P42 External validation of model-based dosing guidelines for vancomycin, gentamicin and tobramycin in critically ill children

Background Pharmacokinetic models are frequently used to simulate dosing strategies for special populations, including critically ill children. The Dutch Pediatric Formulary (DPF) partially bases its guidelines on these models. However, prospective validation of updated dosing regimens is rare. We aimed to identify target attainment and safety of vancomycin, gentamicin and tobramycin after a dose update in the DPF for critically ill neonates and children. Methods Retropsective cohort study in PICU and NICU patients receiving vancomycin, gentamicin or tobramycin between January 2015 and March 2017 in 2 university hospitals. Demographic clinical laboratory and TDM-data were collected. Target (steady state) trough concentrations for vancomycin, gentamicin and tobramycin used were 10–15, ≤1 and ≤1 mg/l, respectively. Target gentamicin peak concentrations used were 8–12 mg/l. Results 486 patients were included in total (165 vancomycin, 97 gentamicin and 224 tobramycin). Trough concentrations of vancomycin, gentamicin and tobramycin were within the target range in 37.5%, 85.3% and 77.2% of patients, respectively. Target attainment of gentamicin peak concentrations in NICU patients was 31%. Non-target trough concentrations were most prevalent in term NICU patients (vancomycin 70%, gentamicin 26% and tobramycin 36.8%). Gentamicin peak concentrations were subtherapeutic in 91% and 45.5% for term and preterm NICU patients, respectively. Creatinine concentrations correlated positively with antibiotic concentrations (correlation coefficient range 0.46–0.54, p≤0.01 in all cohorts). Conclusion Despite recent model-based dosing alterations, sub- and supratherapeutic concentrations of vancomycin, gentamicin and tobramycin are still frequent in critically ill children. Linear dose alterations did offer improvements in target attainment, but did not fully address all relevant covariates that contribute to the large interindividual variation in clearance and/or volume of distribution in these patients. Creatinine clearance was consistently correlated with concentrations of all 3 drugs, but future research is needed to identify whether including this parameter in dosing can improve target attainment and safety. Disclosure(s) Nothing to disclose