Dosing errors of empirical antibiotics in critically ill patients with severe sepsis or septic shock: A prospective observational study.

Background: Effective antibiotic therapy is crucial in sepsis management. Studies have emphasized on antibiotic administration timing more than dosing. We evaluated the frequency and risk factors of antibiotic dosing errors in sepsis.

Materials And Methods: This prospective observational study compared the doses of intravenous empirical antibiotics in the 1 day of sepsis diagnosis with those recommended by Micromedex, considering sepsis source and glomerular filtration rate estimated by the modification of diet in renal disease equation. The doses were classified as under-dosed, over-dosed, or appropriate. We excluded antibiotics which did not need a dose adjustment. The under-dosing rate was retrospectively evaluated according to the creatinine clearance estimated by the Cockcroft-Gault equation.

Results: Between October 1, 2013, and April 30, 2014, 189 patients were evaluated (age 61.6 ± 18.6 years, acute physiology and chronic health evaluation II score 22.8 ± 7.0, 58.7% septic shock) with 415 antibiotic prescriptions (2.2 ± 0.9 prescriptions per patient). Antibiotic dosing was appropriate in 50.8% of patients; under-dosing in 30.7% and overdosing in 25.9%. Under-dosing prevalence was 39.4% when kidney function was assessed by the Cockcroft-Gault equation. Vancomycin was associated with the highest under-dosing rate (39.4%), followed by piperacillin/tazobactam (12.1%). The cohort mortality was 35.4%. Age, shock, chronic kidney disease, and cirrhosis independently predicted mortality on multivariable logistic regression analysis. Antibiotic dosing error was not associated with mortality: Over-versus appropriate dosing (odds ratio [OR], 1.41; 95% confidence interval [CI], 0.57-3.47), under-versus appropriate dosing (OR, 0.57; 95% CI, 0.24-1.35).

Conclusions: Antibiotic dosing errors were common in patients with sepsis. Vancomycin had the highest under-dosing rate. Antibiotic dosing errors were not associated with increased mortality.