method were -0.35 μg/mL (not significant) and 2.27 μg/mL, respectively. The change in MAE was 1.18 μg/mL (P < 0.05). CONCLUSIONS: A simulation based on the Bayesian method using C(2-3) of TEIC is acceptable in clinical settings.
BACKGROUND: The serum level of teicoplanin (TEIC) is immediately elevated following administration of the recommended dose. In this study, the predictability of the serum trough for TEIC was investigated at day 2 or 3 (C(2-3)), and the authors performed a simulation based on the Bayesian method using C(2-3) in Japanese patients. METHODS:Patients whose the serum trough level was measured within 48 hours (C(2-3)) and at steady state (Css) were eligible for the study. C(2-3) was compared with the predicted level based on the population mean method, and Css was compared with the predicted Css based on both the Bayesian method using C(2-3) and the population mean method. Bias and prediction accuracy were evaluated by the mean prediction error and the mean absolute prediction error (MAE), respectively. RESULTS: The observed and predicted C(2-3) values were 13.2 ± 4.2 μg/mL and 10.4 ± 2.1 μg/mL, respectively. The observed Css was 17.1 ± 3.7 μg/mL, and the predicted Css values based on the Bayesian method and the population mean method were 16.8 ± 2.4 μg/mL and 15.3 ± 2.1 μg/mL, respectively. The mean prediction error and MAE for Css based on the population mean method were -1.87 μg/mL (not significant) and 3.45 μg/mL, respectively, and those based on the Bayesian method were -0.35 μg/mL (not significant) and 2.27 μg/mL, respectively. The change in MAE was 1.18 μg/mL (P < 0.05). CONCLUSIONS: A simulation based on the Bayesian method using C(2-3) of TEIC is acceptable in clinical settings.