Background Inappropriate antimicrobial treatment in bloodstream infections is associated with increased mortality rates and health cost.

Purpose To assess the adequacy and modification of empirical treatment, according to preliminary and final microbiological results, in a tertiary care hospital with an ‘on line’ alert system that reports positive blood cultures with Gram staining.

Materials and methods Retrospective observational study using computerised patient microbiological records and the hospital electronic prescription database. Hospitalised patients (excluding Intensive Care Unit) with an ‘on-line’ alert for a positive blood culture were identified over a 30-day period. Dates of blood culture extraction, preliminary Gram-staining alert, and final report (with microorganisms isolated and sensitivity pattern) as well as the patient’s antimicrobial treatment were recorded. Treatment was considered appropriate if causative microorganisms were susceptible at least to one of the antimicrobials prescribed.

Results Thirty patients were included. The mean time to notify the preliminary and final results was 0.86 ± 0.77 and 2.43 ± 0.67 days respectively. The initial staining identified: 70% Gram bacilli, 20% Gram + cocci in clusters, 6.7% Gram + cocci in chains and 3.3% Gram + bacilli. The most frequently isolated pathogens were Escherichia coli (30%) and Staphylococcus epidermidis (20%). Multidrug resistant pathogens were isolated in 20% of patients. Monotherapy was the initial antibiotic choice in 60% of patients (38.7% Piperaciline-Tazobatam, 27.7% Carbapenems, 16.6% Fluoroquinolones, 16.6% other ß-lactam) while combination treatment was administered to the remained 40% patients (50% ß-lactam + Gluco/Lipopeptides, 25% ß-lactam + Aminoglycosides, 25% others). Empirical treatment was adequate in 76.7% of cases (50% for multidrug resistant pathogens). This treatment was modified in 83.4%, 16.7% according to the preliminary result, 3.3% before the final result, 56.7% within 24 h after the final report, and 6.7% later. Treatment was appropriate in 96.7% of patients within 24 h following the final microbiological report.

Conclusions Modification of antibiotic treatment during the post-empirical phase is frequent and achieves high rates of suitability. Although initial information is provided quickly, only some changes are made after a Gram staining alert. Most clinicians wait until the final microbial characterisation.

No conflict of interest.