[Monitoring aminoglycosides in an Intensive Care Unit]

Therapeutic drug monitoring of aminoglycosides in acute myeloid leukaemia patients

International guidelines limit the use of aminoglycosides in febrile neutropenia to severe situations. We retrospectively reviewed the use of aminoglycosides in adult acute myeloid leukaemia patients admitted in 2009. Our guidelines include precise indications (severe sepsis, shock, drug resistance), dosing regimens (once-daily 20 mg/kg/day amikacin, 5 mg/kg/day gentamicin), durations of treatment, drug monitoring timing, and target C(max) concentrations (40 mg/l amikacin, 20 mg/l gentamicin). Thirty-one patients received 46 aminoglycoside courses: 31 amikacin and 15 gentamicin. The mean prescribed dosage was 19 ± 2.8 mg/kg/day for amikacin and 4.7 ± 0.9 mg/kg/day for gentamicin. The mean duration of use was 2.9 days for both drugs. The mean C(max) for amikacin was 47 ± 13 mg/l and for gentamicin was 13.6 ± 7.5 mg/l. In compliant regimens, all amikacin patients and a third of gentamicin patients had adequate C(max). Among 23 isolated pathogens, 65.5% were susceptible to both drugs and 11.5% to amikacin only. This vindicates the 20 mg/kg/day amikacin dosage and suggests a need to increase the gentamicin dosage.

[A priori prediction of gentamicin peak concentrations: Use of a simple and practical tool]

Aminoglycosides are concentration-dependent killing antibiotics, so that it is necessary to obtain elevated peak levels to reach the pharmacokinetic-pharmacodynamic objectives. Because of their nephrotoxicity, this class of antibiotics are frequently underprescribed and giving at an insufficient dosage when prescribed. That is why therapeutical drug monitoring (TDM) of aminoglycosides is recommanded to assess efficacy. Unfortunately, this TDM allows not a dosage adaptation on the first dose of aminoglycosides. The aim of the work was to elaborate a model using simple pharmacokinetics formulae to estimate gentamicin peak levels and so to propose an optimal a priori dosage on the first dose of gentamicin. The final model was tested out on 24 patients treated by gentamicin. The model was able to predict gentamicin peak levels in 62% of patients with a 10% precision and in 79% of patients with a 15% precision. Using this simple and practical tool might avoid under-dosage of gentamicin and clinical failure due to the selection of resistant bacteriae.