Empirical use of antibiotics in a Medical Intensive Care Unit (MICU)

Purpose: Prompt antibiotic administration is life-saving in critically ill patients. Initial antibiotic choices are often empirical, without a definite presence of infection, source or organism. Empirical antibiotic administration may not provide adequate coverage or may also be related to adverse effects and development of antibiotic resistance. Methods: We prospectively followed all patients admitted to the MICU during the three-month period 12/97 through 3/98. Initial antibiotic choices, changes within the first three MICU days, yield of cultures, appropriateness of antibiotic therapy, and outcome were recorded. Results: During the three month period, 33 consecutive patients admitted to the MICU with presumed infection were studied Initial impression of sources included: pneumonia (26), urosepsis (1), bacteremia (4) and meningitis (2). The combination of Vancomycin and ticarcillinclavulanate was used in the majority of patients (48.4%) regardless of suspected source. Antibiotic therapy was changed by the MICU team within the first 24 hours in 57.6%. Cultures provided a definitive diagnosis in 14 cases (42.4%). In 1 case the empirical antibiotic choice did not cover the cultured pathogen. Patients received a mean of 3.6 antibiotics during the first 24 hours. The number of antibiotics was correlated with the presence of circulatory shock (Mann-Whitney Rank sum test, p = 0.001). Overall MICU mortality was 30.3%. MICU mortality was not different in the population where no organisms were isolated compared to the group where a pathogen was identified (p = 0.84). In only 5 out of 14 (35.7%) cases was the antibiotic coverage decreased, once the sensitivities were known. Conclusions: Empirical antibiotic choices and additions during the first 24 hours provided adequate coverage of severe infectious illnesses requiring MICU admission. Patients with greater degree of hemodynamic instability were more likely to be exposed to more than four different antibiotics. Knowledge of the infecting organism and its antibiotic sensitivity resulted in a more specific coverage in only 35.7% of the cases. Clinical Implications: In view of concerns about antibiotic resistance, the choice of empirical antibiotic regimen and the reluctance to narrow coverage require further study.