BACKGROUND: Optimal dosing of vancomycin in critically ill patients receiving renal replacement therapy (RRT) is uncertain due to high pharmacokinetic variability. We aimed to develop individualised vancomycin dosing recommendations that optimise efficacy while minimising toxicity.

METHODS: Prospective, international, pharmacokinetic study enrolling critically ill patients treated with vancomycin and various RRT modalities. A population pharmacokinetic model was developed, externally validated and applied to perform Monte Carlo dosing simulations. We calculated the probability of each dosing regimen to achieve the efficacy target against methicillin-resistant Staphylococcus aureus (ratio of the area under the concentration-time curve to the minimum inhibitory concentration (AUC/MIC) ≥ 400) without exceeding the toxicity threshold (AUC ≥ 700 mg.h/L).

RESULTS: We enrolled 65 critically ill patients from 6 countries receiving continuous RRT (50.8%) or sustained low-efficiency dialysis (49.2%). The model was developed using 1318 pre- and post-filter plasma and effluent concentrations and validated with a new dataset (19 critically ill patients, 124 pre-filter plasma concentrations), with 47.4% patients concomitantly treated with extracorporeal membrane oxygenation. Predictive performance was high (median prediction error =-13.4%). Simulations showed that dosing requirements were dependent on actual body weight and RRT intensity and duration (p < 0.05). An optimised dosing nomogram was subsequently developed considering these clinical characteristics.

CONCLUSIONS: Actual body weight and RRT intensity and duration are the main determinants of vancomycin dosing requirements in critically ill patients treated with RRT. Efficacious, non-toxic dosing may be guided by the optimised nomogram.