Impact of antibiotic changes in empirical therapy on antimicrobial resistance in intensive care unit-acquired infections.

We conducted a one-year prospective study on intensive care unit (ICU)-acquired infections and antimicrobial resistance patterns in an 18-bed medical-surgical ICU of a tertiary-care university hospital. We divided the study into two six-month periods in order to evaluate the impact of antibiotic changes in empirical therapy on antimicrobial resistance profiles of the principal isolated micro-organisms. In the first period no changes were made to the previously applied empirical antibiotic protocol; at the end of this period we found high rates of methicillin resistance (MR) among staphylococci, 93% for Staphylococcus aureus (69 isolates) and 79% for coagulase-negative staphylococci (CNS) (48 isolates), and of multiple drug resistance for Pseudomonas aeruginosa (57 isolates), in particular 67% resistance to piperacillin/tazobactam (PIP/TZ). We therefore decided to substitute PIP/TZ with imipenem in nosocomial pneumonia and with cefepime plus metronidazole in peritonitis. We also considered the previous use of amoxicillin/clavulanate (AM/CL) at admission in critically ill patients inadequate; we therefore advised that no antibiotics should be given unless fever developed and eventually to replace AM/CL with trimethoprim/sulfamethoxazole (TMP/SMX). At the end of this intervention period, we observed a significant decrease of S. aureus MR (93 vs. 73%, P = 0.003) and of P. aeruginosa resistance to PIP/TZ (67 vs. 29%, P < 0.001). A reduction in MR was also seen in CNS (79 vs. 64%, P = 0.09). Other resistance patterns also improved among staphylococci; in contrast P. aeruginosa resistance to imipenem increased in the second period (24 vs. 41%, P = 0.06). A non-premeditated change of antibiotics in empirical therapy, on the basis of detected resistance patterns, provided promising results in reducing some antimicrobial resistance rates. We believe, however, that antibiotic changes must be tailored to local microbiological situation monitoring, and that a repeated rotation is crucial to limit the emergence of new resistance profiles. Furthermore the adoption of this policy should be accompanied by other infection control practices aimed at reducing antimicrobial resistance and nosocomial infection rates. Copyright 2002 The Hospital Infection Society