926

Introduction: β-lactams are widely used as first-line treatment in critically ill patients. It has been shown that creatinine clearance (CrCl) influences the pharmacodynamics of these drugs. Hypothesis: Variations in CrCl are correlated with changes in β-lactams concentrations. Methods: We reviewed the data of all adult patients admitted to our Dept of Intensive Care (ICU) from January 2010 to June 2012, in whom routine therapeutic drug monitoring (TDM) of broad-spectrum β-lactams (ceftazidime, cefepime, piperacillin and meropenem) was performed. Patients were included if they had at least two TDMs during their ICU stay for the same antibiotic and if they were not concomitantly treated by any extracorporeal replacement therapy. Serum drug concentrations were measured by high-performance liquid chromatography (HPLC-UV). Blood samples were taken twice during the elimination phase after a 30-min intravenous drug administration. Antibiotic PKs were calculated using a one-compartment model and the percentage of time spent above four times the MIC (%T>4xMIC) for Pseudomonas aeruginosa as well as antibiotic clearance (ATB-CL) were obtained. CrCl was measured using 24-hours urine collection on the same period than TDM. Results: We included 56 patients (median age 59 years; 40/56 male gender, median APACHE II score on admission 19; 73% treated by mechanical ventilation, ICU mortality 20%). Median time from the start of antibiotic therapy to the first and second TDM was 2 and 5 days, respectively. We found that %T>4xMIC and ATB-CL were significantly correlated with CrCl on the first TDM (Spearman r = -0.41, p=0.002; r = 0.56, p<0.001, respectively) and the second TDM (Spearman r = -0.61, p<0.001; r = 0.62, p<0.001, respectively). Nevertheless, the changes in the %T>4xMIC and ATB-CL were not correlated to the changes in CrCl (Spearman r =0.22, p=0.09 and Spearman r = -0.17, p=0.2). Conclusions: Although CrCl was significantly correlated with the concentrations of broad-spectrum β-lactams, changes in CrCl do not reliably predict variations in drug clearance and pharmacodynamics.