Objectives: To assess the implementation and clinical effectiveness of AUC-guided Vancomycin monitoring in intensive care and toxicology patients, and to examine the performance of the supporting AUC-estimation approaches used in clinical practice.
Methods: A retrospective study was conducted in adult at Intensive Care and Toxicology Department, Nguyen Tri Phuong Hospital patients who used Vancomycin with at least one therapeutic drug monitoring measurement (from August 2023 to May 2024). AUC24 was estimated using PrecisePK® software based on the Rodvold model. Clinical performance was assessed through target attainment rate (AUC24 400-600 mg.h/L). The predictive performance of the popPK and Bayes models was determined by relative bias (%rBias) and relative root mean square error (rRMSE).
Results: The study enrolled 91 patients; the mean age of patients was relatively high (mean 65.7 years); demonstrated low creatinine clearance (mean 45.0 mL/min), with 80.2% of patients had creatinine clearance ≤ 60 mL/min, among them 65.9% received a loading dose, and 63.7% had an appropriate initial maintenance dose. The rate of achieving target AUC24 at the first measurement was 36.3% and 69.8% after dose adjustment. Among patients with creatinine clearance < 60 mL/min, 67.1% had an AUC24 > 600 mg.h/L. The Bayes approach improved prediction accuracy (rBias -10.6%, rRMSE 18.4%) compared with the population model (rBias +17.4%, rRMSE 44.3%).
Conclusion: AUC-guided therapeutic monitoring of Vancomycin improved target attainment in critically ill patients and can be feasibly implemented in clinical practice. Bayes-guided dose adjustment improved target attainment and predictive accuracy, supporting its use as a practical tool for individualized therapy in critically ill patients.