Evaluation of Limited Sampling Strategies for Bayesian Estimation of Daptomycin Area Under the Concentration–Time Curve: A Short Communication

Abstract

Purpose: Increasing evidence supports daptomycin therapeutic drug monitoring. The author's reference center used to perform therapeutic drug monitoring in patients who receive high-dose daptomycin for bone and joint infections, with a three-sample strategy to estimate the daptomycin daily area under the concentration–time curve (AUC). The objective of this study was to evaluate simpler strategies based on only 2 or 1 sample(s). Methods: The authors used the BestDose software to estimate the daptomycin AUC after Bayesian posterior estimation of individual pharmacokinetic (PK) parameters at steady state. The reference AUC (AUCfull) was based on 3 samples obtained predose (T0) and approximately 1 hour (T1) and 6 hours (T6) after the start of a 30-minute infusion of IV daptomycin. It was compared with the AUC based on all possible 2-sample and 1-sample strategies. Bias, imprecision, regression, and Bland–Altman plots were used to assess the performance of the alternative strategies. Results: Data from 77 patients were analyzed. The mean AUCfull value was 936 ± 373 mg·h/L. The best 2-sample strategy was T0 + T6, with a mean prediction bias of 0.13 mg·h/L and absolute imprecision of 3%. The T0 + T1 strategy also performed well with a mean bias of −10 mg·h/L and imprecision of 3%. The best 1-sample strategy was the T6 sample only with a bias of 2.19 mg·h/L and imprecision of 6%. Conclusions: Bayesian estimation of daptomycin AUC based on a two-sample strategy was associated with negligible bias and imprecision compared with the author's usual three-sample strategy. The trough and peak strategy may shorten and simplify patient visits and reduce assay labor and costs.