Precision Daptomycin Dosing: Comparison of 3-, 2-, 1-, and 0-Concentration Sample Strategies.

Background: Daptomycin is a once-daily lipopeptide antibiotic with a narrow therapeutic window. As higher doses (8-12 mg/kg) are increasingly used to treat resistant gram-positive infections, achieving therapeutic exposure while minimizing toxicity has become critical. Understanding daptomycin therapeutic drug monitoring (TDM) and what sampling strategy may be most precise and feasible are key to guiding appropriate dosing. The primary objective of this study was to ascertain the precision and bias of a mid-point sample area under the curve over 24-h (AUC₂₄) determination compared to a peak and trough sample AUC₂₄ determination using both simple calculation-based methods and a Bayesian model versus no TDM.

Methods: Adult patients receiving daptomycin with at least three steady-state concentrations (peak, mid-point, trough) were included in this analysis. A previously published one-compartment population pharmacokinetic model was applied using Bayesian estimation (Monolix Suite 2024R1). AUC₂₄ values were estimated using all three concentrations (AUC_ref), and then re-estimated using individual samples (peak, mid-point, or trough) or paired samples (peak + trough). Performance of each strategy was evaluated using regression analysis with the coefficient of determination for precision (R²) and mean bias.

Results: The cohort included 210 patients (60% male) with a mean (range) age of 66 (18-91) years, body weight 78 (41-140) kg, and BMI 27 (14-51) kg/m². A total of 880 daptomycin concentrations were analyzed in patients receiving a mean (range) daptomycin dose of 9 (5-17) mg/kg. The mean (range) AUC₂₄ was 869 (385-1692) mg·h/L; only 45.7% of patients achieved the target range of 666 to 939 mg·h/L. A single mid-point sample had similar correlation to AUC₂₄ (R² = 0.57) as trough-only (R² = 0.55), and greater precision than peak-only (R² = 0.44). Bayesian estimation with two samples (peak + trough) provided the highest accuracy (R² = 0.87) and lowest bias.

Conclusions: A mid-interval sampling strategy offers a practical alternative to traditional TDM for daptomycin, enabling more consistent AUC estimation when full sampling is not feasible. A two-sample Bayesian approach remains the most accurate, supporting broader implementation of individualized daptomycin dosing.